Translation and Cultural Comparison: Some Epistemological Reflections

In its traditional guise, cross-cultural psychology is a science of comparative measurement. Its chief method is the regulated observation or recording of behaviour in different cultural milieux and its quantitative conversion into a common metric. This method allows for the estimation of differences across cultures on the strength or magnitude of the behaviour. We can therefore speak of one cultural group doing more of Y, or being more Y, than some other group, where Y is a conceptually and operationally defined behaviour taken to apply equally well to both groups. The culture-independence of the definition supports examination of how the magnitude of the behaviour may be determined or conditioned by some cultural variable X. According to this picture, culture (the independent variable) differs from place to place and the features by which it differs produce more or less of the universally defined behaviour (the dependent variable) under study. The hypothetical contingency of ‘culture affecting behaviour’ that underlies this approach is thereby maintained. In this paper, I use the problem of translation to cast doubt on this picture of cross-cultural research and the inferential strength of its comparative method. In opposition, I argue that we never measure the same behaviour across cultures if behaviour is understood as socially significant action. Rather, the specification of analogous actions across cultures is a highly uncertain approximation involving the inevitable projection of one’s own conceptual scheme. I argue further that most social phenomena of interest to cross-cultural psychologists cannot be adequately defined in a manner that divorces them from the local linguistic conventions and normative frameworks through which they are realized as part of cultural life. For this reason, cross-cultural psychology cannot effectively model itself on the natural sciences. It is as much hermeneutics as psychometrics. ‘‘Culture is an observance. Or at least it presupposes an observance.’’ Wittgenstein (1977/1980) To study another culture is to translate. Comparing and equating words and deeds across cultures amounts to establishing correspondences in meaning between disparate conceptual schemes. In this process of harmonization, the question of whether a correspondence is ‘right’ has no unequivocal answer. Working analogies are often achieved through the unconscious projection of one’s own linguistic apparatus, commonsense and scholarly, upon the semantic field of a foreign way of life. Because full-blooded social phenomena are not merely described, but also constituted, by the linguistic rules implicit in the practices that comprise them (Taylor, 1995), even the best translation is never more than a forced alignment of discrepant accounts of social reality. Seen in this light, the comparative method as applied to cultural life takes on an epistemological character distinct in kind from that of the natural sciences. The realism of the natural scientist rests on the belief that the material under investigation exists regardless of our awareness and description of it. To whatever extent conceptual relativism applies to truth claims about the physical world, the latter is nonetheless taken by natural scientists to be what it is with or Social and Personality Psychology Compass 4/4 (2010): 227–237, 10.1111/j.1751-9004.2009.00255.x a 2010 The Author Journal Compilation a 2010 Blackwell Publishing Ltd without them (Putnam, 1990). This metaphysical stance is difficult to defend from within cross-cultural psychology. As social scientists, we take up intentional phenomena that exist by virtue of their conventional meaning in a linguistic community. That is, they exist through the shared conceptions people have of them. Any universal redescription of such phenomena in a physical, nonintentional language throws the baby out with the bathwater. Namely, the content of the phenomena as normative structures of meaning is lost. For this reason, the very marking out of a social action in a foreign culture is best seen as translation, an interpretive practice that occurs at the boundary of more or less divergent conceptual schemes – the researcher’s and that of the society studied. This view of things puts considerable strain on the naı̈ve claim that we are somehow able, through our scientific ingenuity, to discover and measure the same structures of meaning across cultures, or even that there are common meanings to discover and measure. Recognizing cross-cultural translation as an uncertain and often projective approximation demands that we evaluate the inferential logic of our comparative method and the truth claims it generates in a more honest light. Expansion of the above argument begins with the claim that the main challenge for cross-cultural psychology is, and always has been, translation. Making sense of this claim requires an understanding of cross-cultural psychology as a species of scientific inquiry. Rather than risk bias or distortion through offering my own definition of the field, I defer to several of its leading exponents: Cross-cultural psychology, defined broadly as we do here, comprises many ways of studying culture as an important context for human psychological development and behaviour...Crosscultural psychology consists mostly of diverse forms of comparative research (often explicitly and always at least implicitly) in order to discern the influence of various cultural factors, many of them related to ethnicity, on those forms of development and behaviour. In this comparative mode, culture is treated as comprising a set of independent or contextual variables affecting various aspects of individual behaviour...Cross-cultural research typically seeks evidence for such effects (Segall, Lonner, & Berry, 1998, p. 1102). There are more ecumenical definitions on offer that avoid the emphasis on comparison. These, however, tend to be so vaguely inclusive as to obscure important theoretical, epistemological and methodological differences between cross-cultural psychology and alternative approaches to the study of thought and action as cultural phenomena (e.g., cultural psychology, cultural studies, psychological anthropology, comparative sociology, discursive psychology). In fact, the bulk of cross-cultural research published in psychology journals over the past half-century has turned on quantitative comparison of the behaviour of cultural groups. Whether aimed at highlighting similarity, difference, or both, crosscultural psychology has been and continues to be a mainly comparative science. How, then, does cultural comparison entail translation, what problems does this present and what are their implications for our science? The unpacking of this question must be framed around a discussion of how social phenomena are essentially dependent on language and collective agreement and what this means for understanding cultures. Used as a substantive, ‘culture’ refers to a socially coordinated way of life in all its material and subjective aspects. Social coordination, as understood here, applies to a collective of individuals living together within an integrated system of institutionalized roles and practices. Such activity requires shared dispositions for experiencing and understanding oneself, others, and the surrounding world (Bourdieu, 1980/1990). Admittedly, the concept of culture is a convenient abstraction, a theoretical construct that should not blind us to the considerable diversity – even tension, opposition and contradiction – that 228 Translation and Cultural Comparison a 2010 The Author Social and Personality Psychology Compass 4/4 (2010): 227–237, 10.1111/j.1751-9004.2009.00255.x Journal Compilation a 2010 Blackwell Publishing Ltd exists within and across individuals in any society. Nor should the facile reference to culture suggest that its demarcation, or drawing of boundaries around societies that present distinct ways of life, is ever more than a matter of expediency. ‘Cultures’ are defined and framed by people. They are not natural kinds. The dramatic expansion and acceleration in material and symbolic exchange across global regions during the last half-century has made the separation of cultures an even more conventional affair. The most central institution within any collective way of life is the language used to define reality. Language is much more than a representational medium for communicating or inscribing the features of a lifeworld that exists independent of language. It is also the means by which much of that lifeworld is constituted. Although kings and cabbages are equally things in the world, the former depend on language and collective agreement to exist whereas the latter do not. Social reality is more like kings than cabbages, consisting mainly of ‘institutional facts’ that require conventional recognition and symbolic representation (Searle, 1995). There could be no kings without the linguistic conception and acceptance of that conferred status by others. The same could be said of baby showers, graduations, weddings, funerals and just about everything in between. On this theme, Wittgenstein (1958) remarked that understanding a language is understanding the form of life in which it is used. The reverse is equally true: one cannot understand a way of life, or culture, without understanding its primary medium, language. Let us look, then, at how cross-cultural psychologists take up natural language in their comparative work. Imagine a researcher who sets out to compare political attitudes in two societies. The societies possess markedly different cultures and phylogenetically unrelated languages. To measure attitudes in both societies in a way that allows for meaningful comparison of the resulting response distributions, she develops a set of questions written in both languages. The two versions are produced using methods that maximize correspondence in meaning. Of course, strictly equivalent meaning, or synonymy, is not possible across versions because of the semantic dependence of each word on the rest of the language, local background knowledge and the social and pragmatic context of use. Synonymy aside, there are better and worse approximations and the researcher does what she can to produce a pair of competent translations. She confirms her success in part by demonstrating high structural and metric similarity in the response patterns elicited by the two versions. Can we now be assured that she has established near equivalence in meaning? More importantly, can we interpret her numbers in much the same way as we would those of a biologist who compares, say, hepatic iron concentration in the bodies of two different mammalian species? Does the inferential logic of comparative measurement apply just as fittingly to her study as to the biologist’s? I would like to argue that the answer to all three questions is ‘no’. What separates the measurement of ‘the same’ political attitude in two dissimilar cultures from the measurement of the same chemical element in two dissimilar species is a fundamental difference of kind, not a matter of degree. I defend this position below. Quine (1960, 1969) gave epistemological shape to the problem in his argument for the ‘indeterminacy of translation’. He invites us to imagine a field linguist learning the language of a hitherto unknown people. To simplify the argument, he focuses on the translation of ‘occasion sentences,’ statements that depend for their validity on the context (time, place, etc.) in which they are spoken. Quine’s linguist notices that the natives exclaim ‘Gavagai!’ when pointing at rabbits that come into view. This observation leads to the testing and confirmation of the ‘analytic hypothesis’ that gavagai (native language) = there’s a rabbit (English). Testing is carried out by eliciting assent or dissent in various situations where rabbits are either present or absent. The linguist writes down the Translation and Cultural Comparison 229 a 2010 The Author Social and Personality Psychology Compass 4/4 (2010): 227–237, 10.1111/j.1751-9004.2009.00255.x Journal Compilation a 2010 Blackwell Publishing Ltd perceived equivalence in his field manual. Can we conclude that the linguist has successfully ‘discovered’ what the natives mean by gavagai? No, argues Quine, because there is no meaning to be discovered, if meaning is taken to refer to something that exists independent of both English and the native language. Putnam (1988) echoes this position in pointing out that ‘meanings are not objects in a museum, to which words somehow get attached’. If there is no separate language of thought, or ‘mentalese,’ which contains the truth of what the natives really mean by gavagai, then there is nothing outside of natural language to be translated or recoded onto English, French, Swahili, Arabic or – most significantly – the natives’ own spoken language. Instead, as Harman (1999) has argued, the natives can be assumed to think using their own mentally incorporated language and that is all there is to it. If they had been raised instead as fully bilingual, then they would think in two languages, not three. Positing a language of thought to account for a language of communication creates the regressive burden of accounting for the former by positing yet another representational system. Quine’s argument cuts deeper. If there is no mentalese that the linguist and the natives share whereby a meaning common to gavagai and there’s a rabbit is represented, then to what can we appeal in deciding whether gavagai = there’s a rabbit is the correct translation? All we have to go by, suggests Quine, is stimulus meaning, the sensory experiences that prompt assent or dissent to the word taken as an occasion sentence. The presentation of a rabbit prompts assent; the presentation of an elephant or a pitchfork does not. Stimulus meaning is empirical meaning. In the case of ‘stimulus synonymy,’ two occasion sentences have exactly the same stimulus meaning. Does stimulus synonymy ensure synonymy of meaning? No. Stimulus synonymy can exist alongside differences in referential meaning. For example, where the linguist understands there’s a rabbit as referring to an individual, unitary and enduring token of the species, gavagai could be alternatively translated as referring to temporal phases of rabbits, their undetached body parts, rabbits considered collectively rather than individually, the visible spirit of the rabbit and so on. The differences would not be apparent in light of stimulus meaning. The indeterminacy resides in the fact that competing analytical hypotheses can be adopted and used to build up systems of translation that are equally valid by the criterion of empirical meaning yet incompatible with each other. The incompatibility reflects the fact that advanced translation will overrun the possibility of empirical confirmation. As Quine puts it: Rival systems of analytical hypotheses can conform to all speech dispositions within each of the languages concerned and yet dictate, in countless cases, utterly disparate translations; not mere mutual paraphrases, but translations each of which would be excluded by the other system of translation. Two such translations might even be patently contrary in truth value, provided there is no stimulation that would encourage assent to either. (Quine, 1960, pp. 73–74). The choices made would depend on the individuating and quantification apparatus of the linguist’s own language and its projection into the native language to define objects. These choices would be made without thought or deliberation in the only way that seemed natural and sensible to the linguist. In other words, he would be unknowingly imposing his ‘ontological point of view’ on the cultural other (Quine, 1969). In relating the indeterminacy of translation to cross-cultural psychology, I am not claiming that the divergences envisioned by Quine are commonplace. Far from it. Translation in contemporary research is never the work of solitary ‘field linguists’. Nor is it undertaken from ground zero as in Quine’s ‘radical’ example. Researchers typically rely on bilingual interpreters and scientific collaborators who are adept at crossing between languages on the well-worn bridges that have been built up over years of linguistic diffusion and harmonization. 230 Translation and Cultural Comparison a 2010 The Author Social and Personality Psychology Compass 4/4 (2010): 227–237, 10.1111/j.1751-9004.2009.00255.x Journal Compilation a 2010 Blackwell Publishing Ltd What is directly relevant is the doubt that the argument for indeterminacy casts on our intuitive view of interlinguistic translation as hinging upon underlying meanings that are common to mutually foreign symbolic worlds. Although Quine’s argument has inspired considerable debate since its appearance, I know of no convincing refutation of its fundamental challenge to the mentalistic conception of fixed meaning. The import, for crosscultural psychology, is that there is no neutral zone of meaning on which to pitch one’s tent. Such a zone is a fiction. Translation of one language to another entails the penetration of one language into another. The best one can do is to allow for an adversarial balance of imposed structures through reciprocal translation. That is, one can slice up the world created by one language using the symbolic apparatus of another, while also cutting in the other direction with equal force. Practically, however, there is little chance of genuine balance in this regard. Research questions typically arise in and through a particular language. Even linguistically diverse teams of researchers are forced by the pragmatics of collaboration to frame and articulate their purposes and objects of study in a shared theoretical language. The semantics then move from that language into others, colonizing them with its scheme of interlocking references, distinctions and classifications. In this movement, statements generated in the first language to query the reality defined by that language are reconstructed, however painstakingly, using the improvised tools of ‘second’ languages. Given the social, economic and political structure of cross-cultural psychology as a collective endeavour, the first language has overwhelmingly been one of the Western world and the second one of the non-Western world. The cumulative result has been the Westernization of linguistic meaning in theory and research, the best intentions of practitioners notwithstanding. Let us now look at how the problem of translation applies to nonverbal action. One desperate response to the inevitability of projection in linguistic translation is to avoid ‘folk psychological’ description altogether. According to this strategy, we restrict ourselves to studying what psychologists call ‘behaviour’ using a formal observation language that is free of the constitutive work of natural language. Descriptions of this kind require no translation because they are expressed in the universal language of physics, equally applicable to behaviour in any cultural setting. Behaviour itself becomes reduced to matter in motion, shed of any implicit assumptions regarding its subjective significance for those who happen to speak one language as opposed to another. In this way, the ‘dependent variable’ of behaviour is cut away from the ‘independent variable’ of culture (which includes natural language). This separation is loyal to the conceptual duality and doctrine of mechanism implicit in the definition of cross-cultural psychology as the study of ‘how cultural differences affect...individual behaviour’. Without a clear distinction between behaviour and culture, how could we justifiably claim to be comparing a behavioural variable across cultures? What would we be comparing by way of measurement if not a common object? A common object requires a common rubric. If the ‘behaviour’ in question cannot be defined independent of its cultural significance, then the causal contingency at stake in the comparison reduces to a snake swallowing its tail. The only alternative would be to give up the claim that cross-cultural comparison is aimed at discovering causal contingencies. More on that later. The conceivability of a purely physical observation language for studying culture is, unfortunately, a nonstarter. It should be obvious that any language that refers only to natural kinds, motion in time and space, and other more or less self-standing content would fail to address most of the social phenomena that human agents, including cross-cultural psychologists, are concerned with. Describe the lawful microstructure of a baseball game as changes in physical states over time and you have described a very complex physical Translation and Cultural Comparison 231 a 2010 The Author Social and Personality Psychology Compass 4/4 (2010): 227–237, 10.1111/j.1751-9004.2009.00255.x Journal Compilation a 2010 Blackwell Publishing Ltd event, but not a baseball game. The meaning of terms such as home run, strike, flyout, designated hitter and triple play are bound up in the constitutive rules of the game. These rules are formed through language and understood in common by those who play and watch the game. Take away the rules and you lose the meaning of the terms. Take away the terms and you lose the game. More broadly, any description of socially significant action by means of a universal observation language will in most cases fail to capture the intentionality of that action (i.e., what it is about). Human behaviour becomes action when it means something for the actor. This involves choosing to act for a reason, with particular ends in mind. Putting aside the long-standing philosophical debate about whether reasons are genuine causes or just descriptions internal to the action, what is clear enough is that understanding what the natives (this includes us) are up to in a social situation requires knowing how they define the situation and its participants and what each believes himself or herself and the others to be doing. This is a richly intentional conception of what is going on, involving indigenous psychological predicates and the conceptual framework of which they are a part. To illustrate the problem, consider Ryle’s (1971 ⁄2009) anthropologically celebrated example of the winking schoolboys. If we describe the schoolboys’ successive winks as mere contractions of eyelids, we fail to capture the social meaning of their actions, whether it be one boy’s conspiratorial gesture, a second’s parody of the first’s clumsy effort, a third’s practice attempt, or a fourth’s strategic dissimulation. To apprehend what each boy is doing and understand how their deeds differ one from another, we must participate in their community of meaning. We must know the social significance of not only what is said, but also what is performed. There may very well be analogies across languages in how winks are used to communicate, owing perhaps to the physical dimension of the behaviour and the commonalities of our social nature, but that is far from saying that there is a universal language of winking. Winks can only be understood in relation to a socially coordinated way of life. So where does this leave us in describing cultural life? Clearly, any account of nonverbal behaviour as socially significant action, performed by reason of its conventional significance for the actor and the audience, requires interpretation no less than does a linguistic utterance. Flight into a behaviouristic observation language costs us the very explanandum of the inquiry, action as a culturally meaningful event. Thus, it offers no viable alternative to the limits of translation. Whether we are translating words or deeds, we must face up to the same indeterminacy described earlier and the same inevitable projection of meaning. The full burden of translation points to a cardinal difference between the natural and social sciences and returns us to the question of how well the former can serve as a model for the latter. The natural scientist must become versed in the conceptual system of her discipline to understand what is being manipulated and observed in experimental research. That is, she must learn the rules of classification and relation that define the physical reality under investigation. A stock broker dropped into a high-energy physics laboratory would not know what is happening and what the carefully observed numbers and images mean for what is at stake. To understand all that, he would need to become a competent member of the language community of physicists, which is to say that he would need to learn their view of physical reality, know which things count as the same and which as different, and why. In short, he would need to learn how to see things through their conceptual scheme. The social scientist bears the same burden. She too must learn the implicit rules of how reality is structured according to the language of her discipline. However, the reality at 232 Translation and Cultural Comparison a 2010 The Author Social and Personality Psychology Compass 4/4 (2010): 227–237, 10.1111/j.1751-9004.2009.00255.x Journal Compilation a 2010 Blackwell Publishing Ltd issue here is not a constellation of physical properties and their relations, but a world of meaning contained in socially significant activities and artefacts. The leptons, quarks and gauge bosons of the physicist are posited to exist independent of human concerns (the more radical interpretations of relative quantum mechanics notwithstanding). The same cannot be said of promises, shamans and acts of treason as studied by the social scientist. The latter only make sense within a system of institutionalized practices and their rulegoverned, internal meanings – internal because their social significance, or cultural content, is what they are. Their additional significance for the social scientist as examples of this or that theoretical abstraction presupposes a proper understanding of their indigenous meaning. One could not talk sensibly of, say, the place of promises in the unconscious mind of a people, or in some grand pancultural scheme, without first understanding what counts as a promise in the locality and way of life in which it occurs and how promises play out there (Winch, 1958). Otherwise, how would we know with any confidence when a promise is made? The social scientist must share enough of the natives’ point of view to recognize the elements of their cultural life. She cannot study what she cannot see. The natural scientist’s linguistic burden is one-sided, involving negotiation and conceptual alignment with fellow scientists on how to study an objective physical world. The social scientist, in contrast, must reconcile the professional-cum-commonsense view of world that structures her perceptions with the system of meaning of the society she studies in order to bring out the ‘ontologically subjective’ (Searle, 1995) facts of its cultural life. Clearly, we are talking about two very different kinds of science here. The social scientist, then, is playing a double language game. She is carving up social reality according to two potentially incompatible conceptual schemes, one indispensable for knowing what the cultural other is doing and the other indispensable for knowing what she is doing as a scientist. If the latter is limited to a strictly physical description of human action, then the ‘observation sentences’ permitted by this description will fail to fully ‘translate’ the intentional content of the action, as Hempel (2000) and the logical behaviourists came around to recognize. If, instead, the social scientist’s cultural explanations refer to the natives’ intentional content – their propositional attitudes of belief, desire, hope, etc. – then she is defining and explaining action in a very different way. She is now working within what Sellars (1956 ⁄1997) called ‘the logical space of reasons’. To whatever extent reasons for action can be considered causal, they are far from law-bound. A woman who leaves her husband because she believes he is possessed by an evil spirit is not, in acting upon that reason, illustrating a causal contingency that reduces in any clear-cut way to physical laws. The belief that one’s husband is possessed only provides a justification for leaving within the intersubjectively aligned logic of a cultural community. The intentional content of the mental state cannot be said to cause leaving according to the inexorable laws of nature (Davidson, 2001; Hornsby, 1997; McDowell, 1998). Moreover, the conceptual knowledge that underlies the reasoned act – knowing what husbands, spirits, possession, leaving one’s spouse and the dangers of staying are – itself rests upon shared conventions of linguistic use. Public meaning is prior to the private experience of knowing. As Sellars (1956 ⁄1997) observed in his influential attack on sense-data empiricism: Instead of coming to have a concept of something because we have noticed that sort of thing, to have the ability to notice a sort of thing is already to have the concept of that sort of thing, and cannot account for it (p. 87). The upshot for the present argument is this. The cross-cultural psychologist must begin by defining social phenomena according to the constitutive meanings of the people she Translation and Cultural Comparison 233 a 2010 The Author Social and Personality Psychology Compass 4/4 (2010): 227–237, 10.1111/j.1751-9004.2009.00255.x Journal Compilation a 2010 Blackwell Publishing Ltd studies. Comparative insight, theoretical generalization, and pancultural system-building can only apply to bona fide social phenomena that have been brought into clear view (Malcolm, 1968). In trying to understand what the cultural other is doing as reasonable, predictable action, the scientist is not engaged in the discovery of objective and universal cause–effect relations. Rather, she is tracing out the regularities that define the normative system of meaning and rationality in a particular cultural community. This is a largely descriptive endeavour, one that Evans-Pritchard (1954) came to see as an interpretive ‘art’ belonging more to the humanities than the sciences. This implies that the only way to remake cross-cultural psychology as a causal science would be to abandon the intentional idiom altogether in favour of a strictly materialist approach to mind and behaviour focused on neuroscience (P. M. Churchland, 1981; P. S. Churchland, 1986). Such ‘eliminativism,’ however, faces the objection that a purely physical explanation of a slice of behaviour cannot account for its identity as a conventionally defined, integrated action. However successful such an explanation might be at predicting the movement of bodies in space, it is difficult to see how the eliminativist’s nonintentional explanatory framework could deal adequately with events whose only coherence is their internal ‘folk’ significance. Eliminativism’s stark claim that beliefs, desires and other propositional attitudes are theoretical posits that simply do not exist is even more problematic (Stich, 1996). The dependence on local meaning in any convincing account of action brings us back around to the issue of translation across cultures. To the extent that a cross-cultural psychologist attempts to harmonize local cultural meaning with her own analytic scheme, whether through reduction, synonymy, homology or just plain old analogy, she is engaged in a form of translation. The impulse to move beyond the uncertainties and compromises of such translation is understandable. In fact, the desire for a transcendent language of objective explanation capable of encompassing one’s own culture and that of the other while not depending on the conventional meanings of either has been at the heart of cultural research from its beginnings. It appears as the quest for cultural ‘laws,’ the identification of which would grant cultural investigators a respectable seat at the high table of science. The formative years of modern social anthropology were defined by different versions of this quest. Boas (1940) believed the proper aim of anthropology to be the discovery of ‘laws governing the growth of culture’. Comparative and historical research was to be directed at ‘laying clear the complex relations of each individual culture’. Only through a correct understanding of the origins and dynamics of each cultural system could comparative insights be used to identify transcultural processes of continuity and change. Preferring system over history, Malinowski (1944) argued for the reality of key institutions as ‘natural isolates’ best defined by their functional role in fulfilling basic (biological) and derived (cultural) human needs. ‘Concrete analysis’ of this kind, he believed, would provide comparative research with its ‘only valid criteria of cultural identification’. Radcliffe-Brown (1958) had similar faith in the objectivity of the structural and organizational facts of institutionalized life, but located their proper functional analysis in how each contributed to the continuity of the ordered social life of a people. That significance, he argued, could not be found in the reflective understanding of the natives, but required the insights of scientists engaged in a sort of ‘comparative sociology’ of universal laws. Finally, Lévi-Strauss (1963) held that cultural ‘systems of behaviour represent the projection, on the level of conscious and socialized thought, of universal laws which regulate the unconscious activities of the mind’. These laws determine deep structures of meaning identifiable by the ‘semiological science’ of anthropology. These structures could in principle be articulated in a formal symbolism that transcended the explicit ‘social and mental and categories’ of both the natives’ and the scientists’ natural languages. 234 Translation and Cultural Comparison a 2010 The Author Social and Personality Psychology Compass 4/4 (2010): 227–237, 10.1111/j.1751-9004.2009.00255.x Journal Compilation a 2010 Blackwell Publishing Ltd As we near the end of the first decade of the 21st century, we are no closer to identifying any laws of the sort envisioned by the above cultural theorists. Why? Is the fundamental problem the complexity of causal determination that lies behind any cultural fact, as Boas believed? Or is the snag a deeper, conceptual one? My argument suggests the latter. Specifically, cultural or institutional facts, by their nature, cannot be given an external, universal description that does them justice because they depend on the rules of the local ‘language game’ by which they take on their meaning as institutional facts. And if we are tied in this way to conventional meaning in defining social phenomena, we cannot give a fully naturalized, noncultural account of those same phenomena. This means that cross-cultural psychology cannot successfully model itself on the natural sciences and its search for universal laws is a misbegotten aim. The account of translation offered here was intended to make this clear. There are synonymous behaviours across cultures, but no synonymous actions. Described as a physical movement, a wink is a wink is a wink. Described as action, however, a wink opens onto a whole way of life. The intentional content of the behaviour, which makes it mutually intelligible to actor and observer, coheres with the rest of the language within the ‘logical space of reasons’ of a particular cultural community. Meaning in this sense cannot be understood independent of its social context. Actions are inextricably tied to language, social roles, identities and relationships – a world of meaning. Peel these away and you have no action, perhaps not even an articulable behaviour (Goodman, 1978). When we describe our own cultural reality or that of others in an intentional idiom, our account will be underdetermined by what impinges on our senses (Quine, 1961). The social ‘facts’ we observe are an alloy of stimulation and unconscious conceptual structuring. Or, as Goodman (1978) aptly put it, ‘some of the felt stubbornness of fact is the grip of habit’. Ethnography’s systematic documenting of facts, then, can itself be considered covert translation, a projection of our scheme of objects, kinds and events onto foreign symbolic terrains. In arguing against the idea of a private language, Wittgenstein (1958) offered a memorable illustration. In it, he asks us to imagine a group of people carrying around boxes with something inside. Everyone can see only what lies inside his or her own box. All refer to what they see as a ‘beetle’. The word ‘beetle,’ insofar as it functions as a meaningful element of discourse, must have a common meaning for all those who use it. Semantic differences will ‘cancel out’ over time. It does not matter that people see different things altogether in their boxes, that what they see changes over time, or even that some see nothing at all. ‘Beetle’ will nonetheless have a public, shared meaning defined by its use in the language game of the community. The claim that understanding people across cultures is essentially no different than understanding others in one’s own culture falls on the sword of Wittgenstein’s argument. Cultural and linguistic competence requires the convergence of perception and reasoning with the normative conceptual scheme and logic of a community. We understand the actions of others, and our own for that matter, according to this normative structure. What separates our private experience from that of others is irrelevant for communication and social coordination. The unavoidable projection of that normative structure into another culture to give what one sees there a definite form and sense is another matter entirely. It is at best an assimilative translation. In other words, there is a world of difference between translating across cultures and interpreting the actions of others who share our language and practices. In the former case, there is no conventional or institutional basis for commonality of meaning across individuals. In the latter, the public nature of language and the standardization of meaning through long-term social use guarantees mutuality of understanding. Idiosyncrasies of Translation and Cultural Comparison 235 a 2010 The Author Social and Personality Psychology Compass 4/4 (2010): 227–237, 10.1111/j.1751-9004.2009.00255.x Journal Compilation a 2010 Blackwell Publishing Ltd personal meaning remain in play between members of any cultural community, but they are pushed into the background by the social demands of speech and action. So what does all this fuss over the limits of translation allow us to conclude about cross-cultural comparison? First of all, it should not be doubted that what people say and do in another culture is often analogous with our own speech and actions. Cultural and linguistic variation is constrained by the constitutional imperatives of our species and the lawful nature of the physical world we inhabit. If this were not so, we could not have achieved the considerable success we have in predicting behaviour ‘away from home’. As it turns out, systematic analogy goes a long way towards allowing us to know the cultural other. Even so, unless we restrict ourselves to an overly barren observation language, we cannot validly claim that we are comparing things across cultures that have the same meaning. No degree of conceptual abstraction can save us here, as only particular actions can be observed, measured, contrasted and compared on the ground. Similarly, greater psychometric precision and statistical sophistication will not magically transform our analogies into synonymies. At its worst, methodological virtuosity gives a scientistic veneer to our often compromised efforts to infer correspondences and place them on a common metric. We too often forget that the subjective meaning of a behaviour, whether the circling of a number on a Likert scale, the disclosure of personal details to an interviewer, or the solving of a set problem, is what defines that behaviour as a full-bodied action. A case can certainly be made for descriptively or functionally analogous actions across cultures, but in no case does this mean that we are measuring the same underlying thing, because there is no such thing. If our translation is inherently projective, uncertain and possibly inconsistent with rival but equally valid conceptual schemes, so too is the significance we ascribe to any differences we measure across cultures. This limitation recommends a good degree of scientific modesty and the recognition that our conclusions about others may reflect the grooves of our own conceptual categories as much as theirs. Whatever our scientific pretensions as cross-cultural psychologists, we are not the fledgling architects of a Comtean ‘social physics’. According to the view outlined here, a lawful natural science of the social is hardly conceivable. It is more accurate to say that we are engaged in a form of hermeneutics in Rorty’s (1979) sense of the word: an interpretive struggle to make the incommensurable and unfamiliar partly commensurable and a bit more familiar. Cultural comparison as hermeneutics demands the humility of knowing our limits.