Cognitive Processes G F H D E C B A Figure 14.4 An apparatus employed by Tolman, Ritchie, and Kalish (1946) In preliminary training, rats were reinforced for running along the path A-B-C-D-E-F-G to H Another of the various procedures employed to determine whether rats learn specific responses or more general spatial information involved the maze shown in Figure 14.4 Path AB was the starting path The rats ran along the paths B-C-DE-F-G H was a dim light Figure 14.5 shows how the rats were tested Path AD was blocked Path led to the original goal and the dim light, H If the animals were learning a specific response, then presumably they would begin by going left, selecting paths 10 to 18 If they were learning to go to a 11 10 12 BLOCK provides a somewhat different definition of a cognitive map, in that it is evidenced by any orientation based upon computing distance Others suggest that the concept of a cognitive map is not necessary to explain spatial learning According to this view, animals acquire a set of memories of local views of the environment associated with the particular movements that take them from one place to another (e.g., Leonard & McNaughton, 1990) As much as any topic, the ability of animals to go from one place to another brings together the topics of instinct and cognition Consider the indigo bunting, a bird studied by Emlen (1970) The bird migrates over great distances Yet, although migration is a species-typical behavior, specific migratory routes are learned by the bird by its exposure to the star pattern in the sky In what follows it is possible to describe only a few examples of the many procedures that have been used to study map like learning in various species of animals In going from one place to another, animals learn to make specific responses or they acquire more general, cognitively informed spatial information? An early and hotly contested attempt to resolve this issue involved what came to be known as the place versus response controversy This may be illustrated by considering the two cross mazes shown in Figure 14.3 Two groups of rats might be used, both being trained to traverse the maze from each of two different start locations, S1 and S2 The difference is this: The group on the left is rewarded for making a specific response (turning left), going from S1 to G1 and from S2 to G2; the group on the right is rewarded for going to a specific place, from S1 to G1 and from S2 to G1 Thus the response group is rewarded for going to two different places whereas the place group is rewarded for making two different responses, left (S1–G1) and right (S2–G1) As a review (Restle, 1957) of the extensive literature in this area indicated, rats learn to both If trained in a visually rich, well-illuminated environment, the place group is superior to the response group In an impoverished, dimly illuminated environment the response group is superior to the place group 413 H 13 14 15 16 S2 S2 17 18 G1 G2 S1 G1 G2 S1 Figure 14.3 In the cross maze on the left, arrows indicate that animals starting at S1 are reinforced for going to G1 and animals starting at S2 are reinforced for going to G2, thus learning a left-turn response In the cross maze on the right, animals starting at either S1 or S2 are reinforced for going to G1 and thus learn to go to a place (a place response) A Figure 14.5 The apparatus employed in the test phase of the Tolman et al (1946) experiment The most frequently selected path was Path (see Figure 14.6), indicating that the rats learned to go to a place rather than learn a specific response 414 Animal Memory and Cognition 20 A 36% 150 m Hive 160 m 16 B Figure 14.7 The figure shows that bees trained to go from the hive to Place A when moved to a new location (Place B) go from B to A, rather than from B to the hive to A RATS 12 17% 9% 12 % 12 % 6% 4% 4% 4% 2% 12 2% 2% 11 10 PATHS Figure 14.6 The distribution of choices in the test phase of the Tolman et al (1946) experiment specific place (H) then they would select a novel initial response (go to the right) along paths to Of 56 rats that were tested, failed to respond on the test trials Of the remaining 53 rats, 36% selected Path 5; the distribution of choices for the remaining 34 rats is shown in Figure 14.6 If choices from paths to are regarded as novel (right turn instead of a left turn), then more than 87% of the rats selected a novel route different from that learned in initial training This ability to use a novel route to a goal has been examined in bees (J L Gould, 1986) Bees are first trained to go from the hive to Place A located to the west of the hive, as shown in Figure 14.7 Following this training, bees are caught and removed in an opaque box to a new spatial location to the south, Place B Place A is not visible from Place B and the bees have never flown from B to A What will the bees do? Will they return to the hive and then go from the hive to A? Or will they fly directly from B to A? Most bees fly directly from Place B to Place A This behavior, Gould suggests, indicates that the bee is using a map-like representation of its environment Since bees and rats can employ novel routes to reach a goal, it is perhaps not surprising that the chimpanzee can the same In an interesting experiment by Menzel (1973), a chimpanzee was carried on the shoulders of one experimenter around a large area (4,000 m2) while a second experimenter hid food in each of 18 randomly selected locations The chimpanzees, when released, were highly successful in locating the hidden food, retrieving an average 13 of the 18 items In so retrieving the hidden items, the animals neither followed the routes along which they had been taken, nor responded haphazardly and at random Rather, they followed efficient routes that minimized the distance to the items Another indication that chimpanzees can form a map-like representation of the environment comes from a study by Menzel, Premack, and Woodruff (1978) In that study, chimpanzees were shown an impoverished TV picture of where food was hidden in a rather complicated area containing trees, hills, and large objects They were more successful in reaching the goal than were control animals given no information Thus chimpanzees can match the information provided on a flat TV screen to spatial locations in the three-dimensional environment Language Learning in Animals It is quite clear that animals communicate in the sense that certain behaviors in one animal are capable of producing specific and predictable behaviors in another animal Thus, as indicated, a foraging bee’s dance when it returns to the hive indicates to the bee’s conspecifics the location of food (Von Frisch, 1950, 1967) As another example, vervet monkeys can communicate to their conspecifics whether an observed predator is a leopard, a snake, or an eagle (Seyfarth & Cheney, 1990) In the case of the vervet, it is quite clear that some degree of learning is involved For example, young, inexperienced vervets may mistakenly give the eagle call on spying a nondangerous bird Hockett (1960) identified 13 characteristics of human language that can be presented in any system of communication These are shown in Table 14.1 Cognitive Processes TABLE 14.1 Number 10 11 12 13 The Thirteen Design Features of Language Design Feature Vocal auditory channel Broadcast transmission and directional reception Rapid fading (transitoriness) Interchangeability (a speaker can reproduce any linguistic message he or she can understand) Total feedback (the speaker of a language hears everything of linguistic relevance in what he or she says) Specialization (sound waves of speech serve only as signals) Semanticity Arbitrariness Discreteness Displacement Productivity Traditional transmission Duality of patterning Source: Hockett, 1960 Hockett (1960) suggests that characteristics through (vocal auditory channel, broadcast transmission and directional reception, rapid fading [transitoriness], interchangeability, and total feedback) are common to a variety of animal species, including some birds and many mammals Characteristics through (specializing, semanticity, arbitrariness) are to be found in primates Displacement, the ability to consider things that are remote in either space or time, can be found only in humans Other particularly human characteristics are productivity (the ability to say novel things), duality of patterning (which refers to making words from phonemes), and arbitrariness (the fact that words not resemble the objects they signify in any physical sense) Although animals are capable of communication, that they are capable of acquiring language is another matter, and one fraught with continuing controversy On the one hand, there can be little doubt that many of the investigators of ape language, such as the Rumbaughs (e.g., Rumbaugh, 1977) and the Gardners (e.g., 1969), are quite convinced that chimpanzees possess the capacity to acquire language or some aspects of language On the other hand, other investigators of ape language such as Terrace (1979) are equally convinced that no such thing has been demonstrated To complete the picture, importantly, many linguists and other language experts are strongly of the opinion that no animal has come even close to demonstrating the sort of language ability displayed by humans (see, e.g., Healy, 1980; Pinker, 1994) They point out, among other things, that although children acquire complex language skills practically effortlessly, heroics and extensive training efforts are required to get chimpanzees to master relatively simple skills that may, at best, approximate language Some early unsuccessful attempts to teach language to animals involved raising chimpanzees or other primates at 415 home and tutoring them in spoken English (C Hayes, 1951; K J Hayes & Hayes, 1952) It came to be realized that the chimpanzees’ vocal apparatus is not designed for speaking Beatrice and Allen Gardner (1969) overcame the difficulty by teaching sign language to an infant, female chimpanzee named Washoe After 51 months of training, Washoe had acquired 122 signals It was asserted by the Gardners that Washoe could combine signals into phrases of some two to four items An often-cited example is that Washoe gave two signs on seeing a swan: the sign for water followed by the sign for bird Another approach was to get Sarah, a now famous chimpanzee, to place symbols for objects on a board (Premack & Premack, 1983) Sarah was then thought to write sentences on the board For example, Sara was given an apple if she placed on the board the symbols for give and apple Still another approach provided chimpanzees with keys on a computer exhibiting lexigrams The lexigrams were different geometric patterns, each of which represented something such as a request (please) or an item (banana) Lana, one of the chimpanzees employed in this research, could request items by pressing the lexigrams in a particular order: please machine give banana Terrace (1979) trained a young chimpanzee named Nim Chimpsky (a play on the name of the famous linguist Noam Chomsky) Nim was taught sign language As a result of his research, Terrace concluded that there was no evidence suggesting that Nim had learned language Nim, it was concluded, was either imitating his trainer or was merely exhibiting a rather straightforward form of serial learning Terrace’s criticisms brought into question any attempt to teach language to an animal that evolved stringing together items, such as gestures or lexigrams, because these could easily be interpreted as forms of serial learning based on learning simple associations Following Terrace’s telling criticisms, a new tactic was adopted by the Rumbaughs (e.g., Savage-Rumbaugh et al., 1993) Instead of attempting only to get animals to reproduce items such as signals or gestures, they also required the animals to comprehend items For example, Kanzi, a pigmy chimpanzee (or bonobo), might be told “Kanzi go out to the hall and get the ball.” By teaching Kanzi many words, presenting him with many sentences, and issuing commands to him using different word orders (e.g., Savage-Rumbaugh et al., 1993), the researchers concluded that Kanzi had acquired language in a meaningful sense Indeed, in terms of sentence compression Kanzi was said to rival a child, Alia, up until the time she became years old, after which age Alia increasingly surpassed Kanzi The approach involving the ability to comprehend language has also been employed with the dolphin (e.g., Herman, 1986) 416 Animal Memory and Cognition If children are more or less totally deprived of language experience when quite young, they exhibit serious deficiencies in language use later in life (Bickerton, 1998; Candland, 1993) On the other hand, if children are provided with minimum language experienced they use language normally later in life Indeed, as Bickerton (1998) has described, children deprived of normal language manage on their own to create a fully functional language in a single generation This has occurred, for example, when people who speak many different languages are thrown together (as on plantations) and communicate by means of what is called a pidgin Pidgins are very simple ways of communicating that consist of two to three words Children exposed to a pidgin, in the absence of explicit instruction, create a much more sophisticated language called a Creole A similar thing happens when deaf children are exposed to parents or others who may use sign language badly These facts and others (see Pinker, 1994) have caused language experts to suggest that the language gulf between humans and other animals is a matter of kind and not of degree From an evolutionary standpoint, there is no particular reason that our closest relative, the chimpanzee, should possess even a rudimentary language ability Humans and chimpanzees separated from each other—that is, no longer shared a common ancestor—about to million years ago Language ability may have evolved in the human line after we separated from the chimpanzee Perhaps species more closely related to ours, such as the extinct Homo erectus, possessed language, or at least something approximating human language Whatever the truth may be, the language gulf between humans and extant animals such as the chimpanzee is very large and may be one of kind, not degree Pidgins, as indicated, are very simple ways of communicating and consist of a few words being strung together like beads on a string Bickerton (1998) has suggested that animals such as chimpanzees and bonobos are capable, at most, of acquiring a pidgin According to Bickerton, children under years of age use a pidgin, which is acquired on the basis of general intellectual capacity At above years of age, children begin to acquire language on the basis of a mechanism specifically shaped by evolution for this purpose, and often called a language acquisition device In Bickerton’s view, the language capacities of both mature apes and children under years of age are on a par and not represent true language Interestingly, as indicated before, Kanzi, a bonobo, was able to follow verbal commands as well as Alia, a child, up until Alia was years of age, whereupon Alia increasingly surpassed Kanzi in comprehension Experts are not agreed on many important aspects of language that may bear upon animal abilities along this line To consider merely one line of disagreement, Pinker (1994) is of the opinion that language evolved slowly and in stages over the 350,000 or so generations separating our species from the chimpanzee Bickerton (1998) believes that language evolved in two stages, a pidgin stage followed by a Creole-type stage Pinker’s view allows that we might find in animals some intermediate language stage(s) between that of a pidgin and a Creole Bickerton’s view does not allow this and suggests we will never observe in animals a form of language more complicated than a pidgin Both Pinker and Bickerton think that language is a hardwired capacity of our species, independent of general intelligence If, as some (e.g., S J Gould, 1987) think, language is a result of an increase of intelligence in humans, then it should be possible to find in animals increasingly better forms of language correlated with better intelligence Although these are interesting speculations, definite and secure knowledge concerning animal language ability lies in future research Still another view suggesting that language in humans is fundamentally different from that in animals was suggested by Dunbar (1993) Dunbar suggests that language evolved in three principle stages Australopithecus cines, an early ancestor, had a primate system of vocalization similar to that seen in present-day apes such as chimpanzees This could not be called language With Homo erectus, a close relative of our species, a simple kind of language evolved that prompted bonding between individuals About 50,000 or so years ago, our own species, Homo sapiens, developed fully modern language that was employed for more than social boding Essential language developed into a useful means for communicating abstract ideas Thus Dunbar, like Bickerton and Pinker, is of the opinion that as far as language is concerned, the gulf between man and other animas is one of kind and not degree Evolution and Cognition Several characteristics of the approach to animal and human cognition currently employed within experimental psychology are noteworthy Workers in human cognition all but ignore animal learning and cognition Workers in animal cognition, for the most part, employ a limited number of species, most notably rats and pigeons The cognitive capacities isolated for examination in animal cognition are often those similar to those possessed in abundance by humans Do animals possess numerical abilities? Can animals make inferences of the sort If A = B and B = C, then A = C? To what extent human theories of serial learning apply to animals? Are animals, like humans, capable of self-recognition? By virtue of examining these and other questions much useful information has been provided for better understanding both animal Cognitive Processes and human cognition Furthermore, there is ample reason for believing that at least in some cases similar processes are in operation over a broad range of animal species The prime example of this is the sort of associative learning examined in Pavlovian and instrumental conditioning It is probably the case that most (if not all) experimental psychology and perhaps most social scientists accept evolution This being so, Symons (1987) wrote a paper entitled “If we’re Darwinians, what’s the fuss about?” The fuss, according to Symons, is about this: “Perhaps the central issue of psychology is whether the mechanisms of the mind are few, general and simple, on the one hand, or numerous, specific and complex on the other” (p 126) As indicated previously, and speaking generally, many experimental psychologists tend toward the former assumptions whereas many evolutionary biologists and psychologist tend toward the latter, and some incline toward the view that the modules are in communication with each other (Mithen, 1996) To employ metaphors, experimental psychologists might compare the mind to a general-purpose computer, whereas evolutionary biologists and psychologists often compare the mind to a Swiss army knife, a general-purpose tool having a diversity of different functions Evolution and Cognition: Implications At one time, unlike now, animal and human learning and cognition were seen as highly related (see, e.g., McGeoch & Irion, 1952) Workers in the two areas shared a variety of important assumptions and considered, very probably, that choosing animals or humans for study involved little more than a strategic decision If an attitude of common purpose was once the rule of the day, then it appears to be no more On the one hand, many workers in human learning and cognition not appear to regard animal learning and cognition as especially relevant to their concerns It may be that many workers in human learning and cognition see the intellectual gap between our species and others to be very wide, so wide as to be one of kind and not of degree According to this view, animal cognition has little to offer human cognition Interestingly, and perhaps ironically, many workers in animal learning and cognition may hold a contrary opinion, that is, that humans and animals differ cognitively, but only in degree This inference seems to make sense when one examines some of the concerns popular in the study of animal cognition As we have seen, workers in animal cognition are interested in determining the extent to which animals possess language, can use numerical information, imitate others, and so on What these concerns have in common, of course, is that they are things that are done by humans and done very 417 well Needless to say, perhaps, these are legitimate problems, ones well worth studying The attitudes of workers in animal and human cognition may stem from the same source: an incorrect understanding of how evolution shapes cognitive adaptations Both camps appear to accept some version of a continuity view, that as we go from so-called “lower” to “higher” animals, intellectual capacity increases Workers in human learning and cognition, as indicated, appear to believe for whatever reason (e.g., that intermediate species became extinct) that human intellectual capacity, at least in some significant respects, differs from that of animals in kind and therefore that animal cognition may be ignored Workers in animal learning and cognition also accept some version of the continuity view but appear to believe, generally speaking, that intellectual development of various sorts has progressed far enough in animals to make them useful objects of study If the foregoing analysis is correct, then new more useful approaches to learning and cognition are being ignored by both animal and human workers There are certain implications that follow from an evolutionary approach embracing both discontinuity and continuity of intellectual development over species, which may well result in a closer, if not close, reuniting of animal and human learning and cognition On the one hand, although there may well be some capacities unique to humans (e.g., the capacity for language), there are undoubtedly others that are widely shared over species and that, accordingly, should benefit from a comparative approach A prime example here is spatial learning As we saw, spatial learning is well developed in many species, and in some species (including our own), males seem better at spatial learning than females This difference seems related to the different demands placed on males and females in the environment in which they evolved Thus, animals that had to move about a good deal, that had to go from place to place to find food or mates or to escape danger, would be expected to be better at spatial learning than animals that had fewer such demands Note the difference between the emphasis here and that previously described as shared by many workers in human and animal cognition: The emphasis is not on the degree of correspondence between the cognitive capacity of humans versus other animals, but on the relation between particular animals and the specific demands of the environment in which those animals evolved That is, to better understand the degree or kind of cognitive capacity an animal might possess, a first step might well involve a better acquaintance with the problems that animals faced in the environment in which it evolved In Pavlovian conditioning, as indicated, a tone may be presented a few seconds prior to food, and the relation between these events may be learned as indicated by the ... communicate by means of what is called a pidgin Pidgins are very simple ways of communicating that consist of two to three words Children exposed to a pidgin, in the absence of explicit instruction,... capable, at most, of acquiring a pidgin According to Bickerton, children under years of age use a pidgin, which is acquired on the basis of general intellectual capacity At above years of age, children... language on the basis of a mechanism specifically shaped by evolution for this purpose, and often called a language acquisition device In Bickerton’s view, the language capacities of both mature apes