Melissa Kaplan's
Herp Care Collection
Last updated January 1, 2014

Cognitive Ethology: Slayers, Skeptics and Proponents

Marc Bekoff and Colin Allen. 1997. In: Anthropomorphism, Anecdotes and Animals: The Emporers New Clothes? RW Mitchessl, N. Thompson and L. Miles, editors. State University Press of New York State, NY. 313-334ity

The interdisciplinary science of cognitive ethology is concerned with claims about the evolution of cognitive processes. Since behavioral abilities have evolved in response to natural selection pressures, ethologists favor observations and experiments on animals in conditions that are as close as possible to the natural environment where the selection occurred. No longer constrained by psychological behaviorism, cognitive ethologists are interested in comparing thought processes, consciousness, beliefs, and rationality in nonhuman animals (hereafter animals). In addition to situating the study of animal cognition in a comparative and evolutionary framework, cognitive ethologists also maintain that field studies of animals that include careful observation and experimentation can inform studies of animal cognition, and that cognitive ethology will not have to be brought into the laboratory to make it respectable. Furthermore, because cognitive ethology is a comparative science, cognitive ethological studies emphasize broad taxonomic comparisons and do not focus on a few select representatives of limited taxa. Cognitive psychologists, in contrast to cognitive ethologists, typically work on related topics in laboratory settings, and do not emphasize comparative or evolutionary aspects of animal cognition. When cognitive psychologists do make cross-species comparisons, they are typically interested in explaining different behavior patterns in terms of common underlying mechanisms; ethologists, in common with other biologists, are often more concerned with the diversity of solutions that living organisms have found for common problems.

Many different types of research fall under the term "cognitive ethology," and it currently is pointless to try to delimit the boundaries of cognitive ethology; because of the enormous amount of interdisciplinary interest in the area, any stipulative definition of cognitive ethology is likely to become rapidly obsolete. Currently, cognitive ethology faces challenges to its scientific status. Criticism is based on both the subject matter and the methods of cognitive ethology.

In this paper we identify three major groups of people (among some of whose members there are blurred distinctions) with different views on cognitive ethology, namely, slayers, skeptics, and proponents. Our analyses are based on our reading of some published reviews of Donald Griffin's works in cognitive ethology (1976, 1981, 1978, 1984, 1992) and other clearly stated opinions concerning animal cognition, in the sense of attribution of mental states and properties such as beliefs, awareness, and consciousness. Two points need to be made clear at the outset. First, while Griffin's handling of issues such as anthropomorphism and the use of anecdote are often rather superficial (Griffin, 1977; Mitchell, 1986), and some even question whether or not his early works (at least) had much to do with ethology (Hailman, 1978; see Jamieson & Bekoff, 1993 for discussion), we are of the opinion that Griffin's (1976) rekindling of interest in the field that has come to be called cognitive ethology is responsible for the recent surge of interest in the comparative and evolutionary study of animal cognition. Thus our concentration on his work. Second, we ignore, but do not wish to downplay, the important experimental research on cognitive aspects of animal learning (see Galef, 1990; Kamil & Clements, 1990; Pepperberg, 1990; Timberlake, 1990; and Roitblat & von Fersen, 1992 and references therein), although those who do these types of studies may not call themselves cognitive ethologists; space does not allow us to discuss all views on cognitive ethology.

Categorizing views on cognitive ethology in the way we do it here helps us to identify common themes, which in turn helps us to see to what extent genuine dialogue between critics and defenders is possible; analysis of both criticisms and confusions arising from this dialogue will help improve the science. We were surprised by the number and the strength of some of the attacks on cognitive ethology by ethologists, behavioral scientists, and evolutionary biologists (see also Bekoff & Jamieson, 1991). Our purpose is to bring attention to those attacks and to try to understand the sort of argument that underlies them. These arguments include the views that anthropomorphism is unscientific, anecdotes do not constitute legitmate data, attributing beliefs to nonhumans is impossible, and cognitive ethology is a soft science. We are concentrating on the critics because we are sympathetic to cognitive ethology, and categorizing their various views will allow those who are sympathetic to cognitive ethology to recognize common themes and provide appropriate responses. We do not agree with Epstein's (1987, p. 21) claim "that our understanding of the behavior of organisms might advance ever so much faster if we spent more time studying it and less time philosophizing about it." We do, however, believe that the most effective philosophizing about cognitive ethology will occur only if careful attention is paid to the actual empirical work conducted by cognitive ethologists.

SOME VIEWS OF COGNITIVE ETHOLOGY: SLAYERS, SKEPTICS, AND PROPONENTS
For cognitive ethology, the major problems are those that center on methods of data collection and analysis, and on the description, interpretation, and explanation of behavior (Bekoff & Jamieson, 1990a,b). Here we consider how different people's views of cognitive ethology appear to be informed by available evidence, social factors, or some of both. We attempt to locate different views into three major categories, slayers, skeptics, and proponents.

Slayers: Slayers deny any possibility of success in cognitive ethology. In our analyses of their published statements, we have found that they sometimes conflate thedifficultyof doing rigorous cognitive ethological investigations with theimpossibilityof doing so. Slayers also often ignore specific details of work by cognitive ethologists and frequently mount philosophically motivated objections to the possibility of learning anything about animal cognition. Slayers do not believe that cognitive ethological approaches can lead, and have lead, to new and testable hypotheses. They often pick out the most difficult and least accessible phenomena to study (e.g. consciousness) and then conclude that because we can gain little detailed knowledge about this subject, we cannot do better in other areas. Slayers also appeal to parsimony in explanations of animal behavior, but they dismiss the possibility that cognitive explanations can be more parsimonious than noncognitive alternatives, and they deny the utility of cognitive hypotheses for directing empirical research.

Skeptics: Skeptics are often difficult to categorize. They are a bit more open-minded than slayers, and there seems to be greater variation among skeptical views of cognitive ethology than among slayers' opinions. However, some skeptics recognize some past and present successes in cognitive ethology, and remain cautiously optimistic about future successes; in these instances they resemble moderate proponents. Many skeptics appeal to the future of neuroscience, and claim that when we know all there is to know about nervous systems, cognitive ethology will be superfluous (Bekoff, 1993a; it should be noted that Griffin, 1992 also makes strong appeals to neuroscience, but he does not believe that increased knowledge in neurobiology will cause cognitive ethology to disappear). Like slayers, skeptics frequently conflate the difficulty of doing rigorous cognitive ethological investigations with the impossibility of doing so. Skeptics also find folk psychological, anthropomorphic, and cognitive explanations to be off-putting.

Proponents: Proponents recognize the utility of cognitive ethological investigations. They claim that there are already many successes and they see that cognitive ethological approaches have provided new and interesting data that also can inform and motivate further study. Proponents also accept the cautious use of folk psychological and cognitive explanations to build a systematic explanatory framework in conjunction with empirical studies, and do not find anecdotes or anthropomorphism to be thoroughly off-putting. Some proponents are as extreme in their advocacy of cognitive ethology as some slayers are in their opposition. But most proponents are willing to be critical of cognitive ethological research without dooming the field prematurely; if cognitive ethology is to die, it will be of natural causes and not as a result of hasty slayings.

Let us first consider some of what the slayers have to say about cognitive ethology. We realize that we are presenting only snippets from longer statements, but we feel that our abbreviated quotations capture the flavor of the general stance of the people involved.

Heyes (1987a,b) is concerned with what she perceives to be muddled thinking that pervades the area of cognitive ethology, and focuses her attacks on Griffin. She claims that

"Donald Griffin's . . . attempts to enrich the study of animal behaviour by making it cognisant of consciousness . . . deserves attention as an example of the muddled thinking that is increasingly influential in this area . . . " (Heyes, 1987a, p. 107)

Many of her (and others') concerns center on methodological and explanatory problems in studies of animal cognition, issues to which Griffin actually gave little detailed attention. If Griffin contributed to the muddle, it was by his failing to consider these areas critically; this has been pointed out in many of the reviews of Griffin's major books (Mason, 1976; Huntingford, 1985' Whiten, 1992; Yoerg, 1992; Bekoff1993b).

Anthropomorphism can also be off-putting (e.g., Humphrey, 1977; Colgan, 1989), and many people have used their distaste of anthropomorphism to condemn cognitive ethology (Bekoff, 1993a). In John S. Kennedy's (1992) recent book,The New Anthropomorphism, a volume devoted to a grand-scale and rather superficial condemnation of anthropomorphism (see Ridley, 1992) and its supposed negative effect on the field of ethology in general, it is claimed that:

Those who would have us go all the way back to traditional explicit anthropomorphism are still a minority but they show us the way things could go if we are not careful. (p. 5) Anthropomorphism must take its slice of the blame for a sort of malaise that has lately afflicted the subject of ethology as a whole. (p. 55).

In conclusion, I think we can be confident that anthropomorphism will be brought under control, even if it cannot be cured completely. Although it is probably programmed into us genetically as well as being inoculated culturally that does not mean the disease is untreatable. (p. 167)

Humphrey (1977, p. 521) also writes:

Away with critical standards, tight measurements and definitions. If an anthropomorphic explanationfeelsright, try it and see; if it doesn't feel right, try it anyway. (Humphrey, 1977, p. 521)

Kennedy and Humphrey are typical of critics who write as if the only alternatives are an unconstrained, fuzzy-minded use of anthropomorphism on the one hand, and the total elimination of anthropomorphism on the other (see also Estep & Hetts, 1992). But there is a middle position which they ignore (Fisher, 1990, 1991; Lehman, 1992). Anthropomorphism can be useful if it serves heuristically to focus attention on questions about animal behavior that might otherwise be ignored. Anthropomorphism might be used in a rigorous way to assist theory construction (Asquith, 1984) and to motivate empirical research projects (Burghardt, 1991, this volume; quotation below).

Now, let's consider some of Kennedy's other claims. That there is a sort of malaise in the subject of ethology as a whole simply does not jibe with current interest in the field by many diverse people. A glance at any of the newer textbooks in the field as well as Griffin's (1992)Animal Mindsshows that there is a lot of interest in ethology and the stimulating and difficult problems that ethological studies consider. Unlike Kennedy, even some of Griffin's other strongest critics note that there has been, and still is, a great deal of interest in ethology. However, they go on to condemn Griffin for providing merely natural history accounts of the behavior of animals, but fail to recognize that there has been a large amount of experimental work done in cognitive ethology. Griffin does, in fact, write about this research and he should be credited for inspiring much of it directly (Huntingford, 1985). Kennedy also claims, in the total absence of any sort of data, that there is some sort of genetic predisposition to engage in anthropomorphism. This claim is out of character with the rest of his arguments that demand appeal to hard data. Perhaps Kennedy's closing sentences capture the essence of his views:

If scientists, at least finally cease to make the conscious or unconscious assumption that animals have minds, then the consequences can be expected to go beyond the boundaries of the study of animal behaviour. If the age-old mind-body problem comes to be considered as an exclusively human one, instead of indefinitely extended through the animal kingdom, then that problem too will have been brought nearer to a solution. (pp. 167-168)

Kennedy's idea that the mind-body problem might be closer to solution if restricted to humans actually flies in the face of much recent work in the philosophy of mind, where a significant trend is toward naturalizing and even "biologizing" mental properties (e.g. Millikan, 1984). No doubt, Kennedy would regard such moves by philosophers as misguided, but they cannot simply be dismissed by a wave of the hand; at best, his argument depends on a contentious philosophical position. This sort of argument is typical of many slayers, as we shall further illustrate below.

Some slayers use the "hard" (but not necessarily more difficult) sciences as a model for the criticism of "soft" cognitive ethology (Humphrey, 1977; Heyes, 1987a,b; Kennedy, 1992; Williams, 1992), while others tend to be dismissive and conveniently ignore available empirical evidence of animal cognition in their sweeping claims (Ingold, 1988; Colgan, 1989; Zuckerman, 1991). They write:

There can be no historical doubt that behaviourism has advanced ethology as a science, whereas the methods advocated by cognitivists have yet to prove their worth. Until mental concepts are clarified and their need justified by convincing data, cognitive ethology is no advance over the anecdotalism and anthropomorphism which characterized interest in animal behaviour a century ago, and thus should be eschewed. (Colgan, 1989, p. 67)

"A scientist's hunch is acceptable as a start, provided that it leads to a theory that can be rejected in the face of evidence. This has not been achieved in the field of animal cognition. (McFarland, 1989b, pp. 146-147)

A few comments are in order. First, it is notable that the ethologist Patrick Colgan's (1989, p. 67) failure to give an adequate account of available literature has been challenged (McCleery, 1989; see also Houston, 1990), even by slayers (McFarland, 1989a). McCleery (189, p. 1092) notes that "Brevity also characterizes [Colgan's] discussion of cognition", and he continues on to say that Colgan's dismissal of cognitive ethology "scarcely seems warranted on the grounds given here." McFarland (1989a, p. 170) notes that Colgan's "section on cognition is poor. It fails to home in on the essential features of cognition, devotes too much space to fairly irrelevant historical matters, and offers no good examples of cognition in animals." Nonetheless, McFarland (1989b) seems to think that he has reviewed the field of animal cognition and that his conclusion (above quotation) is well-founded. In fact, his chapter entitled "Cognition in Animals" contains very few references to recent work in the field. Others who dismiss the field of cognitive ethology also make little effort to consider available evidence. Zuckerman (1991, p. 46), in his review of Cheney & Seyfarth's (1990) bookHow Monkeys See the World: Inside the Mind of Another Species,exemplifies this dismissive attitude:

Some of the issues they do raise sound profound as set out but, when pursued, turn out to have little intellectual or scientific significance. (Zuckerman, 1991, p. 46)

Heyes (1987a), who is a laboratory psychologist, advises cognitive ethologists to turn to laboratory research if they want to understand animal cognition. She writes:

It is perhaps at this moment that the cognitive ethologist decides to hang up his field glasses, become a cognitive psychologist, and have nothing further to do with talk about consciousness or intention. (Heyes, 1987a, p. 124)

Thus, Heyes denies that evidence gained by observing animals in natural settings is particularly relevant to understanding animal minds. Other slayers, who claim that they need more convincing evidence from the field, rarely tell what evidence would be convincing. Heyes and these other critics generally simply assume that no evidence that could be collected from the field would provide convincing support for attributions of mental states.

Unlike Heyes who thinks that animal cognition can at least be studied in the laboratory, some slayers argue against the study of animal cognition on the basis of a philosophical view about the privacy of the mental (for a well-developed counter-argument see Whiten, 1993), or by the related "other minds" problem. These critics typically do not give specific critiques of actual empirical investigations carried out by cognitive ethologists; rather they dismiss such investigations on philosophical grounds alone. The renowned evolutionary biologist, George C. Williams, writes:

Vitalism today can be recognized in the psycho-physical dualism of some neural and behavioral biologists (e.g. Griffin 1981), who claim that explanations must make use of explicitly mental factors in addition to the merely physical. Griffin's concession (1981, p. 301) that the mental depends on the physical is difficult for me to interpret. If he means total dependence there is no longer any reason to make use of mentalism in biological explanation . . . " (Williams, 1992, p. 4)

In this quotation, Williams reveals his naiveté about contemporary philosophy of mind. In questioning Griffin's notion of dependence, Williams suggests that if the mind depends on physical facts, then it can be eliminated from scientific explanation. In fact, such a view is tenable if you assert that mental properties are strictly identifiable with physical properties (what philosophers call "type-identity" theory), which seems to be what Williams means by the phrase "total dependence". But there are forms of dependence other than type identity that do not support the view that mental states are irrelevant to the explanation of behavior. For example, functionalists like Fodor (1974) hold a token-identity theory that allows a particular mental state to be identified with a particular brain state (just like a particular dollar can be identified with a particular piece of paper) but does not allow the mental and brain state types to be identified because of the possibility that organisms with different nervous systems can be in the same psychological state (just as the type "dollar" cannot be identified with the physical type "piece of paper bearing such and such markings" since there are also silver dollars). Physical predicates and mental predicates appear to carve up the world differently. Whether there are scientific generalizations that require the divisions provided by mental vocabulary is an empirical question that is separate from the dependence of the mental on the physical. Token-identity theory is not the only way to defend the claim that the mental depends on but is not reducible to the physical (e.g. Kim, 1984). Williams' failure to consider alternate conceptions of dependence seriously undermines his argument against Griffin.

Williams (1992, p. 4) also writes:

I am inclined merely to delete it [the mental realm] from biological explanation, because it is an entirely private phenomenon, and biology must deal with the publicly demonstrable. From this quotation, we can construct an argument like this: (1) Mental events are private phenomena. (2) Private phenomena cannot be studied biologically. (3) Therefore, mental events cannot be studied biologically. (4) Cognitive ethology is possible only if mental events can be studied biologically. (5) Therefore, cognitive ethology is not possible.

When analyzed in this way, Williams' argument is seen to depend on another contentious philosophical premise (#1). Slayers often base arguments on claims about the privacy of the mental or skepticism about other minds. But it is ironic that these premises, which can only be defended in nonempirical, philosophical fashion, are produced by critics who would typically regard themselves as hard-nosed empiricists. Cognitive ethologists do empirical work, yet slayers who argue on such philosophical grounds rarely analyze that empirical work to see what it is designed to show and whether it in fact shows what it is designed to show. Instead, they base their arguments on claims that are at least as fraught with interpretive difficulty as the cognitive conclusions they wish to deny. This unwillingness to engage in debate about the actual empirical work of cognitive ethologists gives the impression that many slayers simply barge in, declare victory, and get out without genuinely engaging cognitive ethologists in a dialogue about their work. Williams does not stand alone. McFarland (1989b, p. 146) wonders if we are designed by natural selection to assume that deceitful acts are intentional, and Kennedy claims that the sin of anthropomorphism is programmed into us genetically. If either or both is the case, how are we to know? These claims make for empirical questions that require detailed study.

Rosenberg is an example of a slayer who has paid careful attention to actual work in behavioral biology, specifically the study of play. He says (Rosenberg, 1990):

. . . we have lots of reasons to fear that there can be no evolutionary theory of intentionality. For there can be no scientific theory of intentionality. (p. 183)

Does a cat have the concept of mouse-hunting? . . . The reason we are reluctant to credit the cat with such concepts is that there seems to be no behavior it could engage in that is discriminating enough . . .This leads the tender-hearted among students of animal behavior to insist that though the animal may not have our concept, it has some concept or other . . . The kill-joy argues that short of behavior discriminating enough . . . attributing concepts at all is gratuitous anthropomorphism. . . . When it comes to such abstract concepts as those required to attribute playfulness, in the literal sense, to mammals much below that of the monkey, the kill-joy's position seems hard to deny. (p. 184)

In this passage, Rosenberg belittles the idea that it might be possible to attribute human concepts to nonhumans. His argument against a science of intentionality is based on considerations that seem to conflate the difficulty of specifying the content of intentional states with the impossibility of doing so (Allen, 1992a, Allen & Bekoff, 1994). Specifically, Rosenberg appeals to his inability to imagine what organisms could do to allow the attribution of certain concepts; the difficulty of thinking of suitable experiments is a challenge, not necessarily a barrier to concept attribution (Allen & Hauser, 1991).

Other tactics used by some slayers involve focussing their criticisms on very narrow issues. Cronin thinks that Griffin, a "sentimental softy," and other cognitive ethologists are only concerned with demonstrating cleverness, and hence consciousness. In her recent review of Griffin'sAnimal Minds(1992) she writes:

A Griffin bat is a miniature physics lab. So imagine the consternation among behavioristic ethologists when Mr. Griffin came out a decade ago, with "The Question of Animal Awareness," as a sentimental softy. . . . For Mr. Griffin, all this [cleverness]suggests consciousness. He's wrong. If such cleverness were enough to demonstrate consciousness, scientists could do the job over coffee and philosophers could have packed up their scholarly apparatus years ago. (Cronin, 1992, p. 14, our emphases)

Even McFarland (1989b), who we categorize as a slayer, recognizes that there are indicators of cognition other than the ability to produce clever solutions to environmental problems. Not only is Cronin wrong about slaying the field of cognitive ethology because of the difficulty of dealing with the notion of consciousness (think about all of the other fields of inquiry that would suffer if it were appropriate to base rejection of those fields on singling out their most difficult issues), but she is also wrong to think that demonstrating cleverness is a simple matter. Certainly, the difficult work needed to demonstrate cleverness could not be done over coffee! Furthermore, Cronin conveniently slides from claiming that for Griffin, clevernesssuggestsconsciousness, to claiming that his view is that cleverness is "enough todemonstrateconsciousness" (our emphasis). Even Heyes (1987b) notes that it is not Griffin's program toprovethat animals are conscious. Cronin later goes on to claim that at least chimpanzees are conscious and tells us why. Cronin concludes her scathing review with the following statement:

Well, I know that I am conscious, I know a mere 500,000 generations separate me from my chimpanzee cousin, and I know that evolutionary innovations don't just spring into existence full-blown--certainly not innovations as truly momentous as our hauntingly elusive private world. Cronin places herself on a slippery slope here. Why did she stop with chimpanzees? After all, if evolutionary innovations do not spring into existence full-blown, where did chimpanzee consciousness come from? Her phylogenetic argument cannot be assessed directly; behavioral evidence is needed to help it along.

Slayers' arguments can be based on a number of issues discussed above, and some even throw the debate into the political arena. Thus, Garland Allen (1987) claims:

. . . to argue that we are animals does not mean we are merely animals. (p. 158). It is in this vein that I think that both sociobiology and the study of animal awareness can be seen as part of an evolutionary development of present-day capitalism toward fascism. In the abstract there is nothing inherently fascistic about asking whether animals "think" or have an awareness of their own existence. It is the asking of this question under the present social and economic conditions, and with no procedures for arriving at a rigorous or scientific answer that makes the whole enterprise part and parcel of a fascistic (albeit at this stage a protofascistic) social development. (p. 159) Given the methodological problems inherent in the field of animal consciousness (as well as with sociobiology), pursuit of the problem in the ways exemplified by investigators such as Griffin, can only lead to the sort of confusion and false understanding of the biological nature of human beings, on which the future of fascism can be built. (p. 160)

Slayers may also have something to say about animal welfare. Despite a very large data base demonstrating highly developed cognitive skills in many animals, there are those who ignore research on animal cognition, misinterpret data from studies on humans, and base their conclusions on the moral status of animals using an intuitionistic comparison of animal and human behavior (e.g. Carruthers, 1989; Leahy, 1991; for discussion of these views see Bekoff & Jamieson, 1991; Clark, 1991; Johnson, 1991; Griffin, 1992; Jamieson & Bekoff, 1992a; Singer, 1992).1Thus, Peter Carruthers (1989), who compares the behavior of animals with the behavior of humans who are driving while distracted and humans who suffer from blindsight, writes:

I shall assume that no one would seriously maintain that dogs, cats, sheep, cattle, pigs, or chickens consciously think things to themselves . . .the experiences of all these creatures [are of] the nonconscious variety.(p. 265)

He then claims:

Similarly then in the case of brutes: since their experiences, including their pains, are nonconscious ones, their pains are of no immediate moral concern. Indeed since all the mental states of brutes are nonconscious, their injuries are lacking even in indirect moral concern. Since the disappointments caused to the dog through possession of a broken leg are themselves nonconscious in their turn, they, too, are not appropriate objects of our sympathy. Hence, neither the pain of the broken leg itself, nor its further effects upon the life of the dog, have any rational claim upon our sympathy. (p. 268)

And finally,

And it also follows that there is no moral criticism to be leveled at the majority of people who are indifferent to the pains of factory-farmed animals, which they know to exist but do not themselves observe. (p.269)

All in all, slayers conclude that the study of cognitive ethology is too fraught with difficulty to be worth pursuing, or they simply fail to address the results of cognitive ethological research.

Skeptics' views are more open-minded than those of slayers. Skeptics have some of the same worries about anthropomorphism, anecdote, folk psychological explanations, the inaccessibility of the mental, and the ability of field studies to shed light on animal cognition, but they are not as dismissively forceful. Two representative views are:

. . . folk psychological theory pervades human thinking, remembering, and perceiving and creates a very subtle anthropomorphism that can corrupt the formation of a science of cognitive ethology. (Michel, 1991, p. 253; our emphasis) We argue that attempts to study these processes [intentionality, awareness, and conscious thinking], while revealing impressive behavioral complexity, have proven unsuccessful in establishing the importance of mental experiences in determining animal behavior primarily because of the intractability of the problem. (Yoerg & Kamil, 1991, p. 273)

Alcock (1992) is a notable example in that he does not find the inaccessibility of consciousness to be grounds for dismissing the study of animal cognition (see also Whiten, 1993). In his review of Griffin's (1992) Animal Minds, he wrote:

We need ways in which to test hypotheses in a convincing manner. In this regard Animal Minds disappoints, because it offers no practical guidance on how to test whether consciousness is an all-purpose, problem-solving device widely distributed throughout the animal kingdom. . . . And there are alternative approaches to consciousness not based on the behavioristic principle that thinking cannot be studied because it does not exist. (Alcock, 1992, p. 63)

In his attempt to gain more control in studies of animal cognition, Premack,like Heyes (1987a), wants to see more highly controlled laboratory work. He believes that:

. . . cognition is not exclusively a field phenomenon; it can take place in the laboratory. Indeed, in the case of chimpanzees, advanced cognition would appear to be largely a laboratory phenomenon. For only the chimpanzee who has been specially trained--exposed to the culture of a species more evolved than itself--shows analogical reasoning . . advanced cognition, such as analogical reasoning, is confined to the laboratory. (Premack, 1988, pp. 171-172)

One problem here is that gaining too much control over what may be important variables may result in an impoverished environment that does not allow the animals to use the combination of stimuli that enable them to make assessments of others' behavior or others' minds (Bekoff et al., 1994). Only careful observations under field and captive conditions will permit us to assess just what are the important variables that influence how individuals interact with their social and nonsocial worlds, how these different variables are used, and how they may be combined with one another. Along these lines, de Waal (1991, p. 311) notes that "Certain social phenomena cannot be transferred to the the laboratory," because it would be impossible to recreate the social and nonsocial environments that are responsible for producing and maintaining ongoing encounters, such as those observed in dominance interactions. McGrew (1992, p. 83) also stresses that "No experiment has ever simulated any of the delicate probing tasks such as termite-fishing shown by wild chimpanzees . . . "

The types of explanations that are offered in studies of animal cognition also enter into how cognitive ethological studies are viewed. For example, many slayers and some skeptics favor noncognitive explanations because they believe them more parsimonious and more accurate than cognitive alternatives, and less off-putting to others who do not hold the field of cognitive ethology in high esteem. Snowdon (1991, p. 814) claims that:

It is possible to explore the cognitive capacities of nonhuman animals without recourse to mentalistic concepts such as consciousness, intentionality, and deception. Studies that avoid mentalistic terminology are likely to be more effective in convincing other scientists of the significance of the abilities of nonhuman animals.

Some just make simple claims about the supposed parsimoniousness of noncognitive explanations and move on. Thus, Zabel et al. (1992, p. 129) in their attempts to explain redirected aggression in spotted hyenas (Crocuta crocuta) before quantitatively analyzing it, are of the opinion that:

One must be cautious about inferring complex cognitive processes when simpler explanations will suffice.

However, they admit that the other noncognitive explanations they offer are questionable. The statement by Zabel et al. is a paraphrase of a principle known as "Lloyd Morgan's canon" (Morgan, 1894, p. 53): "In no case may we interpret an action as the outcome of the exercise of a high psychical faculty, if it can be interpreted as the outcome of the exercise of one which stands lower in the psychological scale." It is possible that Morgan's canon, which is concerned with the complexity of processes, should be distinguished from parsimony, which is concerned with the number of processes needed to explain a given behavior (anonymous review). However, unless we are given an explicit standard for judging the complexity of a cognitive process or its place in "the psychological scale," appeals to the canon rest on nothing other than intuitions about relative complexity and need not be counted as scientific. Appeals to parsimony on a case by case basis do not take into account the possibility that cognitive explanations might help scientists come to terms with larger sets of available data that are difficult to understand and also help in the design of future empirical work. (For further discussion of the weakness of the idea that the simplest explanation is always the most parsimonious, see Bennett [1991]. For a comparison of different perspectives on cognitive "versus" more parsimonious explanations, see proponent de Waal's [1991] consideration of the skeptics Kummer et al.'s [1990] views on parsimony.)

Beer (1992, p. 79) also thinks that if cognitive ethology limited its claims for animal awareness to sensation and perception, a practice which could change the vocabulary used in cognitive ethological studies, then "even tough-minded critics would be more receptive." Griffin (1992, p. 11) actually agrees with this point, but even a consideration of simple forms of consciousness is contentious to many slayers and skeptics (see Bekoff, 1993a).

Some other skeptical views include:

Perhaps cognitive ethology will prove to be an advance in the study of behavior in consequence of resuscitating questions of animal awareness. How far back it will be necessary to go to find the more fruitful formulation of these questions remains open to doubt, however: to Darwin and Romanes who believed in animal awareness? or to Descartes, who did not? (Beer, 1992, p. 105) I am afraid that I am somewhat more skeptical than your reviewers on the value of such an approach [cognitive ethology] . . . Well, these are just my thoughts. Clearly, the paper sparked my interest about the issue, as it did for your two reviews . . . I want to see controversial subjects discussed . . . and I do not think that would ever get complete agreement from reviewers on an article like this . . . " (Jane Brockmann, personal communication to the authors)

Beer remains cautiously optimistic, but wants cognitive ethologists to pay close attention to problems that are raised by philosophers of mind; his essay is essential reading for those interested in animal cognition. The second quotation is notable, for it reflects the refreshing open-mindedness of an editor of a major journal in ethology.

Let's now consider the proponents. William Mason's (1976, p. 931) quotation from his review of Griffin's (1976) The Question of Animal Awareness is a good place to start with respect to proponents' views. Mason writes:

That animals are aware can scarcely be questioned. The hows and the whys and wherefores will occupy scientists for many years to come.

Mason's claim is a strong one. Note that in his endorsement of the field, Mason does not qualify his statement by writing "That some animals are aware . . . ". However, he does recognize that animals may differ with respect to levels of development of their cognitive abilities, and at a later date noted that "On the basis of findings such as those reviewed in this paper, I am persuaded that apes and man have entered into a cognitive domain that sets them apart from all other primates" (Mason, 1979, pp. 292-293). Mason's inclusive statement about animal awareness is typical of those who narrowly focus their attention only on primate cognition (Beck, 1982; Bekoff, 1995; Bekoff et al., 1994).

Proponents note that there are already many successes in cognitive ethology and are enthusiastic about the bright future for cognitive ethology (see Cheney & Seyfarth, 1990, 1992; Allen & Hauser, 1991; Burghardt, 1991; de Waal, 1991; Ristau, 1991a; Allen, 1992b; Bekoff & Allen, 1992; Bekoff 1993a, 1995; Bekoff et al., 1994; Jamieson & Bekoff, 1993; Whiten, 1993). Although proponents admit that cognitive ethology has had some failures, they do not single these out and use them to condemn the whole field (Dennett, 1987, p. 271). While skeptics are willing to state what would convince them that cognitive ethology is a viable science, they may be pessimistic about the prospects of cognitive ethology coming up with the goods. Often, however, it is difficult to differentiate between skeptics and moderate proponents who argue that if there is to be a science of cognitive ethology, we must develop empirical methods for applying cognitive terms and making talk about animal minds respectable (e.g. Dennett, 1983, 1987, p. 271, 1991, p. 446; Kummer et al., 1990). Jamieson & Bekoff's (1993) differentiation between weak cognitive ethology, where a cognitive vocabulary can be used to explain, but not to describe behavior, and strong cognitive ethology, where cognitive and affective vocabularies can be used to describe and to explain behavior, highlights some of the difficulties of clearly differentiating between some skeptics and some proponents. Nonetheless, proponents are more willing than are skeptics to rely on methods of classical ethology (Jamieson & Bekoff, 1993) to help them along, whereas skeptics and slayers generally want to see cognitive ethology become consumed by cognitive psychology or by neuroscience.

Proponents are more optimistic in their views about the contributions that the field of cognitive ethology, and its reliance on field work and on comparative ecological and evolutionary studies, can make to the study of animal cognition in terms of opening up new areas of research and reconsidering old data (de Waal, 1991; Bekoff, 1995). As Carolyn Ristau (1991b, p. 102) notes in her attempts to study injury-feigning under field conditions, the cognitive ethological perspective . . . led me to design experiments that I had not otherwise thought to do, that no one else had done, and that revealed complexities in the behavior of the piping plover's distraction display not heretofore appreciated.

The challenge of using ethological ideas in the study of animal cognition is reflected in the following quotation:

At this point, however, cognitive ethologists can console themselves with the knowledge that their discipline is an aspect of the broader field of cognitive studies and conceptually may not be in any worse shape than highly regarded, related fields such as cognitive psychology. We are a long way from understanding the natural history of the mind, but in our view this amounts to a scientific challenge rather that grounds for depression or dismissal. (Jamieson & Bekoff, 1992b,p. 81)

Premack's (1988, see above quotation) views about the absence of high level cognition in the field and its being largely a laboratory phenomenon ignores the complexity of behavior that is found in wild animals (Griffin 1976. 1981, 1984, 1992, Byrne & Whiten 1988, de Waal 1991, Ristau 1991a). While proponents recognize that Griffin has not really made detailed suggestions for experimental studies, this doesn't discourage them from seeking ways to make ideas like Griffin's empirically rigorous (Allen & Hauser, 1991; de Waal, 1991; Ristau, 1991a; Allen, 1992a,b; Whiten, 1992; Allen & Hauser, 1993; Bekoff, 1993a, 1995).

Proponents also share some of the concerns of the slayers and skeptics' with respect to problems associated with the use of anecdote, anthropomorphism, and folk psychological explanations. However, proponents claim that the careful use of anthropomorphic and folk psychological explanations can be helpful in the study of animal cognition, and they also maintain that anecdotes can be used to guide data collection and to suggest new experimental designs (Dennett, 1987; Fentress, 1992). Thus, other proponents write:

Cognitive ethology, rescued from both behaviorism and subjectivism, has much to say about what the life of the animal is really like. It is silent on what it is like to have that life. (Gustafson, 1986, p. 182) Others have picked up where Griffin leaves us by using his collection of anecdotes, his discussion of empirical research, and his ideas, to motivate new and extremely innovative studies, the bases for which might not have been obvious before his work. (Bekoff, 1993a) . . . I have advocated use of a critical anthropomorphism in which various sources of information are used including: natural history, our perceptions, intuitions, feelings, careful behavioral descriptions, identifying with the animal, optimization models, previous studies and so forth in order to generate ideas that may prove useful in gaining understanding and the ability to predict outcomes of planned (experimental) and unplanned interventions . . . " (Burghardt, 1991, p. 73) . . . when a number of anecdotal examples, each with a possible alternative explanation, collectively point to the likelihood of intentional deception, and this is supported by more rigorous tests in the laboratory . . . I would argue that it adds up to a strong case. (Archer, 1992, p. 224)

Perhaps some would assume Griffin to be the strongest proponent of cognitive ethology. After all, his rekindling of interest in the area (Griffin, 1976, 1981, 1978, 1984) is usually given sole credit for restimulating modern interest in cognitive ethology. Furthermore, even in his most recent book, Animal Minds, he has unabashedly launched a number of polemical attacks on his opponents for being closed-minded about the nature of available evidence on animal cognition and on the field of cognitive ethology in general. [These sometimes snide attacks are considered to be unnecessary and off-putting to his critics and supporters alike (Huntingford, 1985; Yoerg, 1992; Bekoff, 1993a).] Toward the end of Animal Minds Griffin (1992, p. 260) writes:

Contrary to the widespread pessimistic opinion that the content of animal thinking is hopelessly inaccessible to scientific inquiry, the communicative signals used by many animals provide empirical data on the basis of which much can reasonably be inferred about their subjective experiences. (our emphasis)

Note that Griffin counters some slayers' and skeptics' concerns about the inaccessibility of animal minds, but he does not make a very strong claim that he or others can ever know the content of animal minds [see italicized phrase above; recall Cronin's (1992) misrepresentation of Griffin's agenda above]. Rather, Griffin, like other proponents, remains open to the possibility that we can learn a lot about animal minds by carefully studying communication and other behavior patterns. He and other proponents want to make the field of cognitive ethology more rigorous on theoretical and empirical grounds.

To Part III