[433]

Michel Verdon

Anthropologue, département d’anthropologie,
Université de Montréal

“Franz Boas :
cultural history for the present,
or obsolete natural history ?”

In Journal of the Royal Anthropological Institute (N.S.) 13, 433-451, 2007. Royal Anthropological Institute.

Abstract [433]

Introduction [433]

The fight against evolutionism and some of its consequences [433]

Downplaying invention, and the resulting atomism [433]

Boas's dissociative method [436]

Cultural hazards and the inspiration from natural history [436]

A natural history of culture and its consequences [438]

Cultures' lack of historicity [438]

Cultures' insubstantial reality [439]

Counter-evidence of an atomistic and natural-historical Boas ? [440]

The indispensable but taxonomic 'genius of a people', and 'quantified' style [440]

Boas and texts [442]

Boas and secondary explanations [443]

The theory's possible intellectual origin [443]

Do secondary explanations have any explanatory value ? [444]

Boas and psychology [445]

Conclusion [447]

Notes [448]

References [449]

Résumé [451]

Abstract

Recently, some neo-Boasian anthropologists have portrayed Boas as an anthropologist with a deep sense of history, of the individual, and of agency. Focusing on Boas's ethnographic practice rather than his theoretical and programmatic statements, I first find an 'atomistic' (opposite of holistic) ethnographer, and a deep convergence between this atomism and Linnaean-type natural history. In Foucault and Jacob's interpretation of natural history, this means studying socio-cultural phenomena through their external manifestations, and removes historicity, and even individual cultures, from Boas's ethnography. Reviewing possible counter-evidence from the holistic Boas (his work on style, meaning, the 'genius of a people,' texts, secondary explanations, and psychology), I retrieve the same natural historian, and the same atomism. All these facets of his practice thus appear as surface manifestations of this underlying episteme, which provides a single interpretative framework making it possible to integrate most of his ethnographic work. Overall, this worldview leaves little, or no, room for individuals and their agency.

Introduction

In recent years, Boas has resurfaced as a source of inspiration for a number of anthropologists. [1] In the process some are presenting him as an anthropologist with a deep sense of history (Bunzl 2004 ; Lewis 2001 ; Orta 2004), of the individual (Lewis 2001 ; Orta 2004), and of agency (Lewis 2001), even anticipating some aspects of postmodernism (Lewis 2001). [2] Is this a credible reading of Boas, or a re-invention to suit new clothes ? I lean towards the latter view, which I will here attempt to substantiate. I do not wish to interpret Boas in the manner of historians, however - many historians have admirably dealt with the German or scientific roots of Boas's main concepts - but in an 'internalist' way, by studying the various aspects of his work to uncover their articulation.

At a first glance, it may seem unproblematic to retrieve a historical Boas, but less so to find a Boas imbued with a deep sense of the individual and of agency. This interpretation of Boas is puzzling if one omits to mention, as did the authors, that it mostly emanates from his books meant for the general public (The mind of primitive man and Anthropology and modern life), as well as his more abstract, and even programmatic, statements. Interestingly enough, by focusing more on Boas's ethnographic practice while obviously taking some of his theoretical statements into account, one gets a very different view.

[434]

In the first part of this article, it will be shown that Boas's critique of evolutionism, possibly more than any other factor, led to his elementaristic (Stocking 1982b [1968] : 207) or atomistic view of culture (by 'atomistic', I mean the opposite of holistic). Basing myself on Foucault and Jacob's interpretation of eighteenth-century natural history as an exercise in the classification of phenomena on the basis of their external attributes (Foucault 1966 ; Jacob 1970), I then argue that Boas's atomistic view of culture neatly converged and dovetailed with a natural-historical perspective in his ethnographic practice ; both the atomism and the natural history worldview apprehend socio-cultural phenomena through their outward manifestations.

The first part of the article deals exclusively with the atomistic Boas, while the second looks at the holistic Boas who went 'behind appearances' by recognizing the role that the 'genius of a people' plays in culture formation ; the Boas who was concerned with style, art, and meaning, interested in texts as an entry into the thought-world of the Indians, and produced a psychological anthropology. We would expect this Boas to escape natural history.

At this juncture, the Foucault-Jacob interpretation of natural history shows its full heuristic potential. I examine every one of these holistic facets and, in every case, find Boas operating like a Linnaean natural historian dealing with objectified cultural productions. Through this demonstration some re-interpretations may be original (possibly those regarding meaning, or the Volksgeist) ; others merely summarize well-known views (the part on texts, for example). From my perspective, however, novel or not, these interpretations are merely 'punctual' if not understood within the article's main innovation : its overarching architecture.

Indeed, beyond individual reconsiderations, this natural-historical interpretation makes it possible to articulate hitherto disconnected dimensions of Boas's anthropology, and to integrate his ethnographic practice into a single interpretative framework. In brief, most of his ethnology's various facets now appear as surface manifestations of this underlying, organizing episteme. Obvious conclusions follow : at the level of his ethnographic practice, I fail to find any holism, and argue that his atomism-CMW-natural history led him away from a true sense of history, of individuals and agency.

The fight against evolutionism
and some of its consequences

Downplaying invention, and the resulting atomism

As early as 1886, Boas started work on the Pacific Northwest Coast and sought to reconstruct the history of its cultures (1974b [1888]). Trained until then in geography, after an early start in physics, he used the techniques of German historical geography for his Northwest Coast project, namely reconstructing the history of cultures without written or archaeological documents from the geographical distribution of cultural traits, tracing 'historical transmission and historical contacts by studies of geographical distribution' (1974g [1906] : 187 ; see also Bunzl 1996 : 53 ; Kluckhohn & Prufer 1959).

Historians have emphasized the influence of historical geography, together with Boas's involvement in museography and 'salvage anthropology', among others, to account for a key feature of his ethnographic practice, namely distributional studies based on a fragmented, or atomized, view of culture that dominated his ethnographic [435] practice well into the late 1910s, according to Benedict (1943 : 28-9), if not completely (Darnell 1997 : 47 ; Voget 1975 : 318 ; White 1963 : 61).

The classical explanations of Boas's atomism none the less leave some questions unanswered. Boas's early, and intense, interest in myths and folklore (evident already in 1886), for instance, seems at odds with this strong atomism. On the one hand, German historical geographers reconstructing history through geographical distributions generally neglected myths and folklore ; on the other, the German anthropologists seriously interested in myths and folklore, namely the 'folk psychologists' (Volkerpsychologe), who also influenced Boas, were not seriously concerned with historical reconstructions. In my view, Boas's choice of myths at the very outset of his Northwest Coast research, as well as his early generalizations about their nature, hint at a very specific agenda. From 1886 to the end of his life, the obvious one seems to have been his increasingly intense opposition to the evolutionists' method and theses (Voget 1975 : 323) and, beyond, the racial discrimination they implied. I thus submit that Boas's choice of myths as his starting-point was methodologically and polemically inspired.

Indeed, Tylor and other evolutionists had already stated that complex socio-cultural phenomena, such as myths and folk-tales, could hardly be explained by independent inventions if found in different areas. Boas concurred, writing about myths in neighbouring populations : '[I]t seems quite incredible that the same complex product should originate twice in a limited territory' (1974c [1899] : 97 ; see also Harris 1968 : 260 ; Voget 1975 : 324, 331). If so, he remarked, they must result from dissemination.

Boas seems to have assumed that most evolutionists explained similarities between 'ethnological phenomena' [3] in terms of independent inventions ; [4] as a consequence, the prominence he gave to diffusion became critical in his opposition to evolutionary theories, and myths and folklore thus provided ideal ammunition against the evolutionist thesis of independent inventions. It also powerfully reinforced his atomism, his 'fragmenting view of culture' (Stocking 1982k [1968] : 213).

Boas further generalized on the very nature of myths. Even in contiguous populations sharing essentially the same lore, he observed, myths showed variations, losing and gaining elements as they 'travelled' ; he then concluded that, although complex, myths proved not to be whole 'organic' phenomena : '[T]he mythologies of the various tribes as we find them now are not organic growths, but have gradually developed and obtained their present form by accretion of foreign material' (1974e [1899] : 96), a conclusion he had already reached by 1891 (1940b [1891]). Why conclude so early that myths are not organic growths ? Presumably because 'organic growths' would have appeared locally through inventions ; as 'accreted phenomena', however, they must have developed through diffusion.

These conclusions on myths led Boas to develop one of his most powerful techniques in his critique of evolutionism. What he achieved so convincingly with myths he repeatedly successfully with equally complex phenomena, such as ceremonials, art, literary productions, and totemism, among others. In every case he showed that the most complex phenomena could be broken down into elements whose geographic distributions did not overlap. From this lack of incidence, or 'dissociation' of the various traits' areas of distribution, he inferred that they could not be organically integrated, and were therefore accreted. I call this method 'dissociative' ; it successfully led to downplaying invention and privileging diffusion. Spier (1931) expertly illustrated the method as it applies to myths, and an article by Boas (1940^ [1903]) on art illustrates it equally well.

[436]

Boas's dissociative method

In 'The decorative art of the North American Indians' (1940d [1903]), Boas uses this technique of dissociation most efficiently. Previous authors (evolutionists) assumed that primitive art evolved from realistic representations towards more conventionalized, geometric forms ; the realistic representations would have therefore remained as the geometric designs' ultimate explanations (or meaning). Others held that technique and material influenced the evolution of design. Overall, previous writers on primitive art had thus tended to associate artistic phenomena in some causal direction, further presupposing this evolution to have occurred independently in the various tribes, through independent inventions. Boas is not convinced.

He first dissociates 'the decorative style applied in ceremonial objects and that employed in articles of every-day use' (1940d [1903] : 547). The former are 'much more realistic than that of ordinary objects', so that 'the reason for the conventionalization of motives can not be solely a technical one' (1940d [1903] : 547). This also raises serious doubts about an alleged evolution from realistic to geometric forms : 'If the style of art were entirely indigenous in a given tribe [read : invented independently], and developed either from conventionalization of realistic designs or from the elaboration of technical motives, we should expect to find a different style and different motives in each tribe' (1940d [1903] : 553-554). In other words, if the various aspects of art - technique, design, and meaning - were associated, we should find systematic cultural variations, pointing to separate inventions. The fact is, we do not.

Using the art of some North American Indians, Boas proves his case by studying design. He focuses on a moccasin design 'so complex that evidently it must have had a single origin' (1940d [1903] : 555). Yet, when investigating the interpretations the Indians give of their designs, he discovers radically different explanations (symbolism), and raises the question : if complex designs cannot have been invented independently, have the simpler ones been ? Here the 'simpler designs' denote the 'component elements' of more complex ones (1940d [1903] : 556) ; Boas reviews their 'indigenous' explanations and finds the same lack of association between design and meaning.

Having found 'styles of interpretations' in some tribes, Boas further shows that their distribution does not coincide with any style of art, and he finally proves 'that the distribution of technique does not agree with that of motives' (1940d [1903] : 560). The fact that art's components -technique, design, interpretation, and style of interpretation - are all dissociated leads him to restate more firmly his two fundamental theses. First, a complex phenomenon such as the art of a tribe is not an organic growth but an accreted phenomenon, the result of the dissemination of techniques, styles, and interpretations. Second, since no particular interpretation is associated to a particular style, it follows that interpretations 'must have developed after the invention or introduction of the design ; that the design is primary, the idea secondary, and that the idea has nothing to do with the historical development of the design itself (1940d [1903] : 555) ; this is a variant of Boas's theory of secondary explanation, which he repeats and generalizes in the last paragraph (1940d [1903] : 563 ; more about secondary explanations below). Thus, in a single dissociative analysis, Boas succeeds in exploding evolutionary sequences rooted in false associations, proving that complex phenomena such as art are accreted.

Cultural hazards and the inspiration from natural history

As dissociation pointed to fragmentation, it reinforced Boas's atomism, which, in turn, produced what Orta (2004) called an 'entropic' vision of culture - culture perceived as [437] a random assemblage of cultural elements. If elements travel, and their geographical distributions rarely coincide, they are therefore unconnected, greatly autonomous, moving about in unpredictable directions, re-assembling with other traits coming from different cultural horizons in a given tribe ; their combination is unsystematic and makes up cultural hazards (theoretically integrated by the genius of a people, as we shall see). Writing on the essential elements of Boas's method and using Boas's work on Tsimshian mythology (1916) as exemplar, Leslie Spier reflected on the role that Boas's theory of secondary explanations played in Boas's anthropology, and insightfully concluded : [5]

This principle of secondary reinterpretation is one of far-reaching implication. By it one may understand why the majority of anthropologists have come to view every culture as a congeries of disconnected traits, associated only by reason of a series of historic accidents, the elements being functionally unrelated, but believed to be related by the bearers of that culture because of the interpretation the traits have undergone. (Spier 1931 : 455, italics added)

This is but one part of the equation. This atomism also went hand in hand with a natural-historical view of the cultural world. Many commentators, and most forcefully Lesser and Kroeber, have emphasized Boas's abiding interest in natural history, and how it influenced his anthropology (Boas 1940a [1887] ; Harris 1968 : 286-287 ; Kroeber 1959 ; Lesser 1981 ; see also Kluckhohn & Prufer 1959 : 39 ; Smith 1959 ; Stocking 1992 :123 ; Voget 1975 : 323, 328). I claim something much more fundamental. The natural history I am referring to is no mere survival from Boas's teens and early university years, a simple penchant for collections and 'the full history of the single phenomenon' that would have coloured his anthropology. I see it not as a direct historical influence, but as a convergence : I submit that Boas understood the cultural world in the manner that Linnaean-type eighteenth-century natural historians apprehended the natural world. I contend that it is a foundational, underlying, and pervasive Boasian episteme, one that fits with his atomistic view of culture ; in fact, I see them as two faces of the same coin.

Foucault and Jacob provided the inspiration for the link between atomism and natural history, with their stimulating interpretations of eighteenth-century natural history (Foucault 1966 ; Jacob 1970) ; to both, biology only surfaced in the nineteenth century with the concept of organization, which implied relating causally the various internal organs of animals. They contrast biology, with its key concept of organization (Cuvier's "correlation of [internal] parts'), to the natural history of the preceding century. According to them, this natural history completely lacked any notion of life and organization, the two epistemologically linked ; eighteenth-century natural historians perceived only living organisms grouped in species to be classified on the basis of their external attributes, or visible structure (where 'structure' does not imply organization but the mere number, shape and relative arrangement of visible parts ; Foucault 1966 : 147 ; see also Foucault 1966 : 144, 149, 156 ; Jacob 1970 : 37, 42, 54 ff).

The convergences are striking. Despite some theoretical claims to the contrary, Boas's atomism and distributional studies presupposed that socio-cultural phenomena's outward manifestations be compared in order to establish similarities or differences. Furthermore, Linnaean-type natural history lacked any notion of organization ; similarly, an entropic view of culture spells the opposite of organization, integration, or system ; an organized, or systemic, entity is one whose parts are correlated, one of functional interdependence, something foreign to Boas's dissociative anthropology, as [438] Spier noted. To that extent Boas's atomism and eighteenth-century natural history are part and parcel of the same episteme. I will thus argue that the main, if not all, the characteristics of Boas's anthropological practice, be they atomistic or allegedly holistic, flowed to a great extent from this epistemic convergence. [6]

A natural history of culture
and its consequences

Cultures' lack of historicity

Boas's understanding of history can be broached through a brief sketch of one of his articles, namely 'The mythology of the Bella Coola Indians' (1974d [1898]), in which he tries to explain the specificity of Bella Coola mythology (mostly its strong local identity, and its systematic nature). He plots some of its various external traits (village endogamy, type of settlement, importance of crests, and so on) against those of its southern and northern neighbours, namely the Coast Salish and Bella Bella, the latter a Kwakiutl group. He assumes the Coast Salish to have a 'simpler form' of social organization than the Bella Bella since they lacked clans criss-crossing villages. After examining the traits' distribution he infers that the Bella Coola were originally of Coast Salish extraction, and moved northwards to the Bella Coola River. He further notices that the Bella Coola also shared a number of traits with their northern Bella Bella neighbours, but he does not discuss the various possible origins of those traits. For instance, he does not raise the possibility that those traits, or some of them at least, might have originated among the Bella Coola and been borrowed by the Bella Bella ; he actually assumes the diffusion to have taken place in one direction only, from the Bella Bella to the Bella Coola. Why ? It can only be surmised, but the available evidence suggests that he must have reached this conclusion from the fact that he considered the Bella Bella to be socially more complex.

Indeed, the mechanisms Boas adduces to account for the diffusion suggest such an interpretation : he explains the diffusion from the Bella Bella to the Bella Coola by the fact that the latter would have felt an inferiority vis-à-vis the Bella Bella because they lacked crests. Wishing to imitate their northern neighbours, they obtained those crests by marrying their women (he makes no mention of women marrying in the other direction). Intermarriage aside, to Boas similarities across cultures, and even what he calls 'parallelism of distribution' (the fact that the distribution of two or more traits sometimes coincide), raise psychological connections, which he answers mostly in terms of imitation.

As far as I can tell, this argument is representative of Boas's various historical reconstructions, all of which are based on distributional studies (see also White 1963). He mapped various phenomena, and the historical links he inferred between cultural elements in neighbouring cultures normally indicated diffusion. As diffusion without in-depth explanations of the reasons why people borrow, save mostly by imitation, one could argue that the 'history' of Boas's 'culture history' program was more or less synonymous with 'movement in space'. Since movement in space takes time, it translates as history. [7] Paradoxically enough, however, this spatialization of time seems to rule out a true sense of history. [8]

This is astonishing on the part of an anthropologist who clearly and explicitly proclaimed his historicism from the beginning to the end of his career (1940c [1896] : 276 ; see also 1940l [1932] : 250 ; 1940m [1936] : 305), and who has been hailed as fundamentally historicist by his own students and most commentators. But, as Spier noted,

[439]

It should be clear that by 'historical' Boas meant only that each cultural trait and configuration must have had a specific antecedent form. This did not involve the need to provide a sweeping picture-in-time ... It sufficed for his purpose to envisage a 'before and after' picture at a particular place and time (1959 : 147 ; see also Voget 1975 : 328).

Boas himself wrote that 'it will be seen that anthropology differs from history, and resembles the natural sciences in its endeavour to disregard the subjective values of historical happenings ; that it tries to consider them objectively, simply as a sequence of events, regardless of their influence upon the course of our civilization' (1974h [1907] : 270).

Is this history ? If we go by Boas's statement, hardly so, but we must acknowledge that Boas was reconstructing the history of societies without written documents, before the advent of archaeological evidence. Even the German historical geographers working without written documents, however, and who most influenced Boas - Ritter and Fischer in particular, according to Kluckhohn and Prufer (1959 : 13-19) - dealt with people through questions of 'migrations, ... conquests, colonization, ... traffic, transportation' (Ritter, in Kluckhohn & Prufer 1959 :14), and of various causal relationships (Fischer's study on the date-palm, in Kluckhohn & Prufer 1959 :18). Boas did not.

If we ignored Ritter and Fischer, could Boas's many historical reconstructions be deemed 'historical' ? Only in a superficial way, I believe. However idiosyncratic an author's understanding of history can be, however 'objective' he or she may wish to be, can we call 'historical' works without historical actors, historical processes, and even minimal causal sequences ? At the end of the road, this is indeed the type of history Boas writes, namely historical 'events' mostly assimilated to borrowings, without strategic, calculating individuals, the 'causes' of which are psychological (except intermarriage) and mostly consist of a psychological predisposition to imitation (taken over from de Tarde's collective psychology). Even at the time Boas was writing, and even without any form of written or archaeological documents, it is difficult to accept that there was no alternative, save the evolutionists', to Boas's history, that it was then impossible to conjecture more convincing historical processes, some minimal sets of causal sequences, and various individual motivations beyond imitation. For various reasons he might not have wished to do so, but it was not an unattainable aim.

A faithful Boasian, Spier was also insightful about Boas's practice. He saw that Boas's 'history' had to do with the 'before-and-after' classification of objective, quantifiable facts, of drawing inferences of contact from static distributions in space. In a word, even to Spier and other contemporaries, Boas's historicism lacked any true depth (see Kroeber 1946 : 8 ; 1952 [1936] ; Radin 1933 : 17).

Cultures' insubstantial reality

In theory, cultural elements assembled in a given 'social place' (among a tribe or a 'people') delineate an individual culture ; Boas recognized their existence, and did write of individual cultures. He was not interested in the mere inventory of a tribe's cultural elements, however, but in reconstructing its history through distributional studies. Distributional studies were also predicated on the existence of individual cultures, but, paradoxically, individual cultures in Boas's ethnography mostly appeared as a means to an end that almost excluded their study, or history. In reality, Boas's geographical distributions charted out traits between neighbouring cultures forming 'cultural areas', or 'regional cultures', as Stocking calls them (1982a [1968] : 157), and he did so on the basis of the traits' external attributes (including myths, beliefs, or ceremonials broken [440] down into fragments). He focused primarily, if not exclusively, on these regional cultures, which he and his students regarded as the set of cultures historically the most closely related.

In his ethnographic practice, except when dealing with texts and some aspects of social organization, individual cultures were thus dissolved into regional ones ; they were 'regionalized', so to speak. [9] As supports for studies rarely focusing directly on them, Boas's individual cultures often appeared more as theoretical necessities than empirical foci. To that extent, they lacked any ontological depth and, therefore, any internal dimension or organization. [10]

Many will counter-argue that this is hardly the whole story. There is another, holistic Boas at work, studying the 'genius of the people' through style, texts, meaning, secondary rationalizations, and creating a psychological anthropology in the process. I will now examine each of these claims.

Counter-evidence of an atomistic
and natural-historical Boas ?

The indispensable but taxonomic 'genius of a people',
and 'quantified' style

Boas also knew from experience that cultures display recognizable styles. As early as 1887, in writings directed against Otis Mason, evolutionist and curator of the US National Museum in Washington, he stressed that cultural traits gained meaning from the style of a culture (1974a [1887] : 62). These references to meaning and style tell of an anthropologist bent on going behind appearances, to the internal dimension of cultures. But how could he move from cultural elements to the style (or patterns, in later writings) and meaning of the whole culture ?

For ten years, Boas remained mute on the source of a tribe's style. He wrote more about it later, however, especially after 1900, 'when his critique of evolutionism was by and large accomplished', although 'his thought on this issue was by no means systematic and has to be extracted from writings that were nominally on somewhat different topics' (Stocking 1974 : 6). What were those thoughts ?

After stating that myths grow by accretion of foreign material, in the quotation mentioned above (p. 435), he immediately added : 'Much of this material must have been adopted ready made, and has been adapted and changed in form according to the genius of the people [German : Volksgeist] who borrowed it' (Boas 1974c [1899] : 97, italics added). [11] The action of this Volksgeist suggests a move beyond atomism and natural history, to something more profound ; does it reveal an organization ?

Boas's 'genius of a people' appears as 'something' acting to assimilate newly borrowed traits by imprinting a style upon them, and to fashion invented traits according to the culture's style. Since it moulds both intrants and cultural productions, it can hardly be the internal dimension of a living organism. One could call it a 'mental operator' processing cultural intrants and outputs. But if Boas closely flirted with such a notion in his study of language because he understood languages' classificatory nature, he never successfully managed to graft a similar intuition to his ethnographic practice.

If not a mental operator, what could it be ? The answer requires a slight detour. What Boas's 'genius of a people' truly does is to make sure that new, mostly borrowed, traits share the style of others within the same culture, since he underplayed the role of invention. This imposes a major distinction. Foucault wrote that natural historians understood structure as the number, shape and arrangement of parts (see p. 437), thereby distinguishing arrangement from organization, and leaving structure aside. A [441] flower arrangement can reveal a style (English, Japanese), but no internal organization, let alone a structure. And similarly with apparently more correlated styles. A melody, or a musical style, is but an arrangement of sounds. One may change them (within some constraints, which have nothing to do with correlations but everything to do with conventions) ; some results might be more cacophonic, but would still display a style.

Likewise with architecture. A Gothic cathedral could be said to have a structure, but the disposition of elements around that structure does not add up to an organization ; it defines a style. Architects could multiply the number of chapels, of doors or stained-glass windows in Gothic cathedrals, or increase their height or length, without altering much of the overall arrangement and while preserving the same style. But there is little we can multiply or enlarge in organized entities such as organisms. This is no longer a matter of arrangement, but of internal organization.

I should thus qualify my initial proposition. The elements that initially go into making a culture in Boas's anthropology are randomly associated but could be 'arranged' through the action of the Volksgeist, thereby creating a style, or what he later referred to as 'pattern', without being systemically organized. But are they ? The answer lies in what Boas had to say on style.

To my knowledge, Boas nowhere described the style of a whole culture, and dealt with style in two contexts only, namely in his study of primitive art and in one article on 'primitive literature' : 'Stylistic aspects of primitive literature' (Boas 1940/ [1925] ). [12] Here, Boas begins by noting some formal aspects of style, such as the fact that songs and tales are universal and must therefore be the primary form of literary activity ; also, the fact that poetry was inseparable from music before the advent of writing. He then emphasizes the difficulty of correctly assessing the 'rhythmic character of the formal prose' in primitive literature, an observation that leads to a first critique against technical explanations of their occurrence.

Having singled out the universal characteristics of primitive style, Boas then looks for particulars :

As soon as we enter into the art forms of a single cultural group, we may observe that there are peculiar features which are not the common property of mankind. This is clearest in certain forms of cultural life that are spread over large areas without reaching universal distribution (1940j [1925] : 495).

In the second part of the article he applies his distributional/dissociative method, noting the presence or absence of proverbs, riddles, animal tales, and epic poetry in various parts of the world, and concludes :

On the ground of the distribution of these types two conclusions may be established : the one that these forms are not necessarily steps in the development of literary forms, but that they occur only under certain conditions ; the other that the forms are not determined by race, but depend upon historical happenings (1940j [1925] : 496).

In other words, the tales agree with the peoples' conditions of life, or their predominant interests and preoccupations (1940/ [1925] : 496 ff). Overall, despite the deft use of dissociation, Boas here confines himself to formal aspects of style and their distribution to undermine evolutionist or racist theses (see also Voget 1975 : 329).

What of art ? Before Primitive art, published in 1927, Boas had already written on the topic in his maturing years (1883-1911, according to Stocking 1974). He either used his [442] method of dissociation to explode conventional interpretations of the evolution of art (1940d [1903] ; see above), or disproved the same evolutionary theses through detailed comparisons of a natural-historical kind (1940c [1908]). In both cases, and even in Primitive art, he did not write about style qua style (Kroeber 1956 : 158 ; Wax 1956). That is, he studied style as he did history, in quantitative terms (techniques, materials used, invention versus imitation) ; thus quantified, broken down in its accreted fragments, style lost its very stylistic quality. This, in my opinion, stems from the fact that Boas understood the Volksgeist in a rather taxonomic manner.

Indeed, Boas's 'genius of a people' transmutes differences (of the various traits entering a culture) into cultural similarity ; it works, so to speak, to create cultural style/pattern and singularity. Thus, through the agency of the Volksgeist, Boas still thought in terms of a classification of traits according to similarities and differences (of style) between elements 'inside' and 'outside' a given culture, [13] on the basis of their visible features. Therefore, the Volksgeist does not 'arrange', organize, or regulate the social organism's multifarious activities ; it more or less creates stylistic inventories.

Torn between dislocated elements and the knowledge that cultures may not be completely patterned but none the less display recognizable styles, Boas evoked the genius of a people but did not describe how it operated. [14] As such, the Volksgeist surges as a theoretical necessity postulated to account for the existence of style. As something indefinite that should have been demonstrated rather than posited, and the workings of which should have been partly understood, it stands as dipetitio principii (as it does for the whole German tradition, from Herder onwards). For this reason I regard it not as a mental operator but as an undetermined 'principle', a 'principle of cultural assimilation and homogenization'. And, as a 'principle', can it have more explanatory value than the 'virtues' or 'principles' of an earlier chemistry or medicine ?

Boas and texts

There are, however, other Boases. He himself urged that we look 'for the common psychological features, not in the outward similarities of ethnic phenomena, but in the similarity of psychological processes' (1974i [1910] : 247, italics added) ; according to others, he ignored outwardly similar effects to find their 'inner meanings' (Jacknis 1985 : 79 ; Stocking 1974 : 4-5), and disregarded outward appearance for process (Jacknis 1996 : 202). And there was, above all, the anthropologist interested in texts as they manifest the native peoples' culture as they see it, who thought up the theory of secondary explanations, and created a strongly psychological anthropology. I shall examine these various other Boases, starting with Boas the text-collector. The focus on texts, if anything, should indicate an interest in individual cultures, and their internal dimension.

The facts seem to belie such an inference, however. Indeed, those texts are first and above all myths, tales, and songs (Jacknis 1996 : 197 ; White 1963 : 23). And why did Boas collect those specific cultural productions ? Because, to him, they constituted 'a presentation of the culture as it appears to the Indian himself. More so, some have argued, because he could fashion them as 'objective' productions (Briggs & Bauman 1999 ; Darnell 2001 : 42 ; Jacknis 1996 :197).

Briggs and Bauman have showed how Boas, even when collaborating with Hunt, constructed these texts, and further 'detextualized' them (1999 : 486, 491), removing all 'discursive traces of his collaborators' active, conscious participation' (1999 : 514), a process they regard as one of 'objectification'.

[443]

Boas's texts were also object-like by their very contents (see Berman 1996 : 236 ; Briggs & Bauman 1999 : 504), and most texts on social organization dealt with customs. Boas seems to have showed little interest in activities as such, or in informal behaviour (Codere 1959 : 61) ; [15] unlike informal behaviour, customs are codified forms of behaviour, atomized bits of 'textualized' behaviours, so to speak. Above all, Boas considered customs, and all ethnological phenomena, as 'mental phenomena' (Boas 1974f [1904] : 23 ; Lyman & O'Brien 2003). They existed in the heads of individuals, and were best reclaimed from the heads of individuals ; thus, in Boas's study of customs, individuals neither acted nor interacted but stood as 'custom-bearers'.

Boas also presented individuals as 'text carriers' rather than cultural actors. Why ? Because texts were logically his best way of repossessing mental phenomena and, by his very definitions, texts related to one thing, namely the mind. Not the mind of this or that particular narrator, but the collective mindset of a people who produced a culture ; at the end of the road, this collective mindset was nothing else than the Volksgeist (Berman 1996 : 220 ; Briggs & Bauman 1999 : 498 ; Bunzl 1996 : 68 ; Codere 1959 : 61 ; Stocking 1982b [1968] : 223, 224 ; Voget 1975 : 335). [16]

These texts were therefore the outward, 'objective' surface manifestations of something internal, yet never grasped, for Boas did not analyse his texts from a cultural point of view. Many commentators have mentioned that the collection of texts was, for him, a manner of letting the natives speak' without the ethnographer's interference (Berman 1996 : 219, 220 ; Boas 1935 : v ; Jacknis 1996 :197 ; Mead 1959 : 32). But for myths, tales, songs, or any other text to give us an insight into a culture, for texts 'to speak', they call for interpretations, either on the part of the narrator or on that of the ethnographer. Boas avoided such interpretations. From Sapir onwards, most commentators thus noted that Boas's texts were culturally unintelligible because they were void of any contextualization and explanation (Berman 1996 : 216, 250 ; Briggs & Bauman 1999 : 512 ; Kroeber 1956 :152 ; Radin 1933 : 8-9 ; Rohner & Rohner 1969 : xxiii ; Sapir 1912 : 197-198 ; White 1963 : 55, 59).

If unexplained, could texts none the less have any explanatory value, elucidating why people performed this or that custom ? Hardly so, because of Boas's thesis of cultural explanations' secondary nature.

Boas and secondary explanations

The theory's possible intellectual origin

As early as 1891 Boas had concluded that myths were made up of accreted fragments, and that their elements 'never had any meaning, at least not among the tribes in whose possession we find them. Therefore they cannot be explained as symbolizing or anthropomorphizing natural phenomena' (1940b [1891] : 445, italics added).

In the 1899 article about the growth of myths also mentioned above (p. 435), Boas developed further his theory of secondary explanations. Unlike anthropologists (read : evolutionists) who associated nature myths and cosmic phenomena, Boas dissociated the two, concluding that myths do not explain 'the phenomena of nature observed by the people to whom the myths belong, [and] that many of them ... never had such a meaning' (1974e [1899] : 96), the very thesis he held in 1891. By some inference which I fail to grasp, he concluded from the above 'that we must give up the attempts at offhand explanation of myth as fanciful' (1974c [1899] : 96).

No doubt, some of Boas's critiques of earlier theories about myths were sorely needed and often hit the nail on the head, but what remains striking is the long list of [444] his exclusions, and the relative lack of qualifications. By 1891, Boas had already concluded that myths neither symbolized nor anthropomorphized natural phenomena. By 1899, he further wrote that they do not explain 'the phenomena of nature observed by the people to whom the myths belong' (1974e [1899] : 96), and possibly never had such a meaning for these people. Furthermore, they could not be considered as fanciful creations. The list of exclusions seems to encompass all the possible explanations of myths at the time, and leaves little room for the possibility that some theories might have accounted for some types of myths, and others for different ones. It simultaneously led Boas to conclude that 'explanations given by the Indians themselves are often secondary, and do not reflect the true origin of the myths' (1974e [1899] : 97), the core of his theory of 'secondary explanations'.

What was this theory ? The thesis that most of our customs have unconscious origins. In a nutshell, Boas held that most customs are rooted in automatic and emotional reactions to some repeated activities - that we act first, and explain most of our actions afterwards (1938 [1911] : 214). These customs thus have unconscious origins, or origins that soon disappeared from consciousness, and the explanations people give of their customs have nothing to do with their true origin, or 'cause' ; hence their secondary nature (1938 [1911] : 205-14 ; 1974e [1899] : 101 ; 1974i [1910] : 248-249). They are rationalizations, in the Freudian sense of the term (Lowie 1937 : 138 ; Stocking 1982k [1968] : 232 ; Wax 1956 : 70).

Why this theory ? Without delving into Boas's deep psychological motivations, we can safely guess at some of his intellectual ones. Interestingly, all the theses on myths excluded seem to have been associated to some evolutionary theory in one way or another. And, again, he elaborated his theory of secondary explanations early in his anthropological career ; the two moves seem not to be unconnected, as Boas himself made quite clear later : 'The essential result of this inquiry is the conclusion that the origin of customs of primitive man must not be looked for in rational processes' (1938 [1911] : 215).

As evolutionists believed in the rational origin of most of our customs, they stood as the obvious target. Indeed, there could be no better theory of human behaviour and culture to elaborate against evolutionists than one asserting the basic non-rationality, if not irrationality, of most of our customs. In this, Boas enthusiastically welcomed de Tarde's theses of humankind's basic non-rational and essentially imitative social behaviour (de Tarde 1890 ; on the profound influence the book had on him, see Kluckhohn & Prufer 1959 : 10 ; Kroeber 1956 :154 ; Lowie 1937 :106-9). Therefore, the theory of secondary explanations was a crucial assumption, and a powerful tool, to prove that no evolutionary sequence can apply to most customs, the roots of which are unconscious and, in fact, irrational.

Do secondary explanations have any explanatory value ?

If secondary explanations cannot shed any light on the origins of customs, could they none the less provide insights into cultures ? In theory, yes. Spier underscored how individuals' re-interpretations of borrowed traits generated the unity of a cultural style (see also Bunzl 1996 : 69). Secondary explanations would thus be culturally patterned and act to 'integrate' cultures' various elements into one totality, giving them 'subjective unity' for the individual (see also Berman 1996 : 69 ; Stocking 1974 : 7 ; 1982b [1968] : 225). This, however, leads us straight back to the genius of a people and, to gain any insight into the Volksgeist through secondary explanations, we would have to focus on the [445] cultural manner of rationalizing. This would necessarily involve the comparative analysis of cultural ways of explanation in two or more individual cultures to understand the cultural logic underpinning them (Berman 1996 : 219), something Boas unfortunately never attempted.

Through texts or secondary explanations, it remains difficult to grasp a sense of internal organization in Boas's ethnography. But what about the other Boases, who went behind appearances to see meaning (Jacknis 1985 : 79 ; Stocking 1974 : 4-5), or to understand process (Jacknis 1996 : 202) ?

Again, both remained programmatic. Admittedly, the study of meaning was then rudimentary. Tylor thought in terms of meaning, especially when dealing with survivals, but he sought their explanation in conjectural evolutionary reconstructions. The study of myths also begged the question of meaning ; some answers were provided (see above) but, towering above his contemporaries, Frazer sought meaning in symbolism. Boas, however, set aside symbolic studies of myths as early as 1891. Only in the case of art did he associate meaning with symbolism, albeit in a minimalist fashion (Boas 1940a¹ [1903]), but he never truly integrated symbolic analysis into his ethnographic practice.

Boas wrote about the various meanings of rattles as early as 1887 (1974a [1887] : 65 ; see also 1974i [1910] : 245) and of masks later on, but this line of reasoning led to an idiosyncratic understanding of meaning. If a rattle was used for religious purposes, it gained a 'religious meaning' ; if used for purely musical purposes, it possessed a 'musical meaning' (see Lowie 1937 : 142). By so arguing, Boas equated meaning with intentionality (or motives, namely why people make and use rattles or masks) and, in most instances, related motives to the object's function, or use. This would explain why he never delved deeply in intentionality : the implicit association between intentionality and functionality made it extremely difficult, if not impossible, truly to study meaning.

The concern with process remained equally programmatic, since Boas never studied it himself (Rohner & Rohner 1969 : xvii), arguably because processes of acculturation also stemmed from the genius of a people, whose workings he did not elucidate. If texts, styles, secondary explanations, and meaning do not refer to anything truly internal save the Geist, this raises the whole issue of Boas's relationship to psychology.

Boas and psychology

If Boas defined the phenomena to be investigated by ethnology as mental phenomena (1974f [1904] : 23), writing of their study as 'folk psychology" and mentioning the work of 'inner' forces in the formation of culture ; if he gave his summa magna the title The mind of primitive man ; if he forcefully repudiated Graebner's diffusionism in the name of psychology, one would indeed expect his ethnology to be psychological, or psychologizing, as many have concluded (Benedict 1943 : 31 ; Darnell 2001 : 42 ; Lowie 1937, among others). [17] Being psychological, it should go behind outward appearances and probe deep into social and cultural phenomena.

Boas often referred to psychology but, sadly, rarely spelled out what he meant by the term. In retrospect, at least four main levels can be identified. First, and most explicit, is the psychology he invoked to explain the origin of customs, and their transmission ; second, the psychology implicitly referred to in his critique of Graebner ; third, his inquiry into 'the mental equipment' of the various races of humanity (culminating in The mind of primitive man) ; and finally the laws, implicitly psychological as he [446] understood cultural phenomena as mental ones (1974f [1904] : 23), underlying the growth of culture (on laws and growth of culture, see 1940c [1896] : 276 ; 1974a [1887] : 64 ; 1974c [1889] : 69 ; 1974h [1907] : 188).

Being the most explicit, the first psychology is the best known, through Boas himself and through Wax's analysis (Wax 1956). This major part of Boas's psychology could be extracted from his theory of secondary explanations but Wax documented it from Boas's Anthropology and modern life (1928 ; Boas was by then 70 years old). He labelled it 'habit psychology', and described it in the following terms :

Men act largely according to habit. The earlier in life the habit is inculcated the more difficult it is to alter, the more automatic is its action, and the stronger are the emotions associated with it. Habit is fundamentally activity, not thought ; and thought about habitual activity is usually rationalization (Wax 1956 : 70 ; see also Boas 1974h [1907] : 280 ; 1974 ; [1910] : 252 ; and all Boas's elaborations on secondary explanation).

This habit psychology, however, operated at three different levels. The one Wax analysed is ontogenetic and applies to the process of socialization ; it would account for cultural determinism. The second is 'phylogenetic', referring to the origins of customs themselves. In this matter Boas held a variant of the above, as we have seen : from strong emotions associated to certain activities emerged unconscious, non-rational customs among primitives (Boas 1940l [1932] : 142-3, quoted in Wax 1956 : 70). To this, Boas added de Tarde's psychology of imitation and emotional associations to explain the 'psychological connections' underpinning similarities between cultures (transmission). In 1910, he clarified the matter, explicitly separating two main dimensions of individual psychology, namely cases when a member of society acts

as a member of a crowd, in which cases his activities are immediately determined by imitation of the activities of his fellows, [and cases in which] ... he may act as individual ; then the influence of the society of which he is a member will make itself felt by habit of action and thought of the individual (1974i [1910] : 245).

If Boas wanted a theory of the subconscious, or the non-rational, why not incorporate Freud ? Again, we can only surmise. Freud understood the subconscious as something quite creative ; it almost achieves at the individual level what Boas assumed the Volksgeist realized at the collective one. Furthermore, it is rooted in family relations, and requires interpretation. If one needed a theory predicated on humankind's essential lack of creativity and imagination, [18] however, Freud was the wrong man. One possible alternative was to declare customary behaviour to have unconscious, and therefore non-rational, origins, while submitting that customs arise out of habit. This further points to human's essential lack of creativity and imagination, and simultaneously bolsters up the case of diffusion over invention to fight evolutionist theses (Lowie 1937 : 108 ; Stocking 1982b[1968] : 226).

At the end of the road, in this type of psychology Boas did not touch upon most psychological processes necessary to study social actors and their interactions. He gave priority to mechanisms, such as spontaneous emotional association to activities, habit, imitation, which regrettably ended up describing social and cultural quasi- automata.

Its very nature makes the second level of psychological processes more elusive. When inveighing against Graebner's neglect of the 'mental' aspect of ethnological [447] phenomena, Boas actually wrote of mental phenomena arising from 'the mental make-up of a society' (1940f [1911] : 300), that is, the Volksgeist.

As to the mental aptitudes, or 'mental processes', of humankind, he selected a few, namely inhibition of impulses, power of attention, logical thinking, and originality (1938 [1911] : 125), all of which refer to mental abilities that could theoretically be tested by some kind of experimental psychology to prove the mental unity of humankind, within narrow variations. Again, it does not refer to any psychology of individual, social or cultural behaviour.

The same applies to the laws of the evolution of mental life'. I do not know that Boas ever spelled them out, but we can suppose that they were the very laws of increasing rationality and individualization that the evolutionists evoked, as he clearly stated in some of his writings (1938 [1911] : 160, 186, 210 ; 1974c [1889] : 68, 69 ; see also Cole 1999 : 277).

Overall, ethnographically and even theoretically speaking, Boas did not seem interested in the inner workings of the mind, in anything that could make us understand in some depth the complexities of social action or cultural productions. His was a psychology of the visible (habit, imitation, aptitudes) or the invisible (Volksgeist).

Conclusion

Boas can be credited with great achievements, such as his relentless fight against racial discrimination and his undermining of evolutionary reconstructions, among many others. But I doubt he could be credited with a sense of history to be retrieved, as well as a sense of the individual as cultural actor.

I have argued that Boas's critique of evolutionary methods and theses fed and reinforced an already fragmented view of culture inherited from his previous experiences, most likely the influence of German historical geography, not to mention his museographic work. I further maintain that this atomism, together with an eighteenth-century natural-historical perspective in his anthropology, were but two faces of the same coin : both apprehend socio-cultural phenomena through their external manifestations. This hypothesis further makes it possible to articulate many facets of Boas's ethnographic practice.

This, much more than his training in physics, would explain why he was above all interested in measurable, or quantifiable, phenomena, but did not manage to integrate their 'arrangements' in his work (all natural historians also failed ; see note 14). Hence the inventory of randomly assembled traits standing in lieu of culture, despite repeated references to style or patterns, a hardly historical historicism, and individual cultures subsumed under regional ones reconstructed in a quantitative fashion ; also, an anthropology of art apprehended in a formal manner, if not quantitatively, with style mostly equated with an inventory of features (Voget 1975 : 335). Hence also meaning mostly understood in terms of intentionality, and intentionality in terms of functionality ; an anthropology of social and cultural quasi-automata acting mostly through habit and imitation, when not from repeated activities and their associated emotions. In the final analysis, this is anthropology without real social or cultural actors, without actions or interactions, without true feelings or power relationships, without anything that could endogenously produce change and historical change in culture and society. I doubt that this part of Boas's ethnography should inspire an ethnography bent on history, the individual, and agency.

[448]

[449]

REFERENCES

Benedict, R. 1943. Franz Boas as ethnologist. American Anthropologist, 45, 27-34.

Berman, J. 1996. 'The culture as it appears to the Indian himself - Boas, George Hunt, and the methods of ethnography. In Volksgeist as method and ethic (ed.) G.W. Stocking, Jr, 215-56. Madison : University of Wisconsin Press.

Boas, F. 1916. Tsimshian mythology (Thirty-first Annual Report). Washington, D.C. : Bureau of American Ethnology.

_____, 1927. Primitive art. Oslo, Norway : H. Aschehoug.

_____, 1928. Anthropology and modern life. New York : WW. Norton.

_____, 1935. Kwakiutl culture as reflected in mythology. New York : The American Folklore Society.

_____, 1938 [1911]. The mind of primitive man. (Revised edition). New York : Free Press.

_____, 1940a [1887]. The study of geography. In Race, language and culture (ed.) F. Boas, 639-647. New York : Free Press.

_____, 1940b [1891]. Dissemination of tales among the natives of North America. In Race, language and culture (ed.) F. Boas, 437-445. New York : Free Press.

_____, 1940c [1896]. The limitations of the comparative method in anthropology. In Race, language and culture (ed.) F. Boas, 270-80. New York : Free Press.

_____, 1940d [1903]. The decorative art of the North American Indians. In Race, language and culture (ed.) F. Boas, 546-63. New York : Free Press.

_____, 1940e [1908]. Decorative designs of Alaskan needlecases : a study in the history of conventional designs. In Race, language and culture (ed.) F. Boas, 564-92. New York : Free Press.

_____, 1940f [1911]. Review of Graebner's Methode der Ethnologic In Race, language and culture (ed.) F. Boas, 295-304. New York : Free Press.

_____, 1940g [1914]. Mythology and folk-tales of the North American Indians. In Race, language and culture (ed.) F. Boas, 451-90. New York : Free Press.

_____, 1940h [1920]. The methods of ethnology. In Race, language and culture (ed.) F. Boas, 280-9. New York : Free Press.

_____, 1940i [1920]. The social organization of the Kwakiutl. In Race, language and culture (ed.) F. Boas, 356-69. New York : Free Press.

_____, 1940j [1925]. Stylistic aspects of primitive literature. In Race, language and culture (ed.) F. Boas, 491-502. New York : Free Press.

_____, 1940k [1929]. Metaphorical expressions in the language of the Kwakiutl Indians. In Race, language and culture (ed.) F. Boas, 232-9. New York : Free Press.

_____, 1940l [1932]. The aims of anthropological research. In Race, language and culture (ed.) F. Boas, 243-259. New York : Free Press.

_____, 1940m [1936]. History and science in anthropology : a reply. In Race, language and culture (ed.) F. Boas, 305-11. New York : Free Press.

_____, 1974a [1887]. The principles of ethnological classification. In The shaping of American anthropology, 1883-1911 : a Franz Boas reader (ed.) G.W. Stocking, Jr, 67-71. Chicago : University Press.

_____, 1974b [1888]. Tylor's 'adhesions' and the distribution of myth-elements. In The shaping of American anthropology, 1883-1911 : a Franz Boas reader (ed.) G.W. Stocking, Jr, 67-71. Chicago : University Press.

_____, 1974c [1889]. The aims of ethnology. In The shaping of American anthropology, 1883-1911 : a Franz Boas reader (ed.) G.W. Stocking, Jr, 67-71. Chicago : University Press.

_____, 1974d [1898]. The mythology of the Bella Coola Indians. In The shaping of American anthropology, 1883-1911 : a Franz Boas reader (ed.) G.W. Stocking, Jr, 148-55. Chicago : University Press.

_____, 1974e [1899]. Fieldwork for the British Association, 1888-1897. In The shaping of American anthropology, 1883-1911 : a Franz Boas reader (ed.) G.W. Stocking, Jr, 88-107. Chicago : University Press.

_____, 1974f [1904]. The history of anthropology. In The shaping of American anthropology, 1883-1911 : a Franz Boas reader (ed.) G.W. Stocking, Jr, 23-35. Chicago : University Press.

_____, 1974g [1906]. Some philological aspects of anthropological research. In The shaping of American anthropology, 1883-1911 : a Franz Boas reader (ed.) G.W. Stocking, Jr, 183-91. Chicago : University Press.

[450]

_____, 1974h [1907]. Anthropology : a lecture delivered at Columbia University in the series on science, philosophy, and art. In The shaping of American anthropology, 2883-2922 : a Franz Boas reader (ed.) G. Stocking, Jr, 267-81. Chicago : University Press.

_____,  1974i [1910]. Psychological problems in anthropology. In The shaping of American anthropology, 2883-2922 : a Franz Boas reader (ed.) G. Stocking, Jr, 243-53. Chicago : University Press. Briggs, C. & R. Bauman 1999. 'The foundation of all future researches' : Franz Boas, George Hunt, Native.

American texts, and the construction of modernity. American Quarterly, 51, 479-528.

Browne, J. 1983. The secular ark : studies in the history of biogeography. New Haven : Yale University Press.

Buckley, T. 1996. 'The little history of pitiful events' : the epistemological and moral contexts of Kroeber's Californian ethnology. In Volksgeist as method and ethic (ed.) G.W. Stocking, Jr, 257-97. Madison : University of Wisconsin Press.

Bunzl, M. 1996. Franz Boas and the Humboldtian tradition. In Volksgeist as method and ethic (ed.) G.W. Stocking, Jr, 117-178. Madison : University of Wisconsin Press.

_____, 2004. Boas, Foucault, and the 'native anthropologist' : notes toward a neo-Boasian anthropology. American Anthropologist 106, 435-442.

Coders, H.  1959.  The  understanding  of the  Kwakiutl.  In   The anthropology of Franz Boas  (ed.) W. Goldschmidt, 61-75 (Vol. 61, Memoir 89). Washington, D.C. : American Anthropological Association.

Cole, D. 1999. Franz Boas : the early years, 2858-2906. Seattle : University of Washington Press.

Darnell, R. 1990. Franz Boas, Edward Sapir, and the Americanist text tradition. Historiographica Linguistica, 17, 129-144.

_____, 1997. The anthropological concept of culture at the end of the Boasian century. Social Analysis, 4, 42-54.

_____, 2001. Invisible genealogies : a history of Americanist anthropology. Lincoln : University of Nebraska Press.

Tarde, G. 1890. Les lois de limitation. Paris : Felix Alcan.

Foucault, M. 1966. Les mots et les choses. Paris : Gallimard.

Harris, M. 1968. The rise of anthropological theory. New York : Thomas Y Crowell.

Jacknis, I. 1985. Franz Boas and exhibits : on the limitations of the museum method in anthropology. In Objects and others (ed.) G.W. Stocking, Jr, 75-11. Madison : University of Wisconsin Press.

_____, 1996. The ethnographic object and the object of ethnology in the early career of Franz Boas. In Volksgeist as method and ethic (ed.) G.W. Stocking, Jr, 215-256. Madison : University of Wisconsin Press.

Jacob, F. 1970. La logique du vivant. Paris : Gallimard.

KLUCKHOHN, C.&O. PRUFER 1959.Influence during the formative years. In The anthropology of Franz Boas (ed.) W. Goldschmidt, 4-28. (Vol. 61, Memoir 89). Washington, D.C. : American Anthropological Association.

Kroeber, A.L. 1946. History and evolution. Southwestern Journal of Anthropology 2,1-15.

_____, 1952 [1936]. History and science in anthropology. In The nature of culture (ed.) A.L. Kroeber, 63-65. Chicago : University Press.

_____, 1956. The place of Boas in anthropology. American Anthropologist, 58, 151-159.

_____, 1959. Preface. In The anthropology of Franz Boas (ed.) W. Goldschmidt, v-vii. (Vol. 61, Memoir 89). Washington, D.C. : American Anthropological Association.

Lesser, A. 1981. Franz Boas. In Totems and taboos (ed.) S. Sylverman, 1-31. New York : Columbia University Press.

Lewis, H.S. 2001. Boas, Darwin, science, and anthropology. Current Anthropology, 42, 381-394.

Lowie, R. 1937. The history of ethnological theory. New York : Holt, Rinehart & Winston.

Lyman, R.L. & M.J. O'Brien 2003. Cultural traits : units of analysis in early twentieth-century anthropology. Journal of Anthropological Research 59, 225-50.

Mead, M. 1959. Apprenticeship under Boas. In The anthropology of Franz Boas (ed.) W Goldschmidt, 29-45 (Vol. 61, Memoir 89). Washington, D.C. : American Anthropological Association.

Orta, A. 2004. The promise of particularism and the theology of culture : limits and lessons of 'neo-Boasianism'. American Anthropologist 106, 473-487.

Radin, P. 1933. The method and theory of ethnology. New York : McGraw-Hill. Rohner, R.P. & E.C.

Rohner. 1969. Franz Boas and the development of North American ethnology and ethnography. In The ethnography of Franz Boas (ed.) R Rohner, xiii-xxx. Chicago : University Press.

Sapir, E. 1912. Review of Boas, Kwakiutl tales. Current Anthropological Literature, 1, 193-198.

Smith, M. 1959. Boas' 'natural history' approach to field method. In The anthropology of Franz Boas (ed.) W. Goldschmidt, 46-60. (Vol. 61, Memoir 89). Washington, D.C. : American Anthropological Association.

Spier, L. 1931. Historical interrelation of culture traits : Franz Boas' study of Tsimshian mythology. In Methods in social science (ed.) S.A. Rice, 449-57. Chicago : University Press.

[451]

_____, 1959. Some central elements in the legacy. In The anthropology of Franz Boas (ed.) W. Goldschmidt, 146-56. (Vol. 61, Memoir 89). Washington, D.C. : American Anthropological Association.

Stocking, G.W., Jr 1974. Introduction : the basic assumptions of Boasian anthropology. In The shaping of American anthropology, 1883-1911 : a Franz Boas reader (ed.) G.W. Stocking, Jr, 1-20. Chicago : University Press.

_____, 1982a [1968]. From physics to ethnology. In Race culture and evolution (ed.) G.W. Stocking, Jr, 133-260. Chicago : University Press.

_____, 1982b [1968]. Franz Boas and the culture concept in historical perspective. In Race, culture and evolution (ed.) G.W. Stocking, Jr, 195-233. Chicago : University Press.

_____, 1992. Ideas and institutions in American anthropology. In The ethnographers magic and other essays (ed.) G.W. Stocking, Jr, 114-77. Madison : University of Wisconsin Press.

Voget, F.W. 1975. A history of ethnology. New York : Holt, Rinehart & Winston.

Wax, M. 1956. The limitations of Boas' anthropology. American Anthropologist 58, 63-74.

White, L. 1963. The ethnography and ethnology of Franz Boas. Austin : The Museum of the University of Texas.

Résumé

Franz Boas : histoire culturelle pour le présent ou histoire naturelle obsolète ?

Quelques anthropologues neé-boasiens ont récemment présenté Boas comme un anthropologue ayant un sens profond de l'histoire, de l'individu et du statut d'agent. En se fondant sur sa pratique ethnographique plutot que sur ses déclarations théoriques et programmatiques, l'auteur dévoile tout d'abord un ethnographe « atomiste » (à l'opposé de holiste) et une convergence profonde entre cet atomisme et une histoire naturelle de type linnéen. Dans ['interprétation de l'histoire naturelle par Foucault et Jacob, cela revient à étudier les phénomènes socioculturels à travers leurs manifestations externes, en éliminant l'historicité, et même les cultures individuelles, de l'ethnographie de Boas. En étudiant les contre-preuves que l’on pourrait trouver dans le Boas holiste (son travail sur le style, la signification, le « génie d'un people », les textes, les explications secondaires et la psychologie), l'auteur retrouve la même histoire naturelle et le même atomisme. Toutes les facettes de la pratique de Boas apparaissent donc comme des manifestations superficielles de cette épistemè sous-jacente, qui crée un seul et même cadre interpretatif permettant d'intégrer la majeure partie de son oeuvre ethnographique. Dans l'ensemble, cette vision du monde n'accorde que peu, voire pas du tout, de place aux individus et à leur intentionnalité.

Michel Verdon obtained his Ph.D. in social anthropology from Cambridge University in 1975, where he taught from 1979 to 1984. Since 1984, he has been Professor of Anthropology at the Université de Montréal. In addition to monographs on Quebec and Ghanaian communities, he has written extensively on the epistemology of social anthropology, of economics, and family history.

Dep. d'anthropologie, Université de Montréal, C.P. 6128, Succursale Centre-ville, Montréal, Québec, Canada H3C 3J7. michel.verdon@umontreal.ca