A shorter version of this review appeared in the Journal of Consciousness Studies, Vol. 13 No. 6 (2006).
Herbert S. Terrace and Janet Metcalfe, eds.
The Missing Link in Cognition: Origins of Self-Reflective Consciousness
Oxford University Press, 2005, 384 pp., £46.00, ISBN 0-19-516156-4 (hbk).
This book originated in a conference on primate cognition, convened by the editors at Columbia University in 2002. The title summarises the stance taken in the editorial introduction, which assumes a wide gap between human mentality and that of other animals (including other primates). Since language is taken to exist only on the human side of the gap, its structure is not considered relevant to the inquiry here. The question is rather where ‘self-reflective consciousness’ could have come from: ‘How did this pinnacle of human mental life—self-reflective consciousness—evolve, and how much does it depend on language?’ (p. xiii). Rather than defining language broadly and looking for primitive forms of it, the approach here is to define ‘metacognition’ broadly and look for ‘metacognitive precursors’ of human-style thinking. The challenge, then, is to investigate the metacognitive abilities of creatures related to us phylogenetically who can’t use language to tell us what they are thinking. Accordingly, the conference brought together several primatologists involved in the study of metacognition, along with theorists in related but more human-centred fields.
This line of inquiry reflects the predilections of Herbert Terrace, who played a prominent and skeptical role in the acrimonious ape-language controversy of the late 1970s. His contribution here continues to offer a ‘killjoy interpretation of an ape's use of symbols’ (p. 101). He argues that human metacognitive and mind-reading skills, beyond age two or so, already exceed those of adult apes, and he accuses other researchers of ‘sheer projection’ when they read ‘intentional meaning’ into symbol use and other ape behaviours (pp. 99-100). But what if such ‘projection’ is inseparable from intersubjective relations? Janet Metcalfe and Hedy Kober seem to imply as much in designating the ‘projectable self’ as a key component of ‘self-reflective consciousness.’ Studies of human ontogeny, such as Katherine Nelson’s in this volume, show that metacognitive skills develop in a social context involving mutual recognition and expectation. If we normally interact by reading intentions into each other, eliminating this factor from observations of apes (if possible at all) might work as a self-fulfilling prophecy, resulting in a study that renders its own hypothesis unfalsifiable and therefore untestable.
In any case, Terrace clearly stands on one side of a debate which Endel Tulving, in the opening essay of this collection, compares to a ‘tug of war.’ On this side, human uniqueness is ‘given’ and the problem is to explain how human mentality could emerge from other forms. The editors refer to the other side of the debate as ‘Darwin's continuity hypothesis,’ since Darwin claimed in his Descent of Man that ‘there is no fundamental difference between man and the higher mammals in their mental faculties,’ any gap between them being ‘filled up by numberless gradations.’ Most of the other primatologists represented here seem to lean in the latter direction; so does philosopher Marcel Kinsbourne, who finds ‘a continuum of self-consciousness that emerges in phylogeny and ontogeny.’ Patricia Kitcher draws similar conclusions, summarising and then challenging the belief of Locke and other modern philosophers ‘that in self-consciousness they had finally found the key to human uniqueness’ (p. 185).
Contributors on the other end of the rope emphasize the difference between humans and others but vary in the features they propose as uniquely human. E. T. Higgins, for instance, argues that humans are ‘unique in being applied motivation scientists’ (p. 170). Tulving also places himself in this camp, naming episodic memory (and the ‘autonoetic’ kind of consciousness associated with it) as uniquely human, while semantic memory is widespread, at least among mammals and birds. Several other contributors to this volume employ the semantic/episodic distinction, but the terms (dating from 1983) are somewhat misleading and seem to be used elsewhere in senses differing somewhat from Tulving's. This review will place them in ‘scare-quotes’ as a reminder that they are used here strictly according to Tulving's current usage. As a first approximation to this sense of ‘semantic’ memory, we could think of it as an animal's access to its internal map of the world, on which the locations of various features or objects are marked. This ‘map’ may have a temporal as well as a spatial dimension: N.S. Clayton and colleagues have shown that scrub jays, who hide and retrieve food items, know not only where specific items are hidden but how long they have been hidden there. If these animals are incapable of ‘episodic’ memory as Tulving claims, then their ‘semantic’ memory must include a time parameter. As for ‘episodic’ memory, the best way to catch the sense of it is to consider a case study (also mentioned by Schacter 1996, 148ff.) which Tulving presents at some length. The patient, here called ‘K.C.’, ‘cannot remember anything that has happened to him’: although his memory for facts (‘semantic’ memory) is near normal, including facts about his own past, he cannot recall the experience of living through events in his life. For instance, K.C. ‘knows that he owned a Black Honda, but does not remember a single trip he ever took in it’ (p. 24). Thus the difference between ‘semantic’ and ‘episodic’ memory can also be expressed as the difference between knowing and remembering. Tulving also speaks of ‘episodic’ memory as ‘mental time travel.’ Being incapable of this, K.C. lacks any sense of subjective or ‘felt time’ (p. 29).
According to Tulving's thesis, such a deficit is exceptional among human adults but is the rule for other animals, who are limited to ‘semantic’ memory. This would entail that scrub jays, like K.C., know about a hidden food item (temporal dimension and all), but do not remember the act of hiding it. Likewise a chimp, lacking ‘episodic’ memory, might know the individual traits of others in his troop and yet be unable to recall past interactions, or imagine future encounters, with those others. This last point is crucial for Tulving, who observes that K.C.’s deficit has rendered him incapable of anticipating his own future; on this he bases a theory that ‘episodic’ memory is a necessary condition for future planning. This circumstance confers the adaptive value which accounts for the selection of ‘episodic’ memory in humans (though that would not explain its absence in other primates). Tulving also claims that his attribution of ‘episodic’ memory exclusively to humans is testable without recourse to linguistic reports of the subject's experience. Though he specifies the criteria for such a test in some detail, he does not describe in concrete terms an actual experiment that would apply these criteria. This leaves his thesis on dubious ground, especially considering that at least one primate (Panzee, described below) seems to have passed Tulving’s test already, as the editors remark in their introduction (p. xix). As Tulving must know about this case, we can infer that he does not consider Panzee to have met his criteria, but his article here does not explain why.
This brings us to the primatologists’ research reports, which furnish this collection with its weightiest content. Some of these also outshine the more theoretical pieces in the care given by the authors to interpreting their empirical results and considering alternative interpretations. All of them investigate various forms of ‘metacognition’—for instance, the ability to judge the reliability of one's own judgements. This is J. David Smith's focus in his studies of ‘uncertainty monitoring’ in monkeys and dolphins. He used perceptual discrimination tasks which included the option of an ‘escape’ response, use of which can plausibly be attributed to uncertainty. The subjects in such tests do appear in some sense to know how reliable their own judgements are, though it does not necessarily follow that they are conscious of those judgements. Similar or related forms of metacognition have been tested, with similar results, in a study reported here by Lisa Son and Nate Kornell, and in another by Robert Hampton.
One limitation of these studies is that they don't demonstrate the kind of ‘offline’ metacognition that would require the object of ‘self-reflective’ attention to be represented in long-term memory. In other words, they present no challenge to the claim that pre-human (including primate) cognition is ‘confined to the here and now in the sense that memory is evoked in relevant circumstances but is not accessible to voluntary recall’ (as Nelson puts it, p. 121). A much stronger challenge to that claim is presented by studies such as Bennett Schwartz's work with a gorilla named King, and especially Charles Menzel's with Panzee, the chimp whose picture graces the dust jacket of the book. As Panzee is the one who (may have) passed Tulving's proposed test for episodic memory, we will need to look at her case in some detail.
Panzee was (at the time of Menzel's study) an 11-year-old female who had already learned over 120 lexigrams (visual symbols), not by trial training but by interacting in a social environment which included their use in everyday contexts. She lived in a double enclosure: an outdoor cage separating her from the surrounding forest (to which she had only visual access), and an inner room affording no view of the outdoors; both parts were furnished with lexigram ‘keyboards,’ which Panzee was accustomed to using (along with gestures) to interact with her caregivers. In Menzel's primary study, the experimenter would stand outside Panzee's enclosure while she was in the outer part of it, show her a test object (a different one in each of the 34 trials), and then hide it in the forest up to 8 meters from the cage, in a different spot each time, covering it with leaf litter and twigs to completely conceal its presence. The experimenter then departed, and the caregivers were not informed about the object or its location, or (in most cases) even that an experiment was going on that day. Sometimes the hiding was done after they had left for the day, thus imposing an overnight delay on Panzee's interaction with them. But in every case, she eventually got a caregiver to retrieve the object for her, by initiating a fairly complex interaction with the caregiver. She would first have to get the caregiver's attention indoors, and then use a combination of gesture, voice and lexigrams to request that he go out to the forest while she went to her outdoor enclosure. From there she would direct him, by pointing and vocalizations, to the spot where he would find the object, typically in less than a minute. It was part of the caregiver's routine to record all this, including the lexigrams used by Panzee, which were always appropriate to the hidden object and not to those used in previous trials. The interval between the hiding observed by Panzee and the retrieval under her direction could be anywhere between a few minutes and several days, and she clearly initiated each of the interactions leading to retrieval. Perhaps everything Panzee does can be explained in terms of ‘semantic’ memory; but perhaps everything humans do can be explained in the same way, if we eliminate the subject's report of his own experience from consideration. (How could we detect the absence of autobiographical memories in Tulving's patient K.C. if he couldn’t verbally answer our questions about what he remembers?)
Josep Call takes another tack in his studies of ‘mental state attribution’, showing that all four kinds of apes (but not dogs) typically ‘seek information’ in situations where they can use it. This serves as a ‘metacognitive index’ because seeking information implies some awareness of a need for it. Call's article is also a good choice to wind up the collection because of its emphasis on social cognition. In this respect he reinforces the contribution of Nelson, who outlines six ‘emerging levels of consciousness in early human development’. If the analogy to evolutionary process is salient, this suggests a compromise solution to the ‘tug of war,’ by replacing both the singularity of a ‘missing link’ and the vague notion of ‘numberless gradations’ with a small number of definite stages, each building on its predecessors. Perhaps the social factor explains why the most compelling evidence for the ‘continuity hypothesis’ comes from studies of apes (King and Panzee) who have lived most of their lives in a context of interaction with humans. Studies of primate societies, many of them presented informally by Frans de Waal in Our Inner Ape, demonstrate a complexity of interaction that would obviously demand cognitive abilities beyond anything documented in The Missing Link—perhaps because so many of the experiments reported here not only ignore social interaction but labor to eliminate it entirely (especially between subject and experimenter), apparently doing this in the name of scientific rigor (as Terrace sees it). The problem here is that if we downplay the social context of ape cognition while emphasizing its importance in the development of human cognition, we stack the deck in favor of human uniqueness. However, a careful reading of this volume may lay this bias bare, and thus open up some deep questions about the supposed objectivity of empirical methods in comparative psychology.
We might even conclude that, on balance, The Missing Link in Cognition supports the continuity hypothesis, though much more cautiously than popular explorations of the territory such as de Waal's. Perhaps this caution would explain why The Missing Link avoids delving into socio-political implications of this inquiry, which de Waal emphasizes—and rightly so, for it may be perilous to ignore those implications, given the current state of the planet. Thomas Berry even asserts (in the opening pages of The Great Work, 1999) that ‘the deepest cause of the present devastation is found in a mode of consciousness that has established a radical discontinuity between the human and other modes of being.’ If The Missing Link in Cognition helps to show how much of our mental life we share with our primate cousins, as I believe it does, then this message is all the more convincing because its lead authors/editors do their best to resist that inference. A book like this can help to shift us toward a better understanding of the human role in the natural order, precisely because it doesn’t preach to the converted.
Gary Fuhrman