ADVANCES IN THE STUDY OF BEHAVIOR V 17

Visit the link below to download the full version of this book:

https://medipdf.com/product/advances-in-the-study-of-behavior-v-17/

Click Download Now

 Advances in
THE STUDY OF
BEHAVIOR
Edited by
JAY S. ROSENBLATT
lnstitute of Animal Behavior
Rutgers University
Newark, New Jersey
COLINBEER
Institute of Animal Behavior
Rutgers University
Newark, New Jersey
MARIE-CLAIRE BUSNEL
UER Biomkdicale
Groupe Gknktique et Comportements
Facultk de Mkdecine Paris V
Paris, France
PETERJ. B. SLATER
Department of Zoology and Marine Biology
University of St. Andrews
Fife, Scotland

VOLUME 17
1987

ACADEMIC PRESS, INC.

Harcourt Brace Jovanovich, Publishers
Orlando San Diego New York Austin
Boston London Sydney Tokyo Toronto
COPYRIGHT @ 1987 BY ACADEMIC PRESS, INC.
ALL RIGHTS RESERVED.
NO PART OF THIS PUBLICATION MAY BE REPRODUCED OR
TRANSMITTED IN ANY FORM OR BY ANY MEANS, ELECTRONIC
OR MECHANICAL, INCLUDING PHOTOCOPY. RECORDING, OR
ANY INFORMATION STORAGE AND RETRIEVAL SYSTEM, WITHOUT
PERMISSION IN WRITING FROM THE PUBLISHER.

ACADEMIC PRESS, INC.

Orlando, Florida 32887

United Kingdom Edition published by

ACADEMIC PRESS INC. (LONDON) LTD.
24-28 Oval Road. London NWI 7DX

LIBRARY
OF CONGRESS CATALOG CARD N U M B E R : 64-8031

ISBN 0-12-004517-6 (alk. paper)

PRINTED IN THE UNITED STATES OF AMERICA

a7 nn np 90 9 B 7 6 s 4 3 2 I
 Contents

Preface ................................................................... vii

Receptive Competencies of Language-Trained Animals

LOUIS M . HERMAN
I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
I1. Language Production versus Language Comprehension . . . . . . . . 2
111. Receptive Competencies of Language-Trained Apes . . . . . . . . . . . 4
IV . Receptive Competencies of Language-Trained Dolphins . . . . . . . . 16
V . Receptive Competencies in Sea Lions ...................... 48
VI . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
References ............................................. 55

Self-Generated Experience and the Development of Lateralized

Neurobehavioral Organization in Infants
GEORGE F . MICHEL
I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
I1. Some Biological Characteristics of Human Handedness . . . . . . . . 62
111. The Expression of Handedness during Infancy . . . . . . . . . . . . . . . 66
IV . Summary Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79

Behavioral Ecology: Theory into Practice

NEIL B . METCALFE AND PAT MONAGHAN

I Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
I1. Control of Foraging ..................................... 86
111. Control of Spacing and Social Behavior ..................... 102
IV . Control of Breeding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
V . Conclusions ............................................ 114
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
V
vi CONTENTS

The Dwarf Mongoose: A Study of Behavior and Social Structure

in Relation to Ecology in a Small, Social Carnivore
0. ANNE E. RASA

I. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12I
11. Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
111. Social Structure and Territory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
IV. Constraining Behavior Patterns and Compensatory Mechanisms . 125
V. Predator Pressure . . . . , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
VI. Discussion ............................................. 148
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 160

Ontogenetic Development of Behavior: The Cricket

Visual World
RAYMOND CAMPAN, GUY BEUGNON,
AND MICHEL LAMBIN

I. Some Theoretical Principles . . . . . . . . . . . . . . . . . . . . . ... . . .. . . 165

11. Psychophysiological Factors Involved in the Selection and Use of
Visual Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
111. Ecological Conditions and Organization of the Visual World . . . . 178
IV. Ontogeny of Behavioral Emergents . . . . . . . . . . . . . . . . . . . . . . , . . 183
V. Rules of Behavior Ontogeny . . . . . . . . . . . . . . . . . . , . . . . . . . . . . . 199
VI. Conclusion: The General Rules . . . . . . . . . . . . . . . . . . . . . . . . . . , . 206
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209

Index ..................................................................... 213

 Preface

The aim of Advances in the Study of Behavior is to serve the increasing

number of scientists who are engaged in the study of animal behavior by present-
ing their theoretical ideas and research to their colleagues and to those in neigh-
boring fields. Since its inception in 1965, this publication has not changed its aim
to serve ". . . as a contribution to the development of cooperation and commu-
nication among scientists in our field." We acknowledge that in the interim new
vigor has been given to traditional fields of animal behavior by their coalescence
with closely related fields and by the closer relationship that now exists between
those studying animal and human subjects. Scientists studying animal behavior
now range from ecologists to evolutionary biologists, geneticists, endo-
crinologists, ethologists, comparative and developmental psychobiologists, and
those doing research in the neurosciences. As the task of developing cooperation
and communication among scientists whose skills and concepts necessarily differ
in accordance with the diversity of phenomena that they study has become more
difficult, the need to do so has become greater. The editors and publisher of
Advances in the Study ofBehavior will continue to provide the means to meet this
need by publishing critical reviews, by inviting extended presentations of signifi-
cant research programs, by encouraging the writing of theoretical syntheses and
reformulations of persistent problems, and by highlighting especially penetrating
research that introduces important new concepts.

vii
This Page Intentionally Left Blank
 ADVANCES IN THE STUDY OF BEHAVIOR. VOL. 17

Receptive Competencies of
Language-Trained Animals

LOUISM. HERMAN
DEPARTMENT OF PSYCHOLOGY AND
KEWALO BASIN MARINE MAMMAL LABORATORY
UNIVERSITY OF HAWAII, HONOLULU, HAWAII 968 14

1. INTRODUCTION

Two major themes appear among the recent critiques of the attempts to teach
languages to great apes: (1) whether apes can use the symbols (words) of a
language as true surrogates for the objects and events they reference, including
objects and events remote in time or space, and (2) whether apes can construct or
understand sentences by using combinations of words as information. The first
theme, recently pursued most vigorously by Savage-Rumbaugh and colleagues
(e.g., Savage-Rumbaugh et al., 1980b, 1983), addresses the semantic aspect of
language-the meaning and reference of words and phrases. The second theme,
identified recently with Terrace and colleagues (e.g., Terrace et al., 1979),
addresses the syntactic aspect-the information that is added to the string of
words comprising a sentence by the structure of the string, such as the ordering
of the words. In human languages, it is most often necessary to account for both
the semantic and the syntactic components of a sentence to interpret a sentence
correctly. Semantics and syntax have been described as the “indispensable core
attributes of any human language” (Paivio and Begg, 1981, p. 25).
The contention of the Savage-Rumbaugh group is that the use of symbols by
apes to name objects or to make requests for foods or other desirables is not of
itself evidence that the symbols function as surrogates for objects and events, as
do words in human languages. Careful analysis of the majority of the work with
apes reveals that the symbols used are often constrained to limited contexts and
are principally a learned means for obtaining a desired outcome. Typically, the
words are not used to “refer” (cf. Terrace, 1985) but with special training may
eventually take on that function (Savage-Rumbaugh and Rumbaugh, 1978).
The syntactic issue, as raised in the review by Terrace ef al. (1979) (also see
Ristau and Robbins, 1982), asked whether the language produced by apes was
grammatical, showing evidence of the use of structure, such as word order, as
I
Copyright Q 1987 by Academic Press. Inc.
All rights of reproduction in m y form reserved.
2 LOUIS M. HERMAN

information. Or, as the question was put directly by Terrace et al. in the title of
their 1979 paper, “Can an Ape Create a Sentence?” It was acknowledged that
apes occasionally generated long strings of words but that successive words
added little or no information, being mainly repetitions, synonyms, or the like,
without elaboration or expansion of meaning. For example, Terrace er al. de-
scribed one sequence produced by Nim, the chimp they tutored in sign language.
To request an orange, Nim produced the following 16 signs: “give orange me
give eat orange me eat orange give me eat orange give me you.” Nim probably
did get the orange.
In other cases of apparently well-formed word sequences produced by apes
additional problems were cited by Terrace et al. (1979) and in other reviews
(e.g., Bronowski and Bellugi, 1970; Fodor et al., 1974; Petitto and Seidenberg,
1979; Ristau and Robbins, 1982; Savage-Rumbaugh et al., 1980a; Seidenberg
and Petitto, 1979, 1981; Terrace, 1979a; Terrace et al., 1981). These problems
included insufficiencies or deficiencies in the reporting of data; the presence of
context cues to guide responding, prompts, or other paralinguistic cues from the
trainers that may have led the ape into a prescribed series of signs; the probable
underreporting of examples that did not conform to a grammatical sequence; the
practice in some cases of deleting “extraneous” or redundant signs in reports of
the sequences produced by the apes; and the overinterpretation of a string of
signs or unique combination of words produced by the apes. Thus, when the
chimp Washoe signed “water bird” on seeing a swan in a lake for the first time,
did she create a novel combination of words and thereby name the swan, or was
she simply giving two independent signs “WATER-BIRD?’’ These crit-
icisms cast grave doubt on the evidence for sentence-processing abilities of apes
and, by implication, on all of the claimed linguistic skills of these animals.

11. LANGUAGE
PRODUCTION
VERSUS
LANGUAGE
COMPREHENSION

Work with apes has emphasized language production, rather than comprehen-
sion. Humans, of course, both produce language and understand it. As noted by
Savage-Rumbaugh and colleagues (Savage-Rumbaugh e l al., 1980a, 1983), by
Seidenberg and Petitto (1979), and by others, the unfortunate assumption was
made in much of the work with apes that production implied comprehension.
Premack, who studied the chimp Sarah and other chimps, using an artificial
language in which words were represented by plastic chips (Premack, 1971,
1976), recognized that production and comprehension may evolve as separate
systems, in child or chimpanzee. Although Premack carried out a number of
studies of comprehension, shortcomings in methodology and data reporting lim-
 RECEPTIVE COMPETENCIES OF LANGUAGE-TRAINED ANIMALS 3

ited the strength of the conclusions that could be drawn from the work (R. A.
Gardner and Gardner, 1978; Ristau and Robbins, 1982; Savage-Rumbaugh er
al., 1980a;Terrace, 1979a). Premack’s work on receptive competencies, as well
as that of other researchers teaching languages to apes, is reviewed in Section 111.
There is evidence for an asymmetry of comprehension and production in the
development and maintenance of language in humans. During early childhood,
comprehension generally precedes and exceeds production (Fraser et al., 1963;
Ingram, 1974), although there may be exceptions (Chapman, 1974; Chapman
and Miller, 1975; also see the review in Bloom, 1974). Production may involve
more complex processes than does comprehension (Bloom, 1974; Schiefelbusch,
1974), giving support to the notion that the two systems, though mutually depen-
dent, are somewhat separate.
Some separation of the systems may remain at adulthood, as suggested by the
specificity of some aphasic disorders. Difficulties in speaking (expressive a-
phasia) generally involve lesions in anterior areas-Broca’s area or other regions
of the frontal lobe-while difficulties in comprehension (receptive aphasia) in-
volve lesions in posterior areas, particularly Wernicke’s area (see the review in
Paivio and Begg, 1981, p. 367). Similarly, syntax and semantics may be sub-
served by different areas: syntactic functions by anterior structures and semantic
functions by posterior structures (Caramazza and Bemdt, 1978).
In studying children’s language, it has proven fruitful to analyze production
and comprehension separately when assessing linguistic competency (e.g.,
Fraser et al., 1963; Ingram, 1974). Tests of comprehension have been particu-
larly useful in the analysis of the grammatical competency of normally develop-
ing children (Chapman and Miller, 1975; Churchill, 1978; Strohner and Nelson,
1974) and in demonstrating language competencies in nonverbal or preverbal
children (Curtis, 1977; Ingram, 1974; Itard, 1932). Seidenberg and Petitto
(1979) commented that in view of the attention given to comprehension in work
with children it was puzzling that so little attention was given to comprehension
in work with apes.
Focus on receptive competencies of animals (or children) has several advan-
tages. The situation in which Comprehension is tested can be described, con-
trolled, and repeated, and objective, quantitative measures of performance can
be obtained. Given the extensive problems in method, assessment, and in-
terpretation of the results of the productive language studies with apes, it is
strategic at this juncture to emphasize receptive competencies. Theoretically, if
production and comprehension involve separate or only partially overlapping
systems in animals, then there may be inherent limitations in one system which
are not present in the other. Consequently, although it is imperative to study both
systems eventually, a current emphasis on comprehension provides a needed
balance to the extensive work on production and contributes toward a better
4 LOUIS M. HERMAN

understanding of the capabilities and limitations of animals in languagelike tasks.

In this article, I review what is known of the receptive competencies of
animals in languagelike tasks. The emphasis is on the extensive work with apes
as well as the recent work with marine mammals, including my own studies of
bottle-nosed dolphins.

111. RECEPTIVE
COMPETENCIES
OF LANGUAGE-TRAINED
APES

A. SIGNLANGUAGE
PROJECTS
Sign language projects (e.g., B. T. Gardner and Gardner, 1971, 1979; R. A.
Gardner and Gardner, 1969; Miles, 1983; Patterson, 1978a,b; Terrace, 1979b)
have given little attention to receptive competencies. Generally, data from these
studies are insufficient for judging the degree to which apes understand the signs
of their trainers. The Gardners (B. T. Gardner and Gardner, 1975), in fact,
disavowed the importance of receptive vocabulary, stating that for assessing the
linguistic ability of the chimp Washoe “our practice is to list items of expressive
vocabulary only” (p. 248). They go on to say that “to include items of receptive
vocabulary [in the list of expressive vocabulary] we would have to add several
score or several hundred additional items, depending on the standards of evi-
dence that one would impose on the data” (p. 248). Further, the Gardners take
exception to forced-choice tests of receptive competencies, claiming that “pro-
ductive tests are methodologically more sound” (p. 256). This claim runs coun-
ter to the extensive criticisms leveled against the methods of productive tests.
It is unfortunate that the Gardners did not test and document the receptive
vocabulary of Washoe or of other chimps they studied, as this might have made
for meaningful comparisons between productive and receptive abilities, enabled
the assessment of sentence understanding, and possibly blunted those criticisms
that tended to summarily dismiss their work as failing to provide convincing
evidence for any language competencies in apes. The only test of receptive
ability systematically applied to signing chimps by the Gardners was the ability
to answer wh- questions appropriately (B. T. Gardner and Gardner, 1975, 1979).
Responses of the chimps Washoe, Pili, Tatu, and Moja to questions such as Who
that? What you want? Where Susan? were tested. A limitation of these tests,
pointed out by Seidenberg and Petitto (1979, 1981), was the policy of scoring
replies as correct if they were from the appropriate grammatical category, such as
nouns as answers to what questions and locatives as answers to where questions.
This scoring criterion is adequate for assessing mastery of a question form
(Brown, 1968), but allows for semantically inappropriate answers. Since the
Gardners provided only examples of the replies of the chimps, it is difficult to
 RECEPTIVE COMPETENCIES OF LANGUAGE-TRAINED ANIMALS 5

judge the degree to which the chimps understood the content as well as the form
of the question. Understanding of form was itself quite variable within and
among chimps, ranging from a low of 12% correct categorizations to a high of
100%.
Despite the Gardners’ caveat against forced-choice receptive tests, Patterson
(Patterson, 1978b; Patterson and Linden, 1981) administered a standard chil-
dren’s forced-choice test of receptive skill (Assessment of Children’s Language
Comprehension Test) to the gorilla Koko. Koko’s task was to look at a card
containing multiple drawings representing objects, attributes, or relationships
among objects and point to a particular drawing in response to verbal or signed
words or both. The first 10 cards of the set tested understanding of single words
(vocabulary) and the remaining 30 tested understanding of phrases, for example,
“point to the bird above the house.” The phrases were at three levels of diffi-
culty, corresponding to the number of “critical elements” (two, three, or four)
referred to by the phrase (e.g., three in the above example: bird, above, and
house). Koko’s performance ranged from 72% correct responses for the 50
vocabulary items (five items per card) to 30-70% correct for phrases (chance
accounting for 20-25%, depending on the number of alternatives on the cards).
The combination of sign plus voice yielded somewhat better scores than did
either alone, and there were no differences between sign-only versus voice-only
testing. Patterson claimed that Koko performed significantly better than chance,
but summed binomial distribution tests that I applied to her data show that this
was true only for the vocabulary test and for three of the nine conditions using
phrases (three levels of critical elements X three replications: sign plus voice,
sign only, and voice only). An additional limitation on the findings is the replica-
tion over language conditions. Although Patterson provided no information on
how these replications were carried out, it appears that Koko went through the 30
phrase cards three times; therefore, her performance during the second and third
replications might have benefited from the previous replication(s). Terrace et al.
(1981) noted that the lack of detail in Patterson’s reports makes it difficult to
attribute Koko’s performance to true comprehension, as opposed to learning sets,
rote drilling, social cueing (which Patterson seems to have controlled for), or the
like. Petitto and Seidenberg (1979), in an article devoted primarily to an evalua-
tion of Patterson’s (1978a) study, did not comment on Patterson’s forced-choice
tests (which might not have been available to them at the time of publication).
However, these authors stressed that the simultaneous use of gestures and spoken
English during normal social interactions with Koko made it difficult to assess
Koko’s understanding of the primary language medium-gestural signs. They
also cautioned that Koko’s performance might have been guided by para-
linguistic cues such as body movements or pointing, a criticism that has been
leveled against all of the signing projects. In fairness to Patterson, her findings of
6 LOUIS M. HERMAN

equivalent performance on comprehension of single vocabulary items with either

gesture or voice cues are given increased credence by recent work by Savage-
Rumbaugh et al. (1989, who reported that the pygmy chimpanzee Kanzi under-
stands some spoken English words, though no specific training for this was given
(see Section III,C) (also see the related report by Fouts et al., 1976). Nev-
ertheless, Patterson’s work as a whole continues to suffer from a spareness of
objective assessment.
Fouts et al. (1976; described in Fouts, 1978) reported understanding by the
chimp Ally of signed commands to select one of five objects and deliver it to one
of five locations. Fouts (1978) stated that some of the commands were novel,
obtained by vocabulary substitutions at the object and location positions of the
sequence of signs. The use of imperatives for testing comprehension is an impor-
tant, objective procedure and was used in our work with dolphins (see Section
IV). Unfortunately, the report by Fouts (1978) failed to provide details of the
training and testing methods, or of the results, so that some of the reservations
noted for Patterson’s work apply here as well.
Fouts, like the Gardners, suggested that the receptive skills of the apes he
tutored were extensive. A similar claim was made for Nim, the ape tutored by
Terrace and colleagues (e.g., Terrace, 1979b). Seidenberg and Petitto (1981)
commented that while extensive receptive vocabularies were claimed for the
chimps Ally and Nim-approximately 130 Ameslan signs for Ally (Fouts er al.,
1978) and 199 signs for Nim (Terrace er al., 1985)-no formal tests were ever
reported.

B , ARTIFICIALLANGUAGE
PROJECTS
I. Premack Project
Projects by Premack (1971, 1976) and by Rumbaugh and associates (Rum-
baugh, 1977) with artificial languages focused to varying degrees on receptive
competency. The most extensive attempt to quantify sentence comprehension in
apes was that by Premack (1976). The chimps Sarah, Peony, and Elizabeth were
tutored in a system in which plastic symbols of arbitrary shape and color were
used to represent objects, properties, and actions within an artificial language.
Sentences could be constructed, by the experimenter or by the chimp, by arrang-
ing the symbols in a linear array on a board. In early tests, the chimps failed to
transfer spontaneously between production and comprehension. For example,
although the chimps could write “Give apple” or “Give banana” to request
those specific foods and also understood the sentences “Take ball” and “Take
block” in the receptive mode, they showed no immediate understanding of the
novel constructions “Take apple” or “Take banana.” After brief, specific
training in transfer from production to comprehension or vice versa, the apes
were able to pass new transfer tests.
 RECEPTIVE COMPETENCIES OF LANCUAGE-TRAINED ANIMALS 7

Premack also carried out extensive studies of the apes’ answers to questions.
Like similar work described for the Gardners, answers to questions are a measure
of receptive competency. In Premack’s studies, a symbol for an interrogative
particle was introduced as a straightforward extension of an already familiar
“same-different” paradigm-for example, A ? A , where the possible answers
are “same” or “different.” The interrogative particle continued to be used in
this form as well as in other question forms that were implicit wh- questions. For
example, “A ? A” was glossed by Premack as “What is A to A?”, while the
symbol string “A same ?” (for which the correct answer is A) was.glossed as
“What is A the same as?”. The task of the chimp was to substitute one of the
available symbols for the question symbol.
Using symbols meaning “same” or “different,” Premack generated addi-
tional Yes-No question forms; for example, “A same A ?” was glossed as “Is
A the same as A?” Symbols meaning Yes and No were available as answers. In
the earlier work, the items compared in the question need not have names (only
their physical identities were at issue). In later work, questions consisting solely
of linguistic elements were constructed, such as “? color-of apple,” a string
consisting of three symbols to be answered by choosing one of two available
color symbols. Another example is “Round ? ball,” to be answered by substitut-
ing the symbol for “color-of,” “shape-of,” or “name-of” for the question
symbol. Premack reported high levels of performance in response to most of
these types of questions.
Several grammatical categories of words were taught to the apes in the com-
prehension mode or by intermixing the production and comprehension modes.
These included negation (e.g., “No Sarah honey bread take”), the demon-
strative (e.g., “Sarah insert this” versus “Sarah insert that”), quantifiers
(“Sarah take some”), and the preposition “on” (“A on B”). Some structurally
novel sentences were also taught in the comprehension mode (e.g., “Sarah take
red dish,” as derived from the previously taught atomic sentences “Sarah take
red” and “Sarah take dish”). Nonetheless, these demonstrations, as well as
additional tests of comprehension (or production) skill presented by Premack,
have been criticized on a number of grounds that, in sum, moderate any strong
conclusions about receptive competencies. The Gardners (R. A. Gardner and
Gardner, 1978) commented on the sparsity of Trial-1 data in Premack’s reports
of transfer and argued that because chimps are capable of rapid intra-problem
learning, without Trial-1 data it is unclear whether or not there was any inter-
problem learning. The Gardners’ contention is supported, for example, by exam-
ination of Table 6.2 in Premack (1976, pp. 120-121). The table reports “noun
contrast” testing of the understanding of new combinations of object and action
terms. Trial-I results are available for only eight of 22 new combinations tested.
Terrace (1979a), as well as Savage-Rumbaugh et al. (1980a), provided exten-
sive critiques of Premack’s work that need not be repeated in detail here. In brief,
8 LOUIS M. HERMAN

however, contextual constraints present during Premack’s experiments could have

guided the apes’ responses. Problems of a given type were clustered within a
testing session, the number of alternative choices available for response was
usually limited to two, and in many cases the problems might have been solved by
nonlinguistic means. For example, the symbols for “Mary,” “give,” and
“Sarah” were used repeatedly within a session in imperative strings such as
“Sarah give Mary X.” Alternative, contrasting symbols were not available during
a session (e.g., alternative trainer names). Sarah, therefore, need not have made
any semantic interpretation of the symbols “Mary,” “give,” and “Sarah” and
could have simply responded to X as in an associative naming exercise by, for
example, giving an apple if X was “apple.”
Because of the attention given in this paper to competencies for processing
syntactic information, it is worthwhile to consider Premack’s evidence for this
ability in apes. This evidence occurs in several contexts in Premack’s work. One
involves the ability to demonstrate an understanding of the relationships among
nonadjacent words in ordered strings. For example, in the string “Sarah apple
pail banana dish insert,” an instruction to Sarah to insert the apple in the pail and
the banana in the dish, “insert” is not adjacent to “apple” or “banana” but
nevertheless directs the action to be taken toward these fruits. Sarah’s ability to
carry out such instructions suggests, therefore, that she can interpret the hier-
archical structure of these sentences. Again, however, there are limitations to
any conclusions since, for sentences of the type listed, the only objects available
to Sarah were the two listed in the sentence, and sentences of the stated type were
clustered within a training session. Therefore, if Sarah understood that the prob-
lem at hand was that of “insertion,” she need not attend to the word “insert.”
Also, as soon as she placed the first object in the correct receptacle, which would
reveal some understanding of those two words and their relationship, the further
operation on the remaining objects could be determined by default. That is, only
one object and one receptacle remain.
Syntactic processing demands also occurred in a prepositional form that per-
mitted the construction of semantically contrasting sentences of the form “A on
B” versus “B on A,” where A and B were names of objects or names of the
colors of cards. The task of the chimps Sarah, Peony, and Elizabeth was to place
the first-named object or first-named color on top of the second. Sarah learned
the task with a fixed set of colors to a level of 75% correct responses, and then
passed a single transfer test to a new pair of colors at the 80% level (eight of ten
correct responses). The other two apes worked with objects rather than colored
cards. Elizabeth required fairly extensive training to learn the problem, but then
passed a single transfer test (nine of ten correct responses). Similar results were
obtained for Peony. Later, Premack demonstrated that the chimps paid attention
to both the A and B names, by providing three alternatives rather than only two.
Performance declined to 70%, but was still well above chance.
Semantically contrasting sentences provide a powerful vehicle for testing syn-