Comparative Psychology and Animal Learning
Jesse E. Purdy, Southwestern University
Michael Domjan, University of Texas at Austin
People have an abiding fascination with animals. Zoos and animals trained for entertainment have been popular for centuries. More recently, nature shows dealing with animal behavior can be seen on TV virtually any time of day, and pet ownership is at an all-time high. Animals can be fun, interesting, and emotionally satisfying. They can also provide important information about learning and cognition and the evolution of behavior. This kind of information can, in turn, provide important insights into what it means to be human. Scientists who have a personal fascination with animals can translate that interest into studying comparative psychology and animal learning and end up knowing more about the human animal as a result.
A Personal Odyssey
I (JEP) have been interested in science and nature most of my life. In high school I told my biology teacher I wanted to become a marine biologist and study the behavior of marine mammals. The teacher informed me that career opportunities in marine biology were nonexistent and suggested I become a physical oceanographer instead. After my first physics course in college, however, I decided to major in psychology. I discovered my passion in two courses, Experimental Psychology and the Psychology of Learning. I enjoyed designing experiments, collecting and analyzing data and I enjoyed thinking about the various theories of learning and how they might be tested. The instructor in Experimental Psychology was Dr. Henry A. Cross who conducted experiments with animals. Dr. Cross was a learning theorist who worked with rats at the time but had also conducted experiments with monkeys under the direction of Harry Harlow. What could be better, I thought, than to study animal learning and behavior in a controlled experimental setting?
I volunteered to work in Cross's laboratory and was delighted when he agreed. I was fascinated with the laboratory. In one study we examined the frustration effect; in another experiment we examined whether frustrative nonreward would stimulate alcohol intake. I worked in Cross's laboratory for three years before enrolling in graduate school. I counted myself lucky when Dr. Cross accepted me into his PhD program at Colorado State University.
Since becoming a faculty member at Southwestern University, I have been able to combine my interests in animal learning, animal behavior, comparative psychology, marine environments, and marine life. Since 1981, I have conducted experiments with a variety of aquatic animals including bass, bettas, carp, cichlids, crabs, cuttlefish, goldfish, koi, minnows, paddlefish, salmon, and shrimp,. The experiments have dealt with predator/prey interactions, basic learning phenomena including behavioral contrast and appetitive and aversive sign tracking, and problems in optimal foraging theory. I have also taken students to Northern Vancouver Island where we have observed killer whales and recorded their vocalizations related to foraging. During the summer of 1999, at the Bamfield Marine Station on Vancouver Island, British Columbia, we played back those sounds and others to salmon to determine their effects on salmon behavior. Plans are being made to extend the work in the summer of 2001.
I have come full circle since high school. I am not a marine biologist studying the behavior of marine animals, but I am a psychologist doing virtually the same thing. I am able to teach and conduct research in a laboratory, and when necessary for my personal sense of well being I am able to get away for 'field research.' A couple of months on an uninhabited island observing killer whales is good therapy. If you value nature, constantly wonder how things work, and are interested in questions about the brain and intelligence, comparative psychology may be for you. It also helps to have an abiding love and appreciation for nonhuman animals.
As one watches animals, it doesn't take long before one sees things that appear to reflect highly 'intelligent' behavior. In a lecture presented at Southwest Texas State University a number of years ago, Dr. Roger Fouts described the following conversation between a lab assistant and a chimpanzee that had learned sign language: As the lab assistant entered the chimp's living quarters, the chimp signed "Bring me orange." At this request, the assistant went to the refrigerator, opened the door, and discovered that there were no oranges. She returned to the chimp and signed "No orange." The chimp immediately signed with increased vigor, "Bring me orange." The lab assistant signed "No orange." The chimp, becoming agitated, signed again for an orange. The assistant, near despair, went to the refrigerator, opened it, looked at the chimpanzee and signed "No oranges." At this point, the chimpanzee signed, "You drive, bring me orange."
Although most of us have not conversed with chimpanzees, we have all observed or heard about episodes that appear to show remarkable intelligence in animals. Aristotle may have been the first to use anecdotes to argue for intelligence in nonhuman animals. In the History of Animals, he wrote: "In the neighborhood of Lake Marotis, it is said, wolves act in concert with the fishermen, and, if the fishermen decline to share with them, they tear the nets to pieces as they lie drying on the shores of the lake" (Barnes, 1984).
Anecdotes like these are fascinating, but what do they tell us about the cognitive abilities of nonhuman animals? Unfortunately, very little. Anecdotes provide captivating descriptions of behavior but the conclusions they imply are without proof. The wolves in Aristotle's story may have torn up the fishing nets for a number of reasons having little to do with revenge. Perhaps they were attracted by the smell of fish on the nets and chewed on the nets after a poor catch left them hungry. Only systematic experimentation with control groups can tell us why animals do the interesting things they do. The anecdotal method cannot accomplish that.
This chapter examines the approaches taken by scientists interested in comparative psychology and animal learning. Since both disciplines are rooted in Darwin's ideas about the evolution of morphological and psychological traits, we begin with a discussion of evolutionary theory and early considerations of the evolution of intelligence. We also describe the influence of ethology and Tinbergen's four questions. In the second part of the paper, we consider the current status of work in comparative psychology and animal learning. In the third and final section, we comment on where research on comparative psychology and animal learning are headed. Included in the last section is a bit of practical advice to students who are interested in going into comparative psychology and animal learning.
From Charles Darwin to C. Lloyd Morgan
Modern studies of comparative psychology and animal learning have their origins in the work of Darwin and Romanes in the late nineteenth century (see Bornstein, 1980; Gottleib, 1979). Both Darwin and Romanes rejected the assumption that human mental life was unique and argued instead for continuity among animal species. Darwin became convinced that nonhuman animals had the same emotions and mental abilities as people (see Darwin's The Descent of Man and Expression of Emotions in Man Animals). He proposed that the difference was quantitative not qualitative. Humans had more powerful intellects and a greater capacity for emotional restraint.
Romanes, who coined the term 'comparative psychology,' foresaw a field of psychological inquiry that compared the mental life of animals with human mental life (Romanes, 1884). Romanes saw a great future for comparative psychology. He believed that every known species of animal possessed intellect, and he argued that the actions of primitive species, in parallel with the homologues of biological structures, had their counterparts in the intelligent behaviors of more complex species, most notably in humans.
Romanes, and others before him, made liberal use of anecdotal evidence to support their claims. Unfortunately, the informality of anecdotes, and their inaccessibility to verification, makes them of little scientific value. Indeed, after the publication of Romanes's book Mental Evolution in Man, the use of the anecdotal method was seriously discredited (Waters, 1934). Guided by 'Morgan's Canon,' comparative psychologists adopted more rigorous empirical methods.
Morgan highlighted the dangers inherent in attempting to think about animals using anthropomorphic human terms. He pointed out that in comparisons between human and animal mental life, there was a tendency to ascribe to animals mental abilities that animals may not possess. He proposed a guiding principle to avoid this mistake: 'In no case may we interpret an action as the outcome of the exercise of a higher psychical faculty, if it can be interpreted as the outcome of the exercise of one which stands lower in the psychological scale' (Morgan, 1901, p. 53). Morgan's Cannon is one of the fundamental axioms of modern comparative psychology.
The Influence of Experimental Psychology
One of the major issues that interested early comparative psychologists was the evolution of intelligence. However, before they could study the evolution of intelligence, they had to decide what constitutes intelligence in animals. The definition early comparative psychologists settled on provided the impetus for the study of animal learning (see Domjan, 1987). Intelligence was defined in terms of behavioral plasticity and the ability to learn or to benefit from one's experiences. Armed with this definition, investigators began to study the evolution of intelligence by examining learning in nonhuman animals.
E. L. Thorndike was one of the first to apply a rigorous scientific methodology to the study of learning. For his doctoral dissertation, he invented the puzzle box technique, which allowed him to track the progress of learning by measuring changes in the latency of his subjects to escape from a box. Thorndike studied learning in cats, chickens, dogs, fish, and monkeys. In interpreting his findings, he rejected anthropomorphic concepts and formulated the Law of Effect, which explained learning in terms of the acquisition of S-R associations (Thorndike, 1911). Thorndike was one of the giants of early comparative psychology. The apparatus he developed for the study of learning (the puzzle box experiments), his strict methodology, and the negative conclusions he drew regarding the presence in animals of ideas, reasoning, and imitation learning all served to legitimize comparative psychology as science (Waters, 1934).
The Influence of Ethology
With research on learning, comparative psychology gained a predominantly experimental and laboratory flavor. However, comparative psychology also was heavily influenced by ethology (Jaynes, 1969). The term 'ethology' was coined in 1859 by Isidore Geoffroy-Saint-Hilaire, who advocated studying the behavior of animals in their natural habitats (Jaynes, 1969). This led ethologists to focus on naturalistic observations. The approach was systematized by Konrad Lorenz, Karl von Frisch, and Niko Tinbergen, who later shared the Nobel Prize for their efforts. Tinbergen (1951, 1963), in particular, spelled out the agenda that guides much of the contemporary research in comparative psychology.
The Ethogram. Tinbergen argued that the scientific study of animal behavior must begin with careful observation and description. We cannot explain why an animal acts in some fashion until we know the animal's ordinary activities. For Tinbergen, the description of behavior was a serious business. Its ultimate aim was 'an accurate picture of the patterns of muscle action' (Tinbergen, 1951, p. 7). Tinbergen championed the use of still and motion pictures, as well as careful study of the muscles involved in a particular behavior. In describing social interactions, he advocated concentrating on those behaviors in one animal that caused behavioral reactions in the other animal.
Ideally, description results in a complete inventory of the behavior patterns of a species. This is called an ethogram. The ethogram can then be used to formulate questions about the causes of behavior. However, Tinbergen cautioned against moving to causal issues before a complete ethogram had been obtained. A study by Baerends (1941) illustrates the strategy. Baerends observed that digger wasps dig a hole in the ground, kill or paralyze a caterpillar, place the caterpillar in the hole, and lay an egg on the caterpillar. Having laid an egg in one nest, the wasp repeats the entire sequence with other nests. After several nests have been set up, the wasp returns to her first nest where the larva has hatched and brings the larva additional caterpillars. The wasp then moves on to the second nest, and so on. Finally, the wasp returns again to the first nest and this time provides six or seven additional caterpillars. She then covers the nest and leaves it forever. The entire sequence is played out over an eight-day period, with the wasp keeping up with as many as five different nests.
How does the wasp know how much food to bring to each nest? Baerends changed the number of caterpillars in the nest by either adding or removing some and found that the number of caterpillars the wasp saw on her first visit of the day to a particular nest determined how many caterpillars she brought to that nest that day. If Baerends removed some of the caterpillars before the first visit, the wasp would bring extra ones; if Baerends added caterpillars, the wasp would bring fewer than the usual number. Changing the number of caterpillars after the first visit of the day did not influence the wasp's behavior.
This study attracted considerable attention because it showed that a wasp could remember what it saw on the first visit to a nest for at least 15 hours. This was remarkable because people at the time thought even orangutans could not remember things for more than about 5 minutes (Maier & Schneirla, 1935). Tinbergen was fond of the wasp study because it showed that with careful observations, and by studying an animal in its own habitat, one can discover remarkable mental abilities, even in the lowly wasp.
After observation and description of an animal's behavior under natural circumstances, one can begin to examine the causes of the behavior. Tinbergen (1951, 1963) pointed out that the causes of behavior can be approached from four different perspectives. These have come to be known as Tinbergen's four questions: 1) How does the behavior develop during the animal's life time, or what is the ontogeny of the behavior? 2) How does the animal's immediate environment activate the physiological mechanisms that generate the behavior, or what are the proximal mechanisms of the behavior? 3) How does the behavior contribute to survival and reproductive success, or what is the adaptive significance of the behavior? 4) What is the evolutionary history or phylogeny of the behavior?
Behavioral Ontogeny. The ontogeny question focuses on how behavior is determined or 'caused' by the progression of events that occur during an animal's life. Tinbergen considered three primary causal factors in this category: maturation, practice, and learning. Maturation refers to growth processes, in receptors, effectors, and the nervous system. Maturation can lead to the emergence of new forms of behavior without specific practice. In contrast, practice effects refer to improvements in an existing behavior brought about by mere repetition of that behavior (or of similar activities).
Learning involves a relatively permanent change in behavioral patterns that occurs as a result of experience. How learning phenomena interact with maturational and other processes was of particular interest to ethologists. Tinbergen (1951) reported, for example, that parent gulls will accept any chick that enters the nest during the first 5 days after their chicks hatch. After that, however, the parents will neglect chicks that were not in the nest earlier, and may even kill them. Evidently, parent gulls learn the identity of their offspring during the first 5 days after their chicks hatch.
Analogous individual recognition learning does not seem to occur during incubation. Although gull eggs differ in color, size, and speckling pattern, incubating hens will accept strange eggs with impunity. Learning during the incubation period seems to be focused on information identifying the spatial location of the nest. If an experimenter removes the eggs from an established nest and places them a few feet away, the gull may continue to sit on the empty nest in full view of the displaced eggs.
The fact that gulls learn about the spatial cues of their nest but not the identity of their eggs makes sense in terms of their natural habitat. During the incubation phase, the hen leaves the nest periodically to obtain food or to scare off or decoy predators, after which she has to be able to find her way back. Ordinarily the location of the eggs and the nest does not change. On the other hand, once the chicks hatch, they soon become mobile. Therefore, the hen has to be able to distinguish her chicks from those of neighboring hens. These differences in what the gulls learn about at different stages of their reproductive cycle illustrate what Tinbergen called special 'dispositions to learn' (Tinbergen, 1951, p. 145). Such specializations in learning remain of considerable contemporary interest (e.g., Domjan, 1997).
Proximate Mechanisms or Immediate Causation. The second question posed by Tinbergen concerns immediate causation or the proximate mechanisms of a behavior. To identify immediate causes, it is important to know what external and internal stimuli are available to the animal. Weakly electric fish, for example, are able to produce and perceive slight changes in an electrical field. These animals use changes in electrical fields to detect prey, conspecifics, and mates, and to regulate their dominance hierarchies (Hopkins, 1977). Most other organisms, including humans, cannot perceive these subtle changes in electric fields. Given the contrasting sensory worlds of diverse species, Tinbergen (1951) noted that investigations of sensation and perception are critical to the study of immediate causation.
Numerous investigations have addressed the external causes of behavior. Tinbergen referred to one class of stimuli that cause behavioral patterns to be emitted as sign stimuli or releasing stimuli. Such stimuli release species typical response patterns. For example, Gallup (1974) examined tonic immobility, which appears to be an innate response that evolved as a defense against predation (Ratner, 1967). During a period of tonic immobility, or death feign, the animal becomes immobile in response to some aspect of the predator. Gallup and his co-workers were able to show that the death feign response in chickens was released primarily by visual presentation of the eyes of a predator. Even artificial eyes mounted on wooden dowels and directed at young chickens produced the tonic immobility response (Gallup, Nash, & Ellison, 1971). The eyes of a chicken hawk, the red breast of a robin, the red belly of a stickleback fish are all examples of external stimuli that elicit certain behaviors.
Adaptive Function. Tinbergen viewed organisms as living in an extremely unstable state and advised scientists to examine how particular behavior patterns aid survival. This question is not as easy as it seems at first glance. Different species solve problems related to survival in different ways. For example, pigeons drink water by sucking whereas most other birds drink water by scooping. Both behaviors are adaptive, and one is no better than the other in an absolute sense. The difference between them is best explained by the evolutionary history of the species involved. Sucking and scooping are convergent solutions based on different evolutionary histories. Tinbergen (1951) suggested that the 'study of convergences helps to show up adaptiveness.'
Tinbergen (1963) distinguished between questions of adaptive significance and immediate causation in the following manner. Immediate causation examines the preceding events that contribute to a behavior. In studying immediate causation, the researcher looks 'back in time' to determine why an animal behaves in a particular fashion. In contrast, in studying adaptive function, the researcher looks 'forward in time' to determine how a particular behavior contributes to the animal's survival or reproductive success. Questions of adaptive significance are more difficult to answer because the researcher has to compare how one behavior, structure, or learned event actually increases reproductive success or survival over another behavior, structure, or learned event.
Evolutionary History. Tinbergen (1951) noted that studies of the evolution of behavior have lagged far behind studies of the evolution of morphological traits. This is probably because describing behavior is more difficult than describing morphology. Distinguishing between innate differences in behavior versus acquired differences is often also difficult, and behavior is not well represented in the fossil record. However, studying fossilized animal tracks (Seilacher, 1967), examining the structure and type of environment in which the fossil was found, and identifying fossilized bones found in the stomachs of fossilized animals can tell us much about the behavior patterns of ancient organisms.
The evolutionary origins of a behavior also may be uncovered by studying behavioral homologies. Tinbergen advocated comparing behaviors in one species with corresponding behaviors in another species, and then determining whether the similarities reflect a common genetic background or convergent evolution. Lorenz provided careful descriptions of behavior in birds and showed how such descriptions could be used to develop classification systems (Lorenz, 1971). Other approaches to uncovering the evolutionary origins of a behavior involve studying behavior genetics, relations between mutations and behavior, and the formation of new species (speciation) and behavior.
Where have these historical antecedents led the study of comparative psychology and animal learning? If we were to ask particular investigators, the answers we would get would be colored by their individual perspectives and biases. A more even-handed description might emerge from a content analysis of what is being published in comparative psychology and animal learning. We elected that approach and examined the articles that appeared from 1990 to 2000 in a prominent journal that contains primarily research on comparative psychology (the Journal of Comparative Psychology or JCP) and a journal that contains primarily research on animal learning (the Journal of Experimental Psychology: Animal Behavior Processes or JEP:ABP). Both of these journals are published by the American Psychological Association and have a long and distinguished history that goes back more than a hundred years. Although they are not the only journals in which research relevant to comparative psychology and animal learning may be found, JCP and JEP:ABP publish primarily the work of psychologists (rather than zoologists or biologists) and therefore provide a good window on the activities of psychologists in these fields.
Perusing current research in comparative psychology and animal learning, one is immediately struck by the wide range of issues that are addressed. There are virtually no limits to what aspect of behavior a particular scientist may choose to study. As a starting point for our content analysis, we were guided by Tinbergen's questions of ontogeny, immediate causation, adaptive function, and evolutionary history. However, we deviated a bit from his definition of these categories to be more consistent with current usage. In contemporary thinking, studies of learning are considered to involve questions of immediate causation rather than ontogeny. We followed this convention and categorized studies of learning as relevant to immediate causation unless an explicit developmental issue was being addressed.
Table 1 summarizes the proportion of the papers published in JCP and JEP:ABP that were devoted to questions of ontogeny, immediate causation, adaptive function, and evolutionary history. Modern studies of comparative psychology focused primarily on behavioral ontogeny and immediate causation. Only 4% of the papers in JCP addressed the adaptive function question, or how a particular behavior contributes to survival and reproductive fitness. Nearly three times as many papers (11%) address the question of evolutionary history. In contrast, studies of animal learning, as exemplified by the research published in JEP:ABP, dealt primarily with questions of immediate causation. Ontogenetic investigations and studies with a focus on adaptive function or evolutionary history hardly ever appeared. Overall, comparative psychologists examined a wider range of activities than students of animal learning and were more concerned with the ontogenetic and phylogenetic context of behavior. In contrast, students of animal learning concentrated on how a behavior is controlled by recently experienced events.
Specific Issues and Examples
Ontogeny. In JCP, most of the studies of behavioral ontogeny focused on either cognition (30%) or social behavior (51%). Interestingly, 70% of the studies of cognitive development involved primates (apes, humans, or monkeys). Investigators reported on issues of cognitive development raised by Piaget, such as ontogeny of sensorimotor intelligence, object permanence, and conservation. Research on the ontogeny of social behavior addressed issues of agonistic behavior in mice, maternity (maternal attachment, parental care by the mother, and maternal aggression) and communication within a social context (song development in birds, affiliative vocalizations in birds and monkeys, grunt communication in humans, and whistle contour development in dolphins).
One of the studies of ontogeny published in JCP characterized the development of tool use in infant chimpanzees (Inoue-Nakamura & Matsuzawa, 1997). These investigators observed chimpanzees in the field at Bossou, Guinea, from 1992 to 1995. On 692 occasions they saw an infant manipulate a nut with a stone. They also recorded 150 episodes of an infant observing an adult manipulate a nut with a stone. The results showed that the wild chimpanzees began to use hammer and anvil stones to crack open oil-palm nuts at three and a half years of age. The infant chimps demonstrated mastery of all the basic actions necessary for nut cracking by two and a half years of age, but they did not combine the responses in an appropriate sequence until they were a year older.
In another study of behavioral ontogeny, Hanson and Coss (1997) examined age differences in the response of California ground squirrels (Spermophilus beecheyi) to predators. California ground squirrels encounter a variety of predators including snakes, mammals (badgers, dogs, coyotes), and birds (hawks and eagles). The degree of danger presented by these predators differs depending on age. For example, adult ground squirrels are in less danger from snakes than infants due to their ability to physically retaliate and their strong resistance to rattlesnake venom.
Hanson and Coss (1997) asked whether the antipredator responses to bird and mammal predators would differ according to age. Juvenile and adult ground squirrels were videotaped responding to a brief presentation of a live dog or a model of a red-tailed hawk in simulated flight. The juveniles did not differentiate between the bird and mammal predators and responded to both as dangerous. In contrast, the adults treated the hawk as of more immediate danger than the dog. In a second experiment, juveniles and adults were presented models of red-tailed hawks, nonthreatening turkey vultures, and crows. At neither age did the ground squirrels discriminate among the bird predators, though adults treated the aerial predators as more dangerous than the juveniles. The authors concluded that learning might play a role in the anti-predatory behavior of California ground squirrels.
Immediate Causation. The second question proposed by Tinbergen concerned immediate causation. More of the articles we looked at dealt with immediate causation than with any of the other four categories of research questions. However, the emphasis on studies of immediate causation was greater in JEP:ABP than in JCP.
Immediate causal mechanisms can be examined for any behavior. Indeed, the studies of immediate causation covered a remarkably large range of behavioral phenomena.
A list of the topics and the number of papers relevant to each topic that we encountered in studies of immediate causation in JCP and JEP:ABP are presented in Table 2. Both comparative and learning psychologists studied perception/attention, memory mechanisms, basic learning processes, and complex cognition. However, there were differences in emphasis. As might be expected, learning and memory phenomena were addressed in much greater detail in JEP:ABP than in JCP. In addition, studies in JEP:ABP dealt with forms of learning (such as drug conditioning and conditioned changes in pain sensitivity) that were not examined by comparative psychologists. In contrast, comparative psychologists examined a much wider range of behaviors. These included various aspects of reproductive behavior, social behavior, foraging, as well as defensive behavior and aggression.
Another interesting point of contrast between comparative psychologists and investigators of animal learning is that comparative psychologists were more likely to tackle forms of cognition that have been primarily associated with human beings. These included topics such as perspective taking, theory of knowledge, and imitation. An important task in studies of these and other forms of animal cognition is to isolate the factors that control the animal's behavior. Only if the immediate causes of the behavior are clearly identified can one decide whether the observed behavior reflects complex cognitive processes or simpler associative and perceptual mechanisms. Studies of imitation clearly illustrate this problem.
In 1949, Fisher and Hinde reported on birds (blue tits) that had been robbing cream from milk bottles delivered to the doors of English homes. Observers speculated that a single blue tit had discovered, probably quite by accident, how to peck through a bottle cap and consume the rich cream that floated on top of the milk. The occurrence of milk theft then gradually spread throughout the whole of England and into Europe through some form of social learning.
Fisher and Hinde argued that animals can acquire information through social interaction and that acquisition of such information may prove to be adaptive. The question is what exactly is being transmitted? Is information about stimuli being transmitted or are the animals learning about responses? Did the first bird to open a milk bottle transmit information about the stimulus (milk bottles are predictive of food) or about the response required to open the bottle (when you encounter these types of objects, open the bottle in the following manner)?
Social learning that occurs because of increased attention to particular stimuli is called 'local enhancement.' Local enhancement was defined by Thorpe (1956) as 'an apparent imitation that resulted in one animal directing another animal's attention to a particular object or to a particular part of the environment' (p. 134). The cream theft behavior exhibited by blue tits may have been an instance of local enhancement in that the behavior of one blue tit served to direct the attention of a naive bird to the milk bottle. Once drawn to the milk bottle, the naive bird may have learned to open it without paying attention to how another bird did that.
Local enhancement is difficult to rule out in studies of imitative learning because the imitated response usually occurs in the context of distinctive cues that may attract more attention when the demonstrator performs its prescribed response. A particularly effective procedure for controlling local enhancement effects is called the 'two-action method'. In the two-action method, different demonstrators respond to a stimulus in distinctive ways. In a recent study, for example, Akins and Zentall (1996) trained domesticated quail to manipulate a small platform (3.8 cm square) by either pecking it or stepping on it. Once the demonstrators learned to either peck or step on the treadle, one group of observers was allowed to watch a demonstrator peck the treadle, and a second group was allowed to watch a demonstrator step on the treadle. The purpose of the experiment was to determine whether observers acquired the specific response they saw their demonstrator perform.
The two-action method assumes that local enhancement will occur with either demonstrated response. In the Akins and Zentall study, pecking the treadle and stepping on it should have drawn attention to the treadle equally. Thus, if seeing the demonstrators manipulate the treadle produced local enhancement, it should not have mattered whether the demonstrators pecked or stepped on the treadle. Contrary to that prediction, however, the response of the observers matched the response form they had seen their demonstrator perform. Quail that observed demonstrators peck the treadle were more likely to peck than to step on the treadle, and quail that observed demonstrators stepping on the treadle were more likely to do that than to peck. This shows that imitative learning can occur independent of local enhancement.
Adaptive Function. Psychologists rarely examined the question of adaptive function - how a particular behavior increased an animal's chances of survival and reproductive fitness. Only 20 papers in JCP focused on this issue and none examined adaptive significance in JEP:ABP. In JCP, the majority of the papers on adaptive function examined reproductive fitness. Investigations addressed paternity advantages accrued from conditioned stimuli, preference for phenotypical variation in mates, acoustic behaviors related to increased survivability and mating success, the strange-male effect, and unusual patterns of copulatory behavior. Other papers examined investigation behavior in snakes, predatory and anti-predatory behavior in snakes and mice, and cues used to increase foraging efficiency in monkeys who forage at night.
In one of the most dramatic studies of reproductive behavior, Karen Hollis and her undergraduate students at Mount Holyoke College asked whether stimuli that reliably predict events of biological significance (classically conditioned stimuli) lead to greater survival or reproductive success (Hollis, Pharr, Dumas, Britton, & Field, 1997). They worked with blue gouramis (Trichogaster trichopterus), a freshwater fish that lives in the tropics. Male blue gourami are territorial and aggressively defend their nest sites. Often this aggressiveness interferes with their ability to attract females and mate because males chase off females as well as males. To determine whether classically conditioned stimuli could have adaptive value in this situation, Hollis presented male blue gourami with a signal followed by presentation of a receptive female fish. Subjects in a second group also received presentations of the stimulus and a receptive female, but these were unpaired.
Following a short training phase, the males were given an opportunity to mate. For one group, the arrival of a female was signaled by presentation of the classically conditioned stimulus. For the other group, the arrival of the female was not signaled. Signaling the arrival of the female had multiple effects. First, males that received the Pavlovian signal showed less aggression towards the females compared to the males in the control group. Second, the classically conditioned males spawned with the females sooner, clasped the females more often, and most importantly, produced far more young than did males in the control group. This experiment provided the first direct demonstration that classically conditioned stimuli can significantly contribute to reproductive success and supported Hollis's contention that classical conditioned stimuli have adaptive value.
Evolutionary History. Perhaps the most difficult of Tinbergen's questions to get evidence for concerns the evolutionary history of a behavior. The evolution of behavior is rarely addressed by any single research report. Rather evolutionary issues are discussed in papers that attempt to integrate the results of studies with various species. As Table 1 indicates, such papers appeared primarily in JCP. Of the 53 papers in JCP concerned with ultimate causation, 32 dealt with laterality or handedness and were conducted with primates. Other papers concerned the heritability of behavior, the evolutionary history of tool use, evolution of foraging and defense behavior, and evolution of communication signals.
What is the point of studying laterality in primates? Comparative psychologists study handedness in an effort to obtain information about the evolution and development of complex cognition. Annett (1985) reported that approximately 90% of the human population is right-handed. Because movements of the right hand are controlled by the left side of the brain, right-handedness reflects a left hemisphere specialization for manual control. Language is also controlled by the left side of the brain. These observations have been used to argue that left-hemispheric specialization for manual control played an important role in the evolution of language and other cognitive functions (Calvin, 1994). The possible connection between handedness and complex cognition has encouraged comparative psychologists to study the extent and the nature of handedness in animals and to use that information to answer questions related to the evolution of intelligence.
Westergaard and Suomi (1996) of the National Institute of Child Health and Human Development examined hand preference in tufted capuchins (Cebus apella) and rhesus macaques (Macaca mulatta). Plastic tubes lined with food were presented to the monkeys. To get the food, the monkeys had to hold the plastic tube with one hand and dig out the food with the other. The tests showed a population-level bias toward use of the right hand in rhesus monkeys but not in capuchin monkeys. In addition, the capuchin monkeys showed greater hand preference on an individual basis than did the rhesus monkeys. Finally, for capuchins, but not for rhesus monkeys, the hand preference was greater for adults than for immature animals.
In a recent follow up, Westergaard, Kuhn, and Suomi (1998) studied the influence of bipedal posture on hand preference in humans and rhesus monkeys. Humans and rhesus monkeys were observed reaching for objects from either a quadrapedal or bipedal posture. Human participants demonstrated a right-hand population-level preference for the right hand regardless of whether they reached from a quadrapedal or bipedal posture. Rhesus macaques showed a left hand population-level preference when reaching from the quadrapedal posture, but a significant shift toward the right hand when reaching from the bipedal posture. When the authors examined preferences for 10 primate species, they discovered that right hand preferences appear to be correlated with bipedal reaching. The authors conclude with the interesting idea that bipedalism 'may have facilitated species-typical right-handedness in humans.'
In 1951 Tinbergen denounced comparative psychology as being too narrow in its focus on learning and in its use of a small number of laboratory species (see also Beach, 1950). He was not amused by the fact that Tolman dedicated his famous 1932 book Purposive Behavior in Animals and Men to the albino rat. Tinbergen commented that 'in spite of the high respect deserved by the interesting work done with rats, one should be a little skeptical of the laboratory rat as a representative of the whole animal kingdom' (Tinbergen, 1951, p. 11.).
Tinbergen's criticism was justified at the time but does not apply to contemporary work in comparative psychology. Between 1990 and 2000, articles in JCP involved research with 132 different species. Each year, an average of 33 different species served in comparative experiments. The most frequently used species were rats (Rattus norvegicus), chimpanzees (Pan troglodyte), humans (Homo sapiens), rhesus macaques (Macaca mulata), capuchin monkeys (Cebus apella), pigeons (Columba livia), and honeybees (Apis mellifera). Primates (humans, apes, and monkeys) were the subjects of investigation in 36% of the papers published in JCP. Studies with birds (23%) and rodents (22%) constituted the next most popular choices. The diversity of species employed was remarkable. From earthworms (Lumbricus terristris) and cuttlefish (Sepia officinalis) to bonobos chimpanzees (Pan paniscus), comparative psychologists are using many different kinds of animals.
The diversity of species that we found in JCP did not carry over to JEP:ABP. In JEP:ABP, primates (including humans) were the subject of 11% of the research reports. In contrast, rats and pigeons accounted for 78% of the papers. However, since 1990 JEP:ABP has also included studies with zebra finches, dolphins, Clark's nutcrackers, chickadees, ferrets, gerbils, honey bees, humming birds, junkos, scrub jays, starlings, and quail.
The Future of Comparative Psychology and Animal Learning
Coming Back Together
Although comparative psychology and animal learning had common origins in the writings of Darwin and Romanes, the two areas subsequently developed in different directions. Animal learning came to focus on the mechanisms of behavior and invested much of its energy in detailed exploration of a limited number of species and a small number of experimental preparations. In contrast, comparative psychology was moved by the influence of ethologists to consider a broad range of behaviors and species, using a variety of experimental techniques. However, investigators of both comparative psychology and animal learning have preferred the laboratory to the field. Between 1990 and 2000, only nine papers in JCP and none in JEP:ABP focused on field observations. Interestingly, of the nine papers, seven were from 1998 and four of these papers were published in 2000. Perhaps we are seeing the beginning of a trend toward greater use of field observation in comparative psychology.
We predict that the future will see comparative psychology and animal learning come back together again. Investigators of animal learning will increasingly turn their attention to the wide range of behaviors that comparative psychologists have been studying. This will lead to more studies of how learning is involved in various forms of social behavior, such as sexual behavior, maternal and paternal care, kin selection, dominance hierarchies, territorial behavior, and predator/prey interactions.
Both comparative psychology and animal learning will continue to be concerned with issues of behavioral evolution. In particular, studies of behavioral evolution should get a special boost from technological advances in biology such as gene splicing and DNA fingerprinting. In addition, we hope that studies of animal learning will include a broader range of species, and species specifically selected to highlight adaptive specializations and evolutionary influences on learning.
It will continue to be important for investigators in comparative psychology and animal learning to be sensitive to how their laboratory paradigms are related to the ecology of their species. As Tinbergen (1951, 1963) pointed out, it is important to know the behavioral patterns of the species being studied. Naturalistic observation can inform investigators of what aspects of behavior can profitably be investigated to address questions related to learning and cognition. Naturalistic observation is also a rich source of hypotheses for later laboratory investigations.
One of the dramatic findings from our content analysis was the relatively little effort being devoted to studies of how certain behaviors or behavioral mechanisms promote survival and reproductive fitness (see Table 1). We hope that studies of the adaptive functions of behavior will attract more of the attention of both comparative psychologists and students of animal learning in the coming years. Tinbergen's four questions were proposed nearly half a century ago and all of them have become well accepted as necessary for a comprehensive account of the causes of a behavior. It may be time to put these sentiments more fully into practice by devoting more effort to studying the functional significance of the behaviors that attract our attention.
Some of the trends we are predicting are already underway. Contemporary interest in mechanisms of memory and other forms of animal cognition are evident in recent issues of JEP:ABP. This represents a shift away from associationistic problems to broader issues of comparative cognition. Some investigators of learning are also starting to move away from the simple lights and tones that were commonly used in traditional studies of learning to more complex, and presumably more ecologically valid, stimuli (Cook, Cavoto, Katz, & Cavoto, 1997; Fetterman, 1996; Spetch, 1995).
Another trend that will be evident in comparative psychology and animal learning is greater interest in the physiological mechanisms of behavior. This in turn will lead to more interdisciplinary research. Traditional boundaries between different fields of science (chemistry, biology, physics) are eroding, and this is accompanied by greater interest in investigating phenomena at multiple levels of analysis. Psychological phenomena like cognition and learning can be described at the level of behavior, or at the level of neural structures and networks, neurotransmitter systems, or cellular mechanisms. In the past, individual investigators worked primarily at one or another level of analysis. It is now time to put all the parts and levels back together. We need to know how events occurring at the cellular level influence the operation of neural systems and how neural networks determine observable behavioral actions. Increasingly investigators will be working in interdisciplinary teams to create an integrated model of animal learning and behavior.
Two Nagging Unresolved (Unresolvable?) Problems
The evolution of cognition and learning has been difficult to study because as one traces back an evolutionary lineage, at some point one has to demonstrate the absence of a phenomenon or skill. However, the failure to find a particular cognitive skill may say more about the ineptitude of the experimenter than the ineptitude of the animal being studied. As experimenters have become more skillful in examining the behavioral potentials of animals, the documented skills of the animals have also increased. For example, early comparisons of pigeons and monkeys suggested that pigeons cannot learn a generalized same-different concept but monkeys can. However, with subsequent improvements in behavioral technology, pigeons have been shown to be perfectly capable of the task (e.g., Wright, Cook, Rivera, Sands, & Delius, 1988). Thus, as the experimenters have become smarter in examining the cognitive capacities of pigeons, they have found pigeons to seem smarter as well.
Interpretation of the evolutionary lineage of a particular cognitive or learning capacity also has been hampered, ironically, by increases in theoretical sophistication. In some cases, seemingly 'intelligent' behavior has turned out to be mediated by fairly elementary associative mechanisms. Recent analyses of the phenomenon of transitive interference illustrate this type of development.
Transitive inference involves forming a mental representation of a sequence of stimuli based on training with only two of the stimuli at a time. Consider for example, five stimuli, which we could label A, B, C, D, and E. During training trials, subjects are exposed to one pair of stimuli at a time. The training pairs are A-B, B-C, C-D, and D-E. Which stimulus is presented on the right or the left in the stimulus display is varied randomly across trials. With each stimulus pair, the subject is reinforced (with food, for example) for selecting the stimulus that is closer to the beginning of the alphabet. Thus, in a choice between A and B, the subject is reinforced for choosing A, but in a choice of B and C, the subject is reinforced for choosing B.
With this kind of training, stimuli (B, C, and D) appear equally often as reinforced and non reinforced stimuli. The main point of the experiment is to determine how the subjects will respond during a test trial when they encounter a novel pair of stimuli, B vs. D for example. Based on how often stimuli B and D were reinforced during training, there should be no basis for selecting between them. However, if the subject learned that the stimuli form a linear sequence A ® E, it should select B over D. Such a choice is considered to be evidence of 'transitive inference' and is presumably based on the subject having extracted the entire stimulus sequence A ® E from its experience with the stimulus pairs during training.
Transitive inference was first documented in chimpanzees and people and was interpreted as demonstrating a complex cognitive skill. More recently, however, transitive inference has been demonstrated in pigeons as well, and this work showed that such seemingly complex cognitive performance can in fact be explained by the fairly elementary associative mechanism of value transfer (Couvillon & Bitterman, 1992; von Fersen, Wynne, Delius, & Staddon, 1991; Zentall & Sherburne, 1994). According to value transfer, a non reinforced stimulus gains some 'value' or associative strength by stimulus generalization if it occurs in combination with another cue that is reinforced. This mechanism predicts that subjects will select B over D following training with adjacent stimulus pairs in a transitive inference procedure because the training pairs permit more transfer or generalization of associative value to stimulus B than to stimulus D.
The history of research on transitive inference illustrates how pigeons have become more capable as experimenters have become more sophisticated in the design of their experiments. This history also illustrates that a phenomenon that appears to reflect the operation of a fairly complex cognitive process (transitive inference) may in fact be produced by simple associative mechanisms once those associative mechanisms are successfully identified. Clever Hans was an early example of cleverness that upon further analysis turned out to be not so clever (see Pfungst, 1965). The future will no doubt continue to challenge us with examples of complex cognition that turn out to be not so complex.
Keeping the Torch Burning
Comparative psychology and animal learning originated as areas of pure research. Some of the findings have formed the basis for applications to education and behavior therapy, but comparative psychology and animal learning continue primarily as scholarly and intellectual endeavors. Students seeking a career in these areas are pretty much restricted to positions at colleges, universities, and research institutes. Unfortunately, not many positions explicitly calling for someone in comparative psychology or animal learning become available each year. Students are more likely to succeed in obtaining employment if they also have strong skills in experimental design, statistics, and various laboratory procedures so that they can fulfill the requirements for positions in general experimental psychology.
Experimentation in comparative psychology and animal learning is expensive and most of it is funded by government agencies. Because of this, future investigators are well advised to hone their writing skills, learn as much as they can about potential sources of funding, and gain experience in writing grant applications.
Whether young or old, all investigators have to justify their work to be successful. One approach to justifying comparative psychology and animal learning is to argue that knowledge in these areas is essential for other important lines of inquiry. For example, one cannot begin to examine the physiological basis of a behavior without first studying the behavior itself. This requires determining how to measure various aspects of the behavior and figuring out how these behavioral parameters are governed by procedural and stimulus variables. Comparative psychology and animal learning provide the behavioral technology that is the foundation for physiological investigations of brain-behavior relationships. Behavioral work with animals is also essential for the development and testing of psychopharmacological agents, for understanding how commonly abused drugs affect behavior, and for research on the psychological effects of toxins in the environment and in food.
Research in comparative psychology
and animal learning is also easy to justify for its own sake. Insight into comparative psychology
does not lead to wealth, but it leads to wisdom. The comparative method is essential
for our understanding of just about everything. One cannot appreciate the size of the Earth without
comparing it to other planets.
One cannot appreciate the warmth of a sunny day except
in comparison to the cold of a winter day.
And, most importantly in the present context, one cannot
comprehend what it is to be human without understanding nonhuman
animal life. Just as astronomy is essential for us to appreciate
how special planet earth is in the greater scheme of celestial
objects, solar systems, and galaxies, so too comparative psychology
is essential for us to understand the place of humanity in the
broader context of animal life and living things in general. As long as we want to know about
ourselves, we will have to continue learning about other animals
Akins, C. K., & Zentall, T. R. (1996). Imitative learning in male Japanese quail (Coturnix japonica) using the two-action method. Journal of Comparative Psychology, 110, 316-320.
Annett, M. (1985). Left, right, hand and brain: The right shift theory. Hillsdale, NJ: Lawrence Erlbaum Inc.
Baerends, G. P. (1941). Fortpflanzungsverhalten und orientierung der grabwespe Ammophila campestris. Jur. Tijdschr. V. Entomol., 84, 68-275.
Barnes, J. (1984). The Complete Works of Aristotle (Vol. 1). Princeton, NJ: Princeton University Press.
Beach, F. A. (1950). The snark was a boojum. American Psychologist. 5, 115-124.
Bornstein, M. H. (1980). Part 1: Principal comparisons. In M. H. Bornstein (Ed.), Comparative Methods in Psychology (pp. 9-15). Hillsdale, NJ: Lawrence Erlbaum.
Calvin, W. H. (1994). The emergence of intelligence. Scientific American, 271, 521-536.
Cook, R. G., Cavoto, B. R., Katz, J. S., & Cavoto, K. K. (1997). Pigeon perception and discrimination of rapidly changing texture stimuli. Journal of Experimental Psychology: Animal Behavior Processes, 23, 390-400.
Couvillon, P. A., & Bitterman, M. E. (1992). A conventional conditioning analysis of transitive inference in pigeons. Journal of Experimental Psychology: Animal Behavior Processes, 18, 308-310.
Darwin, C. (1871/1965). The Descent of Man, and Selection in Relation to Sex. Chicago: University of Chicago Press.
Darwin, C. (1873). On the Expression of the Emotions in Man and Animals. New York:, D. Appleton.
Domjan, M. (1987). Comparative psychology and the study of animal learning. Journal of Comparative Psychology, 101, 237-241.
Domjan, M. (1997). Behavior systems and the demise of equipotentiality: Historical antecedents and evidence from sexual conditioning. In M. E. Bouton and M. S. Fanselow (Eds.) Learning, motivation, and cognition. (pp. 31-51). Washington, D.C.: American Psychological Association.
Fetterman, J. G. (1996). Dimensions of stimulus complexity. Journal of Experimental Psychology: Animal Behavior Processes, 22, 3-18.
Fisher, J., & Hinde, R. A. (1949). The opening of milk bottles by birds. British Birds, 42, 347-357.
Gallup, G. G., Jr. (1974). Animal hypnosis: Factual states of a fictional concept. Psychological Bulletin, 81, 836-853.
Gallup, G. G., Jr., Nash, R. F., & Ellison, A. L. (1971). Tonic immobility as a reaction to predation: Artificial eyes as a fear stimulus for chickens. Bulletin of the Psychonomic Science Society, 23, 79-80.
(1979). Comparative psychology and ethology.
In E. Hearst (Ed.), The First Century of Experimental
Psychology (pp. 147-173). Hillsdale, NJ: Lawrence Erlbaum.
Hanson, M. T., & Coss, R. G. (1997). Age differences in the response of California ground squirrels (Spermophilus beecheyi) to avian and mammalian predators. Journal of Comparative Psychology, 111, 174-184.
Hollis, K. L., Pharr, V. L., Dumas, M. J., Britton, G. B., & Field, J. (1997). Classical conditioning provides paternity advantage for territorial male blue gouramis (Trichogaster trichopterus). Journal of Comparative Psychology, 111, 219-225.
Hopkins, C. D. (1977).
In T. A. Sebeok (Ed.) How
Animals Communicate (pp. 263-289). Bloomington, IN: Indiana University Press.
Inoue-Nakamura, N., & Matsuzawa, T. (1997). Development of stone tool use by wild chimpanzees (Pan troglodytes). Journal of Comparative Psychology, 111, 159-173.
Jaynes, J. (1969). The historical origins of Ethology and Comparative Psychology. Animal Behavior, 17, 601-606.
Lorenz, K. (1971). Comparative studies of the motor patterns of anatinae (1941). (Translated by Robert Martin, in Studies in Animal and Human Behavior, 2. , pp. 14-18, 106-14. Cambridge: Harvard University Press.
Maier, N. R. F. & Schneirla, T. L. (1935). Principles of Animal Psychology. New York: London.
Morgan, C. L. (1901). An Introduction to Comparative Psychology. New York: Scribner's.
Pfungst, O. (1965). Clever Hans: The horse of Mr. Von Osten. New York: Holt, Rinehart, and Winston.
Ratner, S. C. (1967). Comparative aspects of hypnosis. In J. E. Gordon (Ed.) Handbook of Clinical and Experimental Hypnosis (pp. 550-587). New York: Macmillan
Romanes, G. (1884). Animal Intelligence. New York: Appleton.
Romanes, G. (1884). Mental Evolution in Animals. New York: AMS Press.
Seilacher, A. (1967). Fossil behavior. Scientific American, 217, 72-80.
Spetch, M. L. (1995). Overshadowing in landmark learning: Touch-screen studies with pigeons and humans. Journal of Experimental Psychology: Animal Behavior Processes, 21, 166-181.
Thorndike, E. L. (1911). Animal Intelligence: Experimental studies. New York: Macmillan.
Thorpe, W. H. (1956). Learning and instinct in animals. London: Methuen.
Tinbergen, N. (1951). The Study of Instinct. Oxford: The Clarendon Press.
Tinbergen, N. (1963). On aims and methods of ethology. Zeitschrift fur Tierpsychologie, 20: 410-433. Reprinted in L. D. Houck and L. C. Drickamer (Eds.) Foundations of Animal Behavior: Classic papers with commentaries (pp. 114-137). Chicago, IL: The University of Chicago Press.
Von Fersen, Wynne, C. D. L., Delius, J. D., & Staddon, J. E. R. (1991). Transitive inference in pigeons. Journal of Experimental Psychology: Animal Behavior Processes, 17, 334-341.
Waters, R. H. (1934). The historical background of comparative psychology. In F. A. Moss (Ed.) Comparative Psychology (pp. 9 ' 36). New York, NY: Prentice-Hall, Inc.
Westergaard, G. C., Kuhn, H. E.., &
Suomi, S. J. (1998). Bipedal
posture and hand preference in humans and other primates.
Journal of Comparative Psychology, 112, 55-64.
Westergaard, G. C., & Suomi, S. J. (1996). Hand preference for a bimanual task in tufted capuchins (Cebus apella) and rhesus macaques (Macaca mulatta). Journal of Comparative Psychology, 110, 406-411.
Wright, A. A., Cook, R. G., Rivera, J. J., Sands, S. F., & Delius, J. D. (1988). Concept learning by pigeons: Matching-to-sample with trial-unique video picture stimuli. Animal Learning & Behavior, 16, 436-444.
Zentall, T. R., & Sherburne, L. M. (1994). Transfer of value from S+ to S- in a simultaneous discrimination. Journal of Experimental Psychology: Animal Behavior Processes, 20, 176-183.