1
"You Got to Keep It Holy"

Introduction

The trees around the trapline shack were shrouded with animal remains. Skulls, antlers, and cloth packages of smaller bones hung from the trunk and branches, suspended with twine and leather thongs. The skinned carcasses of otters and martens were ranged along the limbs, frozen and twisted so that their naked heads faced the riverbank. Inside, the owner of the shack, a Cree Indian trapper, explained why he hung animal bones and corpses in trees.

I don't scatter my meat all over the place. Like when I skin an animal I don't just throw it outside. I go and hang it someplace. Well, like beaver meat, if you feed it to the dogs, that's here in the bush. But you don't feed a scrub dog outside in town. Those, y' know, dogs running around in town. You don't feed them. You're going to spoil your luck. The next time you go in the bush you don't kill a beaver. Because you're playing with it, playing with the meat if you scatter it all over the place. You got to keep it holy. I mean not to drop your meat, not get blood all over the home. This way you'll be lucky. Pikisitowin ['purity'], ta-pikisit ['someone will be pure']. You don't i-mitawakit ['someone plays with it']. Like if you scatter it, if you don't pay attention what you do with it. Mess around, throw your skins all over the floor, people stepping on them. Animal respect himself, he doesn't want that. You got to try to keep it dean. Away from people. Especially woman. Woman with the rags on, that's terrible if he [woman] starts stepping on your fur. That's bad maskihkiy ['medicine'].

In a well-known passage, Max Weber defined action as "social insofar as, by virtue of the subjective meaning attached to it by the acting individual (or individuals), it takes account of the behavior of

others and is thereby oriented in its course" (1978,1:4). This book forms in part an extended reflection on the limitations of representing foraging or hunting-gathering ethnographically as a mode of social action implicating exclusively human actors. The foragers themselves may experience their productive project quite differently, with determinate consequences for the practical conduct of their foraging. Writing of the Labrador-Quebec Algonquians, Speck (1935a :72) made this observation long ago and with disarming clarity:

To the Montagnais-Naskapi—hunters on the barest subsistence level—the animals of the forest, the tundra, and the waters of the interior and the coast, exist in a specific relation. They have become the objects of engrossing magico-religious activity, for to them hunting is a holy occupation.

Speck identified here a distinctively Algonquian conception of game animals as reactive social others, alternately collaborating in and obstructing the designs of men and women who kill them with guns and traps. In this conception, society embraces rather than excludes animals, and the events of killing and eating them are experienced and talked about as so many ongoing instances of social interaction.

There is nothing unusual in the claim that American Indian foragers ascribe to their animal quarry intellectual, emotional, and spiritual characteristics paralleling in some respects those constitutive of human selves and persons. More novel in this discussion is the conclusion that these definitions of animals traverse the usual ethnographic partitions between "production" and "religion" as, respectively, technical and symbolic practices. Implicit in the theoretical division of ethnological labor is the ascription to Algonquians and other foragers of distinct and incommensurate conceptions of their quarry. Ecologically oriented analyses of boreal Algonquian foraging strategies represent the hunters' perspective on animals as exclusively instrumental. Conversely, Speck represented hunting as a "holy occupation" but was silent on how this sacredness accrued to the technical conduct of production. Cree conceptions of the social and sacred animal are maximally visible to outsiders when embedded in practices conventionally labeled "religious." Neither the structure of ritual enactments—reverent deposition of animal remains in trees, for example—nor the pragmatic intentions of the ritualists are intelligible apart from this definition of animals as active and reactive social subjects. But the social animal figures also in more technical contexts. Some Rock Crees experience animals as social others not only when they sing and bum offerings to them but also

when they search for them, kill them with weapons, butcher their remains, and eat them. Cree discourse itself represents animals concurrently as use values and as social others in all the contexts in which people understand themselves to be interacting with them. Cree maciwin 'hunting' and manitokiwin 'religious practice' are conceptually distinct categories of human practice, but the animal as reactive other figures in both spheres.

In the first section of this book, I attempt to characterize this social animal, to examine the complexities of the Cree human-animal relationship as it is objectified in literature, informal discourse, and ritual enactment. The project thus reconnoiters terrain already traversed by a distinguished community of scholars, pushing the inquiry into an ethnographically little-known comer of northwestern Manitoba and attempting to disclose new perspectives from what people there were kind enough to tell me. This section concludes by suggesting that the Cree representations of the human-animal relationship are profoundly and perhaps necessarily chaotic and disordered. The human and animal categories are themselves continuous rather than discrete, and their interpenetration seems to preclude stable representations of causality or sociality in hunter-prey interactions.

In the second section, I address a different objective, suggesting that aspects of the technical conduct of hunting and trapping cannot be understood or explained without consideration of Cree conceptions of social animals. Both the economic practices of distributing and consuming meat—"social relations of production" in the conventional Marxian idiom—and the technical practice of foraging itself as a "productive force" are irreducibly social (Ingold 1988:274-275). Whether solitary or collective, hunting, like any human labor process, is regulated by social intentionality. Hunting, as the Weberian criterion would suggest, takes careful account of the behavior of others. The social others of whom some Crees take account include not only human foragers but the animals themselves and other beings who regulate hunters' access to them. If we seek a comprehension of Cree decisions about where and when to look for animals, how to hunt and kill them, how many to kill, and how to appropriate them as food, domestic manufactures, and market commodities, the moral commitments and antagonisms that hunters experience with their prey are as pertinent to our understanding as their knowledge of animal habitat and biomass. So much is consistent with the anthropological premise that cultural conceptions taken to be factual are factual in their consequences. Crees

have a different understanding than non-Indians of the essential characteristics of the boreal fauna, and they act on this understanding both in ritual and more narrowly technical contexts. In the concluding chapters, I use documentary sources to address three such contexts: the overhunting of big game animals and furbearers, the subsequent inception of management practices, and the priority of immediate over delayed return foraging strategies. In this way, the analysis of three ricers of Cree foraging serves as the occasion for extending a semiotic reading to two unfamiliar venues: foraging societies and the labor process itself.

The Asiniskawiðiniwak Cree

The word "Cree" refers today to a continuum of culturally and linguistically related native people inhabiting the subarctic boreal forest from Quebec on the east to British Columbia on the west. The term originated as a contraction of "Kiristinon," a French borrowing of Kiristino , the Ojibwa name for a division of Cree-speaking people living south of James Bay in the mid-seventeenth century (Pentland 1981a :227, Bishop 1981b :158-159). The shortened form emerged by the late eighteenth century as a genetic label for other Cree-speaking groups successively encountered by the French and English traders as they moved north and west. The extension of the name reflected the traders' practical knowledge of linguistic similarities rather than any ethnic or political unity acknowledged by the people so denominated. Documentary sources mention numerous named groups associated with particular regions; these names presumably reflect to some degree ethnic and territorial distinctions recognized by the Crees themselves (Pentland 1981a :227-230, 1981b :269-270). ^[1]

Broad ethnic classifications developed by Euro-Canadians and now in common use by academics partially coincide with internal distinctions recognized by Crees. West of Quebec, anthropological usage distinguishes the Plains Crees occupying the southern prairie provinces and adjacent United States from the Western Woods Crees in the boreal forest to the north. The Plains Crees identify themselves as Paskwaw-iyiniwak 'prairie people' and speak varieties of the y -dialect of Cree. The Western Woods Cree (J. G. E. Smith 1981:256) category is comprised of the Swampy, Thickwoods, and Rock divisions, also

groupings recognized by Crees themselves. The Swampy Crees are today n -dialect groups inhabiting both the Hudson Bay lowlands and some inland areas of Manitoba, Ontario, and Saskatchewan. They call themselves Maskekowak 'swamp people', a name attested in 1700 by Bacqueville de La Potherie's reference (1931:258) to the "Mash-kegonhyrinis or Savannahs" living on the Nelson River and perhaps by even earlier forms (Pentland 1981a :227). Thickwoods Crees arc bands who today speak northern varieties of the y -dialect and occupy the boreal forest in Saskatchewan, Alberta, and northwestern British Columbia. They call themselves Sakaw-iyiniwak 'thick woods people', a name attested in 1749 in the translation "Christinaux du Bois fort" (Margry 1879-1888, 6:616). The Rock or Missinippi Crees, with whom this book is primarily concerned, speak the "Woods Cree" or ð-dialect and today occupy the Churchill River drainage in northern Manitoba and Saskatchewan. The people of the Cree communities of Granville Lake and Pukatawagan on the Churchill River in northwestern Manitoba identify themselves in their own language as Nihiðawak 'Crees'. To distinguish themselves from other Cree divisions, they use either Asiniskaw-iðiniwak 'people of the country of abundant rock' or Kiwitinaw-iðiniwak 'northern people'. When speaking English or French, they identify themselves as "Crees." I refer to them hereafter for brevity as "Rock Crees," a gloss that several individuals suggested for Asiniskaw-iðiniwak .

The Swampy, Rock, Thickwoods, and Plains Cree divisions now recognized both by Crees and by anthropologists reflect a complex history of amalgamations and migrations. The project of putting contemporary bands into correspondence with groups noted in early documents is complicated by the fact that group names can refer to three levels of social inclusiveness: broad ethnic divisions, regional bands oriented to particular summer fisheries or trading facilities (later reserves), and hunting groups into which regional bands divided in winter. A Cree of Pukatawagan, Manitoba, might be defined in different contexts as Nihiðaw 'Cree' (as opposed to other Indians or to whites), as Asiniskaw-iðiniw 'Rock Cree' (as opposed to Swampy Cree or Plains Cree), as Pakitawakan-iðiniw 'Pukatawagan person' (as opposed to other regional hands or reserve groups), and as Mwakwa-sakahikan-iðiniw 'Loon Lake person' or member of a group that hunts and traps at Loon Lake (as opposed to other Pukatawagan trapline groups). It is not always dear to which of these levels of social classification the names given in documentary sources refer. Neither, given the fluidity

and mobility of Cree society, is it probable that groups or their names remained stable over time. Contemporary awareness (among non-Indians) of the Rock Crees as a distinct Cree social division derives from Smith's (1975) ethnographic sketch of the Crees of Brochet, Manitoba. ^[2] The missionary, Marius Rossignol (1938, 1939), served at the Oblate missions at Pelican Narrows and Lac la Ronge and identified the "Assiniskawidiniwok" or "Cree of the Rocks" as a separate group. The only earlier mention of the Rock Crees occurs in a list of what Alanson Skinner (1914b ) took to be regional band names of the Plains Crees of Saskatchewan. Thompson's (1962 [1784-1812]:71) eighteenth-century description of the Churchill and Nelson River drainages as the "Stony Region" perhaps reflects Cree usage.

In northern Manitoba today, Rock Crees ordinarily use the ð-dialect No such exact correspondence can be projected into the past, although this dialect was spoken in contemporary Rock Cree territory on the lower Churchill River from at least the late seventeenth century. Henry Kelsey (1929:9) in 1690 used "Nayhaythaways," a nominalized borrowing (with English plural -s ) of Woods Cree nehiðawew 's/he is a Cree or speaks Cree' to refer to the people he encountered inland from the coast in northern Manitoba, probably in the Churchill River drainage. This form became an English designation for Crees in northern Manitoba during the eighteenth century (Pentland 1981b :268). The ð-dialect was also spoken in the seventeenth and eighteenth centuries along the west coast of Hudson and James bays and inland to the east and south in areas of Ontario, Manitoba, and Sakatchewan where it has since been replaced by the n -dialect (Pentland 1978:107-110). During this period, the relationship of dialects to named social divisions is obscure. The single link between the Rock Crees and an earlier named group was provided by Julien Bighetty and Solomon Colomb of Pukatawagan who identified the Misinipiy-iðiniwak as the earliest occupants of the Churchill River. Bighetty represented the Missinippi as making extensive journeys from the source of the Churchill in Alberta to the mouth of the river on Hudson Bay. Following the establishment of the inland trading posts, they settled down at different points along the river in Manitoba and Saskatchewan, merging with other Cree groups to become the Rock Crees. The form is dearly cognate with the "Mishenipee," "Misshenippe," and "Michinipi" who are noted as trading at York Factory in 1715 and Churchill in 1717 (Ray 1974:53,70; Knight 1932:57). Uncertainty about the early location of this group derives from the varying references of misinipiy , literally, great water. Crees

today use it to refer to the Churchill River, although Arthur Ray (1974:70) claims that the traders used it in the 1700s both for the western section upstream from its confluence with the Reindeer River in Saskatchewan and for Reindeer Lake. James Knight (1932:161) referred to "ye great Water Lake" as the home of the "Meshinnepee" in 1717. David Pentland (1978:107) suggests plausibly that "ye great Water lake" was Southern Indian Lake and that its Missinippi inhabitants were speakers of the r -dialect of Cree, subsequently displaced or absorbed by ð-dialect speakers moving West. Southern Indian Lake is today called Iðiniw-sakahikan 'Indian Lake', although this may be a back-translation of the English name. "Missinippi Cree" is emerging in the 1990s as a new social self-designation at Pukatawagan.

Genealogies of members of the Mathias Colomb Band of Pukatawagan and Granville Lake show that they descend primarily from families trading at the Nelson House and Southern Indian Lake posts in the late eighteenth century. Other ancestors came into the area during the later nineteenth century from Churchill, from the area of the second Nelson House post on Three Point Lake, and from Pelican Narrows and Lac la Ronge in Saskatchewan. More broadly, contemporary Rock Crees descend from Woods [ð] dialect-speaking people resident there prehistorically and possibly also from r -dialect groups long resident on the Churchill and from other Woods dialect groups moving west into the area in the 1700s. They identify the Missinippi Crees of the Churchill River as among their ancestors. Cree presence in the areas west of Lake Winnipeg and the Nelson River has conventionally been regarded as the result of mid-seventeenth-century westward expansion prompted by the military advantage of firearms and the search for new trapping territories (see Jenness 1932:264, Mandelbaum 1940:169-187). Smith (1987) recently demonstrated that Cree-speaking people already occupied their present territory at least as far west as Lac la Biche, Alberta, before European penetration. It was the name "Cree" that initially migrated west as traders successively applied it to the Cree-speaking Indians they encountered. Smith's conclusions are consistent with archaeological findings, although the western limits of prehistoric occupation continue to be debated. Most archaeologists concur that the complexes of materials referred to collectively as the Selkirk composite were manufactured by the ancestors of the groups now recognized as Western Woods Cree. Summarizing the archaeological evidence, David Meyer (1987:196) concludes that "Cree occupation of northern Manitoba began by at least the 1200s." The occupation may, of course, have

begun much earlier. The weight of the evidence indicates that the Rock Crees and other Western Woods Cree groups are descendants of populations inhabiting the boreal forest west of the Nelson River and Lake Winnipeg for centuries before the introduction of the fur trade in the late 1600s.

Manitoba Crees in the Early Fur Trade

The Rock Crees' ancestors were hunter-gatherers, sharing with other subarctic and arctic peoples a specialization in meat and only minor utilization of the botanical foods that elsewhere underwrite the reliability of the foraging praxis. Acquisition of European goods and involvement in the trade as fur suppliers to Indian middlemen began in all probability well before the establishment of the Hudson's Bay Company's coastal establishments in the late 1600s. Crees rapidly defined themselves as dependent on such European manufactures as muskets, axes, kettles, metal knives, and ice chisels, and there was also demand for tobacco, tea, ruin, and textiles. The introduced technology rapidly displaced indigenous manufactures fabricated from wood, bone, and stone. Men hunted and trapped moose, barren land caribou, woodland caribou, brown bears, beavers, lynxes, porcupines, and waterfowl. In the fur trade, many species such as beavers and lynxes acquired new significance as commodities. Other furbearers of commercial significance were wolves, wolverines, red foxes, arctic foxes, muskrats, skunks, minks, otters, martens, raccoons, fishers, and weasels; some of these were also eaten and used in domestic manufacture of clothing. Women fished, collected berries and other plant foods in summer, and trapped small game such as hares, ptarmigans, and martens. Women also constructed lodges, prepared hides and pelts for clothing, cooked, cared for young children, retrieved game, cut wood, and drew the toboggans during residential movement.

Cree males valued big game hunting over other subsistence pursuits: a verb for "hunt" referred only to pursuit of moose, caribou, and bears (Thompson 1962 [1784-1812]:83). Moose were hunted all year but especially during the rut and when high winds and deep snow made stalking and pursuit more efficient. During deep snow in spring, Crees sometimes chased down moose on foot and killed them with knives (HBC B.91/a/1 [1805-06]). Caribou were hunted most intensively during spring and fall migrations at river crossings; fences and impoundments were also used. Bears were hunted in their dens in winter.

Beavers were shot during times of open water. After freeze-up in winter, hunters chiseled open beaver lodges after blocking the runways. Dogs were used to track and comer animals in many hunting strategies. Smaller animals were killed with deadfalls, springpole snares, and gun traps. Beginning in the 1790s, steel leg hold traps with castoreum or other baits were set for beavers and other animals in both summer and winter. Men were humiliated when it became necessary to fish (HBC B.141/e/1 [1815], Thompson 1962 [1784-1812]:101) but readily resorted to it when necessary: one hunting group was observed angling for northern pike through the ice in December (HBC B.141/a/2 [1808-09]).

In the mid-1700s, Crees assembled in large groups to travel together to the Hudson Bay coastal forts in summer for trade. Most Crees from the lower Churchill River probably traded at Fort Churchill following its establishment in 1717. In that year, Knight was told by Crees from Southern Indian Lake that they could reach Churchill in four days, whereas a trip to York Factory farther south required at least ten (Knight 1932:161-166). Some bands adjusted to the fur trade by spending the winter hunting in the prairie and parklands south of the boreal forest. Here they traded their used European goods for furs produced by Plains Indian groups and brought these down to Hudson Bay in summer (Ray 1974:51-71, Rich 1960). Following the establishment of inland posts and winter outposts by the Hudson's Bay Company and their Canadian competitors in the 1770s, many Crees no longer made the trip down to the bay. The middleman role was eliminated, but Crees provided the inland posts with meat as well as furs. In 1782, the Crees suffered great population losses from a smallpox epidemic. Thompson (1962 [1784:-1812]:93, see Ray 1974:105-113) estimated that over half the population died, leaving in the "muskrat country" between the Nelson and Churchill rivers a Cree population of 644 persons or 92 families, each with an average of seven members.

The rivalry between the trading companies intensified after 1783 in the "muskrat country" (Morton 1973:441-446), resulting in construction of numerous posts and outposts. ^[3] The Hudson's Bay Company's most permanent establishments during this period were Nelson House on either Nelson or Flatrock Lake and its outpost on Southern Indian Lake. The Cree annual cycle was closely articulated with the opening and dosing of inland posts in the fall and spring. For Crees, summer was the period of maximal population concentration and sedentism, with the same families usually reuniting seasonally at the

same fishing sites. In the aggregate, these families probably composed named regional bands identified with particular rivers or lakes and habitually exploiting the environing territories. Graham (1969 [1767-1791]:171) wrote that the Crees remained around lakes and rivers in the summer subsisting on fish, caribou, and buffalo, "but in winter they move about continually to where provisions are to be had, seldom abiding a fortnight in one place." Thompson (1962 [1784-1812]:79), however, contrasted the Crees who were "scattered by three or four families over a wide extent of forest" with other Indians who sometimes assembled for two or three months, suggesting more mobility and residential dispersal in summer than is usually presumed.

Families assembled at the trading establishments in the fall. George Charles, for example, found ten families waiting for him when he arrived to open Granville House in September 1795. Throughout September, he reported Indians leaving in separate groups for their "winter quarters" (HBC B.83/a [1795]). Before freeze-up, the band divided into multifamily hunting groups that traveled by birchbark canoe to winter hunting areas. Winter residential mobility was frequent, although a large caribou kill sometimes permitted several months of relative sedentism. In winter, hunting groups moved successively from camp to camp on snowshoes, using toboggans to transport equipment and furs. Food storage was minimal, although dried or smoked meat and pemmican products were sometimes prepared and transported. The traders perceived the Crees as usually too indolent to preserve and store food for future use (HBC B.91/a/8 [1822-23]). Both collective and individual hunting strategies were practiced, but meat was always shared among the coresident families, such reciprocity being an irreducible moral axiom. Charles (HBC B.91/a/8 [1822-23]) wrote that "they are a quiet, inoffensive people, and remarkably hospitable, as they seldom or ever fail to offer anything they may possess and generally of the best." In the event of a large kill, the hunting group sometimes moved itself to the kill site.

Crees regarded big game hunting as a more efficient subsistence strategy than fur trapping, although furbearing animals were also eaten. The traders believed that the need or desire to hunt big game for food diverted time from beaver trapping. Men concentrated on trapping during early winter when the furs were prime. Thompson's narrative (1962 [1784-1812]:101-106) of his winter at Reed Lake suggests that Crees scheduled big game and beaver hunting consecutively rather

than concurrently. During beaver trapping, the hunting group sometimes divided into its component nuclear or polygynous families. For example, "Tapapahtum" trapped beavers in fall and early winter, switching to moose hunting when deep snow accumulated. He was then joined by "Wiskahoo" who had also hunted beavers separately. Similarly, "Apistawahshish" parted from others to hunt beavers. In these cases, the constituent families of hunting groups separated in fall to hunt beavers and then came together again for big game hunting later in the winter. Contact between traders and Crees was frequent throughout the winter: Indians continuously visited the posts to trade, and traders frequently traveled to Indian camps to collect furs. Furs were also cached until the Indians visited posts in summer (HBC B.91/a/1 [1805-06]). Each adult traded his or her own furs. In addition to furs, Crees traded fresh meat, dried meat, fat, and moose and caribou hides. With the breakup of ice in late April or early May, canoes were manufactured or retrieved, and the groups converged toward the summer fishing sites.

Winter hunting groups were flexible in composition, the same families tending to hunt together each winter but some shifting temporarily or permanently between groups. In 1822, there were six groups trading at Nelson House, the largest with twenty-eight members and the two smallest with eight (HBC B.91/a/8 [1822-23]). These groups were associated with particular regions of the country, but there is no positive evidence for delimited hunting tracts over which individuals held usufructuary or exclusionary rights (see chap. 9). Hunting groups and perhaps some regional bands possessed male leaders whose authority and influence derived from age and foraging expertise. Group decisions were made by consensus, leaders suggesting or initiating courses of action with which the concurrence of others was voluntary. Most power to control others' decisions and behavior was limited to the authority of men and women over their unmarried junior relatives. Crees viewed with contempt what they conceived as harsh treatment of women by Chipewyan males (Thompson 1962 [1784-1812]:105).

Parents encouraged marriages between first and second cross-cousins, the children and grandchildren of opposite-sex siblings. These unions tended to re-create in successive generations affinal connections among the same hunting groups or the same families within hunting groups, although marriage between previously "unrelated" persons also occurred. Of the Crees trading at York Factory in the 1740s, Drage

wrote, "The consequences of these Marriages are a strict Alliance between the Husband and the Wife's Relations and reciprocally between the Wife's and the Relations of the Husband, as to their assisting each other" (1968 [1748-49], 1:208). After marriage, couples lived and traveled with either spouse's parents and perhaps also with other relatives. A temporary period of residence with the bride's parents was reported by the traders among Crees trading into the coastal forts in the 1700s. Women attempted to space births at three-year intervals to prolong lactation and preclude the burden of carrying two infants; botanical medicines and perhaps other means were sometimes used to induce abortions (Ellis 1967 [1748]:92; Isham 1949 [1743-1749]:104). Eight of the seventy-six men in Charles's census were married polygynously, seven of these to two wives and one to three. Thompson (1962 [1784-1812]:82) wrote that polygyny was rare, the number of co-wives not exceeding three. He associated the practice with widowhood, as men married the wives of deceased male relatives and friends.

In late winter and early spring, Crees sometimes moved to the posts well before open water. In March 1806, John Charles encountered a number of Crees en route to Nelson House for the spring goose hunt (HBC B.91/a/1 [1805-06]). He stated that the Indians all came in with their families to the Indian Lake post in April 1806 (ibid.). The Crees began building birchbark canoes in late April, including the large freight canoes commissioned by the traders (HBC B.83/a [1795]). Crees were inclined to remain in the vicinity of the posts in spring longer than the traders desired: the need for "getting the Indians away" is reiterated in April entries from Nelson House (HBC B.141/a/1 [1802-03]). In 1809, the Crees assembled at Nelson House during April and May, apparently as a response to resource scarcity during spring (HBC B.141/a/2 [1808-09]).

By 1821, when the competing Hudson's Bay and Northwest companies amalgamated, both furbearing animals and big game were greatly reduced in numbers in the Churchill and Nelson River drainages. The District Report of Nelson House for 1825-26 states categorically that the furbearing animals were exhausted, the larger game animals nearly exterminated, and the Crees dependent on the post in winter for provisions and likely to starve without it (HBC B.141/e/2). The Indian Lake and Nelson House posts were dosed in 1823 and 1827 to allow beaver numbers to increase (Fleming 1940:lx ), although posts at Split Lake and Three Points Lake continued to operate in the Nelson River District.

The Late Eighteenth and Nineteenth Centuries

Crees say that Pukatawagan Lake provides an abundant summer and winter fishery and that it has been a gathering place for their people for time out of mind, a status that probably prompted the construction of Northwest Company and Hudson's Bay Company posts there as early as the 1780s. When the Nelson House and Southern Indian Lake posts closed in the 1820s, wading orientations may have shifted west into Sakatchewan toward Lac la Ronge, southwest to The Pas or Cumberland House, or east to the second Nelson House established on Three Points Lake in 1833. Following the resumption of competition in the area, the Hudson's Bay Company reopened their post at Pelican Narrows, Saskatchewan, in 1874. By 1889, Nelson House maintained outposts at Southern Indian and Burntwood lakes (HBC D.25/6). In 1893 (HBC D.25/17), the Pukatawagan Crees were trading both at Pelican Narrows and at the Burntwood outpost. By 1898, an outpost at Pukatawagan was being supplied from Pelican Narrows and run by Louis Morin (HBC B.158/d/11), and this later became a permanent facility. Rossignol (1939:64) identified Pukatawagan together with Lac la Ronge, Pelican Narrows, Burntwood Lake, and Nelson House as the principal summer gathering sites of the Rock Crees around 1900.

An important trajectory of change began in 1878 with the visit of the Oblate priest Bonald from the mission at Pelican Narrows, Saskatchewan. Bonald visited Pukatawagan during the summer in the expectation of finding Indians there, and he was evidently not disappointed since Oblate tradition states that the entire population gathered to embrace Catholicism. For reasons that remain unclear, the majority of the Crees rapidly converted. Following the construction of a church and the presence of a resident priest after 1913, visits to Pukatawagan at Christmas and Easter became customary. In 1918, Father Émile Desormeaux became the permanent priest and remained in the community until his retirement in 1980. Catholic officiation at baptisms, marriages, and funerals became customary. The church proscribed polygyny and attempted to discourage—seemingly with little success—the preference for marriages between cross-cousins. The acquisition of French and/or biblical names for persons of both sexes began during this period. The mission was also significant as a source of emergency food and of certain Western medical supplies and techniques.

In the late 1800s, the pattern of frequent residential movement during winter began to be modified by the construction of log cabins built in imitation of those used by white trappers; these replaced the conical lodge as the usual winter dwelling. This increasing sedentation was made possible by the introduction of dog traction and thus greater winter mobility, greater commitment to food preservation, and the availability of flour and some other foods from the white traders. Log cabin hamlets contained between two and eight houses, typically occupied by members of one or two winter hunting groups. In winter, men sometimes moved between the hamlet and the bush on extended hunting and trapping trips while women and children remained behind. In other cases, the entire family spent periods of the winter at hunting and trapping camps away from the hamlet. Families often built and occupied two or more cabins at different sites in their territories, one site sometimes being used as a fishing camp in summer.

By 1900, a de facto if not de jure hunting territory system was in place in the Pukatawagan area (see Rossignol 1939). The log cabin hamlets were positioned on lakes or rivers whose environs were used in recurrent winters by the same multifamily hunting groups. Crees talk about these territories as central places rather than as bounded and continuous tracts of land. The larger winter log cabin communities used by hunting groups who came together at Pukatawagan in summer were at Granville Lake, Burntwood Lake, High Rock Lake, Sisipuk Lake, Hughes Lake, and Russell Lake; Pukatawagan was itself the site of a log cabin community occupied by one or two hunting groups during the winter. These and other winter communities were the sites of trading outposts operated by the Hudson's Bay Company, Revillon Frères, and independent traders.

Administratively, so-called treaty Indians at Pukatawagan and Granville are members of the Mathias Colomb Band, formed in 1911 from a group formerly included in the Peter Ballentyne Band (from 1901) and before that in the James Roberts Band (now the Lac la Ronge Band), which signed adhesion to Treaty Number 6 at Montreal Lake, Saskatchewan, in 1889. References to the James Roberts and Peter Ballentyne bands in government documents suggest that these were administrative labels for Crees occupying Lac la Ronge and points east along the Churchill River who had not yet settled onto reservations; they probably included from the first Crees with a summer fishing and mission orientation to Pukatawagan (Canada 1889:236, 1900:137,

1901:139). Between 1904 and 1907, for instance, the "Ballantine Band" is described as meeting at Pelican Narrows for its annuities, as without any permanent abode, and as scattered across the country as far as Nelson House to the east (Canada 1906:94, 1907:136, 1908:145). In 1911, the 221 Crees trading at Pukatawagan were formally recognized as a separate band with Mathias Colomb as chief. Reservations were surveyed at Pukatawagan and High Rock Lake in 1929; band members' desires for additional reserves at Granville Lake and other log cabin settlements were belatedly addressed in the 1980s and 1990s.

Pukatawagan Crees first encountered Euro-Canadian education when the Guy Hill Indian Residential School was opened in 1926 at Sturgeon Landing, Saskatchewan. In that year, twenty students attended from Pukatawagan, and until a local school was introduced in the 1950s, all students were sent out to either Guy Hill or comparable institutions. It was not possible for me to determine how provincial and federal educational policies were administered and enforced during this period. Some juveniles attended residential schools while others did not, and those who attended did not complete the same number of grades. Assessments of Guy Hill and other residential schools by former students and their parents range from tempered approval through condemnation of the long separations, hard physical work punishments, and proscription of spoken Cree. The facility was reputedly burned down by an ungrateful student in 1952 and subsequently rebuilt at The Pas. Some students used their acquired knowledge of English and Euro-Canadian culture to exploit opportunities for wage labor that subsequently became available. Until the 1950s, most students returning to Pukatawagan were rapidly resocialized as foragers.

Opportunities for wage labor remained limited through the first half of the present century. Some Crees worked as transporters and in other capacities for the fur trading posts, moving shipments by boat in summer and by tractor train or horses and wagons in winter. Beginning in 1928, construction of the Island Falls hydroelectric dam near the winter settlement of Sandy Bay, Saskatchewan, provided some employment. In the 1930s, improved transportation by plane and tractor train facilitated the extension of commercial fishing into the Pukatawagan area. For some groups, commercial fishing in summer and winter became integrated into the seasonal cycle, alternating with fall and spring trapping. In 1951, mining operations were shifted from Sherridon on the southern margin of the Pukatawagan regional band territory to the

present site of Lynn Lake to the north. Extension of the railroad line north from Sherridon to Lynn Lake between 1951 and 1954 provided employment and radically diminished Pukatawagan's isolation.

The early 1950s saw the development of Pukatawagan from a summer fishing-trading post-mission complex to a micro-urban community occupied throughout the year by most families of the Mathias Colomb Band. The influences resulting in the formation of this community were complex and mutually reinforcing, including diminishing economic returns from trapping, the availability of transfer payments and wage labor, and the introduction of a local school. Perhaps because of the activities of white trappers, Cree conservation measures and territorial practices ceased to function in the 1930s, resulting in furbearer and game depletions. As a result, the provincial Department of Mines and Natural Resources dosed beaver trapping in 1941, subsequently introducing quotas. A depression in fur prices in 1943 combined with game shortages resulted in extreme economic hardship during the 1940s. Beginning in 1944, members of the band in cooperation with provincial officers divided the total area exploited by the band into bounded traplines registered to individuals or families. Despite trap-line registration and the eventual recovery of beavers and other species, income from trapping became increasingly insufficient relative to requirements for food, equipment, clothing, and other commodities obtained from the traders. In 1945, family allowance payments were begun to families with minor children; other federal and provincial payments including old age pensions and emergency rations for the destitute followed. By the late 1940s, many families had shifted their residential orientation toward Pukatawagan, leaving it only for extended periods of trapping in fall and spring. Initially, tents and log cabins were occupied, but the band acquired a tractor and sawmill in the early 1950s, resulting in dusters of occupant-constructed houses linked by a road that was later extended off the reserve to the railroad line. By the mid-1950s, most members of the band were resident in the village throughout most of the year. Perhaps the shrewdest explanations for the "posting" process offered by Crees identify the interaction of family allowance payments with white educational policy. A Department of Indian Affairs grade school staffed by Oblate sisters was established at Pukatawagan in 1952, initially operating between July and April to allow families to disperse to the bush for spring trapping. As several persons noted, families had to send their children to the school if they

wanted to continue to receive family allowance checks, and they stayed at Pukatawagan most of the year to be near their children.

The period between the 1950s and the present has been one of increasing community development and movement away from the subsistence foraging and fur trapping economy as full-time occupations. The Department of Indian Affairs provided new houses in the late 1960s, aligning them on the main road to facilitate connections with the hydroelectric generator; this residential pattern replaced an earlier dispersion of house and tent clusters occupied by related families. Telephone service, an airstrip, a permanent nursing station, and a new school building, now with classes through grade twelve, were introduced in the early 1970s. Both men and women have increasingly sought wage labor both within and outside the community: Crees are employed by the Canadian National Railroad, the Hudson's Bay Company, Manitoba Hydroelectric Company, the nursing station, and the school. During the late 1960s, changes in provincial policy made welfare payments widely available to individuals in reserve communities, an innovation whose effects are still debated by the people of Pukatawagan. During the 1970s and 1980s, the community has increasingly assumed control of its own administration and development, resulting in the creation of many additional jobs in the different branches of the band government.

When I came to Pukatawagan in 1977, trapping, hunting, fishing, and collecting edible and medicinal plants continued to be important for many Pukatawagan residents, although people said that commitment to these extractive practices had decreased since the 1950s. Relatively few families perpetuate the older pattern of residing in cabins or log cabin hamlets on their trapline territories throughout the winter. As has been the case since the time of first contact, subsistence activities have been integrated in various ways into the markets of the majority society. Hunting continues to be combined with fur trapping and commercial fishing, and some men combine both with seasonal wage labor. Most trapping and hunting today is undertaken by all-male groups who travel from their homes on the reserve to their trapline cabins, spending variable periods of time in the bush. Mixed sex and age groups still go out to the trapline for extended periods in April for spring trapping. Beavers, muskrats, and lynxes provide both commercially valuable pelts and staple meats. Bears, woodland caribou, and moose continue to provide meat and also hides employed locally in the manufacture of foot-

wear, mitts, and such prestige garments as beaded jackets. It is probably unnecessary to note at this juncture that Pukatawagan and other Indian reserves I have visited remain distinctively "Cree" in language, culture, and self-identification despite the social and environmental transformations they have experienced in the present century. Attitudes toward the bush and the foraging practices strongly identified with Indian culture are variable among all age groups (cf. Brightman 1990). As one Cree elder put it, "The day will come when the Whiteman will not allow you to hunt. The fish and moose will one day disappear. Hunting is no place for modem people" (Bighetty and Senyk 1986:37). In the words of another, Luc Dumas, "The trapline is just like the Whiteman's bank. It's the Indian bank. There should always be one guy in the family covering it. Kids come back here from school, and there are no jobs. No jobs? Well, they can always start up again with the 'bank.'" Behind such reflections is often the deeply felt conviction that whites and their manufactures are only transient presences in the north, perhaps in the cosmos. If the government collapses, if the supply of goods and transfer payments from "up south" is terminated, some Crees believe that they will return to the bush to hunt and fish.

Granville Lake

In 1986, seven years after initial work in the area, I returned to Pukatawagan and stayed with a Cree family who in many ways typified successful control of the accelerating changes of the period. The mother held a post in the band office, one daughter was a teacher, another daughter was a band constable. The new house was modern, "comfortable" by contemporary Euro-Canadian criteria. Family members code-switched effortlessly between Cree and English. The food was no different from what I had consumed while staying with white friends a few weeks earlier in Winnipeg. One afternoon, there was a visitor, a middle-aged man from Granville Lake, a relative by marriage who had come to visit the nursing station. He and I were equally surprised to see each other, for I had lived with his brother at Granville Lake, seven years earlier. Sakaw-iðiniw is used to refer to y-dialect speaking Crees to the west in Sakatchewan and Alberta, but it possesses another narrower and local meaning. Not all the families of the Mathias Colomb Band moved into Pukatawagan during the centrifugal 1950s. A few remained as full-time occupants of the bush hamlets. There were some at High Rock Lake, some at Burntwood Lake,

and some at Okawimiðihkananihk 'pickerel-oil-rendering-place', a backwater settlement called in English Pickerel Narrows or Granville Lake. The people of these settlements are sometimes called also Sakaw-iðiniwak or "Bush Indians," an epithet with mingled connotations of provincialism and authenticity. Such was the visitor. Food must immediately be served when a senior relative visits. Years ago, a meal of white man's food from cans expressed the giver's esteem more lavishly than moose or beaver. But today, at Pukatawagan, where most food comes as groceries from the Bay, it is "Indian food" that honors guests and is uniquely appropriate for a senior relative from "the bush." No wild game had been served in the house for weeks. But a child was dispatched to a neighbor, and a meal of boiled beaver meat was soon served to the "old man" who appreciatively consumed it as his due.

Granville Lake itself was the site between 1794 and 1796 of the Hudson's Bay Company's Granville House and a competing Canadian post. During the winter of 1804-05, George Charles and David Thompson, respectively, represented the competing companies at "Musquawegun Lake" (Thompson 1962 [1784-1812]:xc , 84). Maskwawikan Sipiy 'Bear-Backbone River' is the Laurie River, and this post was presumably at or near the confluence of the Laurie and Churchill rivers at Granville Lake. By 1900, Granville was among the winter log cabin hamlets occupied by Crees who summered at Pukatawagan and were later included in the Mathias Colomb Band. Revillon Frères operated a post at Granville at this time, and an oust of the Pukatawagan store was built in 1927. When the Colomb reserves were established in 1929, the band requested that Granville and other hamlets be surveyed as separate parcels. By the 1930s, Granville differed from most log cabin hamlets in that members of at least four hunting groups—two Métis groups and two branches of the extensive Bighetty family—resided there simultaneously. Commercial fishing developed as a supplement to hunting and trapping in the 1940s. In the 1970s, Granville was the biggest of the few persisting bush hamlets whose inhabitants had elected not to "move in" to the Pukatawagan reserve. Approximately sixty-five persons lived there, the number constantly fluctuating as men and families moved in and out from their traplines to the settlement. Residents say the community remained unknown to most government agencies until around 1969, and as late as 1977, Department of Indian Affairs and Northern Development (DIAND) personnel informed me confidently that no village existed there. In 1969, the community was "discovered in connection with the Manitoba Hydro-

electric Company's proposed Churchill Diversion Project. Manitoba Hydro's initial plans entailed raising the waters of Granville Lake thirty-five feet, thereby putting the settlement underwater and forcing relocation of the residents to the city of Thompson. The Granville Lake and Southern Indian Lake communities took the case to court, and Manitoba Hydro was eventually compelled to cut back on its plans, raising the lake approximately fifteen feet and leaving the settlement intact. During the 1970s, the village belatedly came to the attention of DIAND: a community hall, a tourist lodge, an ice rink, a garage, an electric generator, one telephone booth, and materials for new housing were introduced. In 1972, a school with grades one through six was built; previously some children went out to Guy Hill residential school at The Pas or lived with relatives and attended school at Pukatawagan. Nursing supplies were deposited with one community member who subsequently gained fame by mistakenly issuing birth control pills to male trappers.

While acquiring some of the facilities of larger reservation communities, Granville remained and remains today relatively isolated. The families resident there are characterized by their commitment to hunting, trapping, and commercial fishing; intermittent guiding for white sport fishermen provided the only wage labor in the community. Granville is inaccessible to commercial travel except by ski- or floatplane and is effectively isolated during breakup and freeze-up. In 1977, it was serviced twice a month by a scheduled mail and supply flight from Lynn Lake. The relative autonomy and the resource opportunities afforded by this isolation were dearly factors in the community's persistence. Residents said that they preferred Granville because of the low population density, the absence of band council regulations (that prohibiting drinking, for example), proximity to traplines, better hunting, trapping, and fishing, distance from the Catholic mission, and, as I was told candidly, the opportunity to interact only minimally with whites. The community is visited periodically by representatives of the Royal Canadian Mounted Police and the Oblate Mission, but during the period of study, the only non-Crees resident in the community were the schoolteacher, the free trader, and me.

Conditions of Work

Whatever else may be implicated in my decision to work in the north, I sought a community where indigenous religion was practiced rather than eulogized and one in which foraging prevailed

over wage labor. I wanted to avoid studying "memory culture." James G. E. Smith had worked with Rock Crees at Brochet and discussed possible field sites in Manitoba with me. I settled on Pukatawagan because no anthropologist had gone there before and because Smith knew that smaller bush settlements such as Granville persisted in the area. I initially wrote to the Mathias Colomb Band Council and to several persons at Granville. No one wrote back, but the letters were useful because some members of both communities appreciated that I had asked to do the research. "We were wondering when you'd show up" was a typical reaction when I finally arrived.

I went first to Pukatawagan, arriving in September 1977. With a population of 1,250 persons, Pukatawagan looked to me like a suburban subdivision dropped into the boreal forest. It had at the time a well-deserved reputation (since dramatically reversed) for social problems and violence, chiefly the consequence of a numerically insignificant but very active group of "problem" families and juveniles. I resided for two months that fall with a white schoolteacher and later with the community development officer and in the home of Luc and Caroline Dumas. I was assimilated into a sector of overlapping Cree-white social life centered around employment by the school. From the outset, I benefited from the interest and support of Paschal Bighetty, then chief of the Mathias Colomb Band, and Sidney Castel of the Manitoba Indian Brotherhood. With Mr. Castel as interpreter, I visited and interviewed a number of older individuals in the community, including Julien and Nancy Bighetty, Selazie Linklater, Solomon Colomb, and Eli Caribou. Mr. Castel was also the first of many individuals from whom I taped the words and phrases that I used to develop some speaking facility in Cree. Initially, I pursued "memory culture" by default, documenting through interviews aspects of the prereserve seasonal round. I was aware from the snowmobile traffic that trappers were continually departing for "the bush," but the people I knew worked at various jobs in town. Those who were in any way visibly "religious" were so in the idioms of Catholicism. During this period, I often inquired about Granville Lake. People were vague or discouraging: "Yeah, there's some people out there." "They're old-fashioned down there." "They don't like white people there."

My offers to bankroll a snowmobile trip downriver to Granville met with no takers. In November, I took the train from Pukatawagan to the mining town of Lynn Lake, timing my arrival to coincide with the scheduled mail flight. I arrived at Granville Lake, Manitoba, encumbered with gear, a minimal speaking knowledge of Cree, and profound

uncertainty. Granville looked like my fantasies of a "remote bush settlement." Twenty or so houses and log cabins were scattered irregularly around a point jutting out into the lake. Uncertain about what to do with me, the Ukrainian trader, Mike Chuiak, suggested I move into the "lodge," a large and not easily heated structure built some years previously by DIAND in the (unrealized) expectation that visiting sport fishermen would rent it. Over the course of the next few days, while under careful and not always unobtrusive scrutiny by many persons in the hamlet, I struck up a casual friendship with Johnny Bighetty, a Cree trapper about forty-four years old who lived with his wife, Sarah, six of their children, and Jim, a Métis from Winnipeg about my own age (twenty-seven) who was learning the trapping business from him. After further scrutiny and an approving assessment of my willingness to cut wood, haul water, snare rabbits, fish through the ice, drink neat rum, and speak (bad) Cree, Johnny invited me into his household, where began my initiation into taking food and other resources from the bush and acting enough like an Indian to get along. It was progressively an initiation into skills, manners, language, generosity, and ultimately into dreams, an education in which many others played crucial roles but in which Johnny's influence was primary. Independently inventing the doctrine we call participant-observation, he declared that anyone who was "writing a book about the Indians" should live with and like an Indian. My specific motivations for seeking this training involved the conventional objectives of increased rapport, the desire to contribute, however minimally, to the household, and the opportunity to observe and participate in hunting and trapping strategies. It is probably unnecessary to note that the manifest benefits were almost entirely mine and that the training so generously given served more to lessen my liability than to allow a positive contribution of any significance. I believe that the course of instruction recapitulated exactly the sequence employed in training small boys: I began with wood and water, progressed to fishing and rabbit snaring, and finally learned some of the rudiments of trapping.

The representations of Cree beliefs and practices that I attempt here are based on what I observed people doing on various occasions and on what they told me, either informally or in the course of discussions consecrated to the project of increasing my understanding. These latter were sometimes generously undertaken at others' initiatives, out of personal liking for me, or because what I was conceived to be doing was thought important. I was once, for example, summoned out of my

sleeping bag at midnight to attend a party at another house where the presiding grandfather had expressed a desire for my presence. More typically, I initiated relationships in which I paid people an hourly wage in exchange for talking to me about particular topics.

These sessions were, at least initially, sometimes bizarre experiences for my Cree preceptors, novel assimilations to the category of "job." No one in Granville Lake or Pukatawagan had ever heard of anthropology. When working with elderly persons, I supplemented the hourly payment with a gift of tobacco, simultaneously establishing a reputation for good manners and obligating them, in their own understanding, to tell me something worthwhile. I rapidly learned certain basic premises of Cree verbal etiquette: that personal questions are impolite, that specific questions are pushy, that long reflective silences should not be interrupted by prompting. The first time I sat down to talk with Jean-Baptiste Merasty, I asked him how the people had divided up the land before the registered trapline system. He gazed into space silently; two minutes became four and then five. I tentatively rephrased the question. Fixing me with a baleful glare, he told me sharply that he was still thinking, and I subsided.

With few exceptions, the discussions were "formal" only in the sense that I arrived with a topical agenda. With some persons and with respect to some topics—genealogy, land tenure, systems of animal classification—I used schedules of prepared questions. Although some valuable information was obtained in this way, most Crees found schedules less to their taste than informal conversation. One long-term benefactor of mine, who after the first few meetings refused to accept anything more than chewing tobacco and rabbits from me in recompense, stated flatly that I would have to pay him a "lot of money" if I wanted to continue plying him with eliciting frames. More usually, because most people disliked batteries of specific questions and because I had doubts about their usefulness, I posed topics in very general terms, and the responses to such questions produced superior material. Depending on the topic, my preceptors would formulate typifications—generalizations about the cultural practices they attributed to their fellow Crees—or, more rarely, refer to their personal experiences. Usually I taped these sessions. Some people did not like the machine; many soon insisted on it. At night, I would review the tapes, identifying the new parameters of uncertainty that each session engendered.

Crees also told me things in less structured circumstances, conversations that developed casually without the use of recording devices in the

course of visiting, working, or traveling. These conversations were, in the happiest circumstances, prompted by instances of productive and ritual practice that I observed or participated in. In such circumstances, Cree "culture" first became manifest to me as a factor in whose terms Crees experienced and acted upon animals in the contexts of daily life and as a scheme for interpreting these contexts. After hearing in an interview that Crees "used to believe" that animals regenerate after death, it interested me to hear a man point out details in the coloring of a dead fox and explain to his son that it was the same animal he had killed earlier the same winter. Since many animal carcasses are brought back to the settlement to be skinned out, butchered if relevant, and consumed, some additional rules concerning speech about animals, dietary rules, and disposal of inedible residue were observable off the trapline. Evidence for persisting and very basic disparities between Cree and Western orientations toward animals became apparent. Freshly plucked porcupine quills were identified as "not dead yet" and reverently suspended from the ceiling. A female lynx in a snare was identified as a beautiful woman of whom the trapper had dreamed.

Two obstacles followed me throughout the work. I was interested in people's personal experiences, but on many topics, direct questions of this kind are bad manners of the most palpable kind. Most frequently, I asked for and was given generalizations and typifications. In the fullness of time, some few of my preceptors selectively disclosed themselves. In some cases, I presumed on friendships to ask such questions. Second, I was interested in the human-animal relationship in all its modalities, and aspects of this relationship are inscribed in manitokiwin , the belief and practice of indigenous religion. Of Crees in the 1790s, Thompson wrote that "what[ever] other people may write as the creed of these natives, I have always found it very difficult to learn their real opinion on what may be called religious subjects. Asking them questions on this head is to no purpose, they will give the answer best adapted to avoid other questions, and please the enquirer" (1962 [1784-1812]:74-75). One of Thompson's Cree hunters explained this reticence readily enough: "You White men always laugh and treat with contempt what we have heard and learned from our fathers, and why should we expose ourself to be laughed at?" (ibid.:79). In the 1820s, Nelson (Brown and Brightman 1988:101) noted that "they [Lac la Ronge Crees] do not chuse that we become too well-informed of all their ceremonies." Things were much the same in 1977, exacerbated by approximately one hundred years of sustained Catholic condemnation.

The effects of the Oblate missionary enterprise on the indigenous religious complex have been profound. Crees say that most missionaries actively discouraged the indigenous religion. From the earliest times, missionary ire focused on sacred drums. Again and again, I was told that the missionaries forcibly confiscated drums, burning them or throwing them in the lake. Alternatively, Crees desiring baptism were required to renounce their drums and other pagan practices.

That little island there [in Pukatawagan Lake], that's where they were first baptized. Priest come in there, he threw all the drums in the lake. Those people, you know, they used to have drums . . . singing . . .

Question: Why'd he throw them in the lake?

Well, he said . . . there's no more of them [drums]. Well, that's . . . when a guy is drumming like that, that's like he's believing something else. Yeah, it's another god. He's singing against . . . [the scriptural deity].

It is not difficult to understand the reactions of the Rock Cree medicine man who told the Oblate Rossignol (1938:67) that "nothing in the missionary's sermons had surprised him, but that the white man's contempt for the pagan observances had shocked him."

Some Crees were uninterested in the new religion. Oyapacikiw 'Aimer', ancestor of the Cursiteur family at Pukatawagan, reacted to an offer of. baptism by threatening to pour water on the minister (ibid.). Most Crees at Pukatawagan accepted Catholicism and "converted" in the late 1800s, reproducing in this context an indigenous receptivity to new forms of spiritual expression. As a result, Catholicism is now Cree "tradition." This new tradition, however, was represented by the priests as incompatible with indigenous spiritual expression. As a result, many persons self-consciously renounced manitokiwin , including the personalized relationship with their dream spirits (cf. Brightman 1989a : 162-163, 167-168). Some earlier priests seem never to have doubted the factuality of the beings worshiped by Crees, seeing in them, in fact, manifestations of the scriptural Devil and his minions. Thus was the entire scope of manitokiwin redefined as satanic worship, perhaps with some Crees actively collaborating in the reinterpretation. Some Crees today talk as though they accept this equation. At the same time, there is predictably reluctance even among devout Catholics to identify honored ancestors as devil worshipers.

But I know there were some good guys in the [pre-Christian] past. There were still some good people living. Like my great grandfather there, he was one of them. And he didn't believe in this evil spirit [the wihtikow , a cannibal monster].

Maybe he did, but he still had enough reason in him that he could oppose these things. There were still some good guys. Even though the Devil was in control of most of them.

In this discourse, the indigenous distinction between good and evil spirit power qualifies the Catholic representation of all indigenous religion as evil.

Some Crees who are practicing Catholics believe that the spirit beings recognized by their pre-Catholic ancestors really exist; they differ as to whether people can still communicate with them. Many Catholics represent indigenous religion itself as something that no longer exists in the present or that exists other places but not at Pukatawagan. I do not know whether any of them really believe this. Some practicing Catholics have experiences and observe rules continuous with the indigenous religion, typically involving aspects of it that do not overtly conflict with Catholic teachings. Some consult Cree medicine specialists and hang up bear skulls, while others do not. Some persons have evolved syntheses equating the superior being Kicimanitow with the scriptural deity, the wihtikow and other monstrous beings with devils, and dream spirits with angels. Such syntheses are in unstable equilibrium with the official Catholic position that all pre-Christian beings and practices are evil and that the two religions cannot be concurrently practiced. Christian and indigenous beliefs are in consequence distributed through the communities—and through the discourses of individuals—in an exceptionally varied fashion. I did not observe the compartmentalization of Christian and indigenous practices by the same persons in different contexts (cf. Rousseau 1952, Tanner 1979:108-109). Many persons, especially those in their thirties and younger, have no visible commitment to either tradition.

Virtually everyone born in Pukatawagan or Granville is baptized as a Catholic, and the entire population has long since been nominally converted. Some persons who practice manitokiwin represent themselves as believing Catholics. Others are privately disparaging of the church and skeptical of its teachings. People were usually reluctant initially to acknowledge manitokiwin , apparently anticipating that a white would disapprove on religious grounds or disparage their beliefs as superstition. A man once told me, with a perfectly straight face, that he had never heard of Crees hanging up animal bones in trees and knew nothing about it; every tree around his shack was festooned with skulls and

antlers. Some were also concerned that they would provoke ridicule from others or that they would be suspected as sorcerers (see Tanner 1979:108), in whose existence most Rock Crees, I think, continue to believe. As a result, some persons arc secretive about their practices. However, at least one man of my acquaintance enjoyed his reputation as a sorcerer. Other factors conditioned what people were willing to tell me, especially about their own experience. Crees expect that misfortune will follow if certain aspects of personal religious experience—a particular song, or a dream, or a part of a dream, for example—should become known to others. Some aspects of such experiences—the identity of a dream spirit, for example—are inferred by others from dietary rules and other details of personal ritual but are seldom verbally communicated. Not least, some of the beings in which I was interested, the misipisiw 'underwater panther', for example, are dangerous to speak about. Those who believe are reluctant to talk about them, especially nohcimihk , "in the bush."

The shamanistic prowess of the ancestors is said to have surpassed the decadent practice of the present, and I am sure I would have been told the same thing two hundred years ago. But there has been a reduction not only of power but of belief and practice. Crees talk about manitokiwin in the past tense or as radically attenuated, and there is no reason to question their assertions. Indigenous spirituality has decisively colored at each stage the sense that Crees make of Catholicism and in this sense reproduces itself in new idioms. In another sense, however, I felt that I was presiding over an inquiry into a moribund religion. Crees say that only a minority of people participate in the shaking lodge, the eat-all feast, traditional curing, and prayer with song and drum and that the vision fast at puberty is no longer practiced at all. This coincides with my own impressions, although there was surely more going on than I knew about. Enough is still known and believed of this religion within the community to provide the basis for its reemergence in the future.

Once in Pukatawagan, I attempted to develop some control of spoken Cree, benefiting from the careful instruction of Sidney Castel and Caroline Caribou, among many others. Initially, I was limited to talking to people in English or to using interpreters to talk to non-English-speaking Crees. I eventually began speaking Cree with people in interviews and other contexts, although many of them understandably preferred to use English. The limits on my control of verb inflec-

tions and vocabulary made it useful to go over taped interviews with bilinguals. Unless they are identified as Cree, the discourses reproduced here were recorded in English.

As befits the epistemological uncertainty of the 1990s, I identify here certain caveats implicated in this and other anthropological projects. I take my representations of what people said and did as giving some evidence for elements of their inner thought and experience but readily acknowledge that this is a multiply problematic strategy. What they did and said was influenced by my presence and questions, and speech and actions are not, of course, transparent with respect to such inner states. Further, my constructions of what they did and said—both at the time and in subsequent phases of digestion—have been, of course, regimented by my own agenda. Moreover, I have attempted to identify in these representations of other people's acts and discourses characteristics of a socially shared if diversified structural configuration whose design may be in certain respects no more available to the consciousness of its participants than the phonological structure of the Cree language. To accord to the material of my observations the maximum resistance to the potentially outlandish opinions that I or other non-Indians could form about them, I here pursue the course, common enough, although not ubiquitous in such work, of attempting to situate my interpretations in the contexts of Cree discourses that initially suggested them. Throughout, I try to follow Bronislaw Malinowski's (1961:3) old but not unfashionable advice to separate "the results of direct observation and of native statements and interpretation" from my own "inferences," without sharing his optimism that the project is possible.

Animal Signs

In this book, I approach questions about Cree human-animal relations from a perspective that is semiotic in multiple respects. It begins from the premise that both technical and spiritual facets of Cree foraging are organized as sign phenomena. In a use consistent with sociolinguistic and anthropological borrowings from Charles S. Peirce, "semiotic" entails analysis of signs as contextually actualized in the practices and experiences of their users. A semiotic perspective differs in this respect from modes of analysis, seemingly moribund in anthropological discourse in the 1990s, that take such practices only as

data from which the formal elements of analytically constructed systems can be abstracted. "Semiotic" thus occurs here in preference to "symbolic," "structuralist," "hermeneutic," and other imaginable substitutes with which it varies freely, often with resonant vagueness, in anthropological usage. I draw, however, on aspects of linguistic and anthropological structuralism, specifically, the definition of linguistic signs—sound images that signify concepts—as arbitrary and oppositive entities. The purpose of the Saussurean (1966 [1915]) argument, convergent in some respects with the findings of a later American relativism, was to explore the consequences of the fact that meanings are not constant across languages and that each language effects a unique classification of conceptual substance or extralinguistic reality. The words and grammatical categories of different languages do not passively and repetitiously encode the same objective discontinuities between external objects, actions, and conditions; they are not different labels for the same independently existing "things." Taking a conventional anthropological example, the meanings of Cree nistis 'my elder brother' and nisim 'my younger sibling' derive from the play of contrasts and resemblances within the lexicon of Cree kin terms and not from their denotational relationship to genealogical kin types. Their meanings are not equivalent to those of English "brother" and "sister" or French "frère" and "soeur." Likewise, the meanings of the Cree linguistic signs pisiskiw 'animal' and iðiniw 'human' or maciw 'he hunts' and pawamiw 'he dreams' are mutually defined in uniquely Cree semiotic environments. To this degree, linguistic signs are oppositive entities, their meanings circumscribed by the meanings of other linguistic signs within semantic environments constituted differently in each language, each social universe. Ferdinand de Saussure's exposition advanced a purely differential concept of meaning on the model of phonemic contrast, the often-cited assertion that "their [signs'] most precise characteristic is in being what the others are not" (Saussure 1966 [1915]:117). As Sergej Karcevskij (1982:51) later observed, "True differentiation presupposes a simultaneous resemblance and difference" (cf. Jakobson 1978:64). Bears, for example, are not significant to Crees only because of their lack of identity with foxes and wolverines. The meaning of "bear" as a linguistic and social sign—and thus the meanings of bears in practical experience—is constituted both by difference and resemblance to other modes of animate being.

Roland Barthes (1972:42) observed of myth that it "has the task of giving an historical intention a natural justification, and making con-

tingency appear eternal." In this respect, ecological explanation in anthropology parallels in certain respects the consciousness people have of their own social practices. Of the linguistic sign, Émile Benveniste (1971:17-18) observed that sound image and concept necessarily evoke one another; their relation is arbitrary "only under the impassive gaze of Sirius" and not in the experience of users. Cultural categories and propositions are understood, probably in most contexts, as more or less transparent representations of reality and experience; put another way, signs are not recognized as such by those who think and act by means of them. Marshall Sahlins (1981:70) observes, "In naive and evidently universal human experience, signs are the names of things 'out there.'" If, as Benveniste shows, the acoustic image and concept united in the linguistic sign are inseparable and thus nonarbitrary in the speakers consciousness, this Perhaps provides the basis for a well-nigh universal folk theory that language is nomenclatural. The mediating effects of the "meaning" are not available to consciousness; People experience words as sounds that name and evoke discrete "things" in the world. In Peirce's terms (1960-1966, 2:157), vocabulary is understood as a diagrammatic icon, a sign whose distinct parts replicate and resemble the distinct parts of its object. As words are felt as names for things, so are cultural propositions felt as factual statements of independently existing states of affairs and social practices represented as realistic and perhaps inevitable accommodations to these. Inedibility is an important aspect of the meaning of "wolverine" for Crees. The social rule does not assert that it is impossible to eat wolverines (starving men are known to do so), but it embodies the "fact" that wolverine meat possesses qualities rendering it undesirable as food. Cree behavior thus enacts and re-creates a claim about wolverine edibility that is not the only one possible "under the impassive gaze of Sirius" but is assumed nonetheless by most Crees to be axiomatic. This naturalization of signs in the consciousness of their users seems to be of very general provenience. Only in privileged contexts—trickster narratives or "reversal" rituals, for example—is the arbitrary character of cultural categorizations and propositions overtly acknowledged or celebrated, and even here the social significance of folly and transgression may be conservative rather than subversive of institutional and cosmic orders (Turner 1982). Like most other people, Crees lack consciousness of external objects and of their experiences with them as the constructs of semiotic mediation by a "culture."

Saussure's exposition took as its object a privileged case of the sign:

that of the word in which a recurrent sound image stands for a recurrent concept. As such, the sign is doubly arbitrary: no necessary bond exists either between the sound image and the concept or between the concept and whatever segment of extralinguistic experience it picks out and categorizes. The second of these relationships has more general implications for the analysis of signs in social life. Any occurrence of a social practice—setting bear traps, for example—implicates multiple constituent objects, actions, processes, and conditions: the trapper's sense of himself as a human agent and as an individual, travel, habitat, weather, bears, trap, bait, setting the traps, and so forth. Some of these elements are lexicalized in Cree, while others are not; all can presumably be referred to at clausal levels of discourse. From the perspective taken here, all of these elements are signs, although they lack the conventionalized referential symbolism of words. Each of them, whether lexicalized or not, is discriminated from semantically related elements by relations of resemblance and contrast. And, like lexical concepts, the elements are arbitrary, not in relation to each other but in their articulation with the coordinates of human biology and the boreal ecosystem. Anthropological writing has overwhelmingly emphasized cultural differences in the classification of tangible objects: colors, animals, plants, and relatives form the usual subject matter. Categories of event and action are equally elements of structure.

I am concerned here with the semiotic structure of Cree animal-human relations at the levels of "structure," of consciously articulated discourses, and of events of practical action in which intentions and dispositions are implicated. In all these modalities, the relations to animals as use-values and social others are organized by socially shared categories, lexicalized or not, of object, action, process, and event which define and circumscribe the world in which Crees understand themselves to be situated. In two of their modalities, cultural signs correspond to the Saussurean distinctions of langue and parole (Sahlins 1981:68-70). As constituents of structures or systems, the semiotic elements of daily life are abstractions from particular people's discourses and practices, postulated as corresponding to the knowledge, both tacit and conscious, that makes coordinated discourse and practice possible. As a linguistic sign, maskwa 'bear' possesses stable meaning as a semantic "type" which it retains from one occasion of its use in speech to the next. And so with more complex cultural signs. As a sign in Cree society, Cree maskwa 'bear' is not identical to or a sensory replica of any particular worldly bear that Crees may talk about or interact with. It is,

rather, an a priori construct that defines what a bear is and does and to which each worldly experience with bears is oriented and assimilated. Similarly, categories of an event in the world—encountering a bear or setting bear traps or dreaming of a bear—possess recurrent meanings as types of experiences the inform the intentionalities and interpretations of their particular token occurrences as unique events.

In another respect, individual bears and particular events of bear trapping are themselves signs. Crees experience and act both ceremonially and productively on the boreal fauna not as objective qualifies but as entities mediated by social meanings. Linguistic signs exist both as types and as tokens in contextualized situations of human speech. On particular occasions of use in speech, "token" occurrences of maskwa , by virtue of the stable association of sound image and concept, can refer to a particular bear in the field of vision, or a bear encountered earlier, or a fictional bear, or the entire class of bears, or, metaphorically, a penis or a menstruating woman. The circumstances of the individual utterance contribute to each token occurrence of the word a residue of meaning that is never precisely equivalent to the context-independent type. So also with events not only of reference to but of interaction with bears. Bears are understood by many Crees as exceptional animals possessing intelligence equaling or exceeding that of humans. It is said that bears, for example, understand spoken Cree, a competence not conventionally generalized to other animals. Consequently, the behavior of bears is likely to be interpreted by Crees as manifesting this intelligence in ways that are not salient to non-Cree observers. Crees are likelier than whites to ascribe a studied deliberateness to the doings of bears. Anything that bears do is likely to be apprehended as a particular token occurrence of a conventional type of ursine acumen. For example,

B. L. cached his clothing and food on a rack for three days before going out to lift his traps. He was worried about bears or wolverines getting into his stuff, so he barked the upright poles on the rack and rubbed them real smooth so no animal could climb up there. When he returned, he was dismayed to find that a bear had been there during his absence. Although the bear couldn't get at the food because it was in the middle of the platform, the bear hooked his claws up over the edge and pulled down all of B.'s clothes. All of his clothes were ripped up into shreds except one new suit of long underwear from the Bay. This suit was spread out on the ground with arms and legs outstretched. It had been carefully arranged. And right on the seat of the underwear, the bear had left a large pile of scats. B. said, "What next?" He said that the bear had gotten mad about the food and gotten even with him on purpose. He said he'd heard about smart bears but he'd never seen one that smart.

This anecdote amused both the Crees and the whites who heard it—but for different reasons. As white humor, the story presents a bear who performs by chance a scatological act suggestive of malicious purpose. The Crees, in contrast, saw nothing accidental in the bears behavior, the humor deriving precisely from malicious intention. Encounters such as these are interpreted in terms of the conventional expectations about ursine potentialities, and each encounter re-creates the expectations. Each context of interaction contains, however, a residue of emergent significance contributed by the circumstances to which the conventions refer. The fate of B.'s underwear was a novel manifestation of ursine intelligence at the same time that it was a culturally conventional one.

The symbolic character of Cree human-animal interactions, both as types and as occurrences, is not exhausted by the reciprocally defining relations of resemblance and difference that exist within and between the two categories. Or, phrased a different way, the constitution of these categories as separate (or as continuous) has everything to do with propositions about the interactions understood to exist between them. Crees typically explain their own foraging practices, including the most esoteric of them, as instrumental procedures that secure desired objectives by taking realistic account of the objective characteristics of animals (and of the Canadian market). From a Western perspective, the technical procedures for killing bears do exactly this. Crees understand maskwa to be a name for something else: bears. Bears them-salves and the events of hunting, killing, retrieving, butchering, eating, and ritualizing them, do not as their primary function refer to other things in this way, although bears and interactions with them can, of course, stand for many things. Bears are significant to Crees as omnivorous scavengers, hibernators, fierce fighters, animals whose meat can satisfy the need to eat, owners of marketable hides, and sources of spiritual power, and in many other respects. An event of catching a bear in a trap may stand for the intrinsic value of trapping as a profession, the trappers technical expertise, the satisfaction of needs for meat and for goods obtained by selling the hide, the opportunity to display generosity by giving away meat, the worldly consequence of a predestining dream, the spiritual "power" of the trapper, and many other possibilities. The categories of animals that Crees understand to exist and the categories of human-animal interaction that they understand themselves to undertake are reciprocally defining and motivating. To trap bears, a man or woman employs different strategies than to trap rabbits;

likewise in more esoteric spheres of interaction where the lavish ritual attention accorded to the bear contrasts with the neglect of the rabbit. The existence of particular modes of interaction is based on the understood character of the animal; reciprocally, the character of the animal is built up in terms of the role it plays in human social life.

As signs in Cree society, humans, animals, and categories of interaction between them are organized by sets of propositions that are themselves complex signs. The existence and meaning of each animal-as-sign is based on overlapping dimensions of resemblance and difference with the others. In an equally basic way, each animal sign is constructed from the aggregate social propositions delineating its instrumental and moral relationships and interactions with the makers and users of the signs. Bears are edible, valuable, powerful, and dangerous. All these attributes encode propositions constitutive of the cultural meaning of "bear," presupposed and re-created in each context of interaction between humans and bears. As Sahlins (1976:64) has shown, cultural propositions about entities may be as arbitrary as the principles of their classification. It is, for example, to some degree arbitrary that Crees distinguish bears from wolverines since societies and languages are imaginable which would not do so. Even more arbitrary is the definition of wolverines—and bears by some Cree individuals—as inedible.

The second part of the book introduces a second sense of "semiotic," one that contrasts not with decontextualized formalism but with the by no means homogeneous congeries of ecological, sociobiological, and Marxian theories that represent social forms as determined by ecosystem variables, human biology, and the labor process. In viewing the labor process and the ecosystem in which Crees themselves participate as semiotic constructs, I suggest that Cree hunting strategies are not now and have never been in the past determined by material forces. The relevant ecosystem is itself a social construction, albeit one with maximal "resistance." The foraging strategies historically attested are only a subset of the possible adjustments that would perpetuate Crees as a biological population in the boreal forest. Other facets of social practice are, in turn, relatively arbitrary with respect to the labor process itself. As the multiple lexical segmentations in different languages demonstrate the arbitrariness of words to external discontinuities, so the multiple possible designs for society in any historical context demonstrate the relative arbitrariness of social forms with respect to their material parameters.

The questioning of such "determination" as a totalizing theory of

forager (or other) societies does not, however, exclude consideration of material coordinates. In qualifying Saussure's doctrine of arbitrariness, Paul Friedrich posed the question of "the degree of congruity between the symbolic system and the external one—particularly the degree to which the former makes use of the latter; material reality may be seen as in part a sort of resource for symbolic reality" (1979:23). Social practices must, of course, comply with material constraints as a necessary precondition for the biological reproduction of the practitioners. The Western Woods Crees survived before and after European contact in a boreal ecosystem characterized by extreme fluctuations in animal biomass and limited access to the botanical foods that elsewhere underwrite the (now well-debated) affluence of tropical foragers. References to famine and death by starvation occur in the earliest documentary references to Crees. It is not surprising, therefore, that Crees act upon animals in terms of market principles (e.g., the price of lynx pelts) and characteristics relevant to effective search and capture. Considerations of habitat, density, aggregation, mobility, and biomass enter into \ animal definitions and classifications, conditioning decisions about hunting. The superiority of a semiotic approach, in contrast to those that would postulate the relevant material properties of resources in advance, derives from its precision in identifying the relevant material parameters of which the foragers take account. Simultaneously, a semiotics of foraging can specify other dimensions of meaning—culturally specific assessments of edibility or availability, for example—that may be of coordinate importance. No foragers forage indiscriminately, none know or act upon all the characteristics of proximal fauna potentially relevant to the provisioning of human populations. Diet breadths and foraging strategies do not passively encode universally recognized material properties. There is always "selection" between alternative possible designs for provisioning society.

The Cree signs iðiniw 'human' and pisiskiw 'animal' and the many signs discriminating modes of interaction between them cannot readily be arrayed as discretely bounded categories in sets of logically inter-related propositions. Neither would such an exposition exhaust the interest of the subject. Much anthropological writing identified as "symbolic" or "structuralist" has been characterized with some justification as unrealistic in its formalism, antisemiotic in its inattention to practice, power, context, history, emotion, contradiction, and human purpose (cf. Ortner 1984). In seeking to make sense of how Crees both talk about and enact human-animal relationships, I proceed by assuming a

dialectic, unstable both synchronically and in short and long durées , between a "structure" of reciprocally defining signs and their predicated interrelations, enactments of these signs in social practices, and the actors' articulated consciousness of their own knowledge and practices. None of these three facets is internally systemic; none is autonomous from or a replication of any of the others. The Cree human-animal relationship is disordered both in the instability of the two categories and in the irreconcilable propositions about their interactions. This is not, I think, a historically engendered chaos arising from the confrontation of the magic animals of Cree manitokiwin with their mute and soulless counterparts in Western biology and scripture. It is probably prehistoric and echoed in the consciousness of other foragers. The animals are endlessly regenerated, and yet they are finite. I am more powerful than the animal because I kill and eat it. The animal is more powerful than I because it can elude me and cause me to starve. The animal is my benefactor and friend. The animal is my victim and adversary. The animal is different from me, and yet it is like me, as much like me as its ancestors were in the earliest time of the world.