Effects 。f Partiaー Reinf。rcement 。f a Discriminative Resp。nse 。n

Memoirs of Osaka Kyoiku Effects University,Scr』,Vol.42, of Partial Response No.1,PP.4ト49 (September,1993)
Reinforcement of a Discriminative
on Extinction in the Turtle:Massed
Trial Acquisition Masato lSHIDA Training
and Yasuko NAKATA
,
Department of Psycho'ogy, Osaka Kyo〃(u{ノ
η'vers"y, (Received Three groups Y-maze;one with continuous They received tinction. of turtles(Geocle〃ly∫reeve∫
with 50%single during groups performed acquisition results, which demonstrated to those of analogous prehensive Key interval of correct than the CR interpretation in a black-white discrimination response group group, determinant of a choice 582
by extinction of the且earned 10-day acquisition no group difference indicated difference during with n.o difference extinction of PREE response.
and 5-day between extinction;the between task in a
and the third
during Rd effect(PREE),are and SA
The
compared
in terms and the response ex-
three
two partial groups. with turtles and rats, and are discussed as a possible Words:turt且es, condition a typical partial reinforcement experiments Japan with 50%random(Rd), followed but a significant better trained the second reinforcement(CR)schedules groups intertrial のwere alternating(SA), on the number Osaka, April 26,1993)
8 trials per day with massed-trial Analyses Kash'wara-C'ty, of the length measurement of
toward.com-
behavior.
partial reinforcement extinction effect, discrimination
1.INTRODUCTION
Weinstock.(1954,1958)gave different.groups 33%,50%,67%,83%,or tween resistance became to extinction a hallmark PREE.The PREE ditional hypotheses grows of rats one 100%reinforcement and experiment as a monotonic percentage described resistance function He of reinforced concerning has been that schedule. the partial of the frequency a 17%,
This
extinction it cannot effect,
be explained the associative of reinforcement be-
acquisition. reinforcement reflects under an inverse'relation trials during as paradoxica璽because to extinction trial per day, found by tra-
strength during which
acquisition(Bush
&Mosteller,1951;Hull,1943).
There reported are a vast number the PREE Lewis,1960 and PREE, especially parison with the underlying fact that some almost Robbins,1971). those rats as experimental experimental studies studies produced with rats and goldfish reward and subjects. condition(for useful rats and mechanisms(Bitterman,1975). between specific goldfish from learning using every Comparative with information sed the similarity of studies under have provided information Most exhaustive some which also the discussion in terms of carry-over
determine whether enabled about about
the com-
generality one who , mechanisms have
reviews,see
evidence Couvillon(1984)is trial variables of them despite or not a PREE of
discusthe
occurs
42
M.IsHIDA and Y. NAKATA
in goldfish.
Although some animals, such efforts marsupials(Papini, subjects ly the similarities other and It is very than interesting the results to a question is assumed to be between lagher(1964)gave turtles studies example,some choice with have turtles is needed
comprehensivebeen found done The present (1964)using a FREE with a Grice a PRE・E PREE(lshida
the
in these studies. ・unw・y・
found in terms discrimination rapidly experiment behavior because to that the FREE other on the formation the black new the present we use a Y-maze placed It is some-
not all studies used discrimination learn-
However, with other
a T-maze(Pavlik&Lehr,1967;
been to continuous alternation N-lengths and group
than middle and used section random a random and then more number reinforcement required of N-R group.
white, respec-
to enter learning with two groups, either
Instead,
rats.
continuous
decrement hypoth-
to extinction, A rewarded variable a rewarded the
as the appar-
them, N-lengths. The precede groups, and they
and between resistance transition, a tank gray ran only tria!s which reinforcement has black, of different scripta on the apparatus used determine an N-R 、exting-
of Wise&Gallagher
to the generalization the number rats
of Pseudemys in a discrimination According of nonrewarded was them Wise&Gallagher trial defines .which instead one of discrimination variables transitions the study Wise&Gallagher as a result groups. after a nonrewarded of the number points, by painting in the gray esis of PREE(Capaldi,1971),sequential of N-R reeve∫", reinforced rats,
to replicate in two sections part.ially reinforced groups. of groups, reinforcement in to'rms with
in Wise&Gallagher
in a spatial in which Geoclemys differs has frequently the formation and.50%random PREE basically end of the tank which as the number task continuously species, study into three turtles were or the white than of experimental w'as divided As regards sta-
in a visual
in a spatial was designed a relative且y However, tively. Their of learn-
Wi・e&G、1-
box. box(McFarland&McGonigle,1967). both of
of their phylogenetic ノ・PP、 ・at・ ・.0・ly a simple situations dichotomy animals.
・th・
in a choice This species because and aquatic a reversed more used, atus, which with
a visua且task(Mackintosh&Holgate,1968;Sutherland,1966).
significantly different used task using investigators learning to find uished single a few studies or not turt且es are boundary in analogous Spear&Pavlik(1966)reported addition was mamma且s found the PREE Spear&Pav且ik,1966)and fined information to discuss a T-maze(Lewis&Cotton,1958;Cotton,Lewis,&Jensen,1959)and failed occurring Only the studies t・・t1・・u・ed・
a visual that they reported discrimination of
Couvillon,&Bitter-
in order by Bitterman(1965,1975), ・t・di・ ・with surprising species.. species whether as suggested Fi…ut・f・i・
(1964).For a variety didelphid
and goldfish that halfof not a single leads rats have by using to find it(Eskin&Bitterman,1961;Gonzalez&Bitterman,1962;Pert&
ing mechanisms, tus which done Diercks,&Capaldi,1961;Wise&Gallagher,1964)but half failed ing with been further among to note Bitterman,1970)though what have honeybees(lshida, rats, birds differences &Kitamura,1988;Murillo, other research octopuses(Papini&Bitterman,1991), using turtles. systematic Mustaca,&∫Bitterman,1988), man,1992)and through toward as toads(Muzio,Segura,&Papini,1992;Ishida&Yoshida,1983), N-length such
trial
is de-
trial.Thus,in
we set up the third group,50%
transitions and less number of
Massed The subjects cause turtles mals. The saurs, or stem ago)to Partial Reinfr)rcement chosen are of considerable are the best extant common reptiles, which have closely related reptiles, survived from 43
in the Turtle
for comparative of the ancient of present-day virtually Response importance representative ancestors the present ordermost here of a Discriminative reptiles birds, and the Triassic unchanged(Romer.1966). analysis that evolved mammals are the cotylo-
period(225-200 Therefore, be-
to mam-
millions turtles years
are the living
to that group.
夏1.METHOD
Subjects
Thirty experimentally dealer in Kyoto, width. They were (28℃)and they imately made housed 6 cm. brick The The in individual These with turtles(Geoclemy∫reevesii)obtained as subjects. illumination. were ×6cm naive served bottom were and home in one corner 9-11 24 cm long,39 the turtle and cm constant wide, water with a commercial
in length with fil且ed with was covered to allow cm in a room were tank from measured located tanks plastic of each was placed tanks home translucent turtles temperature
and 27 cm high;
to a depth 1-3 cm 7-8 cm in
of approx-
of pebbles. A 12×10
to land on.
Apparatus
AY-shaped jects」t maze was filled with tion of food. arms(goal boxes each were given manually to'the when wall started was an angle subject from goal box, of the end when of the subject by 30 watt wall. the sliding reached just above fluorescent could door was from light 60 cm cm 10 cm see a black raised 10 cm the end above box and and was goal
20 cm.
of the end wall. of
food They for turtles)
were box. ejected
On the
glass,9×9cm, was from.
which stopped apparatus was
presented by a stopwatch box.The during which inges-
the two
the the other
of the goal acrylic the bottom the apparatus and starting from the bottom. stimulus goal The to the back recorded above of each and area each,commercial transparent were to train the sub-
to the bifurcation the end wall or white latencies box from from the bottom, Response 10 cm hole,13 a choice them. attached mg used the turt且e's moving the starting pellets(15 about was long), the starting 10 cm long, was the small reached cm separated of 45ヨReward so that a subject the back operated, to facilitate of 50'between high. From manually the subject of each embedded and 20 cm board, box(30 an angle dispenser,which box with of 6 cm of a starting cm】ong)with door, polyvinyl-chloride to a depth consisted 10 cm wide A pellet of light gray water maze sliding goal were end The boxes,38 Avertical parts. made when was was
a head
i1且uminated
experiment.
Procedure
(1)、44叩tation period Before the laboratory for two daily. was given Sunning months, for about For the first half of this period, was eventually reduced the preliminary during which an hour the turtles in a gradual manner training they were began, sunned, per day as long received were were fed, cleaned, as weather ad libitum until they the turtles food;the receiving condition amount 8 pellets adapted and to
handled
permitted.
of pellets
per day to一
44
M.IsHIDA ward the end of adaptation
(2)Preliminary given training a certain both placed were center place was gradually also place. shifted From Day eaten the day during assigned of each of starting to consume training. was Daily On and Day days alternating with three reinforced, sequence terms of average tained to one stimulus black and selected above orders. food 10cm pellet the level spotlight was was If the response Food the subject adopted that when nor spotlight tered the positive move from trial started box from and was was given was Rd illuminated subject was then was box to
introduced
minus the amount
and half the trials were
for the latter group. and Nz)resulting transitions. alternated the CR group,if was the positive same procedure stimulus area was detained from training, white posi-
located either side of
was cor-
about 30 cm
for about with the
pellet food nor
were
area(self-correction
on nonreinforced stimulus after a subject
and neither associated was applied the goal with
the response stimulus and a latency to the positive and
after consuming recorded always respectivepositive entered immediately was con-
board 10sec by the spotlight which SA in accordance door the subject The
in 1.4 in
The transitions, stimulus the sliding 50%
trials per day for ten
the reverse an error removed the subjects
groups:continuous
of N-R was removed 120 sec, an error by hand bases,
position
the starting training, groups, the black when and the positive the
after the last trial of
eight of N-R stimuli terminated incorrect, near placing door ration order by pulling and the subject guided and halfreceived reached side and box within gently with and the spotlight chose given N-lengths(N, the end wall.For The day.
14, feeding
reinforcement(SA;n=10), and the number two from reinforcement of the number other of the spontaneously stimulus the starting and the subject different in this experiment).The a s'ubject food The delivered of the water. was provided. when The in the starting reinforcement. except Position and reached rect, one method negative. placed goal boxes given board the
each as in the feeding
sliding of acquisition of three was trained manner 15 mg on a quasi-random N-length group negative. Gellerman been of average of each which was placed time, in
training.
trials. For the SA two during until Day at least 30 min were and 25 in terms in terms tive eight presented of trial and daily eight runs tank 20, the first day unreinforced, N-length ly. A half of animals had on reinforced in the Rd contained the white was reinforcement(Rd;n=10).They pellets l and30 given alternating and half were reward in the same experimental fashion on both per run. The food,which reinforcement(CR;n=10),50%single random was pellet in the home training training box daily and
session animal At the same box was one provided in quasi-random goal being days, days received
handled of 5-min 5,each counterbalanced. 14, an animal box pre且iminary (3)Experimental were to the inside were for the first four subsequent there. the experiment.
the subjects stimuli at the outset Day pellet the inside in the apparatus, neutral On 10 to Day on the last day of'this given of apparatus or left)being goal with On toward the end of either the subjects fed one shifted was this period was throughout period, the rest of this period. and its position(right was kept of adaptation in the apparatus, the center fed during with During exploration near of'apparatus ration two months of training Reward-free was Animals Following for 14 days. amount of arms. subject 、period.Thisfood training the preliminary and Y. NAKATA
to the PR group
trials, neither
20 sec after it had box. If an animal of 120 sec were area and given en-
failed recorded,
to'
food with
Massed Partial Reinforcementof a Discriminative Response 45
in the Turtle
8
ω山ω 20 α gり山α トO山α¢O O
7
6
5
4
3
▲一 一▲CR
2
ひ 一 〇SA
●一
1
0
12345678910
DAILY Figure 1. Mean 一●.d
number of correct BLOCKS responses quisition(CR;continuous OF 8 TRIALS
for three groups reinforcement, on each Rd;random of 10 daily blocks during partial reinforcement, ac-
SA;single
a且ternation).
spotlight.The daily running Extinction five days, dom The order. ments given training began animals were The were order same given to a subject of animals was determined on the following given eight basic procedure to a subject for incorrect day was applied turtles ness. Finally and nine latency were twenty for each measure tion for three In Fig.1, sponses except .rapidly than quisition were (2,23)=2.59】nor Groups Blocks effect daily 6.00 and the 50%random and were lasted for
run in ran-
that no reinforce-
no corrections number of immobility analyses;eight were were
derived is plotted CR group Rd, in terms groups, to illgroup
in terms increment choice in extinc-
of mean correct on the 10th day of acrespectively. that neither group interaction[F(18,207)<1】was group level of performance after 50 acquisition re-
more
An analy-
difference[F
significant,with
significant【F(9,207)=34.70,ρ<.01】.
the same of
measure.
the discrimination responses and SA showed and from learned ofcorrect due for the CR not any differences in acquisition groups The however, by Trial Blocks being except of course, because are those for the CR, measures, Wise&Gal且agher(1964)reported group here blocks. Mean there decrement reported groups. for repeated and, to statistical Since of the three 7.13,6.67,and sis of variance only Trial results two as in acquisition of experiment subjected for the significant of eight-trial the other were and Rd groups. the.performance on each training ITI. They and DISCUSSION
in the course six turtles groups,the 1-min responses.
removed of the SA with responses III. RESULTS Four after acquisition trials perday for correct randomly。
between trials eventhough the continuous
the・continuous
46
M.IsHIDA group learned formance the discrimination of the present However, the present 89%in the CR correct at the end experiment three to 75%in situation have contrast to small, more rapidly is very groups faster amount reward steak. asymptotic their number fell into a chance the CR with is str{king Analy・i・
tween the CR with tinction, the CR Such and differed rapidly 1961),On study is compatible long hand, the
the PREE, ITI(i.e,,24-hr)unlike those failed a significant Higher outcome found difference resistance in the studies between to extinction with the SA rats(Tyler, days of ex-
For the fifth day of ex。
demonstrated the SA(g=2.44)
the
the PREE beyond
Wise&Gallagher,1964)and
and Murillo et al.,
by Eskin&Bitterman(1961),
several factor of intertrial interval(ITI), the so-called schedule used study ITI supports groups than fac-
for the
as have
paradoxical runway 24-hr Rd possible responsible to find the PREE under in a Rd that
trials in the random
likely and tests showed and the fourth The且atest to find the PREE groups,re一
of significance)be-
Among failed groups
Overall
that of Wise&Gal-
those most the effect. partial the SA(g=3.90).Thus, and with groups,
study,unlike which to be the length the studies SA 50 extinction reviewed because de-
p<.01】,
not from than et al.,1961 out, and and Rd post groups. Pert&Bitterman(1970). in turt且es seems SA in Ishida&Kitamura,1988 pointed re-
group ・t th・ ・e w・ ・e・ig・ifi・ant
the third, partial during is inconsistent correct of CR for 5%level of correct by Muzio,Segura&Papini(1992)who including of extinction. common two of mean in mammals. Tukey's otherstudies response and of PREE found of
about effect【F(4,92)=42.10, in the present in Murillo this study (lshida&Papini,1993)which We made in the present
for the two found the Rd(g=6.34)but number Wise&Gallagher(1964)once effects used of Rd reveal・d中
for the second, with those and 5.98 for the CR, between from troostii the other suggested 100%
large,in
no information subjects PREE mea・u・es significantly of responses・(runway occurrence showed rats in a runway
in diameter, food of the curves the typical no sign of extinction Gonzales&Bitterman(1962), tors groups more extinguished species(Pseudemy∫scripta the type from
receiving in terms than va且ue of q=3.43 significantly reported group.
the reported Blocks differences significantly groups present than cm is plotted divergence follows f・r repeat・d two no significant lagher(1964)which partial with subjects extinction・The more differences(critical showed The for partial they groups trials are 4.33,5.29, and the other tinction, two partial which of studies interaction[F(8,92)=2.91,ρ<.01】. significant and the Rd levels(ranged in their study larger during responses ・f…iance by Days were per-
Wise&Gallagher'study.
a pellet,0.7 because effect【F(2,23)=1∼.45,ρ<'.01】,Trial there acquisition Hill&Wallace,1967;Wagner,1961).
of the three 40 extinction and Groups those The asymptotic A number much daily blocks level earlier. group during ・pectively. seems low speeds this conjecture of correct and Groups reward of eight-trial and means This the performance on・each creased to that.from of reinforcement(e.g., verify group. itself.
In Fig.2, sponse groups. turtles we cannot weight with in both the random relatively the SA)compared of acquisition demonstrated round study,'but than similar reached Wise&Gallagher(1964)reinforced ground and Y. NAKATA
reward
extremely with a
turtles
this view。
at the final daily b!ock
in a SA Wortz&Bitterman,1953;Capaldi&
schedule is a
Massed Partial Reinforcement of a Discriminative Response 47
in the Turtle
8
gり山 ω Z O ﹂ω 四国
7
6
5
\ \
4
トO山国匡O O
.
3
●一 ・
●Rd
2
σ一 ・SSA
1
HCR
0
1 2 3 DAILY BLOCKS
4.5
Figure has explained Rtransitions during and N-length partial groups affecting because the difference over tion than the rat literatures ties, yielded the PREE perhaps gested and unclear because behavior actually the latency of inaccuracy led to higher the extinction could process on the magni-
resistance were about to extinc-
suggested principles not available in measurement,藍atencies especially be-
at the last day of ac-
which explanation were strength
the difference influenced that the same scores as N-
in.Fig.2.
variations, to test the hypothesis such of associative to have schedule for
of 5 daily blocks
variables to discount as can be seen we can offer no satisfactory Although in interpreting insignificant)seems did. Since designed results, the sequential of the Rd to the SA of extinction, in a choice reinforcement in turtles. formation days on each responses extinction.
it is necessary performance was whole reinforcement consistent of correct the generalization.decrement superior tude Inconsistent by assuming However, quisition(though of resistance during properly extinction(Capaldi,1971). tween number three groups OF 8 TRIALS
Minkoff,1967)and 2. Mean apply to tur-
mechanisms of
in the present have by
provided of a discrimination study
useful habit, in-
as sug-
by Mackintosh&Holgate(1968).
REFERENCES
11]Bitterman, M. E.(1965)Phyletic t2】Bitterman, M. E.(1975)The [3]Bush, R. R.,&Mosteller, differences comparative in learning.Amerゴcan analysis of】earning. F.(1951)Amathematical model Psychologist,20,396-410.
Science,188,699-709.
for simple learning. P∫ychological
Reレiew,58,313-323.
[4】Capaldi, H.R. [5]Capaldi, E. J.(1971)Memory and learning:Asequential James(Eds.),/lnimal E.J.,&Minkoff, Memory.New R(1967)Reward viewpoint. York:Academic schedule In W. K. Honig and P.
Press.
effects at a re[atively long intertrial in-
terval. Psychono〃1fc∫cience,6,229-230.
【6】Cotton, J. w., Lewis, D. J.,&Jensen, G. D.(1959)Partial reinforcement effects in a
48
M.ISHIDA T-maze.ノournal(ガ(わ
[7]Couvillon, 用 ρσrα∫'レ
θ(隻P妙
P. A.(1984)Performance rewarded [8】Eskin, and Y, NAKATA
and nonrewarded ∫iological Psychology,52,730-733,
of goldfish(Cara∬'㍑
trials. Journal E. J.,&Bitterman, ∫αμrαf乙`∫)in patterned of Co〃sparative M. E(1961)Partial sequences of
Psychology,98,333-344.
reinforcement in the turt且e.(2μ
αr'εrか ノournal
(ゾExpe「'"ien'α'P∫ycholog:ソ,13,112-116.
【9】Gonzalez, R. C.,&Bitterman, turtle.(∼uarterly, [10】Hill,W. F.,&Wal且ace, factors μ1]Hul且, M. E.(1962)Afurther Journal(ゾEκ
W. P.(1967)Reward in extinction. Psychonom'c C. L(1943)Principles 【12】Ishida, tling response 6ゾ(弛
reinforcement in the
magnitude and number of training trials as joint
Science,7,267-268.
ofbehavior. M.,Couvillon, study of partial ρ8rfη28η'α'Psychology,14,109-112.
New York:Appleton-Century-Crofts.
P. A.,&Bitterman,M. in honeybees(Api∫
E.(1992)Acquisition 〃ielliferのas a function and extinction of a shut-
of the probability of reward. Journal
and its extinction in the turtle, Geoc-
〃47αrα ∫1vεP∫ychology,106,262-269.
【13]Ishida, M.,&Kitamura, le〃rys reevesii.ルfε
[14】Ishida, G.(1988)Partial η70'r∫ ρ1'Osaka M.,&Papini,M. R.(1993)Spaced (Geoclemy∫reevesの. 【15】Ishida, Memoirs M.,&Yoshida, 【17]Lewis, performance in the turtle
effect in a shock-avoidance プAnima'禽
reinforcement:Aselective situation
ンchology,33,43-47.
review of the literature since 1950,
Bulletin,7,1。28.
J. W.(1958)Partial プCompara∫'ve 【18】McFarland, of extinction.ハ
[19】Mackintosh, B.(1967)Frustration V.(1968)Effects shifts.ノourna!of N. R.,Diercks, ∫cripta troo∫'"in and non-response acquisition.
tolerance and incidental learning as
ん瑚7θ,215,786-787.
N.J.,&Holgate, nonreversal reinforcement and Phy∫'olog'ca'P∫},cho'ogy,51,251-254.
D. J。,&McGonigle, determinants 【20】Muri且to, and Instrumental 置
ぜ こ加 ル8∬ 妙, Ser.IV,41,153-161.
Japanese).Annua'(∼
D. J.,&Cotton, ノournal(∼
Kyo'た
S.(1983)[Partia且reinforcement D. J.(1960)Partial Psychological ノ, Ser・IV,37,29-34
training of O∫aka in the toad,Xenopus'aevi∫.」(in [16】Lewis, reinforcement 1くyoiku{ノniver∫1リ
of inconsistent Experi〃zental reinforcement on reversal and
Psychology,76,154-159・
J. K.,&Capaldi, E. J.(1961)Performance a partial-reinforcement of the.turtle,角
α`ゴθ1ηy∫
situation.ノournal(ゾComparatiレ8(隻P加
∫iological
Psychology,54,204-206.
[21]Muzio, R. N.,Segura, reinforcement ing andルlotiレ
【22】Papini,M. 【23】Papini,M. M. R.(1992)The acquisition effect of schedule and magnitude in the toad, arenarum.」Learn。
and extinction Bufo of
αご'on,23,406-429.
R,,&Bitterman,M. 6ゾCo〃ipara"レe E(1991)Appetitive conditioning in Octopu∫(yanea.ノ
α`r朋'
Psychology,105,107-114.'
R.,Mustaca, consummatory A. E,&Bitterman,M. responding 【24】Pavlik, W. Journal of Experimental 125】Pert, E. T.,&Papini, on instrumental of didelphid E.(1988)Successive marsupials.、
B.,&Lehr,・R.(1967)Strength 肋imal of alternative Learning(4c responses negative contrast in the
Behaレfor,16,53-57.
and subsequent choices.
Psychology,74,562-573.
A.,&Bitterman, M. E.(1970)Reward and learning in the turtle. Learning&Mo'1レ
α一
tion,1,121-128.
【26]Robbins, D.(1971)Partial 1960.Psychological [27]Romer, reinforcement:Ase且ective A. S。(1966)The [28】Spear, review of the al[eyway literature since
Bu〃etin,76,415-431.
ve〃ebrate N. E,&Pavlik, W. effects in a T-maze:Between pa'eontology. University B.(1966)Percentage and within of Chicago of reinforcement subjects.ノ
∂umal Press。
and reward of Experimental magnitude
P∫ychology,71,
521-528.
[29】Suther且and, (ゾExperimental 【30]Tyler, N. S.(1966)Partial D. W.,Wortz, partial reinforcement 66 ,
57-65.
reinforcement and breadth of of且earning. Quarterly/ournal
Psycholog:ソ,18,289-302.
E. C,&Bitterman, on resistance M. E.(1953)The to extinction effect of random in the rat・ ・4加erican/ournal and a且ternating
of Psychology・
Massed Partial Reinforcement 【31】Wagner, A. R.(1961)Effects quisition trials on conditioning of a Discriminative of amount Response and percentage and extinction. Jo麗rnal 49
in the Turtle
of reinforcement of Experi〃1ental and number of ac・
Psychology,62,234-
242.
【32】Weinstock, S.(1954)Resistance forcement under widely to extinction spaced trials.ノournal of a running response o/Comparative側d following partial rein-
Phy∫iological Psychology,
47,318-322.
【33】Weinstock, S.(1958)Acquisition and extinction of a partially reinforced 24-hr. intertrial interval.Jo配rna'6ゾ'Comparative and Phy∫iological 【34]Wise, L. M.,&Gallagher, reinforcement the turtle.」
α 〃na!of D. P.(1964)Partial Co川paratル8(4c running response at a
Psychology,56,151-159.
of a discriminative response in
Phy∫iolog'caムP∫ycho'ogy,57,311-313.
カ メにおける弁別反応の部分強化 が消去に及ぼす効果
集中試行習得訓練事態での検討
石 田 雅 人 ・中 田 靖 子
3群 の ク サ ガ メ がY型 迷 路 を 用 い た 白 黒 弁 別 課 題 で 訓 練 を 受 け た 。3群
替(SA),50%ラ
ン ダム(Rd),そ
け た 。 習 得 訓 練 は1日.8試
正 反 応 数 で 分 析 し た 結 果,習
とSAの 両 群 はCR群
して 連 続 強 化(CR)の
去 は 同 条 件 で5日
問 に 有 意 差 は な か っ た が,消
間行 なわ れた。
去 で は そ れ が 認 め られ,Rd
よ り も有 意 に 正 反 応 数 が 多 か っ た 。 た だ し部 分 強 化 両 群(RdとSA)問
な か っ た 。 こ れ は 明 ら か に 部 分 強 化 消 去 効 果(PREE)の
一交
各 群 で あ り,こ れ ら は 訓 練 の の ち 消 去 を受
行 の 集 中 試 行 で10日 間 行 な わ れ,消
得 で は3群
は そ れ ぞ れ50%単
に は差が
生 起 を示 す も の で あ る。 こ の 結 果 は カ メ
を使 っ た 他 の 研 究 及 び ラ ッ トを 用 い た 研 究 と比 較 検 討 さ れ,そ
の 上 で カ メ のPREE生
起 の決 定 因
と して 試 行 間 隔 の 長 さが 示 唆 され た 。 さ ら に 選 択 事 態 を用 い た 結 果 を 一 貫 性 を も って 説 明 す る た
め の 方 策 と して 反 応 測 度 の 問 題 が 論 じ ら れ た 。
キ ー ワ ー ド カ メ,部
分 強 化 消 去 効 果,弁
別