No consensus exists concerning the mechanisms, distribution, or adaptive significance of consciousness. Agreement on any one of these issues would aid in resolving others. Given a reliable behavioral or neuroanatomical test for consciousness, we could map its distribution and describe its evolution. Conversely, if we knew its distribution, we could assess its adaptive value and look for similarly distributed neuroanatomies to help us get at its mechanisms.

Morgan's Canon—the rule that we should avoid attributing humanlike mental states to other animals whenever possible—impedes the use of the comparative method in unraveling this knot. If interpreted in this context as a parsimony criterion, Morgan's Canon is logically equivalent to epiphenomenalism. It is parsimonious if and only if conscious mental events play no causal role in human behavior and human consciousness has no adaptive significance. Rejecting this conclusion entails rejecting the parsimony interpretation of Morgan's Canon.

A monodimensional understanding of both the past and the present has often characterized the historical background provided for scientific disciplines. It scans the past for the antecedents of the current cutting edge. But establishing such genealogical connections is difficult, even on the level of terminology. For example, neither “animal” nor “consciousness” has ever been defined without contention. The main current (and past) problem with the term “animal” is taxonomic: does it include people or not? and if so, are people included in only their physical aspect or in some more encompassing sense? An examination of 18th and 19th-century taxonomical treatments of the great apes shows how convictions and uncertainties on these issues were expressed through classification. With regard to “consciousness” the main problem has been whether that term refers to something shared by humans and other species. An examination of 19th-century attempts to claim that non-primates like the dog resembled humans more closely than did apes shows that this issue too reflects scientific as well as lay discomfort at the awkward proximity of other primates. A final problem with establishing scientific genealogies is that they often assume past consensus among experts on issues where there was actually intense disagreement, complicated by the difficulty of deciding who the experts actually were two or three centuries ago. None of these problems has completely disappeared even now.

Lloyd Morgan's advisory “Canon” on ascribing mental phenomena had an historically unusual and deep impact. It became a dictum defining true psychological science, praised for providing the foundation of a mature, hard science. Seen as a corrective to naive methodology and excessive speculation, it appeared to demand removal of mental qualities in the name of parsimony and rigor. Thus it is also disparaged for illegitimately limiting inquiry. Viewing Morgan's method and reasoning as the reactions of a Darwinian to the problems of psychology provides insight for a science of animal mind. The Canon should be seen as methodological advice for a science caught in the tensions between materialism and subjective mental experience, having to place human mind within phylogenetic continuity. Faced with irresolvable difficulties, behavioral science has oscillated between reduction and unifying integration, typical of debates over broad conceptual issues in evolutionary biology. The reasoning behind Morgan's Canon provides a strategy for balancing these twin pulls of scientific practice.

Our genotype is so similar to those of the African apes, and our last common ancestor with them so recent, that it seems impossible that human and non-human cognition should differ qualitatively. But the outputs of human cognition are unique in their limitless creativity and adaptability. Exaption resolves the apparent paradox. Assume that the power to create symbols emerges from stimulus-stimulus linkages and is latent in many animals, and that the structural side of language emerges from the argument structures inherent in the social calculus associated with reciprocal altruism. These adaptations confer the potential for language. However, creating complex messages requires uniquely long-lasting coherence of neural signals, which depends in turn on the large quantities of neurons unique to Homo. The only difference between human and non-human minds is that we can sustain longer and more complex trains of thought. All else (emotions, rational processes, even consciousness) could be exactly the same.

There are three domains of experience that concern students of behavior: Domain 1. The domain of felt experience, the phenomenological domain. Domain 2. The domain of physiology, the real-time functioning of the brain. Domain 3. The domain of behavioral data, “intersubjectively verifiable” reports and judgments by experimental subjects. Consciousness has meanings in each of these domains. Domain 1 consciousness is beyond the reach of science as public knowledge. Empathy and plausible inference may tell us that our spouse, or our dog, is as conscious as we are. Science cannot. Research in Domains 2 and 3 permits us to infer similarities and differences between human and non-human psychology. Unfortunately, these will never permit us to know ‘what it is like’ to be another creature. An example from the study of motion perception illustrates the point that the fruitless attempt to answer this question can actually impede the objective study of behavioral processes we share with non-human animals.

The possibility of conscious experiences of emotions in non-human animals has been much less explored than that of conscious experiences associated with carrying out complex cognitive tasks. However, no great cognitive powers are needed to feel hunger or pain and it may be that the capacity to feel emotions is widespread in the animal kingdom. Since plants can show surprisingly sophisticated “choice” and “decision-making” mechanisms and yet we would not wish to imply that they are conscious, attribution of emotions to animals has to be done with care. Whether or not an animal possesses anticipatory mechanisms associated with positive and negative reinforcement learning may be a guide as to whether it has evolved emotions.

In a marked departure from past inhibitions about scientific consideration of conscious mental states in animals, the other papers in this symposium review a variety of evidence about what the content of animal consciousness is likely to be. Although fully convincing evidence is not yet available, there are promising opportunities to reduce our current ignorance of what life is like, subjectively, for various animals. For example, recent neurophysiological experiments provide objective evidence about what monkeys are, and are not, conscious of. Versatility of behavior when animals cope with novel and unpredictable challenges strongly suggests simple conscious thinking about alternative actions. Finally, animal communication provides direct and objective, though incomplete and imperfect, evidence about some of their conscious thoughts and feelings.

Researchers often study nonhuman abilities by assuming their subjects form representations about perceived stimuli and then process such information; why then would consciousness be required, and, if required, at what level? Arguments about nonhuman consciousness range from claims of levels comparable to humans to refutation of any need to study such phenomena. We suggest that (a) species exhibit different levels attuned to their ecological niches, and (b) animals, within their maximum possible level, exhibit different extents of awareness appropriate to particular situations, much like humans (presumably conscious) who often act without conscious awareness of factors controlling their behavior. We propose that, to engage in complex information processing, animals likely exhibit perceptual consciousness sensu Natsoulas (1978), i.e., are aware of what is being processed. We discuss these issues and provide examples suggesting perceptual consciousness.

Tests of self-awareness in nonhuman primates have to date been concerned almost entirely with the recognition of an animal's reflection in a mirror. By contrast, we know much less about non-human primates' perception of their place within a social network, or of their understanding of themselves as individuals with unique sets of social relationships. Here we review evidence that monkeys who fail the mirror test may nonetheless behave as if they recognize themselves as distinct individuals, each of whom occupies a unique place in society and has a specific set of relations with others. A free-ranging vervet monkey, baboon, or macaque recognizes other members of his group as individuals. He also recognizes matrilineal kin groups, linear dominance rank orders, and behaves as if he recognizes his own unique place within them. This sense of “social self” in monkeys, however, is markedly different from self-awareness in humans. Although monkeys may behave in ways that accurately place themselves within a social network, they are unaware of the knowledge that allows them to do so: they do not know what they know, cannot reflect on what they know, and cannot become the object of their own attention.

Animal consciousness has long been assumed to be a nonviable arena of investigation. At best, it was thought that any indications of such consciousness, should it exist, would not be interpretable by our species. Recent work in the field of language competencies with bonobos has laid this conception open to serious challenge. This paper reviews this work and the case it makes for our impending capacity to tap the consciousness of a uniquely enculturated group of bonobos who are capable of comprehending human speech and employing a lexical communication system.