The Algonquian Migration from Plateau to Midwest: Linguistics and

J. PETER DENNY University of Western

By means of linguistic research, hypotheses have been developed about the history of the Algonquian , which assert the times and places in which these languages and their ancestor languages were spoken. This historical geography of Algonquian needs to be connected to the archaeology of the same times and places, as has been done, most impressively, for the California languages (Moratto 1984: Chapter 11) and for African languages (Ehret and Posnansky 1982). In this paper,1 I offer a hypothesis linking archaeology and linguistics for one crucial phase of Algonquian history, the entry into the Midwest. The are widely distributed in the northeast quar­ ter of the North American continent, yet linguistic evidence and oral tra­ dition suggests that Proto-Algonquian, the ancestor of the group, was spoken in the Columbia Plateau. (All places named are shown in Fig­ ure 1.) In the present study I argue that the Proto-Algonquians moved east as one culture, and then their language spread and diversified into the daughter languages in the Northeast. I propose an archaeological identity for the starting-point of the Proto-Algonquian migration in the Plateau, and another archaeological identity for its end-point in the Midwest. This study of Algonquian culture history is a part of my research in cross-cultural differences in thinking processes. It is recognized that the habitual cognitive styles which characterize a culture arise at particular

'Thanks are due to Michael Spence and Christopher Ellis for their guidance concerning archaeology and to Chet Creider for advice on historical linguistics. The firstversio n of this paper was read at the 1990 Midwest Archaeological Con­ ference. I am grateful to those who offered valuable comments on that occasion, especially Robert Salzer, Robert Hall and James Griffin. The paper has bene­ fited greatly from suggestions made following its presentation to the Algonquian Conference, especially by Eric Hamp.

103 FIG. 1 MAP OF PLACE NAMES THE ALGONGUIAN MIGRATION 105 points in its past history. The best known case is the development of the de-contextualized thought style of Western Europe in classical Greece, which is first codified in Plato's insistence on studying each thought in and of itself (Havelock 1963). For Amerindian groups, we must try to achieve culture histories which are detailed enough so that we can, in future, discern the development of their intellectual preferences.

Proto-Algonquian in the Midwest An Algonquian migration from the west has long been considered. It is reported in the oral traditions of : the told one of their missionaries, John Heckewelder (1819), that many hundreds of years ago they lived in a very distant country in the western part of the American continent. Also, it has been hypothesized on linguistic grounds: Sapir (1916) argued that since the three languages which had most anciently separated from the Algonquian group, Wiyot, Yurok and Blackfoot, were all found in the west it was likely that the rest of the Algonquians had moved east. To study this migration it is strategic to determine the end-point first, since there are two linguistic methods which scholars have used to work back from the current Algonquian languages to the likely location in the Midwest from which they spread. The first of these is to use the words for biological taxa which are common to all the languages in order to establish the eco­ logical zone in which Proto-Algonquian must have been spoken. This was first done by Siebert (1967) who identified the region between Lakes Huron and Ontario as the Proto-Algonquian homeland in the Northeast; however. there does not seem to be any archaeological unit in this region which might be the Proto-Algonquians. In addition, Snow (1976) has claimed that this pioneering investigation used too many species whose ranges are hard to determine, or are too widespread to be helpful. Snow re-did the analysis and claimed that this technique only limits the Proto-Algonquian homeland to the much larger region of overlap of the larch (tamarack) and the beech, i.e., the mixed forest zone running from Lake to the Atlantic. In this case, other methods will be needed to find the homeland within this vast area. However, an important by-product of these investigations is the assurance, provided by the many words for Northeastern species which are the same in the Algonquian languages, that Algonquian speech arrived in the east as the single ancestral language, Proto-Algonquian, prior to the development of the descendant languages (probably excepting Blackfoot). The large region established by the first linguistic technique can be narrowed by the second one, which involves identifying the region of max­ imum diversity of the languages. Since the only genetic sub-group within Proto-Algonquian as yet established is Proto-Eastern Algonquian (God- 106 J. PETER DENNY

dard 1978a), it is thought that Proto-Algonquian split into the following 12 languages: Blackfoot, , , , Ojibway, Menomini, Potawotami, Fox, , Miami, , and Proto-Eastern Algonquian. The area that holds the largest number of these is the Lake Michigan- River area where 6 of them occur: Menomini, Potawotami, Fox, Illinois, Miami and Shawnee.2 This region is the most likely homeland, according to Dyen's (1956) migration theory, because it requires fewer moves (6) for the other languages to leave this region, than would be required for a larger number of languages to emigrate out of the Plains, the boreal forest, or the Atlantic seaboard regions, given that each of these ends up with fewer languages in it. In a previous study (Denny 1989), I reviewed the various archaeological traditions of the Lake Michigan-Ohio River region for the time period in which Proto-Algonquian was diversifying into its daughter languages, about 1000 to 500 B.C. (Goddard 1978a),3 and found that the twinned Red Ocher and Glacial Kame burial complexes were the only ones which could be used to account for the geographic spread of Algonquian speech.4 If we understand these to be the alliance-building activities of a single culture (Morse and Morse 1964:85), they explain the first wave of expansion which brings Algonquian speech: 1) up both sides of Lake Michigan, poised for later entry into the boreal forest; 2) northeast as far as Lake Champlain, in position for its later spread along the Atlantic Coast; and 3) east as far as the Scioto River.5 During a period of a thousand years, beginning about 1400 B.C., participants in the Red Ocher and Glacial Kame networks

2 Linguists and ethno-historians claim there is substantial evidence for placing Fox-Sauk-Kickapoo- at contact between lower Lake Michigan and lower Lake Huron, with Potawotami just to the north (Goddard 1978b), despite earlier doubts (Wakefield 1966; Fitting 1970). This appears compatible with the detailed hypothesis for their locations in Late Woodland times (after A.D. 1000) given by Stothers and Graves (1985). 3Similar dates are suggested by Siebert (1967) and Haas (1965). 4The alternative proposals are reviewed in detail in Denny (1989). The main one is that Point Peninsula (200 B.C. to A.D. 1000) accounts for the spread of Al­ gonquian speech (Walker 1975; Fiedel 1987, 1989; Seeber 1982). Although Point Peninsula did have influences on the Atlantic Coast (Allen 1982), it dates rather late to align with linguists' estimates that Proto-Algonquian starts splitting up between 1000 B.C. and 500 B.C. More importantly, since the continuities from Point Peninsula to the modern Iroquoian tribes are very clear, Point Peninsula was almost certainly Iroquoian-speaking. Proposals that several of the Lauren- tian traditions (4000-2000 B.C.) were Algonquian-speaking (Tuck 1977; Snow 1980:233) can be ruled out on linguistic grounds: to cover the large geographic ranges involved, even the chain of related Proto-Algonquian dialects that Snow (1977) envisages would have had to diverge into separate languages long before they are believed to have done so. The northeast spread of Proto-Algonquian along the of Lake THE ALGONGUIAN MIGRATION 107

of exchange, alliance and ceremonialism gradually developed bilingualism in Algonquian and eventually abandoned their old languages in favour of Algonquian. Hypotheses about the further spread of Ojibway and Cree into the boreal forest, and of Eastern Algonquian languages along the Atlantic Coast are given in Denny (1989).6

Proto-Algonquian in the Columbia Plateau There are several lines of linguistic and archaeological evidence which make it likely that Proto-Algonquian was spoken in the Columbia Plateau. First, Proto-Algonquian is known to be related genetically to two northern Cal­ ifornia languages, Wiyot and Yurok (Haas 1958; Goddard 1975). Second, speakers of these latter two languages are known to have arrived on the California coast bearing cultures typical of the Plateau region (Moratto 1984:564-565). Third, they are thought to have migrated there recently, the Wiyot about A.D. 900 and the Yurok about A.D. 1100 (Whistler 1979, cited in Moratto 1984:540). Fourth, as argued by Sapir, since the languages most anciently related to Proto-Algonquian, Wiyot and Yurok, occur in the West, it is likely that the split between them and Proto-Algonquian devel­ oped there. Unfortunately, this inference cannot be strengthened by Dyen's migration theory, because Berman (1982) has shown that the split produced only two languages, Proto-Algonquian and Ritwan, the latter splitting sub­ sequently into Wiyot and Yurok. Since there are only two languages, we cannot argue that the movement of Proto-Algonquian eastwards is more probable than the movement of Ritwan westwards. Fifth and finally,ther e are several studies showing phonological and morphological resemblances of Algonquian languages to and Kootenay, both spoken in the north of the Plateau. (Berman 1982; Haas 1965; Denny 1989)." These

Ontario, mediated by Glacial Kame, ensures satisfaction of a final constraint from Snow's (1976) study of biological terms: a word for the harbor seal could enter the language. 6We probably cannot find out what languages were spoken in the Northeast prior to the spread of Algonquian. At best a few words may survive as loanwords into Algonquian. David Pentland (personal communication) has suggested that Cree pi.-sim 'sun', may be an example since it contrasts with the words for sun in the other languages which descend from Proto-Algonquian *ki:sehswa. 7 In my 1989 paper, I suggested that the idiosyncratic morphological resem­ blances that I found between Algonquian and Salishan were evidence that they were genetically related, both being descended from an ancestor language usually called Mosan. However, comment from historical linguists has persuaded me that this is wrong. Eric Hamp (personal communication) points out that morpholog­ ical correlations which are not accompanied by phonological correspondences are only evidence for areal influence. Michael Foster (personal communication) notes that the idiosyncratic morphological resemblances offered by Goddard (1975) to 108 J. PETER DENNY similarities are most likely due to areal influences: the borrowing of struc­ tures back and forth during long residence as neighbours in the Columbia Plateau. Taken together these lines of evidence make it probable that Proto- Algonquian was spoken in the Plateau. To find the archaeological identity of the Proto-Algonquians in the Plateau we need to look for something resembling the archaeological tradi­ tions which are hypothesized to represent the later core of Algonquian in the Midwest, i.e., Red Ocher and Glacial Kame. Such an archaeological unit in the Plateau has been recently identified by Pavesic (1985): the Western Idaho Archaic burial complex has startling similarities to Red Ocher, as set forth below. This complex flourished about 2500 to 2000 B.C., perhaps extending a few hundred years later, and was located on the east side of the south-to-north stretch of the Snake River including the east side tributaries, the Boise, Payette and Weiser Rivers. 'The burial pattern for this Western Idaho complex includes multiple, flexedinterment s with associated artifacts. The burials are commonly placed in high, sandy locations . . . The use of red ochre is common." (Pavesic 1985:67). Besides Olivella shell beads, the di­ agnostic artifacts placed in the burials include four types of points: 1) large turkey-tail blades; 2) large bi-point cache blades (some quite ovate), which are finished but lack wear from use or handling; 3) caches of numerous obsidian blank/pre-forms of ovate-trianguloid shape; and 4) side-notched points of the same shape and size. All of these are illustrated in Volume 11 of the Handbook of North American Indians (Butler 1986:132), and are shown here in Figure 2. Similar side-notched points are found in previous burial units in the Plateau, so that Pavesic (1985:75) regards the firstthre e types as uniquely characteristic of the Western Idaho complex. When these three characteristics are compared to the "three distinctive nuclear traits which we consider to be diagnostic of Red Ocher" (Ritzen- thaler and Quimby 1962:249) there is a remarkable congruence. "The three traits are . . . large ceremonial blades, . . . ovate-trianguloid knives, and . . . turkey tails" (shown in Figure 2 of the present paper). The turkey-tails and ovate-trianguloid points are the same; only the ceremonial blades, derived from Titterington/Sedalia forms (Braun et al 1982:58), are different from the bi-point cache blades of Western Idaho. Pavesic (1985:67, 68) recognized both of these morphological congruences, but drew no inference of cultural connection. Other nuclear traits of Red Ocher which appear as major fea­ tures of the Western Idaho burial complex are: abundant red ocher, flexed burial in sand, and marine shell beads. The only one not characteristic of

support the genetic relation of Wiyot, Yurok and Proto-Algonquian are "far more probative of genetic relationship", than those I offered for Algonquian and Salish, because they show "parallel systems" of morphemes. Sarah Thomason (peronal communication) pointed out various problems I would have to solve in order to correlate Algonquian medials and Salish lexical suffixes. THE ALGONGUIAN MIGRATION 109

Diagnostic artifacts of the Western Idaho Burial complex recovered from the Braden site, a-b, Turkey-tail points: c-d, turkey-tail side and end scrapers; e-f triangular obsidian blanks; g-h, side-notched points; ii, large end-notched silica mineral points; k-l, ovate silica mineral blanks; m, stone abrader; n, stone drill or awl; o, bone point or awl; p, incisor fragment; q, antler haft or socket Length of a 22.0 cm, rest same scale. (Reproduced with permission from Butler, 1986)

Ovate-trianguloid knives of Red Ocher phase from Thiensville site (Redrawn with permission from Ritzenthaler & Quimby 1962) OOoo

Large ceremonial blades of Red Ocher phase from Carey site (Redrawn with permission from Ritzenthaler & Quimby 1962) FIG. 2 POINT TYPES OF WESTERN IDAHO AND RED OCHER 110 J. PETER DENNY

Western Idaho is copper made into artifacts, since it is not available in the Plateau. A detailed comparison of turkey-tail point types between the two tra­ ditions adds further evidence. Pavesic reports that Western Idaho has only the Fulton, Knox, Ross and Spoon types, from those illustrated in Didier's (1967:28) distributional study of Midwestern turkey-tails (see Figure 3). It is especially significant that these four types are derivable, by side-notching only, from bi-points, since bi-points are a diagnostic feature of Western Idaho but not of Red Ocher. This argues that the four bi-point based vari­ eties of turkey-tails found in Western Idaho are the original ones and that the other six types are later developments produced after the move to the Midwest — all of these latter varieties involve a more autonomous develop­ ment of the turkey-tail stem, either by elaborating it (Marshall, Mitchell, Stemmed), or diminishing it (Hebron, Kimmel, Dickson). There are also similarities between the two traditions for many of the "peripheral traits" (Ritzenthaler and Quimby, 1962:249) found in some but not all Red Ocher burials: both occasionally have smoking pipes (although of different form), gorgets (although currently of unclear provenance in the Plateau), and gray-black crystals (of hematite for Western Idaho and galena for Red Ocher). Atlatl weights, such as birdstones, which occur in Red Ocher, do not appear in Western Idaho; perhaps they are also an influence from Titterington/Sedalia. To summarize: oral tradition and linguistic evidence indicate that Proto- Algonquian most likely originated in the Columbia Plateau. Since Red Ocher/Glacial Kame (ca. 1400-400 B.C.) best accounts for the spread of Algonquian speech in the Midwest, it is very significant that there is a prior burial complex in the Columbia Plateau, the Western Idaho Archaic (ca. 2500-1800 B.C.), which strongly resembles Red Ocher. It seems likely that Western Idaho is the Proto-Algonquians in the West, and Red Ocher/ Glacial Kame is the Proto-Algonquians after their migration to the Midwest.8

Origins of Western Idaho and Proto-Algonquian in the Plateau Since linguistic evidence suggests a lengthy past for Algonquian and its ancestor language, Algic, in the Plateau, if we are to equate the Western Idaho complex with the Algonquians, it is important to know whether this burial complex also has deep antecedents in the region. Evidence that it does comes from the careful comparisons which have been made with Cas­ cade phase burials (ca. 4800 B.C.) and the DeMoss burial site (ca. 4000

8 Since all the sites which evidence the lengthy in situ development of both Western Idaho and Red Ocher were accidentally discovered, it is unlikely that we will ever discover archaeological remains of the Proto-Algonquians on the move across the Plains. THE ALGONGUIAN MIGRATION

FIG. 3 TURKEY-TAIL TYPES OF THE RED OCHER PHASE (Modified with permission from Didier, 1967) 112 J. PETER DENNY

B.C.) (Pavesic 1985; Green et al 1986). Certain features of the Western Idaho complex go right back to Cascade phase burials: flexed inhumations, Olivella shell beads, and use of red ochre. However, diagnostic features such as the turkey-tails and ovate-trianguloid cache points do not; nor are the characteristic Cascade points found in the Western Idaho complex. The DeMoss burial, however, indicates the transition from Cascade phase to the Western Idaho complex. It occurs on the northern edge of what is later the Western Idaho region, and contains both Cascade points and incipi­ ent turkey-tails, as well as ovate-trianguloid preforms and large bi-points. Clearly, "there is continuity between the earlier Cascade phase burials and the later Western Idaho burial complex . . . The Cascade association is a verification of a northerly cultural link" (Green et al 1986:40). These archaeological conclusions fit exactly with the view from lin­ guistics: Proto-Algonquian develops gradually out of Algic in the southern Plateau, and has strong affinities to Salishan and Kootenay to the north.

The Migration and Settlement in the Midwest Since linguistic and archaeological evidence show that Proto-Algonquian originated in the Columbia Plateau, and since Algonquian languages spread throughout the Northeast, the Algonquians' arrival and first settlement in the Midwest must be sufficiently large and impressive to the existing popu­ lation so as to account for the extraordinary success of the . A first point relevant to this impressiveness is the nature of the society that set off. Since all the Western Idaho sites are burials, Pavesic (1985:79) can offer only tentative hypotheses about socioeconomic patterns. The most likely staples were camas roots and , both of which are known to encourage quite populous and sedentary lifeways. "A highly integrated and organized society", tentatively thought to be egalitarian, is reflected in the burial wealth. An extensive trading network is also evidenced. It appears that a relatively complex and large-scale gathering-fishing-hunting society is involved. We do not know why an apparently prosperous community living in a favored environment should decide to leave. The most likely hypothesis is invasion by Sahaptin-speakers from further west whose descendants in the region are the Nez Perce. The area in which the Western Idaho complex was located is strategically placed for an eastward migration. It is in the sagebrush-wheatgrass steppe which extends eastward through the Snake River Plain and continues southeastwards across the Continental Divide to the headwaters of the Platte River (Butler 1976). This would be an excellent route east, but for various, rather indirect, reasons to be given below I believe the Proto-Algonquian migrants used a more northerly route. THE ALGONGUIAN MIGRATION 113

I will outline the hypothesized migration route as a whole, and then present evidence in support of various parts of it. Where the Snake River Plain ends at the Divide, I think the migrants turned north and then east along the upper Missouri River, proceeded south to about the James River (in South Dakota), and then went east into the hilly country which gives rise to various southeastwards tributaries of the Mississippi. Here I believe they split into two groups. A northern group migrated approximately down the Cedar- River, across the Mississippi and along the upper Illinois River, settling at the south end of Lake Michigan on the Kankakee River. A southern group moved approximately down the , then southwards on the west side of the Mississippi almost to the mouth of the Illinois, across the Mississippi and the Illinois, settling eventually in the . Given the terminal date of Western Idaho, about 1800 B.C., and the start of Red Ocher about 1400 B.C., the migration must have taken several hundred years in between. Some aspects of this migration hypothesis are supported by the tra­ ditional Delaware migration myth reported by the Moravian missionary John Heckewelder. Heckewelder provides two versions of the story, a long- published one (Heckewelder 1819), and another, slightly more detailed one entitled "A Short Account of the Emigration of the Nation of Indians Call­ ing Themselves Lenni Lennape as Related by Themselves" more recently published for the first time in Weslager (1972:88-90). Another, greatly elaborated version of the Delaware migration myth is found in the Walam Olum (Voegelin et al 1954); however, it cannot be used as evidence in this study because it has not, so far, been authenticated by scholars (Boewe 1987), although it is accepted by some of the Delaware.9

9 The Algonquian text and accompanying pictographs come from a manuscript prepared in 1833 by a biology professor, Constantine Rafinesque, who said they were copies of originals he received from other whites. To date, they have not been traced to any Delaware Indian source. Both the text and the pictographs appear to be mnemonics which give key words and ideas from which a Delaware orator would have created an oral performance. The English text by CF. Voegelin given in the 1954 edition is thus a realization based on translations of the key words. In the 1954 edition the anonymous editor, known to be Eli Ldly who financed the edition, offers the certainly wrong interpretation that the migration described began in . The Walam Olum is most likely a mixture of old and new material, such that the various parts of it can only be believed if external evidence is found for them. Information in it that must be recent are the names of chiefs which are based on loanwords from European languages, such as WaptipaM White Chicken' in verse IV:41 and Wapushuwi 'White Cat' in verse V:3. The former involves tipaas, a loanword for chicken, and the latter puusis, a loanword for cat. 114 J. PETER DENNY

Here is a summary of the main points reported in the two Heckewelder versions:

(1) Many hundred years ago, the resided in a very distant country in the western part of the American continent. (2) They decided to migrate eastward in a body. (3) After many years travel, they arrived at the , and followed it south to a point 100 or more miles north of the Ohio River. (4) There they fell in with some of the Mengwe nation,10 who had also emi­ grated from a very distant country, and had struck upon the Mississippi more to the north. The Mengwe were proceeding eastward toward large lakes, where they would be able to secure themselves against a mighty people inhabiting the country to the southward. (5) Spies sent forward by the Lenape discovered that the country to the east of the Mississippi was inhabited by a very powerful nation called the Talligewi. (6) Although the Talligewi gave permission for the Lenape to pass through to the east, when they began to cross the Mississippi, the Talligewi, seeing their great numbers, attacked them, cutting off their retreat, and preventing that body of their people who had not crossed from doing so. (7) They asked the Mengwe to help, and together, after many bloody bat­ tles, they defeated the Talligewi who fled down the Ohio and Mississippi Rivers, from whence they never returned. (8) The victors divided the country between them, the Mengwe choosing all the lands in the vicinity of the large lakes, and the Lenape taking possession of the whole country to the south of there. (9) For many hundred years, the two nations resided peaceably in this coun­ try and increased very fast.

One can see in the migration myth support for several aspects of the migration that I have outlined previously such as an origin in the west, and a movement east to cross the Mississippi. The division of the migrants into northern and southern groups is also partially supported by the reported alliance with the "Mengwe" who crossed the Mississippi further north and headed east for the large lakes. Also the route of the southern branch down the west side of the Mississippi to cross near the mouth of the Illinois is

In modern Delaware and other Eastern Algonquian languages Mengwe is the word for the (Goddard 1978c:320). It cannot refer to them here since no Iroquois group ever moved east across the Prairies toward the . Apparently Delaware storytellers in recent centuries have applied the name for the recent allies of the Delaware, the Iroquois, to these ancient allies THE ALGONGUIAN MIGRATION 115 attested by the oral tradition.11 Linguistic considerations also support the migration scenario I have out­ lined. First of all, the location of Blackfoot argues for a northern route from the Continental Divide via the Missouri River rather than a southern route via the Platte. Blackfoot is the most deviant of all Algonquian languages, and almost certainly splits earliest from the rest of the family. Traditionally, the Blackfoot have occupied the area north from Yellowstone Lake along the south-to-north stretch of the Missouri and into the Saskatchewan drainage (Janetski 1987:29). It seems likely, therefore, that they stayed behind on the upper Missouri River when the rest of the Algonquians migrated further east. Further linguistic evidence suggests the early development of a group speaking an Algonquian language who remain west of the Mississippi in the Missouri River region. Rhodes (1988) claims that Arapaho, which was even­ tually spoken on the upper Kansas River, diverges from a Central language which only later differentiates into the languages spoken east of the Mis­ sissippi (Menomini/Cheyenne, Cree, Miami, Illinois, Ojibway, Potawotami, Fox/Sauk/Kickapoo/Mascouten, and Shawnee).12 This is congruent with the Delaware migration myth summarized above which claims (point 6) that only part of the population crossed the Mississippi. Archaeological evidence of a group remaining west of the Mississippi is seen in Red Ocher sites such as Elm Point north-west of the mouth of the Missouri; however the further development of the Arapaho is not archaeologically visible at present.13 If future research validates the Walam Olum we will find there many

11 So far as is known to date, Heckewelder gives us a brief summary of an authentic native tradition which we may use as evidence. However, by the 1820s this tradition had been lost due to dislocation and death among some groups of Delaware (Weslager 1978:89, 165), and does not seem to be a part of the tribe's present day oral tradition. The time depth remembered in this tradition, 1800-1400 BC, is about the same as the earliest Vedas of north India. Such a feat of oral memory is possible for a large-scale semi-sedentary gathering-fishing- hunting society of the sort hypothesized for Western Idaho and accepted for Red Ocher/Glacial Kame; in contrast, it is recognized that small, acephalous hunting bands do not preserve lengthy histories. 12From the Delaware oral tradition we can infer that the Central language gen­ erated another daughter language also, which we might call Central Delaware. This probably cannot be detected linguistically because following.the migration of these people into the Atlantic seaboard (probably after A.D. 1000) descnbed in Heckewelder (1819), their language most likely icnerged with an aire*dy^ex­ isting Eastern Algonquian language to produce modern Delaware a member of the Eastern Algonquian grouping. Also, note that Cheyenne split very late from Menomini and the tribe migrated westwards on to the Plains. 13A possibility worth study is that they moved west into the Kansas drainage between 400 B.C. and 100 B.C., as the Black Sand phase was developing in their original territories north of the lower Missouri. 116 J. PETER DENNY details which support the migration as outlined above. First, the original homeland is described as a north-sloping land, which is correct for the south- to-north stretch of the Snake along which Western Idaho occurs. Second, the migrants set off eastwards and upstream which fits the first part of the migration up the Snake valley to the Continental Divide. Third, this part of the journey ends at a place of snowy mountains, hollow wells, and a frozen body of water, which describes salient conditions of the Yellowstone region at the head of the Snake: very high snowfalls, mountains, geyser basins, and Yellowstone Lake frozen over in winter. Fourth, from a lake in the forest they travelled north and then east; this fits Yellowstone Lake and trails from it following the Missouri River. Fifth, their enemies hid in prairie hollows, suggesting movement across the northern Plains. Sixth, they had a chief named White Owl, pointing to travel along the most northerly west-to-east stretch of the Missouri, since the range of the Snowy Owl does not go any further south. Seventh, they turned south away from snow and cold toward nut-producing14 lands; this fitstrave l along the firstsouthward s stretch of the Missouri, leaving the Plains and entering the Prairies where nut trees begin. Eighth, they settled in hills and had an angry split in the tribe in which some people left going eastward; these hills seem to be the north- south range, lying east of the James River, from whose east side rise the Des Moines and the Cedar Rivers. Furthermore, this seems to be the split into those migrating east-southeast eventually to the Kankakee region, and those migrating south of southeast eventually to the Wabash. Ninth, they became separated from some of their tribe at a large north-south river; this seems to be the crossing of the Mississippi leaving behind those who eventually spawn the Arapaho. Tenth, they travelled east over hills to the land of the Talligewi; this fitsth e final movement of the south branch of the migration into the Wabash valley to confront Riverton peoples. Some scholars may think that these many congruences make it unlikely that the Walam Olum is a faked elaboration of the Delaware migration myth, since it would be impossible for a forger, whether Indian or White, to know all these details only revealed by 20th-century research. Other scholars will want to wait for conventional validation. I outline these congruences to facilitate future investigations; however, the Walam Olum evidence has not been used to support any of the positions argued in the present paper.

The Algonquian morpheme -min- used in the words in question refers to fruit of all kinds, both berries and nuts. Since nuts are a staple and berries are not, and since nuts occur in the Prairies but not on the Plains, whereas berries are found everywhere, we can infer that it is nuts which were the valuable distinctive resource encountered on entering the Prairies from the Plains. THE ALGONGUIAN MIGRATION 117

The Archaeology of Algonquian Settlement in the Midwest

The hypothesis under consideration is that the settlement of the Proto- Algonquians in the Midwest shows up as Red Ocher/Glacial Kame. The thousand-year history of these twinned traditions, roughly 1400-400 B.C., is ample time for the Central language they originally spoke to differen­ tiate into the descendant languages listed above. These two burial com­ plexes appear to have originated in a single culture, in the Illinois and adjacent drainages where they both occur, because they have many simi­ larities and few distinctive traits (Morse and Morse 1964). They also seem to be contemporaneous: the earliest widely-accepted date for Red Ocher is 1220 B.C. with late dates around 400 B.C. (Stoltman 1986). The only two dates available for Glacial Kame fall in the middle of this period, 950 B.C. (Spence and Fox 1986) and 980 B.C. (Power 1989). Didier's (1967) distributional study of turkey-tail points, one of the three defining features of Red Ocher, gives considerable support and added detail to the pattern of Algonquian settlement which I have hypothesized. In the south there is a continuous distribution from the mouths of the Illinois and the Missouri, down the east shore of the Mississippi, up both sides of the Ohio and spreading north into the Wabash drainage and the Miami and Scioto valleys. This is most likely the settlement resulting, over several hundred years, from the southern branch of the migration. The heart of this continuous belt is what Didier calls the "Southern " region which includes the three sub-regions containing the hornstone quarries from which the turkey-tails were made. This region stretches from the west side of the Mississippi (starting south of the Kaskaskia) up both sides of the Ohio and into the Wabash drainage. Within this hornstone producing region, Didier recognizes a sub-region, the Wabash drainage, which is the prime focus of the turkey-tail distribution network, and therefore one of the main foci of Red Ocher. In this sub-region only, which contains the famous Harrison County quarries, all 10 varieties of turkey-tails are found. To the north are several discontinuous regions in which turkey-tails oc­ cur. Two of these seem likely to represent the northern migration ending at Lake Michigan: the region as a firststag e across the Missis­ sippi, and the Kankakee region as the main settlement area just south of the lake. The Kankakee region can be identified as the second major focus of Red Ocher, because, as Didier points out, almost all varieties of turkey-tails, except those mostly restricted to the Wabash region, are found there. Also the Kankakee region is the path for the flow of turkey-tails to Didier's three most northerly regions, Lake Winnebago (including Green Bay), Southeast (mid-west shore of Lake Michigan), and Saginaw. By looking at Didier's distribution of turkey-tail varieties we can infer which regions are likely to have received substantial populations of Proto- 118 J. PETER DENNY

Algonquian immigrants, and which ones had existing populations who grad­ ually participated in Red Ocher trade and gradually adopted Algonquian speech. Those regions having substantial numbers of the four original bi- point varieties of turkey-tails found in the Western Idaho Archaic burial complex are the most likely settlement areas for the migrants: these are the whole southern tier of regions from the Illinois-Missouri-Mississippi conflu­ ence through to the Scioto, as well as the Central Illinois and the Kankakee regions. The remaining three most northerly regions have no or almost no occurrences of the bi-point varieties, and are more likely areas of Red Ocher influence on other populations. The most easterly part of the southern tier and likely farthest extent of the original Algonquian migration is Didier's Ohio region, the Miami and Scioto valleys. The turkey-tail distribution there is particularly interesting because it features all three of the elaborated varieties, those in which the turkey-tail stem is given fancier treatment. Perhaps these are early signs, in this third focus of Red Ocher, of an inventive society from which Adena would soon grow. Didier points out other indications that the Central Illinois region may be the fourth major focus of Red Ocher since it seems to participate vigor­ ously in the copper trade with the northern regions and has further elabo­ ration of burial goods in the use of marine shell and galena. Of the northerly regions which did not receive immigrants but did be­ come influenced by Red Ocher and did gradually convert to Algonquian speech, the greatest interest has always been in the Old Copper area to the west of Lake Michigan (Didier's Winnebago and Southeast Wisconsin regions). The general distinction between the earlier era of Laurentian Old Copper and the later Red Ocher era is well-captured by Fogel's (1963) study of the copper trade. The Laurentian pattern of east-west trade in copper, in which the metal flows from quarries along to sites along Lakes Huron, Erie and Ontario, lasts to about 2500 B.C. By 2000 B.C. a new north-south trade pattern is initiated by the Hemphill phase as evi­ denced by the Osceola site on the southwestern fringe of the Old Copper region (Conrad, cited in McElrath et al 1984:39). The proto-Algonquian immigrants in the form of Red Ocher set up their own north-south trade in copper, apparently on an axis a bit to the east of Hemphill. In the context of other reciprocal relations, turkey tails from the hornstone quarries in the very south of the Red Ocher area may have been exchanged for artifacts of copper from the quarries on the northern edge of the Old Copper area. One indicator we have mentioned that the Old Copper area receives only Red Ocher influence, not immigration, is the absence there of the four bi-point derived varieties of turkey-tail points which were the original types made in the Plateau. Instead, all three of the diminished varieties, having reduced THE ALGONGUIAN MIGRATION 119

and unelaborated turkey-tail stems are found there (as well as one of the elaborated varieties, the Mitchell). Further evidence of the indirect influ­ ence of Red Ocher is found in the presence of points which only resemble turkey-tails at the Riverside site (Mason 1981:226). Certain other artifacts which show conservative tendencies reminiscent of the Western Idaho Archaic burial complex itself appear at Riverside (Ma­ son 1981:227): a big block of obsidian from the Yellowstone region and caches of blades, made from Ohio valley hornstone, but in the bi-point'form which was one of the diagnostic features of Western Idaho — these may be turkey-tad preforms. These findings raise the possibility of some later influence from those who remained behind in the Yellowstone region i e the Blackfoot. When the Algonquians arrived in the Midwest, besides Old Copper there were two other well-established foci of funerary ceremonialism associ­ ated with two relatively sedentary and complexly-developed cultures: Tit- terington/Sedalia/Nebo Hill/El Dorado along the lower Missouri and River- ton on the Wabash (Bender 1985:57). The Algonquian migration seems to have skirted the northeastern corner, i.e., the Titterington segment, of the former, and to have eventually displaced the latter. The accommodation of the Algonquians to Titterington is seen in sites such as Elm Point just north-west of the mouth of the Missouri. Originally proposed as a Titterington site, it does not meet the more stringent criteria now in use that Titterington sites display a high frequency of Etley and Wadlow points (Reid 1984:74), as is the case at the Etley site just across the Mississippi. The Elm Point burials are clearly Red Ocher having many turkey-tails and also turkey-tail blanks (i.e., bi-points), but show such Tit­ terington features as limestone slab roofs and an Etley point. It was pointed out earlier that the main adaptation of Red Ocher to Titterington/ Sedalia is the adoption of Sedalia-like cache blades instead of the bi-points found at Elm Point (as well as Riverside). As the Algonquians moved past Titterington and into the Wabash drainage they seem to have come into conflict with the peoples of Riverton culture (1600 B.C. to 1000 B.C.), who had not long before moved north, possibly from the Tennessee valley (Winters 1969). In view of the success­ ful accommodations to Titterington/Sedalia and to the descendants of Old Copper, why did warfare ensue with the Riverton peoples? The answer may lie in their social exclusiveness: Winters (1969:130) describes them as "essentially endogamous" since almost no artifacts of other cultures show up on their sites. According to the Delaware oral tradition, the southern branch of the Algonquians allied themselves with the northern branch, the so-called Mengwe, from the Illinois-Kankakee region, to defeat the Riverton people, the Talligewi, and drive them back southwards. 120 J. PETER DENNY

Given the enormous range of the Red Ocher tradition and its many burial sites it has long been a puzzle that the habitation sites of these people are so hard to find.Winter s (1967), in his survey of the Wabash found very little that was post-Riverton and pre-Havana. Reassuringly, some ovate- trianguloid preforms appear (27: Figure 5, R, S, and T) and some side- notched points that may come from them (U, V, W). The case of the missing habitation sites is discussed very astutely by Morse and Morse (1964:82). They say "in the Red Ocher area, particular habitation sites do not seem to have been reoccupied. The only indication of seasonal permanency are the cemeteries . . . We suggest that seasonal camps or villages could be placed almost anywhere within the areas with no loss in effective exploitation of the environment." Given our hypothesis that the Red Ocher folk are invaders, it is highly pertinent that these authors say: "Another factor might be the intensity of raiding and hence a possible tendency for secretiveness and mobility." Munson (1966) suggests that in the later centuries of Red Ocher their habitation sites were those characterized by Marion Thick pottery and Kramer points. This would cover only the northwesterly two-thirds of their total range, excluding the lower Wabash and the Miami and Scioto valleys.

Conclusion We have looked at evidence from linguistics, archaeology, and oral tradition that points to a migration, sometime between 1800 and 1400 B.C., of Proto- Algonquian speakers from the Columbia Plateau, where they appear as the Western Idaho Archaic burial complex, to the Midwest, where they show up as Red Ocher/Glacial Kame. Finally, we should ask what the Algonquians brought to the Midwest which led to their enormous success in spreading some aspects of their culture, especially Algonquian speech, through most of the Northeast. It seems to have been techniques for social integration which linked many local groups for peaceful exchange, on a scale beyond that which had been previously achieved in the Midwest. As is well known, this contribution of Red Ocher/Glacial Kame was amplified and extended first as Adena/Middlesex and then as the Hopewell interaction sphere. By offering participation in enormous social networks over the millenium of Red Ocher/Glacial Kame, the Proto-Algonquians encouraged other groups to adopt Algonquian speech.

REFERENCES Allen, Patricia 1982 The Oxbow Site: An Archaeological Framework for Northeastern . Pp. 119-133 in Approaches to Algonquian Archae­ ology. Margaret G. Hanna and Brian Kooyman, eds. Calgary: Uni­ versity of Calgary Archaeological Association. THE ALGONGUIAN MIGRATION 121

Bender, Barbara 1985 Emergent Tribal Formations in the American Midcontinent. Amer­ ican Antiquity 50:52-62. Berman, Howard 1982 Two Phonological Innovations in Ritwan. International Journal of American Linguistics 48:412—420. Boewe, Charles 1987 The Walam Olum and Dr. Ward, Again. Indiana Magazine of His­ tory 83:344-359. Braun, David P., James B. Griffin, and Paul F. Titterington 1982 The Snyders Mounds and Five Other Mound Groups in Calhoun County, niinois. University of Michigan Museum of Anthropology Technical Report 13. Ann Arbor. Butler, B. Robert 1976 The Evolution of the Modern Sagebrush-Grass Steppe Biome on the Eastern Snake River Plain. In Holocene Environmental Change in the Great Basin. R. Elston, ed. Nevada Archaeological Survey Research Paper 6. Reno. 1986 Prehistory of the Snake and Salmon River Area. Pp. 127-134 in Handbook of North American Indians, Vol. 11: Great Basin. W.L. D'Azevedo, ed. Washington: Smithsonian Institution. Denny, J. Peter 1989 Algonquian Connections to Salishan and Northeastern Archaeology. Pp. 86-107 in Papers of the Twentieth Algonquian Conference. Wil­ liam Cowan, ed. : Carleton University. Didier, Mary Ellen 1967 A Distributional Study of the Turkey-Tail Point. Wisconsin Archae­ ologist 48:3-73. Dyen, Isidore 1956 Language Distribution and Migration Theory. Language 32:611-626. Ehret, Christopher, and Merrick Posnansky 1982 The Archaeological and Linguistic Reconstruction of African His­ tory. Berkeley: University of California Press. Fiedel, Stuart J. 1987 Algonquian Origins: A Problem in Archaeological-Linguistic Corre­ lation. Archaeology of Eastern 15:1-11. 1989 Middle Woodland Algonquian Expansion: A Refined Model. Ms.

Fitting, James E. 1970 The Archaeology of Michigan. : Doubleday.

Fogel, Ira L. . , , , 1963 The Dispersal of Copper Artifacts in the Late Archaic Period of Prehistoric North America. The Wisconsin Archaeologist 44.129- 180. 122 J. PETER DENNY

Goddard, Ives 1975 Algonquian, Wiyot and Yurok: Proving a Distant Genetic Rela­ tionship. Pp. 249-262 in Linguistics and Anthropology in Honor of CF. Voegelin. M. Dale Kinkage, Kenneth L. Hale, and Oswald Werner, eds. Lisse: Peter de Ridder Press. 1978a Central Algonquian languages. Pp. 583-587 in Handbook of North American Indians, Vol. 15: Northeast; William C. Sturtevant, ed. Washington: Smithsonian Institution. 1978b Mascouten. Pp. 668-672 in Handbook of North American Indians, Vol. 15: Northeast, William C. Sturtevant, ed. Washington: Smith­ sonian Institution. 1978c Synonymy [Appended to William N. Fenton, Northern Iroquoian Culture Patterns]. Pp. 319-321 in Handbook of North American Indians, Vol. 15: Northeast, William C. Sturtevant, ed. Washington: Smithsonian Institution. Green, Thomas J., Max G. Pavesic, James C. Woods, and Gene L. Titmus 1986 The DeMoss Burial Locality: Preliminary Observations. Idaho Ar­ chaeologist 9:31-40. Haas, Mary R. 1958 Algonquian-Ritwan: The End of a Controversy. International Jour­ nal of American Linguistics 24:159-173. 1965 Is Kutenai Related to Algonquian? Canadian Journal of Linguistics 10:77-91. Havelock, Eric 1963 Preface to Plato. Cambridge, Mass.: Harvard University Press. Heckewelder, John 1819 History, Manners and Customs of the Indian Nations. Philadelphia: Abraham Small. Janetski, Joel C. 1987 The Indians of Yellowstone Park. Salt Lake City: University of Utah Press. Mason, Ronald J. 1981 Great Lakes Archaeology. New York: Academic Press. McElrath, Dale L., Thomas E. Emerson, Andrew C. Fortier and James L. Phillips 1984 Late Archaic Period. Pp. 34-58 in Archaeology. Charles J. Bareis and James W. Porter, eds. Urbana: University of Illinois Press. Moratto, Michael J. 1984 California Archaeology. Orlando: Academic Press. Morse, Dan, and Phyllis Morse 1964 1962 Excavations at the Morse Site: A Red Ocher Cemetary in the Illinois Valley. The Wisconsin Archaeologist 45:79-99. Munson, Patrick J. 1966 The Sheets Site: A Late Archaic-Early Woodland Occupation in West-Central Illinois. Michigan Archaeologist 12:111-120. THE ALGONGUIAN MIGRATION 123

Pavesic, Max G. 1985 Cache Blades and Turkey Tails: Piecing Together the Western Idaho Archaic Burial Complex. Pp. 55-89 in Stone Tool Analysis: Essays in Honor of Don E. Crabtree. Mark G. Plew, James C. Woods, and Max G. Pavesic, eds. Albuquerque: University of New Mexico Press. Power, Marjory 1989 The Isle LaMotte Site: A Revisionist's Look at Glacial Kame. Paper read at Society for American Archaeology, Atlanta. Reid, Kenneth C. 1984 Nebo Hill and Late Archaic Prehistory on the Southern Prairie Peninsula. University of Kansas Publications in Anthropology 15. Lawrence. Rhodes, Richard A. 1988 On the Classification of Central Algonquian. Paper read at the 20th Algonquian Conference, Ottawa. Ritzenthaler, Robert E., and George I. Quimby 1962 The Red Ocher Culture of the Upper Great Lakes and Adjacent Areas. Fieldiana: Anthropology 36:243-275. Sapir, Edward 1916 Time Perspective in Aboriginal American Culture: A Study in Method. Geological Survey of Memoir 90. Ottawa. Seeber, Pauleena M. 1982 Eastern Algonquian Prehistory: Correlating Lingusitics and Archae­ ology. Pp. 135-146 in Approaches to Algonquian Archaeology. Mar­ garet G. Hanna and Brian Kooyman, eds. Calgary: University of Calgary Archaeological Association. Siebert, Frank T. 1967 The Original Home of the Proto-Algonquian People Pp. 13-47 in Contributions to Anthropology: Linguistics I (Algonquian). Na­ tional Museum of Canada Bulletin 214. Ottawa. Snow, Dean R. 1976 The Archaeological Implications of the Proto-Algonquian Urheimat. Pp. 339-346 in Papers of the Seventh Algonquian Conference. Wil­ liam Cowan, ed. Ottawa: Carleton University. 1977 Archaeology and Ethnohistory in Eastern New York. Pp. 107-112 in Current Perspectives in Northeastern Archaeology. Robert E. Funk and Charles F. Hayes, eds. Albany: New York State Archaeological Association. 1980 The Archaeology of . New York: Academic Press. Spence, Michael W., and William A. Fox 1986 The Early Woodland Occupations of . Pp. 4-46 in Early Woodland Archaeology. Kenneth B. Farnsworth and Thomas E. Emerson, eds. Kampsville, IL: Center for American Archaeology. Stoltman, James B. 1986 The Archaic Tradition. The Wisconsin Archaeologist 67:207-238. 124 J. PETER DENNY

Stothers, David M., and James R. Graves 1985 The Prairie Penisula Co-Tradition: An Hypothesis for Hopewellian to Upper Mississippi Continuity. Archaeology of Eastern North Amer­ ica 13:153-175. Tuck, James A. 1977 A Look at Laurentian. Pp. 31-40 in Current Perspectives in North­ eastern Archaeology. Robert E. Funk and Charles F. Hayes, eds. Albany: New York State Archaeological Association. Voegelin, Carl F. et al, eds. 1954 Walam Olum or Red Score: The Migration Legend of the Lenni Lenape or Delaware Indians. Indianapolis: Indiana Historical So­ ciety. Wakefield, Francis 1966 The Elusive . Michigan History 50:228-234. Walker, Willard 1975 The Proto-Algonquians. Pp. 633-648 in Linguistics and Anthropol­ ogy in Honor of CF. Voegelin. M. Dale Kinkade, Kenneth L. Hale, and Oswald Werner, eds. Lisse: Peter de Ridder Press. Weslager, CA. 1972 The Delaware Indians: A History. New Brunswick: Rutgers Uni­ versity Press. 1978 The Delaware Indian Westward Migration. Wallingford, CT: Middle Atlantic Press. Whistler, K.A. 1979 Linguistic Prehistory in the Northwest California Culture Area. Pp. 11- 26 in A Study of Cultural Resources in Redwood National Park. P. McW. Bickel, ed. Denver: Report to the National Park Service. Winters, Howard D. 1967 An Archaeological Survey of the Wabash Valley in Illinois. Rlinois State Museum Reports of Investigations 10. Springfield. 1969 The Riverton Culture, niinois State Museum Reports of Investiga­ tions 13. Springfield.