PII: S0304 422X(99)00025 X


1 POETICS ELSEVIER (2000) 195-231 Poetics 27 www.elsevier.nl/locate/poetic What do all day?" do animals of labor, bodies, The division class in popular thinking imagination * the and totemic Martin* Levi John Sociology, Rutgers State University of New Jersey, 54 Joyce Kilmer Avenue, Department of NJ 08854-8045, Piscataway, USA Abstract article relatively new methods of the analysis of qualitative data This investigate uses to in relation animal species and occupation between the popular imagination, socio-logical the in the world of children's literature, in order to test a claim that the class specifically habitus that the division of labor, erasing the contingent nature of class domination, does naturalizes divisions simply the internalization of objective social via into a subjective social not arise but rather begins with the application of a totemic logic which maps differences vision, people thereby differences between animals, between exaggerating and naturalizing onto social Children evidently instructed in the reality of class them. and the logic of are bodies structure they have any first-hand acquaintance before these social processes; indeed, by with working the embodied relations of class domination into the role play and role learning of the pre-school years, make it difficult for them to have any unmediated first-hand experience we by would these habitual distinctions. © 2000 Published against Elsevier Science that militate All rights reserved. B.V. Animals; Totemism; Class body; Busytown; Symbolic domination; Division of Keywords: labor I. The problem of the division of labor Introduction: us considered the first and last major topics Let by Emile Durkheim, consider the namely on the moral significance of work division of Labor (Durkheim, 1933 his An earlier version of this paper was presented at the 1997 Rutgers-Princeton Conference: Dialogues in Culture Cognition. I would like to thank the participants, especially Paul DiMaggio, Viviana and I Zelizer, Zerubavel, for their comments and suggestions. Eviatar would like to thank Lucien Preuss and for his helpful comments on and discussions of the methodology involved, John Mohr for his comments Random and both Les Editions Minuit and and House for permission to reprint illustrations. suggestions, 0304-422X/00/$ - see front matter © 2000 Published by Elsevier Science B.V. All rights reserved. PII: S0304-422X(99)00025-X

2 196 J.L. / Poetics 27 (2000) 195-231 Martin and his on Australian totemism (Durkheim, 1954 [1912]). These pro- [1893]) work often jects representative of Durkheim in two extremely different seen are being young Durkheim 'old' Durkheim, though the the Durkheim 'young' and modes, the than the old. In this understanding, appears goes from being more stodgy Durkheim Comtian to being the first positivist phenomenologist (see Parsons, last the social But both of these projects really orbit around the 1968: central set of 306f.). same namely the nature of divisions between classes of persons in society; concerns, (1) moral these subjectively understood; (3) the are and social impli- how divisions (2) this. Indeed, it may be that cations approach could have profited from of Durkheim's deliberate of these themes. This integration begins with the intellectual more paper posed by the division of labor, at least as it first appeared to Durkheim, and problem then considers problem of totemism with which Durkheim ended his researches. the political theory seen the division of labor as inseparable from Enlightenment had 1937 dependence upon one another (Smith, persons [1776]: 14-16), a mutual of the Rousseau (1967 [1755]) claimed had dependence human beings to which reduced sycophants, begin to associate for who improvement and end associ- craven mutual to gossip, backbite, and put on airs. Durkheim's (1933 [1893]) task was first ating foremost image improve the and of the division of labor, by linking interdepen- to feelings, to if not the highest, moral higher, establishing a new type of dence the order within the very division of labor that was previously seen as morally moral to ~ issue of the moral significance of the division of labor was key The corrosive. Durkheim's while today's readers may sociolatry: ascance at this blatant huck- look stering of the division of labor, at least Durkheim took the issue head on (as opposed to Marx, denied the necessity of the division of labor: one labors, fishes, then who Indeed, Durkheim forced to do consciously and sociologically philosophizes). was to stratification to do unconsciously; since seem favor socialism most systems what required the rehabilitation of the over of labor, Durkheim had individualism division come with a general justifying up of social divisions. to explanation evaluation of society, in other words, tends to turn on how we interpret the Our of the Like division first Enlightenment thinkers, we are skeptical of any labor. 2 division not based on a natural division, is and hence there are always social that to naturalize such social divisions. This is a more subtle issue pressures simply than the of inequality (which can be accomplished simply by linking it legitimizing to inexplicable decree or 'systems' imperatives); divine social divisions must seem the 1 Hence his brilliant rhetorical move of taking the emotionally loaded distinction of 'organic' as opposed to tied by T6nnies (1957 [1887]), from his first paragraph on, to Gemeinschafi 'mechanical', 'good' Gesellschafl reversing the order, so that the and kind of social organization was and respectively, the old one, but the new. not Rousseau ends [1755]: 246) 2 his Discourse on the origin of inequality as follows: "Moral [i.e. (1967 sanctioned inequality, solely by positive right [i.e. authorized, by political laws], clashes with nat- social] ural right, whenever it is not in proportion to physical [i.e. natural] inequality; a distinction which suffi- of ciently we are to think of that kind what inequality which obtains in all civilized nations, determines since it is evidently against the law of nature that children should command old men, and fools lead the while wise, that a handful should gorge themselves with superfluities, and the starving masses lack the barest necessities of life".

3 J.L. Martin Poetics 27 (2000) 195-231 197 / natural as division between night and day, or between east and west - in other as the natural words, arbitrary social divisions which we have naturalized to as as other that creations. are no longer perceived as social extent More specifi- such an they of need be made natural in two senses divisions the word; they must such cally, to both in nature and in 'second-nature'. The first requires be the social rooted that between seem to reflect an underlying division in nature (and there- division persons the non- requires that this social division be experienced as legitimate); fore second expected, i.e. navigable - - second-nature. problematic understood, a (1984 Bourdieu argued that such [1979]) naturalizing process is only Pierre has be expected, since we learn about types of persons in such a way as to make the to of seem division both navigable and natural, namely by incorporating an affec- labor charged representation social space in the form of habitus. What Bourdieu tively of the dialectic internalization of externality and the externalization of "the of calls more, simply, of incorporation and objectification" (Bourdieu, internality, or 1977 72, his, cf. pp. 84, 90) is emphasis beautiful formal vision of the process [1972]: a our minds internalize the objective distinctions of the social world, allow- whereby us skillfully ing - and with objective strategy, though not necessarily subjective to friction - world. The lack of this between cognitive instru- strategizing navigate expressly designed to fit the objects they are to catch and manipulate serves ments proof positive that there is no alternative conceptual scheme worthy of consider- as issue - rather, it prevents such an ation from even being brought to conscious- or ness. But the formal elegance of this answer is linked to its drawback - in this case, Bourdieu offers substantive statement about how real people actually learn real no divisions. In dialectical vision is silent regarding the fact that much of what particular, this about children. social world we learn indirectly, as know 3 That is, we are we the we by or another, about the social world; means do not, as full fledged taught, one learn it ourselves. The question of how these distinctions are actually inter- adults, then attention involve nalized to the mind of the child, not just the abstract must labor More learn about the division of children in a number of adult. specifically, such as approved role play. Preschoolers are generally given a dress-up cor- ways, containing ner, other things clothing representing different occupations, so among (1962 the that as George Herbert Mead children [1934]: 159) says, try on one may, self after another through play. But even before this stage, children are presented with affectively representations of the division of labor, both in disorganized charged organized via by adults, and in remarks fashion through literature, fashion occasional and written. Is it possible that such representations of the oral of labor con- division tribute the naturalizing of the division of labor? to explore this, I wish to turn to To 3 While Bourdieu emphasizes the necessary homology between the inculcation of habitus via formal to schooling of his attention) and the pre-existing one formed via the family, he has tends subject (the reduce pre-school development of habitus to this primary pedagogic action whereby linguistic capi- the tal is developed in some youngsters (e.g. Bourdieu and Passeron 1990 [1970]: 42f.). But it is not simply outlines cultural which is formed early, it is the capital of the ability to make affectively charged dis- tinctions.

4 198 J.L. / Poetics 27 (2000) 195-231 Martin other element Durkheim's thinking mentioned above, namely the meaning of the of totemic logic. Totemic 2. logic of contribution the study of the division to labor; his major was Durkheim's first the study of totemic logic, which we might last as the idealization of was to see per Durkheim was struck by two things se. he didn't try to division social (which as Boas, 1916: 323, pointed out) - that many societies keep the Ameri- separate, in and were divided up into sub-groups, each subgroup identified with a cas Australia nature), (less plant or other feature of a and that the persons animal often certain that they were related in some often to the animal that was the 'totem' stressed way their Durkheim (1954 [1912]; also subsection. and Mauss, 1963 of Durkheim concluded from this that the first mental classes were nothing but these [1903]) divisions, social hence when the Bororo say, 'we are parrots', they really mean and they are subset of the class of parrots (at least to the extent that the Bororo that a notion by class). 4 Most anthropologists, even those greatly influenced a have of from a claims here, leading to dissented debate over the nature of Durkheim, his that Evans-Pritchard [1956]), for example, argued (1974 the point was totemic logic. the person and the animal stood in that similar relation to something else; Rad- a cliffe-Brown argued that it was a spiritual relationship; Levy-Bruhl (1985 (1952) type 77, that it was a different 366-368) of logic based on partici- [1910]: 91,247, - when the Bororo say they are parrots, pation do not mean 'Bororo = Parrot' they the we do, though they do indeed way an actual identity. And f'mally later affirm writers, like Barry Bames (1981), completed the circle by implying that there is no difference between logic and our logic at all - who knows, maybe they are totemic parrots. Barnes analysis, however, simply repeated seemingly ethnocentrifugal at mistake that totemic classification was thinking some level original of Durkheim's scientific classification - that to the Bororo saying equivalent are parrots' is to 'we same as saying 'parrots are birds' thing either Westerners or the Bororo). the (for "The Bororo sincerely imagines himself to be a parrot [...] The Trumai 4 genuinely thought to be are aquatic (Durkheim and Mauss 1963 [1903]: 6f.). The alleged identity of the Bororo and the animals" has the a staple of sociologies of knowledge - in fact, parrots reason the Bororo refer to them- become long selves (or macaws or red araras) is quite a parrots and interesting story, one that has nothing to as do with the generalities of relativist socio-philosophies First of all, the red macaws to which the Bororo assert identity not 'totems' in the classic sense (the Bororo have somewhat atypical relations between are pets, and IAvi-Strauss 1966 [1962]: 99]); instead, they are see the only pets kept by the animals taboos, Furthermore, they are kept by women - Bororo. well fed, groomed, given proper names from "admired, the matrilineage, taken on trips", etc. (Shepard, owner's Second, the Bororo are matrilocal - the 1996). newly married man moves to his wife's uncle's house, which is ruled by the women, there to dwell the among "By controlling food and procreation, kin. women bind masculine loyalties and fetter her their freedom just as surely as they domesticate macaws". Hence, when the Bororo say they are social "macaws", are making a familiar and ironic comment on their own "they circumstances by means of reference to the macaw's situation". (Shepard, 1996)

5 J.L. Martin Poetics 27 (2000) 195-231 199 / only do unremarkable ethnozoologies coexist with totemic arrangements Not rather with totems, similar statements equating a class of persons with a in societies but alongside may in contemporary Western society found our quite animals be class of Consider the admittedly unusual different of the work zoological classification. case Sheldon, notorious Somatotyper (inventor the the ectomorph/meso- William of of scheme) whose (1954) Atlas of men assigned each morph/endomorph classification type precise body animal equivalent, within a more general type. 5 Thus discrete a with a Marsupials, would be 'Opossums', of the help- 4-2-4, somatype subdivision description: "Delicate, furtive marsupials who hunt beetles at night, or ful appended such scraps can find [...] Nearly innocent of mesomorphy, their most whatever they (Sheldon, against is to 'play 'possum'" attack 1954: 193). The effective defense of some types ('The Florida spotted exact,_'tude kitty', a sub-type of under-the-barn the general classification UNDER-THE-BARN KITTIES) is similar to the more classification of North American tribes (Blackbirds with white meticulous some with heads, etc.). 6 blackbirds red heads, Bororo, then, Sheldon might also say that a number of persons were Like the par- but certainly understood this neither in a he sense, nor in a ritual sense rots, literal suggested by Radcliffe-Brown, 1952, and others). Instead, it is clear that he (as used the categorization as a metaphorical system, which had two properties. The animal was was content of any animal metaphor the used to direct the interpretive first that of the reader's eye - that is, the faculties attributes are used to train the vehicle's reader seeing the tenor. After learning that in certain type of body is a leaping a gazelle, one looks at that body in a different way from when one is told that it is a squatting frog. musculature assumes interpretative prominence, some com- Certain across persons highlighted and others ignored. The second property of monalties are the system formal; it is simply wholly analogic hyperbole of metaphorical is this between human beings to differences between likening That is, differences animals. sub-groups one species are to each other of two different species are. The two as of this analogic hyperbole is wholly separable from the particular strength metaphoric content any particular vehicle and any particular tenor. linking - hypothesis is fundamentally analogic hyperbole totemism a kind of The that through exaggeration - offers an extremely simple explanation for one explication of namely thinking, half the division of a group into sub-groups each linked totemic the a of animal or plant. While bracketing to other question of the spiritual link type to the totem (as well as the awkward existence of totems like the wind), this account suggests that the root of totemic logic might be an extremely simple motion of at up species, level in the tree of life - from sub-species to going from species to one from to genus, family, etc. 7 genus that recently notoriety as Sheldon was re-discovered re-achieved many Ivy League schools had all it of its entering classes - including many of today's power elite - pose naked for Sheldon's team of anthropometric photographers. This example is from 6 clan of the Omaha. a 7 Lrvi-Strauss (1963 [1962l: 13, 16, 77, 95, espec. 89-91, also 86) made this basic point about this about formal of totemism, namely that it is property relations between groups, not the groups them-

6 200 J.L. / Poetics 27 (2000) 195-231 Martin totemic movement analogic hyperbole - ascending a level to accentuate This of difference - well suited to perform the task of accounting for and systematize seems it of an idealization of the differentiation Indeed, the social group social divisions. is differentiation) as of the nature of this explanation so breathtakingly bold well (as an in construction that it makes Durkheim's tortuous idealization of the yet simple created the solidarity' division of labor look positively amateurish in 'organic by interest it an application of Durkheim's such in the formation Further, is comparison. categories of social thought, especially the category of class, to of problem of the the naturalization the division of labor, that has motivated of work. the Bourdieu's Class bodies and animal bodies 3. started with the problem of naturalizing the division We labor, that is, how of does happen that we believe that it is natural for certain persons to do certain tasks it not others? has argued that this happens through the correlative shap- and Bourdieu the fit and the eye (habitus), so that our bodies of our jobs and, moreover, ing body employment do to others, who conclude that our so is a 'natural' result of appear to physical endowment. We can then unproblematically internalize our interpretive our of understanding objectively structured world from these "embodied, and hence the The social 1996 [1989]: 150). (Bourdieu, butcher shown as naturalized, differences" of the key illustration of this point in Distinction (see part 1) is not merely a rep- Fig. resentative a particular occupation, he is of representation of it, a concept intuited a via the sensory capacities) His body and bearing, marked by the traces of his trajec- tory through space, do not merely 'correspond' to his job, they are an embod- social theory of division of labor. How could this man be anything but a butcher - ied the meat. Bourdieu body", writes is (1984 [1979]: 192f.), "a social product he "The 'person', is tangible manifestation of the only is commonly perceived as which the most natural expression of innermost nature". As a social product, it corresponds the its in location to social space; "Thus one can begin to map out a universe bearer's to class (biological accidents apart) tends which reproduce in its specific of bodies, the universe of the social structure". logic appealing from suffers This two empirical problems: the first is that it is approach reliably that differences between bodies are the enough related to 'the job' unlikely as to allow persons to naturalize the division of labor by simply internalizing objec- tive distinctions. is, every thin butcher should delegitimize the existing social That eye, were for an additional training of the not that is, the instillation of a order, it regarding appropriate bodies that allows for the preconception of the dis- rejection selves, it is about a logic of transformations (A:B: that But he went further, and tried to reinte- :C:D). grate the two senses of totemism, ritual identification and social division, which used form as the key to upon unlock of animals chosen. I do not draw content that second portion of his work here. the 8 More exactly, Bourdieu's account corresponds exactly to Kant's (1950 [1787]) invocation of the that presence apriori structuring to 'sensibility' (Sinnlichkeit), but with the proviso of these aprioris are objective-but-contingent social constructs, not Kant's subjective-but-necessary universals.

7 J.L. Martin Poetics 27 (2000) 195-231 201 / 1. The for the job. Fig. body from (Butcher Distinction, used by permission of Les Editions Minuit.) Pierre picture Bourdieu, as as a focus on the expected and 'examples'. This pre- cases crepant 'exceptions' allows one to divide the world into vision, and exceptions, leading which examples a in which preconceptions tend to be validated no matter how skewed to circularity to might towards their rejection, is unlikely seem arise simply by internaliz- the data objective differences. But if ing is instilled, it allows for a kind of selective it abstraction leads the same bodily features to have different 'meanings' in dif- which contexts. ferent problem is that, as said above, second construction of at least skeletal The the of these systems of classification must take place before one really has forms wide social that is, during childhood. How then might we develop a pre-vision experience, abstract class the ability to selectively bodies, certain features from certain bod- for ies, and thus develop the naturalizing logic of class bodies? Further, how could we develop this children? 9 This brings us to a key question of the expanded visual as the (Grady, how the culturally specific training of namely eye is sociology 1996), into socialization. woven is here that the analogic hyperbole of totemism might come into play. Children It fascinated are and sensitive to animals (see Myers, 1998: 3), which are certainly by written, a of literature, both oral and part for children. Further, as Fernandez stock 9 Adults, too, may at times need to reinforce the distinctions located in habitus with arguments regard- ing the established distance between types of persons, but they are likely to use scientific dis- naturally Harwood (for example here, see Schiebinger, 1987). an (1979) provides a succinct summary of course the patterns by which science which intends to legitimize inequality stresses differentiation as opposed science to of human types. The literature on the use of integration to naturalize differences is too large to refer to here.

8 202 J.L. / Poetics 27 (2000) 195-231 Martin 46) has adults use animal metaphors as terms of endearment for chil- (1972: argued, very dren linking animals and their attributes to the developing early from ages, to (Thus preschoolers' dreams, and seem predominate wane in self. animals sense of children are able to develop importance forms of symbolically rep- only after other to see Foulkes, 1982: 48, themselves, 82, 115) themselves resenting 60, is even a fairly serious degree of evidence that animals Finally, seized there are by as the first material for categorization. In the words of Paul Shep- upon children "Children 10), exponent of this view, chief respond spontaneously to (1996: the ard of nature and the names and movements the animals because animals were details of are) path to categorical thought and, the the terms of a philosophy (and eventually, cosmology". 1° Contrary to the myth, popular among social scientists, that it is or impossible to 'likeness' and that all categorical labels come from lan- perceive 11 children able to perceive species with a fair degree of regularity before guage, are language. such characteristics [that distinguish species] understand "Because they to people, birds and other animals are the perfect are of diver- visible embodiments in and unity in diversity. Anciently unity to category making as the sity connected for the cognitive act itself, they constitute the child's practice ground in archetypes categories the balance making poles of likeness and difference" Shepard, (1996: that taxonomic "It only that each species occupies a not space that serves to model 48). is but each is framed in relationship to others by behavior and personal- categorizing, (Shepard, (1998 [1978]: 63; cf. 249, 253,255). ity" their actually day admirers, Durkheim and Mauss Unlike would have con- latter perception that pre-social (and prelinguistic) a of likeness between animals ceded was possible; they simply maintained that such perception of likeness was not the same as their position in a system of classification, in which objects understanding be placed nested sets of greater or lesser inclusiveness. This abstract, hierar- may in found argued Mauss (1963 [1903]), is not and in the nat- structure, Durkheim chical though mere 'likeness' is. That is, it ural certainly true that one can, in world, is fashion, that one bottle-nosed dolphin is perceive same type of thing unmediated the another bottle-nosed dolphin, and that either of these two as is somewhat individuals like not entirely like) a porpoise, and that both porpoises and dolphins are fairly (but and a and, finally, that the dolphin, porpoise, whale, whale are a little bit like right tuna and swordfish, and a little bit like cows and seals. But this system of pair- like wise of 'likeness' is nothing at degrees like our developed classification, in which all say, Flipper is a dolphin, and a toothed whale (along with the porpoise and the nar- whal), and whale (along with the right whale), and a mammal (along with the a for My not do justice to Shepard's actual claims, does he would distinguish the use of ani- t0 exposition in totemic society properly so called, mals society appropriate to hunter-gatherers, from what he terms a 'domestic' the class-rended society that arises with society, and herding. Only in the latter are cultivation group differences truly naturalized: "Class culture perverts totemic thought by replacing metaphor with homonym" (1998 cf. 145, 149, 162-170). [1978]: suffice This will be explicated in tl forthcoming essay; point it to say that ethnobiology does not find a the type of cross-cutting division of the natural world into different species associated with the claims of Durkheimians like Barnes (1981) and Douglas (1986). See Berlin (1978).

9 J.L. Martin Poetics 27 (2000) 195-231 203 / Degrees of likeness become absolute likeness at different levels of COW). pair-wise analysis. and [1903]) argued that in order to internalize such a hier- Mauss Durkheim (1963 that we not exist in nature, one must first experience such a archical does structure, as other which, predictably, they identified realm society. (More some scheme in nested system which arises from repeated fission of a family tree - particularly, the repeatedly experienced people as they negotiate a divided social world in one by - logical to actual social authority corresponds gives a template the hierarchy which interpretation of natural categories.) But it is not clear for it is easier to per- the why classes human society than among the in of animals. While one may ceive world that it is unlikely that the identity of whale, platypus, and bat as 'mam- acknowledge is mals' seen, it does not seem beyond reason that the identity of cow, goat, easily sheep can perceived, especially as they are all in a similar functional relation and be (domesticated, producing, grazing, hoofed quadrapeds). humans milk-and-meat to is as least as reasonable that the natural organization of animals Indeed, it would the for the organization of society than template reverse ~2 - Durkheim, as a form the in modern France, had little reason to obsessively sociologist and contemplate study animals, instead obsessively contemplated society. Hunter-gatherers and herders and this did And children may - in reverse. respect - satisfy Durkheim's probably the in being more like hunter-gatherers then twentieth-century sociologists. ontogenesis conclude, there is every reason to believe that children can perceive not only To species, likenesses members of the same between but also the differences the between species, and that they are extraordinarily attentive to these differences at the time in they are entering verbal thought. Thus, they may use differences which animals as template for categorical thinking in general, and to construct between the example, between persons in particular. For of in his differences types categorical of Asturian children, Femandez (1974: 124) ethnographic this: study confirms come quickly perceive a difference between to cow and a cat, chicken, "Children a donkey, etc., [...] In their earliest search for identity, these children turn dog, the [...] towards of some animals rather than others, choosing animals that imitation space the portions in the quality desirable of their culture". occupy more other words, children employ the analogic hyperbole of totemism as a way to In produce the of classification, which is then applied to the social realm, an 'category' 1998 reversal Durkheim's logic (cf. Shepard, of [1978]: 130). But this totemic exact logic, we recall, is wholly formal - no particular content is attached to analogic hyperbole. This I hazard, may be supplied by adults, partially answering our content, as eye. the actual processes involved in the training of the Bourdieuian question to supplying literature content, through In and informal interpretations of animal this the behavior, the possible applications of shape categorical structure of ani- adults J2 While it goes beyond the bounds of this paper to fully explore this question, it is worth noting that animals Shepard has emphasized that the system of 99) is not only used ontogenetically by each (1996: child in the production of cognitive categories, but phylogenetically by evolving human intelligence (in think- stark of the Durkheimian script): "Animals were among the first objects of classificatory reversal became ing. follows that interspecies concepts It the model for our social definitions".

10 204 J.L. / Poetics 27 (2000) 195-231 Martin to the world. (This has been explicitly noted by children's writers such mals social Lewis, 1963: More particularly, we can derive a hypothesis to test, as C.S. 463.) instruct use logic in children's literature can totemic the socializ- the of namely that to the reality of class bodies. In a ing circle, animals may be endowed eye as typical attributes from types of persons, and abstracted these animals used bodily with then to highlight and naturalize differences between persons. as examples of literature in children's use to accentuate differences between peo- The animals well ple in classical fairy tales and fables, animals represent characteris- known is - 1996: persons stereotyped and allegorical fashion (Shepard, a 76). Thus the of tics in of Aesop's fables involves modal interaction of representatives of two dif- type the species; around 2% involve conversations between members of the same ferent only modernization But we retrospectively call what involved the substitution species. ~3 a new form of differentiation between persons for the of dominant differ- previously entiation character, namely occupational role. 14 What children need to learn - and of they are in the early years of life through explicit curricular foci and the what taught to materials instruction and for play - has both do with the mean- of for presentation this form of division. As Shepard (1996: 84) ing noted, "It is not the roles in of has that being learned, but the more fundamental are that there are roles, and life lesson because we are a poetic species, are best grasped these, indirectly". The from this perhaps somewhat lengthy chain of reasoning is that we conclusion of use to train the developing social eye logic children so that they will may totemic the class bodies discussed by Bourdieu (1984 [1979]). Testing such a recognize might out claim of the bounds of plausibility, but I hope to offer at least pro- seem now support, an analysis of what is through the most popular work for chil- visional dren on the division of labor in the world (and has been since I was a child), Richard Scarry's work Busytown. ~5 Richard Scary has written many books about the same on This is course to be taken as a rough finding. It comes from my analysis of a collection of 207 ~3 of 1964), There largest collection I could find. (Aesop, are no clear boundaries as to what fables fables the a Aesop's consensus as to whether or not such alone person ever existed), and there are some are (let which are probably of later origin. Fables may involve any fables permutation of gods, humans, possible animals, and inanimate objects as actors, leading to a difficulty in coming up with a simple table plants, animals fable those with two different but are clearly most frequent. of types, More accurately, there was a transitional period in which 14 and occupations were inte- characteristics grally in what were to become animal folktales. related the animal represents not the pure charac- Here ter as in Aesop, nor the pure occupation (as we shall see below), but the more ambiguous Stand (here see Btir'Ocz, In Shepard's (1996: 1996: 95) words, "The folktale is an internal dialogue by which 1997). parody society members in a language where animals its not only classes and values, but a inculcates traits, hierarchical notions, and moral principles". It might also be said that if the categorical personality of system exerts symbolic violence, foiktales are often symbolic sabotage. By making, for exam- habitus regular a stand for the gluttonous bureaucrat, a ple, association pointed to by Shepard, the dominated pig not only 'inculcate' its members in a language, they innoculate themselves against the attempted sym- superiority bolic the dominating, who would fain naturalize their of in their bodies (cf. Shepard, violence 1998 [1978]: 172f.). ~5 It is perhaps worthy of note that one of Scarry's early works was a collection of Jean de la accurate Fontaine's of Aesop's fables (Retan and Risom, 1997: 47). It is difficult to get version data on the sales for children's books; see figures discussion of Scarry's popularity below.

11 J.L. Martin Poetics 27 (2000) 195-231 205 / town, establishing consistency of geography and lineage despite his high mythical a makes productivity like an amateur. Several of these books deal Faulkner that look of all but most famous is What do people do division day (Scarry, with the labor, t6 1968). Do a book entitled 'What Do People The All Day' is that there odd about thing people, only animals. L7 But that, of course, are the point of totemic logic - peo- no is are because they are different. And it is this difference that forms the ple animals, What WDPDAD), people do all day (hence of which is about the variety subject do division jobs typical town with an advanced a of labor. In Busytown animals of in the same language, get along fine, and they don't eat each other, but neither speak they do - pigs stay with pigs and have pig children. Most importantly, dif- interbreed species are represented in different occupations, and - here is the ferent differentially - the occupations generally have no relation to point physical attrib- important these a the Thus the airplane pilot is not of bird, but a fox. Occasionally utes animals. are made for special jobs - the mine is run by gophers, one lumberjack exceptions only a is (but beaver. But the star characters - those with names or special jobs one) do is display such a 'natural' relation of species to job. And yet there - a rela- not indeed bring 'naturalizing' relation. To tion, out the socio-logic of this relation, I a aggregate propose a formal analysis to the apply information contained in the dis- to persion of species across occupations. 4. Methods stated above, is about a variety of species living in one town hold- As WDPDAD who variety One may therefore code each animal jobs. is performing a job a of ing of the job performed, and his or in species. Doing so leads to an N of 272, terms her among species of animal and 20 different occupational categories. To distributed 132 losing information regarding the socio-logic of the avoid between species relation and I used as fine a coding scheme as possible. Appendix A describes the cod- jobs, extremely in Clearly, this is an detail. sparse table; conventional methods ing more analysis that rely on asymptotic statistics are inapplicable. To test, for example, of whether simply not the distribution of species across occupations is consistent or must with hypothesis of independence, one null perform an extremely computa- a tionally-intensive version of an exact test. An application of the algorithm devised by Mehta Patel (1983, 1986, incorporated into the SAS package) indicated that and fact significance null hypothesis must be rejected at a decisive .001 level of in - this But other there in a significant interaction between species and occupation. words, is ~6 pig would be Postman most important other one. the 17 This point, however, was not lost on the sales department at Random House - upon receipt of the 'What manuscript, pressure to change the title to was do animals do all day?'; responded Walter there Retan, Scarry's connection as editor in chief of children's books, defending the title: "They're not ani- a mals, people. If you use the word animals, readers will expect they're book of true facts about animal activities" (Retan and Risom, 1997: 75).

12 206 J.L. / Poetics 27 (2000) 195-231 Martin kind of What is the socio-logic of the relation between species and what interaction? occupation? that trying to uncover the logic of the dispersion of species across we Given are as logic to knowing in advance what this map, is, we have the occupational opposed than of if we let the data themselves speak chance if we try to code better a success 18 That is, we will attempt to 'cluster' or aggregate these jobs not on the them apriori. what of mean in the human world, but what they 'mean' in the Busytown basis they This that they divide and distinguish species. how meaning is formally - is, world the duality of species and job - a species 'means' based jobs it can hold, on what a 'means' what species tend to hold it. job At the simplest level, we could while 19 into a single category all those jobs that were held only by member of a sin- combine species, gle 'Rabbit jobs'. Greater difficulty, however, enters when we want to e.g. jobs that basically the same distribution of species: 40% pig, 40% rac- combine have the mouse. one could simply test course, hypothesis of independence 20% coon, Of rows corresponding to any two jobs in the occupation by species matrix, between the agglomerate progressively pairs which are most likely to be judged inde- and those (1981) This been used by Goodman has and Breiger (1994). How- pendent. method the sparse nature of this table ever, such tests inapplicable, as they depend on makes asymptotic to retrieve p-values. statistics comes common similar problems when it have to the analysis of Other techniques data. While it is possible to construct an exact test version of the Good- sparse approach, it is extremely computationally intense for a table of this size, man/Breiger be. it and clear how useful the results would is 2° But there is a different, though not related, method, which is not based on tests of significance but on actual information contained in table, and thus small Ns, though they increase standard errors, do the invalidate the Furthermore, this method is computationaUy trivial (hav- not results. theoretically form ties in to a and and methodologically coher- closed solutions) ing to examining cross-classified tables, namely the entropic ent approach approach For it is unclear whether we should begin example, a conventional coding system, such as 18 with jobs by skill requirements, or by industry, or whether we should we select a classification dividing more inter-species tied to scheme differences, such as those that require opposable thumbs concretely those that not. and do 141), duality first pointed to by Simmel This [1922]: was who argued that we could just as 19 (1955 see an individual as an intersection well the groups to which she belonged, as see a group as the union of of who belong to it. Breiger (1974) individuals such an approach, and further extensions have formalized been used (e.g. Mohr and Duquenne, 1997) to explicate the social meaning of a categorical system. Such an approach sense when we cannot be sure that we understand the meaning of a categorical sys- makes which a we do not wish to throw away information and might be contained in a joint distri- tem priori, of this and another categorical system. (Such a bution of information would occur were we to con- loss dense codings without attention to this joint distribution.) our the case at hand, I am using the In distribution of species across jobs to understand what jobs 'mean'; as will be made clear below, this is us to in such a way that the information that species gives done about the meaning of jobs is con- be densed most efficiently. 20 This is because there are somewhat arbitrary decisions that have to be made regarding how to con- exact vert discrete probability distribution retrieved from an the test to a continuous and comparable number.

13 J.L. Martin Poetics 27 (2000) 195-231 207 / on Shannon's [1948]) work on the transmission of information. Recent based (1963 this tradition greater attention from and use by sociologists than work in deserves the been has case. yet detailed discussion of this approach, but what is most important for Space forbids a entropy is while the first use of Shannon's that measure as the basis here our purposes system for statistical analysis (Kullback, 1959; cf. Luce, 1960) was asymptoti- of a to loglinear Goodman equivalent system, which is more flexible, clear, and cally the in and have computational advantages to certain circumstances known, well seems 133, 135; 1971: 169), Preuss (1980) has put forward a set of mea- (Goodman, 1970: the relation between the various dimensions of a polychotomous table which sures of small to based measures, and less sensitive to superior Ns. 2~ In a recent are chi-square Preuss and Vorkauf (1997: especially 157-159) build upon this work, and approach provide method for determining the optimal grouping of categories which maxi- a the 'terseness' the table, or the ratio of output of information to input. This mizes of then effi- measure of, in Preuss and Vorkauf's (1997: 141) words, "the is terseness a the of of a repository of knowledge".22 Preuss and Vorkauf ciency cross-classification data have user-friendly program to analyze extremely (TAXIS). For the pur- also an of brevity, I will not give a technical discussion of the measures poses but an used, impressionistic (though one which is never technically inaccurate). Interested one can the cited works for more detail. readers consult can think of the raw data as being contained in a table with R rows (in our We This case, of 132 occupations) and C columns (in our case, 20 species). consist these table 'information', and the more contains individual animals are in the dispersed table, the more information there is. That is, if all individuals were in one cell (one species and job), why look at a 132 by 20 table? Second, each variable has a one of information with it. Third, we can measure the association degree associated of two the inverse of the amount as information lost when the variables between to predict the other. This is the heart of the using approach - to one informational the test of independence with an interpretable measure of all-or-nothing replace the of association between two nominal variables. degree however, has gone farther, and used these Preuss, to do the following: to statistics (1) the rows23; (2) make a dendogram-like structure of the aggregation of order a In (Martin, 1999) I have argued that work thermodynamic version of entropy should be 21 previous to Shannon's entropy for contingency table analysis preferred the N is small relative to the num- when ber cells (also see Yuan and Kesavan, of 140). Preuss (personal communication), however, argues 1997: that the Shannon measure is to be preferred, at least when we conceive of the data as being generated as a sample a generative process with fixed (but unknown) probabilities. from by In approach, the total dependence can be 'measured' loglinear the chi-square, but this is a 22 the quantity that is always largest when no rows metric-less collapsed and hence has to be compared to are the of freedom. But the relation between degrees chi-square and the degrees of freedom always the involves the sample size, and hence the same distribution has different 'meanings' depending on the from sample there is no way to separate measures of strength Thus measures of significant infer- size. ence using chi-square type statistics. 23 All this applies equally to columns; I will refer only to rows for brevity. However, unlike the Good- and man/Breiger Preuss's approach does not simultaneously order or agglomerate rows approach, columns. One must do one first, and then the other.

14 208 J.L. / Poetics 27 (2000) 195-231 Martin (3) choose optimal stopping point in the agglomerative process. First, to rows; an rows, order take all R(R-1)/2 pairs of rows, and examine the degree to we the can us the same column information, and then arrange give rows so that which they the note are to each other. It is important to alike that this in no way most pairs next of a linear or monotonic arrangement of rows in terms of assures latent us some that be correlated with the column distribution. This is in contrast to variable might (permute of used to 'diagonalize' a table routine the rows and columns results a the to maximize the number of observations so roughly along the line con- as falling the [1,1] and [R,C]) 24, or cell one-dimensional Goodman RC(I) model necting a 1979). For example, if there were (Goodman, latent dimensions that affected the two distribution species into jobs, one 'income received' and the other 'education of and the of jobs happened to be C-shaped within the two- required', distribution space, occupations might be arranged from income-by-education the dimensional high education] to [high income, medium [medium-high to income, education] income, education] to [low income, low education] to [medium- [medium medium income, high education] (see Fig. 2 for an example). One would see a monoto- low increase in nic education nor in income across rows. neither for that of rows utilized a statistic ordering each pair determining how Recall the they were in their column distribution. We can then begin to agglomerate close pairs of by first combining those closest to one another, then those rows second-closest. (Of rows may be agglomerated course, super-rows consisting of previously into agglomerated rows.) This can be continued until all rows are merged into one. The question, then, where to stop. Many clustering techniques are forced to use ad-hoc is of thumb. and Vorkauf (1997), however, provide two meaningful crite- rules Preuss first in the 'terseness' - the tersest table is one The which the fewest rows are is ria. a successfully a column. It is basically identify measure of how exclusively to needed column is linked to one row and each row to one column, over all columns and each The criterion stopping rows. leads to what Preuss and Vorkauf call the 'eas- second in partition, maximizes the gamma introduced which Preuss (1980) - the iest' one with which one can pick out ('retrieve') a row given only column information ease vice-versa. or retrievability of a column would be how efficiently one could The retrievability select only row information. The given is bounded by the terseness, it and the 'easiest' partition is hence always as compact or more compact than the 'tersest' partition. the tersest partition is usually to be preferred on theoretical While like one the 'easiest' partition in a case use ours, where we are mak- grounds, might the first pass through finely coded data (Preuss ing Partners, 1993: 28). In the and analyses follow, both standards are employed to create which as one is partitions, nested within the other. To conclude, I will apply a method of ordering and collapsing the rows (occupa- tions) which to preserve whatever information exists in the table regarding attempts jobs, socio-logic the dispersion of species into the while simplifying this table so of that we may understand this logic. 32), 24 a method is apparently used by Bourdieu (1996 [1989]: Such though it is unsaid precisely how the diagonalization was accomplished.

15 J.L. Martin Poetics 27 (2000) 195-231 209 / ] I High income i i.--;-.- " ... ". j:ii -"2i :" ... - 15 ° ... ... i ~6 ,.. High Low ql v I education education i7i 8~ 12 ill 11191111 Low I income along Hypothetical 2. columns that is not ordered arrangement either of two latent dimensions. Fig. of numbers represent rows in a table, each (The corresponding to an occupation with a certain average row education and a certain income received. An 'ordering' of the rows according to a nearest- required principle not from 1 to 12. This would neighbor correspond to either an increase or a decrease stretches order.) either the latent dimensions that structure the of along 5. Analyses 5.1. The totemic categories I will begin with a data-reductive analysis using the methods just introduced, then supplement this additional comparisons involving other codings, and then sum- with presents the an interpretable form. Table 1 in the results of the order- marize findings and both the tersest and easiest partition ing the occupations (it was previously of determined with the exception of the single that and the single skunk, both of bug which had the same job, there was no purchase to combining species). All 132 occu- each pations in order, as well as the cluster in which they were placed in listed are of partitions. After each occupation are these numbers, the first, running from 1 two to 33, is the number of the cluster to which it belongs in the 'tersest' partition, and cluster the (from 1 to 18) is the second number to which the occupation belongs in

16 210 J.L. / Poetics 27 (2000) 195-231 Martin 'easiest' partition will call these 'minor' and 'major' categories for purposes of the (I since are until the sixth cluster, same the minor categories These exposition). the combined in the sixth major category. Blank lines separate major cate- 6-18 are Further, the represented in each minor category are listed in descend- gories. species major order importance, and all of the proportionate categories that are greater ing of a single job are than a name summing up the dominant species. Finally, the given algorithm out to be indifferent to the ordering of the rows within minor cate- turned other, (that permutation is as good as any any as long as minor partitions gories is, not crossed). 25 I have therefore taken are liberty in some cases of arranging the the rows as to facilitate interpretation. so Table 1 Re-arranged ordering and categorization of occupations Occupation Minor Major Composition Name partition (in (in major) partition minor) 1 washer 1 Window Plumber 1 1 R 1 1 Boat builders asst A 1 1 operator gate Manual B RABBIT Police chief 1 1 B 1 1 reporter Newspaper ONLY I Farm 1 1 hands (child) T 1 1 Laundress 1 1 Beautician 2 Frozen dessert Pig and Monkey - sales 2 assistant 3 3 Spinner's Dressmaker 3 3 3 Weaver 3 3 3 Seamstress 3 Secretary 3 3 3 attendant Flight 3 3 Nurse Poet 3 3 Violinist 3 3 3 Dancing teacher 3 C boy 3 delivery Bicycle 3 CAT ONLY A Sales 3 3 T Grocer 3 3 Carpenter 3 3 Shoemaker 3 3 3 3 repair Watch pilot 3 3 Helicopter captain Ship's 3 3 3 pilot Ship's 3 25 This indifference to order within minor categories is not necessarily the case; it comes from the sparseness of these data.

17 J.L. Martin 27 (2000) 195-231 211 / Poetics (continues) 1 Table Name Minor Composition Occupation Major major) partition (in partition minor) (in Police 3 3 officer labor 3 Semi-skilled 3 4 Cat and Goat - 4 worker Highway R Chemical delivery 5 5 A 5 5 author Fiction C ONLY RACCOON Gas station attend 5 5 C 5 5 (skilled) Miller O operator 5 5 Pump O N Furniture 6 Raccoon, Cat, maker Owl 6 6 engineer Train 7 Chemical 6 worker Cat, Rabbit, Fox 7 Food service 7 6 Sawmill worker 6 8 Bear, Leopard Dog, Rabbit, Cat, cutter 9 6 Rabbit Raccoon, Cat, Fabric Lumberjack 10 6 Cat, Mouse, Rabbit, Fox, Beaver, Dog, Bear Farmer 6 11 control 6 traffic Air 11 dominated; and some Pig also Cat 6 11 Banker Dog, Raccoon, Mouse, Rabbit, Owl, Goat, Skunk Baker 11 6 and Bug, Fox, 11 engineer Operating 6 worker 12 6 Coon, Postal Pig, Cat, Fox, Beaver, Dog, Leopard, Tiger Weaver's assistant 13 6 6 Raccoon, Dog, Airplane mechanic 13 G 6 asst 13 Surveyor's Mouse R dominated; and 6 also Porter 13 Pig, Bear Cat, E 13 Truck driver 6 A 6 Miller T 13 (semiskilled) Materials 13 6 handler E Beaver Cat, Rabbit, 6 14 worker co. Dog, Paper R 6 Tailor 15 and Dog Rabbit Airplane janitor 16 6 P vendor 16 Dog I Cat and 6 Ticket Fuel truck driver 16 6 G 16 Carpenters 6 assistant pilot 17 6 Airplane Fox Raccoon Dog, also dominated; Mason's 6 18 assistant Mason 18 6 worker 18 6 Stone laborer 18 Unskilled 6 Janitor 18 6 Street cleaner 18 6 Switchman 18 6 18 Fire fighter 6

18 212 Martin Poetics 27 (2000) 195-231 J.L. / (continues) 1 Table Minor Major Composition Name Occupation (in minor) (in partition partition major) painter i 8 6 House Weather forecaster 18 6 18 6 (79.4%); also some PIG Surveyor 6 Cat Dog, Raccoon, 18 dispatcher Police 18 6 Fishermen Dye workers 18 6 18 6 Chef builder Boat 18 6 Barber 6 18 salesman 18 6 Automobile 6 18 ground Airport crew Construction worker G 19 7 Mouse; R M and Dog Electrician 19 7 O also Fox E Policeman 19 7 7 tech 19 Supervising U A T S 20 Sailor 7 E E ONLY MOUSE Lighthouse 7 operator 20 20 7 Airplane navigator R Gardener 20 7 Ship's navigator 21 8 21 worker Railroad 8 Conductor 21 8 8 Sales, clothing 21 8 Cashier 21 Waiter 21 8 Chimney sweep 21 8 21 driver Taxi 8 D Mover 21 8 Baggage ONLY O DOG handler 21 8 Truck sales 21 8 G Tailor's helper 21 8 Photographer 21 8 Portrait 21 8 artist Newsstand operator 21 8 Newspaper editor 21 8 Police sergeant 21 8 8 21 Detective 21 8 Druggist Anesthesiologist 21 8 Ambulance driver 21 8 F Blacksmith 22 9 O ONLY FOX Electrical 22 worker 9 X Realtor 22 9 9 22 operator Telephone

19 J.L. Martin Poetics 27 (2000) 195-231 213 / 1 (continues) Table Major Minor Composition Occupation Name minor) partition (in partition (in major) Fox 10 Bear G B 23 thief Probable and R E Postmaster 24 10 BEAR 24 T A Contractor/engineer 10 ONLY R R 25 Mayor Fox and Leopard L P 11 26 11 Fire A chief E 26 LEOPARD ONLY O R Dentist 11 D - Detached related: but Air dispatcher 27 12 O 27 12 OWL ONLY W Eye doctor 27 teacher L Music 12 13 BEAVER ONLY 28 Miner BEAVER 13 Printer 28 Totally detached: 29 14 TIGER Omitted Spinner to Door sales 30 15 OCTOPUS Omitted door 31 16 Omitted Physician LION 17 Thief GORILLA 32 Omitted 33 HYENA Omitted photographer Wildlife 18 out that It all the rows are actually well-ordered; some rows are com- turns not separable from the main table. 26 That is, if there is an pletely O held by an occupation animal species S, and no other S holds a job held by any other species, and no of put other job O, there is no way to holds O in relation to other jobs. At the species end of Table 1, we see five detached jobs held by five detached species. These will be omitted further consideration. There were also two clusters that were 'semi- from from the of the table. That is to say, they were placed next to each detached' rest large closely but their relation to the related, ordering comprising the body of other as 1 was uncertain. I shall discuss Table reasons for this ambiguous position below. the We interpret the ordering of these rows in terms of the animals which hold can the them. this is in principle different from that possible tactic of using the rows Note to order the animals. Here we are interested not so much in what animals are 'clos- est' to another in terms of being candidates for aggregation (since no such aggre- one were understanding but rather in gations the totemic nature of occupational useful), interpretation stratification table. To begin our this of this totemic ordering, we in may begin by simply eyeballing the table, and looking for meaningful tendencies - algorithm 26 separability can be detected by permuting the order of the table before the Such is applied, a and whether this leads to examining difference in the final order.

20 214 J.L. / Poetics 27 (2000) 195-231 Martin disproportionate representation one type of job for one animal class. Starting a of top with 1, we see a cluster of rabbit jobs, contiguous to (ignoring the of the Table of dessert single occupation of frozen position sales) cat jobs. slightly anomalous the see jobs rabbit to cat jobs? Well, we considered at the end of the are Why similar a few female jobs (as well as child labor and the 'tame' job of rabbit jobs newspaper The jobs pick up from here, with a cluster of fabric-related female reporting). cat to then of personal-service female jobs, leading cluster artistic jobs, and and a jobs, the 'tame' jobs of delivery boy and grocer, then, some petit bourgeois occu- via to But also may hold semi-skilled laborer jobs, cats bring them near to pations. which raccoons. the raccoon jobs we From to the major category of GREATER PIG, which come could called without too much error 'America's Working Man' (with all the con- be of the consciousness uncovered by Halle, 1984). The minor cat- notations particular go work Raccoon- and Cat-related jobs of industrial here to the more egories from 13) jobs and transporting things (minor category lifting to the minor cat- 'doggy' of of pig jobs proper: the dominated fraction of the dominated category, to use egory language. a we see Bourdieu's number of sub-categories that disproportion- Here then employ involving stone, sanitation, and jobs the somewhat irrespon- ately pigs: (as it turns out from the text) fire-fighters and house-painters. sible to GREATER Next PIG major category is the GREATER MOUSE, involving the we jobs some mice and dogs. From here shared move on to DOG proper: note by that dog jobs involve transportation, personal service (especially picking up and moving things: handling is the occupational equivalent of 'fetch!') to some baggage requiring the of trust only a good dog is worthy of: police work, dispens- jobs kind (both anesthesiologist an drugs and pharmacist, the latter being the most as ing in America, according to Gallup), and driving ambulances. occupation trusted in quick succession, we see FOX, BEAR, and LEOPARD. Now we may not Finally, as the to see be reasons for the contiguous placement of these categories next to able feminization other could when we noted the shared we of rabbit and cat jobs. each as we must recall that the likeness of But bear, and leopard is in part due to their fox, common from other jobs, which differences come up in later analysis. may Finally, we come to the semi-detached owl and beaver jobs. Two of the owl jobs seem to to do with eyes: this makes them hard to integrate with the other owl have of who in the petit-bourgeois end found GREATER PIG. The beaver workers are are similar in that they do not fit in jobs the area of GREATER PIG where other to beavers found. Thus both owls and are are related to the dominant classifi- beavers catory scheme, but they cannot be arranged in that scheme's order - they float classifica- above, the self-employed who float above current occupational unlike not tion schemes. 5.2. Authority, skill, sector, and species This analysis seems to offer us a limited sense of the totemic logic underlying that What do all day. It suggests people the 'working man' jobs of GREATER do PIG jobs contrasted both to the feminized jobs of cats and rabbits, and also to the are

21 J.L. Martin Poetics 27 (2000) 195-231 215 / predators (i.e. fox, bear, and leopard, whom we find at the opposite end of the of the the order and finally to owls, whom I have compared to the self rabbit from jobs), each indeed To test this logic, I coded case? occupation on four this the employed. Is first is whether it involves authority variables. another, accepting the 27 The over another, no authority relation. The second is or sector of the econ- of authority the or extractive (e.g. mining, agriculture, lumber); secondary or omy: primary indus- tertiary service; sales; transportation; government; or 'underground' (i.e. trial; or with effort to avoid confusing the job made the sector (thus one Every was criminal). a frozen dessert truck is in service, not transportation). The third variable who drives semi- degree skill required by the job: no skill (for example, ditch digging), of the is (for asphalt paving), skilled example, engineer, chemical worker), skilled (operating (doctor, lawyer), or sales/service/other. The last variable professional sex. With is these and some illustrative examples, we can explicate the logic retrieved codings, the ordering. we look at all characters in the sample (as opposed to the by Here which I presented in Table 1), though in the interests of space table reduced was presentation the separable species (with from exception of the symbol- eliminate the however, interesting totals at the bottom, The give the distribution across ically lion). 272 characters, to facilitate comparison to the sample as a whole. all of distribution regarding the First of authority, it is clear that foxes in particu- all, be and general, are most likely to in in positions of command (see lar, predators 2). The mayor is a fox, the airplane pilot is a fox, etc. Indeed, when Table mayor the of neighboring town ('Workville') appears, he turns out to be a leopard. It the is almost to see the two mayors, chilling clutching bags of money they side-by-side, have collected as tax revenue to spend on public works. On the opposite side are mice, who occupy any position of authority, despite their diversity of occu- cannot 2s Pigs raccoons tend to be at the bottom of any authority relation. Owls pation. and authority in are totally disconnected from they relations - evidently, unusual that are residual craftsmen not yet proletarianized by they 29 are industrialization. women resoundingly likely to be cats, 3° are cats - both male and Second, and - are most likely to have clerical or service jobs in the service sector (see female 3 and 4). In other words, women have certain jobs not because they are Tables type from but because they are a certain occupations, of being - in this barred men's cat - which naturally tends to occupy dead end white collar positions. Dogs case are also in service sector jobs, though, as we have seen, this comes disproportionately but not being found in feminized occupations, their because of their quality of from personal service. Trustworthy creatures, they make excellent public servants (espe- 27 The of occupations, however, was done before the results of the partition, as a minimal check coding to contagion, it was impossible that employ independent coders. against given The problem with using mice is that, as Blount (1975: 152) points out, "perhaps 28 is easier to imag- it ine as] members of their own hidden social systems [mice to think that when out of sight they might and be a part of a miniature mirror world". This seems to be the case here - Scarry lets mice run whole as minor a ship at sea), thereby obscuring their relation to other animals (e.g. they drop out from worlds the larger society to fill their microcosms. 29 I suspect that the owl body type and face reminded Scarry of a lovable old German woodworker. the 50 cats make up around a fifth of While coded characters, they comprise over half of the females.

22 216 J.L. / Poetics 27 (2000) 195-231 Martin 2 Table species by authority Crosstabulation of authority or authority? Species Have under Total Has Under No authority authority authority 11 2 Rabbit Count 7 20 55.0% 10.0% 100% 35.0% Percent 23 30 3 56 Cat Count 41.1% 53.6% 100% Percent 5.4% i Raccoon 1 25 2 Count 12 4.0% 100% 48.0% Percent 48.0% Bug/Skunk 2 Count 2 100% 100% Percent Pig Count 25 21 3 49 51.0% 42.9 % 6.1% 100% Percent Count 9 9 18 Mouse 50.0% 50.0% Percent 100% 18 Dog 54 34 Count 2 3.7% 100% 33.3% Percent 63.0% 7 6 3 Fox Count 16 43.8% 18.8% 100% Percent 37.5% Count 4 1 6 Bear 1 16.7% 66.7% 16.7% Percent 100% Leopard 2 1 2 5 Count 40.0% 40.0% 100% Percent 20.0% Lion Count 1 1 100% 100% Percent Count 5 Owl 5 100% Percent 100% Count 4 3 Beaver 7 Percent 57.1% 42.9% 100% Total (all 272) Count 112 142 18 272 Percent 41.2% 6.6% 100% 52.2% policemen), and we have seen they are disproportionately likely to have as cially involve porting things from one place to another. that jobs results dovetail with the ordering of jobs in Table These and they also make 1, reference some interpretable characteristics of the different species. But how come to Recall working was identified with the pig? the that we found the minor cate- mass gory of pig jobs to be disproportionately unskilled and demeaning, such as ditch dig- ging or (though Table 4 demonstrates that pigs are by no means restricted sanitation can to Now in some cases, we jobs31). understand the socio-logic behind the such 3t One will notice that the distribution of raccoons in Table 4 (and Table 2 for that matter) is almost are identical that of pigs; but the two animals to actually not quite interchangeable, as Table 3 makes

23 J.L. Martin Poetics 27 (2000) 195-231 / 217 3 Table Crosstabulation sector of by species Species Sector Total Primary Govern- Under- Manu- Sales Service Trans- ground ment facture portation 9 5 1 1 20 Rabbit Count 4 25.0% 5.0% 5.0% 100% 45.0% 20.0% Percent 3 23 15 1 12 2 56 Cat Count 41.1% 26.8% 1.8% 21.4% 3.6% 100% Percent 5.4% 1 3 6 2 25 Count Raccoon 13 52.0% 12.0% 24.0% 8.0% 100% Percent 4.0% 2 2 Bug/ Count 100% Percent 100% Skunk 5 18 Pig 2 6 9 49 Count 9 10.2% 36.7% 18.4% 4.1% 12.2% 18.4% Percent 100% Mouse 1 7 2 6 2 18 Count 5.6% 38.9% 33.3% 11.1% 100% Percent 11.1% 1 17 2 15 4 54 Count 15 Dog 27.8% 31.5% 3.7% 27.8% Percent 100% 1.9% 7.4% Count 1 6 3 3 2 1 16 Fox 6.3% 37.5% 18.8% 18.8% 12.5% 6.3% 100% Percent Count Bear 2 1 1 1 6 1 16.7% 33.3% 16.7% 16.7% 100% Percent 16.7% 1 1 3 5 Leopard Count 20.0% 20.0% 60.0% Percent 100% Lion 1 1 Count 100% 100% Percent Count 1 1 2 1 5 Owl Percent 20.0% 20.0% 40.0% 20.0% 100% Beaver Count 2 1 7 4 57.1% 28.6% 100% Percent 14.3% 23 59 7 51 28 3 272 Count 101 Total (All 8.5% 37.1% 21.7% 2.6% 18.8% 10.3% 1.1% 100% 272) Percent of species for the job. Thus the fidelity of dogs - an actual trait based on choice wolf sociality further bred into the species - makes them ideal for law-enforcement. and proverbial them of foxes also suits The for politics. Even the connection cunning interpretable; between and the female is readily feline in our culture cats are the closely linked with the declawed female - they are seen as soft, indoor, useless ani- clear. Raccoons to have jobs with little in the way of skilled requirements, but they are highly clus- tend in in industrial sector, while pigs are found the the extractive industries and government (e.g. san- tered itation engineer) - this probably stems from the fact that raccoons are well known to have high motor also control their paws, which well suits them for putting things together, but it with seems to indicate a bit more self-reliance than that possessed by the pig.

24 218 J.L. / Poetics 27 (2000) 195-231 Martin 4 Table species by skill of labor Crosstabulation of labor? Species Skilled Total Sales, Pro- Semi- service, Unskilled Skilled skilled fessional other 5 6 2 Rabbit 20 Count 1 6 25.0% 30.0% 10.0% 30.0% 100% 5.0% Percent 2 18 18 5 13 Cat Count 56 3.6% 32.1% 8.9% 23.2% 100% Percent 32.1% 5 5 1 4 25 Count 10 Raccoon 40.0% 20.0% 4.0% Percent 100% 20.0% 16.0% Count 2 2 Bug/ Skunk Percent 100% 100% Count 5 16 11 Pig 15 49 2 Percent 32.7% 22.4% 4.1% 30.6% 100% 10.2% Count 1 5 1 4 18 Mouse 7 38.9% 5.6% 22.2% 100% 5.6% 27.8% Percent 6 17 11 4 Dog 54 Count 16 11.1% 31.5% 20.4% 7.4% 29.6% 100% Percent Count 2 8 2 Fox 16 4 Percent 50.0% 12.5% 25.0% 100% 12.5% Count 1 2 6 Bear 3 50.0% 16.7% 33.3% 100% Percent Count 1 Leopard 3 5 1 20.0% Percent 60.0% 100% 20.0% Lion Count 1 1 Percent 100% 100% Owl Count 1 2 5 2 40.0% 20.0% 100% Percent 40.0% 5 7 Count 2 Beaver Percent 71.4% 28.6% 100% 20 86 72 20 Total 272 Count 74 272) Percent 7.4% 31.6% 26.5% 7.4% 27.2% 100% (All ... though of course cats mals not appear in this guise in earlier times, when they did were outdoor, rat-trappers, and gendered male. nasty, ... pig, presents a The of a conundrum however, unless one sees the pig as an bit unconscious parody of the working class body type - the body without breeding, tending to polite boundaries and destine its bearer for simple tasks. 32 As overflow thick/thin (1984 has argued, the opposition [1979]) is reliably connected to Bourdieu the dichotomy between mental and manual laborers: it cannot be accidental that jobs for involving of stone are PIG jobs, manipulation the thickness and heaviness of the 32 As Shepard (1996: 298f.) says, "pigs have come to represent the degraded status connoting 'animal' or 'flesh' or 'body'".

25 J.L. Martin Poetics 27 (2000) 195-231 219 / stone has naturalizing affinity for the thickness and heaviness of the pig's body. the a pigs' Thus the occupational hierarchy comes from their ability to rep- position the in body polar resent a type. of the pig legitimizes class stigmatization is seen in the dispropor- That the use foolish pigs that of cause in this work. (Indeed, when Scarry tionate number mishaps accidents, character would specialize in foolish who Mr. Frumble introduced later a fumble -, he made him a pig.) Scarry is well - for his filling almost every i.e. known with outrageous accident or another, and he has claimed that he used ani- page some could characters to people so that he opposed portray such accidents with- as as mals seen as sadistic. But it is not just out animal that finds itself in these being any accidents. 16% of the pigs are involved Fully a mishap of some other, humiliating in more than any other far and unlike the other species, they are the cause of species, the 75% of the time, as opposed to the innocent victim. Less than 2% of the mishap animals, in are the cause of a mishap. But pigs fall into bread dough, other contrast, into ('Daddy toss pancakes out of windows. The main pig character fall they rivers, they himself food, getting so stuffs he breaks his children's beds - with Pig') heavy into Pig's bed, while Mommy lies snoring obliviously. move he fully appreciate this depiction, compare the serious professionalism of To the other in this scene to the absurd antics of the three pigs (one of whom firefighter smashes Fig. a door with his axe without seeing if it is even locked) (see later down By But this would mean nothing. 3). when one surveys Busytown, as in the itself, have overview pages 4-5, one begins to on an explanation of why the street given cleaner and the ditch digger are pigs. It is an answer to the question posed by the inequity of division of labor that is simpler - only slightly - than that of Davis the Moore (1945), then again, this is a book for children. and but can glosses by our findings, but supported we make sense interpretive are These scheme as a whole? Let us arrange of major classifications in a Durkheimian the the attentive the fact (discussed above) that to ordering algorithm used does schema, the necessarily correlate with a single dimension. We can summarize the important not of 2-4 findings in Table 5, which will allow us to understand the order of Tables order retrieved 1. Fig. 4 plots this Table in a three-dimensional totemic species from in which the relative positions of the species provide a key both to space, hier- their archical and what Evans-Pritchard (1950 [1940]: 109f) called 'social dis- relations in tance': between groups of persons distance a social system, expressed in "the terms of values". The order of animals is stretched into a U-shape. Animals at the bottom of U may be considered to be dominated, animals on top dominating. As the placement shall there is a logic to the horizontal while of species, it is not to we see, interpreted as a single latent dimension. The dominated animals are be raccoon, bug/skunk, and mouse. Note that Table 5 demonstrates that the vertical dimen- pig, of sion quite well with the percent correlates the species that are professionals, here who have authority, and who have skilled jobs. There are two exceptions which help us understand multidimensionality of the schema. The first is the position of the expected. - have less authority than would be mice As we have seen, this is they characteristic of mice: they are small things, fearful, and not able to command like dog, the The second exception is the predators. whose lack of skill seems out of

26 195-231 220 / Poetics 27 (2000) Martin J.L. / Fig. 3. Pig firemen. (From What do people do all day? by Richard Scarry, copyright 1968, 1979. Used by the kind penni., sion of Random House.)

27 J.L. Martin Poetics 27 (2000) 195-231 221 / with his position. It seems that his faithfulness has place relatively non-dominated ! been rewarded Table 5 used be Interpret Fig. 1 to Data to of percent having Ratio of percent Percent female Percent Species Ratio others skilled to percent Professional to authority over authority unskilled percent under others of RABBIT 10.0 15 6 .29 .12 8 8.9 21 CAT 8 1 4.0 .08 RACCOON BUG/ 0.0 .00 (semiskilled only) 0 SKUNK 4.1 2.2 .12 0 PIG MOUSE .00 4.6 0 5.6 (0/50) 7.4 6 .12 1.8 DOG 6 FOX* .43 infinite 12.5 (50/0) BEAR* 1.00 infinite 0 0.0 (17/0) LEOPARD* 1.00 0 20.0 undefined (0/0) undefined infinite 20.0 OWL 20 (40/0) (0/0) BEAVER* 0 0 0.0 infinite (0/57) (29/0) This species is disproportionately * in government. concentrated Owls beavers, as we have seen, and jobs that are classified as GREATER hold PIG jobs, yet OWL ONLY jobs and BEAVER ONLY jobs do not fit in the order of animals. They somewhat above it, as they do not quite find their place in the float occupational structure. I contended, were pre-industrial craftsmen modem Owls, beavers not were proletarianized: so too yet seem uniquely wedded to who wholly extractive industries of early industrial capitalism, and unlike the other workers, the have branched out into other sectors such as sales and service which not make would them in their occupational distribution comparable the other dominated animals. to Their presence therefore induces a partial third-dimension, one having to do with time, in space corresponding to industrial evolution. Hence the dashed line is to the the read coming be of the page; the further out, as further back in economic out development is the species. The U-shape of the main distribution indicates two paths of upward mobility (and paths it separation of these two the that induces the horizontal differentiation), is though to is not without significance that the one pictorially higher in fact leads it

28 222 / Poetics 27 (2000) 195-231 Martin J.L. T domination Increasing Increasing Increasing skill, authority autonomy, and industrial government employment employment . I Pre-industrial @ Fig. 4. table. arrangement recreated from master Totemic

29 J.L. Martin Poetics 27 (2000) 195-231 223 / positions of status. The stark opposition of these two wings is just as higher absolute import full the domination of pigs and raccoons itself. The rightward wing as of is increasing rapidly and professionalism, but only increasing author- involves not skill we involvement. government to fox to leopard, From go from the hum- and ity dog servant to the power elite. (It ble also not insignificant that the major dog state is in world is a policeman, 'Sergeant Murphy': thus dogs do actu- character Scarry's the mobility path of upward represent via Tammany Hall.) ally Irish also route quite different - it leftward involves increasing authority (mod- The is but what is most notable is that it is the route of the 'tame', culminating in erately), rabbit. the most central rabbit is Stiches the tailor, with his innumerable chil- The Scarry unconsciously upon stereotypes of the timid European immi- dren: draws a more for this character, and probably generally, this leftward route of Jew, grant, appropriate to such types normally considered tame, paradigmati- seems mobility women. This is a route of mobility, but it is obviously limited - there is no path cally the fraction to of the dominant class, the large cats, despite the evolu- dominating closeness domestic cats and leopards. tionary between sum, the totemic organization retrieved by the entropic analysis adds a richness In our of to how Scarry's analogic hyperbole operates, for it allows us understanding species only see how individual to exaggerate characteristics associated with not jobs, but how (in strict accordance with the explication of totemic logic given by L6vi-Strauss, 1963 the relations between species map onto the relation [1962]), persons of sorts. The dominated position of the pig, while 'natu- between different at appear a result of the internal porcine to of his hammy body, is as ralized' nature same an indexical relation of time what Bourdieu (1996 [1989]: the difference, referred to as "the reciprocal externality of positions". 2) the just learns not Hence what people do all day, but what kinds of people child portray what things. Precisely by refusing to of humans as humans, which do kinds require explicit decisions as to the particular would for the particular job, and person a choice of race, ethnicity, and other difficult signs, Scarry is able to provide visible children the key to unlock the objectively coded world of occupational stratification. Of course is often (though not always) unintended by the author, but this is all this more hyperbole to take the results of this exercise in irresistible analogic the reason 33 production unconscious nature of the seriously. does not decrease its effec- The matrix tiveness up a provisional habitus, a setting of dispositions tied to cate- in gories, that allows the child who has no knowledge of real people to see that sanita- performed tion labor are connected - connected by virtue of being unskilled by and the same type of person, the goofy, thick, clown. 3~ At other times, the choice of species is quite deliberate - thus Scarry clearly understands that the goat who farms references the Scandinavian visage of American Gothic.

30 224 J.L, / Poetics 27 (2000) 195-231 Martin Conclusions 6. society turn man according to the settings in which he deploys "Does not into his different men as there are varieties in zoology? as activity many a a workman, an administrator, a between soldier, The differences a scholar, a statesman, a merchant, a sailor, a lawyer, and an idler, pauper are as great - although more difficult just grasp - as those priest a to wolf, a lion, a donkey, a raven, between shark, a sea-cow, a sheep, etc. a a always been and there always will be social species as there There have species." La com~die humaine 34) zoological (Balzac, are retrieved by formal analysis are not only consonant with the The patterns theoret- discussion the means by which a kind of totemic logic can naturalize the divi- ical of of labor, they are sion consistent with an informal analysis of the illustrations also we seen. Despite this impressive internal consistency, however, there are two have objections to thesis: the first is that it simply reads too much into this obvious my (WDPDAD); have second is that I work picked one unrepresenta- light-hearted the far out work Regarding the former, I am of from the first to engage tive many. on the terrain of his portrayal of the division of labor, and the Scarry internalization of types flowing from the use of animals. First of all, Scarry's very use of character is Scarry's in terms of the division of people into categories. animals explained remarks first "[At the time of Scarry's hagiographer book, in 1963], many chil- that and books illustrations of white children had adults. But Scarry used humor- dren's ous animals in his illustrations. He avoided representing only one group of people" (Berg, 1994:11). himself seconded this: "I like to draw animals. 35 And I Scarry that children identify more closely with pictures of animals than they can think can of of child. They see an illustration pictures a blond girl or a dark- another with than who know is someone other they themselves, and competition boy haired in" (Berg, 1994: 21). The challenges of a diverse society prompted Scarry to creeps animals yet avoid offense, use describe the range of human types. to legitimizing am first to see Scarry's figures as the a form of social stratifi- Nor I Scarry received much criticism from women's groups for portraying women cation. traditional in like that of the housewife. More importantly, Scarry himself saw roles has such as relevant. "Their criticism reading been valid in some instances", a Scarry is quoted as saying. "And I'm trying to improve my [N.B.] image" (Berg, 1994: 27). 'improved his image' by improving that of women - he not only Scarry works, women 'modem' roles in his later more he even revised his first portrayed in success, Richard Scarry's best word book ever to portray female major in a animals wider of roles (Retan and Risom, 1997: 50f.) - for variety on the second example, 34 Cited in Shepard (1996: 113). 3.s Indeed, Scarry also liked animals - when he had achieved success, he made yearly safaris in Kenya, (Retan to lions, giraffes, zebras, and hippopotamuses elephants, and Risom, 1997:113). But note see none of these dominate his works, and indeed only lions and hippos are regularly seen. Why? Presum- are ably only these successfully reference types of people (lions because trustworthy dominant males, hippos are somewhat ridiculous middle-age women).

31 J.L. Martin Poetics 27 (2000) 195-231 225 / cover the pushing the baby carriage has become a male, the male edition's female a traffic female, cop etc. work is of labor or no, Scarry's division - and there- question whether, The other - is representative. Now of course, fore is only one book from one these results this Richard is the world's best Scarry children's picture book However, author. selling over 300 books selling more than 100 million copies world wide. author, If with 36 was pick any single author for detailed analysis, to test my claim regarding one to of of division of labor via a totemic logic which makes use naturalization the the it Indeed, certainly be Scarry. hyperbole, after beginning this analogic would symbolism, found article on class and animal I by the Swedish anthro- research one Orvar Lrfgren. While the work is a social history of the place of pologist in the pets context Swedish class formation, 37 it begins with an anecdote from the author's of at a his three year old son is disappointed to find that the firemen are life: parade, "It recalls me some time", pigs. LOfgren (1985: 184), "before I realized not took picture-book had over-exposed to the bizarre been world of Richard Scarry, that he pigs do not man the police station but the fire department". (We have of where uncovered reason for this.) course the he is only one author, then, Scarry's sampling weight is prodigious. Of Though 'bizarre' Scarry as are we all (if not unique, as Lrfgren unkindly calls him), course, is what sets him apart is the way he handles but point, not his interest in it. A very this different worthy of more attention than I treatment, give it here, would be DeBrun- can hoff's (1963 [1935]) Babar the king. Briefly, in this book about the ever-popular ele- phant, we Babar instructing his subjects in the building of a city, to be named follow Babar's wife. has returned from a sojourn among humans, and has gifts which after He modernist distribute construction of the overplanned the city, Celestville. will after he construction of this Pachydermal Brazilia, Babar After the presents, in the distributes special apparently constructed solely for this building (see Fig. 5). "He a purpose a gift to each elephant and also serviceable clothes gives for work-days and suitable beautiful clothes for holidays. After thanking their King most heartily, the ele- rich [1935]: all dancing with glee" (DeBrunhoff, 1963 home 14). But this is a phants go misleading. The elephants, who until this time bit been naked quadrupeds, have emerge this room as clothed bipeds - each from a job and a new identity. They with do not return to four legs once they stop their obsequious dance; they remain now protohumans. It an inversion of Scarry's logic - animals of the same species must is different like be differentiated; humans are to from other animals become humans their interspecies variation is greater due to the division of labor. 38 because In Scarry alone, ~6 sold over 4 million books world-wide, over 60 million sold entire. 1980 classes ~7 201f.) demonstrates that the British middle (1985: seized upon birds as 'paragons of L6fgren bourgeois virtues' - unlike other animals they were clean, focused on their nuclear families, and rela- tively discrete sexual matters. Those that were not were castigated as bad birds (more particularly, the in evaluative or of birds) in remarkably blatant 'Cockneys' discussions in supposedly educa- 'Irishmen' tional works. ~8 The point that Babar presents animals as differentiated only when they become civilized (human- essay ized) recently been noted in a somewhat strained has by Kohl (1995: 7, 10, 17). Interestingly, Kohl

32 226 J.L. / Poetics 27 (2000) 195-231 Martin 5. From of nature to civil society in Celesteville. Fig. state by the by Babar DeBrunhoff, copyright 1935, 1963.) Used king, the kind permission of (From Jean House.) Random make no mistake, this differentiation is And paramount interest to DeBrunhoff. of He a long (and rather dull) description of how the division of labor works has ("If Barbacol tailor] wants a statue for [the mantelpiece, he asks Podular [the sculp- his tor] to carve one for him, and when Podular's coat is worn out Barbacol makes a new one", 1963 [1935]: 24), even down to reciprocity between the DeBrunhoff, the officer guards them from enemies, and army street cleaner who sweeps up who also notes that DeBrunhoff's son Laurent, who has carried on the Babar books, has made quite clear that between he the seriousness of the implications of the Babar series regarding the relations understands he races sexes, and that, like Scan'y, and is trying to 'update Babar's image'.

33 J.L. Martin Poetics 27 (2000) 195-231 227 / streets. These identities become all encompassing for the de-ani- their occupational - malized enforced parade to honor their dictator, Babar has all the at elephants the he true by their trade, so that fashion can survey in corporatist elephants organized as they pass before the social organs him. does a summer make, as Aesop says; not not even two. But swallow One perhaps evaluates my particular claims here, the development of however cognitive one the that the division of labor cannot be forever shrugged off with a schema naturalize primary image, be studied as part of must socialization. 39 We ontological but general fully understand the importance of the class probably of the butcher cannot physique Bourdieu's [1979]) example 'The body fits the (1984 (Fig. 1) - nor why in job' picked that particular example - without comparing it to Bourdieu butcher in the Busytown Fig. 6). No person could simply internalize the subjective correlative (see the objective of butcher, a position and trajectory that leave readable to position in the There assuredly are, even body. France, endomorphic, in butcher's traces ectomorphic butchers, though this mesomorphic, may be uneven. and distribution as and Mauss (1963 [1903]) Durkheim something 'more' is Furthermore, argued, to transform mere likeness into categorical identity, and it is such categorical needed that to required identity truly naturalize the division of labor. We have seen evi- is of that occur through the totemism can analogic hyperbole, by mapping dence this between people onto differences between animals. Children may be pre- differences to use the formal characteristics of the animal kingdom to develop the disposed be empirically of classification which can then sense used to order the social valid world. But the content of that mapping is to some degree determined by the social- ized form which structured oppositions between animals are presented in litera- in a form bears the imprint not of the natural divisions between animals, but ture, which naturalizing of division of labor. of the the note on coding A: Appendix A were coded if the species and occupation All be determined. In some characters could characters no clear occupation (e.g. Wild had Hiccup); in others, characters were cases, Bill behind others so as to make shielded unclear what species they were. In at least one case it (e.g. vendor on p. 47), the animal could have been a young wolf, a fat mole, or a newstand same dog. were very few such husky Some care was taken to make sure that the There cases. character not coded twice; unnamed characters were was to be the same within a assumed story (the dog fireman getting off the truck is, unless otherwise indicated by varietal differ- ences, the as the dog fireman in the firehouse), but to be different across stories. (A cat same in cat story is not, unless otherwise indicated, the same as the ditchdigger ditchdigger in one request), different As one can a by examining the raw data (available upon story.) there see were few (or no) very of the latter duplication. examples 39 One of the most important ways in which this socialization takes place, in which one generation can whatever rapidly introject into the next generation and crucial information it has distilled from condense time immemorial, is to make it see-able (see the lovely discussion at the end of Crosby, 1997). We of should expect any naturalization of the division therefore labor to appear most infallibly in printed materials.

34 228 J.L. / Poetics 27 (2000) 195-231 Martin 6. The fits the job II: A Busytown butcher. Fig. body by do all day? by Richard Scarry, copyright 1968, 1979. Used do the kind permis- people What (From Random House.) of sion were observed doing one Often that might be only one portion of a self- animals task businessman's In that case, it was assumed job. the animal was the business- employed that and hence the driver of the boat builder man, was classified as a boat builder. But when truck the driver was working for a larger company, he was classified as a truck driver. Also, truck different some collection of different animals perform a tasks that might best be in cases of as part of one occupation (e.g. farming thought tasks that might be assigned to a includes truck an operating engineer, an unskilled driver, If the context seemed to suggest laborer). that the workers were interchangeable, they were just coded as belonging to that more general occupation (e.g. When one species had a monopoly on a certain activity or company farmer). all mice staff, the beavers in the mine), ship's were classified as 'sailors' or 'miners', (the as of their particular activity (explosives, drill, irrespective as is was believed that little pump), light be shed on the division of would when it was due to ethnic takeover. labor Only workers were coded - excluded were of course tramps and hobos, but also wives not the in labor force. A brief note about them is in order. Wives were found for paid follow- the ing pig (Mommypig), cat (Huckle's mom), species: (Blacksmith Fox's wife), bear (Chief fox Road Engineer's wife and Besty's mom), and rabbit (Stiches' wife). Other women identified a were pig, mouse, dog (the tiny wife of cat, plane passenger), rabbit, and one fox and the rac-

35 J.L. Martin Poetics 27 (2000) 195-231 229 / on the page. 4° In other words, the distribution of non-working women does not coon first association of with cats, but is not restricted to that association. challenge the women References Fables Aesop, translated by S.A. 1964. New York: Penguin. Handford. Aesop, of cognition. 1981. conventional component in knowledge and the Philosophy of the Barry, Barnes, On 11,303-333. Social Sciences 1994. to Scarry: A tribute Julie, the young at heart. Minneapolis, MN: Abdo. Berg, Richard 1978. cat- classification. In: E. Rosch and B. Lloyed (eds.), Cognition and Brent, Berlin, Ethnobiological Hillsdale, Earlbaum. 9-26. egorization, NJ: 1975. Animal land. New York: Morrow. Blount, Margaret, American 1916. origin of totemism. The Anthropologist 18, 319-326. Boas, Franz, Jrzsef, 1997. Stand reconstructed: Brrrcz, closure and institutional change. Sociological Contingent Theory 215-248. 15, Pierre, 1977 Outline of a theory of practice, translated by Richard Nice. Cambridge: Bourdieu, [1972]. Press. Univerity Cambridge 1984 [1979]. Distinction: A social critique of the judgement Bourdieu, taste, translated by Pierre, of Nice. Cambridge, MA: Harvard University Press. Richard Pierre, 1996 [1989]. The state nobility, translated by Lauretta C. Clough. Stanford, KY: Stan- Bourdieu, University ford Press. Pierre, and Passeron, 1990 [1970]. Reproduction in education, society, and cul- Jean-Claude Bourdieu, edition, translated by Richard Nice. London: Sage. second ture, Ronald L., 1974. Breiger, duality of persons and groups. Social Forces 53, 181-190. The Breiger, L., 1994. Dual aggregation on the Ronaid of relational homogeneity. Paper presented at basis the International Social Networks Conference, New Orleans. Crosby, Alfred 1997. The measure of reality. Cambridge: Cambridge University Press. W., Kingsley and E. Moore, 1945. Some Principles of Stratification. American Sociological Davis, Wilbert 309-321. 7, Review 1963 [1935]. Babar the king, translated by Merle S. Haas. New DeBrunhoff, Random. Jean, York: Mary, How institutions think. 1986. Syracuse University Press. Douglas, Syracuse: Emile, 1933 [1893]. The division of labor in society, translated by George Simpson. Durkheim, Glencoe, 1L: Press. Free Joseph Emile, The elementary forms of religious life, translated by [1912]. Ward Durkheim, 1954 Glencoe, IL: The Free Press. Swain. Emile and Marcel Mauss, 1963 Durkheim, Primitive classification, translated by Rodney Need- [1903]. ham. IL: University of Chicago Press. Chicago, E.E., Evans-Pritchard, [1940]. The Nuer. Oxford: Oxford University Press. 1950 Evans-Pritchard, E.E., 1974 [1956]. Nuer religion. New York: Oxford University Press. Fernandez, James 1972. Persuasions and performances: Of the beast in every body ... and the W., of DeadaluslO1, 39-60. metaphors everyman. James W., 1974. The mission of metaphor in expressive culture. Fernandez, Anthropology 15, Current 119-145. David, 1982. Children's dreams: Foulkes, studies. New York: Wiley. Longitudinal Goodman, Leo, 1970. The multivariate analysis of qualitative data: Interactions among multiple classi- fications. Journal the American Statistical Association 65, 226-256. of contingency Leo, The analysis of multidimensional Goodman, tables: Stepwise procedures and 1971. direct estimation methods for building models for multiple classifications. Technometrics 13, 33-61. Alfalfa 40 this book there is no wife of In the farmer (goat); she does appear in other works.

36 230 J.L. / Poetics 27 (2000) 195-231 Martin Leo, 1979. models for the analysis of association in cross-classifications having Goodman, Simple Journal ordered American Statistical Association 74, 537-552. of categories. the a Criteria whether certain categories in determining cross-classification table 1981. for Goodman, Leo, with special reference to occupational categories in an occupational should be combined, mobility Journal Sociology 87, 612-650. table. American of Sociology The 1996. visual sociology. Visual scope 11, 10-24. John, Grady, of 1984. America's working man. Chicago, IL: University of Chicago Press. Halle, David, 1979. Harwood, environment, and the legitimation of social policy. In: B. Barnes Jonathan, Heredity, Hills, Shapin order, 231-248. Beverly Natural CA: Sage. S. and (eds.), 1950 [1787]. Critique of pure reason, translated by Norman Kemp Smith. London: Kant, Immanuel, and Company. Macmillan 1995. Press. we burn Babar? New York: The New Kohl, Herbert, Should 1959. Information theory and statistics. New York: Wiley. Kullback, Solomon, Claude, 1963 [ 1962]. Totemism, translated by Rodney L6vi-Strauss, Boston: Beacon. Needham. L6vi-Strauss, 1966 [1962]. The savage mind. London: Weidenfeld. Claude, Lucien, 1985 How natives think, translated by Lilian A. Clarke. Princeton, NJ: Levy-Bruhl, [1910]. Press. Princeton University Horn On ways of writing for C.S., The three Book Magazine, October, 459-469. 1963. Lewis, children. 1985. Our friends in nature: L6fgren, and animal symbolism. Ethnos 50, 184-213. Orvar, Class R. 1960. The theory of selective information Duncan, some of its behavioral applications: Part Luce, and of developments in mathematical psychology. Glencoe, IL: I Press. Free Martin, Levi, 1999. Entropic measures of belief system constraint. Social Science Research 28, John 111-134. from 1962 [1934]. Mind, self, and society Herbert, the standpoint of a social behaviorist. Mead, George IL: University of Chicago Press. Chicago, in Cyrus and Mehta, R. Patel, 1983. A network algorithm for performing Fisher's exact test R. r x Nitin c tables. Journal of the contingency Statistical Association 78, 427-434. American Mehta, Cyrus R. and Nitin R. Patel, 1986. FEXACT: A FORTRAN subroutine for Fisher's exact test on unordered r c contingency tables. ACM Transactions on Mathematical Software 12, 154-161. x John W. Vincent Duquenne, 1997. The duality of culture and practice: Poverty relief in New Mohr, and 1888-1917. York Society 26, 305-356. Theory City and 1998. and animals. Boulder, CO: Westview. Gene, Myers, Children 1968. The structure of social action, vol. Parsons, New York: The Free Press. Talcott, I. Lucien 1980. A class of statistics based G., the information concept. Communications in Sta- Preuss, on - Theoretical and Methodological A9(I 5), tistics 1563-1585. Preuss Partners, 1993. TAXIS: A knowledge base. Zurich. and Lucien and Vorkauf, 1997. The Knowledge Content of Statistical Data. Psychometrika Preuss, Helmut 133-161. 62: 1952. Structure and function A.R., primitive society. London: Cohen. Radcliffe-Brown, in Walter and Ole Risom, Retan, The busy, busy world of Richard Scarry. New York: Abrams. 1997. Rousseau, 1967 [1755]. Discourse on the origin of inequality, Jean-Jacques, by Lester G. Crocker. edited New York: Washington Square. Scarry, Richard, 1968. What do people do all day? New York: Random House. Schiebinger, Lorna, Skeletons in the Closet: The first illustrations of the female skeleton in eigh- 1987. The anatomy. Gallagher and T. Laqueur (eds.), C. making of the modern body, teenth-century In: Berkeley, CA: University of California. 42-82. E. Claude Shannon, [1948]. The mathematical theory of communication, reprinted by Claude E., 1963 Shannon Warren Weaver. Urbana, IL: and University of Illinois Press. The Sheldon, William H., 1954. Atlas of men: A Guide for the somatotyping the adult male at all ages, with the collaboration C. Wesley Dupertuis and Eugene McDermott. New York: Harper. of made Shepard, The others: How animals 1996. us human. Washington DC: Island. Paul, Shepard, Paul, 1998 [1978], Thinking animals: Animals and the development of human intelligence. Athens, University of Georgia Press. GA: Simmel, Georg, 1955 [1922]. Conflict and the web of group affiliations, translated by Reinhard Bendix and Kurt H. Wolff. Glencoe, IL: Free Press.

37 J.L. Martin Poetics 27 (2000) 195-231 231 / Adam, 1937 An inquiry into the nature and causes of the wealth of nations. New York: [1776]. Smith, Library. The Modem (Gemeinschaft 1975 Community and Ttinnies, [1887]. und Gesellschaft), translated Ferdinand, society Charles P. Loomis. East Lansing, MI: Michigan by University Press. State Yuan, and Kesavan, H.K., 1997. Bayesian estimation of Shannon entropy. Communications in Sta- Lin Theory 26, Methods tistics, 139-148. and Department Levi is an Assistant John in the Martin of Sociology at Rutgers University. He Professor is interested in formal properties of cultural and belief systems and the ways in which social networks and social shape them. Other active interests include the numerical analysis of qualitative institutions and change relation between economic data, and sexual decision-making. Recent articles include the sociological knowledge heteronomy in classical and theory', (Sociological Theory, 1998); 'Authoritative 'Structures of power in naturally occurring groups' (Social Networks, 1998); 'The myth of the con- (Theory sumption the rise of the desiring subject and and Society, 1999), and (with James economy Wiley) 'Algebraic representations of beliefs and attitudes: Partial order models for item responses' (Sociological Methodology, 1999).

Related documents