Edward Sapir’s Time Perspective in Aboriginal American Culture is a methodological essay published by the Geological Survey of Canada in 1916, written for anthropologists who wanted to know how old anything in Native American culture actually was. It is not a survey of peoples or artifacts but a working manual of inference: a set of techniques for squeezing chronology out of evidence that carries no dates. Sapir, a linguist by training and temperament, ranges across documentary records, native testimony, archaeology, myth, and language itself, asking at every turn how a flat description of culture can be given depth in time.
“Cultural anthropology is more and more rapidly getting to realize itself as a strictly historical science.”
The book’s governing premise is that ethnology belongs among the historical sciences. Culture, for Sapir, is intelligible only as the end-point of specific sequences reaching into a remote past, and the anthropologist’s habit of falling back on folk-psychological generalization is a symptom of having no chronology to work with. The task he sets is to read time perspective into American culture the way a viewer reads space out of a photograph—translating a two-dimensional record into three-dimensional reality. He is candid about the limits: absolute dates are rare and come with margins of years or millennia, sequence often has to stand in for date, and the units are generalized classes rather than individual persons.
The Problem of Time in a Flat Picture
Sapir opens by insisting that cultural anthropology has come to see itself as a strictly historical science: its data make sense only as the end-points of specific sequences reaching back into a remote past. Why, then, have anthropologists so often drifted toward folk-psychological generalization? He gives three reasons—the lack of direct chronological guides for primitive cultures, the lesser complexity of those cultures, and the greater diversity of cultures, which lends a “truer perspective.” None of this excuses the failure. Primitive culture must be worked out historically, in terms of actual happenings with definite sequence, localization, and mutual relations.
The task, as he frames it, is to inject chronology into descriptive fact—to translate a two-dimensional photograph into three-dimensional reality, reading time perspective into the flat picture of American culture the way space perspective is read from a photograph. The limitations are real. Absolute chronology is available only within margins ranging from years to millennia; often one can establish sequence without dates; and the events recoverable are generalized rather than individual—a “class of shamans,” or the Haida and Tsimshian treated as units even though their relations are borne by individuals, house-groups, and clans. These are quantitative, not qualitative, corrections. Two examples fix the scale of the problem: whether the movement of Siouan tribes like the Omaha and Ponca to the western plains preceded or followed the adoption of the calumet ritual, a question bearing on the Pawnee Hako ceremony; and the relative chronology among the Nootka thunderbird myth, house paintings, dances, personal names, and “thundering” for wealth, compared with the Kwakiutl. The evidence, he announces, divides into direct and inferential.
Documents, Legends, and the Long-Exposure Chart
Direct evidence begins with documentary records—the Jesuit Relations, Cook’s Voyages. More than four centuries have passed since discovery, and cultures have not stood still; comparing statements from different periods yields maximal and minimal dates. Wissler dated the spread of the horse; Mooney traced the Cheyenne from southern Minnesota to eastern Colorado and Wyoming; the Norse mention of kayaks gives that boat a minimal date; Maximilian and Catlin give a minimal date for Plains age societies such as the Mandan’s; and a bird-shaped Nootka rattle collected on Cook’s voyage, now in the British Museum, is identical with specimens still in use. Negative documentary evidence demands caution: the failure of early writers to mention floral beadwork, moose hair, or porcupine-quill designs among the Huron, Ojibwa, and Cree hints at a recent, European-influenced origin, and the absence of sails in old canoe illustrations suggests recent West Coast sail use—an observation Sapir credits to Dr. C. F. Newcombe. Absence alone, he stresses, is not demonstration.
Native testimony offers another route: statements embedded in formal legends or personal information about the age and sequence of events. The distant past must be handled with reserve, since history and myth merge there, but Sapir thinks the tendency to make light of Indian migration accounts has gone too far. West Coast accounts of village-to-village clan movements “ring of history”; Hopi clan legends of pueblo accessions deserve credence; the Tewa pueblo of Hano in Tusayan demonstrably traces its origin to the Rio Grande valley; and the northern provenience of Nahuatl-speaking Aztec legends is confirmed because Nahuatl’s remoter relatives—Cora-Huichol, Piman, Shoshonean—lie north of the Aztec historical home. The Tsimshian claim to have derived secret societies from the Northern Kwakiutl, corroborated, illuminates the chronology of secret-society spread. The Nootka acknowledge a Kwakiutl origin for specific dances and songs in their Wolf Ritual but not for the ritual itself—suggesting the earliest Kwakiutl influence antedates the Tsimshian borrowing of the whole complex. The Southern Paiute say their mourning ceremony with its peculiar songs came from the Yuman tribes to the west, and the Bear dance more recently from the Utes to the north. Culture, Sapir writes, is comparable not to an ordinary photograph but to a long-exposure star chart, the planets’ movement betrayed by short lines. A special type of testimony dates native monuments by aboriginal chronology—Aztec and Maya—almost the only direct fixed dates in remote aboriginal America. The oldest Maya dates reach to late classical times by our reckoning, far short of the total span needed and no help on earliest occupation, yet invaluable as a measure of vast time perspectives, since at the earliest datable period a highly complex culture already existed. Rejecting reputed finds of Tertiary man in America does not impair the certainty of a vast lapse of time.
Layers in the Ground
Stratified archaeological testimony rests on translating the successive deposition of artifacts and skeletal remains into a chronological cultural and racial sequence. Properly speaking it is inferential, but the inferences are so clear that Sapir treats it as direct. It has produced brilliant results in prehistoric Europe and western Asia; in America the results are meagre—perhaps, he suggests, because of the continent’s conditions—and the discussion passes to the inferential evidence, beginning with the seriation of cultural elements in order of complexity.
The Trap of Simple-to-Complex Seriation
The tacit assumption behind seriation is that development proceeds from simple to complex, so simpler forms are read as older. The single-figure Nootka totem pole becomes older than the superimposed-figure poles of the Haida and Tsimshian; the two-piece fire-drill of western tribes older than the Eskimo bow-drill or the Iroquois pump-drill; the unorganized shamanism of the Eastern Cree older than the Ojibwa and Menomini Midewiwin; the simple suitor myth older than its clan-legend form; detached amulets older than medicine bundles with rituals. Seriation also aligns elements by other logical criteria, such as geometric before realistic designs. Sapir’s warning is blunt: without outside chronological evidence an opposite bias could reverse the series, or one might dismiss it as subjective and historically fortuitous. Seriation suits evolutionary rather than strictly historical interpretation, ignores local and tribal differences and mutual influence, and yields a pseudo-historical construction. His illustration is the Iroquois and Wyandot, with exogamous animal-named clans and individual name sets grouped into two exogamous phratries, beside the neighbouring Mississauga Ojibwa, who had exogamous animal- and plant-named clans with individual names but no phratries. Seriation would call the Mississauga type older; the Ojibwa organization instead suggests the Mississauga borrowed a simplified form from the Iroquois, making the clan-phratric type older—though, a footnote concedes, this does not prove the Iroquois phratry older than the clan. Seriation works best for simple-to-complex sequences in artifacts and industrial processes, especially within a single tribe or restricted area.
What Must Come First
Under cultural associations Sapir places the principle of necessary presupposition: elements presupposed by others are earlier. The buffalo-skin tipi of the Plains presupposes a developed technique of fleshing, dehairing, and rendering hides pliable, presumably with deer brains soaked in water, so skin dressing is older than the tipi—though whether the conical pole-foundation tipi, such as the northern Algonkian birch-bark lodge, is also later than skin dressing is not necessarily involved. The woven rabbit-skin blanket and the rabbit throwing-stick are among the oldest North American elements; since other methods such as snaring exist, the throwing-stick’s greater age is not absolutely required, but Sapir hazards it as older, which would be confirmed if the technique originated in the southern plateaus, say among Shoshonean tribes, and spread north and east—of which there is no proof. Eskimo circle-and-dot and concentric-circle designs, producible practically only by the drill, are younger than the drill. Blackfoot medicine-bundle rituals centre on a manitou experience and are therefore much more recent than the typical American manitou experience. The caution: an element may be borrowed without its chronological antecedent—cultivated tobacco used as a religious offering may be purchased without tobacco cultivation—or the antecedent may be replaced by an equivalent, so a sequence holds only for part of an area, as a design arising with one technique may be adapted to another.
When One Element Lives Inside Another
A non-necessary sequence appears when one element forms the subject matter of another; if the reference is integral rather than accessory, the subject is older. Plains Sun Dance self-torture is an old practice specialized in a definite setting, probably older than the Sun Dance complex, its age further indicated by its occurrence in other connexions among the Kwakiutl and Nootka, though independent origin is not inconceivable. References in myths to implements, customs, or beliefs grow more telling the more frequent and stereotyped they are: the frequent whaling references in Nootka family legends show whaling older than a certain type of family legend. Conversely, the persistent failure of an element to appear in a representative set of myths is good negative evidence for recent origin—the Nootka Ts’ayeq doctoring ceremony, never mentioned in legends despite its religious importance, was likely introduced later than the Wolf ritual or the whaling rituals, confirmed by the more northern Nootka tribes lacking the Ts’ayeq. Place names and individual names serve as gauges: more Nootka individual names refer to whaling and whaling feasts than to Wolf Ritual dances or potlatching, indicating greater age for the former; the large proportion of Hopi individual names referring to corn culture implies agriculture practised for a very long time. Any well-defined style or traditional mode of treatment is apt to embody an old element.
Loose Fit, Late Arrival
Relative firmness of association gives another lever: the firmer the association, the older the element; the looser, the more recent, at least in that connexion. Composite complexes like myths and rituals can thus be resolved into time sequences. The Blackfoot Beaver bundle ritual, loosely superimposed on the Sun Dance complex, is later than the Sun Dance. In the Nootka Wolf Ritual dance complex most dances have nothing to do with the essential nucleus; two wolf dances are probably oldest; other typical, generally performed dances are more recent than the wolf nucleus but fairly old; and a large number of externally related dances are late, including the Cannibal dance, known from other evidence to be a recent acquisition from the Kwakiutl.
Words as Clues to Age
Sapir turns to the linguistic record as a way of dating culture, and he sorts the evidence into two broad channels: the analysis of individual words and the study of how language groups are distributed and how far they have diverged. The first channel rests on a simple observation about form. A term that can be broken down into meaningful parts—a descriptive compound—usually betrays a secondary or late formation, while a simple, irreducible word suggests antiquity, so long as borrowed terms have been cleared away. English makes the point plainly: bow, arrow, spear, wheel, plough, king, knight belong to the remote layer, whereas railroad, insulator, battleship, submarine, percolator, capitalist, attorney-general are transparently built from pieces and so are recent. The caution is that a word may once have been descriptive and lost its transparency through phonetic wear—king descends from cyning, a derivative of kin. Tsimshian offers a native illustration: dzabk, “crest,” resolves into dzab, “to make,” plus a mediopassive -k, while pte-x, “phratry,” resists analysis altogether, which suggests that crest representation came after phratry groupings were already in place; the Nass River dialect confirms it, keeping *pte-q\** for phratry but using a different word, ‘ayukus, for crest. Among the Nootka, yatsmi’lhsi, “attendant at a feast,” is analysable as “one who walks about in the house,” whereas the Kwakiutl ‘dlku is not—pointing to an earlier feasting ceremony among the Kwakiutl.
Place-names follow the same logic, since long occupation wears names down into conventional, unanalysable labels. Many Nootka village names resist analysis, while Paiute and Ojibwa topographic names are readily interpreted. Mt. Shasta is nm-nts-an lak-gai, “white mountain,” in Hupa but the unanalysable wa\galu-* in Yana, which argues that the Hupa are newcomers to northern California beside the Yana.
Sapir then hedges the method. Comparing equivalent words across languages is risky, because languages differ in how readily they form descriptive terms—Chinook and Takelma abound in radicals, Athabaskan has fewer radicals but greater synthesis—and in how fast they wear down phonetically. Athabaskan “mitten” is transparently “hand-bag” (Chipewyan la-djis, Hare lla-dji, Loucheux nle-djic), and the same compound reappears in Navaho la-djic. Taboo reshapes terminology: the Hupa use the descriptive xas-lm tau, “he who frequents riffles,” for crane instead of the regular Athabaskan stem del. Meaning shifts too—English needle and Hell, the latter once a cold domain ruled by a female deity; Athabaskan tlel means “matches” in Chasta Costa and Navaho but properly “fire-drill,” as Chipewyan shows. Specialization narrows sense: carpet-sweeper beside spinster, formerly “one who spins” and now “unmarried female of somewhat advanced age,” preserved in the rare agentive -ster of huckster, songster, Baxter, Webster. Nootka lutcha’, “buying a woman,” properly denotes the bridegroom’s distribution of property but now covers the whole marriage complex, songs included. A richer, more ramified vocabulary—especially where terms are non-descriptive—marks greater antiquity: cattle and dairy terms against orange-growing terms; sea-mammal hunting among Nootka and Eskimo; canoeing on the West Coast and among Eastern Algonkin; Iroquois agriculture; Plateau root-gathering; against the thin vocabularies of Ute decorative art or the Ghost Dance.
Grammar supplies a further test. Archaic, unproductive processes touching culture words imply age—English oxen, kine, calves/to calve; Nootka irregular plurals fya’wrli, “chiefs,” qaqo’l, “slaves,” >a’t:entl, “dogs,” which imply an ancient chief’s class, slavery, and dog domestication; Tsimshian sdm’i’gid/s9mgigad, “chief.” The criterion works only positively, never negatively, because analogy levels irregularities away—sheep, oxen beside horses, goats; a child’s mans, runned, brang; worked for wrought—and languages differ in how much irregularity they tolerate, Takelma against Yana and Paiute. Affixes with concrete significance likewise imply age: Yurok numeral classifiers for woodpecker-scalps and obsidian blades; Salish and Tsimshian canoe classifiers, the Tsimshian using distinct stems for “one” and “two” with canoes; Nootka classifiers for fathoms, spans, finger-widths, board-lengths; Nootka suffixes -o-‘il, -to’la, -inl naming potlatch procedure. Negative evidence stays weak, especially across structurally dissimilar languages like Hupa and Yurok.
Distribution, Borrowing, and the Direction of Diffusion
Wider distribution of a culture word suggests greater age, though borrowed elements are easily renamed, so caution is needed; linguistic evidence is generally less liable to misreading on two counts. Culture elements often make it hard to tell diffusion from common retention, yet native and borrowed words can frequently be distinguished, and that distinction lets inherited terms be separated from terms borrowed by alien groups. Words may also diffuse within a genetic group, judged by the same criteria applied more delicately to dialectic features. Even where origin is doubtful, a minimum date can be fixed from linguistic features, which matters chronologically. A corollary follows: linguistic evidence can show not only that a word is foreign but which language it came from, assigning a culture element’s origin to one tribe rather than another and fixing the order of assimilation.
The evidence is morphological or phonetic, or both. A word beyond a certain length must be at least partly analysable into stem and formative; if it is not, foreign origin is suspected. English hurricane, moccasin, tomato are unanalysable and longer than normal English stems, hence borrowed—clearing the path for a culture history of the moccasin and the tomato. Nootka tlo’kwa’na, the wolf ritual, is a native case: normal Nootka stems are monosyllabic (consonant-vowel-consonant), rarely two syllables, never three; tlo’kwa’na looks like tlo’kw- plus -a’na, but neither piece means anything. Kwakiutl dlo-gwala, a wolf dance, analyses cleanly into dlo’gw-, “to be powerful,” plus durative -(a)la. The inference: elements of the Nootka wolf ritual were taken over from the neighbouring Kwakiutl. Nootka topa’ti, “inherited privilege,” looks foreign too, though unconnected with any foreign element; if native it would be of great antiquity, since no descriptive meaning attaches to it. The most instructive cases trace to a specific language and so fix the direction of diffusion.
Morphology sometimes fails, especially with short unanalysable words: diffusion may be certain while direction is not. The word for “tobacco” among Diegueño (up), Shasta (o~p), and Takelma (o’upl) shows cultivated tobacco spreading gradually over western North America, but no single origin can be named; a south-to-north spread would rest on other than linguistic evidence. A widespread word for “dog”—Nahuatl chichi, Yana cucu, Takelma tsixi—poses the same problem.
Where morphology fails, phonetics may serve, since languages replace foreign sounds with their nearest equivalents, giving a deceptive native look. English rum becomes lam in Lower Umpqua and na-ma in Nootka, neither having r and Nootka lacking l. Nootka tlo’kwa’na, borrowed from Kwakiutl dlo-gwala, shows un-Nootka dl, gw, l replaced by tl, kw, n. Foreign sounds are sometimes preserved in demonstrably foreign words, letting alien provenience be fixed for others: the English voiced sibilant j never occurs in native words, only in foreign ones (chiefly French, Latin, Greek), from original j (rouge) or zy (pleasure, erasure, aphasia). Garage‘s j-sound and accent point to French origin, and though the French debt for automobile and garage rests on more direct evidence, phonetics alone strongly suggests it. Sounds native in some positions occur only in foreign words in others: z and dj are common medially and finally (as, fleas, chosen; edge, fledgling) but initially only in foreign words (zeal, zoology; Jew, just, John), which helps date judge, jury, general and their culture concepts. Combinations like -ps- mark foreign provenience (rhapsody, apse, Cyclops, lapse), while lips and sips resolve into a p-stem plus s-suffix.
American examples follow. In Haida m is rare and initially absent from native words; mat, “mountain goat,” related to Tsimshian mati, is clearly a loan from Tsimshian, not the reverse, and the mountain goat’s absence on the Queen Charlotte Islands strengthens but is not necessary to the argument. If the word is borrowed, the mountain-goat crest was likely borrowed from the Tsimshian as well, confirmed by other testimony; chronologically the crest is later among the Haida than among the Tsimshian.
The Argument Against Easy Evolution
The most bracing stretch of the book dismantles the assumption that simple forms are old forms. Seriating cultural elements from simple to complex, Sapir shows, produces a pseudo-history: it suits evolutionary theory rather than historical fact, ignores local differences and mutual borrowing, and could be reversed by an opposite bias without any evidence to choose between them. His example is the Iroquois and Wyandot clan-phratry system beside the neighbouring Mississauga, who had clans but no phratries. Seriation would call the Mississauga arrangement older; the borrowing record suggests the opposite, that the Mississauga took a simplified form from the Iroquois. The lesson recurs throughout: an element’s simplicity tells you almost nothing until you know who was in contact with whom.
What the Criteria Are For
Sapir’s inferential toolkit is best understood as a set of questions to ask of any trait. Does it presuppose other techniques—does the buffalo-skin tipi require hide dressing, does the drilled Eskimo design require the drill? Does it appear as subject matter inside another element, as self-torture appears within the Sun Dance, or whaling inside Nootka family legends? How firmly is it bound to the complex it accompanies? How widely is it spread, and how broken is its distribution? Each question yields a relative age, never a date, and each comes with cautions: borrowed elements arrive without their antecedents, complexes outlive their parts, and diffusion rates vary wildly between, say, humorous tales and secret medicines. The culture area, he adds, is a descriptive convenience, not a historical unit; its defining traits are of very different ages, and the Plains complex may be a recent overlay on older Woodland and Eskimo foundations.
Language as a Clock
Sapir’s longest and most personal argument concerns linguistic evidence, and it is here that his confidence runs highest. A word’s shape betrays its age: analyzable descriptive terms like battleship or percolator are late formations, while irreducible terms like bow or king reach back further, once borrowed words are set aside. Place names behave the same way—Nootka village names resist analysis while Paiute names parse easily, and Mount Shasta is transparent in Hupa but opaque in Yana, hinting that the Hupa are relative newcomers. Morphology and phonetics can even identify the lender: the Nootka word for the wolf ritual looks foreign, and its Kwakiutl counterpart is analyzable, pointing to borrowing from the north; Haida mat, with its un-Haida initial m, comes from Tsimshian, carrying the mountain-goat crest along with it. The distribution of whole language families then reconstructs the direction of tribal movements—Algonkin differentiation pushing the centre of gravity west of the Great Lakes, Athabaskan pointing to a northern homeland, Aleut and the Alaskan dialects placing Eskimo dispersion in Alaska.
A Method Offered, Not a System
Sapir closes by disclaiming completeness: he wants to suggest avenues, not legislate a procedure. He grants that direct evidence—datable documents, stratified archaeology—is generally superior, and that ethnological data outrank linguistic data in significance; his insistence on language reflects his own familiarity and a pedagogue’s sense that Americanists undervalue it. The one principle he presses hardest is negative: no single criterion is valid in a given case, and the doctrinaire’s clean constructions never match history’s actualities. The book’s lasting contribution is that warning, together with the demonstration that a culture element can never be safely torn from its psychological and geographical setting.
