More on Migration in Prehistory: Accommodating New Evidence in the Northern Iroquoian Case

by Dean R. Snow
More on Migration in Prehistory: Accommodating New Evidence in the Northern Iroquoian Case
Dean R. Snow
American Antiquity
Start Page: 
End Page: 
Select license: 
Select License



Dean R. Snow

Crawford und Smith have developed important new evidence thut beurs on the h~pothesis that the horthern Iroquoians migrated into the lo~jer Great Lakes region sometimr ajjer A.D. 900. Clarification of'the Princess Point Comple,~ in Onrurio forces u revision of'the h~pothesis. While an Appalachiun origin jflr the Northern Iroquoians and their subsequent migration is not rejected, neM' evidence strongly suggests that !he population ship rook place three centuries earlier than I previous!,' proposed. The situation calls ,for both further refinement of puleodemogruphic theorjl and new empirical research into

Owasco and other earlier Northern Iroquoiun comple.xes.

Crlrwjflrd y Smith hun desarrollado nuevu e importunte evidenciu con respecto a la hipdtesis de que 10s iroqueses del norte emigraron hacia el sur de lu regidn de 10s Grundes Lagos alrededor de YO0 d.C. Un reerumen del complejo Princess Point en Ontario amerita la revisidn de esta hiphtesis. Si hien no se rechaza un origen en 10s Apaluches para 10s iroqueses del norte ni .se descarta su migracidn suhsiguiente, la nueva evidenciu sugiere fitertemente que el camhio de pohlucihn ocurrib !re.\ siglos antes de lo anteriormente propuesto. La situacidn e-rige un mayor refinumiento de la teoriu puleodemogrrifica J' de las nuevas investiguciones sohre el Owusco y otros complejos de 10s iroqueses del norte.

very much appreciate the opportunity to com- complex to be in the A.D. 800-1000 period. Later ment on the article written by Gary Crawford researchers reassigned the two western expresand David Smith (this issue). They have sions to other traditions and extended the age of undertaken some much needed research in their the remaining Grand River expression of Princess effort to clarify the Princess Point Complex, and Point to A.D. 600 or 650 (Fox 1990). their recent findings necessarily force a major Furthermore, while Stothers (1976: 139, 158) rethinking of the evidence for Iroquoian origins made the Porteous site on the Grand River a com- and the advent of maize horticulture in the Lower ponent of the Princess Point Complex, Noble and Great Lakes region. That said, while the new find- Kenyon (1972) argued that it was probably an ings necessitate a revision in the specifics of the early Glen Meyer village site. So far as I am migration hypothesis I proposed in 1995, I con-aware, everyone agrees that Glen Meyer is an clude that they provide additional support for a archaeological construct found in roughly the migration hypothesis of some kind. The following same area as Princess Point, that it is later than discussion includes a revised hypothesis. Princess Point, and that it was associated with

The Princess Point Complex is currently a very Northern Iroquoian people. different construct from that initially defined by My main problem in applying the migration Stothers (1976, 1977) two decades ago. As cur- hypothesis to the Ontario evidence was in decid- rently used, the complex is comprised of sites ing whether Princess Point was part of the devel- found along the lower Grand River of southern opmental continuum begun by Iroquoian Ontario and the nearby shorelines of western immigrants and evolving into Glen Meyer and Lake Ontario and northeastern Lake Erie. later phases, or (alternatively) that it was an inde- Stothers initially referred to this as only one of pendent Middle Woodland period development three "foci" of Princess Point culture, the other that was not directly ancestral to Glen Meyer. I two lying westward in two areas between Lake was influenced by evidence that charred maize Huron and Lake Erie. He estimated the age of the kernels once attributed to Princess Point occupa-

Dean R. Snow .Department of Anthropology, Pennsylvania State University, University Park, PA 16802

American Antiquity, 61(4), 1996, pp. 791-796. 
Copyright C by the Society for American Archaeology 

tions at key sites were by 1990 thought by at least some investigators to have come from later occu- pations (Fox 1990: 178). Although Princess Point seemed to have appeared suddenly and without predecessors in the Grand River area (Stothers

1977: 155-158), from published accounts there seemed also to be a discontinuity between Princess Point and Glen Meyer. Given that Princess Point sites did not include the large com- pact village sites usually associated with early Iroquoians, that their ages often predated A.D. 900, and that doubt had been cast on the associa- tion of charred maize with Princess Point compo- nents, I decided that Princess Point was probably outside the continuum of Iroquoian development, not inside it. Crawford and Smith are quite right in saying that this conclusion depended on either negative or inconclusive evidence at the time I came to it.

Crawford and Smith's new data have changed matters considerably. They have shown with as much certainty as one can expect from archaeol- ogy that charred maize is indeed associated with Princess Point sites and that AMS dates carried out on the maize kernels indicate that the sites were occupied by at least A.D. 600, perhaps ear- lier. Furthermore, their recent work has shown that at least the Lone Pine and Grand Banks sites are the large compact village sites that we previ- ously thought were missing from the Princess Point Complex.

The findings of Crawford and Smith put Princess Point within and at the beginning of the long continuum leading to the historic Ontario Iroquois. Furthermore, they indicate that Princess Point is at least as old as, and perhaps older than, Clemson's Island culture in central Pennsylvania. If Stewart's dating of Clemson's Island sites is correct, the later Owasco sites in New York could be derivative from Clemson's Island but Princess Point cannot be. Thus the specific hypothesis that Owasco, Glen Meyer, Pickering, and other later Iroquoian expressions all derived from Clemson's Island must be modified.

Although the 1995 version of the migration hypothesis must be modified the anomalies that I cited in 1995 still persist. Matrilocality and matri- lineality do not develop slowly over time, they usually appear suddenly and as a feature of migrating communities having means of subsis- tence that allow for relatively large compact per- manent villages. The communities are typically large relative to those of the population(s) being displaced by the migrating one. Anomalies also still persist in Iroquoian historical linguistics. The speakers of Proto-Northern-Iroquoian knew about maize and had a vocabulary that strongly suggests origins in the Appalachian uplands. Moreover the site distribution data and ceramic data that I argued pointed to discontinuity in the sequence have also not gone away. The difference in the lat- ter is that Smith's reanalysis of Princess Point ceramics indicates that they are similar in basic construction to later Iroquoian ceramics and not fundamentally similar to Point Peninsula ceramics (Smith 1995). I inferred too much from Williamson's (1990295-298) statements contrasting Princess Point and Early Iroquoian ves- sels. The discontinuity is now clearly at the boundary between Princess Point and Point Peninsula, around A.D. 600 or earlier.

Where I must continue to disagree with Crawford and Smith is on the nature of the asso- ciation of horticulture, matrilocality, migration, and compact villages. Horticulture need not be associated with matrilocality, as cases from inte- rior Alaska attest. However, matrilocality and migration are often associated with compact vil- lages, and the latter must be supported by means of subsistence that are productive enough to allow their persistence over the long term. Thus, while matrilocality is often associated with horticulture, horticulture does not in itself produce matrilocal- ity (Divale 1984). This is contrary to assumptions in much earlier literature (e.g., Trigger 1978). I do not now conclude that Princess Point was just another Middle Woodland complex with a little maize added. What I do conclude is that the new Princess Point evidence puts the Iroquoian intru- sion into Ontario about three centuries earlier than I had previously thought. The new data have not erased the discontinuity in the sequence, they have pushed it back. Migration is still necessary to explain discontinuity in the archaeological record.

I am grateful that migration is no longer a taboo subject. It is one of a small number of essential demographic processes, and its denial does not facilitate realistic archaeological infer- ence in the long run (Anthony 1990). Bogucki's very recent synthesis of the evidence for the spread of agriculture into Europe is additional evidence of the new appreciation for a realistic assessment of paleodemography. If farming entered Europe from Anatolia by way of Greece, then a major issue becomes whether the process involved colonization or the spread of domesti- cated plants and animals to indigenous foraging groups. The answer appears to be that both processes were going on, and that archaeologists can often detect which one characterized any par- ticular case. Colonizing farmers moved into flood plain habitats not much used by indigenous for- agers, and their sites appear without precedent in the seventh millennium B.C. Elsewhere in Greece there is evidence for continuity on forager sites in terms of most of their artifactual inventories, with the addition of domesticated plants and animals. In the former case there is archaeological discon- tinuity, in the latter there is continuity, and the two

can be distinguished (Bogucki 1996).

Returning to the Iroquoian case, there is addi- tional evidence from the Eastern Woodlands that deserves mention. It is clear from maps published by Anderson (1991) that the spatial distribution of archaeological phases in the Eastern Woodlands was very spotty over the last thousand years of prehistory. A recent revision of those maps con- firms that there were large buffer zones in the region and that core areas defined by site clusters shifted in space over time (Milner et al. 1992). Anderson's (1 994) examination of Savannah River chiefdoms reveals cycles of subregional flo- rescence and abandonment that can only be explained in terms of dynamic demographic processes that must have included migration. The signatures of those processes are repeated many times over in the region. Of specific interest to the Iroquoian case is the sudden appearance of the Cashie phase in North Carolina sometime around

A.D. 800 (Anderson 1991 :12; Phelps 1983:43). This phase persists in modified form and without discontinuity into the seventeenth century, and is identified with the historic Tuscarora, Meherrin, and Nottoway, all of them Northern Iroquoian speakers. This branch of Northern Iroquoians rep- resents, on linguistic grounds, the oldest diver- gence from Proto-Northern-lroquoian. All other differentiation of Northern Iroquoian languages must have taken place after the separation of ancestral Tuscarora, Meherrin, and Nottoway. The associated archaeological evidence (i.e., the Cashie phase) indicates that this must have taken place by at least A.D. 800, perhaps a century or two earlier. Northern Iroquoian glottochronology, such as it is, supports this inference.

Glottochronology continues to be alternately used and condemned by archaeologists and lin- guists. Although it has worked in specific cases, it has failed spectacularly in others. One problem appears to be that the persistence of linguistic homogeneity over large areas is density depen- dent (Shaul 1986). A thin population of foragers might maintain a widespread common language for centuries, and in this kind of situation one should expect glottochronology to fail. While lin- guistic diversification may be maladaptive for low-density populations, it may characterize locally dense populations of horticulturalists. Unfortunately, historical linguists have typically given up on glottochronology rather than make an attempt to refine it by finding ways to control for the density effect and other possible variables. Based on archaeological tests, glottochronology does appear to work reasonably well for the last millennium of prehistory in the portion of the Eastern Woodlands lying south of the subarctic (Fiedel 1987, 1991). My hope is that historical linguists will take a fresh look at demographic factors and attempt to refine a technique that in my judgment is not hopelessly flawed.

The new evidence from Princess Point sites suggests that the breakup of Northern Iroquoian languages must have begun at least three centuries earlier than I previously thought. In other words, it now appears likely that Princess Point, Clemson's Island and Cashie all represent early Northern Iroquoian speech communities. On the basis of current evidence, Owasco still looks like it is derivative from Clemson's Island for the accepted dates for Early Owasco sites in south- central New York average around A.D. 900. However, there are at least four radiocarbon dates from Owasco occupations in that area that cali- brate to earlier than A.D. 900 (Funk 1993: 158-171; Wurst and Versaggi 1993). The

earliest of these is a date of 1425 k150 B.P. (QC- 1001), which calibrates to 644 cal A.D. with the program CALIB 3.0.3 (Stuiver and Reimer 1993). Funk, Wurst, and Versaggi discount this very early date as well as three dates calibrating to the eighth century A.D. However, this rationalization of early dates might have more to do with expec- tations arising from the generally accepted ortho- dox New York sequence than with a critical appraisal of the age determinations, the samples from which they derived, and their associations. I have argued elsewhere (Snow 1995:65) that the Hunter's Home phase, the last in the Point Peninsula tradition, is a spurious construct. If that phase never existed and it is actually based on an archaeological amalgam of earlier Point Peninsula and intrusive Early Owasco remains as I have argued, then perhaps the growing number of radiocarbon dates calibrating to the period A.D. 644-900 should be taken at face value rather than dismissed. If the Hunter's Home phase never existed, then Owasco sites in south-central New York dating to the eighth century or even earlier should not be unacceptable. Indeed, the early dates are consistent with dates for early Clemson's Island sites and (thanks to Crawford and Smith) what we now know to be similar dates for Princess Point sites.

If Owasco eventually proves to be as old as Clemson's Island and Princess Point, then it too cannot be derivative from Clemson's Island. My hypothesis will have to be modified again, and I will do it without regret. Either way, a more basic question remains: where did all of these complexes (including Cashie) come from'? What per- sists from my original hypothesis is the argument that they cannot have come from the Point Peninsula tradition. What is new is that some and perhaps all of them arrived earlier than we had realized.

What is thus also new is that the Princess Point evidence makes it likely that the initial spread of horticulture (and horticulturalists) into southern Ontario was not simply a consequence of the onset of the Medieval warm epoch, for this climatic episode did not begin until ca. A.D. 1000 (Ingram et al. 198 1). If Owasco proves to be similarly older than currently thought, then we have additional evidence leading to the same observation.

I remain mindful that general acceptance of the migration hypothesis, or some further revision of it that might be made necessary by still more new data, depends on prior acceptance in princi- ple of some important demographic processes. The spottiness of Middle and Late Woodland pop- ulations and their mobility over the long term are necessary assumptions. So too is the assumption that there was linguistic and other cultural conti- nuity over time within the shifting clusters mapped by Anderson ( 1991)and others (Milner et al. 1993). That said, it is worth noting that the his- toric Cheyenne case provides both a supportive example and a cautionary tale. In the seventeenth century the Cheyenne were wild rice gatherers living in temporary camps in what is now north- ern Minnesota. By the middle of the eighteenth century they were sedentary maize horticultural- ists living in southern Minnesota and the Dakotas. By the early nineteenth century they were mounted buffalo hunters living in nomadic tipi dwellings on the Plains. Archaeologists have pre- tended for too long that this kind of movement over time and space was a special and generally unprecedented consequence of the rapid displace- ments, population declines, and technological changes brought on by contact with Europeans. It may be that normal demographic processes were accelerated andlor magnified by European contact, but they were surely not created by it without precedent. The same criticism can be leveled at traditional archaeological inferences regarding the presumed lack of demographic dynamics in northern Europe before contact with the Roman armies that supposedly both set them in motion and documented the process. I argue that pale- odemography is not exempt from uniformitarian principles.

The cautionary part of the Cheyenne case has to do with the rapid, perhaps even revolutionary, transformation that Cheyenne culture went through at least twice in only two centuries. One could not hope to track such an evolution archae- ologically without the assistance of documentary sources. There are at least two great discontinu- ities in the Cheyenne archaeological record despite documented cultural continuity over time (Moore 1987:82-87). Moreover, the transforma- tions they experienced each involved pervasive cultural reconstruction from elements derived from two or more antecedent societies, a rhizotic form of ethnogenesis that is very different from the cladistic model often used to describe the inferred development and diversification of cultures over time (Moore 1994). We would not necessarily be able to assume continuity in various Northern Iroquoian sequences after A.D. 600 were it not present in the archaeological record. Fortunately for us that continuity is present. More important, perhaps, is that we cannot necessarily expect to find a clear common archaeological ancestor for the related Princess Point, Clemson's Island Owasco, and Cashie complexes. Lexical items in Proto-Northern-Iroquoian and the deeper relationship between Northern Iroquoian languages and the sole surviving Southern Iroquoian language (Cherokee) all point to origins in the Appalachians. Proto-Northern-Iroquoian has words for red oak, hickory, maple, elm, basswood and pine, but words for northern species like birch, hemlock, fir, and tamarack were added later to descendant languages. E~~~ the historic distribution of those languages points to origins in the central Appalachians. Unfortunately, their

adaptive radiation out of their homeland might have involved speed and transformational change like that of the historic Cheyenne, leaving us at least for now with no means to identify their common archaeological origins with reasonable certainty. We may search for them, and we might even find them, but we should not be surprised if we do not.

I suspect and hope that Crawford and Smith will find much to agree with in all of this. They have made an important empirical contribution to Iroquoian archaeology. What seems to me to be most needed now is further refinement of the theoretical issues I have raised and new empirical studies to better define the beginnings of Owasco in New York. If the prehistoric Iroquoians force us to deal with those issues more productively, they too will have done another service for archaeology.

References Cited

Anderson, D. G.

1991 Examining Prehistoric Settlement Distribution in

Eastern North America. Archaeolog>, of Eustern North

America 19:1-22.

1994 The Savunnuh River Chiefaoms: Political Change in

the Late Prehistoric Sourheast. University of Alabama Press. Tuscaloosa. Anthony, D. W. 1990 Migration in Archeology: The Baby and the Bathwater. American Anthropologisr 92:895-914. Bogucki, P. 1996 The Spread of Early Farming in Europe. American Scientist 84:242-253. Divale, W. 1984 Matrilocul Residence in Pre-Literate Socieh.. UMI Research Press, Ann Arbor. Fiedel, S. J.

1987 Algonquian Origins: A Problem in Archeological-Linguistic Correlation. Arc,haeology of Eustern ~Vorth America 15.1-11.

1991 Correlating Archaeology and Linguistics: The Algonquian Case. Man in the Northeast 41 :9-32. Fox, W. A. 1990 The Middle to Late Woodland Transition. In The Archaeology ofSouthern Ontario ro A.D. 1650, edited by

C. J. Ellis and N. Ferris, pp. 171-188. Occasional Publication of the London Chapter No. 5. Ontario Archaeological Society, London.

Funk, R. E. 1993 Archaeological Investigations in the Upper Susquehanna Vallet. New York Srute, vol. 1. Monographs in Archaeolog;,~Persimmon Press, I~M., G. F ~ ~ ~ ~, ~ ,

and T,~wigley1981 Past Climates and Their Impact on Man: A Review. In CIim~feand Histor), edited by T. Wigley, M. Ingram, and

G. Farmer, pp. 3-50. Cambridge University Press, Cambridge. Milner, G. R.,-D. G. Anderson, and M. T. Smith

1992 The Distribution of Eastern Woodlands Peoples at the Prehistoric and Historic Interface. Paper presented at the 57th Annual Meeting of the Society for American Archaeology, Pittsburgh, Pennsylvania.

Moore, J. H. 1987 The Cheyenne Nurion: A Sociul and Demographic Histoy. University of Nebraska Press, Lincoln. 1994 Putting Anthropology Back Together Again: The Ethnogenetic Critique of Cladistic Theory. American Anrhropologisr 96:925-948.

Noble, W. C., and I. T. Kenyon 1972 Porteous (AqHb-1): A Probable Early Glen Meyer Village in Brant County, Ontario. Ontario Archueolog>, 19:ll-38.

Phelps, D. S. 1983 Archaeology of the North Carolina Coast and Coastal Plain: Problems and Hypotheses. In The Prehistory of North Carolinu: An Archaeologicul S>,mposium, pp. 1-51, edited by M. A. Mathis and J. J. Crow. North Carolina Department of Cultural Resources, Division of Archives and History, Raleigh.

Shaul, D. L. 1986 Linguistic Adaptation and the Great Basin. American Antiquih. 51:415416. Smith, D. G.

1995 Cord-Marked Pottery and the Early Late Woodland in the Northeast. Paper presented at the 60th Annual Meeting of the Society for American Archaeology, Minneapolis, Minnesota.

Snow, D. R. 1995 Migration in Prehistory: The Northern Iroquoian Case. American Antiquih. 60:59-79.

Stothers, D. M.

1976 The Princess Point Complex: A Reg~onal Represen- tative of an Early Late Woodland Horizon in the Great Lakes Area. In The Late Prehistor? of rhr Luke Erie Drainage Basin: A 1972 Symposium Revised. pp. 137-161, edited by D. S. Brose. Cleveland Museum of Natural History. Cleveland.

1977 The Princess Point C'ompler. Mercury Series Paper No. 58. Archaeological Survey of Canada, National Museum of Man. Ottawa. Ontario.

Stuiver, M., and P. J. Reimer 1993 Extended "C Data Base and Revised Calib 3.0 "C Calibration Program. Radioc,arhon 35:215-230. Trlgger, B. G. 1978 lroquoian Matriliny Penris~~lvan~tr



Williamson, R. F. 1990 The Early Iroquoian Period of Southern Ontario. In The Archaeolog~, of' Southern Onrurio ro A.D. 1650, edited by C. J. Ellis and N. Ferris, pp. 291-320. Occasional Publication of the London Chapter No. 5. Ontario Archaeological Society, London.

Wurst. L . and N. M. Versaggi 1993 Under the Asphalt: The Archaeology of the Binghamton Mall Project. Unpublished report of the Public Archaeology Facility, Binghamton University. On file at Binghamton University. SUNY.

Received IClny 13. 1996: uccepred .\./a), 13. 1996

VISA and Mastercard accepted. Shipping Ph: 822-5959, Fax: 1-800-668-0821 6344 Memorial Road, Vancouver BC V6T http://w.ubcpress .ubc. ca


uses his love of the north and his fascination with its

people to bring the Palaeo-Eskimos back to life. From the

hints they left behind, McGhee reconstructs their lives at

the margins -he describes how they hunted, stored

food, and made clothes and tools. In addition, the many

art objects found at the sites, preserved by the extreme

cold of their environment, permit a unique glance at an

ancient people's religion and view of the world.

Richly illustrated with maps and photographs of many of the artifacts and the landscape in which they were found, this book presents an evocative picture of the first Arctic inhabitants.

'A tale of epicproportions ... will appeal to anyone with

an interest in the North and its history.'

Iohrz Wilson, Quill & Quire

256 pages, hc $35.95



  • Recommend Us