8
"You Got to Eat the Whole Works"

An Eat-All Feast at Watt Lake, 1977

In December 1977, together with his partner Rick, I joined Edouard C. on his trapline. We stayed for four nights in a log cabin that Edouard had built and maintained for some years. During the days, Edouard and Rick set beaver snares through the ice and a variety of snares, leg-hold and conibear traps for lynx, fox, marten, and otter. On the third day, in the early afternoon, we crossed the snow-mobile trail of Edouard's relative, Marcel, whose land adjoined to the west. We followed the skiddoo trail into Marcel's camp, two cabins, a lodge, and a canvas tent on a bluff on the shores of Watt Lake.

Marcel was one of the oldest active trappers then resident at Granville. Like Edouard, he alternated residence in Granville with trips to the trapline; unlike him, he was inclined to remain out in his cabin on the line for several weeks at a time. With Marcel were his wife, Pelagie, his father, Antoine, his son, Lawrence, and his son's wife, Pauline. During the course of our visit, Marcel invited Edouard to come to the house again on the afternoon of the following day. The conversation was in Cree, and at that stage, my comprehension was minimal. I heard the now familiar wimistikosiwak 'white people' but was uncertain of the import. I later surmised that the appropriateness of Rick's and my presence was being debated. Antoine was probably instrumental in the decision that we should return with Edouard the following day. I had given the old man rabbits at Granville, an insignificant gift that nonetheless inspired an improvised "grandfather"/"grandson" rela-

tionship. That night, Edouard said, "Marcel is going to do a whole set of beavers. We'll eat good tomorrow." I subsequently learned that Marcel had experienced a dream some weeks earlier which prompted him to trap out a large beaver colony he had built up over several years and to sponsor a feast with the resulting surplus of meat. We had been invited to wihkohtowin , a feast of game at which a large surplus is prepared and entirely consumed. I had read and written of these feasts as practiced by Eastern Crees but was unaware that they occurred in Manitoba.

The next day, we reached Marcel's camp in the growing darkness of late afternoon. Marcel's cabin faced southeast toward Watt Lake. The door was set slightly off center toward the right and was sheltered by a plywood vestibule where stove wood and equipment were stored. Considerable effort had gone into cleaning and straightening the cabin. The previous day's collection of clothing, dirty dishes, snowshoes, shell boxes, and other paraphernalia had been neatly stowed or hung on pegs along the wall. The floor had been swept, and several chairs and makeshift seats (an upended pail, an unsplit log of stove wood, the seat from a skiddoo, grub boxes) had been arranged around the walls. Like the cabin, the appearance of the participants reflected formality. Marcel and Lawrence had shaved, and Marcel wore a necktie.

Opposite the door along the back wall of the cabin was a table made of a large sheet of plywood nailed onto two sawhorses. Over this table a clean plastic tablecloth had now been laid and the surface covered with plates, knives and forks, cups, pails, and metal pans. Immediately to the left of the table along the back wall was a cast-iron wood stove that had years ago been brought into the bush, no doubt with considerable effort, from the old Hudson's Bay Company outpost at Granville. On the stove, two large kettles contained four of the beavers that were being served; these had been cut into pieces and put to boil. Another large kettle contained an enormous quantity of macaroni and cheese. Outside the cabin, two more beavers, together with the tails of the four that were boiled, were being roasted on ponasks over a fire in a small cooking tent. Antoine, rather than either of the two women present, was supervising the cooking of the beavers and macaroni. Shortly after we arrived, however, Pelagie and Pauline began mixing and baking bannocks in the oven of the stove.

There were, as usual, no greetings, but small talk about the weather began after we seated ourselves along the wall to the right of the table. Edouard helped himself to tea, and Rick and I followed his example. Antoine Periodically went outside to check on the roasting meat. The

roasting beavers and the tails were brought in about a half hour later and placed in pans on the table. Marcel and his father busied themselves with cutting up the beavers. Edouard suggested that if Rick or I wanted to go outside, we should do it now because it would not be good manners to leave after we started eating. When Antoine came in the second time from the cooking tipi outside, he carefully closed and secured with a wedge the outside door to the vestibule before closing the inner door. The quantity of food that finally appeared on the table was formidable. The beavers were eventually arranged in five containers. The first pan contained the heads. A second very large kettle contained the segments of the boiled beavers. The roasted beavers were cut and placed in a large metal pan. The beaver tails were cut into pieces, the largest of which was a section of the bone with meat adhering to it, and placed in a fourth tray. Another pan contained sections of beaver fat that had been either roasted separately or separated out while Marcel and his father prepared the meat at the table. The broth from the kettle in which the beavers were boiled was placed in a smaller kettle. Bannocks were piled on a plate, and the macaroni kettle remained on the stove.

When the food was ready to be served, Marcel took a chair in front of the serving table and Antoine seated himself to his right in front of the stove, both men aligning their backs to the rear wall of the cabin. Lawrence sat to their right, his back to the left wall (facing in) of the cabin, facing Edouard, Rick, and me, similarly seated with our backs to the right wall. Edouard sat farthest to the back, followed by me, with Kick closest to the door. The two women sat with their backs to the front wall of the cabin, facing Marcel and Antoine. At this point, Marcel decided that it would be better to position some of the food in the center of the room between us and sent his son outside for another sheet of plywood that was soon propped up on pails. The cloth was removed from the table and placed over the impromptu serving board, which was about two feet off the floor. The kettles containing the beaver heads and the soup were placed on the table.

Marcel rose and began speaking in Cree. As I later learned, he said some words to the effect that he was pleased to have his family present and thanked his father for preparing the feast food. He also thanked his wife for helping to prepare the cabin. He asked that we try to help him eat all the food that had been prepared. He paused and then, speaking much more quietly, expressed his thanks for or to the beavers and all the other animals that he had trapped that fall. He then cut small pieces from each of the kettles and put them into a plate with small portions of

macaroni and bannock. He opened the lid of the stove and silently emptied the pan into the fire.

Marcel then took the kettle containing the "soup" in which the beavers were boiled and gave it to his father who drank from it and passed it to Lawrence; he drank, and the kettle then passed successively to Edouard, me, Rick, Pelagie, and Pauline. While it went around, Marcel prepared plates for each of us, serving successively Antoine, Edouard, Lawrence, me, Rick, Pauline, and Pelagie. Of the six beaver heads, Marcel and Lawrence each ate one, and two each went to Antoine and Edouard. Rick and I were given one or two of the disarticulated lower jaws. The women were not served meat from the head or the pieces of the anterior quarters of the beaver, including the forelegs. Otherwise, everyone received a portion of each kind of food available. Tea was served with the meal.

Conversation went on uninterruptedly throughout the meal, Edouard and Marcel switching from Cree to English for the benefit of Rick and me, translating a long and funny story told by Antoine about someone who got the shits from a wihkohtowin . Whenever anyone emptied their plate or pan, Marcel or Antoine retrieved it, piled it high with meat, bannock, macaroni, and fat, and returned it. Periodically, pans and vessels containing lean beaver roasted and boiled, roasted fat, beaver tail, "soup," and macaroni were passed. These always went around the room in a clockwise direction, beginning with Marcel and winding up with Marcel's father. I rapidly became full but continued to eat at a measured pace while Antoine looked on approvingly. Antoine instructed me through Edouard to "eat good so you'll make them be lucky." Pauline, who managed to look both respectful and as though she would rather be somewhere else, was the least enthusiastic eater present. At several points, people went outside, carefully closing the inner door behind them before opening the vestibule. I ate five plates of beaver meat, fat, macaroni, and bannock and ended the feast feeling nauseous.

Not all of the boiled beaver meat was or seemingly could be consumed, although I believe more than three-fourths of it was. Marcel and Antoine collected all the plates and other containers and busied themselves wiping or licking them clean. The bones were collected and emptied into the fire in the metal stove together with the remaining beaver meat and the cloths used to clean the dishes and cooking vessels. The skulls and larger bones were boiled in a kettle to remove the remaining meat and then fished out with a stick and wrapped in a clean

rag for later deposition on trees. The kettle in which they were boiled was then covered and set off on the back of the stove. The women stowed away the cooking gear and the remains of the macaroni and bannock and then served more tea and also jelly doughnuts, of which everyone ate at least one. After the cleanup, Marcel extracted a suitcase from under the bed and removed a single-headed drum perhaps two inches in depth and two feet in diameter. The drum was given to Antoine who held it vertically in front of him, drumming with a carved stick, and singing so softly that I could not have understood him even if the words had been English. The drum passed successively to Marcel and to Edouard, each of whom sang softly while the others smoked and talked. Although not all the food had been eaten, Marcel's wihkohtowin was unquestionably successful in the most important of respects: except for the skulls and other bones, not a trace of the animals taken from the beaver lodge in the preceding days remained. All had gone into our stomachs or into the fire; no meat remained in the camp.

Historical Background

Every inclination of theirs is brutal; they are naturally gluttons, knowing no other beatitude in life than eating and drinking. Their brutality is remarked even in their games and diversions which are always preceded and followed by feasts. There are farewell feasts, complimentary feasts, war, peace, death, health, and marriage feasts. In their banquets they pass days and nights, especially when they make feasts they call "eat-all" for no one is permitted to leave until he has swallowed everything. (LeClercq 1881[1691]:222)

This uncharitable testimonial, set down by the Recollect missionary LeCaron in 1618, is perhaps the earliest reference to the Algonquian ceremonial game feast of the type I chanced to experience at Watt Lake in 1977. LeCaron here described the Laurentian Montagnais near Tadoussac, and it is among the groups referred to today as Montagnais and Eastern Cree that the practice has been most extensively described (Turner 1894; Speck 1935a ; Henriksen 1973; Rogers 1963; Tanner 1979). There exist, however, scattered earlier references to the practice among Western Woods Crees in Manitoba and Saskatchewan. The earliest source is Bacqueville de La Potherie (1931:231-232), who described the Crees trading at York Factory between 1697 and 1713 and referred briefly to a mortuary feast at which the guests were required to

consume all the prepared food. Isham (1949 [1743-1749]:76) wrote of the Crees trading into York Factory in the 1740s:

They have particular Days they make feasts of which is at a time when several tribes meets together, at such a time, one treats another tell all their provender is gone, Eating from morning to night,—and itt's to be observ'd he who Keeps the feast obliges Every one to Eat what is alotted him, and not to make waste, or to give any of his Companians any.

Graham (1969 [1767-1791]:165-166) described eat-all feasts as practiced at York Factory and Churchill in the late 1700s. Married men and women with children received invitation sticks from the sponsor, while others attended without such formalities; everyone arrived carefully groomed and dressed. The seating arrangement was circular, men, women, and the children and unmarried women successively forming three concentric circles outward from the central fire. The host provided a speech of welcome, while a younger hunter performed the office of server. "Every person must eat what he gets, none is to remain or be carried away." Overfull feasters sometimes hired others to consume the remains of their portion. The feast concluded with dancing and singing. Drage's (1968 [1748-49], 1:219-220) observations on York Factory Crees in 1746-1749 indicate that collective feasts in general were ceremonial affairs even when the leftover food could be carried away by the guests and eaten later. Like eat-all feasts, these occasions were characterized by formal invitations, serving procedures, and seating arrangements; Drage observed also a second seating for women after the men had eaten and the performance of dances and hunting songs. Daniel Harmon (1905 [1821]:313) observed the eat-all feast among Crees in the first decades of the nineteenth century, also distinguishing it from those in which the guests removed the surplus.

Richardson (Franklin 1823:122-123) provided a detailed description of a feast at Cumberland House in 1819 at which he was evidently a participant. The feast was held to celebrate the coming of spring and took place in a specially constructed oblong lodge with the door facing the west. The interior of the lodge was cleaned, supplied with three equidistant fires, and ornamented with wooden effigies of the spirit "Kepoochikawn" and other beings which were oriented to face the doorway. In contrast to some Rock Cree eat-all feasts and to earlier Eastern Cree practice, the door was intentionally left open and unobstructed. Women were categorically excluded. Guests were invited by a "slave" and attended in formal dress, the older men being seated toward the rear of the lodge and the younger toward the door. The sponsor's

speech was a prayer for successful hunting and included "an invocation to all the animals in the land." The feast of marrow, pemmican, and berries was served to the guests by one of the younger men. "This was done in new dishes of birch bark and the utmost diligence was displayed in emptying them, it being considered extremely improper in a man to leave any part of that which is placed before him on such occasions" (ibid.:123). The feast concluded with smoking, singing, and dancing.

Nelson also attended feasts of this kind among the Crees with whom he traded in the early 1800s and left a synthetic description, emphasizing once again the obligations of the guest to consume all that was given to him:

Some of them [feasts] arc very grand and ceremoni[o]us: the tit bits of the animal only, as the head, heart, and liver, tongue and paws, when of a Bear: It is only the Great men that arc allowed to cat of these: others again, besides the above, the brisket, rump, and ribbs and very seldom a woman is allowed to partake of them, particularly if it is un festin à tout manger , i.e., to cat the whole; tho' there may be sufficient for 2 or 3 times the number of Guests, all must be eaten before day; though in certain cases the Feaster is obliged, and commonly does take part back, providing a knife, a bit of tobacco, or something else attend with the dish. (Brown and Brightman 1988:100-101)

Nelson availed himself of this expedient on at least one occasion:

The Feaster was uneasy and said he would have been proud had we eaten all, for in that case his dreamed would have been propitious: we were obliged to dance also; but when I could stan[d] no more I gave him my knife and a bit of Tobacco and walked off leaving him to settle with his God as well as he could. (ibid.:101)

The eat-all feast was and continues to be a Maussian "totality," simultaneously implicating economic, political, social, aesthetic, and religious principles. Within it are also aggregated the most disparate intentions and dispositions, both among humans and between them and animals.

A Composite Wihkohtowin

Raconteurs at Pukatawagan and Granville Lake recalled eat-all feasts more elaborate than the one I attended and provided the information on which the composite account is based. Characteristical-

ly, there were major discrepancies between different accounts. I use the present tense below, although Crees say that few eat-all feasts are sponsored today. Some Crees consider the feast to be anti-Catholic, and it is also associated with sorcery. Several persons were visibly surprised when told that I had attended such an event.

Crees use the word wihkohtowin to refer to the eat-all feast and to other collective meals associated with indigenous religion. The verbs wihkohtow 'someone makes an eat-all feast' and wihkohtohiw 'make an eat-all feast for someone' also occur, the latter with connotations of hostile competition. Eat-all feasts are distinguished from makosiwin , a general term for collective meals shared by two or more family units. Such events are prompted by any number of circumstances, occur both in the bush and on the reservation, and may involve two families or the entire reservation community: "makosiwin , that's 'party'. Like Christmas. When you feed people in a big bunch. That's what it means, i-makosit ['he/she gives a party-feast'], when you're putting on a party. Another word is i-yisakit . The real word is i-makosit , you're putting on a party. You're feeding people. Lots of food. A few drinks." Birthdays, religious holidays, reunions, and marriages are the principal contexts for makosiwin .

Wihkohtowin , in contrast, is given in response to dreams, to express the sponsor's gratitude for some good fortune, to celebrate a boy's first kill of a moose or caribou, when an abundant kill is made, to mark the first kill of a species in the fall, often when a bear is killed, and to inaugurate the beginning of the fall and spring seasons. Edward Rogers (1963:65) and Tanner (1979:163) discuss similar prompting circumstances in Eastern Cree communities. In the past, eat-all feasts were typically initiated by the senior active man in a winter hunting group, sometimes by two or more such men. The winter group often included two or more families, in which case preparations involved cooperation between the family of the sponsor and the other co-resident families. Efforts were also made to extend invitations to other winter bands of the same totim , or bilateral kindred, as the sponsor, or, more generally, to anyone who could attend. During the period of summer aggregation at Pukatawagan, invitations were extended selectively, but people say that few eat-all feasts were held at this time. Today, I am told, eat-all feasts occur exclusively in trapline camps or in bush hamlets.

Most feasts in living memory were held either in log cabins, in the traditional three-pole lodge, or in rectangular tents of Euro-Canadian

derivation. One individual stated that during his childhood, feasts were held in the sapohtotawan , an exceptionally long and narrow lodge built specifically for the purpose. This form is cognate with Eastern Cree sapatowan , (Speck 1935a :103), which referred both to the feast and to the large lodge in which it was held and which might also be a multifamily dwelling. Prior to the feast, the vicinity of the dwelling is cleaned of refuse and dog excrement, and the feast lodge itself is cleaned and put in order.

Through rules regulating eligibility to attend, seating arrangements, priority in serving, and quality of food served, the eat-all feast sub-categorizes its participants by vectors of gender, age, and marital status. Some Crees recall eat-all feasts at which adult males were the only participants but state that both women and children are usually present. In the more recent past, no feasts are recalled from which either women or children were categorically excluded. Adolescent and menstruating women, it is said, "wouldn't go." During the Eastern Cree caribou feast described by Henriksen (1978:36-37), women and children entered to eat at two specified intervals, remaining outside for most of the meal. A more complex system was observed by LeJeune (1897a :217-219, 279-293) among the Laurentian Montagnais in the 1630s. Eat-all feasts and feasts in general were primarily male affairs; women and children were generally excluded, although widows attended non-eat-alls. In the case of a bear feast, widows and married women with children ate at a preliminary sitting after which the men convened to consume the rest. Adolescent, unmarried, and childless women were required to leave the camp entirely. The significance of marital and parental status is also indicated by Graham's (1969 [1767-1791]:165) description of Cree feasts near York Factory: formal invitations were extended only to couples with children, and the unmarried women and children formed a third circle around the married women who sat in turn in back of their husbands.

Preparation of the feast food normally occurs prior to the arrival of guests, contrary to the practice in a Naskapi feast where the food was prepared by the participants during the course of the ritual (Henriksen 1973). Bears, beavers, moose, caribou, lynxes, and geese are the focal foods consumed at eat-all feasts. Bannock, bread, macaroni, and sweet bakery goods figure as supplementary dishes, while tea and occasionally home brew are the customary beverages. Sometimes feasts include meat from only one species, often a single animal. Meat is served boiled or roasted. Fat is served in the form of liquid grease, hardened slabs, and

roasted pieces. The tibia of moose, caribou, and bears are crushed, the marrow extracted, and the bone splinters and ends boiled to yield marrow fat. The head, legs, and some organs of the animal are served separately (cf. Tanner 1979, Henriksen 1973:36-37).

The feast formally begins with the entrance of the guests. Those attending wash and wear clean clothing if available. In some cases, they bring plates, cups, and eating utensils with them to the feast dwelling. Crees describe different seating arrangements. Seating may be in a circle or a rectangle with the participants facing inward. In one account, women and children form a second circle around the men, while in another they sit along the walls of the front of the cabin while the men occupy the rear opposite the door. The latter arrangement corresponds most closely to the Watt Lake seating where women faced the rear of the house and the two oldest men faced the door. In male-only feasts, the same correlation of the rear and front of the cabin with notional or actual higher and lower status is reproduced in a separation of senior from junior men, as in Richardson's desciption. More prosaically, most accounts describe family groups of men, women, and children seated in dusters next to one another. Some people simply say that "people sit wherever they want."

Before the food is served, the sponsor addresses the guests briefly, thanking them for coming and instructing them that all the food has to be consumed. In some instances, the sponsor also addresses a discourse to one or more ahcak beings, requesting that hunting and trapping success and other benefits be forthcoming. Whether or not these beings are formally addressed, portions of each of the prepared foods are burned in the stove. The recipients are addressed as "grandfather."

The cooked foods are placed in kettles, pans, or other containers in front of the "owner" and server. Food is given to seated guests by the server or placed on plates that are then handed around the circle to the guests. In some cases, a container of food is itself handed around, each individual removing a portion and passing it on. Food travels clockwise, a direction intended to replicate movement from north to east to south to west. The order of serving sometimes interferes with the clockwise movement of food. The status distinctions regulating attendance are reproduced also in the order of serving. In cases where there is insufficient space to seat everyone simultaneously, older men eat before younger men and adult men before women and children. When all are seated together, the same distinctions organize the order in which indi-

viduals are served. Dietary rules to which women arc subject reproduce those conventionally observed. Compliance with such rules is said to be more punctilious at feasts than at secular meals, and certain meats ordinarily eaten by women perhaps shift into the male-only category at feasts. At the same time that the feast subcategorizes the participants, it figuratively unites them through a kind of hypercommensality. It is said that the feast will not secure positive effects unless each guest eats part of each variety of food for which he or she is eligible.

In some eat-all feasts, including the feast at Watt Lake which I attended, the entrance to the lodge or house is closed, and after this point the lodge is ideally kept sealed. The sponsor in such cases in-strum the guests not to enter or leave the lodge. In the event that a participant needs to leave to defecate or vomit, the prepared foods are covered with cloths or placed temporarily in sealed containers. This aspect of the feast is not elsewhere mentioned in connection with Western Woods Crees but is amply documented among the Eastern Crees prior to the middle of the present century. Clouston's account (Davies 1963:36-37) of an Eastern Cree eat-all feast makes dear that the food was carefully covered each time the lodge was opened. The caribou marrow and bear meat were covered with clean skins while the guests entered. "Every person was then ordered to keep in the tent and all the holes on the tent carefully closed up." Later, when pans were brought from outside, the food was again temporarily covered.

In wihkohtowin , meat cannot be removed from the cabin or lodge. Ideally, all is consumed, but if this proves impossible, it is kept in the kettles or pans and guests can return later to eat it. If food still remains, it is then burned. Again, this practice is documented among the Quebec groups. Clouston observed, "None of the feast durst be taken out of the tent" (ibid.). A faux pas by Lucien Turner at an Ungava band feast in the 1880s gives additional information:

I soon departed, and attempted to take the remnant of the pemmican with me. This was instantly forbidden, and information given me that by so doing I should cause all the deer to desert the vicinity and thus make the people starve. (Turner 1894:323)

In the recent past, Eastern Crees have replaced this rule with the requirement that uneaten food be carefully wrapped before guests remove it from the feast lodge or cabin. More generally, in the Quebec-Labrador area, all cooked food carried outside dwellings is ideally

covered (Rogers 1963:67, Cooper 1934:34, Tanner 1975:298). The idea that the food should remain in the lodge is strongly expressed by Rock Crees.

It's dirty to take that kind of food around [outside]. Makes it get dirty. So the old guy [the feast giver] tells them they gotta' eat all of it. Eat the whole thing up.

Q: What do they do if there's too much food to eat?

You can't take the food out of the house. Lose your luck. Maybe get nothing to eat any more. Where they make it, I suppose [?] So they can bum it. But you got . . . you're supposed to eat the whole thing.

The careful disposal of uneaten food, bones, crumbs, and other residue at the Watt Lake "eat-all" parallels Eastern Cree practice. Tanner's (1979:166) statement about the Mistassini, that "it is as if every morsel must be accounted for," is apt. Like other large bones from game animals, the bones of animals eaten in wihkohtowin are either burned, placed into the water, or, especially in the case of skulls and antlers, suspended in trees. Smaller bones and even very minor residue like crumbs are burned in the stove. This concern with residue extended even to the cloths used to clean the dishes and vessels used in cooking and eating; these also were put into the fire. The singing of individual hunting songs with which the Watt Lake eat-all feast concluded is seemingly the attenuated form of more elaborate musical celebrations that formerly involved collective dances.

The Feast as Gift

Of the ceremonial meals of the Crees at Lac la Ronge in the 1820s, Nelson wrote,

We denominate these Feasts, and from their own Term it would seem they so mean; but I consider this again as a premature interpretation which I have not leisure to explain: I consider them rather as sacrifices —indeed they may perhaps rather be esteemed as partaking of both. (Brown and Brightman 1988:100)

Wihkohtowin is first of all a sacrifice and partakes of the sacrificial logic delineated by Henri Hubert and Marcel Mauss (1964:97): "This procedure consists in establishing a means of communication between the sacred and profane worlds through the mediation of a victim, that is, of a thing that in the course of the ceremony is destroyed." The significance

of the victim in this scheme derives from its successive identification both with the sacrificers and the deity to whom it is given over. Because of its attachment to both, the animal victim serves as a conduit, effecting through its ingestion a communion with the deity and simultaneously committing the latter to a delayed obligation, "for the gods who give and repay are there to give something great in exchange for something small" (ibid.:15).

Already identified with animals, the meat, as a gift from humans to nonhumans, acquires further significance through its symbolic identification with the humans who have become the owners of it. Notionally, the meat becomes the "property" of humans when the animal is killed. Since the meat is typically interpreted as a gift from the animal to its hunters, the sacrifice is directly reciprocative, a sharing of the food with the beings from which it came. The identification of the meat with the sacrificers is effected both by the cooking process that transforms it into a cultural product and by its ritual manipulation inside the feast cabin.

The cabin or other site becomes a space in which facets of human society are celebrated and reified. Specifically, the feast is a carefully orchestrated sequence of symbolic actions whose referents are drawn from different domains of Cree political, economic, and social practice. These actions symbolize human society or, more specifically, nihiðawiwin , the particular modality of social existence in which Crees understand themselves to participate. It is in this respect surely significant that the eat-all feast and such other visible manifestations of manitokiwin as drumming and the shaking lodge have—like Cree marriage practices, clothing, hairstyles, and language—been the subject of a sustained cultural critique by Euro-Canadians whose effect has been the identical one of objectification. The lodge is cleaned, and the participants wear clean clothes. The feast is the site of elaborate cooking procedures and of aesthetic performance: song, drumming, and, in the past, dances. The seating arrangements, the division of the meat, and the order in which it is served symbolize and presuppose the status distinctions organizing relations between men and women, seniors and juniors, unmarried and married, parents and the childless. At the same time that the feast is a sustained diacritic of status differentials, it stands for community and for the primordial and axiomatic morality of sharing meat. The commensality and extravagant sharing (for more food is provided than can easily be eaten) of the eat-all feast are symbols that refer both to their quotidien counterparts and to the essential notions

of sociality with which they are associated. The feast is a gift given by one man to others or a collective event in which multiple families cooperate. The rules effect a hypercommensality. At the Watt Lake eat-all, the kettle containing the "soup" was passed from hand to hand, each participant drinking from the same vessel. Similar sharing of a single dish is reported from Eastern Crees (Skinner 1914a :204-205; Speck 1935a :104,107; Rogers 1963:43). Further, it is required that all participants eat each kind of prepared food for which they are eligible.

The eat-all feast is in all these respects the antithesis of liminal rites de passage or of reversal rituals that suspend or invert conventional practices. The meat is identified with the social condition of the sacrificers not only through cooking but metonymically as the focus of a rite in which Cree society is ceremonially enacted. The relevance of a distinction between human and animal (or more broadly nonhuman) spheres is suggested by a curious piece of exegetical testimony. Some thirty years or more before the oppositional values of nature and culture became a structuralist aria, Speck (1935a :234) interviewed a Montagnais who interpreted distinct designs on his pack strap as representing "animals" and what Speck interpreted as "the artificial world of man." It is the latter artificial world that is enacted in the feast lodge.

The practices of closing up the lodge or cabin from within and covering the meat indicate an obvious concern with the creation of a privileged space. Unlike the shaking lodge, which is erected in the bush, the eat-all feast typically occurs within an existing dwelling, possibly one proximate to others in a bush hamlet. The feast thus occurs otinahk , in town, or kapisiwinihk 'in camp' rather than nohcimihk , in the bush. The movement of animals in production is from the bush into camp or town and then into individual dwellings. As noted earlier, the bush is considered to embody pikisitowin , cleanliness and holiness, in contrast to human spaces, which contain wiðipisiwin 'dirt'/'pollution'. The closure of the feast dwelling interposes a barrier between pihtokamihk 'inside' and waðawitamihk 'outside' (Tanner 1979:90). One meaning of the closure is to secure a clean space for the meat and for the ahcak beings to whom it is made over. "Outside" becomes, from this point of view, the polluted human space around the feast dwelling.

The dwelling is simultaneously a conduit within which transfers of food and more immaterial commodities between humans and ahcak beings can be effected. From this perspective, "inside" is a sanctified human space and "outside" is the sphere associated with the sacrificial recipients. Tanner's analysis (ibid.:93) suggests that Mistassini Crees

(Quebec) situate the "spirit world" to which sacrifices are addressed in the upper air. Rock Crees do not identify or situate cosmographically a single spirit world. Rituals involving spirit-human communication almost invariably presuppose barriers or distance, but these may involve underwater, underground, and remote terrestrial domains as well as the upper air. The game rulers, for example, are somewhat vaguely said to "live in the bush, but in a particular place." The closure of the lodge may be understood to create contiguity between the interior and a more generalized domain of ahcak beings which traverses and perhaps invisibly parallels and coincides with the space conventionally perceptible to humans. At the same time, the lodge must leave a passageway open, either for the entrance and departure of ahcak beings or for the transmission of the sacrifice. When the structure is sealed, the only visible opening is the smokehole in the nineteenth-century lodge or the stovepipe that now replaces it in lodges, tents, cabins, and houses. There perhaps exists a parallel here with the shaking lodge that spirit beings are said to enter through apertures at the top. These openings are, of course, vertically aligned with ispimihk 'above', or what one Cree called in English "the top of the earth." Crees do not elucidate a theory of burning in sacrifice, beyond saying that that conversion of meat into smoke "feeds" the sacrificial recipients, allowing them to appropriate an invisible simulacrum of the meal. It is interesting to compare Richardson's account of an eat-all feast at Cumberland House, Saskatchewan, where the entrance to the lodge was kept open throughout the meal (Franklin 1823:122-123). The door opened on the west, and wooden effigies of ahcak being were positioned inside the lodge to face the same direction; the meal was arranged so that no obstructions were interposed between these images, the doorway, and the west. More needs to be known of the symbolism of doorways in boreal Algonquian cultures, but here the doorway is seemingly a transformation of the stovepipe or smokehole. It is also significant that Rock Crees are careful to avoid physical contact between the doorway and slain animals introduced into a house.

The most observable moment of sacrifice in the feast is pakitinikiwin , the burning of a portion of each variety of feast food in the stove. These offerings sometimes accompany other meals. The most singular aspect of this practice to me was the minuscule quantity of food made over. In wihkohtowin , however, these gifts are peripheral to another economically convenient doctrine: the ahcak beings invisibly appropriate both the meat eaten by the feasters and the implements employed.

You kill a bear. And after you eat the whole bear, whatever you use on the bear—dishes, cups, whatever you use—when you're eating the bear, after you eat the whole bear, after it's gone, everything's gone, that's when he picks up the ones you were using. He's got it in his pack and he goes back living again and he's got all your junk. He goes back in his den next year and comes back to life. But he's got all your junk with him. All the food, the cups and plates, anything you use eating it. It's got everything.

I was told by one person that the ahcak beings are immaterially present in the lodge, by another that they ate the food "somewhere else," and by a third that they are simultaneously near and far; most people did not know. The idea that the food eaten by the feasters is taken by spirits is probably of wide subarctic provenience. It is not specified in Eastern Cree ethnography, but Landes's (1968:34) southwestern Ojibwa teachers understood bear feasts in this way. The wihkohtowin juxtaposes in a single act two practices that W. Robertson Smith (1956) imagined as successive stages in the evolution of sacrifice: the feast in which sacrificer and deity are commensals and a separate offering is made over to the latter. And it is consistent with the logic of Algonquian economic sociality that the spirits should be represented as commensals of humans. It is not only gifts of meat between families but events of collective eating that symbolize and create friendship and reciprocal dependence. In the eat-all feast, the hunter feeds the spirits in the act of feeding himself. Indeed, the assistance of the spirit is what allows the feaster to eat excessive quantifies of food:

He's [the feaster] not eating, it's not him that's eating. By the time he's done eating, he'll have nothing in the gut. Somebody else is eating it. It's disappearing as you eat. It's not in your belly after you're through eating. i-mamahtawisit , that's a witchcraft. Well, somebody's got to be . . . you can't eat one hindquarter on the beaver. You'll be full. Never mind the rest of it. Fifty pounds of meat there. A big beaver. You got to cat the whole works.

Just as secular reciprocity subsumes both gifts of food separately consumed and shared food collectively consumed, and just as the eat-all feast is simultaneously a gift from a sponsor to others and a collective meal, the sacrificial movement encompasses both a separate gift of food placed in the fire and a collective meal with the spirits.

Crees identify four kinds of entities as the recipients of food in wihkohtowin : the "soul" of the game animals being eaten, living game animals of the same species, game rulers such as Mistamisk 'Great Beaver', and the pawakan of the man who makes the feast. The latter

two arc unexceptional, since it is assumed that ahcak beings desire offerings of cooked meat. The notion that the animals derive satisfaction from the meat of their own worldly bodies is more enigmatic. One naturally reflects on the logic of a system in which powerful beings superior to man can be persuaded to discharge benefits in return for a sacrificial victim over which they already exercise dominion or with which they are, indeed, identical. On this paradox, Robertson Smith (1957:393) cited Jehovah: "I will take no bullock out of thy house, nor he-goats from thy fold; for every beast of the forest is mine, and the cattle on a thousand hills." Ingold (1987:247, 253), in a discussion of sacrifice, seeks to resolve this ambiguity that is central to Cree and perhaps other boreal Algonquian sacrificial practices: "An animal, quite simply, cannot simultaneously be the focus of veneration and an object [i.e., victim] of sacrifice, for towards the former man is the subordinate party, towards the latter he is dominant." Ingold concludes that the recipients of sacrifice in forager and pastoralist societies are everywhere beings of the type of the game ruler, not the souls of the animals themselves. Crees do not conceive events of sacrifice this systematically, and neither are the power relationships unambiguous or continuous. In Cree religious thought, animals or humans endlessly exchange dominant and subordinate statuses from one foraging event to the next. The question of characterizing one or the other as "dominant" could only be approached by induction from multiple discourses in which Cree foragers articulated their subjective understanding of their spiritual control over hunting. There is every reason to assume substantial variation in these understandings between different individuals and for single individuals over time as they revalue their serf-conceptions in the light of worldly successes and failures.

More paradoxical is the cannibalistic implication that animals derive gustatory satisfaction from consumption of their own flesh. But this is precisely what some Rock Crees understand to occur. "Oh, sure, they [animal souls] take all the food," I was told by a man who seemed surprised that I would ask. "It's like they're there in the house with you. But at the same time, they're way off somewhere else." Again the ethnography of the Eastern Crees is consistent, confirming this doctrine with respect to the burnt offerings: "Certain parts of the bear's flesh are burnt ('given to its soul to eat') including a small piece of its heart" (Skinner 1914a :204). At a Mistassini beaver feast, the names of different species were called out as the portions of food were placed in the

fire, suggesting also that the flesh of one species may be used to feed the souls of others (Lips 1947:389). The most explicit expression of this autocannibalistic doctrine known to me occurs in a myth told by the Salishan-speaking Thompson Indians of British Columbia in which deer provision themselves by hunting each other, killing and eating the animal forms of their companions who then regenerate and participate in the meal (Thompson 1966:169-173). In wihkohtowin , the animal is simultaneously sacrificial victim (as cooked meat) and among the sacrificial recipients. The problem of the identification of the meat with the deities is from this point of view already solved. All this suggests an implicit distinction between the infrahuman essence of the animal, which dines with the feasters, and the transient bodily form of the animal manifest in worldly experience.

Another facet of the feasters' hospitality remains to be discussed. Crees say that the eat-all feast was formerly the occasion for collective singing and dancing, a fact confirmed by the documentary sources. The hunting songs individually sung with the drum by Edouard, Antoine, and Marcel were reflexes of more substantial musical entertainments intended to honor and please the animals and other beings. Thompson (1962 [1784-1812]:81) wrote, "All their dances have a religious tendency, they are not, as with us, dances of mere pleasure, of the joyous countenance: they are grave, each dancer considers it a religious rite for some purpose; their motions are slow and graceful; yet I have sometimes seen occasional dances of a gay character." The earlier reports from Hudson Bay indicate that eat-all feasts—and feasts in general—usually concluded with collective singing and dancing (Drage 1968 [1748-49],1:219-220; Graham 1969 [1767-1791]:165-166, Franklin 1823:124). Dances might involve either all-male, all-female, or mixed participation. The animals and other ahcak beings are both fed and courteously entertained.

Communion

Hubert and Mauss (1964:39-40) wrote that sacrificial victims, when eaten by the sacrificers, transferred to them "the religious qualities that the sacrificial operation had kindled within them." Likewise, Robertson Smith's (1957) reconstruction of the early stages of Semitic sacrifice, whatever its relevance to the Semites, has Algonquian

resonances since it postulates a theology in which, as in wihkohtowin , sacred animals were at once sacrificial victims and the commensals of the sacrificers. According to Robertson Smith, food was understood to be constantly refashioning the flesh and other physical substance of the eater. For him, then, the sacrificial meal unites men with sacred animal both socially as commensals and physiologically through ingestion. By eating the sacred animal, the sacrificer "creates or keeps alive a living bond or union between the worshippers and their god. . . . The notion that by eating the flesh, or particularly by drinking the blood, of another living being, a man absorbs its life or nature into his own is one which appears among primitive people in many forms" (ibid.:313). And, as he might have added, among Catholics, a parallel Crees acknowledge when reflecting on the mass.

Apart from the obvious advantage of having a good meal, the rationale for forging such consubstantial bonds with deity is rather underspecified in the classic theories of sacrifice. The Algonquian testimony is quite specific. That Crees understand food (in most circumstances) as constitutive of the eaters' flesh is dear from certain of their foraging strategies, for example, the practice of harvesting hares only after they have fattened themselves on jackpine needle bait. Eating is thus the transformation of animal flesh into human flesh. As such, it is more than a means of materially renewing life. It also facilitates communication with animal beings through dreams and allows humans to acquire characteristics of animals that, it is said, they would otherwise lack. In a happy circularity, eating meat confers aptitudes and experiences congenial to the project of hunting successfully and thus being able to eat more meat. In this respect, the eat-all feast rehearses perhaps the most compelling of reasons for eating gods: the appropriation of the gods' superior endowments. In this respect also, the feast is continuous with the logic of the dietary rules to which individuals subject themselves and with that of other eating practices in nonfeast contexts. That eaters appropriate qualities that have been imparted to the meat by the sacrificial practice of the feast itself is not dearly evident; the meat may always contain such benefits. At the same time, the special cooking methods used and the rules requiring total consumption suggest that the meat at eat-all feasts is in exceptionally sacred condition.

I have already made reference (chap. 4) to the idea that a man both amplifies his dream guardian's "knowledge" of him and facilitates "getting in touch" with it by eating meat from an animal of the same species. A more general conception was recorded from an Eastern Cree:

"To eat that which is alive [meat], is certain to bring on animal visitors in dreams" (Speck 1935a :190). At a more immediate level, there is evidence that animal flesh contains within it just those endowments that allow humans to function most effectively as hunters. I was told by one man that he drank raw moose blood from freshly killed animals in the hope that this would impart to him in reverse the "invisibility" of the stealthy moose to its stalkers. I was not explicitly told that the food consumed in eat-all feasts has such effects, but there is evidence, admittedly paraphrased, that Eastern Cree feasters hope that this is so. At Eastmain bear feasts, the heart was eaten "in order that he [hunter] may acquire the cunning and courage of his victim" (Skinner 1914a :204). Likewise, four Montagnais stated that they ate raw marrow from the metapodial bones of caribou "in order to obtain the same courage (vaillance ), the same nerves, as this caribou had" (Harper 1964:58). These cases focus attention on particular meats, undoubtedly aligning with the intersecting male/female, older/younger, and idiosyncratic dietary classifications adumbrated earlier. The Eastern Cree literature also refers to rules proscribing the use of utensils while eating certain meats at eat-all feasts (Davies 1963:37, Skinner 1914a :204-205, Comeau 1923:87), conjecturally the inverse of the prescribed utensils that insulate pubescent women from the forces of the external world (cf. Lévi-Strauss 1969:336). It should be noted that this appropriation of animal attributes is not secured only by ingestion; charms of diverse kinds and animal hide clothing (cf. Skinner 1911:54) effect the identical goals through contiguity. Beyond the pragmatic desire to assimilate the courage, strength, endurance, stealth, and warmth of animals, these Algonquian doctrines of consubtantiation intersect with more imaginative themes—those exemplified by the trickster's animal impersonations, for example (Brightman 1988:79-80,86-87)—and perhaps with more diffuse moral and spiritual concerns.

Negative reciprocity

Wihkohtowin epitomizes the dominant model of reciprocity between humans and animals, but it connotes also the relationship of adversarial conquest. If the sacrifice symbolically conjoins reciprocity between hunter and prey with a ceremonial reciprocity

among hunters, it also associates the ambiguities and tensions of human food sharing with the adversarial relationship of hunter to animal victim. Meat-sharing practices, I am told, have undergone changes in the course of the present century. Prior to the 1950s, Crees spent most of the year in small multifamily groups in the bush. Older people described a system in which any large game animal was apportioned among the proximate households, either in a formal division or by casual gift.

The mode of distribution of meat following the capture of animals by Crees is today conditioned by the social context of the settlement in which it is butchered and consumed. When two or more families are co-resident in the bush, meat obtained cooperatively is divided among the producers. Other meat is usually maintained "privately," each family storing the surplus of its own kills. Notionally, each person "owns" the meat he or she hunts or traps. In fact, each person is subject to rules dictating that the meat must be shared with co-resident families. What is objectively "owned" is the right to consume meat one has produced, to give it to other families whose supplies are significantly lower (or who simply happen not to have quantities of that kind of meat on hand), and to receive meat in like circumstances. Gifts of food are readily exchanged between the commensal units, resulting in a general distribution of the surplus over time. At Granville Lake, for example, fish retrieved from Johnny Bighetty's fishnet circulated rapidly around the hamlet, whenever it was lifted. Fish were given immediately on the lake to those who helped lift the net; these, in turn, shared with other households. Feasts are another occasion during which the meat produced by one household is shared with others.

Enormous prestige accrues to kills of big game, and potentially any gift of such meat can carry the connotation that meat givers are superior to takers, the more so as the persons involved are differentially accomplished (cf. Tanner 1979:177). Recipients probably resent those who consistently hunt and trap more successfully than themselves. They are, in any case, imagined as liable to such feelings since jealousy over hunting skill is said to provoke sorcery. Resentment is also sometimes expressed toward those who consistently receive without commensurate reciprocation, often in the form of a criticism of the meat takers' lack of initiative or skill. A third conflictual aspect of meat sharing derives from the sedentism and increased density of post-1950s settlements. A hunter who brings meat either to the reserve at Pukatawagan or to a larger

bush settlement such as Granville must make decisions from among multiple eligible recipients. Attempts are sometimes made to conceal both the meat and the identity of the specific recipients from among all the possible ones. These attempts at secrecy are often unsuccessful, leading sometimes to resentments that are only assuaged by later gifts of meat. Prior to the sedentism of the 1950s, this problem must have been negligible during winter when the number of co-resident families entitled to gifts of meat was limited and each family was engaged in foraging. During summer aggregation, comparable problems perhaps existed. To these ambiguities must be added the supposition that there are surely some persons in some contexts who do not wish to share what they have killed.

The social tensions provoked by sharing should not be exaggerated. I have never encountered more generous people in my life than at Pukatawagan and Granville Lake. People visibly take satisfaction from giving, receiving, and reciprocating gifts of meat, equipment, and services. It is my impression that most Crees experience this sharing not as an onerous genuflection to official morality but as one of life's pleasures. The ambiguities nonetheless exist: the acts of giving and receiving meat that mean and create equality and community may also engender hierarchy and distance. They persist despite the presence of diverse mechanisms whose effect is to deny differentials in hunting skill: the categorical proscription of bragging and the use of mediating parties to transfer meat from one household to another. Lee (1979:244-249) documents similar measures among the !Kung. In the conventional eat-all feast, there is no apparent reference to these conflictual aspects of quotidian food sharing, unless the prestige diacritics of the dietary rules can be considered as such. The feast seeks to create between humans and ahcak beings the generalized reciprocity among humans to which it refers and of which it is a ceremonious instance.

The associations of food sharing with competition and hierarchy are, however, directly expressed in a distinct kind of eat-all feast that is symbolically the antithesis of the one described thus far. The feast becomes not a celebration of community but a shamanistic duel between human adversaries. The significance shifts from how much meat the collectivity can cooperatively consume to which of two contending parties can eat the most. Nelson recorded a myth describing the conflict of the hero Nehanimis with the monstrous "Hairy Heart beings" (chap. 5). The opponents attempt to overcome one another in simultaneous competitive eat-all feasts.

"Now," said Nayhanimis, addressing his family, "we must take 20 beavers, one for each man of them (meaning the hairy breasts [hearts]) and make a feast. If it turns out that we be able to eat these twenty Beaver, and they not, then we shall be superior to them and have the upper hand." The Beaver were cooked accordingly: he took his rattler which he shook to the tune of his songs—performed the usual ceremonies, and they eat the whole twenty beaver with ease. . . . The Hairy Breasts on their return did the same as Nayhanimis and cooked also 20 B[eaver] thinking that his [Nehanimis's] band did really consist of that number. They eat but everyone was already full and yet more than ¾ of the feast remained. (Brown and Brightman 1988:77)

Here, each beaver in the feast symbolizes a member of the opposing party, and consumption symbolizes the domination that it magically ensures. The myth illustrates the association of the eat-all feast with competition and aggression, recalling the more amiable competitive eat-all feasts of some Great Lakes Algonquians. It is interesting that the transitive verb wihkohtohiw 'feast someone' has connotations of this murderous competition absent both from the corresponding intransitive and the noun that refer to the conventional feast.

Like a guy might cook two beavers. Two swam. He give you one swan and he have one swan. One swan apiece. You got to eat your swan, the whole thing. If you don't, well, you're not sure to see next year. You make sure you wouldn't see next year. You're going to die between . . . in no time at all. You're a dead man. If you do eat the whole thing, well . . . you're alright. The guy is trying to kill you. If he can't eat it, well he's a dead man. After about a month after, he'll die. Just because . . . who's going to be the big okimaw ['leader'], you know? He's trying to get you. Either you or him. If he don't eat the whole thing, it means he can't do it.

I do not know whether Crees actually participated in competitive feasts with enemies, but there is little reason to doubt that such challenges were sometimes issued and accepted. In these cases, the meat is a medium for rivalry between human beings. The symbolism, however, is consonant with the adversarial model of hunting. The meat, as in the Nehanimis myth, stands for the opponent, and the latter's destruction is effected by devouring it. These duels between human adversaries suggest that consuming all the food prepared at conventional eat-all feasts possesses also the significance of a collective and aggressive act of magical control over animal adversaries. For the effectiveness of the feast is dearly proportional to the quantity eaten. Each animal eaten may, from this perspective, correspond to and prefigure the successful kill of an individual animal, or the same animal, in subsequent hunting,

just as each beaver represents a member of the opposed party in the Nehanimis myth. To eat all in the feast is thus a coercive act that prefigures domination of animals and secures in advance the event of eating well in the future. It is tempting to see also in the closure of the lodge an anticipatory capture of the animal in the future. Crees employed fences and impoundments in caribou hunting, surround techniques in stalking, and, of course, a variety of traps. Possibly the "imprisonment" of the animal in the lodge magically precludes its worldly escape from such devices and stratagems.

If the gift of meat between families sometimes connotes the superiority of the giver, this is consonant with the ideology of reciprocity in human-animal relationships that represent the hunters as the dependents of superordinate animal benefactors. The previous chapter suggested that the adversarial model provides an alternative to this self-conception of dependency. There is no indication that the prestations made over to animals in the eat-all feast connote human superiority. There is, however, reason to suppose that the closure of the lodge possesses the additional meaning of concealment. The entire complex of "respect" practices that the eat-all feast epitomizes has as its function to redefine as voluntary reciprocity events of hunting that the animals may experience as exploitation: theft of life, body, and autonomy. Speck (1935a :202-203) first raised the question of the polysemy of the closure in Eastern Cree feasts:

It is hard to see what the main religious aim of this feast may be. Whether intended to placate spirits of slain animals or whether it is a breaking forth of indulgence in carnal appetites—since we observe that measures of deceit are taken to preclude animal spirits from witnessing it—it is certainly an orgy of repletion.

In the same context, Speck cites exegetical testimony that "the openings of the lodge are kept closed lest the animal spirits who are thronging outside enter or obtain a glimpse of the proceedings inside." Three aspects of the feast—the closed lodge, the eat-all practice itself, and, in latter day Eastern Cree feasts, the covering of cooked food taken outside the lodge (cf. Tanner 1979:167)—all have the effect of concealing events in which animal flesh is eaten. When Turner (1894:323), for example, attempted to remove food from a Naskapi feast lodge at Ungava, he was told that to do so would result in the caribou deserting the vicinity. Tanner's (1975:308) first analysis of Eastern Cree feasts proposed that they implicate two discrete categories of spirits, benign beings

who are the recipients of sacrifices and the hostile game rulers from whom the feast must be concealed. It is, however, clear that in other Eastern Cree groups the feasts are understood to be required by the game rulers, not objectionable to them, as in the Naskapi feast observed by Henriksen (1973:35). Tanners later analysis (1979:173-176) of concealment strategies as expressing more general adversarial elements in human-animal relations is identical to the interpretation I offer here. The animal friend is within the cabin, partaking of the feast. The animal victim is outside where it can be kept in ignorance of its fate. They are really the same being. Rather than presupposing discrete classes of spirits, the paradoxical juxtaposition of display and concealment seems attuned to the disparate and ultimately unknowable moral evaluations that animals and their rulers bring to the hunt. In wihkohtowin , eating is reciprocity and communion: the ingestion of meat simultaneously feeds the spirits and suffuses the human body with their essence. But eating is also domination, and the eat-all feast is concurrently a display of eating as reciprocity and a concealment of eating as exploitation. One naturally reflects on the significance of covering the blood at kill sites with moss or snow, likewise a practice that gratifies the animal friend and tricks the animal victim. Tanner (1975:306) suggested, following Jean Briggs, that the practices of eating all the food and wrapping it outside the feast dwelling parallel the tactics available to the owner of a surplus who desires not to share with others: it must be secretly consumed or concealed. The analogy is apt, but, less than concealing a surplus from animals with whom the feasters do not wish to share, the sealed tent, the empty plates, or the covered vessel conceal the transformation of the animal into food and its ingestion by humans. Eating is no less culpable than killing.

The strategy of concealment raises the question of the degree of awareness of human comportment that Crees ascribe to animals. With respect to certain other spirit beings, there is no question: Crees conceive themselves as subject to a surveillance reminiscent of the panopticon.

"How durst thou doubt anything I say—knowest thou not how clearly and distinctly objects are discovered and seen in a plain, from an eminence; and my abode is in the regions above—I see every object as distinctly as you see at your feet, doubt then no more, and never hereafter call our Power into question."

"Aye!" replied some of the other spirits. "We not only see all that you do, however secret and hid you think yourselves, but we also hear every word you utter." (Brown and Brightman 1988:41-42)

Similar omniscience on the part of animals seems to be presupposed in all the objectifications of respect: rules of address and reference, mortuary deposition, dietary practices, music, and the eat-all feast itself.

Eating and Regeneration

Beyond reciprocity and duplicity, wihkohtowin is an instrumental involvement of humans in the reproduction of the game animals. Both the sealing of the cabin and the disposal of the inedible remains are implicated in this regeneration. Since the feast food is covered when the door is opened and uncovered only when the door is sealed, everything appears as if some immaterial or intangible property of the meat itself is intended to be immobilized or contained within the cooking vessel and the cabin. Further, the meat is carefully channeled to prescribed destinations, the entire conduct of the ritual serving to ensure that the meat is either further contained in the feasters' stomachs or disposed of in the fire. Speck (1935a :209) makes this association, whether from his own inference or Eastern Cree exegesis is not clear.

At the feast, all the holes in the covering of the lodge, whether it was of bark, skin, or canvas, were carefully closed so that the spirits of the beasts whose flesh was being consumed would not go out and in consequence, fail to return to life in the others of their kind to be reborn. Only after all the meat had been eaten would the lodge be opened.

Here the spirits conjoined with the flesh are preserved from dispersal and nonreproduction by being confined within the lodge. The most obvious analogy is with menstrual blood, which, like the meat-soul of the animals, "flows out" of an open container to the outside where it becomes a "wasted" substance that does not result in reproduction. The eat-all lodge is itself a simulated womb within which the game animals are prepared for the event spoken of as akwanaham otoskana 'they cover their bones' in which they regenerate as mature animals or are reborn in fetal form. Concern that the random physical dispersal of animal products may have negative consequences is expressed in other contexts. The Cree discourse at the beginning of the book inveighs against "scattering" meat, blood, and furs around a dwelling or settlement.

The attention focused in most feasts on grease, fat, and soups from water in which meat or bones have been boiled reflects Cree tastes but is

also directed toward regeneration. The indigenous theory of reproduction identifies menstrual blood as the substance from which the flesh and blood of the fetus are built up. Simultaneously, the fetus feeds on and forms its bones from semen. Menstrual blood that is retained in the womb eventuates, like the flesh contained in the lodge, in reproduction.

They got this special way of eating the moose. They mix some of that [raw] marrow from the long bones with the fat from the bones. Kinda' like pemmican. They smash that bone to make fat. Boil it up in a big kettle. And then they mix it. And they say you're . . . making a little moose. You make it over again from its bones by mixing it like that.

Inferentially, the marrow fat is semen and marrow is menstrual blood in this equivalence that can be generalized to diverse fat products and soups, on the one hand, and to all boiled and roasted meat, on the other. This may explain the importance attached to the soup and to the roasted fat that were served separately from the meat at the Watt Lake wihkohtowin . Henriksen's (1973) detailed account of an Eastern Cree eat-all feast seemingly exemplifies the same scheme: marrow from caribou metapodial bones was mixed both with boiled meat and with fat boiled from the marrow and bone splinters. Ultimately, not only hunting but also cooking and eating are metaphors of sex and reproduction.

Simulated regeneration was probably represented in an event at an Eastern Cree feast in 1820 which occurred immediately after the guests arrived and before the tent was sealed.

A painted deerskin [caribou] was then wrapped around a child who could crawl but not walk. The child thus wrapped was laid outside the door, and its mother began to call it to her. The child was so wrapped that it could only move like an earthworm and though it had only about two feet to crawl, about ten minutes elapsed before it could move that distance. The innocent cried in vain for assistance, though within reach of its mother's arm where she sat calling it to her. As soon as the child had got inside the door, the Indian began to beat the drum and sing for a little time. (Davies 1963:37)

The child's labored progression from the outside through the doorway and into the lodge symbolized both the quotidian introduction of caribou meat into the household and the ingress of animal souls invited to partake in the feast. That the child is induced to crawl into the lodge under its own power suggests perhaps a simulation of the caribou's voluntary entrance. Put another way, the child prefigures caribou that will come to hunters with the same determination with which the child seeks to rejoin its mother. The significance of moving "like an earth-

worm" is obscure, unless the actors desired to impart a similar immobility to their prey. The closure of the lodge after the child's entrance is metaphorically a capture. As a metaphor of regeneration, the child represents an animal soul that is recovered with flesh within the lodge, sent outside, and then brought in again. This particular theatrical moment microcosmically represents the whole ceremony it precedes and also the larger cycle of human-animal reciprocity in which it is embedded. Wihkohtowin reclothes the animal with flesh. It returns to the bush, there subsequently to be killed and once again brought into the lodge in the capacity of honored guest.

While the symbolism of rebirth is not directly implicated in the digestion of meat or the burning of offerings and residue in the stove, both practices are implicitly connected with regeneration. Eating, burning, and mortuary deposition of bones are the only appropriate ways of disposing of animals, aside from the uses made of skins and some bones. All must be appropriately disposed of or the lethal intervention of hunters in the animal's cycle of existence becomes permanent instead of momentary. The life force of the animal is imagined to be as partible as the meat in which it inheres, and everything must be concentrated, kept together, and rechanneled.

They say that it just comes up again and again. It will die and go back to life again. But then if you get it in your trap. Or you shoot it. Well, that's taking it away. You've spoiled it. Like there's holes in the skin now. Or you cut it apart for the skin. It's [animal] all ruined. So you have to be careful to put k back. Careful how you cut it up. Because it won't come back like before. Not if you don't treat the bones right. It'll stay dead forever.

The Auspicious Famine

The eat-all feast is simultaneously communion, reciprocity, trickery, and regeneration. Any given feast is contextualized as one moment in a perpetual series of transactions between humans and game animals. Each eat-all feast necessarily refers to two other moments and possesses in this sense a dual time reference to the past and to the future. The feast simultaneously reciprocates nonhuman others for the gifts of meat already procured—especially those eaten in the ritual—and initiates or anticipates subsequent prestations of the same order.

Like any gift of meat between humans, it simultaneously discharges an existing obligation of the giver to the recipient and re-creates a new obligation. In the dominant ideology of reciprocity, it is animals and the game rulers who are the meat givers in the two events of hunting that the eat-all feast—as a gift initiated by humans—concurrently reciprocates and anticipates. These gifts from animals to humans are themselves transactions interpreted as responses to previous gifts given by humans. In the first moment, the animals are "given" to the hunter, a prestation that may be prefigured in dreams or through other media but that is consummated when an animal is shot or taken dead from a trap. Hunters receive the animal corpses both as gifts and guests, eating the meat and employing the skins in trade or domestic manufacture.

The eat-all feast represents a reactive communicatory moment initiated by humans. The techniques that are used to induce contiguity between humans and the animal sphere are precisely those that differentiate the eat-all from secular meals: collective attendance, the sealing of the lodge, the special treatment of the food, music, and the enactment of status differentials. These operations place both the lodge and the meat—as sacrificial victim—into a sacred condition appropriate to communication with ahcak beings. The eating of the feast food is a complex act in which multiple functions and influences intersect. Through the food, from the nonhuman others to the hunters flow modalities of mamahtawisiwin , or power. Eating is also regeneration, the mingling of fat and meat in cooking and consumption prefiguring conception. Eating is also giving, for the animals and spirits partake in the feast as invisible commensals.

As an instance of reciprocity, the feast is doubly significant. First, through burnt offerings and the act of consumption itself, the food and other objects are made over as gifts to ahcak beings. In this respect, to eat all is to give all. But it is also to receive all. Crees say that wasting usable animal products is offensive to the animals and their rulers (cf. Strong 1929:284-285, Speck 1935a :83). In dealings with animals, total utilization is to gracious acceptance as waste is to refusal: uneaten food connotes a gift taken lightly or an unwillingness to receive. The eat-all feast thus discharges in a particularly conspicuous way the obligation to receive. Here also, relations between hunter and prey parallel relations between hunters. I never committed the error of refusing food but had occasion to observe the negative social effects of refusing alcohol. LeJeune (1897b :249) had similar experiences with the Montagnais in 1634:

It is giving a sort of insult to a savage to refuse the pieces which he offers you. A certain one, seeing that I had declined what my host had offered me to cat, said to me, "Thou dost not love him, since thou refusest him." I told him it was not our custom to eat at all hours; but nevertheless, I would take what he would give me if he did not give it to me quite so often. They all began to laugh; and an old woman said to me that, if I wished to be loved by their tribe, I must eat a great deal.

The eat-all feast is simultaneously an event of receiving and of reciprocation, relative to the gifts of animals that are consumed within it. It is in this respect that the magnitude of the feast becomes important: the more the feasters eat, the greater the magnitude both of their graciousness and of their return gift.

As one moment of reciprocity with nonhumans, wihkohtowin should structure economic relations between camp and bush such that the animals occupy the role of meat giver in future interactions. This is, above all, the raison d'être of the feast, and it defines the most fundamental meaning of the act of eating everything and destroying all remains. Lévi-Strauss (1966a :225) summarized the significance of the sacrificial victim in the Hubert-Mauss scheme as follows:

Sacrifice seeks to establish a desired connection between two initially separate domains. . . . It claims to be able to do this by first bringing together the two domains through a sacralized victim (an ambiguous object in effect attaching to both) and then eliminating this connecting term. The sacrifice thus creates a lack of contiguity, and by the purposive nature of the prayer, it induces (or is supposed to induce) a compensating continuity to arise on the plane where the initial deficiency experienced by the sacrificer traced the path which leads to the deity, in advance, and, as it were, by a dotted line.

It is through the destruction of the victim that human sacrificer and divine recipient exchange roles. Contiguity between giver and receiver is effected through the victim, an offering made over by the sacrificers. When the victim is destroyed, the contact is severed, ideally inducing the recipient to discharge future benefactions. In wihkohtowin , hunters initiate contiguity between human and animal domains, and the sacrificial victim is the meat. Already attached to the animal domain, the meat is identified with the sacrificers by being cooked and ceremonially eaten. The importance of bears derives from a similar logic: transitional between categories, they are especially appropriate mediators between hunters and animals. Contiguity is created both through the ingestion of animal attributes in the meat and through the commensality and sharing that the feast effects. The feast redefines the meat, originally a

gift from animals to hunters, as a gift from hunters to animals. Both the contiguity and the statuses of human host and animal guest continue as long as the meat remains. When the last of the meat is destroyed, contiguity is broken, and the statuses are exchanged. The absolute destruction of the meat, primarily by ingestion, ideally provokes resumption by the animals and game rulers of the status of generous benefactors. If the quantity of food consumed is important, the fact that none must remain is essential to the logic of sacrifice.

There is another meaning to the destruction of the meat-victim in wihkohtowin , consistent with but distinguishable from its position in the global scheme developed by Hubert and Mauss. It is a more markedly local meaning in that it implicates both the anticipated benefaction of successful hunting and the principle of generalized sharing that suffuses Algonquian economics. The feast begins with an abundant surplus of food and terminates with the absolute absence of food. The scarcity may be objectively realized since I am told that at some feasts everything edible in the camp was consumed. In other cases, only the food cooked for the feast is consumed, the people retaining other supplies. Or, as among Eastern Cree, the absence of food may be simulated by concealing the remains of the feast. In any case, the feast concludes in a genuine or simulated condition of famine. All the food is disposed of, and the hunters are thus left figuratively bankrupt. Gifts of meat between hunters occur automatically whenever disparities exist between stocks on hand. "Only a very bad man," I was told succinctly in Cree, "fails to share food with other people." Even more morally axiomatic is the obligation to share food with the destitute, those who have no food at all. In this connection, consider the formal resemblance of kitimakisiw 'someone is destitute' to kitimakiŒimiw 'relieve someone's distress'. By devouring all the food on hand, hunters manufacture a condition of indigence that animals and other beings in the expanded moral society will feel themselves bound to remedy. The discrepancy between the infinitely renewable supply of animal flesh in the bush and the poverty of the hunters is ceremoniously dramatized. Further, the hunters represent themselves as hungry beings in the throes of famine. The feast premises a moral universe in which the ahcak beings will not see people starve. "It is believed," Speck wrote (1935a :95), "that caribou spirits constantly visit the hunter in his dreams when food is exhausted." By fabricating a condition of scarcity, a simulated famine, wihkohtowin seeks to evoke the moral response appropriate to the tragedy of genuine starvation.