PDF file, 12 MB - 39. Jahrestagung der Deutschen Gesellschaft für

0
1
0
0
N
F
U
O
P
D
&
K
&
I
1
0
1
1
L
E
I
I
N
I
K
O
0
0
0
D
R
D
H
A
U
E
O
D
I
N F
1
A
0
1
O
K
C
I
O
R
D
I
N
O
0 0
0
1 0
0 L
0 U
N 1
O F
I
I E
U
I E
E R
F O
R M
1
1
D
0
1
C
O
F
N
N
R
U
R
A
0
0
1
1
R
M
A
E
G
N
M
T
1
0
0
0
0
R
I
N
I
O
U
G
A
I
1
1
1
1
N
O
I
N
F
R
N
1
1
1
R
H
0
F
N
N
M
G
I
T I
O N
0
1
1
T
A
H
E
S
D
A
I
N
O
1
1
0
1
0
U
C
L
K
0
0
1
U
R
N
M
R
O
I
N F
F O
N
&
0
1
F
1
T
L
N
G
U
C
O
R
K
S
0
0
0
R
1
R
L
T
R
H
R
M
O
P
0
0
0
0
N
0
R
H
G
A
M
A
D
R
1
0
1
N
S
0
0
1
1
1
D
N
I
I
H
1
1
R
0
F
A
G D
R H
O A
C L
A T
T I O
I E R
A C H
0
1
U
1
E
U
K
C
I
N
I
N
U
L
0
N
0
F
C
I
N
O
H
S
O
1
1
O
E
M
H
I
N
N
P
N
S
N G
I C
1
1
0
0
H
1
D
D
S
H
0
U
0
1
U
1
A
K
C
U
L
P R
S
H E
1
0
A
O
H
0
O
N
T
N
I
A
P
0
1
I
F
1
R
R
D
H
D
C
C
R
K
1
R
1
O
O
T
T
A
R
N
H
H
A
O
1
0
1
A
F
O
N
O
M
F
0 1 1 1
1 1 0 1
N 1 1 1
D 1 N
O A I D
F O A I
N S H R
U S C M
N D A
O K O
S P R A
L I C H E
C H L I C
D I E R U
1
1
R
1
K
1
O
N
N
I
C
0
1
0
H
N
N
I
P
S
T
H
&
H E
N G
39. Jahrestagung der
Deutschen Gesellschaft für Sprachwissenschaft
08.–10. März 2017
Universität des Saarlandes
I
A
E
A
D
G
D
N
1
N
1
1
1
0
N
N
N
N
I
O
A
L
1
L
U
0
1
1
F
D
C
D
N
I
O
R
0
R
1
1
U
0
O R
I
O D
E N
G
A I
U S
G N
R T
O 0
R 0
1 1
0 1
1 0
M
N
I
C
A
H
O
D
F
U
N
0
R
0
A
F
N
O
N
P
I
K
1
1
1
0
1
T
O
G
D
G
N
A
N
D
R
0
0
1
0
I O N
R M A
I N
I N G
U R M
U O I
I I
I D A
A U I
0 N H
1 0 1
1 0 0
1 0 0
1 1 1
A N
T I O
F O R
I N
A I
E N
F E N
O R U
H E D
U H 1
D 1 0
1 0 D
1 1 1
0 1 0
D
N
M
F
C
I
N
I
K
O
L
F
1
0
A
O
H
F
N
H
N
F
0
0
0
0
L
E
T
R
A
N
S
C
1
O
1
A
1
0
I
N
I
M
C
D
L
O
C
A
0
I
0
1
N
C
O
A
H
K
R
F
M
I
1
1
0
0
G U I S
O D I N
N L I
T I O
L N S P
O U R N
G T H R
O A I I
R O 0 H
D K H A
R T F E
1 N 0 0
0 0 1
1 0 0 1
T
G
N
N
H
S
I
F
U
0
0
1
1
1
I C
L
G U
T
U N
C T
H C
E C
N L
R T
1 O
1 R
N 1
0 1
I
I
I
D
H
O
M
R
1
F
U
A
0
E
N
S
C
N
R
N
T
0
N
O
0
0
1
N
G
T
L
F
M
D
N
O
S
A
O
1
1
C
U
I
I
O
N
K
N
F
1
O D I N G
I S T I C
C A N D
N G U I S
K O I T
D A N S
N C M N P
S H R O I
O A I 1 N
F E C M N
D 1 N 1 H
0 1 1 1 R 0
0 0 0 1 1 1
0 1 1 1 1 0
39. Jahrestagung
der Deutschen Gesellschaft
für Sprachwissenschaft
Information &
sprachliche Kodierung
8.–10. März 2017
Universität des Saarlandes
Druck:
COD Büroservice GmbH
Bleichstraße 22–24
66111 Saarbrücken
Tel. 0681/39353-0
E-Mail: [email protected]
Satz:
Dieser Band wurde mit XELATEX und den Fonts Skolar (Sans) PE gesetzt.
Haftungsausschluss:
Die digitale Version dieses Tagungsbandes enthält Hyperlinks, die auf externe Internetangebote verweisen. Wir übernehmen keine Haftung für eventuelle Datenschutz- und sonstige Rechtsverletzungen in anderen Internetangeboten, auf die wir einen Link gesetzt haben. Für die Inhalte der von uns verlinkten Fremdangebote sind die jeweiligen Herausgeber verantwortlich. Vor
dem Einrichten von Links sind die Webseiten der anderen Anbieter mit großer Sorgfalt und nach bestem Wissen und Gewissen geprüft worden. Es kann
jedoch keine Gewähr für die Vollständigkeit und Richtigkeit von Informationen auf verlinkten Seiten übernommen werden.
Federführende Organisation
Ingo Reich & Augustin Speyer
Organisatorische Unterstützung
Denise Mayer, WuT, Universität des Saarlandes
Organisationsteam
Anne-Kathrin Balo, Vera Demberg
Christoph Clodo, Luise Ehrmantraut
Ellen Geibel-Stutz, Remus Gergel
Peter Gluting, Stefanie Haberzettl
Nele Hartung, Eva Horch
Lucia Hubig, Natascha Immesberger
Sergey Kulakov, Robin Lemke
Helena Raber, Philipp Rauth
Asad Sayeed, Lisa Schäfer
Jessica Schmidt, Josef Schu
Julia Schüler, Julia Stark
Sophia Voigtmann, Jonathan Watkins
Magdalena Wojtecka
Mit freundlicher Unterstützung des SFB 1102 »Information Density
and Linguistic Encoding« (Sprecherin: Elke Teich)
grußwort der organisatoren
Sehr geehrte Damen und Herren, liebe Kolleginnen und Kollegen,
das erste und bisher einzige Mal fand die Jahrestagung der Deutschen Gesellschaft für Sprachwissenschaft e.V. (DGfS) in Saarbrücken im Jahre 1990 statt,
vom 28. Februar bis 02. März. Zu dieser Zeit gingen die meisten Mitglieder des
diesjährigen Organisationsteams noch in die Schule (und manche sogar noch
in den Kindergarten) und wussten mit Sprachwissenschaft vermutlich noch
wenig bis nichts anzufangen. Nun, 27 Jahre später, hat sich die Situation zum
Glück etwas geändert, und wir freuen uns alle sehr, Sie bei der 39. Auflage der
Jahrestagung der DGfS in Saarbrücken begrüßen zu dürfen.
Die Situation ist dabei nicht ganz unähnlich zu der in Konstanz vor einem
Jahr. Wie die Konstanzer sind auch wir knapp an einem Jubiläumsjahr der
DGfS vorbeigeschrammt. Wie die Konstanzer haben aber auch wir nicht ganz
unbedeutende lokale Jubiläen vorzuweisen: So wurde exakt heute vor 70 Jahren (am 8. März 1947) die Vorläuferin der Universität des Saarlandes eröffnet, das Centre Universitaire d’Études Supé rieures de Hombourg, eine medizinische Hochschule, damals noch institutionell angebunden an die französische
Universität Nancy. Die Gründung der Universität des Saarlandes selbst, mit der
Einrichtung des Campus Saarbrücken in den Räumlichkeiten der ehemaligen
Below-Kaserne, ließ dann noch gut ein Jahr auf sich warten. Nochmal knapp
9 Jahre und eine Volksbefragung später wurde dann das Saarland als zehntes
Bundesland Teil der Bundesrepublik Deutschland. Aus diesem Grund freuen
wir uns, dieses Jahr auch »60 Jahre Saarland« feiern zu dürfen.
Noch eine Parallelität zu den Konstanzern besteht darin, dass sich auch das
Saarland in einer Art Drei-Länder-Eck befindet, mit Frankreich im Südwesten und Luxemburg im Nordwesten. Darüber hinaus verläuft mitten durch das
Saarland eine wichtige Dialektgrenze – die das/dat-Linie –, die insbesondere
das Rheinfränkische vom Moselfränkischen trennt. Diese tagtäglich er- und
gelebte sprachliche Vielfalt mag einer der Gründe sein, wieso die Universität
des Saarlandes einen sprachwissenschaftlichen Schwerpunkt mit den Sprachwissenschaften der Philologien, der Computerlinguistik, der Psycholinguistik,
der Phonetik und den Übersetzungswissenschaften ausgebildet hat. Zeichen
der Breite und Stärke dieses sprachwissenschaftlichen Schwerpunktes ist der
2014 eingeworbene SFB 1102 Information Density and Linguistic Encoding, in dessen Zentrum informationstheoretische Konzepte wie Surprisal oder Entropie
1
stehen sowie deren Relevanz für die Wahl zwischen alternativen sprachlichen
Kodierungen. Das Rahmenthema der diesjährigen Jahrestagung
Information und sprachliche Kodierung
lehnt sich an diese Fragestellung an, hebt sie aber auf eine abstraktere Ebene, um auch anderen theoretischen Ansätzen Raum zu geben. In diesem Rahmen werden in einer Postersession der Sektion Computerlinguistik und in
dreizehn Arbeitsgruppen Fragestellungen diskutiert, die von der optimalen
Verteilung von Information über einen sprachlichen Ausdruck hinweg über
alternative sprachliche Realisierungsformen semantisch äquivalenter Aussagen bis hin zur Register-Abhängigkeit grammatischer Phänomene reicht. Wir
freuen uns alle auf interessante Vorträge und lebhafte Diskussionen, auf ein
Wiedersehen mit alten und das Kennenlernen von neuen Kolleginnen und Kollegen.
Genießen wir also die kommenden drei Tage!
Ingo Reich & Augustin Speyer
(stellvertretend für das Organisationsteam)
2
danksagungen
Die Organisatorinnen und Organisatoren bedanken sich herzlich bei den folgenden Sponsoren für die Unterstützung bei der Finanzierung der Tagung (in
alphabetischer Reihenfolge):
• John Benjamins Publishing Company db
• Helmut Buske Verlag GmbH
• Cornelsen Verlag GmbH
• Walter de Gruyter GmbH
• Deutsche Forschungsgemeinschaft (DFG)
• Dr. Ute Hempen Verlag
• DUDEN Bibliographisches Institut GmbH
• Frank & Timme GmbH Verlag für wissenschaftliche Literatur
• ibidem-Verlag
• SFB 1102: Information Density and Linguistic Encoding
• IUDICIUM Verlag GmbH
• Peter Lang GmbH
• Lin|gu|is|tik Portal für Sprachwissenschaft
• Liverpool University Press
• J. B. Metzler Verlag
• Narr Francke Attempto Verlag GmbH
• Philipp Reclam jun. Verlag GmbH
• Saarländische Staatskanzlei
• Saarländisches Landesinstitut für Pädagogik und Medien
• Saarländisches Ministerium für Finanzen und Europa
• Erich Schmidt Verlag GmbH & Co. KG
• Stauffenburg Verlag GmbH
• Franz Steiner Verlag
• Universität des Saarlandes
• Universitätsgesellschaft des Saarlandes
• V&R unipress GmbH
3
• Volksbank Westliche Saar plus eG
• Verlage Westermann Gruppe, Schöningh
• Waxmann Verlag GmbH
• Universitätsverlag Winter GmbH
(Stand bei Redaktionsschluss)
Die Organisatorinnen und Organisatoren der Lehramtsinitiative bedanken
sich ganz herzlich für die Unterstützung des Landesinstituts für Pädagogik und
Medien des Saarlandes.
Das Team des Tagungsbandes bedankt sich ganz herzlich bei den Organisatorinnen und Organisatoren der DGfS-Jahrestagung 2016 in Konstanz, die
uns den LATEX-Code des Tagungsbandes 2016 zur Verfügung gestellt haben. Der
Code dieses Bandes wird selbstverständlich auch allen Interessierten weitergegeben. Zu diesem Zweck einfach bei Philipp Rauth melden
([email protected]).
4
sponsoren
5
Inhaltsverzeichnis
Grußwort der Organisatoren . . . . . . . . . . . . . . . . . . . . . .
Danksagungen . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Allgemeine Informationen
Informationen zur Tagung . . . . . . .
Anreise nach Saarbrücken . . . . . . .
Anfahrt zum Campus . . . . . . . . . .
Campusplan und Gebäudepläne . . . .
Raumübersicht . . . . . . . . . . . . .
Essen und Trinken auf dem Campus . .
Gastronomie in der Stadt . . . . . . . .
Sehenswürdigkeiten und Ausgehtipps
Ausflüge in die Umgebung . . . . . . .
1
3
9
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Programmübersicht
Allgemeines Programm . . . . . . . . . . . . . . . . . . . . . . . . .
AG-Programme . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11
16
18
22
26
28
31
41
42
45
47
49
Plenarvorträge
75
Arbeitsgruppen
87
AG 1 – Informationstheoretisch basierte Modellierung sprachlicher
Variation im Kontext . . . . . . . . . . . . . . . . . . . . . . .
AG 2 – Information structuring in discourse . . . . . . . . . . . . .
AG 3 – Sekundäre Information und ihre sprachliche Kodierung . . .
AG 4 – Encoding language and linguistic information in historical
corpora . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
AG 5 – Sprachliche Kodierungs-Asymmetrien, Gebrauchsfrequenz
und Informativität . . . . . . . . . . . . . . . . . . . . . . . .
AG 6 – Prosody in syntactic encoding . . . . . . . . . . . . . . . . .
89
106
127
145
156
187
7
AG 7 – Wen wurmt der Ohrwurm? – An interdisciplinary, crosslingual perspective on the role of constituents in multi-word
expressions . . . . . . . . . . . . . . . . . . . . . . . . . . . .
AG 8 – Non-canonical verb positioning in main clauses . . . . . . .
AG 9 – Towards an ontology of modal flavors . . . . . . . . . . . . .
AG 10 – Polysemy and coercion of clause-embedding predicates . . .
AG 11 – Coercion Across Linguistic Fields (CALF) . . . . . . . . . . .
AG 12 – Morphologische Variation – Theorie und Empirie . . . . . .
AG 13 – Register in linguistic theory: Modeling functional variation
207
221
240
253
265
281
301
Postersession der Sektion Computerlinguistik
319
Tutorium der Sektion Computerlinguistik
343
Doktorandenforum
347
Infotag der Lehramtsinitiative der DGfS
353
Tagung der Arbeitsgemeinschaft Linguistische Pragmatik
363
DFG-Informationsveranstaltung
369
Personenverzeichnis
383
Gesamtübersicht der Arbeitsgruppensitzungen
391
8
Allgemeine Informationen
Allgemeine Informationen
i
informationen zur tagung
Veranstalter
Universität des Saarlandes, FR Germanistik
Deutsche Gesellschaft für Sprachwissenschaft (DGfS)
Wissenschaftliche Leitung
Prof. Dr. Ingo Reich (Neuere Deutsche Sprachwissenschaft)
Prof. Dr. Augustin Speyer (Neuere Deutsche Sprachwissenschaft)
Organisatorische Unterstützung
Denise Mayer, WuT, Universität des Saarlandes
Homepage
dgfs2017.uni-saarland.de
Tagungsort
Universität des Saarlandes
Campus
66123 Saarbrücken
www.uni-saarland.de
Tagungsbüro: Anmeldung und Information
Die Anmeldung erfolgt ab Mittwoch, 08.03.2017 im Tagungsbüro, das sich
in Raum 0.21 im Tagungsgebäude B4 1 befindet. Dort erhalten Sie Ihre
Teilnahmeunterlagen und alle wichtigen Informationen zur Tagung. Das
Tagungsbüro ist vom 08.03.–10.03.2017 besetzt und dient als zentrale Anlaufstelle für alle Fragen. Hier finden Sie außerdem das Fundbüro und eine
Ansprechperson für Notfälle (Erste Hilfe).
11
Allgemeine Informationen
i
Öffnungszeiten des Tagungsbüros
Mittwoch, 08. 03. 2017
08:00–18:00 Uhr
Donnerstag, 09. 03. 2017 08:00–18:00 Uhr
Freitag, 10. 03. 2017
08:00–14:00 Uhr
Telefon
In dringenden Fällen können Sie das Tagungsbüro während der oben genannten Öffnungszeiten unter folgender Telefonnummer erreichen:
 0681 3024084
Internetzugang
Für Hochschulangehörige besteht grundsätzlich der Internetzugang über
Eduroam (www.eduroam.org). Sollten Sie keinen Zugriff auf Eduroam haben,
können Sie beim Tagungsbüro einen individuellen Zugang über die Universität des Saarlandes beantragen.
Uni Saar App
Die „Uni Saar“ App für iPhones und alle Android-Geräte erlaubt einen schnellen und übersichtlichen Zugang zu nützlichen Informationen rund um die
Universität des Saarlandes, z. B. Speisepläne der Mensa, Busabfahrtszeiten
(inkl. Verspätungsradar) oder einen interaktiven Lageplan des Campus mit
Ortungsfunktion.
www.st.cs.uni-saarland.de/uniapp/
Kopieren und drucken
Für kurzfristige Ausdrucke oder Kopien Ihrer Unterlagen vor Ort steht Ihnen
die Fotostelle der Saarländischen Universitäts- und Landesbibliothek (SULB),
Gebäude B1 1, zur Verfügung. Sie ist von 09:00–16:00 Uhr geöffnet.
www.sulb.uni-saarland.de/service/fotostelle/
Erste Hilfe
Im Tagungsbüro (B4 1, 0.21) finden Sie stets eine Ansprechperson für Notfälle.
Außerdem kann über das Konferenz-Personal jederzeit telefonisch Hilfe
angefordert werden.
12
Allgemeine Informationen
Gepäck
Während der Tagung besteht die Möglichkeit, Gepäck zu deponieren. Wenden
Sie sich hierzu bitte an das Tagungsbüro.
Barrierefreiheit
Die Räumlichkeiten der Konferenz in den Gebäuden B3 1, B3 2 und B4 1 sind
barrierefrei zugänglich. Über den Standort von barrierefrei zugänglichen
Eingängen, Toiletten, Parkplätzen und einen barrierearmen Weg zur Mensa
können Sie sich auf dem Campusplan auf der Innenseite des hinteren Umschlags des Tagungsbands informieren.
Kinderbetreuung
Während der Tagung ist eine Kinderbetreuung verfügbar. Nehmen Sie dazu
im Vorhinein bitte Kontakt mit der lokalen Organisation auf.
Anmeldung
Die Online-Anmeldung zur DGfS 2017 erfolgt über die Internetseite der DGfS
unter dem Menüpunkt „Anmeldung“. Vor Ort kann die Teilnahmegebühr
außerdem noch in Barzahlung und per EC-/Kreditkarte entrichtet werden.
Dort erhalten Sie weiterhin Ihre Teilnahmeunterlagen und alle wichtigen
Informationen zur Tagung.
Teilnahmegebühren
Der Frühbucherrabatt ist bei Registrierung bis zum 31.01.2017 gültig. Bei Registrierung nach dem 31.01.2017 erhöht sich die Konferenzgebühr um 5,00 Euro. Die Konferenzgebühr hängt darüber hinaus von der Mitgliedschaft in der
DGfS und der Verfügbarkeit eines Einkommens ab.
DGfS-Mitglieder – mit regulärem DGfS-Beitrag:
DGfS-Mitglieder – mit reduziertem DGfS-Beitrag:
Nicht-Mitglieder – mit Einkommen:
Nicht-Mitglieder – ohne Einkommen:
Frühbucher
50,– EUR
35,– EUR
70,– EUR
40,– EUR
regulär
55,– EUR
40,– EUR
75,– EUR
45,– EUR
13
i
Allgemeine Informationen
i
Bankverbindung
Institut:
Kontoinhaber:
Kontonummer:
Bankleitzahl:
IBAN:
BIC:
Bank1Saar eG
Universität des Saarlandes
977 180 08
591 900 00
DE94 5919 0000 0097 7180 08
SABADE5S
Rahmenprogramm
– Warming Up
Das Warming Up findet am Dienstag, 07. 03. 2017, ab 19:00 Uhr im Stiefelbräu (Am Stiefel 2, 66111 Saarbrücken) statt. Das traditionsreiche Saarbrücker Gasthaus bietet klassische saarländische Küche.
www.der-stiefel.de/stiefelbraeu
– Empfang im Rathaus
Am Mittwoch, 08. 03. 2017, wird ab 19:00 Uhr im Rathaus der Stadt Saarbrücken ein Sektempfang stattfinden. Die Anzahl der Plätze beim Empfang ist auf ca. 200 begrenzt.
– Geselliger Abend / Conference Dinner
Der Gesellige Abend findet am Donnerstag, 09. 03. 2017, ab 19:30 Uhr
in der Bar Celona statt. Die Bar Celona liegt in der Innenstadt an der
Berliner Promenade mit Blick auf die Saar.
Die Anzahl der Plätze beim Geselligen Abend ist auf ca. 300 begrenzt.
Bitte geben Sie bei der Registrierung an, ob Sie an dem Geselligen Abend
teilnehmen möchten. Sie erhalten bei der Registrierung ein Ticket, das
beim Einlass kontrolliert werden wird. Für das leibliche Wohl wird
es ein mediterranes Buffet geben (35,– EUR pro Person, Getränke auf
eigene Rechnung).
http://celona.de/mein-celona/details/cafe-bar-celonasaarbruecken
14
Allgemeine Informationen
Geldautomaten
Auf dem Campusgelände befinden sich vier Geldautomaten, deren Standorte
auch auf dem Campusplan auf der Innenseite des hinteren Umschlags des Tagungsbands verzeichnet sind (ec-Zeichen).
Sparkasse, im Erdgeschoss der Mensa des Studentenwerks, Gebäude D4 1.
Postbank, vor Gebäude A5 4 (gegenüber Campus Center, Gebäude A4 4).
Volksbank (Bank 1 Saar), im Campus Center, Gebäude A4 4.
Santander, im Campus Center, Gebäude A4 4.
Mobilfunknetz
Durch die Nähe zu Frankreich kann es vorkommen, dass bei automatischer
Netzwahl ein französischer Netzbetreiber ausgewählt wird, wodurch höhere
Kosten entstehen können. Dies kann vermieden werden, indem die automatische Netzwahl deaktiviert und der Netzbetreiber manuell ausgewählt wird.
Fachausstellung
Besuchen Sie auch die Fachausstellung der Verlage im Foyer des Tagungsgebäudes B4 1.
Taxiruf
Taxi Saarbrücken e.G.
Saarbrücker Taxigenossenschaft e.G.
 0681 33 0 33
 0681 55 0 00
15
i
Allgemeine Informationen
i
anreise nach saarbrücken
Mit dem Flugzeug
Flüge von Berlin oder Hamburg nach Saarbrücken dauern rund 1 Std. 15 Min.
Der Flughafen Saarbrücken liegt ca. 12 km vom Uni-Campus entfernt. Die Busfahrt (Linie R10, Umstieg in der Innenstadt) zum Campus dauert zwischen 45
Min. und 1 Stunde und kostet 2,60 Euro. Die Fahrt mit dem Taxi dauert 15–20
Min. und kostet ca. 20 Euro.
Nicht aus Deutschland anreisende Teilnehmer können eine Anreise über
Frankfurt oder Paris in Betracht ziehen, die per ICE/TGV-Schnellzug nach
Saarbrücken angebunden sind (Dauer ca. 2 Std.).
Mit der Bahn
Es bestehen direkte Verbindungen zum Saarbrücker Hauptbahnhof mit dem
ICE/TGV aus Frankfurt/Mannheim und aus Paris. Einzelne Direktverbindungen gibt es auch aus Richtung Dresden/Leipzig und Salzburg/München nach
Saarbrücken. Aus Richtung Mainz, Trier und Straßburg gelangen Sie mit dem
Regionalverkehr nach Saarbrücken.
Mit dem Auto
Aus Richtung Mannheim/Karlsruhe: Autobahn A6 bis zur Ausfahrt „St. Ingbert
West“, von dort der Beschilderung „Universität“ folgen.
Aus Richtung Koblenz/Trier: A1 bis Autobahnkreuz Saarbrücken, dort auf A8
Richtung Karlsruhe wechseln, am Dreieck Friedrichsthal auf die A623 Richtung Saarbrücken/Frankreich bis zur Abfahrt Sulzbach. Durch Sulzbach fahren, an der großen Kreuzung geradeaus auf die L126, dann der Beschilderung
„Universität“ folgen.
Aus Paris/Metz bzw. Straßburg: Autobahn A4 bis zur Ausfahrt 40, dort auf die
A320/E50 Richtung Saarbrücken wechseln, weiter geradeaus auf die A6 Richtung Mannheim bis zur Ausfahrt „St. Ingbert West“, von dort der Beschilderung „Universität“ folgen.
Aus Luxemburg: Autobahn A620 Richtung Saarbrücken, am Dreieck Saarbrücken auf die A6 Richtung Mannheim. An der Ausfahrt „St. Ingbert West“ abfahren und der Beschilderung „Universität“ folgen. Das ist zwar ein Umweg,
erspart Ihnen aber den recht komplizierten Weg durch die Innenstadt.
16
Allgemeine Informationen
Mit dem Auto – Parken
Theoretisch können Sie im Campus-Innenbereich parken (Parkzone A). Tatsächlich stehen hier nur sehr wenige Parkplätze zur Verfügung und die Parkgebühren sind sehr hoch. Wir empfehlen daher, in einem der Parkhäuser (siehe Campusplan) zu parken.
Nach Ziehen eines Tickets an der Schrankenanlage können Sie im CampusInnenbereich parken. Die Ausfahrt erfolgt nach vorheriger Bezahlung der
Parkgebühr an einem der Kassenautomaten.
Auch bei kostenlosem Parken oder Durchfahrt muss ein Parkticket gezogen
und bei der Ausfahrt, ohne vorherige Bezahlung, wieder in die Schrankenanlage eingeben werden. Sollte der Parkschein verloren gehen, müssen 15,– EUR
bezahlt werden.
Parkgebühren für den Campus-Innenbereich (Parkzone A):
Die ersten 60 Min. Kurzparken oder für eine Durchfahrt sind kostenlos. Die
zweite Stunde kostet 4,– EUR, danach jede angefangene Stunde 2,– EUR. Die
Maximalkosten belaufen sich auf 15,– EUR pro Tag.
Parken in den Parkhäusern P2 und P3
In den Parkhäusern P2 (Uni Mitte) und P3 (Uni Ost) sind die ersten 60 Min.
kostenlos. Danach kostet jede angefangene Stunde 1,– EUR. Die Maximalkosten belaufen sich auf 3,– EUR pro Tag.
17
i
Allgemeine Informationen
i
anfahrt zum campus
Busanbindung des Campus Saarbrücken
Die Konferenz wird auf dem Hauptcampus der Universität des Saarlandes
stattfinden, der sich außerhalb des Stadtzentrums befindet. Der Campus
ist vom Hauptbahnhof und vom Stadtzentrum („Wilhelm-Heinrich-Brücke“,
„Rathaus“, „Haus der Zufkunft“) aus mit mehreren Buslinien in 10–15 Minuten zu erreichen.
Kostenloses Konferenzticket
Die Benutzung der öffentlichen Verkehrsmittel in Saarbrücken ist für alle Konferenzteilnehmer über ein kostenloses Konferenzticket abgedeckt. Online regristrierte Teilnehmer erhalten bereits vor der Anreise einen Nachweis, den
sie bei ihrer ersten Fahrt zum Campus vorlegen können. Bei der Anmeldung
vor Ort erhalten Sie ein Namensschild mit dem Logo der Saarbahn. Dieses dient
als Konferenzticket.
Fußweg zum Tagungsgebäude
Die Tagung wird in den Gebäuden B3 1 (Geschichts- und Kulturwissenschaften), B3 2 (Bibliothek Geschichts- und Kulturwissenschaften) und B4 1 (Rechtsund Wirtschaftswissenschaften) auf dem Campus stattfinden. Der Fußweg
von der Bushaltestelle „Universität Campus“ zu den Konferenzgebäuden ist
auf dem Campusplan auf der hinteren Umschlagsseite des Tagungsbandes verzeichnet.
Fahrplanauskunft
Alle Busverbindungen finden Sie auf der Internetseite des Saarländischen
Verkehrsverbundes (SaarVV) www.saarfahrplan.de, bzw. in der Handy-App
„Saarfahrplan“ oder über die App der Deutschen Bahn „DB Navigator“. Geben
Sie „Universität Campus“ als Ziel an.
Abfahrtszeiten in den Stoßzeiten
Auf den folgenden Seiten finden Sie eine Auflistung der Busabfahrtszeiten
zum Campus und zurück in die Innenstadt zu den Stoßzeiten der DGfSTagung. Außerhalb der Stoßzeiten verkehrt rund alle zehn Minuten ein Bus
zum Campus bzw. zurück in die Innenstadt.
18
Allgemeine Informationen
Hauptbahnhof → Universität Campus
Ab Hauptbahnhof verkehrt die Linie 124 direkt zum Campusgelände. Alternativ können Sie auch mit der Saarbahn (S1, verkehrt alle 7,5 Min.) zwei Stationen zur Johanneskirche/zum Rathaus fahren und dort in die anderen Buslinien (101, 102, 109) zum Campus umsteigen.
Linie
124
S1 (Umstieg)
Richtung
Universität Busterminal
Brebach, Kleinblittersdorf,
Sarreguemines
Dauer
0:14
0:18-0:21
Abfahrt
08:01
08:03
08:16
08:18
08:31
08:33
08:46
08:48
09:03
09:16
09:18
09:33
09:46
Linie
124
S1, 101
124
S1, 102
124
S1, 101
124
S1, 102
S1, 101
124
S1, 102
S1, 101
124
Dauer
00:14
00:18
00:14
00:21
00:14
00:18
00:14
00:21
00:18
00:14
00:21
00:18
00:14
Ankunft
08:15
08:21
08:30
08:39
08:45
08:51
09:00
09:09
09:21
09:30
09:39
09:51
10:00
19
i
Allgemeine Informationen
i
Rathaus (Johanneskirche) → Universität Campus
An der Haltstelle „Rathaus“ verkehren drei Buslinien (101, 102, 109) direkt zum Campusgelände. In unmittelbarer Nähe (ca. 60 m) befindet sich
die Saarbahn-Haltestelle „Johanneskirche“, wo Sie Anschluss aus Richtung
Hauptbahnhof haben.
Linie
101
102
109
Richtung
Dudweiler Dudoplatz
Dudweiler Dudoplatz
Universität Busterminal
Dauer
0:12
0:15
0:15
Abfahrt
08:09
08:11
08:24
08:39
08:45
08:54
09:09
09:15
09:24
09:39
09:45
Linie
101
109
102
101
109
102
101
109
102
101
109
Dauer
00:12
00:15
00:15
00:12
00:15
00:15
00:12
00:15
00:15
00:12
00:15
20
Ankunft
08:21
08:26
08:39
08:51
09:00
09:09
09:21
09:30
09:39
09:51
10:00
Allgemeine Informationen
Universität Campus → Innenstadt, Hauptbahnhof
Von der Haltestelle „Universität Campus“ verkehren drei Buslinien (101, 102,
109) direkt in die Innenstadt zur Haltestelle „Johanneskirche“, von wo Sie Anschluss zur Saarbahn (S1, verkehrt alle 7,5 Min.) bzw. zu anderen Buslinien haben. Die Buslinie 124 fährt vom Campus direkt zum „Hauptbahnhof “, hält aber
auch an der Haltestelle „Haus der Zukunft“, die nur ca. 200 m Fußweg von der
Haltestelle „Johanneskirche“ entfernt ist. Sie können also mit allen vier Linien
in die Innenstadt fahren.
Linie
101
102
109
Dauer
0:13
0:15
0:15
124
Richtung
Füllengarten Siedlung
Altenkessel Talstraße
Rabbiner-Rülf-Platz,
Goldene Bremm
Hauptbahnhof, Betriebshof
Abfahrt
17:07
17:17
17:20
17:27
17:35
17:37
17:47
17:53
17:57
18:05
18:07
18:17
18:20
18:27
18:35
18:37
18:47
18:50
18:57
19:05
19:07
19:17
19:20
19:27
Linie
109
102
124
101
124
109
102
124
101
124
109
102
124
101
124
109
102
124
101
124
109
102
124
101
Dauer
00:15
00:15
00:15
00:13
00:15
00:15
00:15
00:15
00:13
00:15
00:15
00:15
00:15
00:13
00:15
00:15
00:15
00:15
00:13
00:15
00:15
00:15
00:15
00:13
Ankunft
17:22
17:32
17:35 (Hbf.)
17:40
17:50 (Hbf.)
17:52
18:02
18:08 (Hbf.)
18:10
18:20 (Hbf.)
18:22
18:32
18:35 (Hbf.)
18:40
18:50 (Hbf.)
18:52
19:02
19:05 (Hbf.)
19:10
19:20 (Hbf.)
19:22
19:32
19:35 (Hbf.)
19:40
0:15
21
i
Allgemeine Informationen
i
campusplan und gebäudepläne
Die DGfS-Jahrestagung 2017 findet auf dem Campus Saarbrücken der Universität des Saarlandes statt. An der Universität werden die Gebäude mit einer Kombination aus Buchstaben (Bereich) und Zahlen (Gebäudeensemble, Gebäude)
benannt. Die Tagungsgebäude tragen die Bezeichnungen B3 1 (Geschichtsund Kulturwissenschaften), B3 2 (Bibliothek Geschichts- und Kulturwissenschaften) und B4 1 (Audimax; Rechts- und Wirtschaftswissenschaften). Die
Gebäude befinden sich in unmittelbarer Nachbarschaft. Der Großteil der Arbeitsgruppen, das Tagungsbüro, die Kaffeepausen, die Plenarvorträge und die
Verlagsausstellung befinden sich im Gebäude B4 1.
Der Fußweg von der Bushaltestelle „Universität Campus“ vor dem Campus
Center zum Tagungsgebäude ist auf der Übersichtskarte auf der nächsten Seite und noch einmal in Farbe auf der hinteren Umschlagsseite des Tagungsbands verzeichnet.
22
Allgemeine Informationen
i
23
Allgemeine Informationen
i
Übersicht Tagungsgebäude B4 1
Der folgende Plan zeigt eine Übersicht über das Tagungsgebäude B4 1. Das
Tagungsbüro (Anmeldung, Gepäckaufbewahrung) befindet sich in Raum 0.21
im hinteren Bereich des Gebäudes. Die Plenarvorträge finden im Audimax
(0.01), die AGs 5 bis 13 finden in den Seminarräumen 0.04 bis 0.07 bzw.
0.22 bis 0.26 statt. Die Verlagsausstellung, die Posterausstellung CL und die
Pausenverpflegung befinden sich im Foyer des Gebäudes.
24
Allgemeine Informationen
Übersicht Tagungsgebäude B3 1 und B3 2
Der folgende Plan zeigt eine Übersicht über die beiden Tagungsgebäude B3 1
und B3 2. Die AGs 1 bis 4 finden in den Seminarräumen 0.12 bis 0.14 (B3 1)
bzw. 0.03 (B3 2) statt. Die Verlagsausstellung, die Posterausstellung CL und
die Pausenverpflegung befinden sich im Foyer des Gebäudes B4 1.
25
i
Allgemeine Informationen
i
raumübersicht
Anmeldung
B4 1, 0.21
Tagungsbüro
B4 1, 0.21
Kaffeepausen
B4 1, Foyer
Gepäckaufbewahrung
B4 1, 0.21
Verlagsausstellung
B4 1, Foyer
Dienstag, 07. 03. 2017
Lehramtsinitiative
C5 1–3
CL-Tutorium
A2 2, 2.14
Doktorandenforum
C9 3
(Jägerheim)
ALP-Tagung
B3 2, 0.03
Mittwoch, 08. 03. 2017 bis Freitag, 10. 03. 2017
Plenarvorträge
B4 1, 0.01
(Audimax)
AG 1
Informationstheoretisch basierte Modellierung
sprachlicher Variation im Kontext
B3 2, 0.03
AG 2
Information structuring in discourse
B3 1, 0.14
AG 3
Sekundäre Information und ihre sprachliche
Kodierung
B3 1, 0.13
26
Allgemeine Informationen
AG 4
Encoding language and linguistic information
in historical corpora
B3 1, 0.12
AG 5
Sprachliche Kodierungs-Asymmetrien,
Gebrauchsfrequenz und Informativität
B4 1, 0.26
AG 6
Prosody in syntactic encoding
B4 1, 0.25
AG 7
Wen wurmt der Ohrwurm? – An
interdisciplinary, cross-lingual perspective on
the role of constituents in multi-word
expressions
B4 1, 0.24
AG 8
Non-canonical verb positioning in main clauses
B4 1, 0.23
AG 9
Towards an ontology of modal flavors
B4 1, 0.22
AG 10
Polysemy and coercion of clause-embedding
predicates
B4 1, 0.07
AG 11
Coercion Across Linguistic Fields
B4 1, 0.06
AG 12
Morphologische Variation – Theorie und
Empirie
B4 1, 0.05
AG 13
Register in linguistic theory: Modeling
functional variation
B4 1, 0.04
DGfS-Mitgliederversammlung
B4 1, 0.18
CL-Mitgliederversammlung
B3 1, 0.11
Posterausstellung CL
B4 1, Foyer
i
27
Allgemeine Informationen
i
essen und trinken auf dem campus
Verpflegung in den Kaffeepausen
Während der Konferenz werden im Foyer des Tagungsgebäudes B4 1 verschiedene Erfrischungen (Kaffee, Tee, Kaltgetränke, kleiner Imbiss) angeboten.
Mensa des Studentenwerks
In der Mensa des Studentenwerks (Gebäude D4 1) erhalten Sie jeweils von
11:30–14:15 Uhr verschiedene Mittagsgerichte (Selbstzahler).
Bei Aufgang A (rot) wird das Komplettmenü (Vorspeisensuppe, Beilagensalat,
Hauptgericht, Nachspeise), bei Aufgang B (blau) das vegetarische Komplettmenü (Beilagensalat, Hauptgericht, Nachspeise) angeboten. Ebenfalls über Aufgang B erreichen Sie den Freeflow, wo Sie die Wahl haben zwischen einem
Low-Fat Gericht, einem Fischgericht, einem preiswerten Tellergericht sowie
vegetarischen und regionalen Gerichten. Außerdem befindet sich dort ein Salatbüffet. Aufgang C (gelb) ist in der vorlesungsfreien Zeit geschlossen.
Die Gerichte von Freeflow können Sie an den Kassen bezahlen, die Barzahlung akzeptieren. Für das Komplett- und vegetarische Menü müssen Sie zuerst
an den Barkassen im Voraus bezahlen und erhalten dann einen Bon für die Essensausgabe, da dort nur bargeldlos mit dem Studentenausweis bezahlt werden kann.
Den Speiseplan können Sie in der „Uni Saar“-App einsehen oder auf den Internetseiten des Studentenwerks www.studentenwerk-saarland.de (Essen >
Mensa Saarbrücken).
In der Tagungsmappe, die Sie bei der Anmeldung vor Ort erhalten, finden Sie
zusätzlich ein Informationsblatt zu der Funktionsweise der Mensa des Studentenwerks.
Gastronomie auf dem Campus
Neben der Mensa gibt es noch zahlreiche weitere gastronomische Angebote
auf dem Campus:
– Mensacafé
im Erdgeschoss der Mensa des Studentenwerks, Gebäude D4 1, Frühstücksbüffet, Mittagessen (auch vegane Gerichte im Angebot), Heißund Kaltgetränke, Backwaren, Öffnungszeiten: 07:45–15:00 Uhr.
– Sportlertreff (Mensa des Landessportverbands)
auf dem Gelände der Hermann-Neuberger-Sportschule (schließt im
28
Allgemeine Informationen
Südwesten an den Campus an), Frühstück (07:00–09:00 Uhr), Mittagessen (11:30–13:30 Uhr), Abendessen (18:00–20:00 Uhr).
www.lsvs.de
– Ausländer-Café (AC)
in Gebäude A3 2, täglich zwei wechselnde Stammessen, feste Speisekarte, auch vegetarische und vegane Gerichte.
www.ac-bistro.de
– Juristen-Café
in Gebäude B4 2, Frühstück, heiße Theke, Stammessen, Öffnungszeiten:
07:30–15:00 Uhr.
www.juristen-cafe.de
– Philo-Café
in Gebäude C5 2, persische und internationale Küche, Bio, Halal, vegan,
vegetarisch, Öffnungszeiten: 07:30–17:00 Uhr.
– Canossa
im Untergeschoss der Mensa des Studentenwerks, Gebäude D4 1, Pizza,
Pasta, Salate, Öffnungszeiten: 10:30–22:00 Uhr.
www.canossa.de
– Café Unique
im Campus Center, Gebäude A4 4, Heiß- und Kaltgetränke, belegte Brötchen, Backwaren, Salate, Öffnungszeiten: 07:00–18:00 Uhr.
www.unique-cafe.de
– Fastfood Heroes
neben Uni-Markt, in Gebäude C5 5, Mittagessen, Fastfood, Gegrilltes
(auch vegetarisch), Öffnungszeiten: ab 11:30.
Supermarkt
Auf dem Campus (Gebäude C5 5) gibt es einen Supermarkt (Uni-Markt), in
dem man sich neben Kosmetikartikeln, Schreibwaren, Briefmarken oder Zeitschriften auch mit belegten Brötchen, Salaten, Backwaren, Heiß- und Kaltgetränken, Milchprodukten oder Süßigkeiten eindecken kann. Öffnungszeiten:
08:00–17:00 Uhr.
29
i
Allgemeine Informationen
i
30
Allgemeine Informationen
i
gastronomie in der stadt
€ = Preisspanne unter 10 EUR, €€ = Preisspanne 11–20 EUR, €€€ = Preisspanne
21–40 EUR, €€€€ = Preisspanne gehoben
Französische und gesternte Küche
Name, Adresse & Kontakt
Preis
GästeHaus Klaus Erfort ***
Mainzer Str. 95, 66121 Saarbrücken
 http://www.gaestehaus-erfort.de  0681 9582682
€€€€
La Bastille
Kronenstr. 1b, 66111 Saarbrücken
 https://www.restaurant-labastille.de  0681 31064
€€
Tempelier
Kappenstr. 9, 66111 Saarbrücken
 0681 68894101
€€
Tomate 2
Schloßstr. 2, 66117 Saarbrücken
 http://www.tomate2.de  0681 57846
€€
Deutsche und saarländische Küche
Name, Adresse & Kontakt
Preis
Haus Brück
Mainzer Str. 6, 66111 Saarbrücken
 haus-brueck.de  0681 9508800
€€€
Fürst Ludwig
Am Ludwigsplatz 13, 66117 Saarbrücken
 http://www.fuerst-ludwig.de  0681 52573
€
Herzenslust
Nauwieserplatz 5, 66111 Saarbrücken
 herzenslust-saar.de  0681 68832126
€€€
31
Allgemeine Informationen
i
32
Name, Adresse & Kontakt
Preis
Die Kartoffel
St. Johanner Markt 32, 66111 Saarbrücken
 0681 36217
€€
Café Kostbar
Nauwieserstr. 19, 66111 Saarbrücken
 http://www.cafekostbar.de  0681 374360
€€
Krottenschenke
Saarstraße 13, 66111 Saarbrücken
 0681 9101990
€
Odeon
St. Johanner Markt 33, 66111 Saarbrücken
 0681 372469
€€
Oro
St. Johanner Markt 7, 66111 Saarbrücken
 http://oro-online.de  0681 9388663
€€€
Ratskeller
Rathausplatz 1, 66111 Saarbrücken
 http://www.ratskeller-saarbruecken.de
 0681 9101708
€€
Schnokeloch
Kappenstr. 6, 66111 Saarbrücken
 http://www.schnokeloch.de  0681 33397
€€
Schwimmschiff
Berliner Promenade 18, 66111 Saarbrücken
 0681 371498
€€€
Stiefel (Bräu & Restaurant)
Am Stiefel 2, 66111 Saarbrücken
 http://www.der-stiefel.de  0681 936450
€€
Tante Maja
St. Johanner Markt 8, 66111 Saarbrücken
 http://tantemaja.de  0681 30589
€€
Allgemeine Informationen
Name, Adresse & Kontakt
Preis
Tellerrand
Am Stiefel 1, 66111 Saarbrücken
 www.facebook.com/tellerrand.saarbruecken/
 0681 95813022
€€
Gasthaus Zahm
Saarstr. 6, 66111 Saarbrücken
 0681 9591317
€€
i
Mediterrane Küche
Name, Adresse & Kontakt
Preis
Bar Celona
Berliner Promenade 5, 66111 Saarbrücken
 https://celona.de/  0681 93866523
€
Dubrovnik
Kupfergasse 5, 66111 Saarbrücken
 http://www.dubrovnik-sb.de  0681 33752
€€€
Café Especial
Kronenstr. 1, 66111 Saarbrücken
 http://www.cafe-especial.com  0681 3906619
€€
Fruit de Mer
Bleichstr. 28, 66111 Saarbrücken
 http://www.fruits-de-mer.online  0681 31416
€€
Iguana
Mainzer Str. 2, 66111 Saarbrücken
 http://iguana-restaurant.de  0681 397744
€€
die konkrete Utopie
Hohenzollernstr. 79, 66117 Saarbrücken
 http://www.diekonkreteutopie.de  0681 9385794
€€
Langenfeld
St. Johanner Markt 5, 66111 Saarbrücken
 http://www.langenfeld-cafe.de/  0681 5953669
€€
33
Allgemeine Informationen
i
Name, Adresse & Kontakt
Preis
L’Osteria
Trierer Str. 33, 66111 Saarbrücken
 http://losteria.de/restaurant/saarbruecken/
 0681 99274820
€€
Roma
Hafenstr. 12, 66111 Saarbrücken
 http://www.roma-saarbruecken.de  0681 45470
€€€
s’Olivo
Mainzer Str. 10, 66111 Saarbrücken
 http://www.leidinger-saarbruecken.de/restaurants/
solivo/  0681 9327-0
€€€
Sur
Eisenbahnstr. 6, 66117 Saarbrücken
 http://www.sur-picadas-bar.de  01516 5775437
€€
To Steki
Am Kieselhumes 42b, 66123 Saarbrücken
 http://to-steki.com/Willkommen  0681 36710
€€
Trattoria Toskana
Fröschengasse 18–20, 66111 Saarbrücken
 http://www.original-trattoria-toscana.de/
 0681 9101895
€€€
Viva Zapata
Mainzer Str. 8, 66111 Saarbrücken
 www.vivazapata.info  0681 375647
€€
34
Allgemeine Informationen
i
Asiatische Küche
Name, Adresse & Kontakt
Preis
Café Bali
Rotenbergstr. 10, 66111 Saarbrücken
 http://cafebali.de  0681 3799313
€€ Hashimoto (Restaurant)
Cecilienstr. 7, 66111 Saarbrücken
 http://hashimoto-saar.de  0681 398034
€€€
Hashimoto (Brasserie)
Fröschengasse 1, 66111 Saarbrücken
 http://hashimoto-saar.de  0681 3906563
€€€€
Hokkaido
Mainzer Str. 152, 66121 Saarbrücken
 http://www.hokkaido-sb.de/  0681 9681100
€€
Kimdo
Mainzer Str. 61, 66121 Saarbrücken
 www.kimdo-restaurant.de  0681 9685343
€€€
Krua Thai
Mainzer Str. 71, 66121 Saarbrücken
 0681 64695
€€€
Mei Thai
Kappenstr. 7, 66111 Saarbrücken
 http://meithai-saar.de  0681 3908202
€€€
Oishii
Berliner Promenade 17 - 19, 66111 Saarbrücken
 http://www.okinii.de/standorte/saarbrucken/
 0681 9066876
€€€
Osaka
Dudweilerstr. 1, 66111 Saarbrücken
 http://www.sushi-saarbruecken.de  0681 3799066
€€
Siam
Mainzer Str. 22, 66111 Saarbrücken
 http://www.restaurant-siam.de  0681 68507001
€€
35
Allgemeine Informationen
i
Name, Adresse & Kontakt
Preis
Tbilissi [Georgische Küche]
Saarstr. 13, 66111 Saarbrücken
 0681 91005645
€€
Angloamerikanische Küche
Name, Adresse & Kontakt
Preis
Baker Street
Mainzer Str. 8, 66111 Saarbrücken
www.bakerstreetsb.de  0681 95812454
€€€
Buffalo Steakhouse
Betzenstr. 10, 66111 Saarbrücken
 http://www.buffalo-saar.de/  0681 32772
€€€
Burgerei
Fröschengasse 2, 66111 Saarbrücken
 http://www.dieburgerei.com  0681 50062400
€
Cafés
36
Name, Adresse & Kontakt
Preis
Alex
Saarstr. 15, 66111 Saarbrücken
 www.dein-alex.de/saarbruecken  0681 37995950
€€
The Bakery
Gerberstr. 7, 66111 Saarbrücken
 0681 95818570
€€
Gelümmel und Getümmel (Kindercafé )
Nauwieserstr. 35, 66111 Saarbrücken
 http://www.geluemmel.de  0681 95123575
€€
Kulturcafé
St. Johanner Markt 24, 66111 Saarbrücken
 http://www.kulturcafe-saarbruecken.de/
 0681 3799200
€€
Allgemeine Informationen
Name, Adresse & Kontakt
Preis
Madame
Mainzer Str. 4, 66111 Saarbrücken
 0681 32963
€€
Nauwies
Nauwieserstr. 22, 66111 Saarbrücken
 http://nauwies.de/  0681 35743
€€
Platanen-Café
Am Staden, 66121 Saarbrücken
 http://www.derstaden.de/  0681 67757
€€
Tesorito
Cecilienstraße 16, 66111 Saarbrücken
 http://www.cafe-am-schloss.com  0681 582621
€€
Café Thonet
Katholisch-Kirch-Str. 24, 66111 Saarbrücken
 www.facebook.com/thonet.sb
€€
Ubu Roi
Cecilienstr. 15, 66111 Saarbrücken
 01573 0141037
€€
Café am Schloss
Schlossplatz, 66121 Saarbrücken
 https://tesorito.de/  0681 9101655
€€
Tante Anna
Türkenstr. 3, 66111 Saarbrücken
 0681 398178
€€
i
37
Allgemeine Informationen
i
Cocktails und Longdrinks
38
Name, Adresse & Kontakt
Preis
Arnie & Jules
Berliner Promenade 19, 66111 Saarbrücken
 http://arnieandjules.de/  0176 31148365
€€
Esplanade
Nauwieserstr. 5, 66111 Saarbrücken
 www.esplanade-sb.de  0681 8596566
€€€
Feinkost Schmitt
Nassauerstr. 14, 66111 Saarbrücken
 www.feinkost-schmitt.de
€€
Green Buddha
Bleichstr. 7–9, 66111 Saarbrücken
 0681 9101543
€€
Kurze Eck
Nauwieserstr. 15, 66111 Saarbrücken
 http://www.unserviertel.de/KurzeEck/eck.htm
 0681 36864
€€
Manhattan
Türkenstr. 1, 66111 Saarbrücken
 http://www.manhattan-sb.de/  0681 3798250
€€
Milk
St. Johanner Markt 17–19, 66111 Saarbrücken
 0681 9101760
€€
Mono
Nauwieserstr. 38, 66111 Saarbrücken
 0681 9066967
€€
Nautilus
Försterstraße 17, 66111 Saarbrücken
 0681 76042110
€€
Odeon
St. Johanner Markt 33, 66111 Saarbrücken
 0681 372469
€€
Allgemeine Informationen
Name, Adresse & Kontakt
Preis
Ovid
Kaltenbachstr. 6, 66111 Saarbrücken
 0681 371317
€€
Sankt J
St. Johanner Markt 3, 66111 Saarbrücken
 0681 371198
€€
Tante Maja
St. Johanner Markt 8, 66111 Saarbrücken
 http://tantemaja.de  0681 30589
€€
Shotz
Fröschengasse 6, 66111 Saarbrücken
 http://shotz-bar.com/locations/saarbruecken/
€
i
Weinlokale
Name, Adresse & Kontakt
Preis
Hauck Alte Feuerwache
Am Landwehrplatz 1, 66111 Saarbrücken
 https://fine-time.de/hauck  0681 9381640
€€
Weinstube
Bleichstr. 32, 66111 Saarbrücken
 0681 3908849
€€
„Die Winzer“ – Kunst- und Kulturclub
Martin-Luther-Str. 5, 66111 Saarbrücken
 http://www.diewinzer.com  0151 51146862
€€
Pubs
Name, Adresse & Kontakt
Preis
Baker Street
Mainzer Str. 8, 66111 Saarbrücken
 www.bakerstreetsb.de  0681 95812454
€€€
39
Allgemeine Informationen
i
Name, Adresse & Kontakt
Preis
Old Murphy’s
St. Johanner Markt 11, 66111 Saarbrücken
 http://www.oldmurphys.de/  0681 91032217
€€
Wally’s Irish Pub
Katholisch-Kirch-Str. 1, 66111 Saarbrücken
 www.wallys-irishpub.de/  0681 50060563
€€
40
Allgemeine Informationen
i
sehenswürdigkeiten und ausgehtipps
St. Johanner Markt
Der St. Johanner Markt mit seinen Boutiquen, Kneipen, Bistros und Restaurants ist das Herzstück des Saarbrücker Lebens. Hier trifft man sich oder man
bummelt durch die malerischen Gässchen rund um den Markplatz. Seit 1978
ist der Altstadtbereich Fußgängerzone. Vom barocken Marktbrunnen (1759
von Stengel entworfen) gibt es eine Sichtachse zum Schloss, früher eine weitere zur Ludwigskirche und von dieser aus eine zurück zum Schloss – das sog.
„Stengelsche Dreieck“.
Saarbrücker Schloss
Aus der Burg „Castellum Sarrabrucca“ entwickelte sich im 17. Jahrhundert ein
Renaissanceschloss, von dem heute noch unterirdische Anlagen vorhanden
sind. Nach dessen Zerstörung ließ Fürst Wilhelm Heinrich im 18. Jahrhundert
durch seinen Baumeister Stengel eine neue barocke Residenz errichten.
Ludwigskirche
Die Ludwigskirche, als Hauptstück einer „Place-Royale“-Architektur, ist die
Krönung des unermüdlichen Schaffens von Baumeister Stengel. Sie gilt als eine der stilreinsten und schönsten evangelischen Barockkirchen in Deutschland, vergleichbar mit dem Michel in Hamburg oder der Frauenkirche in Dresden. Zusammen mit dem Ludwigsplatz, den umliegenden Palais und Beamtenhäusern bildet sie ein einzigartiges Barockensemble, das 1775 fertiggestellt
wurde.
Nauwieserviertel
Die „Nauwies“ ist seit einiger Zeit das Ausgehviertel des Saarlandes. Die Gründerzeitbauten beherbergen zahlreiche Studentenkneipen, Bierwirtschaften,
Künstlerateliers und kleine Geschäfte. In den Abendstunden sind die Straßen
bevölkert von Saarbrückern und Saarländern aus dem Umland, die hier ihr
Feierabendbier trinken oder sich mit Freunden treffen. Ein Programmkino
(Kino 8 21 ), ein Kultur- und Werkhof (Nauwieser 19), eine Spielstätte des Staatstheaters (Alte Feuerwache) oder das freie Szenetheater „Theater im Viertel“
sorgen dafür, dass auch der kulturelle Aspekt nicht zu kurz kommt.
Weitere Informationen unter www.saarbruecken.de/tourismus.
Quelle: Internetauftritt der Landeshauptstadt Saarbrücken
41
Allgemeine Informationen
i
ausflüge in die umgebung
Weltkulturerbe Völklinger Hütte (ca. 13 km)
1986 stillgelegt und 1994 von der UNESCO zum Weltkulturerbe erhoben, ist die
Völklinger Hütte das weltweit einzige authentisch erhaltene Eisenwerk aus
der Blütezeit der Eisen- und Stahlindustrie. Ein 5000m langer Parcours führt
die Besucher durch ein eindrucksvolles Zeugnis von Ingenieurskunst und Industriekultur des frühen 20. Jahrhunderts: gigantische Maschinen in der Gebläsehalle, die sechs Hochöfen im frei begehbaren Hüttenpark oder den einzigartigen Schräge-Aufzug.
Saarschleife und Baumwipfelpfad (ca. 60 km)
Die Saarschleife ist das wohl bekannteste Postkartenmotiv des Saarlandes. Die
beste Sicht hat man vom Aussichtspunkt „Cloef “. Seit einigen Monaten kann
man von hier aus auch in bis zu 23 Meter Höhe über dem Waldboden in unberührter Natur auf dem Baumwipfelpfad spazieren gehen. An Lichtungen vorbei führt der Weg hinauf in die mächtigen Wipfel, bevor er in den architektonisch einmaligen Aussichtsturm abzweigt.
Kulturpark Bliesbruck-Reinheim (ca. 25 km)
An der deutsch-französischen Grenze, im schönen Tal der Blies, ist ein einzigartiges Projekt grenzüberschreitender Zusammenarbeit entstanden. Langjährige archäologische Forschungen bezeugen eine über 8000-jährige wechselvolle Geschichte. Unter anderem wird hier ein keltisches Fürstinnengrab
ausgestellt, einer der bedeutendsten Grabfunde aus keltischer Zeit in Mitteleuropa.
Metz (ca. 70 km)
Historisches Stadtbild; Centre Pompidou Metz (moderne und zeitgenössische
Kunst)
Trier (ca. 100 km)
Weltkulturerbestätten aus römischer Zeit (Porta Nigra, Dom, Basilika, Kaiserthermen, Amphitheater, Römerbrücke)
Luxemburg (ca. 100 km)
Historisches Stadtbild; europäische Institutionen
Quelle: Internetauftritte der Stadt Saarbrücken und des Kulturparks Bliesbruck-Reinheim
42
Sprachwissenschaft
bei Winter
ptashnyk, stefaniya
beckert, ronny
wolf-farré, patrick
wolny, matthias (Hg.)
Gegenwärtige
Sprachkontakte im
Kontext der Migration
2016. 344 Seiten. (Schriften des
Europäischen Zentrums für Sprachwissenschaften (EZS), Band 5)
Geb. € 45,–
isbn 978-3-8253-6551-6
Universitätsverlag
winter
Heidelberg
girnth, heiko
hofmann, andy alexander
Politolinguistik
2016. 74 Seiten. (Literaturhinweise
zur Linguistik, Band 4)
E-Book: € 14,–
isbn 978-3-8253-7590-4
helmer, henrike
Analepsen in
der Interaktion
Semantische und sequenzielle
Eigenschaften von Topik-Drop
im gesprochenen Deutsch
2016. 274 Seiten, 35 Abbildungen.
(OraLingua, Band 13)
Geb. € 45,–
isbn 978-3-8253-6577-6
petkova, marina
Multiples
Code-Switching:
wendelstein, britta
Gesprochene Sprache im
Vorfeld der AlzheimerDemenz
Linguistische Analysen im Verlauf von präklinischen Stadien
bis zur leichten Demenz
2016. 200 Seiten, xv Seiten Anhang,
38 Abbildungen, 9 Tabellen.
Kart. € 39,–
isbn 978-3-8253-6596-7
2016. ca. 308 Seiten. (OraLingua,
Band 14)
Geb. ca. € 48,–
isbn 978-3-8253-6528-8
peterson, john
Sprache und Migration
2015. x, 101 Seiten. (Kurze Einführungen in die germanistische
Linguistik, KEGLI, Band 18)
Kart. € 13,–
isbn 978-3-8253-6454-0
www.winter-verlag.de
Ein Sprachkontaktphänomen
am Beispiel der Deutschschweiz
Die Fernsehberichterstattung
zur ‚Euro 08‘ und andere Vorkommenskontexte aus interaktionsanalytischer Perspektive
Programmübersicht
Programm
allgemeines programm
Programm
Dienstag, 07. 03. 2017
09:00–18:00
Tagung der Arbeitsgemeinschaft
Linguistische Pragmatik
B3 2, 0.03
10:00–17:00
Computerlinguistik Tutorium
A2 2, 2.14
09:00–17:30
Doktorandenforum
C9 3
14:30–19:00
Lehramtsinitiative
C5 1–3
ab 19:00
Warming Up
Stiefelbräu
Mittwoch, 08. 03. 2017
08:00–09:00
Anmeldung
B4 1, 0.21
(Tagungsbüro)
09:00–09:30
Begrüßung mit einem Grußwort der
Ministerpräsidentin des Saarlandes,
Annegret Kramp-Karrenbauer
B4 1, 0.01
(Audimax)
09:30–10:30
Plenarvortrag: Manfred Krifka
Focus in answers and questions in
Commitment-Space Semantics
B4 1, 0.01
(Audimax)
10:30–11:00
Verleihung des Wilhelm von
Humboldt-Preises
B4 1, 0.01
(Audimax)
11:00–11:30
Kaffeepause
B4 1, Foyer
11:30–12:30
Plenarvortrag:
Anthony Kroch & Beatrice Santorini
Detecting grammatical properties in usage
data
B4 1, 0.01
(Audimax)
12:30–13:45
Mittagspause
Mitgliederversammlung der Sektion
Computerlinguistik
B3 1, 0.11
Arbeitsgruppensitzungen
B3 1, B3 2, B4 1
13:45–15:45
47
Programm
15:45–16:30
Programm
Kaffeepause
Postersession Computerlinguistik (Teil 1)
B4 1, Foyer
16:30–18:00
Arbeitsgruppensitzungen
B3 1, B3 2, B4 1
19:00
Sektempfang im Rathaus
Rathausplatz 1, 66111 Saarbrücken
(Haltestelle „Johanneskirche“)
Donnerstag, 09. 03. 2017
09:00–10:30
Arbeitsgruppensitzungen
B3 1, B3 2, B4 1
10:30–11:15
Kaffeepause
Postersession Computerlinguistik (Teil 2)
B4 1, Foyer
11:15–12:45
Arbeitsgruppensitzungen
B3 1, B3 2, B4 1
12:45–13:45
Mittagspause
Postersession Computerlinguistik (Teil 3)
DFG-Informationsveranstaltung
B4 1, Foyer
B3 1, 0.11
13:45–14:45
Arbeitsgruppensitzungen
B3 1, B3 2, B4 1
15:00–18:00
Mitgliederversammlung der DGfS
B4 1, 0.18
19:00
Geselliger Abend im „Café & Bar Celona“
Berliner Promenade 5, 66111 Saarbrücken
Freitag, 10. 03. 2017
09:00–10:00
Plenarvortrag: John Goldsmith
Learning morphology
B4 1, 0.01
(Audimax)
10:00–11:00
Plenarvortrag: Matthew Crocker
A neurocomputational model of surprisal in
comprehension
B4 1, 0.01
(Audimax)
11:00–11:30
Kaffeepause
B4 1, Foyer
11:30–14:00
Arbeitsgruppensitzungen
B3 1, B3 2, B4 1
48
Programm
ag-programme
Programm
AG 1
Informationstheoretisch basierte Modellierung
sprachlicher Variation im Kontext
Elke Teich, Bernd Möbius & Vera Demberg
Raum: B3 2, 0.03
Mittwoch, 08. 03. 2017
13:45–14:15
Elke Teich, Bernd Möbius & Vera Demberg
Introduction
14:15–15:15
Hinrich Schütze
Embeddings or visions of the future of the lexicon
15:15–15.45
Stefania Degaetano-Ortlieb, Merel Scholmann, Stefan
Fischer, Elke Teich & Vera Demberg Saarbrücken
An information-theoretic account on the diachronic
development of discourse connectors in scientific writing
15:45–16:30
Kaffeepause
16:30–17:00
Natalia Levshina
Does information density help? A multifactorial Bayesian
analysis of help + (to) Vinf in twenty varieties of English
17:00–17:30
Fabian Tomaschek & Benjamin V. Tucker
Modeling segmental durations using the Naive Discriminative
Learner
17:30–18:00
Jessie Nixon & Catherine Best
High-variability distributions lower perceptual certainty
during acoustic cue acquisition
49
Programm
Donnerstag, 09. 03. 2017
09:00–09:30
Geertje van Bergen
Expectation management in interaction: Discourse particles
signal surprisal of upcoming referents
09:30–10:00
Ekaterina Kravtchenko, Ashutosh Modi, Vera Demberg,
Ivan Titov & Manfred Pinkal
Does UID affect rate of pronominalization?
10:00–10:30
Olga Seminck & Pascal Amsill
Predicting processing cost of anaphora resolution
10:30–11:15
Kaffeepause
11:15–11:45
Ayush Jain, Vishal Singh, Sumeet Agarval & Rajakrishnan
Raijkumar
Uniform Information Density models for language production:
A comparative study of Hindi and English
11:45–12:15
Elli Tourtouri, Francesca Delogu & Matthew Crocker
Over-specifications efficiently manage referential entropy in
situated communication
12:15–12:45
Eric Meinhardt
Non-stationarity and other critical mathematical problems
for channel coding-based explanations of variation in
language production
12:45–13:45
Mittagspause
13:45–14:45
Tal Linzen
Entropy in language comprehension
Programm
Freitag, 10. 03. 2017
11:30–12:00
Maria Piñango, Ashwini Deo & Muye Zhang
Accessing locative interpretations in ”have” sentences:
Context-sensitive variability
12:00–12:30
Aaron Steven White & Kyle Rawlins
Entropy predicts uncertainty in subcategorization frame
distributions
50
Programm
12:30–13:00
Robin Lemke, Ingo Reich & Eva Horch
Information density constrains article omission. An
experimental approach
13:00–13:30
David Howcroft, Cynthia A. Johnson & Rory Turnbull
German morphosyntactic change is consistent with an optimal
encoding hypothesis
13:30–14:00
Abschlussdiskussion
Information structuring in discourse
AG 2
Anke Holler, Katja Suckow, Barbara Hemforth &
Israel de la Fuente
Raum: B3 1, 0.14
Mittwoch, 08. 03. 2017
13:45–14:15
Arndt Riester, Lisa Brunetti & Kordula De Kuthy
Principles of information-structure and discourse-structure
analysis
14:15–14:45
Jet Hoek, Jacqueline Evers-Vermeul & Ted J.M. Sanders
Complete and independent? Reconsidering discourse
segmentation basics
14:45–15:45
Nicholas Asher
Goals, coherence and information structure in dialogue
15:45–16:30
Kaffeepause
16:30–17:00
Daniel Altshuler & Dag Haug
The semantics of provisional, temporal anaphors and
cataphors
17:00–17:30
Sophia Döring
Modal particles and their influence on discourse structure
17:30–18:00
Manfred Stede
Segmentation and topic annotation of German newspaper
editorials
51
Programm
Programm
Donnerstag, 09. 03. 2017
09:00–09:30
Ciro Greco & Liliane Haegeman
Discourse frame setters and the syntax of subject-initial V2 in
Standard Dutch and West Flemish
09:30–10:00
Rosemarie Lühr
Zum Ausdruck von (non-)at-issueness in alten Sprachen
10:00–10:30
Cassandra Freiberg
Identifying basic units of discourse structure in corpus
languages: The case of Ancient Greek
10:30–11:15
Kaffeepause
11:15–11:45
Marco Coniglio, Roland Hinterhölzl
Discourse structure of relative constructions: a crosslinguistic
and diachronic study on the interaction between mood, syntax
and event structure
11:45–12:15
Claudia Poschmann
Embedded NRRCs and discourse structure
12:15–12:45
Clare Patterson & Claudia Felser
Cleft focus and accessibility: Online vs. offline differences
12:45–13:45
Mittagspause
13:45–14:45
Hannah Rohde
Small building blocks, multiple threads, and large
repercussions
Programm
Freitag, 10. 03. 2017
11:30–12:00
Sofiana Chiriacescu
Script knowledge effects on information structuring
12:00–12:30
Merel Scholman & Vera Demberg
The influence of context on the interpretation of the segments
in a discourse relation
12:30–13:00
Umut Özge & Klaus von Heusinger
Inferrable and partitive indefinites in topic position
52
Programm
13:00–13:30
13:30–14:00
Tommaso Raso, Maryualê Malvessi Mittmann & Frederico
Amorim Cavalcante
A Cross-Linguistic study on Topic within the framework of the
Language into Act Theory
Matthijs Westera
How the symmetry problem solves the symmetry problem
Alternate speakers
Markus Bader & Yvonne Portele
German pronouns in discourse: Information structure versus surface properties
Jennifer S. Cole & Stefan Baumann
Accounting for context and variability in a prominence-based model of discourse
meaning
AG 3
Sekundäre Information und ihre sprachliche
Kodierung
Daniel Gutzmann, Katharina Turgay
Raum: B3 1, 0.13
Mittwoch, 08. 03. 2017
13:45–14:15
Daniel Gutzmann & Katharina Turgay
Secondary information and linguistic encoding – an
introduction
14:15–14.45
Stefan Hinterwimmer
The Bavarian discourse particle fei as a marker of
non-at-issueness
14:45–15:15
Holden Härtl
The name-informing and the distancing use of sogenannt
(‘so-called’). A pragmatic account
15:15–15:45
Laura Dörre & Josef Bayer
The processing of secondary information conveyed by German
modal particles
53
Programm
Programm
Programm
15:45–16:30
Kaffeepause
16:30–17:00
Elena Castroviejo Miró & Berit Gehrke
Non-truth-conditional intensification. The case of ‘good’
17:00–17:30
Osamu Sawada
Interpretations of the embedded expressive motto in Japanese:
Varieties of meaning and projectivity
17:30–18:00
Claudia Borgonovo
Though as a marker of humbleness
Donnerstag, 09. 03. 2017
09:00–09:30
Dirk Kindermann
Fragmented Contexts
09:30–10:00
Ana Aguilar Guevara
Literal and enriched meaning of sentences with weak definites
and bare singulars
10:00–10:30
Patricia Amaral
Full NPs as personal pronouns: Reference, truth-conditional
meaning, and use-conditional content
10:30–11:15
Kaffeepause
11:15–12:15
Judith Tonhauser
Relating not-at-issueness to the question under discussion
12:15–12:45
Mira Grubic
Additives and accommodation
12:45–13:45
Mittagspause
13:45–14:15
Elsi Kaiser
(Un)expected secondary content in Finnish: additives and
scalars
14:15–14:45
Agata Renans, Nadine Bade & Joseph P. DeVeaugh-Geiss
Presupposition triggers in a cross-linguistic perspective:
maximize presupposition vs. obligatory implicatures in
Ga(Kwa)
54
Programm
Freitag, 10. 03. 2017
11:30–12:30
Robert Henderson & Eric McCready
Dog-whistles and the at-issue/non-at-issue distinction
12:30–13:00
Corinna Trabandt, Alex Thiel, Emanuela Sanfelici & Petra
Schulz
Appositive interpretation of relative clauses – Is prosody the
cue?
13:00–13:30
Alexander Haselow
Ad-hoc shifts from primary to secondary information in
spontaneous speech
13:30–14:00
Mikaela Petkova-Kessanlis
Parenthesen und ihre Funktionen in didaktisch aufbereiteten
linguistischen Texten
Programm
Alternate speakers
Kalle Müller
Satzadverbiale und (non-)at-issueness
AG 4
Encoding language and linguistic information in
historical corpora
Kerstin Eckart, Carolin Odebrecht
Raum: B3 1, 0.12
Donnerstag, 09. 03. 2017
11:15–12:15
Mathilde Hennig
Basic categories in multi layered grammatical annotation
12:15–12:45
Svetlana Petrova
Particle verb constructions in historical German and what
corpus studies reveal about them
12:45–13:45
Mittagspause
55
Programm
13:45–14:15
Lisa Dücker, Stefan Hartmann & Renata Szczepaniak
Annotating a multiregional diachronic corpus of Early New
High German handwritten texts
Programm
Freitag, 10. 03. 2017
11:30–12:00
Maarten Janssen
TEITOK: Combining language and linguistic information
without compromise
12:00–12:30
Zarah Weiß & Gohar Schnelle
Annotation of an Early New High German Corpus: The
LangBank Pipeline
12:30–13:00
Cătălina Mărănduc, Cenel-Augusto Perez, Ludmila
Malahov & Alexandru Colesnicov
A diachronic corpus for Romanian (RoDia)
13:00–13:30
Vera Faßhauer & Henry Seidel
Limitations and possibilities of Early New High German text
annotation:private ducal correspondences in Early Modern
Germany
13:30–14:00
Nicoletta Puddu
Encoding sociolinguistic variables in a corpus of Medieval
Sardinian texts
56
Programm
AG 5
Sprachliche Kodierungs-Asymmetrien,
Gebrauchsfrequenz und Informativität
Martin Haspelmath, MPI-SHH Jena & Universität Leipzig
Programm
Raum: B4 1, 0.26
Mittwoch, 08. 03. 2017
13:45–14:45
Gertraud Fenk-Oczlon
Frequency, coding asymmetries, and the constant flow of
linguistic information
14:45–15.15
Laura Becker & Matías Guzmán Naranjo
Coding asymmetries, frequency and predictability: The case of
to vs from
15:15–15:45
Alice Blumenthal-Dramé & Bernd Kortmann
Causal and concessive relations: Typology meets cognition
15:45–16:30
Kaffeepause
16:30–17:00
Livio Gaeta
Diachrony as a source of coding asymmetries
17:00–17:30
Geoffrey Khan
Asymmetry in the historical development of the copula in
Neo-Aramaic
17:30–18:00
Simon Kasper
The asymmetry between morphology and word order with
respect to informativity
Donnerstag, 09. 03. 2017
09:00–09:30
Sander Lestrade
Simulating the development of encoding asymmetries in
argument marking
09:30–10:00
Natalia Levshina
Explaining coding asymmetries: Frequency or informativity?
10:00–10:30
Matti Miestamo
Making sense of the asymmetry between affirmation and
negation
57
Programm
Programm
10:30–11:15
Kaffeepause
11:15–11:45
Na’ama Pat-El
The Semitic Perfect and the problem of zero subjects
11:45–12:15
Dirk Pijpops & Freek Van de Velde
Processing shapes grammar: But whose processing are we
talking about?
12:15–12:45
Ulrike Schneider & Britta Mondorf
Why bring is doing the splits: Exploring transitivity as an
explanatory factor for coding asymmetries
12:45–13:45
Mittagspause
13:45–14:15
Martin Haspelmath
On the scope of the form-frequency correspondence principle
14:15–14:45
Karsten Schmidtke-Bode
On the optionality of boundary markers (and pro-forms) of
subordinate clauses
Freitag, 10. 03. 2017
11:30–12:00
Ilja Seržant
Towards functional motivation for the reduced third person
indexing
12:00–12:30
Helen Sims-Williams
A diachronic mechanism for form-frequency asymmetries in
inflectional paradigms
12:30–13:00
Eva van Lier & Marlou van Rijn
Differential possessive marking of arguments in action
nominalizations: A typological survey
13:00–13:30
Jingting Ye
Coding asymmetry between independent and dependent
pronominal possessors: A cross-linguistic study
13:30–14:00
Natalia Zaika
Markedness disharmony in Basque
58
Programm
Prosody in syntactic encoding
AG 6
Gerrit Kentner, Joost Kremers
Raum: B4 1, 0.25
Programm
Mittwoch, 08. 03. 2017
13:45–14:15
Gerrit Kentner & Joost Kremers
Introduction to „prosody in syntactic encoding“
14:15–14.45
Katy Carlson
Comparative constructions are strongly affected by focus
structure
14:45–15:15
Tina Bögel
Ambiguities at the interface: production and comprehension
15:15–15:45
Anna Dannenberg, Stefan Werner, Vainio Martti & Suni
Antti
Prosodic and syntactic structures in spontaneous speech: a
wavelet-based approach to prosodic modelling
15:45–16:30
Kaffeepause
16:30–17:00
Luciana Lucente
Prosodic boundaries constraints by discursive elements
17:00–17:30
Laetitia Leonarduzzi & Sophie Herment
How Prosody and syntax interact: the case of english it-clefts
17:30–18:00
Daniela Kolbe-Hanna & Judith Manzoni
Prosody as a determinant of the syntactic status of „I think“
Donnerstag, 09. 03. 2017
09:00–10:00
Arto Anttila
Sentence stress in presidential speeches
10:00–10:30
Stavros Skopeteas
Clitic placement and implications for the syntax-prosody
mapping
10:30–11:15
Kaffeepause
59
Programm
11:15–11:45
Elyse Jamieson
Prosody and tag question forms in Glasgow Scots
11:45–12:15
Johannes Heim
Syntactic integration of sentential intonation
12:15–12:45
Marta Wierzba
The ordering of syntax-prosody-interpretation mapping rules
12:45–13:45
Mittagspause
13:45–14:15
Volker Struckmeier
Towards a non-centralized, subtractive architecture of
grammar
14:15–14:45
Manuela Korth
Syntax and prosody in parallel systems
Programm
Freitag, 10. 03. 2017
11:30–12:00
Hisao Tokizaki & Jiro Inaba
Prosodic constraint on prenominal modification
12:00–12:30
Nomi Erteschik-Shir, Gunlög Josefsson & Björn Köhnlein
Prosody determines word order: the case of Mainland
Scandinavian object shift
12:30–13:00
Farhat Jabeen & Miriam Butt
Urdu/Hindi Polar kya at the syntax-pragmatics-prosody
interface
13:00–13:30
Emmanuel Schang, François Nemo & Fanny Krimou
Uttered sentences, prosody and word order
13:30–14:00
Janina Reinhardt
The place of où: where’s the difference?
60
Programm
AG 7
Wen wurmt der Ohrwurm? – An interdisciplinary,
cross-lingual perspective on the role of constituents
in multi-word expressions
Programm
Sabine Schulte im Walde, Eva Smolka
Raum: B4 1, 0.24
Mittwoch, 08. 03. 2017
13:45–14:45
Gary Libben
Morphological superposition and the nature of the mental
lexicon
14:45–15:15
Stefanie Rößler, Thomas Weskott & Anke Holler
N1-accessibility as a matter of compound processing
15:15–15:45
Serkan Uygun & Ayse Gürel
Factors affecting the processing of compounds in the second
language
15:45–16:30
Kaffeepause
16:30–17:00
Anna Hätty & Michael Dorna
Exploring the impact of transparency and productivity of
multiword term constituents on single-word term
identification
17:00–17:30
Saskia E. Lensink & Harald Baayen
Multi-word units in a discriminative framework
17:30–18:00
Melanie Bell & Martin Schäfer
Semantic entropy measures and the semantic transparency of
noun-noun compounds
Donnerstag, 09. 03. 2017
9:00–9:30
Giannina Iordachioaia, Lonneke van der Plas & Glorianna
Jagfeld
The role of the head in the interpretation of deverbal
compounds
61
Programm
Programm
9:30–10:00
Inga Henneke
Semantic transparency and variation in nominal syntagmatic
compounds in Romance languages
10:00–10:30
Sandro Pezzele, Maria Silvia Micheli & Elisabetta Jezek
The different meanings of ‚a‘: Capturing qualia relations of
Italian complex nominals with distributional semantics
10:30–11:15
Kaffeepause
11:15–11:45
Fabienne Cap
Approximating compound compositionality based on word
alignments
11:45–12:15
Marco S. G. Senaldi, Gianluca E. Lebani & Alessandro
Lenci
Exploring idiomaticity with distributional semantics and
entropy
12:15–12:45
Corina Dima, Jianqiang Ma & Erhard Hinrichs
Evaluating semantic composition of German compounds
12:45–13:45
Mittagspause
13:45–14:45
Marco Marelli
Understanding compound words: A new perspective from
compositional systems in distributional semantics
Non-canonical verb positioning in main clauses
AG 8
Mailin Antomo, Sonja Müller
Raum: B4 1, 0.23
Mittwoch, 08. 03. 2017
13:45–14:15
Mailin Antomo & Sonja Müller
Introduction
14:15–15.15
Hubert Truckenbrodt
Verb position and speech acts in German
15:15–15:45
Volker Struckmeier & Sebastian Kaiser
Only just CP? Rethinking classification criteria for sentence
types theories
62
Programm
15:45–16:30
Kaffeepause
16:30–17:00
Imke Driemel
Variable verb positions in German exclamatives
17:00–17:30
Rita Finkbeiner
Verb-final w-clauses in headlines
17:30–18:00
Nathalie Staratschek
Non-canonic verb positioning in disintegrated verb-final
weil-clauses in German
Programm
Donnerstag, 09. 03. 2017
09:00–09:30
Sonja Müller
On the role of doch in V1- and Wo-VE-clauses in German
09:30–10:00
Janina Beutler
V1-declaratives and assertion
10:00–10:30
Ulrike Demske
The brief history of V-final root clauses in Early New High
German
10:30–11:15
Kaffeepause
11:15–11:45
N. N.
wird noch bekanntgegeben
11:45–12:15
Katja Barnickel
Obligatory V1-order in German SLF-coordination
12:15–12:45
Julia Bacskai-Atkari
Clause typing in main clauses and V1 conditionals in
Germanic
12:45–13:45
Mittagspause
13:45–14:45
Heike Wiese
The hidden life of V3: an overlooked word order option in
Germanic „V2 languages“
63
Programm
Freitag, 10. 03. 2017
11:30–12:00
Gereon Müller
Structure removal in complex prefields
12:00–12:30
Werner Frey
A syntactic condition for supposed multiple fronting in
German
12:30–13:00
Oliver Bunk
Insights into the processing of non-canonical sentence
structures in German: The case of V3 matrix declaratives in
informal Standard German
13:00–13:30
Liliane Haegeman & Ciro Greco
Frame setters and V3 patterns in West Flemish
13:30–14:00
Artemis Alexiadou & Terje Lohndal
V3 in Germanic: A comparison of urban vernaculars and
heritage languages
Programm
Towards an ontology of modal flavors
AG 9
Ryan Bochnak, Anne Mucha, Kilu von Prince
Raum: B4 1, 0.22
Mittwoch, 08. 03. 2017
13:45–14:45
Aynat Rubinstein
Flavors of existential / possessive modals
14:45–15:15
Matthew Mandelkern & Jonathan Philipps
Force: topicalization, context-sensitivity, and morality
15:15–15:45
Eva Csipak
The early production of conditionals
15:45–16:30
Kaffeepause
16:30–17:00
Jakob Maché
Decomposing modal flavours
64
Programm
17:00–17:30
Vera Hohaus & Jozina Vander Klok
Weak necessity modals and modal flavor: The view from
Paciran Javanese
17:30–18:00
Lisa Matthewson & Hubert Truckenbrodt
Modal flavor/modal force interactions in German
Programm
Donnerstag, 09. 03. 2017
9:00–9:30
Fabian Bross & Daniel Hole
Swabian, German, Chinese and German Sign Language:
multi-source convergence on a cartographic array of modal
flavors
9:30–10:00
Dana Kratochvílová
The (in)stability of modal flavors: The case of English modals
and their Spanish equivalents
10:00–10:30
Ana Werkmann Horvat
Modal force and flavor as semantic restrictors of possible
double modal combinations in Croatian
10:30–11:15
Kaffeepause
11:15–11:45
Lavi Wolf
Modal concord is not modal concord
11:45–12:15
Adam Marushak
Veritic semantics for epistemic modals
12:15–12:45
Dietmar Zaefferer
The ontological cookbook of modal categories: There are more
flavors than you think
65
Programm
AG 10
Polysemy and coercion of clause-embedding
predicates
Marie-Luise Popp, Barbara Stiebels
Programm
Raum: B4 1, 0.07
Mittwoch, 08. 03. 2017
13:45–14:45
Marie Luise Popp & Barbara Stiebels
Introduction: The role of polysemy and coercion in
clause-embedding predicates (with aspecial focus on
NEG-Raising)
14:45–15.45
Elizabeth Bogal-Allbritten
Basic pieces, complex meanings: Building attitudes in Navajo
and beyond
15:45–16:30
Kaffeepause
16:30–17:30
Alda Mari
Systematic polysemy of non-factive epistemic and fiction verbs
in Italian: evidence from mood variation
17:30–18:00
Michelle Li
Mood selection of two visual perception verbs in Hong Kong
Cantonese
Donnerstag, 09. 03. 2017
09:00–10:00
A. Marlijn Meijer
How believe so is different from believing it
10:00–10:30
Julian Rentzsch & Liljana Mitkovska
Complementation strategies with the verb ‘know’ in Balkan
Turkish compared to Standard Turkish and Macedonian
10:30–11:15
Kaffeepause
11:15–12:15
Stephen Wechsler
Clause embedding sound emission verbs
12:15–12:45
Natalia Serdobolskaya
Against classification of complement-taking predicates: the
case for mental verbs
66
Programm
12:45–13:45
Mittagspause
13:45–14:45
Kerstin Schwabe
‘Chameleon predicates’: variation in argument realization
Programm
Freitag, 10. 03. 2017
11:30–12:30
Itamar Kastner
Coercing propositional anaphora
12:30–13:30
Marie-Louise Lind Sørensen & Kasper Boye
Utterance-predicate complementation
13:30–14:00
Final discussion
Coercion Across Linguistic Fields (CALF)
AG 11
Hana Filip, Laura Kallmeyer
Raum: B4 1, 0.06
Mittwoch, 08. 03. 2017
13.45–14.15
Hana Filip & Laura Kallmeyer
Introduction
14.15–14.45
Paul Dekker
Coercion as a Methodological Tool
14.45–15.45
Margaret Grant, Sonia Michniewicz & Jessica Rett
Immediate commitment, but no evidence for a coercion cost, in
individual/degree polysemy
16.30–17.00
Maria Victoria Escandell-Vidal
Evidential effects and mismatch resolution
17.00–17.30
Silvia Gumiel-Molina, Norberto Moreno-Quibén & Isabel
Pérez-Jiménez
Interpreting quantifiers: the case of ’ligeramente’ + A in
Spanish
17.30–18.00
Carmen Dobrovie-Sorin & Marta Donazzan
Mass-Count Meaning Shifts and the Mass-Count Distinction
67
Programm
Donnerstag, 09. 03. 2017
9.00–10.00
Nicholas Asher
Two takes on coercion and co-composition: combining
distributional and formal semantics
10.00–10.30
Peter Sutton & Hana Filip
Mass- to-count coercion in ‘granular’ nouns
10.30–11.15
Kaffeepause
11.15–11.45
Andreas Blümel & Hans-Joachim Particke
Revisiting German wieder: A restitutive prefix and its coerced
object
11.45–12.45
Eva Csipak & Sarah Zobel,
Postnominal temporal adverbs in the German prefield
12.45–13.45
Mittagspause
13.45–14.15
Masanori Deguchi
Phonological Coercion in Pawnee
14.15–14.45
Ruben van de Vijver, Vicky Tsouni & Kim Strütjen
Coercion in loanword adaptation
Programm
Freitag, 10. 03. 2017
11.30–12.00
William Babonnaud, Laura Kallmeyer & Rainer Osswald
Polysemy and Coercion – A Frame-based Approach Using LTAG
and Hybrid Logic
12.00–12.30
Elisabetta Jezek
Empirical evidence for the study of coercion mechanisms in
predicate-argument composition
12.30–13.00
José Manuel Igoa & María del Carmen Horno
An Experimental Study on Coercion in Spanish Adjectival
Phrases
13.00–14.00
Final discussion
68
Programm
Morphologische Variation – Theorie und Empirie
AG 12
Antje Dammel, Oliver Schallert
Raum: B4 1, 0.05
Programm
Mittwoch, 08. 03. 2017
13:45–14:45
Hans-Olav Enger
Thoughts on morphemes and variation in a Scandinavian
perspective
14:45–15:15
Anja Hasse
The interaction of morphological and phonological variation.
A case study on Zurich German
15:15–15:45
Alfred Lameli & Alexander Werth
Phonotaktik jenseits der Silbe – Quantitative Analysen zur
Relevanz phonotaktischer Strukturen für die Morphologie
15:45–16:30
Kaffeepause
16:30–17:00
Pia Bergmann
/t/-Realisierungen in komplexen Wörtern im Deutschen – Ein
Fall für hybride Modelle?
17:00–17:30
Lars Bülow, Hannes Scheutz & Dominik Wallner
Zum Wandel des Drei-Formen-Plurals im
salzburgisch-bayerischen Grenzgebiet? Eine Pilotstudie zur
intraindividuellen morphologischen Variation
17:30–18:00
Hanna Fischer
Veni, vidi, vici − Ich bin gekommen, habe gesehen und gesiegt?
Perfektexpansion und Präteritumschwund im Deutschen
Donnerstag, 09. 03. 2017
9:00–9:30
Christian Zimmer
Das No Blur Principle und nominalmorphologische Variation
im Deutschen
9:30–10:00
Arjen Versloot, Elzbieta Adamczyk & Eric Hoekstra
A dynamic Systems Approach to Morphological irregularity
69
Programm
Programm
10:00–10:30
Dankmar Enke & Roland Mühlenbernd
Case marking variation – an evolutionary perspective
10:30–11:15
Kaffeepause
11:15–11:45
Sophie Ellsäßer
Content and form of Upper German case paradigms. A
formalist approach to usage-based data?
11:45–12:15
Tanja Ackermann
Das possessive -s im Deutschen: Entwicklung, Variation und
theoretischer Status
12:15–12:45
Astrid Niebuhr
Mit einem stetigem Anstieg – Variation in der Adjektivflexion
12:45–13:45
Mittagspause
13:45–14:15
Helmut Weiß & Seyna Maria Carlucci-Dirani
Stark oder schwach? – Oder: wie die Informationsstruktur
morpho-syntaktische Variationen steuert
14:15–14:45
Stephanie Leser-Cronau
Variation of agreement forms: investigations on lexical
hybrids in German dialects
Freitag, 10. 03. 2017
11:30–12:00
Göz Kaufmann
Complex heads in Pomeranian: Some morphosysntactic
considerations
12:00–12:30
Caroline Döhmer
Dat hätt ké inten anescht gemaach ginn – IPP und supinale
Formen bei luxemburgischen Modalverben
12:30–13:00
Ann-Marie Moser
Negative concord im Alemannischen: Eine morphologische
Erklärung
13:00–14:00
Michele Loporcaro
Concurrent gender systems in Europe: a new look at the
Asturian neuter
70
Programm
AG 13
Register in linguistic theory: Modeling functional
variation
Aria Adli, Anke Lüdeling
Programm
Raum: B4 1, 0.04
Mittwoch, 08. 03. 2017
13:45–14:15
Anke Lüdeling & Aria Adli
Register and variation: An introduction
14:15–15:15
Elisabeth Verhoeven & Nico Lehmann
Recursive embedding and register variation
15:15–15.45
Jason Grafmiller
Register specificity in the English genitive alternation. Do
variable cues reflect different grammars?
15:45–16:30
Kaffeepause
16:30–17:00
Stella Neumann, Stefan Evert & Gert De Sutter
Register-specific interference in translation
17:00–17:30
Kerstin Kunz, Erich Steiner, Ekaterina
Lapshinova-Koltunski, José Martinez & Katrin Menzel
Patterns of cohesion as dependent variables in a contrastive
study of registers in English and German
Donnerstag, 09. 03. 2017
09:00–10:00
Felix Bildhauer & Roland Schäfer
Automatic register annotation for linguistic research?
10:00–10:30
Thomas Haider
A functional stylistics for genre and register
10:30–11:15
Kaffeepause
11:15–11:45
Jürgen Trouvain & Friederike Kern
Prosodic aspects of style and register of live sports
commentaries in radio and television
11:45–12:15
Johanna Stahnke
Prosodic variation in French: self-repairs in conceptual
distance and proximity
71
Programm
Programm
12:15–12:45
Markus Egg
Register dependency of deliberate metaphor
12:45–13:45
Mittagspause
13:45–14:15
Richard Waltereit
Assessing the role of intra-speaker variation for language
change
14:15–14:45
Gohar Schnelle & Karin Donhauser
Register variation in OHG: evidence for register-based
variation in the recordings of OHG
Freitag, 10. 03. 2017
11:30–12:30
Benedikt Szmrecsanyi
The register-specificity of variation grammars
12:30–13:00
Ines Rehbein
Register variation in the use of DRDs in argumentative texts
13:00–13:30
Tatjana Scheffler
Register variation across social media
13:30–14:00
Roland Meyer & Luka Szusich
Functional style categories vs. bottom-up corpus analysis:
Empirical adequacy and usefulness of register ascriptions in
contemporary Russian
72
PAPERBACKS OF
SELECT TITLES
NEW AT DE GRUYTER MOUTON
ff Paperbacks
of select titles 24 months after the hardcover appears
120 titles available
ff Titles of special interest to students and the general public
ff Over
Select Titles (Complete List: www.degruyter.com/paperback)
RRP € 24.95 [D] / US$ 35.00 / £ 18.99
ISBN 978-1-5015-1267-4
RRP € 19.95 [D] / US$ 19.95 / £ 14.99
ISBN 978-3-11-048211-9
RRP € 24.95 [D] / US$ 24.95 / £ 18.99
ISBN 978-3-11-048478-6
RRP € 19.95 [D] / US$ 19.95 / £ 14.99
ISBN 978-3-11-048492-2
RRP € 19.95 [D] / US$ 28.00 / £ 14.99
ISBN 978-3-11-048329-1
RRP € 19.95 [D] / US$ 19.95 / £ 14.99
ISBN 978-3-11-048535-6
RRP € 29.95 [D] / US$ 29.95 / £ 22.99
ISBN 978-3-11-048808-1
RRP € 19.95 [D] / US$ 19.95 / £ 14.99
ISBN 978-3-11-048805-0
degruyter.com/paperback
For more information on paperbacks at De Gruyter,
check our website or contact your editor!
Plenarvorträge
Plenarvorträge
Focus in answers and questions in Commitment-Space Semantics
Mittwoch
08.03.2017
09:30 – 10:30
B4 1, 0.01
Manfred Krifka
Humboldt-Universität zu Berlin, ZAS
[email protected]
PV
Focus has long been seen as one of the dimensions of information structure,
used to identify parts of an utterance that come with non-realized alternatives
that are relevant for the interpretation of the whole utterance. The classical use
of focus is in answers to constituent questions, where the focus of the answer
corresponds to the wh-constituent of the question, as in Which flower did Mary
give to John? — Mary gave the ROSE to John. But focus also occurs in polarity
questions, cf. Did Mary give the TULIP to John?, which may be answered by a
simple Yes, or by a combination of No and another sentence, e.g. Mary gave the
ROSE to John.
In this talk, I will show how both focus in answers and focus in polarity questions can be interpreted as indicating alternatives. For polarity questions, it
will be essential to couch the proposal in the framework of Commitment-Space
Semantics, developed by the author in recent work. Commitment Spaces are
common grounds with their possible continuations; a question does not add
information to the common ground but restricts the possible continuations to
those that answer the question. This allows for the notion of a monopolar question, a polarity questions that propose just one proposition, not two or more,
as in other semantic frameworks. I argue that Did Mary give the TULIP to John?
is such a biased monopolar question that wants to elicit from the addressee the
assertion of Mary gave the tulip to John. Focus indicates that there are alternative elicitations of assertions of the form Mary gave X to John. I will show that
the disjunction of these alternative elicitations are identical to the meaning of
the constituent question What / Which flower did Mary give to John? If the addressee negates the intended assertion proposed by Did Mary give the TULIP to
John?, the discourse state falls back to the disjunction of the alternative elicited
assertions, i.e. the meaning of the constituent question, and the answer carries
focus, as predicted.
I will also consider focus in wh-questions, as in What did MARY give to John?,
arguing that focus indicates alternative wh-questions here. This focus is typical of contrastive topics, which indicate a partial answer to a complex question
under discussion.
77
Plenarvorträge
Detecting grammatical properties in usage data
Mittwoch
08.03.2017
11:30 – 12:30
B4 1, 0.01
Anthony Kroch
University of Pennsylvania
Beatrice Santorini
University of Pennsylvania
[email protected]
[email protected]
PV
A well-known limitation on the utility of corpus data for linguistic research
is the absence of negative evidence, just the evidence that is readily available
in the data of acceptability judgments. Of course, in the case of historical investigations, judgment data is simply unavailable. In this situation, it is tempting but dangerous to assume that non-occurring configurations are ungrammatical. A better approach, widely adopted with the growing availability of
digital corpora, especially annotated ones, is to make use of the frequency
information in the corpora to infer properties of the grammars underlying
observed usage patterns. The most obvious patterns are diachronic developments in which a form either arises from nothing or disappears, and it has always been assumed that these cases reflect grammatical change. A more ambitious use of frequency information has been work in the spirit of Kroch’s “Constant Rate Effect” (Kroch 1989). In such work, evidence is assembled to show
that distinct linguistic environments sharing a common innovative grammatical feature will evolve together over time. The CRE has by now been replicated
sufficiently often to be accepted as reliable. Most recently, for example, Zimmermann (2015) has carried out a large scale replication involving a dataset of
more than 50K instances of the English do-support environment.
Less well-known than the CRE is the pattern reported in Santorini (1993),
Taylor (1994) and elsewhere where we see that grammatical options (for example, “extraposition”) that are not undergoing change tend to be stable in
their frequency of use in corpus data. This work also reports that grammatically independent options, like the extraposition of one XP or another or both
in a clause that contains two such phrases, tend to be statistically independent;
that is to say, if the probability of one occurrence of extraposition is p, then the
probability of two occurrences will be approximately p · p. Largely ignored in
the literature, however, is the contrapositive implication that when options
are statistically linked, we have evidence of grammatical linkage.
In this paper, we present evidence from four languages for which we have
parsed historical corpora: English, French, Icelandic and Yiddish (Kroch and
Taylor 2000, Martineau and et al. 2009, Wallenberg et al. 2011, Santorini 2008)
78
Plenarvorträge
of statistical linkage with grammatical implications and also of the loss of such
linkage over the course of time. The data on which we rely is word order inside VP, where we find that these languages undergo a shift from XV to VX in
multiple stages, two of which can only be distinguished by the presence versus absence of statistical linkage between different word order options. The
pattern we have found, stated within the framework of antisymmetric syntax,
is that XV surface word order has sources in leftward movements of two distinct types: (1) remnant scrambling of VP with the verb itself stranded in v and
(2) scrambling of an XP argument/adjunct with VP remaining in situ. Since,
under option 2, more than one XP may scramble and since, under option 1,
XPs can be stranded after the verb via a sequence of XP scrambling followed
by remnant VP scrambling, an identical range of surface orders is produced
by the two options. Only quantitative evidence allows us to distinguish them.
Concretely, we find the following quantitative patterns in our languages:
1. From their earliest attested periods, the languages exhibit leftward
movement of single XPs across the verb, as expected under the XP scrambling option.
2. At the same time, in the earlier periods, the frequency of verb-final order
in clauses with multiple XPs in pre-verbal position is much higher than
expected, given the frequencies of single XP movement.
3. As reported for Ancient Greek in Taylor (1994), the frequencies of leftward movement of single XPs of a given syntactic type are largely independent of the presence of other XPs in the clause.
4. After initial periods with an excess of multiple XP in pre-verbal position,
the frequency of XP>V orders declines in all four languages to that predicted by the rates of single XP scrambling.
From these results, we conclude that the loss of surface OV order in our languages proceeds in three stages. In the first, which antecedes our earliest
records, the remnant scrambling of VP begins to be lost, leading to an alternation between XP>V and V>XP surface orders. At this time, XP>V order in
clauses with one VP- internal constituent becomes ambiguous between a VPmovement derivation and one in which single XPs scramble leftward. In the
second stage, the VP-movement option disappears so that XP>V order is always derived by XP scrambling. Finally, XP scrambling itself disappears or becomes restricted to quantificational expressions.
References: • Theresa Biberauer. 2003. Reconsidering the EPP and Spec-TP in Germanic. In Luisa
Astruc and Marc Richards, editors, Cambridge Occasional Papers in Linguistics (COPiL), number 1, pp.
79
PV
Plenarvorträge
PV
100–120. Department of Linguistics, University of Cambridge. • Roland Hinterhöltzl. 2006. Scrambling, Remnant Movement, and Restructuring in West Germanic. Oxford University Press, Oxford and
Cambridge, MA. • Anthony Kroch. 1989. Reflexes of grammar in patterns of language change. Language Variation and Change, 1:199–244. • Anthony Kroch and Ann Taylor. 2000. Penn-Helsinki Parsed
Corpus of Middle English. http://www.ling.upenn.edu/hist-corpora/, second edition. • France Martineau and Anthony Kroch et al. 2009. Corpus MCVF, Modé liser le changement: les voies du français. University of Ottawa, first edition. • Beatrice Santorini. 1993. The rate of phrase structure change in the
history of Yiddish. Language Variation and Change, 5:257–283. • Beatrice Santorini. 2008. Penn Yiddish Corpus. Contact author for access. • Ann Taylor. 1994. The change from SOV to SVO in Ancient
Greek. Language Variation and Change, 6: 1–37. • Joel C. Wallenberg, Anton Karl Ingason, Einar Freyr
Sigurðsson, and Eiríkur Rögnvaldsson. 2011. Icelandic Parsed Historical Corpus (IcePaHC). University
of Iceland. version 9. • Richard Zimmermann. November 2015. A syntactic change with lots of data: The
rise of ‘do’-support with possessive ‘have’ in American english. Manchester LEL research seminar.
80
Plenarvorträge
Learning morphology
Freitag
10.03.2017
09:00 – 10:00
B4 1, 0.01
John Goldsmith
University of Chicago
[email protected]
PV
All linguists are familiar with the experience of explaining to non-linguists
that the goal of linguistics is to make explicit what native speakers know implicitly. We understand that this task (the task of turning the implicit into
something that is both formal and explicit) is anything but trivial. Tacit or implicit knowledge is wonderful and of great value (the knowledge of language
that native speakers have), but it is not easy to subject it to analysis—though
we linguists believe that there is a great reward that comes from the analysis.
The same thing can be said, I believe, about the linguist’s knowledge of how to
construct grammars, and the task of making this process explicit and open to
public scrutiny. My goal in this talk is to illustrate what can be learned from
the effort to develop algorithms for producing grammars from data.
I will begin with a brief introduction to word discovery based on Minimum
Description Length (MDL) analysis (Rissanen 1989, de Marcken 1996), and
show how the errors we observe there lead to the study of the automatic learning of morphological structure. The first steps of morphological analysis are
segmentation and classification, and these steps are followed by building a
morphological grammar (such as a phrase-structure grammar, for example).
Each step we take to build an algorithm to embody our analytic knowledge
as linguists teaches us in two ways: on the one hand, it reassures us that we do
succeed in important ways in analyzing even languages we have never seen,
but on the other, it makes us very aware of how difficult it is to analyze one
part of a language without assuming other parts of the language to have already been successfully analyzed. There are three very difficult questions that
emerge from this effort:
1. How do we evaluate how well a grammar models a particular set of data?
2. How do we evaluate and compare two different grammars that handle
the same data?
3. How is it possible for an algorithm to seek a better analysis than its current analysis?
81
Plenarvorträge
PV
That question has three subparts: What does it even mean for an algorithm
to propose something new? How can an algorithm “look at” part of its analysis and recognize that it should be dissatisfied with it? And how can it tell if a
change in its analysis is an improvement or not?
These questions relate very directly to the question of what kind of innate
knowledge linguists’ work supports. Is the knowledge that the learning algorithm (or the child) must be endowed with of the same sort as the knowledge
that she will be learning as she learns her grammar? If the answer is yes, then
this kind of innate knowledge fits Leibniz’s picture of innate knowledge as enthymeme, whereby the discovery of innate knowledge is just like discovering
implicit major premises in people’s arguments. If the answer is no (and the answer is no), then this kind of knowledge is akin to Kant’s notion of a category
(though it is not static in the way Kant’s is): the knowledge that is needed in
order to answer the 3 questions above all involve serious questions regarding
information and complexity, notions that are preconditions for any kind of
concrete knowledge about the world.
Much of the relevant work in this area has been done under the rubric of machine learning, most notably the subdomain referred to as unsupervised learning. Among the most important elements found there is the ability to quantify
the amount of information that is left unexplained in an analysis. This is important not because we wish to leave information unexplained, but because it
allows us to compare two different analyses of the same data, and determine
which one leaves less information unexplained.
These lofty ideas will be illustrated in concrete problems of learning the morphology of English, French, and Swahili. Review articles can be found in Goldsmith 2010 and Goldsmith, Lee, and Xanthos 2017. Chater et al (2015) presents
a different aspect of this concern.
References: • Rissanen, Jorma (1989): Stochastic Complexity in Statistical Analysis. World Scientific
Publishing. • de Marcken, Carl. (1996): Unsupervised Language Acquisition. PhD dissertation, MIT.
arXiv:cmp-lg/9611002. • Goldsmith, John (2010): Segmentation and morphology. The Handbook
of Computational Linguistics and Natural Language Processing, 364-393. Wiley-Blackwell. • Goldsmith, John, Jackson Lee, and Aris Xanthos. (2017): Computational approaches to morphology. Annual Review of Linguistics. • Chater, Nick, Alex Clark, John Goldsmith, and Amy Perfors (2015): Empiricism and Language Learnability. OUP.
82
Plenarvorträge
A neurocomputational model of surprisal in comprehension
Freitag
10.03.2017
10:00 – 11:00
B4 1, 0.01
Matthew Crocker
Universität des Saarlandes
[email protected]
PV
People continuously assign meaning to the linguistic signal on a more or
less word-by-word basis, building rich semantic representations of what has
been encountered so far, which in turn conditions expectations about what is
likely—or not—to follow. Surprisal Theory (Hale, 2001; Levy, 2008) has been
particularly successful in offering a broad account of word-by-word processing difficulty. The theory asserts that the processing difficulty incurred by a
word is inversely proportional to the expectancy (or surprisal) of a word, as
estimated by probabilistic language models. Such models are limited, however, in that they assume expectancy is determined by linguistic experience
alone, making it difficult to accommodate the influence of world and situational knowledge.
To address this issue, we1 present a neurocomputational model of language
processing that seamlessly integrates linguistic experience and probabilistic world knowledge in online comprehension. The model is a simple recurrent network (SRN) that is trained to map sentences onto rich probabilistic
meaning representations that are derived from a Distributed Situation-state
Space (DSS: Frank et al, 2003). Crucially, these meaning representations go
beyond literal propositional content and capture inferences driven by knowledge about the world (cf. a mental model (Johnson-Laird, 1983), or situation
model (Zwaan, 1998)). The model is trained to construct the DSS representation for a sentence on an incremental, word-by-word basis, which results in
the model being inherently sensitive to the frequency with which sentences
are mapped onto specific DSSs (cf. Mayberry et al, 2009). Furthermore, the incremental construction of DSS representations allows for the computation of
online surprisal based on the likelihood of the sentence meaning for the just
processed word, given the sentence meaning up until before the word was encountered.
We demonstrate that our online surprisal metric integrates both the likelihood of a situation model (DSS)—thereby reflecting world knowledge—as
well as linguistic experience. This ‘comprehension-centric’ characterisation
1
This abstract presents joint work with Harm Brouwer and Noortje Venhuizen.
83
Plenarvorträge
PV
of surprisal thus provides a more general index of the effort involved in mapping from the linguistic signal to rich and knowledge-driven situation models: Not only can this sentence-to-meaning mapping capture established surprisal phenomena reflecting linguistic experience, it also offers the potential
for surprisal-based explanations of a range of findings that have demonstrated
the importance of knowledge-, discourse-, and script-driven influences on
processing difficulty.
Moreover, we will outline how our model can be related to neurophysiological indices of processing difficulty. Previous work has linked word-induced
surprisal to the N400 component of the ERP signal (e.g., Frank et al. 2015). Our
meaning-based notion of surprisal, however, reflects how unexpected an update is for the unfolding interpretation. This suggests a close link to the P600,
which has been argued to index of the effort involved in updating the interpretation (Brouwer et al, in press). By supporting the straight-forward computation of meaning-based surprisal, as well as providing a direct linking hypothesis to ERP correlates, this work sheds much needed light on the neural
mechanisms that underlie the otherwise still abstract notion of surprisal.
References: • Brouwer, H., Crocker, M.W., Venhuizen, N., Hoeks, J. (in press): A Neurocomputational Model of the N400 and the P600 in Language Processing. Cognitive Science. • Frank, S., Koppen, M., Noordman, L., Vonk, W. (2003): Modeling knowledge-based inferences in story comprehension. Cognitive Science 27(6):875–910. • Frank, S., Otten, L., Galli, G., Vigliocco, G. (2015): The
ERP response to the amount of information conveyed by words in sentences. Brain and language
140:1–11. • Hale, J. (2001): A probabilistic Earley parser as a psycholinguistic model. In: Proc. of the
ACL 1–8. • Johnson-Laird, P.N. (1983): Mental Models. • Levy, R. (2008): Expectation-based syntactic comprehension. Cognition 106(3):1126–1177. • Mayberry, M., Crocker, M.W., Knoeferle, P. (2009).
Learning to Attend: A Connectionist Model of Situated Language Comprehension. Cognitive Science
33(1):449–496. • Zwaan, R., Radvansky, G. (1998): Situation models in language comprehension and
memory. Psych. Bul. 123:162–185.
84
e-book
978 90 272 6688 0 EUR 95.00
New Journals
/ 978 90 272 6688 0 USD 143.00
The Agenda Setting Journal
Theory, Practice, Critique
Editor: Salma I. Ghanem
DePaul University
The Agenda Setting Journal: Theory, Practice, Critique focuses on the theoretical
developments that continue in agenda setting and how the theory is applied to
areas outside of mass communication. The journal also represents the growth
and maturity of the communication field as it is also the first and only to-date
theory-based journal in the communication discipline.
ISSN 2452-0063 | e-ISSN 2452-0071
Language and Linguistics
Editor-in-Chief: Jo-wang Lin
Academia Sinica
This is an academic publication of the Institute of Linguistics at Academia
Sinica. Established in 2000, it publishes research in general and theoretical
linguistics on the languages of East Asia and the Pacific region, including
Sino-Tibetan, Austronesian, and the Austroasiatic and Altaic language families.
John Benjamins Publishing Company is the official publisher
as of Volume 18 (2017).
issn 1606-822X | e-issn 2309-5067
Language Ecology
General Editors: Umberto Ansaldo and Lisa Lim
University of Hong Kong
The ecology of language is a framework for the study of language as
conceptualised primarily in Einar Haugen’s 1971/72 work, where he defines
language ecology as “the study of interactions between any given language and
its environment”.
issn 2452-1949 | e-issn 2452-2147
Pragmatics
Quarterly Publication of the International
Pragmatics Association (IPrA)
Editor-in-Chief: Helmut Gruber
University of Vienna
Pragmatics is the peer-reviewed quarterly journal of the International Pragmatics
Association (IPrA), which was established in 1986 to represent the field of
linguistic pragmatics, broadly conceived as the interdisciplinary (cognitive,
social, cultural) science of language use.
John Benjamins Publishing Company is the official publisher
as of Volume 27 (2017).
ISSN 1018-2101 | E-ISSN 2406-4238
JOHN BENJAMINS PUBLISHING COMPANY
www.benjamins.com [email protected]
Arbeitsgruppen
AG1
Informationstheoretisch basierte Modellierung sprachlicher
Variation im Kontext
Elke Teich, Bernd Möbius & Vera Demberg
Universität des Saarlandes, Universität des Saarlandes & Universität des
Saarlandes
[email protected], [email protected], [email protected]
Raum: B3 2, 0.03
Short description
Sprache approximiert eine für die Kommunikation optimale Kodierung, indem sie Anpassungen an kontextuelle Bedingungen (Kontext, Situationskontext) ermöglicht. Wir fokussieren informationstheoretisch-basierte Ansätze zur Modellierung von unterschiedlichen sprachlichen Prozessen (menschliches Sprachverstehen, maschinelle Sprachverarbeitung, Sprachwandel
u. –evolution) auf allen linguistischen Ebenen (Phonetik, Lexik, Syntax, Semantik, Diskurs) unter besonderer Berücksichtigung der Beziehung von
sprachlicher Variation im Kontext.
89
AG 1 · Informationstheoretisch basierte Modellierung sprachlicher Variation im Kontext
Embeddings or visions of the future of the lexicon
Mittwoch
08.03.2017
14:15 – 15:15
B3 2, 0.03
Hinrich Schütze
LMU München
[email protected]
AG1
One way to model the predictability of a linguistic unit alpha in context gamma
is to represent alpha and gamma as vectors a and c in high-dimensional space
and estimate P(alpha|gamma) based on a and c. These types of vector representations are commonly referred to as embeddings. I will give a general talk
about embeddings – covering issues like learning, granularity and compositionality – and how they relate to the traditional lexicon in computational linguistics. At the end, I will turn to the question of what the implications are for
models of predictability.
90
AG 1 · Gebäude B3 2, Raum 0.03
An information-theoretic account on the diachronic development of
discourse connectors in scientific writing
Mittwoch
08.03.2017
15:15 – 15:45
B3 2, 0.03
Stefania Degaetano-Ortlieb, Merel Scholman, Stefan Fischer, Elke Teich,
Vera Demberg
(Universität des Saarlandes)
[email protected]
We investigate the diachronic development of discourse connectors (DC) in scientific writing, considering their linguistic encoding and looking at whether
there is a change with respect to shorter vs. longer encodings over time driven
by information-theoretic effects.
Our theoretical framework is based on register theory (cf. Quirk et al. 1985):
a register being characterized by distributions of lexico-grammatical features
according to field (discourse topic), tenor (relation between participants) and
mode of discourse (text-forming function). As DCs create cohesive ties between units in a discourse, they contribute to mode. Given psycholinguistic
evidence on the correlation between variation in linguistic encoding and information density (cf. Aylett and Turk 2004), we assume that in scientific writing,
features related to mode are more likely to be realized in less dense encodings
than field features (assumed to be realized in denser encodings (cf. Halliday
1988)) to balance information density in the discourse. Thus, we expect DCs
with a short linguistic encoding (e.g. but) to decrease over time, while those
with longer encodings will increase (e.g. on the other hand).
We analyze diachronic tendencies of DCs in scientific and general language
corpora from 1650 to present. Preliminary results indicate a decrease of short
DCs in scientific writing, while longer ones increase.
References: • Aylett, Matthew and Alice Turk (2004). The Smooth Signal Redundancy Hypothesis: A
Functional Explanation for Relationships between Redundancy, Prosodic Prominence and Duration
in Spontaneous Speech. Language and Speech 47 (1): 31–56. • Quirk, Randolph, Sidney Greenbaum,
Geoffrey Leech, and Jan Svartvik (1985). A Comprehensive Grammar of the English Language. London:
Longman. • Halliday, M. A. K. (1988). On the Language of Physical Science. In Registers of Written English: Situational Factors and Linguistic Features, edited by Ghadessy, Mohsen, 162–177. London: Pinter.
91
AG1
AG 1 · Informationstheoretisch basierte Modellierung sprachlicher Variation im Kontext
Does information density help? A Bayesian analysis of
help + (to) Vinf in varieties of English
Mittwoch
08.03.2017
16:30 – 17:00
B3 2, 0.03
Natalia Levshina
Leipzig University
[email protected]
This study deals with the constructional variants help + (to) Vinf, as exemplified by the sentences (1a) and (1b):
AG1
(1)
a.
b.
Mary helped John write the letter.
Mary helped John to write the letter.
In many previous accounts, the choice between the variants was explained semantically. For example, the use of the bare or marked infinitive is assumed to
be related to the degree of the subject’s involvement in the event represented
by the infinitive (Dixon 1991). In addition, such factors as cognitive complexity, avoidance of identity (horror aequi) and the inflectional form of help, have
been shown to constrain the choice between the bare or marked infinitive
(Lohrmann 2011). In this study, I test an alternative explanation based on the
principles of economy (e.g. Haiman 1983) and Uniform Information Density
(UID) (Jaeger 2010). The higher the predictability of the infinitive in a given
context, the higher are the chances of the bare infinitive being used.
The analyses are based on Davies’ (2013) corpus of Global Web-based
English, which represents twenty countries where English is spoken. I fit
Bayesian mixed-effects binomial regression models, with the type of the infinitive as the response variable, the above-mentioned contextual variables as
fixed effects and the infinitives and text IDs as random intercepts.
References: • Davies, M. (2013): Corpus of Global Web-Based English: 1.9 billion words from speakers
in 20 countries. • Dixon, R.M.W. (1991): A New Approach to English Grammar, on Semantic Principles. Oxford: Clarendon Press. • Jaeger, T.F. (2010): Redundancy and reduction: speakers manage
syntactic information density. Cognitive Psychology 61: 23–62. • Haiman, J. (1983): Iconic and economic motivation. Language 59(4): 781–819. • Lohmann, A. (2011): Help vs. help to – a multifactorial, mixed-effects account of infinitive marker omission. English Language and Linguistics 15(3):
499–521.
92
AG 1 · Gebäude B3 2, Raum 0.03
Modeling segmental durations using the Naive Discriminative
Learner
Fabian Tomaschek
Universität Tübingen
Benjamin V. Tucker
University of Alberta
[email protected]
[email protected]
Mittwoch
08.03.2017
17:00 – 17:30
B3 2, 0.03
It is well known that acoustic durations of words and segments are codetermined by lexical properties such as frequency of occurrence (Zipf 1935,
Bell et al. 2003, Gahl, 2008; Cohen Priva, 2015). The current study explores the
use of a predictor measuring learning for modeling segment durations in German: weights and activations from the Naive Discriminative Learner Model
(NDL: Baayen et al., 2011). Given a set of input units (diphones and triphones),
the network estimates their connection or association strengths to a set of output units (word forms), providing a measure of how well word forms are associated with their phonetic markup.
The project investigates how the goodness of fit of the learning measures
provided by NDL varies with the complexity of the cue-to-outcome structure.
Also, we intend to investigate how multi-collinearities present in the data affect the results and how to cope with them. Finally, we will compare the goodness of fit of learning measures with traditional predictors such as frequencies
and transitional probabilities.
References: • Baayen, R. Harald et al. (2011): An amorphous model for morphological processing in
visual comprehension based on naive discriminative learning. Psychological review, 118 (3). p.438–481
• Gahl, S. (2008): ”Thyme” and ”Time” are not homophones. Word durations in spontaneous speech.
Language, 84 (3): p. 474–496. • Cohen Priva, U. (2015): Informativity affects consonant duration and
deletion rates. Laboratory Phonology, 6 (2). p. 243–278. • Zipf, G. (1935): The psycho-biology of language.
An introduction to dynamic philology. Cambridge: MIT Press.
93
AG1
AG 1 · Informationstheoretisch basierte Modellierung sprachlicher Variation im Kontext
Mittwoch
08.03.2017
17:30 – 18:00
B3 2, 0.03
AG1
High-variability distributions increase perceptual uncertainty during
acoustic cue acquisition
Jessie S. Nixon
University of Potsdam
Catherine T. Best
Western Sydney University
[email protected]
[email protected]
Much statistical learning research has focused on effects of the number of
peaks in input distributions. Recent research has examined how statistical
variance in acoustic cue distributions affects perceptual uncertainty in native
speech perception (Clayards, Tanenhaus, Aslin & Jacobs, 2008; Nixon, van Rij,
Mok, Baayen & Chen, 2016). The present study investigates how statistical variance affects acquisition of a non-native cue, namely tone.
In a visual world eyetracking experiment, participants saw four pictures of
common objects and heard a Cantonese word. Auditory stimuli consisted of
a 12-step continuum of increasing pitch, presented in a bimodal distribution.
Distribution peaks corresponded to prototypical high and mid tones, respectively. Statistical variance (the width of the distribution) varied between participants: high- vs. low-variance.
Eye movements (Euclidean distance from target) were analysed with
generalised additive mixed models (GAMMs). Models showed a significant
condition-by-pitch nonlinear interaction over time (p < 0.001); and a significant condition-by-pitch nonlinear interaction over the experiment (i.e. trial;
p < 0.001). Over the course of the experiment, and over the trial, fixations became closer to the target in the low-variance condition.
With more distinctive pitch cues in the low-variance condition, participants
were better able to distinguish between tones. In early trials, distance from target was actually larger in the low-variance condition, especially for the high
tone. However, by the second half of the experiment, participants were already
closer to the target than those in the high-variance condition. These results
provide evidence that in implicit statistical learning of new acoustic dimensions, within-category acoustic variation hinders acquisition; learning is enhanced by low within-category variability.
References: • Clayards, M. et al. (2008): Perception of speech reflects optimal use of probabilistic
speech cues. Cognition. 108(3), 804–809. • Nixon, J.S., van Rij, J., Mok, P., Baayen, R.H. and Chen, Y.
(2016): The temporal dynamics of perceptual uncertainty: eye movement evidence from Cantonese
segment and tone perception. Journal of Memory and Language. 90, 103–125.
94
AG 1 · Gebäude B3 2, Raum 0.03
Expectation management in interaction: Discourse particles signal
surprisal of upcoming referents
Donnerstag
09.03.2017
09:00 – 09:30
B3 2, 0.03
Geertje van Bergen
Max Planck Institute for Psycholinguistics
[email protected]
This study focuses on expectation management in conversation via highlevel discourse cues, namely the Dutch common-ground managing discourse
particles (DPs) eigenlijk (‘actually’,‘in fact’) and inderdaad (‘indeed’). In a visual world eye-tracking experiment, I investigate whether and when listeners integrate the complex interpersonal information encoded in expectationmanaging DPs during incremental language understanding. It is hypothesized
that inderdaad signals upcoming low surprisal, whereas eigenlijk signals upcoming high surprisal. As such, DPs are expected to modulate the listener’s
predictions about upcoming referents.
Dutch participants were exposed to short conversations, consisting of a constraining context followed by a question and an answer in which a critical word
was replaced by a beep. They were instructed to click on the picture that best
fit the answer. Displays included a Target (low surprisal given the context),
a Competitor (high surprisal but related to the context) and two Distractors
(unrelated to the context); answers contained a discourse particle (inderdaad
vs. eigenlijk) or a control word.
Targets were clicked more often and faster in the inderdaad-condition
compared to the control condition; in the eigenlijk-condition, Target clicks
were less frequent and responses were slower. Gaze patterns show that immediately upon encountering the discourse particle, Target fixations significantly increased in the inderdaad-condition and decreased in the eigenlijk-condition compared to the control condition. Findings suggest that
expectation-managing DPs can modulate context-based linguistic expectations during incremental language understanding, and hence improve communicative efficiency. From a methodological perspective, DPs provide a useful means to further investigate the nature of discourse-based expectations.
95
AG1
AG 1 · Informationstheoretisch basierte Modellierung sprachlicher Variation im Kontext
Donnerstag
09.03.2017
09:30 – 10:00
B3 2, 0.03
Does UID affect rate of pronominalization?
Ekaterina Kravtchenko1 , Ashutosh Modi1 , Vera Demberg1 , Ivan Titov2 ,
Manfred Pinkal1
(1 Universität des Saarlandes, 2 University of Amsterdam)
{eskrav|ashutosh|vera|pinkal}@coli.uni-sb.de; [email protected]
AG1
A large body of experimental evidence supports the Uniform Information Density hypothesis (UID), which shows that high predictability of linguistic units
leads to use of shorter variants, while low predictability is correlated with
longer variants (Jaeger & Buz, 2016). However, the range of linguistic phenomena that UID extends to remains unclear. UID predicts that pronouns should
be used for more predictable referents, while longer expressions (names or
complex NPs) should be used for those less predictable.
Existing experimental evidence is mixed. Some studies report an effect of
referent predictability on referring expression (RE) type: Tily & Piantadosi
(2009) and Kravtchenko (2014), both corpus studies, found that shorter REs
are favored over longer ones in more predictable contexts. Rohde & Kehler
(2014) manipulated RE predictability in sequences like ”Peter admired / impressed Mary. __”, by choosing subject- vs. object-biased verbs, and seeing
how discourse participants were referred to in prompted continuations. They
found no effect of predictability on RE choice.
Here we report two new studies addressing this question. Our first study
attempts to replicate Tily & Piantadosi (2009), but with a significantly larger
corpus from a more colloquial data register. We find no significant effect of
predictability on RE type, after accounting for structural biases.
Our second study attempts to replicate Rohde & Kehler (2014), but looks at a
broader range of REs. In our sequences we alternated names with long NPs, as
people may be more motivated to reduce the latter. We successfully replicate
Rohde & Kehler, but find no effect of predictability with long NPs.
These experiments suggest that UID may not affect RE choice, and may extend to limited phenomena beyond the level of surface form predictability.
References: • Jaeger, T. F. & E. Buz (2016): Signal reduction and linguistic encoding. Handbook of
psycholinguistics. • Kravtchenko, E. (2014): Predictability and syntactic production: Evidence from
subject omission in Russian. Proceedings of CogSci 2014, 785–790. • Rohde, H. & A. Kehler (2014): Grammatical and Information-Structural Influences on Pronoun Production. Language, Cognition, and Neuroscience 29(8), 912–927. • Tily, H. & S. T. Piantadosi (2009): Refer efficiently: Use less informative
expressions for more predictable meanings. In PRE-CogSci 2009.
96
AG 1 · Gebäude B3 2, Raum 0.03
Predicting processing cost of anaphora resolution
Olga Seminck
Université Paris Diderot
Pascal Amsili
Université Paris Diderot
[email protected]
[email protected]
Donnerstag
09.03.2017
10:00 – 10:30
B3 2, 0.03
Anaphora resolution, for human speakers, can be more or less costly depending on various factors like ambiguity, syntactic complexity and semantic plausibility. The variation of cost has been measured by many studies in
psycholinguistics, through experimental paradigms like self-paced reading,
or eye-tracking. Our project aims at devising a system, inspired by current
NLP coreference resolution systems, that can predict a processing cost for
anaphora resolution, which can be evaluated by running our system on human
data coming from psycholinguistic experiments, or eye-tracking corpora e.g.
the Dundee Corpus (Kennedy 2003). Inspired by surprisal theory (Hale 2001)
and the entropy reduction hypothesis (Hale 2006), we propose a continuous,
incremental measure that assigns processing cost to anaphora. Our measure
reflects how certain a probabilistic anaphora resolution system is about its decisions. To do so, with a simple anaphora resolution tool, we compute a probability distribution over all antecedent candidates of an anaphor and calculate
entropy over it. We hypothesize that the entropy over this distribution can be
seen as the processing cost of the resolution of the anaphor. So the smaller the
entropy, the less processing cost that is predicted. A first study we conducted
on two biases that were discovered by psycholinguists (Subject Assignment
Strategy and Parallel Function Hypothesis (e.g. Crawley 1990)) showed that
our model was able to simulate human performance in these matters: it assigned the pronouns in a way comparable to human participants and the cost
it predicted corresponded to reading times recorded in self-paced reading experiments.
References: • Rosalind A Crawley et al. (1990). The use of heuristic strategies in the interpretation
of pronouns. In: Journal of Psycholinguistic Research 19.4, 245–264. • John Hale (2001). A probabilistic
Earley parser as a psycholinguistic model. In: Proceedings of the second meeting of the North American
Chapter of the Association for Computational Linguistics on Language technologies. Association for Computational Linguistics, 1–8. • John Hale (2006). Uncertainty about the rest of the sentence. In: Cognitive Science 30.4, 643–672. • Alan Kennedy et al. (2003). The dundee corpus. In: Proceedings of the 12th
European conference on eye movement.
97
AG1
AG 1 · Informationstheoretisch basierte Modellierung sprachlicher Variation im Kontext
Uniform Information Density models for language production:
A comparative study of Hindi and English
Donnerstag
09.03.2017
11:15 – 11:45
B3 2, 0.03
Ayush Jain, Vishal Singh, Sumeet Agarwal, Rajakrishnan Rajkumar
(Indian Institute of Technology Delhi)
{ee1120439, ee1120493, sumeet, raja}@iitd.ac.in
AG1
In this paper, the extent to which language production is governed by the Uniform Information Density (UID) hypothesis (Jaeger 2010) is analysed for Hindi
and English. The hypothesis states that a speaker tries distributing the information across a sentence in the most uniform fashion possible so as to communicate efficiently, analogously to communication in a noisy channel (Shannon
1948). We examine the effect of word-order change on the information distribution of a sentence and its effect on production choice.
Several quantitative UID models were defined, based on the variation of
lexical N-gram scores within a sentence. As alternatives, the UID measures
were also estimated using syntactic data (part-of-speech tags) instead of lexical data, and only at chunk boundaries rather than at all words. In order to
assess the validity of the UID hypothesis for language production, the information orthogonal to that captured by a basic N-gram score was estimated by
performing a binary classification (reference corpus sentences vs. variants)
and comparing the classification accuracy in the presence and absence of UID
scores.
While using lexical UID measures, segregation of corpus and non-corpus
sentences was poor for both English and Hindi. However for syntactic UID
measures, the segregation was much better for English, though for Hindi it
was still poor. On addition of normalized UID measures to a baseline feature
set consisting of N-gram scores, the classification accuracy for the Hindi sentences increased slightly, suggesting that perhaps the notion of UID might in
some cases be informative about production choices even for Hindi. The difference in the results obtained for Hindi and English is possibly because of greater
flexibility of word order in Hindi.
References: • T. Florian Jaeger. Redundancy and reduction: Speakers manage syntactic information
density. Cognitive Psychology, 61(1):23–62, 2010. • C. E. Shannon. A mathematical theory of communication. The Bell System Technical Journal, 27(3):379–423, July 1948. • C. E. Shannon. A mathematical
theory of communication. The Bell System Technical Journal, 27(4):623–656, Oct 1948.
98
AG 1 · Gebäude B3 2, Raum 0.03
Over-specifications efficiently manage referential entropy in
situated communication
Elli N. Tourtouri
Universität des
Saarlandes
Francesca Delogu
Universität des
Saarlandes
Matthew Crocker
Universität des
Saarlandes
[email protected]
[email protected]
[email protected]
Donnerstag
09.03.2017
11:45 – 12:15
B3 2, 0.03
When referring to objects in their visual environment, speakers often use
more information than required for referent identification, thereby violating Grice’s Quantity maxim [1]. Previous research provided contradicting evidence suggesting either that such over-specifications (OS) hinder referential
processing [2, 3], or that their processing does not differ from that of minimal
specifications (MS) [4]. We provide evidence that OS in fact aids listeners in
their effort to identify the visual target, especially when the number of referential candidates is increased.
In an eye-tracking experiment, participants were presented with arrays of
objects and heard OS or MS descriptions of a target, reducing referential entropy more or less gradually. We hypothesised that OS would result in faster
and easier target identification than MS, especially when referential candidates are eliminated more gradually. We find that, when referential entropy
is reduced less gradually, OS results in equally fast and easy target identification as MS, while OS facilitates reference resolution when more visual search
is required.
From these results we conclude, firstly, that situated comprehension is sensitive to the distributional properties of the visual environment, as well as the
encoding of the referring expression. Secondly, we argue that the prime determinant of processing effort in situated communication is the efficient modulation of visual search complexity through the gradual reduction of referential entropy across the utterance, even when this results in violations of the
gricean maxim.
References: • [1] Grice (1975). In Cole & Morgan (Eds.), 41–58. • [2] Engelhardt et al. (2011). Brain
Cogn, 77, 304–331. • [3] Davies & Katsos (2013). J pragmat, 49, 78–106. • [4] Arts et al. (2011). J pragmat,
43, 361–374.
99
AG1
AG 1 · Informationstheoretisch basierte Modellierung sprachlicher Variation im Kontext
Donnerstag
09.03.2017
12:15 – 12:45
B3 2, 0.03
Mathematical barriers to channel coding-based explanations of
variation in language production
Eric Meinhardt
University of California, San Diego
[email protected]
AG1
In the pursuit of computational-level analyses of language production, three
influential theories (Genzel & Charniak, 2002; Aylett & Turk, 2004; Jaeger,
2010) have offered information theoretic explanations of variation. The common eponymous idea of all three theories is roughly that, from timestep to
timestep the amount of information transmitted in natural language signal
sequences should be approximately constant, ceteris paribus. They claim this
property of language is implied by a rationality assumption about human communication and the noisy channel coding theorem (‘NCCT’) of Shannon (1948).
I demonstrate that this property is not, in fact, a consequence of the NCCT
through two lines of reasoning.
First, these theories have not appreciated ways in which natural language
represents a channel with significantly different mathematical properties
from Shannon’s, or that these differences require much more sophisticated
and possibly novel proofs. Second, Shannon’s results and the component mathematical objects do not (presently) have clear relevance to incremental online
choices in a language-like encoding. For example, entropy rates are asymptotic
in the length of the signal sequence, whereas natural language features relatively short signals. As well, in Shannon’s channel, there are no choice points
for senders – each message maps to one signal. Finally, channel capacity is a
maximum over all possible distributions on the source. This is of clear interest to
an engineer without knowledge of the future source distribution, but it is not
immediately obvious why or how which alternative distributions over word or
sound sequences are a relevant part of an upper bound on the performance of
a speaker in a particular situation.
References: • Aylett, M. & Turk, A. (2004). The smooth signal redundancy hypothesis. Language and
Speech, 47 (Pt 1), 31-56. • Genzel, D. & Charniak, E. (2002). Entropy Rate Constancy in Text. Proceedings of the 40th Annual Meeting of the ACL, 199–206. • Jaeger, T.F. (2010). Redundancy and reduction:
Speakers manage syntactic information density. Cognitive Psychology, 61 (1), 23–62. • Shannon, C.
(1948). A Mathematical Theory of Communication. Bell System Technical Journal, 27 (3), 379–423.
100
AG 1 · Gebäude B3 2, Raum 0.03
Entropy in language comprehension
Donnerstag
09.03.2017
13:45 – 14:45
B3 2, 0.03
Tal Linzen
Ecole Normale Supé rieure, Paris
[email protected]
Predictability effects are pervasive in language comprehension. These effects
support the view that language users maintain probability distributions over
potential representations and use those distributions to allocate more resources to process likely upcoming representations. While it is fairly established that language comprehension is modulated by the probability of the incoming linguistic unit, it is unclear if it is also affected by the way in which
probability mass is distributed among potential outcomes.
This talk will investigate whether this question can be addressed using the
entropy of the distribution over representations. I will explore implications of
existing proposals that have largely been overlooked, and suggest new linking hypotheses that use entropy to predict psycholinguistic measurements.
These entropy-based linking hypotheses will be used to analyse empirical data
drawn from two domains: reading times in sentence comprehension and electromagnetic recordings during spoken word recognition. I will show that entropy can be profitably used to predict responses in both areas, though not
necessarily using the same linking hypotheses; surprisal cannot be subsumed
under entropy in either case. In summary, entropy can be a useful component
of linking hypotheses in studying language comprehension, but only when its
properties and potential limitations are taken into consideration.
101
AG1
AG 1 · Informationstheoretisch basierte Modellierung sprachlicher Variation im Kontext
Locative have and context modulation
Freitag
10.03.2017
11:30 – 12:00
B3 2, 0.03
Maria M. Piñango
Yale University
Ashwini Deo
Ohio State University
Muye Zhang
Yale University
[email protected]
[email protected]
[email protected]
It’s been claimed that in the absence of a preposisional phrase, have sentences
are said to be either ungrammatical ((1a) vs. (1b)) or receive a non-locative interpretation ((1c) vs. (1d)).
AG1
(1)
a. *The tree has a nest.
b. [The tree]i has a nest in iti . (locative)
c. The mall has a children’s park (only non-loc. reading possible)
d. [The mall]i has a children’s park near iti . (locative)
We test the hypothesis that have lexically captures lexico-conceptual space
that ranges from pure incidental proximity to inalienable possession. Accordingly, have sentences are semantically compatible with locative and nonlocative (possessive) interpretations. The alleged unacceptability of sentences
like (1a) arises not from grammatical constraints but from (a) the competing
salience of the possessive meaning of have, and (b) the absence of supportive
context for the possible (though not frequent) locative interpretation. If this
is correct then we expect speaker responses to PP-less locative have sentences
to be when supported by (local) linguistic locative context. An acceptability
study (N=100) bears this prediction out: when presented with PP-less have
sentences like (1a) with a prior location-invoking context, subjects rate them
significantly higher (within the acceptable range) than when presented with
the same sentence in isolation or in possession-supporting contexts. We interpret the upward shift in acceptability ratings in the presence of the right contextual conditions to signal an increase in certainty (decrease in surprisal) on
part of the hearer. Thus, in the semantic domain, reduction in surprisal can be
understood as the increase in the speaker’s confidence (measurable in acceptability ratings) that communication is succeeding.
102
AG 1 · Gebäude B3 2, Raum 0.03
Entropy predicts uncertainty in subcategorization frame
distributions
Aaron Steven White
Johns Hopkins University
Kyle Rawlins
Johns Hopkins University
[email protected]
[email protected]
Freitag
10.03.2017
12:00 – 12:30
B3 2, 0.03
Syntactic bootstrapping approaches to verb-learning posit that learners use
a verb’s subcategorization frame (SCF) distribution to learn its meaning (Landau & Gleitman 1985). Most computational models of syntactic boot- strapping
assume that this is carried out via some form of verb clustering based on byverb SCF distribution entropy (Alishahi & Stevenson 2008, White 2015). Recent work in online sentence processing raises a potential problem for this
view: while comprehenders are sensitive to surprisal (Hale 2001) and entropy
reduction (Hale 2006) in online processing, they are not sensitive to a verb’s
SCF distribution entropy (Linzen & Jaeger 2015).
We show that, while SCF distribution entropy is not deployed online, it is
nonetheless encoded, yielding traces in the variability found across participants’ responses on offline measures. To establish this, we derive a measure
of SCF uncertainty using White & Rawlins’s (2016) large-scale acceptability
judgment dataset and a measure of SCF entropy using Korhonen et al.’s (2006)
VALEX dataset for 1000 different verbs. We show that, controlling for frequency, there is a reliable positive relationship between uncertainty and entropy, suggesting that computational models of syntactic bootstrapping are licensed in assuming access to SCF entropy.
References: • Alishahi, A., and S. Stevenson. 2008. A computational model of early argument structure acquisition. Cognitive Science 32:789–834. • Hale, J. 2001. A Probabilistic Earley Parser As a Psycholinguistic Model. In Proceedings of the Second Meeting of the North American Chapter of the Association
for Computational Linguistics on Language Technologies. • Hale, J. 2006. Uncertainty about the rest of
the sentence. Cognitive Science 30:643–672. • Korhonen, A., Y. Krymolowski, and T. Briscoe. 2006. A
large subcategorization lexicon for natural language processing applications. In Proceedings of LREC.
• Landau, B., and L.R. Gleitman. 1985. Language and experience: Evidence from the blind child. Harvard
University Press. • Linzen, T., and T.F. Jaeger. 2016. Uncertainty and expectation in sentence processing: evidence from subcategorization distributions. Cognitive science 40(6): 1382–1411. • White,
A.S. 2015. Information and incrementality in syntactic bootstrapping. Doctoral Dissertation, University
of Maryland. • White, A.S., and K. Rawlins. 2016. A computational model of S-selection. In Semantics
and Linguistic Theory 26.
103
AG1
AG 1 · Informationstheoretisch basierte Modellierung sprachlicher Variation im Kontext
Information density constrains article omission
Freitag
10.03.2017
12:30 – 13:00
B3 2, 0.03
AG1
Eva Horch
Universität des
Saarlandes
Robin Lemke
Universität des
Saarlandes
Ingo Reich
Universität des
Saarlandes
[email protected]
[email protected]
[email protected]
Article omission (AO) can be observed in specific text types, e.g in newspaper
headlines (1a), while in standard (written) German the article before the NP
Elfmeterschießen needs to be realized. AO is not obligatory in headlines though,
so what drives the choice for or against AO?
(1)
Portugal nach ∅ Elfmeterschießen im Halbfinale
‘Portugal after ∅ penalty shoot-out in semi-finals’
We pursue the hypothesis that, AO is guided by Uniform Information Density
(UID, Jaeger, 2010) where it is allowed by grammar (Reich, to appear). UID implements a preference for distributing information uniformly across the utterance: As articles lower the information on the head noun, UID predicts a
stronger preference for AO the more predictable the noun is.
(2)
Franziskus unterstützt (das/∅) (Projekt/Klage) gegen Kinderarbeit
‘Franziskus supports (the/∅) (project/action) against child.labor’
We investigated the effect of surprisal on AO with an acceptability rating study
on postverbal nouns in constructed headlines as (2). Items appeared either
with a predictable (Projekt, S = 4.27) or an unpredictable noun (Klage, S = 11.12)
preceded by an article in 50% of the trials. Noun surprisal was computed from
verb-noun pairs extracted from DeReKo (Kupietz & Keibel, 2009). A significant interaction between Surprisal and AO (z = 2.9, p < .01) indicates that
AO is preferred more strongly the less predictable the noun is. This confirms
our hypothesis and is in line with previous research by Jaeger (2010) on complementizer deletion.
References: • Jaeger, Florian (2010): Redundancy and reduction: Speakers manage syntactic information
density, Cognitive Psychology 61, 23–62. • Reich, Ingo (to appear): On the omission of articles and
copulae in German newspaper headlines. Linguistic Variation 17(2).
104
AG 1 · Gebäude B3 2, Raum 0.03
German morphosyntactic change is consistent with an optimal
encoding hypothesis
David M. Howcroft
Universität des
Saarlandes
Freitag
10.03.2017
13:00 – 13:30
B3 2, 0.03
Cynthia A. Johnson
Universiteit Gent
Rory Turnbull
École Normale Supé rieure
[email protected]
[email protected]
[email protected]
The assumption that language approximates an optimal code for human communication has implications for diachronic linguistics, predicting that the disruptions due to language change result in pressure to re-optimize the code.
We examine these implications in the context of morphosyntactic change between Middle High German (MHG) and New High German (NHG). The adjectival paradigm of MHG was bipartite (strong, weak) but developed into a tripartite system in NHG (strong, weak, mixed). The choice of adjective form depends on the morphosyntactic context; e.g., the weak form is used with the definite article, but after an indefinite article the strong form is used in MHG and
the mixed form in NHG. Introducing a new paradigm looks like an increase in
systemic complexity, but the novel distinctions re-optimize the code such that
overall complexity need not increase in successive stages of the language.
We examine the relative entropy of the adjectival systems in MHG and NHG,
using the predictability of adjective forms given their preceding context and
the predictability of the context forms given the adjective form. These measures allow us to examine the balance of information within the noun phrase
before and after the shift to the tripartite system, showing how the distribution of information across the words of the noun phrase changes while situating our findings in terms of the optimality of the resulting code.
Our results provide information theoretic support for claiming that strong
forms ‘compensate’ for the lack of information from articles (synchronically;
cf. Durrell 2002) and, furthermore, suggest that the morphosyntactic changes
observed between MHG and NHG, far from being a process of increasing complexity, are consistent with an optimal encoding hypothesis.
References: • Durrell, M. (2002): Hammer’s German Grammar and Usage. London: McGraw-Hill.
105
AG 2
Information structuring in discourse
Anke Holler, Katja Suckow, Barbara Hemforth & Israel de la Fuente
Georg-August-Universität Göttingen, Georg-August-Universität Göttingen,
Université Paris Diderot / CNRS, University of Lille
[email protected], [email protected],
[email protected], [email protected]
Raum: B3 1, 0.14
Short description
Although the need to model the relation between linguistic features of utterances and discourse structure is commonly acknowledged, there is still much
debate about what ought to be the appropriate level of analysis of discourse
segmentation and what the criteria to identify units of discourse structure are.
This working session investigates the interplay of linguistic cues and discourse
segmentation empirically and theoretically; taking historic, cross-linguistic,
experimental and computational perspectives into account.
106
AG 2 · Gebäude B3 1, Raum 0.14
Principles of information-structure and discourse analysis
Lisa Brunetti
Université
Paris-Diderot/CNRS
[email protected]
Kordula De Kuthy
Universität Tübingen
Arndt Riester
Universität Stuttgart
[email protected]
[email protected]
Mittwoch
08.03.2017
13:45 – 14:15
B3 1, 0.14
stuttgart.de
paris-diderot.fr
We present a pragmatic, i.e. meaning-based, method for the informationstructural analysis of corpus data, which is built on the idea that for any assertion contained in a text there is an implicit Question under Discussion (QUD)
that determines which parts of the assertion are focused or backgrounded.
The discussed method is largely in accordance with Alternative Semantics (e.g.
Rooth 1992). At the same time, the method makes strong predictions concerning the structure of discourse. In our discourse trees – which structurally differ in a systematic way from analyses of SDRT or RST – the terminal nodes of
a tree represent (A)ssertions and non-terminal nodes represent (Q)uestions.
An annotated corpus example is given in (1) (for space reasons, the tree is represented by > symbols).
(1)
(Snowden interview, ARD TV, Jan 2014)
A14 : When you are on the inside and you go into work everyday and you the
power you have.
Q15 : {What power do [employees of the NSA] have?}
> Q15.1 : {Whom can you wire tap?}
> > A15.1 : [YouT can wire tap [the President of the United States]F ]∼,
> > A15.1′′ : [youT can wire tap [a Federal Judge]F ]∼
QUD identification, which has often been waived as arbitrary or circular, is shown to be a highly constrained process. Principles (derived from
Schwarzschild 1999 and Büring 2008) involve, among others:
(i) Q-A-Congruence (QUDs must be answerable by their respective answers in the text).
(ii) Q-Givenness (implicit QUDs may only consist of given material).
(iii) Maximize-Q-Anaphoricity (QUDs should contain as much given material as possible).
Further constraints govern the treatment of simple parallelisms (contrastive
focus) and complex parallelisms (contrastive topic + focus; Büring 2003).
References: • D. Büring (2003). On D-trees, beans, and B-accents. Linguistics & Philosophy 26(5)
• D. Büring (2008). What’s New (and what’s Given) in the theory of focus? In Berkeley Linguistics
107
AG2
AG 2 · Information structuring in discourse
Society • M. Rooth (1992). A Theory of focus interpretation. Natural Language Semantics 1(1) • R.
Schwarzschild (1999). GIVENness, AvoidF, and other constraints on the placement of accent. Natural
Language Semantics 7(2).
Mittwoch
08.03.2017
14:15 – 14:45
B3 1, 0.14
Complete and independent? Reconsidering discourse segmentation
basics
Jet Hoek
Utrecht University
[email protected]
Jacqueline EversVermeul
Utrecht University
Ted J.M. Sanders
Utrecht University
[email protected]
[email protected]
AG2
This paper theoretically approaches discourse segmentation and focuses on
two issues concerning segmentation that were proposed by Mann and Thompson (1988) in their introduction of RST, but that have been implemented in
many other discourse annotation approaches as well: 1) the treatment of segmentation and annotation as a two-step process, which prevents the circularity of a process in which annotation and segmentation are intertwined
(Taboada & Mann 2006), and 2) the completeness constraint, which poses that
the segmentation of a text has to include all elements of that text. Taking the
clause as the syntactic basis for the identification of discourse segments, we
discuss English fragments, mainly from the Europarl corpus (Koehn 2005),
that present segmentation difficulties. We propose that accurate segmentation
is at least in part dependent on the propositional content of text fragments, and
that completely separating segmentation and annotation can be at the expense
of the quality of the segmentation. For fragments with embedded clauses, e.g.,
clausal complements, multiple segmentation options should be considered.
Using the interpretation of a text fragment can help to distinguish between
distinct syntactic constructions that have identical surface structures.
We argue in favor of amending Mann and Thompson’s (1988) completeness
constraint to pertain only to the propositional content of a discourse. Stance
markers, which are not part of the propositional content of the text (Biber
& Finegan 1989), may for instance be left out. Determining whether a stance
marker should be included in a text segment can be done by considering the
interpretation of the text.
References: • Biber, D. & Finegan, E. (1989). Styles of stance in English: Lexical and grammatical
marking of evidentiality and affect.Text 9(1),93-124. • Koehn, P. (2005). Europarl: A parallel cor-
108
AG 2 · Gebäude B3 1, Raum 0.14
pus for statistical machine translation. MT Summit X. • Mann, W.C. & Thompson, S.A. (1988). RST:
Toward a functional theory of text organization. Text 8(3), 243-281. • Taboada, M. & Mann, W.C.,
(2006). RST: Looking back and moving ahead. Discourse Studies 8(3), 423-459.
Goals, coherence and information structure in dialogue
Mittwoch
08.03.2017
14:45 – 15:45
B3 1, 0.14
Nicholas Asher
Université Paul Sabatier Toulouse
[email protected]
While information structure is often conceived as a sentence local phenomenon, in my talk I am going to take a look at some more global influences
on information structuring in dialogue. I will review how theories of discourse
structure approach this topic and then talk about how speakers’ ideas about
how they want the conversation to go might contribute to particular information structuring devices. The use of presupposition and appositions are one
such device, but there are others.
The semantics of provisional, temporal anaphors and cataphors
Daniel Altshuler
Hampshire College
Dag Trygve Truslew Haug
University of Oslo
[email protected]
[email protected]
Intro: A holistic theory of discourse interpretation should provide: (i) a way
of segmenting discourse units (DUs) and (ii) relating them, as well as (iii) a
semantics for both the linguistic expressions making up the DUs and their relations. Ever since Hobbs (1979), anaphora resolution has been a key guide for
what (i)-(iii) should be like. For example, SDRT addresses many of the challenges imposed by (i)-(iii) by modeling anaphoric dependencies. The goal of
this talk is to investigate an outstanding issue noted by Asher and Lascarides
(2003): given a discourse context C and two DUs π1 , π2 to be related by a relation R, it may be that C makes R(π1 , π2 ) the most plausible inference, but
an extension of C may make it more plausible that a distinct R′ (π1 , π2 ) is pre-
109
AG2
Mittwoch
08.03.2017
16:30 – 17:00
B3 1, 0.14
AG 2 · Information structuring in discourse
ferred. That is, inferring relations between DUs is non-monotonic. And since
these relations entail temporal constraints, anaphoric connections between
eventualities often undergo revision. This is especially apparent in the French
novella, Sylvie, where, famously, the reader chooses a resolution strategy that
she later finds to be wrong. Modeling this revision is both complicated and
important. The complication can be summed up as follows: What semantics
should we posit for temporal anaphors and cataphors found in Sylvie such that
we could model their resolution as being provisional? This question is important because it is at the heart of (iii). To the best of our knowledge, however, it
has not been addressed. The goal of the talk is to provide an answer. We adopt
SDRT for analyzing the discourse structure of the key passages in Sylvie and
extend Haug’s (2014) PCDRT to deal with the temporal anaphors and cataphors.
While our analysis is a long shot from what the systematic analysis of texts like
Sylvie requires, we nevertheless motivate a future project to merge SDRT and
PCDRT, which would enrich our understanding of how (iii) relates to (i) and
(ii).
AG2
References: • Asher, N. & Lascarides, A. (2003): Logics of conversation • Hobbs, J. (1979), Coherence and coreference. Cognitive Science 3, 67–90. • Haug, D. (2014), Partial dynamic semantics for
anaphora: Compositionality with syntactic coindexation. Journal of Semantics 31, 457–511.
Modal particles and their influence on discourse structure
Mittwoch
08.03.2017
17:00 – 17:30
B3 1, 0.14
Sophia Döring
Humboldt-Universität zu Berlin
[email protected]
In the present work, the interaction of modal particles (MPs) and discourse
structure (as it is modelled in Rhetorical Structure Theory in Mann & Thompson 1988) is investigated, which – despite the intensive research on MPs in
the last decades – has been mainly neglected so far. For a number of MPs, it
is claimed that they have a function with respect to common ground management (cf. Karagjosova 2004, Repp 2013), but it is never spelled out what influence this has on the structure of discourse. The talk presents two quantitative
studies. In a corpus of German parliament speeches, all sentences containing a
MP (ja, doch, eben, halt, wohl and schon) were annotated for their discourse relations (DRs). The corpus study revealed that – for instance – ja occurred significantly more often than would be expected (on the basis of the general distribu-
110
AG 2 · Gebäude B3 1, Raum 0.14
tion of DRs in this text type) in discourse units that constitute Background relations. It occurred significantly less often than expected in Elaboration relations. With ja, a speaker indicates that a proposition should be already known
to the addressee. By using ja in Background, the speaker can mark information as not new and therefore increase the effect the DR has. More than that,
the analysis also shows that ja can trigger the reading of information as backgrounded – even when no Background relation is present. The finding that ja
hardly occurs in Elaboration is in line with this argumentation: In Elaboration, a speaker usually introduces new information. Therefore, the meaning
of ja is not compatible with the function of this DR. This new perspective on
MPs show how speakers can make use of MPs to advise the addressee how to
file incoming information. They can be used to indicate how a proposition is
related to (an)other proposition(s) and anchor information in discourse structure in a certain way.
References: • Karagjosova, E. (2004): The meaning and function of German modal particles. DKFI;
Saarland University 10, 31–36. • Mann, W., Thompson, S. (1988): Rhetorical Structure Theory: A theory
of text organisation. Text. 8(3), 243–281. • Repp, S. (2013): Common ground management: modal
particles, illocutionary negation and VERUM. In: Beyond Expressives–Explorations in Use-Conditional
Meaning, 231–274.
Segmentation and topic annotation of German newspaper editorials
Manfred Stede
Universität Potsdam
[email protected]
Attempts to reliably identify aboutness topics in authentic language have
shown that this task is notoriously difficult (Cook/Bildhauer 2013). We recently presented an annotation study (Stede/Mamprin 2016) where the annotator agreement shows some improvement over the state of the art, and we released a new annotation layer of aboutness topic on the Potsdam Commentary
Corpus (Stede/Neumann 2014).
In the present work, we use that data for an initial qualitative study, which
looks at the relationships between topics and subjects. A crucial factor here
is segmentation, which in our approach consists of a largely structure-driven
“generic” discourse segmentation, followed by a task-specific one (here: information structure) that filters for certain segment types. E.g., for subordinate
111
AG2
Mittwoch
08.03.2017
17:30 – 18:00
B3 1, 0.14
AG 2 · Information structuring in discourse
clauses, we follow Matic et al (2014) in distinguishing between d-subordination
(one complex proposition with the matrix clause) and ad-subordination (two
distinct propositions).
Studying 20 texts with 316 discourse segments and 169 aboutness topics,
we find that 119 (70%) coincide with subjects. The reasons for disjoint topics/subjects sometimes are structural (40%), while for the majority, there appears to be an underlying pragmatic choice (e.g., change of discourse topic).
Almost half of the discourse segments in the data are thetic (topicless), and we
provide a classification of the reasons.
References: • Cook, P. and Bildhauer, F. (2013): Annotating information structure. The case of ”topic”.
In: Dialogue and Discourse 4(2):118–141. • Matic, D., van Gijn, R. and van Valin, R. (2014): Overview.
In: van Gijn, R., Hammond, J., Matic, D., van Putten, S. and Galucio, A.V. (eds.): Information structure
and reference tracking in complex sentences. Amsterdam: John Benjamins. • Stede, M. and Neumann,
A. (2014): Potsdam Commentary Corpus 2.0: Annotation for discourse research. In: Proc. of LREC,
Reykjavik. • Stede, M. and Mamprin, S. (2016): Information structure in the Potsdam Commentary
Corpus: Topics. In: Proc. of LREC, Portoroz.
AG2
Discourse frame setters and the syntax of subject initial V2 in Dutch
Donnerstag
09.03.2017
09:00 – 09:30
B3 1, 0.14
Ciro Greco
Ghent University
Liliane Haegeman
Ghent University
[email protected]
[email protected]
This paper examines the relation between clause internal syntax, in particular the syntax of subject initial V2 patterns in Dutch and its dialects, and its effect on V3 patterns. Both Standard Dutch (StD) and West Flemish (WF) allow a
range of V3 patterns in which a ‘main clause external’ constituent (Broekhuis
and Corver 2016) combines with a bona fide V2 clause. Relevance conditionals
(1) are one example:
(1)
Als je honger hebt, er ligt brood in de kast.
If you hunger have, there is bread in the cupboard
‘If you are hungry there is bread in the cupboard.’
However, an unexpected asymmetry arises between WF and StD with respect to the compatibility of a clause external frame setter with subject-initial
declaratives.
112
AG 2 · Gebäude B3 1, Raum 0.14
(2)
√
Als mijn tekst klaar is, ik zal je hem opsturen. *StD/ WF
when my text ready is, I will you it
up-send
‘When my text is ready, I’ll send it to you.’
Based on a general proposal for the syntax of discourse frame setters, we will
account for the micro-variation observed in (2) and in so doing shed light on
the interface between syntax, semantics and discourse. Our account uses two
core ingredients: (i) we develop a syntax-to-discourse mapping for frame setters (esp. temporal and conditional clauses); (ii) we argue that StD and WF differ with respect to the derivation of subject initial V2 declaratives, which interacts with frame setters at the interface (as also shown in Mikkelsen 2015).
References: • Broekhuis, H. and N. Corver (2016): Dutch Syntax. Verb phrases. Amsterdam University
Press • Mikkelsen, L (2015): VP anaphora and verb second order in Danish. Journal of Linguistics 51,
595–643.
AG2
Zum Ausdruck von (non-)at-issueness in alten Sprachen
Rosemarie Lühr
Humboldt-Universität zu Berlin
Donnerstag
09.03.2017
09:30 – 10:00
B3 1, 0.14
rosemarie.lü[email protected]
Nach Koev (o.J.) hängt der informationstrukturelle Status von appositiven Relativsätzen von der linearen Anordnung im Satz ab. Wenn solche Sätze in der
Satzmitte erscheinen, sind sie not-at-issue, am Ende aber können sie at-issue
sein. So kann der Hörer den appositiven Inhalt nur in (1b) bestreiten:
(1)
a.
b.
Messi, who once scored a goal with his hand, won the Ballon d’Or.
Everybody admires Messi, who once scored a goal with his hand.
Dass derartige Strukturen “sensitive to linear position” sind, bestätigen die altindogermanischen Sprachen. Vgl. Sallust, Catilina 48:
(2)
avaritia
pecuniae
studium
habet
quam
nemo
Habgier
Geld
Streben
hat
welches kein
sapiens concupivit
Weiser
begehrt
‘Die Habgier beinhaltet das Streben nach Geld, das kein Weiser begehrt’
113
AG 2 · Information structuring in discourse
Gleiches gilt für weitere “Subordinating Relations”: Partizipialkonstruktionen, dass-Sätze (Explikativsätze) und Parenthesen. Untersucht werden
die syntaktischen Merkmale dieser “Subordinating Relations” und ihre
Anordnung im Satz. Die Beschreibung ist korpusbasiert. Das Datenmaterial
stammt aus dem Hethitischen, Altindischen, Altgriechischen und Lateinischen. Für die Deutung der (not-)at-issue-Fälle wird auf den “Common Ground”
Bezug genommen, auf vorherige Äußerungen, den situativen Kontext und
Weltwissen.
References: • Antomo, M. I. (2012): Abhängige Sätze in einem fragebasierten Diskursmodell. Phil.
Diss. Göttingen. • Koev, T. (2016): Appositive Relative Clauses, At-issueness, and Timing in Discourse.
Handout.
Identifying basic discourse units in ancient Greek
AG2
Cassandra Freiberg
VU Amsterdam, FSU Jena, MPI SHH Jena
[email protected]
Donnerstag
09.03.2017
10:00 – 10:30
B3 1, 0.14
The aim of this paper is to discuss whether the Intonation Unit (IU) is a valid
basic discourse unit for the corpus language Ancient Greek. To this end the linguistic cues heretofore adduced in favour of IU-segmentation will be critically
reviewed. These most prominently include particles and other clitic elements
which are held to have an inherent P2-tendency within the IU. This view, however, has recently been challenged by Goldstein 2015, who proposes to integrate the clitic elements into a clause-based syntactico-semantic framework
with the internal order phrasal (PH) > sentential (S) > clausal (C) clitic, cf. (1),
thus potentially reducing the IU to a mere epiphenomenal and only descriptively adequate rather than basic element of Ancient Greek discourse structure.
(1)
[tēn menPH garS proterēn hēmerēn] panta sphiC kaka ekhein. [tēn
[the while for previous day]
all
them bad to have [the
dePH
tote pareousan] panta agatha.
whereas then being present] all
good
‘[For on the previous day], everythying was bad for them.
[During the present (day)], however, everything (has been) good.’
(Hdt. 1.126.4; analysis and translation by Goldstein 2015: 88 (4.1))
114
AG 2 · Gebäude B3 1, Raum 0.14
References: • Allan, R.J. (2006): Herodotus’ historiën als sprekend leesboek: Herodotus tussen
oraliteit en geletterdheid. In: Lampas 39, 19–32. • Bakker, E.J. (1997): Poetry in Speech. Orality and
Homeric Discourse. Cornell University Press. • Dik, H. (1995): Word Order in Ancient Greek. A Pragmatic
Account of Word Order Variation in Herodotus. Gieben. • Goldstein, D.M. (2015): Classical Greek Syntax.
Wackernagel’s Law in Herodotus. Brill. • Lühr, R. (2007): Information Structure in Ancient Greek. In:
Anita Steube (ed.): The Discourse Potential of Underspecified Structures. de Gruyter, 487–512. • Matić,
D. (2003): Topic, Focus, and Discourse Structure. Ancient Greek Word Order. In: Studies in Language
27, 573–633. • Scheppers, F. (2011): The Colon Hypothesis. Word Order, Discourse Segmentation and Discourse Coherence in Ancient Greek. VUBPRESS. • Slings, S.R. (1992): Written and Spoken Language: An
Exercise in the Pragmatics of the Greek Sentence. In: Classical Philology 87, 95–109.
Discourse structure of relative constructions
Marco Coniglio
University of Göttingen
Roland Hinterhölzl
University of Venice
[email protected]
[email protected]
Donnerstag
09.03.2017
11:15 – 11:45
B3 1, 0.14
AG2
In Romance languages, mood distinctions are paired with differences in the
interpretation of the head noun of relative clauses (RCs). In (1a) – for Italian –
Gianni is looking for a specific woman, while in (1b) a speaker- and hearer-new
discourse referent is established. In Modern German, the same distinction can
be expressed with Vfinal (Vf-RCs) and V2-RCs, respectively (Ebert et al. 2007).
While the Vf-RC in (2a) is ambiguous between a de re and a de dicto reading, the
V2-RC in (2b) is only compatible with a de re interpretation of the head noun:
(1)
a.
b.
(2)
a.
b.
Gianni cerca una donna che ha
gli occhi blu. (de re)
Gianni looks.for a woman that has.Ind the eyes blue
Gianni cerca una donna che abbia gli occhi blu. (de dicto)
Gianni looks.for a woman that has.Sub the eyes blue
Hans sucht eine Frau, die blaue Augen hat. (de re/de dicto)
Hans looks.for a
woman who blue eyes has
Hans sucht eine Frau, die hat blaue Augen. (de re/*de dicto)
Hans looks.for a
woman who has blue eyes
German lost mood distinctions in RCs in the course of its history. However, Old
High German (OHG) not only allowed for V2 as well as for Vf orders, but also
productively used the relevant mood distinctions:
(3)
Huuer ist dher dhiz al ni chisehe […]? (I, De Fide, 8, 3)
who is who this all Neg sees.Sub
This paper will tackle the question how verb position and mood distinctions
115
AG 2 · Information structuring in discourse
interact in OHG RCs and will present a study on the syntax, IS and discoursestructural interpretation of OHG RCs and their diachronic development.
References: • Ebert, C., C. Endriss & H.-M. Gärtner (2007): An Information Structural Account of
German Integrated Verb Second Clauses. In Semantic Compositionality, ed. Frank Richter & Manfred
Sailer, 415–434.
Embedded NRRCs and discourse structure
Donnerstag
09.03.2017
11:45 – 12:15
B3 1, 0.14
Claudia Poschmann
Goethe Universität Frankfurt a.M.
[email protected]
AG2
Contrary to standard assumptions (McCawley 1982, Potts 2005), Schlenker
(2013) argues that non-restrictive relative clauses (NRRCs) in English can be
embedded semantically under operators of their matrix clause. In a wide scope
reading of the NRRC in (1), Gerd is saved as soon as he reaches Dr. Meier. In
an embedded reading scenario, however, Gerd is saved only if he reaches Dr.
Meier and Dr. Meier happens to give him the right anti-dot.
(1)
Wenn Gerd rechtzeitig Dr. Meier erreicht, der ihm das passende
Gegengift verabreicht, kann er gerettet werden.
If Gerd reaches Dr. Meier in time, who gives him the right anti-dote, can he be
saved
In two experiments, we tested the availability of these embedded readings in
German depending on the clause-type of the embedded construction (NRRCs, and-conjunctions, V2-parenthesis, postponed matrix clauses) and the
predicate type (event vs. state). The results of both experiments indicate
that NRRCs, unlike the corresponding V2-constructions, indeed can be interpreted as embedded in German. This embeddability, however, is dependent
on the predicate type. Only NRRCs with event predicate rated significantly
higher than the V2-constructions. We will discuss several explanations why
the predicate-type of an NRRC might affect its embeddability. One option:
Event predicates allow the NRRC to stand in a coordinating discourse relation (Asher & Lascarides 2003) to the proposition expressed by its host-clause
(Holler 2005) and are thus more easily conjoinable to the host’s proposition
than NRRCs with subordinating discourse relations.
116
AG 2 · Gebäude B3 1, Raum 0.14
References: • Asher, N. & Lascarides, A. (2003): Logics of Conversation. Cambridge University Press,
Cambridge. • Holler, A. (2005): Weiterführende Relativsätze. Akademie Verlag, Berlin. • McCawley,
J.D.(1982): Parentheticals and discontinuous Constituent Structure. In: Linguistic Inquiry, 13(1), 91
–106. • Poschmann, Claudia (2013): Attaching NRRCs to plural quantificational Heads. In: Proceedings
of Amsterdam Colloquium, 171–183. • Potts, Chris (2005): The Logic of Conventional Implicatures. Oxford
University Press, Oxford. • Schlenker, Philippe (2013): Supplements without Bidimensionalism [expanded version]. Manuscript, www.semanticsarchive.net.
Cleft focus and accessibility: Online vs. offline differences
Clare Patterson
Universität Potsdam
Claudia Felser
Universität Potsdam
[email protected]
[email protected]
Donnerstag
09.03.2017
12:15 – 12:45
B3 1, 0.14
Clefting makes a noun more accessible for pronoun reference, but this effect
is critically dependent on discourse units (DUs). When the clefted noun appears in the same DU as the pronoun, the boost in accessibility is reversed. This
reversal has been dubbed the anti-focus effect [1; 2]. This study investigates
whether and how the anti-focus effect emerges during online processing. An
eye-tracking experiment was carried out, manipulating the focus type of the
first noun phrase (NP1), see (1).
(1)
a.
b.
c.
Herr Müller erklärte der Lehrerin am Freitag, dass er nicht mitfahren könne.
‘Mr Müller explained to the teacher (fem.) on Friday that he could
not come along.’
Es war Herr Müller, der der Lehrerin am Freitag erklärte…
‘It was Mr Müller who explained to the teacher (fem.) on Friday…’
Ausgerechnet Herr Müller erklärte der Lehrerin am Freitag…
‘Of all people Mr Müller explained to the teacher (fem.) on Friday…’
There was an early advantage for NP1 reference when the pronoun refers to
the clefted antecedent, in line with memory accounts of easier retrieval for
focused NPs. Later measures, however, did not show a reversal of this pattern. The contrast with the previous offline results and participants’ own postexperiment ratings suggests that the anti-focus effect, which is necessary to
override the initial advantage for the clefted antecedent, may emerge at a late
stage, well after the whole DU has been processed.
117
AG2
AG 2 · Information structuring in discourse
References: • [1] Colonna, S., Schimke, S., & Hemforth, B. (2012). Information structure effects on
anaphora resolution in German and French: A crosslinguistic study of pronoun resolution. Linguistics, 50(5), 991–1013. • [2] de la Fuente, I. (2015, December). Putting pronoun resolution in context: The
role of syntax, semantics, and pragmatics in pronoun interpretation. Université Sorbonne-Paris Cité and
Université Paris Diderot – Paris 7, Paris, France.
Donnerstag
09.03.2017
13:45 – 14:45
B3 1, 0.14
Small building blocks, multiple threads, and large repercussions
Hannah Rohde
University of Edinburgh
[email protected]
AG2
In this talk, I ask two fundamental questions about where discourse relations
can hold: First, within sentences, can small subsentential elements stand in
a discourse relation with other elements of the sentence? Second, across sentences, can multiple relations be inferred to hold between the same pair of
sentences? These questions and their repercussions extend beyond the analysis of discourse structure itself into syntactic parsing, pronoun interpretation,
and the meaning of particular discourse connectives. The studies I report draw
on a range of methodologies, from story continuations to reading time to eyetracking to large-scale crowd-sourced annotation.
When we point to examples of discourse relations, we typically point to relations that hold between sentences. These intersentential connections include
a(n admittedly controversial) number of different types: Some relations rely
on causal reasoning; others depend on the inference of parallelism; still others
describe a sequence of events, etc. However, the size of the segments between
which relations can be inferred need not be restricted to full sentences. If a discourse relation operates within the sentence—for example, between a matrix
and relative clause—this creates an opportunity for structural factors to interact with pragmatics in a way that is not possible for intersentential relations.
The studies reported here draw attention to the interdependence of syntactic
parsing and discourse parsing.
In a parallel vein, when we talk about discourse relations, we typically divide the space of possibilities into cases in which the relation is marked explicitly and cases in which the relation is left to inference. Based on a new largescale dataset with over 50 discourse adverbials, I show that this is not an exclusive or: A conjunction can often be inferred alongside an explicit discourse
adverbial, revealing in some cases multiple concurrent relations.
118
AG 2 · Gebäude B3 1, Raum 0.14
Lastly, I consider different discourse frameworks and the parallels that can
be drawn between models that rely on an inventory of relevance relations and
models that posit the inference of questions under discussion. As an example, I test how intonation can help signal that a particular question is being
addressed, thereby changing expectations about the operative discourse relation and indirectly influencing other pragmatic phenomena.
Script knowledge effects on information structuring
Freitag
10.03.2017
11:30 – 12:00
B3 1, 0.14
Sofiana Chiriacescu
Transilvania Universität aus Brasov
[email protected]
The knowledge that humans possess about the world is not randomly stored in
memory, but is packed into units or scripts, which include information about
objects, situations and sequences of situations (Bartlett 1932, Schank & Abelson 1977, Alba & Hasher 1983). Situations can be more or less structured and
are organized into variables, each of which can contain values. In the present
study we used a multi-sentence story-continuation task to explore to what extend script knowledge activated by different verb types impacts referential
continuity and the structure of events.
Results show that discourses have both local and global coherence structures and that language has different means to give structure to both. First,
participants build expectations about who will be mentioned next based on
the verb introduced in the target sentence (Rohde 2008). Second, script knowledge gives rise to various probabilistic expectations with respect to event sequences and who will be mentioned next in the global discourse. Upon reading
a story introducing a superordinate event (e.g. a rescuing event), hearers activate the subevents belonging to the superordinate event (e.g. cause of rescue).
However, superordinate or general events differ with respect to how specified they are. Results show that the more particularized in terms of containing subevents an event is, the more similar participants’ continuations are, as
participants fill out the activated and missing subevents. On the contrary, if
participants are not bound to filling out particular subevents, their continuations are more heterogeneous.
References: • Alba, J.W., & Hasher, L. (1983): Is memory schematic? Psychological Bulletin 93, 203–231.
119
AG2
AG 2 · Information structuring in discourse
• Bartlett, F.C. (1932): Remembering. Cambridge UP. • Rohde, H. (2008): Coherence-driven effects in
sentence and discourse processing. University of California, San Diego. • Schank, R.C., & Abelson, R.P.
(1977): Scripts, plans, goals and understanding. Psychology Press.
Freitag
10.03.2017
12:00 – 12:30
B3 1, 0.14
AG2
The influence of context on the interpretation of the segments in a
discourse relation
Merel Scholman
Universität des Saarlandes
Vera Demberg
Universität des Saarlandes
[email protected]
[email protected]
Discourse annotation frameworks stipulate different segmentation rules, but
the basic notion is the same: the segments of a relation should convey sufficient information for the intended discourse relation to be inferred. This assumption is also implemented in most automatic discourse relation classifiers.
However, there is also a general consensus in the field that annotators need to
take into account the context of a relation during discourse annotation (see
Lascarides, Asher & Oberlander, 1992; Song, 2010).
The current contribution experimentally examines the influence of context on the interpretation of a discourse relation, with a specific focus on
whether there is an interaction between characteristics of the segment and
the presence of context. In the experiment, participants were asked to insert a
connective from a predefined list between two segments. One group only saw
the two segments of the relation, while the other group saw the two segments
including two preceding and one following context sentence.
Distributions of inserted connectives differed between the conditions for 8%
of all items. Manual inspection of these items revealed that presence of context
lead to higher annotator agreement when
• the first segment of a relation refers to an entity or event in the context,
or introduces important background information
• the first segment consists of a deranked subordinate clause attaching to
the context
• the context sentence following the relation expands on the second.
We plan to present a more detailed discussion of the results, using examples to
illustrate the findings. We will also discuss possible explanations for the lack
of effect of context, and the implications for discourse annotation.
120
AG 2 · Gebäude B3 1, Raum 0.14
References: • Lascarides, A., Asher, N., & Oberlander, J. (1992, June). Inferring discourse relations
in context. In ACL Proceedings. • Song, L. (2010). The Role of Context in Discourse Analysis. Journal of
Language Teaching and Research, 1(6), 876-879.
Inferrable and partitive indefinites in topic position
Umut Özge
Middle East Technical University
Klaus von Heusinger
University of Cologne
[email protected]
[email protected]
Freitag
10.03.2017
12:30 – 13:00
B3 1, 0.14
Indefinite noun phrases can appear in topic position only when they are
strong, specific, or discourse linked (Portner and Yabushita 2001). We argue
that the type of discourse linking itself also influences the acceptability. We
distinguish two types of linking: (i) inferential or conceptual linking, the defining property of inferrable indefinites, and (ii) contextual linking, the defining
property for partitives. Both parameters have two values (+/–), which yields
four conditions:
(1)
a.
-concept
(none)
-concept
b.
+concept
c.
-concept
after the scandal. A car dealer…
(infr)
-concept
+concept
+concept
The government started a support
scheme for art. A sculptor…
(part)
+concept
d.
The police started an investigation
There was a group of people standing
near the entrance. A boy…
(both)
There was a staff meeting in the
newspaper. A reporter…
We used Turkish as our target language, due to its clear indication of topichood and indefiniteness. We tested the acceptability of an indefnite direct
object in contexts like (1a-d) in topic, i.e. sentence-initial, position and compared it to the acceptability of that indefinite in the canonical, i.e. immediately pre-verbal, position. Our results demonstrated that partitive indefinites
(PART) are accepted in topic position, while unlinked indefinites (NONE) and
inferrable indefinites (INFR) are not. Surprisingly, when the partitive relation
was also coupled by an inference (BOTH), the indefinite behaves like an inferrable indefinite, rather than a partitive, i.e. it was not accepted in topic po-
121
AG2
AG 2 · Information structuring in discourse
sition. We account for this observation by proposing that the conceptual relation that licenses the inference overrides the contextually given part-whole
relation expressed by a partitive.
References: • Portner, P. and Yabushita, K. (2001): Specific indefinites and the information structure
theory of topics. Journal of Semantics. 18, 271–297.
A cross-linguistic study on Topic within the framework of the
language into Act Theory
Freitag
10.03.2017
13:00 – 13:30
B3 1, 0.14
Tommaso Raso
Universidade Federal de
Minas Gerais
[email protected]
AG2
Maryualê
Malvessi Mittmann
Univ.do Vale do Itajaí
Centro Univ. FACVEST
Federico
Amorim Cavalcante
Universidade Federal de
Minas Gerais
[email protected]
[email protected]
This paper aims to show how: (i) Utterances and information units can be
parsed through prosodic criteria, and (ii) Topic function, defined with functional, prosodic and distributional criteria, can be consistently identified and
compared across languages, namely, Italian (IT), Brazilian Portuguese (BP)
and American English (AE). For cross-linguistic comparison we selected language samples of approximately 30.000 words and 20 texts. The IT sample is
derived from the C-ORAL-ROM (Cresti & Moneglia 2005), the BP sample is
derived from C-ORAL-BRASIL (Raso & Mello 2012) and the AE sample comes
from a re-segmentation of parts of the Santa Barbara Corpus of Spoken American English (Du Bois et al 2005). Prosodic boundaries signal the segmentation of spontaneous speech into tone units and utterances (the minimal stretch
of speech that is autonomously interpretable both pragmatically and prosodically in discourse). Each tone unit correlates with one Information Unit (IU).
IUs are identified through three criteria: function, prosodic profile and distribution (Cresti 2000; Moneglia & Raso 2014). The cross-linguistic analysis
shows that the concept of Topic applied in the study is consistent across different languages. That is, TOP IUs were identified in the 3 samples using the same
operational criteria described by Cavalcante (2016).
References: • Cresti, E. (2000): Corpus di Italiano parlato. Accademia della Crusca. • Cresti, E.
& Moneglia, M. (2005): C-ORAL-ROM. John Benjamins. • Du Bois, J. et al (2005): Santa Barbara
corpus of spoken American English, Parts 1–4. Linguistic Data Consortium. • Raso, T. & Mello, H.
(2012): C-ORAL-BRASIL I. UFMG. • Moneglia, M. & Raso, T. (2014): Notes on Language into Act
122
AG 2 · Gebäude B3 1, Raum 0.14
Theory. In: Raso, T. & Mello, H: Spoken Corpora and Linguistic Studies. John Benjamins, 468–94.
• Cavalcante, F. (2016): The topic unit in spontaneous american english. UFMG. Available from
<http://hdl.handle.net/1843/MGSS-A7GQ48>.
How the Symmetry Problem solves the Symmetry Problem
Freitag
10.03.2017
13:30 – 14:00
B3 1, 0.14
Matthijs Westera
University of Amsterdam, ILLC
[email protected]
A well-known challenge for accounts of exhaustivity (or “scalar”) implications
is the Symmetry Problem (for an early statement see Kroch 1972):
• Symmetry Problem: if relevance is closed under negation, then excluding all relevant, stronger alternatives leads to a contradiction.
What would make this a problem can be understood in two ways:
I. Foundational: Relevance is necessarily, fundamentally symmetrical.
II. Empirical: Relevance is sometimes symmetrical in cases where exhaustivity is implied.
The first understanding is the most common, but, following Horn (1989), we
argue the contrary: relevance is quite typically asymmetrical. Instead, the Symmetry Problem must be approached as an empirical puzzle – namely that exhaustivity occurs despite explicitly symmetrical interests, e.g.:
(1)
a.
b.
c.
A: I need to know, for each of these five individuals,
whether they were present or absent.
B: John was there, and Bill was there.
A: Wow, only two! That’s disappointing!
I present a new, precise explanation of what is going on here, thereby solving the empirical Symmetry Problem. Its central component is the assumption
that speakers can implicitly raise new questions under discussion, provided
these are part of a discourse strategy for a previous question (Roberts, 1996).
I show that the explanation is superior to existing complexity-based or scalebased attempts to break the symmetry. Surprisingly, it works for any account
of exhaustivity (pragmatic or grammatical) that runs into the symmetry problem – the problem, in a sense, solves itself.
123
AG2
AG 2 · Information structuring in discourse
References: • Horn, L.R. (1989). A natural history of negation. University of Chicago Press. • Kroch,
A. (1972). Lexical and Inferred Meanings for Some Time Adverbs. Quarterly Progress Report of the MIT
Research Laboratory of Electronics 104.
• Roberts, C. (1996). Information structure in discourse. In Yoon & Kathol (eds.), Ohio State University
Working Papers in Linguistics 49:91-136.
Alternate
B3 1, 0.14
AG2
German pronouns in discourse: Information structure versus surface
properties
Markus Bader
Goethe-Universität Frankfurt
Yvonne Portele
Goethe-Universität Frankfurt
[email protected]
[email protected]
We present an ongoing study investigating German p(ersonal)-pronouns and
d(emonstrative)-pronouns. In a prototypical example, the p-pronoun preferentially refers to the first NP and the d-pronoun to the second NP:
(1)
Max will einen Freund treffen. Aber er/der ist krank geworden.
M. wants a
friend meet
But he/d-pro is sick become
‘Peter wants to meet a friend. But he became sick.’
For pronoun resolution, a range of heuristics in terms of information structure, syntactic functions and linear positions have been proposed (summarized in Ellert, 2013). However, the various properties have not been systematically varied in prior research. A major aim of our research was therefore to systematically disentangle the contribution of the different information sources.
Experimental results on the interpretation and production of p- and dpronouns as well as corpus data converge on the conclusion that all of the three
properties mentioned above are necessary to capture the use of d-pronouns, as
stated in the following conclusion.
(2)
A d-pronoun refers to the antecedent that is least prominent in terms
of at least ±subject, ±initial and ±discourse-old.
We discuss our data with regard to the model proposed in Kehler & Rohde
(2013). This model combines heuristics with expectations deriving from world
knowledge and discourse coherence. We are currently running additional experiments to explore the consequences of this theory for the interpretation of
pronouns in German.
124
AG 2 · Gebäude B3 1, Raum 0.14
References: • Ellert, M. (2013): Information structure affects the resolution of the subject pronouns
er and der in spoken German discourse. Discours 12(12), 3–24. • Kehler, A.; Rohde, H.(2013): A probabilistic reconciliation of coherence-driven and centering-driven theories of pronoun interpretation.
Theoretical Linguistics 39(1-2), 1–37.
Accounting for context and variability in a prominence-based model
of discourse meaning
Stefan Baumann
IfL-Phonetik, Universität zu Köln
Jennifer Cole
Northwestern University
[email protected]
[email protected]
Prosody conveys discourse meaning through the encoding of focus and information status, but these factors alone are not sufficient to model observed
variation in the production and comprehension of prosody (see Cole 2015),
especially for prominence. We propose a comprehensive, predictive model
of discourse prominence and its prosodic realization, where multiple factors
incrementally determine the prominence value of each word in an utterance (Fig.1). Categorical and gradient factors are expressed by independent
prominence scales, either lexically determined (Part-of-Speech, semantic
weight) or based on the discourse context of the utterance in which the
word occurs (givenness and pragmatic focus). Prominence values from these
scales combine with syntactic structure in shaping the prosodic structure of
an utterance – locating phrasal juncture and nuclear accents. Within each
resulting intonation phrase, the prominence relations between words are
spelled out, e.g. in a metrical grid. Finally, the adjusted prominence values are
mapped onto accent types and gradient phonetic parameters. This last step is
needed to account for the probabilistic nature of accent type distribution as
well as speaker-specific variation.
Figure 1: Sketch of a comprehensive
prominence-based
model of discourse meaning
(for West Germanic languages)
References: • Cole, J. (2015): Prosody in context. Language, Cognition and Neuroscience 30(1–2), 1–31.
125
Alternate
B3 1, 0.14
AG3
Sekundäre Information und ihre sprachliche Kodierung
Daniel Gutzmann & Katharina Turgay
Universität zu Köln & Universität Landau
[email protected], [email protected]
Raum: B3 1, 0.13
Short description
Ausdrücke und Konstruktionen, die Neben- oder Hintergrundinformationen kommunizieren, nehmen einen gesonderten Status hinsichtlich der
Diskursentfaltung ein. Dies soll in dieser AG untersucht werden, sowie die
Frage, welche Arten dieser sekundären Informationen existieren, ob sie einheitlich charakterisiert werden können und wie sie sprachlich kodiert werden
(Relativsätze, Appositionen, Herausstellungen, Interjektionen, expressive Adjektive etc.).
127
AG 3 · Sekundäre Information und ihre sprachliche Kodierung
Mittwoch
08.03.2017
14:15 – 14:45
B3 1, 0.13
The Bavarian discourse particle fei as a marker of non-at-issueness
Stefan Hinterwimmer
Universität zu Köln
[email protected]
This paper deals with the interpretation of the Bavarian discourse particle fei,
which has traditionally been described as an emphatic particle (see, for example, Schlieben-Lange 1979) and whose first discussion in modern linguistic
terms is found in Thoma (2009). I will show that fei is special among discourse
particles in the following sense: It not only makes a contribution that is interpreted at a level distinct from the level where at issue content (Potts 2005) is
interpreted – as is standard for discourse particles (see Gutzmann 2015 and
the references therein) –, but also exclusively relates to propositions that do
not have entered the Common Ground via being the at issue content of an assertion made by the addressee. This is shown by contrasts like the following
one:
AG3
(1)
A: I’ll ask Franz whether he wants to buy my old bicycle.
I would like to have at least 200 euros for it.
B: Da Franz is fei koi Depp.
The
is
no idiot
(2)
A: Franz is such an idiot.
B: So a Schmarrn. Da Franz is (?? fei) koi Depp.
Such a nonsense. The
is
no idiot
Intuitively, fei is used by the speaker in order to direct the addressee’s attention to a conflict between her own beliefs and the addressee’s beliefs that is
not salient at the point where the sentence containing fei is uttered in the following sense: The proposition p believed by the addressee that contradicts the
proposition q believed by the speaker has not been made a topic of the ongoing
conversation.
References: • Gutzmann, D. (2015): Use-conditional meaning. • Schlieben-Lange, B. (1979): Bairisch
eh-halt-fei. • Thoma, S. (2009). To p or to ¬p. The Bavarian Particle fei as Polarity Discourse Particle.
128
AG 3 · Gebäude B3 1, Raum 0.13
The name-informing and the distancing use of
sogenannt (‘so-called’). A pragmatic account
Mittwoch
08.03.2017
14:45 – 15:15
B3 1, 0.13
Holden Härtl
Universität Kassel
[email protected]
Name-mentioning constructions involving sogenannt are instances of pure
quotation (Davidson 1979; Quine 1981). They adopt two distinct interpretations,
a name-informing interpretation, see (1a), and a distanced interpretation (1b):
(1)
a.
b.
Der Doktor diagnostizierte eine sogenannte “Sepsis”.
Das sogenannte “Hotel” erwies sich als üble Absteige.
In the paper, we will propose a unitary semantic analysis for sogenannt, in
which the expression is treated as polysemous. The so in sogenannt, in its function as a demonstrative anaphor, will be argued to operate as a pointer to the
lexical shape of a name, thus binding the Name argument of the underlying
verbal root of sogenannt. The varying interpretations arise as the result of an interplay between sogenannt’s primary semantic content and pragmatic factors.
In particular, we will claim a relevance-based implicature to be effective in a
sogenannt-construction with highly conventionalized nouns like Hotel, giving
rise to the distanced interpretation we observe in cases like (1b). Distancing sogenannt-constructions will be analyzed as an instance of verbal irony. As such,
they echo a preceding utterance of the mentioned name (Wilson 2013). Crucially, sogenannt-constructions will be claimed to convey not-at-issue content
(Tonhauser 2012): (i) the speaker asserts himself/herself to oppose the semantic appropriateness of the mentioned name and (ii) evaluates the denotatum
negatively. A careful investigation of the empirical facts will shed light on the
tenability of this claim.
References: • Davidson, D. (1979): Quotation. In: Theory and Decision 11(1), 27–40. • Quine, W.O. (1981):
Mathematical Logic (Revised Edition). Harvard University Press. • Tonhauser, J. (2012): Diagnosing
(Not-)At-Issue Content: In: Proceedings of SULA 6, 239–254. • Wilson, D. (2013): Irony Comprehension:
A Developmental Perspective. In: Journal of Pragmatics 59, 40–56.
129
AG3
AG 3 · Sekundäre Information und ihre sprachliche Kodierung
Mittwoch
08.03.2017
15:15 – 15:45
B3 1, 0.13
AG3
The processing of secondary information conveyed by German modal
particles
Laura Dörre
Universität Konstanz
Josef Bayer
Universität Konstanz
[email protected]
[email protected]
Modal particles (MPs) convey secondary information by expressing the
speaker’s attitude about him-/herself and his/her model of the addressee
with respect to the propositional content of the utterance (Bayer & Obenauer
2011). As such, they display an expressive meaning (cf. Kratzer 2004, Gutzmann 2013). Every MP has a counterpart which affects the propositional content. In contrast, MPs take the propositional content of the sentence as their
argument and operate at a different meaning dimension (Gutzmann 2013). We
ask whether this two-dimensionality can be made visible by psycholinguistic
methods. If the MP conveys an additional meaning, it should lead to higher
processing costs. In a series of self-paced reading studies and an auditory sentence completion task, we used minimal pairs (1, 2) containing up to ten particles (boldface). We created contexts (underlined) that triggered either the MP
(1) or the counterpart meaning (2).
(1)
Jetzt ist Maria mal wieder total pleite, warum hat sie bloss das Kleid gekauft?
‘Maria is completely broke again, why did she buy the dress (I can’t find a reason)’
(2)
Maria hätte sich viel mehr leisten können, warum hat sie bloss das Kleid gekauft?
‘Maria was able to afford much more, why did she buy only the dress?’
We found a processing advantage of the counterpart meaning after the particle
and of the MP meaning towards the end of the sentence. The secondary meaning seems to be activated in an additional processing step, supporting the existence of a two-dimensional meaning representation. Furthermore, we found
differences between MPs: MP bloss was processed faster than MP nur and there
was a preference for bloss to occur with an MP-intonation, which was absent
for nur. This suggests that bloss is conventionally marked as conveying expressivity. These results offer promising insights into the meaning representation
of MPs and build the foundation for further psycholinguistic studies on the
processing of secondary information.
References: • Bayer, J. & Obenauer, H.-G. (1989), The Linguistic Review 28, 449–491. • Gutzmann, D.
(2013): In: Beyond Expressives: Explorations in Use-Conditional Meaning, 1–58, Leiden: Brill. • Kratzer,
A. (2004). Theoretical Linguistics 30, 123–136.
130
AG 3 · Gebäude B3 1, Raum 0.13
Non-truth-conditional intensification. The case of ‘good’
Elena Castroviejo
Ikerbasque & UPV/EHU
Berit Gehrke
CNRS-LLF / Paris Diderot
[email protected]
[email protected]
Mittwoch
08.03.2017
16:30 – 17:00
B3 1, 0.13
We address the intensifying use of Catalan bon ‘good’, which emphasizes the
property denoted by the modified N and thus yields intensification and potentially the expression of an emotive attitude (indicated with ‘!’ (1a)); it is unavailable in negative environments (1b).
(1)
a.
b.
Hem tingut un bon ensurt!
we had a good shock
‘≈ We had a big shock!’
(#No) he
menjat un bon tros de pa.
neg have.I eaten a good piece of bread
‘I have (#not) eaten a good piece of bread.’
We treat intensifying good as a more restrictive version of subsective good
(2a), which yields intensification through a monotonic inference encoded as
non-truth-conditional content. This amounts to the comment that any individuals in the extension of N that are ordered higher on the scale, also count as
good Ns (inspired by Nouwen’s 2011 analysis of evaluative adverbs like amazingly), and we formalize this as in (2b).
(2)
a.
b.
[[bonint ]] = λPe,t λxe : ∀y, z ∈ P[y > z ∨ z > y].(good-as(P))(x)
∀y[P(y) ∧ y > x → (good-as(P))(y)]
Treating the monotonicity inference as non-truth-conditionally conveyed accounts for the positive polarity behavior of intensifying good (recall (1b)):
if the at-issue content is negated, then the non-truth-conditional meaning,
which cannot be detached, yields a falsity. That is, if x is not a good piece of
bread, then it does not follow that any larger piece will count as good. This,
we put forth, is the reason why evaluativity-based intensifiers behave as positive polarity items. They share the expression of intensification as a secondary
meaning, which yields falsity if the at-issue part is challenged by sentential
operators.
References: • Nouwen, R. (2011). Degree modifiers and monotonicity. In Egré & Klinedinst (eds.):
Vagueness and Language Use. Palgrave McMillan.
131
AG3
AG 3 · Sekundäre Information und ihre sprachliche Kodierung
Interpretations of the embedded expressive motto in Japanese:
Varieties of meaning and projectivity
Mittwoch
08.03.2017
17:00 – 17:30
B3 1, 0.13
Osamu Sawada
Mie University
[email protected]
AG3
Recent studies of expressives have shown that when expressives such as damn
in English are embedded in the complement of an attitude predicate, they not
only can be speaker-oriented, but also can be non-speaker oriented (Amaral
et al. 2007; Harris & Potts 2009; Tonhauser et al., 2013). Amaral et al. (2007)
and Harris and Potts (2009) have suggested this phenomenon is an instance
of indexicality/pragmatic phenomenon (cf. Schlenker’s semantic binding approach).
In this paper, I will investigate the interpretations of embedded expressives
from new data, i.e., the Japanese comparative expressive motto, and argue that
the interpretation of the embedded motto is not purely pragmatic (not just
a matter of indexicality), and that both semantic and pragmatic mechanisms
can be involved. The expressive motto conventionally implicates that “the expected degree is much greater than a current degree” and often pragmatically
triggers a speaker’s complaint (Sawada 2014). What is interesting about motto
is its variation in meaning and projectivity. When it is embedded under an attitude predicate, it is always subject-oriented and its meaning is within the
scope of an attitude predicate (it is an expressive in the subject’s belief). However, if a deontic modal is inserted in the main clause, both a speaker-oriented
reading, which is a conventional implicature (CI), and a subject-oriented reading, which is an at-issue meaning, become available.
I argue that (i) there can be a dimensional shift from a CI to a secondary
at-issue entailment at a clausal level in a non-speaker-oriented reading, and
that (ii) there is a type, a dependent projective content which requires a consistency between at-issue and CI meanings (including a judge).
References: • Amaral et al. (2007): Review of the logic of conventional implicatures by Chris Potts.
L&P 30, 707–749. • Harris, J. and Potts, C. (2009): Perspective-shifting with appositives and expressive. L&P 32, 523–552. • Sawada, O. (2014): An utterance situation-based comparison. Linguistics and
Philosophy 37, 205–248.
132
AG 3 · Gebäude B3 1, Raum 0.13
Though as a marker of pragmatic humbleness
Mittwoch
08.03.2017
17:30 – 18:00
B3 1, 0.13
Claudia Borgonovo
Université Laval
[email protected]
Abbot (2016) and Simons et al. (2010) predict that languages should contain
within their inventory an item that marks a portion of the content as not at issue but they do not give an example of such an item. The aim of this paper is to
show that the conjunction (al)though marks what it subordinates as not at issue
content. I show that concessive clauses (CCs) may be fully integrated syntactically with the matrix clause; they are central adverbial clauses, but contrary to
other CACs their content is not at issue: they can be dismissed, they outscope
negation and other relevant operators, they can never be focused (neither informationally nor contrastively) and they can never be an answer to any question. The reason for this is not syntactic. The focal allergy of CCs (and maybe
their inherent not at issue-ness) may be derived from the semantic contribution CCs make. CCs express uncauses, inoperant causes: excluding inoperant
causes for the benefit of yet another inoperant cause, the point of focusing, is a
pointless move, if we communicate to eliminate possible ways the world can be
in order to zero in of the way(s) the world actually is. CCs cannot be conceived
of as the only true alternative in a focus set.
References: • Abbot, B. (2016). An information packaging approach to presuppositions and conventional implicatures. Topoi 35(1), 9–21. • Simons, M., J. Tonhauser, D. Beaver and C. Roberts (2010):
What projects and why. Proceedings of SALT 20, 309–321.
133
AG3
AG 3 · Sekundäre Information und ihre sprachliche Kodierung
Fragmented contexts
Donnerstag
09.03.2017
09:00 – 09:30
B3 1, 0.13
Dirk Kindermann
Karl-Franzens-Universität Graz
[email protected]
Stalnaker’s influential conception of the conversational context as common
ground models it in terms of the information mutually taken for granted by
all interlocutors (e.g., Stalnaker 1978). But Stalnaker’s account has a downside: All information in the common ground – whether it was contributed as
at-issue content or non-at-issue content – is equally available, or accessible, to
all participants. However, a number of seemingly unrelated phenomena motivate the thesis that all information is not equally available:
1. Questions & Answers: The same presupposed information may be available as a correct (partial or complete) answer to one question not to another, to which it is also a correct answer.
AG3
2. Presupposition accommodation in disagreements: In disagreements,
speakers are often willing to temporarily accept presuppositions of their
opponents. Such presuppositions are only available for as long as the parties jointly accept one view. Different presuppositions become available
with the temporary joint acceptance of the other view.
3. Making information available: Under some conditions, asserting information that is already in the common ground is a felicitous move and
serves a conversational purpose.
In this talk, I will argue that adding the notion of available information to
the model can account for the above phenomena. On the model, the common
ground is ‘fragmented’: the common ground does not form a single consistent,
deductively closed set of presuppositions, but is rather organized into a number of fragments.(Cf. Stalnaker 1984, Cherniak 1986, Elga & Rayo 2014 for mental fragmentation.) Information in a single fragment is jointly available relative to a linguistic task. Linguistic tasks, in turn, are individuated by questionsunder-discussion (cf. Roberts 1996).
References: • Cherniak. C. (1986): Minimal Rationality. Cambridge, MA: MIT Press. • Elga, A. and
Rayo, A. (2016). Fragmentation and Information Access. Unpublished Ms. • Roberts, C. (1996). Information Structure in Discourse: Towards an Integrates Formal Theory of Pragmatics. In: Working
Papers in Linguistics, 49, 91–136 • Stalnaker, R. (1978): Assertion. Reprinted in Stalnaker, R. (1999).
Context and Content. Oxford: OUP, 78–95. • Stalnaker, R. (1984). Inquiry. Cambridge, MA: MIT Press.
134
AG 3 · Gebäude B3 1, Raum 0.13
Literal and enriched meaning of sentences with weak definites and
bare singulars*
Donnerstag
09.03.2017
09:30 – 10:00
B3 1, 0.13
Ana Aguilar Guevara
National Autonomous University of Mexico
[email protected]
Sentences with weak definites (1) and bare singulars (2) in complementary
distribution convey two kinds of content: the literal meaning (LM), which is
evidently derived from the combination of the meaning of the sentences’ constituents, and the enriched meaning (EM), which cannot be directly attributed
to any constituent:
(1)
Lola went to the store.
LM = Lola went to a store.
EM = Lola went to do some shopping.
(2)
Theo went to church.
LM = Theo went to a church.
EM = Theo went to attend Mass
AG3
This talk examines the semantic-pragmatic nature of LMs and EMs and concludes that, whereas LMs should be considered truth-conditional content,
EMs should be treated as being partly truth-conditional content and partly
conversational implicature. This behaviour is accounted by Aguilar-Guevara
and Zwarts’ [2011, 2013] analysis of weak definites and its extension to bare
singulars, according to which the meaning of (1) is that: (a) Lola is the agent of
an event of going to a location exemplified by the store kind, and (b) the event
is part of the stereotypical usages associated with the kind. (a) corresponds
to the LM that Lola went to a store. (b) accounts for the EM that Lola went
to do some shopping, as this is a usual stereotypical purpose associated with
stores. However, that Lola went there to do some general shopping and nothing else still stereotypically associated with stores is not really stated in the
logical form, but rather an inference conversationally derived.
*
This research was sponsored by the Program UNAM-DGAPA-PAPIIT (project IA401116 “Definitud
regular y defectiva en la lengua natural”)
135
AG 3 · Sekundäre Information und ihre sprachliche Kodierung
Donnerstag
09.03.2017
10:00 – 10:30
B3 1, 0.13
Full NPs as personal pronouns: Reference, truth-conditional meaning,
and use-conditional content
Patrícia Amaral
Indiana University
[email protected]
This paper focuses on the behavior of full NPs used as 2nd person singular personal pronouns that trigger 3rd person singular verb agreement in European
Portuguese, as in (1). Their interpretation is analyzed in terms of two dimensions of meaning: primary (two types of truth-conditional meaning) and secondary (or use-conditional, in the sense of Gutzmann 2015).
(1)
AG3
(Addressing the interlocutor) O Pedro/o Pai/o senhor quer um café?
‘Do you [lit. the Pedro/the father/the sir] want a coffee?’
Besides having the deictic function of referring to the addressee in a specific
conversation, these NPs have descriptive content; they denote a property of
the addressee. Additionally, they have use-conditional meaning; they encode
the social relationship between the interlocutors (in [1], the speaker shows respect and social distance towards the addressee). I propose that these forms
are functional mixed use-conditional items displaying the features [+2d], [+f]
in Gutzmann’s system; the argument of the descriptive content of the NP in the
truth-conditional dimension is reused at the use-conditional level.
This paper shows that items displaying a multidimensional meaning can
have more than one type of truth-conditional content. As for the useconditional component, I discuss to what extent it can be considered backgrounded, as it can easily be refuted and become the main topic of the conversation.
References: • Bhat, S. (2007): Pronouns. Oxford: OUP. • Gutzmann, D. (2015): Use-conditional Meaning. Studies in Multidimensional Semantics. Oxford: OUP. • Pountain, C. (2003): Pragmatic and structural reflections on the expression of the second person notion in Romance, with special reference
to Spanish and Portuguese. In: Bulletin of Spanish Studies 80(2), 145–60.
136
AG 3 · Gebäude B3 1, Raum 0.13
Relating (not-)at-issueness to the Question Under Discussion
Donnerstag
09.03.2017
11:15 – 12:15
B3 1, 0.13
Judith Tonhauser
The Ohio State University
[email protected]
[Based on joint work with David Beaver, Craige Roberts and Mandy Simons]
An important component of understanding the meaning of a speaker’s utterance is identifying how the utterance was intended to contribute to the current topic of the discourse, i.e., identifying which utterance content is at-issue
and which is not. Researchers employ a variety of diagnostics to distinguish atissue and not-at-issue content, including versions of diagnostics that ask participants to choose or interpret assent/dissent responses, as in (1a), to judge
what is being asked in a question like (1b), or to judge the acceptability of utterances like those in (1c); see, e.g., Amaral et al. 2007, Jayez 2009, Xue & Onea
2011, Tonhauser 2012, Destruel et al. 2015 and Syrett & Koev 2015.
(1)
a.
b.
c.
A: Waldo stopped wearing a red and white striped shirt.
B: Yes, that’s true. / No, that’s not true.
Has Waldo stopped wearing a red and white striped shirt?
Waldo stopped wearing a red and white striped shirt because {he
was being made fun of / #loved the pattern}.
How is (not-)at-issue content formally characterized so that these diverse diagnostics can be taken to distinguish at-issue from not-at-issue content? In this
talk, this question is addressed based on a formal characterization of at-issue
content as ordinary semantic content that is relevant to the Question Under
Discussion, as proposed in Beaver et al. 2017.
References: • Beaver, Roberts, Simons & Tonhauser (2017): Questions under Discussion: Where information structure meets projective content. Annual Review of Linguistics 3. • Syrett & Koev (2015):
Experimental evidence for the truth conditional contribution and shifting information status of appositives, JoS 32. • Tonhauser (2012): Diagnosing (not-)at-issue content. Proceedings of SULA 6.
137
AG3
AG 3 · Sekundäre Information und ihre sprachliche Kodierung
Additives and accommodation
Donnerstag
09.03.2017
12:15 – 12:45
B3 1, 0.13
Mira Grubic
Universität Potsdam
[email protected]
The presupposition of additive particles like too is usually assumed to be especially hard to accommodate, due to its anaphoricity (Kripke 2009). There is
however no consensus on what is responsible for this anaphoricity; whether it
is the existence of a salient other individual which is required to be in the discourse context (Geurts & van der Sandt 2004), or whether the whole parallel
proposition needs to be salient (e.g. Beaver & Zeevat 2007).
The former account predicts that the presupposition of the additive particle
in (1) is similar to that of the possessive in (2): both presuppose the existence
of a salient individual (i), and a possessor-possessed relation (ii).
AG3
(1)
(i)
(ii)
Peter has a sister, too.
there is a salient other individual
that individual has a sister
(2)
(i)
(ii)
her sister
there is a salient female individual
that individual has a sister
The prediction would be that (i) is hard to accommodate, whereas (ii) can be
accommodated. Thus, in a context like (3), both should be fine.
(3)
Peter and Mary talked about how it was to grow up on the countryside.
a. During the discussion, it was revealed that he has a sister, too.
b. During the discussion, it was revealed that he knows her sister.
We present the results of an experiment in which this is tested for the German
additive auch and discuss the relevance of our findings for the taxonomy of
projective meaning discussed in Tonhauser et al. (2013). One relevant issue is
the question of what the different triggers grouped together in the same class
have in common, whether there is any way to reduce one to the other. Additives, in particular, have been proposed to underlyingly involve pronoun resolution (cf. also Heim 1992), as have anaphoric definite descriptions (Schwarz
2009). In other work, the anaphoricity of additives has been related to their
association with focus: the salient antecedent is independently required due
to the Givenness of the backgrounded material (Ruys 2015). In this discussion
whether the anaphoricity of additives is due to pronoun resolution or due to
Givenness, our talk makes a contribution in favour of the former account.
138
AG 3 · Gebäude B3 1, Raum 0.13
(Un)expected secondary content in Finnish: Additives and scalars
Elsi Kaiser
University of Southern California
Donnerstag
09.03.2017
13:45 – 14:15
B3 1, 0.13
[email protected]
This research investigates the meaning of focus-sensitive additive and scalar
clitics and particles in Finnish. Additives and scalars are crosslinguistically regarded as contributing presuppositional/non-at-issue information, which is
often viewed as ‘secondary information.’ In Finnish, additive and scalar particles (myös ‘too/also’ and jopapos /edesneg ‘even’) show parallel patterns with different parts-of-speech, but the meaning contribution of additive and scalar
clitics (-kin/-kAAn; positive/negative) differ strikingly depending on whether
they are attached to nouns/adjectives vs. verbs. On nouns and adjectives, the
clitic [-kin/-kAAn] has an additive too/also-meaning and a scalar even-meaning.
Strikingly, when the clitic occurs on verbs, it can make reference to the entire
event (unlike scalar particles jopa/edes, which, when associated with focused
verbs, provide information about the verb likelihood, not the whole event).
Moreover, with verb-modifying clitics the likelihood scale is not ordered as
it is with noun-modifying clitics.
I propose an analysis for noun-attached clitics, identify differences between
verb- and noun-attached clitics, and propose a possible source for these differences. I suggest that differences in the non-at-issue meaning conveyed by
noun-attached vs. verb-attached clitics stem from verb-attached clitics providing secondary information not about the verb but rather about the entire
event. Furthermore, the verb/noun asymmetry poses some challenges for traditional analyses of even which derive its behavior from likelihood scales (see
also Greenberg, to appear), but are in line with approaches emphasizing the
discourse management function of additive/scalar expressions (e.g. Zimmermann 2014).
References: • Greenberg, Y. to appear. A novel problem for the likelihood-based semantics of even.
Semantics and Pragmatics. • Zimmermann, M. 2014. Conventional vs Free Association with Focus.
Talk at Focus Sensitivity from a Cross-Linguistic Perspective.
139
AG3
AG 3 · Sekundäre Information und ihre sprachliche Kodierung
Donnerstag
09.03.2017
14:15 – 14:45
B3 1, 0.13
Presupposition triggers in a cross-linguistic perspective:
Maximize Presupposition vs. Obligatory Implicatures in Ga (Kwa)
Agata Renans
Ulster University
Nadine Bade
Universität Tübingen
[email protected]
nadine.bade@uni-
Joseph
P. DeVeaugh-Geiss
Universität Potsdam
tuebingen.de
[email protected]
AG3
This paper presents an experimental investigation of the obligatory occurrence of presuppositional additives in Ga (Kwa), an under-researched Ghanian
language, in comparison to previous studies on German.
Additives are obligatory when their presupposition is verified by the context, as in (1). The obligatory insertion of presupposition triggers has been explained by exploiting the Maximize Presupposition (MP) principle, i.e., presuppose as much as possible (Heim 1991), and Obligatory Implicatures (OI),
i.e., the stronger the exhaustivity, the more obligatory the additive (Bade 2016,
based on Krifka 1999, Saeboe 2004).
(1)
a. John came to the party. b. Bill did, #(too).
For MP, no contextual factors beyond whether a presupposition holds are predicted to play a role in the insertion of a trigger. By contrast, according to OI
the insertion of the additive should depend on whether an exhaustivity implicature is made prominent in the discourse: the stronger the exhaustivity, the
more obligatory the additive. We tested experimentally the hypothesis that
obligatory additives are related to the strength of exhaustivity in Ga (Kwa),
and compare the results to previous experiments in German using a different
methodology.
The results of the experiment show that Ga exploits the MP principle. This
data contrast with an experiment in German, the results of which support the
OI principle. Comparing the results for both languages points to previously
unattested cross-linguistic variation in pragmatics.
References: • Bade, N. (2015): Obligatory Presupposition Triggers in Discourse. Universität Tübingen PhD Thesis • Heim, I. (1991): Artikel und Definitheit. In: Semantik: ein itnernationales Handbuch
der Zeitgenössischen Forschung, 487–535 • Renans, A (2016): Modeling the exhaustivity effect of clefts:
evidence from Ga (Kwa). In: Proceedings of Sinn und Bedeutung 20.
140
AG 3 · Gebäude B3 1, Raum 0.13
Dog-whistles and the at-issue/non-at-issue distinction
Robert Henderson
University of Arizona
Nadine Bade
Universität Tübingen
[email protected]
nadine.bade@uni-
Eric McCready
Aoyama Gakuin
University
tuebingen.de
[email protected]
Freitag
10.03.2017
11:30 – 12:30
B3 1, 0.13
George Bush’s 2003 State of the Union address contains the following line.
(1)
Yet there’s power—wonder-working power—in the goodness and idealism and faith of the American people.
To most people this sounds like, at worst, a civil-religious banality, but to a certain segment of the population the phrase wonder-working power is intimately
connected to their conception and worship of Jesus. When someone says (1),
they hear (2).
(2)
Yet there’s power—Christian power—in the goodness and idealism and
faith of the American people.
Albertson (2015) shows experimentally that examples like (1) do in fact improve a speaker’s appeal to religious voters, while slipping right by unreligious
voters, unlike uncoded religious appeals like (2), which are punished by nonreligious voters.
Stanley (2015) argues that dog-whistle language like (1) involves a conventional non-at-issue component. After arguing against a not-at-issue account,
we develop our own positive proposal along the lines of McCready 2012, which
considers how speakers converge on whether certain expressives have positive or negative evaluative content. In addition, we will consider properties of
dog-whistle language that are of interest to pragmatic theory, in particular,
the fact that dog-whistle language is only semi-lexicalized and not wholly cooperative.
References: • Albertson, B. L. (2015). Dog-whistle politics: Multivocal communication and religious
appeals. Political Behavior, 37(1), 3-26. • McCready, E. (2012). Emotive equilibria. Linguistics and
Philosophy, 35(3), 243-283s. • Stanley, J. (2015): How Propaganda Works. Princeton University Press.
141
AG3
AG 3 · Sekundäre Information und ihre sprachliche Kodierung
Freitag
10.03.2017
12:30 – 13:00
B3 1, 0.13
Appositive interpretation of relative clauses – Is prosody the cue?
Corinna Trabandt, Alexander Thiel, Emanuela Sanfelici, Petra Schulz
(Goethe Universität Frankfurt)
c.trabandt;thiel;sanfelici;[email protected]
Relative clauses (RCs) in their appositive (1a) but not in their restrictive reading (1b) convey secondary, non-at-issue (NAI) information.
(1)
a.
b.
AG3
Robbi adores Frege, who (as you may know) wrote Über Sinn
und Bedeutung.
Robbi adores the man who wrote Über Sinn und Bedeutung.
In languages like German, where both types of RCs are introduced by the same
form of relative pronoun, the comma intonation, i.e. a non-integrated prosodic
contour of the RC (Emonds 1976), is predicted to be crucial in detecting an appositive reading (Potts 2005). However, studies on the prosodic realization of
RCs in German indicate that non-integrated prosody may not constitute a reliable cue for NAI interpretations (Birkner 2008, Schubö et al. 2015, Trabandt
2016). Furthermore, the role of other cues suggested in the literature, e.g. type
of head noun, has not been investigated experimentally. In two experiments
we investigated the influence of prosody and head noun type (definite DP vs.
bare plural) as potential cues for NAI interpretations of structurally ambiguous RCs in 32 German-speaking adults. The results demonstrate that prosody
is not a sufficient cue to trigger appositive (NAI) interpretations of ambiguous
RCs. Instead, the type of head noun crucially influences the readiness to derive
appositive interpretations along with prosody.
References: • Birkner, K. (2008). Relativ(satz)konstruktionen im gesprochenen Deutsch. Berlin: de
Gruyter. • Emonds, J. (1976). A transformational approach to English syntax. New York: Academic Press.
• Potts, C. (2005). The logic of conventional implicatures. Oxford: OUP. • Schubö, F., Roth, A., Haase, V.,
& Féry, C. (2015). Experimental investigations on the prosodic realization of restrictive and appositive relative clauses in German. Lingua 154. 65–86 • Trabandt, C. (2016). On the acquisition of restrictive
and appositive relative clauses. PhD dissertation, University of Frankfurt.
142
AG 3 · Gebäude B3 1, Raum 0.13
Ad-hoc shifts from primary to secondary information in spontaneous
speech
Freitag
10.03.2017
13:00 – 13:30
B3 1, 0.13
Alexander Haselow
Universität Rostock
[email protected]
Since thinking and structural planning occur in a quasi-simultaneous way in
spontaneous speech, speakers are often caught in a situation in which they
need to deal with two concurring tasks, such as shifting from main or ‘primary’ to background or ‘secondary’ information. An example for such shifts
is provided in (1), where the speaker interrupts a quotative construction-inprogress in order to provide the addressee with background information on
one of the persons talked about.
(1)
so he said San:dra who is this woman who I’d (.) sort of been surreptitiously introduced to in London Zoo the way one is to one’s father’s
mistress uhm (.) was going to have a baby.
(International Corpus of English – Great Britain, S1A-075)
This talk will focus on the different strategies chosen by speakers to utter secondary information within a unit of talk-in-progress in which an already initiated syntactic structure dealing with primary information is temporarily suspended in order to provide secondary information. I will discuss the question
of how speakers of English “get into” and “out” of secondary information in
the course of the production of a single unit of talk, and how speakers resume
the syntactic “thread” after its temporary suspension. The discussion lends
support to more recent findings in syntactic processing (e.g. Phillips 2003)
that real-time speech processing is based on the planning and production of
smaller syntactic segments that are built incrementally in real time, and appears to be guided by the mental activation level of concepts and ideas (Ferreira 2005).
References: • Phillips, C. (2003): Linear order and constituency. In: Linguistic Inquiry 34(1), 37–90.
• Ferreira, F. (2005): Psycholinguistics, formal grammars, and cognitive science. In: The Linguistic
Review 22, 365–380.
143
AG3
AG 3 · Sekundäre Information und ihre sprachliche Kodierung
Parenthesen und ihre Funktionen in didaktisch aufbereiteten
linguistischen Texten
Freitag
10.03.2017
13:30 – 14:00
B3 1, 0.13
Mikaela Petkova-Kessanlis
St.-Kliment-Ochridski Universität Sofia
[email protected]
In der Fachsprachen- und Wissenschaftsspracheforschung wird Fach- bzw.
Wissenschaftstexten mehrheitlich sprachliche Prägnanz als charakteristische Stileigenschaft zugeschrieben. Die zentralen Strategien textueller Reduktion in der Wissenschaftssprache sind laut Kretzenbacher (1991: 119) auf
der Ebene der Syntax zu finden; der Parenthese wird eine relativ hohe Frequenz bescheinigt. Über die textinterne(n) (illokutive(n)) Funktion(en) der
in wissenschaftlichen Texten verwendeten Parenthesen ist allerdings wenig
bekannt.
Der Beitrag setzt sich zum Ziel, sowohl satzwertige als auch nicht-satzwertige Parenthesen, die mittels Klammern und Gedankenstichen als Sekundär- bzw. Nebeninformationen grafisch markiert sind, in didaktisch aufbereiteten wissenschaftlichen Texten deskriptiv zu erfassen.
Leitend für die Untersuchung sind folgende Fragen: Kann Parenthesen,
die in wissenschaftlichen Texten verwendet werden, ein Handlungsstatus
zugeschrieben werden? Wenn (mehrheitlich) ja, welchen Handlungstypen
sind diese Handlungen zuzuordnen? In welchem hierarchischen Verhältnis
stehen die mittels Parenthesen realisierten Handlungen zu der Illokution des
Trägersatzes (vgl. den Ansatz von Bassarak 1985). Handelt es sich um selbstständige Nebenhandlungen oder um subsidiäre Handlungen, die den kommunikativen Erfolg der Illokution im Trägersatz unterstützen (sollen)? Aus
diesen Erkenntnissen sollen die kommunikativen Funktionen der Parenthesen abgeleitet werden.
References: • Bassarak, A. (1985): Zu den Beziehungen zwischen Parenthesen und Trägersätzen.
ZPSK 38, 368–375. • Kretzenbacher, H. L. (1991): Syntax des wissenschaftlichen Textes. Fachsprache
13/3-4, 118–137. • Kügelgen, R. (2003): Parenthesen handlungstheoretisch betrachtet. In: Hoffmann,
L. Funktionale Syntax. Die pragmatische Perspektive, de Gruyter, 208–230.
144
AG4
Encoding language and linguistic information in historical
corpora
Kerstin Eckart & Carolin Odebrecht
Universität Stuttgart, Humboldt-Universität zu Berlin
[email protected], [email protected]
Raum: B3 1, 0.12
Short description
Historical corpora have been established as an empirical digital base for various types of linguistic studies. Annotations on these corpora have to balance
between a diplomatic representation of historical text and its linguistic analysis. This requires a linguistic modelling of annotations to develop annotation
guidelines and concepts, as well as assignment methods and corpus architectures. The working group thus focuses on established and new approaches
which address these requirements for a structured exploration of historical
corpus data.
145
AG 4 · Encoding language and linguistic information in historical corpora
Basic categories in multi layered grammatical annotation
Donnerstag
09.03.2017
11:15 – 12:15
B3 1, 0.12
Mathilde Hennig
Justus-Liebig-Universität Gießen
[email protected]
AG4
The paper presents reflections on the interplay of a multi layered grammatical annotation of New High German texts and the research on basic grammatical desiderata in this period of German language history, based on the annotations of Kajuk (= Kasseler Junktionskorpus) and the annotations in the project
“Basic Syntactic Structures of New High German – about the grammatical foundation of a New High German reference corpus” (carried out in Gießen and
Kassel, funded by the German Research Foundation as a long term project). The
paper argues towards the use of the possibilities of modern multi layered annotation by annotating atomic basic categories on various layers of annotation.
The advantage of annotating atomic basic categories (as opposed to hybrid tags
that merge several informations) is, that while using a corpus for grammatical
analysis on different research questions, the combination of categories can be
choosen freely. Thus the corpus can not only be used for different research
questions, it also becomes compatible with different theoretical interests. Although the annotation of basic atomic categories obviously requires profound
grammatical analysis, it provides the possibility to divide the process of annotation and analysis with regard to theoretical and empirical delicate fields of
research such as German Felderstruktur or elliptical structures. The paper gives
examples for this taken from Kajuk and also provides an outlook on the annotation methods used in the new project on the grammar of New High German.
References: • Kajuk = http://www1.uni-giessen.de/kajuk/index.htm; https://korpling.german.huberlin.de/annis3/ • grammatical foundation of a New High German reference corpus = shortly via
https://www.uni-giessen.de/fbz/fb05/germanistik/absprache/sprachtheorie/team/hennig/forschung
146
AG 4 · Gebäude B3 1, Raum 0.12
Particle verb constructions in historical German and what corpus
studies reveal about them
Donnerstag
09.03.2017
12:15 – 12:45
B3 1, 0.12
Svetlana Petrova
Bergische Universität Wuppertal
[email protected]
Particle verb constructions are a challenging topic not only in modern German (modG) but in its historical stages as well. Standard historical dictionaries
list compounds consisting of a preposition or an adverbs and a verbal base as
lexical items in the vocabulary of the period (ababîzen ‘to bite up’, durehbîzzen
‘bite through’, ûzspîwan ‘spit out’ etc.), but it is unclear whether the first constituent has the status of a prefix bound to the verbal base, or rather behaves
like a separable particle stranded in main clauses (syntactic separability) or
separated from the verb by the infinitive zu-morpheme or the past-participle
morpheme ge- (morphological separability). In addition, it has to be asked if
the potential equivalents of the modG separable particles share their syntactic behaviour, among all their ban from scrambling in the middlefield (unless
contrastively used) and their occasional possibility to move to the prefield of
the clause (see Lüdeling 2001). Searching the recently launched Referenzkorpus Altdeutsch, we discover that there are cases in which the first constituent
actually never occurs in distance from the verb (e.g. ubarfaran ‘to cross’, ubarwintan ‘to overcome’, duruhbîzzan ‘to bite through’ etc.). In other cases, there
is evidence for morphological and syntactic separability, even in translations
close to the word order of their original. Also, like in modG, these items may
appear in the prefield of the clause, especially if they originate from the class
of directional adverbs, but against modG, these elements can also appear in
the middlefield of the clause without bearing contrastive interpretation. This
implies that the particle verb constructions attested in OHG are not exactely
like their modG counterparts, and also that the tag PTKVZ is too broad, covering both separable particles as well as the equivalents of modG non-separable
verbal prefixes.
References: • Lüdeling, A. (2001): On Particle Verbs and Similar Constructions in German. Stanford
• https://korpling.org/annis3/ddd/
147
AG4
AG 4 · Encoding language and linguistic information in historical corpora
Donnerstag
09.03.2017
13:45 – 14:15
B3 1, 0.12
AG4
Annotating a multiregional diachronic corpus of Early New High
German handwritten texts
Lisa Dücker
Universität Hamburg
Stefan Hartmann
Universität Hamburg
Renata Szczepaniak
Universität Hamburg
[email protected]
stefan.hartmann@uni-
renata.szczepaniak@uni-
hamburg.de
hamburg.de
This talk discusses the methodologies and annotation guidelines employed
in encoding a corpus of 57 Early New High German protocols of witch trials.
To our knowledge, this corpus represents the first attempt to make spontaneously produced, handwritten German texts available as a linguistic corpus.
The corpus was compiled in a project on the interplay of semantic and syntactic factors in the evolution of sentence-internal capitalization (Satzinterne
Großschreibung, SiGS for short) in German.
Handwritten texts entail specific challenges for the linguistic encoding of
the data, such as defining sentence or word boundaries (Szczepaniak & Barteld
2016). In this talk, we will give a brief overview of how these challenges were
approached. However, our main focus will be on the semantic annotation of
the data. As animacy plays a major role in the development of noun capitalization (cf. Bergmann & Nerius 1998), all nouns were encoded for animacy. Semantic role annotation is another major step in encoding the corpus data. In
a first step, we use a binary coding scheme (“agent” vs. “non-agent”), which
will be further extended in a subsequent phase of the project. However, the
SiGS corpus is not only a valuable resource for studying graphemic change.
The data can also be used to pursue, for instance, morphological or syntactic
research questions. Therefore, the data will eventually be made available via
ANNIS. In sum, we believe that the SiGS corpus is an important addition to the
German corpus landscape and that the decisions made during the encoding
process can also prove insightful for other researchers dealing with historical
language data.
References: • Bergmann, R. & D. Nerius (1998): Die Entwicklung der Großschreibung im Deutschen
von 1500-1700. Heidelberg. • Szczepaniak, R. & F. Barteld (2016): Hexenverhörprotokolle als sprachhistorisches Korpus. In S. Kwekkeboom & S. Waldenberger (eds.), PerspektivWechsel 43–70. Berlin:
Schmidt.
148
AG 4 · Gebäude B3 1, Raum 0.12
TEITOK: Combining language and linguistic information without
compromise
Freitag
10.03.2017
11:30 – 12:00
B3 1, 0.12
Maarten Janssen
CELGA-ILTEC, Universidade de Coimbra
[email protected]
TEITOK is an online XML-based corpus framework using a strategy to combine the needs of various users of (historic) corpora, by allowing multiple orthographic realizations of the same text, such as a paleographic and a regularized form to serve both philological and computational needs. It uses an inheritance tree for forms, so that only distinct forms have to be stored explicitly.
The different forms are modelled as features over XML nodes (tokens). This
facilitates not merely having two different forms, but as many as required, allowing to also include a diplomatic form, a transliterated form, etc.
In such multi-layered texts, not only the orthography changes, but also the
number of tokens: normalisation can both merge and split words. And for
linguistic processing, it is often required to split contractions into multiple
words, or groups multiple words into a MWE. To solve this, TEITOK first of all
allows spaces inside tokens at any level. And secondly, it allows a single token
to contain more than one grammatical word inside, where the linguistic tags
are modelled over the grammatical words. In this fashion, the “word” doutra
can be normalized to de outra, and then provided with two grammatical words,
the first a preposition, and the second a pronoun.
TEITOK adds the token nodes inline in TEI documents that can contain rich
typographic annotations, notes, facsimile images, etc. And TEITOK allows adding standoff annotation files for complex annotations. The framework contains a collection of GUI tools to manage the resulting XML files, which become
so heavily annotated that editing the raw files becomes unfeasible. It also contains a POS tagger that tags directly over XML files, and can be used with existing parameters, or easily trained on the corpus itself, and in the interface
all annotations can easily be corrected. It also provides a searchable version
of the corpus using Corpus Workbench. The resulting corpora have proven to
be of use not only for corpus/computational linguists, but also for a range of
other researchers.
149
AG4
AG 4 · Encoding language and linguistic information in historical corpora
Syntactical Annotation of an Early New High German Corpus:
Pipeline of the LangBank-Corpus
Freitag
10.03.2017
12:00 – 12:30
B3 1, 0.12
AG4
Zarah Weiß
Universität Tübingen
Gohar Schnelle
Humboldt Universität zu Berlin
[email protected]
[email protected]
Within the framework of the LangBank project, we are developing a semi-automatic pipeline for syntactic annotations and complexity analyses of Early New
High German (ENHG). Currently, we use diplomatically transcribed ENHG
data and their orthographic and morphological normalization, as provided by
the RIDGES corpus (Odebrecht et al. submitted). We use the normalized data
as a proxy for the diplomatic ENHG to perform Natural Language Processing
(NLP) with contemporary German models, thereby circumventing the lack of
stable NLP tools for ENHG. This leads to good performance for most NLP tools.
However, the lack of definite patterns of sentence delimitation in ENHG prohibits stable automatic sentence segmentation, which is mandatory for most
further NLP. Therefore, we manually segment the data into parseable sentential units, which we defined using a non-graphematic, linguistically and pragmatically motivated approach (Weiß & Schnelle forthcoming). It allows us to
obtain dependency, constituency, and topological field parses, based on which
we calculate over 200 features of linguistic complexity. For this, we use Weiß
& Meurers’ (submitted) system for measuring complexity of morphological,
lexical, clausal, sentential, cohesion, coherence, and deagentivation domains,
which we are going to extend by measures of specific ENHG constructions.
Complexity features and all parses are used as new annotation layers. We are
also developing a method to additionally derive parses of the diplomatic layer
from the normalized layer’s parses. All annotations are reintegrated to RIDGES
and exported to ANNIS via the Pepper tool (Krause & Zeles 2014) for query and
visualisation and will be made publicly available.
References: • Krause, T. & A. Zeldes (2014): ANNIS3: A New Architecture for Generic Corpus Query
and Visualization. Digital Scholarship in the Humanities 33(1), 118–139. • Odebrecht, C., M. Belz, A.
Zeldes, A. Lüdeling, T. Krause (Submitted): RIDGES Herbology - Designing a Diachronic Multi-Layer
Corpus. • Weiß, Z. & G. Schnelle (To appear): Sentence Segmentation Guidelines for Early New High German. • Weiß, Z. & D. Meurers (Submitted): Fine-Grained Linguistic Modeling of Textual Complexity
Improves German L1 Grade Level Assessment. COLING Workshop on ”Computational Linguistics for Linguistic Complexity”.
150
AG 4 · Gebäude B3 1, Raum 0.12
A Diacronic Corpus for Romanian (RoDia)
Freitag
10.03.2017
12:30 – 13:00
B3 1, 0.12
Cătălina Mărănduc1 , Cenel-Augusto Perez1 , Ludmila Malahov2 &
Alexandru Colesnicov2
1
Al. I. Cuza University, 2 Academy of Sciences of Moldova
[email protected], [email protected], [email protected],
[email protected]
This paper discusses the evolution of a Romanian corpus of the Dependency
Treebank type, built at the Al. I.Cuza University of Iași. The corpus has rich
annotation and balanced structure. Having the intention to participate at the
PROIEL project, which aligns the oldest Latin, Greek, Slavonic and Armenian
New Testaments, we chose to annotate the first printed Romanian NT at Alba
Iulia (1648). The print of the book contains a lot of peculiarities difficult to process, described in the paper. We began by the automated processing of its first
2,000 sentences in classical syntactic annotation over the previous morphological annotation. We applied the tools for Contemporary Romanian to a fragment in modern Latin script. But the first edition is written in Old Cyrillic alphabet, used to print old books in Romania and Moldova (where the same language is spoken). A first fragment of the Alba Iulia NT has been transformed
in editable Cyrillic text by an OCR program built by the Computer Scientists in
Republic of Moldova. The editable text in the Cyrillic script has been checked
by the computational linguists from Iași (Romania), comparing it with the
printed old book, then it has been wrapped in the checked XML format, and
the form with Latin letters, obtained in the second step of the OCR processing,
has been compared with the second edition of the book. The entirely annotated
and checked book will be used for extracting an old lexicon to be introduced
in a POS-tagger able to annotate Old Romanian written with Latin or Cyrillic
letters.
References: • Davies M. (2010) Creating Useful Historical Corpora: in Diacronía de las Lenguas Iberoromances: 137-166. • Dipper, S. Faulstich, L. Leser, U. Ludeling A. (2010) Challenges in Modelling a
Richly Annotated Diachronic Corpus of German. Proceedings of LREC. • Haug, D. T. T., Jøhndal, M. L.
(2008) Creating a Parallel Treebank of the Old Indo-European Bible Translations. Proceedings of the
Second Workshop on Language Technology for Cultural Heritage Data 27-34.
151
AG4
AG 4 · Encoding language and linguistic information in historical corpora
Freitag
10.03.2017
13:00 – 13:30
B3 1, 0.12
Development and annotation of a newspaper corpus as part of a
doctoral thesis on text structure and cohesion in news items from
the 17th and 18th centuries
Katrin Goldschmidt
Rheinische Friedrich-Wilhelms-Universität Bonn
[email protected]
AG4
Historical news items in contrast to contemporary ones are less identifiable
by typographic means. Moreover, some news items are syntactically and / or
thematically linked to each other, i.e. by repetition of single event-related entities (person, location, time, action) in successive news items (Fritz & Straßner
1996). In order to examine how text and event structure contribute to the constitution and linking of historical news items, a corpus of historical newspapers has been developed, which enables either (text)linguistic or journalistic
questions.
In the course of corpus development 7 German newspaper issues (1609-1767)
were transcribed, subdivided in ca. 430 news items by three annotators, segmented into sentences (ca. 1,600 sentences), and ca. 30,000 token were tagged
with parts of speech. The annotation of textual macro structures (such as sources, cited documents or comments) and event-related entities is based on a
multilevel annotation scheme, that allows the annotation of complex spans
(i.e. entities as discontinuous phrases) and relations between feature values
(i.e. part-whole relations).
Investigating the hypothesis that news item boundaries are typically
marked by punctation and typographic means, the presentation will provide
an overview of the corpus and some evaluation possibilities with the analysis
platform ANNIS (Krause & Zeldes 2014).
References: • Fritz, G. & E. Straßner (eds.) (1996): Die Sprache der ersten deutschen Wochenzeitungen im 17. Jahrhundert. Tübingen: Narr. • Krause, T. & A. Zeldes (2014): ANNIS3: A New Architecture
for Generic Corpus Query and Visualization. Literary and Linguistic Computing.
152
AG 4 · Gebäude B3 1, Raum 0.12
Encoding sociolinguistic variables in a corpus of Medieval Sardinian
texts
Nicoletta Puddu
University of Cagliari
[email protected]
Medieval texts are characterized by a high degree of graphical and linguistic variation and by limited and partial documentation. On the one hand, these
features pose clear problems in the creation of a corpus which, by definition,
should be balanced, representative and normalized. On the other hand, the peculiar variation which emerges from Medieval texts can be fruitfully investigated in a socio-historical perspective. The lack of standardization can indeed
provide clues of historical, regional and social variation, and the widespread
presence of code mixing can reveal repertoires of text producers, shedding
light on their level of literacy and on the distribution of the varieties in the community (Mazzon 2015). In this perspective, Sornicola (2015) clearly demonstrates the importance of providing an adequate sociolinguistic characterization of the document compilers in order to correctly interpret variation in Medieval texts.
In this contribution we will deal with the problem of annotating the condaghes, Sardinian medieval texts that record transactions of churches and monasteries and often contain transcriptions of legal disputes. We will show that a
combined XML annotation of extra-linguistic and philological information according to TEIP5 can help the correct meaning to be assigned to the variation
phenomena present in the texts. In particular, we shall focus on the interference of Italian in the condaghes, which can be assigned to different levels of
text production, and, as a consequence, to different levels of annotation.
References: • Mazzon, G. (2015): Historical sociolinguistics and history of English. In Molinelli, P.&
I. Putzu (eds.), Modelli epistemologici, metodologie della ricerca e qualità del dato. Milano: FrancoAngeli,
50-68. • Sornicola, R. (2015): Curiales, Notarii, Presbyteri nella Campania Alto-Medievale. In Consani, C. (ed.), Contatto interlinguistico fra presente e passato. Milano: Edizioni Universitarie Lettere,
Economia e Diritto, 237-282.
153
Freitag
10.03.2017
13:30 – 14:00
B3 1, 0.12
Jetzt in Neuauflage:
Rainer Dietrich/Johannes Gerwien
Gisela Klann-Delius
Eine Einführung
3., aktualisierte und erweiterte Aufl. 2016
270 Seiten, mit Abb. und Grafiken, € 24,95
ISBN 978-3-476-02644-6 (print)
ISBN 978-3-476-05494-4 (eBook)
Eine Einführung
3., aktualisierte und erweiterte Aufl. 2016
VII, 224 Seiten, mit Abb. und Grafiken, € 24,95
ISBN 978-3-476-02632-3 (print)
ISBN 978-3-476-05473-9 (eBook)
Psycholinguistik
Welche kognitiven Voraussetzungen und Prozesse machen die menschliche Sprachfähigkeit
aus? Die bewährte Einführung präsentiert die
Teilgebiete Sprachliches Wissen, Spracherwerb,
Sprechen, Sprachverstehen sowie Störung und
Krankheit des Sprachsystems. Eine umfangreiche
Bibliographie und ein Sachregister schließen den
Band ab. Die dritte Auflage wurde vollständig
neu bearbeitet sowie aktualisiert und berücksichtigt insbesondere neuere neurowissenschaftliche Erkenntnisse.
Spracherwerb
Der Spracherwerb ist ein Phänomen, das die
Philosophie sowie die Natur- und Geisteswissenschaften zu immer neuen Erklärungsansätzen
gebracht hat. Die Einführung charakterisiert die
Stadien des Erstspracherwerbs von der Lautentwicklung bis zum Erwerb von Syntax und von
diskursiven Fähigkeiten. Weitere Kapitel präsentieren die zentralen Erklärungsmodelle des
Spracherwerbs. Für die dritte Auflage wurde der
Band aktualisiert und um neuere Forschungsergebnisse sowie Literaturhinweise erweitert.
www.metzlerverlag.de
AG5
Sprachliche Kodierungs-Asymmetrien, Gebrauchsfrequenz und
Informativität
Martin Haspelmath
MPI-SHH Jena und Universität Leipzig
[email protected]
Raum: B4 1, 0.26
Short description
Asymmetrische Muster in der Morphosyntax sind ein Kernproblem der Grammatik und wurden früher oft mit dem Begriff Markiertheit beschrieben.
Asymmetrische Paare wie Nominativ/Akkusativ, 3./2. Person, Singular/
Plural, Präsens/Futur, Aktiv/Passiv, Lokativ/Ablativ werden manchmal auch
mit Ikonizität erklärt, aber in dieser AG steht der Gesichtspunkt der Informativität im Vordergrund, denn es gibt viele Hinweise darauf, dass Gebrauchsfrequenz und Vorhersagbarkeit zentral für ihr Verständnis sind.
156
AG 5 · Gebäude B4 1, Raum 0.26
Frequency, coding asymmetries, and the constant flow of linguistic
information
Mittwoch
08.03.2017
13:45 – 14:45
B4 1, 0.26
Gertraud Fenk-Oczlon
Alpen-Adria-Universität Klagenfurt
[email protected]
This presentation starts with some general remarks on the interrelationship
between frequency and cognition: Relative frequency – overall token frequency as well as relative frequency within specific contexts – affects first of all cognitive processes. Higher frequency of use results in higher familiarity, faster
accessibility or higher availability, and thus also in lower cognitive costs of
speech production and perception. These cognitive mechanisms on their part
influence linguistic variables such as length of morphological forms and even
word order: More frequently units tend to be shorter (i) and to be placed before
less frequently used units (ii).
I will present empirical results showing frequency-dependent asymmetric
coding of morphological forms as well as frequency-dependent asymmetric ordering of linguistic units. Then I will argue – based on information theoretic
considerations – that both rules (i) and (ii) contribute to a rather even distribution of information over the time, i.e. to a roughly constant flow of linguistic
information:
(i) In terms of information theory, high relative frequency is related to low
informational content. An element carrying a smaller amount of information
can be processed within a shorter time. This means: Less time or less structural
complexity for communicating less information. The proportionality function
between information content and length of units provides an approximation
to a constant flow of linguistic information.
(ii) As a sentence continues, the remaining words get more and more predictable – the number of possible and plausible continuations decreases, and
so does the (subjective) information. To place informationally rich elements in
a position, which is per se characterized by high information, would produce
peaks of cognitive overload. An appropriate strategy to avoid such peaks is the
tendency to begin a sentence or a clause with those words having a higher predictability in this context, e.g. with (groups of ) words referring to (groups of)
words of the preceding sentence, and with terms coding concepts activated by
this preceding sentence. This tendency would explain, among other things, the
rule “old before new”, “topic before comment”, or “subject before object”.
157
AG5
AG 5 · Sprachliche Kodierungs-Asymmetrien, Gebrauchsfrequenz und Informativität
Coding asymmetries, frequency and predictability:
the case of to vs from
Mittwoch
08.03.2017
14:45 – 15:15
B4 1, 0.26
AG5
Laura Becker
Universität Leipzig
Matías Guzmán Naranjo
Universität Leipzig
[email protected]
[email protected]
Different asymmetries with respect to the frequency/complexity/length of
an expression from the domain of spatial relations have been claimed to hold
cross-linguistically. One such case is the goal-source asymmetry (eg. Nam
2004, Stefanowitsch and Rohde 2004, Lakusta 2005, Lakusta and Landau
2005, Markovskaya 2006, Regier and Zheng 2007, Papafragou 2010): the goal
being more relevant and prominent, it has been argued that motion to is expressed more frequently and in a less complex way than motion from. Different motivating mechanisms, e.g. iconicity, markedness, or diachronic development, have been proposed to be responsible for the attested effect. An alternative that has recently gained popularity for explaining different kinds of coding asymmetries is that these asymmetries are the product of frequency effects
(Haspelmath 2008). In a coding asymmetry, the less frequent form will be at
least as complex (in length or number of markers, etc.) as the more frequent
form. The explanation given is that the more frequent forms are more predictable, and thus they need to be less marked for speakers to identify them.
However, this explanation requires two hypotheses that have not yet been
explored. First, it assumes that higher frequency always translates into higher
predictability, and second, it assumes that more predictable forms will be shorter. So far, typological work on coding asymmetries has not directly explored
these claims.
Addressing the goal-source asymmetry, we propose an empirical approach
to the relation between frequency, iconicity, and predictability. We show that
frequency and predictability of an expression are not linked in a trivial way,
since the latter depends on frequencies of both the whole expression (verb +
preposition) and the individual frequencies of verbs and prepositions. Results
also show that the semantics of the verbal context influence the predictability
of the motion type, so that iconicity could play a role independent from frequency.
In order to address the expression of motion to and from in Spanish and Portuguese, we selected 12 verbs that can combine with prepositions expressing
158
AG 5 · Gebäude B4 1, Raum 0.26
both motion types. The verbs were divided into two groups: (i) those that feature inherent deictic semantics and imply directionality (Spanish ir ‘go’, venir
‘come’, llegar ‘arrive’, salir ‘exit, leave’, entrar ‘enter’, llevar ‘take’, traer ‘bring’,
as well as their Portuguese counterparts); (i) manner motion verbs that are
neutral to directionality and deixis (Spanish andar ‘stroll’, correr ‘run’, caminar ‘walk’, cargar ‘carry’, viajar ‘travel’, nadar ‘swim’, as well as the Portuguese
counterparts). As for the expression of direction, we included the prepositions
a and para for motion to, as well as de and desde coding motion from.
We extracted the frequency for each verb + preposition pair from the Corpus
del Español Web and Corpus do Português Web (Davies 2015-2016, Davies and
Ferreira 2015-2016), and for the individual frequencies of the verb and preposition on their own. We then calculated the directional attraction, known as ∆ p,
from the verb to the preposition (Gries 2013). This measure is calculated as the
probability of the preposition given that the verb is present, minus the probability of the preposition given that the verb is absent. Under the predictability
hypothesis we would expect to find the shorter expressions a and de to be more
predictable than longer ones para and desde.
We find that overall, in accordance with the goal-source asymmetry, a
‘to’ was the most predictable of the four prepositions, but desde ‘from’ was
much more predictable than de ‘from, of ’. Additionally, although de is more
frequent than a (144487746 and 54479710 hits respectively), verb preposition
pairs were more frequent with a than with de. Thus, we find no correlation
between length of the expression and predictability of the expression, but
we do find a correlation between the raw frequencies of the expressions and
their length. An important exception was salir which was the only verb that
attracted de more than a (∆ p 0.17 vs 0.11, respectively). Since salir is strongly
conceptualized as a source motion verb, this attraction asymmetry speaks for
an iconicity effect. If frequency plays a role in marking asymmetries, this is
not linked to predictability.
directional
neutral
∆p
directional
∆p
neutral
mean(freq)
mean(freq)
a
0.2701
a
0.0707
a
656804.143
a
21511.400
de
-0.0009
de
-0.0009
de
85348.857
de
12638.400
desde
0.0034
desde
0.0034
desde
8590.429
desde
1015.000
para
-0.0021
para
-0.0008
para
9992.857
para
1776.400
References: • Davies, M. (2015-2016): Corpus del Español: 2 billion words. Available online at www.
corpusdelespanol.org. • Davies, M. and M. Ferreira (2015-2016). Corpus do Português: 1 billion words.
Available online at http://www.corpusdoportugues.org. • Gries, S. T. (2013). 50-something years of
159
AG5
AG 5 · Sprachliche Kodierungs-Asymmetrien, Gebrauchsfrequenz und Informativität
work on collocations. In: International Journal of Corpus Linguistics 18(1), 137–165. • Haspelmath, M.
(2008). Frequency vs. iconicity in explaining grammatical asymmetries. In: Cognitive Linguistics 19(1),
1–33. • Lakusta, L. (2005). Source and goal asymmetries in non-linguistic motion event representations.
PhD thesis. Baltimore, Maryland: Johns Hopkins University. • Lakusta, L. and B. Landau (2005).
Starting at the end: the importance of goals in spatial language. In: Cognition 96, 1-33. • Markovskaya,
E. (2006). Goal-Source Asymmetry and Russian Spatial Prefixes. In: Tromsø Working Papers in Linguistics. Ed. by P. Svenonius. 33(2), 200–219. • Nam, S. (2004). Goal and source: Asymmetry in their
syntax and semantics. Talk given at: ‘Workshop on Event Structures’, Leipzig University, Leipzig,
March 17-19. • Papafragou, A. (2010). Source-goal asymmetries in motion representation: Implications for language production and comprehension. In: Cognitive Science, 34(6), 1064–1092. • Regier,
T. and M. Zheng (2007). Attention to Endpoints: A Cross-Linguistic Constraint on Spatial Meaning.
In: Cognitive Science 31, 705–719. • Stefanowitsch, A. and A. Rohde (2004). The goal bias in the encoding of motion events. In: Studies in Linguistic Motivation. Ed. by G. Radden and K.-U. Panther. Berlin:
Mouton de Gruyter, 249–268.
Causal and concessive relations: Typology meets cognition
Mittwoch
08.03.2017
15:15 – 15:45
B4 1, 0.26
AG5
Alice Blumenthal-Dramé
Universität Freiburg
Bernd Kortmann
Universität Freiburg
[email protected]
[email protected]
This talk deals with the cognitive foundations of well-established crosslinguistic asymmetries between causal and concessive relations. The talk is
organized into two major parts. Part one provides an overview of typologically recurrent asymmetries that have long been attributed to the iconicity of
complexity. In particular, it will be shown that while there is a general tendency for concessive relations to be marked overtly, causal relations are more
often left implicit. Likewise, compared to causal connectives, concessive connectives tend to be morphologically more complex, to be acquired later in ontogeny, and to emerge later in diachrony. Finally, unlike causal relations, concessive relations do not give rise to online interpretative augmentation or to
diachronic semantic change (König & Siemund 2000; Kortmann 1991, 1997).
Part 2 is devoted to testing the claim that these asymmetries reflect differences in cognitive complexity. More specifically, we present a self-paced reading experiment comparing the processing of interclausal causal and concessive relations in native speakers of English. The experiment explores the following hypotheses:
1. Implicit concessivity is more disruptive to discourse processing than implicit causality.
2. Concessive connectors provide a larger cognitive benefit than causal
160
AG 5 · Gebäude B4 1, Raum 0.26
ones (Xu, Jiang & Zhou 2015).
3. Concessive connectors are more informative than causal ones (in the
sense of setting up stronger online expectations regarding the semantics of upcoming discourse; cf. Xiang & Kuperberg 2015).
Besides measuring reading times, we track EEG components and other physiological responses (skin conductance, pupil dilation) that have been related
to the generation and satisfaction of online predictions in order to explore
whether potential differences in cognitive complexity can be put down to differences in informativeness (Lewis & Bastiaansen, 2015). Overall, this talk
aims to make a step towards illuminating the relationship between typological generalizations and the cognition of individual language users.
References: • König, E., & Siemund, P. (2000). Causal and concessive clauses: Formal and semantic
relations. • In: E. Couper-Kuhlen & B. Kortmann (Eds.), Cause – condition – concession – contrast (pp.
341–360). Mouton de Gruyter. • Kortmann, B. (1991). Free Adjuncts and Absolutes in English: Problems of
Control and Interpretation. Psychology Press. • Kortmann, B. (1997). Adverbial Subordination: A Typology and History of Adverbial Subordinators Based on European Languages. Walter de Gruyter. • Lewis,
A. G., & Bastiaansen, M. (2015). A predictive coding framework for rapid neural dynamics during
sentence-level language comprehension. Cortex, 68, 155-168. • Xiang, M., & Kuperberg, G. (2015). Reversing expectations during discourse comprehension. Language, Cognition and Neuroscience, 30(6),
648-672. • Xu, X., Jiang, X., & Zhou, X. (2015). When a causal assumption is not satisfied by reality: differential brain responses to concessive and causal relations during sentence comprehension.
Language, Cognition and Neuroscience, 30(6), 704-715.
Diachrony as a source of coding asymmetries
Livio Gaeta
Università di Torino
AG5
[email protected]
While much discussion on the role of the coding asymmetries relies on synchronic factors such as the nature of processing or more general cognitive, e.g.
perceptual, factors, in my paper I will focus on diachrony as a source of coding asymmetries. This is not meant to deny the relevance of the synchronic
perspective, but the stress on diachrony may help us understand that a number of phenomena can be structurally accounted for, i.e. are the way they are,
because of their origin from earlier structural environments.
First and foremost, coding asymmetries arise as a consequence of grammaticalization processes. Accordingly, the increase of expressivity of the derived
161
Mittwoch
08.03.2017
16:30 – 17:00
B4 1, 0.26
AG 5 · Sprachliche Kodierungs-Asymmetrien, Gebrauchsfrequenz und Informativität
AG5
form results from the stepwise process of reduction of earlier lexemes subsequently grammaticalized as morphological markers. To make just one example, the grammaticalization of the suffix -erweise to form adverbs in German komisch ‘funny’ → komischerweise ‘funnily’, which has taken place quite
recently and displays an increasing productivity in the last century, restores a
coding asymmetry which was present in Old High German, as testified by adverb formations like snël ‘quick’ → snëllo ‘quickly’, and was lost in subsequent
times due to the general reduction of unstressed final vowels. Clearly, the increased expressivity aimed at by the speakers creating new formulas subsequently developed into grammaticalized structures, as suggested by Haspelmath (1999), explains the asymmetric coding since the more expressive formula also contains an additional semantic nuance which is subsequently expanded into a grammatical construction. While grammaticalization systematically ends up with asymmetric coding, exceptions can arise, namely antisymmetric coding in which the base form is longer than the derived one, usually as a consequence of what has been called exaptation in Gaeta (2016), namely the refunctionalization of extant complex structures which turn out to be
reused to convey a different meaning. Among others, the case of Tsakonian
Greek future forms can be mentioned which go back to earlier presents, exapted into future forms in concomitance with the grammaticalization of new presents (cf. Haspelmath 1998). This example shows that grammaticalization is
not (teleologically!) bound to give rise to asymmetric coding, but follows its
own way, sometimes creating an antisymmetric coding.
A second diachronic source of asymmetric coding depends on what Wurzel
(1984) has called system adequacy, namely the tendency of a (morphological)
system to develop internal consistency by increasing the strength of its systemdefining structural properties, for instance enforcing the reach of extra-morphologically motivated inflectional paradigms. Accordingly, the attraction of
the verb brauchen ‘to need’ towards the sphere of modals brings about the rise
of asymmetric coding insofar as a new singular / plural opposition in the 3rd
person of the present braucht / brauch[ņ] > brauch / brauch[ņ] is created adopting the typical inflectional model of the modals, e.g. soll / soll[ņ] ‘must’. Notice
that phonological erosion cannot serve as an explanation because the homonymic 2nd person plural form braucht is not affected by the change. Also in
this case the increase of system adequacy is not necessarily bound to give rise
to asymmetric coding, as shown by cases in which antisymmetric coding is
created like in Milanese where the subtractive plural marking displayed by
the feminine a-class of scarpa / scarp ‘shoe(s)’ is extended to cases formerly
displaying a zero plural: + carn / carn ‘meat(s)’ > carna / carn (cf. Gaeta 2015).
A third diachronic source of asymmetric coding relates to what has been
162
AG 5 · Gebäude B4 1, Raum 0.26
called the principle of maximal differentiation in Di Meola (2002), whereby
the layering of the same form in different functions due its grammaticalization into a new function is subsequently solved by developing a new shape for
the new function. For instance, in Early Modern German the grammaticalization of the article as a relative pronoun gave rise to a process of formal differentiation from the article source morpheme. In particular, asymmetric coding
characterizes the nominative / genitive form respectively of the masculine der
/ dessen and of the feminine die / deren or derer in pre- or postnominal position
of the relative pronoun with respect the corresponding forms of the article der
/ des vs. die / der. In this case, the formal differentiation has remolded morphological material coming from the analogical matching with the so-called weak
adjectives (cf. Bærentzen 1995).
Finally, a further diachronic source of asymmetric coding is borrowing, as
shown for instance by the case of the German loan suffix -ieren, coming from
the reanalysis of the inflectional ending of Middle French verbs like galoper ‘to
gallop’ > galopieren ‘to gallop’. This brought into the German word formation
system a new way for building denominal verbs like Gast ‘guest’ → gastieren
‘to give a performance as a guest’ besides the native conversions like Zigeuner
‘gypsy’ → zigeunern ‘to lead the life of a vagabond’.
References: • Bærentzen, P. (1995), Zum Gebrauch der Pronominalformen deren und derer im heutigen Deutsch. Beiträge zur Geschichte der deutschen Sprache und Literatur 117: 199-217. • Gaeta, L. (2015),
Anti-relevant, contra-iconic but system-adequate: on unexpected inflectional changes. Paper presented at the 22nd International Conference on Historical Linguistics, Naples 27.-31.7.2015. • Gaeta, L.
(2016) Co-opting exaptation in a theory of language change. In: M. Norde et al. (eds.), Exaptation
and Language Change. Amsterdam etc., 57-92. • Di Meola, C. (2002), Präpositionale Rektionsalternation unter dem Gesichtspunkt der Grammatikalisierung. In: H. Cuyckens et al. (eds.), Perspectives on
Prepositions. Tübingen, 101-129. • Haspelmath, M. (1998), The Semantic Development of Old Presents:
New futures and sub- junctives without grammaticalization. Diachronica 15.1: 29-62. • Haspelmath,
M. (1999), Why is Grammaticalization Irreversible? Linguistics 37.6: 1043- 1068. • Wurzel, W. U.
(1984), Flexionsmorphologie und Natürlichkeit. Berlin.
AG5
163
AG 5 · Sprachliche Kodierungs-Asymmetrien, Gebrauchsfrequenz und Informativität
Asymmetry in the historical development of the copula in
Neo-Aramaic
Mittwoch
08.03.2017
17:00 – 17:30
B4 1, 0.26
Geoffrey Khan
University of Cambridge
[email protected]
The focus of this paper will be the dialect group known as North-Eastern
Neo-Aramaic (NENA) (Khan 2007). NENA dialects, most of which are now
highly endangered, were spoken until recently in south-eastern Turkey,
northern Iraq and western Iran. They have been in intense contact with nonSemitic languages for many centuries, mainly Kurdish, an Iranian language.
The paper is concerned with the historical development of a pronominal copula to a verbal copula in the NENA dialects, which has been stimulated by replication of the model of the verbal copula of Kurdish. The various NENA dialects
(many of them only recently documented) exhibit different degrees of development along the path of change from pronominal to verbal copula, which result
in asymmetries within the individual dialects in the distribution of pronominal and verbal morphological exponents within the paradigms of the copula.
Most dialects are in the process of transition and exhibit asymmetries in the
copula paradigms, which consist of both pronominal and verbal forms. It is significant that the shift of pronominal to verbal exponents of the copula conform
to implicational hierarchies, which include the following:
(i) 1st and 2nd person > 3rd person
(ii) negative > positive
AG5
(iii) non-present > present
Categories on the left of these hierarchies have a greater tendency to have verbal exponents than categories on the right. The categories 3rd person, positive
and present are recognized as being the unmarked members of these hierarchies (Greenberg 1966; Haspelmath 2006, etc.). So, their lesser tendency to
develop into verbal forms can be correlated with their unmarked status. The
paper will examine various possible explanations for these asymmetries in
category shift. One possible explanation is that the unmarked status of the
categories in question can be equated simply with frequency of occurrence.
These categories would resist change and gain autonomy due to their higher
frequency of occurrence (Bybee 2010; Haspelmath 2008). Frequency of occur-
164
AG 5 · Gebäude B4 1, Raum 0.26
rence does not, however, satisfactorily explain other hierarchical tendencies
that can be identified in the shift to verbal exponents in the copula paradigms
of the NENA dialects. An alternative explanation is, therefore, proposed. This
is based on the notion that replication of the Kurdish verbal model is facilitated
by congruence between feature specifications of the Kurdish model schema
and those of the replicated schema of the NENA dialects. Replication of the
Kurdish verbal model takes place primarily when there is congruence between
specified features in Kurdish and NENA. When a category or structure is underspecified in NENA or Kurdish, or both, then the NENA dialect does not
so easily replicate the Kurdish model, which results in the retention of nonverbal copula forms.
References: • Bybee, J.L. (2010): Language Usage and Cognition. Cambridge: Cambridge University
Press. • Greenberg, J.H. (1966): Language Universals with Special Reference to Feature Hierarchies. The
Hague: Mouton. • Haspelmath, M. (2006): ‘Against Markedness (and What to Replace It with)’. Journal of Linguistics 42: 25–70. • idem (2008): ‘Frequency vs. Iconicity in Explaining Grammatical Asymmetries’. Cognitive Linguistics 19 (1): 1–33. • Khan, G. (2007): ‘The North Eastern Neo-Aramaic Dialects’. Journal of Semitic Studies 52: 1–20.
The asymmetry between morphology and word order with respect to
informativity
Mittwoch
08.03.2017
17:30 – 18:00
B4 1, 0.26
Simon Kasper
Philipps-Universität Marburg
[email protected]
In the presentation I will argue that
(a) word order (of S and O in the sense of Dryer 2013) is in important respects motivated by extra-linguistic matters of general event cognition
guiding language users’ expectations and predictions in (incremental)
interpretation, and that
(b) morphological markers are language-inherent means of either (i) redundantly satisfying the predictions for word orders during the time-course
of (incremental) interpretation or (ii) relevantly overriding the predictions for word orders.
The rationale behind (a) is firstly based on the observation of an actor preference in language comprehension. Neurophysiological studies point to the
165
AG5
AG 5 · Sprachliche Kodierungs-Asymmetrien, Gebrauchsfrequenz und Informativität
AG5
conclusion that in incremental interpretation language users tend to interpret
the first ambiguous NP in a clause as the actor of the event (cf. BornkesselSchlesewsky & Schlesewsky 2009).
It is secondly based on the observation that a similar but obviously more general interpretation principle is present in (non-linguistic) event perception.
According to social psychological studies, and based on the assumption that
agentivity cannot be directly perceived, cognizers try to identify the causer
or responsible causer (= agent) of an event as fast as possible and choose the
first one that affords (responsible) causation. In terms of human ecology, the
quick identification of the (responsible) causer in an ongoing event allows the
prediction of the most probable outcome of that event and a quick behavioral
adaptation to it, i.e. this “Responsible Causer Preference” is the most efficient
way of mediating cognizers’ perceptions and actions (cf. Kasper [submitted]).
If language, as I will argue, mediates between someone’s perception and somebody else’s action, the most efficient way of verbalizing events is in a diagrammatically iconic S-O structure (cf. Kasper 2015).
That morphological markers are asymmetric with respect to word order, as
(b) states, follows directly from the aforementioned. Since the “zero hypothesis” of language users is that the word order of an utterance will conform to
the Responsible Causer Preference, the functions of morphological markers
require a fundamental reassessment, from both an offline, static, competencebased perspective and an incremental, dynamic, performance-based perspective. From the former perspective the morphological case-markers in a sentence like ...dass der Chefkoch den Eintopf gewürzt hat (‘that the chef.NOM seasoned the stew.ACC’) are redundant information in that they merely actuate an
interpretation which was the default prediction anyway. Thus, they have no informational value. From the latter, the incremental, perspective the morphological markers do have a function: not to distinguish meanings, of course, but
to confirm the default expectation of an S-O structure even before the speech
act is ended. This is what (i) in hypothesis (b) says.
The back side of the asymmetry, referring to (ii) in hypothesis (b), shows
up in sentences like ...dass den Eintopf der Chefkoch gewürzt hat (lit. ‘that the
stew.ACC seasoned the chef.NOM’). Here, case markers become relevant from
the offline perspective in that they contradict, and in fact override the word order expectation; and they not merely contradict the prediction in incremental
interpretation but also require a reanalysis which produces cognitive costs.
Finally, the default S-O expectation does not only determine the informational relevance or redundancy of morphological markers, but is itself functional, namely as a strategy for interpreting morphologically ambiguous sentences or those without morphological markers. In the presentation I will pre-
166
AG 5 · Gebäude B4 1, Raum 0.26
sent historical, typological, and neurophysiological evidence for the asymmetry hypothesis presented in (a) and (b).
References: • Bornkessel-Schlesewsky, I./Schlesewsky, M. (2009): The role of prominence information in real-time comprehension of transitive constructions. A cross-linguistic approach. In: Language and Linguistics Compass 3(1), 19–58. • Dryer, M.S. (2013): Order of Subject, Object and Verb.
In: Dryer, M.S./Haspelmath, M. (eds.): The World Atlas of Language Structures Online. Leipzig: Max
Planck Institute for Evolutionary Anthropology. (Available online at http://wals.info/chapter/81, Accessed on 2016-06-17.) • Kasper, S. (2015): Instruction Grammar. From Perception via Grammar to Action.
Berlin/ Boston: de Gruyter. • Kasper, Simon (submitted): Frequency and iconicity revisited. Towards
an integrative ecological perspective. In: Herbeck, P./Tschugmell, N./Wolf, J. (eds.): Living Economies.
Berlin/Boston: de Gruyter.
Simulating the development of encoding asymmetries in argument
marking
Donnerstag
09.03.2017
09:00 – 09:30
B4 1, 0.26
Sander Lestrade
Centre for Language Studies, Radboud University Nijmegen
[email protected]
How are motivating factors (such as Jakobson’s markedness, Haiman’s iconicity of complexity, and Zipfian communicative efficiency) turned into linguistic conventions? In his ERC proposal, Haspelmath (2015) suggests that “higherfrequency items are more predictable than lower-frequency items, and predictable content need not be expressed overtly or can be expressed by shorter
forms.” Still, it is unclear what exactly are the mechanisms that create and
maintain efficient structures language.
I have developed a computational model of event communication in which
language evolution, or rather, at present, the development of argument marking, can be simulated. In such a model it is necessary to be maximally explicit
about the mechanisms involved. I slightly deviate from Haspelmath’s suggestion and tease apart the hypothesized effects of frequency and predictability,
assuming that predictable parts of meaning need not be said explicitly, and
that frequent words are more likely to be used again (because of a lower activation threshold) and pronounced less carefully (because of pronunciation
automatization). In this talk I will introduce the model showing how asymmetrical argument marking may emerge in a protolanguage from these assumptions.
Agents communicate about automatically generated events in their virtual
167
AG5
AG 5 · Sprachliche Kodierungs-Asymmetrien, Gebrauchsfrequenz und Informativität
AG5
world. The speaker has to find an adequate wording for a target event that
is sufficiently distinctive given the situational context in which other events
are ongoing too (i.e., if there are similar distractor objects, referential expressions have to be more specific; cf. Levelt 1989). This also involves making clear
the distribution of predicate roles over the event participants in the communicated event if necessary (contrast man book read, which is non-ambiguous in
the absence of marking, with man woman see). If the hearer correctly identifies
the event the speaker is talking about, the agents mark the successful usage
of the words that constitute the utterance, remember the exact meaning for
which the words were used, and next either switch turns to go on with their
conversation or end it, after which two new agents are randomly selected for
a new conversation (cf. Steels 1997).
Over time, words may grammaticalize (Heine and Kuteva 2007). Two important mechanisms in this process are erosion (frequent forms being pronounced
sloppily and eventually becoming represented accordingly; Nettle 1999) and
desemanticization (frequent meanings becoming more general as a function of
the different contexts in which they are used; Bybee 2010). If a meaning becomes more general, it can be used in even more contexts, and if a form becomes too short to stand on its own, it is suffixed to its host (Bybee, 1985).
As relative frequency plays a role in word activation, items that have previously been used for role disambiguation are more likely to be considered again.
And as there are only two roles to be kept apart, a previously used role marker
is often found good enough. As a result, its frequency of usage increases further, as well as the variation of its usage contexts. Because of the former, its
form is likely to erode; because of the latter, its meaning is likely to bleach.
Thus, it may end up as a (differentially used) case marker eventually.
Also the asymmetrical development of person marking follows relatively
straightforwardly. Speech participants figure in many events. Because of
the resulting frequent and varied usage, words referring to local persons are
more prone to erosion and desemanticization. Differently from role markers,
however, which do not have a referential function, once the form of a referential expression becomes too short to refer properly, a more expressive copy
has to be recruited. The erstwhile local pronoun attaches to the verb indexing
the person of its helper (Ariel, 1999). In contrast, third person pronouns
often cannot be used, as they would lead to ambiguity because of the object
distractors in the situational context.
References: • Ariel, M. (1999): “The development of person agreement markers: From pronouns to
higher accessibility markers”. In: Usage based models of language, 197-260. • Bybee, J.L. (1985): Morphology. a study of the relation between meaning and form. John Benjamins. • Bybee, J.L. (2010): Lan-
168
AG 5 · Gebäude B4 1, Raum 0.26
guage, Usage and Cognition. Cambridge University Press. • Haspelmath, M. (2015): “Grammatical
Universals”. (http://research.uni-leipzig.de/unicodas/grammatical-universals/) • Heine, B. and T.
Kuteva (2007): The genesis of grammar: a reconstruction. Oxford University Press. • Levelt, W. J. M.
(1989): Speaking. From intention to articulation. MIT. • Nettle, D. (1999): Linguistic diversity. Oxford
University Press. • Steels, L. (1997): “Constructing and sharing perceptual distinctions.” Machine
Learning. In: ECML-97, 4-13.
Explaining coding asymmetries: Frequency or informativity?
Donnerstag
09.03.2017
09:30 – 10:00
B4 1, 0.26
Natalia Levshina
Leipzig University
natalia.levshina@ uni-leipzig.de
It is well known that more frequent grammatical categories tend to be less
formally marked than less frequent ones Greenberg (1966). For example, the
number markers of singular nouns languages are either as long as or shorter
than those of plural nouns. However, more recent computational linguistic
studies (e.g. Piantadosi et al. 2011) demonstrate that formal length is in fact
better predicted by average informativity of a word (i.e. the inverse of the
word’s average conditional probability given the preceding context) than by
its context-independent frequency. These findings support the theory of uniform informational density as a means of optimization of human communication (Jaeger 2010). The present paper aims to answer the question whether linguistic coding asymmetries can be better explained by differences in relative
frequency (e.g. Greenberg 1966; Haspelmath 2008) or by those in informativity.
This study focuses on three well-known cases of coding asymmetries: singular/plural nouns, absolute/comparative forms of adjectives and cardinal/ordinal numerals. On the basis of the Google books n-grams in English, French,
German, Italian, Russian and Spanish, I compute the normalized and relative
frequencies, as well as the average contextual predictability scores based on
n-grams with different n for samples of word forms representing the grammatical categories. Goodness-of-fit measures of mixed-effect logistic regression models are used to estimate how helpful frequency and informativity are
in predicting whether a given word form belongs to one or the other category.
References: • Greenberg, J. (1966): Language universals, with special reference to feature hierarchies.
The Hague: Mouton. • Haspelmath, M. (2008): Frequencies vs. iconicity in explaining grammatical asymmetries. Cognitive Linguistics 19 (1), 1–33. • Jaeger T.F. (2010): Redundancy and reduction:
169
AG5
AG 5 · Sprachliche Kodierungs-Asymmetrien, Gebrauchsfrequenz und Informativität
Speakers manage syntactic information density. Cognitive Psychology 61, 23–62. • Piantadosi, S., H.
Tily & E. Gibson. (2011): Word lengths are optimized for efficient communication. PNAS 108(9).
Donnerstag
09.03.2017
10:00 – 10:30
B4 1, 0.26
Making sense of the asymmetry between affirmation and negation
Matti Miestamo
University of Helsinki
[email protected]
AG5
Negation and affirmation are known to show asymmetric morphosyntactic behaviour across languages. According to all classical markedness criteria – structural coding, behavioural potential and frequency, to use Croft’s
(2003) terms – negation appears as marked vis-à-vis affirmation. In Miestamo
(2005), the asymmetry between affirmatives and negatives was examined in
a broad typological perspective. A distinction was made between symmetric
and asymmetric negation according to whether the only difference between
affirmatives and negatives is the presence of the negative marker(s) (symmetric negation) or whether affirmatives and negatives differ in other ways, too
(asymmetric negation). The asymmetry between affirmation and negation can
manifest itself in the construction or in the paradigm: constructional asymmetry is about the structural differences between a negative clause and its affirmative counterpart, whereas paradigmatic asymmetry is about the paradigmatic choices (e.g., TAM distinctions) available in negatives vs. affirmatives.
Furthermore, asymmetric negation can be divided into subtypes according to
which grammatical domain is affected by negation and how. These subtypes
include A/Fin in which the finiteness of the verb(s) is affected, A/NonReal
in which negatives are obligatorily marked with non-realis categories, and
A/Cat/Neutr in which grammatical category distinctions are neutralized under negation, i.e. some categories (e.g., TAM, person, number) available in the
affirmative are blocked under negation.
Possible explanations for the cross-linguistically recurring asymmetries
were discussed in Miestamo 2005. For the asymmetry affecting finiteness
(A/Fin), in which negative clauses are typically construed as stative predications, the proposed explanation lies in the stative character of negation. For the
asymmetry in the marking of reality status (A/NonReal), it is natural to seek
the explanation in the unreal semantics of negation. Finally, for the neutralization of category distinctions (A/Cat/Neutr), a possible explanation is found
170
AG 5 · Gebäude B4 1, Raum 0.26
in the discourse context of negation: negatives are typically used in contexts
in which the corresponding affirmative is somehow present for the discourse
participants (cf. Givón 1978) and the pragmatic need for the explicit marking of
the properties of the event (e.g., TAM) is thereby reduced. An additional or alternative motivation for the neutralization of distinctions can be found in the
properties of non-events: regarding tense, for example, it is not always easy or
relevant to place a negated event in time. Finally, motivations can be sought in
the effects of frequency: negatives being much less frequent in discourse than
affirmatives, languages do not as easily grammaticalize and preserve distinctions in negatives as they do in affirmatives (cf. Haspelmath 2008). These motivations are relevant as alternative/additional motivations for Types A/Fin and
A/NonReal as well, to the extent these involve reduced marking of grammatical distinctions under negation.
The asymmetry between affirmation and negation has many different manifestations and different factors can be recruited to explain them. For some
aspects of the asymmetry we can propose alternative explanations that can
also be seen as working simultaneously towards a similar effect. In this talk,
I will discuss the different motivations and their relationship in explaining
the cross-linguistic patterns. I will also make reference to diachronic developments in individual languages or language families to see whether we can find
concrete support for the proposed motivations. Furthermore, I will use corpus data to evaluate the proposed effects of the discourse context of negation
on the structure of negatives.
References: • Croft, W. (2003): Typology and universals. 2nd ed. Cambridge University Press. • Givón,
T. (1978): Negation in language: Pragmatics, function, ontology. In: Syntax and Semantics. Vol. 9. Pragmatics, Peter Cole (ed.), 69–112. Academic Press. • Haspelmath, M. (2008): Frequency vs. iconicity
in explaining grammatical asymmetries. In: Cognitive Linguistics 19(1), 1–33. • Miestamo, M. (2005):
Standard negation: The negation of declarative verbal main clauses in a typological perspective. Mouton de
Gruyter.
AG5
171
AG 5 · Sprachliche Kodierungs-Asymmetrien, Gebrauchsfrequenz und Informativität
The Semitic Perfect and the problem of zero subjects
Donnerstag
09.03.2017
11:15 – 11:45
B4 1, 0.26
Na’ama Pat-El
University of Texas, Austin
[email protected]
AG5
In many languages third person subject markers on verbs are frequently
zero, while first and second person are overt (Cysouw 2003; Siewierska 2009).
There have been several explana- tions for this phenomenon. Some suggest
that it is a function of frequency: third person pronouns are less frequent in
discourse (e.g., Bybee 1985). According to this explanation third person subjects do not develop in the first place. Others point to accessibility hierarchy,
i.e., the more accessible a referent is, the less grammatical encoding it requires.
Thus, first and second persons, which are more accessible, constantly regenerate, while the third person markers do not. Proponents of this explanation
assume that third person subjects do develop, but are subsequently lost and reduce to zero (Givón 1976). Another possible explanation takes an iconicity approach: third persons are semantically unmarked relative to first and second
persons and are therefore also morphologically unmarked (Siewierska 2009).
Siewierska distinguishes between absolute zero, where a language lacks any
form in a particular paradigm to mark a person, and paradigmatic zero, where
one slot in the paradigm, mostly third person singular, may have zero, but not
others. Third person markers may be expressed as zero in adnominal possessive position, object or subject. She examines the frequency and cause for zero
morpheme in all three and suggests that subjects tend to be of the paradigmatic
type; she further concludes that a loss scenario is more likely to explain zero
third person subjects. Koch (1995) suggested that synchronically in a paradigm
one value is unmarked and least specific. Since singular is semantically the unmarked value of the number paradigm and third is semantically the unmarked
value of the person paradigm, the third person singular is the unmarked value
in any paradigm that marks such distinctions, i.e., subject, possessor etc. On
the basis of this, he proposed a diachronic principle, according to which “[a]
word-form which expresses by means of a non-zero marker a property which
is typologically expected to be coded by zero is liable to be reanalysed as containing a zero marker”. According to Koch, therefore, morphological change
consists of a drive to semantically create more iconic coding.
In this talk, I will present several cases of verbal paradigms developed in the
172
AG 5 · Gebäude B4 1, Raum 0.26
Semitic family, where third person forms are not marked for person, while
first and second persons are overt and their source is fairly transparent. The
third person pronoun in Semitic is a distal demonstrative and does not have
distinct forms for subject and oblique. These functions are therefore expressed
syntactically: the subject is independent, while the oblique is suffixed (Huehnergard & Pat-El 2012). In verbal paradigms based on canonical subjects (“nominative”), the 3rd person cannot be marked and the paradigm is asymmetric;
however, in verbal paradigms where the person is based on non-canonical subjects (“oblique”), it is marked, and the paradigm is symmetric. This distribution suggests that the reason for zero of 3 person in verbal paradigms is primarily structural, not semantic or pragmatic.
References: • Bybee, J. L. (1985). Morphology. Amsterdam: Benjamins. • Cysouw, M. (2003). The
paradigmatic structure of person marking. Oxford: OUP • Givón, T. (1976). Topic, pronoun and grammatical agreement. In: Subject and topic. Ed. by Ch. Li., 151–188. New York: Academic Press. • Huehnergard, J. & N. Pat-El (2012). Third person possessive suffixes as definite articles in Semitic. Journal
of Historical Linguistics 2.1: 25–51. • Koch, Harold (1995). The creation of morphological zeroes. In:
Yearbook of Morphology 1994, 31–71. • Siewierska, Anna (2009). Person asymmetries in zero expression and grammatical function. In: Essais de linguistique gé né rale et de typologie linguistique offerts au
Professeur Denis Creissels à l’occasion de ses 65 ans. Ed. by F. Floricic, 425–438. Paris.
Processing shapes grammar. But whose processing are we talking
about?
Dirk Pijpops
Research Foundation Flanders,
University of Leuven
Donnerstag
09.03.2017
11:45 – 12:15
B4 1, 0.26
Freek Van de Velde
University of Leuven
[email protected]
[email protected]
AG5
Processing shapes grammatical organisation, including asymmetric coding
with a marked vs. unmarked alternance (Hawkins 2004), but it is unclear
whether the processing considerations at issue are those of speakers or of addressees. Hawkins’s model is framed as benefiting the addressee, though he
remarks that it equally benefits the speaker (2004: 24-25). Glossing over parsing and production is legitimate as long as speakers’ and addressees’ motivations are aligned, but this is not always the case. The idea that language has to
seek an optimal balance between the often opposite demands of both speech
act participants is old, harking back at least to Georg von der Gabelentz in the
19th century. So eventually, we will have to decide which of the two speech
173
AG 5 · Sprachliche Kodierungs-Asymmetrien, Gebrauchsfrequenz und Informativität
AG5
act participants has the upper hand in the processing-driven organisation of
grammar.
On the one hand, there is evidence for an addressee-oriented view:
Hawkins’s ‘Minimize Domains’ principle, stating that the syntactic structure
should be recognisable in as short a span as possible, benefits the addressee, as
the speaker is never unsure about the syntactic structure. Likewise, Rohdenburg’s (1996) Complexity Principle stating that in complex structures more explicit encoding is used is only beneficial to the addressee. If the structure is
already complex, adding extra grammatical encoding arguably burdens the
speaker’s performance even more. On the other hand, it is not self-evident
that speakers should be concerned with their addressees’ needs forfeiting
their own. Speaker’s altruism is evolutionarily implausible (Kirby 1999). Levinson (2000) also stresses the speaker’s needs in his neo-Gricean approach. As
Levinson points out, the bottleneck in human communication is at the production side: decoding is much faster and more effortless than encoding (Levinson 2000: 28), so that taking inferential short-cuts to add layers of meaning on top of what is truth-conditionally encoded is especially helpful for the
speaker. Adding extra material in the overtly coded variant in an alternance
(e.g. zero- vs. that-complementation in English) goes against the rationale to
prioritize production efficiency over parsing speed. Hawkins’s principle ‘Minimize Forms’ also seems first and foremost serve the speaker’s comfort. True,
reducing forms also adds to the parsing effort, as the form-function pair of the
extra encoding has to be stored in the hearer’s brain, but given the ease with
which inferencing is accomplished (Levinson 2000), and given the vast storage capacities of the human mind (Dąbrowska 2014: 626), the extra speaker’s
efforts outweigh the extra addressees’ efforts.
In our paper, we will adduce quantitative data from a close-up case study
that can shed light in the debate over speaker vs. addressee processing. The
case study deals with the direct object vs. prepositional object alternance in
Dutch verbs, like zoeken (naar) ‘search (for)’. A corpus study reveal that the
prepositional variant is used more often when the object is syntactically complex. This can be explained in two ways: first, the preposition can function as a
signpost to help the addressee decode the message. This would be in line with
Rohdenburg’s Complexity Principle, and would point to a hearer-driven processing account. Second, the use of a preposition allows the object to be extraposed (or ‘exbraciated’). This would be beneficial to the speaker, who can
postpone the expression of the complex object at the end of the clause, when
all other issues have been resolved, avoiding centre-embedding. On the basis
of corpus investigation, we will tease apart both explanations. Of special interest are cases such as (1), where the head noun of the object is not extraposed
174
AG 5 · Gebäude B4 1, Raum 0.26
(to the right of gezocht ‘search-PST.PTCP’), but the submodifying complement
clause is. If the use of the prepositional variant is especially favoured in this
context, this would be an argument for the first explanation. Here, the processing difficulty of the discontinuous object may be alleviated for the hearer
by adding the extra signpost.
(1)
De meesten van ons hebben (naar) manieren gezocht om de
The most
of us have
ways
searched for the
(to)
dilemma’s van de conflicten in hun relaties en hun jeugd
dilemmas of
the conflicts
in their relations and their youth
dilemmas te boven te komen.
dilemmas to above to come
‘Most of us have searched (for) ways to overcome the dilemmas of the
conflicts in their relations and their youth.’ (SoNaR, Oostdijk et al. 2013)
References: • Dąbrowska, E. (2014): Recycling utterances: a speaker’s guide to sentence processing.
In: Cognitive Linguistics 25(4): 617-653. • Hawkins, J. (2004): Efficiency and complexity in grammars.
Cambridge: Cambridge University Press. • Kirby, S. (1999): Function, selection and innateness. The
emergence of language universals. Oxford: Oxford University Press. • Levinson, S. (2000): Presumptive
meanings: the theory of generalized conversational implicatures. Cambridge, MA: MIT Press. • Oostdijk,
N., M. Reynaert, V. Hoste & I. Schuurman. (2013): The Construction of a 500-Million-Word Reference
Corpus of Contemporary Written Dutch. In: Theory and Applications of Natural Language Processing.
219-247.
Donnerstag
09.03.2017
12:15 – 12:45
B4 1, 0.26
Why bring is doing the splits
Exploring transitivity as an explanatory factor for coding
asymmetries.
Ulrike Schneider
Universität Mainz
Britta Mondorf
Universität Mainz
[email protected]
[email protected]
AG5
This paper explores the constructional split between causative and noncausative bring. In English, the possibilities of morphosyntactic modification
of the causative bring construction have become rather limited. This is best illustrated by a comparison between English and German. While the latter still
permits uses such as (1), the English equivalent (2) has become impossible. The
modal, negated, reflexive use in (3), on the other hand, is alive and kicking.
175
AG 5 · Sprachliche Kodierungs-Asymmetrien, Gebrauchsfrequenz und Informativität
(1)
(2)
(3)
Er brachte Deutschland zum Lachen: […] (Nordwest-Zeitung Online 6
Oct 2014)
*He brought Germany to laugh.
He couldn’t bring himself to laugh.
We first present a diachronic analysis of the development this causative
construction underwent in English. Employing a 76-million-word corpus of
British English novels spanning six centuries, we show that the causative bring
construction has started out as a morphosyntactically freely modifiable construction, accepting all options listed in (4). But by the 20th century, the construction has come to be almost exclusively restricted to modal, negated reflexive uses.
(4)
AG5
Animate/inanimate
agent
Full NP/reflexive as
patient
Realis/modal
verb phrase
Affirmative/negative
verb phrase
Finite/non-finite
verb phrase
Active/passive
construction
Secondly, we explore the reasons for this enormous restriction of the construction and particularly for the direction the development has taken. Our primary
explanatory factor is transitivity as defined by Hopper and Thompson (1980),
who advocate a gradual, semantic concept of transitivity – the more effectively
an action is transferred to a patient, the more transitive is the clause. Among
the six factors we investigated (see (4)), it is always the former which makes a
clause more transitive and the latter which ‘detransitivises’. Thus our results
show that the transitivity of the construction has been systematically declining.
For this presentation, we also analyse causative bring’s ‘huge neighbours’
– non-causative bring as well as causative make and get – in order to explore the reasons for this detransitivisation process. We hypothesise that,
firstly, in order to avoid ambiguous bring-constructions which can have both a
causative and a locative reading, English simply split the constructional territory between the two constructional variants. Secondly, we hypothesise that
processing-related factors can explain why it was causative bring which took
the beating: Firstly, in detransitivised constructions, we have an asymmetry
between more coding material and less action, i.e. a lower semantic load. This
176
AG 5 · Gebäude B4 1, Raum 0.26
discrepancy between greater formal complexity, or explicitness, and less informativeness, or semantic load, heightens hidden/cognitive complexity. Secondly, usage frequency explains that if one construction had to shoulder this
burden, it had to be causative bring, leaving the less demanding and more efficient options to its much more frequent brother.
References: • Hopper, Paul J. and Sandra A. Thompson (1980): “Transitivity in Grammar and Discourse.” Language 56 (2). 251-99.
On the scope of the form-frequency correspondence principle
Donnerstag
09.03.2017
13:45 – 14:15
B4 1, 0.26
Martin Haspelmath
MPI-SHH Jena & Leipzig University
[email protected]
In earlier work (Haspelmath 2008; Haspelmath et al. 2014), I made a strong
claim:
(1)
The form-frequency correspondence principle in grammar
When two minimally different grammatical patterns (i.e. patterns that
form an opposition) occur with significantly different frequencies, the
less frequent pattern tends to be overtly coded (or coded with more coding material), while the more frequent pattern tends to be zero-coded
(or coded with less coding material).
This is illustrated by a wide range of coding asymmetries: singular/plural, nominative/accusative, affirmative/negative, cardinal/ordinal,
present/future, active/passive, same-subject/different-subject, and many
others (as first identified and catalogued by Greenberg 1966).
In addition, I also claimed that the reverse also holds: all systematic form
asymmetries correspond to frequency asymmetries. However, since cases of
ellipsis also qualify as form asymmetries and they are due to contextual predictability rather than to frequency, the more general explanatory statement
is that form asymmetries are due to predictability, whether contextual or
frequency-based.
In this presentation, am going to discuss a few general issues that arise for
this research programme, using concrete examples: (A) What kinds of factors
can override the preference for short coding, leading to symmetric coding? (B)
177
AG5
AG 5 · Sprachliche Kodierungs-Asymmetrien, Gebrauchsfrequenz und Informativität
Can we treat non-occurrence of a pattern as a special case of coding asymmetry
(e.g. the non-existence of certain relativization patterns)? (C) Given that the
frequency asymmetries are claimed for meanings, rather than specific forms,
is there a way to measure frequency of meanings independently of frequency
of forms? (D) Can we find a general factor underlying frequency asymmetries
(maybe some kind of “cognitive asymmetry”), which might also explain the
form asymmetries? Is there thus a possible alternative to the proposed causal
chain (frequency > predictability > shortness of coding)?
References: • Greenberg, J. H. (1966) Language universals, with special reference to feature hierarchies. The Hague: Mouton. • Haspelmath, M. (2008) Creating economical morphosyntactic patterns
in language change. In J. Good (ed.), Linguistic universals and language change, 185–214. Oxford:
OUP. • Haspelmath, M. et al. (2014) Coding causal-noncausal verb alternations. Journal of Linguistics 50(3). 587–625.
On the optionality of boundary markers (and pro-forms) of
subordinate clauses
Donnerstag
09.03.2017
14:15 – 14:45
B4 1, 0.26
Karsten Schmidtke-Bode
Universität Leipzig
[email protected]
AG5
In the spirit of Hawkins’ (2004) ‘Performance-Grammar-Correspondence
Principle’, this talk sets out to explore usage patterns and grammatical conventions in a typologically understudied domain of clause combining: In many
languages, one can find alternations between overt and null realizations of
morphological material at the boundary of the subordinate clause (as in (1))
or, less commonly, between overt and null realizations of a pro-form of the
subordinate clause in the matrix (as in (2)):
(1)
a.
b.
I know (that) Robert is going to the concert tonight.
S-Pip-waš (hi=)Pal-saP-aktina
3-say-PST DEP=STAT-FUT-come
‘She said (that) she was going to come.’
Barbareño Chumash (Wash 2001: 89)
(2)
He knew (it) that they were not mad at him.
(BNC)
The aim of the talk is to provide a first cross-linguistic survey of these phe-
178
AG 5 · Gebäude B4 1, Raum 0.26
nomena in selected types of complex sentences (notably complement, relative and purpose clauses), and to elucidate recurrent principles that govern
these kinds of differential coding. Among other things, the observed variability turns out to be sensitive to argument sharing across the two clauses, to the
position of the subordinate clause, persistence effects of grammaticalization
and, crucially, to the predictability of the subordinate clause in the context
of specific material in the matrix (e.g. a certain matrix predicate). The latter
factor is fully in line with proposals that speakers exploit their probabilistic
knowledge of the relative surprisal and mutual informativity of co-occurring
linguistic units (cf., e.g., Levy and Jaeger 2007, Jaeger 2010) and that the resulting economical patterns tend to conventionalize into grammatical constraints
(Haspelmath 2008). However, the talk will also explore alternative (or additional) explanations for the occurrence of the less economical variant, such as
more socio-communicatively grounded accounts (e.g. McGregor 2013) and the
effects of analogy and information structure (e.g. Bergh 1997).
References: • Bergh, G. (1997): Vacuous extraposition from object in English. Studia Neophilologica
69, 37–41. • Hawkins, J. A. (2004). Efficiency and Complexity in Grammars. Oxford University Press.
• Haspelmath, M. (2008). Creating economical morphosyntactic patterns in language change. In:
Language Universals and Language Change. Ed. J. Good. Oxford University Press, 185–214. • Jaeger, T. F.
(2010). Redundancy and reduction: Speakers manage information density. Cognitive Psychology 61(1),
23–62. • Levy, R. and T. F. Jaeger (2007). Speakers optimize information density through syntactic
reduction. In: Proceedings of the Twentieth Annual Conference on Neural Information Processing Systems.
Eds. B. Schölkopf, J. C. Platt and T. Hoffman. MIT Press, 849–856. • McGregor, W. B. (2013). Optionality in grammar and language use. Linguistics 51(6), 1147–1204. • Wash, S. (2001). Adverbial Clauses in
Barbareño Chumash Narrative Discourse. PhD dissertation, University of California at Santa Barbara.
Towards functional motivation for the reduced third person indexing
AG5
Ilja A. Seržant
Universität Leipzig
[email protected]
Intro. Personal indexes are sometimes coded asymmetrically. In this case,
more frequently it is the third person that is coded with less material (or
zero) than other person indexes (Bybee 1985: 53; Siewierska 2013). Reduced
third person indexing has been devoted much attention in the literature (Ariel
2000; Siewirska 2010; Bickel et al. 2015). The aim of the paper is to explore
some of the mechanisms that might be responsible for this asymmetry. I exam-
179
Freitag
10.03.2017
11:30 – 12:00
B4 1, 0.26
AG 5 · Sprachliche Kodierungs-Asymmetrien, Gebrauchsfrequenz und Informativität
AG5
ine a small set of frequency data to detect usage asymmetries which correlate
with this coding asymmetry and, hence, might motivate the coding asymmetry.
The diachronic development of pronominal indexes into a full-fledged agreement may be schematically described as proceeding along three stages described by Creissels (2006) for Bantu languages: Stage I (the pronominal markers are in complementary distribution with full NPs), Stage II (pronominal
markers are obligatory even with full NPs but still constitute referring expressions) and Stage III (pronominal markers are obligatory but are no longer referring and need a full NP for a reference = agreement) (Creissels 2006: 44-45).
Data. I use a small collection of texts from Lithuanian (Baltic, IndoEuropean) which – while not having achieved Creissels’ Stage III – exhibits
strong reduction of the third person indexes (diachronically and synchronically). This data shows asymmetry in usage preferences: the 1st & 2nd person
subject indexes occur much more frequently with “pro-drop” than the 3rd person index when referring to a continuous topic. What is more, when referring
to a shifted topic, the 3rd person index with “pro-drop” is extremely rare (3
out of 36 instances of 3rd p. topic shift) while 1st and 2nd person indexes are
as frequent with the “pro-drop” as with independent pronouns.
Discussion. Thus, 1st & 2nd indexes unequivocally pertain to Stage II while
the 3rd person index can be characterized as predominantly Stage III; the 3rd
person is one stage ahead here. The reason for this is that 1st & 2nd are complete expressions with no need of supporting information (such as previous
discourse or a full NP) as regards reference tracking. The range of possible
supporting information is very limited with 1st & 2nd indexes – in fact, only
independent personal pronouns usually can co-occur with the 1st & 2nd person indexes; these are, however, semantically co-referential and may provide
only some emphatic effect or serve as foci (in an argument-focus). In turn, the
potential range of possible meanings and referents of the third person is too
broad to be coherently interpreted and thus is likely to rely on some additional
contextual support: either on the preceding discourse or, crucially, on a full
NP in the same clause. Semantically, the third person index is much weaker
a statement than 1st & 2nd person indexes: it is referentially a variable that is
assigned a referent situationally while the first and second person indexes represent solid, individual reference. This asymmetry is somewhat reminiscent of
the difference between nouns (≈ 3rd person indexes) vs. proper names (≈1st &
2nd indexes): while the former are not lexically referring and often have some
referential support (such as articles, classifiers, modifiers, etc.) the latter are
rather used barely and have lexically unique reference.
Preliminary claim. This functional asymmetry contributes to the fre-
180
AG 5 · Gebäude B4 1, Raum 0.26
quency asymmetry: third person index is expected to co-occur with full NPs
much more frequently than would the 1st and 2nd person indexes at Stage I
and II. Hence, the occurrences in which the third person index repeats the
information already contained in the full NP should be much more frequent
than the occurrences of “redundantly” used 1st and 2nd person indexes. In
turn, the over-use of the 3rd p. index with an overt subject NP in contrast to
the 1st & 2nd p. indexes leads to the “de-emphasizing” and “de-stressing” of
its original meaning (cf. Givón 2001: 421) and, finally, to its redundancy. This
expectation is indeed confirmed by the Lithuanian data. Redundant markers
are more likely to be dropped or shortened than the non-redundant ones due
to economy; hence, the asymmetric coding.
References: • Ariel, M. (2000): The development of person agreement markers: from pronouns to
higher accessibility markers. In Michael Barlow & Suzanne Kemmer (eds.), Usage-based models of
language, 197–260. Stanford: CSLI. • Bickel, B., A. Witzlack-Makarevich, T. Zakharko, G. Iemmolo.
(2015): Exploring diachronic universals of agreement: alignment patterns and zero marking across
person categories. In Fleischer, J., E. Rieken, P. Widmer (eds.) Agreement from a Diachronic Perspective,
29–52. Berlin: De Gruyter Mouton. • Bybee, J. L. (1985): Morphology: A Study of the Relation between
Meaning and Form. Amsterdam/Philadelphia: John Benjamins. • Creissels, D. (2006): A typology of
subject and object markers in African languages. In F. K. Erhard Voeltz (ed.), Studies in African Linguistic Typology, 43-70. Amsterdam: John Benjamins. • Siewierska, A. (2010): Person asymmetries in
zero expression and grammatical function, in: Franck F. (ed.), Essais de de typologie et de linguistique
gé né rale: Mé langes offerts à Denis Creissels. Lyons: ENS ÉDITIONS. 471-485. • Siewierska, A. (2013):
Third person zero of verbal person marking. In The World Atlas of Language Structures Online, M. S.
Dryer and M. Haspelmath (eds). Leipzig: Max Planck Institute for Evolutionary Anthropology. (Available online at http://wals.info/chapter/103, Accessed on 03 10 2014)
A diachronic mechanism for form-frequency asymmetries in
inflectional paradigms
AG5
Helen Sims-Williams
University of Surrey
[email protected]
Cross-linguistically there exists a negative relationship between the relative
frequency of values for inflectional features, and the length of the forms which
realise them – thus singular forms tend to be shorter than plurals, third person
forms tend to be shorter than other persons, and so on (Greenberg 1966). This
paper considers a possible diachronic explanation for this recurring structural
characteristic of inflectional paradigms, starting from the observation that the
following two factors influence morphological change:
181
Freitag
10.03.2017
12:00 – 12:30
B4 1, 0.26
AG 5 · Sprachliche Kodierungs-Asymmetrien, Gebrauchsfrequenz und Informativität
(a) Frequency – lower frequency forms in inflectional paradigms are more
likely to be replaced in morphological change, while higher frequency forms
are more likely to serve as analogical bases for the remodelling of other forms
(Mańczak 1980).
(b) Openness – operations which add phonological material to an analogical
base are more open than those which subtract or replace material, in that they
can be extended to a greater number of lexemes, because they do not require
that a particular sequence exists in the input to be subtracted or replaced. For
example, an alternating pattern like foot – feet can only be extended to words
containing a particular vowel, whereas the pattern of sister – sisters can be extended to any noun (Bybee 1995).
Together, (a) and (b) should promote changes which derive less frequent
members of paradigms by adding phonological material to more frequent
members. In turn, this should concentrate phonological length and morphological complexity in the least frequent paradigm cells over time.
The paper tests the contribution of this diachronic explanation to the synchronic relationship between form length and frequency, using a computational simulation of morphological change.
References: • Bybee, J. (1995). Regular morphology and the lexicon. Language and cognitive processes,
10(5), 425-455. • Greenberg, J. H. (1966). Universals of language. Cambridge: MIT Press. • Mańczak
(1980). Laws of analogy. In Fisiak, J. (Ed.), Historical Morphology (pp. 183-88). Berlin: de Gruyter.
Differential possessive marking of arguments in action
nominalizations: a typological survey
AG5
Freitag
10.03.2017
12:30 – 13:00
B4 1, 0.26
Eva van Lier
University of Amsterdam
Marlou van Rijn
University of Amsterdam
[email protected]
[email protected]
Traditionally, the distribution of alienable versus inalienable possessive
marking over the core arguments of action nominalizations is accounted for
in terms of the semantic factor of ‘control’: The relationship between agentive
(SA and A) arguments coded as possessors and their ‘possessum’ – the event
denoted by the verb – is viewed as controlled, and hence the alienable construction is used. By contrast, patientive arguments (SP and P) are viewed as having
no control over the relationship to their predicate and are therefore encoded as
inalienable possessors (Capell 1949: 172ff; Seiler 1983a: 22, 1983b; Koptjevskaja-
182
AG 5 · Gebäude B4 1, Raum 0.26
Tamm 1993: 210ff; Palmer 2011).
However, early studies on Polynesian languages already show that in many
cases agentive arguments of action nominalizations (SA and/or A) can be encoded as inalienable possessors (Chung 1973; Clark 1981). Moreover, in a pilot study on Central-Eastern Oceanic languages, we found that the effect of
control on the possessive coding of arguments in nominalizations is relative
rather than absolute. It can be described in terms of a hierarchy of argument
types, as given in (1) below: If in a particular language an argument on this
hierarchy may be encoded with an inalienable possessive construction, then
all arguments to its left will either also take inalienable possessive coding or
sentential coding, but not alienable coding.
(1)
P > SP > SA > A
Our account of differential possessive marking in nominalizations implies predictions concerning the issue of coding asymmetries, since it is well-known
that alienable possessive marking is cross-linguistically more formally complex than inalienable possessive marking (Haspelmath 2008 and references
therein). Therefore, in this presentation, we test:
(i) whether the generalization in (1) holds for a world-wide sample of ca.
80 languages, using data from relevant WALS chapters (KoptjevskajaTamm 2013; Nichols & Bickel 2013) and from our own earlier work on
possessive constructions;
(ii) whether the data found in (i) are in line with predictions concerning coding asymmetries in possessive constructions with non-derived possessum nouns.
We evaluate the results in light of the various explanations proposed for possessive coding asymmetries. Our data support an account in terms of iconicity,
since, arguably, P arguments are in a closer conceptual relation with their verbal predicate than A arguments (cf. Croft 2008). In addition, we will draw attention to the role of transitivity – as the above account of iconicity of distance
does not apply to S arguments – and to the fact that many language-specific
patterns do not appear to be straightforwardly compatible with the general explanatory principle sketched above; the latter point suggests that while typological generalizations about coding asymmetries can be drawn, this does not
necessarily mean that individual patterns are shaped by identical functional
factors.
References: • Capell, A. (1949). The concept of ownership in the languages of Australia and the Pacific.
Southwestern Journal of Anthropology 5(3). 169–189. • Chung, S.(1973). The syntax of nominalizations
183
AG5
AG 5 · Sprachliche Kodierungs-Asymmetrien, Gebrauchsfrequenz und Informativität
in Polynesian. Oceanic Linguistics 12. 641–686. • Clark, R. 1981. Inside and outside Polynesian nominalizations. In: J. Hollyman & A. Pawley (eds.), Studies in Pacific languages and cultures in honour of
Bruce Biggs, 65–81. Auckland: Linguistics Society of New Zealand. • Croft, W. (2008). On iconicity of
distance. Cognitive Linguistics 19(1). 49–57. • Haspelmath, M. (2008). Frequency vs. iconicity in the
explanation of grammatical asymmetries, Cognitive Linguistics 19(1). 1–33. • Koptjevskaja-Tamm, M.
(1993). Nominalizations. London: Routledge. • Koptjevskaja-Tamm, M. (2013). Action Nominal Constructions. In: M.S. Dryer & M. Haspelmath (eds.), The world atlas of language structures online. Leipzig:
MPI EVA. • Nichols, J. & B. Bickel (2013). Possessive classification. In M.S. Dryer & M. Haspelmath
(eds.), The world atlas of language structures online. Leipzig: MPI EVA. • Palmer, B. (2011). Subject-indexing and possessive morphology in Northwest Solomonic. Linguistics 49(4). 685–747. • Seiler, H.
(1983a). Possession as an operational dimension of language. Tübingen: Gunter Narr. • Seiler, H. (1983b).
Possessivity, subject and object. Studies in Language 7(1). 89–117.
Freitag
10.03.2017
13:00 – 13:30
B4 1, 0.26
Coding asymmetry between independent and dependent pronominal
possessors: A cross-linguistic study
Jingting Ye
Fudan University
[email protected]
This study carries out a typological survey of independent and dependent
pronominal possessor forms, showing that although there are different types
of coding in different languages, it seems to be a universal tendency that independent pronominal possessor forms are either longer than or identical in
length with dependent forms. This universal tendency is illustrated by the following examples:
(1)
AG5
(2)
wò -làà
wò-dÉ-y
2PL.POSS-village 2PL.POSS-Substantivizer
‘your village’ ‘yours’
o-u
vae o-u
POSS-2SG leg POSS-2SG
‘your leg’ ‘yours’
Zialo (Babaev 2010: 158, 65)
Vaeakau-Taumako
(Næss & Hovdhaugen 2011:329, 392)
According to my investigation of more than 30 languages from different language families, the above-mentioned universal tendency holds true no matter
these pronominal possessor forms are full pronouns or person indexes.
The diachronic source of independent pronominal possessors can shed light
on the coding asymmetry between independent pronominal possessors and
184
AG 5 · Gebäude B4 1, Raum 0.26
dependent ones. The independent pronominal possessor is mostly formed either by adding a substantivizer to the dependent one or by the dependent form
itself. Consequently, the form of independent possessor is either longer or
identical to the dependent one. The reason why the dependent possessor was
not further reduced when expressed independently lies in the fact that the reduced form is not informative enough to show the possessive relation.
The above-mentioned universal tendency can also be explained by the formfrequency correspondence proposed by Haspelmath (2008). Since the usage
of independent pronominal possessors is mostly restricted to certain constructions, for instance comparative constructions (‘Yours is bigger than mine’) and
equative constructions (‘It is yours’), they are less frequent than dependent
pronominal possessors, which can occur as a part of noun phrases in most syntactic positions. This leads to the result that the independent pronominal possessors are never shorter than the more frequent forms of dependent pronominal possessors.
References: • Babaev, Kirill (2010): Zialo: the Newly-Discovered Mande Language of Guinea. München:
Lincom. • Haspelmath, Martin (2008): Frequency vs. iconicity in explaining grammatical asymmetries. In: Cognitive Linguistics 19(1), 1–33. • Næss, Åshild & Hovdhaugen, Even (2011): A Grammar of
Vaeakau-Taumako. Berlin: Mouton de Gruyter.
Markedness Disharmony in Basque
Freitag
10.03.2017
13:30 – 14:00
B4 1, 0.26
Natalia M. Zaika
Institute for Linguistic Studies, Russian Academy of Sciences
[email protected]
The paper is devoted to several cases of markedness disharmony in Basque.
This notion is used to refer to the cases in which some criteria of markedness
contradict other ones, cf. the notion of markedness harmony (“if something is
marked in way A, then it is also marked in way B”) in (Zeevat s.a.: 1), and is
close to the notion of markedness reversal that is «“marked” behaviour of categories that are usually unmarked» under certain circumstances; these cases
are explained by frequency (Haspelmath 2006). In this paper, I investigate the
cases in which formal markedness (zero coding vs. explicit coding) contradicts
statistical markedness, the formally more marked form being more frequent.
The first type of markedness disharmony is due to phonological change.
Zero marking of plural vs. explicit marking of singular in the genitive and
185
AG5
AG 5 · Sprachliche Kodierungs-Asymmetrien, Gebrauchsfrequenz und Informativität
some other cases derived from the genitive, cf. lagun-a-ren friend-SG-GEN
‘friend’s’ vs. lagun-Ø-en friend-PL-GEN ‘friends’’, can be explained as a result
of vowel contraction. Another case of markedness disharmony is the second
person singular present tense form (Ø-haiz /ais/ 2SG-AUX.PRS ‘(thou) art’,
where the zero marker of the second person is opposed to the explicit marker
of the first person, cf. n-aiz 1SG-AUX.PRS ‘I am’) is due to the dropping of the
first consonant. Markedness disharmony due to historical reasons can be observed in Russian (the zero genitive plural for some declension classes); in English (the third singular present marker -s vs. the zero marker of the other
persons and numbers); in Old French (the oblique case plural); in Spanish
(the zero marker of the first person singular imperfective); and in Georgian
(the zero second person singular subject marker). Interestingly enough, in all
the cases of this type, the zero marker is never the only allomorph of its morpheme.
The second type of markedness disharmony is linked to cases where phonological reduction seems never to have taken place, and zero morphemes have
emerged in another way, cf. (Bybee 1994: 240). One such case is the statistically
unmarked (the most frequent) non-finite verbal form, which is the perfect participle, cf. sar-tu enter-PFV ‘entered’ vs. the radical form sar-Ø enter-RAD ‘enter’. As both the participle and the radical are most often used as a constituent
of a finite verb, for the indicative (more frequent) and for the subjunctive (less
frequent) mood correspondingly, in this case, markedness disharmony could
be explained by the slightly reformulated principle of markedness complementarity, introduced by Shapiro, stating that “oppositely marked stems and
desinences attract, identically marked stems and desinences repel” (Shapiro
1983: 146).
Another case of markedness disharmony of the second type is the formally
marked definite singular form of noun phrases (mendi-a mountain-DEF.SG
‘a/the mountain’) vs. the unmarked indefinite form (mendi-Ø mountain-INDF
‘(a) mountain’), which is extremely limited both in terms of frequency and
context. This contrast is not typical for all cases: in locative cases the least
marked form is the definite singular, cf. mendi-ta-ra mountain-INDF-ALL ‘to
a mountain’ vs. mendi-Ø-ra mountain-DEF.SG-ALL ‘to a/the mountain’. Actually, in this case, formally marked and unmarked forms do not represent a
well-formed morphological opposition, as two grammatical categories (definiteness and number) are involved in it.
References: • Bybee, J. (1994): The Grammaticization of Zero. Asymmetries in Tense and Aspect Systems. In: Perspectives on Grammaticalization, W. Pagliuca (ed.). John Benjamins, 235–254.
• Haspelmath, M. (2006): Against Markedness (and what to replace it with). In: Journal of Linguistics 42(1), 25–70. • Shapiro, M. (1983): The Sense of Grammar. Language as Semeiotic. Indiana University Press. • Zeevat, Henk. (s.a.): Markednesses (https://www.uva.nl/binaries/content/documents/
personalpages/z/e/h.w.zeevat/en/tab-one/tab-one/cpitem%5B40%5D/asset?1355372856802).
186
AG6
Prosody in syntactic encoding
Gerrit Kentner & Joost Kremers
Goethe-Universität Frankfurt, Georg-August-Universität Göttingen
[email protected], [email protected]
Raum: B4 1, 0.25
Short description
This workshop focuses on the interplay between syntax and prosody in linguistic encoding, specifically examining the extent to which prosody affects
syntax, and vice versa. In light of the assumption that language production
and perception involves recourse to grammatical knowledge, we especially ask
how the grammar has to be conceptualized to be in a position to explain prosodic/phonological influences on sentence structure.
187
AG 6 · Prosody in syntactic encoding
Comparatives are strongly affected by focus structure
Mittwoch
08.03.2017
14:15 – 14:45
B4 1, 0.25
Katy Carlson
Morehead State University
[email protected]
Comparative constructions as in (1) have many possible syntactic continuations, including bare NPs (a), VP Ellipsis (b), and full clauses (c). This project
explores their processing and use by examining the frequency of different
comparative structures within the Corpus of Contemporary American English (COCA) and conducting experiments on the interpretation preferences
of comparative bare NP ellipsis. The corpus data shows that ellipsis structures are much more frequent than full clauses, with bare NP ellipsis most
frequent (50% of the data). This suggests that the repetition involved in complete clauses, and the deaccenting that would be needed in the prosody, makes
clauses dispreferred compared to structures that retain mostly contrastive information.
(1)
AG6
Tasha called Bella more often than...
{a. Sonya / b. Sonya did / c. Sonya called Bella}.
Interestingly, 80% of bare NPs in the corpus contrast with the subject of the
previous clause, but bare NPs preferentially contrast with the object in processing (as shown in written and auditory questionnaires). Most theories of
comparative ellipsis propose a complete syntactic clause even for bare NP ellipsis (e.g., Lechner 2008), so structural economy should not favor an object
contrast. The frequency of subject NP contrasts in the corpus also fails to explain the processing bias. Since overt contrastive accents on the subject (Tasha)
or object (Bella) do strongly affect the preferred interpretation, we suggest
that the default expectation of focus on the last argument in the first clause
accounts for the object bias in processing. Thus both the syntactic structures
produced and the interpretation of ambiguous examples can be partly tied to
the prosodic structure of comparatives.
188
AG 6 · Gebäude B4 1, Raum 0.25
Ambiguities at the interface: production and comprehension
Mittwoch
08.03.2017
14:45 – 15:15
B4 1, 0.25
Tina Bögel
Universität Konstanz
[email protected]
Over the last few decades, several theories on how syntactic phrasing influences prosodic phrasing have been proposed. However, it is equally assumed
that prosodic phrasing is not only determined by syntactic structure, but
by other modules as well, e.g., information structure. Furthermore, prosodic
phrasing often seems to undergo a postsyntactic ‘rephrasing’ process to meet
certain well-formedness constraints (a.o., Selkirk 1995). The resulting mismatches between syntax and prosody raise the question as to how the influence of syntax on prosody (during production) is in fact reversible (during
comprehension).
The idea of grammar as consisting of different modules with their own principles and parameters has been adopted into several frameworks. This talk will
discuss the interaction between the two modules of syntax and prosody in Lexical Functional Grammar (LFG; Kaplan & Bresnan (1985)) with respect to German case ambiguities. In the following example, the syncretism between the
feminine forms of the dative and the genitive articles leads to an ambiguity in
the subordinate clause’s second DP der Gräfin.
(1) Alle
Everyone
[der
waren
überrascht
was
surprised
Diener]DP1
[der
dass
that
Gräfin]DP2
zuhörte.
the.masc.nom servant
the.fem.gen/dat Countess
listened
‘Everyone was surprised that [the Countess’ servant listened // the servant listened to the Countess].’
Based on experimental evidence, it will be shown that it is indeed crucial to
distinguish between the two processes of comprehension (prosody → syntax)
and production (syntax → prosody) and that the modular framework of LFG
allows for a straightforward modeling of this difference at the syntax–prosody
interface.
References: • Selkirk, E.O. 1995. The prosodic structure of function words. In Beckmann, Dickey
and Urbanczyk (eds.), Papers in Optimality Theory, University of Massachusetts. • Kaplan, R.M. and
J. Bresnan. 1982. Lexical-Functional Grammar: A formal system for grammatical representation. In
Bresnan (ed.) The Mental Representation of Grammatical Relations. Cambridge, MA [a.o.]: MIT Press.
189
AG6
AG 6 · Prosody in syntactic encoding
Prosodic and syntactic structures in spontaneous speech: a
wavelet-based approach to prosodic modelling
Mittwoch
08.03.2017
15:15 – 15:45
B4 1, 0.25
Anna Dannenberg1 , Stefan Werner2 , Martti Vainio1 & Antti Suni1
1
University of Helsinki, 2 University of Eastern Finland
{anna.dannenberg|martti.vainio|antti.suni}@helsinki.fi, [email protected]
AG6
For analysing the prosodic structure of spoken language, especially of spontaneous speech, effective automatic applications have been infrequent. For instance, prosody is often assumed to have some kind of a hierarchical structure,
but prosodic patterns have been hard to observe or visualise directly.
Our solution is a Continuous Wavelet Transform (CWT). The method applies
the weighted sum of f0, energy and segmental durations to represent prosodic
signals in a two-dimensional time-scale plane akin to spectrograms. This can
be further enhanced with lines of maximum amplitude to produce a visual representation of the prosodic hierarchies of speech.
We have used a CWT based tool to analyse the prosodic structure of spontaneous speech. The results have been compared with grammatical analysis
of the same data, achieved both automatically and manually, to examine the
relation of prosody and syntax in spoken language.
Our study of English and Finnish speech data demonstrates a significantly
high rate of co-occurrence between prosodic and syntactic boundaries. This
result proposes a connection between prosody and syntax. The internal hierarchical structures of prosodic and syntactic units, on the contrary, appear to
have practically no resemblance to each other. This may be due to our choice
of syntactic model, based on traditional phrase structure grammar, that is not
well suited for analysing unplanned spontaneous speech.
It thus seems that CWT is a plausible method for modelling prosodic structures of spoken language, but new grammatical approaches are required to
deal with syntactic characteristics of spontaneous speech. We are looking forward to see if CWT based models of prosodic hierarchy can help to find new
perspectives to the grammar of spoken language.
References: • Mallat, S. (1999): A Wavelet Tour of Signal Processing. Academic Press. • Suni, A., Aalto,
D. and Vainio, M. (2015): Hierarchical representation of prosody for statistical speech synthesis.
arXiv:1510.01949. • Dannenberg, A., Werner, S. and Vainio, M. (2016): Prosodic and syntactic structures in spontaneous English speech. In: Proceedings of Speech Prosody 2016, Boston, USA, 59–63.
190
AG 6 · Gebäude B4 1, Raum 0.25
Prosodic boundaries constraints by discursive elements
Mittwoch
08.03.2017
16:30 – 17:00
B4 1, 0.25
Luciana Lucente
University Federal of Alagoas, Brazil
[email protected]
From the second part of the 20th century the conception of speech processing and production has been predominantly conceived under a perspective in
which the prosody is submitted to the syntactic domain. In a dynamic perspective, the syntactic analysis is considered just one of the linguistic elements
that works in parallel. In this work, the dynamic model of speech production
adopted (Barbosa, 2007; Lucente 2012) considers the speech a product of the
dynamic activity. In a nutshell, the refered model assumes that information
from superior linguistic levels are encoded in lexical gestures in parallel with
accentual and syllabic coupled oscillators. These oscillators work on speech
rhythm, and when connected with the gestures’ duration, give as result the
prosody. Following this theoretical perspective, it is proposed here an analysis
of prosodic boundaries and their alignment with discursive breaks, in order to
observe whether syntactic or discourse elements are involved in determining
these boundaries. The discursive segmentation of a spontaneous speech corpus was made according to a computational model (Grosz and Sidner, 1986)
that focuses on attention and intention of the speakers to the conversational
segmentation. Based on the assumptions of speech production model, a Praat
script was developed to detect the prosodic boundaries – finals and intermediates – and intonational prominences. When the discursive segmentation and
the automatic detection of prosodic boundaries are put together is possible to
observe that the boundaries aligned to the discursive breaks are mostly motivated by discoursive elements.
References: • Barbosa, P. A. (2007) From syntax to acoustic duration: a dynamical model of speech
rhythm production. Speech Communication. 49 (1-2), 725-742 • Grosz, B., Sidner, C. (1986) Attention,
intention and structureof discourse. Computational Linguistics, v12(3), 175-204 • Lucente, L. (2012)
Aspectos dinâmicos da fala e da entoação no português brasileiro. Ph.D. Thesis. Campinas.
191
AG6
AG 6 · Prosody in syntactic encoding
How prosody and syntax interact: the case of English it-clefts
Mittwoch
08.03.2017
17:00 – 17:30
B4 1, 0.25
Laetitia Leonarduzzi
Aix Marseille Université
Sophie Herment
Aix Marseille Université
[email protected]
[email protected]
The aim of this presentation is to bring in factual data concerning the
prosody of cleft sentences in English. It is based on 148 it-clefts drawn from
the oral component of the ICE-GB corpus. For each cleft we determined several
features: information-structural features for both the focused (highlighted) element and the presupposed part (discourse status and “informativeness”, i.e.
relevance at this point of discourse); the discourse function of the cleft; the
type of main clause; contrast and emphasis. The prosody of clefts is here analysed through Halliday’s three Ts (Tonality, Tonicity, Tones) (Halliday, 1967).
In some cases at least prosody and syntax tightly combine to yield meaning:
for instance, a cleft produced in one Intonation Phrase with a Nucleus on the
focused element invariably induces contrast, and we claim that this is due to
the interaction between the type of syntactic structure and the prosodic pattern. Here focus and prominence coincide. Tones, when non-neutral, seem on
the contrary to act independently from syntax (though not completely) to add
meaning: Falling-Rising and High Falling tones are used for contrast (whatever the prosodic pattern) and emphasis. This is true also of tonality, whose
role is to indicate the informativeness of each part of the cleft (here we join
Lambrecht, 2001). Finally, tonicity, when non-canonical, may (indirectly) rebound on syntax to modify what we shall call the scope of the presupposition.
So there seems to be evidence for a two-way model of the syntax-phonology interface (Selkirk 1986). Prosody and syntax interact in very complex ways, and
their relationship differs according to which part of prosody is at stake.
AG6
References: • Halliday, M.A.K., (1967): Intonation and Grammar in British English. The Hague-Paris:
Mouton. • Lambrecht, K. (2001): A framework for the analysis of cleft constructions. In: Linguistics 3(3), 463-516. • Selkirk, E. (1986): On derived domains in sentence phonology. In: Phonology 3,
371-405.
192
AG 6 · Gebäude B4 1, Raum 0.25
Prosody as a determinant of the syntactic status of I think
Daniela Kolbe-Hanna
Universität Trier
Judith Manzoni
Université du Luxembourg
[email protected]
[email protected]
Mittwoch
08.03.2017
17:30 – 18:00
B4 1, 0.25
Parenthetical I think as in (1a) is a frequent comment clause in English that is
syntactically subordinate to the main clause (see, e.g. Biber et al. 972-982). In
initial position it fills the typical position of a matrix clause, but recent studies
suggest that – with (1b) or without (1c) a complementizer (that) – it is unclear
syntactically whether I think is a comment clause or a matrix clause (see, e.g.
Dehé and Wichmann (2010b)).
(1)
a.
b.
c.
Lord Scarman, I think, was right.
I think that Lord Scarman was right
I think Lord Scarman was right
(ICE-GB s1b-033)
According to Dehé and Wichmann (2010) in the matrix clause I has prosodic
prominence, whereas prominent think is related to comment clause function.
Thus, when a syntactic matrix clause is reinterpreted as comment clause,
prosody overrides syntactic structure. However, in medial position (as exemplified in (1a), I think remains a comment clause even if I is stressed: In this
case, syntax seems to override prosody. Thus there appears to be a continuing
transfer of information between both prosody and syntax.
In order to examine the interrelation between prosody and syntactic hierarchy in sentences introduced by I think we draw on the Buckeye Corpus (http://
buckeyecorpus.osu.edu/) for auditory and acoustic analyses. For this purpose,
we (i) investigate the correlation between prosody and presence or absence of
that, (ii) describe the intonation patterns of I think and (iii) measure acoustic
movements in their relation to the range of pitch in the phrase.
References: • Dehé, N. and A. Wichmann (2010): Sentence-initial I think (that) and I believe (that):
Prosodic evidence for use as main clause, comment clause and discourse marker. In: Studies in Language 34(1), 36–74.
193
AG6
AG 6 · Prosody in syntactic encoding
Sentence stress in presidential speeches
Donnerstag
09.03.2017
09:00 – 10:00
B4 1, 0.25
Arto Anttila
Stanford University
[email protected]
English phrasal stress is rule-governed, but variable. A sentence like
(1)
the savings of many years in thousands of families are gone
(F.D. Roosevelt, Inaugural Address, March 4, 1933, Sentence 19)
can be performed with different stress contours. Both the regularities and
the variation require an explanation. We explore the view that regularities in
phrasal stress come from stress rules such as the Nuclear Stress Rule and the
Compound Stress Rule that operate on syntactic structures (received wisdom),
while variability partly depends on ambiguity at the level of lexical phonology (new proposal). For example, words like in may or may not be lexically
stressed. Since phrasal stress is a function of lexical stresses and their mode of
combination, variation results.
On the empirical side, we report on our ongoing study of rhythm in presidential speeches. Building on data made available by the American Presidency Project (Peters and Woolley 1999-2017), syntactic analysis by the Stanford Parser (Chen and Manning 2014), automatic metrical analysis by MetricalTree (Dozat 2015), and native speaker stress judgments collected using the
web application MetricGold (Shapiro 2016) we compare the theoretically predicted stress patterns to the actually experienced stress patterns, exploring
the interaction of stress, lexical frequency, and syntactic linearization. In particular, we consider the hypothesis that informative words tend to be placed in
positions where they are highlighted by phrasal stress (Bolinger 1972, Cohen
Priva 2012).
AG6
References: • Bolinger, D. (1972): Accent is predictable (if you‘re a mind reader). Language 48:
633-644. • Chen, D. & Manning, C. D. (2014): A Fast and Accurate Dependency Parser using Neural
Networks. EMNLP 2014. • Cohen Priva, U. (2012): Sign and signal: Deriving linguistic generalizations
from information utility. Ph.D. thesis, Stanford University. • Dozat, T. (2015): MetricalTree. Software
package. Stanford University. • Shapiro, N. (2016): MetricGold. Software package. Stanford University.
194
AG 6 · Gebäude B4 1, Raum 0.25
Clitic placement and syntax-prosody mapping
Donnerstag
09.03.2017
10:00 – 10:30
B4 1, 0.25
Stavros Skopeteas
Universität Bielefeld
[email protected]
A class of enclitics in several Mayan languages occurs at the right edge of a
prosodic constituents of the major phrase level (Aissen 1992; Skopeteas 2009).
The enclitics in the following examples illustrate this phenomenon in Yucatec
Maya. Definite noun phrases are obligatorily enclosed by an enclitic; see (1).
This enclitic does not necessarily appear adjacent to the noun phrase but at
the right edge of the encompassing major phrase; see (2).
(1)
táan u wen-el
le xibpal *(-a’/-o’/-e’)
PROG A.3 sleep-INCMPL DEF man:child-D1/D2/D3
‘The boy (here/ there/ afore mentioned) is sleeping.’
(2)
k-u
xíimbat-ik le h-mèen hun-túul h-k’ìin-o’.
IPFV-A.3 visit-INCMPL DEF M-shaman one-CL.AN M-priest-D2
‘A priest visits the shaman.’
This talk presents evidence from several classes of enclitics in several syntactic configurations (focus vs. topic constructions; restrictive vs. non-restrictive
relative clauses; different types of clausal complements) and reaches the following generalizations: (a) the placement of a subclass of enclitics is determined by the prosodic domains and can be accounted for in terms of mapping
rules between prosody and syntactic linearizations; (b) another class of enclitics are direct exponents of prosodic boundaries; (c) the choice of syntactic
construction in conflict situations implies a mechanism that anticipates the
possible conflicts that will arise after the completion of the syntax-to-prosody
mapping.
References: • Aissen, J. (1992): Topic and Focus in Mayan. Language 68.1, 43–80. • Skopeteas, S.
(2009): Syntax-phonology interface and clitic placement in Mayan languages. In Torrens et al. (eds.),
Movement and Clitics. Newcastle: Cambridge Scholars Publishing.
195
AG6
AG 6 · Prosody in syntactic encoding
Prosody and tag question forms in Glasgow Scots
Donnerstag
09.03.2017
11:15 – 11:45
B4 1, 0.25
Elyse Jamieson
University of Edinburgh
[email protected]
Wiltschko and Heim (to appear) discuss the role of prosody with regard to English confirmational particles like right and eh. The authors argue rising intonation is syntacticized in a “Call on Addressee” (CoA) position (Beyssade
and Marandin, 2006) in the Grounding domain. The particles sit in lower
Ground projections, and combine with the intonation to give the “confirmational” meaning.
(1)
[CoA / [GroundA eh [GroundS [CP ….]]]]
Confirmational particles and tag questions seemingly carry out similar functions; however, English tag questions do not always have rising intonation (e.g.
Ladd 1981).
Here, I will explore data from Glasgow Scots, first noted in Thoms et al.
(2013). Glasgow Scots has a particle, -int, used only in tag questions and exclamatives.
(2)
AG6
They wur leavin, wint they?
they were leaving, weren’t they
‘They were leaving, weren’t they?’
(Thoms et al. 2013)
I will present the results of a grammaticality judgment experiment testing two
hypotheses: firstly, that –int is only available in “confirmation” tag questions
(Ladd 1981). Secondly, that although it looks like negation, –int is a CoA marker,
in complementary distribution with rising intonation. If this does turn out to
be the case, it will lend support to Wiltschko and Heim’s claim that prosody can
be syntacticized; it will also prove interesting for the relationship between tag
questions and confirmation particles.
References: • Beyssade, C. and Marandin, J-M. (2006) ”The speech act assignment problem revisited:
Disentangling speaker’s commitment from speaker’s call on addressee” In Selected Papers of CSSP 2005
37–68 • Ladd, R. (1981) ”A first look at the semantics and pragmatics of negative questions and tag
questions” CLS 17 164–173. • Thoms, G., Adger, D., Heycock, C. and Smith, J. (2013) ”Remarks on negation in varieties of Scots” Handout, Cambridge workshop on English dialects. • Wiltschko, M. and
Heim, J. (to appear) The syntax of confirmationals: A neo-performative analysis.
196
AG 6 · Gebäude B4 1, Raum 0.25
Syntactic integration of sentential intonation
Donnerstag
09.03.2017
11:45 – 12:15
B4 1, 0.25
Johannes Heim
University of British Columbia
[email protected]
This paper proposes a reorganization of the left periphery that situates force
in a discourse-oriented, complex projection above CP to include prosodic information in the derivation of speech acts. This is necessary to arrive at different
speech act (SA) interpretations among constructions of identical clause type
and sentential intonation exemplified below:
(1)
a.
b.
c.
It is raining. (↘) [falling declarative, SA: assertion]
It is raining? (↗) [rising declarative, SA: question]
It is raining. (↗) [uptalk, SA: assertion]
The declarative sentences in (1) show that intonation does not map onto a
unique speech act, and speech acts do not map onto specific intonational contours. Only a combination of structural and prosodic properties can derive the
different interpretations. I propose to expand the ‘syntactiza-tion of discourse’
(Speas & Tenny 2003) to include prosodic information in a discourse-related
projection replacing the traditional ForceP. Force and an Intonation occur in
the specifier positions of a complex speech act phrase with a speaker- and an
addressee-oriented projection (Lam 2014).
SaP
Intonation
Sa’
Hearer
saP
Force
AG6
sa’
Speaker
CP
References: • Lam, Z. (2014). A Complex ForceP for Speaker- and Addressee-oriented Discourse Particles in Cantonese. SCL 35: 61-80. • Speas, M. & Tenny, C. (2003). Configurational properties of
point of view roles. AiG 1, 315-345.
197
AG 6 · Prosody in syntactic encoding
The ordering of syntax-prosody-interpretation mapping rules
Donnerstag
09.03.2017
12:15 – 12:45
B4 1, 0.25
Marta Wierzba
Universität Potsdam
[email protected]
Sentence stress on the object is compatible with a broad focus reading in
German, in both canonical verb-final structures and derived object-initial V2
clauses. This parallel can be modeled by assuming that mapping rules determining the relation between syntax, prosody, and interpretation can apply to
traces/deleted copies (as suggested by Selkirk 1995 for English and by Korth
2014 for German). The use of object-initial sentences is restricted, however:
the subject needs to be discourse-given, as shown in (2).
AG6
(1)
What’s happening?
Ich denke, dass [Peter ein Buch liest]foc .
I think that Peter a book reads
‘I think that Peter is reading a book.’
(2)
a.
b.
Why is the teacher surprised? — #Ein Buch liest [Peter]new .
Why is Peter’s teacher surprised? — Ein Buch liest [Peter]given .
I will present evidence for the patterns in (1)-(2) stemming from acceptability
rating experiments as well as related data for contrastive topic (CTs). The data
suggest that for the purpose of the mapping between prosody and the focus/CT,
access to the full syntactic structure including previous stages of the derivation is necessary. The mapping between prosody and givenness, on the other
hand, seems to interact with the arguably ‘late’ process of postnuclear compression, which does not affect the metrical structure but merely the phonetic
realization of pitch accents (cf. Kügler & Féry 2016). Taken together, these considerations point towards an architecture of grammar with ‘early’ and ‘late’ interface mapping rules (preceding/following the transformation to a phonetic
signal).
focus / CT mapping
(accessing traces)
→
postnuclear
compression
→
givenness mapping
(surface-oriented)
References: • Korth, M. (2014): Von der Syntax zur Prosodie. • Kügler, F. & C. Féry. (2016): Postfocal Downstep in German. • Selkirk, E. O. (1984): Phonology and Syntax: The Relation between Sound and
Structure. • Selkirk, E. O. (1995): Sentence Prosody: Intonation, Stress, and Phrasing.
198
AG 6 · Gebäude B4 1, Raum 0.25
Towards a non-centralized, subtractive architecture of grammar
Volker Struckmeier
University of Cologne
Donnerstag
09.03.2017
13:45 – 14:15
B4 1, 0.25
[email protected]
Established architectures of grammar often give syntax precedence over
”the interfaces”. I argue that scrambling in German warrants the assumption
that syntax as well as the semantic and prosodic interfaces express independent restrictions, with no precedence of one system over the others. I propose
a subtractive architecture of grammar which can represent stacked, independent restrictions without precedence relations.
In many (e.g., cartographic) syntactic analyses, information structural (e.g.
topic or antifocus) heads trigger scrambling movements. Proposals of this
kind, I argue, are highly problematic: Theoretically, the stipulation of such a
trigger is a circular device which yields no insights as to why the head causes
movement in a given language (cf. Struckmeier, to appear), why the head takes
its specific position, or why the category constitutes a syntactic head in the language at all. Empirically, our experimental findings show that subjects a) judge
sentences significantly differently under acoustic or written presentations, b)
are often unable to intuit semantic (!) triggers for scrambling under written
presentations, and c) often judge word order changes by their semantic or
prosodic effects – and not by syntactic positions crossed (directly contradicting cartographic tenets).
To represent these findings, a subtractive grammar combines a) syntactic
contraints on formally possible structures, b) semantic restrictions on interpretable structures, and c) prosodic factors that assess prosodic constellations.
Since these constraints are applied independently, we find mismatches: Semantic constraints can enforce bad prosody, and prosodic factors may cause
semantic intransparency, I show. De-centralized models, unlike centralized
ones, can accomodate and explain these findings without circularity, I claim.
References: • Struckmeier, V. to appear: ”Against information structure heads: A relational analysis
of German scrambling”, Glossa.
199
AG6
AG 6 · Prosody in syntactic encoding
Syntax and prosody in parallel systems
Donnerstag
09.03.2017
14:15 – 14:45
B4 1, 0.25
Manuela Korth
Johannes Gutenberg-Universität Mainz
[email protected]
The talk is intended to show how a modular grammar with parallel architecture can deal with mismatches and reciprocal influences at the syntaxprosody interface. The first part of the talk looks at mismatches at the DP level
in structures like (1). It will be argued that phonological conditions (based on
stress) as well as interface conditions (based on syntactic relationship) lead to
the indicated prosodic phrasing. Such structures can be made more isomorphic by speakers through an adjustment of the syntactic and the informationstructural encoding, which results in a different prosodic structure.
(1)
a.
b.
[dass [[der
JÄger des
KÖnigs] starb]
that the.NOM hunter the.GEN king
died
[dass der JÄger] [des KÖnigs starb]
syntactic phrasing
prosodic phrasing
The second part of the talk concentrates on pre- and postnominal modifying
genitives like (2) and their alternatives. The syntactic case feature on the modifier in (2) is not realized phonologically because of a degemination process.
Thus speakers avoid such phrases and use prenominal datives or postnominal
PPs instead.
(2)
a.
b.
AG6
Jonas’
Buch
Jonas.GEN book
das
Buch Jonas’
the.NOM book Jonas.GEN
Phonological reduction like degemination in (2) or erosion as part of grammaticalization processes can lead to a loss of grammatical encoding, because
phonology does not care about the realization of syntactic features. So phonology takes an indirect influence on syntax here. In a modular grammar with
parallel architecture, the interface component can function as a control system, which links the outcomes of the individual modules, evaluates them and
informs the syntactic component about a non-optimal structure, so that an alternative can be chosen instead.
200
AG 6 · Gebäude B4 1, Raum 0.25
Prosodic constraint on prenominal modification
Hisao Tokizaki
Sapporo University
Jiro Inaba
Tokyo University
[email protected]
[email protected]
Freitag
10.03.2017
11:30 – 12:00
B4 1, 0.25
It has been argued that the head of a prenominal modifier must be adjacent
to the head noun (Grosu & Horvath 2006, Haider 2010).
(1)
a. a [baby [sleeping [on the sofa]]]
b. *a [[sleeping [on the sofa]] baby]
(English)
(2)
a.
(German)
ein [[[ in Saarbrücken] wohnhafter] Professor]
a
in Saarbrücken living
professor
b. *ein [[wohnhafter [in Saarbrücken]] Professor]
However, the head-to-head adjacency condition cannot be assumed in the minimalist framework with no linear order in syntactic derivation. Moreover, the
adjacency condition wrongly rules out head-initial phrasal compounds in English and German and head-initial modifiers in Russian.
(3)
a.
b.
[[over-[the-fence]] gossip]
der [’Fit-[statt-fett]’]-Bürowettbewerb
‘the fit-over-fat office contest’
(4)
[[gotovyi [na vse]]
student]
ready
on everything student
‘a student ready for anything’
(English)
(German)
(Russian)
Instead of the adjacency condition, we propose a prosodic constraint to the
effect that the modifier and the noun it modifies cannot be separated by a
prosodic boundary. In (1a) and (2a) the noun phrase has no prosodic boundary
between the modifier and the head it modifies because of its uni-directional
branching structure (Tokizaki 1999). We argue that a phrasal compound may
not have a prosodic boundary at its right edge and that the proposed constraint may be relaxed in languages without determiners, such as Russian (cf.
Bošković 2008).
References: • Bošković, Ž. (2008): What will you have, DP or NP? NELS 37, 101–114. • Grosu, A. &
J. Horvath (2006) Reply to Bhatt and Pancheva’s “Late Merger of Degree Clauses”: The Irrelevance
of (Non)conservativity. Linguistic Inquiry 37, 457–483. • Haider, H. (2010) The syntax of German. CUP.
• Tokizaki, H. (1999) Prosodic Phrasing and Bare Phrase Structure. NELS 29, 381-395.
201
AG6
AG 6 · Prosody in syntactic encoding
Freitag
10.03.2017
12:00 – 12:30
B4 1, 0.25
Prosody determines word order: the case of Mainland Scandinavian
object shift
Nomi Erteschik-Shir
Ben-Gurion Univ.
[email protected]
Gunlög Josefsson
Lund University
[email protected]
Björn Köhnlein
Ohio State University
[email protected]
In this paper we argue for an analysis of Object Shift (OS) in Mainland Scandinavian (MSc) in which the a weak object pronoun is placed to the left of an
adverb instead of in the canonical position for objects to the right of the sentence adverb. We argue following Bennet et al.’s proposal that elements may
move in the phonology in cases of prosodic repair that OS applies as a prosodic
repair when the weak pronoun is left unincorporated when preceded by an
adverb.
Whether or not OS is obligatory varies among the MSc languages and varieties. For instance, OS is obligatory in Standard Danish but optional in a number of southern Danish dialects, for example the dialect spoken on the island
of Ærø. In Swedish OS is optional in most dialects. We observe that varieties
with optional OS also have a tone accent contrast. We argue that the in-situ
word order is licensed in these dialects because tonal accent creates a higher
prosodic domain licensing the incorporation of the weak pronoun in situ.
Syntactic accounts of OS are problematic: There is no obvious way of linking
the occurrence OS to Verb-movement upon which it is dependent. OS has no
semantic or even information structural motivation nor is there any obvious
syntactic motivation and there is no way to syntactically constrain optionality
of OS by making reference to language/dialect specific prosodic properties.
Our analysis opens up an exciting area of research examining which cases
of movement belong in Syntax and which do not.
AG6
References: • Bennett, R., Elfner, E., and McCloskey, J. (2016): Lightest to the Right: An Apparently
Anomalous Displacement in Irish. Linguistic Inquiry 47:169-234. • Myrberg, S., and Riad, T. (2015):
The prosodic hierarchy of Swedish. Nordic Journal of Linguistics 38:115-147. • Selkirk, E. (2011): The
syntax-phonology interface. In The Handbook of Phonological Theory. Blackwell.
202
AG 6 · Gebäude B4 1, Raum 0.25
Urdu/Hindi polar kya at the syntax-pragmatics-prosody interface
Farhat Jabeen
Universität Konstanz
Miriam Butt
Universität Konstanz
[email protected]
[email protected]
This talk focuses on Urdu/Hindi polar kya ‘what’ from the perspective of
Lexical-Functional Grammar (LFG). Polar questions in Urdu/Hindi take the
syntactic form of declaratives. Intonation distinguishes between a declarative
and an interrogative. Polar questions can also be overtly optionally marked via
kya ‘what’ at different positions in a sentence. Bhatt and Dayal (2014) suggest
that these different possible positions result from topicalization, so that all the
material to the left of kya is given and thus not open for questioning.
This analysis is, however, not supported by our data. We conducted an experiment via an on-line acceptability judgement questionnaire and found that
regardless of the position of polar kya either the subject or the object were available for questioning/correction and hence not necessarily given/presupposed.
We also conducted a follow up intonational study and found that if a given constituent was prosodically focused then that constituent was also available to
be corrected. We were able to determine that Bhatt and Dayal’s (2014) analysis
only pertains to wide focus context, i.e., neither the subject nor the object are
focused.
Our data show that prosodic information overlays the syntactic information.
Prosody interacts with syntax, but does not derive from it (see also ShattuckHufnagel and Turk 1996). We show how this can be modelled elegantly within
the modular, mutually constraining architecture of LFG. We also propose an
alternative analysis of polar kya that sees it as an operator which takes the
questioned proposition ?p (Biezma and Rawlins 2012) of a polar question and
renders it as a modal statement of possibility. This analysis is motivated by the
results of an additional corpus study of Bollywood scripts, which shows that
the polar kya appears in contexts expressing questions about whether a certain action or situation is possible.
In conclusion, Urdu/Hindi polar kya presents a complex interaction across
the prosody-syntax-pragmatics interfaces. Attempting to analyze its effects
without integrating an understanding of the intonational properties of the language will yield only a partial understanding of the patterns.
203
Freitag
10.03.2017
12:30 – 13:00
B4 1, 0.25
AG6
AG 6 · Prosody in syntactic encoding
Uttered sentences, prosody and word order
Freitag
10.03.2017
13:00 – 13:30
B4 1, 0.25
Emmanuel Schang
Université d’Orlé ans
emmanuel.schang@univ-
François Nemo
Université d’Orlé ans
Fanny Krimou
Université d’Orlé ans
[email protected] [email protected]
orleans.fr
orleans.fr
Consider the following examples:
AG6
(1)
Tu veux arrêter.
(4)
Veux-tu arrêter !
(2)
Tu veux arrêter ?
(5)
Tu veux arrêter !
(3)
Veux-tu arrêter ?
(6)
Veux-tu !
A long tradition describes the Subject-Verb-Inversion in (3) as an illustration of verb movement (to C) triggered by the interrogative force. But French
also allows for interrogatives with the standard order, as in (2), with prosody
being then the only factor differenciating it from (1) and that (4) to (6) also
trigger an imperative/exclamative reading, with (4) contrasting with (3) only
through prosodic marking, (4) contrasting with both (1) and (2), and (6) having an elliptical syntax and the indexical interpretation “Will you stop!”. To
sum up, prosody alone can tease apart declarative/imperative/interrogative
sentences. Our aim will be to question the way semantic interpretation, wordorder and prosodic contours are associated one with another. Based on an ongoing corpus study, we shall discuss: i) the fact that prosody challenges the
classical semantic distinction between sentences and utterances; ii) the existence of stable prosodic contours shared by various clause-types; iii) the fact
that the diversity of prosodic realizations among “imperative imperative sentences” appears to match more fine-grained levels than classical illocutionary
acts such as advising, allowing, requesting, ordering, challenging, encouraging, etc. ; iv) the theoretical consequences for syntax and the modeling of the
relation between syntactic, phonetic and semantic form.
References: • Elordieta Gorka & Pilar Prieto (eds.) (2012) Prosody and Meaning. Berlin: De Gruyter
Mouton. • Lacheret, A., & Victorri, B. (2002). La période intonative comme unité d’analyse pour
l’étude du français parlé: modélisation prosodique et enjeux linguistiques. In Verbum (Vol. 1, No. 24,
pp. 55-72). • Rupp, L. (2002). The syntax of imperatives in English and Germanic: Word order variation in
the Minimalist framework. Springer.
204
AG 6 · Gebäude B4 1, Raum 0.25
The position of où : Where’s the difference?
Freitag
10.03.2017
13:30 – 14:00
B4 1, 0.25
Janina Reinhardt
Universität Konstanz
[email protected]
Morphosyntactic variation in French interrogatives has long been the focus
of investigation (see e.g. Pohl 1965, Coveney 1995, Boucher 2010). However, significantly less research has been done on the mapping between morphosyntactic variants and intonational forms.
In this paper, I will investigate the prosody of ‘where’-questions in four
French audio books. In contrast to Delattre 1966 and in line with DelaisRoussarie 2015, about half of the questions were not realised with a simple fall
(L*L%). Indeed, one third were even realised with a simple rise (H*H%). This
evidence casts doubt on the assumption of a falling default pattern for constituent questions. Sentences (1) and (2) were produced with falling as well as
with rising intonation, leaving no space to explain variation by the fronting or
non-fronting of the question word either (cf. Déprez et al. 2012).
(1)
Tu étais où ?
(2)
Où va-t-on ?
A closer look at the context suggests that the falling patterns may be due to a
secondary meaning of reproach, or the mere expression of a thought or resignation. If intonational patterns can be mapped, it is thus to pragmatic meaning
rather than to word order.
References: • Coveney, A. (1995): The use of the QU- final interrogative structure in spoken French.
JFLS 5(2), 143–171. • Delais-Roussarie, E. et al. (2015): Intonational Phonology of French: Developing & Prieto, P., Intonation in Romance, 63–100. • Delattre, P. (1966): Les Dix Intonations de base du
français. The French Review 40(1), 1–14. • Déprez, V. et al. (2012): The interaction of syntax, prosody,
and discourse in licensing French wh-in-situ questions. Lingua 124, 4–19. • Pohl, J. (1965): Observations sur les formes d’interrogation dans la langue parlée et dans la langue écrite non littéraire.
In: Actes du Xe Congrè s 1965, 501–516. • Audio sources: Alex (Camille Verhœven 2) (written by P.
Lemaitre, read by P. Résimont) • Le temps est assassin (M. Bussi, J. Basecqz) • Si c’était à refaire
(M. Lévy, M. Marchese) • Total Khéops (J.-C. Izzo, A. M. Mancels).
205
AG7
Wen wurmt der Ohrwurm? – An interdisciplinary, cross-lingual
perspective on the role of constituents in multi-word
expressions
Sabine Schulte im Walde & Eva Smolka
Universität Stuttgart, Universität Konstanz
[email protected], [email protected]
Raum: B4 1, 0.24
Short description
The processing and representation of multi-word expressions, ranging from
noun compounds (e.g. „hogwash“ in English, and „Ohrwurm“ in German) to
particle verbs (e.g. „give up“ in English, and „aufgeben“ in German) has remained an unsettled issue. In this workshop, we aim to shed light on the interaction of constituent properties and compound transparency across languages and disciplines integrating linguistic, psycholinguistic, corpus-based
and computational studies.
207
AG 7 · Wen wurmt der Ohrwurm?
Mittwoch
08.03.2017
13:45 – 14:45
B4 1, 0.24
Morphological superposition and the nature of the mental lexicon
Gary Libben
Brock University
[email protected]
In recent years, research on the mental lexicon and on lexical processing has
been undergoing a paradigm shift. As a result of new methodologies and new
approaches to statistical analysis, we are able to adopt a psychocentric approach to research on lexical and morphological knowledge and to view the
mental lexicon as a cognitive system that is characterized by dynamicity.
I will discuss new behavioural methods for the investigation of complex
and compound word processing. These include primed progressive demasking, typing analysis, binaural presentation, and word recognition of real-time
written production. I present data that have used these techniques to study
the processing of compound words and will discuss the extent to which they
support a conceptualization of morphological representation that I term “morphological superposition”. The terminological metaphor of morphological superposition is derived from a construct of early 20th century quantum physics.
I argue that this metaphor can provide a framework with which it is possible
to capture the manner in which morphological structure in lexical processing
is dependent on particular task demands.
Mittwoch
08.03.2017
14:45 – 15:15
B4 1, 0.24
AG7
N1-accessibility as a matter of compound processing
Stefanie Rößler
University of Göttingen
Thomas Weskott
University of Göttingen
Anke Holler
University of Göttingen
[email protected]
[email protected]
[email protected]
The first constituent in noun-noun compounds (N1) appears to be unaccessible for pronominal anaphora (e.g. Postal 1969). However, experimental findings have challenged this constraint. We present three psycholinguistic experiments with German compounds suggesting that distinct factors can contribute to render the N1 accessible for anaphora. The investigated factors are:
208
AG 7 · Gebäude B4 1, Raum 0.24
A - animacy of the N1 (animate vs. inanimate, e.g. dog bowl vs. plastic bowl);
B - semantic relation (have vs. for, e.g. can stock vs. can opener); C - spatiotemporal contiguity (stc) (+stc vs. -stc, e.g. car accident vs. car insurance);
D - compound structure (root vs. synthetic compounds, e.g. roof garden
vs. roof greening)
Exp. 1 (sentence completion) tested for the factors A and B. Participants completed sentence fragments starting with a pronoun with the N1 as antecedent.
Our prediction that in the conditions ’animate’ and ’have’ there will be more
N1-references was borne out; in addition, we found an interaction of both factors. Exp. 2 tested for the factors C and D in the same paradigm as Exp. 1. We
predicted that N1-references increase in the conditions ’+stc’ and ’synthetic
compound’. While the prediction for the main effect of factor C was borne out,
the effect of factor D ran against our prediction. Further we conducted an eyetracking during reading experiment focusing on the apparently subtle factor
D (synthetic vs. root vs. monolexeme). This time the prediction concerning factor D was borne out.
Our data highlight the interplay of different factors that have to be integrated by processing models. The interactions provide us with a better understanding of how this integration might work, and that the effect of subtle structural factors might be buried beneath world-knowledge factors. Finally, we
want to discuss how our results can be brought in line with theoretical frameworks (e.g. Marantz 1997).
References: • Marantz, M. (1997): No escape from syntax: Don’t try morphological analysis in the
privacy of your own lexicon. In: University of Pennsylvania Working Papers in Linguistics 4(2), 201–225.
• Postal, P. M. (1969): Anaphoric Islands. In: Binnick, R. et al. (eds.): Papers from the Fifth Regional
Meeting of the Chicago Linguistic Society. Chicago, April 18–19, 1969. Chicago: University of Chicago,
205–239.
Factors affecting the processing of compounds in the L2
Serkan Uygun
Yeditepe University
Ayşe Gürel
Boğaziçi University
[email protected]
[email protected]
Mittwoch
08.03.2017
15:15 – 15:45
B4 1, 0.24
Compound processing has a particular place in the psycholinguistic literature since it contributes to our understanding of the mental representation/processing of multimorphemic words and allows us to examine the role
209
AG7
AG 7 · Wen wurmt der Ohrwurm?
of factors such as constituency, frequency, and semantic transparency in processing complex words. Previous studies involving different languages have
revealed the role of semantic transparency and headedness in compound processing (Jarema et al., 1999; Libben et al., 2003). In second language (L2) acquisition, the effects of semantic transparency and headedness are found to vary
on the basis of L2 proficiency (Wang, 2010).
The present study investigates the processing of nominal compounds in
L2 Turkish, a language with right-headed and productive compounding. In a
masked priming experiment, 35 advanced, 36 intermediate level learners of
Turkish with L1 English and 73 Turkish monolinguals were tested. The stimuli consisted of 10 transparent-transparent, 10 partially-opaque compounds,
10 pseudocompounds and 60 monomorphemic words together with 90 nonwords matched on length and frequency.
A 2x3x3 Mixed ANOVA for the RTs revealed that for monolinguals, semantic transparency plays a role in constituent activation as both constituents
are accessed in partially-opaque compounds and only the head is accessed in
transparent-transparent compounds. L2 participants, however, exhibited no
priming effects irrespective of their proficiency level, implying that neither
headedness nor transparency plays a role in processing L2 compounds in late
learners.
References: • Jarema, G., Busson, C., Nikolova, R., & Libben, G. (1999): Processing compounds: A
cross-linguistic study. Brain and Language 68, 362–369. • Libben, G., Gibson, M., Yoon, Y.B., & Sandra,
D. (2003): Compound fracture: The role of semantic transparency and morphological headedness.
Brain and Language 84, 50–64. • Wang, M. (2010): Bilingual compound processing: The effects of
constituent frequency and semantic transparency. Writing Systems Research 2(2), 117–137.
Mittwoch
08.03.2017
16:30 – 17:00
B4 1, 0.24
AG7
Exploring the impact of transparency and productivity of multiword
term constituents on single-word term identification
Anna Hätty
Robert Bosch GmbH
Michael Dorna
Robert Bosch GmbH
[email protected]
[email protected]
Terms are expressions that characterize specialized domains. They comprise
both single- (SWT) and multiword terms (MWT). For the scoring of multiword expressions as terms their components are often taken into account (e.g.
Zhang, 2012), and transparency plays a role for translation and synonym ex-
210
AG 7 · Gebäude B4 1, Raum 0.24
traction of MWTs. Our goal is to exploit the relation of MWTs to their constituents in order to identify SWTs among the constituents. SWTs are often
less specific than MWTs and it is harder to score them for termhood. We hypothesize that SWTs which are frequently found in diverse complex terms and
which contribute to the meaning of the MWTs are more likely to be terms as
well. Therefore we approach this problem by taking the constituents of multiword terms as term candidates. We then investigate the influence of their
productivity and transparency within these MWTs on the prediction of termhood. We use the ACL RD-TEC (Zadeh and Handschuh, 2014), a corpus for
the evaluation of term extraction in the field of Computational Linguistics. We
address transparency with a vector space model by computing the similarity
of compound and constituent vectors. We show that transparency variance
for highly productive heads influences their prediction as terms. We investigate the interplay of productivity, transparency, variance of transparency
in constituent families and frequency by training a classification model with
those features. Finally, we compare this approach with the modified C-value
for SWTs by Barrón-Cedeño et al. (2009).
References: • Barrón-Cedeño, A., Sierra, G., Drouin, P., & Ananiadou, S (2009). An Improved Automatic Term Recognition Method for Spanish. CICLing ’09 • Zadeh, B. Q., & Handschuh, S. (2014).
The ACL RD-TEC: a dataset for benchmarking terminology extraction and classification in computational linguistics. CompuTerm ’14 • Zhang, C., Niu, Z., Jiang, P., & Fu, H. (2012). Domain-specific
term extraction from free texts. FSKD ’12, 1290-1293, IEEE.
Multi-word units in a discriminative framework
Saskia E. Lensink
Leiden University
R. Harald Baayen
Universität Tübingen
[email protected]
[email protected]
Mittwoch
08.03.2017
17:00 – 17:30
B4 1, 0.24
Phrasal frequency effects (Bannard & Matthews, 2008) have often been taken
as evidence for the existence of representations for multi-word units. However, Baayen et al. (2013) were able to simulate phrasal frequency effects using
a naive discriminative learning (NDL) network that does not rely on representations for multi-word units, by connecting letter digraphs to the constituent
words of the multi-word units.
Even though it may sometimes be possible to model phrasal frequency effects without relying on representations of multi-word units, it is not clear
211
AG7
AG 7 · Wen wurmt der Ohrwurm?
whether this is the most optimal architecture to use. Therefore, we built an
NDL network where a layer of input cues, consisting of the constituent words
of a set of 300 semantically transparent trigrams, and a layer of outcomes, the
trigrams themselves, are connected.
The validity of this NDL architecture was tested against two experimental
data sets where stimuli consisted of these 300 trigrams. Statistical analyzes
showed that the NDL measures performed to comparable levels as traditional
frequency measures in explaining the empirical data. This testifies to the usefulness of an NDL architecture with full-form representations for semantically
transparent multi-word units, and with their constituent words used as learning cues.
One problem noted by Baayen et al. (2013) is that it is unclear how multiword units can be distinguished from each other. Our discriminative network
provides a first tentative solution to this question. We will also argue that a discriminatory perspective clarifies why multi-word units have to be short (with
a most five words).
References: • Baayen, R. H., Hendrix, P., & Ramscar, M. (2013). Sidestepping the combinatorial explosion: An explanation of n-gram frequency effects based on naive discriminative learning. Language and Speech, 56(3), 329-347. • Bannard, C., & Matthews, D. (2008). Stored word sequences in
language learning the effect of familiarity on children’s repetition of four-word combinations. Psychological science, 19(3), 241-248.
Mittwoch
08.03.2017
17:30 – 18:00
B4 1, 0.24
AG7
Semantic entropy measures and the semantic transparency of noun
noun compounds
Melanie J. Bell
Anglia Ruskin University
Martin Schäfer
Universität Jena
[email protected]
[email protected]
Recently, two different entropy measures based on the relational structure of
English compounds have been used in studies of semantic transparency and
lexical decision times. Pham & Baayen (2013) show that the entropy of semantic relations in the modifier family is negatively correlated with semantic transparency. Schmidtke et al. (2015) find that the relational entropy for
individual compounds is correlated with lexical decision time. However, neither study takes the ambiguity of the compound constituents into account. Our
contribution addresses this gap. In a model of semantic transparency, we show
212
AG 7 · Gebäude B4 1, Raum 0.24
that the relational entropy of the head family interacts with a second entropy
measure, namely synset entropy: uncertainty about the reading of a given constituent in terms of homonymy and polysemy.
We used the publically available compound dataset from Bell & Schäfer
(2016), which is annotated with semantic relation and WordNet synset (sense)
of the constituents. We calculated the entropy of the probability distributions
of the synsets and semantic relations for all modifier and head constituent families and modelled the semantic transparency ratings collected by Reddy et al.
(2011). Our final model of compound transparency shows an interaction between the two entropy measures based on the head constituent families. When
the synset entropy is low, perceived transparency is unaffected by relational
entropy; when the synset entropy is high, perceived transparency is negatively
correlated with relational entropy, mirroring the finding by Pham & Baayen
(2013) for modifier families.
These findings suggest that relation entropy is not negatively correlated
with semantic transparency across the board. If the reading of the head is
easily predictable (low synset entropy), it makes little difference to perceived
transparency whether the relation entropy is high or low. But with increasing synset entropy, the effect of relation entropy becomes increasingly pronounced. When the reading of the head is uncertain, the compound appears
more transparent if the relation is easily predictable across all possible senses
than if it is not. One way of interpreting this result is that in cases of high synset
entropy, low relation entropy masks constituent ambiguity, while high relation entropy necessitates greater activation of different readings to arrive at
even a ‘gist’ interpretation, thereby increasing processing effort and lowering
perceived transparency.
References: • Bell, M. J. & Schäfer, M. (2016): Modelling semantic transparency. Morphology 26(2),
157–199. • Pham, H. & Baayen, R. H. (2013). Semantic relations and compound transparency: A regression study in CARIN theory. Psihologija 46 (4), 455–478. • Reddy, S., McCarthy, D. & Manandhar,
S. (2011). An Empirical Study on Compositionality in Compound Nouns . In: Proceedings of the 5th
IJCNLP 2011. • Schmidtke, D., Kuperman, V., Gagné, C.L. & Spalding, Th. L. (2016). Competition between conceptual relations affects compound recognition: the role of entropy. Psychon Bull Rev 23(2),
556–570.
AG7
213
AG 7 · Wen wurmt der Ohrwurm?
Donnerstag
09.03.2017
09:00 – 09:30
B4 1, 0.24
AG7
The role of the head in the interpretation of deverbal compounds
Gianina Iordăchioaia
Universität Stuttgart
Lonneke van der Plas
University of Malta
Glorianna Jagfeld
Universität Stuttgart
gianina
lonneke.vanderplas
glorianna.jagfeld
@ifla.uni-stuttgart.de
@um.edu.mt
@gmail.com
We present an interdisciplinary study on the correlation between the transparency of NN deverbal compounds (DCs; e.g., task assignment) and the ambiguity of their deverbal heads as predicted by their morphosyntax. Theoretical hypotheses are tested with computational tools and resources. We start
with Grimshaw’s (1990) observation that deverbal nouns are ambiguous between Argument-Structure-Nominal (ASN) readings, which inherit verbal arguments (e.g., the assignment of difficult problems), and the less verbal and more
lexicalized Result Nominal readings (RNs: cf. two-page assignment). Following Grimshaw, our hypothesis is that the presence of ASN properties in the
head triggers a direct object interpretation on the non-head of a DC (i.e., the
verb’s lowest argument). If the head is a RN, it should allow a broader range
of context-dependent semantic relations just like primary compounds such as
chocolate box (cf. Borer 2013).
We selected a varied yet controlled set of DCs from the Gigaword corpus
(Napoles et al. 2012) for which we ran an annotation effort with three native
speakers. To determine the ASN-hood of DC heads we constructed 7 indicative
patterns inspired by Grimshaw’s ASN properties and collected evidence from
Gigaword. To test our hypotheses we ran a MaxEnt classifier.
We identified two properties of the deverbal heads with high predictive
power in the interpretation of DCs: a head’s predilection for DC-contexts and
its frequent realization of internal arguments outside DCs (i.e., its predilection for the ASN-reading). Both predict an object interpretation of the nonhead, which, especially in the case of the latter property, confirms Grimshaw
(1990)’s claim and our hypothesis that DCs have some event structure hosting
internal arguments. These experiments are a first attempt to identify the patterns that underlie the interpretation of deverbal compounds. Further work
is necessary to determine the interdependence of the individual features and
the contribution of the remaining features.
References: • Borer, H. 2013. Taking Form. Oxford: OUP. • Grimshaw, J. 1990. Argument Structure.
Cambridge, MA: MIT Press. • Napoles, C., M. Gormley, and B. Van Durme. 2012. Annotated Gigaword.
Proceedings of AKBC-WEKEX 2012.
214
AG 7 · Gebäude B4 1, Raum 0.24
Semantic transparency and variation in nominal syntagmatic
compounds in Romance languages
Donnerstag
09.03.2017
09:30 – 10:00
B4 1, 0.24
Inga Hennecke
Universität Tübingen
[email protected]
The scientific debate about syntagmatic compounds of the type N+Prep+N is
characterized by terminological insecurity and very heterogeneous attempts
of classification. The status of multi-word units such as sp. bicicleta de montaña,
fr. livre pour enfants or port. histó ria em quadrinhos is discussed controversially,
as they lie at the interface of lexicon and syntax. In fact, in Romance languages,
syntagmatic compounds appear to be very productive and they may vary in
their degree of lexicalization, idiomaticity and fixedness.
The present talk aims to concentrate on the role and semantic transparency
of the prepositional element in syntagmatic compounds. It has been argued,
that prepositions in syntagmatic compounds have lost their initial meaning
and are now fully opaque and lexicalized items. In a first part, I will analyze
the alternation and variation of the prepositional element. This includes the
omission of the constituent as in fr. stylo à bille - stylo-bille and sp. tren de mercancías - tren mercancías, as well as the alternation between different prepositional elements, such as sp. esmalte de/para uñas or fr. flûte de/à champagne. In a
second part, I will present and discuss a quantitative and qualitative analysis of
the role of the prepositional element in word-formation processes by means of
the TenTen Corpora of French, Spanish and Portuguese. The results may shed
light on the role of the prepositional element in syntagmatic compounds, its
transparency and its interchangeability.
References: • Booij, G. (2010): Construction morphology. Oxford: Oxford University Press. • Bisetto,
A.; Scalise, S. (2005): The classification of compounds. In: Lingue e Linguaggio, IV(2), 319–332. • Guevara, E. R. (2012): Spanish compounds. In: W. L. Wetzels (Hg.). On Romance Compounds. Berlin: De
Gruyter, 175-195. • Villoing, F. (2012): French compounds. In: W. L. Wetzels (Hg.). On Romance Compounds. Berlin: De Gruyter, 29-60.
AG7
215
AG 7 · Wen wurmt der Ohrwurm?
Donnerstag
09.03.2017
10:00 – 10:30
B4 1, 0.24
The different meanings of ‘a’: Capturing qualia relations of Italian
complex nominals with distributional semantics
Sandro Pezzelle
University of Trento
Elisabetta Ježek
University of Pavia
Maria Silvia Micheli
University of Pavia
[email protected]
[email protected]
[email protected]
This paper examines the semantic role of the preposition ‘a’ in NaN Italian
complex nominals using a distributional semantic approach. Starting from the
assumption that ‘a’ may introduce one the following qualia relations (Pustejovsky, 1995) - Formal (F, introducing taxonomic information, such as shape
in cacciavite a stella ‘star screwdriver’), Constitutive (C, introducing information on parts, as in codice a barre ‘barcode’), Telic (T, introducing information
on purpose and function, as in barca a vela ‘sailboat’) - we verified whether the
difference in the semantic contribution of ‘a’ in T (‘a-telic’), F (‘a-formal’) or
C (‘a-constitutive’) NaNs is confirmed by a semantic analysis performed using
vector models. We generated meaning representations for each preposition ‘a’
using a distributional semantic approach. First, we extracted all NaNs with
frequency > 5 from the 1.7B tokens itWaC corpus (Baroni et al., 2009). Then,
two of the authors annotated them with F, C, or T according to the scheme in
Bouillon et al. (2012). In total, annotators agreed on 66 NaNs (19 C, 21 F, and 26
T). Finally, we generated meaning representations for both NaNs and single
Ns by training a word2vec model by Mikolov et al. (2013) on the whole corpus.
Meaning representations for each preposition ‘a’ were obtained by subtracting the vector resulting from the sum of the nouns (e.g. barca+vela) from the
NaN vector barca_a_vela). The resulting vectors were then used for running
a cluster analysis. With 3 clusters, ‘a-telic’ clustered together (78%), with ‘aformal’ forming a relatively defined cluster (52%) and ‘a-constitutive’ being
almost equally distributed among the clusters. With 2 clusters, the distinction
turns out to be much clearer, with ‘a-telic’ items (76% in cluster 1) clearly distinguished from the ‘a-non-telic’ (83% in cluster 2). Interestingly, all the ‘anon-telic’ clustered with ‘a-telic’ are constitutive.
AG7
References: • Baroni M. et al. (2009). The WaCky wide web: a collection of very large linguistically
processed web-crawled corpora. • Bouillon P. et al. (2012). Annotating qualia relations in Italian and
French complex nominals. • Mikolov T. et al. (2013). Efficient estimation of word representations in
vector space. • Pustejovsky J. (1995) The Generative Lexicon. Cambridge, MIT Press.
216
AG 7 · Gebäude B4 1, Raum 0.24
Approximating compound compositionality
based on word alignments
Donnerstag
09.03.2017
11:15 – 11:45
B4 1, 0.24
Fabienne Cap
Uppsala University
[email protected]
Introduction We approximate the compositionality of German noun-noun
compounds using statistical word alignments, based on (Villada Moirón and
Tiedemann, 2016). Our hypothesis is that compositional constructions are
translated similarly by human translators, whereas non-compositional constructions exhibit more variance. When training a statistical word alignment
this greater variance leads to a large number of different alignments, which
we use to determine the compositionality of a construction.
Experimental Setup We split all noun-noun compounds occuring in the German Europarl corpus (Koehn, 2005) and then run statistical word alignment
on the English and the modified German corpus. We then calculate the translational entropy (TE) score (Villada Moirón and Tiedemann, 2016) and sort the
compounds in descending order so that compounds with the greatest likelihood of being non-compositional appear at the top of the list. First, the TEscores of both components are weighted equally, but different weightings are
investigated. More lists are produced, sorted according to the TE-score of either modifiers or heads.
Results In Figure 1(a) we show some examples from our lists with the modifier Auge, which show that TE scores correlate well with compositionality. 1(b)
illustrates the greater variance in the non-compositional Augenzwinkern compared to Augenschäden. Figure 1:
Compound
TE
Word
Auge|Maß
3.428
Auge
Auge|Höhe
2.236
(Zwinkern)
Auge|Zwinkern
1.748
Auge
Auge|Schäden
0.637
(Schäden)
(a) TE scores
Alignments
=
nod (2), cheek (1), a (1), glint (1),
blind eye (1), personalise (1)
=
eye (3)
AG7
(b) Word alignments for Auge
217
AG 7 · Wen wurmt der Ohrwurm?
References: • Koehn, P. (2005): Europarl: a parallel corpus for statistical machine translation. In
Proceedings of the MT Summit. • Villada Moirón, B. and Tiedemann, J. (2006): Identifying idiomatic
Expressions using automatic word alignment. In: Proceedings of the EACL 2006 MWE Workshop.
Donnerstag
09.03.2017
11:45 – 12:15
B4 1, 0.24
AG7
Exploring idiomaticity with variant-based distributional measures
and Shannon’s entropy
Marco S. G. Senaldi
SNS, Pisa
Gianluca E. Lebani
University of Pisa
Alessandro Lenci
University of Pisa
[email protected]
[email protected]
[email protected]
The goal of this research is to investigate whether we can take advantage of the
syntactic and lexical fixedness of idiomatic expressions to devise corpus-based
indices of idiomaticity and compositionality and whether these measures can
actually predict human ratings of idiom syntactic flexibility.
First of all we describe a method for automatically distinguishing potential
idioms from only literal combinations via compositionality indices that leverage the greater lexical rigidity of idioms. Starting from two sets of idiomatic
and literal Italian verbal constructions and adjective-noun pairs, we generated
a series of lexical variants out of them, replacing their constituents with semantically related words. We then represented both the original targets and
their variants as vectors in a distributional space and calculated cosine similarity between a given target and its variants, expecting idiomatic vectors to
result less similar to the vectors of their variants with respect to the literal expression vectors. All in all, this proved to be the case, showing that focusing on
the limited exchangeability of the constituents is an effective way to compute
the idiomaticity degree of a given word combination.
In the second part of our study, participants to a CrowdFlower questionnaire
gave 1-7 acceptability scores to sentences containing Italian verbal idiomatic
and literal combinations in different syntactic variants. We then modeled the
human ratings with a hierarchical regression analysis via corpus-based measures computed for the same idioms. These included all the aforementioned
compositionality indices and other formal flexibility measures which used
Shannon’s Entropy to calculate the idiom variability with regard to various parameters, such as the constituents morphology, the presence and type of determiners, etc. Promising results in this regression analysis support the cognitive
plausibility of our computational indices to explain the way speakers process
idioms.
218
AG 7 · Gebäude B4 1, Raum 0.24
Evaluating semantic composition of German compounds
Corina Dima
Universität Tübingen
Jianqiang Ma
Universität Tübingen
Erhard Hinrichs
Universität Tübingen
[email protected]
[email protected]
[email protected]
Donnerstag
09.03.2017
12:15 – 12:45
B4 1, 0.24
Evaluation of composition models. Creating meaningful, reusable representations for multi-word expressions remains an open problem for distributional semantics. This work quantitatively evaluates composition functions
that can construct a composite representation for German compounds e.g.
Apfelbaum ‘apple tree’ from the representations of their constituents, e.g. Apfel
and Baum (see Dima, 2015). The composite representation of a compound
should ideally be indistinguishable from its observed representation. The observed representations of both the compounds and their constituent words are
built using the GloVe method and a 10 billion token raw-text corpus. We use
the rank metric to evaluate 12 composition functions on a frequency-filtered
subset of the compounds available in GermaNet 9.0. The representation of the
head (model 1) was used as a strong baseline, which was slightly outperformed
by the weighted vector addition (model 6). The Wmask model (model 12) produced the best results.
Transparency and composition. Given that the meaning of nontransparent compounds cannot be inferred from that of their constituents,
how does transparency affect the performance of composition functions?
To answer this question, we compared the human judgments for compound
transparency (Schulte im Walde et al., 2013) with the composition results of
model 12 for a subset of the compounds. We found that for less transparent
compounds, the composition yielded lower quality representations than for
more transparent ones. While this is to be expected, the composition functions
also struggled with transparent compounds whose constituents have either
a metaphoric meaning (e.g. Schneemann ‘snowman’) or multiple senses (e.g.
Kaffeemühle ‘coffee grinder’). As future work, we will address these issues by
(1) identifying opaque compounds and building their representations directly
and (2) using sense-aware word representations.
References: • Dima, C. (2015): Reverse-engineering Language: A Study on the Semantic Compositionality of German Compounds. In: Proceedings of EMNLP, 17–21. • Schulte im Walde, S., Müller, S.,
& Roller, S. (2013) Exploring Vector Space Models to Predict the Compositionality of German Noun-Noun Compounds. In: Proceedings of *SEM, 255–265.
219
AG7
AG 7 · Wen wurmt der Ohrwurm?
Understanding compound words: a new perspective from
compositional systems in distributional semantics
Donnerstag
09.03.2017
13:45 – 14:45
B4 1, 0.24
Marco Marelli
Ghent University
[email protected]
In the present work I discuss CAOSS (Compounding as Abstract Processes in
Semantic Space), a model that aims at capturing the semantic dynamics of
compound processing in a data-driven framework.
In CAOSS, word meanings are represented as vectors encoding lexical cooccurrences in a reference corpus (e.g., the meaning of snow will be based on
how often snow appears with other words), according to the tenets of distributional semantics (e.g., Landauer & Dumais, 1997). A combinatorial procedure
is induced following Guevara (2010): given two vectors (constituent words)
u and v, their composed representation (the compound) can be computed as
c = M ∗ u + H ∗ v, where M and H are weight matrices estimated from corpus
examples. The matrices are trained using least squares regression, having the
vectors of the constituents as independent words (car and wash, rail and way)
as inputs and the vectors of example compounds (carwash, railway) as outputs,
so that the similarity between M ∗ u + H ∗ v and c is maximized. In other words,
the matrices are defined in order to recreate the compound examples as accurately as possible. Once the two weight matrices are estimated, they can be
applied to any word pair in order to obtain a meaning representation for their
combination.
CAOSS is shown to correctly predict effects related to the processing of novel
compounds, and in particular the impact of relational information. Moreover,
model predictions are useful for the comprehension of the role of semantic
transparency in the processing of familiar compounds. Taken together, the
model simulations indicate that a compositional perspective on compoundword meaning is crucial for understading the processing of both novel and familiar combinations.
References: • Landauer, T. K., & Dumais, S. T. (1997). A solution to Plato’s problem: The latent semantic analysis theory of acquisition, induction, and representation of knowledge. Psychological Review,
104(2), 211. • Guevara, E. (2010). A regression model of adjective-noun compositionality in distributional semantics. In Proceedings of the 2010 Workshop on GEometrical Models of Natural Language Semantics (pp. 33-37). Association for Computational Linguistics.
220
AG8
Non-canonical verb positioning in main clauses
Mailin Antomo & Sonja Müller
Georg-August-Universität Göttingen, Universität Wuppertal
[email protected], [email protected]
Raum: B4 1, 0.23
Short description
Not only dependent clauses display verb-order variation, there are also different options for positioning the finite verb in main clauses. This workshop intends to study the formal and interpretative properties of main clauses which
do not display the word order which is canonically expected of them across
different languages as well as different historical stages within one language.
221
AG 8 · Non-canonical verb positioning in main clauses
AG8
Verb position and speech acts in German
Mittwoch
08.03.2017
14:15 – 15:15
B4 1, 0.23
Hubert Truckenbrodt
ZAS and HU Berlin
[email protected]
The talk discusses developments on unembedded German V-final clauses and
postulated semantic correlates of V-to-C movement in German.
An analysis of unembedded V-final clauses (e.g. Ob es regnet?, Wenn ich das
wüsste!) is motivated, which builds on Oppenrieder (1989) and Truckenbrodt
(2013a,b): The left periphery here specifies kinds of embedded clauses (e.g.
interrogative CPQ , dass-/wenn-sentence), but no speech act potential. A series of restrictions on these sentence types can be derived from the hypothesis that the content of unembedded V-final clauses is interpreted anaphorically/factive. The speech act proper seems to be constituted by the intonational
and gestural accompaniments that these clauses specifically require. In this
analysis, intonation and gestures take a propositional argument, the content
of the sentence.
The discussion of V-in-C-clauses is based on joint work with Frank Sode and
includes development of Lohnstein (2000), Truckenbrodt (2006a,b), and Sode
(2014). Using the root phenomenon of V1-parentheticals (Steinbach 2007), a
representation of root-clauses at the syntax-semantics interface is first hypothesized. In this representation, root clauses carry a semantically interpreted perspective index in C and V1-parentheticals employ this index as a
docking site. In addition, the index seems to carry a semantically interpreted
feature that interacts directly with verbal mood (Sode 2014) and it seems to
interact with V-to-C movement in V2-clauses. These hypotheses represent a
partial understanding of the role of V-to-C in the speech acts of V2-clauses.
References: • Lohnstein, H. (2000): Satzmodus kompositionell. Akademie-V. • Oppenrieder, W.
(1989): Selbständige Verb-letzt-Sätze, in Zur Intonation von Modus und Fokus im Deutschen, H. Altmann
(ed.). • Sode, F. (2014): Zur Semantik und Pragmatik des Konjunktivs der Indirektheit im Deutschen, Doktorarbeit, HU Berlin. • Steinbach, M. (2007) Integrated parentheticals as assertional complements,
in Parentheticals, N. Dehé & Y. Kavalova (eds.) • Truckenbrodt, H. (2006a,b) in Theoretical Linguistics
32. • Truckenbrodt, H. (2013a,b) in Satztypen des Deutschen, J. Maibauer, M. Steinbach & H. Altmann
(eds.).
222
AG 8 · Gebäude B4 1, Raum 0.23
AG8
Only just CP? Rethinking classification criteria for sentence types
theories
Volker Struckmeier
Universität zu Köln
Sebastian Kaiser
Formerly Universität zu Köln
[email protected]
[email protected]
We argue that sentence type theories must not be defined too narrowly syntactic, but must take into account lexical material outside the left periphery (for
German: modal particles) and prosodic markers.
V-to-C movement in German is often labeled as main clause formation. Thus,
V-last orders may be mislabelled as (in-) subordinated (e.g. Evans 2007, Lohnstein 2000): V-last exclamatives, e.g., are not (in-) subordinated, and neither
are deliberative V-last questions. Minimal changes in prosody or modal particle use, however, can lead to subordination with syntactically near-identical
structures. V1 exclamatives and conditionals are not questions, despite their
CP syntax – but changes in prosody can bring about the interrogative interpretation.
We argue that empty prefields denote that the clauses’ truth values are not
evaluated vis-a-vis a world under discussion (cf., similarly, Reis & Wöllstein
2010). Boundary tones (L% vs. H%) differentiate truth value assessments. Accent tones (H* vs. L*) signal whether the proposition is to be added to the CG
or not (Truckenbrodt 2013, Kaiser 2014). Exclamative accents (e.g. L+H*) distinguish clause types and so do modal particles (Struckmeier 2014), we show.
Multi-factor sentence type theories of this type will help avoid the miscategorizations of purely syntactic approaches, we hope.
References: • Evans, N. (2007): “Insubordination and its uses.” In: Finiteness, 366-431. • Kaiser,
S. (2014): Interpretation selbständiger Sätze im Diskurs. • Lohnstein, H. (2000): Satzmodus – kompositionell. • Reis, M. & A. Wöllstein (2010): Zur Grammatik (vor allem) konditionaler V1-Gefüge im
Deutschen. ZS 29, 111-179. • Struckmeier, V. (2014): “Ja doch wohl C?” Studia Linguistica 68, 16-48.
• Truckenbrodt, H. (2013): “Satztyp, Prosodie und Intonation”. In: Satztypen des Deutschen, 570-601.
223
Mittwoch
08.03.2017
15:15 – 15:45
B4 1, 0.23
AG 8 · Non-canonical verb positioning in main clauses
AG8
Variable verb positions in German exclamatives
Mittwoch
08.03.2017
16:30 – 17:00
B4 1, 0.23
Imke Driemel
Universität Leipzig
[email protected]
German exclamatives come in all sorts of syntactic forms, including variable
verb positions, see (1a) and (1b).
(1)
a.
b.
Wen die alles kennt!
who she all knows
‘How many people she knows!’
Wen kennt die alles!
who knows she all
‘How many people she knows!’
Prior accounts attribute T-to-C movement in German to the ability either to
carry assertional force (Gärtner 2002), to carry unambiguous sentence force
(Schwabe 2007), or to carry sentence force at all (Lohnstein 2000). Since (1a)
and (1b) clearly come with exclamative sentence force, none of the accounts
seem to be transferrable to exclamatives. Neither can the verb positions be
related to an at-issue/non-at-issue distinction (Antomo 2015) since exclamatives are commonly taken to be factive (D’Avis 2002, Abels 2010). Thus, I suggest an analysis along the lines of a speech act encoding syntax (Haegeman and
Hill 2013) that takes the addressee of an exclamative into account: in V-final
wh-exclamatives the speaker merely wants to express his surprise whereas in
V2 wh-exclamatives the speaker wants the addressee to be surprised as well.
Following Truckenbrodt (2004, 2006), I assume that an addressee feature on
exclamative C is responsible for T-to-C movement in German V2-exclamatives.
This addressee requirement can be implemented as a presupposition – granted
that felicity conditions such as the preparatory and sincerity condition (Searle
1969) can be encoded as presuppositions on speech acts that have to be fulfilled in order for them to be successful (see also Roguska 2008). The account
provides an explanation for the apparent “optionality” of verb positions since
there are hardly any specific contexts in which the speaker either wants or
does not want the addressee to be surprised.
References: • Antomo, M. (2015): Abhängige Sätze in einem fragebasierten Diskursmodell. PhD thesis 10, 31–36. • Gärtner, H.M. (1996): On the force of V2 declaratives. ThL 28. • Truckenbrodt, H. (2006):
On the semantic motivation of syntactic verb movement to C in German.ThL 32.
224
AG 8 · Gebäude B4 1, Raum 0.23
AG8
Verb-final w-clauses in headlines
Mittwoch
08.03.2017
17:00 – 17:30
B4 1, 0.23
Rita Finkbeiner
Johannes Gutenberg-Universität Mainz
[email protected]
Genre is frequently referred to as a potential licensing condition for certain
sentence types, such as V1-declaratives in jokes. A less well-known case are
verb-final w-clauses in headlines, cf. (1).
(1)
Wo Bayerns Millionäre wohnen
where Bavaria’s millionaires live
‘Where Bavaria’s millionaires live’
While (1) allows for embedding under a class of functionally equivalent verbs
(e.g. lesen/erfahren ‘read’/‘get to know’), it is not elliptical, but is used as independent utterance with its own illocutionary potential (cf. Oppenrieder 1989).
However, this sentence type does not seem to fit into any of the three established classes of non-embedded verb-final w-clauses in German, as given in
Truckenbrodt (2013).
One of the crucial features of (1) is that it is restricted to headlines. Its illocution may best be described as a special kind of assertion, by which the writer
gives notice of a longer text to follow. Therefore, one may consider it being a
candidate of the controversial class of verb-final declaratives (cf. Pafel 2016).
In this talk, I will investigate in more detail the formal and interpretative
properties of verb-final w-headlines. Their properties will be tested against
properties of related sentence types, such as interrogatives and declaratives,
in order to shed more light on the question where to locate w-headlines in a system of sentence (and utterance) types. The focus will be on the question what
role genre plays as a licensing condition for verb-final position in w-headlines.
References: • Oppenrieder, W. (1989): Selbständige Verb-Letzt-Sätze. In: Altmann, H. et al. (eds.):
Zur Intonation von Modus und Fokus im Deutschen. Tübingen, 163-244. • Pafel, J. (2016): Satztyp
und kommunikative Intention. In: Finkbeiner, R./Meibauer, J. (eds.): Satztypen und Konstruktionen.
Berlin, 406-432. • Truckenbrodt, H. (2013): Selbständige Verbletzt-Sätze. In: Meibauer, J. et al. (eds.):
Satztypen des Deutschen. Berlin, 232-246.
225
AG 8 · Non-canonical verb positioning in main clauses
AG8
Mittwoch
08.03.2017
17:30 – 18:00
B4 1, 0.23
Non-canonic verb positioning in disintegrated verb-final weil-clauses
in German
Nathalie Staraschek
Bergische Universität Wuppertal
[email protected]
(1)
Weil Paul nie den Rasen mäht.
because Paul never the lawn mows
‘because Paul never maws the lawn’
The talk deals with the necessary conditions for verb-final weil-clauses (disWVL) in German, which are neither integrated in respect to their syntax nor
their information-structure and carry an assertional force (cf.(1)). One aspect
is, that these phrases may only encode a causal relation to a non-presupposed
proposition, which seems to be due to the fact that their disintegration results
in a complete interpretation of the antecedent before their own meaning is deciphered1 . Within contexts as described there are two ways for these causal
clauses to be informative, the proposition of their antecedent not being presupposed. They can relate to just a causal relation or encode an additional nonpresupposed proposition, therefor being not yet or already integrated into the
common ground (CG), the latter not being true for the disintegrated V2-variants.
Possible questions to be dealt with: Are these phrases purely syntactic variations? If not, is the idea that the propositions of disWVL are not put on top of
the metaphorical table (cf. Bruce/Farkas (2010)), but listed as publicly committed to by the speaker, adequate? Should this be true, would this and if how,
be reflected in terms of integration of the non-presupposed proposition into
future CG-states?
References: • Antomo, M./Steinbach, M. (2010), Desintegration und Interpretation: Weil-V2-Sätze
an der Schnittstelle zwischen Syntax, Semantik und Pragmatik, in: Zeitschrift für Sprachwissenschaft
29, 1-37 • Bruce, K.B./Farkas, D.(2010), On Reacting to Assertions and Polar Questions, Journal of
Semantics 27, 81-118
1
The same condition holds for weil-verb-second clauses. This constitutes an essential contrast to
integrated causal clauses (iWVL), which may encode and assert solely the causal relation of two
presupposed propositions. This difference is due to their different syntactic and pragmatic status,
which force iWVL into being interpreted in a conjoined discourse update with their antecedent.
About the notion, that disintegrated causal clauses are processed separately from the antecedent
cf. Antomo/Steinbach (2010:28).
226
AG 8 · Gebäude B4 1, Raum 0.23
AG8
On the role of doch in V1- and Wo-VE-clauses in German
Donnerstag
09.03.2017
09:00 – 09:30
B4 1, 0.23
Sonja Müller
Bergische Universität Wuppertal
[email protected]
Particular characteristics and meanings/uses have been attributed to V1- and
Wo-VE-sentences (cf. Önnerfors 1997, Pasch 1999, Günthner 2002 e.g.): They
are claimed to receive a causal/concessive interpretation. V1-clauses are said
to be interpreted causally, Wo-VE-clauses concessively. The modal particle
doch is supposed to be obligatory or at least very typical for these sentence
types. This fact has been related to the assumption that both sentences presuppose their contents. However, at the same time, one component of the meaning
usually ascribed to doch (contradiction, adversativity) cannot be made out. For
that reason, the question is still unsolved why doch favours this environment
so strongly.
By referring to corpus data and acceptability judgments, I will question such
characteristics. In particular, I will argue against the presupposed status of the
utterances’ contents and that this is why doch occurs so often in this context.
The strict association of V1 & a causal interpretation and Wo-VE & a concessive reading turns out to be too strong an assumption. With Önnerfors (1997),
I will claim that doch is indirectly responsible for the causality by assuming
that a causal default interpretation is decisive. In contrast to his analysis, my
modelling of doch provides an explanation for doch facilitating (even if not coding) the causal reading. Based on the contribution I attribute to the particle, its
meaning can be assumed to be transparently present. Above that, it also allows
to derive certain stylistic effects (expressivity, emotional involvement) which
other authors have vaguely referred to, but which has not been spelled out so
far.
References: • Günthner, S. (2002): Zum kausalen und konzessiven Gebrauch des Konnektors wo im
gesprochenen Umgangsdeutsch. ZGL 30/3, 310–341. • Önnerfors, O. (1997): Verb-erst-Deklarativsätze:
Grammatik und Pragmatik. Almqvist & Wiksell International. • Pasch, R. (1999): Der subordinierende
Konnektor wo: kausal und konzessive? In: Freudenberg-Findeisen, R. (ed.), Ausdrucksgrammatik vs.
Inhalts-grammatik Linguistische und didaktische Aspekte der Grammatik, München: IUDICIUM Verlag,
139–154.
227
AG 8 · Non-canonical verb positioning in main clauses
AG8
V1-declaratives and assertion
Donnerstag
09.03.2017
09:30 – 10:00
B4 1, 0.23
Janina Beutler
Bergische Universität Wuppertal
[email protected]
In my presentation I will argue that verb first declarative sentences do not
carry assertion (in the same way) as verb second declaratives do, which
presents itself when taking selected phenomena into consideration:
(1)
A: Hat Fritz den Hund gefüttert?
B: *Ja, hat Fritz den Hund gefüttert.
B: Ja, Fritz hat den Hund gefüttert.
question-answer-pairs
(2)
A: Erna hat ihren Ranzen noch nicht gepackt.
B: Doch, HAT Erna ihren Ranzen gepackt.
B: Doch, Erna HAT ihren Ranzen gepackt.
(3)
a. ?Kommt ja ein Mann in die Bar.
b. Ein Mann kommt ja in die Bar.
verum-focus
assertive modal particles
Beyond the misguiding terminology of the “declarative clause”, it will be
brought forward that there is no evidence, concerning German, of an ASSERT
operator localised through a syntactic force projection, being responsible for
identifying a sentence as an assertion.
References: • Gärtner, H.-M. (2002): On the Force of V2 Declaratives. In: Theoretical Linguistics. 28.1, 33-42. • Lohnstein, H. (2015): V2 als Bindungs-Phänomen. Internationaler
Workshop zur Verbzweit-Stellung, Bergische Universität Wuppertal, 24.-26.Juli 2015. Url:
http://www.linguistik.uni-wuppertal.de/v2- workshop-2015/
handouts/lohnstein.pdf. • Önnerfors, O. (1997): Verb-erst-Deklarativsätze – Grammatik und Pragmatik.
Lund: Almqvist. • Reis, M. (2000): Anmerkungen zu Verb-erst-Satz-Typen im Deutschen. In: R.
Thieroff et al. (eds.): Deutsche Grammatik in Theorie und Praxis. Tübingen: Niemeyer, 215-227.
228
AG 8 · Gebäude B4 1, Raum 0.23
AG8
V-final root clauses in Early New High German?
Donnerstag
09.03.2017
10:00 – 10:30
B4 1, 0.23
Ulrike Demske
Universität Potsdam
[email protected]
The structural asymmetry between V2 main clauses and V-final subclauses is
well attested for all historical stages of German. The rise and loss of V-final
root clauses throughout the period of Early New High German (1350 – 1650)
is hence unexpected from a diachronic point of view. Maurer (1926) was the
first to point out the word order pattern in question, followed by Behaghel
(1932), Lötscher (2000) and others. They agree that the emergence of V-final
root clauses in Early New High German is due to Latin influence. DemskeNeumann (1990) and Senyuk (2014) on the other hand argue that examples
like (1) are instances of subclauses, considering the position of the finite verb
as an unambiguous marker for the syntactic dependency of the clauses at issue.
(1)
Unnd do sy also bey ein ander in allermengklichs abwesens warend,
empfieng herczog Wilhalmen Agleyen. Darauff sy czů im sprach:
In the present talk, I will consider V-final clauses introduced by a demonstrative, either by a pronominal adverb as in (1) or by a demonstrative DP as in
(2), suggesting that they are syntactically dependent but pragmatically independent of their matrix clause. The variation between V2 and V-final clauses
introduced by a demonstrative is accounted for in terms of different discourse
functions (foreground vs. background).
(2)
Do sprachen die freydigen zů der klaren junckfrawen, das sy sich mit
e
kurczen worten erklagte, wie sy wolt, und den tod darnach lytte. Des
selben urlaubs die magt auß der maßen fro ward.
Contrary to the widely held view of the alleged Latin influence, it will be argued that word order patterns as (1) and (2) are motivated language-internally,
focussing on the distribution of the pronominal adverb darauf ’then’.
The loss V-final patterns during the 18th century was presumably triggered
by further sharpening the structural asymmetry between V2 main clauses and
V-final subclauses, including the rise of w-adverbs.
229
AG 8 · Non-canonical verb positioning in main clauses
AG8
Obligatory V1-order in German SLF-coordination
Donnerstag
09.03.2017
11:45 – 12:15
B4 1, 0.23
Katja Barnickel
Universität Leipzig
[email protected]
One of the main characteristics of SLF-coordination (Höhle 1990, Subjektlücke
in finiten Sätzen subject gap in finite sentences’) is a missing subject in the second conjunct (1). Semantically, the subject of the first conjunct is also understood as the subject of the second conjunct. Nevertheless, the second conjunct,
which is a declarative main clause, has obligatorily to be in verb first order.
Verb second order, where an object (2-a) or an adverb (2-b) has been topicalised to the prefield, as well as verb final order (2-c) are ungrammatical. This
is a notorious problematic and recalcitrant phenomenon that resists a conclusive and convincing analysis so far.
(1)
[Einen Wagen kaufte Hans] und [baute sofort
einen Unfall].
a
car
bought Hans and built immediately an
accident
‘Hans bought a car and caused an accident immediately.’ (B&H 1998)
(2)
a. *[Einen Wagen kaufte Hans] und [einen Unfall2 baute sofort t2 ].
b. *[Einen Wagen kaufte Hans] und [sofort2 baute t2 einen Unfall].
c. *[Einen Wagen kaufte Hans] und [sofort einen Unfall baute].
I claim that the obligatory V1-order follows naturally from putting together
two independently proposed processes: a) biclausal coordinated structures
originate out of monoclausal structures (Weisser 2015) and b) syntactic structure can be removed and remerged (Müller 2015, Heck 2015). I will show that
there was a point at the derivation where the subject of the first conjunct actually was the subject of the second conjunct and hence triggered canonical V2order. Since there is evidence against ATB-movement and ellipsis approaches,
I propose that the subject was removed from the second conjunct later on and
was remerged within the first conjunct.
References: • Büring, D. & Hartmann, K. (1998): Asymmetrische Koordination. Linguistiche Berichte
174. • Heck, F. (2015): Non-monotonic derivations. Ms, U Leipzig. • Höhle, T. (1990): Assumptions
about Asymmetric Coordination. Grammar in Progress. Glow Essays for H. van Riemsdijk. Dordrecht.
• Müller, G. (2015): Structure Removal: A New Approach to Conflicting Representations. Ms, U
Leipzig. • Weisser, P. (2015): Derived Coordination. A Minimalist Perspective on Clause Chains, Converbs
and Asymmetric Coordination. De Gruyter.
230
AG 8 · Gebäude B4 1, Raum 0.23
AG8
Clause typing in main clauses and V1 conditionals in Germanic
Donnerstag
09.03.2017
12:15 – 12:45
B4 1, 0.23
Julia Bacskai-Atkari
Universität Potsdam
[email protected]
Concentrating on German, I examine the left periphery of V1 main clauses (polar interrogatives, V1 conditionals, and V1 declaratives) and their relation to ordinary V2 clauses. I investigate whether the first position, [Spec,CP], is filled,
and why verb movement takes place even in the absence of overt material in
[Spec,CP]. I argue that verb movement occurs because the [fin] feature of the
C head must be lexicalised. While there is in this sense an overtness requirement on C in German main clauses, there is no such requirement on the specifier: some constituent has to be there due to an [edge] feature that is always
present in main clauses, yet this XP does not have to be overt. However, zero
elements (clause-typing operators and anaphors) are licensed only under certain conditions, which is why the language normally surfaces as V2. Contrary
to Zwart (2005), I claim that zero elements in [Spec,CP] in main clauses are
not postulated but are semantically and syntactically motivated, yet I follow
Fanselow (2009) in assuming that verb movement to C and the (overt/covert)
filling of [Spec,CP] are not inseparable phenomena.
Polar interrogatives contain an operator corresponding to whether; a covert
operator in German yields surface V1. In V1 conditionals and declaratives, the
anaphoric elements dann and so may appear in [Spec,CP], and they are licensed
by a preceding clause (the subclause in conditionals and an independent clause
in declaratives). As these anaphors are recoverable from the context, their
zero counterpart is licensed as well, yielding surface V1. Importantly, V1 conditionals/declaratives cannot be uttered in an “out of the blue” context: the
anaphor (overt or covert) needs a preceding proposition as an antecedent.
Thus, V1 main clauses are licensed if the zero operator/anaphor is pragmatically felicitous and semantically recoverable; verb movement is syntactically
triggered by lexicalising [fin] in C regularly.
References: • Fanselow, G. (2009): Bootstrapping verb movement and the clausal architecture of
German (and other languages). In: A. Alexiadou et al. (eds.) Advances in comparative Germanic syntax. John Benjamins. 85–118. • Zwart, C. J-W. (2005): Verb second as a function of Merge. In: M. den
Dikken and Ch. M. Tortora (eds.) The function of function words and functional categories. John Benjamins. 11–40.
231
AG 8 · Non-canonical verb positioning in main clauses
AG8
Donnerstag
09.03.2017
13:45 – 14:45
B4 1, 0.23
The hidden life of V3: an overlooked word order option in Germanic
“V2 languages”
Heike Wiese
Universität Potsdam
[email protected]
In German examples as (2), there are two constitutents, rather than one, occupying the “forefield” in front of the finite verb, in deviation of the Germanic
V2 rule for root declaratives (Wiese 2013, te Velde to appear).
(1)
In diesen Forschergruppen die Kohärenz ist einfach so wahnsinnig
in these research.groups
the coherence is simply so crazily
wichtig.
important
‘In these research groups, coherence is just terribly important.’
Accordingly, such examples are typically starred as “ungrammatical” in discussions of word order options in Germanic languages. Yet, over the last years,
similar data has been attested in a range of Germanic “V2” languages, indicating a systematic – if comparably low-frequency – option (Wiese 2009; Walkden to appear).
Based on corpus data and experimental evidence, I will show that what we
find here is a genuine V3 option, distinct both from SVO and from patterns
analysed as putative multiple forefields (Müller 2005; cf. also Winkler 2014). I
will suggest that this might be a diachronically old pattern that never got lost,
but got overlooked in analyses, underlining the importance of taking into account the whole gamut of language variation for grammatical theory, rather
than only standard-close language use.
References: • Müller, S. (2005): Zur Analyse der scheinbar mehrfachen Vorfeldbesetzung. In: Linguistische Berichte 203, 29-62. • te Velde, J. (to appear): Temporal adverbs in the Kiezdeutsch left
periphery: combining late merge with deaccentuation for V3. In: Studia Linguistica. • Walkden, G.
(to appear): Language contact and V3 in Germanic varieties new and old. In: Journal of Comparative
Germanic Linguistics. • Wiese, H. (2009). Grammatical innovation in multiethnic urban Europe. In:
Lingua 119: 782-806. • Wiese, H. (2013): What can new urban dialects tell us about internal language
dynamics? In: Linguistische Berichte SI 19, 208-245. • Winkler, J. (2014): Verbdrittstellung im Deutschen.
Wiss. Verlag Berlin.
232
AG 8 · Gebäude B4 1, Raum 0.23
AG8
Structure removal in complex prefields
Freitag
10.03.2017
11:30 – 12:00
B4 1, 0.23
Gereon Müller
Universität Leipzig
[email protected]
In complex prefield constructions in German of the type in (1), it looks as
though more than one constituent can occupy the position in front of the finite verb in declarative root clauses.
(1)
Fast alles
im Sitzen bewältigte Joaquim Rodgriguez auf dem
almost everythingacc seated
managed Joaquim Rodgriguez on the
Weg zum Gipfel
way to the peak
There are two kinds of analysis: In one approach, there are multiple constituents in the (possibly split) SpecC domain (2; Lötscher (1985), Eisenberg
(1999), Wurmbrand (2004), Speyer (2008)). In the other approach, complex
prefields are single VP constituents lacking an overt V head ((2b); Fanselow
(1991; 1993), Müller (1998), Müller, St. (2005; 2015)).
(2)
a.
b.
[ CP XP1 [ C′ YP2 [ C′ C [ TP ... t1 ... t2 ...
[ CP [ VP0 XP1 [ V′ YP2 [ V – ]]] [ C′ C [ TP ... t0 ... ]]]
I argue that there is empirical evidence for both views. Well-known arguments
for single constituency involve (i) a clause-mate condition; (ii) order restrictions; and (iii) lack of an upper bound of affected items. However, there is
also substantial evidence for multiple constituency. In addition to existing
arguments from (i) left dislocation and (ii) extraposition, I present new arguments involving (iii) freezing effects; (iv) Barss’ generalization effects; (v)
weak crossover effects; (vi) negative polarity items; and (vii) idioms. I develop
a derivational, minimalist analysis based on an independently motivated operation Remove that is the exact counterpart of the operation Merge (Chomsky
(2001)), and that I take to underlie various constructions that demand conflicting structure assignments. On this view, complex prefields involve both simple VPs (at early stages of the derivation) and multiple constituents (after removal of the VP projection). Finally, I suggest that the information-structural
restrictions on complex prefields (Bildhauer & Cook (2010)) can trigger Remove operations as a last resort.
233
AG 8 · Non-canonical verb positioning in main clauses
AG8
Freitag
10.03.2017
12:00 – 12:30
B4 1, 0.23
A syntactic condition for supposed multiple fronting in German
Werner Frey
ZAS Berlin
[email protected]
After sorting out the truly challenging cases of (supposed) multiple fronting
(MF) in German, the talk will shortly discuss the two approaches available to
account for them: (i) the V2-constraint for German is given up to allow for two
and more independent constituents to appear in front of the finite verb in an
independent clause, (ii) the V2-constraint is upheld, MF being analysed as a
vp/VP constituent projected by an empty verb (e.g. Müller, G. 1998, Müller,
S. 2015). The second approach is more plausible and will be the starting point.
However, it still has severe shortcomings: for example, it highly overgenerates,
the licensing of the empty verb remains unclear, and left dislocations involving MF show unexpected forms of the resumptive.
Independently from MF, Frey (2015) argues that more XPs are incorporated
into the verbal complex in German than usually thought. Building on this
work, I will propose the following syntactic MF constraint (MFC): the rightmost constituent of a MF-construction has to be an incorporated XP. The talk
will show that MFC significantly improves the empirical adequacy of approach
(ii) above and that it refines our understanding of MF. I will also discuss why
only an incorporated XP can licence the empty head which makes MF possible.
References: • Frey, W. (2015): NP-Incorporation in German. In: O. Borik & B. Gehrke (eds.): The Syntax and Semantics of Pseudo-Incorporation. Leiden: Brill, 227-263. • Müller, G. (1998) Incomplete Category Fronting. Dordrecht: Kluwer. • Müller, S. (2015): German Clause Structure: An Analysis with Special
Consideration of So-Called Multiple Frontings. With Contributions by F. Bildhauer & P. Cook. Language
Science Press, Berlin.
234
AG 8 · Gebäude B4 1, Raum 0.23
AG8
Insights into the processing of non-canonical sentence structures in
German: The case of V3 matrix declaratives in informal Standard
German
Oliver Bunk
Universität Potsdam
[email protected]
German matrix declaratives are usually considered to be V2 structures. Exceptions from this V2 constraint have been analyzed in different contexts, e.g. left
dislocation, free topics, topic drop or V1 declaratives. In this paper I investigate
the processing of non-canonical sentences displaying the structure adverbial
– subject – finite verb (cf. 1).
(1)
Jetzt wir fahren zurück.
now we drive back
’Now we drive back.’
(KiDKo: Mo05WD)
While these sentences have been discussed in terms of usage and structure in
corpus studies (cf. Schalowski to appear; Wiese & Rehbein 2015), I present data
from a self-paced reading experiment. The data shows that variables such as
frequency have similar effects on the processing of Adv-S-fin sentences, like
they have been observed in other processing experiments concerning noncanonical word order phenomena (e.g. Kaiser & Trueswell 2004). Therefore I
argue that Adv-S-Vfin must not be treated as an ungrammatical structure that
is prohibited by the language system but as a structure that has specific properties and a full syntactic representation.
References: • Kaiser, E. & Trueswell, J. (2004): The role of discourse context in the processing of a
flexible word-order language. Cognition 94.2. 113–147. • Schalowski, S. (to appear): From an adverbial
to a discourse connective. The function of ‘dann’ and ‘danach’ in non-canonical prefields of German.
In: Fried, M. & Leheckova, E. (eds.), Connectives as a functional category: between clauses and discourse
units, Amsterdam: John Benjamins. • Wiese, H. & Rehbein, I. (2015): Coherence in new urban dialects:
A case study. Lingua 172–173. 45–61.
235
Freitag
10.03.2017
12:30 – 13:00
B4 1, 0.23
AG 8 · Non-canonical verb positioning in main clauses
AG8
Frame setters and V3 patterns in West Flemish
Freitag
10.03.2017
13:00 – 13:30
B4 1, 0.23
Liliane Haegeman
Ghent University
Ciro Greco
Ghent University
[email protected]
[email protected]
Both Standard Dutch (StD) and West Flemich (WF) allow a range of apparent V2 violations in which an extrasentential constituent combines with a V2
clause. StD and WF differ in that the former disallows this pattern when a
frame setting adjunct precedes a subject initial V2 clause:, while the pattern
is acceptable in spoken WF:
(1)
Als mijn tekst klaar is, ik zal je hem opsturen.*StD/✓WF
when my text ready is, I will you it
up-send
‘When my text is ready, I’ll send it to you.’
This pattern, which is reminiscent of the Kiezdeutch data in Freywald et al
((2011), is attested and has been noted in the descriptive dialect literature, but
has by and large been ignored or set aside in the formal literature. After inventorizing some of the main properties of the pattern, we provide an analysis
proposing that the frame setting adjunct in (1) is merged in an extra sentential
discourse projection (‘FrameP’).
Much in the spirit of Mikkelsen (2015)’s work on Danish, we will argue that
the microvariation between StD and WF observed in (1) hinges upon a difference in the derivation of subject initial V2 clauses.
Our account will be shown to capture the temporal interpretation associated
with the frame setting adjunct in periphrastic tenses in (1).
Time permitting, we will briefly discuss the microvariation between WF V3
patterns such (1) and similar patterns in French Flemish reported in literature
(Ryckeboer 2004, Vanacker 1977).
References: • Freywald, U, K. Mayr, T. Özçelik, & H. Wiese. 2011. Kiezdeutsch as a multiethnolect.
Ethnic Styles of Speaking in European Metropolitan Areas ed. F. Kern & M. Selting, 45–73. Amsterdam:
John Benjamins • Mikkelsen, L. 2015. VP anaphora and verb second order in Danish. Journal of Linguistics 51, 595-643 • Ryckeboer, H. 2004. Fransvlaams. Lannoo: Tielt • Vanacker, V.F. 1977. Syntactische overeenkomsten tussen Frans-Vlaamse en Westvlaamse dialecten. De Franse Nederlanden. Les
Pays Bas Français. Jaarboek. Rekkem: Ons Erfdeel, 206-216
236
AG 8 · Gebäude B4 1, Raum 0.23
V3 in Germanic: A comparison of urban vernaculars and heritage
languages
Artemis Alexiadou
Humboldt & ZAS
Terje Lohndal
NTNU & UIT
[email protected]
[email protected]
It is well known that varieties of Germanic do not display a strict V2 system
whereby the finite verb is in the second position in main clauses. In this paper,
we want to compare Germanic heritage languages with Germanic urban vernaculars, demonstrating that similar processes and structures are in place in
both varieties. This, we claim, provides important information about the nature of bilingual grammars in general, and of verb placement specifically.
Walkden (in press) provides a detailed overview of verb placement in
learner varieties of various Germanic languages. Building on Walkden, we argue that the same reasoning can be extended to other varieties of Germanic
languages, most notably heritage languages. This comparison suggests that the
factors favoring V3 in Germanic are uniform, involving adjunct initial clauses.
To this end, we will present data from American Norwegian showing surprisingly similar data.
American Norwegian is a heritage language spoken in the US. Data have
been collected by Haugen (1953), Hjelde (1992), and more recently through the
spoken corpus CANS (Johannessen et al. 2015). Eide & Hjelde (2015) investigate V2 in American Norwegian based on the corpus. (1) provides an example
of V3.
(1)
nå je fløtte nerri her, kjinner alle
her, veit du.
now I move down here, know everyone here, know you
‘Now I’m moving down here, I know everyone here, you know.’
(Eide & Hjelde 2015: 86)
We provide an analysis of V3 patterns and the relationship to V2, demonstrating that V2 structures appear more easily with adjuncts as the initial constituent, for which we provide a formal analysis.
References: • Haugen, E. (1953): The Norwegian Language in America. Indiana UP • Walkden, G.
(in press): Language contact and V3 in Germanic varieties new and old. Journal of Comparative Germanic Linguistics 20.
237
Freitag
10.03.2017
13:30 – 14:00
B4 1, 0.23
narr. Wissen mit Profil.
ISTEN
U
G
N
I
L
R
NKT FÜ
U
P
F
F
E
R
e.de
h
c
DER T
a
r
p
s
n
r r t-i
w w w. v e r n
a
AG9
Towards an ontology of modal flavors
Ryan Bochnak, Anne Mucha & Kilu von Prince
Universität Leipzig/Universität Konstanz, Institut für Deutsche Sprache,
Mannheim, Humboldt-Universität zu Berlin
[email protected], [email protected], [email protected]
Raum: B4 1, 0.22
Short description
Our understanding of modal meanings is crucially based on the notion of various modal flavors, which distinguish, for example, between epistemic and deontic readings. However, neither within nor across linguistic subfields is there
any consensus about the exact ontology of those modal flavors. This workshop
aims to provide a forum for researchers in formal semantics, typology, syntax,
language description, psycholinguistics and language acquisition to address
these issues in the analysis of linguistic modality, in order to gain a better understanding of the role of modal flavors in grammar and cognition.
240
AG 9 · Gebäude B4 1, Raum 0.22
Flavors of existential/possessive modals
AG9
Aynat Rubinstein
The Hebrew University of Jerusalem
[email protected]
In many unrelated languages, expressions of possession give rise to modal
meanings in which they express obligation (e.g., English have as in I have a
bottle of wine vs. have to as in I have to save it for a special occasion; Bybee et al.
1994; Bhatt 1998; Bjorkman and Cowper 2016). Modern Hebrew seems to conform to this crosslinguistic tendency with the existential copula yeš ‘there is’
(Boneh 2013; also eyn ‘there is not’), although there are two wrinkles: the construction is not fully possessive, and adding a possessor phrase gives rise to a
specific type of goal-oriented modality, not obligation. Moreover, earlier varieties of Hebrew constitute a counterexample to the crosslinguistic possessionobligation generalization. While modal uses of the copula are documented
throughout the history of Hebrew, these uses differ from the modern ones in
terms of both force and flavor. In terms of force, historical yeš can express possibility and not necessity. In terms of flavor, it can refer to non-normative possibilities like abilities (Ben-Ḥayyim 1953/1992, Shehadeh 1991).
This talk aims to provide a semantic analysis of the modern modal existential/possessive constructions in Hebrew that is informed by their trajectory
of historical change, focusing on what looks like a change in modal force and
morphosyntactic loss of the possessor phrase. The profile of modal flavor in
the different constructions sheds light on the grammatical underpinnings of
possessive modality crosslinguistically.
References: • Ben-Ḥayyim, Z. (1992): The struggle for a language. The Academy of the Hebrew Language. In Hebrew. • Bhatt, R. (1998): Obligation and possession. In Papers from the UPenn/MIT
roundtable on argument structure and aspect, ed. Heidi Harley, volume 32 of MIT Working Papers in Linguistics, 21–40. MITWPL. • Bjorkman, B. and E. Cowper (2016): Possession and necessity: from individuals to worlds. Lingua 182(2016), 30–48. • Boneh, N. (2013): Mood and modality: Modern Hebrew.
In Encyclopedia of Hebrew language and linguistics, ed. Geoffrey Khan, volume 2, 693–703. Brill.
241
Mittwoch
08.03.2017
13:45 – 14:45
B4 1, 0.22
AG 9 · Towards an ontology of modal flavors
Force: Topicalization, context-sensitivity, and morality
AG9
Mittwoch
08.03.2017
14:45 – 15:15
B4 1, 0.22
Matthew Mandelkern
Massachusetts Institute of Technology
Jonathan Phillips
Harvard University
[email protected]
[email protected]
Recent studies have shown that moral assessments have surprising and
widespread ramifications on judgments in many other apparently distinct
conceptual domains. We attempt to improve our understanding of this phenomenon with a case study on force. We begin by reviewing the basic phenomenon: subjects are less likely to agree that an individual was forced to
do φ if φ is judged to be a morally bad action. Then we introduce new data
which show that these judgments are sensitive to topicalization: in a scenario
in which A in some sense compels B to do a morally bad action φ, subjects are
much more willing to agree with (1a) than with (1b):
(1)
a. A forced B to φ.
b. B was forced by A to φ.
Why does topicalizing B in this way affect judgments? We get a foothold on this
problem by demonstrating that these judgments are also subject to counterintuitive order effects. We then explain the overall pattern of data by arguing for
two principles: First, we suggest that immoral actions lead us to ignore substantial parts of the causal background and to instead conceptualize agents as
free to act in a variety of ways. Second, however, when elements of the causal
background are made salient, e.g. by topicalization, we can no longer ignore
them, and thus tend to judge agents as less free to do otherwise. We model this
by positing that judgments about force depend on a contextually given parameter, which is sensitive to how much information about the causal background
we are taking into account. We suggest that this account of the role of morality in judgments about ‘force’ extends broadly to the array of concepts which
make up the core of our practical lives.
References: • Knobe, J. (2010). Person as scientist, person as moralist. Behavioral and Brain Sciences,
33(04):315-329. • Mandelkern, M., Schultheis, G., and Boylan, D. (2015). I believe I can φ. In 20th
Amsterdam Colloquium, pages 256-265. • Phillips, J. and Knobe, J. (2009). Moral judgments and intuitions about freedom. Psychological Inquiry, 20(1):30-36. • Phillips, J. S. (2015). The Psychological
Representation of Modality. PhD thesis, Yale University.
242
AG 9 · Gebäude B4 1, Raum 0.22
The early production of conditionals
AG9
Eva Csipak
Universität Konstanz
[email protected]
This talk investigates whether insights into the acquisition of modal verbs
can be extended to conditionals by looking at production data of monolingual
German-speaking children aged 1;4 to 3;8 (Szagun 2001).
Conditional sentences of the form if p, q are generally assumed have a modal
meaning, with the antecedent proposition p serving as the restrictor of the
worlds quantified over in the consequent q (Lewis 1973; Kratzer 1986; much
subsequent work). Modals exhibit modal flavour which falls into root and nonroot varities (Hoffmann 1966). Nonroot modals deal with possibilities stemming from the speaker’s knowledge, whereas root modals deal with possibilities stemming from relevant facts in the world. Bare conditionals (those without an overt modal in the consequent) have a generic, law-like flavour (von
Fintel 1997). Thus they are expected to behave like root modals. Language acquisition research shows that children acquire root modality in modal verbs
before nonroot modality (Papafragou 1998).
By searching the Szagun subcorpus of the Child Language Data Exchange
System for cases of children producing the word wenn ‘if ’ and classifying the resulting hits as having root or nonroot modality, it is shown that German monolingual children productively use root modality, especially deontic modality,
before age 3;8. Conditionals with nonroot modality are also produced, albeit
not productively. The figures should be read as follows: root modality – 1; nonroot modality – 2; ambiguous cases – 9. The acquisition of conditionals follows
the same pattern as the acquisition of modal verbs.
References: • von Fintel, K. (1997): Bare plurals, bare conditionals, and only. JoS 14:1-56. • Papafragou,
A. (1998): The acquisition of modality. M&L 13:370-399. • Szagun, G. (2001): Learning different regularities. First Language 21:109-141.
243
Mittwoch
08.03.2017
15:15 – 15:45
B4 1, 0.22
AG 9 · Towards an ontology of modal flavors
Decomposing modal flavours
AG9
Mittwoch
08.03.2017
16:30 – 17:00
B4 1, 0.22
Jakob Maché
Ọbáfẹ́mi Awó lọ́wọ̀ University Ilé -Ifẹ̀
[email protected]
This paper shows that besides modal force, modal base and ordering source,
there are two further important aspects which are essential to correctly analyse modal readings: the modal source and the sentential mood.
As pointed out in the corpus study carried out by Mache (2013), there are
a couple of environments in which reportative modal verbs occur and their
epistemic cognates fail to: (i) in nominalisations, (ii) in adverbial infinitives
(cf. (1)), (iii) embedded under tense auxiliaries and (iv) optative operators.
(1)
Schließlich gab Sabine Marker nach und setzte ihre Unterschrift auf
finally
gave Sabine Marker after and put her signature
on
die Erklärung, ohne gewusst haben zu wollen, was sie da
the declaration without know-ppp have-inf to want-inf what she there
unterzeichnet.
(DeReKo: NUN10/OKT.03036)
signes
‘Finally, Sabine Marker complied and put her signature under the declaration and now she claims that she did not know what she was signing.’
In this paper, it will be shown that any modal operator introduces a variable
for a modal source which has to be bound in the most local configuration. With
circumstantial and reportative modal verbs, the modal source will be bound by
a (covert) argument of the modal predicate. The variable of the modal source is
bound within the proposition. With the variable bound, the reportative modal
can be embedded in nominalisations, adverbial infinitives, optatives and under tense auxiliaries. With epistemic modal verbs however, the modal source
is identified by the referent who utters the proposition. An epistemic modal
verb cannot be embedded under most operators as long the variable of the
modal source is not bound. Moreover, it will be shown how some epistemic
modal verbs are affected by subjunctive operators, much the way Rubinstein
(2012) observed.
References: • Maché, J. (2013): On Black Magic. PhD-thesis FU Berlin. • Rubinstein, A. (2012): Roots
of Modality. PhD-thesis UMass.
244
AG 9 · Gebäude B4 1, Raum 0.22
Weak necessity modals and modal flavor:
The view from Paciran Javanese
Vera Hohaus
Eberhard Karls Universität Tübingen
[email protected]
AG9
Jozina Vander Klok
University of British Columbia,
Vancouver
Mittwoch
08.03.2017
17:00 – 17:30
B4 1, 0.22
[email protected]
Considering that modality concerns two dimensions, modal force and modal
flavor, the question arises whether the distinction of force between strong and
weak necessity is independent from the type of modal flavor. That is, would
natural language allow for changes in modal flavor based on secondary priorities? We investigate this question based on data from primary fieldwork
on Paciran Javanese (a Javanese dialect spoken in East Java, Indonesia), and
propose that modal flavor remains invariant in the modal force change from
strong to weak.
Weak necessity modals in Paciran Javanese are morphologically complex:
They are transparently built by combining a necessity modal with the suffix
–ne, as in (1). We suggest that –ne overtly realizes a context-dependent secondary ordering source, which refines the ranking of the possible worlds quantified over by the modal (Kratzer 1991), drawing on analyses proposed in von
Fintel & Iatridou (2008) and Rubinstein (2012).
(1)
Sampean kudu-ne
ora mbengok-mbengok.
2SG
root.nec-ne neg red-av.shout
‘You should not shout!’
At Logical Form, –ne attach within the restrictor of the modal, modifying the
set of worlds determined by the accessiblity relation and the first ordering
source. We therefore correctly predict that -ne may weaken the modal force
but does not alter the modal flavor of the modal.
References: • von Fintel, K. & S. Iatridou (2008): The Composition of Weak Necessity Modals. In: J.
Guéron & J. Lecarme (eds.), Time and Modality, 115–141. • Kratzer, A. (1991): Modality. In: A. von Stechow & M. Herwig (eds.), Semantik, 639-650. • Rubinstein, A. (2012): Roots of Modality. Ph.D. thesis,
UMass Amherst.
245
AG 9 · Towards an ontology of modal flavors
Modal flavor/modal force interactions in German
AG9
Mittwoch
08.03.2017
17:30 – 18:00
B4 1, 0.22
Lisa Matthewson
UBC Vancouver
Hubert Truckenbrodt
ZAS and HU Berlin
[email protected]
[email protected]
Von Fintel and Iatridou (2008) show for a range of languages that counterfactual morphology, added to a strong necessity modal like must, leads to the
weak necessity (WN) reading of English ought. We generalize the analysis of
Rubinstein (2012) for this phenomenon and analyze German müssen and sollen
against this background, building on Ehrich (2001), among others. German
shows interaction with modal flavor and allows WN readings only in the synthetic form [modalKONJ.II ], not with analytic forms [würdeKONJ.II + modal]. For
müssen we find the expected WN reading with epistemic flavor as in (1b).
(1)
a. Peter muss in der Küche sein.
(only option given the evidence)
b. Peter müsste in der Küche sein.
(not only option)
c. #Peter würde in der Küche sein müssen.
(würde + modal: *WN)
However, the deontic use of müssen does not have such a WN alternative, as
shown in (2b). We find instead a WN politeness reading, as indicated in (2b).
The analytic form is again not possible, cf. (2c).
(2)
A: Wie komme ich nach Amherst?
a.
b.
B: Du musst Rt. 9 nehmen.
(presented as only road to A.)
B: Du müsstest Rt. 9 nehmen.
(presented as only road to
A., Konj.II adds an element like ‘if you don’t mind’) ̸= You ought to
take Rt. 9. (pres. as best but not only road t.A.)
c. #B: Du würdest Rt. 9 nehmen müssen. (würde + modal: *WN)
The analytic form [würde + modal] is possible where Konj.II is licensed by a
counterfactual conditional and does not have a WN reading:
(3)
Wenn er ein Auto hätte, würde er es anmelden müssen.
References: • Ehrich, V. (2001): Was nicht müssen und nicht können (nicht) bedeuten können (...).
LB Sonderheft 9, 149-176. • von Fintel, K. & S. Iatridou (2008): How to say ought in foreign. In Time
and Modality, Dordrecht, Springer Netherlands, 115-141. • Rubinstein, A. (2012): Roots of Modality.
Ph.D. dissertation, UMass, Amherst.
246
AG 9 · Gebäude B4 1, Raum 0.22
Swabian, German, Chinese and German Sign Language: Multi-source
convergence on a cartographic array of modal flavors
AG9
Daniel Hole
Universität Stuttgart
Fabian Bross
Universität Stuttgart
[email protected]
[email protected]
Donnerstag
09.03.2017
09:00 – 09:30
B4 1, 0.22
The aim of this talk is to provide novel cross-linguistic evidence for an essentially cartographic theory (Cinque 1999) of modal meanings in several unrelated languages: Middle Swabian, Standard German; Chinese and German
Sign Language. We will argue for a modal flavor that projects between deontic
and circumstantial modality. We dub this flavor ‘design modality’ and leave it
open as to whether it fully coincides with, or covers only a subdomain of, Rubinstein’s (2012) goal-oriented modality. We argue that the different shades
of modality conventionalize in the (morpho-)syntax as a function of their respective ordering sources/modal anchors (Kratzer 1991, Hacquard 2006). We
consider the (presumably incomplete) shades of modality and their characterizations in (1). The order provided is one that, we assume, holds in the syntax
as well:1
(1)
a.
b.
c.
d.
e.
f.
g.
epistemic ‘What can or must hold in view of what the speaker knows?’
[deictic tense]
bouletic/volitional ‘W.c.o.m.h.i.v.o.w. the subject wants?’
deontic ‘Wcomhivow the asymmetric power relations are like?’
design ‘Wcomhivow the relevant participant was designed for?’
circumstantial ‘Wcomhivo causality affecting the relevant participant?’
root ‘Wcomhivo the inherent properties of the modal anchor?’
We will provide evidence to the effect that moving down the hierarchy in (1)
makes a (morpho-)syntactic difference for each step in at least one of the languages surveyed.
References: • Cinque, G. (1999): Adverbs and functional heads. A cross-linguistic perspective. New York:
Oxford University Press. • Hacquard, V. (2006): Aspects of modality (Doctoral dissertation, Massachusetts Institute of Technology). • Kratzer, A. (1991): Modality. In: von Stechow & Wunderlich, D.
(eds.): Semantics: An International Handbook of Contemporary Research. Berlin: de Gruyter. 639-650.
• Rubinstein, A. (2012): Roots of modality (Doctoral dissertation, University of Massachusetts
Amherst).
1
W.c.o.m.h.i.v.o.w. = ‘What can or must hold in view of what’
247
AG 9 · Towards an ontology of modal flavors
AG9
The (in)stability of modal flavors: The case of English modals and
their Spanish equivalents
Dana Kratochvílová
Charles University
Donnerstag
09.03.2017
09:30 – 10:00
B4 1, 0.22
[email protected]
In English, the main instrument for expressing speaker’s attitude toward the
content of his utterance are the modal verbs. In Spanish, the center of the
modal system is formed by verbal moods that can co-occur with modal verbs.
We may distinguish two levels of this co-occurrence: cooperation and competition.
The fact that Spanish modal verbs can be subjected do tense and, more importantly, modal inflection offers to Spanish speakers a possibility of double
modalization which is, generally, impossible in English (cooperation).
On the other hand, the strongly fusional character of the Spanish verbal system often leads to a complete omission of modal verbs and their substitution by
other resources for expressing modality, especially, by the subjunctive (competition).
Our research provides a contrastive study of English constructions that contain a modal verb and their respective Spanish equivalents. Using the parallel
corpus InterCorp, we analyze a wide range of Spanish translations of English
modals, concentrating on their frequency, on the criteria that lead to their selection and on the level of their cooperation / competition with other modal
resources (complete substitution of the English modal by the subjunctive, conditional or epistemic future, modal and tense inflection of the modal verb or
its combination with lexical expressions of modality).
This analysis enables us to outline some of the main differences between
Spanish and English modality in general and helps us to determine the position of Spanish modal verbs in the Spanish modal systems. It also shows to
what extent a concrete modal flavor expressed in English can be translated
into Spanish, reflecting, thus, their (in)stability.
References: • Rosen, A. - Vavřín, M. (2016): Korpus InterCorp – Spanish, English, version 9
(www.korpus.cz).
248
AG 9 · Gebäude B4 1, Raum 0.22
Modal force and flavour as semantic restrictors of possible double
modal combinations in Croatian
AG9
Ana Werkmann Horvat
University of Oxford
Donnerstag
09.03.2017
10:00 – 10:30
B4 1, 0.22
[email protected]
The primary aim of this paper is to explore the manner in which modal force
and flavour impose restrictions on possible double modal (DM) combinations
in Croatian.
DM data shows that modal auxiliaries combine on the basis of certain semantic restrictions that occur between different modal flavour groups. While
some of these semantic restrictions have already been reported in the literature, such as epistemics scoping over non-epistemics (Nauze 2008; Butler
2003), and priority modals scoping over circumstantials (Nauze 2008); Croatian data offers a new insight into semantic restrictions within the circumstantial group of modals. Croatian DM data suggests that what has traditionally been considered one modal flavour, i.e. circumstantial, ought to be considered three different grammatically distinct flavours, namely pure possibility,
ability and disposition (Coates 1983, Kratzer 1981, Palmer 2001). They create
a scope hierarchy, similarly to the above defined restrictions. Data shows that
pure possibility scopes over ability and disposition, while ability scopes over
disposition.
Croatian data also shows that there is a constraint on the order of the modal
force, i.e. necessity scopes over possibility (Cinque, 1999; Butler, 2003). This
constraint holds for all flavours, but crucially only within a flavour group, and
not between flavour groups.
A result of this research is a hierarchical analysis which takes into account
both modal flavour and modal force as crucial pieces of modal meaning and
challenges the way we traditionally define modal flavours within the nonepistemic group.
References: • Butler, J. (2003). A minimalist treatment of modality. Lingua, 113(10), 967–996.
• Cinque, G. (1999). Adverbs and functional heads: A cross-linguistic perspective. Oxford: Oxford University Press. • Coates, J. (1983). The semantics of the modal auxiliaries. London: Croom Helm. • Kratzer,
A. (1981). The notional category of modality. Words, worlds, and contexts, 38-74. • Nauze, F. D. (2008).
Modality in typological perspective. PhD thesis, Amsterdam: Institute for Logic, Language and Computation. • Palmer, F. R. (2001). Mood and modality. Cambridge: Cambridge University Press.
249
AG 9 · Towards an ontology of modal flavors
Modal concord is not modal concord
AG9
Donnerstag
09.03.2017
11:15 – 11:45
B4 1, 0.22
Lavi Wolf
Ben Gurion University of the Negev
[email protected]
Modal flavors are classically divided into epistemic and root. Interestingly,
while root modals have an inner division, e.g. deontic, ability, bouletic, aletic
etc. epistemic modals are considered a class of their own. This is, however, not
necessarily the case. This talk discusses a phenomenon that points to a division
between two types of epistemic modals, termed Modal Concord (MC). When
modal adverbs interact with modal auxiliaries of the same quantificational
force and apparent modal flavor, one of the modals seems to become semantically vacuous:
(1)
Jane might possibly be in the office.
The interpretation of (1), is interestingly not what we would expect from two
modals of the same flavor, i.e. a possibility of a possibility. Rather, there seems
to be a single possibility here.
The proposed theory is that the so-called MC is actually not MC. Instead, it reflects an interaction between two different flavors within epistemic modality.
Might is a truth-conditional epistemic modal and possibly is use-conditional.
The formal account is a speech act based probabilistic account in which truth
conditional epistemic modals modify the propositional content and use conditional epistemic modals modify the degree of illocutionary force by which the
speech act is performed.
References: • Anand, P., & Brasoveanu, A. (2010). Modal concord as modal modification. In Proceedings of Sinn und Bedeutung, vol. 14 (pp. 19–36). • Geurts, B., & Huitink, J. (2006). Modal concord. In P.
Dekker & H. Zeijlstra (Eds.), Concord Phenomena and the Syntax Semantics Interface, ESSLLI. Malaga.
• Grosz, P. (2010). Grading modality: A new approach to modal concord and its relatives. In Proceedings of Sinn und Bedeutung, vol. 14, pp. 185-201. 2010. • Gutzmann, D. (2015). Use-conditional meaning:
Studies in multidimensional semantics. Oxford: Oxford University Press. • Huitink, J. (2012). Modal
Concord: A Case Study of Dutch. Journal of Semantics, 29(3), 403–437. • Potts, C. (2007). The expressive dimension. Theoretical Linguistics, 33, 165–198. • Zeijlstra, H. (2007). Modal concord. In Proceedings of SALT Vol. 17 (pp. 317–332).
250
AG 9 · Gebäude B4 1, Raum 0.22
Veritic semantics for epistemic modals
AG9
Adam Marushak
University of Pittsburgh
[email protected]
Epistemic modals have the name they do because they are often thought to
describe how things stand with bodies of knowledge. My aim in this talk is to
argue that this semantics is false: epistemic modals have no special semantic
ties to knowledge, evidence, or any other epistemological concept.
My argument proceeds in two stages. First, I’ll raise a dilemma for any
knowledge-describing semantics for epistemic modals: such a semantics either fails to explain the unembeddability of epistemic contradictions, or fails
to explain the embeddability of certain sentences explicitly describing epistemic reasons.
I’ll then use this dilemma to motivate an alternative, veritic semantics for
epistemic modals, according to which epistemic modals describe how things
stand with the contextually relevant truths.
The result is a significant departure from the orthodox account of the epistemic flavor of modality: it is reason claims, not epistemic modals, that describe knowledge.
The ontological cookbook of modal categories: There are more
flavors than you think
Donnerstag
09.03.2017
11:45 – 12:15
B4 1, 0.22
Donnerstag
09.03.2017
12:15 – 12:45
B4 1, 0.22
Dietmar Zaefferer
Ludwig-Maximilians-Universität München
[email protected]
modal categories via speaker’s subjectivity (Palmer 2001) or a vague appeal to
necessity and possibility (Kratzer 1991) is insufficient. We define modal categories as consisting of four operators with different forces: strong, weak, antistrong, antiweak. Whereas the set of modal flavors is open, only a few of them
are central, among them the two that link up with the major sentence moods:
251
AG 9 · Towards an ontology of modal flavors
volitionality and epistemicity.
Strong epistemic operators map their argument P to the proposition that the
epistemic subject subjectively knows that p, where p is either P (transparent
case) or one of the alternatives coded by P (opaque case). Assuming that the
German perfecto-present wissen codes epistemicity in this sense predicts that
it has both a transparent and an opaque reading.
Strong volitional operators map their argument p to the proposition that the
volitional subject realistically aims at bringing about p. Assuming that the German modals wollen and sollen code volitionality in this sense predicts that applied to arguments that cannot be realistically aimed at they either result in
nonsense or invite metonymic coercion.
(1)
Eva will mit 23 Jahren ihre Prüfung /a. bestehen /b. bestanden haben
Eva wants to /a. pass /b. have passed her exam at the age of 23
If Eva is 21 years old (1) a. is read as plain future, and (1) b. as future perfect: In
both cases she aims at having the degree within two years.
If Eva is 31 (1) a. is nonsense, and (1) b. gets an epistemic reading: She aims
at making people think she got her degree eight years ago.
Putting together these building blocks we argue that the core sentence
moods are best defined as coding either material volition (imperative) or epistemic volition: transparent (declarative) or opaque (interrogative).
References: • Kratzer, A. (1991): Modality. In: Hdbook of Semantics. De Gruyter, 639-650. • Palmer,
F. R. (2001). Mood and Modality. Cambridge University Press.
252
AG10
Polysemy and coercion of clause-embedding predicates
Marie-Luise Popp & Barbara Stiebels
Universität Leipzig, Universität Leipzig
[email protected], [email protected]
Raum: B4 1, 0.07
Short description
The workshop will address the role of polysemy of clause-embedding predicates (CEPs; e.g., tell, ask) for the distributional properties of these predicates
(e.g., complementation patterns, NEG-raising, control vs. raising, mood selection, restructuring/clause union) and the role of coercion in adding clausal arguments to non-clause-embedding verbs or in enhancing the syntactic flexibility of CEPs.
253
AG 10 · Polysemy and coercion of clause-embedding predicates
Mittwoch
08.03.2017
13:45 – 14:45
B4 1, 0.07
Introduction: The role of polysemy and coercion in clause-embedding
predicates
AG10
Marie-Luise Popp
Universität Leipzig
Barbara Stiebels
Universität Leipzig
[email protected]
[email protected]
In our introduction we will address some cases that highlight the role of polysemy and coercion for the complementation patterns of clause-embedding
predicates (CEPs).
We will exemplify the role of the complementation type with vague or polysemous predicates in languages with small inventories of clause-embedding
predicates and contrast these data with German.
We will discuss the role of coercion for the syntactic flexibility of CEPs. In
particular, we will demonstrate how coercion feeds the licensing of V2 complements in German. Unlike infinitival complements whose licensing is often
modulated via modals, modal expressions and aspectual markers in the infinitival complement, V2 licensing is frequently enhanced via coercion of the
matrix predicate (e.g., establishing a speech act predicate reading). This coercion mechanism is related to the fact that German allows the conflation of a
manner component (e.g., sound emission) with other meaning aspects (e.g.,
movement, speech act). Based on Troyke-Lekschas’ (2013) study we will show
that the usualization of coerced predicates paves the way from less integrated
clausal complements (direct speech, V2) to more integrated structures.
Based on a small-scale cross-linguistic study of NEG-raising predicates
(Popp 2016) we will demonstrate that predicates that exhibit a ‘hope’/‘expect’
polysemy (e.g., Spanish, Lithuanian or Swahili), only the ‘expect’ reading allows NEG raising; ‘expect’ predicates in other languages function as “strong”
NEG raisers, whereas ‘hope’ predicates only function in some languages as
NEG raisers.
References: • Popp, M.-L. (2016): NEG-raising in cross-linguistic perspective. MA thesis. University of
Leipzig. • Troyke-Lekschas, S. (2013): Korpuslinguistische Untersuchungen zum Phänomen der Satzeinbettung bei deutschen Geräuschverben. MA thesis, HU Berlin.
254
AG 10 · Gebäude B4 1, Raum 0.07
Basic pieces, complex meanings: Building attitudes in Navajo
Mittwoch
08.03.2017
14:45 – 15:45
B4 1, 0.07
Elizabeth Bogal-Allbritten
University of Gothenburg
[email protected]
AG10
It is crosslinguistically widely-attested for different attitude reports to be distinguished chiefly by their verbs, e.g. think/denken, vs. want/wollen. In familiar
accounts, the attitude verb determines the meaning of the attitude report in its
entirety. More recently, however, Kratzer (2006, 2013) and Moulton (2009)
have argued that key semantic aspects of English and German attitude reports
come from embedded material, not attitude verbs.
I present fieldwork data from Navajo in support of Kratzer and Moulton’s
compositional account. Navajo sentences in (1) express either belief or desire ((1c) is string ambiguous). Unlike their English translations, however, the
Navajo sentences all contain the same verb, nízin. I show that nízin is not lexically ambiguous between meanings on par with familiar entries of think and
want. Rather, Navajo is a limiting case in the empirical landscape predicted
by Kratzer and Moulton: the embedded clause determines all attitude-related
meaning and nízin only adds the attitude holder.
(1)
a. [Nisneez] nízin.
1sg.tall
3sg.nizin
‘S/he thinks I am tall.’
b. [Nisneez laanaa] nízin.
1sg.tall desire 3sg.nizin
‘S/he wishes I were tall.’
c. [Nisneez dooleeł] nízin.
1sg.tall future 3sg.nizin
(i) ‘S/he thinks I will be tall.’ (ii) ‘S/he wants me to be tall.’
Crucially, clauses embedded by nízin can also function as main clauses with
meanings intuitively related to attitudes. E.g., unembedded nisneez (compare
(1a)) expresses an assertion (‘I am tall’) whereas nisneez dooleeł is ambiguous
much like (1c), expressing an assertion (‘I will be tall’) or a priority (‘I need to
be tall’). I propose that Navajo builds beliefs and desire from nízin and operators (assertion, priority) used beyond attitude reports.
References: • Kratzer, A. (2006): Decomposing attitude verbs. Talk at the Hebrew University
Jerusalem. • Kratzer, A. (2013): Modality and the semantics of embedding. Amsterdam Colloquium.
• Moulton, K. (2009): Natural selection and the syntax of clausal complementation, UMass dissertation.
255
AG 10 · Polysemy and coercion of clause-embedding predicates
Mittwoch
08.03.2017
16:30 – 17:30
B4 1, 0.07
Systematic polysemy of non-factive epistemic and fiction verbs in
Italian: evidence from mood variation
Alda Mari
Institut Jean Nicod, ENS/EHESS/PSL/UMR 8129
AG10
[email protected]
Empirical puzzle. Semantic approaches of mood choice (e.g. Quer, 1998; Giannakidou, 1999; Farkas, 2003; Villalta, 2008; Anand and Hacquard, 2013) predict belief predicates to be indicative selectors by assuming a clash between
(i) Hintikka (1962) semantics for belief using a homogeneous doxastic modal
base, and (ii) the constraint according to which the subjunctive is triggered
by the presence of alternatives. Italian credere (believe), licenses both the subjunctive and the indicative and is a notable exception to this cross-linguistic
generalization. We newly note that fiction predicates in Italian also license the
subjunctive, contrary to the expectations.
Proposal. We will maintain that the subjunctive in the embedded clause is
triggered by polar alternatives in the modal base of the matrix predicate. Languages that license the subjunctive under belief and fiction verbs allow us to
see a systematic polysemy previously gone unnoticed. We newly propose that
belief and fiction verbs cross-linguistically are systematically polysemous between a hintikkean-type of meaning (which triggers the indicative) and what
we call an ‘inquisitive’ meaning. This triggers the subjunctive as several new
pieces of data show. It is a multilayered meaning that features a presupposition of epistemic uncertainty and conveys doxastic certainity in the assertion.
We show that fictional predicates are ambiguous in a parallel way, and select
the subjunctive whenever they feature an epistemic uncertainity layer in the
presupposition.
References: • Anand, P. & V. Hacquard (2013): Epistemics and Attitudes. S&P 6: 1-59. • Farkas, D.
(2003): Assertion, belief and mood choice. Esslli. • Giannakidou, A. (1999): Affective dependencies.
L&P 22: 367-421 • Mari, A. (2016): Assertability conditions of epistemic (and fictional) attitudes and
mood variation. SALT 26: 61-81. • Portner, P. (forthcoming): Mood. OUP. • Quer, J. (1998): Moods at
the interface. PhD. • Villalta, E. (2008): Mood and gradability. L&P 31: 467-522
256
AG 10 · Gebäude B4 1, Raum 0.07
Mood selection of two visual perception verbs in Hong Kong
Cantonese
Mittwoch
08.03.2017
17:30 – 18:00
B4 1, 0.07
Michelle Li
Caritas Institute of Higher Education
AG10
[email protected]
Cantonese belongs to the Yue dialect group of Chinese. This paper examines
the polyfunctionality of two visual perception verbs in Hong Kong Cantonese
– gin3 ‘see’ and tai2 ‘look (at)’. While gin3 refers to successful perception, tai2
is an activity verb. These verbs follow the well-known semantic extension
of perception to cognition (Viberg 1984). They also function as complementtaking predicates referring to different cognitive states and activities and select clausal complements of different modal properties. Gin3 can mean ‘aware,
know, notice, realize’ which selects realis complements expressing a high degree of certainty as illustrated in (1).
(1)
ngo5 gin3 koei5 mei6 hou2 faan1 so2ji5 mou5 giu3 keoi5 lai4
1SG see 3SG not good back so
not call 3SG come
‘Seeing that she had not yet recovered, I didn’t ask her to come.’
Tai2 refers to the act of thinking whose complements have an irrealis interpretation which not available to gin3 as shown in (2).
(2)
nei5 *gin3/tai2 keoi5 wui5 m4 wui5 jeng4?
2SG *see/look 3SG will not will win
‘Do you think she will win?’
It is argued that the difference in modal properties of these two verbs can be attributed to their basic use as perception verbs, despite weakening of this sense
in the process of semantic extension.
References: • Viberg, Å. (1984): The verbs of perception: a typological study. In Brian Utterworth,
Bernard Comrie & Östen Dahl (eds.). Explanations for Language Universals, 123-162. Berlin: Mouton.
257
AG 10 · Polysemy and coercion of clause-embedding predicates
How believing so is different from believing it
Donnerstag
09.03.2017
09:00 – 10:00
B4 1, 0.07
A. Marlijn Meijer
Humboldt-Universität zu Berlin
[email protected]
AG10
This study focuses on the English propositional anaphor so, in comparison to
it; see 1. Native speakers report that in embedded responses to questions, such
as 1, using so is better than using it. Needham’s (2012) corpus study shows that
antecedents of so mostly are questions. Furthermore, the distribution of so is
quite restricted: it cannot occur with verbs such as regret or resent. Therefore, it
has been suggested that so only occurs with non-factives (Kiparsky & Kiparsky
1971); that so does not presuppose that the referent is true (Cornish 1992); or
that so refers to the question under discussion, to which the speaker is not
committed (Needham 2012). However, Bhatt’s (2010) finding, that so can occur with know in certain contexts, e.g. in 2, is problematic for these theories.
I argue that so presupposes that its referent is still under discussion and
thus is not part of the common ground (CG), at the time of the occurrence of
the eventuality of the predicate that so combines with. For it, I follow Moulton (2015) in assuming that it refers to salient propositional content, without speculating on whether the associated proposition is part of the CG or
not. This explains the context-sensitive distribution of so, as well as the finding that so more is often used in response to questions. Following Farkas &
Bruce (2009:24), I assume that propositions denoting polar questions are not
part of the CG until the ‘asker’ (implicitly) signals agreement with the answer,
whereas affirmation of an assertion can happen ‘unsignalled’. Reference to asserted propositions by so is thus only possible in restricted environments; for
example in rejecting responses such as I don’t believe so or if the assertion was
not confirmed yet by other speakers (e.g. 2).
1. A: Is John coming tonight?
B: I believe {so | ? it}.
2. It will rain tomorrow. I know so, because I checked the weather report.
References: • Cornish, F. (1992): So Be It: The Discourse-Semantic Roles of So and It. JoS. • Farkas,
D. F. & K. B. Bruce (2009): On reacting to assertions and polar questions. JoS. • Moulton (2015): CPs:
Copies and Compositionality. LI. • Needham, S. M. (2012): Propositional anaphora in English. PhD thesis, Carleton University Ottawa.
258
AG 10 · Gebäude B4 1, Raum 0.07
Complementation strategies with the verb ‘know’ in Balkan Turkish
compared to Standard Turkish and Macedonian
Julian Rentzsch
FON University Skopje/JGU Mainz
Liljana Mitkovska
FON University Skopje
[email protected]
[email protected]
Donnerstag
09.03.2017
10:00 – 10:30
B4 1, 0.07
AG10
Standard Macedonian and Standard Turkish differ greatly in how they link
clausal complements to verba dicendi et sentiendi: Macedonian uses finite
verb forms and free complementizers in complement clauses, while Standard
Turkish generally prefers verbal nouns, i.e. bound non-finite verb forms. Both
languages possess means to encode the opposition [±factual]. The Western
Rumelian Turkish (WRT) dialects in Macedonia use finite complementation
strategies much more frequently than ST, although non-finite strategies occur
as well in certain settings.
The paper will investigate and compare the complementation strategies attested with the corresponding equivalents for the verb ‘know’ in Slavic and in
the Turkish varieties spoken in Macedonia. The focus will be on how the semantic polysemy of the verb in the language varieties examined affects the CC
patterns. Thus in the course of the paper we discuss semantic issues related
to the opposition [±factual] and to shades of meaning within the non-factual
domain (e.g. the development to know → to be able). We address the question to
which degree language contact, internal developments and universal tendencies have contributed to the emergence of the picture in WRT and demonstrate
how structures in a language can adjust to reflect features of very different
structures of contact languages.
References: • Dixon, R.M.W. & Alexandra Y. Aikhenvald (eds.) (2006): Complementation. A cross-linguistic typology. Oxford: Oxford University Press. • Friedman, Victor A. (2003): Turkish in Macedonia
and beyond. Studies in contact, typology and other phenomena in the Balkans and the Caucasus. (Turcologica 52.) Wiesbaden: Harrassowitz. • Noonan, Michael (1985 [2007]): Complementation. In: Shopen,
Timothy (ed.). Language typology and syntactic description. Vol. 2: Complex constructions. Cambridge:
Cambridge University Press, 52–150.
259
AG 10 · Polysemy and coercion of clause-embedding predicates
Clause embedding sound emission verbs
Donnerstag
09.03.2017
11:15 – 12:15
B4 1, 0.07
Stephen Wechsler
The University of Texas
[email protected]
AG10
Vocal sound emission verbs like scream often take clausal complements:
(1)
I remember screaming that I wanted him back every night for weeks.
(interview on ABC News, 2013)
In this usage the sounds produced must be verbal utterances whose semantic
content the clausal complement expresses. The clause cannot report, e.g., the
content of a non-linguistic scream:
(2)
The infant screamed {when / ?#that} he was hungry.
In (1), a verb meaning ‘Vocalize in a type x manner’ is extended to also mean
‘Say (content of complement clause) by vocalizing in a type x manner’. What
gives rise to this polysemy pattern?
One source of polysemy is sense indeterminacy: multiple alternative word
senses are consistent with a given reference context (cf. Quine’s 1960 “gavagai problem”; and Erk et al. 2012). Word senses pick out regularities across
observed situations, so polysemy results when there are multiple correlated
regularities. We hypothesize that the stronger the correlation between senses
s1 and s2 across described situations, the more likely that a word meant to denote s1 is interpreted as denoting s2 and thus that s1 and s2 are senses of a
single polysemous word. E.g. autohyponymy is common because all upwardentailing contexts consistent with s1 are also consistent with s1’s hypernyms,
and all downward-entailing contexts are consistent with s1’s hyponyms. Turning to verbs where s1 is ‘vocalize in a type x manner’, we expect s2 to be an
activity that correlates with type x vocalizations. Whenever the vocalizations
are human speech sounds (as in (1)) then the specific correlating s2 is the act
of saying something by making a type x sound.
References: • Quine, Willard Van Orman (1960): Word and object. Cambridge, MA: MIT Press. • Erk,
Katrin, Diana McCarthy, & Nicholas Gaylord (2012): Measuring word meaning in context. Computational Linguistics, November, 501–44.
260
AG 10 · Gebäude B4 1, Raum 0.07
Against classifications of complement-taking predicates: the case of
mental verbs
Natalia Serdobolskaya
Russian State University for the Humanities
Moscow State University of Education
Donnerstag
09.03.2017
12:15 – 12:45
B4 1, 0.07
AG10
[email protected]
While describing complementation, many grammars and special studies use
semantic classifications of complement-taking predicates (CTPs) as a descriptive tool. The claim of my paper is that such an approach is not effective due to
polysemy of CTPs, which influence the range of complement clause types the
CTPs take. In support of my claim I consider the data of several non-related
languages.
I focus on the verbs with the meaning ‘think, believe, seem’, which have
been classified as verbs of cognition in Givón (1980), propositional attitude in
Noonan (1985), propositional attitude (positive) in Hengeveld (2008–2009),
thinking in Dixon, Aikhenvald (2006). I describe the polysemy patterns of
CTPs meaning ‘think, believe, seem’ and the complement types they can take.
The described meaning shows is polysemous with the meanings of CTPs belonging to different classes, including mental (‘dream’, ‘remember’), emotive
(‘hope’, ‘fear’), perception (‘feel’), wishing (‘want’), intention CTPs (‘to be going to’).
I show that an adequate view on CTPs and the distribution of complement
types in a given language may be obtained through an analysis of basic meanings of CTPs and cannot rely on a priori classifications of CTPs or semantic
types of complements. This entails the necessity of a lexical typological research even in works on complementation that strive to limit themselves to
syntax.
References: • Dixon, R.M,W. & A. Aikhenvald (eds.) (2006): Complementation. A
cross-linguistic typology. Oxford University Press. • Givón, T. (1980): The binding hierarchy and the typology of complements. Studies in language 4(4), 333–377.
• Hengeveld, K. (2008–2009): The typological questionnary on complement clauses.
[https://www.eva.mpg.de/lingua/tools-at-lingboard/questionnaire/complementclauses_description.php] • Noonan, M. (1985): Complementation. In T. Shopen (ed.). Language
typology and syntactic description 2: Complex constructions. Cambridge University Press, 42–140.
261
AG 10 · Polysemy and coercion of clause-embedding predicates
‘Chameleon verbs’: variation in argument realization
Donnerstag
09.03.2017
13:45 – 14:45
B4 1, 0.07
Kerstin Schwabe
ZAS Berlin
[email protected]
AG10
There are polysemous German matrix predicates the argument realization
of which varies while the argument structure remains unchanged – e.g.
(es/darüber) diskutieren ‘discuss’ and (es/davon) hören ‘hear’. Their propositional argument either exhibits accusative or oblique case. Whereas accusative case expresses a direct relationship between the matrix subject and
the embedded clause, the prepositional case indicates that something else is
involved in the relationship. Most of the predicates denote an utterance event
if used with an accusative propositional argument. If such a predicate denotes
additionally an emotive attitude (mosern ‘grumble’), it can be used with the
prepositional correlate darüber. Then, it relates to a proposition that follows
from the matrix subject’s knowledge, that is, to a fact. As for predicates that
denote mental activities (sinnieren ‘ponder’) or the manner or purpose of an
utterance (polemisieren ’polemicize’), they relate to propositions that are contingent with or contradict the matrix subject’s knowledge if used with the
propositional correlate. Another subtype of ’chameleons’ consists of verbs like
wissen ‘know’ or hören ‘hear’. If they co-occur with an es-correlate, their embedded proposition is a fact. If they are used with the prepositional correlate
davon, the embedded proposition follows from what the matrix subject knows
or has heard. A further subgroup consists of decision predicates like abstimmen
‘vote’. Their oblique argument relates to the question that has to be decided on
whereas their accusative argument denotes the result of the decision.
262
AG 10 · Gebäude B4 1, Raum 0.07
Coercing propositional anaphora
Freitag
10.03.2017
11:30 – 12:30
B4 1, 0.07
Itamar Kastner
Humboldt-Universität zu Berlin
[email protected]
AG10
The meaning of a clause-embedding predicate is often determined by its
complement: know that ̸= know how (e.g. Grimshaw 1979). The current presentation discusses how meaning is conditioned by the syntactic category of the
complement: CP (clausal) or DP (nominal), mirroring the contrast between a
proposition and an entity. Taking into account predicates that exceptionally
coerce anaphoric complements, we are led to a theory of complementation that
has consequences beyond clausal embedding.
Definite DP complements are salient in the discourse whereas CP complements serve to introduce a new topic to the discourse. Once this distinction
is implemented in the syntax, the resulting system is able to make a number
of predictions regarding the factivity, presupposition and interpretation of CP
and DP complements (Kastner 2015).
Recently, however, Elliott (2016) has called attention to a set of propositional
anaphora: nominals with operators whose interpretation depends on another
DP in the clause or in the immediately accessible discourse. These include everything, the same NP and it. Strikingly, these DPs can serve as complements to
a handful of verbs that do not otherwise take DPs as complements, in particular think.
I propose that a clause embedded under propositional anaphors is coerced
into an entity whose propositional content must be filled in by an immediately
accessible proposition. As far as selection goes, these results imply that any
verb is compatible with a DP complement as long as the semantics is satisfied.
This much is already known from cognate objects, where unergative verbs are
able to take a restricted set of internal arguments.
In sum, under specific conditions it is possible to coerce a propositional
anaphor DP even if the verb is otherwise incompatible with nominal complements. This supports a theory separating syntactic selection from semantic
computation, where a verbal root constrains the latter but not the former.
References: • Grimshaw, J. (1979): Complement Selection and the Lexicon. Linguistic Inquiry 10,
279–326. • Kastner, I. (2015): Factivity mirrors interpretation: The selectional requirements of presuppositional verbs. Lingua 164, 156–188. • Elliott, P. D. (2016): Explaining DPs vs. CPs without syntax.
In: Proceedings of CLS 52.
263
AG 10 · Polysemy and coercion of clause-embedding predicates
Utterance-predicate complementation
Freitag
10.03.2017
12:30 – 13:30
B4 1, 0.07
Marie-Luise Lind Sørensen
University of Copenhagen
Kasper Boye
University of Copenhagen
[email protected]
[email protected]
In a classic work on complementation, Noonan (2007) restricts the term “utterance predicate“ to predicates such as ‘say’, ‘tell’ and ‘ask’ in constructions
with a propositional complement, as in (1a). In contrast, constructions like (1b)
are taken to involve a distinct type of predicate referred to as “manipulative”
(examples from Noonan 2007:120).
(1)
a.
b.
Floyd told Zeke that Roscoe burried the mash.
Floyd told Zeke to bury the mash.
Thus, the term “utterance predicate” is defined as a cover term for predicates
that describe assertions or polar questions, but not for predicates that describe directive speech acts. Likewise, Cristofaro’s (2013) study of utterancepredicate complementation excludes constructions like (1b). In this paper, we
argue that these approaches to utterance predicates are less than ideal. Firstly,
they miss generalizations across predicate types: for example, perception- and
knowledge-predicate complements often exhibit contrasts that are both morphosyntactically and semantically similar to that in (1). The goals of this paper
are: 1) to document the parallel between complements of utterance-predicates
and other predicate types, 2) to test Cristofaro’s claim that utterance-predicate
complements tend to be balanced, and 3) to argue that assertive and question
predicates take propositional complements,while directive predicates take
state-of-affairs complements (Boye 2012: 410-411). The study is based on data
from a genetically stratified sample of 173 languages.
References: • Boye, K. (2010): Reference and clausal perception-verb complements. Linguistics 48(2), 391-430. • Cristofaro, S. (2013): Utterance Complement Clauses. The World Atlas
of Language Structures Online. Leipzig: Max Planck Institute for Evolutionary Anthropology.
http://wals.info/chapter/128 (19 November, 2015). • Noonan, M. (2007): Complex constructions. In:
Timothy Shopen (ed.), Language Typology and Syntactic Description, vol. 2. Cambridge: Cambridge University Press.
264
AG11
Coercion Across Linguistic Fields (CALF)
Hana Filip & Laura Kallmeyer
Heinrich-Heine-Universität Düsseldorf, Heinrich-Heine-Universität
Düsseldorf
[email protected], [email protected]
Raum: B4 1, 0.06
Short description
The goal of this workshop is to discuss constraints on the workings of coercion
operators that are needed in different parts of grammar, not only in semantics,
pragmatics, but also in morphology, syntax, phonology and in computational
linguistics. What are the similarities and differences in the coercive operations
across these different domains?
265
AG 11 · Coercion Across Linguistic Fields (CALF)
Coercion as a methodological tool
Mittwoch
08.03.2017
14:15 – 14:45
B4 1, 0.06
Paul Dekker
University of Amsterdam
[email protected]
AG11
Proposals for semantic analysis are often based on intuitive judgments by the
intended theoretical audiences about the (un-)grammaticality, (non-) wellformedness, (in-)felicity, or (im-)propriety of certain linguistic constructions.
Suspending an evaluation of the soundness of such an empirical basis, it can
hardly escape notice that such judgments are not uniform. It appears to be impossible to distinguish, e.g., ungrammaticality from sheer oddness if not from
within a theoretical framework, which the semantic proposal often is intended
to contribute to. Also, the marks of imperfection, ranging from ∗ and ? to ∗∗
and ??? , often have to remain without unambiguous explanation. Often, but
not often enough, these marks come with a ceteris paribus qualification, “all
other things being equal”, a qualification that is widely accepted, but scientifically suicidal if referenced to a non-existent theory of interpretation specifying what these other things are.
However, once we give up the hypothesis that natural language can be totally captured by a system of structural rules, the ceteris paribus assumption
can be turned inside out, and figure as a methodological tool that may provide
support for specific semantic hypotheses. For if an explanation of a certain infelicity (ungrammaticality, …) consists in the presence or absence of a particular semantic feature of an example deemed infelicitous (ungrammatical, …), a
critical and creative linguist can construe a context that coerces absence of presence of precisely that semantic feature. If the proposed explanation is sound,
the infelicity (ungrammaticality, …) disappears.
Such a test or tool, stated as an “interference principle” in Dekker (2014),
builds on the assumption that many semantic features, like that of countability, animicity, telicity, perfection, present, …, are not rigidly encoded in the
grammar, and therefore in principle subject to contextual coercion. The principle can be seen to have been at work in actual semantic theorizing. With my
talk I want to illustrate its workings and outline its scope.
References: • Dekker, P. (2014): The Live Principle of Compositionality, in: D. Gutzmann et al., Approaches to Meaning, Brill, Leiden.
266
AG 11 · Gebäude B4 1, Raum 0.06
Immediate commitment, but no evidence for a coercion cost,
in individual/degree polysemy
Margaret Grant
Humboldt Universität zu
Berlin
[email protected]
Mittwoch
08.03.2017
14:45 – 15:45
B4 1, 0.06
Sonia Michniewicz
University of Toronto
Jessica Rett
University of California
Los Angeles
[email protected]
[email protected]
We present two experiments measuring eye movements during reading that
investigate the processing of individual and degree interpretations of DPs. We
conclude that there is evidence for immediate commitment to a single representation during processing (unlike other types of polysemy, see Frisson
2009), however we fail to find evidence for a cost to processing a dispreferred
or potentially semantically enriched representation (unlike other types of semantic coercion, see Traxler et al. 2002).
Rett (2014) shows that DPs have an available degree reading in the appropriate context (e.g., 1), and proposes that these degree readings are related to the
canonical individual readings via a null measure phrase operator.
(1)
Four/the/many pizzas ...are vegetarian. (ind.) ...is too many. (deg.)
Experiment 1 tested whether readers commit to a single analysis during realtime processing, and if so whether their preferred interpretation is constant
across DP types (definites, numerals and many-DPs). We measured reading
times on sentence continuations that were compatible with either a degree or
an individual interpretation. Our results reveal immediate commitment during processing, although interpretations differ across DP types. With individual continuations, there was an advantage for definite DPs, and with degree
continuations, there was an advantage for numeral DPs.
Experiment 2 tested sentences with predicates that require a degree interpretation of the subsequent definite DP (e.g., increase) or ones that require an
individual interpretation (e.g., soak). Eye movement measures fail to show a
penalty for processing the definite DP following a degree predicate.
References: • Frisson, S. (2009): Semantic underspecification in language processing. L&LC, 3(1):
111-127. • Rett, J. (2014): The polysemy of measurement. Lingua, 143: 242-266. • Traxler, M. J., Pickering, M. J., and McElree, B. (2002): Coercion in sentence processing: Evidence from eye-movements
and self-paced reading. JML, 47: 530-547.
267
AG11
AG 11 · Coercion Across Linguistic Fields (CALF)
Evidential effects and mismatch resolution
Mittwoch
08.03.2017
16:30 – 17:00
B4 1, 0.06
Victoria Escandell-Vidal
UNED
[email protected]
Consider the Spanish examples in (1) and (2):
(1a)
María
es
María
be(ser).prs.3sg
‘María is young.’
(2a)
Juan va a venir
mañana.
Juan go.come.prs.3sg tomorrow.
‘Juan will come tomorrow.’
AG11
joven.
young
(1b)
joven.
María está
María be(estar).prs.3sg young
‘María is young [I saw her.]’
(2b)
Juan venía
mañana.
Juan come.pst.3sg tomorrow.
‘Juan will come tomorrow.
[Someone said so]’
The states-of-affairs conveyed in (1) and (2) are basically the same; however,
the b. examples have an additional interpretive feature: (1b) is systematically
understood as conveying that the speaker is the direct source for the assertion,
whereas in (2b) the content has to be obligatorily attributed to someone else.
In neither of the two cases is there any overt indicator responsible for the evidential reading.
The aim of this talk is to argue that this additional evidential content arises
as the result of a pragmatic process of mismatch resolution. In both cases, an
acquisition-of-information event has to be inferred to avoid the conflict in aspectual and temporal anchoring. The mismatch obtains under very specific
conditions and is solved in a fully predictable way.
The analysis of these phenomena has implications for the design and properties of grammar, and provides new insights on the relations between linguistic
form and interpretation.
References: • Davis, Ch. et al. (2007): The pragmatic values of evidential sentences. SALT 17, 71-88.
• de Swart, H. (2011): Mismatches and coercion. In C. Maienborn et al. (eds.). Semantics. Berlin: Mouton de Gruyter, 574-596.
268
AG 11 · Gebäude B4 1, Raum 0.06
Interpreting quantifiers: the case of ligeramente + A in Spanish
Silvia Gumiel-Molina
Universidad de Alcalá
[email protected]
Norberto MorenoQuibén
Universidad de Alcalá
Mittwoch
08.03.2017
17:00 – 17:30
B4 1, 0.06
Isabel Pérez-Jiménez
Universidad de Alcalá
[email protected]
[email protected]
The aim of this talk is to provide a description of the distribution of adjectives with the quantifier ligeramente (‘slightly’) in Spanish. Kennedy & McNally (2005) propose that degree modifiers are sensitive to the dimension of
the scale whereas Sassoon & Toledo (2012) propose that the element accessed
by the quantifier is not the dimension of the scale but the standard of comparison. In this talk, we assume the proposal developed in Gumiel, Moreno & Pérez
(2015), in which we propose that the distribution on copular verbs in Spanish
depends on the comparison class of the adjective: Estar appears whenever a
gradable adjective merges with a within-individual comparison class; ser appears when a gradable adjective merges with a between-individuals comparison class. We also assume that the pragmatic component can access the semantics of the syntactic structure obtained by an intrusion that consist of repairing
malformed logical forms. This pragmatic intrusion needs a syntactic element
to be triggered, such as for-phrases or the nature of the subject, which improve
the unacceptability of certain sequences ligeramente + adjective. We argue that
ligeramente is a minimizer from the semantic point of view, therefore, it will
be compatible with those adjectives that allow to obtain a minimal standard.
In those cases in which the semantics of the minimizer is incompatible with
the adjective, by lacking the minimal point, the pragmatic intrusion facilitates
the well formedness of the construction by interpreting a functional standard
(Kagan & Alexeyenko 2010, Bylinina 2012).
References: • Bylinina, L. (2012). Functional standards and the absolute/relative distinction. Proceedings of Sinn & Bedeutung 16, ed. A. Guevara, A. Chernilovskaya, R. Nowen, 1:141-157. Cambridge, MA:
MITWPL. • Gumiel-Molina, S., N. Moreno-Quibén & I. Pérez-Jiménez (2015). “Comparison classes
and the relative/absolute distinction: a degree-based compositional account of the ser/estar alternation in Spanish”. Natural Language & Linguistic Theory 33(3). 955–1001. • Kagan, O. & S. Alexeyenko
(2011). “Degree modification in Russian morphology: The case of the suffix-ovat”. Proceedings of Sinn
& Bedeutung, vol. 15, 321–335. • Kennedy, C. & L. McNally (2005). “Scale Structure, Degree Modification, and the Semantics of Gradable Predicates”. Language 81(2). 345–381. • Toledo, A. & G.W. Sassoon
(2011). “Absolute vs. Relative Adjectives - Variance Within vs. Between Individuals”. Proceedings of
SALT 21. 135–154.
269
AG11
AG 11 · Coercion Across Linguistic Fields (CALF)
Mass-count shifts and the mass-count distinction
Mittwoch
08.03.2017
17:30 – 18:00
B4 1, 0.06
Carmen Dobrovie-Sorin
CNRS/Université Paris 7
Marta Donazzan
Universität zu Köln
[email protected]
[email protected]
Sentences such as (1) in English have been analysed in the literature as cases
of coercion of a count N (apple) into a mass one, cf. the well-known Universal
Grinder Hypothesis (Pelletier 1979).
AG11
(1)
There is apple in the salad.
A different analysis is offered by Bale & Barner (2009), which relies on their
generalizations stated in (2)-(3):
(2)
No term that can be used in count syntax can also be used in a mass
syntax to denote individuals.
(3)
Some mass nouns (in the context of use) have individuals in their denotation and others do not.
B&B’s explanation of (2)-(3) relies on three assumptions: (i) nouns of the furniture-type are derived from roots that denote individuated join semi-lattices,
i.e., join semi-lattices that have individuals at their bottom, whereas all other
nouns are derived from roots that denote non-individuated ones; (ii) count
interpretations are obtained via a count functional head that turns a nonindividuated join semi-lattice into an individuated one; (iii) the mass functional head denotes the identity function. As a consequence, according to B&B
(2009), the sentence in (1) would not be obtained by coercing the count N into
mass, but it would rather display the baseline use of the NP apple in English,
i.e. its use when no functional structure is added on top of it. In this talk, we
argue that Gen(2) faces cross-linguistic challenges, in particular with respect
to languages that exhibit General Number (Corbett 2000), and we propose an
analysis of the cross-linguistic variation in the coercion effects triggered by
mass-count use of NPs. Our proposal relies on different algebraic structures
for the domains of mass and count, and on a cross-linguistic parametric difference at the interface with morphosyntax, specifically in the requirement
of morphosyntactic marking of Number.
270
AG 11 · Gebäude B4 1, Raum 0.06
References: • Bale, A. & Barner, D. (2009): The interpretation of functional heads. Using comparatives to explore the mass/count distinction. Journal of Semantics 26, 217-252. • Corbett, G. (2000):
Number. CUP • Pelletier, F. (1979). Non-singular reference: some preliminaries. Mass terms: some
philosophical problems. 1-14.
Two takes on coercion and co-composition:
combining distributional and formal semantics
Donnerstag
09.03.2017
09:00 – 10:00
B4 1, 0.06
Nicholas Asher
CNRS/IRIT
AG11
[email protected]
In this talk, I explore an integration of a formal semantic approach to lexical
meaning and an approach based on distributional methods. First, I sketch the
outlines of a formal semantic theory (from Asher 2011) that has the promise
to combine the virtues of both formal and distributional frameworks. I then
proceed to develop an algebraic interpretation of that formal semantic theory
and show how at least two kinds of distributional models make this interpretation concrete. Focusing on the case of adjective-noun composition, I compare
several distributional models with respect to the semantic information that
a formal semantic theory would need, and we show how to integrate the information provided by distributional models back into the formal semantic
framework.
References: • Asher, N. (2011): Lexical Meaning in Context: A Web of Words. Cambridge: CUP.
Mass-to-count coercion in ‘granular’ nouns
Donnerstag
09.03.2017
10:00 – 10:30
B4 1, 0.06
Peter Sutton
Hana Filip
Heinrich-Heine-Universität Düsseldorf Heinrich-Heine-Universität Düsseldorf
[email protected]
[email protected]
Background: Nouns which denote entities that come in grains, granules,
flakes, display cross- and intralinguistic variation in mass/count lexicalisation
patterns (count oats, lentil-s vs. mass oatmeal, čočka (‘lentil’, Czech)). ‘Granular’
271
AG 11 · Coercion Across Linguistic Fields (CALF)
AG11
Ns exhibit puzzling restrictions regarding mass-to-count coercion. E.g., rice
can be coerced into a portion (e.g. bowls) count reading: Even with two rices between four of us, it didn’t all get eaten...1 ), but strongly resists coercion into an
individual unit reading: ??Three rice(s) fell of my fork. (Int: grain(s) of rice). In
this paper, we will provide a formal model of such puzzling restrictions on coercion.
Proposal: The account is based on but adapts Sutton & Filip (2016a,b) and is
partly inspired by Chierchia (2010). All predicates are interpreted at precisification contexts πi ∈ Π which model extension changes in predicates across
contexts of use. For example, single grains of rice count as rice in some contexts (e.g. food allergy), but are too little in quantity to count as rice in others
(e.g. making dinner). All mass nouns are saturated with the null precisifica∩
tion context in the lexicon, π0 : Pπ0 = Pπi ∈Π (at which only Ps that are Ps at
all contexts are in P at π0 ). Single grains can be denoted using explicit unit extracting classifiers (e.g grain of). This requires shifting the precisification context to one at which single grains are accessible. Portions can be denoted using
explicit container classifier expressions (e.g. bowl of). These classifiers do not
require a context shift. Classifiers triggered by numericals modifying mass Ns
cannot induce the operation of re-writing the precisification context in the
lexicon, hence single grains are inaccessible via mass-to-count coercion while
whole portions are not.
References: • Chierchia, G. (2010): Mass nouns, vagueness and semantic variation. Synthese
174:99–149. • Sutton, P., Filip, H., (2016a): Mass/count variation, a mereological, two-dimensional
semantics. Forthcoming in: The Baltic International Yearbook of Cognition, Logic and Communication. • Sutton, P., Filip, H., (2016b): Counting in context. SALT 26, 350–370.
Revisiting wieder: a restitutive prefix and its coerced object
Donnerstag
09.03.2017
11:15 – 11:45
B4 1, 0.06
Andreas Blümel
Universität Göttingen
Hans-Joachim Particke
Universität Göttingen
[email protected]
[email protected]
Starting with von Stechow’s (1996) structural account of the repetitiverestitutive ambiguity of the German adverb wieder ‘again’, we investigate its hitherto unstudied verbal prefix counterpart as in wiederauferstehen ‘resurrect’. As it is string-identical to the syntactic adverb plus verb,
1
http://durhamfoodanddrink.com/gilesgate-tandoori-durham/ (accessed 10.2016)
272
AG 11 · Gebäude B4 1, Raum 0.06
we use deverbal nominalizations like Wiederauferstehung ‘resurrection’ to
avoid confusion. Wieder-prefixes occur for the most part with prefixed or
particle verbs and often is the prefix/particle obligatory, see for instance
Wieder*(er)öffnung ‘reopening’. A few cases without additional prefix/particle
exist (e.g. Wiederkehren ‘returning’). After a series of tests, we observed that
only restitutive readings are available, suggesting that prefixal wieder does
not modify the event denoted by the verb but rather a structurally represented result state (pace Lieber’s 2004:147 account of English verbal prefix re). This observation raises the expectation that wieder-prefixation is unavailable with activity verbs (unlike the homophonous syntactic adverb). Minimal
pairs like *Wiederschlafen vs. Wiedereinschlafen confirm this prediction. Furthermore, prefixal wieder requires an object (transitive object or object of unaccusative, cf. Horn 1980 on English re-). As many German verbs appear to
form unaccusatives by means of prefixes or particles, their preponderance
with wieder is not surprising. We adopt a DM-style analysis of Marantz (2007)
according to which the restitutive prefix directly selects the (underlying) object DP or pro (in nominalizations) respectively. Semantically, the DP/pro is
coerced from an expression of type e into a change of state event. Wieder adds
the presupposition that the end state, which is part of the coerced change-ofstate denotation of the DP, has existed before. Various units in the structure
name the result state. After an informal semantic characterization, the crossand intra-linguistic (un)availability of forms is considered.
AG11
References: • Horn, L. (1980): Affixation and the Unaccusative Hypothesis. In: CLS 16, 134–146.
• Lieber, R. (2004): Morphology and Lexical Semantics. Cambridge: CUP. • Marantz, A. (2007): Restitutive re- and the first phase syntax/semantics of the VP. Ms. • Von Stechow, A. (1996): The Different
Readings of Wieder ‘Again’: A Structural Account. In: Journal of Semantics 13, 87–138.
Postnominal temporal adverbs in the German prefield
Eva Csipak
Universität Konstanz
Sarah Zobel
Universitäten Göttingen/Tübingen
[email protected]
[email protected]
Donnerstag
09.03.2017
11:45 – 12:45
B4 1, 0.06
Temporal adverbs like gestern (‘yesterday’) can occur post-nominally in the
German prefield (e.g. Alexiadou et al. 2007, Bücking 2012, Gunkel & Schlotthauer 2012). Adverbs in this position are nominal modifiers and do not temporally locate the event described by the verbal predicate. Hence, sentences
273
AG 11 · Coercion Across Linguistic Fields (CALF)
with postposed temporal adverbs in the prefield (1a) and sentences with the
adverb in the middle field (1b) are not synonymous.
(1)
a.
b.
AG11
{Das Mädchen / der Unfall} gestern war schlimm.
the girl
the accident yesterday was awful
{Das Mädchen / der Unfall} war gestern schlimm.
the girl
the accident was yesterday awful
In (1a), gestern locates times associated with the individuals/events that are
described by girl/accident. Hence, it functions as a restrictive modifier of the
noun. In (1b), gestern locates the state of being awful; the girl and the accident
are uniquely identified independently.
Temporal adverbs are lexically specified as modifiers of the temporal domain and, hence, combine freely with temporal nouns (e.g., der Nachmittag
gestern, ‘yesterday afternoon’). With individual nouns and event nouns, we argue, temporal adverbs can only combine via coercion. To capture this combinatoric flexibility, we adopt the system developed in Asher 2011. Hence, temporal
adverbs, like gestern in (2), license dependent types.
(2)
λP.λt.λπ.yesterday(t, π ∗ argyesterday
: time − ϵ(hd(P))) ∧ P(π)(t)
1
The times that are inferred with individual and event nouns need to be intimately connected to the individuals/events that are described by that noun.
For events, these are their run-times (τ (e)). For individuals, we coerce the runtimes of events in which the individual participated.
References: • Alexiadou, A. L. Haegeman & M. Stavrou (2007): Noun Phrase in the Generative Perspective. De Gruyter. • Asher, N. (2011): Lexical meaning in context: a web of words. Cambridge University
Press. • Bücking, S. (2012): Kompositional flexibel. Stauffenburg. • Gunkel & Schlotthauer (2012): Adnominale Adverbien im europäischen Vergleich. In: Deutsch im Sprachvergleich, 273–300.
Phonological coercion in Pawnee
Donnerstag
09.03.2017
13:45 – 14:15
B4 1, 0.06
Masanori Deguchi
Western Washington University
[email protected]
This study examines certain phonological coercion processes in Pawnee, a
North Caddoan language (Parks 1976). As shown in the contrast between (1)
and (2), the voicelss alveolar stop /t/ is coerced as an affricate [c] when fol-
274
AG 11 · Gebäude B4 1, Raum 0.06
lowed by a homorganic consonant.
(1)
a.
b.
c.
/t/+/t/ → [ct]: /ta-t+tau:t-∅/ → [tactaPu] ‘I stole it.’
/t/+/s/ → [ct]: /ta-t+sa-∅/ → [tacta] ‘I’m lying.’
/t/+/c/ → [ct]: /ta-t+cak-∅/ → [tactat] ‘I shot it.’
(2)
a.
b.
/t/+/p/ → [tp]: /ta-t+paks/ → [ta-tpaks] ‘I hit the boy…’
/t/+/k/ → [tk]: /ta-t+kawi/ → [tatkawi] ‘I’m grinding it.’
I demonstrate that /t/ undergoes coercion in order to improve the transition
between syllables. Specifically, I argue that /t/ in the coda is coerced as [c],
rendering it more sonorous than the onset of the following syllable. Based on
this observation, I propose a syllable contact constraint (Vennemann 1988) defined in terms of sonority (Davis & Shin 1990). Since this constraint makes
reference to the relative sonority between segments, it accounts for not only
the coercion of /t/ into [c] (in the coda) but also the coercion of /c/ into [t]
(in the onset) as in (1c). I further propose that this constraint be conjoined with
a constraint that millitates against homorganic clusters (Smolensky 1995). As
support, I discuss predictions for the coercion patterns in geminates.
In summary, the phonological coercion discussed above is triggered for syllable contact, but it does not target a specific segment or a specific syllable position; while coda segments are usually coerced, onset segments are also coerced
occasionally.
References: • Davis, S. & S.-H. Shin (1990): The Syllable Contact Constraint in Korean. Journal of East
Asian Linguistics 8, 285–312. • Parks, D. (1976): A Grammar of Pawnee. Garland Publishing. • Smolensky, P. (1995): On the Structure of the Constraint Component Con of UG. Handout, UCLA. • Vennemann,
T. (1988): Preference Law for Syllable Structure. Mouton de Gruyter.
275
AG11
AG 11 · Coercion Across Linguistic Fields (CALF)
Donnerstag
09.03.2017
14:15 – 14:45
B4 1, 0.06
AG11
Coercion in loanword adaptation
Ruben van de Vijver
Heinrich-HeineUniversität
Düsseldorf
Vicky Tsouni
Heinrich-HeineUniversität
Düsseldorf
Kim Strütjen
Heinrich-HeineUniversität
Düsseldorf
[email protected]
[email protected]
[email protected]
Loan words are often adapted to fit the phonotactics of the borrowing language.
Such adaptation is a phonological instantiation of coercion. We will present
the results of a production experiment and a nonce word learning experiment
with Japanese speakers.
Japanese has a very restrictive syllable phonotactics. Codas can only be
nasals or geminates that are homorganic with the following onset. Sphinx, for
example, is borrowed as [sWfiNkWsW] (Dupoux et al. 1999). The clusters are bro˚
˚
>
ken up by vowels. Zeitgeist is borrowed as [tsaitogaisWto]. The epenthetic
˚
vowel after t is [o]–not [W]–for phonotactic reasons (Kubozono 2015). Perceptually the vowels are real (Dupoux et al. 1999). Kwon (2017) found that more
proficient speakers of the host language adapt loans less.
We investigated the properties of the epenthetic vowels in production and
whether they are used to store words in memory; both as a function of language experience. We will conduct two experiments, both in Germany and in
Japan. In the first one, we will present Japanese participants with audio recordings of German bisyllabic nonce words with intervocalic consonant clusters
that are illegal in Japanese (e.g. okto). The participants are asked to produce
each nonce in a carrier sentence. The clusters are then phonetically analyzed
to investigate whether there is a vowel and its quality. In the word learn experiment, we will associate the nonce words with fantasy animals. We will then
investigate whether the nonce okto is stored as [okto] or as [okWto]. We expect
˚
that the Japanese speakers in Germany are less likely to need coerced vowels
than the ones in Japan.
References: • Dupoux, E., K. Kakehi, Y. Hirose, C. Pallier and J. Mehler (1999): Epenthetic vowels in
Japanese: A perceptual Illusion? In: Journal of Experimental Psychology: Human Perception and Performance 25(6), 1568–1578. • Kubozono, H. (2015): Loanword phonology. In: Handbook of Japanese Phonetics and Phonology, H. Kubozono (Ed.), Mouton de Gruyter. • Kwon, H. (2017): Language experience,
speech perception and loanword adaptation: Variable adaptation of English word-final plosives into
Korean. In: Journal of Phonetics 60, 1–19.
276
AG 11 · Gebäude B4 1, Raum 0.06
Polysemy and coercion –
A frame-based approach using LTAG and hybrid logic
William Babonnaud
ENS Cachan, Université
Paris-Saclay
william.babonnaud@ens-
Laura Kallmeyer
Heinrich-HeineUniversität
Düsseldorf
Rainer Osswald
Heinrich-HeineUniversität
Düsseldorf
cachan.fr
[email protected]
[email protected]
Freitag
10.03.2017
11:30 – 12:00
B4 1, 0.06
In this work∗ , we propose an analysis of polysemy and coercion phenomena
within Lexicalized Tree Adjoining Grammar (LTAG) and frame semantics.
Consider for instance the inherent polysemy of book between a physical object reading (“The book is heavy”) and an information content reading (“The
book is interesting”). Following Pustejovsky (1998), our frame structure for
book contains two nodes information and phys-obj respectively, related via a
content attribute (see frame on the right). Read can select such a dot object,
perception
perc-comp
agent
reading
ment-comp
book
stimulus
orderedoverlap
content
information
content
content
information
comprehension
but also enables coercion of its complement from information (“story”) or physobj (“blackboard”). In our analysis, the reading frame has a perceptual and a
mental component, related by an ordered-overlap. The former has an attribute
stimulus, and the latter has an attribute content that is identical to the content of the stimulus (see the left frame above). The syntax-semantics interface
specifies that the object can contribute either the stimulus value or its content.
Combining our two sample frames, the book node therefore necessarily unifies
with the stimulus and the two information nodes unify as well.
Further details and examples will be discussed in the talk.
∗
This work was supported by the German Research Foundation (DFG) within
CRC 991 and by ENS Cachan, Université Paris-Saclay.
References: • Pustejovsky, James (1998): The semantics of lexical underspecification. Folia Linguistica
32(3-4), 323–348.
277
AG11
AG 11 · Coercion Across Linguistic Fields (CALF)
Empirical evidence for the study of coercion mechanisms
in predicate-argument composition
Freitag
10.03.2017
12:00 – 12:30
B4 1, 0.06
Elisabetta Jezek
Università di Pavia
[email protected]
AG11
Within the context of building an Italian resource of corpus-driven typed
predicate argument structured (T-PAS, such as (1) [Human]-subj] raggiunge
‘reaches’ [[Location]-obj]) for linguistic analysis and NLP tasks (Jezek et al.
2014), we are building a repository of corpus-derived type mismatches in argument positions (for example in (1) “Abbiamo raggiunto l’isola alle 5” ‘We
reached the island at 5’ (matching) vs. “Ho raggiunto il semaforo e ho svoltato
a destra ‘I reached the traffic light and turned right’ (mismatch)), meant to
provide a structured dataset for studying semantic type coercions (Asher 2011,
Pustejovsky and Jezek 2008, Lauwers and Willems (2011)) both quantitatively
and qualitatively. Two main issues we are currently addressing are the following:
– examine how argument semantic type coercions are spread among verb
classes, i.e. which verbs tend to be more coercive than others. Preliminary results show that these include: aspectual verbs, communication verbs, perception
verbs, avoid verbs, forbid verbs, verbs of desire, directed motion verbs, verbs of motion using a vehicle;
– collect empirical data to clarify whether coercion mechanisms differ from
mechanisms of systematic meaning modulation licensed by the interplay of
lexical information, ontological complexity and contextual selection (Jezek
and Vieu 2014).
In the presentation, I will provide empirical evidence on both points above
and discuss how empirical evidence can be integrated in a formal model of
context-sensitive lexical semantics.
References: • Asher, N. (2011) Lexical meaning in context. CUP. • Jezek, E. and V. Quochi (2010):
Capturing Coercions in Texts: a First Annotation Exercise. In: LREC 2010. • Jezek, E. and L. Vieu
(2014) Distributional analysis of copredication: Towards distinguishing systematic polysemy from
coercion. In Proceedings of CLiC-it. • Jezek E., B. Magnini, A. Feltracco, A. Bianchini and O. Popescu
(2014): T-PAS: a resource of corpus-derived Typed Predicate Argument Structures for linguistic analysis and semantic processing. In: LREC 2014. • Lauwers P. and D. Willems (2011): Coercion: Definition
and Challenges. In: Linguistics. • Pustejovsky, J. and E. Jezek (2008): Semantic Coercion in Language:
Beyond Distributional Analysis. In: IJL.
278
AG 11 · Gebäude B4 1, Raum 0.06
An experimental study on coercion in Spanish adjectival phrases
José Manuel Igoa
Universidad Autó noma de Madrid
Maria del Carmen Horno
Universidad de Zaragoza
[email protected]
[email protected]
Scalar adjectival phrases may be classified in two broad categories: bounded
(or absolute), with adjectives such as dry, wet or full, and open (or relative),
with adjectives like tall, fat or clever. The former have a maximum and/or minimum degree and an internal standard of comparison, whilst the latter lack
specific limits and have a contextual standard of comparison. From a lexicalist standpoint, the difference lies in the distinction between two kinds of
scalar adjectives (Kennedy & McNally, 2005): absolute adjectives, which combine with quantifiers like totally, completely or slightly, and relative adjectives,
which usually combine with a quantifier like very.
The current experimental study was designed to test this. We created two
lists of sentences, one with absolute and the other with relative adjectives,
combined with two different quantifiers, slightly versus very, and put the sentences to test by means of a self-paced reading task. As a control condition, we
made up another list of sentences where the same adjectives under a metaphorical reading were combined with the “expected” quantifier.
Results showed significant differences between literal and metaphoric uses
of all adjectives. As for the combination with quantifiers, open adjectives
showed significantly longer reading times for adjectives with unexpected
quantifiers, which is interpreted as evidence for coercion. On the contrary,
bounded adjectives show no differences in reading times as a function of quantifiers, which seems to indicate that there is no coercion in this case. Thus, the
current results purport to show that absolute adjectives acquire their standard
of comparison in syntax, whilst relative adjectives do so on the basis of lexical
information.
References: • Kennedy, C. & L. McNally (2005) Scale structure, degree modification, and the semantics of gradable predicates. Language 81, 345-381.
279
Freitag
10.03.2017
12:30 – 13:00
B4 1, 0.06
AG12
Morphologische Variation – Theorie und Empirie
Antje Dammel & Oliver Schallert
Universität Freiburg, Ludwig-Maximilians-Universität München
[email protected], [email protected]
Raum: B4 1, 0.05
Short description
Die Kernfrage dieser AG ist, ob und wie sich moderne morphologische
Theorien in der Modellierung und Erklärung empirischer morphologischer
Variation fruchtbar machen lassen. Diskussionsbeiträge verschiedener theoretischer Ausrichtungen (z.B. Konstruktionsmorphologie, Paradigm Function Morphology) und Perspektiven auf morphologische Variation und ihre
Schnittstellen (areal, sozial, diachron, kognitiv) sind willkommen.
281
AG 12 · Morphologische Variation – Theorie und Empirie
Morphomes and variation: a Scandinavian perspective
Mittwoch
08.03.2017
13:45 – 14:45
B4 1, 0.05
Hans-Olav Enger
Universitetet i Oslo
[email protected]
AG12
Aronoff ’s (1994) morphome has been found useful also in diachrony, in particular in the study of Romance languages (e.g. Maiden 2016, Loporcaro 2013). Recently, however, this “morphomic approach” has been criticised (e.g. by Bowern 2015). In my talk, I’ll submit that some of this criticism is misplaced. The
arguments involve Scandinavian mostly, but some will involve Romance.
I wish to reflect on how some historical changes proceed, to see if the morphome concept is useful there. The indefinite plural of a subset of neuters has
been changed, in varieties of Danish, Swedish and Norwegian, from having no
suffix to having the suffix -r, and this is usually analysed as analogy from feminines. In this way, a mixed (heteroclitic) paradigm arises. However, some varieties that had got -r in these neuters innovate again, replacing -r by -n, and
this innovation must be given another account – but, intriguingly, it targets
the same neuters. Also in many varieties, there is a next step after the introduction of -r, a change involving the definite plural. Another case we shall go into
is the introduction, in the definite sg., of the old masculine suffix -a in some
few feminines in East Norwegian. Again, this results in a “mixed” paradigm.
The diachronic evidence indicates that we should not see inflectional classes as
‘flags’ (cf. e.g. Dammel et al. 2010, Enger 2010). I shall ask what implications,
if any, this has for the morphome.
An over-zealous hunt for morphomes may lead us to neglect semantic relations, yet some semantic relations traditionally postulated are perhaps rather
morphomic. If time allows, we shall consider that, too.
References: • Aronoff, M. (1994): Morphology by itself. MIT. • Bowern, C. (2015): Diachrony. In:
The Oxford Handbook of Inflection. OUP, 233–250. • Dammel, A. (2011): Konjugationsklassenwandel.
De Gruyter. • Dammel, A. et al. (2010): Strong verb category levelling. JGL 22(4), 337–359. • Enger, H.-O. (2010): How do words change their inflection class? LgSci 32(3), 366–379. • Loporcaro, M.
(2013): Morphomes in Sardinian Verb inflection. In: The boundaries of pure morphology. OUP, 137–160.
• Maiden, M. (2016): Some lessons from history. In: The morphome debate. OUP, 33–64.
282
AG 12 · Gebäude B4 1, Raum 0.05
The interaction of morphological and phonological variation
A case study on Zurich German
Mittwoch
08.03.2017
14:45 – 15:15
B4 1, 0.05
Anja Hasse
Universität Zürich / Surrey Morphology Group
[email protected]
Zurich German shows an interesting case of competing forms. The dative
cell of the indefinite article exhibits several forms: masc./neutr. eme and
emene, fem. ere and enere. These cell-mates, in turn, occur in several phonological variants. Besides the forms mentioned above, the forms can be reduced
for instance by aphaeresis (e.g. fem. nere as opposed to enere). Both the factors conditioning the morphological and the phonological variation are still
largely unknown. Despite being considered rare and diachronically unstable,
cf. Fehringer (2004: 285–28-6), morphological doublets have been reported in
several studies, cf. Fehringer (inter alia 2004) on Standard German, Thornton
(inter alia 2012) on Italian. While those studies are all based on written data,
studies on spoken language are still widely missing. Thus, understanding the
factors conditioning the preference of one form over another enables us not
only to explain the distribution of these competing forms in Zurich German,
but also to get an insight on overabundance in spoken language.
In my talk, I present results from a study on spontaneous spoken data of
Zurich German and discuss to what extent the morphological and the phonological variation are governed by the same phonological and syntactic factors.
In doing so, I compare overabundance, as a pure morphological phenomenon,
to phonological variation.
References: • Fehringer, C. (2004): How Stable are Morphological Doublets. A Case Study of @ ∼ Ø
Variants in Dutch and German. Journal of Germanic Linguistics 16(4), 285–329. • Thornton, A. (2011):
Overabundance (Multiple Forms Realizing the Same Cell): A Non-canonical Phenomenon in Italian
Verb Morphology. In: Morphological Autonomy. Perspectives from Romance Inflectional Morphology. OUP,
358–381. • Thornton, A. (2012): Reduction and maintenance of overabundance. A case study on Italian verb paradigms. Word Structure 5(2), 183–207.
283
AG12
AG 12 · Morphologische Variation – Theorie und Empirie
Mittwoch
08.03.2017
15:15 – 15:45
B4 1, 0.05
AG12
Phonotaktik jenseits der Silbe – Quantitative Analysen zur Relevanz
phonotaktischer Strukturen für die Morphologie
Alfred Lameli
Forschungszentrum Deutscher
Sprachatlas, Marburg
Alexander Werth
Forschungszentrum Deutscher
Sprachatlas, Marburg
[email protected]
[email protected]
Während die Silbe klassischerweise als diejenige Domäne gilt, auf der phonotaktische Regeln appliziert werden, ist wenig darüber bekannt, inwiefern
auch Morpheme zur Strukturierung von Lautverbindungen beitragen können.
Bereits Trubetzkoy (1939: 225) vermutet, dass dem Morphem diesbezüglich
eine besondere Relevanz zukommt, doch fehlt es an empirischen Untersuchungen, die zeigen, dass die Verkettung von Lauten innerhalb von Morphemen anderen Regularitäten folgt als innerhalb von z. B. Silben oder Wörtern.
Wir möchten an diesem Punkt ansetzen und die morphologische Relevanz
phonotaktischer Muster aus zwei Perspektiven heraus betrachten. Anhand
von natürlich-sprachlichen und phonetisch fein transkribierten Daten aus
182 Orten untersuchen wir auf der einen Seite, welchen Erklärungsmehrwert
das Morphem für die Strukturiertheit phonotaktischer Muster im Deutschen
besitzt. Darüber hinaus wollen wir im Vortrag zeigen, dass der Phonotaktik
einsilbiger Wörter im Deutschen ein indexikalischer Wert für die Differenzierung von Wortklassen und Wortarten zukommen kann. So werden in den
Daten z. B. einsilbige Grammeme deutlich häufiger mit einer minimalen CVStruktur realisiert als Lexeme. Diese Tendenz verstärkt sich deutlich, wenn
statt CV-Skelett eine feinere Differenzierung der Laute in Vokale, Obstruenten, Nasale und Liquide vorgenommen wird. Gleichzeitig lässt sich zeigen,
dass bestimmte Silbenstrukturtypen signifikant häufiger mit bestimmten
Wortarten auftreten, wobei teilweise sogar ein komplementäres Verhältnis für
die Präferenzen phonotaktischer Strukturen bei verschiedenen Wortarten zu
verzeichnen ist.
References: • Trubetzkoy, N. S. (1939): Grundzüge der Phonologie. Prag.
284
AG 12 · Gebäude B4 1, Raum 0.05
/t/-realization in German – A case for hybrid models?
Mittwoch
08.03.2017
16:30 – 17:00
B4 1, 0.05
Pia Bergmann
Universität Duisburg-Essen
[email protected]
Reduction and deletion of /t/ has been the centre of interest of several empirical studies in the recent past, displaying a complex interplay of phonological,
morphological and other factors in sound production (cf. e.g. Zimmerer et al.
2014). The proven relevance of usage-based and postlexical factors for the realization of words calls into question some basic assumptions of “traditional”
accounts of Lexical Phonology which suppose a division between lexical and
postlexical processes in word production (cf. e.g. Wiese 2000). To address this
problem, hybrid models have come into play, aiming to reconcile systematic
usage-based effects including phonetic detail with more formal processes and
abstract representations (cf. e.g. Hinskens et al. 2014). Against this backdrop,
the paper presents the results of a corpus study of spontaneous speech of ca.
600 instances of /t/ in complex German words. The dependent variables included categorical /t/-deletion as well as gradient durational reduction of the
cluster. Next to controlling for several possible influences for the realization of
/t/, the independent factors were token frequency and semantic transparency
of the complex word, as well as type frequency and semantic bleaching of
the first constituent. The results underline the reductional effect of token frequency on gradient duration as well as on categorical deletion of /t/. However,
gradient reductions and categorical deletions do not always react to the same
factors. All in all, the results can be seen to support the call for hybrid models
where on the one hand some gradient and categorical effects should be separated, while on the other hand usage-based and word-based information can
exert a systematic influence on sound production in complex words.
References: • Hinskens, F. et al. (2014): Grammar or lexicon. Or Grammar and Lexicon? Rule-based
and usage-based approaches to phonological variation. Lingua 142, 1–26. • Wiese, R. (2000): The
phonology of German. 2. ed. OUP. • Zimmerer, F. et al. (2014): Phonological and morphological constraints on German /t/-deletions. Journal of Phonetics 45, 64–-75.
285
AG12
AG 12 · Morphologische Variation – Theorie und Empirie
Mittwoch
08.03.2017
17:00 – 17:30
B4 1, 0.05
AG12
Zum Wandel des Drei-Formen-Plurals im salzburgisch-bayerischen
Grenzgebiet. Eine Pilotstudie zur intraindividuellen
morphologischen Variation
Lars Bülow
Universität Salzburg
Hannes Scheutz
Universität Salzburg
Dominik Wallner
Universität Salzburg
[email protected]
[email protected]
[email protected]
Im Untersuchungsgebiet ist im Basisdialekt noch der Drei-Formen-Plural in
der Verbkonjugation anzutreffen (1a). Die Verhältnisse sind allerdings auch
intraindividuell betrachtet instabil. Dabei ist insbesondere ein Wandel vom
Drei- zum Zwei-Formen-Plural zu beobachten. Häufig wird die Endung der
1. P. Pl. auf die 3. P. Pl. ausgedehnt (1b). Es kommt aber auch zu Übertragungen der 3. P. Pl. auf die 1. P. Pl. (1c). Bei manchen Sprechern ist sogar der
sprachgeschichtlich „verdrehte“ Zustand belegt.
(1)
a.
mi(a) keem-an
es kem-dds
se keem-and
b.
mi(a) keem-an
es kem-dds
se keem-an
c.
mi(a) keem-and
es kem-dds
se keem-and
In unserer Analyse geht es um genau diese intraindividuelle Variation, auf
deren Relevanz für die Prozesse der Sprachveränderung erstmals die Complex
Dynamic Systems Theory (CDST) aufmerksam gemacht hat. Die CDST betrachtet den Idiolekt als ein dynamisches und komplexes adaptives System (Ellis 2011), dessen Variation über die Zeit nicht als „Rauschen“ in den Daten,
sondern als eine wichtige Informationsquelle betrachtet werden muss. Wir
vergleichen erstens Aufnahmen von denselben Sprechern, die mehr als 10
Jahre auseinanderliegen. Diese stammen aus einem EuRegio-Projekt (Scheutz
2007) und aktuellen Erhebungen, die im Rahmen des SFB-Projekts „Deutsch
in Österreich“ durchgeführt wurden. Zweitens untersuchen wir die intraindividuelle Variation innerhalb kürzerer Zeitintervalle. Dafür werden fünf
Sprecher über einen Zeitraum von zwei Monaten fünfmal mit Hilfe eines
Sprachproduktionstests untersucht.
References: • Ellis, N.C. (2011): The Emergence of Language as a Complex Adaptive System. In: The
Routledge Handbook of Applied Linguistics. Routledge, 654–667. • Scheutz, H. (2007): Drent und herent.
EuRegio.
286
AG 12 · Gebäude B4 1, Raum 0.05
Veni, vidi, vici –
Ich bin gekommen, habe gesehen und gesiegt?
Perfektexpansion und Präteritumschwund im Deutschen
Mittwoch
08.03.2017
17:30 – 18:00
B4 1, 0.05
Hanna Fischer
Philipps-Universität Marburg
[email protected]
Das deutsche Perfekt (ich habe geschrieben) hat in seiner Verwendung stark
expandiert und das Präteritum (ich schrieb) zu großen Teilen marginalisiert.
Dieser Prozess wird umso deutlicher, wenn wir einen Blick auf die gesprochenen Varietäten (v.a. Dialekte) werfen. In den oberdt. Dialekten führte die Perfektexpansion zu einem Schwund der Präteritumformen. Dieser Prozess lässt
sich zum einen anhand dialektologischer Dokumente in seiner arealen Struktur beschreiben. Zum anderen kann anhand historischer und kontrastiver
Studien der Expansionsweg der Perfektform rekonstruiert werden, so dass
die Mechanismen hinter dieser für das dt. Verbalsystem so wichtigen Entwicklung sichtbar werden.
Der Vortrag geht anhand von Sprachkarten auf die Arealstruktur des
sich im Raum abbildenden Präteritumschwunds ein und benennt die Faktoren der verbspezifischen Formendistribution (u.a. Tokenfrequenz, Ausdrucksverfahren, Verbalsemantik). In einem zweiten Schritt wird der Expansionsprozess der Perfektform nachgezeichnet und in Form eines Modells operationalisiert, das die Bedeutungsbereiche zwischen „Perfektbedeutung“ und „Präteritalbedeutung“ mithilfe der Kategorien temporale. Verortung, Definitheit der temporalen. Verankerung und Gegenwartsrelevanz abgrenzt. Das Modell wird im Anschluss exemplarisch auf das regionalsprachliche Korpus des Projekts Regionalsprache.de (REDE) angewendet. Zur Frage
steht, ob die ober- und mitteldeutschent. Perfektformen stärker expandiert
haben als die niederdeutschent. Perfektformen und damit die Arealstruktur
in den Sprachkarten erklären.
References: • Lindgren, K. (1957): Über den oberdt. Präteritumschwund. Helsinki. • Rowley, A. (1983):
Das Präteritum in den heutigen dt. Dialekten. ZDL 50(2), 161–182. • Dentler, S. (1997): Zur Perfekterneuerung im Mittelhochdt. Göteborg. • Sapp, C. (2009): Syncope as the cause of Präteritumschwund. JGL 21(4), 419–450. • Amft, C. (2013): Das präteritale Konzept im Frühneuhochdt. Uppsala.
• Fischer, H. (2015): Präteritumschwund in den Dialekten Hessens. In: Dt. Dialekte. Steiner, 107–133.
287
AG12
AG 12 · Morphologische Variation – Theorie und Empirie
Das No-Blur-Principle und nominalmorphologische
Variation im Deutschen
Donnerstag
09.03.2017
09:00 – 09:30
B4 1, 0.05
Christian Zimmer
Freie Universität Berlin
[email protected]
AG12
Nach Carstairs-McCarthy (1994) hat jedes Allomorph eine eindeutige intralinguistische Bedeutung. Dies formuliert er in seinem No-Blur-Principle
folgendermaßen: „Within any set of competing inflectional affixal realizations for the same paradigmatic cell, no more than one can fail to identify
inflection class unambiguously“ (Carstairs-McCarthy 1994: 742). Demnach
sei jedes Morph entweder class-default (und käme in mehreren Flexionsbzw. Makroklassen vor) oder class-identifier (und käme in nur einer einzigen Flexions- bzw. Makroklasse vor), wobei es nicht mehr als einen classdefault geben könne (vgl. hierzu auch Enger 2007, 2013, 2016). Dies ist im
Großen und Ganzen mit der deutschen Nominalflexion kompatibel – wenn
man allerdings auch Variation berücksichtigt, ergeben sich Probleme (z.B.
viele Punk-s; des Punk-Ø vs. des Punk-s). In meinem Vortrag möchte ich mich
von dieser Beobachtung ausgehend dem empirischen und theoretischen Status des No-Blur-Principles widmen, wobei einige weiterführende Fragen aufgeworfen und diskutiert werden, die in diesem Zusammenhang relevant sind. So
stellt sich z.B. die Frage, wie der Terminus Flexionsklasse (bzw. auch der Begriff Makroklasse) definiert wird und wie eine solche Definition mit morphologischer Variation kompatibel ist. Bei den empirischen Fragen werde ich mich
auf umfangreiche Untersuchungen zur Synchronie und Diachronie der relevanten morphologischen Varianten in den Korpora DECOW2012, DeReKo und
DTA sowie auf experimentell erhobene Akzeptabilitätsurteile beziehen.
References: • Carstairs-McCarthy, A. (1994): Inflection classes, gender, and the principle of contrast.
Language 70(4), 737–788. • Enger, H.-O. (2007): The No Blur Principle meets Norwegian dialects. Studia Linguistica 61, 279–309. • Enger, H.-O. (2013): Vocabular Clarity meets Faroese noun declensions.
Folia Linguistica 47(2), 345–373. • Enger, H.-O. (2016): The No Blur Principle and Faroese conjugation.
Beiträge zur Geschichte der deutschen Sprache und Literatur (PBB) 138(1), 1–29.
288
AG 12 · Gebäude B4 1, Raum 0.05
A dynamic systems approach to morphological irregularity
Arjen Versloot
Universiteit van
Amsterdam
Elżbieta Adamczyk
Universität Wuppertal
Donnerstag
09.03.2017
09:30 – 10:00
B4 1, 0.05
Eric Hoekstra
Fryske Akademy
[email protected] [email protected]
[email protected]
The study of stability of linguistic structures is confronted with two seemingly
contradictory tendencies, namely:
• The persistence of “marked” (= low type frequency) inflectional types,
such as irregular plurals or strong and irregular verbs;
• the continuous analogical pressure of morphological patterns with high
type frequency, leading to regularization.
In the traditional interpretation of structuralism, irregularity and diachronic
change are always a disturbance of a given state, which, ideally, should be in
balance. Irregularity is considered to be a burden to the system, its emergence
counter-intuitive and its only temporary “‘stability”’ unexplained. Rather
than looking for explanations of the deviations from the stable, regular system, one should aim for a model of language that explains both regularities
and apparent “‘irregularities”’ by means of the same set of underlying mechanisms. The two mentioned contradictions become logical and can be explained
in a dynamic systems approach to language (Beckner et al. 2009). In such
systems, regularities and (temporary) stability are products of dynamic selforganisation.
In this paper we will present such a model using the data from English and
Frisian nominal morphology. The controlling factors in this model are absolute and relative frequencies of the plural form, as well as its salience. These
three factors interact with each other and sometimes show sometimes a nonlinear feedback pattern, which is a typical prerequisite for dynamic systems
behaviour.
References: • Beckner, C. et al. (2009): Language Is a Complex Adaptive System: Position Paper. Language Learning 59(2), 1–26.
289
AG12
AG 12 · Morphologische Variation – Theorie und Empirie
Donnerstag
09.03.2017
10:00 – 10:30
B4 1, 0.05
Case marking variation – an evolutionary perspective
Dankmar Enke
Ludwig-Maximilians-Universität
München
Roland Mühlenbernd
Eberhard Karls Universität Tübingen
[email protected]
[email protected]
AG12
Jäger (2007) developed a game-theoretic model that involves the formalization of case marking strategies in terms of encoding-decoding processes. The
analytic component of Jäger’s study is the detection of evolutionarily stable
case marking systems under evolutionary dynamics. His analysis shows that
exactly those case marking systems are evolutionarily stable (Maynard Smith
1982) – thus particularly resistant against forces of change –, which are predominant among the languages of the world (cf. Bickel & Nichols 2008). Their
frequent presence can therefore be explained on the basis of insights from
evolutionary game theory, such as the notions of evolutionary stability. However, the model struggles to motivate the strategies and parameters in the cognitive constraints of the individual agents, as determined by so-called “processing cues” (cf. the Competition Model by MacWhinney & Bates 1989). The
interaction of these cues varies across languages (cf. ibid.) and the languagespecific weighting (cue strength) depends on cue validity (measuring how helpful a given cue actually is in determining an interpretation). Together, the
Competition model and a game-theoretic model for communication will allow
for providing new insights into the evolutionary emergence and development
of ideal actor identification strategies, from both a neurocognitive and a communicative perspective. In our talk, we will present a game-theoretic model that
extends Jäger’s (2007) work by implementing processing cues in order to determine whether the model can detect the same case systems as evolutionarily
stable as are also detected by the original model. In total, the talk will give new
insights into ranges and boundaries of the diversity and change of case marking systems.
References: • Bates, E. & B. MacWhinney (1989): Functionalism and the competition model. In
B. MacWhinney & E. Bates (eds.), The Crosslinguistic Study of Sentence Processing. CUP, 3–73. • Bickel,
B. & J. Nichols (2008): Case marking and alignment. In A. Malchukov & A. Spencer (eds.), The Oxford
Handbook of Case. OUP, 304–321. • Jäger, G. (2007): Evolutionary game theory and typology: A case
study. Language 83(1), 74–109.
290
AG 12 · Gebäude B4 1, Raum 0.05
Content and form of Upper German case paradigms.
A formalist approach to usage-based data?
Donnerstag
09.03.2017
11:15 – 11:45
B4 1, 0.05
Sophie Ellsäßer
Albert-Ludwigs-Universität Freiburg
[email protected]
By its focus on describing divergences of content and form paradigms,
Paradigm Function Morphology (PFM) (see, e.g. Stump 2015) is an interesting
approach for modelling and describing inflectional paradigms with complex
syncretic patterns. Since case marking in German dialects represents such
complex systems, it provides a good testing ground for the theory.
Based on a corpus of spoken Upper German dialects (Ruoff 1984), I attempt
an in- depth analysis of closely related case systems in an exemplarily selected
dialect area. The texts in the corpus represent idiolectal systems of individual
speakers. All potentially case marking word forms, such as definite and indefinite articles, pronouns and adjectives, are analysed. In this way, I aim at identifying the areal variation of case marking patterns on the one hand, but also
to investigate patterns shared by all idiolects, revealing general principles of
case marking on the other hand.
When applying PFM to the data gathered in the study, special attention has
to be given to the properties of dialectal systems such as the geographical aspect of variation, which is but rarely implemented in the theory yet. An issue
to cope with in particular is the decision to assume either one global or several different content paradigms (that is, the set of morphosyntactic properties providing canonical case forms). Thus, apart from outlining the design of
my project and presenting first results, I will especially address problematic
issues arising from modelling usage-based data of spoken language in its geographical variation from a general perspective beyond PFM.
References: • Ruoff, A. (1984): Alltagstexte I. Transkriptionen von Tonbandaufnahmen aus Baden-Württemberg und Bayrisch-Schwaben. Niemeyer. • Stump, G. (2015): Inflectional Paradigms. Content and
Form at the Syntax-Morphology Interface. CUP.
291
AG12
AG 12 · Morphologische Variation – Theorie und Empirie
Das possessive -s im Deutschen:
Entwicklung, Variation und theoretischer Status
Donnerstag
09.03.2017
11:45 – 12:15
B4 1, 0.05
Tanja Ackermann
Freie Universität Berlin
[email protected]
AG12
Hinsichtlich der Kasusmarkierung haben im onymischen Bereich in den letzten Jahrhunderten sowohl Allomorphieabbau als auch Markerschwund in
starkem Maße stattgefunden (vgl. z.B. Nübling 2012). Im Genitiv hat sich zum
Neuhochdeutschen hin mit dem invarianten -s ein überstabiler Marker entwickelt, der seit dem späten 18. Jh. allerdings dort schwindet, wo der Genitiv an einem anderen Element eindeutig markiert ist (z.B.: der Abschied de-s
Lukas Podolski-Ø). In adnominalen Konstruktionen, in denen der Name artikellos auftritt (z.B.: Lukas Podolskis Abschied), ist das -s hingegen stabil und erfährt aktuell sogar eine Ausbreitung. Solche pränominalen Possessivphrasen
sind bislang vor allem aus theoretisch-syntaktischer Perspektive beschrieben
worden (vgl. z.B. Demske 2001, Fuß 2011); eine theoretisch-morphologische
Beschreibung des s-Markers steht hingegen noch aus. Anhand von diachronen
und synchronen Korpusdaten (DTA und DECOW2012) soll im Vortrag gezeigt
werden, wie sich das -s von einem Genitivflexiv mit Allomorphie zu einem
invarianten Marker entwickelt hat, der gegenwärtig zwischen einem noch
auf Wortebene operierenden Flexiv und einem auf Phrasenebene operierenden Marker schwankt. Auf Basis der empirischen Befunde soll diskutiert werden, welche Konsequenzen sich für verschiedene theoretische (synchrone
und diachrone) Modellierungen der pränominalen Possessivphrasen ergeben:
Haben wir es mit Degrammatikalisierung oder Exaptation zu tun? Was bedeutet eine Analyse als Klitikon vs. Flexiv für Modellierungen, die das -s als
Kopf analysieren? Hilft uns die Annahme einer s-construction?
References: • Demske, U. (2001): Merkmale und Relationen. De Gruyter. • Fuß, E. (2011): Eigennamen
und adnominaler Genitiv im Deutschen. Linguistische Berichte 225, 19–42. • Nübling, D. (2012): Auf
dem Weg zu Nicht-Flektierbaren: Die Deflexion der deutschen Eigennamen diachron und synchron.
In: Nicht-flektierte und nicht-flektierbare Wortarten. De Gruyter, 224–246.
292
AG 12 · Gebäude B4 1, Raum 0.05
Mit einem stetigem Anstieg – Variation in der Adjektivflexion
Donnerstag
09.03.2017
12:15 – 12:45
B4 1, 0.05
Astrid Niebuhr
Philipps-Universität Marburg
[email protected]
Thema des Vortrags ist ein an der Schnittstelle von Morphologie und Syntax angesiedeltes Variationsphänomen in der Adjektivflexion des Deutschen.
Die Norm ist diesbezüglich eindeutig: „Wenn dem Adjektiv ein Artikelwort
mit Flexionsendung vorangeht, wird das Adjektiv schwach flektiert, sonst
stark“ (Duden -Grammatik 2009: 363). Dennoch treten vermehrt Dativkonstruktionen wie vor einem kleinem Publikum und sogar solche mit dem bestimmten Artikel auf, obwohl z.B. Sahel (2011) für die Entwicklung ähnlicher
Variationsphänomene in der NP klare Tendenzen zur Monoflexion diagnostiziert. Das Auftreten starker Formen nach dem bestimmten Artikel ist zudem
insofern erstaunlich, als diese Kombination – von der Umbruchsituation im
Frühneuhochdeutschen abgesehen (vgl. Demske 2000: 82–84ff.) – keinerlei
historisches Vorbild hat. Und während z.B. die Parallel- und Wechselflexion
bei zwei koordinierten attributiven Adjektiven ohne Artikel verschiedentlich
untersucht wird (vgl. z.B. Nübling 2011), fehlt eine Untersuchung der Konstruktion mit Artikel bislang.
Im Vortrag wird dem Tatbestand auf Basis eines aktuellen Zeitungskorpus nachgegangen, denn es gilt herauszufinden, welche (z.B. phonologischen,
morphologischen oder pragmatischen) Faktoren das Auftreten der starken
Flexionsendung nach Artikel begünstigen. Außerdem soll geprüft werden,
inwiefern theoretische Zugänge, z.B. die Konstruktionsgrammatik, zur Erklärung des Phänomens beitragen können.
References: • Demske, U. (2000): Merkmale und Relationen. De Gruyter. • Duden-Grammatik (2009).
Dudenverlag. • Nübling, D. (2011): Unter großem persönlichem oder persönlichen Einsatz? Der
sprachliche Zweifelsfall adjektivischer Parallel- vs. Wechselflexion als Beispiel für aktuellen grammatischen Wandel. In: Grammatik – Lehren, Lernen, Verstehen. De Gruyter. • Sahel, S. (2011):
Monoflexion als Erklärung für Variation in der Nominalphrasenflexion des Deutschen. In: Grammatik und Korpora 2009. Narr.
293
AG12
AG 12 · Morphologische Variation – Theorie und Empirie
Stark oder schwach? – Oder: wie die Informationsstruktur
morpho-syntaktische Variationen steuert
Donnerstag
09.03.2017
13:45 – 14:15
B4 1, 0.05
AG12
Helmut Weiß
Goethe-Universität Frankfurt
Seyna Carlucci-Dirani
Goethe-Universität Frankfurt
[email protected]
[email protected]
Wortarten wie Personalpronomen und Definitartikel weisen in deutschen
Dialekten systematisch mindestens zwei verschiedene morphologische Reihen auf (volle und reduzierte Formen), deren syntaktische Distribution von
pragmatischen Faktoren abhängt. Wir möchten zeigen, dass die Distribution
der verschiedenen Formen bei beiden den gleichen Restriktionen unterliegt.
Dafür konzentrieren wir uns insbesondere auf Phänomene wie Deixis, Kontrastfokus, Topik-Shift und Relativsätze, in denen ausschließlich die jeweiligen vollen Formen auftreten können.
Der Grund für die Distribution starker/voller und schwacher/reduzierter
Formen liegt in deren Struktur. Ausschließlich erstere werden in D0 basisgeneriert, während letztere eine Kopfposition unterhalb davon ausbuchstabieren (je nach Ansatz n0 /AgrD0 /φ0 , cf. Wiltschko 1998, Roberts 2010, Weiß
2015, Trutkowski & Weiß 2016). Syntaktisch präsent sind zwar immer volle
DPs (kontra Cardinaletti & Starke 1999), die Artikel und Pronomen buchstabieren aber unterschiedliche Teile davon aus – was letztlich deren informationsstrukturelle Unterschiede erklärt.
References: • Cardinalleti, A. & M. /Starke , M. (1999): The typology of structural deficiency:
On the three grammatical classes. In: Clitics in the languages of Europe. De Gruyter, 145–-233.
• Roberts, I. (2010): Agreement and head movement: clitics, incorporation, and defective goals. MIT Press.
• Trutkowski, E. & H. Weiß (2016): When Personal Pronouns Compete with Relative Pronouns. In:
The Impact of Pronominal Form on Interpretation. De Gruyter, 135–166. • Weiß, H. (2015): When the
subject follows the object. On a curiosity in the syntax of personal pronouns in some German dialects. Journal of Comparative Germanic Linguistics 18, 65–92. • Wiltschko, M. (1998): On the Syntax
and Semantics of (Relative) Pronouns and Determiners. Journal of Comparative Germanic Linguistics 2,
143–181.
294
AG 12 · Gebäude B4 1, Raum 0.05
Variation of agreement forms:
Investigations on lexical hybrids in German dialects
Donnerstag
09.03.2017
14:15 – 14:45
B4 1, 0.05
Stephanie Leser-Cronau
Philipps-Universität Marburg
[email protected]
The term lexical hybrid (Corbett 2006: 213–215ff.) designates nouns with inconsistent agreement patterns concerning number or gender. The current presentation focuses on lexical hybrids with an internal conflict between grammatical gender and biological sex. One famous example is the German lexeme
Mädchen (‘girl’), which is biologically female but grammatically neuter. Agreement targets of Mädchen may either show agreement with the biological sex
(semantic agreement) or the grammatical gender (formal agreement). Corbett
(1979) established a hierarchy based on agreement patterns from multiple languages allowing a prediction about the possibility of semantic agreement. This
agreement hierarchy is shown below:
(1)
attributive > predicate > relative pronoun > personal pronoun
The possibility for semantic agreement increases consistently from the left to
the right. If semantic agreement is possible in a given position, it should also be
possible for all elements to the right of that position. There are also a couple of
other factors influencing the choice of formal or semantic agreement, like e.g.
the linear distance between controller and target. All previous studies have in
common that they analyze written standard language. In contrast, the current
presentation follows a novel approach by using oral data of dialect speakers.
The project Syntax of Hessian Dialects (SyHD) collected written and oral data
from so called NORMs and NORFs (non-mobile , older rural males/females).
In addition, a corpus investigation is used to test as to whether German dialects behave differently from the standard language concerning agreement
patterns.
References: • Corbett, G. G. (1979): The Agreement Hierarchy. Journal of Linguistics 2(15), 203–224.
• Corbett, G. G. (2006): Agreement. CUP.
295
AG12
AG 12 · Morphologische Variation – Theorie und Empirie
Complex heads in Pomeranian: Some morphosysntactic
considerations
Freitag
10.03.2017
11:30 – 12:00
B4 1, 0.05
Göz Kaufmann
Universität Freiburg
[email protected]
Translating the sentence yesterday I could have sold the ring, several Brazilian
speakers of Pomeranian produced the following two variants:
(1)
jestern hätt ik kütt de Fingerring verköft hewe
yesterday had I could the ring
sold
have
(2)
jestern hätt kütt ik der Fingerring verkauft hat hewe
yesterday had could I the ring
sold
had have
AG12
While the verbal sequence in (1) is common to West Germanic varieties, the
translation in (2), which features two verbal elements in C0 , is a rarum for
which our data suggest a theoretically intriguing explanation that is based on
the informants’ behavior with regard to two other morphosyntactic phenomena. The first one is the fact that the informants responsible for (2) – but not the
ones responsible for (1) – produce the cluster V3-V1-V2 in dependent clauses,
for example in the translation of the conditional clause if he really had wanted
to write this letter […]:
(3)
wenn her auf ehrlich de Karte schriewe hätt wutt
if
he in earnest the letter write
had wanted
If we derive the cluster schriewe hätt wutt by head movement and right adjunction of V2 wutt to V1 hätt in I0 , the fact that informants who produce (3) also
produce (2) can be explained by assuming that complex heads such as hätt wutt
in (3) or hätt kütt in (2) form morphologically bound units and are moved as
a whole from I0 to C0 in root contexts. The other phenomenon in (2), entirely
absent from the data of the informants responsible for (1), is the doubling in
had hewe (‘had have’). Our explanation in this case is that the non-finite features of the participle kütt in (2) clash with the finiteness requirements of
C0 . Therefore, these morphological features may have to be produced clausefinally with a lexically light element such as hewe (‘have’).
296
AG 12 · Gebäude B4 1, Raum 0.05
Dat hätt kéinten anescht gemaach ginn – IPP und
supinale Formen bei luxemburgischen Modalverben
Freitag
10.03.2017
12:00 – 12:30
B4 1, 0.05
Caroline Döhmer
Universität Luxemburg
[email protected]
In diesem Beitrag geht es um den IPP-Effekt (Ersatzinfinitiv) von Modalverben
im Luxemburgischen, denn hier stehen nicht nur „echte“ Infinitive zur Verfügung, sondern auch modifizierte Formen, so genannte Supina. Der Beispielsatz aus dem Titel soll dieses Phänomen im Luxemburgischen (LUX) sowie im
Standarddeutschen (ST-DT) verdeutlichen.
(1)
LUX: Dat hätt kéinten anescht gemaach ginn.
(Inf.: kënnen, Supinum: ké inten)
(2)
ST-DT: Das hätte anders gemacht werden können.
(Inf.: können)
AG12
Für die vorliegende Untersuchung wurden über 8000 drei- und viergliedrige
Verbalkomplexe des Typs AUX + MOD + V bzw. AUX + MOD + AUX + V in
Haupt- und Nebensätzen aus einem umfangreichen Korpus (ca. 80 Mio. Wortformen) ausgewertet. Dabei lassen sich zwei zentrale Beobachtungen machen:
(A) Aus quantitativer Sicht zeigt sich, dass supinale Formen (Typ ké inten) im
Schnitt etwa ein Viertel der IPP-Konstruktionen ausmachen. (B) Aus qualitativer Sicht lässt sich eine gewisse Modussensitivität der Supina erkennen, d.h.,
dass sich deren morphologische Form offenkundig nach dem Modus des Perfektauxiliars richtet. Es wird zu überprüfen sein, wie diese Supina paradigmatisch festzuhalten sind (Stump 2015) und welche morphologische Basis
für die unterschiedlichen Verbformen ausschlaggebend ist. Darüber hinaus
bleibt auch zu klären, unter welchen (weiteren) syntaktischen Bedingungen
die supinalen Formen auftreten können bzw. blockiert sind.
References: • Schallert, O. (2014a): Zur Syntax der Ersatzinfinitivkonstruktion: Typologie und Variation.
Stauffenburg. • Stump, G. (2015): Inflectional Paradigms. Content and Form at the Syntax-Morphology
Interface. CUP.
297
AG 12 · Morphologische Variation – Theorie und Empirie
Freitag
10.03.2017
12:30 – 13:00
B4 1, 0.05
Negative concord im Alemannischen: Eine morphologische Erklärung
Ann-Marie Moser
Ludwig-Maximilians-Universität München
[email protected]
Negative concord tritt in den deutschen Dialekten sowohl als negative doubling
als auch negative spread auf. Folgendes Beispiel zeigt negative spread im Alemannischen:
(1)
AG12
Miir säit me jo nie nüt
Mir sagt man ja nie nichts
‘Mir sagt man ja nie etwas’
(Muster/Bürkli 2001:206)
Während diese Form im gesamten alemannischen Sprachgebiet belegt ist,
wird dagegen negative doubling, also negative concord in Form von Satznegation
und negativem Indefinitum, kaum bis gar nicht akzeptiert. Ergebnisse aus aktuellen Syntaxprojekten, Dialektgrammatiken oder auch Kommentare in der
Forschungsliteratur bezeugen diese Diskrepanz. Der Kontrast ist besonders
auffällig, wenn man diese Auftretensfrequenz mit jener anderer deutscher Dialekte (Bairisch, Ost- und Westmitteldeutsch, Niederdeutsch) vergleicht (vgl.
Moser 2015). Bisherige Erklärungen zu negative concord (z.B. Jäger 2008, Weiß
1998, Zeijlstra 2004) nehmen auf syntaktischer Ebene eine (c)overte Satznegation an. Ich möchte hingegen einen Ansatz vorstellen und auf das Alemannische anwenden, den Tubau (2016) für die britischen Varietäten entwickelt hat: Negative spread wird hier nicht mithilfe der Abhängigkeit von einer
coverten Satznegation, sondern mit der Existenz von zwei verschiedenen morphologischen bzw. /lexikalischen Varianten erklärt, die sich in ihrer (Nicht–
)Interpretierbarkeit des [Neg]-Merkmals unterscheiden.
References: • Jäger, A. (2008): History of German negation. John Benjamins. • Moser, A. (2015): Doppelte Negation in den deutschen Dialekten. [unpublizierte Abschlussarbeit] • Muster, H. P. & B. Flaig
Bürkli (2001): Baselbieter Wörterbuch. Christoph-Merian-Verl. • Tubau, S. (2016): Lexical variation
and Negative Concord in Traditional Dialects of British English. Journal of Comparative Germanic Linguistics 19, 143–177. • Weiß, H. (1998): Syntax des Bairischen. Niemeyer. • Zeijlstra, H. (2004): Sentential Negation and Negative Concord. LOT Publications.
298
AG 12 · Gebäude B4 1, Raum 0.05
Concurrent gender systems in Europe:
a new look at the Asturian neuter
Freitag
10.03.2017
13:00 – 14:00
B4 1, 0.05
Michele Loporcaro
Universität Zurich
[email protected]
Recent work in Canonical Typology (Fedden & Corbett 2016) pursues a quest,
started in Corbett (1991: 184–188), for languages with more than one gender
system, with examples mostly found in languages that have been maintained
to possess gender alongside classifiers, since “the traditional division between
gender and classifiers as fulfilling similar functions in languages of different
types is ever harder to maintain” (p. 1). Languages with concurrent systems occur in Australia, Papua New Guinea and the Americas. The closest one gets to
Romance so far is Michif, a French/Cree mixed language of Ontario (cf. Bakker
1997: 109 for the data and Corbett 2012: 176 for discussion), while no Romance
language has been analyzed as featuring two gender systems: thus, implicitly,
the room for variation in time and space in Romance has been held to be constrained by the (unitary) three-valued gender system inherited from Latin. In
this paper, I argue that Central Asturian has changed in a more radical way,
becoming a language with two concurrent gender systems. This language famously has a kind of neuter agreement (marked with an o-ending on nonprenominal adjectives) that stirred an unrelenting debate, with traditional
analyses – which called it a (neuter) gender value –rightly criticized, by proponents of alternative analyses because of the mismatch with the common
Romance binary gender agreement seen on determiners: e.g. el café frio vs. la
tsiche frio ‘cold coffee/milk’. Linking theory (i.e., recent research in Canonical
Typology) with empirical evidence, here, automatically yields the best analysis, leading to the conclusion that a) the Asturian neuter is a gender value,
but b) within a second gender system, concurrent with the binary common
Romance one.
References: • Bakker, P. (1997): A language of our own. The genesis of Michif. OUP. • Corbett, G.G. (1991):
Gender. CUP. • Fedden, S. & G.G. Corbett (2016): Gender and classifiers as concurrent systems: a first typology. Ms. University of Sidney and /Surrey Morphology Group.
299
AG13
Register in linguistic theory: Modeling functional variation
Aria Adli & Anke Lüdeling
Universität zu Köln, Humboldt-Universität zu Berlin
[email protected], [email protected]
Raum: B4 1, 0.04
Short description
This workshop deals with the variationist modelling of register variation. The
term register is used here to describe a variety of a language that is associated with particular functional or situational features, thus describing intraspeaker variation. Variation exists on each linguistic level (phonology, morphology, syntax, lexicon, etc.). We welcome variationist research on qualitative and quantitative aspects of synchronic and diachronic register variation.
301
AG 13 · Register in linguistic theory: Modeling functional variation
Recursive embedding and register variation
Mittwoch
08.03.2017
14:15 – 15:15
B4 1, 0.04
AG13
Elisabeth Verhoeven
Humboldt-Universität zu Berlin
Nico Lehmann
Humboldt-Universität zu Berlin
[email protected]
[email protected]
Recursive embedding is a core property of the language faculty. We know a
lot about the cognitive limitations in parsing recursive structures (e.g. Roeper
& Speas, eds. 2015), but less so about the determinants of their occurrence
in naturalistic communication. In this paper we explore the use and limits
of recursive embedding in spoken discourse in public vs. non-public speech
situations in German. We report results of an ongoing study on the depth of
(self-)embedding of C, V, and N projections, i.e. we measured instances of [C [C
…]], [[…V] V], and [N [N …]]. The analyzed data stem from the Datenbank für
gesprochenes Deutsch (DGD) Grundstrukturen: Freiburger Korpus and FOLK
(both available at http://agd.ids-mannheim.de, IDS Mannheim) from which
22 conversations were selected and classified according to contextual factors
forming a public and a non-public subcorpus of 11 conversations of about 1000
tokens each. The public subcorpus features argumentative public conversation with four to nine unacquainted speakers whereas the non-public subcorpus features argumentative talk from social settings with two to three fairly familiar speakers per conversation. The investigated texts contain further properties that cannot be kept invariant in a study of natural data but which may influence the choice of a particular structure (Biber & Conrad 2009). Hence, a set
of 10 features relating to the communicative setting (speaker symmetry, communicative role of participant, discourse topic, etc.) have been annotated and
integrated in the model as random factors. The findings confirm the hypothesis that speakers use significantly more (self-)embeddings in all three studied
projections in public registers compared to non-public registers. In addition,
we found evidence that the depth of recursion between different projections
is correlated across speakers, which suggests that [α [α …]] structures are an
entity of grammar that is relevant for speaker’s reality, i.e., may be selected for
particular purposes.
References: • Biber, D. & Conrad, S. (2009): Register, genre, and style. Cambridge: CUP. • Roeper, T. &
Speas, M. eds. (2015): Recursion: Complexity and cognition. Springer.
302
AG 13 · Gebäude B4 1, Raum 0.04
Register specificity in the English genitive alternation.
Do variable cues reflect different grammars?
Mittwoch
08.03.2017
15:15 – 15:45
B4 1, 0.04
Jason Grafmiller
University of Leuven
[email protected]
We examine variability in the genitive alternation (the president’s decision vs.
the decision of the president) across five written registers of 20th century American English, focusing on quantitative differences in the constraints on writers’ deployment of this variable in different stylistic contexts. We argue that
variation in written registers of this kind is a clear example of complex ‘style
switching’ (Rickford, 2014), and discuss how our findings speak to the relationship between grammatical representation and quantitative variability in
constraints across styles.
We draw on data from the Brown and Frown corpora, sampling a subset
of registers in the early 1960s (Brown) and 1990s (Frown): Press reportage,
Non-fiction (memoirs), Learned, General fiction, and Adventure fiction. We
extracted 5098 genitive tokens from the two corpora (Brown N = 2497; Frown
N = 2601), annotating for factors known to condition the choice of genitive
variant (see e.g. Grafmiller, 2014). Such factors include: the length (in words)
of both constituents; the semantic relation between possessor and possessum;
the presence of a final sibilant on the possessor; and the animacy, frequency,
givenness, and NP type (common vs. proper) of the possessor. Data were analyzed using mixed-effects logistic regression.
We find sizable differences among registers in the influence of possessor animacy, which has a significantly weaker effect in Press than in other registers.
Additionally, the relative ranking of this constraint is lower in Press than in
other registers, and in general, the relative constraint rankings vary noticeably across registers. Within registers, there is little intra-author variability
in the constraint effects, however, inter-author rates of genitive use vary considerably across the registers. Fiction writers vary the most, while journalists
vary the least.
We interpret these patterns in Press writing as reflections of journalists’
semi-conscious move toward more economical and colloquial modes of expression (Biber, 2003). But does the variability we observe imply that we are dealing with distinct, register-specific grammars, a la Guy (2015)? We believe that
303
AG13
AG 13 · Register in linguistic theory: Modeling functional variation
within an experience-based, probabilistic approach to grammar, the multiplegrammar model is ill-defined, and argue instead for an alternative model in
which situational/stylistic cues can directly shape the influence of internal
constraints within a single ‘grammar’.
References: • Biber, D. (2003): Compressed noun-phrase structures in newspaper discourse. In: New
Media Language, 169–181. • Grafmiller, J. (2014): Variation in English genitives across modality and
genres. English Language and Linguistics 18(3), 471–496. • Guy, G.R. (2015): Coherence, constraints and
quantities. NWAV 44. • Rickford, J. (2014): Situation: Stylistic variation in sociolinguistic corpora and
theory. Language and Linguistics Compass 8(11), 590–603.
Mittwoch
08.03.2017
16:30 – 17:00
B4 1, 0.04
Register-specific interference in translation
Stella Neumann
Stefan Evert
RWTH Aachen University FAU Erlangen- Nürnberg
[email protected]
[email protected]
Gert De Sutter
Universiteit Gent
[email protected]
aachen.de
AG13
This paper discusses a quantitative corpus analysis of register-specific translator behaviour, in terms of the extent to which interference of the source
language is admitted into the translation. Evert & Neumann (in press) provide corpus evidence for source language interference and differences depending on translation direction, arguably due to the prestige of the languages involved. However, they do not further differentiate the findings according to
register, although there is growing evidence that translators are susceptible
to register conventions (e.g. Delaere 2015). This is in line with assumptions
about language being essentially organised into registers as subsystems that
capture the probabilistically distributed features of a language (Halliday 1991).
Accordingly, a feature may be highly likely to occur in one register and virtually blocked in another. These assumptions would predict that translators
adapt to the specific situational context of both the source and target register
and that they modulate the extent to which they allow the source language
to interfere or adapt to register conventions in the target language. We will
use the multivariate methodology developed by Diwersy et al. (2014) to examine the influence of register on interference. We use lexico-grammatical features capturing dimensions of register extracted semi-automatically from the
English-German CroCo Corpus and from the Dutch Parallel Corpus so as to test
whether the findings reflect more general patterns across language pairs. The
304
AG 13 · Gebäude B4 1, Raum 0.04
findings will also shed light on more general claims about the role of register
in language.
References: • Delaere, I. (2015): Do translators walk the line? University of Ghent. • Diwersy, S., Evert,
S., & Neumann, S. (2014): A weakly supervised multivariate approach to the study of language variation. In: Aggregating Dialectology, Typology, and Register Analysis. de Gruyter, 174–204. • Evert, S., &
Neumann, S. (in press): The impact of translation direction on characteristics of translated texts. In:
Empirical Translation Studies. New theoretical and methodological traditions. de Gruyter. • Halliday, M.
A. K. (1991): Towards probabilistic interpretations. In: Functional and Systemic Linguistics. Approaches
and Uses. Mouton de Gruyter, 39–61.
Patterns of cohesion as dependent variables in a contrastive study of
registers in English and German
Mittwoch
08.03.2017
17:00 – 17:30
B4 1, 0.04
Kerstin Kunz, Erich Steiner, Ekaterina Lapshinova-Koltunski,
José Martinez, Katrin Menzel
[email protected]
In studies of register so far, the focus has usually been on variation in lexicogrammar, and within one language (exceptions include Biber 1995). We
argue in favour of adding cohesion as a linguistic level to the modelling of
register, and we argue that lexicogrammatical properties can to some extent
be understood as interacting with principles of cohesion in different texttypes. Our talk will report on empirical testing of our hypotheses, relying on
a linguistically annotated corpus of English and German texts in a number of
parallel registers. The talk is based on the ongoing GECCo project funded by
the Deutsche Forschungsgemeinschaft (http://www.gecco.uni-saarland.
de/GECCo/index.html).
The features we investigate are derived from five main types of cohesion
(Halliday & Hasan 1976) as linguistic variables and their concrete instantiated
variants. These variants include cohesive devices (e.g. proforms, ellipses, conjunctions and lexical items), and properties of co-reference chains and lexical
chains (e.g. number of and distance between elements in chains, types of semantic relations (cf. Kunz et al in press; Kunz et al 2016)). Our findings demonstrate that cohesion is an important linguistic dimension for modeling functional variation. They provide important information on how registers vary in
the development of topics in texts, the types of semantic relations established,
the frequency and strength of cohesive relations, and the degree of variation
in cohesion.
305
AG13
AG 13 · Register in linguistic theory: Modeling functional variation
References: • Biber, D. (1995): Dimensions of register variation. A cross-linguistic comparison. Cambridge University Press • Halliday, M. A. K. and Hasan, R. (1976): Cohesion in English. Longman.
• Kunz, K., Degaetano-Ortlieb, S., Lapshinova-Koltunski, E., Menzel, K. and Steiner, E. (in press):
English-German contrasts in cohesion and implications for translation.Empirical Translation Studies.
New Theoretical and Methodological Traditions. Mouton de Gruyter. • Kunz, K., Lapshinova-Koltunski,E., Martinez-Martinez, J.M. (2016): Beyond Identity Coreference: Contrasting Indicators of Textual Coherence in English and German. Proceedings of CORBON at NAACL-HLT2016, San Diego, 16 June.
Automatic register annotation for linguistic research?
Donnerstag
09.03.2017
09:00 – 10:00
B4 1, 0.04
AG13
Felix Bildhauer
IDS Mannheim
Roland Schäfer
Freie Universität Berlin
[email protected]
[email protected]
Corpus linguists often require document-level meta data such as document
registers. Since manual annotation of registers is infeasible for very large
corpora (such as crawled web corpora), the only viable alternative is automatic classification. Most approaches to automatic register annotation rely
on a (usually high) number of linguistic features extracted from the documents. In our talk, we explore the usefulness of automatically annotated register categories for models of alternation phenomena in morpho-syntax. An obvious conceptual problem of such models is circularity if the registers are operationalized in terms of document-internal morpho-syntactic features. However, we discuss a technical aspect, namely that models can be of much higher
quality if the raw features are used instead of aggregated register categories.
To this end, we conducted two case studies on German: a) case variation after prepositions and b) inflection of adjectives after pronominal adjectives. We
created an ad-hoc corpus of approx. 0.6 m documents/ 0.4 bn tokens sampled
from the DECOW16 web corpus (Schäfer and Bildhauer, 2012) and the DeReKo
corpus (Kupietz et al., 2010) and extracted a large number of features at the
document level. First, we used these as predictors in a generalized linear model
(GLM) modeling the two alternation phenomena. Second, we used the features
to induce document categories in the corpus, then fitting a number of alternative models with these categories as predictors. Additionally, we used Monte
Carlo simulations to demonstrate how aggregation of features into register categories can systematically affect the quality of GLMs. We compare the quality of the different models thus obtained and discuss the implications of our
findings for using automatically annotated high-level categories in research
306
AG 13 · Gebäude B4 1, Raum 0.04
on grammatical and morphological alternation phenomena.
References: • Kupietz, M., Belica, C., Keibel, H. and Witt, A. (2010). The German Reference Corpus
DeReKo: A Primordial Sample for Linguistic Research. In N. Calzolari, et al. (eds.), Proceedings of
LREC’10, pages 1848–1854, Valletta, Malta: ELRA. • Schäfer, R. and Bildhauer, F. (2012). Building large
corpora from the web using a new efficient tool chain. In N. Calzolari, et al. (eds.), Proceedings of
LREC’12, pages 486–493, Istanbul: ELRA.
A functional stylistics for register and genre
Thomas Haider
Universität Heidelberg
Alexis Palmer
University of North Texas
[email protected]
[email protected]
Donnerstag
09.03.2017
10:00 – 10:30
B4 1, 0.04
Work on register for German is rather scarce, compared to English. We narrow this gap by (i) developing a theoretically grounded comparative typology
for genre and register analysis, (ii) compiling a corpus of German register and
genre out of DeReKo (Kupietz et al., 2010), (iii) developing high accuracy machine learning algorithms for supervised text genre classification, (iv) providing an extensive description of prototypical register classes through the agglomeration of style feature loadings. We thus contribute both to linguistic
theory building and to research on genre adaptation of NLP methods within
the LiMo project.
Register can be understood as the function between situational context and
linguistic analysis (Biber & Conrad, 2009) or as functional stylistics i.e. the distribution of style features in communicative areas (Jakobson, 1960). In designing
the corpus typology (taxonomy) we make use of prototype theory (Rosch, 1973)
that affords the embedding of a basic-level category (communicative purpose)
within a super-level (social text use) and a sub-genre (topic modelling). We
model the prototypical text classes using dimensions such as document size, a
‘natural’ distribution of topics, or the dominant register of a genre.
We then develop a machine learning system for register and genre classification. We focus on a high-quality feature extraction pipeline which implements style features through numerous morphosyntactic annotations, lexicons of linguistic nature (i.a. verb classes), psycholinguistic word norms (i.a.
concreteness), and topic models. We identify the best models for supervised
text genre classification, achieving results comparable to the state-of-the-art
for English. Finally, we show that our work is also applicable to English. Here,
307
AG13
AG 13 · Register in linguistic theory: Modeling functional variation
we are able to identify pervasive functional dimensions using PCA (to wit, interpersonal and narrative dimensions in literary genre).
References: • Biber, D. and Conrad, S. (2009). Register, genre, and style. Cambridge University Press.
• Jakobson, R. (1960). Closing statement: Linguistics and poetics. Style in language, 350:377. • Kupietz,
M., Belica, C., Keibel, H., and Witt, A. (2010). The german reference corpus DeReKo: A primordial
sample for linguistic research. In LREC. • Rosch, E. H. (1973). Natural categories. Cognitive psychology, 4(3):328–350.
Prosodic aspects of style and register of
live sports commentaries in radio and television
Donnerstag
09.03.2017
11:15 – 11:45
B4 1, 0.04
AG13
Jürgen Trouvain
Universität des Saarlandes
Friederike Kern
Universität Bielefeld
[email protected]
[email protected]
In phonetic research, there is raising acknowledgment with regard to the diversity of authentic speech that goes beyond highly controlled read speech.
Different communicative situations require an adaptation of phonetic and
linguistic resources that leads to various forms of non-scripted speech. Live
sports commentaries provide good examples for studying such adaptions empirically. For instance, horse race commentaries are characterized by a dramatic rise of pitch and perceived increasing tempo as well as extreme syntactic
restrictions.
Studies have shown that live football commentaries feature narrative, predramatic and dramatic phases that can be distinguished by different speech
styles with typical phonetic and syntactic characteristics. E.g., the climax of
the dramatic phase usually leads to an affective goal comment often produced
as roaring with extremely high pitch. Dramatic phases that usually occur before goal scenes are characterized by a shortening and simplification of syntactic phrases co-occuring with a predominant use of pre-fabricated lexical
units. In addition, dramatic phrases are marked prosodically by raising pitch
and tempo (articulation rate and pausing). This prosodic behaviour serves
as a communicative tool to signal increasing affect. It overrides the linguistic prosodic structure: we can observe sharp pitch rising at places where we
would expect falling pitch contours (at the end of declaratives) or no marked
pitch movement (at non-accented syllables).
In our current research, we investigate the phonetic variation found in dra-
308
AG 13 · Gebäude B4 1, Raum 0.04
matic phases of live football commentaries of the same game (Germany v. England at world cup 2010) across languages (German vs. English), media (radio
vs. TV) and media culture (public vs. private stations). The results of acoustic and perceptual analysis reveal the general tendency that sets of prosodic
features, with pitch as the most important, are typical for dramatic speech.
Prosodic variation in French:
self-repairs in conceptual distance and proximity
Donnerstag
09.03.2017
11:45 – 12:15
B4 1, 0.04
Johanna Stahnke
Bergische Universität Wuppertal
[email protected]
This contribution investigates the conversational functions and prosodic
forms of French self-repairs in conceptual distance and proximity (Koch &
Oesterreicher 1985), relating to the questions of how register variation can
be modeled and how it may lead to language change. While in paraphrases
the relation between repairable and repair is semantic equivalence, semantic difference is established in corrections (Gülich & Kotschi 1996). The intonational structure of repairs is deaccented for paraphrases and overaccented
for corrections (Morel & Danon-Boileau 1998). A total of 379 repairs from
two sample corpora was coded for function (paraphrase, correction), register (distance, proximity) and prosodic structure (standard, de-, overaccentuation). 35% of proximity corrections are deaccented; only 23% are overaccented.
Deaccentuation is significantly influenced by conceptual variation (p<.001),
with proximity as the favoring factor. By contrast, functional variation does
not have a significant effect (p=.229). To account for these results, a model
of speaker-strategic routinization is proposed (Detges & Waltereit, to appear).
Corrections are more disruptive than paraphrases with regard to continuous
discourse flow and turn-maintaining, and these problems are specifically related to proximity. Speakers in proximity possibly encode corrections as paraphrases by violating conversational maxims. This type of variation may eventually lead to linguistic change, as has been shown for lexical elements which
develop into discourse markers.
References: • Detges, U. & Waltereit, R. (to appear): Grammaticalization and pragmaticalization. In:
Fischer, S. & Gabriel, C. (eds.): Manual of Grammatical Interfaces in Romance. De Gruyter. • Gülich, E. &
Kotschi, T. (1996): Textherstellungsverfahren in mündlicher Kommunikation. In: Motsch, W. (ed.):
Ebenen der Textstruktur. Niemeyer, 37-80. • Koch, P. & Oesterreicher, W. (1985): Sprache der Nähe
– Sprache der Distanz. Mündlichkeit und Schriftlichkeit im Spannungsfeld von Sprachtheorie und
Sprachgeschichte. In: Romanistisches Jahrbuch 36, 15-43. • Morel, M.-A. & Danon-Boileau, L. (1998):
309
Grammaire de l’intonation. Ophrys.
AG13
AG 13 · Register in linguistic theory: Modeling functional variation
Register-dependency of deliberate metaphor
Donnerstag
09.03.2017
12:15 – 12:45
B4 1, 0.04
Markus Egg
Humboldt-Universität zu Berlin
[email protected]
AG13
Topic of my talk is the dependency of metaphor on register (Deignan et al.
2013): Register is functionally determined linguistic variety; and the function
and motivation of metaphor (as a non-literal mode of communication) has
been the focus of an intensive debate. I will investigate this dependency by
comparing a specific kind of metaphor, viz., deliberate metaphor (DM), in different registers.
Following Lakoff and Johnson (1981), metaphor introduces target domains
(TDs) by mapping the structure of another, the source domain (SD) onto the
respective TD, e.g., in the well-known metaphor of life as a journey. This process often goes unnoticed both in the production and processing of metaphor,
but not in DMs, which are recognised by their hearers/readers (Steen 2009).
DM intends to change the addressee’s perspective on a topic (Steen 2008),
and therefore is one way of alienation (Schklowski 1971). Alienation through
metaphor brings together quite dissimilar domains, thus the mapping of the
SD structure onto the TD introduces an unfamiliar perspective on the TD.
DM dependency on register will be investigated in a corpus-based approach,
investigating texts of two domains, viz., sermons, and academic lectures. The
texts are taken from the genre of didactic discourse, because DM – though otherwise quite rare – appears in didactic discourse with a higher frequency than
usual (Beger 2011; Beger 2015).
The variable is the use of metaphor as defined above for the reference to
key concepts in the texts. The variants are different instantiations of metaphor
for these, in particular, either as deliberate or as non-deliberate metaphor, or
in terms of literal, non-metaphorical expressions. As subvariants, different
ways of indicating DM in form and content (Steen 2009; Krennmayr 2011) will
be distinguished, among them signalled metaphor (similes, expressions like
as it were, inverted commas...), direct metaphor (perhaps excluding clichéd
phrases like he is a pig), and extended metaphor. We will also distinguish the
source domain of the metaphor, following the ‘Master Metaphor List’ (Lakoff
et al. 1991) of the Cognitive Linguistics Group. The presentation targets the
questions a, c, and e as listed in the call for papers.
310
AG 13 · Gebäude B4 1, Raum 0.04
References: • Beger, A. (2011). Deliberate metaphors? An exploration of the choice and functions of metaphors in US-American college lectures. metaphorik.de 20, 39–60. • Beger, A. (2015).
Metaphors in psychology genres. Counseling vs. academic lectures. In B. Herrmann and T. Sardinha
(eds), Metaphor in specialist discourse, 53–75. Amsterdam: Benjamins. • Deignan, A., J. Littlemore, & E. Semino (2013). Figurative language, genre and register. Cambridge: Cambridge University Press. • Krennmayr, T. (2011). Metaphor in newspapers. Ph.D. thesis, Vrije Universiteit Amsterdam. • Lakoff, G., J. Espenson, & A. Schwartz (1991). Master metaphor list. available from
http://araw.mede.uic.edu/~alansz/metaphor/METAPHORLIST.pdf. • Lakoff, G. & M. Johnson
(1981). Metaphors we live by. Chicago: University of Chicago Press. • Schklowski, V. (1971). Kunst als
Verfahren. In J. Striedter (ed.), Russischer Formalismus. Texte zur allgemeinen Literaturtheorie und zur
Theorie der Prosa, 3–35. München: Fink. • Steen, G. (2008). The paradox of metaphor: Why we need
a threedimensional model of metaphor. Metaphor and Symbol 23, 213–241. • Steen, G. (2009). Deliberate metaphor affords conscious metaphorical cognition. Cognitive Semiotics 5, 179–197.
On the role of intra-speaker variation for language change
Donnerstag
09.03.2017
13:45 – 14:15
B4 1, 0.04
Richard Waltereit
Newcastle University
[email protected]
Intra-speaker variation as implied by register has long been shown to exist in
the “styles” used in variationist analysis (conversation, reading, word-lists),
i.e., controlled by the researcher. In his search for intra-speaker variation
within conversation, Labov (2013) found that the most vernacular speech is in
oral narrative of events that the speaker has personal knowledge of, and even
more so when those events revolve around three “universal centres of interest”
that invite “dramatic” verbalization: (i) death and danger of death (including
violence, illness, etc.); (ii) sex (including marriage, affairs, etc.); (iii) moral indignation (blame, injustice, social norms etc.).
This finding is of interest to diachronic linguistics. A popular assumption
in historical linguistics is that innovations may originate in potentially hyperbolic, dramatic speech and then spread, by rhetorical devaluation, across the
speech community (e.g. Haspelmath 1999, Detges & Waltereit 2002). The implication of Labov’s findings is that it may be the other way round, namely that
dramatic speech is simply the most vernacular register, and that its features
may spread through the community by ordinary linguistic diffusion without
the need to invoke any rhetorical devaluation. A prediction of this hypothesis is that new variants of a variable have a higher concentration in discourse
contexts that relate to the three centres of interest referred to above than elsewhere. In my talk, I will assess this hypothesis for the spread of French bi-
311
AG13
AG 13 · Register in linguistic theory: Modeling functional variation
partite negation (e.g. ne…pas) at the expense of simple negation (ne) in Old and
Middle French (cf. Detges & Waltereit 2002, Hansen 2009).
References: • Detges, U. & R. Waltereit (2002): Grammaticalization vs. reanalysis: a semantic-pragmatic account of functional change in grammar. Zeitschrift für Sprachwissenschaft 21, 151-195.
• Hansen, M.-B. M. (2009): The grammaticalization of negative reinforcers in Old and Middle
French: a discourse-functional approach. In: M.-B. M. Hansen & J. Visconti (eds.), Current trends in
diachronic semantics and pragmatics, Bingley, 227-251. • Haspelmath, M. (1999): Why is grammaticalization irreversible? Linguistics 37, 1043–1068. • Labov, W. (2013): The language of life and death. The
transformation of experience in oral narrative. Cambridge.
Register variation in OHG: evidences for
register based variation in the recordings of OHG
Donnerstag
09.03.2017
14:15 – 14:45
B4 1, 0.04
AG13
Gohar Schnelle
Humboldt-Universität zu Berlin
Karin Donhauser
Humboldt-Universität zu Berlin
[email protected]
[email protected]
In our talk we will present the results of a pilot study designed to find out
whether register is a relevant parameter when it comes to explaining the
distribution of variants in OHG recording. As a textbase we use the Evangelienharmonie of Otfrid von Weißenburg. This text is particularily suitable
for our study because it was written by a single author (Otfrid von Weißenburg) in a specific place (Weißenburg) and within a short period of time (between 863 and 871). Data is collected with a corpus-based approach from
a register-annotated version of the DDD-Referenzkorpus Altdeutsch (Donhauser et al. 2016)) using the search and visualization architecture ANNIS
(Krause & Zeldes (2016)). Different core-linguistic features are then quantitatively analyzed according to Biber (2009) in order to explore the existence
of patterns distributed in correlation with register. Our study focuses on syntactical phenomena such as verb position in main clauses (V1, V2, VL), realization of pronominal subjects (pro-drop, non-pro-drop), position of attributive
adjectives (N ADJ, ADJ N), realization of verbal complements (infinitives, subordinated clauses).
References: • Biber, D. (2009). Multi-Dimensional Approaches. In: Lüdeling, A. & Kytö, M.
(ed.) Corpus Linguistics. An International Handbook. Vol 2. Mouton de Gruyter, 822-855. • Donhauser, K. (2015). Das Referenzkorpus Altdeutsch. Das Konzept, die Realisierung und die neuen
Möglichkeiten. In: Gippert, J. & Gehrke, R. (ed.) Historical Corpora. Challenges and Perspektives.
Narr, 35-49. • Donhauser, K.; Gippert, J. & Lühr, R. ddd-ad (Version 0.1), Humboldt-Univer-
312
AG 13 · Gebäude B4 1, Raum 0.04
sität zu Berlin. http://hdl.handle.net/11022/0000-0000-7FC2-7 • Fleischer, J.(2006). Zur
Methodologie althochdeutscher Syntaxforschung. In: Beiträge zur Geschichte der deutschen Sprache
und Literatur (PBB) 128(1):25-69. • Krause, T. & Zeldes, A. (2016). ANNIS3: A new architecture for generic corpus query and visualization. Digital Scholarship in the Humanities 2016 (31).
http://dsh.oxfordjournals.org/content/31/1/118 • Schlachter, E. (2012). Syntax und Informationsstruktur im Althochdeutschen. Untersuchungen am Beispiel der Isidor-Gruppe. Winter.
The register-specificity of variation grammars
Freitag
10.03.2017
11:30 – 12:30
B4 1, 0.04
Benedikt Szmrecsanyi
KU Leuven
[email protected]
Variationist linguistics is concerned with how language users choose between
alternative ways of saying or writing the same thing. The set of relevant probabilistic constraints on a particular variation phenomenon constitutes what I
call a “variation grammar”.
The talk addresses the extent to which language users may have available
different variation grammars for different types of situational context (“registers”). Thus, the crucial question is whether or not our linguistic choice making processes differ depending on whether we engage in e.g. informal conversation or write blog entries. This issue is under-researched but loaded theoretically: variationist sociolinguists tend to believe that “internal constraints […]
are normally independent of social and stylistic factors” (Labov 2010:265), but
preliminary evidence (Grafmiller 2014) suggests that there may actually be a
good deal of register-specificity.
I outline a methodology to address this matter empirically, drawing on both
observational corpus evidence (i.e. customary multivariate modeling of linguistic choice making in naturalistic production data) and supplementary rating task experiments in the spirit of Bresnan (2007). To exemplify, I will discuss a case study covering syntactic variation in multiple spoken and written
registers in varieties of English around the world (see Szmrecsanyi et al. 2016).
References: • Bresnan, J. (2007): Is syntactic knowledge probabilistic? Experiments with the English
dative alternation. In: Roots: Linguistics in Search of Its Evidential Base, 75–96. • Grafmiller, J. (2014).
Variation in English genitives across modality and genres. English Language and Linguistics 18, 471–496.
• Labov, W. (2010): Principles of linguistic change. Vol. 3: Cognitive and cultural factors. Wiley-Blackwell.
• Szmrecsanyi, B., Jason G., Heller, B. & Röthlisberger, M. (2016). Around the world in three alternations: Modeling syntactic variation in varieties of English. English World-Wide 37, 109-137.
313
AG13
AG 13 · Register in linguistic theory: Modeling functional variation
Register variation in the use of DRDs in argumentative texts
Freitag
10.03.2017
12:30 – 13:00
B4 1, 0.04
Ines Rehbein
Leibniz ScienceCampus, IDS Mannheim / Universität Heidelberg
[email protected]
AG13
This contribution studies the use of discourse relational devices (DRDs) in argumentative texts in different communicative settings. By DRDs we refer to
linguistic devices that are used to structure and organise the discourse, such
as discourse connectives, adverbials, or linguistic cue phrases. DRDs are highly
ambiguous and polyfunctional and can vary across different dimensions, depending on the medium (spoken vs. written), the discourse situation (monologic vs. dialogic, formal vs. informal), the purpose of communication, and
more. Recent work on argumentation mining has pointed out the important
role of DRDs for analysing argumentation structure (Eckert-Kohler et al. 2015).
Our main interest is in investigating how the different dimensions of variation
can impact the linguistic behaviour of an individual speaker during the production of argumentative texts. The data we use in our analysis are political
articles, interviews and talks by the same author, Noam Chomsky. Our data
covers spoken and written texts and ranges from highly edited to less edited,
including monologic as well as dialogic data.
We follow the tradition of Biber’s register analysis (Biber 1995) and perform
a Principal Component Analysis (PCA), based on raw counts of word forms of
DRDs, to identify the main variables of variance. The PCA allows us to identify DRDs typical for written articles and those that are used in the less edited,
dialogic interview data.
To investigate how the different distribution of DRDs reflects different
strategies used to pursue a communicative purpose, we provide an analysis
of discourse relations according to Prasad et al. (2008), with a focus on causal
relations such as Cause and Result which play a crucial role in argumentation.
References: • Biber, D. (1995): Dimensions of register variation: A cross-linguistic comparison. Cambridge: Cambridge University Press. Amsterdam: Elsevier. • Eckle-Kohler, J., Kluge, R., & Gurevych,
I. (2015): On the Role of Discourse Markers for Discriminating Claims and Premises in Argumentative
Discourse. In: LREC 2015, 2249–2255. Lisbon, Portugal. • Prasad, R., et al. (2008): The Penn Discourse
Treebank 2.0. In: LREC 2008, 2961–2968. Marrakech, Morocco.
314
AG 13 · Gebäude B4 1, Raum 0.04
Register variation across social media
Freitag
10.03.2017
13:00 – 13:30
B4 1, 0.04
Tatjana Scheffler
Universität Potsdam
[email protected]
It is known that speakers can vary their use of linguistic features across different channels and discourse situations, but little previous research has addressed the question which linguistic phenomena are affected. In one exception, Tagliamonte & Denis (2008) study linguistic features of teens in spoken
and instant-messaging conversations, who show a mix of conservative and innovative behaviors in text messaging vs. speech. Our contribution investigates
intra-speaker variation across different social media platforms, which eliminates the differences wrt. transmission mode (oral vs. written). Research questions are: (i) Does the same speaker’s language differ across social media platforms? (ii) Are the observed differences of the same kinds, i.e., can we observe
the formation of register norms? (iii) Which phenomena show variant vs. stable expression across the different registers? We analyze tweets and blog posts
for the same speakers with respect to the following frequent “non-standard”
phenomena: across-the-board capitalization, letter duplication, intensifiers,
abbreviations, novel sound and emotion words, sentence particles, and neologisms. We observe both inter-speaker variation in the usage of the above features, as well as intra-speaker variation across the two types of platforms. All
speakers make much greater use of non-standard linguistic features in their
tweets than in their (full-length) blog posts. In particular, we find: (i) Each
speaker in our study adapts their language use to the platform in both conscious and unconscious ways. (ii) The direction of the observed difference in
feature frequency is always the same, indicating the development of register
norms. (iii) The studied non-standard features all vary across the platforms,
though other phenomena such as word order (Rehbein, 2014) have been shown
to be more stable in previous work.
References: • Tagliamonte, S. A., & Denis, D. (2008). Linguistic ruin? LOL! Instant messaging and
teen language. American speech 83(1), 3-34. • Rehbein, I. (2014). Using Twitter for linguistic purposes:
three case studies. Slides, DGfS 2014.
315
AG13
AG 13 · Register in linguistic theory: Modeling functional variation
Factor analysis of Russian register and linguistic variation
Freitag
10.03.2017
13:30 – 14:00
B4 1, 0.04
Roland Meyer
Humboldt-Universität zu Berlin
Luka Szucsich
Humboldt-Universität zu Berlin
[email protected]
[email protected]
Slavic register phenomena have traditionally been captured by a fixed set of
functional styles (FS). With modern linguistic corpora, registers may be derived bottom-up by a factor analysis of the (non-)occurrence of selected linguistic features across text types (Biber 1995; Baayen 2009, ch.5). In this paper,
we undertake such an analysis on the hand-corrected “gold standard” corpus
of Russian (1.6 Mio tokens), annotated with morphosyntactic tags and (textwide) ascriptions of FS and text types (article, critique, discussion, speech
etc.) –- http://ruscorpora.ru/en/corpora-usage.html. FS-related candidate properties were derived from the relevant literature (28 clearly identifiable features from a set of >600, based, among others, on Kožina et al.
2010). They included (i) lexical occurrences (e.g. particles and adverbs), (ii)
certain morphemes (e.g. in internationalisms), (iii) POS categories (e.g. nouns,
gerunds). A factor analysis with 3 factors was run on these categories across
text types, dimension scores were calculated for factors above a certain threshold, and text types were ranked according to their dimension scores (cf. Biber
1995). Preliminary conclusions are: (i) Factor 1 (main dimension of variation)
supports an interpretation like „reporting actions“ vs. „static, argumentative“.
The written/spoken divide does not seem to be at stake here, discussions and
conversations being at opposite ends of the scale. This is but one of the relevant
distinctions which have been overlooked in the traditional taxonomy of FS. (ii)
Traditional FS cannot explain the opposite scalar positions of certain linguistic features – e.g., AdvP (gerunds) and –acija internationalisms should both
characterize written, especially scientific FS, contrary to their loadings. The
bottom-up register distinctions will be compared to the distributions of welldefined linguistic variables, such as subtypes of sentential noun modifiers.
References: • Biber, D. (1995): Dimensions of register variation. • Baayen, H. (2009): Analyzing
linguistic data. CUP. • Kožina, M.N. et al. (2011): Stilistika russkogo jazyka. Nauka. • Lapteva, O.A.
(2003): Živaja russkaja reč’ s teleėkrana. URSS.
316
Sprachwissenschaft im Stauffenburg Verlag
Elisabeth Leiss /
Sonja Zeman (Hrsg.)
Die Zukunft von
Grammatik –
Die Grammatik der
Zukunkft
Festschrift für Werner
Abraham anlässlich
seines 80. Geburtstags
Studien zur deutschen Grammatik, Band 92
Anfang 2017, ca. 340 Seiten, kart.
ISBN 978-3-95809-543-4 € 64,–
Der Band umfasst Beiträge zu einem Symposium über die
„Zukunft von Grammatik – Die Grammatik der Zukunft“, das
vom 12. bis 13. Februar 2016 zu Ehren des 80. Geburtstags von Werner Abraham an der LMU München abgehalten wurde. Thematisiert wird der Stellenwert von Grammatikschreibung, Grammatiktheorie und Universalgrammatik in
einer Zeit, die sich zunehmend von der Beschreibung und
Erklärung grammatischer Strukturen sowie von übereinzelsprachlichen Generalisierungen abwendet.
Iris Meißner / Eva Lia Wyss (Hrsg.)
Begründen – Erklären –
Argumentieren
Konzepte und Modellierung
in der angewandten Linguistik
Stauffenburg Linguistik, Band 93
Frühjahr 2017, ca. 270 Seiten, kart.
ISBN 978-3-95809-514-4 ca. € 39,80
Sandra Döring /
Jochen Geilfuß-Wolfgang (Hrsg.)
Probleme der
syntaktischen Kategorisierung:
Einzelgänger, Außenseiter und mehr
Stauffenburg Linguistik, Band 90
2016, ca. 330 Seiten, kart.
ISBN 978-3-95809-511-3­ ca. € 49,80
Dmitrij Dobrovol’skij
Kognitive Aspekte
der Idiom-Semantik
Studien zum Thesaurus deutscher Idiome
Eurogermanistik, Band 8
2., aktualisierte und erweiterte Auflage 2016
399 Seiten, kart.
ISBN 978-3-86057-368-6 € 68,–
Olga Heindl
Aspekt und
Genitivobjekt
Eine kontrastiv-typologische
Untersuchung zweier
Phänomene der historischen
germanischen Syntax
Stauffenburg Mediävistik, Band 1
Ende 2016, ca. 285 Seiten, kart.
ISBN 978-3-95809-850-3 € 49,80
Erwin Tschirner / Olaf Bärenfänger /
Jupp Möhring (Hrsg.)
Deutsch als fremde Bildungssprache
Das Spannungsfeld von Fachwissen,
sprachlicher Kompetenz, Diagnostik
und Didaktik
Deutsch als Fremd- und Zweitsprache,
Schriften des Herder-Instituts, Band 7
2016, 266 Seiten, kart.
ISBN 978-3-95809-071-2 € 39,80
Bildungssprachliche Kompetenzen in der Fremd- und Zweitsprache Deutsch stellen für Schüler und Studierende eine
der wichtigsten Voraussetzungen für ihren fachlichen Erfolg dar. Die Beiträge diese Bandes unterbreiten Vorschläge, wie bildungssprachlicher Bedarf empirisch ermittelt
und beschrieben werden kann, wie Tests und Diagnoseinstrumente zur Messung bildungssprachlicher Kompetenz
gestaltet sein sollten und wie sprachliche Fertigkeiten an
Schule und Universität gezielt gefördert werden können.
Stauffenburg Verlag GmbH
Postfach 25 25 D-72015 Tübingen www.stauffenburg.de
Postersession
der Sektion Computerlinguistik
Organisation: Vera Demberg, Asad Sayeed
Ort: Gebäude B4 1 (Foyer)
CL-Postersession
Mittwoch, 08. 03. 2017
15:45–16:30
Jagoda Bruni, Daniel Duran & Grzegorz Dogil
Simulating Language Change in Tswana
Max Ionov & Christian Chiarcos
A vector-based phonological search for cognates across
dictionaries
Anke Lüdeling, Carolin Odebrecht, Gohar Schnelle &
Laura Perlitz
Die Normalisierung von Komposita in frühneuhochdeutschen
Texten am Beispiel des RIDGES-Korpus
Felix Bildhauer & Roland Schäfer
A lexico-grammatical document annotation framework for
very large German corpora
Asad Sayeed, Pavel Shkadzko & Vera Demberg
A large corpus automatically annotated with semantic role
information
Heike Zinsmeister & Jessica Katharina Sohl
How to make manual annotation more efficient – Ensemble
dependency parsing as a pre-processing step to obtain
high-quality annotation in an efficient way
CL Poster
Donnerstag, 09. 03. 2017
10:30–11:15
Jannik Strötgen
Multilingual and Domain-sensitive Temporal Tagging with
HeidelTime
Laura Bostan & Jonathan Oberländer
Color distributions in German poetry
Asad Sayeed, Xudong Hong & Vera Demberg
Roleo: distributional space visualisation for thematic fit
modeling
Holger Grumt Suárez, Natali Karlova-Bourbonus &
Henning Lobin
Semi-automatische TEI Repräsentation der Diskursstruktur
eines deutschsprachigen Blogkorpus
321
CL-Postersession
Petra Steiner
Using contextual information for deep-level morphological
analysis
Zarah Weiß, Detmar Meurers, Anke Lüdeling, Uwe
Springmann, Brian McWhinney & John Kowalski
A digital infrastructure to study Latin and historical German
12:45–13:45
Christian Chiarcos, Ralf Plate & Maria Sukhareva
A clash of methods? Comparing quantitative and qualitative
studies of word order flexibility in historical German
Mohammad Fazleh Elahi, Dimitra Anastasiou & Hui Shi
A unified approach to dialogue model for situated referential
grounding
Laura Bostan & Jonathan Oberländer
Distributional and neural gastronomics
Markus Gärtner & Kerstin Eckart
In support of self-assessment – exploiting available
information from tools
Stephan Druskat, Thomas Krause & Carolin Odebrecht
Agile creation of multi-layer corpora with corpus-tools.org
CL Poster
322
CL-Postersession
Simulating language change in Tswana
Jagoda Bruni
Universität Stuttgart
Daniel Duran
Universität Stuttgart
Grzegorz Dogil
Universität Stuttgart
[email protected]
[email protected]
[email protected]
stuttgart.de
stuttgart.de
stuttgart.de
Mittwoch
08.03.2017
15:45 – 16:30
B4 1, Foyer
This study describes the influence of socio-political changes in the South
African phylum on the current shape of Tswana, a Bantu language of the Sotho
group. Tswana has phonetically marked post-nasal devoicing which has been
described in detail by Zsiga et al. (2006) and Coetzee & Pretorius (2010). It
has been well-documented that post-nasal devoicing (/mb/ → [mp]) is a phonetically unintuitive phenomenon which costs more articulatory effort than
producing sequences of nasals followed by voiced stops (Wesbury & Keating,
1984). We use simulations based on Social Impact Theory (Nettle, 1999). In particular, we model social dynamics of Tswana speakers by assigning them either
to small world (parochial) communication networks or whole world intensive
code switching networks typical of the present linguistic situation in SA. With
our in-house instantiation of Wedel’s (2004) exemplar model, we examine the
behavior of contrasting voicing realizations across speakers. The model simulates emergence and maintenance of contrast in the context of speaker/hearer
interactions. Exemplar-based categories compete for assignment and storage
of incoming percepts and the production process is biased towards gesture reuse. By employing this model, we can ascertain, via simulation, how the contrasting realizations can emerge and stabilize within a generation, inspect the
selection processes which yield these realizations, and examine the acoustic
changes which bring about the contrast.
References: • Coetzee, A.W. and Pretorius, R. (2010): Phonetically grounded phonology and sound
change: The case of Tswana labial plosives. J. of Phonetics, 38(3), 404–421. • Nettle, D. (1999): Using
Social Impact Theory to simulate language change. Lingua, 108(2–3), 95–117. • Wedel, A. (2004): Category competition drives contrast maintenance within an exemplar-based production/perception
loop. In: Proc. of ACL Special Interest Group in Computational Phonology. 1–10. • Westbury, J.R. and
Keating, P.A. (1986): On the naturalness of stop consonant voicing. J. of Linguistics, 22(1), 145. • Zsiga,
E., Gouskova, M. and Tlale, O. (2006): On the status of voiced stops in Tswana: Against *ND. In: The
Proc. of NELS 36. 721–734.
323
CL Poster
CL-Postersession
Mittwoch
08.03.2017
15:45 – 16:30
B4 1, Foyer
CL Poster
A vector-based phonological search for cognates across dictionaries
Max Ionov
Goethe-Universität Frankfurt
Christian Chiarcos
Goethe-Universität Frankfurt
[email protected]
[email protected]
We present an ongoing work on searching phonologically similar words in related languages in and across dictionaries. This is a part of a larger project devoted to unravelling both synchronic and diachronic lexical connections in related less-resourced languages.
The task of searching for phonologically similar words has two applications:
in one language, applied to a corpus, it can cluster inflectional forms of one lexeme, which is extremely useful for low-resourced languages with high inflection and no POS-taggers available. In several related languages, applied across
dictionaries or wordlists, it can detect possible cognates – words with the common etymological origin. The latter is a common method for dialectometry (e.g.
Heeringa et al. 2006), but it was also applied to the field of historical linguistics
(List and Moran 2013).
We present an approach that employs PHOIBLE dataset, the universal
phonological inventory (Moran et al. 2014) and vector-based phoneme representation and compare it with several well-known approaches, starting from
simple yet popular minimum edit distance approach (Holman et al. 2011 inter
alia) to more sophisticated approaches like LexStat (List 2012).
Using linguistic insight, we examine the limitations of automatic approaches and propose directions for overcoming them.
References: • Heeringa, W., Kleiweg, P., Gooskens, C., Nerbonne, J. (2006): Evaluation of String
Distance Algorithms for Dialectology. Proceedings of the Workshop on Linguistic Distances, 51–62.
• Holman, E.W., Brown C.H., Wichmann S., Müller A., Velupillai V., Hammarström H., Sauppe S.,
Jung H., Bakker D., Brown P., Belyaev O., Urban M., Mailhammer R., List J.-M., Egorov D. (2011): Automated Dating of the World’s Language Families Based on Lexical Similarity. Current Anthropology
52(6), 841–875. • List, J.-M. 2012: LexStat. Automatic detection of cognates in multilingual wordlists.
Proceedings of the EACL 2012 Joint Workshop of Visualization of Linguistic Patterns and Uncovering Language History from Multilingual Resources, 117–125. • List, J.-M., Moran S. 2013: An Open
Source Toolkit for Quantitative Historical Linguistics. Proceedings of the 51st Annual Meeting of
the ACL: System Demonstrations, 13–18. • Moran S., McCloy D., Wright R. 2014: PHOIBLE Online,
http://phoible.org/.
324
CL-Postersession
Die Normalisierung von Komposita in fnhd. Texten am Beispiel des
RIDGES-Korpus
Mittwoch
08.03.2017
15:45 – 16:30
B4 1, Foyer
Anke Lüdeling, Carolin Odebrecht, Laura Perlitz, Gohar Schnelle
(Humboldt-Universität zu Berlin, Institut für deutsche Sprache und Linguistik,
Lehrstuhl für Korpuslinguistik und Morphologie)
Anke.Lüdeling;odebreca;perlitzl;[email protected]
Wie lassen sich Komposita im Frühneuhochdeutschen variationistisch untersuchen? In dieser Sprachstufe existiert keine graphematische Norm, die die
Getrennt- und Zusammenschreibung von Komposita regelt. Demzufolge ist
ihre zweifelsfreie Identifikation nicht möglich, da konzeptuelle Abgrenzungsprobleme gegenüber Syntagmen wie z.B. Nominalphrasen mit Adjektivattributen (leinen Tue chlein vs. Leinentüchlein) oder vorangestellten Genitivattributen (Widers Teuffels Biſſz vs. Teufelsbiss) bestehen (vgl. Pavlov 1983,
105). Um Komposita in einem Korpus systematisch suchen und analysieren
zu können, ist daher eine Normalisierung notwendig. Dies zeigen wir anhand
von RIDGES, einem diachrones Kräuterkundekorpus. Auf folgenden linguistischen Ebenen wird normalisiert (vgl. Belz et al. 2016): graphematisch (Teuffels Biſſz → Teufelsbiss); phonologisch (wieh=rauch → Weihrauch); morphologisch
(Kolfewer → Kohlenfeuer). Bei ambigen Fällen greift das Prinzip der Konservativität (auff das Stro das leinen tue chlein → das leinene Tüchlein). Weitere
Annotationen helfen, verschiedene Forschungsfragen über Komposita zu untersuchen, z.B. die Getrennt- und Zusammenschreibung (vgl. Perlitz 2014, 3132).
References: • Belz et al. (2016): Annotationsrichtlinien zu Ridges Herbology Version 5.0, HU Berlin.
https:// hu.berlin/ridges_annotationsrichtlinien_v5. • Lüdeling, A.; Odebrecht, C.; Zeldes, A.:
RIDGES-Herbology (Version 5.0), Humboldt-Universität zu Berlin. https://hu.berlin/ridges. • Odebrecht, C., Belz, M., Zeldes, A., Lüdeling, A., Krause, T. (eingereicht): RIDGES Herbology - Designing a
Diachronic Multi-Layer Corpus. • Pavlov, V. M. (1983): Zur Ausbildung der Norm der deutschen Literatursprache (1470- 1730). Berlin: Akademie-Verlag. • Perlitz, L. (2014): Konkurrenz zwischen Wortbildung
und Syntax - historische Entwicklung von Benennung (Bachelorarbeit, Humboldt-Universität zu Berlin).
http://edoc. hu-berlin.de/master/perlitz-laura-2014-08-08/PDF/perlitz.pdf.
325
CL Poster
CL-Postersession
Mittwoch
08.03.2017
15:45 – 16:30
B4 1, Foyer
CL Poster
COReX und COReCO: A lexico-grammatical document annotation
framework for large German corpora
Felix Bildhauer
Institut für Deutsche Sprache,
Mannheim
Roland Schäfer
Freie Universität Berlin
(DFG SCHA1916/1-1)
[email protected]
[email protected]
As an unstructured collection of documents, very large modern corpora
would be of little use for many linguists. The automatic creation of linguistically relevant meta data is therefore crucial. We present an open-source annotation framework developed for the German DECOW and DeReKo corpora that
provides (1) topical and (2) grammatical text categorization. The classification
methods are based on document-internal features because this is the most viable approach for automatic classification, and it has conceptual advantages.
Classifying documents by internally defined text types rather than situationally defined categories such as register or genre (Lee 2001) avoids many and potentially insolvable conceptual problems of determining and operationalizing
the true set of register and genre categories (see the less than satisfying results
in Biber & Egbert 2016). We classify documents by (1) distributions of words
and (2) distributions of grammatical features.
For the lexical classification, we use topic modeling algorithms (e. g., Latent Dirichlet Allocation; Blei et al. 2003) combined with supervised machine
learning to annotate documents with a coarse-grained set of twenty easily interpretable topic domains (such as Politics or Sports). For the grammatical classification, we annotate each document with automatically extracted features
similar to Biber’s (1988) features. We generate over thirty per-document features such as the density of modal verbs, genitives, or passive constructions.
Users have access to the raw distributions of these features and aggregated
categories obtained by clustering.
References: • Biber, Douglas. 1988. Variation across speech and writing. Cambridge, MA: Cambridge
university Press. • Biber, Douglas and Egbert, Jesse. 2016. Using Grammatical Features for Automatic
Register Identification in an Unrestricted Corpus of Documents from the Open Web. J. Res. Des. &
Stat. in Ling. & Comm. Sc. 2, 3–36. • Blei, David M., Ng, Andrew Y. and Jordan, Michael I. 2003. Latent dirichlet allocation. J. M. L. Res. 3, 993–1022. • Lee, David. 2001. Genres, registers, text types,
domains, and styles: Claryfying the concepts and navigating a path through the BNC jungle. L. Learn.
& Tech. 5(3), 37–72.
326
CL-Postersession
A large corpus automatically annotated with semantic role
information
Asad Sayeed
Universität des
Saarlandes
Pavel Shkadzko
Universität des
Saarlandes
Vera Demberg
Universität des
Saarlandes
[email protected]
[email protected]
[email protected]
Mittwoch
08.03.2017
15:45 – 16:30
B4 1, Foyer
We present Rollenwechsel-English (RW-eng), large, automatically labeled
corpus based on the ukWaC web crawl corpus and the British National Corpus (BNC). Our automatic annotations contain predicate-argument relation
information at a sentence level. The basic annonation is performed by a combination of semantic role labelling using the SENNA SRL tool and MALT dependency parsing. SENNA provides PropBank-style role labelling over whole
phrases through a sequence-labelling model. We use MALT parses to identify
predicate-head relations within the text spans found by SENNA.
The large size of this corpus (approx. 78 million sentences) makes it useful for distributional semantic modeling, in which semantic relations are required, but human annotation at scale is unrealistic to obtain. It has already
been successfully used in large-scale models of thematic fit/selectional preferences, both count-based and neural. The presence of both whole phrases and
identified heads allows for richer semantic modelling applications.
The XML-formatted, UTF8-compliant corpus lists every automaticallyidentified predicate per sentence that is found in the source corpora and
with these predicates, each role-filling phrase, including the automaticallydiscoved head. Head-finding is performed by a cascading series of heuristics,
and the found heads are listed with the heuristic used to identify them. RW-EN
is available at http://rollen.mmci.uni-saarland.de/RW-eng/.
References: • Collobert, R., Weston, J., Bottou, L., Karlen, M., Kavukcuoglu, K., Kuksa, P. (2011): Natural language processing (almost) from scratch. JMLR. • Sayeed, A., Demberg, V., Shkadzko, P. (2015):
An exploration of semantic features in an unsupervised thematic fit evaluation framework. IJCoL:
1(1), 25–40. • Tilk, O., Demberg, V., Sayeed, A., Klakow, D., Thater, S. (2016): Event participant modelling with neural networks. EMNLP.
327
CL Poster
CL-Postersession
Mittwoch
08.03.2017
15:45 – 16:30
B4 1, Foyer
How to make manual annotation more efficient
- Ensemble Dependency Parsing as a preprocessing step to obtain
high-quality annotation in an efficient way
Heike Zinsmeister
Universität Hamburg
Jessica Sohl
Universität Hamburg
[email protected]
[email protected]
CL Poster
The “multistrategy approach” of ensemble tagging is a method of tagging texts
automatically by a cluster of taggers, resulting from training different learning
algorithms on the same data (von Halteren et al. 2001, p. 201, e.g., Rehbein et
al. 2014).
In this poster, we will present ongoing work of using this method for dependency parsing as a pre-processing step for manual annotation of German
textbook texts. An ensemble of three parsers is used to annotate sentences automatically and to identify sentences that need to be checked manually. We
deviate from the original ensemble idea by including a rule-based system in
the cluster in addition to the machine learning approaches. We will present
qualitative analyses of sentence types that the ensemble fails to parse correctly
and provide cross-validation results for two training corpora (Hamburg Dependency Treebank (Menzel et al. 2014), TIGER Corpus v2.2 dependency conversion (Seeker & Kuhn 2012)) and the accuracy of adapting the trained ensemble to our target domain of textbook texts.
References: • Foth, Kilian, Köhn, Arne, Beuck, Niels, Menzel, Wolfgang (2014): Because Size Does
Matter: The Hamburg Dependency Treebank. In: Proceedings of the Language Resources and Evaluation
Conference 2014 / European Language Resources Association. • Rehbein, Ines, Schalowski, Sören, Wiese,
Heike (2014): The KiezDeutsch Korpus (KiDKo) Release 1.0. In: Proceedings of the 9th International Conference on Language Resources and Evaluation, p. 3927-3934. Reykjavík, Iceland. • Seeker, Wolfgang,
Kuhn, Jonas (2012): Making Ellipses Explicit in Dependency Conversion for a German Treebank. In:
Proceedings of the 8th International Conference on Language Resources and Evaluation, p. 3132-3139, Istanbul, Turkey. • Van Halteren, Hans, Daelemans, Walter, Zavrel, Jakub (2001): Improving Accuracy in
Word Class Tagging through the Combination of Machine Learning Systems. In: Computational Linguistics 27(2), p. 199-229.
328
CL-Postersession
Multilingual Domain-sensitive Temporal Tagging with HeidelTime
Jannik Strötgen
Max Planck Institute for Informatics, Saarland Informatics Campus
Donnerstag
09.03.2017
10:30 – 11:15
B4 1, Foyer
[email protected]
Due to the prevalence of temporal expressions in diverse types of documents
and the importance of temporal information in any information space, the detection of temporal expressions and the normalization of their semantics to
some standard format is an important task in NLP. This task is known as temporal tagging, and many types of applications, e.g., in information retrieval,
question answering, and digital humanities, can benefit from the output of
temporal taggers to provide more meaningful and useful results.
Research on temporal tagging has focused on processing English news articles for quite a long time. Only more recently, challenges of texts of other
domains have also been studied (e.g., clinical documents, literary narratives,
and colloquial text), and further languages have been addressed. A general
overview on temporal tagging is given by Strötgen & Gertz (2016).
In this system demonstration, we present the temporal tagger HeidelTime.
It uses Timex3 tags, which are defined in the temporal markup language
TimeML (Pustejovsky et al., 2005), to annotate dates (July 2017), times (9
pm), durations (two days), and set expressions (twice a week). HeidelTime is a rule-based system initially developed for English news articles, but
later extended to process more languages and to tackle the challenges of different domains. It is publicly available (https://github.com/HeidelTime/
heideltime/) and constantly maintained, e.g., it was recently extended with
automatically created language resources for more than 200 languages in addition to the 13 languages for which resources have been manually developed
by several researchers.
The main goals of this demonstration are (i) to show how to use HeidelTime
for multilingual, domain-sensitive temporal tagging, (ii) to explain how HeidelTime’s language resources can be adapted without even touching HeidelTime’s source code, and (iii) to discuss further application scenarios that can
benefit of a temporal tagger’s output.
References: • Pustejovsky et al. (2005): Temporal and Event Information in Natural Language Text.
Language Resources and Evaluation, 39(2-3):123–164. • Strötgen & Gertz (2016): Domain-Sensitive
Temporal Tagging. Synthesis Lectures on Human Language Technologies, Morgan & Claypool Publishers.
329
CL Poster
CL-Postersession
Color distributions in German poetry
Donnerstag
09.03.2017
10:30 – 11:15
B4 1, Foyer
CL Poster
Laura Bostan
Trento University
Jonathan Oberländer
Trento University
[email protected]
[email protected]
We investigate the usefulness of color frequencies in identifying the period
in which a poetic work was created. We predict that these features alone aren’t
able to discriminate time periods with great accuracy, but by adding them a
classifier can get better results.
McManus (1983) showed that the distribution of color words in English and
Chinese poetry correlates with the order of evolution of color words across
languages (Berlin and Kay, 1963). Later, McManus (1997) continued the experiment computationally on a larger English dataset. We also intend to repeat
this study for German poetry, and possibly for other languages.
We crawled all poems available on Project Gutenberg-DE1 .The number of different authors was 299, most of which are German, Austrian or Swiss. The rest
are of different nationalities (with their works translated to German). After
the crawling, we counted the color word distribution for each author using a
set of regular expressions.
We then balanced the raw counts by dividing the occurrence count of a color
to the count of all word occurrences for each author. Using Wikipedia, the list
of authors was automatically enriched with information about their years of
birth and death. We intend to further enrich it with a label of literary period.
As we were interested in seeing whether the color distribution changes over
time, we created a small visualization2 to explore the data, but no obvious development can be seen.
We are currently working on extending the experiments to include new
datasets in languages other than German. Here, interesting future work would
be to investigate the relationship between poetry in different source languages
and different literary currents.
References: • McManus, I. C. (1983): Basic colour terms. Language and Speech 26, 247–252. • McManus, I. C. (1997): Half-a-million basic colour words: Berlin and Kay and the usage of colour words
in literature and science. Perception 26, 367–370. • Berlin B., Kay P. (1969): Basic Color Terms: Their
Universality and Evolution. Berkeley, CA.
1
2
http://gutenberg.spiegel.de/genre/gedicht-poem
https://jsfiddle.net/4w6vx11s/1/embedded/result/
330
CL-Postersession
Roleo: distributional space visualisation for thematic fit modeling
Asad Sayeed
Universität des
Saarlandes
Xudong Hong
Universität des
Saarlandes
Vera Demberg
Universität des
Saarlandes
[email protected]
[email protected]
[email protected]
We demonstrate Roleo, a web tool for visualizing selectional preferences.
Different structured vector space models trained on large amounts of data
(e.g., Distributional Memory, word embeddings) are in the back end of the
tool and allow us to query different thematic role fillers of a given verb. For
example, one can ask for agent, patient, or instrument of “eat” and can compare prototypical role fillers to a given role filler (e.g., “sausage” as a patient
or “knife” as an instrument). Our tool provides a variety of options for visualizing such queries and comparing the query results of different models. These
vector space models can be used for language applications involving the prediction of predicates and event participants.
Vector-space models differ in dimensionality and many other parameters;
visualization helps us compare and hypothesize about fine-grained differences between models on this task. Roleo is a Django-based web tool that allows users to qualitatively compare different approaches to constructing distributional spaces for the selectional preferences task. Roleo’s basic paradigm
is to construct a representative role-filler for a given verb-role from the vector
space, and then surround this centroid with candidate noun fillers projected
down from higher dimensions (via, e.g., singular value decomposition) to a
two-dimensional canvas.
The source code for Roleo is provided at
http://github.com/tony-hong/roleo and can be downloaded and easily
customized to include new kinds of distributional spaces. Roleo can be used
directly at http://roleo.coli.uni-saarland.de/.
References: • Baroni, M., Lenci, A. (2010): Distributional memory: a general framework for corpus-based semantics. Computational Linguistics: 36(4):673–721. • Sayeed, A., Demberg, V., Shkadzko,
P. (2015): An exploration of semantic features in an unsupervised thematic fit evaluation framework.
IJCoL: 1(1), 25–40. • Sayeed, A., Greenberg, C., Demberg, V. (2016): Thematic fit evaluation: an aspect
of selectional preferences. ACL RepEval. • Tilk, O., Demberg, V., Sayeed, A., Klakow, D., Thater, S.
(2016): Event participant modelling with neural networks. EMNLP.
331
Donnerstag
09.03.2017
10:30 – 11:15
B4 1, Foyer
CL Poster
CL-Postersession
Donnerstag
09.03.2017
10:30 – 11:15
B4 1, Foyer
Semi-automatische TEI Repräsentation der Diskursstruktur eines
deutschsprachigen Blogkorpus
Holger Grumt Suárez
Natali KarlovaHenning Lobin
Justus-Liebig-Universität
Bourbonus
Justus-Liebig-Universität
Gießen
Justus-Liebig-Universität
Gießen
Gießen
[email protected]@unigiessen.de
Natali.Karlova-
giessen.de
[email protected]
CL Poster
Das Poster „Semi-automatische TEI Repräsentation der Diskursstruktur eines
deutschsprachigen Blogkorpus“ informiert über das Vorgehen sowie die ersten Forschungszwischenergebnisse und die weiteren Ziele der Kompilierung
und Annotation eines deutschsprachigen Blogkorpus.
Bislang existiert kein Standard für das Repräsentieren von sogenannten
Computer-Mediated Communication-Daten (kurz CMC), allerdings arbeitet
die TEI CMC Special Interest Group (vgl. Beißwenger 2016) seit 2013 an einem
Schema für die Repräsentation von CMC-Genres. Unser Forschungsvorhaben
leistet einen Beitrag zur Standardisierung der Text Encoding Initiative (TEI)
für die CMC. Das Hauptziel des Vorhabens umfasst die semi-automatische
Kompilation sowie die Repräsentation der Blogdiskursstruktur. Dabei sollen
die Relationen zwischen den textuellen und multimodalen Elementen (Blogbeiträge, Kommentare, Hyperlinks, Bilder und Töne) und den verschiedenen
Textproduzenten (Blogger, Kommentatoren) abgebildet werden. Die Grundlage des Korpus bildet das Wissenschaftsblogportal SciLogs – Tagebücher
der Wissenschaft (SciLogs 2016). Zusammenfassend soll das Poster nicht nur
unser Vorhaben vorstellen, sondern auch einen Einblick in unser grundsätzliches Vorgehen bei der Erstellung eines CMC-Korpus geben.
References: • Abendroth-Timmer, D. et al. (2014). Corpus d’apprentissage INFRAL (Interculturel Franco-Allemand en Ligne). Banque de corpus CoMeRe. Ortolang.fr: Nancy. https://hdl.handle.net/11403/
comere/cmr-infral. • Beißwenger, M. (2016). SIG:Computer- Mediated Communication. Http://wiki.
teic.org/index.php/SIG: Computer-Mediated_Communication. • SciLogs (2016). SciLogs. Tagebücher
der Wissenschaft. Spektrum der Wissenschaft. • WebCorp (2013). Birmingham Blog Corpus. WebCorp:
Linguist’s Search Engine. Birmingham City University.
332
CL-Postersession
Using Contextual Information for
Deep-Level Morphological Analysis
Donnerstag
09.03.2017
10:30 – 11:15
B4 1, Foyer
Petra Steiner
Institut für Deutsche Sprache
[email protected]
For detailed semantic processing, e.g. in information retrieval or the building
of terminology databases, the recognition of hierarchical structures is a prerequisite. But concerning these structures, many word forms are leading to
ambiguous structure interpretations. Using ontologies of specific domains can
be helpful, if available (see Bretschneider & Zillner 2015 for compound splitting). Information about general semantic similarities (e.g. Ziering et al. 2016)
does not take into account the ambiguities of linguistic forms if drawn from
large corpora. By contrast, the methodological framework of this investigation
builds on the hypothesis that morphological analysis can be improved by the
specific contextual information of the lexical items without necessitating an
ontology or other semantic networks.
Steiner & Ruppenhofer (2015) and Steiner (2016) developed a method for
building parts of morphological structures by using counts from a morphological database and a corpus for computing the weighting measures, thereby using a wide notion of context. The current approach uses more restricted context definitions, as it works with frequencies of smaller and more homogenous
texts and corpora. We use a corpus of a small domain, gather information from
wider and narrower contexts and show to what extent these can improve morphological analyses.
References: • Bretschneider, C. & S. Zillner (2015): Semantic Splitting of German Medical Compounds. In: Král, P. & Matoušek, V., eds. Text, Speech, and Dialogue: Proceedings of the 18th International
Conference, TSD 2015, Pilsen, Czech Republic, September 14–17, 2015. Cham: Springer International Publishing. 207–215. • Steiner, P. (2016): Kontextbasiertes morphologisches Parsing. Poster presentation at
the 38rd Annual Conference of the German Linguistic Society (DGfS 2016), University of Konstanz, February
24–26, 2016. • Steiner, P. & J. Ruppenhofer (2015): Growing Trees from Morphs: Towards Data-Driven
Morphological Parsing. In: Fisseni, B., B. Schröder & T. Zesch, eds. Proceedings of the International Conference of the German Society for Computational Linguistics and Language Technology (GSCL 2015), University of Duisburg-Essen, Germany, September 30 – October 2, 2015. 49–57. • Ziering, P., S. Müller & L. van
der Plas. (2016): Top a Splitter: Using Distributional Semantics for Improving Compound Splitting.
In: Proceedings of the 12th Workshop on Multiword Expressions, Berlin, Germany, August 7–12, 2016. 50–55.
333
CL Poster
CL-Postersession
A Digital Infrastructure to Study Latin and Historical German
Donnerstag
09.03.2017
10:30 – 11:15
B4 1, Foyer
CL Poster
Zarah Weiß
Universität Tübingen
Gohar Schnelle
Humboldt Universität zu Berlin
[email protected]
[email protected]
We present the current state of the LangBank project, which is developing
a web-based corpus infrastructure dedicated to support the study of Latin
and Early New High German (ENHG). It provides authentic language input
enhanced by computational and corpus linguistic methods: We are working
on providing in-line translations and adjusting Weiß & Meurers’ (submitted)
complexity analysis to Latin and ENHG. The resulting multi-layer corpora
may be queried for grammatical constructions, and texts may be grouped together based on similar complexity or vocabulary. Currently, we are designing a small, but expandable data basis: For Latin, we augmented the standardized editions of widely taught classical texts from the LatinLit corpus (Almas
& Beaulieu 2016). For ENHG, we use diplomatic and normalized texts with
highly variable word order, grammar, and spelling from the RIDGES corpus
(Odebrecht et al. submitted). We addressed the lack of standardized punctuation in ENHG by introducing our own guidelines for manual, non-graphematic
sentence segmentation (Weiß & Schnelle to appear). Also, we investigate the
applicability of automatic normalization approaches to augment the data basis. We started to design two interfaces for our resource, which we will make
freely accessible online: For complex linguistic queries, we converted all annotations layers to the ANNIS format using Pepper (Krause & Zeldes 2014). For
assisted reading, we are working on another interface featuring in-line translations, vocabulary information, and text selection based on text complexity
and topic or specific linguistic constructions.
References: • Almas, B. & M.-C. Beaulieu (2016): The Perseids Platform: Scholarship for all! Digital Classics Outside the Echo-Chamber: Teaching, Knowledge, 171–186. • Krause, T. & A. Zeldes (2014):
ANNIS3: A New Architecture for Generic Corpus Query and Visualization. Digital Scholarship in the
Humanities 33(1), 118–139. • Odebrecht, C., M. Belz, A. Zeldes, A. Lüdeling, T. Krause (Submitted):
RIDGES Herbology - Designing a Diachronic Multi-Layer Corpus. • Weiß, Z. & G. Schnelle (To appear): Sentence Segmentation Guidelines for Early New High German. • Weiß, Z. & D. Meurers (Submitted): Fine-Grained Linguistic Modeling of Textual Complexity Improves German L1 Grade Level
Assessment. COLING Workshop on ”Computational Linguistics for Linguistic Complexity”.
334
CL-Postersession
A clash of methods? Comparing quantitative and qualitative studies
of word order flexibility in historical German
Christian Chiarcos
Goethe-Universität
Frankfurt
Ralf Plate
Universität Trier
[email protected]
Donnerstag
09.03.2017
12:45 – 13:45
B4 1, Foyer
Maria Sukhareva
Goethe-Universität
Frankfurt
[email protected]
[email protected]
frankfurt.de
frankfurt.de
Qualitative and quantitative methods frequently seem to point to different results. This abstract represents an ongoing effort to investigate how the results
obtained by the two methods can be brought in accordance with each other on
the example of German word order. In its historical dimension, ordering preferences for direct (DO) and indirect objects (IO) have been studied by (Speyer,
2011) with a corpus-based, but qualitative methodology. Speyer retrieved
corpus samples using a pre-established list of verbs, covering the Early Modern High German (EMHG) and Modern High German (DE) periods, using a
philologically sound and balanced sample of texts from different genres. He
observed an apparent increase of word order flexibility since the 16th c., with
earlier samples indicating a higher degree of stability.
We describe results of a quantitative pilot study of the same phenomenon.
We sampled both Speyer’s texts and larger corpora and extended the analysis
to Old High German, Old Saxon, Old English and Gothic.
In DE middle field, the DO tends to follow the IO. However, Old English, Old
Saxon and Gothic show 40%-50% “non-canonical” DO-IO order, and remaining at high levels until Middle High German. The drop in usage of IO-DO constructions in favor of canonical DO-IO starts in the 14th c. and continues till
the 17th century. Our quantitative, but partially heuristic analysis confirms
Speyer’s flexibility minimum in the 14th -15th c.c. but fails to confirm a significant increase in word order flexibility since then: Though the decrease continues over the 15th-16th c.c., the difference is not statistically significant and
we could not confirm that this decrease continued later on. Our poster elaborates on these findings and discusses possible explanations for these apparent
deviations.
References: • Speyer, A. (2011). Die Freiheit der Mittelfeldabfolge im Deutschen. Ein modernes
Phänomen. Beiträge zur Geschichte der deutschen Sprache und Literatur PBB, 133:14–31.
335
CL Poster
CL-Postersession
A unified approach to dialogue model for situated referential
grounding
Donnerstag
09.03.2017
12:45 – 13:45
B4 1, Foyer
Mohammad Fazleh Elahi Dimitra Anastasiou
Ludwig Maximilian
Luxemburg Institute of
University of Munich
Science and Technology
[email protected]
Hui Shi
Universität Bremen
[email protected]
[email protected]
muenchen.de
CL Poster
This poster presents a spoken dialogue system, which combines the AgentOriented Dialogue Management (AODM) model (Ross & Bateman, 2009a) and
the generalized dialogue modelling approach proposed in Shi et al. (2011) for
referential grounding. Benefiting from theories, the approach deals with user
description and the mental state (of the participants) to enable effective dialogues to make the referential task successful. The aim of the dialogue system
is to identify objects that user refers to in the environment. To demonstrate
the application of this approach to human and robot interaction, the model is
implemented on the backbone of situated dialogues (Ross & Bateman, 2009b).
In this framework, the user spoken utterance is converted to text and then
sent to the language analyser module applying CCG parsing. The parse is then
analysed by “linguistically motivated ontology” (Bateman et al, 2010) that provides spatial semantics. The system then relates the language with the physical
world. The dialogue management then plans dialogues that are then given to
the language generator. Unlike the dialogue graph based approach; the unified
approach is capable of enabling clarification dialogues based on object description and spatial relations.
References: • Bateman, J. A., Hois, J., Ross, R., & Tenbrink, T. (2010). A linguistic ontology of space
for natural language processing. Artificial Intelligence, 174 (14), 1027-1071, September 2010. • Ross,
R., & Bateman, J. A. (2009a). Agency & Information State in Situated Dialogues: Analysis & Computational Modelling. Proceedings of DiaHolmia 2009 Workshop on the Semantics and Pragmatics
of Dialogue. • Ross, R., & Bateman J. A. (2009b). DAISIE: Information state dialogues for situated
systems.12th International Conference on Text, Speech and Dialogue (TSD), 2009, September 13-17,
2009. • Shi, H., Jian, C., & Rachuy, C. (2011). Evaluation of a unified dialogue model for human-computer interaction. In International Journal of Computational Linguistics and Applications (IJCLA),
Vol 2, S. 155-173. Bahri.
336
CL-Postersession
Distributional and Neural Gastronomics
Laura Bostan
Trento University
Jonathan Oberländer
Trento University
[email protected]
[email protected]
Donnerstag
09.03.2017
12:45 – 13:45
B4 1, Foyer
The goal of this project is to explore the potential for computational creativity in gastronomy by developing a data-driven recipe generation system. The
completed system should be able to learn events from a corpus of recipes, learn
which events follow which by looking at ordered event co-occurrences, and
then sequence events to form the basis of a new recipe using a model trained
on event sequences of previous recipes.
We used Open Recipes1 , a database of recipes, automatically collected by
crawler scripts.This dataset was enriched with cooking instructions by crawling the original websites of the recipes. After getting a collection of cleaned
recipes; ingredients, modifiers, units, utensils and events were extracted.
Cleaned ingredients were hard to obtain, because even though they are specifically listed per recipe normalizing them was difficult because ingredient
names are intertwined with amounts, units, and other noise. Events of cooking are simply verbs that take an ingredient as an argument. This requires previous POS-tagging and parsing of the cooking instructions. We noted that the
approach used by Kiddon et al. (2015) is more elaborate and we plan to follow
their heuristics for extracting events.
We then created a semantic vector space using co-occurrence counts (ingredients co-occurring with ingredients, modifiers, amounts, and units),
reweighted the resulting matrix with PPMI and reduced the dimensions to 100
using SVD. These enriched learned representations are used as the inputs to
our generator. We experimented with building our generator using different
neural models (char/word based LSTM language models, seq2seq) on the possible events given the ingredients.
We have thought of different ways of evaluating the quality of the automatically generated recipes, such as manual linguistic analysis of the output recipe,
blind taste tests, or Turing tests comparing “real” recipes with generated ones
and with competitor systems.
References: • Kiddon C., et al. (2015): Mise en Place: Unsupervised Interpretation of Instructional
Recipes. Proceedings of the 2015 EMNLP.
1
https://github.com/fictivekin/openrecipes
337
CL Poster
CL-Postersession
Donnerstag
09.03.2017
12:45 – 13:45
B4 1, Foyer
CL Poster
In support of self-assessment – exploiting available information
from tools
Markus Gärtner
Universität Stuttgart
Kerstin Eckart
Universität Stuttgart
[email protected]
[email protected]
Many automatic tools for natural language processing tend to produce less reliable results if presented with data which differs from an expected standard.
Users in search of the most suitable tool for their out-of-domain data have to
either perform some additional error detection after processing (cf. Dickinson,
2015, for an automatic approach) or invest significant manual effort into the
detailed evaluation of several tools.
Interestingly enough, automatic tools are often internally aware of a relative reliability of their output, since they make use of probabilities and forced
guessing to decide on a single analysis. Such information can be understood
as the internal confidence for the complete analysis, e.g. n-best lists of outputs
correspond to a ranking according to confidence.
We argue for an approach where tool output is transparent with respect to
its internal confidence estimation (e.g. BitPar; Schmid 2004). A single confidence value does not have to cover the whole analysis, but can refer to subparts,
such as a dependency relation or a specific label.
We suggest that tool output includes confidence values as an additional annotation layer which will be an advantage when handling large data sets that
rely on automatic annotation: Users can restrict their queries to more reliable
parts or find interesting cases by inspecting parts with low confidence. That
is, transparent confidence estimation will not increase the quality of the output as such, but its usability in that it helps users to assess if an analysis, or a
specific part of it, is sufficiently reliable.
Moreover, providing confidence values as annotation raises the awareness
with regard to reliability, which we think can foster the application of stateof-the-art tools on out-of-domain data also in related fields such as the Digital
Humanities.
References: • Dickinson, M. (2015): Detection of annotation errors in corpora. Language and Linguistics Compass 9(3), 119–138. • Schmid, H. (2004): Efficient parsing of highly ambiguous context-free
grammars with bit vectors. In: Proceedings of the 20th International Conference on Computational Linguistics (COLING 2004). Geneva, Switzerland.
338
CL-Postersession
Agile creation of multi-layer corpora with corpus-tools.org
Stephan Druskat
HU Berlin
Thomas Krause
HU Berlin
Carolin Odebrecht
HU Berlin
[email protected]
[email protected]
[email protected]
Donnerstag
09.03.2017
12:45 – 13:45
B4 1, Foyer
Agile corpus creation replaces the linear corpus creation process with iterative cycles of query, schema edits, annotation and analysis. We demonstrate
corpus-tools.org, a suite of generic tools tailored to the agile creation of multilayer corpora. It consists of Salt, a graph-based meta model and API for linguistic data; Pepper, a conversion platform; Atomic, an extensible annotation
software; ANNIS, a search and visualization architecture for multi-layer corpora. As of now, Atomic lacks search capabilities for agile workflows. ANNIS
provides a search system based on annotation graphs, and the ANNIS Query
Language (AQL). ANNIS, however, has been optimised for linear workflows,
which graphANNIS (https://git.io/vijrI), a new C++-based implementation, will change. It will also make ANNIS self-contained, dropping the dependency to a separate database installation. graphANNIS supports a large subset
of AQL, aligns its data representation more closely with the Salt model, and
provides a Java API. Its encapsulation allows for graphANNIS to be embedded
in Atomic, as its search engine. While Atomic will be responsible for storage
of corpus data, graphANNIS provides an additional index which is updated
whenever a document is changed. For search tasks, Atomic will provide a GUI
section for AQL queries. These will be parsed by the ANNIS AQL parser and
passed to the graphANNIS search system, which will return the Salt IDs of
the matched nodes, in turn used in Atomic to present the results. This setup
will provide corpus-tools.org with capabilities for agile multi-layer corpus creation.
References: • Druskat, S.; Krause, T.; Odebrecht, C. (preprint): Agile creation of multilayer corpora
with corpus-tools.org. https://doi.org/10.5281/zenodo.157166. • Voormann, H.; Gut, U. (2008):
Agile corpus creation. CLLT, 4(2), 235–251. • Zipser, F.; Romary, L. (2010): A model oriented approach
to the mapping of annotation formats using standards. In: Proceedings of LREC 2010, Valletta. • Zipser,
F.; Zeldes, A.; Ritz, J.; Romary, L.; Leser, U. (2011): Pepper: Handling a multiverse of formats. Poster,
DGfS 2011, Göttingen. • Druskat, S.; Bierkandt, L.; Gast, V.; Rzymski, C.; Zipser, F. (2014): Atomic: an
open-source software platform for multi-level corpus annotation. In: Proceedings of KONVENS 2014,
228–234. • Krause, T.; Zeldes, A. (2016): ANNIS3: A new architecture for generic corpus query and
visualization. DSH, 31(1), 118–139.
339
CL Poster
CL Poster
Linguistische Berichte
Ingo Reich, Augustin Speyer
(eds.)
Franz d’Avis,
Horst Lohnstein (Hg.)
Co- and sub­
ordination in
German and
other languages
Normalität
in der Sprache
Augustin Speyer, Ingo Reich: Introduction:
Alternations in co- and subordinations
Mailin Antomo: Marking (not-)at-issue content
by using verb order variation in German
Ermenegildo Bidese, Alessandra Tomaselli: The
decline of asymmetric word order in Cimbrian
subordination and the special case of umbrómm
Patrick Brandt, Beata Trawinski, Angelika Wöllstein:
(Anti-)Control in German: evidence from com­
parative, corpus- and psycholinguistic studies
Nicholas Catasso, Roland Hinterhölzl: On the question
of subordination or coordination in V2-relatives
in German
Christian Fortmann: Da capo je-desto – On the com­
parative conditional construction in German
Werner Frey: On some correlations between interpre­
tative and formal properties of causal clauses
Ulrike Freywald: Clause integration and verb position
in German
Robert Külpmann, Vilma Symanczyk Joppe: Argument
omission in imperative-declarative conjunctions
Rosemarie Lühr: Causal clauses in Old Indo-European
languages
Stefanie Pitsch: Syntax and semantics of causal
nachdem-clauses in German
Marga Reis: Consecutive so…V2-clauses in German
Sophie von Wietersheim: Variable binding as evidence
for clausal attachment
Linguistische Berichte, Sonderheft 21
Ca. 350 Seiten. 978-3-87548-746-6
Kartoniert, ca. Euro 68,00
(für LB-Abonnenten ca. Euro 58,00)
Lieferbar
buske.de
Franz d’Avis und Horst Lohnstein: Einleitung
Franz d’Avis: Konzessivität und Normalvorstellungen
Eva Breindl: Konnexion in lernersprachlichen und in
muttersprachlichen argumentativen deutschen
Texten
Frederike Eggs: Das personale Indefinitum man
Daniel Gutzmann und Katharina Turgay: Normal­
exklamationen – normal!
Holden Härtl: Normality at the boundary between
word-formation and syntax
Julia Kolkmann: What makes a default interpretation?
Considerations from English attributive posses­
sion
Horst Lohnstein: Normalität und sprachliche Form
Sonja Müller: Normalität im Diskurs – Implikations­
verstärkung in (halt eben-/eben halt-)Asserti­
onen
Sven Müller: Konzessivität und Normalvorstellungen
Olav Müller-Reichau: Normalitätseffekte in Existenz­
sätzen des Russischen
Linguistische Berichte, Sonderheft 22
Ca. 350 Seiten. 978-3-87548-789-3
Kartoniert, ca. Euro 68,00
(für LB-Abonnenten ca. Euro 58,00)
Lieferbar
Die Sonderhefte der Linguistischen Berichte
geben einen Überblick über die wesent­lichen
Probleme und Entwicklungen von Theoriebereichen
(z.B. Phonologie, Semantik, Dialektsyntax, Sprach­
erwerb, historische Linguistik und Pragmatik) und
informieren über den aktuellen Stand der Forschung.
Tutorium
der Sektion Computerlinguistik
CL Tutorium
Automatische Syntaxanalyse für diachrone und synchrone
Linguistik: Vorverarbeitung, Tagging, Parsing
Gerold Schneider, Kyoko Sugisaki
Universität Konstanz und Universität Zürich, Universität Zürich
[email protected]
Raum: A2 2, 2.14
Zeit: Dienstag, 07.03.2017, 10:00–17:00 Uhr
Short description
Nach einer kurzen Einführung in Parsingtheorie zeigen wir schrittweise, wie
man Texte vorverarbeitet, und syntaktisch analysiert, und geben einen Ausblick auf Anwendungen. Wir beschränken uns auf Englisch und Deutsch als
Objektsprachen.
Die Vorverarbeitung beinhaltet bei historischen Texten Standardisierung
der Schreibweise (Spelling Normalisation), wozu wir die Tools VARD fürs Englische und NORMA fürs Deutsch vorstellen. Der nächste Schritt, Wortartenzuweisung mit einem Tagger, verlangt als Eingageformat z.B. ein Wort pro
Zeile, so für den Tree-Tagger, den wir anwenden. Für die syntaktische Analyse
stellen wir mehrere Dependenzgrammatik-Parser vor und wenden sie an. Wir
diskutieren die noch junge Forschung zur Parseranwendung und -anpassung
auf historische Daten. Das Parsing kann auch durch den Einsatz eines Supertaggers verbessert werden, welcher nicht nur Wortarten, sondern auch
wahrscheinliche Dependenz-Markierungen vorschlägt.
345
CL Tutorium
Doktorandenforum
Organisation: Julia Schüler
Ort: Gebäude C9 3 (Jägerheim)
PhD forum
Introduction to statistics with R
Mindaugas Mozuraitis
Raum: C9 3 (Jägerheim)
Zeit: Dienstag, 07.03.2017, 09:00–17:30 Uhr
Short description
This workshop will introduce you to the basic statistical tools you will need for
the analysis of experimental data (e.g., behavioral experiments in psycholinguistics, perception studies in phonetics, acceptability rating studies in experimental linguistics). We will then cover the most common statistical methods
used in speech and language research such as
• different types of t-tests
• different types of chi-square tests
• correlation
The workshop will combine brief theoretical introductions from the instructor
with hands-on exercises using the free statistical analysis software R (cran.rproject.org). The focus will be on learning how to explore and organize data
for the analyses, run the analyses, interpret R output, and report the findings.
At the conclusion of this workshop, you will be able to:
• Describe the logic behind common statistical procedures.
• Select the appropriate statistical procedure given the type of data (e.g.,
continuous vs. categorical), distribution of the data (e.g, skewed vs.
symmetric), and experimental design (e.g., independent vs. repeatedmeasures).
349
PhD
Doktorandenforum
• Test whether your data complies with assumptions required for different statistical analyses through various data exploration techniques.
• Choose between most common transformations in cases where assumptions of parametric statistical test(s) is/are violated and some of the nonparametric alternatives that are available.
• Do basic manipulations and visualizations of data, and data analysis in
R.
Schedule
09:00–10:00
Session 1
10:00–10:30
Coffee break
10:30–11:30
Session 2
11:30–12:30
Lunch break
12:30–14:00
Session 3
14:00–14:30
Coffee break
14:30–15:30
Session 4
15:30–17:30
Individual consultations
PhD
350
Sprache – Politik – Gesellschaft
Eric Wallis
Kampagnensprache
Wie Greenpeace mit Sprachkritik den Umweltdiskurs
beeinflusst
2016. 442 Seiten, Hardcover, 978-3-944312-36-1, € 39,00. SPG Bd. 17
In der EU ist es seit 2004 verboten, genveränderte Nahrungsmittel
ohne entsprechende Produktkennzeichnungen zu verkaufen. Erlaubt blieb jedoch, genveränderte Pflanzen als Viehfutter zu nutzen,
ohne Endprodukte wie z.B. Milch zu deklarieren. Um darauf hinzuweisen, startete die Umweltorganisation Greenpeace eine Kampagne gegen den Milchhersteller Müller-Milch. Der Streit beider Akteure dauerte mehrere
Jahre. Anhand dieses Streits zeichnet Eric Wallis nach, wie sich gegnerische Sichtweisen
sprachlich verbreiten, denn Sprache ist standpunktgebunden. Während einer Kampagne
treffen die Standpunkte öffentlich aufeinander, sodass andere Akteure in den Streit hineingezogen werden. In dieser Dynamik sich aneinander orientierender Akteure verbreitet sich das Wissen, wie in diesem Fall die Idee, sich gegen die Milch von mit Genfutter
gefütterten Kühen zu entscheiden. So lässt sich Sprachwandel u. a. als ein Symptom der
Etablierung neuer und des Verschwindens alter Sichtweisen erklären.
Jana Reissen-Kosch
Identifikationsangebote der rechten Szene im Netz
Linguistische Analyse persuasiver Online-Kommunikation
2016. 212 Seiten, Hardcover, 978-3-944312-34-7, € 36,00. SPG Bd. 19
Eine Auseinandersetzung mit Rechtsextremismus im Alltag führt
nicht an den Rand, sondern in die Mitte der Gesellschaft: Rechts­
extreme Gruppierungen haben heute in den verschiedensten
Lebens­welten einen Platz gefunden und bieten unterschiedlichen
Zielgruppen Mitmach-Potenzial. Dieses manifestiert sich unter
anderem auf sprachlicher Ebene und wird auch über das Internet
beworben, weshalb sich die Frage stellt, wie rechtsextremistische
Internetauftritte aus linguistischer Perspektive auf ihre Zielgruppenorientierung hin untersucht werden können. Im Rahmen einer Pilotstudie ent­wickelt die Arbeit einen Ansatz,
der Grundannahmen der Politolinguistik mit einem marketingstrategischen ZielgruppenModell verknüpft und für die Analyse fruchtbar macht.
Als Ergebnis lassen sich verschiedene Werteprofile identifizieren, die durch die analysierten Websites vorrangig angesprochen werden und die wiederum mit gestalterischen
Merkmalen dieser Websites verknüpft werden können.
Dr. Ute Hempen Verlag, Clausewitzstr. 12, 28211 Bremen,
Tel. 0421/3479901, www.hempen-verlag.de, [email protected]
Infotag der Lehramtsinitiative der DGfS
Organisation: Stefanie Haberzettl, Magdalena Wojtecka & Lucia Hubig
Ort: Gebäude C5 1, C5 2, C5 3
Lehramtsinitiative
Neu zugewanderte Schülerinnen und Schüler sprachsensibel
unterrichten und fördern
Ansprechpartnerinnen:
Stefanie Haberzettl, Magdalena Wojtecka, Lucia Hubig
[email protected], [email protected],
[email protected]
Raum: Plenarvortrag: C5 1 (Musiksaal), Workshops: C5 2, C5 3
Zeit: Dienstag, 07.03.2017, 14:30–19:00 Uhr
Short description
Neu zugewanderte Schülerinnen und Schüler stellen eine große Herausforderung für die Lehrkräfte sowohl an Grundschulen als auch an weiterführenden Schulen dar. Diese Kinder und Jugendlichen haben im Unterricht
immer eine Doppelaufgabe zu leisten: Sie müssen nicht nur die neuen, zu
einem großen Teil sprachlich vermittelten Inhalte verarbeiten, sondern auch
die Sprache zu diesen Inhalten erst lernen. Um sie dabei gezielt und erfolgreich zu unterstützen, muss man sie als Lehrkraft sprachsensibel unterrichten
und fördern. Diese Vorgehensweise ist aber nur mit ausreichenden linguistischen Kenntnissen möglich. Der Infotag der Lehramtsinitiative im Rahmen
der Jahrestagung der Deutschen Gesellschaft für Sprachwissenschaft am 7.
März 2017 an der Universität des Saarlandes bietet daher Lehramtsstudierenden, ReferendarInnen und schon im Schuldienst befindlichen Lehrkräften
die Möglichkeit, sich mit ausgewählten Bereichen des sprachsensiblen Unterrichts auseinanderzusetzen.
Im Rahmen des Plenarvortrags, für den Gabriele Kniffka gewonnen werden konnte, werden die TeilnehmerInnen in die Thematik des sprachsensiblen
Vorgehens im Unterricht mit neu zugewanderten SchülerInnen eingeführt.
355
LAI
Lehramtsinitiative
Die anschließenden Workshops decken ein breites Spektrum des sprachsensiblen Unterrichts und der Sprachförderung für Seiteneinsteiger ab und sind
so ausgewählt, dass sowohl GrundschullehrerInnen als auch Lehrkräfte an
den weiterführenden Schulen angesprochen werden. Folgende Themenbereiche werden in acht Workshops (siehe auf den folgenden Seiten) sprachwissenschaftlich und sprachdidaktisch von den eingeladenen ExpertInnen mit
den TeilnehmerInnen erarbeitet:
• Berücksichtigung der Herkunftssprache und der Sprachbiographie
• Alphabetisierung
• Entwicklung eines Sprachförderkonzeptes
• Systematische DaZ-Förderung an der Grundschule
• Wortschatzförderung
• Sprachförderung in der naturwissenschaftlichen Fächern in der Sek 1
Plenarvortrag
Gabriele Kniffka (Freiburg)
Neu zugewanderte Schülerinnen und Schüler sprachsensibel unterrichten und
fördern
LAI
356
Lehramtsinitiative
Programm
14:30–15:00
Ankommen, Schulbuchausstellung im
Foyer
C5 1
(Musiksaal)
15:00–16:00
Plenarvortrag:
Gabriele Kniffka, PH Freiburg
Neu zugewanderte Schülerinnen und Schüler
sprachsensibel unterrichten und fördern
C5 1
(Musiksaal)
16:00–16:30
Kaffeepause
16:30–19:00
Erkan Gürsoy
Herkunftssprachen und Sprachbiographie
C5 3, 1.20
Giulio Pagonis
Systematische DaZ-Förderung in der
Grundschule
C5 3, 2.06
Katja Hirschmann
Alphabetisierung von neu zugewanderten
SchülerInnen
C5 3, 2.09
Anja Müller
Wortschatzförderung
C5 2, 1.12
Elisabeth Venohr
Sprachförderung von Seiteneinsteigern im
naturwissenschaftlichen Unterricht der Sek 1
C5 2, 1.28
Barbara Kiefer
Sprachbildender Fachunterricht:
Operatoren im Fokus
C5 3, 3.24
Roland Nenno & Sandra Steinmetz
Fachtexte vereinfachen – eine sprachfördernde
Maßnahme?
C5 2, 1.08
wird noch bekanntgegeben
Sprachförderkonzepte an Regelschulen
C5 3, 4.25
LAI
357
Lehramtsinitiative
Neu zugewanderte Schülerinnen und Schüler sprachsensibel
unterrichten und fördern
Dienstag
07.03.2017
15:00 – 16:00
C5 1
(Musiksaal)
Gabriele Kniffka
PH Freiburg
Die Integration neu zugewanderter Schülerinnen und Schüler in das deutsche
Schulsystem stellt alle an diesem Prozess Beteiligten vor große Herausforderungen unterschiedlicher Art. Eine der größten Herausforderungen
bildet die sprachliche Vorbereitung auf den Unterricht: Eine erfolgreiche Partizipation am Unterricht erfordert sprachlicher Kompetenzen in bildungsbzw. fachsprachlichen Varietäten des Deutschen. Diese müssen von Anfang
an vermittelt und im Regelunterricht weiter ausgebaut werden: Zur Vorbereitung auf die Regelklasse bedarf es daher eines fachsensiblen Sprachunterrichts, nach dem Wechsel in die Regelklasse eines sprachsensiblen Fachunterrichts. Im Vortrag werden Möglichkeiten der Sprachbildung in Vorbereitungsklassen und im Regelunterricht aufgezeigt.
Dienstag
07.03.2017
16:30 – 19:00
C5 3, 1.20
Herkunftssprachen und Sprachbiographie
Erkan Gürsoy
Universität Essen-Duisburg, ProDaZ
LAI
Im Workshop werden Erhebungsverfahren vorgestellt, die die Herkunftssprachen und Sprachbiographien von mehrsprachigen Kindern
und Jugendlichen erfassen sollen. Im Modellprojekt ProDaZ (www.unidue.de/prodaz) und weiteren Projekten wurden diese Erhebungsverfahren
empirisch erprobt und in Theorie-Praxis-Projekten mit Schulen eingesetzt.
Die Ergebnisse zeigen, welche Grundlagen zur differenzierten Erfassung von
Mehrsprachigkeit beachtet werden müssen und worin auch Grenzen liegen.
Inwiefern solche Verfahren für neu zugewanderte Schülerinnen und Schüler
möglich sind, wird ebenfalls thematisiert.
358
Lehramtsinitiative
Systematische DaZ-Förderung in der Grundschule
Dienstag
07.03.2017
16:30 – 19:00
C5 3, 2.06
Giulio Pagonis
Universität Heidelberg
Eine Annahme der Sprachlern- und -lehrforschung lautet, dass Sprachvermittlung bei Kindern (z.B. die Vermittlung von Wortschatz, grammatischen Strukturen oder Textkompetenz) umso besser gelingt, je stärker dabei
Prozesse und Strategien des natürlichen kindlichen Spracherwerbs berücksichtigt werden.
In dem Workshop werden deshalb zunächst Charakteristika einer systematischen Sprachvermittlung erarbeitet. Für ausgewählte Bereiche des
Deutschen als Zweitsprache wird dabei aufgezeigt, wie sprachliche Systeme
natürlicherweise von Kindern erworben werden. Anschließend werden in
einem praxisorientierten Teil ausgehend von konkreten Materialien didaktische Vorgehensweisen diskutiert, mit denen Erwerbsprozesse bei Kindern im
Primarbereich gezielt unterstützt werden können.
Alphabetisierung von neu zugewanderten SchülerInnen
Dienstag
07.03.2017
16:30 – 19:00
C5 3, 2.09
Katja Hirschmann
HTW Saarbrücken
Ziel des Workshops ist es, verschiedener Methoden im Alphabetisierungsunterricht zu erarbeiten. Die TeilnehmerInnen lernen mithilfe abwechslungsreicher Materialien und gängiger Lehrwerke die praxisnahe Anwendung der Alphabetisierungsmethoden in der Arbeit mit neu zugewanderten
Schülern und Schülerinnen kennen.
LAI
359
Lehramtsinitiative
Wortschatzförderung
Dienstag
07.03.2017
16:30 – 19:00
C5 2, 1.12
Anja Müller
Universität Mainz
Zwei Themen bilden die Schwerpunkte des Workshops: Zu Beginn soll ausgehend von sprachwissenschaftlichen und spracherwerbstheoretischen Erkenntnissen geklärt werden, was Wortschatz ist und wie Wortschatz aufgebaut wird. Darauf basierend werden im zweiten Teil des Workshops Bereiche/Ebenen und Methoden der Wortschatzförderung abgeleitet und Ideen
für den Unterricht entwickelt und diskutiert.
Dienstag
07.03.2017
16:30 – 19:00
C5 2, 1.28
Sprachförderung von Seiteneinsteigern im naturwissenschaftlichen
Unterricht der Sek 1
Elisabeth Venohr
Saarbrücken / Katowice
LAI
Kinder und Jugendliche mit Migrationshintergrund haben einen spezifischen Sprachförderbedarf, der stark von der jeweiligen Erwerbsbiografie abhängig ist (Alter der Einreise, Herkunftssprache, familiäre Rahmenbedingungen usw.). Seiteneinsteiger bilden hier eine besondere Zielgruppe, da bereits eine schulische Sozialisation „mitgebracht“ wird (hier: schriftsprachliche
Kompetenz in der Muttersprache).
Da Fachinhalte immer auch sprachlich vermittelt werden, geht es zunächst
um die Spezifik der schulischen Fachsprache bzw. Bildungssprache, insbesondere um Aufgabenstellungen und deren Versprachlichung, die den Zugang
zu inhaltlichem Wissen erschweren können. In einem zweiten Schritt sollen
die Prinzipien des sprachsensiblen Unterrichts und die „verdeckten Sprachschwierigkeiten“ an Beispieltexten (Schulbuchtexte, Lernertexte usw.) illustriert und in praktischen Übungen auf die Lernziele in den MINT-Fächern in
der Sekundarstufe 1 umgesetzt werden.
360
Lehramtsinitiative
Ziele des Workshops: typische Lernerfehler und Indikatoren für Spracherwerbsstufe in Deutsch als Zweitsprache identifizieren können, Operatoren
in den MINT-Fächern auf ihren fremd-/fachsprachlichen Schwierigkeitsgrad
hin beurteilen, Verständlichkeit von Aufgabenstellungen optimieren, Verzahnung von Regel- und Sprachunterricht erkennen und gestalten.
Sprachbildender Fachunterricht: Operatoren im Fokus
Dienstag
07.03.2017
16:30 – 19:00
C5 3, 3.24
Barbara Kiefer
LPM Saarbrücken
Warum ist Sprache für das fachliche Lernen wichtig? Fachlernen und Sprachlernen sind untrennbar miteinander verbunden, daher muss der Gebrauch
der Bildungssprache von allen Schüler/innen erlernt und trainiert werden. Im
Workshop werden Grundlagen und Methoden des sprachsensiblen Fachunterrichts aufgezeigt und am Beispiel der Operatoren konkretisiert.
Dienstag
07.03.2017
16:30 – 19:00
C5 2, 1.08
Fachtexte vereinfachen – eine sprachfördernde Maßnahme?
Roland Nenno
Universität des Saarlandes
Sandra Steinmetz
Universität des Saarlandes
Wir sehen uns Merkmale von Fachtexten an und analysieren 1 oder
2 Textbeispiele aus Lehrwerken im Hinblick auf diese Merkmale. In
einem Folgeschritt vereinfachen wir die Texte exemplarisch (nach versch.
Niveaustufen). Der Workshop richtet sich hauptsächlich an LehrerInnen der
Sek 1 und Sek 2.
LAI
361
Lehramtsinitiative
Sprachförderkonzepte an Regelschulen
Dienstag
07.03.2017
16:30 – 19:00
C5 3, 4.25
wird noch bekanntgegeben
Ein Sprachförderkonzept kann nur Schritt für Schritt entwickelt und aufgebaut werden. Ziel des Workshops ist es anhand eines erfolgreichen Schulbeispiels zu zeigen, wie allgemeine Sprachbildung und gezielte Sprachförderung in den Schulalltag integriert werden können. Dabei werden erprobte Maßnahmen und konkrete didaktische Konzepte vorgestellt. Zusätzlich werden auch zahlreiche Tipps und Anregungen gegeben, die bei der Entwicklung eins Sprachförderkonzepts zu berücksichtigen sind.
362
Tagung der
Arbeitsgemeinschaft Linguistische Pragmatik
(ALP e.V.)
ALP
ALP
An den Grenzen der Pragmatik
Organisatoren:
Bettina Bock, Philipp Dreesen, Konstanze Marx, Simon Meier & Robert
Mroczynski
Raum: B3 2, 0.03
Zeit: Dienstag, 07.03.2017, 09:00–18:00 Uhr
Short description
Als linguistische Teildisziplin ist die Pragmatik längst vollständig etabliert
und wirkt ihrerseits im Fach integrierend. Entsprechend kann sich die Pragmatik nicht mehr als revolutionärer Gegenentwurf zu einer sich einseitig
entwickelnden Systemlinguistik verstehen. Auch aktuelle Grammatiken und
Wörterbücher ordnen ihre Belege mittlerweile pragmatisch ein, und neuere
Forschungsrichtungen wie die Grammatikalisierungstheorie, die Konstruktionsgrammatik oder die Korpuslinguistik, die gerade die Interdependenzen
von Sprachsystem und Sprachgebrauch betonen, fügen sich ohnehin nicht
einer allzu strikten Zweiteilung des linguistischen Gegenstandsbereichs.
Ein Aspekt dieser Entwicklung ist, dass – wenigstens im deutschen
Sprachraum – die klassischen gegenstandsbezogenen Bestimmungen der
Pragmatik als synchron angewandte Theorie der Deixis, Sprechakte, Implikaturen usw. in den Hintergrund rücken zugunsten eines Verständnisses von Pragmatik als einer funktionalen Perspektive auf Sprache und
Sprachgebrauch, die quer zu den üblichen Teildisziplinen der Linguistik steht.
Strikte Eingrenzungen des Gegenstandsbereichs der Pragmatik, aber auch die
üblichen Gründungsmythen etwa um die Entdeckungen der Ordinary Language Philosophy verlieren so an Plausibilität.
Dieser Befund einer gewissen Entgrenzung der linguistischen Pragmatik
(die auch mit Identitätsfragen einhergeht) soll Anlass sein, auf der kom-
365
ALP-Tagung
ALP
menden ALP-Tagung gerade nach den Grenzen der Pragmatik zu fragen.
Lassen sich überhaupt Grenzen ziehen, wo wären sie zu ziehen und wie ließe
sich das begründen? Dabei wollen wir u. a. folgende Achsen unterscheiden, an
deren Enden solche Grenzen zu erwarten sind:
1. Fachgeschichtlich: Wie weit lassen sich über die üblichen Gründungsdaten der Pragmatik hinaus in der Wissenschaftsgeschichte Konzepte
und Theorien zurückverfolgen, die als (proto-)pragmatisch bezeichnet werden können (vgl. Nerlich/Clarke 1995, Meier 2016)? Kann man
gegenwärtig bereits von einer postpragmatischen Phase sprechen (vgl.
Dreesen 2015: 61), etwa angesichts diskurslinguistischer Kritik (vgl.
etwa Spitzmüller/Warnke 2011: 48, 51) an für die Pragmatik so wichtigen Begriffen wie Intention, Gemeintes und Außersprachliches?
2. Gegenstandsbezogen: Was sind aus welchen Gründen genuin pragmatische Gegenstände und wo hört die Zuständigkeit der Pragmatik sinnvollerweise auf? Wird durch den Verzicht auf klare Gegenstandsbestimmungen nicht gleichsam unter der Hand und in anderem Gewand
der Vorwurf des „pragmatic wastebasket“ (Bar-Hillel 1971) wieder aktuell, wenn alle nicht explizit an semantischen oder grammatischen
Fragestellungen interessierten (interdisziplinären) Forschungsrichtungen kurzerhand der Pragmatik zugerechnet werden? Welche möglicherweise ungewöhnlichen Gegenstände werden derzeit aus pragmalinguistischer Sicht erforscht?
3. Methodisch: Gibt es eine konsensfähige spezifisch pragmatische
Methodik und Methodologie? Wie ließen sich ihre Spezifika bestimmen, etwa angesichts der Tatsache, dass auch die Grammatikforschung
längst korpusbasiert mit Sprachgebrauchsdaten arbeitet? Bedarf es
innerhalb der Pragmatik einer Neubestimmung ihrer theoretischmethodischen Prämissen?
Diese Fragen sollen auf der kommenden ALP-Tagung diskutiert werden.
Besonders willkommen sind Vorträge mit konzeptioneller Orientierung,
wobei die zu entwickelnden Bestimmungen über die Grenzen der Pragmatik
selbstverständlich anhand von empirischen Fallstudien veranschaulicht werden können. Wir laden alle Vortragenden ein, zwei bis drei Thesen in der Form
„Pragmatik ist (nicht) …“ zu formulieren und zur Diskussion zu stellen.
Keynote Speaker
Helmuth Feilke (Gießen)
366
ALP-Tagung
ALP
programmübersicht
08:45–09:00
Begrüßung und Einführung
09:00–09:45
Helmuth Feilke (Gießen)
“Im nächsten Moment” – Bühlers ‘Radieschen’ und die
pragmatische Scheidung möglicher Welten in der Sprache
09:45–10:15
Rita Finkbeiner (Mainz)
Grammatik vs. Pragmatik: Eine Programmatik
10:15–10:45
Tilo Weber (Halle)
Von der linguistischen Pragmatik zu einer pragmatisch
orientierten Linguistik
10:45–11:15
Kaffeepause
11:15–11:45
Daniel Schmidt-Brücken (Bremen)
Selbstwidersprüche zwischen Pragmatik, Semantik und
Grammatik
11:45–12:15
Wolf-Andreas Liebert und Pamela Steen (Koblenz)
Grenzen der Pragmatik am Beispiel der
Mensch-Tier-Kommunikation und des religiösen Sprechens
12:15–13:45
Mittagspause
13:45–14:15
Konstanze Marx (Mannheim)
Ist die Internetlinguistik eine pragmatische Disziplin? Eine
Problematisierung
14:15–14:45
Andreas Rothenhöfer (Bremen)
Von der Protopragmatik der Emotion. Erklärungsansätze
zwischen diskurslinguistischem Zentralmodell und
Wastebasket im Wastebasket
14:45–15:15
Frank Liedtke und Lena Rosenbaum (Leipzig)
Interjektionen als Randphänomen
15:15–15:45
Kaffeepause
15:45–16:15
Marie-Luis Merten (Paderborn)
Pragmatisch geprägt: Konstruktionen-in-Praktiken-Analyse
367
ALP-Tagung
ALP
16:15–16:45
François Nemo (Orléans)
Prosodic marking of contributional orientation in utterances:
Reconsidering the linguistic/pragmatic interface
16:45–17:00
Kurze Pause
17:00–17:30
Beatrix Weber (Dresden)
Zur Verwendung des Begriffs ‘Pragmatik’
17:30–18:00
Katharina Böhnert (Aachen) und Ilka Lemke (Bochum)
Pragmatik in der Schule – Plädoyer zur Nutzung des
didaktischen Potenzials pragmatischer
Untersuchungsgegenstände
18:00–18:15
Abschlussdiskussion
18:30
Mitgliederversammlung der Arbeitsgemeinschaft
Linguistische Pragmatik e. V.
Posterpräsentationen
Khrystyna Dyakiv (Lwiw, Ukraine)
Deviationen in Videointerviews. Pragmalinguistische Aspekte
Matthias Knopp (Köln)
Bedeutungspotenzial und aktuelle Referenz: Das mentale Lexikon beim medial
schriftlichen Sprachgebrauch
368
DFG-Informationsveranstaltung
DFG
DFG Info
Info
Talk and Question Time:
DFG funding opportunities for Linguists
Helga Weyerts-Schweda (DFG Bonn) &
Ulrike Demske (Spokesperson of the Review Board on linguistics
[Fachkollegium 104])
Raum: B3 1, 0.11
Zeit: Donnerstag, 09. 03. 2017, 12:45–13:45 Uhr
Short description
How to apply for DFG funding? What characterises a good DFG proposal? How
does the review process work and what are the chances of getting funded? Why
do applications fail and which mistakes can be avoided? These and other questions will be addressed from the perspective of the DFG and the Review Board
on linguistics.
For further questions individual contact time with Helga Weyerts-Schweda
can be arranged, either on Thursday, 9 March from 02:00–06:00 p.m. or on
Friday, 10 March from 11:30 a.m.–12:30 p.m.
371
DFG Info
Texte und Materialien
zur Sprachwissenschaft
978-3-15-019348-8 · € 13,80
Paperback · Format 15 x 21,5 cm
978-3-15-011056-0 · € 16,95
978-3-15-018807-1 · € 6,80
978-3-15-006922-6 · € 6,80
www.reclam.de
978-3-15-018241-3 · € 14,80
Reclam
notizen
.
374
.
375
.
376
.
377
.
378
379
Zeitschrift für Wortbildung /
Journal of Word Formation
Z
eitschrift für Wortbildung / Journal of Word Formation
(ZWJW) is an open access and double-blind peer reviewed
international journal published by Peter Lang. ZWJW publishes papers with respect to any language and linguistic
field, e.g. morphology, syntax, lexicology, phonology, semantics, pragmatics, language history, typology, dialectology, language acquisition, language contact.
T
he journal is published online, two volumes are released
each year. It contains original papers, reviews and general
information such as announcements of conferences, meetings,
workshops, etc. Special issues devoted to important topics in
word formation will occasionally be published. Manuscripts
are accepted in English, German, French and Spanish; every manuscript must have an English
abstract of max. 350 characters (including spaces).
Please send your manuscript to the editor-in-chief:
Prof. Dr. Petra M. Vogel
E-mail: [email protected]
Website: https://www.uni-siegen.de/phil/germanistik/mitarbeiter/vogel_petra_m/
Editorship:
Advisory Board:
Petra M. Vogel (University of Siegen)
Elke Donalies (Institut für Deutsche Sprache,
Mannheim)
Ludwig Eichinger (Institut für Deutsche
Sprache, Mannheim)
Mechthild Habermann (Erlangen-Nürnberg
University)
Jörg Meibauer (Mainz University)
Barbara Schlücker (University of Bonn)
Hans-Joachim Solms (Halle-Wittenberg
University)
Pavol Štekauer (Pavol Jozef Šafárik University
Košice)
Salvador Valera Hernández (University
of Granada)
Werner Abraham (University of Vienna
& Munich University)
Aleksandra Bagasheva (Sofia University)
Irmhild Barz (University of Leipzig)
Geert Booij (University of Leiden)
Jan Čermák (Charles University Prague)
Gerrit J. Dimmendaal (University of Cologne)
Jan Don (University of Amsterdam)
Nanna Fuhrhop (University of Oldenburg)
Livio Gaeta (University of Turin)
Luise Kempf (Mainz University)
Lívia Körtvélyessy (Pavol Jozef Šafárik
University Košice)
Elisabeth Leiss (Munich University)
Hans Christian Luschützky (University of Vienna)
Francesca Masini (University of Bologna)
Franz Rainer (Vienna University of Economics
and Business)
Anna Thornton (University of L’Aquila)
Carola Trips (University of Mannheim)
Personenverzeichnis
Ackermann, Tanja, 292
Adamczyk, Elżbieta, 289
Adli, Aria, 301
Agarwal, Sumeet, 98
Aguilar Guevara, Ana, 135
Alexiadou, Artemis, 237
Altshuler, Daniel, 109
Amaral, Patrícia, 136
Amsili, Pascal, 97
Anastasiou, Dimitra, 336
Antomo, Mailin, 221
Anttila, Arto, 194
Asher, Nicholas, 109, 271
Bögel, Tina, 189
Baayen, R. Harald, 211
Babonnaud, William, 277
Bacskai-Atkari, Julia, 231
Bade, Nadine, 140, 141
Bader, Markus, 124
Balo, Anne-Kathrin, 5
Barnickel, Katja, 230
Baumann, Stefan, 125
Bayer, Josef, 130
Becker, Laura, 158
Bell, Melanie J., 212
Bergmann, Pia, 285
Best, Catherine T., 94
Beutler, Janina, 228
Bildhauer, Felix, 306, 326
Blümel, Andreas, 272
Blumenthal-Dramé, Alice, 160
Bochnak, Ryan, 240
Bock, Bettina, 365
Bogal-Allbritten, Elizabeth, 255
Borgonovo, Claudia, 133
Bostan, Laura, 330, 337
Boye, Kasper, 264
Bross, Fabian, 247
Brunetti, Lisa, 107
Bruni, Jagoda, 323
Bunk, Oliver, 235
Butt, Miriam, 203
Bülow, Lars, 286
Cap, Fabienne, 217
Carlson, Katy, 188
Carlucci-Dirani, Seyna, 294
Castroviejo, Elena, 131
Cavalcante, Federico Amorim, 122
Chiarcos, Christian, 324, 335
Chiriacescu, Sofiana, 119
Clodo, Christoph, 5
Cole, Jennifer, 125
Colesnicov, Alexandru, 151
Coniglio, Marco, 115
Crocker, Matthew, 83, 99
Csipak, Eva, 243, 273
Dammel, Antje, 281
Dannenberg, Anna, 190
De Kuthy, Kordula, 107
de la Fuente, Israel, 106
De Sutter, Gert, 304
Degaetano-Ortlieb, Stefania, 91
Deguchi, Masanori, 274
Dekker, Paul, 266
del Carmen Horno, Maria, 279
385
Index
Personenverzeichnis
Index
Delogu, Francesca, 99
Demberg, Vera, 5, 89, 91, 96, 120,
327, 331
Demske, Ulrike, 229, 371
Deo, Ashwini, 102
DeVeaugh-Geiss, Joseph P., 140
Dima, Corina, 219
Dobrovie-Sorin, Carmen, 270
Dogil, Grzegorz, 323
Donazzan, Marta, 270
Donhauser, Karin, 312
Dorna, Michael, 210
Dreesen, Philipp, 365
Driemel, Imke, 224
Druskat, Stephan, 339
Duran, Daniel, 323
Döhmer, Caroline, 297
Döring, Sophia, 110
Dörre, Laura, 130
Dücker, Lisa, 148
Frey, Werner, 234
Eckart, Kerstin, 145, 338
Egg, Markus, 310
Ehrmantraut, Luise, 5
Ellsäßer, Sophie, 291
Enger, Hans-Olav, 282
Enke, Dankmar, 290
Erteschik-Shir, Nomi, 202
Escandell-Vidal, Victoria, 268
Evers-Vermeul, Jacqueline, 108
Evert, Stefan, 304
Haberzettl, Stefanie, 5, 355
Haegeman, Liliane, 112, 236
Haider, Thomas, 307
Hartmann, Stefan, 148
Hartung, Nele, 5
Haselow, Alexander, 143
Haspelmath, Martin, 156, 177
Hasse, Anja, 283
Haug, Dag Trygve Truslew, 109
Heim, Johannes, 197
Hemforth, Barbara, 106
Henderson, Robert, 141
Hennecke, Inga, 215
Hennig, Mathilde, 146
Herment, Sophie, 192
Hinrichs, Erhard, 219
Hinterhölzl, Roland, 115
Hinterwimmer, Stefan, 128
Hirschmann, Katja, 359
Fazleh Elahi, Mohammad, 336
Felser, Claudia, 117
Fenk-Oczlon , Gertraud, 157
Filip, Hana, 265, 271
Finkbeiner, Rita, 225
Fischer, Hanna, 287
Fischer, Stefan, 91
Freiberg, Cassandra, 114
386
Gärtner, Markus, 338
Gaeta, Livio, 161
Gehrke, Berit, 131
Geibel-Stutz, Ellen, 5
Gergel, Remus, 5
Gluting, Peter, 5
Goldschmidt, Katrin, 152
Goldsmith, John, 81
Grafmiller, Jason, 303
Grant, Margaret, 267
Greco, Ciro, 112, 236
Grubic, Mira, 138
Grumt Suárez, Holger, 332
Gumiel-Molina, Silvia, 269
Gutzmann, Daniel, 127
Guzmán Naranjo, Matías, 158
Gürel, Ayşe, 209
Gürsoy, Erkan, 358
Personenverzeichnis
Hoek, Jet, 108
Hoekstra, Eric, 289
Hohaus, Vera, 245
Hole, Daniel, 247
Holler, Anke, 106, 208
Hong, Xudong, 331
Horch, Eva, 5, 104
Howcroft, David M., 105
Hubig, Lucia, 5, 355
Härtl, Holden, 129
Hätty, Anna, 210
Igoa, José Manuel, 279
Immesberger, Natascha, 5
Inaba, Jiro, 201
Ionov, Max, 324
Iordăchioaia, Gianina, 214
Jabeen, Farhat, 203
Jagfeld, Glorianna, 214
Jain, Ayush, 98
Jamieson, Elyse, 196
Janssen, Maarten, 149
Jezek, Elisabetta, 278
Ježek, Elisabetta, 216
Johnson, Cynthia A., 105
Josefsson, Gunlög, 202
Kaiser, Elsi, 139
Kaiser, Sebastian, 223
Kallmeyer, Laura, 265, 277
Karlova-Bourbonus, Natali, 332
Kasper, Simon, 165
Kastner, Itamar, 263
Kaufmann, Göz, 296
Kentner, Gerrit, 187
Kern, Friederike, 308
Khan, Geoffrey, 164
Kiefer, Barbara, 361
Kindermann, Dirk, 134
Kniffka, Gabriele, 358
Kolbe-Hanna, Daniela, 193
Korth, Manuela, 200
Kortmann, Bernd, 160
Kramp-Karrenbauer, Annegret,
47
Kratochvílová, Dana, 248
Krause, Thomas, 339
Kravtchenko, Ekaterina, 96
Kremers, Joost, 187
Krifka, Manfred, 77
Krimou, Fanny, 204
Kroch, Anthony, 78
Kulakov, Sergey, 5
Kunz, Kerstin, 305
Köhnlein, Björn, 202
Lühr, Rosemarie, 113
Lameli, Alfred, 284
Lapshinova-Koltunski, Ekaterina,
305
Lebani, Gianluca E., 218
Lehmann, Nico, 302
Lemke, Robin, 5, 104
Lenci, Alessandro, 218
Lensink, Saskia E., 211
Leonarduzzi, Laetitia, 192
Leser-Cronau, Stephanie, 295
Lestrade, Sander, 167
Levshina, Natalia, 92, 169
Li, Michelle, 257
Libben, Gary, 208
Lind Sørensen, Marie-Luise, 264
Linzen, Tal, 101
Lobin, Henning, 332
Lohndal, Terje, 237
Loporcaro, Michele, 299
Lucente, Luciana, 191
Lüdeling, Anke, 301, 325
387
Index
Personenverzeichnis
Index
Müller, Gereon, 233
Mühlenbernd, Roland, 290
Ma, Jianqiang, 219
Maché, Jakob, 244
Malahov, Ludmila, 151
Mandelkern, Matthew, 242
Manzoni, Judith, 193
Marelli, Marco, 220
Mari, Alda, 256
Martinez, José, 305
Marushak, Adam, 251
Marx, Konstanze, 365
Matthewson, Lisa, 246
Mayer, Denise, 5, 11
McCready, Eric, 141
Meier, Simon, 365
Meijer, A. Marlijn, 258
Meinhardt, Eric, 100
Menzel, Katrin, 305
Meyer, Roland, 316
Micheli, Maria Silvia, 216
Michniewicz, Sonia, 267
Miestamo, Matti, 170
Mitkovska, Liljana, 259
Mittmann, Maryualê Malvessi,
122
Modi, Ashutosh, 96
Mondorf, Britta, 175
Moreno-Quibén, Norberto, 269
Moser, Ann-Marie, 298
Mozuraitis, Mindaugas, 349
Mroczynski, Robert, 365
Mucha, Anne, 240
Möbius, Bernd, 89
Müller, Anja, 360
Müller, Sonja, 221, 227
Mărănduc, Cătălina, 151
Nemo, François, 204
Nenno, Roland, 361
388
Neumann, Stella, 304
Niebuhr, Astrid, 293
Nixon, Jessie S., 94
Oberländer, Jonathan, 330, 337
Odebrecht, Carolin, 145, 325, 339
Osswald, Rainer, 277
Özge, Umut, 121
Pagonis, Giulio, 359
Palmer, Alexis, 307
Particke, Hans-Joachim, 272
Pat-El, Na’ama, 172
Patterson, Clare, 117
Perez, Cenel-Augusto, 151
Perlitz, Laura, 325
Petkova-Kessanlis, Mikaela, 144
Petrova, Svetlana, 147
Pezzelle, Sandro, 216
Phillips, Jonathan, 242
Piñango, Maria M., 102
Pijpops, Dirk, 173
Pinkal, Manfred, 96
Plate, Ralf, 335
Popp, Marie-Luise, 253, 254
Portele, Yvonne, 124
Poschmann, Claudia, 116
Puddu, Nicoletta, 153
Pérez-Jiménez, Isabel, 269
Raber, Helena, 5
Rajkumar, Rajakrishnan, 98
Raso, Tommaso, 122
Rauth, Philipp, 4, 5
Rawlins, Kyle, 103
Rehbein, Ines, 314
Reich, Ingo, 5, 11, 104
Reinhardt, Janina , 205
Renans, Agata, 140
Rentzsch, Julian, 259
Personenverzeichnis
Rett, Jessica, 267
Riester, Arndt, 107
Rohde, Hannah, 118
Rubinstein, Aynat, 241
Rößler, Stefanie, 208
Sanders, Ted J.M., 108
Sanfelici, Emanuela, 142
Santorini, Beatrice, 78
Sawada, Osamu, 132
Sayeed, Asad, 5, 327, 331
Schallert, Oliver, 281
Schang, Emmanuel, 204
Scheffler, Tatjana, 315
Scheutz, Hannes, 286
Schmidt, Jessica, 5
Schmidtke-Bode, Karsten, 178
Schneider, Gerold, 345
Schneider, Ulrike, 175
Schnelle, Gohar, 150, 312, 325, 334
Scholman, Merel, 91, 120
Schu, Josef, 5
Schulte im Walde, Sabine, 207
Schulz, Petra, 142
Schwabe, Kerstin, 262
Schäfer, Lisa, 5
Schäfer, Martin, 212
Schäfer, Roland, 306, 326
Schüler, Julia, 5, 347
Schütze, Hinrich, 90
Seminck, Olga, 97
Senaldi, Marco S. G., 218
Serdobolskaya, Natalia, 261
Seržant, Ilja A., 179
Shi, Hui, 336
Shkadzko, Pavel, 327
Sims-Williams, Helen, 181
Singh, Vishal, 98
Skopeteas, Stavros, 195
Smolka, Eva, 207
Sohl, Jessica, 328
Speyer, Augustin, 5, 11
Stahnke, Johanna, 309
Staraschek, Nathalie, 226
Stark, Julia, 5
Stede, Manfred, 111
Steiner, Erich, 305
Steiner, Petra, 333
Steinmetz, Sandra, 361
Stiebels, Barbara, 253, 254
Strötgen, Jannik, 329
Strütjen, Kim, 276
Struckmeier, Volker, 199, 223
Suckow, Katja, 106
Sugisaki, Kyoko, 345
Sukhareva, Maria, 335
Suni, Antti, 190
Sutton, Peter, 271
Szczepaniak, Renata , 148
Szmrecsanyi, Benedikt, 313
Szucsich, Luka, 316
Index
Teich, Elke, 5, 89, 91
Thiel, Alexander, 142
Titov, Ivan, 96
Tokizaki, Hisao, 201
Tomaschek, Fabian, 93
Tonhauser, Judith, 137
Tourtouri, Elli N., 99
Trabandt, Corinna, 142
Trouvain, Jürgen, 308
Truckenbrodt, Hubert, 222, 246
Tsouni, Vicky, 276
Tucker, Benjamin V., 93
Turgay, Katharina, 127
Turnbull, Rory, 105
Uygun, Serkan, 209
Vainio, Martti, 190
389
Personenverzeichnis
Index
van Bergen, Geertje, 95
Van de Velde, Freek, 173
van de Vijver, Ruben, 276
van der Plas, Lonneke, 214
van Lier, Eva, 182
van Rijn, Marlou, 182
Vander Klok, Jozina, 245
Venohr, Elisabeth, 360
Verhoeven, Elisabeth, 302
Versloot, Arjen, 289
Voigtmann, Sophia, 5
von Heusinger, Klaus, 121
von Prince, Kilu, 240
Wallner, Dominik, 286
Waltereit, Richard, 311
Watkins, Jonathan, 5
Wechsler, Stephen, 260
Weiß, Helmut, 294
Weiß, Zarah, 150, 334
Werkmann Horvat, Ana, 249
Werner, Stefan, 190
Werth, Alexander, 284
Weskott, Thomas, 208
Westera, Matthijs, 123
Weyerts-Schweda, Helga, 371
White, Aaron Steven, 103
Wierzba, Marta, 198
Wiese, Heike, 232
Wojtecka, Magdalena, 5, 355
Wolf, Lavi, 250
Ye, Jingting, 184
Zaefferer, Dietmar, 251
Zaika, Natalia M., 185
Zhang, Muye, 102
Zimmer, Christian, 288
Zinsmeister, Heike, 328
Zobel, Sarah, 273
390
Gesamtübersicht der Arbeitsgruppensitzungen
B3 2
0.03
Teich et
al.
Schütze
Schütze
DegaetanoOrtlieb
et al.
Raum
13:45–
14:15
14:15–
14:45
14:45–
15:15
15:15–
15:45
Levshina
Tomaschek
& Tucker
Nixon &
Best
16:30–
17:00
17:00–
17:30
17:30–
18:00
15:45–
16:30
1
AG
Stede
Döring
Altshuler
& Haug
Asher
Asher
Hoek et
al.
Riester
et al.
B3 1
0.14
2
Mittwoch, 08. 03. 2017
Borgonovo
Sawada
Miró &
Gehrke
Dörre &
Bayer
Härtl
Hinterwimmer
Gutzmann
& Turgay
B3 1
0.13
3
–
–
–
–
–
–
–
B3 1
0.12
4
Bögel
Carlson
Kentner
&
Kremers
B4 1
0.25
6
Kasper
Khan
Gaeta
Uygun &
Gürel
Rößler et
al.
Libben
Libben
B4 1
0.24
7
Struckmeier &
Kaiser
Truckenbrodt
Truckenbrodt
Antomo
& Müller
B4 1
0.23
8
KolbeHanna &
Manzoni
Leonarduzzi &
Herment
Lucente
Bell
&
Schäfer
Lensink
& Baayen
Hätty &
Dorna
Staratschek
Finkbeiner
Driemel
Kaffeepause & Postersession
Blumenthal- DannenDramé
berg et
& Kortal.
mann
Becker &
Guzmán
Naranjo
FenkOczlon
FenkOczlon
B4 1
0.26
5
Matthewson
&
Truckenbrodt
Hohaus
& Vander Klok
Maché
Csipak
Mandelkern &
Philipps
Rubinstein
Rubinstein
B4 1
0.22
9
Li
Mari
Mari
BogalAllbritten
BogalAllbritten
Popp &
Stiebels
Popp &
Stiebels
B4 1
0.07
10
DobrovieSorin &
Donazzan
Gumiel
Molina
et al.
Escandell
Vidal
Grant et
al.
Grant et
al.
Dekker
Filip &
Kallmeyer
B4 1
0.06
11
Fischer
Bülow et
al.
Bergmann
Lameli &
Werth
Hasse
Enger
Enger
B4 1
0.05
12
–
Kunz et
al.
Neumann
et al.
Grafmiller
Verhoeven
&
Lehmann
Verhoeven
&
Lehmann
Lüdeling &
Adli
B4 1
0.04
13
AG-Gesamtübersicht
393
AG-Gesamtübersicht
2
B3 1
0.13
3
B3 1
0.12
4
B4 1
0.26
5
Anttila
B4 1
0.25
6
Iordachioaia
et al.
B4 1
0.24
7
Beutler
Müller
B4 1
0.23
8
Kratochvílová
Bross &
Hole
B4 1
0.22
9
Rentzsch
&
Mitkovska
Meijer
Meijer
B4 1
0.07
10
Sutton &
Filip
Asher
Asher
B4 1
0.06
11
Enke &
Mühlenbernd
Versloot
et al.
Zimmer
B4 1
0.05
12
Haider
Bildhauer
&
Schäfer
Bildhauer
&
Schäfer
B4 1
0.04
13
Donnerstag, 09. 03. 2017
1
B3 1
0.14
Lestrade
Hennecke
Werkmann
Horvat
AG
B3 2
0.03
Kindermann –
Anttila
Demske
Raum
Greco &
Haegeman
Levshina
Pezzele
et al.
van
Bergen
–
Skopeteas
9:00–
9:30
Guevara
Miestamo
Kravtchenko Lühr
et al.
–
9:30–
10:00
Amaral
Seminck
& Amsill
Wolf
Wechsler
Blümel &
Particke
Ellsäßer
Trouvain
& Kern
Freiberg
10:00–
10:30
N.N.
Kaffeepause & Postersession
Cap
Stahnke
10:30–
11:15
Jamieson
Ackermann
Pat-El
Csipak &
Zobel
Egg
12:15–
12:45
Hennig
Wechsler
Niebuhr
Waltereit
12:45–
13:45
Tonhauser
Marushak
Csipak &
Zobel
Weiß &
CarlucciDirani
Schnelle
& Donhauser
Jain et al.
Barnickel
Serdobolskaya
Deguchi
LeserCronau
11:15–
11:45
Senaldi
et al.
Zaefferer
Schwabe
van de
Vijver et
al.
Coniglio
& Hinterhölzl
Heim
BacskaiAtkari
–
Schwabe
Hennig
Dima et
al.
Wiese
–
Tonhauser
Wierzba
Marelli
Wiese
Poschmann
Schneider
& Mondorf
Struckmeier
Marelli
Tourtouri
et al.
Petrova
Haspelmath
Korth
11:45–
12:15
Dücker
et al.
SchmidtkeBode
Pijpops
& Van de
Velde
Kaiser
–
Grubic
Rohde
Renans
et al.
Patterson
& Felser
Linzen
Rohde
Meinhardt
13:45–
14:15
Linzen
Mittagspause & Postersession
14:15–
14:45
394
1
B3 2
0.03
Piñango
et al.
White &
Rawlins
Horch et
al.
Howcroft
et al.
Abschlussdiskussion
AG
Raum
11:30–
12:00
12:00–
12:30
12:30–
13:00
13:00–
13:30
13:30–
14:00
et
Westera
Raso
al.
Özge
&
von
Heusinger
Scholman
& Demberg
Chiriacescu
B3 1
0.14
2
Freitag, 10. 03. 2017
PetkovaKessanlis
Haselow
Trabandt
et al.
Henderson
&
McCready
Henderson
&
McCready
B3 1
0.13
3
van Lier
&
van
Rijn
SimsWilliams
Seržant
B4 1
0.26
5
Puddu
Zaika
Goldschmidt Ye
Mărănduc
et al.
Weiß &
Schnelle
Janssen
B3 1
0.12
4
Reinhardt
Schang
et al.
Jabeen &
Butt
ErteschikShir et
al.
Tokizaki
& Inaba
B4 1
0.25
6
–
–
–
–
–
B4 1
0.24
7
Alexiadou
& Lohndal
Haegeman
& Greco
Bunk
Frey
Müller
B4 1
0.23
8
–
–
–
–
–
B4 1
0.22
9
Final discussion
Lind
Sørensen
& Boye
Lind
Sørensen
& Boye
Kastner
Kastner
B4 1
0.07
10
Final discussion
Final discussion
Igoa &
Horno
Jezek
Babonnaud
et al.
B4 1
0.06
11
Loporcaro
Loporcaro
Moser
Döhmer
Kaufmann
B4 1
0.05
12
Meyer &
Szusich
Scheffler
Rehbein
Szmrecsanyi
Szmrecsanyi
B4 1
0.04
13
AG-Gesamtübersicht
395