Abstract
The conceptual foundations and ontology of cognitive neuroscience are rarely analysed in cross-cultural perspective, although they are manifestly the outcome of historical currents in specifically Western psychological science. How robust such concepts are, and how generalizable to other cultures, is thus quite problematic. Users of empirical techniques in imaging neuroscience are now actively exploring such topics as attention, volition, emotion and empathy, but with little awareness of how well or badly these concepts can be translated. This essay addresses issues of cultural bias and the potentially misleading use of extended metaphors in the typical deployment of mentalistic terminology, and suggests that there may be alternative conceptualizations, perhaps inspired by phenomenology, which would have less cultural baggage. Ultimately, the most scientifically useful ontology for interpreting and predicting human action may result from an integration of high quality ethnographic reports of mentalistic concepts and terminology found in other cultures. Social and cultural anthropologists are urged to prioritize the identification of such concepts during their fieldwork experience.
Keywords
Introduction
Neuroanthropology and the desirability of cross-cultural robustness for cognitive neuroscience hypotheses
In this essay I shall be considering the concepts, terminology and ontology of what is generally termed ‘cognitive neuroscience’. This rapidly expanding and prestigious field of research attempts to discover reliable relationships between observable human behaviour, postulated models of thought processes (conscious or unconscious), and measurable localized brain activity. The parent discipline, cognitive science, grew up in the 1950s as a reaction to the sterility of behavioural psychology, reintroducing the concepts of mind and meaning into attempts to understand and interpret human action.
Cognitive models often result from behavioural experimentation. If such models successfully predict brain activity measurements, they are considered to be scientifically established, and thus usable as the basis of further modelling. Typical examples are described in a highly instructive review by Price (2000), who shows how cognitive language models, sets of linked boxes connected by arrows, have developed over the last two centuries.
It should be hardly necessary to point out that for the validated hypotheses of cognitive neuroscience to be considered steps forward in our understanding of ourselves, they should make sense in a range of human cultures, and, more importantly, they should be meaningful where studies include human subjects from a range of cultures. Without this, we might be simply massaging our own cultural prejudices or, worse still, reifying a pointlessly divisive ideological perspective.
The argument of this paper has the following steps. Firstly, I will consider the degree to which many concepts of cognitive neuroscience have been derived from an ontology that is not necessarily shared across cultures, and why this may have arisen. I will then discuss the diverse ontological status of terminologies in the subdisciplines relevant to brain science. Next, I will provide illustrative examples of the sometimes poorly validated metaphorical extensions of everyday language that are taken for granted in the field of cognitve neuroscience. I suggest that the phenomenology of Husserl and Merleau-Ponty offers an alternative conceptualization of the relationship between perception and action. Finally, in an attempt to escape from the Whorfian dilemma that we can only think discursively about that for which we can find words, I propose an intensified study of the mentalistic terminology in everyday use in other cultures, in the hopes that commonalities may be discoverable that offer a more secure, less culture-bound ontology for cognitive neuroscience.
It is not the purpose of the present paper to summarize comprehensively the philosophical debate that has been developing over the last few centuries regarding appropriate concepts and terminology for describing the workings of the human mind. Indeed, the present paper is by no means the first critique even from within neuroscience of the constructs of cognitive neuroscience. In their controversial books Philosophical Foundations of Neuroscience and History of Cognitive Neuroscience, the neurophysiologist MR Bennett and the philosopher PMS Hacker have already attacked the conceptual incoherence associated with the metaphorical use in neuroscientific explanations of such terms such as ‘representation’ and ‘map’. Recent work by the group of the senior neurobiochemist RG Shulman (van Eijsden P et al. 2009) also attacks the standard cognitivist model. The response to these criticisms by cognitive neuroscientists so far suggests that they have elided in their own thinking the distinction between literal and metaphorical meaning to such an extent that they can no longer recognize the sweeping implicit assumptions that this entails. This paper should thus be seen as a call to arms for ethnographers who, far from becoming dependent on cognitive scientists for their analytic terminology, may be able to provide a far more robust set of concepts based on much wider human consensus.
It is easy to overestimate the generality of the mentalistic constructs of cognitive science in remarking the diversity of national origins of cognitive scientists, and the diversity of countries which pursue research in this discipline. The field is currently dominated by American, British, German, Italian, French and Japanese researchers, with important contributions from Australia and Canada. Chinese research in this area is rapidly developing. However, cognitive science is only just emerging from its historical origins in American and British behaviorist psychology, with its insistence that human action can only be explained scientifically by correlating behavioural output (‘behaviour’) with experimental input (‘perception’). Cognitive science adds ‘mechanisms’ in the head to account for such observable relationships, which can at least be localized through the techniques of functional neuroimaging, even if their operational details remain almost completely obscure. Combining the search terms ‘mechanism’ and ‘cognition’ results in 3248 hits on MedLine, characteristic examples being Carhart-Harris (2010) and Friedman et al. (2011). Cognitive scientists are rarely interested in cultural differences, and while broad distinctions between ways of thinking of Western and East Asian cultures have recently received considerable attention (e.g. Goh and Park 2009; Uchida and Kitayama 2009; Varnum et al. 2010), there is no sign yet that this has led to fundamental revisions of psychological concepts. These use primary Western psychological constructs and methodologies to identify systematic differences regarding how East Asians and Westerners see themselves socially. The emerging discipline of ‘social cognitive neuroscience’ bases itself firmly on traditional Western mentalistic concepts, and the cultural appropriateness of the type of experimental paradigms used in such cross-cultural research is rarely questioned.
A relatively survey of cognitive neuroscience literature provides not a single example of a study in which a ‘psychological’ concept traditional to a non-Western culture has been central to the investigation. The closest approach to this is the work of Singer (e.g. Leiberg et al. 2011), which takes some of its inspiration from the Tibetan Buddhist concept of compassion meditation. However, strikingly, no traditional Buddhist authors appear in the list of references for this paper, which cites only secondary sources. Neither of these research directions includes the systematic ethnopsychology which is the topic of the present paper, which can only be honestly performed using the methodology of participant observer ethnographic fieldwork. The desirable bridge between cognitive psychology and ethnopsychology apparently does not yet exist.
The high-impact British and American journals that publish cognitive science research have a striking reductionist-materialist bias. Given their dominance and command over academic recognition, alternative formulations of human brain activity that owe more to James than to Skinner have been relatively slow to emerge. Fresh approaches to the mind-brain problem, in the hands of Varela and Gallagher, to name but two authors inspired by the phenomenology of Merleau-Ponty, are gradually finding grudging acceptance. For a review see Ellis (2006). Especially in American cognitive neuroscience, however, materialist reductionism is often an unquestioned article of dogma (see Bickle 2007) rather than a pragmatic strategy for discovering simple and law-like generalizations about such complex assemblages of adaptive matter as the human brain. It is worth commenting, as Midgley puts it (2004: 36), that reductionists feel that ‘there is still something unofficial about everyday language, because it speaks of entities such as human beings and homes and prisons that are not in the repertoire of physics, entities that cannot, therefore, be quite real’. This view problematizes common-sense ontologies and conceptualizations of our selves and our relationships, attempting to embody generalizations in technical terminologies that may be far from ordinary lived experience. This will be discussed further below in the context of ethnography.
Shore, in his penetrating analysis of the direction and potential of cognitive anthropology (1996: 317), comments perceptively that in the 1960s and 1970s
there was a subtle but critical shift in the significance of ‘ethno’ from ethnoscience to ethnopsychology. … The issue was no longer the study of cultural variability in basic human cognitive processes but the use of that variability to deconstruct and undermine the generalizing enterprise of scientific psychology itself. The effect was to throw into question the relevance of the categories of academic psychology for other cultures. Anthropology and psychology went their separate ways.
Neuroanthropology, conceived afresh by Dominguez et al. (2009a, 2009b) as a discipline bridging cultural anthropology and cognitive neuroscience, provides an arena for well-informed and even-handed discussion of these important topics. Built into its definition is a dialogue between these disciplines, in which cognitive neuroscience can provide realistic constraints on the variety of cultures. This should allow ethnography to become yet more empirical, while at the same time cultural anthropology can refine and generalize the constructs of neuropsychology, making them stronger and more likely to relate to the underlying neuronal mechanisms and algorithms that can be supported by a mammalian brain.
The terminologies of cognitive neuroscience
As an integral component of modern cognitive neuroscience, the emerging discipline of imaging neuroscience uses specialized language drawn from at least four well-established branches of science: physics, neurology (including neuroanatomy), physiology, and of course psychology. Examples of physics-based terms are fMRI, EEG, BOLD contrast, perfusion, action potential, and structural connectivity. Examples of neurological terms are hippocampus, thalamus, Brodmann area, gyrus, sulcus, cortical layer, cortical column, neuron, synapse, epilepsy, neurodegeneration, and so forth. Cognitive neuroscience also uses physiological terms such as motor control, action, and response. These definitions of physical, physiological and neurological terms are relatively unproblematic, because they apply to concrete objects or physical processes which can be identified by scientists with appropriate expertise, and described to others with any educational and cultural background. They support an ontology that is consistent across the parent disciplines of physics, physiology and neurology.
This is not the case, however, for many of the terms used in psychology. As a science, this discipline was established largely by brilliant Germans in the 19th and early 20th centuries, such as Weber, Fechner, Wundt, Wernicke and, in a certain sense, Freud. This makes the following example all the more striking. Consider a dialogue between a British psychologist (A) and an English-speaking German (B) without university training.
B: What is your profession? A: I am a psychologist. B: And what exactly do you study? A: The human mind. B: The human what?
The point of this example is that the German language, although it is etymologically very closely related to English, does not contain a word that precisely translates the English concept ‘mind’. In German, ‘mind’ is translated as Geist or Verstand. Geist also means spirit, ghost, and wit, while Verstand also means reason. Indeed, many European languages have no term that covers precisely the same range of meanings as ‘mind’. Perhaps the Spanish ‘mente’ or the Greek ‘nous’ come closest. French, which is also very close etymologically to English, uses the word ‘esprit’, but this also covers the senses of ‘spirit’, ‘wit’ and ‘personality’, none of which is included in the English ‘mind’. Japanese has five words that can be roughly translated as ‘mind’, but when a close equivalent is needed the term ‘maindo’ is used, simply a borrowing from English, because all the indigenous terms (omoi, nentou, kokoro, seishin, munesanzun) carry other unwanted connotations.
Native English speakers are so comfortable with the idea that such a thing as ‘mind’ has ontological validity, that is to say objective existence, that whole books are written with optimistic titles such as ‘How the Mind Works’. However, as Taylor Carman (2008: 37) points out, glossing Merleau-Ponty, ‘other minds are not directly perceptible, and my awareness of them must instead be founded on an analogical transfer from my perception of myself as an embodied consciousness’. This suggests that mind itself, like redness, is a ‘quale’ to use the term in Dennett's sense (1988). The great neuropsychologist Antonio Damasio, in a flourish that can only be described as poetic, tells us that ‘minds are a subtle flowing combination of actual images and recalled images, in ever-changing proportions’ (Damasio 2000: 71), and that ‘minds can be either conscious or non-conscious’ (2000: 72). Maybe this is a word that can mean whatever we want it to mean.
Psychological language is full of such terms describing attributes of a person, which have to be regarded as abstractions. The list is long. Consider perception, consciousness, intention, attention, memory, altruism, conflict, empathy, attachment, cognition, emotion, representations and arousal. All of these terms are used widely in ordinary English speech and writing. We use them confidently, as if we know exactly what we mean by them. But none of them, in general usage, refer to anything to which we can point, and they are typically difficult to quantify. Imagine, for instance, trying to describe the colour or smell of a person's intention. Even ostensive definition of material objects, as Tallis (2010) elegantly argues, has its philosophical problems, but these pale into insignificance compared with defining mentalistic concepts.
A further problem arises from the etymology of many of these terms, which can often be dead metaphors (as CS Lewis described them in 1962). For instance, attention derives from Latin ‘attendere’, to give heed to, literally, ‘to stretch toward,’ from ad- ‘to’ and tendere, ‘stretch’. This association carries with it a context of agency, making it harder to conceive of a state of passive attention, for example. It is rare for metaphors to stay dead, and unwanted associations can strongly bias subsequent use of terms.
Some researchers take pains to provide acceptable practical definitions of psychological concepts, notably Damasio in regard to ‘consciousness’ and ‘emotion’ (2000). Others take them for granted, as if their presence in wide literatures somehow guarantees their validity. (There is also a wide literature on the Hindu god Ganesh, but this does not in itself ensure the existence of such an entity.) Regarding ‘consciousness’, the novelist and philosopher Michael Frayn comments perceptively, ‘we can't even say what sort of thing it is. It lurks like an outlaw on the boundary between philosophy and science, and continually slips away across the frontier to elude capture by the forces of law and order from either side’ (2006: 400).
Not all terms used in psychology are so intrinsically vague. Words such as ‘choice’, ‘conditioning’, and indeed ‘working memory’ can be empirically defined and thus quantitatively characterized. And, on the other hand, many words need to remain vague for them to be flexible enough for use in ordinary language. Indeed, efforts to constrain their definition can suffer from two drawbacks: they become reified, so their significands become treated as ontologically unquestionable; and they become corrupted insofar as they are thus co-opted into an implicit or explicit ideology, serving as instruments of political power. When this occurs, such words lose much of their scientific utility. When underlying neurobiological processes are asserted by authorities to exhaust the content of commonly used words, such as ‘will’, to take a contemporary example, one must invoke George Orwell. Such authors as Mark Hallett, Wolf Singer and Patricia Churchland take it for granted that Western concepts as intention and emotion have ontological validity, and then search for neural correlates, which they can of course find, which can then in turn legitimate a restricted range of meaning. But this does not mean that the concepts themselves are sound. Any data set can be fitted with an indefinite number of theories. It is not within the scope of this essay to elaborate this point, except to mention the most salient example, the use of the term ‘behaviour’ in relation to human action, which implicitly but ineluctably discounts human agency.
Cognitive neuroscience analyses of the generality and validity of mentalistic terminology
It is not rare to find careful attempts to define terms in cognitive science. It is also recognized that this may be problematic. Consider again the term ‘attention’. William James (1890) remarked: ‘Everybody knows what attention is.’ Within academic Western cultures, this is probably true. Wikipedia defines attention as ‘the cognitive process of selectively concentrating on one aspect of the environment while ignoring other things. Attention has also been referred to as the allocation of processing resources’. It is striking how little congruity there is between these two definitions, both of which are entirely consistent with the use of the term in imaging neuroscience research. The first is simply circular, defining one familiar mentalistic term in terms of several others that are equally familiar, while slipping in the buzzword ‘cognitive’ to sound more scientific. The second tries to avoid mentalistic terms, but drags in two tendentious metaphors, ‘processing’ and ‘resources’. Here the mind is simply conceived of as matter organized into a kind of factory, with an invisible manager who allocates a limited number of machines for processing the raw material of perceptual experience. How well would a person with no cultural experience of factories appreciate this definition?
The Wikipedia article continues:
Attention remains a major area of investigation within education, psychology and neuroscience. Many of the major debates of James’ time remain unresolved. For example, although most scientists accept that attention can be split, strong proof has remained elusive. And there is still no widely accepted definition of attention more concrete than that given in the James quote above. This lack of progress has led many observers to speculate that attention refers to many separate processes without a common mechanism.
Another contestable concept is ‘intention’. A deeply debated term among philosophers, it has become crucial to neuropsychological understanding of such conditions as autism, for which the term ‘Theory of Mind’ (ToM) has been coined. This denotes the normal human capacity to infer the ‘beliefs’ and ‘intentions’ of others (which may include animals and spiritual beings) from their actions and utterances. Sufferers from autism are considered to be lacking in ToM. An alternative word for the same capacity is ‘mentalizing’, favoured by Uta Frith (Happé and Frith 1996). Over 200 empirical studies, including imaging neuroscience studies, have been conducted using this concept (e.g. Gallagher et al. 2000, Castelli et al. 2000, Gobbini et al. 2007, Pelphrey and Carter 2008, Desmurget and Sirigu 2009). Kaiser and Weisbrod (2007) have reflected on intentionality more deeply than most neuropsychologists, leading them to the phenomenological approach of Husserl (1973, 1991), who has proposed a layered account of intentionality, starting with the basic organization of experience in time, space and body, and rising to a final level of intersubjectivity. In an experimental neuroimaging study of intention, Gobbini et al. (2007) distinguish between the reading of motor action as goal-directed, on the one hand, and the representation of covert mental states that may predict behaviour, on the other, suggesting that these may involve distinct neural systems. The term ‘prospective memory’, denoting our capacity to remember what we intend to do, is gaining prominence in cognitive neuroscience studies, although it remains entirely obscure how we make the necessary distinction between what we actually have done, and what we intend to do.
It is not possible here to provide a critical review of all such studies, but it is clear that the term ‘intention’ is used in several different ways in cognitive science, and has undergone several different attempts at deconstruction. Ultimately, what will be of most help to brain science in understanding human action is a description of areas of cross-cultural consensus regarding intentionality, since concepts derived from such data are likely to have firmer ontological grounding, and are thus more likely to be associated with specifiable operations of neuronal assemblies.
Perhaps surprisingly, the concept of perception is itself problematic, and it is becoming more so as a result of recent neuroscientific research. Visual perception, which has been the most thoroughly studied, can be treated as the classic example. Brain scientists since Richard Gregory (1966) have come to recognize that vision is pre-emptive: the relatively thin stream of action potentials that arise from the eye and the next synaptic node, the lateral geniculate nucleus, is dwarfed by the rich flood of neuronal signals coming from upstream brain areas with progressively more specialized feature detection capacities. What constitutes an act of perception is apparently not a passive logical deduction of the identity of what is visible, but a confirmation that the brain's Bayesian guess is correct (Wolpert 2007). This is almost reminiscent of the view of the ancient philosophers, such as Plato and Euclid, who held that the eye itself illuminates what it sees. Our brains form these guesses from what they have retained through probabilistic learning of previous experience, a learning which is implemented through the material adaptation of neurons, each evolved to be appropriately adaptive, taking easily the imprint of experience (Turner 2002: Turner and Whitehead 2008). Learning, plasticity and memory are merely different labels for the same biochemical and biophysical processes. Indeed, as Frayn (2006: 35) puts it, ‘perception and memory have imperceptibly merged’.
At this level, not only perception and memory fail to be distinct. Perception and action themselves can be argued to be deeply intertwined. Classical cognitive psychology assumes three stages of ‘information processing’: perception, cognition and then action, in a kind of sandwich where perception and action do not interact directly. The buzzword ‘cognition’, whose meaning is (alas) somewhat obscure, is typically used to describe the ‘processing’ required to convert perceptual representations into action, characterized as a ‘mapping’ between ‘sensory’ and ‘motor’ codes.
A contrasting account known as the theory of common coding, developed by Wolfgang Prinz and followers (e.g. 2005), claims instead that perception and action are directly linked by a common computational code, giving parity to perception and action. Actions are assumed to be coded in terms of their perceivable effects, and perception of an action should activate action representations to the degree that the perceived and the represented action are similar. Thus observed, executed and imagined actions are represented commensurately in our brains. Specific predictions can then be made regarding these action and perceptual representations, that representations for observed and executed actions should rely on a shared neural substrate, and that action should be facilitated by observation of the same action directly beforehand (and vice versa). Furthermore, interference effects are predicted when action and perception attempt to access shared representations simultaneously. The implication is that at a neural level the distinction between action and perception is hard to make. Since, like the concept of mind itself, this distinction feels ‘natural’ to those of us who are raised in Western culture, the question is again raised: how do other cultures classify these mental operations?
This question can only be addressed by careful ethnography, and not by cognitive psychology. It is already possible that buried in many monographs lie clear and comprehensive accounts of emic classifications of action and perception. If so, anthropological readers of this paper, expert in particular cultures, are invited to provide examples. Explicit focus on this question may enable guide cognitive psychology to become more empirical.
The relevance of phenomenology
We have already alluded to the work of Husserl and Merleau-Ponty. In the context of this essay, their philosophical stance known as ‘phenomenology’ takes on a particular importance, because it represents an alternative, and perhaps more parsimonious, way of characterizing what goes on inside our bodies (now including the brain), that avoids some of the mysteriousness of mentalistic language. Since the basic insights of phenomenology are not drawn from contemporary neuroscience, the very fact of their existence gives some hope that well-informed, thoroughgoing ethnopsychology – defined as ethnographies of the mentalistic terminology employed in other cultures – will enable more robust, more generalizable sets of concepts that can be operationalized with the aid of neuroscientific techniques such as neuroimaging.
‘Merleau-Ponty's central philosophical idea is that perception is a bodily phenomenon, not a mental event occurring at the end of a chain of physical causes and effects, as Descartes supposed. It is the body that perceives, not the mind’ (Carman 2008: 26). The phenomenological approach is thus clearly compatible with materialist neuroscience, but it avoids the behaviourist mistake of excluding personal agency. Merleau-Ponty is by no means an ally of behaviourism:
The introduction of Phenomenology of Perception … contains Merleau-Ponty's critique of traditional psychology and its uncritical reliance on abstract concepts such as sensation, association, memory, attention and judgment. It is crucial to appreciate, however, that in criticizing such concepts, he is not denying that our mental life is indeed rich and complex in ways that virtually force us to avail ourselves of words like these in describing it. (Carman 2008: 44)
To attempt to avoid the Cartesian error, as many neuropsychologists now do, by claiming that the brain is the organ of the mind gives the mind a strange ontological status. It is now characterized as an abstract concept that somehow controls a physical object. This may turn out to be the best possible pragmatic description of what takes place – but it would be bad science to start with this assumption.
Ethnographic and linguistic evidence regarding the strength or weakness of cognitivist terminology
In anthropology, according to Duranti (1993):
such analytical concepts as truth, interpretation, and intentionality are hardly discussed in the ethnographic literature. The relationship between people, their thoughts, and their actions is a major topic of ethnographic description, but anthropologists tend to avoid philosophical discussions about such relationships. Confronted with what they consider to be too abstract and culturally deprived characterizations of human thought and human action, most ethnographers tend to stay away from debates about philosophy of mind or philosophy of language.
However, once one dips into ethnographic writing, cultural differences regarding intentionality, for instance, are easy to find. Take, for instance, the geographically not very distant Sudanese societies of the Azande, as described by Evans-Pritchard (1937), and the Dinka, described by Lienhardt (1961). Azande people, as in other societies with strongly-held witchcraft beliefs, consider many events that we would ascribe to chance to be the result of the personal intention of a human witch. By extreme contrast, Dinka people prefer to ascribe many events and circumstances to divine powers, and not to human intention at all. Even one's own acts can be attributed to external powers, a theme often found in traditions of religious mysticism.
Maurice Bloch (1998) has addressed the forms of memory in the Malagasy society he studied, finding interestingly contradictory collective and individual accounts of events important in his informants’ pasts. For members of Western cultures, there is at least a pretence that there is an ‘objective’ history, so that wildly discrepant accounts are frowned on. Of course, in practice, history books change (an extreme example can be found in the complete exchange of history books after the fall of communism in East Germany). It is clear that across cultures even the concept of memory is problematic.
Conclusion: How cognitive science needs ethnopsychology
If we want to understand our human minds (as they are defined in English-speaking cultures), and their relationship to our bodies and brains, we need to start by recognizing human diversity. The opaqueness of many cultures, overcome only by years of ethnographic fieldwork, should be enough evidence that we do not all think alike. But thus far in its development cognitive neuroscience has employed a limited and stereotyped set of mentalistic concepts, forcing all volunteer human subjects onto a Procrustean scanner bed, regardless of their various socially-ingrained conceptions of themselves, and of what we would call their memories, their intentions and beliefs, their emotions and sympathies, and their loves and hates.
A serious and systematic inventory is urgently needed of mentalistic concepts and terminology found in other cultures, widely ranging from pastoralists, through sub-Saharan African societies, through South American jungle cultures, to highly-developed but very foreign civilizations such as the Japanese. Such an enquiry fits easily under the term ‘ethnopsychology’, an endeavour which has previously led anthropology away from psychology, as Shore points out (see above). There is a vast resource of published good quality fieldwork which could be sifted with this goal in mind.
What should be sought is not a ‘least common denominator’ set of mentalistic constructs, because, as with colour categories (Levinson 2000), it is possible that some societies have little or no use for any such constructs. Instead, what is needed is to find more of a broad-brush consensus, identifying concepts that are easily defined and demonstrated, and that allow reasonably precise specification in most cultures without doing violence to their usage in general parlance. As has been claimed for colour terminology, it is at least possible that such constructs would reflect universal human brain structures specialized for particular aspects of human social activity.
How can this be done? Of course, the starting point has to be language. The first challenge will be to find cross-culturally generalizable ways of showing one's interlocutor what one means. When dealing with mentalistic terminology, this will already need considerable familiarity with the forms of communication of the other person's culture. It takes gifted ethnographers years to unpack the rich semantics of ritual symbols (e.g. Turner 1967). To be confident that one has appropriately matched Western and local concepts of mental attributes can be expected to require as much hard work and repeated careful cross-checking. Once one is reasonably sure that one is using the local terms correctly, further work will be needed to establish the social contexts in which such terms are used, and the importance of the terms in driving social action – all part of their meaning.
The reward, in gaining stronger concepts, may be enormous for cognitive neuroscience. Such a study may help to lead anthropology back, in partnership with a re-formed neuropsychology based on cognitive neuroscience, to its rightful place as the most empirical of social sciences.