ALLITERATION AS A SOFT WARE, PART I: A SHORT RETROSPECT FOR THE QUANTITAIVE ANALYSIS OF PIERS PLOWMAN SASAKIMichio 0.0.Some twenty years ago, I felt always uneasy as a newly−4edged instructor in Historical English Grammar at a Japanese college whenever the stude早ts tried to squeeze an a茸swer from me with Suρh innocent questions as‘‘Tell us how the style of alliteration was developed by those ancient people, and why.”or“What do you think面as the formal beuaty of the alliteration, if it had any P” Fortunately or unfortuna,tely, lingusitics could not be explanatory either about the causal.rela− tion of its birth or about the characteristics of its formal beauty. That is, I was not q・・1i且・d t・gi・6・・ci¢・ti且・ally・・n・i・t・・t・xp1・n・ti・n t・th・g・n・・i・・f th・alliter− ative poetry as a geologi6t could⑳about the birth 6f a mohntain. The students seemed to understand the limitations to theoretical fra皿ework of linguistics.quite easily,りut I still remained uneasy, because I cQuld not help feeling frustrated by my own inability that might have hindered a possible growth of the linguistic sicence. Another frustration was that European scholars of Gefmanic.4escent seemed to h・v・di・cussed th・f・・m・f・11it・・ati・n u・u・lly with・・id・1・ng gl・nce at th・p・eti・ ・tyl¢・・f M・diterrab・an・・igi・th・t h・d・1・eady・ecured th・msevl・・a・t・n母・・d f・・ literary taste even in Germanic countries after the Renaissance. A1七hough nobody can deny. 狽?e ef五cacy of a colhparative study.especially in order to reveal a structural peculiarity, it would not be fair for the descendants of Germanic warriors to weigh. 狽?eir own heritage always with some similari土ie串to the Classical literature. Thεre must be a way to explain the significahce of a literary heritage 「 quite independently of other cultures.. @It w6uld. become possible, however, only .after a theoretical frameWork of applied.linguistics had largely extended its possibilities. @to give a deeper insight in‡o humanity than ever. 1・lfeady b・1i・v・d th・t・ny・f.th・・ld her・i・p・・m・皐・d bee・…m・nqn verb・1.. property of a society preserved dearly by auralっral trans士nlsslons a坦ong many ・n・nym・u・p・・pl・,・・m・tter wh・t certai・ty・・uld b…c・ib・d t・it・・uth・rship・. If・・J w・uld p・ssibly b・all・w・母t・・upP・・e th・t the aliterati・n h・d been・w・・k− ing、d・vice t・prev・nt・preci6耳・messag・f・・m the er・・i・n.th・t与・d h・pP・n・d anytime partic亡la士1y in those days when people had.not bεen able to depend on a stable storage like letters. This hypothesis soon found a distant green light in the sprbut of information theory and its application to psychology. That is, o早ce some regularities were found about human informatiop processing, we were able to form aInodel on which a few regularities in the stylistic structure of verbal message would 58 Artes Liberale3 No.21,1978 b・an・lyzed t・ace・蜘・xt・nt・Th・n, Mill・・’s吻・碧φ・伽・励〃・酬伽θク.吻細, 6z〃。 was reproted tQ astound us with the ullexpect6d sirnplicity of oμr own percep− ti・n(Miller 1956)・Thi・.・ee血・d t・b・th・且・・t・lu・f・・th・t・・k, becau・e重t 7w・uld possibly且nd i士s relationship寅ith the quantitative regularity gf.syllabic arrangm611ts in w与atever peotic style there. h毎d been on ear士h. Then.,.we would be able.to explain a numefically defined structure Qf poetry as a soft宙are.for the aura1−6ra1 .transr[11SS1011. 0・1・Anoth町psychQlogical problem that.could not be avoided when we were 9・i・gt・9・apPl・with.・. pec“1i…tyl♀lik・‡h・alliterati・n w・・th・t・f・鰯. c認. Cer⑳ly・there h・d al・eady bee・numer…exp・・i㈹t・l rep・・t・th・t t・ld・・th・ ・貸ecti・孕ess・f・・u・i・th・・ec・11・f・・y・tim・1・・i・・1・di・g・verb・1…,・nd th・y seemed to lnake a powerful ally when“task involYed an inquiry about the function ・f・lli重erativ・.・・und th・t h・d.b6・・lik・ly t・gi・… u・f・r recall t・it・f・11・wi・g string of syllables. There were, however, also many丘ndings about the distrubanc6 ・fthe cu・・re・embling…h・ther in the recall・f・1in6・・v・・b・1血essag・lik・5乃・ s〃∫∫θα∫舵♂♂εoπ魏ξsθα論076. As far as工was goinεto fabricat曾ahypothesis that would satisfy my supposition above, it might be bet七er for me to neglect the meance ・fth・1・tter・b・t thi・w・uld・1・・m・k・i‡P・ssibl・th・t・士・・ult母・t th…yw・uld b・ crumbled away into dust.anytime at a slight push of this once ignored foe. Furthermore, we never.could disregard the organizing function of meaningfulness that had always been estimated consid6rably stronger thaII that of a sheer formal constltutent like cue. W虹at then suggested the ide毎of a possible modelling about the alliterative poetry was unexpectedly the theqry of Pulse Code Modulation(PCM) ln our mlcrowave telephony, where an ever−bending curve Qf a speaker’s physical v61ce was converted illto many rectangular blocks with the indices representing th・ir re・pectiv・h・ight・iゆi…ydigit・th・t・・uld b・・6・t in a・t・ccat・・f p・1・e・ without any.interference of various Iloises twisting round the message in the air (B・111968・48−53)・.If w・..・re・lw・y・li・b1・tQ・ew・ite an grigi・・l mess・g・.dh・i・g ・n・u・al一・・al t・an・皿iS・i・n・・h・・bee・rev・al・d by m・・y・f thg seri・1 rep・・d・Cti・n tests(Cf. Paul I959), we have to regard the mea垣ng of a verbal message as its most sensltlve receptor to the noises brought forth for皿aspeaker’島value syStem or a social prejudice of his group. That is, the most effective organizing ingredient for the recall of a message is nothing but what makes it most susceptible to the inter− ference・f・em・nti・n・i・e・・If ther6 w・・e m・・n・t・・e・t・ai・th・a・bit…yrew・it・ .work in those days of minstrels, one of them c6廿1d be to control the stylistic ・晦ce・f・giv・n messag・・1・r ex・mp1・, by giving…n・i・t・nt・u・t・af・w・t・i・g・ ・f・yll・bl・・i・・li・・j・・t lik・i・d・ki・g・n・utli…frect・ng・1・r w・v・with bi…y ヒhgits for PCM transmissiqn. 0.2. The first step in the task wag to define.the unique function of syllable, because there was of.cou七se no poetic systle that did not depend.on a numerical r・g・1・・ity・f.Syllψi・arr・hg・m・nt・. At th・t ti畑, th・丘・・l d・v・1・pm・nt。f Bl・・m・丘・ldi・n・t・u・t・・ali・m W・S. o・・mpt・d・・m・6h 1・・g・1y by th6 f・・m・1・n・lyse・ of emic.constitutents of language that a sub−emic unit like syllable was sti1116ft to negligellce.. The aρtua1. elimination of syllable from the emic systeln alrea4y 1 59 SAsAKI:AlliteratiQn as a Sof七Ware, Par七1 ・uggest・d th・t it. 浴E・n・t・u・i七・h・…terized by it・. f・・m・1・niq・・ness th・圃ect・d m・a・i・g・1ik・th・t・f ph…m…血・・ph・m・,. b浮煤E・ther・n・th・t m・d・a早1nte「一 mediate unit between the phoneme and the morpheme for anothedndispensable pur一 り・・e.P・ssibly, it i・f・・.th・q・・ntit・ti・・p・・cessi・g・f・peerh・・u・d・withi・・ thre・h・ld f…u・perceptl・n. Th・w・y f・・m Miller’・舳・g・t・th・m・ki・g・f ・yl・bl・w・・n・t 6・f・・a・might be expect・d・Already・Miller supP・・ed th・t J・k・b口 son’s distinctive features were bundled int6 a phoneme with his span of absolute j・dg・m・・t(cherry, H・ll・,&J・k・b,・・1953),・nd・dmitt・d th・t th6re c・uld be a ,ecu,sive chuhki・g・f the chu・k・とhunk¢d・nce b・f・re・Th・t i・・if. w・・pPly. th・ recursive chunki・g twice t・th・di・tin・tiv・f・・ture・, we c・n・a・ily・bt・in t与・ ・yll・bl・,・・d…m・・e apPli・・ti・n m・y p・・d・・俘・q・・ntit・ti・・ly・eg・1・t・d・t・1・g ・f・yll・b1・呂whi・h w・m・y・all th・・yll・戸i・ph・“・e as sh・w・i・Fig・・e l b・1・w・ djstinctive phonemes syllabies 艶atures 4142 S 一1Sn 轟 F一一含ls2 8 1 1 幽 Buffer(Tempo_ rary Memory) ____一」」 _ マ s phoneme sy.llabie ___1 SP syHabic phrase Figure 1. The model of recursive chunkings required an inductive examination 。,peci・lly・b・ut the real・i・6・・f・xi・ti・g・yll・b1・・i…d・ゆt・・ify it・apP1・i・bility・ Thahks to.the development Qf statistical or computational linguistics∴I could 。a,ily g・t・f・w b・t・u茄・i・nt rep・・t・th・t・・ver f・il・d t・gi・・.・・.・num・・i・・11y classified list of the島trings of Enεlish sounds(Roberts l965). As far as we put .th・m・gi・&1 number sev・n p1・…mi・耳・tw・t・th・upPer thre・h・ld・f b・th ph・nlc 。。n、titu七。ntS 6f・・y11・b1・and・y11・bi・・h…f・.mq・ph・m・・the e又i・土・n・e・f recursive chu面ng w・・a・cert・in・d by the crude reality・Thi・was rep・・t・d with・ 1。w unt。,ti丘・d・pecul・ti・n・ab・ut.m・t・i・・t・u・ture・・f p・rt・y i・g・neral(S・・aki 1967).. .Up t・thi・1…1・f m・・ph・m・, l w・・p・ssibly・11・w・d t・f・ll・w Mill・・’・di「6c− ti。n, becau・e th・units I h・d treat・d・・f・・w・・e un・nim・u・ly m・anlngless in th。m、ev1・・untll th・1・・g・・t・n・assum・d. m・ani・g・t th・且・・1・t乱9・・Of・・urse・it was to leave ro6m forざn extensi611.of the meaningiess models that工had aユready prepared th・n・ti・n・f・yllψi・ph・ar・whi・葺・・uld b・d・耳・・d q・it・irrel・v・nt to 即・・ni・gfulness.1・f・・t, we c・uld.・a・ily see th・t・・t・・ph・i・Old.エcel・ndi・ P・・t・y図been arrang・d with・ight.h・1f−1i…c・皿P・・ed.・f f・u・t・・i・・yll・bl・・ re・p・・ti・・1y.. Sd, it w・・n・t imp・ssib1・t・・upP・・e th・t th・・t・・ph・it・elf h・d been 60 Artes Liberales No.21,1978 .包reali・ati…fthe cum・1・tive ch・nki・g・・エ…d・・t・exp1・re thi・p・ssibility, Iw・・already g・i・g本・m・k・a・n・ani・gless ln・d・1・f verb・l mess・g・・t。dd。d with c“・・i・・f・i・th・P・th・t i師・・1d f・・ilit・t・a・・t・m・m・・i・ati・n(S・・aki l968). B・yρ・dth・.1・v・1・f m・fph・m・・h・w・ver, w・u・u・lly・ann・t di・ti・g・i・h whi・h one of the magical nulnber. seven・the cue, Qr the meahingfulness does really supPort ou「「ecall・f・Yerb・l messag・a・f・・a・it rem・i・・m・ani・gf・1・v・n i・p・・t. The contradictions享n the relevant psychologica1.findings工have just described in O・1・i・・vit・bly i・士ertwi・・d h・r・t・1・av・m・u・・b1・t・9・any f・・ther..Alth・ugh I wa 秩@sure f「・m th・f・reg・i・g・x・mi・・ti・・th・t…ht・・ll・d l・ngth・f・yll・bi・ph・a,e would b・t・k・n・ut e・・ily by・・imp1・・tim・lus−re・p・n・e t・・t.・・aq。・ntit。tive universal f・・all th・p・・ti・・tyles・n…th, i・・u1蜘t decid・y・t wh。ther I、h。uld attach a greater importance to a recall−stimulating function of the cue than to its anothe「fun・ti・・.th・t i・t・・a・t・d…aki・d・f i・d・x, with・gi・・…nt・xt。。l m,an− ing・Seemi・glγ・th・f・・mer w・・i・…p・・at・4 i・t・th・1・tter. Th,盃登erence, h・w・ver・might 1・ad t・th・tw・di登erent m・d・1・th・t。。nt,adi。t。d,a。h。ther at least i・・n・・f th・i・essenti・1・・Th・・nly w・y・ut, therefdre, w・・t・d・ap,y。h。. logi・al・xp・・im・nt th・t w・uld be c・mp・t・nt f・r ex・皿i・i・g the strength、。f。11 the f・・t…inv・1・・d here re・pecti・・1y, and th・n, t・ki・g acc・unt・f it、 re、ult、, t。 m。k, a model most adequate to explain the peculiar structure of the alliteration. 1・0・lt w…f・・亡rse imp・ssib1・f・・the experim・・t t・u・e a re・1・lliter・ti・・p・・t,y 「ec・・d・d i・A・gl・一S・x・n・・.Middl・E・gli・h th・t h・d・lready bec・m, an unk。。w。 fore嬉n language even for English与peaking people. Besides, the shbjects I could ask f・「h・lp were vi・t・・11ylimit・d t・J・p・nese c・11・ge st・d・nt・. If l w・・g・i。g t。 depend on the available subjects, I had to compose the m6ssages that could be 「et「ieved by their intuition・.I was probably justi丘ed, ho宙ever, in making the messages wlth th・ph・n・甲i・ry11・bl…fJ・p・n・・e l・・g・ag・, if・nly I・・uld prep・・e Some measu「abl・di登ere・ce・i・(A)m・ani・gf・1・ess,(B)1・ngth・f・y皿註bi・ph・a,e, and(C)・…t…ti・fy th・f・・t・rs an・lyzed・ut・f. th・・11iterati・・p・et・y. Thu・・lm・d・33一(27)messages i…der t・ad・pt・.1・ti・・quare a・th。 m。,t .approprlate.4 NO 724 for the experiment. In C鉢tegory A that represented meaning− fulness, the messages were classi丘ed illto th ree subcategories:(A 1)meaningless,(A2). partially meaningfulβnd(A3)meaningful. Category B had also three subcategories in the length of a$yllabic phrase:(B I) 12−syllabic,(B2)8−syllabic, and (B3)4− syllabic・Amessage unanimously consisted of twenty four syllabls that were classified into two l2−syllabic phrases for B 1, three 8−syllab ic ones for B2, and six 4−syllabic ones for B3, resp6ctively. Category.q showed whether a syllab ic phrase was cu・d by・ph…mi・・yll・b1・(C1)・・u・d by・ph・n6m・.(C2),・・with・・t・u・(C3). C・n・id・・i・g th・t C・nly referred t・wh・ther syll・bi・ph…e・i・・mess・g・h・d・ common cue・「n・t・it w・uld・imply requi・e lu・t tw・・ub・at・g・・i・・u・der th・pl・in P・i・・ipl・・f bi…y・electi…B・tl・ince th・J・P・n・S・1・・g・ゴ9・W・・i・・g・ly・ccupi。d by syllables of V and CV types with a few exceptions of syllabic consonants that m・k・a且・it・・y・t・m・f・夕n・bi・alph・b・t, P・・pl・16・m th・・yll・bl・・with・ut・plit. ひ の tlng theln mto smaller constituents even in a school education. The lninimum u血it of Ietters is also syHabi6. It was quite possib1㊧, therefore, that the 歯aiable SAsAKI:Allitera七ion as a Sof七Ware, Part I 61 subjects would not grasp a significant e琴istence of a sound unless it was beyond the level of syllable. Then, I had to prepare two different. levels of cue for fear that a phoneme would get.through a subject’s consciousness, leaving no impression at all or developing.to another syllable with anQther imaginary soun母・ 1.1.The.messages were, for example, as follows. Category A (1)A1−B 1−C1(meanillgless l2−syllabic phrases cued by a phonemic syllable) NOH:AYOYAK:OROHEMOTOATANAINOH:OTIOTEYAH:ARO WA YU NOTI. (2) A2−B 1−C 1(p訂tially Ineaningful 12−syllabic phrases.cued by a phonemic sy1. 1able) :τ㌔4NEハ4.4 K1(sowing)TI KE NI KA TO SE SO HI/L4且4 KO y.4 (tabac− cosnit)RO I RI MA NA.KE TA SO. (3)A3−B l−C l(meaningful 12−syllabic phrases cued by a phonemic syllable) SIB.4R.4Kσン4}琵4N、4K.4:Z’:r!1KEPO/51NP.41(;07’ODEルτ04:τ’:τン1 NO.(1.haveh’t seen you for a long time. Was there any trouble that prevented you from seeing meP) Category.B (4)A2−B 1−C2(partially Ineaningfu112−syllabic phrases cued by a phoneme) H.4RE,MO No(swelling)N. Ku Ko HE.NI MI BA No侭。 Nルfo No (real thing)HA SAUTO NAHI SE SL (5) A2−B2−C2(⇔artially r耳eaningfu18−syllabic phrases cued by a phoneme) TO N NE Rσ(tunnel)MU HO YO MA/:τ「E N KO o(weather)Yo sI YA NAμ「1 ZI RE KE(waving hair)RE SA MO HE. (6)A2−B3−C2(parti母11y meaningfu14−syllabic phrases cued by a phone血e) ∫ON(10ss)TE M工/∫E研.4(care)RI HA/5.41(rhinoceros)ME KOISσTO(strike) .HO A/SE K1.(cough)MO HO/∬RO.(white)SU KA.. Category C. (7)A3−B3−C1(meaningfu14−syllabic phrases cued by.a phonemic syllable) sσ7「ODE MO/5σRσN24 R.4/5σGσNIMO/5σsσME TE/5ひ1slN/sσ BE KI D浸.(If you.are going to go on a strike or make some other struggle set−ups, you have to drive the campaign for that immediately.) (8) A3−B−C2(meaningful 4−sy11ゆic phrases cued by.aphoneme)50 RE N且R.4/ 5.4S50 Kσ/51.(;O TO Nl/sE Eエ)∠451/50 N N24 DO/5σRσN!4 yO. (Then, get busy right away. Don’t take a. P0ss any more.) (9)A3−B串一C3(meaningful 4−syllabic phrases without c廿e) KoNos!11/0sEzloμTTEルω/MσKooGオ/DooNIMo/N護1∼σMオ1, (As the case stands now, he won’t bb able to claエn himself, even if you are going to flatter him.) Just as in the written form串above, every meaningful word was delivered to the subjgcts as a staccato of splitted syllables. At each.boundary of syllabic phrases that is marked with an oblqiue line.in the above, a tinkle of bell was sent to inform them of the renewal of ph±asal unit. 1・2・The subj ects were.199 freshmen ih 1970 at Iwate Medical University whose 62 Artes Liberales No,21,1978 ・g・・v・・i・df・・m・ighteen t・tw・nty一・ev・h,30・45 i・.・Yerag・・. After・・impl・fre6− recall test, they were c!assified into tw6nty−sもven groups so that each group would h・v・th…m・av…g・・c6re・bt・i・・d f・・m.th6 t・・t・み9…p・…i・t・d・f fg・r subjects・Each group was assigned to listen thr6e kinds of mes6age that belonged more or Iess to the sam6 subactegory. Although three groups had to deal with the same messages, t取e order of messages in transmission was different for each group. E・・h・ubject・1i・t…d. t・th・assig・・⑳essag・・th・・ugh hl・h・adph・n・・i… ordinary Lab booth, to which he. ヤad already got accustorn6d in his weekly drills of modern. foreign languages. A message was played only once, and. after a short pause he had to wri‡e it in Japanese syllabic charact.ers at a㌻ink:lihg signa16f chime. Between the messages a 3−lninute passage of classical music珊as played for relaxation. The results obtained were reported with detailed data ill l971(Sasaki et al.1971),where the analysis of variallce of overall re6ults showed an unquestionable predominance of the meaningfulness over the other七wo factors, while B2 vs. B3 and C.1vs. C2 also showed a noticeable approximation to a statistically signi倉cant level. Then, another analysis of variance applied to the pooled data with latin square arrangment brought a convincing signi且cance to the above tendencies of B.and C factors・That is, it was revealed that(1)human in且rmation processing of a verbal message was greatly facilitated by a numeri6al chunking of syllables in around the span of absolute judgel耳ent, and (2)asyllabic.cue given to each syllabic phrase exerted a strollger valence fQr a subjecfs recall than the rest of the same category. Achain composed of the shortest 4−syllabic phrases, however, brought forth a con− siderable disturbance.produced obviously by a mutual similarity of given cues. 1・3・エn・・d・・t・see士he e避ect・・f th・・e t蜘f・⑳rs in multipl・, re6urrent percep− tions, another experiment was carried out on.the eighteeh messages exceゴt nine menaingful ones that had already proved to b俘overwhelrningly powerful in the first experiment. The.9ubjects were 180 freshmen in 1971 at 工wate Medical U・iversity whg・ang・d i・・g・f・・m・ighteen t・thi・ty, with m・an・g・20.02. After a.simple free.recall test, they were classi且ed into eighteell gr6ups with the salne averag・.・c・re・T・n・tud・nt・i・ag・・up w・・e a・sign・d t・1i・t・n t・・nly・n・ message且ve times and write it in Ja亘anese syllabib characters at.atinkling signal of ・him・d・・i・g.・a・h p・u・e b・tween the rep・ati・g.mess・ges. Wh・t th・y.h・d written before was lnasked by the other side 6f a turned page to prevent them from a casual reference. The results were reproted in 1973(Kanno et al.1973), wh6re the findings of the丘rst experiments were ascertained once more. Roughly speaking, B・耳dCf・・t・rs are u孕d・廿bt・dly re1・v・nt. t・the recall・f・v・・b・1 mess・g9, respectively. In splte of the outward success in corro1)orating the significances of three fac士ors in relatlon to oロr memory, their difEerences in actual strength led me to a vaguely ♀xpected prospect that I could not leave a supPosed processing of alliterated mes− sage solely on the function of cues any more, Although the absolute judgement h・dbee・・b・i・u・ly re・・9・ized亀・aq・・ntit・ti・・u・iversal f…u・p・・cessi・g.・f a metrlc mes呂age, its significance was far beyond the structural peculairity of allitera− tion, It was an essential for a11.the poetical styles on earth a島had been SAsAKI:Alliteration as a Sof七Ware, Part I 63 presupposed in O.1., where the alliteration made only a tiny fraction. That.was, however, the only effect I could certify experimentally as valid on this level of stimulus−response psychology.. she r年st of three factors were still Ieft on what.kind of model I could then make in order to explain the alliteration.. 1.4.Acloser examination of retrieved Inessages revealed that皿eaningfulness often induced the昌ubjects to.do arbitrary re脚rite works, though the large ahlount of whole score covered the partial losses on the surface of numerical.data. On the other hand, cues ga,ve the studentS no七 〇nly a siエnilarity disturbance but a’lso a marked promillence around Which they could gather up some of the assigned recalL That is, both meaningfuiness and cue have a Janus−like f耳nction, renledial as well as poisonous for our processing of a verbal message.. If thes曾two factors eYer supported an aura1−oral transmission of.aheroic poem, there had to be an active coding to cover a poison⑳s function of one with a remedial function of. another, though it had.been deviSed. quite unconsiciously at the beginning・ There, I had to abandon a still且ickering hypQthesis described in the first half of O.1.that could probably be called the Drawer Theory, because it h母d supposed a cue to function as a label of a drawer contalning a syllabic phrase numerically controlled under the absolute judgement.1.nstead,1. settled myself on the supposltlon s廿g− gested in the latter half of.0.1. which might be called the Tongue Twister Theory,. where the alllteratlon had been supPQsed to be a codlng devlce to turn a message lnto akind of tongue twister. To quote a famihar example here agaiH, who ever dares to rewrite 5加εθZ♂s sθαsん6μ∫o%彦乃θsθαs加7θP I.t has long been kept uncorrupte(! because of its tongue−twiSterness. To assume such a coding accorded.well with one of my early suppositions that. there might be a process of selective choice in an oral.reproduction of alliterative.poern through which the cues Would e妊ectively narrow both the granlmatical and semantic.latitudes of a given text(Sasaki l968: 64−5), 2.0. The diffe士ence between the two hypo亡heses may better be illustrated by the figure below that士epresents the three rec七ahgular syllabic arrangments. ?or.Pα ω昭s、厚706gα76乃〃βψ64.舐プ珈(Bewo耳lf 1.64). In the figures,(a)shows a full line still unsplitted tQ surpass the threshold for immediate chunking of syllables. The Iine is then segmented into two half− 質. lines in(b)in order to control.the size of a syllabic.phrase below the threshold, and an alliterating cue.has to be put to som6where in each column, for example, a5 an index to predict a following word or string of syllables that satisfies.a g工ven grammatical and semantic c6ntext. This is a model.for the Tongue Twister Theory.. Ft1rther, if we arrange the syllabic blocks again so as to put an alliterating. syllable on the.top of each column, we obtain a model for the Draw6r Theory as in (c). If I wa孚Inerely running mad after the g缶ciency in a qualltitative processing alone, I was doomed to run ashore by the char耳n of(c)that was really a match.for Sirell as far as I stuck to the. silhilarity to PCM.工was fortunate enough, however, to retrea‡afloat from the Water’s.edge at(b). T草erefore, the only possible interac− tion b¢tween A and C factors wa呂acomplementary one outlined in the explana一 64 Artes Liberales .No,21,1978 ぬ 一一 y晦7θ5乃0147一一一一一一一一ニー一一一一一 乃 乃 乃 乃 .乃 一一一一一一●’ 一一一一一[’ 一一一一一一◎! (a) (b) (c) Figure 2. tlon glven to(b)above. Tha,t is, not only a sernantic predictioll enabled a cue to d・aw・w・・d 6・ph・a・e, b・t.・n・・bit・a・y士・w・it・w・・k b・・ught by・p・・ti。ul・。 1nterpretatlon Inight.also be checked by the distribution of a cue that would ・登ecti・・ly narr・w the rang・・f・・elective ch・ice. lf the c・mp1・m・nt・・y i・tera・ti。。 has t・b・f…h・1・t・d・・毎P・・9・ammi・g・r s・ft ware b・ilt・1ready i・th・・tyl・・f ・llit・・ati・n・it w・uld need tw・P・・P・・iti・n・th・t・re i・terd・p・nd・・t・n,a,h。th。士. If th・m・ani・g req・i・ed here i・X, select・w6・d・・w・・d g・・up with・u・y (1) to fulfill its context. If th・w・・d・・w・・d g・・up req・ired here i・th・・n・with・u・Y,・upP6・e a (2) meaning X and.an arrangement of words that satis丘es it. This hypothesis was reproted at the 29th Annual Meeting of Tohoku English 1974. Literary Society in 2.1. The supposltlon above had then no other way to bg validated ex6ept a possible co ll・ti・n with・n・v・i・ble reality・If th・・e曽…n・lliterati・・p6・m with ●fdi荘erent・diti・n・・an・x・minati・n・f th・i・di登erence・w・uld p・・b− varlous coples、o ably give substance to good or bad of the hypothesis, provided that each copy was agenulne re且ection of its oral.version under a stand ardized orthography..It was. of course imp ossible. The only a11iterative poem that has a certain.amount o壬 different copies is P づ〃・P肋彫・紬・tth・i・d晦ence・are c…idered.t・b・m・・tly scribal according to th especialists whQ have beell deeply concerned in its.textual problems. .lt w…h・w・ver・b・y・nd d・ubt th・t the religi・u・p・em w・uld m。k, a best ・ampl・becau・e・f th・v・・i・ty・f it・c・pi・・th・t・p・ead・ver m・ny di・t・i・t・.。nd peri・d・・if・ply l w…II・w・d t・P・t・・th・9・aphi・di登erences i・・t・ad・f.ph。n,tiC on6・・Alth・ugh・・m・・f th・di登erences the c・pies sh・w・d i・th・i…th・9・aphy h・d t・b・phg・・tic ev・・i・th・f・i・t light・f・u・hi・t・・i・al h・g・i・tics“・d di・lect・1・gy, w・ 事 がS4sAKI:.Alliteration1・as a Soft Ware, Part I 65 cannot explicitly distlnguish the rest of Iargely graphemic quality fronl those purely phonetlc variants. Hence, if I. was.going to use.the samples in.orde士to get a possible validation for the built−in soft ware, the task had to. be premised on the assumptiQn that the same hypothesis was theh going tQ be apPlied to the scribal transmission as. ィstep farther beyond an aural−oral pr6totype. The only dif− ference I..had to expect there was that the number of possiblβvariants for a word wo耳ld infallibly be inc士eased by tねe number.of exclusively.orthographic ones. The increase, of course, did Ilot make a prdhibition against“n appllcation of the model to.a. 翌窒奄狽狽?n version, because the more the burden of variability increased, the more the strength qf an organizing system woμld be exanilned accurately. 1.n spite of a few contradictions.in the chronological aspe¢t of its textua1 ¢oblemsJ chose珠version as a hypothetica!orlginal of.the一酬〃s P♂oω粥α% .according tQ thO exhausti>e analySis of Kane(1くane 1960)..Even.if thereΨas a lllistake in hls 6upposition, it would not.aflect my. work:at all from the premise that amisplaced prototype of the trans且guring poem c㎝1dわ6 regarded as a forln restored at the traget of a regressive.transmission from it61ater for血just like an Indo−EurΦean倉arental.Word that had been rερonstructed’by Comparative Lingusitics. Th律t is, whether a. 狽窒≠獅唐高奄唐唐奄盾氏@had been directed to plus or mihus in co−ordinates of elapsed time did notエnatter in this case, if only一彦was admitted into the working notions.(Cf. Sas印ki 1968:5&一9)l What was deeply involved in the. problem above was that, although.it. was quite obvious that every one. 盾? various verSions had pot necessarily. @been .derived direρtly from A, I had to depend on the Inodel of.aradial transmission emittQd only from the supposed original,. simply because there was no possible way to且nd or fabricate a complex network that had士eally g6t the me$sage士elayed among Inany scribes. . .. 』 2.2.In order to simplify the pfocedure.for quεntitativ6 treatment.of textual d置erences,.工prepar♀aset of variables that cou14 be measured within the span of a full line. T葺ey we士e as follows. 7:.Sum of words in a lille. 盟: Number of words in a line that have at least an・.(》rthographlc.variant, .irrespective.ol number of variants one 6f them may have, .む/7: Variant ratio that sho甲s the propori‡io耳of varying words to the.suni of words in a line. ≧4/診: Alliterated varia戸t ratig that shows theかぞoportion of alliterated varyihg 曽ords to the su]〔n.of varying words in. a line. 5ψ:Alliterated synonymity rati6 tha士.show串thg PrQportion of.. aUi七erated varying words unchanged in their meanings to.the sum of varying words in a line. ”一、4/汐: Unalliterated variant r年tio.th母t shows the. proporヰion of unalliterate(i’ varying words to the 6u血of varying wordS in a line. .. 1 L 8錫/汐:Unalliterated synonymity ratiQ.that shows.the.proportion of. unalliterated varying words uichanged in theif meanings to the shm ofサarying words in a llne。 脇粥砺:Maximum number 6f varia亘ts for an alliterated word in a lhle. ● 66 Artes Lib6rales No.21,1978 脇田砺:Maximum. 獅浮高b?r of.variants for an unalliterated word in. a lin61 1f I wds going to ihc1旦de a word added anew to a Iine in t#e category 6f varian本 only because it.couldわe copsidered a derivative from zero, a sum of.varying words in a line might.so血etllnes.exceed total sum of wqrds in‡he original to swing the ・・ti・・h・n imp・Φr・f・a・tig・・S…只・h年・・w m・mber寅・・n・glec#・d i…unt wh・n it h・d.…q・i・・1・nt i・th・t・xt・・pP6・ed t・b…igi・・1,0・.. 狽?・gther ha・d, if・・ alliterated word had.derived an unalliterated variant., the.derivati寸e was included in the gategory of alliterated variant t6 denote. only its derivation. That is, no 士natter what from a derivative had finally assumed, it was classi且ed by its origin.in order to se6 t取e strength of.alliterative restriction in its true proportion, Similarly, if a line had two dif[βrent alliterative sounds,.one that was same or smaller in nロm− ber of allit6rated words. was neglected to avoid an ove士6島tim3tion of the spell.of alliteration. Then,.e母ch line of A versiQh was. going tO be reprinted on the toP.of a card with 合very variant below its equivalent. in the. supposed origina1..Source of each variant was marked with t虹e familiar code for the existing copy in a column on right hand(Cf. Kane l960:1−18). For exampl合,1.50f Prologue was expressed as follows. Tα協1. Varian七s for l.5, Prologue. 5, But on a may mo「wellyn19 on maluerne hilles R vpon E aPQn L 0.〃Z. Mayes .VHJ Morwnllynge V 皿ornynge UDChHJKWM mo「rwe RH2 morn LE vpOI1 TDChH2 maluarne malueron ChL J hulles. DVL hullis H hellys M The values of variables here are in the following亡able. 丁励Zβ2. Variables for l.5, Prologue. 7 勿 ”/7 浸1汐 5ψ (び一遵)/び sゆ 8 6 .0.75 0.50 0.50 0.50 0.50 . Hρre…麟i・f・uτf・・物・・陶㎎・・ゆ・.吻鰯i・three f・r殿・・..Alth・亡9与 the quanti且ρation of 22651ines in al!for an overall computatiori is still on the way with soine inquir璃. 艪盾浮ソan adequate.statistic.method,..I am practi6ally. con且dellt ・f』th・マ・lidity・f th・T・n9耳・TwiSter Th…y・becaur・n・・th・r. t・nd・n・y t蜘th・ supPosed one aPP白ared to break th年s亡pefiority of alli七erated synony並i‡y. ratio so far at least. 盾氏@the s亡rfac60f arithmetic means.. SAsAKI:Alliteration as.. a SQft Wafe, Part I ..67 1have nothihg more to say in tレis interiin rep6rt except my deepfelt gratitude t・M・・Y・ki・Ohkゆ・,. P・・fe・・…fP・y・h・1・gy,・nd Mrs・Mi・W・t・n・bと.(f・・merly K・・h・)・Lゆ・f・t・・y.Assi・t・nt・b・th・t th・lw・t・M・di・・1 U・iversity f・・.th・i・ valuable Co−operation i耳the e琴periments to which their students also contributed very much.as subjects. Referellces 1..Attneave, F .4勿1勿観。%(ゾ1顧07辮α劾%Tんθαワ≠o P砂。乃oZo8γ. N.Y.:Holt, Rinehart 、 &Winston,1959. 2.・ Audley・RJ〆‘A stochastic皿odel for the act of choiceノ’・乙80づらM’θ醜040∼08γ απ4 P雇10soク勿ノ(ゾS画6%08,耳agel, E., Suppes, P.〉.&Tarski, A., eds,,. Stanford: Stanford U.P.,1962,391−9, 3. Bar七let七, EC. R6〃z6〃zろ6短%81 .4 S彦観む 勿 E擁)θ擁〃36π≠α♂ απ諺 50擁αZ −P5ア碗oJo8ツ・ London:.C&mbridge U.P.,1932. .4. Be11, D..A. Z蛎07窺α渉¢仮丁ゐθoη伽4∫6∫E銘8勿θ67吻g.4クμづ。α渉づ。πs. Londo且:Sir Isaac pitr耳a耳.&Sons,1968, 5.Bliss, A.J.㍑θM6〃θ.(ゾBθoω働ぴ.oxfQrd:B“sil Blackwell&Mo仕,1958. 6.Broadbent, D.E Pθ7ρ6卿。%α痛Co粥甥襯づ。α彦ゼ。%. LQndon:Pergamo11,1958. 7..Broers, A.C. and Zw餓n, E.J.,‘‘The information value of initial let七ers in identi且ca七ion of words,”ノ’o%γ%諺(ゾ7θ7ろ認L6α7溺πgα”476乃α♂. B8ゐα面07,5(1966),441−6. 8.Brown, R and Hidum, D.C.,‘‘Expect昂nQy and the perception of syllablesノ’五朋g%α8θ,. 32 (1956), 41..1−9. 9.cherry, E cりHalle、 M,,and J3kobson, R,‘‘Towards t耳e.logical descrlption of languages in their phonemic aspect,”.乙α%g拗86,20(1953),34−46. 10.Earhafd, M.,‘‘Cued recall and free recall as a function of the number of items per. cロe”,ノ’oτ〃ηα♂(ゾ7θ7δαZ」Lβαη¢づ%8α7¢47θ7わαZBθゐα四つ07,6(1967),257−63. 11・K・n・,G・P廓P」・ω吻・%・丁田浸7〃・繊1・L・唄・n・U・fL・耳d・n,1960・ 『 12.Kanno, M., Sasqki, M. and Ohkubo, Y.,‘‘A psychological s七udy of allitera七ion; experi.mental supPlement to S解θα嘱M8脚η, II,η浸槻襯♂丑6ρ07’(ゾ1ωα渉6 M「θ漉偽Z σ%づりθγ5つ砂,SoんooJ(ゾ」しづろθ7αZ/4γ’sα3¢4∫6つ6ηoθ5,8.(1973),.95−104. 13≧ Ka七〇血a,α g囎㈱痂%8α編M脇oγ伽%81 N.Y.:Hafner,.1940. 14.McL鼠ughlin, J.c, A G7砂んθ鰯。 Pゐ。πθ編。 s彦%吻q∫α M客思1θ一E囎あ5ゐ Mαηπ507ψ’・. The Hague:Mou七6n,.1963. 15.M.illerl G.A.,“The magica1.number seven plus or minus two;.sQme lin斗its on our capacity 6f.processing ihfo士mation”, P5ッ。んoZo860α♂Rθ廊ω,昏3(1956),81−87. 16. Oiji, T. Tんθ黙認。%(ゾーP留餐PZoω〃zαπ,.Tokyo=.Shinozaki−shorin,1968. 17・P・rry, J・.Tん・助・乃・♂・甜σH%脚πc・翅隅癩・励箆・L・ndqn・u・f L・nd・n,1♀67・ 18.. oaul,1.H.,‘‘S七udies in re田骨mbering,”一P5γ罐δ♂08¢oαZ ls錫β5,1−2(1959), Monograph .II, 膨 19.Roberts, A.H.A.,オS彦翻ε≠初1躍ηα砂5づ∫〔ゾ4耀γ伽πEπ8Z勧i The.Hague;Mohton, L 1965. 20. Ross, A.S.C.,‘‘Philological pfobability proble血,’.’∫o%γπαZ(ゾRgッ認S脇彦ゼ∫’づ。αJ Soo¢θ妙, B 22 (1950), 19−50. 21.. rasaki, M.,‘‘A hypothesis on並sychologipal reality of syll耳bleノ’オ露朋αZ.R砂。π(ゾ ∫ωの彦θMθ4づ。αZ.σ%づ”θ75づ吃ソ,∫oん60乙(ゾ」しづわθ7α1∠47’sα%4Soゴθっ¢oθ5,2(1967.),41−6. 22.S母saki, M.,“Style and memory:alliteratiQn as a modulation. of verbal message”, 2わづ4.,3 (1968), 55−71. 23.Sasaki, M., Kanno, M,, and.Ohkubo,. Y.,‘‘A psychological study.of alliterati6n; experi血e耳tal supplement to S砂」診㈱4 M粥。η,1ノ’伽4.,6(1971>,77−S7. 2.4..Skeat, WW.五壇8伽4’εP繍彦舵Ploω勉α%α編Rづ伽γ諺.. F乃θ丑θ翻6∬. London』1 0xford U.P.,1886. 25..Stevick, R.D♂∫砂7α58g吻θ%観5, Mθ彦θ7,α銘躍.動θMα%zφ507つμ(ゾ’Bθoω翫弄 The Hag亡e: MoUton,1968兜
© Copyright 2024 ExpyDoc