As the titles of his works indicate, Munn's unit of psychological analysis expanded over the years. This expansion culminated in the 1971 work, which included chapters not only on the "Rise of Animal Intelligence" but also on "The Emergence of Man"; "Cultural Evolution"; and "The Shaping of Modern Minds."

This work was intended for a "general reader" audience and, as presented under the careful editorship of Leonard Carmichael (another eminent figure of the period), is clearly indicative of the best efforts and furthest reaches of that tradition.

"What I have attempted here is to provide a meaningful, yet not too detailed, synthesis of what can be told today about the evolution of mind from lower animals to man and, in man, from his beginnings until he reached his so-called 'civilized' status. This account embraces cultural as well as biological evolution, and it deals, among other things, with the shaping of modern minds" (Munn, 1971, preface, p. vii).

There are various rationales for selecting and posting the following "Rise of Animal Intelligence" chapter from Munn's (1971) work. One historical-disciplinary rationale is that the chapter provides concise descriptions of the primary empirical tools used by animal psychologists during the first three quarters of the 20th century including: delayed reaction, detour problems, double alternation tasks, and discrimination learning tasks. In the course of doing so, Munn summarizes the research findings produced by many of the names now associated with that research tradition including: Walter Hunter, Wolfgang Köhler, Robert M. Yerkes, Carl J. Warden, Keith & Cathy Hayes, Harry F. Harlow, Winthrop N. Kellogg, David Premack, and Jane Goodall.

A related theoretical rationale is that the chapter is representative of the 'real upper limit' of the intellectual tools used in that same tradition. More specifically, although the opening section begins with commentary referring to "unlearned and learned behavior" it becomes clear (successively), throughout the succeeding sections that Munn is struggling to distance himself theoretically from the old mental continuity doctrine (of Reflex, Instinct and Reason). This theoretical point is particularly evident in the section on "continuity and discontinuity" of mental evolution where Munn explicitly sides with the views of Henry Nissen (1951). With this endorsement, Munn sheds the mistakes of his own early to mid-career and takes his rightful place among a long tradition of relatively under-recognized 'emergentist or discontinuity' theorists including: G. H. Lewes; William James; John Dewey; C. Lloyd Morgan; Lev Vygotsky; Alex Novikoff; A. R. Luria, and A. N. Leontyev. As such, this chapter provides an instructive basis for comparison and contrast with the Activity Theory approach of A. N. Leontyev (1981) posted elsewhere (see "The problem of the origin of sensation" and "An outline of the evolution of the psyche").

The intellectual modesty of the chapter is also worthy of note. While Munn attempts to point toward some of the important rough interdisciplinary guideposts "bearing upon the general problem of mental evolution" (including the distinction between biological and cultural evolution), he makes no pretense to provide psychology with the full-formed requisite terminological or practical distinctions that will ultimately be adequate to that disciplinary task. Occasional lapses of terminological specificity are therefore encountered. For instance, Munn clearly overextends the term "symbolic processes" to cover the admittedly vast and varied differential capacities of animals to "reflect" upon (i.e., retain, hold in mind, or consider) past experiences. If we are to take to heart his eventual endorsement of quantitative and qualitative change in mental capacity, that over-stretching of an otherwise useful term must be considered an unfortunate vagary of this particular chapter. To the extent, however, that Munn was attempting to provide a general entrée level account, the chapter (and the 1971 work itself) is successful as far as it goes.

The ongoing disciplinary challenge for comparative psychology (in all of its varied manifestations and applications) is to redress the sometimes radical disconnect between appeal to empirical rigor and the inadequate theoretical tools typically being utilized. Munn's late-career effort is just one of many historiographic exemplars of how we might go about rising to that challenge.

Reference:

Munn. N.L. (1980). Being and Becoming. Adelaide Australia: Adelaide University Union Press.

Chapter Four

THE RISE OF ANIMAL INTELLIGENCE

In our earlier discussion of the animal mind we called attention to the role of unlearned behavior (tropistic, reflex, and instinctive) in the adaptations of animals to aspect of their environment. Tropisms are especially prevalent in insects, although they are also found in many other animals, even lower mammals such as the rat. All animals with a synaptic nervous system have [sensory-motor] reflexes. These relatively automatic responses of particular organs play an important part in [bodily] adjustment, even at the human level. Reflexes may be modified (conditioned). Pavlov and others have made the claim that all learned behavior has, as its base, the conditioning of reflexes. Instinctive behavior, often defined as an unlearned patterning of reflexes, is also evident in all animals having a synaptic nervous system, except possibly the higher primates. The presence of instincts in humans is seriously questioned, because learned behavior (habit) dominates so many aspects of human life.

Unlearned behavior is, for the most part, highly adaptive. [p. 117] It often suggests the presence of intelligence.... But this is "built-in" intelligence, if we can call it intelligence at all, and not an outcome of individual achievement. Moreover, the behavior is more stereotyped than variable, flexible, or versatile. It has evolved in relation to relatively fixed conditions of life, and environmental changes tend to disrupt it. Individually acquired intelligence, on the other hand, is characterized by versatility in adjusting to changed conditions....

Although there are undoubtedly unlearned aspects of mind (and intelligence), our best evidence of mental evolution comes from the ability of organisms to depart from unlearned modes of adjustment by adopting learned adaptations...

TESTS OF LEARNING ABILITY

In efforts to compare the learning ability of animals at different levels of evolution, psychologists have devised a large variety of tests. Some of the simpler ones are reversal of a tropism (as in the cockroach), modification of reflexes (as when an animal is conditioned so that light evokes salivation), inducing the animal to press a lever which releases food (as in operant conditioning), training the animal to [p. 118] discriminate (light versus dark, triangle versus circle, and so on), and maze training on paths varying in complexity from a single T or Y unit to a succession of such units with many blind alleys between entrance and exit.

The use of such tests as these has shown [unequivocally] that most of the animals tested are capable of learning. It is some of the simpler invertebrates whose learning capacity is in question. The behavior of many of these, even the paramecium, has been modified by experimental procedures designed to test for learning ability, but there is considerable controversy as to whether learning had been demonstrated, or whether the modifications are unlearned reactions to changed environmental conditions. [fn-1] Many vertebrates, representing fishes, amphibia, reptiles, birds, and mammals, have been successful in conditioning, discrimination, and maze learning tests. Generally speaking, there is more rapid learning (fewer trials and less time required) as we go from lower vertebrates to mammals. In discriminating learning and maze tests there are also fewer errors. Moreover, the complexity of what can be learned increases as the mammalian level is approached.

From lower mammals like the rat up to man we find that discrimination problems and mazes do not make very good tests of intelligence. They involve such elementary aspects of learning that they do not sufficiently differentiate between the more and less intelligent animals. In maze learning, for example, all the animal must do is make the proper turn, in response to the proper stimuli, at the proper time, and in the proper sequence. The reader may be surprised to learn that white rats can even best college students in this sort of learning; in fact, they have done so repeatedly. We know that college students are more intelligent than rats, so we are forced to conclude that while mazes are good tests of sheer modifiability, there is much about mental capacity that they do not reveal. [p. 119]

SYMBOLIC PROCESSES

With the coming of mammals something beyond sheer ability to be modified and to make adjustments is added to mental life. This has already been referred to as the symbolic process, a process enabling organisms to recall something previously experienced but now absent. The ability to learn is a necessary antecedent, as is the ability to retain what has been learned.

A symbol is something which represents or can serve as a substitute for an object, a situation, or an event. Such an association between a symbol and what it symbolizes must be learned, as when the child learns that such and such an object is a cat. Human beings learn to associate images and gestures, as well as words, with particular experiences. Then they use these substitutes to recall (or think of) the experiences after they have passed. Thus, a man may recall the wife he left at home, the women he might have married, or anything else that he has experienced. This ability to acquire symbols which represent past experience is the basis of all higher mental processes. Its emergence was a great forward leap in mental evolution.

The Delayed Reaction

Rudimentary aspects of the ability to recall past experiences are revealed in the so-called delayed reaction test. What we find out in a delayed reaction test is whether the animal can respond to, or in terms of, something within it, perhaps some modification of its brain, represents the absent stimulus. A necessary precaution in such experiments is to make certain that no present stimulus is guiding its response, for response to stimuli present at the time of response is not only easy for most animals but also no evidence that a symbolic [(i.e., 'representative, retentive, or memory')] process is involved.

In one test of delayed reaction the animal is placed in a restraining cage made of wire mesh. Facing this are three identical exit doors, each with its own electric grid and light bulb, [p. 120] as illustrated... In each test, only one bulb is lit. The light may appear in any one of the three positions, its location on a particular trial being determined by chance, so that the light, and not the position in which it appears, will be associated with the correct exit. The animal responds when the restraining door is lifted. When it must learn initially is that going to the light gives access to food, while responding to the unlighted exits is followed by an electric shock. After the lighted exit has been selected consistently for a time, a series of delayed-reaction trials is given.

So far we have considered only the situation in which the light was lit at the time of the animal's release and it thus served as an external cue telling the animal where to go. In delayed reaction tests, however, the light is turned on for a moment, then off, and it is off at the moment of release. What we want to know at this point is whether the animal remembers where the light was.

When tested in this way, a rat goes to the correct exit so [p. 121] long as it can point its head toward it while waiting for release and then follow its nose. This could mean that the stimuli guiding the [delayed] response are muscular -that muscle tensions rather than symbolic processes are bridging the interval from the time the light goes off until release occurs. The animal could therefore be responding to stimuli (muscular cues) present at the time of response even though the light itself is off. Such cues may be broken up by turning the animal around before releasing it. In the case of the rat, this is followed by no better than chance accuracy, so we are forced to conclude that, in this test, the animal has failed to demonstrate the presence of a symbolic process [(even as we have defined it)].

The raccoon, comparatively a very wise animal, has passed the same test with flying colors, with delays of up to 25 seconds. Instead of pointing, a raccoon spends the delay interval pacing back and forth. Upon release it goes directly to the exit where the light was. Something other than muscle tension, something in the raccoon's nervous system which represents the absent light, must bridge the interval. But it cannot do so for longer than 25 seconds. Intervals longer than this are followed by chance accuracy.

Although rats have failed this delayed reaction test, they have been successful on somewhat simpler tests involving the same principle. In a delayed alternation test, for example, the basic problem is to turn to the right or left after emerging from a central alley of a T-shaped box in a sequence such as right, left, right, left, and so on. If a delay is introduced after each turn, for how long an interval can the animal remember the preceding turn? Rats have responded correctly in this type of situation with delays as long as 10-15 seconds.

Many other tests of delayed reaction have also been used with rats, dogs, cats, and raccoons as well as primates. Rats, cats, and dogs have, in some types of delayed reaction tests, delayed successfully for several minutes, but the methods of testing them have differed so much that there is no basis for comparing their performances.

Below the mammalian level there is no good evidence of [p. 122] the type of memory involved in delayed reaction tests. The octopus has failed such tests. One experiment with minnows seemed to show evidence of delayed reaction, but a repetition of the experiment, in which additional controls were introduced, yielded negative results. Naturalists sometimes observe what appears to be delayed response in animals, but the absence of experimental controls makes it difficult to know whether the response is truly symbolic or whether the animal is reacting to stimuli not evident to the observer.

Monkeys and chimpanzees do exceptionally well on test of delayed reaction. Moreover, they can be tested more directly, without preliminary training (see illustration...). The animal sits in its cage. In front of it are two cups which look and smell alike, one to the right, the other to the left After getting the animal's attention, we slip a piece of food under one cup, then introduce a screen which hides the cups during the delay interval. During this interval, the animal may sit in its cage, sleep, or be taken for a walk. Later the screen is removed and the animal is released. If the monkey selects the correct cup, it gets to eat the food. Selection of the incorrect cup is unrewarded. From trial to trial the baited cup varies in right-left position according to a chance sequence. Since the cups look alike, since both are smeared with the same food so as to smell alike, and since they vary in random sequence from right to left, there is no external cue to guide the animal. In order to respond correctly, it must remember where it saw the food placed. Monkeys remember correctly for intervals up to 24 hours. In a slightly different situation, chimpanzees have remembered for as long as 48 hours.

The problem is sometimes complicated by using several pairs of cups. After seeing the food placed under one cup of each pair, a monkey goes from one pair to the next, selecting the correct cup in each. When monkeys and chimpanzees are compared on this test, the chimpanzees come out way ahead. They remember more pairs than the monkeys, and they remember them after longer intervals.

What is perhaps more interesting than the fact that a [p. 123] [p. 124] monkey or chimpanzee remembers the correct position of the food is that fact that it also remembers what was hidden. The monkey takes and eats whatever it saw placed under the cup. But what will happen if we surreptitiously substitute a less preferred food -say, lettuce, after banana has been placed under the cup? When such a substitution is made during the interval, the monkey picks up the correct cup, rejects the lettuce (which is accepted when the animal has seen it place), and hunt around for the missing banana. Quite often it also has a temper tantrum.

The ability to retain a symbolic representation of past experience is an extremely important step in mental evolution because it prepares the way for understanding and thinking. Once the organism can think of an object or event, it can begin to "put tow and two together," to solve problems by reasoning instead of by overt trial and error. The difference between a haphazard or trial and error solution and a solution based on understanding, or insight, is illustrated by detour problems.

Detour Problems

In a relatively simple detour test the animal is placed behind a wire mesh or glass screen through which it can see food on the other side, as illustrated... To reach the food the animal must go around the barrier, which involves turning away from the food at first. A chicken or rat respond directly to the barrier, as if trying to get through it to the food. After this approach fails, there is much random running about. Sometimes the animal runs back and forth along the screen and reaches the food accidentally in the course of such activity. Put back behind the screen, the chicken or rat behaves pretty much as before making direct approaches to the screen. However, it may get to the food a little faster than before by relinquishing some of its inadequate responses. After repeated [trials], the animal learns to circumvent the barrier without loss of time and effort.

Contrast this with the behaviour of a monkey or a chimpanzee tested in a similar situation. It goes around the barrier to [p. 125] the food without delay. It can do this because it is not bound by what is physically present but can reconstruct the situation while looking it over and, one may assume, in some way [p. 126] thinking about it. It "puts two and two together" and perceives possible moves. Primates thus learn by observing and thinking about what they observe as well as by acting. They demonstrate what Wolfgang Köhler, the famous Gestalt psychologist, called insight.

Consider a detour problem of greater complexity. Professor Köhler put a banana high about a chimpanzee in a cage which also contained possible aids with which to reach this delicacy. These included sticks, boxes, and a hanging rope. The question was: Will the chimpanzee have sufficient insight to make effective use of such objects? Sometimes it does. Chimpanzees have been known to stack boxes, uses long sticks somewhat as a pole vaulter does, and to swing on a rope Tarzan fashion. One chimpanzee got the banana by suddenly scrambling up on the Professor's shoulder. Professor Yerkes, who considered chimpanzees almost human, liked to tell about a chimpanzee to whom he presented a problem having, in his opinion, three possible solutions, only to have the animal solve it by a fourth way.

Chimpanzees have been known to join two bamboo poles by poking one into the end of the other to make a stick long enough to reach otherwise inaccessible objects (see illustration....). They are very fond of ants, and it is not uncommon for those in captivity to lick a piece of straw, put it down until ants swarm over it and get stuck, then lick the ants off. Jane Goodall has observed wild chimpanzees poking sticks into termites' nests, leaving them there for a moment, then drawing them out covered with termites, which are then removed with lips and tongue... She also reports observations made by others showing that chimpanzees use sticks to get honey out of an underground bees' nest and use a stone to break open the kernel of the palm nut. In their ant-seeking activities, chimpanzees break sticks to make them of suitable length. Here, according to Jane Goodall, are "the first examples of free-ranging non-human primates actually making very crude tools. [fn-3] [p. 127] [p. 128]

All of these examples involve some form of detour behavior where the animal circumvents the barrier either by locomotion around it or by using instruments of one kind or another. Various mammals, from the rat up, have been credited with some degree of insightful behavior in certain relatively simple situations, but no lower animal has even approached the level of insight and resourcefulness exhibited by the primates.

The Double Alternation Test

Even more convincing proof of [representative] symbolic ability in animals is found in problems designed to be impossible of solution except through reasoning. One of these, Walter S. Hunter's double alternation problem, has been used to test a large number of animals ranging from rats to human beings. There are various forms of this test, but the one most widely used is that illustrated ... Note the T-shaped arrangement of alleys. The subject starts at the foot of the T, goes to the head, and, turning into the right or left arm of the T, returns by a passage to the starting point. In order to earn a reward (and avoid electric shocks for making wrong turns), the animal must go through the run four times, making a right turn at the head of the T the first time around, another right turn on the second trip, a left turn on the third trip, and another left turn on the fourth trip, This right, right, left, left sequence of alternations is arbitrarily set.

Under certain circumstances this would be a simple problem which any mammal might solve. We could, for example, have two lights at the choice point, a bright one for right turns and a dim one for left urns. Then the animal would only need to learn the meaning of these signals. We would have a bright, bright, dim, dim sequence, and most animals would soon learn to respond right, right, left, left. In the problem under consideration, however, there are no such lights. Indeed there are no external signs of any kind which could signal the turn to be mad. In order to solve this problem the animal has no alternative but to learn the principle involved. [p. 129] [p. 130]

Note, moreover, that there are no muscle cues which might guide the animal. If the sequence were right, left, right, left -signal alternation- there might be such cues. Having gone to the right could leave muscle tensions which might serve as stimuli for a left turn. Having gone to the left could provide the cue, similarly, for a right turn. But in the double alternation problem there are no such cues. The same muscle tensions precede now a right turn, now a left.

Thus, if an animal is to learn this problem at all, it must figure it out, using something to represent the sequence required. Human beings use words, saying something like, "Oh I get it. You go twice to the right and twice to the left." We do not know what symbols are used by animals which solve this problem, but it is inconceivable that they could succeed without using symbols of some kind. We say this because it is necessary to remember the preceding turn, as well as what the sequence calls for, while approaching the choice point.

On this test a rat fails completely, even after 1000 trials spread over several months. [fn-4] raccoons, on the other hand, solve the problem in 500 to 800 trials, and cats and dogs do about as well. Monkeys and chimpanzees learn this type of problem in about 100 trials. Children under three years of age have failed it, but beyond this age, it is learned with fewer trials in successively older groups of children. The average number of trials required by a group of 38 children ranging in age from three to thirteen was approximately 15. On the same test, 25 college students required an average of 6 trials.

Ability to extend the rrll series of turns beyond one cycle is another indication of the level of mastery achieved. After learning the rrll sequence, some raccoons extended the series to rrllrr. Dogs and cats failed to do this. Monkeys appear to [p. 131] have no difficulty extending the series up to at least 8 additional turns, or rrllrrllrrll. Human subjects can continue giving the appropriate turns until stopped. They say to themselves,"right, right, left, left, right, right, left, left, " and so on.

Language is not essential in leaning and extending the double alternation problem. If it were, animals could not learn it.; But language certainly helps. As we said earlier, words are excellent substitutes for past experiences. They not only represent experiences, but they also bridge gaps to establish relationships and to formulate principles, such as that of the double alternation problem.

Acquiring Learning Sets

Another widely used test of animal intelligence involves discrimination procedures, but it goes beyond mere discrimination to investigate what Professor Harry Harlow calls learning sets. The development of such sets is sometimes so rapid as to suggest that the animal has insight into what is required of it.

The use of discrimination procedures in studying animal sensitivity has already been discussed. Our earlier interest in such procedures was limited to the question of whether or not the animal could discriminate between brightness, colors, figures, odors, or other stimulus pairs. We were not concerned with discrimination learning as such. If we had been, we should have discussed such aspects as trials to learn and the number of errors made prior to mastery. Actually, this would not have led us far, since discrimination learning, like maze learning, does not change significantly as we go from the lower vertebrates to primates. All of these animals learn discrimination problems easily, providing they have the requisite sensory capacity. Thus, discrimination learning in itself is too elementary to serve as a basis for comparing the intelligence levels of different vertebrates.

However, in testing for learning sets we begin with a discrimination problem. This may be a visual one in which positional cues are controlled so that the animal must [p. 132] discriminate in terms of what it sees. [fn-5] Suppose we begin with a problem requiring the animal to select a triangle instead of the cross that is paired with it. After our subject has learned this discrimination to an accuracy of 90 per cent in 20 consecutive trials, we present another pair of stimulus patterns, say a vertically striped versus a horizontally striped figure. When this is learned, another stimulus pair is presented, and so on, with many different stimulus pairs being presented. What we want to know is how rapidly the animal can learn new pairs. If the number of trials required is reduced as successive discriminations are learned, then we have evidence that the animals developing a learning set -that it is "learning to learn." This presupposes that the stimulus pairs are not themselves successively easier to discriminate.

The ultimate level of achievement in this type of test is reached when each new pair of stimuli is consistently discriminated after the first trial. One trial is necessary to indicate which pattern is correct. The animal with insight into what is going on makes a selection and then, in terms of the outcome, continues selecting on that basis. If its fist selection is correct, it thereafter selects that stimulus pattern; if incorrect, it subsequently selects the other.

In the experiment with an Indian elephant..., 330 trials were required to master the initial discrimination -between a black circle and a black cross. As new pairs of stimulus patterns were introduced, the number of trials required was reduced. Only 10 trials were required for the forth pair. With sixteen more pairs, the elephant could do no better than it did on its fourth test. Tested similarly, monkeys often achieve one-trial learning after many stimulus pairs have been learned.

In the examples given, each discrimination was mastered to a given level of proficiency before the next stimulus pair was introduced. A more widely used procedure is to present [p. 133] one stimulus pair for a particular number of trials (perhaps 50), then present another pair, regardless of how well the animal is doing. This continues, with a substitution being made after each block of trials, until the animal's maximum level of performance can be ascertained. When this procedure is used, the investigator is interested in observing whether accuracy improves in successive blocks of trials. Cats, raccoons, and other mammals have shown slight improvement. [p. 134]

For example, in later blocks of trials, cats have reached an accuracy of 70 per cent and raccoons 60 per cent. Monkeys, on the other hand, reached an accuracy of 90 per cent. Some of them exhibited one-trial learning after the first few stimulus pairs were presented. Chimpanzees and gorillas do no better than monkeys on such tests.

Concepts

Discrimination problems like those already described may be extended still further to provide tests of abstraction, generalization, or concept formation. The more complicated of such tests call for a high level of reasoning ability. Concepts represent the basic similarities between somewhat diverse things. Thus, many figures which differ from each other in color, size, and overall configuration have in common the characteristic known as triangularity. This characteristic is abstracted from the many particulars, hence all are called "triangles." Human beings have words like "triangularity" to represent such concepts. Animals do not have words, but some of them have been induced to respond on a conceptual basis, as when an equivalent response is made to many different figures which have in common three sides and three angles. When an animal does this, we say that it has developed a concept of triangularity.

Rats, dogs, raccoons, monkeys, and chimpanzees have demonstrated ability to respond in this way when given a sufficient amount of suitable training. However, no infra-primate has learned concepts at a higher level of abstraction. It is with such concepts that we now concern ourselves. All of the tests to be considered here were carried out in Professor Harry Harlow's primate laboratory at the University of Wisconsin, using the Wisconsin general Test Apparatus.

One of the simplest concepts involved in these studies is the "oddity" concept. As shown in the illustration... , a monkey is presented with three tests objects, in this case two circular and one T-shaped, or one circular and two T-shaped objects. In the first instance the monkey is rewarded if he selects the T; in the second, if he selects the circular [p. 135] object. From one presentation to another the items may differ from those illustrated -the size may be odd, the color odd, or the form odd, and the odd object will be in different positions relative to the other two objects. Learning to select the non-odd from such arrangements of objects is an even more difficult problem.

A further complication of the oddity-nonoddity problem is as follows: The test tray, let us say, holds three objects -red triangle, red cube, and cream cube. From the standpoint of color, the cream cube is odd. In terms of shape, the read triangle is odd. However, the test tray itself differed in color from one trial to another. If it was orange, the odd color was correct. Thus, the type of oddity to be selected was conditional upon the color of the test tray on which the objects were presented. Monkeys learned this problem, but they often required as many as 6000 trials before seeing the point and responding appropriately. Chimpanzees have not done as well as monkeys on such testes, but this is attributed to personality differences rather than to lower intelligence. It has been said that "chimpanzees are extremely sensitive to the slightest changes in an otherwise familiar situation, a new pair of shoes worn by the regular caretaker may [p. 136] temporarily transform the animal's attitude from friendly, confident approach to one one of wary avoidance." [fn-6] There is also an "enormous effect" when the stimulus is displaced an inch or so from its usual position. No doubt all of this is, in itself, a mark of high intelligence, although distracting when there are problems to be attended to.

Some aspects of Social Behavior

Social interaction at a relatively high level of complexity involving imitation and various forms of cooperative behavior is sometimes observed in monkeys and apes. Since this reveals an advanced animal intelligence, it also has a place in the present context.

Even animals as low in the scale as rats may be taught to imitate simple acts, as when one animal learns to follow another in order to get a reward. Cooperation at an elementary level may also be leaned. Rats, for example, have learned to take turns on a platform connected with a shocking device, in this way being able to avoid shock while eating. However, spontaneously aroused imitation and cooperation have not been elicited experimentally except at the primate level.

When one imitates, he learns something by observing another's performance, and when this takes place more or less suddenly, without special training -that is to say, spontaneously- it is usually regarded as a sign that higher processes are involved. The animal must observe a performance and, after such observation, use a symbolic representation of what it has observed as the basis for imitation.

Observational learning of this nature was studied in rhesus and cebus monkeys by Professor Carl J. Warden of Columbia University. He used two identical cages side by side, with one monkey in each. Everything which happened in one cage could be seen from the other. Each cage contained a puzzle device which the monkey could open by pulling a chain, turning a knob, lifting a latch, or carrying out some [p. 137] relatively simple act. The monkey in one cage had already been trained to perform the necessary act. The question was: Would the untrained monkey, after observing such an act, immediately open its puzzle box in a similar fashion? As one test was completed, a new opening device was slipped over the front of the puzzle boxes. In this way, 24 different tests were given. Rhesus and cebus monkeys imitated within 60 seconds of seeing the trained animal's performance. One of the rhesus monkeys was successful in 23 of the 24 tests.

Viki, a home-raised chimpanzee, imitated problems of greater complexity than the above when told "Do this," then given a demonstration. Some problems were: inserting a pencil in a pencil sharpener and turning the handle, taking the lid off a can with a screwdriver, and displacing a string by leverage with a stick, thus opening a puzzle box. Professor and Mrs. Keith Hayes, who reared Viki, report that here performance equaled that of two-to three-year-old children given a comparable test. [fn-7]

Two types of cooperative problem solving were observed in chimpanzees by Dr. Meredith Crawford. One of these involved pulling on ropes in unison so that access to food could be gained (see illustration...). The ropes were attached to a box loaded with pieces of iron so that it was too heavy for a chimpanzee to pull in alone. Preliminary training was given, such as teaching both chimpanzees to pull on command. However, the real test came when a pair were on their own. Under these circumstances each chimpanzee coordinated its pulling by watching the other. But it sometimes happened that one chimpanzee failed to pull until solicited and induced to do so by the other, more enthusiastic member of the pair. Often, when both were pulling, one suddenly let go of the rope. When this happened, the partner, acing surprised, gave a sideward glance in it s direction, then by various gestures, tried to induce it to pull. Sometimes this was successful and sometimes not. The soliciting chimpanzee could not indicate what it wanted [p. 138] [p. 139] from the other, and chimpanzees do not point. So the gestures and attempts of the soliciting animal to turn its partner around when he was headed away from the rope were saying nothing more specific, if anything, than "Help me do something."

The other test of cooperative behavior required that yellow , green, red, and blue panels be pushed in that order to release food. Each animal was trained to push the panels, then the partners were separated by a grill with yellow and red panels on one side and green and blue on the other. First the chimpanzee with the yellow panel had to push it. Then the partner had to push its green panel. This had to be followed, in order, by pushing of the red and blue panels. Four of the chimpanzees given this test cooperated, each watching for the appropriate response from its partner before responding itself. Two of them solicited when the partner did not respond. They did this by reaching through the grill and turning the animal in the proper direction, or by pushing him.

LANGUAGE

Soliciting behavior of the kind observed in the foregoing experiments comes as close to language as any animal has been able to achieve with its own untutored resources.

In their natural state, chimpanzees vocalize when emotionally aroused, and they communicate with gestures, but for some reason they have never been known, without special training, to bridge the gap between these limited forms of communication and conventionalized vocal or gestural signals such as all languages use. While their vocalization cover a wide range, these have no symbolic reference to aspects of the environment or to the animal's past experience.

This deficiency goes back, in the fist instance, to the animal's failure to invent a language as primitive men did. But it is apparent, also, in the chimpanzee's failure to imitate, without elaborate training, an already given language. The normal human child acquires the language of those [p. 140] around it with apparent spontaneity, but a chimpanzee, even with the intelligence of a two- to three-year-old child and treated like a child in a human family situation, makes no progress in this direction unless subjected to an elaborate conditioning program. Even the, progress is slow and extremely limited. One chimpanzee has been trained to use a few sounds resembling words, another has been taught to communicate with human beings by using conventionalized gestures selected for this purpose, and still another has leaned to communicate with it trainers by arranging various colored plastic shapes in certain sequences.

Some problems involved in inducing chimpanzees to communicate linguistically are brought out in the following discussion, which also compares linguistic acquisition in chimpanzees and children.

Linguistic Acquisition in Chimpanzee and Child

Viki, the chimpanzee whose imitation has already been mentioned, was taught to "say" a few "words," but teaching her to do so was a very arduous process (see illustration...). The chief difficulty, at the outset, was to get here to make any sounds other than inborn reflex vocalizations elicited by excitement. Except when emotionally aroused, Viki was silent. A human infant of the same age would have been babbling for months, spontaneously and without any obvious relation to what was going on around it.

Since Viki had no urge to vocalize, Dr. and Mrs. Hayes decided to make her "speak" for here supper, as a dog might be taught to bark for it. This met with no success at first. When asked to speak, Viki looked at the milk held in front of her but said nothing. If Mrs. Hayes moved away, however, she gave worried little sounds (oo oo). This vocalizing, although only an emotional reaction, was immediately rewarded: Viki got some milk. While drinking the mild, she mad "sputtered food barks," which brought more milk. After five weeks of further training, a new sound (ahhh), [p. 141] [p.142] accompanied by facial contortions and a tense, preoccupied look, began to appear. When she made this sound, Viki reached for her milk, and she was rewarded. Thereafter, the command to speak brought an ahhh. Now Viki was ready to learn her first word, Mama. Mrs. Hayes trained her to say this by pressing her lips together and releasing them as she said ahhh for food. After a few weeks of such training, it was no longer necessary to touch Viki's lips, she was saying Mama by herself. The word Papa was laboriously added to her vocabulary after she had learned to imitate a Bronx cheer. Softer and shorter p's in succession were required, then the the repetition of two p's in succession. This produced something approximating the sound Papa. The word cup was acquired by learning to repeat the sounds k and p in rapid succession. Later still, Viki learned ch for a drink. She also learned to click her teeth for a ride in the car. This is as far as Viki went in learning to speak. Moreover, she never came to use her words for social purposes or for egocentric expression. She spoke only when there was no other way of getting what she wanted. There was no evidence that she had any insight into the meaning of language.

How different it is when children learn to speak! It is not necessary to subject them to special training. They not only vocalize spontaneously (babble), they also learning to imitate, with gradually increasing accuracy, the words used by those around them. This process is aided by the favorable responses of others to their vocalizing and by the fact that as their imitations get better, they are able to communicate their wishes more easily. But nobody has to command them to speak, move heir lips in appropriate ways, or provide special rewards. Children soon "catch on" to the fact that everything has a name and the that they can use words to represent aspects of their world.

The sudden insight which is often a part of this process was dramatically revealed in Helen Keller's learning of her fist word. Helen had two handicaps -blindness and deafness- but she was a very bright child. Her teacher, Ann Sullivan, communicated with her tactually, through a manual alphabet. One morning during Helen's seventh year she [p. 143] made an extremely important discover, and her brightness enabled here to capitalize on it almost immediately.

Ann Sullivan tells how the child asked the name for water by pointing to it and patting the teacher's hand. Miss Sullivan spelled the word in the manual alphabet. Later, when they went out to the well for water, she made Helen hold her mug under the pump spout while she spelled "w-a-t-e-r" into the child's free hand. Says Miss Sullivan:

The word, coming so close upon the sensation of cold water rushing over her hand, seemed to startle her. She dropped he mug and stood as one transfixed. A new light came into her face. She spelled "water" several times. Then she dropped on the ground and asked for its name and pointed to the pump and trellis, and suddenly turning round she asked for my name. I spelled "Teacher." Just then the nurse brought Helen's little sister into the pump-house and Helen spelled "baby" and pointed to the nurse. All the way back to the house she was highly excited, and learned the name of every object she touched, so that in a few hours she had added thirty new words to her vocabulary.

The next day Helen was like "a radiant fairy," going from object to object naming it. Everything has a name is "like an intellectual revolution," as the German philosopher Ernst Cassirer pointed out in his An Essay on Man. "The child begins to see the world in anew light. It has learned the use of words, not merely as mechanical signs and signals, but as an entirely new instrument of thought" [fn-9]

Insight such as this is beyond the reach of any animal. Animals react only to sounds, not to the symbolic meaning of words. The late Professor Edward L. Thorndike once demonstrated the point with an experiment on cats. He had trained the animals so that when he said "I must feed those cats," they dashed to the food box, even when he put no food in it. One day, to test their understanding of his words, [p. 144] he exclaimed at the cat's mealtime: "Today is Tuesday." The cats instantly sped to the box. The words "My name is Thorndike" evoked the same response.

Here, then, is an important and puzzling question. Why did the apes get so far in symbolic development, as revealed by delayed reaction, double alternation, and other such tests, yet fail to bridge the gap between nonlinguistic symbols and those of speech? Why are they "almost human," yet without speech or anything comparable with it? Have they nothing to say? This is a possibility, but one linguist has pointed out that having nothing to say does not stop us from speaking! The difficulty does not lie with the ape's vocal range. This does not lend itself readily to the making of human sounds, but it includes enough sounds to serve a linguistic function. There is no evidence, however, that apes have a vocal language of their own. As suggested earlier, in speaking of Viki, apes make no sounds at all except emotional ones. These would have no communicative significance beyond telling another of their kind that its mate is angry, frightened, or on the prowl.

Some importance may be attached to the fact that apes fail to vocalize spontaneously, as babies do when they babble by the hour. Recall that Viki could not be trained to "speak" until she had fist been induced to vocalize unemotionally. Man probably vocalized in a babbling fashion until he got the idea that sounds can be used to represent objects and events and are thus useful for communicative purposes.

The apes' linguistic backwardness may be attributed to insufficient insight and relatively poor symbolizing ability. If it were to develop into a talking animal, it would need a very high level of insight, exceptional ability to store and retrieve information, and great versatility in reasoning -in "putting two and two together."

There are no doubt several neurological and anatomical reasons for the ape's failure to invent a language. Recall that the ape's brain weighs only one third as much as our own, yet it has an equally large (or larger) body to control. [fn-10] This means that its brain is largely concerned with sensory and [p. 145] motor functions. The association areas are small in comparison with the areas specialized for sensory and motor processes. Even the frontal lobes, so prominent in the human brain, are relatively small.

Teaching Sign Language to a Chimpanzee

Chimpanzees gesture spontaneously, although their movements have no known linguistic significance under natural conditions. Professor Winthrop N. Kellogg [fn-11] asks whether this spontaneous use of gestures by chimpanzees could not be developed into something more. Might not "an intelligent animal learn a series of regular or standardized signals -as a sort of semaphore system?" Kellogg continues by pointing out that "Even though a chimp may lack the laryngeal structure or neural speech centers of man, it does not necessarily follow that it has deficiencies in general motor activity" which would make it unable to "communicate back and front in a series of hand movements, arm signals, and postures." When Kellogg's statement appeared, tow investigators at the University of Nevada had already begun an experiment designed to teach a chimpanzee a selected group of signals from the American Sign Language (ASL), which finds wide use in communication among the deaf. " These investigators, R. Allen Gardner and Beatrice T. Gardner, selected for their experiment a wild-born female chimpanzee estimated to be eight to fourteen months old. They named her Washoe.

After a period of adaptation to a human environment, Washoe was gradually trained, by various "friends and playmates" as well as "providers and protectors," to use the [p. 146] selected signals, which were introduced into games and other activities calculated to "result in maximum interaction." Each of the participants had already learned the signals, and these, rather than oral communication, were used in the interactions with Washoe.

Part of the training comprised "Do this" games (See illustration...). One of the rewards for correct imitation of a signal was tickling. Imitation on command was difficult to obtain. The Gardners say:

It was not until the 16th month of the project that we achieved any degree of control over Washoe's imitation of gestures. Eventually we got to the point where she would imitate a simple gesture, such as pulling at her ears, or a series of such gestures -first we make a gesture, then she imitates, then we make a second gesture, she imitates the second gesture, and so on- for the reward of being tickled. [fn-12]

A prompting method was also used to introduce new signals and to correct errors. This involved "repeating in exaggeratedly correct form,the sign she had just made," until she repeated it herself "in more correct form." Pressing too hard along these lines, however, made Washoe depart from what she had been doing -to "ask for something entirely different, run away, go into a tantrum, or even bite her tutor."

Delayed imitation also played a role in Washoe's acquiring of appropriate signs, for example, the sign for "toothbrush."

A part of the daily routine has been to brush her teeth after every meal. When this routine was first introduced Washoe generally resisted it. She gradually come to resist with less and less fuss, and after many months she would even help or sometimes brush her teeth herself.... One day, in the 10th month project, Washoe was visiting the Gardners' home and found [p. 147] her way into the bathroom. She climbed up on the counter, looked at our mug full of toothbrushes, and signed "toothbrush." At the time we believed that Washoe understood the sign but we had not seen her use it. She had no reason to ask for the toothbrushes, because they were well within her reach, and it is most unlikely that she was asking to have her teeth brushed. This was our fist observation, and one of the clearest examples, of behavior in which Washoe seemed to name an object or an event for no obvious motive other than communication. [fn-13]

By the 14th month, after imitative prompting and other procedures, she called for her toothbrush by making the proper sign when the meal was finished. Delayed imitation also appeared in acquisition of the sign for "flower." Early in the experiment, when Washoe showed an interest in flowers and pictures of flowers, the sign for flower was given by her tutors, and she was induced to imitate it. But there was no spontaneous use of the sign until the 15th month, when she made it as she and a companion approached a [p. 148] flower garden. The investigators capitalized on this spontaneous signing, as they had done in the case of "toothbrush," and the sign eventually appeared quite reliably in response to a variety of flowers and their pictorial reproductions.

A gestural form of babbling was encouraged and utilized in teaching new signs. Take, for example, acquisition of the sigh for "funny." In this, as used by Washoe, the tip of the index finger touches the nose and a snort is given. This sign first appeared "as a spontaneous babble that lent itself readily to a simple imitation game -first Washoe signed "funny," then we did, and, then she did, and so on. We would laugh and smile during the interchanges that she initiated, and initiate the game ourselves when something funny happened. Eventually Washoe came to use the 'funny' sign spontaneously in roughly appropriate situations." [fn-14]

Instrumental conditioning with successive approximations to a desired response also played a part in sign acquisitions. As an illustration of this, take the sign for "more." As we have already said, Washoe liked to be tickled and was rewarded by tickling. When tickling stopped, she often indicated that she wanted more by taking the tutor's hands and placing them in the appropriate positions on her body. The Gardners say that:

The meaning of these gestures was unmistakable, but since we were not studying our human ability to interpret her chimpanzee gestures we decided to shape an arbitrary response that she could use to ask for more tickling. We noted that, when being tickled, she tended to bring her arms together to cover the place being tickled. The result was a very crude approximation to the ASL sign for "more." Thus we would stop tickling and then pull Washoe's arms away from her body. When we released her arms and threatened to resume tackling, she tended to bring her hands together again. From time to time we would stop tickling and wait for her to put her hands together by herself. At first, any approximation to the "more" sign, however, crude, was rewarded. Later, we required closer [p. 149] approximations and introduced imitative prompting. Soon, a very good version of the "more" sign could be obtained, but it was quite specific to the tickling situation. [fn-15]

Later, the Gardners attempted to extend the "more" sign to a wider range of situations. It came to be used for more pushing across the floor in a laundry basket, more swinging of Washoe in the arms, for more feeding, and so on. Finally, Washoe "transferred the 'more' sign to all activities, including feeding. The transfer was usually spontaneous, occurring when there was some pause in a desired activity or when some object was removed. Often we ourselves were not sure that Washoe wanted 'more' until she signed to us." [fn-16] In a somewhat comparable way, Washoe learned the sign for "open," using it first as a request to open the door,. then to open many things -as well as turn on the water faucet.

After subjection to such procedures for 22 months, Washoe had acquired a vocabulary of 34 signs. In addition to those already mentioned there were the signs for open, tickle, go, out, hurry, hear-listen, drink, hurt, sorry, please, food-eat, cover-blanket, you, napkin-bib, brush, hat, shoes, pants, clothes, cat, come-gimme, key, baby, and clean. These were all used frequently, in appropriate contexts. Acquisition was initially slow, but it accelerated as training continued. For example, there were 4 spontaneously aroused signs in the first seven months, 9 new ones in the second seven months, and 21 additions in the third seven months. Some difficult differentiations were involved, like the difference between "flower" and "smell." Transfer from one context to other appropriate contexts has been common, like the "dog" sign for a real or pictured dog to the sound of a dog barking. After she had acquired a repertoire of about 10 signs, Washoe began to use combinations of them, such as "gimme drink please," "go in," "open key" (locked door), "listen eat" (alarm clock signaling mealtime). Acquisition of "I-me" and "you" has led to combinations resembling short sentences.

This experiment continues, and the eventual outcome [p. 150] remains to be seen. Concerning this, the investigators say, "In terms of the eventual level of communication that a chimpanzee might be able to attain, the most promising results have been spontaneous naming, spontaneous transfer to new referents, and spontaneous combinations and recombinations of signs." [fn-17]

Why did Washoe do so much better than Viki in acquiring and using linguistic symbols? It seems unlikely that she was that much more intelligent. It seems unlikely, also, that better training techniques were used with Washoe, except for the fact that gestures are more readily shaped than vocalizations. The chimpanzee is limited when it comes to spontaneous vocalization, as we saw in the case of Viki, but the range of gestures and their spontaneous arousal is not so limited, and it appears that this gives the answer to our question. Vocally, a chimpanzee appears to have relatively little to say. Gesturally, however, it can, with appropriate training, be given the wherewithal to "say" a great deal. It can also be taught to communicate through appropriate use of visual signs.

Use of Visual Signs by a Chimpanzee

Sarah, a seven-year-old chimpanzee, learned to communicate with Professor David Premack and his assistants by using chips of various shapes, sizes, colors, and textures. [fn-18] The chips were made of plastic and backed by metal so that they would adhere to a magnetized "language board" on which the chimpanzee had to place them, sometimes singly and sometimes in meaningful combinations, in order to "say" what was require to get a reward. Some of the most easily described chips and the words represented by them were: blue triangle (apple), small purplish square (banana), blue star (the verb insert), a yellow figure shaped somewhat like an arrow (equal to), the same shape colored red (not equal to), a blue propeller-like shape (yes) and a similarly shaped [p. 151] gray chip (no). The chips represented various verbs, personal names, colors, foods, objects, concepts, adjectives, and adverbs.

First Sarah was presented with a readily accessible banana, then she was required to learn that an otherwise inaccessible banana could be gained by putting a nearby chip on the language board. When she had learned to "ask" for the banana in this way, she was presented with other fruits and their chips. After having learned to use this signs appropriately, Sarah was taught to use chips representing verbs like insert and give as well as various additional nouns. The next step was to require an appropriate sequence, such as "Mary give apple Sarah," "Sarah insert banana pail," and so forth. That is to say, the chips representing Mary, give, apple, and Sarah had to be place in this order on the language board. A further step was to teach such things as "Apple name of ...." and "Apple not name of...." in relation to the presented fruit.

This experiment had run for two years and was continuing when the brief report so far available was written. It appears that during this time Sarah had learned 120 "words" and also such language functions as construction of sentences and asking and answering questions. According to Professor Premack, Sarah not only used the chips in meaningful combination but also "understood" what she was doing, a evidenced by the fact that she at one time devised a "sentence-completion test" for here trainer.

There is an obvious gap between man's invention of language, oral, gestural, and written and the ape's failure in this respect. And there is a marked difference in the ability of an ape and a child of normal intelligence to acquire the linguistic symbols already available in a human environment. The child picks up these symbols quite readily, and its vocabulary grows apace. The ape, on the other hand, acquires very few of the symbols available to it, and even these it acquires only after laborious tutoring. All of this is relevant to an issue with which we shall end this chapter on the evolution of intelligence. [p. 152]

CONTINUITY OR DISCONTINUITY?

Was mental evolution a process in which successive small increments built one upon another until the level of human mentality had been achieved? Theorists who argue for the continuity of mental evolution say that it proceeded in just that way. But there are others who hold that mental evolution moved ahead in leaps rather than, or in addition to, continuous increments. In the succeeding discussion we shall see that evolution, in its various aspects, reveals both continuities and discontinuities. We shall see, too, that the invention of language as a means of communication comes as close as anything to supporting the argument or discontinuity.

The continuity viewpoint is clearly expressed by Professor Harry Harlow, who says; "If we are to explain learning in terms of modern evolutionary theory, there should be continuity from the simplest to the most complex forms of learning. The appearance of a radically new kind of learning at any evolutionary point or period, including that during which man developed, is not in keeping with modern gene theory," which Harlow interprets as the theory that "evolution takes place by natural selection among multiple mutations, each of which produces some small organic change." [fn-19]

But gene changes may be small or large. When small, they may not become evident structurally until enough of them, or a certain combination, sufficiently changes the genic balance. Then the structural outcome may be markedly divergent from previous structure even though each gene mutation is, in itself, very small. We must also keep in mind that many mutations are recessive, hence they do not find structural expression unless the animal carrying them breeds with others having a similar genetic constitution. When this happens, there may be a suddenly appearing change in structure; [p. 153] even a large one. [fn-20] If inbreeding does not occur, the mutations may never become apparent. Moreover, many mutations which found structural expression in the past may have been lost, as suggested by the many extinct animals. Such a loss would produce gaps in the line of descent. Thus, we are not forced by modern gene theory to suppose that mental evolution was marked by continuity. There appears to be continuity interspersed with discontinuity at certain points.

The presence of discontinuity was stressed by Professor D'Arcy Thompson, who say particular evidence of it in the transition from invertebrates to vertebrates. The problem exists, to a degree, even as we go from one vertebrate level to another. According to Thompson, "Darwinian evolution has not taught us how birds descend from reptiles, mammals from earlier quadrupeds, quadrupeds from fishes, nor vertebrates from the invertebrate stock." But he was not presenting an argument against evolutionary descent, for "we may fail to find the actual links between the vertebrate groups, but yet their resemblance and their relationship, real though indefinable, are plain to see; there are gaps between the groups, but we can see, so to speak, across the gap." However, "We cannot transform an invertebrate into a vertebrate, nor a coelenterate into a worm, by any simple and legitimate transformation, nor by anything short of reduction to elementary principles." Thompson says that his argument "indicates, if it does not prove, that... mutations, occurring on a comparatively few definite lines, or plain alternatives, of physicomathematical possibility, are likely to repeat themselves; that the 'higher' protozoa, for instance, may have sprung not from or through one another but severally from the simpler forms; or that the worm-type, to take another [p. 154] example, may have come into being again and again. [fn-21]

Commenting on Thompson's views, the late Professor Henry W. Nissen [1951] said that "this emphasis on large and abrupt variations in evolution is representative of most modern biological thinking," and "its implications for comparative psychology is that in behavior also we may expect to find discontinuity -qualitative rather than merely quantitative changes- as we pass from the lower to the higher animal forms." [fn-22]

In discussing the problem before us, Professor Dobzhansky, who sees both discontinuity and continuity in evolution says:

Man is not simply a very clever ape. On the contrary, he possesses some faculties that occur in other animals only as rudiments, if at all. Quantum evolution, emergence of novel adaptive designs, may involve breaks in the evolutionary continuity when the differences between the ancestors and the descendants increase so rapidly that they are perceived as differences in kind. Antecedents of the new designs may, nevertheless, be detected in the old one. We must equally resist the temptation to regard man either as something completely unlike any animal or as something devoid of all novelty.... For example, legs in land-living vertebrates were new organs, since fishes for which the land vertebrates descended had no legs. Comparative anatomy shows, however, that the extremities of the land vertebrates arose from the paired fins of their fishlike ancestors. [fn-23]

Modern evolutionary theory thus lends support to the idea that there was continuity in the evolution of some psychological processes and discontinuity in the evolution of others. What aspects of intelligence, then, appear to fall into the continuous category and what into the discontinuous? [p. 155]

We see the clearest evidence for continuity in the simpler, or basic, processes -those upon which sheer survival is contingent. The natural environment, in itself, places limitations on the kinds of mutants that can survive. Survival requires at least minimal sensitivity to environmental changes and ability to respond to aspects of the environment in such a way as to satisfy basic physiological needs, such as the need for food, water, and shelter from the elements. Therefore one would expect to find all animals with some degree of sensitivity. We know, in fact, that there is a more or less continuous line of development from the unspecialized sensitivity of the ameba to the specialized sensitivity of human beings. In some respects, this trend can be considered purely quantitative -as increasing increments of the same sort of sensitivity. There is gradually increasing sensitivity o brightness. An increasing amount of detail in visual configurations becomes evident to the organism. Sounds of decreasing loudness can be heard. And so on, with many other aspects of sensitivity. On the other hand, discontinuity is suggested when color vision emerges. This suddenly-appearing sensitivity to the wavelength properties of light cannot, it seems, be derived from some quantitative transformation of brightness vision. [fn-24] The emergence of a depth dimension in the vision of primates is perhaps another example of sensory discontinuity.

In motor functions there is likewise a certain degree of continuity and also discontinuity. Adaptation requires locomotor ability, and all animals move around to some extent, although with different degrees of facility and by use of different locomotor mechanisms. Thus, in some respects the evolution of locomotor ability from ameba to higher forms is continuous. One will recall the previously mentioned [p. 156] structural change from fins to legs, a transformation regarded by Dobzhansky as basically continuous and quantitative. But when a terrestrial animal gives rise to one which moves through the air, we have something essentially new in locomotion, despite the similarities (homologies) in limb and wing structures.

Some psychologists se evidence of discontinuity between unlearned and learned behavior, i.e., between instinct and habit. Professor Gaston Viaud, referring to discontinuity in mental evolution, says "The milestones in the evolution of intelligence are the emergence from instinctive reactions and the appearance of conceptual thought." [fn-25] The emergence from instinctive reactions is cited as the beginning of intelligence. However, the idea that there is a basic dichotomy between instinctive and intelligence, or between instinct and habit, has been defended and refuted many times. The argument seems rather fruitless because both learned and unlearned behavior persist throughout most of the animal world, with inborn modes of response predominating t lower levels and acquired ones at higher levels. Learning itself may be viewed as a modification of unlearned reactions and earlier learned reactions. Thus, there is no clear break between instinctive and learned behavior.

The evolution of learning ability is marked by apparent continuity, at least up to the mammalian level. Pavlovians and many others regard learning as reducible, in the last analysis, to conditioning, which is viewed, essentially, as the acquisition of new stimulus-response connections. We have already observed that the ability to learn discrimination habits and avoidance of blind alleys in a maze shows more or less gradual improvement until the mammalian level is reached. This is a quantitative development, evidenced by faster learning, fewer errors, decreased effort, and an increase in the complexity of sensorimotor skills that the animal can learn. In this elementary type of learning, with its dependence upon conditioning and trial and error, there is [p. 157] no further advance from rat to man. This does not mean that the simpler learning processes ceased to exist beyond the rat level, but only that they were supplemented by processes essentially different in kind and more characteristically mammalian. These processes have been referred to, in general, as symbolic: more specifically, as insight, recall memory, concept formation, generalizing, and reasoning. All have in common the fact that, to some degree, they transcend the immediate stimulating circumstances by bringing to bear upon them the fruits of past experience. In their more complex aspects they comprise what Professor Viaud and many others refer to as conceptual thought.

It is possible that the emergence of symbolic processes had to wait upon the evolution of the cerebral cortex, which first became clearly apparent in mammals. In lower vertebrates, as was pointed out in earlier discussions, the cerebrum is rudimentary and so largely involved in olfaction that it is often referred to as a "smell brain."

Professor Bitterman and collaborators at Bryn Mawr College believe that they have fond evidence for discontinuity in learning between the fish and bird levels, with birds (pigeons) behaving at a typically mammalian (rat) level despite the absence of a well developed cerebrum. These investigators used several learning problems, but we shall consider the one which comes closest to being a test of insight -that involving successive reversals of a visual discrimination. This will be recognized by the reader as a learning-set (or learning-how-to-learn) problem. Our earlier examples of learning set involved successive pairs of stimulus patterns, and the animal was credited with insight if it showed more or less sudden improvement in tits learning of successive discriminations. In reversal learning, on the other hand, the same stimulus patterns (for example, vertical versus horizontal black and white stripes) are used throughout. The animal is trained with one pattern rewarded (say, horizontal), then the other (vertical) is rewarded, making the former incorrect. When trained with the vertical pattern rewarded is finished, another reversal occurs, making the [p. 158] horizontal again correct. Such reversing of cues continues, and the investigator seeks evidence of progressive improvement with successive reversals. On problems like this, monkeys sometimes reach a one-trial level of performance. As in learning successive discriminations, they make no more errors after the first trial with the new reversal. This is taken to mean that they have insight into what is going on. On the other hand, animals which show progressive improvement, even thought they do not reach the one-trial level, may be credited with a certain degree of insight.

On problems involving reversal of visual cues, fish and turtles showed no progressive improvement. In fact, after 18 reversals their performance was poorer than in the original learning and earlier reversals. The situation was quite different with pigeons. Their performance go progressively worse in the early reversals, then began to improve, until after 18 reversals, they were making relatively few errors. Rats also demonstrated progressive improvement after the first few rehearsals. Thus, Bitterman concluded that his experiments had tapped "an intellectual capability of higher animals that is not at all developed in the fish." [fn-26]

Psychologists have had little or no success in finding evidence of recall memory and reasoning below the mammalian level. This is a clearer discontinuity than can be found in the development of learning sets based on reversal problems. Our earlier discussion of symbolic processes showed that rats have what might be called rudimentary symbolic processes, as revealed in delayed reaction and reasoning tests. These processes become increasingly evident as the primate level is approached. They also reach high levels of complexity, as in the solving of complex delayed reaction, double alternation, and conceptual learning problems by monkeys and chimpanzees.

Although apes have symbolic processes basically comparable[p.159] to man's, their failure to invent speech is perhaps the most evident discontinuity in the evolution of mental life. Nevertheless, Professor Harry Harlow questions whether this difference is as large as it seems. He refers to the "common error of assuming that the particular human traits of language and culture imply the existence of some vast intellectual gap between man and other animals." Harlow believes that "a relatively small intellectual gain by man over the anthropoid apes" may have given rise to "the development of symbolic language and also culture." [fn-27] He does not profess to know what this "small intellectual gain" may have been, but reference is made to the possibility that the ape's "language inadequacies basically result from the failure to possesses certain unlearned responses." More specifically, apes fail to engage in spontaneous vocalizing, or what, in human beings, is called babbling. We have seen from the researches of the Gardners and David Premack, however, that certain unlearned gestures and visual signs may be utilized by the chimpanzee for linguistic communication when suitable training procedures are used. This is a "small intellectual gain" not found in the vocal behavior of chimpanzees, but it still leaves a large gap between the communicative abilities of apes and men. Indeed, Noam Chomsky thinks that human linguistic communication is based on "entirely different principles" and "associated with a specific type of mental organization" not possessed by sub-humans. "There seems to be no substance," he says, "to the view that human language is simply a more complex instance of something to be found elsewhere in the animal world. This poses a problem for the biologist, since, if true, it is an example of true 'emergence' -the appearance of a qualitatively different phenomenon at a specific stage of complexity of organization." [fn-28] [p. 160]

By bridging the gap between animal communication and human language, however this "emergence" may have occurred, the human species can be said to have entered a "new world." By communicating linguistically, they had the means to think more clearly, to reflect upon the past, and to project themselves into the future. Language opened up a new level of mastery over the environment, both social and geographical, and every aspect of human life since its invention has felt the impact.

fn-1 See N.L. Munn, The Evolution of Growth of Human Behavior (2nd ed.), Houghton Mifflin, 1965, pp. 131-137; W.N. Kellogg, "Worms, Dogs, and Paramecia," Science, 1958, Vol. 127, p. 166; and M.S. Katz and W.A. Deterline, "Apparent Learning in the Paramecium," Journal of Comparative and Physiological Psychology, 1958, pp. 243-248.

fn-2 Jane Goodall, "chimpanzees of the Gombe Stream Reserve," Chapter 12 in Irven DeVore (Ed.), Primate Behavior, Holt, Rinehart and Winston, 1965.

fn-3 Goodall, 1965, p. 473.

fn-4 Rats have performed the double alternation when gradually "educated" to it by subsidiary procedures, but they do not succeed on the test described here. On this point, see N.L. Munn's Handbook of Psychological Research on the Rat, Houghton Mifflin, 1950, pp. 210-212.

fn-5 This is important because experiments on learning set in which positional cues can be learned do not reveal much difference in ability from lower to higher vertebrates.

fn-6 L.P. Gardner and H.W. Nissen, "Simple Discrimination Behavior of Young Chimpanzees: Comparisons with Human Aments and Domestic Animals," Journal of Genetic Psychology, 1948, Vol. 72, p. 161.

fn-7 A popular presentation of the experiments with Viki appears in Cathy Hayes' The Ape in Our House, Harper, 1951.

fn-8 Helen Keller, The Story of My Life, Doubleday, 1903, p. 315.

fn-9 Ernst Cassirer, An Essay on Man, Yale University Press, 1944, p. 35.

fn-10 It is interesting to observe that human infants of 15 months, an age when speech development is normally well under way, have an average brain weight of 944 grams, while the brain weight of a mature chimpanzee is no greater than 450 grams. Human brain weights at various ages are given in J.L. Conel, The Postnatal Development of the Human Cerebral Cortex, Vols. 1-8, Harvard University Press, 1939-1967.

fn-11 W.N. Kellogg, "Communication and Language in the Home-Raised Chimpanzee," Science, 1968, pp. 423-427 (quotation from p. 426).

fn-12 R. Allen Gardner and B.T. Gardner, "Teaching Sign Language to a Chimpanzee," Science, 1969, Vol. 165, pp. 664-672 (quotations from p. 666).

fn-13 Gardner and Gardner, p. 667.

fn-14 Gardner and Gardner, p. 667.

fn-15 Gardner and Gardner, p. 669

fn-16 Gardner and Gardner, pp. 669-670.

fn-17 Gardner and Gardner, p. 672.

fn-18 D. Premack, "The Education of Sarah," Psychology Today, September, 1970, pp. 54-58.

fn-19 Harry F. Harlow, "The Evolution of Learning," in Anne Roe and George Gaylord Simpson (Eds.), Behavior and Evolution, Yale University Press, 1958, pp. 277-278.

fn-20 Inbreeding and its effects on mutations are exceedingly complex, even under the relatively restricted conditions imposed upon laboratory colonies of mice. For a detained discussion of the mutations observed in mice and the effects of various breeding systems upon their transmission, see Earl L. Green (Ed.), Biology of the Laboratory Mouse (2nd ed.), McGraw-Hill, 1966.

fn-21 D'Arcy W. Thompson, On Growth and Form, Cambridge University Press, 1942, pp. 1093-1095.

fn-22 H.W. Nissen, "Phylogenetic Comparison," in S.S. Stevens (Ed.), Handbook of Experimental Psychology, Wiley, 1951, p. 348.

fn-23 Theodosius Dobzhansky, Mankind Evolving, Yale University Press, 1962.

fn-24 One must keep in mind that some processes have undergone independent evolution in different evolutionary lines. For example, different mechanisms underlie color vision in insects, birds, and primates. Spatial vision in owls has as perhaps its only similarity with that of primates the fact that the eyes, being toward the front of the head, can get overlapping images of what lies before them.

fn-25 G. Viaud, Intelligence: Its Evolution and Forms. Harper, 1960, p. 116.

fn-26 M.E. Bitterman, "The Evolution of Intelligence," Scientific American, 1965, Vol. 212, p. 96. For a more detailed report of these experiments, see M.E. Bitterman, "Phyletic Differences in Learning," American Psychologist, 1965, Vol. 20, pp. 395-410.

fn-27 H.F. Harlow, "The Evolution of Learning," 1958, p. 278.

fn-28 Noam Chomsky, Language and Mind, Harcourt, Brace and World, 1968, p. 62.