“There is no such thing as philosophy-free science; there is only science whose philosophical baggage is taken on board without examination.” (Dennett 1995)
To those who would dismiss anthropology’s engagement with questions of culture: read Wittgenstein’s Philosophical Investigations and then try arguing that culture and language play no role in the construction of models and representations of the world. (A poll of North American philosophers recognized Investigations as the single most important philosophical work of the 20th century (Lackey 1999). Kuhn’s Structure, in the top 10, also centralizes the role of culture and language in scientific models and representations.)
An inherent feature of the practice of observation in empirical sciences is its dependence upon our perception and how that perception is ordered into description (necessarily via language) and communicated to others. This obviously makes the practice of empirical science heavily dependent upon language and pre-existing concepts.
Preceding Wittgenstein, Wilhelm von Humboldt was one of the earliest documented figures in the European intellectual tradition to express the idea that there is a close relationship between social groups, languages, and cultures and their representations of reality (both perceptual and conceptual) via his concept of Weltansicht (roughly translatable as ‘linguistic worldview’). This condition has implications for all scientific and philosophical enquiry (Wittgenstein, Kuhn). Further, the very choice of what is considered a valuable field of enquiry to engage upon will reflect the culture, representations and values of the society that makes that choice. Is it any surprise that 4th century BCE Indian investigations of economics (as ‘the science of wealth’) were expounded upon in a work that also discussed ‘the science of government’ and statecraft more generally (Trautmann 2012)?
In their attempts to represent or model processes in the world, open-ended enquiries (in contrast to dogmas) must continuously struggle to disentangle their observations and proposed models from their inherited linguistic, perceptual, conceptual and values-based biases. This is necessary as part of the broader bootstrapping that science must carry out, via responding to methodological criticism, new observations and counter-evidence, continuously re-calibrating methods, models and frames of reference. As Dennett puts it: “There is no such thing as philosophy-free science; there is only science whose philosophical baggage is taken on board without examination.” (Dennett 1995)
Though a constraint for all scientific enquiry, the baggage of inherited language and representations has particularly significant implications in the behavioural and social sciences. In the physical sciences, those subjects of enquiry that are rarely encountered or observed in the common daily use of language (and are thus often counterintuitive) – e.g. the relative movements of astronomical bodies, the chemical or electromagnetic properties of substances, the structures of a cell – may require the coining of novel terms and models to conceptualize their subject matter. These may or may not overlap greatly with preconceived ideas prevalent in the culture of the observer.
Early European astronomers working within a geocentric model were at least partly constrained by established cultural representations (e.g. from religious heritage), and there was great philosophical and cultural conflict over the emergence of heliocentric models. On the other hand, early observations and theories about electromagnetism were not so greatly constrained by existing conceptions of these phenomena from either religious representations or other aspects of the linguistic and cultural background (of course the corollary of this is that the absence of pre-existing conceptions inevitably makes entirely new domains of phenomena correspondingly difficult to perceive or investigate ex nihilo). Geertz (1976), following Kohut, calls phenomena that are outside of our common experience ‘experience-distant’ concepts.
Observations of behaviour on the other hand, especially those made of we humans ourselves (individually and collectively), are more readily observed through and described by the kinds of everyday existing common language, concepts and behavioral models (folk-psychological models if you will) that accompany our social interactions and evaluations (what Geertz calls ‘experience-near’ concepts). In the study of social and behavioural phenomena then, our models and working theories are particularly encumbered by our linguistic and cultural milieu; they are consequently particularly vulnerable to constraint and bias via the particular culture that encompasses them; i.e. vulnerable to cultural bias.
It is an open question as to whether fully escaping these limitations is possible. What we can at least conclude is that practitioners in the behavioral and social sciences (economics, psychology, anthropology, sociology and more) – unless taking vigilant steps to mitigate these inherent sources of cultural bias – are very unlikely to be able to make observations, descriptions, theories and frameworks generally applicable to all human cultures.
An example of a discipline of social and behavioural study that has become acutely aware of this fundamental problem is anthropology. This is partly because the field of enquiry entails that anthropologists deliberately attempt to explore the linguistic, conceptual, behavioural and social variations between cultural groups, and also face the concomitant problem of considering how to ‘translate’ these between languages and cultures. Ethnographic methods of fieldwork, pioneered by anthropologists such Boas and Malinowski are an example of an attempt to mitigate cultural bias in the observer/scientist.
Given that Boas’ cautions were first outlined over 100 years ago, one might expect that most behavioural and social sciences are by now well aware of these pitfalls and at least attempt to control or compensate for them. However, although there has been some progress in these areas (ethnographic and participant-observer methods, reflexivity) there is still much work to be done. In psychology for example, significant debate only arose around the need to address these issues in the 1980s and 1990s (e.g. Katz 1985) and methodologies to compensate are still being debated. Mainstream economics with its inherent analysis of various kinds of value has barely begun to address these issues.
Let us now turn to a concrete example which both illustrates the cultural bias problem, and illustrates how incorporating criticism is integral to bootstrapping in science. Schneider’s critical work in the late 1960s through mid 1980s of the anthropological use of ‘kinship’ concepts demonstrated that, even in this comparatively culture-aware field of enquiry, cultural bias existed long after Boas’ work. To highlight the bias, Schneider noted a distinction between the traditional anthropological conception of kinship relationships as intrinsically ‘given’ and inalienable (‘from birth’), and an alternative view of kinship relationships as created, constituted and maintained by a process of interaction, or ‘doing’. With characteristic directness, Schneider included his own earlier work in his critique. He revisited his own analysis of the citamangen / fak relationship in Yap society, which he had formerly treated as a ‘father / son’ relationship, to contrast these two alternative conceptions;
“The crucial point is this: in the relationship between citamangen and fak the stress in the definition of the relationship is more on doing than on being. That is, it is more what the citamangen does for fak and what fak does for citamangen that makes or constitutes the relationship. This is demonstrated, first, in the ability to terminate absolutely the relationship where there is a failure in the doing, when the fak fails to do what he is supposed to do; and second, in the reversal of terms so that the old, dependent man becomes fak, to the young man, tam.
The European and the anthropological notion of consanguinity, of blood relationship and descent, rest on precisely the opposite kind of value. It rests more on the state of being, on the sharing of certain inherent and therefore inalienable attributes, on the biogenetic relationship which is represented by one or another variant of the symbol of “blood” (consanguinity), or on “birth,” on qualities rather than on performance. We have tried to impose this definition of a kind of relation on all peoples, insisting that kinship consists in relations of consanguinity and that kinship as consanguinity is a universal condition.” (Schneider 1984)
As well as traditional anthropological accounts, Schneider also included sociobiology in his critique:
“Finally, the most recent, explicit, detailed, and developed commitment to the premise that Blood Is Thicker Than Water is made by the socio-biologists. They do this in numerous publications that need not be quoted here since they are so well known.” (Schneider 1984)
To their credit, anthropologists responded quickly by engaging in debate about Schneider’s critique, and in many cases, adjusting their methodology and models: observing the distinction between ‘being’/blood vs. ‘doing’/performance has been one of the central orientations of kinship studies ever since. One might think that the sociobiologists too might have usefully responded to this critique of their cultural bias by showing that they had learned the lesson and understood the problems that the work of Humboldt, Boas, Wittgenstein and many others had suggested, and introduced some reflexivity and caution into their analyses and models.
Instead, human sociobiologists responded to this and related critiques (e.g. Sahlins 1976, Kitcher 1985) by retrenching, rebranding (they now go by the name of ‘Evolutionary Psychologists’) and derogating all perspectives that attempt to account for culture (which they refer to as the ‘Standard Social Science Model’ e.g. Pinker 1999). They continue to boldly advance particular models of ‘human nature’ (sometimes arguing for these via experimenting on small groups of US college students) often in the face of contrary evidence, and essentially closing their ears to the rest of science and philosophy. Unsurprisingly, the observations made and models advanced often bear striking resemblance to particular behaviours, practices, and folk-theories common in Euro-American cultures.
That they have not taken Schneider’s critique seriously can be clearly seen in how they have continued to conceptualize social relationships, kinship and cooperation. Both earlier human sociobiologists and current Evolutionary Psychologists have typically claimed that biological theories about the evolution of social behaviour can be correctly interpreted to make predictions about how such behaviour is mediated (in both humans and other animals). For example, a recent experiment conducted on humans by Robin Dunbar and colleagues (Madsen et al. 2007) was, as they understood it, designed “to test the prediction that altruistic behaviour is mediated by Hamilton’s rule” and more specifically “If participants follow Hamilton’s rule, investment (time for which the position was held) should increase with the recipient’s relatedness to the participant. In effect, we tested whether investment flows differentially down channels of relatedness.” From their results, they concluded that “human altruistic behaviour is mediated by Hamilton’s rule… humans behave in such a way as to maximize inclusive fitness: they are more willing to benefit closer relatives than more distantly related individuals.” (Madsen et al. 2007).
As well as ignoring the counter-evidence advanced by Schneider, Sahlins and others, the fundamental problem with this position is that the biological theory it appeals to does not make the prediction that organisms behave in such a way as to maximize inclusive fitness. The position thus achieves the rare feat of being simultaneously; empirically disproven; logically fallacious; and still frequently employed.
The evolutionary biology theory they refer to (inclusive fitness theory), necessarily takes the form of an ‘ultimate’ explanation; it is a model specifying a covariance criterion relevant to the evolution of certain social traits. Hamilton’s rule (C<Br) is in fact a short form of this criterion. What is misunderstood by claims such as those made by Dunbar and colleagues is that evolutionary explanations and proximate explanations are distinct in biological analysis and should not be conflated (Tinbergen 1963). Inclusive fitness theory is explicitly an evolutionary explanation, and not a proximate one. Confusing these distinct forms of explanation amounts to claiming that behaviour is goal-driven to achieve certain outcomes; it is fallacious and a type of reductionism. Even a cursory review of findings about the proximate expression of social behaviours in mammals and primates should make it clear that these are not contingent upon genetic relatedness per se, but are instead mediated by context-based and proximity-based cues. Social traits that have taken this proximate form in evolutionarily typical environments are nevertheless compatible with the covariance criterion suggested by inclusive fitness theory.
A more careful interpretation of the biological theory along these lines (Holland 2012) reveals that there is in fact compatibility between biological and social science disciplines regarding social bonding and kinship; between the correctly interpreted biological position; the position in attachment psychology (which has recently begun incorporating cultural diversity); and the themes that emerge from many ethnographic accounts conducted by cultural anthropologists. Why have Evolutionary Psychologists not taken on board the kinds of critiques that Schneider (and others) have outlined, and instead stuck to their previous unsupported and incorrect models of social behaviour? It seems likely that these incorrect models remain prevalent precisely because they reflect the cultural biases that Schneider outlined.
A constructive treatment of the biology that attempts to minimize cultural bias thus actually reinforces the scientific value of the ethnographic accounts that cultural anthropologists have carefully produced over several decades. Constructing an essentialised model of ‘human nature’ from narrow cultural particulars (Euro-American or otherwise) does not constitute science; it is closer to cultural colonialism. It bears a similar relationship to evolutionary biology as social Darwinism did in the 19th century. In any analysis intended to shed light on proposed universals of the human condition, reflexivity and a culture-aware approach are essential.
As further examples of the merits of culture-aware approaches gradually find their way into social and behavioural sciences, we might hope that the valuable insights of von Humboldt, Boas, Wittgenstein and others begin to be taken more seriously.
(an abridged version of this post is on ethnography.com)
Dennett, 1995 Darwin’s Dangerous Idea
Wittgenstein, 1953 Philosophical Investigations
Lackey, 1999 What are the Modern Classics?
Kuhn, 1962 The Structure of Scientific Revolutions
von Humboldt, 1836 On the Diversity of Human Language
Trautmann, 2012 Arthashastra: The Science of Wealth
Geertz, 1976 From the Native’s point of view
Boas, 1920 The Methods of Ethnology
Malinowski, 1922 Argonauts of the Western Pacific
Katz, 1985 The sociopolitical nature of counseling
Schneider, 1984 A critique of the study of kinship
Sahlins, 1976 The use and abuse of biology
Kitcher, 1985 Vaulting ambition
Pinker, 1999 The blank slate
Madsen et al., 2007 Kinship and altruism a cross-cultural experimental study
Hamilton, 1964 The genetical evolution of social behaviour
Tinbergen, 1963 On Aims and Methods in Ethology
Holland, 2012 Social Bonding and Nurture Kinship