The Settlement of the Americas

CURRENT ANTHROPOLOGY Vol. 27, No. 5, December 1986 ? 1986 by The Wenner-GrenFoundation for Anthropological Research, all rightsreserved 001 1-3204/86/2705-0004$2.25

The Settlementof the Americas: A Comparison of the Linguistic, Dental, and Genetic Evidence

byJoseph H. Greenberg,Christy G. TurnerII, and StephenL. Zegura

A CONSIDERATION OF THE BIOLOGICAL evidence regarding the restmainly on the absence of human skeletalremains earlier hominid settlementof the Americas leads to the following basic than the terminalPleistocene, the biologicalresemblances be- conclusions: (1) It did not occur prior to the terminal Pleis- tween Amerindianand Asian populations,and the fact that tocene. (2) It was carried out by evolutionarily modern Homo human dental variationin the Americasis less than that in sapiens sapiens. (3) It had its origin in migration fromAsia via Asia, suggestingthe relativerecency of Americansettlement. the Bering Strait. These views have had wide acceptance and If the foregoingstatements are taken as a set of working hypotheses,a series of major questionsof greaterspecificity immediatelyarises. These includethe numberof migrations to be postulated,the ethnicand archaeologicalidentity of each, JOSEPHH. GREENBERGis Ray Lyman Wilbur Professorof the Social Sciences in Anthropologyemeritus at StanfordUniversity and theirrelative and absolutechronology. In developinghy- (Stanford,Calif. 94305, U.S.A.). Born in 1915, he did his under- pothesesregarding these questions, we will turnto thelinguis- graduatework at Columbia (B.A., 1936) and receiveda Ph.D. in tic,dental, and geneticevidence, in thatorder. The threelines anthropologyfrom Northwestern (1940). He has been electedto the ofevidence agree that the Americas were settled by threesepa- National Academyof Sciences,the American Academy of Arts and rate populationmovements whose identitycan be most pre- Sciences, and the American PhilosophicalSociety. His research ciselyexpressed in linguisticterms as Amerind,Na-Dene, and interestsare chieflyin linguistics,including typology and universals and the historicalclassification of languages. Amonghis publica- Aleut-Eskimo. tionsare The Languageof Africa (Bloomington: Indiana University These threetypes of evidence are logicallyindependent in Press, 1963); AnthropologicalLinguistics (New York: Random the sense thatthere are no logical constraintsoperating across House, 1967),and thefour-volume collection Universals of Human anypair ofthem. It is obvious,for example, that the markers n Language (Stanford:Stanford University Press, 1978). forthe first person singular and m forthe second person singu- CHRISTYG. TURNERII, Professorof Anthropology,Arizona State lar, bothof which are, in fact,very widespread in theAmerind University(Tempe, Ariz. 85228 U.S.A.), was born in 1933 and linguisticstock, have no necessaryconnection with the pres- educated at the Universityof Arizona (B.A., 1957; M.A., 1958) ence ofA and B genesat thesame geneticlocus or thepresence and the Universityof Wisconsin,Madison (Ph.D., 1967). Awards or absence of incisorshoveling. There is, however,one source includea 1970-71 fellowshipat the Centerfor Advanced Studyin of bias thatneeds to be takeninto account. If the investigator the Behavioral Sciences, Stanford; IREX Senior Scholar to in one fieldis aware ofthe conclusions proposed in another,he U.S.S.R. Academy of Sciences (1980-81, 1984); Arizona State Uni- versityResearch Professor(1984-85); and Museum of Northern or she may be influencedby this knowledgein developinga Arizona Life Fellow (1985). His more than 100 publicationsare theory.It need not, of course,have this effect.The ultimate based on archeologicalexcavations in the Aleutians since 1962, testis whetherthe scientistis able to justifythe conclusions on skeletalstudies in North and South America,Asia, Europe, and the basis of data drawnexclusively from the field itself. In the the U.S.S.R., and otherresearch. presentinstance the linguistic researcher, on theone hand, and STEPHENL. ZEGURAis Associate Professorof Anthropologyand the geneticand dental investigators,on the other,carried out Genetics at the Universityof Arizona (Tucson, Arizona 85721, theirwork independently. Since dentaland geneticstudies be- U.S.A.). Born in 1943, he was educated at StanfordUniversity long to the same basic field,physical anthropology, it is not (B.A., 1961) and the Universityof Wisconsin(M.S., 1969; Ph.D., surprisingthat each was aware of relevanttheories in the 1971). His researchinterests are humangenetics, evolutionary theother'sdomain. That theirconclusions were reached exclu- ory,biological systematics,statistics, and the generalhuman biol- sivelyon thebasis of internalevidence within the field itself is, ogy of Native Americanpopulations. Among his recentpublica- tionsare "The InitialPeopling of the Americas: An Overviewfrom however,sufficient indication of theirindependence. the Perspectiveof PhysicalAnthropology" (Acta Anthropogenetica 8:1-21), "MultipleObservers, Humidity, and Choice of Precision Statistics:Factors Influencing Craniometric Data Quality"(Ameri- THE LINGUISTIC EVIDENCE can Journal of Physical Anthropology61:85-95), and "The In- fluenceof Recent Advances in EvolutionaryBiology on Human There is only one comprehensiveclassification of the indige- PopulationBiology" (Symposia Thracica 6). nous languages of the Americas(Greenberg 1960, n.d.). The The presentpaper was submittedin finalform 15 iv 86. major alternativeand one which has become increasinglyin-

Vol. 27 * No. 5 * December1986

This content downloaded from 128.143.023.241 on August 23, 2016 15:09:18 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). fluentialin the recentperiod would involvethe acceptanceof catalogingof numeroussupposedly independent stocks with- somethinglike 200 independentlinguistic stocks (Loukotka out any attemptat synthesis.The most recentwork in that 1968, Campbell and Mithun 1979a). What is being tacitly traditionis Loukotka (1968), which lists 118 independent claimed by the proponentsof this pointof view is that,if we families.Some of these are extinctand have leftno linguistic take any threeof these stocksat random,we will neverfind tracesexcept, in some instances,in place names,and a feware thattwo of themare significantlymore similar to each otherin only known froma handfulof lexical itemsand are now ex- regardto traitsrelevant to a historicalgenetic classification. If tinct.Even subtractingthese, the numberof distinctfamilies this claim is accepted, two alternativetypes of nonlinguistic remainsvery large. historicalconclusions are possible. One is that each of these In a paper deliveredin 1956 (Greenberg1960), Greenberg representsa separate migration,thus requiringa trafficcon- tentativelyadvanced the hypothesisthat all the languagesof trollerat theBering Strait. The otheris thatthere are relatively South America fall into three groups, Ge-Pano-Carib,An- fewmigrations or even a singleone but thatthe time elapsed is dean-Equatorial,and Chibchan-Paezan,the last of which,in so vast thatall tracesof affinityamong any of thegroups have accordancewith the views of earlierscholars, even Loukotka, been effaced. had a large extensioninto the neighboringareas of Central A briefsurvey of the generaltrends in the classificationof America.This was part of a largerhypothesis that all threeof Americanlanguages will help to put the presentproposal in theseSouth American groupings along witha major portionof historicalperspective. Initially, it maybe notedthat, except for the languages of North America belonged to one large thework of Greenbergand thatof Swadesh whichpreceded it Amerindfamily. Details regardingthe membershipand the but did notresult in a preciseclassification, linguistic classifica- internalsubgroupings of the newlyproposed South American tion has proceededseparately in North and South America, families were given. The only groups not belonging to with some Central American languages being included in Amerindwere stated to be Na-Dene and Aleut-Eskimo,so eitherinstance when relationsacross the border were obvious. that all the languages of the Americasfell into threegroups. For NorthAmerica the pointof departurehas been Powell The paper itselfwas a briefone in whichthe classificationof (1891), whichdistinguished 58 stocksin NorthAmerica, some South American languages was presentedin tabular form of which extendedinto Mexico. A numberof Americanan- withoutsupporting evidence and other groups were merely thropologicallinguists, including Harrington, Dixon, Kroeber, named. Whorf,Trager, and Sapir, beginningearly in the 20thcentury This paper was not publisheduntil four years later. In the includedmany of Powell's stocksin moreextensive groupings meantimeLamb (1959) independently,but withoutincluding of whichthe mostimportant were Hokan, Penutian,and Az- SouthAmerica, arrived at a classificationidentical with that of teco-Tanoan. This period of consolidationreached its climax Greenberg,namely, that the languages of North America in Sapir (1929), a classificationin whichsix basic stocksnorth could be divided geneticallyinto three groups: Aleut-Eskimo, of Mexico were distinguished,with, in some instances,mem- Na-Dene, and the rest.As withGreenberg, no supportingevi- berssouth of the United States border. Sapir's six families were dence was presented. I, Eskimo-Aleut;II, Algonkin-Wakashan;III, Na-Dene; IV, In a laterpaper (1979), Greenbergpresented the same basic Penutian; V, Hokan-Coahuiltecan;and VI, Azteco-Tanoan. thesis in a somewhat revised forminsofar as it related to In theperiod following its publication and up to about 25-30 Amerindand included details on NorthAmerica. Some sup- yearsago, some investigators,who acceptedthe basic premise portinggrammatical evidence was presentedin regard to thatthere was a verylimited number of stocks north of Mexico Amerind.Moreover, Aleut-Eskimo,for which various rela- and also the validityof at least a substantialportion of Sapir's tionshipsin northernAsia had oftenbeen suggested,was as- classification,began to publishstudies in whichevidence was sertedto belongto a Eurasiatic groupingconsisting of Aleut- presentedfor relationshipscutting across Sapir's groupings. Eskimo, Chukotian, Gilyak (Nivkhi), Japanese, Korean, Instancesof theseare Haas (1958), linkinga portionof Sapir's Ainu, Altaic, Uralic-Yukaghir,and Indo-European. Hokan-Coahuiltecan,namely, the Gulflanguages, to theAlgic Beginningabout 1960, Greenbergbegan to compilea vast group(Algonkin, Wiyot, and Yurok), belongingto Sapir's Al- data base on the vocabularyand grammarof Amerindlan- gonkin-Wakashan;Shipley (1957), comparing Yukian in guages. The vocabularydata are containedin 23 notebooks GroupV of Sapir's classificationwith Penutian (IV); and New- reproducedin the StanfordLibrary and available on interli- man (1964), seekingto prove the affiliationof Zuni (put in VI braryloan. The detailed resultsin Greenberg(1986) include witha query)to the Penutiangroup. etymologiesfor each of the 11 subgroupsof Amerind,totaling It is perhaps partly,at least, because of these and similar approximately1,900, and 300 involvingtwo or moreof these hypothesesthat the last two decades have witnesseda skepti- 11 subgroups. A chapter on grammaticalmarkers contains cism among a substantialportion of specialistsin American approximately100 items. Indian languages even of groupingswhich had hithertobeen The 11 subgroupsare (1) Macro-Ge, (2) Macro-Panoan,(3) consideredsecure, e.g., Hokan in the narrowersense. There Macro-Carib, (4) Equatorial, (5) Macro-Tucanoan, (6) An- have been negative articles, notably Levine (1979), which dean, (7) Chibchan-Paezan,(8) CentralAmerind, (9) Hokan, seeks to disprovethe affiliationof Na-Dene with Haida, its (10) Penutian, and (11) Almosan-Keresiouan.These groups mostdistant branch. This trendreaches its climax in Campbell may be brieflyoutlined in the followingmanner: Macro-Ge is and Mithun(1979a), which, in spiteof itstitle The Languagesof essentiallyas in Greenberg(1960) withthe inclusionof Nam- Native America,embraces only North America with some ex- bicuara. Macro-Panoan is also virtuallyidentical with that tensionsinto Mexico and CentralAmerica. In theirintroduc- subfamilyas outlinedin the same source,as are Macro-Carib torycontribution the editorslist 62 separate stocks,a larger and Andean. The formerEquatorial is splitinto two groups, numberthan that of Powell. In several instancesdoubts are one consistingof the languages most closelyrelated to Tuca- expressedconcerning the validity of even someof these, so that noan and the other of the remainder.Chibchan-Paezan in- the numberis probablyeven larger.It is clear, however,that cludessome additionalmembers, notably Huarpe in thesouth, some of the contributorsto the volume do not espouse this Tarascan in the north,and Timucua, an extinctlanguage of generalpoint of view. Florida includedin thePaezan branchand mostclosely related As indicatedearlier, the classificationof the languages of to the languages of Venezuela. Central Amerindhas three South Americahas been carriedout in isolationfrom those of coordinatebranches: Oto-Mangue, Uto-Aztecan, and Kiowa- NorthAmerica. Certain low-level stocks of extensive member- Tanoan. Hokan is essentially "traditional" Hokan- ship were early recognized,e.g., Arawakan, Carib, and Ge. Coahuiltecan,Sapir's V. 1. Penutianembraces, besides the lan- However, the virtually exclusive traditionhas been the guages generally reckoned as Penutian (including here

478 CURRENT ANTHROPOLOGY

This content downloaded from 128.143.023.241 on August 23, 2016 15:09:18 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). Mexican Penutian),Yukian and the Gulflanguages. Northern Greenberg,Turner, and Zegura: SETTLEMENT OF AMERICAS Amerindhas two branches: one of these is identical with Sapir's Algonkin-Wakashanand the otheris Keresiouan,con- likelihoodthat theyare branchesof a singlelinguistic stock. sistingof Keresan, Siouan, Yuchi, Caddoan, and Iroquoian. The determinationof still deeper relationshipsregarding the There are some clear subgroupingsamong the 11 groups. affiliationsof Eskimo-Aleutand Na-Dene would, of course, One consistsof Macro-Ge, Macro-Panoan, and Macro-Carib constituteimportant evidence concerningthe point of Old and can be equated with the Ge-Pano-Carib of Greenberg World originof thesegroups. There are some indicationsthat (1960). A second consistsof Equatorial and Macro-Tucanoan, Eurasiaticand Amerindare closergenetically than either is to thusforming a substantialportion of Equatorial-Andeanof the Na-Dene. On the basis of a suggestionof Sapir, Shafer(1952) formerclassification. A northerngroup consists of Hokan, presentedsome evidence for a connectionof Na-Dene with Penutian,and Keresiouan.There may be a special connection Sino-Tibetan.Both of theseproposals require further investi- between Andean and Chibchan-Paezan. Central Amerind gation.If theyshould prove to be true,they would pointto a seems to stand somewhatapart fromthe rest. moresoutherly ultimate place of originin Asia. The followinghistorical inferences concerning the settlement Three linesof evidencewhich can be used to arriveat abso- of the New World may be derivedfrom the above classifica- lutedates are prehistory(by C14and othermeans), phylogenies tion: There were threemigrations, or at least only threeleft based on geneticdata, and glottochronology,a purely linguis- linguistictraces. The oldestis probablyAmerind, since it cen- tic technique.The last will be consideredhere in connection ters fartherto the south than the othersand shows greater with the linguisticevidence, but, for reasons which will ap- internaldifferentiation. On the other hand, the absence of pear, it is the least cogentof the three. sharperlinguistic differentiation in the northsuggests a rela- Glottochronology,the one available methodfor dating the tivelyrapid spread allowing no opportunityfor the develop- separationof linguistic stocks in theabsence of written records, mentof one or morestrongly divergent subgroups. Should it, was firstdevised about 1950 (Hymes 1960). By examiningthe however,prove valid that CentralAmerind forms a separate rate of retentionof a specificlist of 200 words it was deter- subgroupfrom all the rest,there would have to be a period minedthat slightly more than .80 of thislist was retainedover relativelyfar to the northduring which this divergencetook 1,000years. If two languagesdiverged from a singleancestral place. language and evolved independently,after 1,000 years they Na-Dene has deeperinternal divisions and is geographically would be expectedto have relatedforms in approximately.802 less peripheralthan Aleut-Eskimo.In particular,the break of the list. The process was hypothesizedto follow a decay betweenHaida and the remainderof Na-Dene requiresa con- functionso that aftern millenniathe proportionof cognate siderabletime depth. The internaldivisions of Na-Dene are formsbetween two languageswould be .82n* (1a) Athapaskan-Eyak, (1 b) Tlingit, and (2) Haida. The Unfortunately,there are several major problemswith this spread of the Athapaskansfrom interior Northwest Canada method.To beginwith, the documented cases suggestthat the and Alaska to California,Oregon, and the Southwestmust be probableerror is quite high. Second, the separationdates for veryrecent, probably within the last 1,000 years. The largest longperiods are seriouslyunderestimated by theassumption of numberof distinctsubgroups of Athapaskanis in the north- a homogeneousrate of retentionof the entirelist; it is well west,and its nearestrelative is Eyak, spokennow or formerly establishedthat certain words such as personalpronouns have in two southernAlaskan coastal areas, theCopper RiverDelta a far higherretention rate than others. An alternativein- and Yakutat. The geographicallocations of Tlingitand Haida homogeneousrate was proposedby Joos (1964). This was ar- furtherreinforce the thesisthat the originsof Na-Dene are to rived at by splittingthe originallist into eightsublists with be soughtin thegeneral area ofsoutheastern Alaska and north- differingretention rates. Because of incompletedata, a shorter ern BritishColumbia. list of 100 words with a retentionrate of .86 per millennium Aleut-Eskimois probablythe mostrecent. Its primaryinter- was moreoften used, and even thislist was oftenincomplete. nal divisionis less deep thanthat within Na-Dene. One indica- The rate constanthas also been revised and recalculateda tion is that no one has publiclydoubted the relationshipbe- numberof times. Moreover,when the time of separationis tween Aleut and Eskimo whereas, as has been seen, the great(i.e., approaches 10,000 years) differencesof judgment affiliationof Haida to the restof Na-Dene has been seriously regardingeven one cognatewill produce substantially different questioned.Both the internaland the externalevidence point results.Because of theseand otherproblems, the methodhas to the originalhabitat of the Aleut-Eskimoas being on the been almostcompletely abandoned forpurposes of the dating southwesterncoast of Alaska. WithinEskimo itself,the vast of timesof origin of ancestrallanguages, though it is stillocca- Inuitdistribution from Central Alaska to Greenlandwith shal- sionallyused for arrivingat the subgroupingsof a linguistic low internaldifferences suggests a veryrecent migration from stock (Hattori 1973, Miller 1984). For the linguisticfamilies the far westernend of the presentdistribution. The other withwhich we are concernedhere, no substantiallynew counts branch,Yuit, is foundin the centraland southwesterncoastal have been publishedduring the last two decades, althoughold areas of Alaska. It is also spokenin Siberia. If Sirenik,which data and reinterpretationshave been discussedand reworked in certainrespects is drasticallydifferent from the restof Es- (Krauss 1973). kimo, should prove to have separate geneticstatus, it even With all these reservationsin mind, we may considerthe becomes plausible that Eskimo or even Aleut-Eskimoorigi- relevantresults. Dates forthe Aleut-Eskimodivergence have nated in northeasternSiberia. However, the positionof Aleut ranged from2,900 to 5,600 B.P. and have tended to cluster suggestsrather an origineast of Bering Straitwith Siberian about 4,000 B.P. This variabilitydepends for the mostpart on Eskimo as a subsequentreflux. The externalevidence, with changesin the mathematicalassumptions of glottochronology Aleut-Eskimoon the extremeeastern end of a vast Eurasiatic ratherthan a moreextended or reliabledata base. In regardto distribution,likewise supports an originin eitherthe extreme Na-Dene, Swadesh arrivedat 9,000 B.P. as the date of Proto- northwestof the Americas or northeasternSiberia. Although it Na-Dene by comparingHaida, the most distant language, seems mostprobable linguistically that the Aleut-Eskimomi- with otherlanguages of the group; however,in subsequent grationwas themost recent, it is defensibleto hypothesizethat publicationshe drasticallyreduced the date to 4,700 B.P. the Proto-Aleut-Eskimocommunity arrived in the extreme (1959, 1962). The earlier date would seem more probable. northwestas a stillinternally undifferentiated unit beforethe Krauss (1973) estimates5,000 B.P. forthe Tlingit-Athapaskan comingof the Na-Dene. divergence,a much closer relationshipthan that of Haida to We have hypothesizedthat the threelinguistic stocks repre- theother Na-Dene languages. We are even less secureregard- sentseparate migrations. They differgreatly, and thereis little ing dates forAmerind, which would have to be based on com-

Vol. 27 * No. 5 * December1986 479

This content downloaded from 128.143.023.241 on August 23, 2016 15:09:18 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). parisons of languages that belong to differentsubgroups in provenienceof the 24 series. Table 2 and figure1 summarize Greenberg'sclassification. Here again thework has been done sample informationfor each series. by Swadesh, but resultssubsequent to Swadesh (1958) are un- Four observationsare noteworthy:(1) All New World usable because the actual cognatepercentages and the sizes of groupsresemble each othermore than they do mostOld World the listsused are not given. Moreover,these unpublished lists populations(a small MMD indicatesgreater similarity than a have been subjectto mathematicalmanipulations which have largervalue). (2) Dental variationis greaterin thenorth than in not been specifiedin detail. In addition,Swadesh tended to thesouth. (3) New Worldgroups are morelike Asians than like see more related formsthan otherlinguists. His figuresfor Europeans. (4) Aleut-Eskimos,Greater Northwest Coast In- Amerindlanguages belongingto differentsubgroups in the dians (Na-Dene), and all otherIndians (Macro-Indian)form Greenbergclassification cluster at about 7,000-10,000 B.P., threeNew World dental clusters. which when adjusted by the Joos functionbecomes 9,000- From thesefour observations and otherinformation, a den- 11,000 B.P. Our opinion is that forAmerind we are dealing tal hypothesishas been developed for the peopling of the with a time period probablygreater than 11,000 years and Americasthat envisionsthree distinct late Pleistocenemigra- beyondthe limits of glottochronology(Greenberg, Turner, and tions fromSiberia (Turner 1971, 1983a, 1983b, 1985, 1986). Zegura 1985). Startingwith the firstobservation, it is clear that all Native Americangroups had theirancestral origin in one Northeast Asian populationsystem, since all are similarand possesswhat THE DENTAL EVIDENCE Turnerhas termedthe Sinodontdental pattern rather than the SoutheastAsian Sundadont or the European dental pattern Excellenceof preservation, the existence of numerous indepen- (Turner 1983a). This interpretationis consistentwith other denttraits, genetic determination, evolutionary conservatism, biologicalevidence such as broad facial form,straight black between-groupvariation, and ease of observationare some of hair, Mongoloid sacral spot, and numerousother anatomical the qualities that make teeth a major source of direct and and biochemicalfeatures that indicate a commonAsian ances- diachronicinformation on past and presentpopulation his- tryfor all Native Americans(Hrdlicka 1913, Laughlin 1963, tories and relationships.Depending on the taxonomiclevel Stewart1973). Dental morphologyis in perfectaccord with all under consideration,differing sets of dental morphological otherbiological evidence pointing to an Asian origin.The teeth traitscan be considered.At the macroevolutionarylevel, that are even moregeographically specific, as theyreject Southeast- is, above species, size of canine, cingulumform, and other ernAsia as a possiblehomeland. Additional dental data point traitsare usefulfor assessingaffinity and descentwithin the to NorthChina as the ultimateancestral homeland for all Na- primateorder. At themicroevolutionary level, thatis, withina tive Americans(Turner 1985). narrowlydefined population such as theAleut-Eskimo system, Because New Worlddental variation is greaterin the north othertraits like lower firstmolar root numberand degreeof than in the south (for example, the Aleut-EskimoMMD of incisorshoveling are valuable for recognizingchanges over 0.040 is greaterthan the 0.011 MMD for South America- veryshort periods of time.It is at the mesoevolutionarylevel, Mesoamerica),it is evidentthat the peoplingof the Americas by whichwe mean taxonomicunits above the regionalgroup proceededfrom Alaska southward.Variation should be great- and below thespecies, that a batteryof 28 keycrown and root est wherepopulations have residedthe longest period of time. traitshas been developedto addressspecifically the question of This findingis consistentwith archaeological (Jennings 1978, Native Americanorigins. Examples of thisbattery include in- West 1981, Wormington 1983) and paleoenvironmental cisorshoveling and Carabelli'strait, as well as less well-known findings(Hopkins 1979, Hopkins et al. 1982) thatindicate hu- featuressuch as the Uto-Aztecanpremolar variant (Morris, man colonizationof the New World by way of the now- Hughes, and Dahlberg 1978). Anatomical variation within submergedBering land bridge.Only in Alaska have all three each of thesetraits is scoredwith a standardreference plaque of the earliestwell-founded New World stone tool traditions or otherstrict observation criteria. The fullbattery of traits can been discovered:(1) bifaciallychipped basally thinned or fluted be foundin Turner(1985), along withdichotomized frequen- points,absence of microblades(early component of Dry Creek cies ofeach forapproximately 9,000 variouslycomplete prehis- [Powers,Guthrie, and Hoffecker1983]); (2) bifaciallychipped toric Native Americancrania examined by Turner between but unflutedpoints, presence of microblades(later component 1975 and 1984.1 of Dry Creek, Groundhog Bay, various Denali sites, and Dental variation within and between the Americas and others [West 1981, Ackerman 1983]); (3) macro- to micro- Eurasia. Table 1 providesa rankedordering of multivariate blades, absence of bifaciallychipped tools (Anangula, Aleu- mean measures of divergence(MMDs) for the entireNew tians[Laughlin 1963]). Outside ofAlaska and westernCanada Worldand comparativeOld Worldpopulation samples. These onlythe first tradition has been found(Dumond 1980,Stanford newlycalculated MMDs differslightly from previously pub- 1982, Bryan 1983). lished values because of minorchanges in the dichotomizing The thirdobservation, namely, that New World teethare breakpointsfor some traits and increasedsample sizes for some more like those of easternAsia than like those of Europe, groups,particularly Eskimo. All but 23 of the 276 MMDs are providesa solid basis forchallenging the archaeologicalview statisticallysignificant, and 18 of the 23 nonsignificantvalues thatPaleo-Indians originated in Europe because theirmethods are caused by the Athapaskansample, whichhas the smallest of stonetool manufacturingwere like thoseused by late Pleis- mean numberof individuals per trait and the least certain toceneCro-Magnon hunters such as the Kostienkior Sunghir tribes.The fourEuropean samples of table 1 and figure1 are verysimilar among themselvesand least like the New World groups.There is no supportin thisgenetically sensible spatial 1 The dentalinformation presented here was gatheredand analyzed patternfor theorizingthat Native Americansoriginated in with help fromthe National GeographicSociety, National Science Europe or thatthey are someform of European-Asianhybridi- Foundation,IREX (InternationalResearch and Exchanges Board), zation. Examiningthe MMDs betweenSouth America and all U.S.S.R. Academyof Science,and ArizonaState University.Some of othergroups neatly shows thatdivergence is substantiallycor- the archaeologicalinformation is fromTurner's excavations in Alaska relatedwith geographicdistance. The same is generallytrue and museum studies throughoutthe Americasand in the U.S.S.R. forall otherNew World groups,even the Athapaskans. Linda Nuss assistedwith data processing.This is contributionno. 25 to his Peopling of the Pacific Basin and AdjoiningAreas Series. A This geographic-distance/dental-divergencerelationship major dental reportis in preparation,in which all findingsand ac- also indicatesthe chiefprobable cause of dental microevolu- knowledgmentswill be given. tion. Althoughregularity in the MMD distributionis sugges-

480 CURRENT ANTHROPOLOGY

This content downloaded from 128.143.023.241 on August 23, 2016 15:09:18 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). TABLE 1

RANKED ORDERING OF MEAN MEASURES OF DIVERGENCE (28 TRAITS) South America Mesoamerica SouthwestU.S. .011 Mesoamerica .011 South America .027 E U.S. & Canada .016 California .025 California .032 NW U.S. & Canada .032 E U.S. & Canada .037 E U.S. & Canada .033 ArchaicCanada .043 SW U.S. .048 SW U.S. .037 California .056 ArchaicCanada .070 ArchaicCanada .043 South America .058 NW U.S. & Canada .081 GulfAlaska .048 Mesoamerica .062 GulfAlaska .086 NW U.S. & Canada .048 Athapaskan .082 Athapaskan .115 NE Siberia .049 GulfAlaska .100 NE Siberia .117 Athapaskan .064 NE Siberia .117 Amur .142 Aleut .080 Japan .120 Japan .151 Amur .095 Aleut .122 Aleut .152 Japan .104 Eskimo .132 Eskimo .159 Eskimo .114 Hiogo Japan .152 Hiogo Japan .189 Hiogo Japan .127 Amur .164 Japan recent .196 Japan recent .128 Japan recent .177 Urga & Mongol 2 .216 Urga & Mongol 2 .132 Urga & Mongol 2 .199 An-yang .255 An-yang .136 An-yang .250 RecentThai .273 RecentThai .173 RecentThai .298 Early Thai .320 Early Thai .213 Early Thai .461 NW Europe .466 NW Europe .372 NW Europe .594 Poundbury .609 Poundbury .494 Poundbury .632 Danish Neolithic .651 Holland .533 Danish Neolithic .637 Holland .652 Danish Neolithic .555 Holland California E U.S. & Canada NW U.S. & Canada .012 E U.S. & Canada .012 California .013 GulfAlaska .016 South America .027 SW U.S. .023 ArchaicCanada .025 Mesoamerica .032 South America .024 Athapaskan .037 SW U.S. .035 ArchaicCanada .032 SW U.S. .050 ArchaicCanada .037 Mesoamerica .046 E U.S. & Canada .054 NW U.S. & Canada .046 NW U.S. & Canada .049 Eskimo .064 GulfAlaska .058 Athapaskan .054 California .094 Athapaskan .064 GulfAlaska .056 NE Siberia .101 NE Siberia .096 NE Siberia .058 Aleut .111 Japan .102 Japan .058 South America .138 Hiogo Japan .129 Aleut .081 Japan .143 Aleut .141 Urga & Mongol 2 .086 Mesoamerica .145 Eskimo .144 An-yang .112 Urga & Mongol 2 .151 Amur .146 Eskimo .113 HiogoJapan .158 Japan recent .146 Hiogo Japan .113 Amur .169 An-yang .155 Amur .118 Japan recent .178 Urga & Mongol 2 .168 Japan recent .125 An-yang .233 RecentThai .193 RecentThai .170 RecentThai .286 Early Thai .219 Early Thai .225 Early Thai .467 NW Europe .420 NW Europe .409 NW Europe .593 Poundbury .544 Poundbury .516 Poundbury .616 Danish Neolithic .551 Danish Neolithic .539 Danish Neolithic .662 Holland .612 Holland .571 Holland ArchaicCanada Athapaskan GulfAlaska .023 NW U.S. & Canada .024 NW U.S. & Canada .013 NW U.S. & Canada .033 SW U.S. .031 Aleut .024 NE Siberia .035 E U.S. & Canada .047 GulfAlaska .032 Eskimo .050 California .048 SW U.S. .037 Aleut .056 South America .058 ArchaicCanada .047 Athapaskan .057 GulfAlaska .058 E U.S. & Canada .049 SW U.S. .058 Athapaskan .070 Urga & Mongol 2 .057 ArchaicCanada .069 NE Siberia .071 Japan .062 South America .070 Mesoamerica .072 Amur .064 E U.S. & Canada .083 Eskimo .074 NE Siberia .064 California .092 Aleut .084 South America .073 Japan .096 Japan .090 An-yang .081 Mesoamerica .129 Amur .093 Eskimo .093 Amur .144 Hiogo Japan .094 California .093 Hiogo Japan .145 An-yang .100 Hiogo Japan .098 Japan recent .157 Urga & Mongol 2 .117 Mesoamerica .101 Urga & Mongol 2 .158Japan recent .120 Japan recent .145 An-yang .188 RecentThai .141 RecentThai .190 RecentThai .222 Early Thai .180 Early Thai .240 Early Thai .396 NW Europe .379 NW Europe .366 NW Europe .513 Poundbury .481 Poundbury .484 Poundbury .524 Danish Neolithic .520 Holland .506 Danish Neolithic .570 Holland .527 Danish Neolithic .531 Holland

Vol. 27 *No. 5 *December 1986 481

Aleut Eskimo NE Siberia .031 Athapaskan .031 NE Siberia .024 GulfAlaska .032 NE Siberia .032 GulfAlaska .031 Eskimo .037 GulfAlaska .040 Aleut .032 Aleut .040 Eskimo .049 NW U.S. & Canada .056 NW U.S. & Canada .058 NW U.S. & Canada .083 ArchaicCanada .064 SW U.S. .092 ArchaicCanada .090 Hiogo Japan .069 ArchaicCanada .095 SW U.S. .093 Athapaskan .072 Japan .107 Japan .101 Japan .074 Hiogo Japan .112 Hiogo Japan .102 Japan recent .074 Athapaskan .114 Amur .104 SW U.S. .096 E U.S. & Canada .115 Japan recent .106 Amur .096 Japan recent .118 Urga & Mongol 2 .132 South America .100 South America .122 South America .145 California .101 California .129 E. U.S. & Canada .146 E U.S. & Canada .106 Amur .142 Mesoamerica .159 Mesoamerica .115 Mesoamerica .143 California .165 Urga & Mongol 2 .140 Urga & Mongol 2 .161 An-yang .179 RecentThai .141 An-yang .203 RecentThai .190 An-yang .148 RecentThai .248 Early Thai .250 Early Thai .190 Early Thai .381 NW Europe .352 NW Europe .309 NW Europe .479 Poundbury .431 Poundbury .401 Poundbury .519 Holland .472 Danish Neolithic .421 Danish Neolithic .521 Danish Neolithic .500 Holland .472 Holland Amur Urga & Mongol 2 An-yang .062 Japan .023 Japan .027 Japan .072 Athapaskan .031 An-yang .031 Urga & Mongol 2 .090 Hiogo Japan .067 Hiogo Japan .066 Hiogo Japan .093 GulfAlaska .070 Athapaskan .090 Athapaskan .095 Japan recent .084 Japan recent .105 Japan recent .106 NE Siberia .101 GulfAlaska .125 NW U.S. & Canada .106 Eskimo .112 NW U.S. & Canada .131 RecentThai .113 NW U.S. & Canada .118 Aleut .136 SW U.S. .114 Aleut .126 RecentThai .138 Early Thailand .117 South America .129 Amur .141 NE Siberia .127 SW U.S. .132 SW U.S. .144 E U.S. & Canada .129 Urga & Mongol 2 .134 Early Thailand .145 ArchaicCanada .129 ArchaicCanada .140 NE Siberia .145 GulfAlaska .151 Mesoamerica .141 E U.S. & Canada .161 Aleut .151 California .157 ArchaicCanada .169 California .155 RecentThai .165 Eskimo .169 Amur .155 E U.S. & Canada .177 South America .190 Eskimo .169 An-yang .178 California .199 South America .241 Early Thailand .216 Mesoamerica .255 Mesoamerica .354 NW Europe .335 NW Europe .441 NW Europe .484 Poundbury .410 Danish Neolithic .508 Danish Neolithic .511 Danish Neolithic .424 Poundbury .519 Poundbury .514 Holland .476 Holland .614 Holland Japan Hiogo Japan Japan recent .000 Hiogo Japan .000 Japan recent .000 Hiogo Japan .018 Japan recent .000 Japan .018 Japan .023 Urga & Mongol 2 .054 RecentThai .058 RecentThai .027 An-yang .066 An-yang .084 Urga & Mongol 2 .062 Amur .067 Urga & Mongol 2 .095 Amur .071 Athapaskan .074 NE Siberia .096 NE Siberia .072 NE Siberia .090 Eskimo .098 GulfAlaska .072 RecentThai .090 Amur .102 Eskimo .073 GulfAlaska .093 GulfAlaska .105 An-yang .080 SW U.S. .100 Athapaskan .115 Aleut .081 NW U.S. & Canada .101 Early Thai .118 NW U.S. & Canada .096 ArchaicCanada .112 Aleut .120 Athapaskan .101 Eskimo .113 NW U.S. & Canada .123 Early Thailand .102 E U.S. & Canada .114 SW U.S. .128 SW U.S. .107 Aleut .138 California .158 California .111 California .144 ArchaicCanada .164 South America .112 Early Thai .146 E U.S. & Canada .168 ArchaicCanada .120 South America .152 South America .168 E U.S. & Canada .152 Mesoamerica . 189 Mesoamerica . 196 Mesoamerica .293 NW Europe .233 Europe .227 NW Europe .390 Danish Neolithic .303 Danish Neolithic .322 Danish Neolithic .407 Poundbury .304 Poundbury .329 Poundbury .473 Holland .376 Holland .380 Holland

482 C UR R ENT A N THR OPO L OG Y

Recent Thai Early Thai NW Europe .013 Early Thai .013 RecentThai .019 Danish Neolithic .054 Hiogo Japan .101 Hiogo Japan .024 Poundbury .058 Japan Recent .112 Japan .028 Holland .072 Japan .123 NW Europe .123 Early Thai .126 Urga & Mongol 2 .123 Japan recent .135 RecentThai .131 An-yang .134 Urga & Mongol 2 .227 Japan recent .135 NW Europe .135 Danish Neolithic .233 Hiogo Japan .141 Athapaskan .138 An-yang .293 Japan .148 NE Siberia .171 Poundbury .309 NE Siberia .155 Amur .180 Athapaskan .335 Urga & Mongol 2 .165 Danish Neolithic .190 NE Siberia .352 Eskimo .170 NW U.S. & Canada .213 SW U.S. .354 Amur .173 SW U.S. .219 E U.S. & Canada .366 GulfAlaska .179 Eskimo .222 ArchaicCanada .372 SW U.S. .188 ArchaicCanada .224 Holland .379 Athapaskan .190 GulfAlaska .225 NW U.S. & Canada .381 Aleut .193 E U.S. & Canada .240 GulfAlaska .396 ArchaicCanada .197 Poundbury .241 Amur .409 NW U.S. & Canada .203 Aleut .248 Aleut .420 E U.S. & Canada .233 California .250 Eskimo .441 An-yang .250 South America .286 California .461 South America .263 Holland .298 South America .466 Mesoamerica .273 Mesoamerica .320 Mesoamerica .467 California Poundbury Holland Danish Neolithic .004 Holland .004 Poundbury .009 Poundbury .009 Danish Neolithic .028 NW Europe .019 NW Europe .024 NW Europe .046 Danish Neolithic .046 Holland .171 Early Thai .224 Early Thai .135 Early Thai .197 RecentThai .263 RecentThai .165 RecentThai .304 Hiogo Japan .376 Hiogo Japan .303 Hiogo Japan .329 Japan recent .380 Japan recent .322 Japan recent .401 NE Siberia .472 NE Siberia .390 Japan .407 Japan .473 Japan .410 Urga & Mongol 2 .424 Urga & Mongol 2 .484 Urga & Mongol 2 .421 NE Siberia .431 Eskimo .500 Eskimo .472 Eskimo .479 Aleut .514 Amur .506 GulfAlaska .481 Athapaskan .519 Aleut .508 An-yang .484 GulfAlaska .520 Athapaskan .511 Amur .484 Amur .531 GulfAlaska .521 Aleut .494 SW U.S. .555 SW U.S. .524 ArchaicCanada .513 ArchaicCanada .570 ArchaicCanada .527 Athapaskan .516 NW U.S. & Canada .571 NW U.S. & Canada .533 SW U.S. .519 An-yang .612 E U.S. & Canada .539 NW U.S. & Canada .544 E U.S. & Canada .614 An-yang .551 E U.S. & Canada .593 California .637 South America .616 California .594 South America .651 Mesoamerica .632 South America .609 Mesoamerica .662 California .652 Mesoamerica

NOTE: Traitsand scoring procedures are in Turner(1985). Some of the dichotomizing breakpoints differ in thepresent paper. tiveof clinal variation, which is commonlyattributed to gradu- Northand South AmericanIndians (Macro-Indian).This can ated selection, there are no identifiablepotential dental be appreciated by inspectingthe dental dendogram(fig. 1) selectivepressures from north to southin theAmericas or east based on the 28-traitMMD matrixclustered with the un- to west in Eurasia. Moreover,no convincingselective advan- weighted-pair-group,arithmetic-averages method. Aleuts, tage has yet been identifiedfor these dental traits,certainly Eskimos, and Northeast Siberian Eskimos, Chukchi, and none that gives reproductiveadvantage proportionalto the Koryakform one division.The clusteringprocedure has linked magnitudeof the MMD range.None ofthese traits is knownto theAthapaskans with the BeringSea Mongoloidseven though be associatedwith any otherphysical or biochemicaltrait that the smallestAthapaskan MMD (0.024, table 1) is with the selectionmight have been workingon. Given the small band NorthwestCoast Indians. A second clusteris the Greater sizes thatmust have been involvedin the colonizingof Siberia NorthwestCoast groupshown as Gulf of Alaska and North- and the Americasand the factthat all musthave been rather west U.S. & Canada in figure1. Most otherNorth and South closelyrelated, population structure must be the underlying AmericanIndians make up thethird division. Archaic Canada conditionfor the distance-divergencerelationship. The small and SouthwestUnited States clusteranomalously, probably Asian-AmericanMMDs supporta rapid expanding-frontcol- because bothare pulled northwardby havingsome "Na-Dene" onizationmodel like thatproposed by Martin(1984). crania in each series.This is almostcertain for the Southwest, New World dentalvariation forms three clusters: (1) Aleut- since it includescrania fromPecos and San CristobalPueblos Eskimo, (2) GreaterNorthwest Coast or Na-Dene (in table 1 thathad Apache and Navajo contact,trade, and intermarriage this division is made up of Southwest United States and (Gunnerson1979). Archaic Canada is weightednumerically Canada, Gulf of Alaska, and Athapaskan),and (3) all other towardsa greaterinterior-western than eastern representation.

Vol. 27 * No. 5 * December1986 483

This content downloaded from 128.143.023.241 on August 23, 2016 15:09:18 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). TABLE 2 JAPAN DENTAL GROUP COMPOSITION AND SAMPLE SIZES HIOGO JAPAN JAPAN: RECENT GROUP SAMPLE SIZE MEAN S.D. AN-YANG CHINESE 1. South America: Panama 1 254-1,203 689.1 259.0 and 2, Ecuador (Ayalan, URGA+ MONGOL.2 Cotocollao, Santa Elena, THAI: RECENT Valdivia/Chanduy),Peru (Preceramic,Paloma, Peru 1 EARLY THAILAND and 2), Bolivia, Chile ALE UT (Patagonia, Herradura, Cuchipuy,Punta Teatinos), I ATHAPASKAN Brazil (Lagoa Santa 1-3, Corondo, Sambaqui north NE SIBERIA and south,Minas Gerais) ALASKA + GREENLAND ESKIMO 2. Mesoamerica:Mexico 66-367 222.1 90.0 (Coahuila, Tlatelolco, SOUTH AMERICA Tehuacan, Cuicuilco, MESOAMERICA Chichen Itza) 3. SouthwestUnited States: 172-628 429.1 104.0 EASTERN U.S +CANADA Canyon, Utah (Cottonwood CALIFORNIA Grand Gulch), Arizona (Canyonsde Chellyand del ARCHAIC CANADA Muerto,Kayenta, Chevez Pass, Point of Pines, Grass- SOUTHWEST U.S. hopper,Mogollon), New GULF OF ALASKA Mexico (San Cristobal,Pecos) 4. California:Southern, Hum- 40-292 137.5 65.5 NORTHWEST U.S+ CANADA boldt County,Sacramento AMUR County,Alameda County, Archaic(Tranquility, SJo-68) POUNDBURY ENGLAND 5. Eastern United States and 140-708 453.8 140.2 HOLLAND Canada: Iroquois (Toronto, Roebuck),Maryland (Nan- DANISH NEOLITHIC jemoy/Juhle,Ossuaries 2 and ______N W. EUROPE 4), Arkansas(Quapaw, Togo, Vernon Golightly,Wapanoca, FIG. 1. Dentaldendrogram. Paul, Nodena), Alabama 6. NorthwestUnited States and 5 7-475 250.0 127.1 Canada: NorthernMaritime 1-4 (Namu), Central Na-Dene-speakersmay have been rangingwell outsideAlaska Maritime1-3, Gulf of Geor- in Archaictimes (Dumond 1969). An importantdental quality gia and Puget Sound 1-3, of the GreaterNorthwest Coast groupis the intermediatefre- Lower Columbia River quencies of traits when compared with Aleut-Eskimoand 7. ArchaicCanada: Saskatche- 18-124 59.9 26.9 Macro-Indian.This qualityhas been discussedin detail else- wan, Quebec to be due to the ancestral 8. Athapaskan:Yukon, Inter- 0-100 41.9 27.9 where(Turner 1985). It is thought MountainFraser, Apache Siberiancondition instead of some formof New Worldclinal 9. Gulfof Alaska: Kachemak, 42-170 109.0 30.7 selection,past hybridizationbetween Paleo-Indian and ances- Kodiak (Uyak), Alaska tral Aleut-Eskimo,or samplingerror. Peninsula Dental correspondenceswith language and prehistory.From 71.8 10. Aleut: Eastern,Western 17-273 122.9 the similaramount of divergencebetween Southeast Asia and 11. Eskimo: St. Lawrence, Point 117-786 370.4 197.3 Hope, Point Barrow 1 and Northeast Asia (Recent Thai/RecentJapan, 0.06 MMD), 2, Mackenzie, Southampton, withinEurope (Holland/DanishNeolithic, 0.046 MMD), or SmithSound, East and West withinMacro-Indian (South America/ArchaicCanada, 0.056 Greenland MMD), it is clear that the peopling of the Americaswas a 12. NortheastSiberia: Ekven; 22-264 103.5 65.0 relativelyrecent event (see also Turnerand Bird 1981). More- Uelen, Chukchi; Eastern it has been that the worldwiderate of dental Siberia ("Koryak") over, proposed 13. Amur:Ulchi, Goldi, Orochi, 14-111 51.9 25.4 microevolutionis about 0.01 MMD/1,000years (Turner 1985, Negedal, Tungus, Gilyak n.d.), and inspectionof table 1 shows that when this rate is 14. Urga and Mongol 2 14-150 81.5 34.3 applied to the Macro-Indian samples, theyhave been sepa- 15. An-Yang (NorthChina) 8-224 127.6 46.2 ratedfrom Northeast Asian, on the average,for 14,000years. 16. Japan 20-138 84.7 35.9 This correspondswell withthe widelyheld view thatthe first 17. Hiogo Japan 13-96 60.4 25.6 Paleo- 18. Japan recent 43-110 82.4 16.5 Americanswere the Clovis-culturebig-game-hunting 19. Thai recent(Bangkok) 43-133 90.4 24.0 Indians who reached the southwesternUnited States 12,000 20. Early Thailand: Don Klang/ 43-237 139.9 46.1 yearsago. While thereare claimsfor pre-Clovis occupation in Ban Tong, Ban Kao, Non the New World going back to 30,000 or more years, most Nok Tha, Ban Na Di, Cen- archaeology(Owen 1984) and the dental divergencerate are tral Thailand, Ban Chiang thata groupentered the 82.2 33.1 unsupportive.It is possible pre-Clovis 21. NorthwestEurope: 19-136 recent (U.S.S.R.) Lapp, Reindeer New Worldwhich did notleave anydescendants among Island, Karilian Peninsula, Indians,but this is nota parsimoniousway to explainaway the U.S.S.R. Upper Paleolithic Clovis-dental-timerelationship. Given the provenreproduc- 22. Danish Neolithic 11-68 42.2 15.5 tive success of the Paleo-Indian colonizersand theirdogs, let 23. Holland: Dorestad de Huel, 22-96 53.8 20.7 alone thatof the manysubsequent successful animal and plant Lent would have to 24. Poundbury(England) 34-119 82.1 22.7 colonizersof the New World,what mechanism

484 CURRENT ANTHROPOLOGY

This content downloaded from 128.143.023.241 on August 23, 2016 15:09:18 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). be proposedto explainthe reproductive failure of the hypothet- Greenberg,Turner, and Zegura: SETTLEMENT OF AMERICAS ical pre-Clovispeople? The pointto be made hereis thatthere is no body of agreed-uponevidence that requires a pre-Clovis tributedto threedistinct Siberian migrations, a briefreview of migration,and the small amountof Macro-Indiandental di- NortheastAsian prehistorywill help show the robustnessof vergencesupports a relativelylate initialpeopling of the New this hypothesis.Turner (1983b and elsewhere), following World by an absolutelysmall originalfounding group from Laughlin (1963), proposed that the Aleut-Eskimomaritime which all later Macro-Indianlanguages and culturalsystems traditionbegan about 15,000 years ago in the regionof the evolved in situ. lower Amur River basin includingHokkaido because of the AncestralAleut-Eskimo entered as a distinctpopulation al- similaritiesof this region's blade tools (Derevyanko 1969; ready differentiatedin Siberia and pursuingan ecologically Vasilievsky1973; Yoshizaki, personalcommunication) to early isolatedcoastal lifeway.This propositionhas been fullydevel- Aleutian blade artifactsand some dental similaritiesto later oped by Laughlin(1975, 1980, and elsewhere)and need notbe Aleut teeth.The routefrom the lower Amur to Anangula in repeated here except to emphasize that Macro-Indian and the easternAleutians was by way of the Beringland bridge's Aleut-Eskimomust have had a commonancestor, suggested to southerncoast, as proposedby Laughlin (1963). have lived in North China about 20,000 years ago (Turner The ancestorsof Paleo-Indians probablyexited Siberia by 1985 and elsewhere).Given the hundredsof derivedMacro- way of the Lena River basin, crossingto Alaska in search of Indian languages and the relativelyfew proposedfor Aleut- the few large or small animals thatcould surviveon the cold, Eskimo, it would appear that the latter reached North dry steppelikeland bridge. No archaeologicalor dental evi- America much later than the former.However, the known dence for ancestralPaleo-Indian has yet been found in the severityof the arctic environment,the many diverse ar- Lena drainagesystem, but thisroute can be proposedon the chaeologicalcultures of Alaska, small populationsize, and the basis of the manysignificant dental trait frequency differences patternof linearcoastal settlementequally suggestthat fewer betweenAleut-Eskimo and Macro-Indian(16/2 7 [59.3%]; Tur- languagescould evolve and the language extinctionrate was ner 1985:78). For thesedifferences to evolve requireskeeping greaterin the far norththan elsewherein the New World. In these two major populationsseparate aftertheir original de- addition,Aleut-Eskimo teeth are just about as divergentfrom partureas a singlenorthward-expanding North China popula- Chineseand Japaneseas are Macro-Indians'-an unlikelycon- tion. A Siberian exit for ancestral Paleo-Indians cannot be ditionif Aleut-Eskimo reached Alaska around5,000 yearsago proposedany farther west than the Lena Basin, say, theOb or as was thoughtbefore it was appreciatedthat the Hypsither- Yenisei drainages,because it would put these Sinodontswell mal rise in Holocene sea level musthave destroyedmost evi- into European Cro-Magnon territoryas establishedby the dence of Alaskan coastal occupationbefore that time (Turner 18,000-year-oldMal'ta teeth and artifacts(Turner 1983b). 1985). European Cro-Magnon people and culture could have ex- There is a remarkablygood fitbetween Greenberg's Macro- tendedeast to about Lake Baikal on geographicgrounds alone, Indian and Aleut-Eskimolinguistic divisions and the dental since much of northwesternEurasia has relativelylow topo- clustersso far proposed.However, the fitbetween Na-Dene- graphic relief,whereas northeasternEurasia is characteris- speakersand the GreaterNorthwest Coast or Na-Dene dental ticallymountainous. During the same period at the Upper groupis not as precise. The Na-Dene-speakershave a much Cave of Zhoukoudian, Sinodont peoples were present,and morelimited geographic distribution than does thedental clus- North China tribeshad already begun to develop the North ter.There are severalpossible explanations, although it should China generalized "microlithic"tool traditionfrom which not be overlookedthat Greenbergand Turner both envision Americanstone tool typescan be derived(Turner, Gai, and the Gulf of Alaska coast as the centerof Na-Dene language Stanfordn.d.). and dental groups. First, the Na-Dene language distribution Mochanov (1978) has recognizedlate PleistoceneDiuktai may have shrunka greatdeal in the last threeor fourmillen- people betweenthe Lena and AmurBasins whose stonetools nia. The teeth,particularly those from Namu (Carlson 1983), were the same types as those used by the AmericanPaleo- are well dated as prehistoricand providea diachronicbasis for Arcticor Denali people just slightlylater in time(West 1981). proposinga moresoutherly limit for the Na-Dene speechcom- Thus, it is in the NortheastSiberian regionbounded by the munitythan existstoday on the NorthwestCoast. In this re- Lena and AmurBasins thatthe Na-Dene dentalgroup differ- gard it is noteworthythat Athapaskan-speakingIndians had entiated-itspattern of intermediacy between Aleut-Eskimo and reached northernCalifornia by at least protohistorictimes Macro-Indian.These riverine-forestDiuktai/Na-Dene people (Gould 1978). Second, the great amountof NorthwestCoast enteredAlaska just beforethe final flooding of theland bridge Indian trading,slaving, and exogamous social organization as boreal forestreplaced arcticsteppe vegetation(Colinvaux may have spread Na-Dene-sourceddental genes well beyond 1981, Hamilton 1982) and seeminglystimulated the small the languageboundary. This has been suggestedfor the more numbersof Macro-Indiansto move southwardout of eastern northerlylink-up of the SouthwestUnited States and Archaic Beringia. Canada dental groups.Whatever underlies the lack of precise The majorityof the 14,000-year-oldand youngerKamchat- fitbetween the GreaterNorthwest Coast dentaland Na-Dene kan Ushki artifactsfound and reportedby Dikov (1979), some languageunits, it hardlydetracts from the basic factthat both of whichhave been examinedby R. Ackerman,R. Carlson,P. Greenbergand Turner recognize three geographicallycor- Hobler, Turner, and others,seem to belong to the Diuktai relateddivisions of Native Americans.Supportive of thisden- assemblageand bear no convincingresemblance to Clovis ar- tal and linguisticcorrelation is the distributionof the three tifacts.This is anotherreason for proposing the Lena Basin as previouslymentioned early New World archaeologicalstone the exitroute for ancestral Paleo-Indians. Thus, thereis some tool traditions-Paleo-Indian, Denali or Paleo-Arctic,and Siberian archaeologicaland dental evidence supportingthe early Aleutian. Paleo-Indian is distributedthroughout the three-migrationhypothesis for the peopling of the New World. Macro-Indian dentitionand language area. Paleo-Arcticoc- In sum, New World dental variationmatches the North curs chieflyin Na-Dene territory.Early Aleutianis limitedto Asian Sinodontpattern, is greaterin the norththan in the thisisland chain, the Alaska Peninsula,and possiblyone site south,has a divergenceschedule corresponding to Clovis, and northof the Brooks Range. This archaeological,linguistic, and formsthree clusters which correlate highly with linguistic and dentaldistribution correspondence is simplytoo substantialto archaeological distributions.These facts, taken together, attributeto chance. stronglysuggest three late-Pleistocene migrations from Siberia Since the New World dentaland linguisticdivisions are at- to Alaska.

Vol. 27 * No. 5 * December1986 485

This content downloaded from 128.143.023.241 on August 23, 2016 15:09:18 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). THE GENETIC EVIDENCE also preserveDNA underconditions of anaerobia, neutral pH, and highion concentrations.The possibilitythat direct genetic Wiley (1981) creditsthe German entomologistWilli Hennig comparisonsof past and presenthuman populations could help withcodifying a seriesof ideas thatbegan a major revolution clarifymajor populationshifts such as the initialpeopling of in systematicbiology, the fieldof phylogeneticsystematics or the New World would be yetanother unexpected dividend of cladistics.Some of the more importantHennigian principles basic molecular biological research and recombinantDNA which underliethe reconstructionof populationhistory at all technology(Agee et al. n.d.). In a similarvein, Lowenstein evolutionarylevels are as follows(Wiley 1981:1): (1985) reviewsthe accomplishmentsof proteinanalysis using the radioimmunoassaymethod in decipheringtaxonomic 1. The relationshipsleading to thecohesion of and extinctor- rela- living tionshipsof ganismsare genealogical("blood") relationships. fossil,extinct, and living organisms.Since 1977 2. Such relationshipsexist for individuals within populations, be- the woollymammoth, Stellar's sea cow, the Tasmanian wolf, tweenpopulations, and betweenspecies. thequagga, thePiltdown hoax remains,various plant remains, 3. All othertypes of relationship(i.e., phenotypicand genetic)are bloodstains, and shrunkenhuman heads have all yielded phenomenacorrelated with genealogical descent and thusare best moleculardata usefulfor species identification. In manycases understoodwithin the contextof genealogicaldescent with new informationon dates of evolutionarydivergence also re- modification(quite literally "evolution"). sulted.One clearimplication of theseapplications is thatwhat 4. The genealogicalrelationships among populations and species may mightbe called molecular paleontology,molecular paleoan- be recovered(discovered) for by searching particularcharacters thropology,or moleculararchaeology promises to provide whichdocument these relationships. crucial testsof hypothesesbased on geneticdata gatheredfrom In general,organisms are similarbiologically because theyare living populationsor on morphologicaldata gatheredfrom related;however, organisms are not necessarilyclosely related skeletalremains and fossilizedmaterial (Lowenstein 1985). just because theyare similar.Genealogical descent or common One interpretationof the presentgenetic evidence forthe ancestrycauses biologicalsimilarity. Unfortunately, evolution- peoplingof the New World involvesa tripartitegrouping of ary biologistsand all scientistsinterested in organicevolution Native Americansconcordant with the groupings derived from are faced with an enormousproblem. One actuallyjudges linguisticand dental data: Amerind,Na-Dene, and Aleut- similarity(whether phenotypic or genetic),and fromthese Eskimo. This representsa hypothesisthat can be testedwith judgments about biological similaritiesand differencesbe- more and bettergenetic data. The total body of geneticdata tween individualsand populationsinferences must be made reviewedhere included gene and/orgenotype frequency distri- concerninggenealogical relationships. Sometimes these infer- butions, haplotype frequencies,restriction-fragment-length ences turnout to be wrongbecause of evolutionaryconver- polymorphismfrequencies (RFLPs), geneticdistance matrices, gence,unrecognized gene flow,or incorrectcharacter polarity dendrogram representations,oligovariate plots, synthetic assessment.In othercases thehistory of a particularallele in a gene-frequencymaps, phylogenetictrees, and microevolutio- populationmay confoundany predictiveassociation between naryscenarios based on theprimary data. Most of thedata are geneticsimilarity and relatedness,as Schwartzand Armitage serologicalin nature,involving blood-group antigens, serum (1983) have clearlydemonstrated in marmotcolonies. Like- proteins,erythrocyte enzymes, immunoglobulins, and leuko- wise, microevolutionarycomponents of human population cyte antigens.Additional genetic data come fromrestriction structuresuch as geneticdrift, inbreeding, and gene flowcan endonucleaseanalysis of mitochondrialDNA and fromthe interactwith natural selection to obscurethe genealogical rela- genetic epidemiologicalanalyses of disease data associated tionshipsof Native Americans.Amerinds have probablybeen withthe New World Syndrome.Data and/orinterpretive ma- in theAmericas for somewhere between 500 and 1,000genera- terialwere takenfrom the followingsources: Agee et al. n.d., tions. Tryingto decipherthese 12,000+ years of population Alekseyev(1979), Constans et al. (1985), Crawford(1984), historyis a verycomplex undertaking, especially with a data Crawfordand Enciso (1982), Crawfordet al. (1981), Dykes, base of 20th-centurygenetic data. Crawford,and Polesky(1983), Eriksson,Lehmann, and Simp- Froma biologicalviewpoint there are at presentfew alterna- son (1980), Ferrellet al. (1981), Fitch and Neel (1969), Green- tives.The adventof C14accelerator mass-spectrometry dating berg,Turner, and Zegura (1985), Harper (1980), Harper and has sounded the death knellfor all the major claims of great Laughlin(1982), Lampl and Blumberg(1979), Laughlin(1963, antiquityfor human skeletons in the New World(Taylor et al. 1966), Mourant, Kopec, and Domaniewska-Sobczak(1976), 1985, Zegura 1984). Even Bada (1985) now admits that the Neel (1976), Neel and Salzano (1966), Nei and Roychoudhury skeletonsoriginally thought to be representativeof Upper (1982), O'Rourke, Suarez, and Crouse (1985), Rychkovand Pleistocenehumans in the New World are more likely all Sheremet'eva(1980), Schell et al. (1978), Scott (1979), Scott Holocene in age. We are leftwith a soberingrealization: "it and Wright (1983), Spuhler (1979), Suarez, Crouse, and now appears thatcurrently the oldest human skeletons directly O'Rourke(1985), Suarez, O'Rourke,and Crouse(1985), Suker- dated by C-14 analysisbased on an organicfraction are from nik et al. (1981); Szathmary(1977, 1979a, 1979b, 1981, 1983, the Wilsall (Anzick)site in Montana withan age of 10,600 + 1985),Szathmary and Auger(1983), Szathmaryand Ossenberg 300 C-14 years B.P." (Taylor et al. 1985:138). An unknown (1978), Szathmary,Ferrell, and Gershowitz(1983), Williams et numberof millenniaare thus unaccountedfor in the New al. (1985), Wallace, Garrison, and Knowler (1985), Weiss World skeletal record (Irving 1985). Bonnichsenand Bolen (1985a, 1985b), Weiss, Ferrell,and Hania (1984), and Zegura (1985) reportedfinding the oldesthuman hair in theAmericas (1984). at False Cougar Cave, also in Montana. These hairspecimens It should be emphasizedthat we view the interpretationof occur stratigraphicallybelow charcoal dated at 10,530 ? 140 the geneticdata as secondarysupport for the primaryinfer- B.P. (which could make themroughly contemporaneous with ences based on linguisticand dentaldata. Otherauthors have theAnzick skeletons), although the hair may actually be as old oftendrawn very differentconclusions based on portionsof as 14,590years B.P. thesesame data (Crawfordet al. 1981; Ferrellet al. 1981; Nei Perhaps the most excitingrecent development in Amerind and Roychoudhury1982; Spuhler 1979; Szathmary1979a, biological research has been the isolation of mitochondrial 1979b). There is, in fact, littleagreement among those most DNA fromthe brain tissue of 8,000-year-oldskeletons from closelyinvolved with the actual data collectionand analysis the Windoverarchaeological site in Florida (Agee et al. n.d.). regardingthe numberor identityof the major groupingsof It appears thatDNA can be recoveredfrom nearly any tissue Native Americans based on genetic data (Crawford 1984; preservedthrough rapid drying.In addition,tissues preserved Szathmary, personal.communication; Weiss 1985b). With in saturatedenvironments like the Windover "muck" pond can these caveats in mind, we will now review what the genetic

486 CURRENT ANTHROPOLOGY

This content downloaded from 128.143.023.241 on August 23, 2016 15:09:18 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). evidence reveals about the groupingor clusteringof Native Greenberg,Turner, and Zegura: SETTLEMENT OF AMERICAS Americanpopulations. Historically,as geneticdata began to accumulate,the tradi- can Indians and Eskimosare notas greatas theprevious litera- tional positionof two major groups(American Indian versus turesuggests. Szathmary (1979b) has also proposedsuccessive Aleut-Eskimo)became the paradigmaticinterpretation of the waves of migrationas a possible explanationfor her dendro- geneticdiversity in the Americas (Laughlin 1963). Then in gramresults. The non-Na-Denesubarctic Algonkin groups in 1978 Szathmaryand Ossenbergused bothgenetic and skeletal her analyses generallyform a clusterof theirown, separate data to questionthis conventional wisdom in a paper entitled fromthe Na-Dene clusterand farremoved from the majority "Are the Biological Differencesbetween North American In- of the Eskimo populations,while the Na-Dene are theAmeri- dians and Eskimos Truly Profound?"In the meantime,other can Indian groupwith the closestrelationship to the Eskimos investigatorshad taken up the challengeof tryingto decide if based on currentgenetic similarities (Szathmary 1977, 1979a, and how American Indian geneticvariation could be par- 1979b, 1985). In addition,Szathmary has opined thatthe Na- titioned(Mourant, Kopec, and Domaniewska-Sobczak1976, Dene may have been closerto the groupsthat gave rise to the Neel 1976). Those who emphasizeda Pan-American Indian Eskimos than any otherIndian or Siberian population.As a geneticidentity saw no need for partitioningthe American cautionarynote, it should be pointedout that Ferrellet al. Indians because AmericanIndians were all membersof one (1981), using almost the same geneticdata as Szathmaryand "race." Others regarded local partitioningas essential but Ossenberg(1978), make a case fora much cleanerseparation pointed to the many examples of presumedgenetic drift in betweenAleut-Eskimos and NorthAmerican Indians. Indeed, South AmericanIndian gene pools, which,in turn,led to the it was an attemptto integrateSzathmary's (1979b) and Szath- questioningof the phylogeneticutility of such partitioning. mary and Ossenberg's (1978) resultswith those of Spuhler These doubtshave subsequentlyreceived a generaltheoretical (1979) and to reconcilethis new informationwith preceding endorsementfrom Relethford and Lees (1982:124-25), who genetic data and interpretationsof Native American pop- state: "In sum, it appears that the basic differencebetween ulation historythat led Zegura to formulatea tripartite- monogenicand polygenicsystems is theirrate of responsesto partitioninghypothesis and to explorepossible scenariosfor changesin gene frequencies,such that serologicaldata show thepeopling of theAmericas consistent with an Aleut-Eskimo/ driftand recentmigration to a greaterextent, while metrics Na-Dene/non-Na-DeneIndian trichotomy. tend to show long-termhistorical relationships." The gene- Meanwhile,Harper (1980) presenteda modelfor the origins frequencydistributions compiled by Mourantet al. (1976) pre- and divergenceof Native American populations based on senta confusinggenetic kaleidoscope. On the one hand, there erythrocyteenzyme and serum proteindata. His model in- is a markedchange in the frequenciesof the A, B, and Diego cludesancestral connections and hypotheticaldivergence dates antigensat approximatelythe latitudeof the United States- for the Athapaskans (Na-Dene) and the Aleut-Eskimosbut Mexican border, which argues against any simplisticPan- does not explicitlyinclude information on non-AthapaskanIn- Americangenetic identity. For other geneticsystems, how- dians. Accordingto his scenario,the organizationof the gene ever, CentralAmerica seems to representa transitionalzone pool responsiblefor the Native Americanpopulation system betweenNorth and SouthAmerica rather than a disjunctionor began about 19,000 years ago. The Athapaskans(Na-Dene) bottleneck(Greenberg, Turner, and Zegura 1985, Mourantet and the Aleut-Eskimosare depictedas having had a period al. 1976). of common ancestry,with the eventual divergenceof the Between the two classificatoryextremes of Pan-American Athapaskanlineage from the seminal population occurring ap- unityand groupingby extensivelocal partitioningcome those proximately15,000 yearsB.P. By about 10,000years B.P. the theorieswhich propose a major dichotomyof possible mi- non-Athapaskanbranch had evolved into an entitycalled the croevolutionaryimport within the AmericanIndian popula- Bering Sea Mongoloids. This populationquickly bifurcated tion system.For example, Nei and Roychoudhury(1982) see intothe Aleuts and Eskimos,the subsequent bifurcation of the thedichotomy as NorthAmerican Indians versusSouth Amer- Eskimo lineage into Yuit (Yupik) and Inuit (Inupiaq) taking ican Indians, Crawford(1984) sees it as Middle and South place about 5,000 years ago (Harper and Laughlin 1982). Al- AmericanIndians versusa later-arrivingNorth American Na- thoughthese dates do notfit the chronologiesdeveloped in the tivepopulation system, and we see it as theNa-Dene versusall linguisticsand dental anthropologysections, the omissionof the rest of the Amerinds(whether North, Central,or South any discussionof the non-AthapaskanIndians rendersdirect American).Nei and Roychoudhury's(1982) resultscan easily comparisonsfutile. We are encouraged,nevertheless, by the be reinterpretedto agree withour frameworkonce thelinguis- definiteconcordances between linguistic diversity and genetic tic affiliationsof theirstudy groups are made explicit(Green- diversityin modernEskimo populations.For instance,Inuit berget al. 1985). Unfortunately,the data on serumvitamin D communitiesshow a highdegree of genetic homogeneity which bindingprotein (the Gc system)presented by Constanset al. parallels their linguistichomogeneity, while the Yuit show (1985) graphicallyillustrate that this is not always the case much greatergenetic and linguisticdiversity (Crawford and (althoughthe combined Gc and PGM1 resultsin Dykes, Craw- Enciso 1982, Scottand Wright1983). These linguistic/genetic ford, and Polesky [1983:143] are in accord with our concordancescorroborate theories about therelative recency of trichotomy). the eastward Thule (Inuit) migrationand the greatertime Spuhler's(1979) studyis the mostextensive in theliterature depthof Eskimo habitationin southwesternAlaska based on restrictedto Native North American population affinities. archaeologicaldata. When allocated by gene frequenciesinto language phyla in a From a geneticperspective the hypothesisof threeseparate stepwise discriminant-functionanalysis, 82% of the Aleut- migrationsleading to a tripartitedivision of modernNative Eskimo groups(9 of 11), 67% of the Na-Dene groups(8 of 12) Americansis stillwithout strong confirmation. This is whywe and 52% of the non-Na-Dene Indian groups(15 of 29) were regardthe geneticevidence as supplementaryrather than pri- correctlyassigned to the proper language phylum. Inter- mary.The bestdirect genetic support comes from the Williams estingly,no non-Na-DeneIndian languagegroup had a higher et al. (1985) studyof immunoglobulinG (antibody)allotypes. percentageof correctallocation than the Na-Dene, underscor- Accordingto theseresearchers, when current Gm distributions ing the geneticdistinctiveness of the Na-Dene withrespect to are analyzed with respectto the three-migrationhypothesis, otherNorth American language groups. there are three distinctGm distributionsfor the postulated Szathmary(1977, 1979a, 1979b, 1981,1985) and Szathmary migrants:Gm'2' and Gm' forthe Paleo-Indians;Gm1, and Ossenberg(1978) have presenteddetailed evidence for the Gm1221, and Gm1'""3 for the Na-Dene;GmI;2' and and positionthat the biologicaldifferences between North Ameni- Gm"""l3 forthe Aleut-Eskimo. Wallace et al. (1985) have also

Vol. 27 * No. 5 * December1986 487

This content downloaded from 128.143.023.241 on August 23, 2016 15:09:18 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). provided an experimentaldesign to test the three-migration tions.It correctssuch past excessesas the routineproposal of hypothesisagainst a plausible alternativeusing mitochondrial remoteconnections as hunchesfor further investigation to re- DNA. Unfortunately,the necessary Na-Dene and Aleut- duce linguisticdiversity. Regrettably, many such connections Eskimo data have not yet been collected,but theywill be. became frozenin the subsequentliterature without investiga- Likewise, when more data are available fromNa-Dene and tion. Methods for establishingdistant linguistic genetic rela- especiallyfrom Aleut-Eskimo and Siberian groups on their tionshipsare clear (Bright1984; Campbelland Kaufman1981, geneticsusceptibility to theNew Worldsyndrome of metabolic 1983), and when theyhave been applied to formerproposals diseases (adult-onsetor non-insulin-dependentdiabetes mel- many have had to be abandoned (Campbell and Mithun litus,obesity, cholesterol gallstone formation, gall bladdercan- 1979b). Greenberg's"Amerind" will fall among these, given cer, and othermiddle-digestive-system tumors), migratory hy- theevidence and methodshe has used. It is unfortunatethat in pothesesincongruent with these data can be discarded(Weiss his "lumping"he has notheeded the demands for proof that led 1985a, Weiss et al. 1984). to the "splitting"tendency that characterizes current Amerin- AlthoughSuarez, O'Rourke, and Crouse (1985:226) warn dian linguistics.The widespreadfirst-person n and less wide- against "fanciful interpretation"of their syntheticgene- spread second-personm markers, the only evidence men- frequencymaps based on Native North Americandata, by tioned, have been recognized from the beginningwithout theirown admission the overall topographyof two of their significantimpact on classification (consider the many figurespreserves the Eskimo-Indiandichotomy as well as the "Amerind"languages lacking one or the otheror bothand the distinctrelationship between Northern and SouthernAthapas- non-Amerindlanguage containingthem, and consider the kan-speaking(Na-Dene) groups. The maps also supportthe questions of sufficientproof and of other explanationsfor notionthat the Eskimo populationsare relativelydistinct from these). Many proponentsof the "major alternative"may be the Na-Dene groups(Suarez et al. 1985:223).Their maps cer- sympatheticto the belief that (many) AmericanIndian lan- tainlydo not disconfirmour three-migrationhypothesis, but guages may have a commonorigin; however, they opine, cur- neitherdo theystrongly confirm it in theopinion of the authors rentlyaccepted methodsand evidencecannot demonstrate it. as expressedin a companionarticle (Suarez et al. 1985:238)on Given the audacityof this"Amerind" proposal, neglect of im- the relationshipbetween heterozygosity and culturalcomplex- portantrecent work relevant to the claimsis notto be excused ity. Perhaps Weiss (1985b:491) has summarizedthe present lightly(cf. Bright1984; Campbell and Kaufman 1981, 1983; state of geneticknowledge most succinctly: Campbell and Migliazza 1986; Klein and Stark 1985; Su'arez 1981, plus the extensive work within individual language Markerssuch as theGm system do tendto discriminate toa reasonable families extentamong Eskimo-Aleut, Paleo-Indian, and Athapaskanpeo- and geographicalareas). ple. . . . For less definitivemarkers, the cumulativeevidence of many The dental and geneticcorrelations are unconvincing-at loci supportssomewhat similar relationships.... Of course,Arctic worstirrelevant, at best consistentwith other interpretations. peoplesare not species, and they are not so easilytreated phylogeneti- Repetitionof the obvious seemsrequired: there is no determin- cally.However, there seems to be sometruth in thepolychotomized isticconnection between language and genepools or culture.A descriptionofArctic peoples, and this may imply aspects of the histor- singlelanguage can be spoken by a geneticallyand/or cultur- ical processeswhich generated their differences. ally diversecommunity; a culturallyand/or genetically homo- geneouspopulation can speak morethan one language. That is, language shiftand multilingualismare facts of linguistic CONCLUSIONS (and cultural)life; genes neithercause nor caterto them. Thus, the"Aleut-Eskimo" dental cluster fits linguistics, but, The threelines of evidence, linguistic, dental, and genetic,lead as the authors note, the "Greater NorthwestCoast" (called to closelysimilar divisions of the indigenous New Worldpopu- misleadingly"Na-Dene") clusterdoes not match "Na-Dene" lation into three groups; Amerind, Na-Dene, and Aleut- language geography.The NorthwestCoast has few attested Eskimo. In our opinion,the mostreasonable historical inter- Na-Dene groupsand manyothers. It is a notoriouslinguistic pretationis that these threerepresent three migrations from (and cultural)diffusion area, withmultilingualism, borrowing Asia. The orderand chronologyremain, however, less certain, of linguistictraits, slaving, and intermarriage.Here, language especiallythe relativechronological priority of the Na-Dene and genetictraits should not be expectedto match. and the Aleut-Eskimo. Similarly,the genetic data they regard as "still without strongconfirmation" and so as "supplementary."Specialists disagree,with conflictinginterpretations from the same evidence-hardly strongsupport. Conclusion:neither their linguistic classification nor its den- Comments tal/geneticcorrelation is supported,the conclusionsabout migrationsare unwarranted,and the whole speculativeventure by LYLE CAMPBELL should be abandoned. Indeed, the linguisticclassification Departmentof Anthropology, State Universityof New York, shouldbe shouteddown in ordernot to confusenonspecialists Albany,N.Y. 12222, U.S.A. 22 VI 86 or detractfrom the real contributionslinguistics can make to This articleis distressing.I object to muchbut concentrateon prehistory. (1) Greenberg'sclassification of AmericanIndian languages and (2) nonlinguisticcorrelations. Most at issue is Greenberg'shypothesized "Amerind," in- by JAMES A. Fox cludingall AmericanIndian languagessave Eskimo-Aleut(es- Departmentof Anthropology,Stanford University,Stan- tablishedlong ago) and Na-Dene (whereAthapaskan is clear, ford, Calif. 94305, U.S.A. 28 VII 86 othersless so). "Amerind"is discountedby nearlyall special- Greenberghas been workingon the generalclassification of ists. AmericanIndian languages for over 25 years, latterlywith Greenberget al. assertthat I and others,as proponentsof NSF support,but publicationof his findingshas been limited the "major alternative"classification, with ca. 200 indepen- to briefsummaries and an invitation(Greenberg 1979) to con- dent linguisticgroups, have "tacitlyclaimed" that "we will sult his microfilmednotebooks through interlibrary loan. It is neverfind that two . . . are significantlymore similar to each expectablethat this should be so; anyonewho has struggled otherin . . . geneticclassification." Not so; this "alternative" with the linguisticand bibliographiccomplexities of even a representspresent knowledge, a healthyrecognition of limita- singleAmerind stock will appreciatethe magnitudeof Green-

488 CURRENT ANTHROPOLOGY

This content downloaded from 128.143.023.241 on August 23, 2016 15:09:18 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). berg'stask. That we now have a completework in the offing Greenberg,Turner, and Zegura: SETTLEMENT OF AMERICAS (Greenbergn.d.) is an excitingdevelopment in Americanhis- toricallinguistics, as theinferences drawn by Greenberg,Tur- Mayan problems.Ironically, this "disputed" relationship is so ner,and Zegura suggest,and in Old Worldlinguistics as well, shallow thatit is not even mentionedin thesummary outlined in view of the proposedinclusion of Aleut-Eskimo in a revolu- by Greenberg. tionary"Euro-Asiatic." To anticipatethe debate beforethe Such mechanisticand unrealisticexpectations ignore the his- evidenceis in, could Greenbergbe right? toryand scientificvalue of classification.Once comparable He is, firstof all, in good company;Sapir (1968 [1916]:454- data became available to scholars,many of the most significant 55) proposed long ago that his Eskimo-Aleutand Na-Dene classifications(Finno-Ugric, Indo-European, Austronesian, stocks representseparate, and late, migrationsinto North Sino-Tibetan)were made almostimmediately, on inspectionof America (thoughhe did not group the remaininglanguages lexicallists and grammaticalparadigms. The Mayan language intoa singlestock). Indeed, as Sapir argued,if (as it seems)the family,for example, was recognizedby the brilliantSpanish timedepth of Americansettlement is not greatenough to ac- missionarylinguists of the 16thcentury long beforeeven cur- count for its linguisticdiversity, the case for a multiple- sorysystematic attention to correspondencesbegan, and subse- migrationtheory of Amerindorigins is compelling. quent scholarshiphas neitherdramatically upset the member- Greenbergis also a superbhistorical linguist. His pioneering ship nor achieved consensuson subgroupingwithin Mayan. and monumentalwork on language universalsand theirdy- Most Mayanists still accept a subgroupingbased more on namicexplanations and the successof his generalclassification quantityof lexical resemblancesthan on systematicpositive ofAfrican languages (after initially virulent criticism) warrant innovations.This is not to advocate shoddycomparison but to an expectationof sophisticatedand tellingarguments. So far, emphasize that wider classification,rather than being some he has emphasizedtwo comparativeprinciples: grammatical final product that must await perfectionwithin the known correspondencesand mass comparison.Both are criticalto the subgroup,is an integraland dynamicpart of the comparative comparativemethod, and neitherhas been widelyapplied to methoditself. It is reasonableto expectthat at leastsome wider AmericanIndian languages. relationshipsshould become apparent on inspectionin thefirst His casual referencein the presentarticle to a "widespread" significantapplication of mass comparisonto the Amerind presenceof first-personn- and second-personm- pronominal data. Therefore,while even preliminaryjudgment must await prefixesin General Amerind,along with his insightfuland the evidence and argumentation,I do not view Greenberg's more detailed treatment(Greenberg 1979) of the proposed proposalas eitherrash or improbable,and I expectAmerican GeneralAmerind third-person preconsonantal/prevocalic i-It- Indian historicallinguistics, including my own work,to profit alternationas an explanationfor problems arising from inter- fromboth the hypothesesand the example. nal reconstructionin certainSouth Americanlanguages, sug- While I pretendno special competencein the dental or ge- gestthat we may expectmuch of thisin his forthcomingbook. neticaspects of the issue, I do wish to pointout a discrepancy The lattercorrespondence, by theway, figuredprominently in in the authors'culture-historic interpretations of linguisticand Teeter's (1964:1029) acceptance of the special relationshipof dentalvariation. Greenberg (presumably) notes the enormous Wiyotto Algonquian,but I (Fox 1978:35-36) have pointedout diversityof his General Amerindas comparedwith Na-Dene thatit is so widelyfound in Latin AmericanIndian languages and therelatively greater diversity of Na-Dene comparedwith that it can only supportthe Wiyot-Algonquianrelationship, Aleut-Eskimoand concludesthat the GeneralAmerind migra- not the subgrouping.Incidentally, while the moststraightfor- tion was first,Na-Dene probablysecond, and Aleut-Eskimo ward reconstructionof the Proto-Mayanthird-person precon- last. This followslogically from principles pioneered in linguis- sonantal/prevocalicpronominal prefixes is *u-/*r-,evidence of tics by Sapir-degree of variationis generallyproportional to a previousthird-person prefix *i- is foundin the manyMayan the amountof timeavailable forthe variationto evolve. Tur- kinand body-partterms beginning with i, obviouslyby fossili- ner,however, citing the same principle,interprets precisely the zation of the pronounin cases of inalienablepossession (Fox oppositepattern of dental variation(greater in the norththan 1985:406); a pre-Proto-Mayansystem *i-l*r-is remarkably in the south) as evidence that the northhas been populated similarto the systemreconstructed by Greenberg. longerthan the south. Grantedthe threefoldclustering, why Greenberg'sinsistence on mass comparisonis also valid. theopposite patterns of variation? One suspectsgenetic mixing Since the explanation by Verner of certain exceptionsto and languageextinction in the north(Boas warnedus!), but if Grimm's Law by referenceto data fromrelated but non- thatis thecase, thedental variation is notprimarily a function Germaniclanguages, the value ofexternal comparison, though of microevolutionin place over a long periodof time. it places extraordinarydemands on scholarship,has not been Finally, the equation of Na-Dene with the GreaterNorth- disputed.The chances of survivalof a reflexof some etymon west Coast, an area of greatlinguistic, cultural, and presum- and of evidencepertaining to its reconstructionincrease with ably geneticdiffusion, is simplistic;it may be right,as thatis thenumber of surviving stocks and withthe genetic distance of the centerof Na-Dene diffusion,but the pointwill have to be such stocks fromeach other;mass comparisondramatically demonstrated. increasesthe chances of identifyingsuch a reflexand of prop- erlyassessing the relevanceof borrowing. As Greenbergpoints out, Amerind-stockspecialists' com- by W. S. LAUGHLIN mon demandsfor detailed lists of sound correspondencesand Departmentof Ecology and EvolutionaryBiology, Univer- completecomparative analyses as proofof proposedrelation- sityof Connecticut,Storrs, Conn. 06268, U.S.A. 3 VI 86 ships and classificatoryschemes are stronglycorrelated with a Three knowledgeablescholars present an informativediscus- virtualcessation of the classificatorygeneralization begun by sion of the peoplingof Americain whichthey argue forthree Sapir. Such demands are extravaganteven forone stock;in- divisionsin America and for threeseparate migrationsfrom deed, in Mayan, a venerablefocus of linguisticattention, they Siberia. This paper usefullyfocuses attention on the central have notyet been met.Amerind historical linguists tend to stay pointin contemporarystudies-whether the singlegene pool "safely"within the confinesof obvious stocks and postpone divergedbefore the migrantsleft Siberia or afterthey reached lookingafield, Verner's example notwithstanding. Attempts to America. explorethe special relationshipof Mayan to itspatently nearest The discussionof glottochronologycorrectly indicates a high relatedstock (Mixe-Zoquean), forexample, have been vigor- degreeof ambiguityand is neatlysummed up withthe obser- ouslyattacked-not withoutjustification, but to the exclusion vation that the relevanttime period is beyond the limitsof of pursuingthis potentialsource of explanationsfor infra- glottochronology.This accords well with the fact that many

Vol. 27 * No. 5 * December1986 489

This content downloaded from 128.143.023.241 on August 23, 2016 15:09:18 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). linguistshave searched for connectionsin Siberia but have and arctic biology. The authorsthink the evidence in their foundno counterparts.Compounding 11 subgroupsinto three, fieldssupports the three-migrationmodel for the settlementof witha separatemigration for each, Greenbergcogently notes the Americasand agree that the descendantsof the founding thatit is defensibleto hypothesizethat the Proto-Aleut-Eskimo groups are most precisely defined "in linguisticterms as communityarrived as an internallyundifferentiated unit be- Amerind,Na-Dene, and Aleut-Eskimo."Although their per- forethe comingof the Na-Dene. He discardsthe possibilityof spectivesare offeredin sectionssubtitled "The LinguisticEvi- one migrationwith elapsed time so great that all traces of dence,""The Dental Evidence," and "The GeneticEvidence," affinityamong any of the groupshave been effaced. verylittle evidence is presented.Rather, we are givenconclu- The dental evidenceis displayedin a dendrogramthat car- sions based on research,their own and that of others,pub- ries no hintof a tripledivision but ratheris eloquentevidence lished elsewhere.This is legitimateif one regardsGreenberg, of a single migration,with minorsubdivisions in America. Turner,and Zegura'sjoint effort as a "positionpaper" that will Clearly,dental evidence comprehends greater time depth than generate more hypothesis-testingresearch on their much- linguisticevidence. Turnerproves the Asiatic affinitiesof In- debatedsubject. I thinkthis is necessary,for the material pre- dians and also disprovesa European component.He is also sentedhere does not convinceme thatthere have indeedbeen realisticin suggestinga timedepth of some 20,000 yearsfor a threewaves of migrationinto the Americas. seminalpopulation in NorthChina, a date in good agreement While Greenberg'sanalysis of Native Americanlanguages with Harper's 19,000 years for "unmigratedAmerican In- may be correct,one has to suspend judgmentuntil his book dians." Thus, his dental, geographic,and archaeologicalevi- (Greenbergn.d.) is published and one can examine his evidence focuseson the centralproblem of whetherdivergence dence. This paper leaves me withthe impression that the proof took place in Siberia or in America. is based on grammarand "etymologies."The former,espe- The geneticdata covera broad span of kindsof geneticloci, ciallyif it consistsof unusual grammaticalforms rather than a ways ofgrouping them, and migrationhypotheses. Zegura rec- possible borrowedoverlay, may be indicativeof relationship ognizesthat the hypothesis of three separate migrations leading but does not constituteproof of relationship.The latter,"ety- to a tripartitedivision of Indians is still without strong mologies,"is moreproblematic. Greenberg needs to show that confirmationand thereforeviews the geneticevidence as sup- thereare a sufficientnumber of correspondences(i.e., system- plementaryrather than primary.Another reason for this view atic sound shifts)among the languages to permitthe recon- is foundin his treatmentof Spuhler's(1979) study,for which he structionof a largenumber of morphemes.If such morphemes cites the fact that when allocated by gene frequenciesinto formthe basis of his etymologies,then his argumentcan be languagephyla, 82% ofAleut-Eskimo groups (9 of 11), 67% of accepted. Without demonstrationthat the etymologiesare Na-Dene groups(8 of 12), and 52% ofnon-Na-Dene groups (15 based on cognatesrather than borrowings,his conclusionsare of 29) were correctlyassigned to the properlanguage phylum. unsupported. Interestingly,a chi-square test reveals no significantdifference Turner'scontribution repeats the classificationhe has pro- betweenright and wrongassignments for these three groups. posed elsewhere(Turner 1985) that equates GreaterNorthwest Alaska, not the Bering Strait,was obviouslythe gateway Coast with Na-Dene. In contrastwith his earlierwork, his intothe New World,with the Beringland bridgea fundamen- discussion here carefullydistinguishes between "Na-Dene- tal partof Alaska. The glacial maximumoccurred some 18,000 speakers"and "Na-Dene dentalgroup," for the two are notthe years ago, and the bridge was broached by risingsea level same. The GreaterNorthwest Coast group,for example, in- some 14,000years ago. The bridgeitself was one ofthe lowest, cludes Kachemak,Kodiak, and Alaska Peninsulasamples that flattest,and vegetationallymost forlornareas in the world, are likelyEskimoan in originand samplesfrom the lower Col- rangingfrom depauperate tundra in thesouth to polardesert in umbia Riverthat are non-Na-DeneIndian. Frankly,I wishwe the north,except for the southern coast, with marine resources had rulesakin to theinternational code ofzoological nomencla- and driftwood.Fish, seals, walrus,white whales, drift whales, ture that would prevent the assignmentof a name that birds,and caribouwere theprimary resources for the presum- identifiesa particulargroup to somethingthat does not belong ably small (300) migrationthat followedthe coast. Descen- to that group. Most would agree thatuse of a linguisticlabel dants of the small group that trickledup the Yukon River indicates linguisticrelationship and that for Amerindians 15,000 years ago, when its mouthlay betweenSt. Lawrence northof Mexico linguisticrelationship implies probable biolog- Island and the PribilofIslands, reached the Magellanic area ical relationship(Spuhler 1972, 1979). Turner's "Na-Dene" some 11,000 years ago. The othersmall divisionof the single in fact includes representativesof what Greenberg calls migrationcontinued on the coast, and subsequentlytheir de- "Amerind"and "Aleut-Eskimo."How does this,then, consti- scendantsoccupied the entirecoast fromAttu to Angmags- tuteevidence for a second wave of migrationby the ancestors salik. of the Na-Dene (Greenberg'sand Zegura's meaning of the The differencesbetween American populations are notlarge term)?It does not. Nevertheless,Turner would have us believe enoughto postulatemore than one migration;the taxonomic thatthere was such a migrationand thatthe teethof non-Na- categoryof American Indian easilyembraces all ofthem. From Dene-speakersof the GreaterNorthwest Coast look like those thestandpoint of evolutionary biology they appear appropriate of the Na-Dene because the latter'sgenes have spread exten- to a 15,000-yeartime depth with divergencein America,in- sively along the PacificCoast fromCalifornia to the Gulf of cludingthe now liquidatedland bridge.A singlesmall migra- Alaska. If this reallyoccurred, genetic distances should also tion some 16,000 years ago appears most parsimonious.Re- reveal the claimed extensiveand widespreadgene flow,yet I searcherswho flirtwith trinitiesshould be remindedthat foundthat the Nootka ofVancouver Island, theNa-Dene (rep- Eskimoshave walked on waterfor 10,000 years. They wait for resentedby Haida, Tlingit,and NorthernAthapaskan), and it to freeze,and when on thinice theyavoid creatingunneces- SouthAlaskan Eskimos (representedby Kodiak Islandersand sarywaves. threeKoniag isolates fromthe mainland)did not clusterto- gether(Szathmary 1979). Turner's equating the label "Na- Dene" withthe GreaterNorthwest Coast groupsuggests that by EMOKE J. E. SZATHMARY he is not preparedto question,let alone to reject,the three- Departmentof Anthropology, McMaster University,Hamil- migrationhypothesis whatever the resultsof his calculations. ton, Ont., Canada L85 4L9. 11 VII 86 Rather,he interpretshis analyticresults in thelight of a preex- What is usefulabout this articleis that it bringstogether the istinghypothesis that he simplyassumes to be true. viewpointson the peoplingof the Americasof threeindivid- Zegura shows restraintin his summationof the geneticevi- uals knownfor their work in linguistics,dental anthropology, denceconcerning the peopling of the Americas. Nevertheless, I

490 CURRENT ANTHROPOLOGY

This content downloaded from 128.143.023.241 on August 23, 2016 15:09:18 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). disagreewith his conclusionthat the Gm allotypedistributions Greenberg,Turner, and Zegura: SETTLEMENT OF AMERICAS (Williamset al. 1985) providedirect support for the tripartite- peoplingmodel. First,I disputethe notionthat Gm'2,2' does ers will not be Amerindianspecialists. Was the analysisbased not occur in Eskimos exceptas a productof admixture.The on strictcomparative methodology, or were other,less widely allotypeis presentin 11 of 15 Eskimo populationstested, in- acceptedmethods used to establishhigher relationships? If the cluding groups on St. Lawrence Island (Ferrell et al. 1981) latter,what were thesemethods? How much actual historical and in NorthAlaska (Matsumotoet al. 1982), South Alaska reconstructionwas done, and how much of the study was (M. S. Schanfield,unpublished data on Eskimosalong the Kus- based on surfacecomparisons of words fromthe various lan- kokwimRiver), Canada, and Greenland(Williams et al. 1985). guages? What sortsof statisticaltechniques were used? What Furthermore,it is presentin polymorphicfrequencies in 8 of steps were taken to screen out resemblancesresulting from the 11 groupsin whichit occursand is also foundin theChuk- language contact? Althoughspace may have been limited, chi (Sukernikand Osipova 1982). These data alone call into some of the space devoted to the historyof Amerindianlan- questionthe assumptionthat this gene has been absent from guage classificationand the historyof the developmentof the unadmixedEskimos at any timein theirhistory. Secondly, the three-grouptheory might better have been devoted to a few probabilityof geneticrelationship between any two sets of sample cases showingwhat kind of evidenceunder what kind populationsis based not on the presenceor absence of any one ofanalysis establishes a linkbetween stocks that other scholars allele in eitherbut on the overall probabilityprovided by all claim to be unrelated. A few examples of the formshy- the alleles at the given locus. It is worthnoting that cluster pothesizedto be relatedand theirreconstructed parent forms analysisof the geneticdistances derived fromthe same Gm would give the reader an idea of how controversialthe evi- data as used by Williamset al. (1985) shows thatAthapaskan- dence is. speakers(Haida and TlingitGm distributionsare unknown) The cautionurged in thisarticle with respect to theresults of are consistentlyintermixed with Eskimos and Chukchi(Szath- the glottochronologydating techniqueshould be taken seri- mary1986). OtherIndians are consistentlyseparate. This cer- ously. The method assumes a constant rate of linguistic tainlydoes not supporta tripartite-migrationmodel but lends change, but the actual rate can be speeded or slowed by lan- itselfto the notionthat Athapaskan-speakersare genetically guagecontact. We cannotassume that in thesewaves ofmigra- closerto Eskimos than are otherIndians. tion people limited themselvesto unoccupied areas. There Whetherone can make a case for more than one time of musthave been contactof some sortwhen one wave pushed entryof Asiatic hunting bands intothe Americas depends upon into areas stilloccupied by the last wave. (This problemper- the kinds of evidence one musters.I am on recordas saying tains as well to the dental and geneticdata.) that currentlyavailable geneticdata are consistentwith the Even withoutthe problemof one wave pushinginto an- view that some groupsmust have arrivedin the New World other,multilingualism is the orderof the day in most tradi- early enoughto have become separatedby the glaciersfrom tionalsocieties. If one branchof a languagegroup is in contact bands that roamed across Beringia (Szathmary1984, 1985). with an unrelatedgroup, it is likelyto change fasterthan its This certainlyimplies more than one timeof entry,but it does sisterlanguages and in a differentdirection. It may thus ap- not commitme to supportinga bipartite,tripartite, or mul- pear by glottochronologyto have divergedfrom its sisters at an tipartitemodel of occupation. In myopinion, postulation of the earlierdate thanis actuallycorrect. Renewed contact between precisenumber of "waves" is an exercisein hypothesisgenera- branchesof the same groupcan resultin borrowingsthat will tion. May therealways be creativeindividuals who propose make thebranches appear to have divergedmore recently than models,and may therealways be scientistswhose testingwill is really the case. Extremecontact situationssuch as those finallyallow us to selectthe scenariothat is mostlikely. which resultin the formationof a pidgin can throwoff the resultsof glottochronology entirely and mayeven createa false bridgebetween two unrelatedlanguage stocks. An application ofglottochronology to Englishand the English-basedpidgin of by KENNETH M. WEISS and ELLEN WOOLFORD Papua New Guinea, Tok Pisin, gives a separationdate of Departmentof Anthropology, The PennsylvaniaState Uni- somethinglike 2,000 B.P., but Tok Pisinhas notexisted a tenth versity,University Park, Pa. 16802, U.S.A. 24 VI 86 ofthat time. Moreover, the fact that Tok Pisincontains words Greenberget al. arguefrom linguistic, dental, and geneticevi- fromMalay and Tolai as well as English mightconvince a dencethat there exist three distinct groups of American natives scholar fromthe futurethat these stocks were closely con- and thatthese are the resultof threeseparate waves of immi- nected. Finally,there is one otherfactor that, if presentin a grationinto the New Worldfrom Asia. Althoughit is popular culture,can upset datingestimates. If the culturehas a taboo to decrytypological sorting of human populations, one or both on usingwords that sound like a dead relative'sname, lexical of these logicallyindependent claims may in some senses be substitutioncan occur virtuallyovernight (and substitute correct.Nevertheless, there are manypotential problems with wordsmay be borrowedfrom adjacent languages), and thereis the interpretationof such data that should be noted. Even no guarantee that words on the basic list used by glotto- nonexpertscan imaginethe difficultiesthat would be encoun- chronologywill be especiallyprotected. tered by anthropologistsin a distantfuture trying to recon- Chretien(1962) and othershave pointedout in mathematical structthe immigrationwaves that producedthe currentlin- termswhy glottochronologybecomes extremelyunreliable af- guisticand geneticdiversity in theAmericas. People ofAfrican ter only a few thousandyears, even if the rate of language ancestryspeak Englishin some areas, Spanish in others,and change is constant.Even if each of two daughterlanguages creolesin stillother areas, and manyAmerindian-speakers in loses, say, 14% of its vocabularyper 1,000 years,there is no NorthAmerica and much of Spanish-speakingLatin America knownway to determinewhether the same or differentwords are a thoroughEuropean-Amerindian and in some places also are being lost in each language. As a result,after only 1,000 Black geneticmixture. years,two daughterlanguages could have anywherefrom 72% In the linguisticssection, readers are givenvery little infor- to 86% common vocabulary. After2,000 years, the range mation beyond a bare presentationof the conclusions,and could be 48% to 74% or even greaterif some of the same words thereis no way to evaluate theseresults without going to the thatwere lost in thelast millenniumwere changed again. Even sourcescited. A discussionof the methodof linguisticanalysis whenthe most probable loss is calculated,the probableranges used, parallelingthat given in thegenetics and dentalsections, of commonvocabulary after each successivemillennium over- would have been appropriate,especially as the linguisticevi- lap more and more, and after5,000 years the overlap is so dence is allegedto be the strongestof the threeand mostread- greatthat the methodseems quite untrustworthy.Yet the lan-

Vol. 27 * No. 5 * December1986 491 This content downloaded from 128.143.023.241 on August 23, 2016 15:09:18 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). guage groups in question appear to have at least that much large-scalepopulation movements within the Americaslong timedepth. afterthe originalsettlement. The questionof the relativeorder of migrationof the three Isolation by distanceamong groupswith a long historyof groupsinto the Americasfrom Asia dependsto a greatextent habitationof a singlelocal area can producegenerally the same on the statusof Haida in the Na-Dene group.Yet at least two kind of diversityas is observed,especially if a certainamount alternativehypotheses might account for the unexpected depth of populationmovement and expansion or contractionover of the divisionbetween this language and the otherNa-Dene longtime periods occurs. Thus, even ifthere is a generalthree- groups. If this divisionoccurred in the Old World, the Na- way divisionof arctic peoples, this proves neitherthat they Dene would representmigrations of two relatedgroups, and have a three-partphylogenetic relationship nor that any such the distance between the Haida and otherNa-Dene groups relationshipas existsis due to separatewaves of immigration. could not be used to date the divergenceof this group in the Human variation,both culturaland biological,is substan- New World. This is not an impossiblescenario, since if there tial everywherein the world and was so in the "ethnographic was pressureon one groupto migrate,there could as well have present."It may be a mistaketo assume thatthere ever have been pressureon two nearbygroups to move at about thesame been monomorphicAmerindian ancestral stocks. It is difficult time. Alternatively,the Haida language mighthave encoun- to distinguishthe effectsof extrusionfrom a severeBeringian teredan extremecontact situation that caused it to changevery bottleneckand of a shiftingmosaic withinNorth America over rapidlyaway fromthe other Na-Dene groups.The factthat the 20,000 years. However, even the three-wavetheory depends dentaltraits of the Na-Dene extendfarther than the language on a shiftingmosaic-in NortheastAsia-to providethe dif- does could be accounted for if one portionof this wave of ferentwaves to send across Beringia. migrantswere absorbedby groupsthey encountered while the How can one tell,then, whether today's diversity developed rest,although not absorbed,had sufficientcontact to change here or in Asia? More data are needed, especiallyfrom the theirlanguage radically. (It may be relevantthat a pidgin, Asian side ofthe Bering Strait. Perhaps the best would be clear ChinookJargon, was spoken in the PacificNorthwest before polymorphic-DNA-haplotypedata fromboth sides of the Ber- European contact [Thomason 1983].) Under eitherof these ing Strait. Groups of closelylinked genes could providethe hypotheses,the split between Haida versus Tlingit and materialfrom which to constructtrue trees of identifyby de- Athapaskan-Eyakwould not indicate a time depth in the scentamong the identified haplotypes and to determinewhether Americasas greatas 9,000 B.P. thesecorrespond closely to phylogeniesbased on othercharac- Genes (markersor those underlyingdental traits)may not teristics,such as language or dental traits. change as rapidlyas language traits,but many of the same contactphenomena would be expectedto have a distorting,or even misleading,effect on subsequent population relationships. Tree comparisonsare difficult,and much work by Szathmaryand Spuhlerhas clearlyshown that the correspon- Reply dence of language, genetic,and morphological"trees" is any- thingbut perfect,as Greenberget al. note.The justificationfor by JOSEPH H. GREENBERG, CHRISTY G. TURNER II, concludingthat the languageevidence, probably as volatileas and STEPHEN L. ZEGURA any, is morereliable is not clear. Stanford,Calif., U.S.A. 1 viii 86 The standarderrors on geneticdistances and phylogenetic Our commentswill firstbe addressed to the specificissues splittimes are so largerelative to thedistances themselves that raisedin regardto thelinguistic, genetic, and dentalevidence. theyare notuseful in dealingwith subspecies. When there is no Afterthat the broader questions concerning the interrelation of sharp division(i.e., when thereis contactbetween the diver- theselines of investigationin theirbearing on the peoplingof sifyinggroups), "separation" times lose muchof theirmeaning the Americaswill be considered. (Weiss and Maruyama 1976), as indeed they should in that In the linguisticsection, as commentatorspoint out, thereis such separationwill have been incomplete.What are thestan- an absence of concreteevidence to supportthe conclusions dard errorson time estimatesderived fromlanguage differ- advanced. One reasonfor this is thatwe expectedGreenberg's ences? The archeologicalevidence would suggest a branch (n.d.) book to be available well beforethe appearanceof this depthof about 12,000to 15,000years for Na-Dene and Paleo- article. The book is now scheduled forJanuary 1987. Thus Indian, with Eskimo-Aleutsarriving about 5,000 or more StanfordUniversity Press was muchslower and CURRENTAN- years ago, thoughthe archeological(Dumond 1979) and cul- THROPOLOGYfar swifter than anticipated.The otherreason is tural(Townsend 1979) evidence suggests a less rigiddivisibility thatthe linguistic evidence is trulymassive and it was impossi- of these arctic peoples. With glottochronologyincapable of ble to do it justice withina relativelybrief article of broad reachingso far back in time,how can language comparisons scope. Greenberg'sforthcoming book will contain close to confidentlybe claimed to relateto such distantevents? 2,000 etymologiesand over 100 sectionsdealing with individ- It is also unclear why the authors consider the genetic ual grammaticalpoints. Some ofthese are intendedto establish markerdata to be less reliable than the dental data, when particularsubgroups of Amerind,and othersshow a distribu- presumablythe dental traitsthemselves are chosen because tionover a numberof subgroups and are evidencefor Amerind theyare geneticallydetermined. Despite what was foundby as a whole. There is also a chapteron Na-Dene. Borrowingis Relethfordand Lees (1982), quantitativetraits are not always accepted as an explanationwhere it is plausible, but it can moreinformative relative to phylogeniesthan single-locus ones neverbe an overall explanationof a mass of resemblancesin (Rogersand Harpending1983). If therehas been geneflow, as basic items,lexical and grammatical,with varied and often is reflectedin the geneticmarker pattern, why is thisnot also vast distributionsover an extendedarea. For agreementsin reflectedin the dentaltraits? One reasonmight be selection- morphologicalirregularities appearing in widelyseparated re- butif that has occurred,the rationale is lostfor most historical- gions, as illustratedby Fox's comments,borrowing is simply phylogeneticanalysis, since selectioncan obscure migration ruled out. and driftpatterns. Indeed, O'Rourke, Suarez, and Crouse First-personn and second-personm are but a small part of (1986) and Piazza, Menotti,and Cavalli-Sforza(1981) believe theevidence. They were mentionedsimply as an illustrationof that gene-frequencypatterns reflect climatic selection, result- thelogical independence of linguistic, genetic, and dentaltraits ing in latitudinalpatterns which may be highlyrelevant in the and notto prove,by themselves,the existence and limitsof the Americancase. It is also probable that these gene-frequency Amerind stock. Nevertheless,even these two items often patternsreflect patterns of admixturewith Europeans and foundtogether in the same language in numerouscases from

492 CURRENT ANTHROPOLOGY

This content downloaded from 128.143.023.241 on August 23, 2016 15:09:18 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). Chile to Canada and embracingall 11 subgroupsof Amerind Greenberg,Turner, and Zegura: SETTLEMENT OF AMERICAS are sufficientto show thatthere cannot be somethinglike 200 independentstocks in the Americas. There mightbe a few yet Hrozny, who firstconvinced the learned world of this, accidents,but borrowingcan be virtuallyruled out. There is, describes the reasons for accepting it as follows (1917:vii): to our knowledge,not a single authenticatedinstance of the "Everyonewho wishesto interpretthe Bogazkby texts ... will borrowingof a first-or second-personpronoun. Thus forbor- like the authorcome to the same conclusionon the basis of rowingto be an overall explanationan utterlyimprobable instances like the fact that wadar means 'water' . . . , that its eventwould have to have been repeatedscores and scoresof genitiveis wedenas, . . . that the Hittitepresent is inflected times.This is indeedto strainat a gnatand swallow a camel. jami, jasi, jazi, jaweni, jatteni,janzi. " Thus it is thegrammat- What the n/mexample shows is thatwe have the nucleusof ical irregularityof the nominativein r as againstthe oblique one or a verysmall numberof stocks.Further evidence serves cases of n, an irregularityalso foundin otherIndo-European to resolvethis question and may add furtherlanguages, since languages,that is cited,as well as the close resemblanceof the even first-and second-personpronouns are not eternal.Were verbinflections to thoseof Sanskrit and Latin. In therest of the one to put n/mon a map of theworld and thenmlt, one would passage he mentionspronouns and a numberof commonlex- finda tremendousclustering of the firstset in the New World ical items. and of the second in Europe and northernAsia extendingto Hrozny does not presentthe table of phoneticcorrespon- Aleut and Eskimo. Nonrandomphenomena require an expla- dences thathas become de rigeurin the pages of the Interna- nation,and the geneticis the only reasonableone in this in- tionalJournal of American Linguistics, nor has anyonesince. stance. The reason is simple;the new Hittitedata revolutionizedour It was also impossiblein an articleof this scope to discussthe ideas concerningthe sound system of Indo-European. The con- methodologyof linguistic classification.Greenberg (1957: siderable consensus which existed prior to the discoveryof chap. 5 and elsewhere)has alreadydiscussed this topic, and in Hittiteand the otherAnatolian languages was shattered,and the firstchapter of his forthcomingbook it is treatedin much at present there are very few points on which all Indo- greaterdetail. Not all of the issues can be coveredhere within Europeanistsagree. reasonablecompass. It is possibleto say, however,as is more Delbruecksaw this possibilityclearly when in anotherpas- thanhinted at in Fox's comments,that a vastlyoversimplified sage ofthe book citedearlier he said thatsound laws wereonly versionof what is called the comparativemethod has become provisionaland could be transformed(umgestaltet) by new currentin AmericanIndian studies. data. The questionis obviouslytoo complexto be treatedin The comparativemethod is associated with the school of detail here, but it is to be hoped that at least the seeds of Young Grammarianswhich flourished in the '70s and '80s of reasonabledoubt have been plantedin theminds of those who the last centuryin Germany. Indo-European has been the have accepted presentdogma uncriticallyand that theywill model,and it is nottoo muchto say thata considerablemajor- read Greenberg'sforthcoming methodological chapter with an ityof historical linguists have been Indo-Europeanistsand that open mind. the methodsof the Young Grammarianshave not been modi- Withrespect to thosecomments directed specifically toward fied in essentialssince theirtime. If one were to name the thegenetics section of our paper,we findlittle that we disagree leadingmembers of thisschool, even in a shortlist the names withor have not alreadyaddressed somewhere in thetext. We of Leskien, Brugmann,Delbrueck, and Paul would surely shareWeiss and Woolford'sconcern with the possible confound- figure.Of these only Delbrueck, the collaboratorwith Brug- ing effectsof populationstructure and selectionon historical- mann on the celebratedGrundriss, wrote a generaltext which phylogeneticanalysis. Incidentally, one reasonwe considerthe includes the question of linguisticclassification. Those who geneticdata less reliablethan the dental data is methodologi- assume thatregular sound change,phonetic correspondences, cal. All the dental data were collectedby a singleindividual and reconstructionwere consideredby thisschool, which first and thereforesubject to no interobservererror effects, whereas elaboratedthese methods in theirmodern form, as relevantto the geneticdata representan amalgamationof diversesources provingrelationships, much less as embodyinga methodof often with unknown reproducibility.We also concur with geneticclassification, may be shockedby the followingstate- Weissand Woolford'scall formore genetic data (especiallyfrom mentof Delbrueck(1904:121-22; I translatebut add the Ger- restriction-enzymestudies of DNA) collectedfrom both sides man originalsof certainkey terms): of the Bering Strait. These nuclear and mitochondrialDNA fragmentsmay provide the crucial genetictest of our three- My startingpoint is thatspecific result of comparativelinguistics migrationhypothesis, wherein substantial population diversity whichis notin doubt and cannot be indoubt. It wasproven [erwiesen] by Bopp and othersthat the so-called Indo-European languages are originatedin Asia, againstLaughlin's plausible alternative of a related.The proof[Beweis] was producedby the juxtaposition single small migration15,000-16,000 years ago with subse- [Nebeneinanderstellung]ofwords and forms of similar meaning. When quent diversificationtaking place in the Americas.It is clear oneconsiders that in theselanguages the formation of theinflectional thatSzathmary would heartilyendorse such a test. formsof the verb, nouns, and pronouns agree in essentials and likewise Szathmary,however, does bringup one substantiveissue for thatan extraordinarynumber of inflected and uninflected words agree whichwe have no immediateanswer. She questionsour use of intheir lexical parts, the assumption of chance agreement must appear the Gm allotypedata presentedin Williams et al. (1985) as absurd. directsupport for our tripartite-peoplingmodel. We have ac- Delbrueck avoids even the term for "comparison" (Ver- ceptedthose researchers' analysis as valid. Specifically,we ten- gleichung)as beingtoo elaboratefor this stage. Bopp, to whom tativelyagree with their conclusion that "when the current Gm he refers,is generallyagreed to have foundedthe comparative distributionsare analyzed with respectto the three-migration studyof Indo-European in 1816. The topicof his workwas the hypothesis,there are threedistinct Gm distributionsfor the conjugationalsystems of the verb in Sanskrit,Avestan, Latin, postulatedmigrants: Gm1;21, Gm' 2;21for the Paleo-Indians . . .; Greek, and Germanic. He did not use regularsound laws, Gml;21, Gm" 2;21,and Gm', 1,13 forthe second wave ofNa-Dene because theconcept did notyet exist, nor did he reconstruct.In hunters. . . ; and Gm1;2' and Gm';","13 forthe Eskimo-Aleut subsequentworks, he graduallyadded all the languagesthat migration9,000 years ago" (Williamset al. 1985:1). Since we were thenavailable and that are now recognizedto be Indo- have never discussed Gm" 2'2' distributionin detail, Szath- European usingthe same methods. mary'scomments about polymorphicfrequencies, admixture, The mostimportant subsequent addition was Hittite,which and the correctnessof the assumptionthat "this gene has been only became known fromthe cuneiformtablets of Bogazkoy absentfrom unadmixed Eskimos at any timein theirhistory" in Turkeyin the 20thcentury.. No one questionsits affiliation, pertainto the Williams et al. analysis. There is clearlydis-

Vol. 27 * No. 5 * December1986

This content downloaded from 128.143.023.241 on August 23, 2016 15:09:18 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). agreementbetween Williams and Szathmary.According to tionof Laughlin,the commentatorsseem willingto allow that Williamset al. (p. 17), "theseallotypic markers cannot support some of the New World dental and geneticvariation first theconclusions of Szathmaryand Ossenberg(1978) and Szath- evolved in Siberia. Study of a good East Siberian landform mary(1979, 1981) that the Na-Dene are closerto the Eskimo map willshow how implausibleit would have been fora norththan are otherNative Americans.Each categoryhas its own ward-expandinglate-Pleistocene Siberian population not to unique distribution."As Szathmaryreports here, she has since have brokeninto several relativelyisolated small population includedthe data used byWilliams et al. in a clusteranalysis of systems,each derivedfrom a limitednumber of founders.As a 16-locusgenetic-distance data base and concludedfrom its indicatedin our paper, at least two East Siberianpopulation resultsthat the Athapaskan-speakers(Na-Dene) are genet- systemscan be proposed on archaeologicalgrounds-blade- icallycloser to Eskimosthan are otherIndians. Thus, thequal- making sea-mammal-huntingand fishingfolk of the lower itative distinctionsemphasized by Williams et al. seem less Amurand Hokkaido and theterrestrial- and riverine-resource- secure when multivariateanalyses are carriedout with addi- based Diuktai people betweenthe Amur and theLena basin. It tional loci. Even if the Na-Dene are geneticallycloser to the seems well establishedthat membersof both of these groups Aleut-Eskimosthan are the restof theAmerinds, this by itself reached Alaska. The all-importantstratigraphy of the Dry does not logically falsifyour qualitative trichotomyor our Creek site near Fairbanks, Alaska, demonstratesthat both three-wavemodel (perhaps the Na-Dene and Aleut-Eskimo were almostcertainly preceded by Paleo-Indians.If all Native shared a most recentancestor in Asia beforethe bifurcation Americanvariation arose froma singlefounding population, event).Nevertheless, if the geneticconnections between these then why is the Uto-Aztecanpremolar never found in Na- populationswere to prove substantialand not due to gene Dene, NorthwestCoast, GreaterNorthwest Coast (or what- flow,Laughlin's marvelouslast two sentenceswould take on ever label one chooses to identifythe far westernprehistoric morethan an analogical referentand an Americandivergence Canadian and Alaskan people), or Aleut-Eskimo crania? would be considerablymore probable. Laughlinis correctthat the dental cladogram does notconvinc- The diversityof commentatoropinion usefullyillustrates inglyreveal three migrations. This particularanalysis does not thatthere are stillopportunities for research into Native Amer- show thethree New Worlddental clusters as well as have other ican cultural,linguistic, and biologicalmicroevolution, varia- analysesthat do not includeAsian samples(see Turner1986). tion, and origins.Meanwhile thereis a need for a working For thisreason the actual mean-measure-of-divergencevalues consensuson the basic elementsof this relativelysimple and are provided.Study of this matrix will revealthree New World shortchapter in late Quaternaryhuman population history. It dental clusters. is worthtrying to establishsome formof workingagreement, Correspondencesbetween New Worldlanguages, dentition, sincethe New Worldcase has considerablepotential as a mod- and genetics.The essenceof our paper is thatthree researchers elingtool for reconstruction of the more complex biological and workingmore or less independentlyfind three sets of the New culturalevolution in the Old World. We will use the commen- World data to have reasonable spatial correspondencesand, tators'remarks to identifywhere consensusseems near and when held up againstthe archaeologicalrecord of Siberiaand whereadditional work and thoughtare apparentlyneeded. the Americas,to suggestthree waves of late Pleistocenemi- Ultimateorigin. Laughlin agrees that the homelandof all grantsexploiting three different Beringian niches. The correla- Native Americans was in Eastern Asia, probably North tion betweenbiology and language has persistedbecause the China, and not Europe. No commentatordisagrees on this New World isolation until European contact preventedthe point. sortof massive disruption which followed contact. That varia- Timing.Again, Laughlin agrees with the suggesteddepar- tionin language,teeth, genetics, and archaeologycan be inter- turedate of about 20,000 yearsago, and no commentatordis- pretedto hypothesizethree migrations must be morethan coin- agrees. cidence. Models of one, two, or scores of migrationsare Numberof migrations. All commentatorshave strongviews incapable of accommodatingthis diversity of independentin- on thenumber of migrations, but, with the exception of Camp- formation.Turner (1983) has evaluated modelsof one to four bell, who seems to have troublerecognizing the absurdityof migrations,but only the three-wavescenario stands up. For many migrationsif New World languages are not classified thisreason we maintainthat it is thebest all-round explanatory into a few large groupings,the othersallow that onlyone to device for the New World biological and culturalvariation three migrationwaves are needed to explain New World underreview. Had that variation,particularly the biological, genetic,cultural, linguistic, and other variation. Laughlin's correlatedwith some environmentalelement, it is doubtful single-migrationproposal is an importantone because it recog- thatwe could have called on the peoplingof the Americasas nizes the potentialrichness of the late PleistoceneBering land an explanatorymechanism; instead, some selective agency bridgeriver systems, whereas nearly all otherBeringian schol- would likelybe underdiscussion. Until a differentevolution- ars have focusedon reconstructingthe terrestrialhabitat and aryscenario better explains our similarlinguistic, genetic, and game possibilitiesfor Paleo-Indian food sources. Laughlin is to dentalclassifications, a multiple-,preferably three-, migration be encouragedto startassembling evidence for this revision of hypothesismost adequately accounts for the data presentedby his (1963) classic two-wavehypothesis. the commentatorsand ourselves. Weiss and Woolfordthoughtfully explore some of the more theoreticalaspects of the three-migrationhypothesis. Given that Weiss (with Maruyama 1976) earlier reached negative conclusionsabout reconstructingPleistocene and Holocene racial historyon the basis of geneticevidence, it is understand- able that a residualnegative outlook persists. However, it is ReferencesCited not directedat the dentalevidence, because Weiss and Wool- ACKERMAN, R. E. 1983. "Observationson prehistory:Gulf of Alaska, fordseemingly understand the diachronicpower of paleontol- Bering Sea, and Asia duringthe late Pleistocene-earlyHolocene ogy. Unlike geneticmarkers in livingpopulations, which are epochs," in Late Pleistocene and early Holocene culturalconnec- limitedto synchronic"possible" evolutionaryscenarios, ar- tionsof Asia and America.Edited by R. S. Vasilievsky,pp. 49-56. chaeologicallyobtained dental remains are theactual recordof Novosibirsk:U.S. S. R. Academyof Sciences,Siberian Branch. human populationhistory and evolution.Thus, we hope for AGEE, F., W. BALLINGER, D. DICKEL, G. DORAN, W. HAUSWIRTH, and P. LAIPIS. n.d. 8,000-year-oldhuman brain tissuefrom Win- more archaeologicaland skeletaldata as well as the new ge- dover Archaeological Site: Anatomical, cellular, and molecular neticinformation they seek. analysis.In preparation. Divergencebefore or afterleaving Siberia. Withthe excep- ALEKSEYEV, V. P. 1979. The geneticstructure of Asiatic Eskimos and

494 CURRENT ANTHROPOLOGY

This content downloaded from 128.143.023.241 on August 23, 2016 15:09:18 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). coastal Chukchiscompared to thatof AmericanArctic populations. Greenberg,Turner, and Zegura: SETTLEMENT OF AMERICAS ArcticAnthropology 16:147-64. BADA, J. L. 1985. "The datingof Paleoindianskeletons from Califor- Fox, JAMES A. 1978. Proto-Mayanaccent, morpheme structure condi- nia: Implicationsfor the colonizationof the Americas,"in Woman, tions, and velar innovations.Ph.D. diss., Universityof Chicago, poet, scientist:Essays in New World anthropologyhonoring Dr. Chicago, Ill. [JAF] Emma Louise Davis. Edited by T. C. Blackburn, pp. 1-5. Los . 1985. Kinship terminologyand social process: Two Mayan Altos,Caiif.: Ballena Press. etymologies.International Journal of American Linguistics 51:405- BONNICHSEN, R., and C. W. BOLEN. 1985. "A hair, faunal, and 7. [JAF] flakedstone assemblage: A Holocene and Late Pleistocenerecord GOULD, R. A. 1978. "Tolowa," in Handbook of NorthAmerican In- fromFalse Cougar Cave, Montana," in Woman,poet, scientist: dians, vol. 8, California. Edited by R. F. Heizer, pp. 128-36. Essays in New World anthropologyhonoring Dr. Emma Louise Washington:Smithsonian Institution. Davis. Edited by T. C. Blackburn, pp. 6-15. Los Altos, Calif.: GREENBERG, JOSEPH H. 1957. Essays in linguistics.Chicago: Univer- Ballena Press. sityof Chicago Press. BRIGHT, WILLIAM. 1984. AmericanIndian linguisticsand literature. . 1960. "The generalclassification of Centraland SouthAmeri- Berlin:Mouton. [LC] can languages,"in Selected papers of the FifthInternational Con- BRYAN, A. L. 1983. "South America," in Early man in the New gress of AnthropologicalSciences, Philadelphia, Sept. 1-6, 1956. World.Edited by R. Shutler,Jr., pp. 137-46. BeverlyHills: Sage. Edited byA. F. C. Wallace, pp. 791-94. Philadelphia:University of CAMPBELL, LYLE, and TERRENCE KAUFMAN. 1981. On Mesoameri- PennsylvaniaPress. can linguistics.American Anthropologist 82:850-57. [LC] . 1979. "The classificationof AmericanIndian languages,"in . 1983. Mesoamericanhistorical linguistics and distantgenetic Papers of the 1978 Mid-AmericaLinguistics Conference at Okla- relationships:Getting it straight.American Anthropologist 85:362- homa. Edited by Ralph E. Cooley,Mervin R. Barnes, and JohnA. 72. [LC] Dunn, pp. 7-22. Norman: Universityof Oklahoma. CAMPBELL, LYLE, and ERNST MIGLIAZZA. 1986. Panoramageneral de . n.d. Languagein theAmericas. Stanford: Stanford University las lenguas indigenas de las Americas. (Historia General en Press. In press. America.) Caracas: Instituto Panamericana de Geografia e GREENBERG, J. H., C. G. TURNER II, and S. L. ZEGURA. 1985. Historia. [LC] Convergenceof evidence for the peopling of the Americas. Col- CAMPBELL, L., and M. MITHUN. Editors. 1979a. The languagesof legiumAntropologicum 9:33-42. nativeAmerica. Austin: University of Texas Press. GUNNERSON, J. H. 1979. "Southern Athapaskan archeology,"in . 1979b. "Introduction,"in The languagesof native America. Handbookof North American Indians, vol. 9, Southwest.Edited by Edited by L. Campbell and M. Mithun,pp. 3-69. Austin:Univer- A. Ortiz, pp. 162-69. Washington:Smithsonian Institution. sityof Texas Press. [LC] HAAS,M. 1958. A new linguisticrelation in NorthAmerica: Algon- CARLSON, R. L. 1983. "Early culturesof the NorthwestCoast of kian and the Gulflanguages. Southwestern Journal of Anthropology North America," in Late Pleistocene and early Holocene cultur- 14:231-64. al connectionsof Asia and America. Edited by R. S. Vasilievsky, HAMILTON, T. D. 1982. A late Pleistoceneglacial chronologyfor the pp. 61-71. Novosibirsk:U.S.S.R. Academyof Sciences, Siberian southern Brooks Range: Stratigraphic record and regional Branch. significance.Geological Society ofAmerica Bulletin 93:700-716. CHRETIEN, C. D. 1962. Mathematicalmodels of glottochronology. HARPER, A. B. 1980. Originsand divergenceof Aleuts, Eskimos, and Language 38:11-37. [KMW, EW] AmericanIndians. Annals ofHuman Biology 7:547-54. COLINVAUX,P. 1981. Historicalecology in Beringia:The southland HARPER, A. B., and W. S. LAUGHLIN. 1982. "Inquiries into the bridgecoast at St. Paul Island. QuaternaryResearch 16:18-36. peoplingof the New World: Developmentof ideas and recentad- CONSTANS, J. S. HAZOUT, R. M. GARRUTO, D. C. GAJDUSEK, and vances," in A historyof Americanphysical anthropology1930- E. K. SPEES. 1985. Populationdistribution of the humanvitamin- 1980. Edited by F. Spencer, pp. 281-304. New York: Academic D-bindingprotein: Anthropological considerations. American Jour- Press. nal ofPhysical Anthropology 68:107-22. HATTORI, S. 1973. "Japanesedialects," in Currenttrends in linguis- CRAWFORD,M. 1984. Interview.Mammoth Trumpet 1(2):1, 3. tics, vol. 11. Edited by T. A. Sebeok, pp. 368-400. The Hague: CRAWFORD, M. H., and V. B. ENCISO. 1982. "Population structure of Mouton. circumpolargroups of Siberia,Alaska, Canada, and Greenland,"in HOPKINS, D. M. 1979. "Landscape and climateof Beringiaduring late Currentdevelopments in anthropologicalgenetics, vol. 2, Ecology Pleistoceneand Holocene time," in The firstAmericans: Origins, and populationstructure. Edited by M. H. Crawfordand J. H. affinities,and adaptations. Edited by W. S. Laughlin and A. B. Mielke, pp. 51-59. New York: PlenumPress. Harper, pp. 15-41. New York: Gustav Fischer. CRAWFORD, M. H., J. H. MIELKE, E. J. DEVOR, D. D. DYKES, and HOPKINS, D. M., J.V. MATTHEWS, JR., C. E. SCHWEGER, and S. B. H. F. POLESKY. 1981. Populationstructure of Alaskan and Siberian YOUNG. Editors. 1982. Paleoecologyof Beringia. New York: Aca- indigenouscommunities. American Journal of Physical Anthropol- demic Press. ogy55:167-85. HRDLItKA, A. 1913. "The derivationand probableplace of originof DELBRUECK, ANTON. 1904. 4th edition. Einleitung in das Sprach- the North AmericanIndian." Proceedingsof XVIII International studium.Leipzig: Breitkopfand Haertel. Congressof Americanists, pp. 57-62. DEREVYANKO, A. P. 1969. The Novopetrovkablade cultureon the HROZNV, FRIEDRICH. 1917. Die Sprache der Hethiter.Leipzig: Hin- middleAmur. Arctic Anthropology 6:119-27. richs. DIKov, N. N. 1979. Ancientcultures of northeastern Asia: Asia join- HYMES, D. H. 1960. Lexicostatisticsso far. CURRENT ANTHROPOLOGY ingAmerica in ancient times.Moscow: Nauka. 1:3-44. DUMOND,D. E. 1969. Toward a prehistoryof the Na-Dene, with a IRVING, W. 1985. Context and chronologyof early man in the generalcomment on populationmovements among nomadichunt- Americas.Annual Review ofAnthropology 14:529-55. ers. AmericanAnthropologist 71:857-63. JENNINGS, J. D. 1978. "Origins," in Ancient native Americans. . 1979. Eskimo-Indianrelationships: A view fromprehistory. Edited byJ. D. Jennings,pp. 1-41. San Francisco:W. H. Freeman. ArcticAnthropology 16:2-22. [KMW, EW] Joos, M. 1964. "Glottochronologywith retention-rate inhomogeneity. " 1980. The archeologyof Alaska and the peoplingof America. Proceedingsof the Ninth International Congress of Linguists, Cam- Science 209:984-91. bridge,Mass., August27-31. Edited by H. G. Lunt, p. 237. The DYKES, D. D., M. H. CRAWFORD, and H. F. POLESKY. 1983. Popula- Hague: Mouton. tion distributionin North and Central Americaof PGM1 and Gc KLEIN, HARRIET, and LOUISE STARK. 1985. South AmericanIndian subtypesas determinedby isoelectricfocusing (IEF). American languages: Retrospectand prospect. Austin: Universityof Texas Journalof Physical Anthropology62:137-45. Press. [LC] ERIKSSON, A. W., W. LEHMANN, and N. E. SIMPSON. 1980. "Genetic KRAUSS,M. 1973. "Na-Dene," in Currenttrends in linguistics,vol. 10. studieson circumpolarpopulations," The humanbiology of circum- Edited by T. A. Sebeok, pp. 903-78. The Hague: Mouton. polarpopulations. Edited by F. A. Milan, pp. 81-168. Cambridge: LAMB, S. H. 1959. Some proposals for linguistictaxonomy. An- CambridgeUniversity Press. thropologicalLinguistics 1(2):33-49. FERRELL, R. E., R. CHAKRABORTY,H. GERSHOWITZ, W. S. LAUGH- LAMPL, M., and B. S. BLUMBERG. 1979. "Blood polymorphismsand LIN, and W. J. SCHULL. 1981. The St. Lawrence Island Eskimos: the originsof New World populations,"in The firstAmericans: Geneticvariation and geneticdistance. American Journal of Physi- Origins,affinities, and adaptations.Edited by W. S. Laughlinand cal Anthropology55:35 1-58. A. B. Harper, pp. 107-23. New York: Gustav Fischer. FITCH, W. M., and J. V. NEEL. 1969. The phylogeneticrelationship LAUGHLIN, W. S. 1963. Eskimosand Aleuts:Their origins and evolu- of some Indian tribes of Central and South America. American tion. Science 142:633-45. Journalof Human Genetics21:384-97. .1966. "Genetical and anthropological characteristics of Arctic

Vol. 27 * No. 5 * December1986

This content downloaded from 128.143.023.241 on August 23, 2016 15:09:18 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). populations,"in The biologyof human adapatability.Edited by Edited by David G. Mandelbaum. Berkeleyand Los Angeles:Uni- P. T. Baker and J. S. Weiner, pp. 469-95. Oxford:Clarendon versityof CaliforniaPress. [JAF] Press. SCHELL, L. M., S. S. AGARWAL, B. S. BLUMBERG, H. LEVY, P. H. . 1975. Aleuts:Ecosystem, Holocene history,and Siberianori- BENNETT, W. S. LAUGHLIN, and J. P. MARTIN. 1978. Distribution gin. Science 189:507-15. of albumin variants Naskapi and Mexico among Aleuts, Frobisher . 1980. Aleuts:Survivors of the Bering land bridge.New York: Bay Eskimos, and Micmac, Naskapi, Mohawk, Omaha, and Holt, Rinehartand Winston. Apache Indians. American Journal of Physical Anthropology LEVINE,R. D. 1979. Haida and Na-Dene: A new look at theevidence. 49:111-18. InternationalJournal of American Linguistics 45:157-70. SCHWARTZ, 0. A., and K. B. ARMITAGE. 1983. Problemsin theuse of LOUKOTKA,C. 1968. Classificationof South American languages. genetic similarity to show relatedness. Evolution 37:417-20. Edited by J. Wilbert.Los Angeles: Universityof CaliforniaLatin SCOTT, E. M. 1979. Geneticdiversity of AthabascanIndians. Annals AmericanCenter. ofHuman Biology 6:241-47. LOWENSTEIN,J. M. 1985. Molecular approachesto the identification SCOTT, E. M., and R. C. WRIGHT. 1983. Genetic diversity of Central of species. AmericanScientist 73:541-47. Yupik Eskimos. Human Biology55:409-15. MARTIN,P. S. 1984. "Prehistoricoverkill: The global model," in SHAFER, R. 1952. Athapaskan and Sino-Tibetan. International Jour- Quaternaryextinctions: A prehistoricrevolution. Edited by P. S. nal ofAmerican Linguistics 18:178-81. Martinand Richard G. Klein, pp. 354-403. Tucson: Universityof SHIPLEY, W. 1957. Some Yukian-Penutianlexical resemblances.In- Arizona Press. ternationalJournal of American Linguistics 23:269-74. MATSUMOTO,H., T. MIYAZAKI,N. ISHIDA,and K. KATAYAMA.1982. SPUHLER, J. S. 1972. "Genetic, linguistic, and geographical distances Mongoloidpopulations from the viewpointsof Gm patterns.Japa- in nativeNorth America," in The assessmentof population affinities nese Journalof Human Genetics27:271-82. [EJES] in man. Edited by J. S. Weinerand J. Huizinga, pp. 72-95. Lon- MILLER, W. R. 1984. The classificationof the Uto-Aztecanlanguages don: OxfordUniversity Press. [EJES] based on lexical evidence. InternationalJournal of American Lin- . 1979. "Geneticdistances, trees, and maps of NorthAmerican guistics50:1-24. Indians," in The firstAmericans: Origins, affinities, and adapta- MOCHANOV, Yu. A. 1978. "Stratigraphyand absolute chronologyof tions. Edited by W. S. Laughlin and A. B. Harper, pp. 135-83. the Paleolithicof NortheastAsia," in Early man in Americafrom a New York: Gustav Fischer. circum-Pacificperspective. Edited by A. L. Bryan,pp. 54-66. Ed- STANFORD, D. J. 1982. "A critical review of archeological evidence monton:Archaeological Researches International. relatingto the antiquityof human occupationof the New World," MORRIS,D. H., S. G. HUGHES,and A. A. DAHLBERG.1978. "Uto- in Plains Indian studies: A collectionof essays in honor of John Aztecan premolar:The anthropologyof a dentaltrait," in Develop- C. Ewers and Waldo R. Wedel. Edited by D. H. Ubelaker and ment,function, and evolutionof teeth. Edited by P. M. Butlerand H. J. Viola, pp. 202-18. SmithsonianContributions to Anthropol- K. A. Joysey,pp. 69-79. New York: Academic Press. ogy 30. MOURANT,A. E., A. C. KOPEC, and K. DOMANIEWSKA-SOBCZAK. STEWART, T. D. 1973. The people ofAmerica. New York: Scribner. 1976. The distributionof the human blood groupsand otherpoly- SUAREZ, B. K., J. D. CROUSE, and D. H. O'RoURKE. 1985. Genetic morphisms.London: OxfordUniversity Press. variation in North Amerindian populations: The geography of gene NEEL, J. V. 1976. Applicationsof multiple-variableanalysis to ques- frequencies.American Journal of Physical Anthropology 67:221-32. tions of Amerindianrelationships. Paper preparedfor the annual SUAREZ, B. K., D. H. O'RoURKE, and J. D. CROUSE. 1985. Genetic meetingof the AmericanAssociation for the Advancementof Sci- variation in North Amerindian populations: Association with ence, Boston, February18. socioculturalcomplexity. American Journal of Physical Anthropol- NEEL, J. V., and F. M. SALZANO. 1966. "A prospectusfor genetic ogy 67:233-39. studieson the AmericanIndians," in The biologyof humanadapt- SUAREZ, JORGE. 1981. "South AmericanIndian languages,"in Ency- ability.Edited by P. T. Baker and J. S. Weiner,pp. 245-74. Ox- clopaedia Britannica, 15thedition, vol. 17. [LC] ford:Clarendon Press. SUKERNIK, R. I., S. V. LEMZA, T. M. KARAPHET, and L. P. OSIPOVA. NEI, M., and A. K. ROYCHOUDHURY.1982. "Geneticrelationship and 1981. Reindeer Chukchi and Siberian Eskimos: Studies on blood evolutionof humanraces," in Evolutionarybiology, vol. 14. Edited groups,serum proteins, and red cell enzymeswith regard to genetic by M. K. Hecht, B. Wallace, and G. T. Prance, pp. 1-59. New heterogeneity.American Journal of Physical Anthropology55:12 1- York: PlenumPress. 28. NEWMAN,S. S. 1964. Comparisonof Zuni and CaliforniaPenutian. SUKERNIK, R. I., and L. P. OSIPOVA. 1982. Gm and Km immuno- InternationalJournal of American Linguistics 30:1-13. globulinallotypes in ReindeerChukchi and SiberianEskimos. Hu- O'RoURKE, D. H., B. K. SUAREZ, and J. D. CROUSE. 1985. Genetic man Genetics61:148-53. [EJES] variationin North Amerindianpopulations: Covariance with cli- SWADESH, M. 1958. "Some new glottochronological dates for Amerin- mate. AmericanJournal of Physical Anthropology67:241-50. dian linguisticgroups." Proceedings of the 32nd Congressof Ameri- OWEN, R. C. 1984. "The Americas: The case against an Ice-Age canists, Copenhagen, 1956, pp. 670-74. Copenhagen: Munksgaard. human population,"in The originsof modernhumans: A world 1959. Mapas de clasificaci6nlingiiistica de Mexico y las surveyof the fossil evidence. Edited by F. H. Smithand F. Spencer. Americas. Cuadernos del Instituto de Historia, Serie Antropol6- New York: A. R. Liss. gica 8. PIAZZA,A., P. MENOZZI,and L. L. CAVALLI-SFORZA.1981. Synthetic .1962. "Afinidades de las lenguas Amerindias." Proceedings of gene frequencymaps of man and selectiveeffects of climate.Pro- the34th InternationalCongress of Americanists, Vienna, pp. 729- ceedings of the National Academy of Sciences, U.S.A. 78:2638- 38. Horner-Wien: Berger. 42. [KMW, EW] SZATHMARY, E. J. E. 1977. "Geneticcharacteristics of Athapaskan- POWELL,J. W. 1891. Indian linguisticfamilies of Americanorth of speakers:The problemof geneticrelationships," in Prehistoryof the Mexico. Bureau of AmericanEthnology Annual Report7. NorthAmerican sub-Arctic: The Athapaskanquestion. Edited by POWERS,W. R., R. D. GUTHRIE,and J.F. HOFFECKER.1983. Dry J. W. Helmer,S. Van Dyke, and F. J. Kense, pp. 111-20. Calgary: Creek: Archaeologyand paleoecologyof a late PleistoceneAlaskan University of Calgary Press. huntingcamp. Paper preparedfor Reportto U.S. National Park . 1979a. Eskimo and Indian contact: Examination of cranio- Service. MS. metric,anthropometric, and geneticevidence. Arctic Anthropology RELETHFORD,J. H., and F. C. LEES. 1982. The use of quantitative 16:23-48. traitsin the studyof human population structure. Yearbook of Phys- . 1979b. "Blood groupsof Siberians,Eskimos, and Subarctic ical Anthropology25:113-32. and NorthwestCoast Indians: The problemof originsand genetic ROGERS, A. R., and H. C. HARPENDING. 1983. Populationstructure relationships,"in Thefirst Americans: Origins, affinities, and adap- and quantitativecharacters. Genetics 105:985-1002. [KMW, tations.Edited by W. S. Laughlinand A. B. Harper,pp. 185-209. EW] New York: Gustav Fischer. RYCHKOV, Yu. G., and V. A. SHEREMET'EVA. 1980. "The geneticsof . 1981. Genetic markers in Siberian and northernNorth Ameri- circumpolarpopulations of Eurasia relatedto theproblem of human can populations.Yearbook of Physical Anthropology24:37-73. adaptation," in The human biology of circumpolarpopulations. . 1983. Dogrib Indians of the NorthwestTerritories, Canada: Edited by F. A. Milan, pp. 37-80. Cambridge:Cambridge Univer- Genetic diversityand genetic relationshipamong sub-ArcticIn- sityPress. dians. Human Biology 10:147-62. SAPIR,E. 1929. "Centraland North Americanlanguages," in Ency- .1984. Peoplingof northern North America: Clues fromgenetic clopaedia Britannica, 14thedition. studies. Acta Anthropogenetica8:79-110. [EJES] .1968 (1916). "Time perspectivein aboriginalAmerican cul- .1985. "Peoplingof NorthAmerica: Clues fromgenetic stud- ture: A studyin method,"in Selected writingsof Edward Sapir. ies," in Out ofAsia: Peoplingthe Americasand thePaci.fic. Edited

496 CURRENT ANTHROPOLOGY

This content downloaded from 128.143.023.241 on August 23, 2016 15:09:18 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c). by R. L. Kirk and E. J. E. Szathmary,pp. 79-104. Canberra: Greenberg,Turner, and Zegura: SETTLEMENT OF AMERICAS Journalof PacificHistory. . 1986. The assessmentof populationaffinity. Paper read at the and E. Szathmary,pp. 31-78. Canberra:Journal of PacificHistory. 69th annual meetingof the AmericanAssociation of PhysicalAn- . 1986. The firstAmericans: The dental evidence. National thropologists,Albuquerque, N.M., April. [EJES] GeographicResearch 2:37-46. SZATHMARY,E. J. E., and F. AUGER. 1983. "Biologicaldistances and . n.d. Dentochronologicalseparation estimates for Pacific Rim geneticrelationships within Algonkians," in Boreal forestadapta- populations.MS. tions. Edited by A. T. Steegmann,Jr., pp. 289-315. New York: TURNER,C. G., II, and J. BIRD. 1981. Dentitionof Chilean Paleo- PlenumPress. Indians and peoplingof the Americas.Science 212:1053-55. SZATHMARY,E. J. E., R. E. FERRELL,and H. GERSHOWITZ.1983. TURNER, C. G., II, P. GAI, and D. J. STANFORD. n.d. The North Geneticdifferentiation in Dogrib Indians: Serumprotein and eryth- China originof AmericanIndians. MS. rocyteenzyme variation. American Journal of PhysicalAnthropol- VASILIEVSKY, R. S. 1973. On the problemof the originof the ancient ogy62:249-54. culturesof the sea hunterson the North Pacificcoast. Paper pre- SZATHMARY, E. J. E., and N. S. OSSENBERG. 1978. Are thebiological sented at the 9th InternationalCongress of Anthropologicaland differencesbetween North AmericanIndians and Eskimos truly EthnologicalSciences, Chicago, Ill. profound?CURRENT ANTHROPOLOGY 19:673-701. WALLACE, D. C., K. GARRISON, and W. C. KNOWLER. 1985. Dra- TAYLOR, R. E., L. A. PAYEN, C. A. PRIOR, P. J. SLOTA, JR., R. matic foundereffects in Amerindianmitochondrial DNAs. Ameri- GILLESPIE, J. A. J. GOWLETT, R. C. M. HEDGES, A. J. T. JULL, can Journalof Physical Anthropology68:149-55. T. H. ZABEL, D. J. DONAHUE, and R. BERGER. 1985. Major revi- WEISS, K. M. 1985a. Diseases of environmentaltransition in Amerin- sionsin the Pleistoceneage assignmentsfor North American human dian and relatedpeoples. CollegiumAntropologicum 9:49-61. skeletonsby C-14 acceleratormass spectrometry:None older than . 1985b.Comment on: Anthropologyin theArctic: A critiqueof 11,000 C-14 years B.P. AmericanAntiquity 50:136-40. racial typologyand normativetheory, by Debra L. Schindler.CUR- TEETER, KARL V. 1964. "Algonquianlanguages and geneticrelation- RENT ANTHROPOLOGY26:490-91. ship." Proceedingsof the Ninth International Congress of Linguists, WEISS, K. M., R. E. FERRELL, and C. L. HANIA. 1984. A New World Cambridge,Mass., August27-31. Edited by H. G. Lunt, pp. 1026- syndromeof metabolic diseases with a geneticand evolutionary 34. The Hague: Mouton. [JAF] basis. Yearbookof PhysicalAnthropology 27:153-78. THOMASON, G. 1983. ChinookJargon in areal and historicalcontext. WEISS, K. M., and T. MARUYAMA. 1976. Archaeology,population Language 59:820-68. [KMW, EW] genetics,and studiesof humanracial ancestry.American Journal of TOWNSEND, J. B. 1979. Indian or Eskimo? Interactionand identity PhysicalAnthropology 44:31-49. [KMW, EW] in southernAlaska. Arctic Anthropology16:160-82. [KMW, WEST, F. H. 1981. The archaeologyof Beringia. New York: Columbia EW] UniversityPress. TURNER, C. G., II. 1971. Three-rootedmandibular first permanent WILEY, E. 0. 1981. Phylogenetics:The theoryand practiceof phy- molars and the question of American Indian origins.American logeneticsystematics. New York: Wiley. Journalof Physical Anthropology 34:229-41. WILLIAMS, R. C., A. G. STEINBERG, H. GERSHOWITZ, P. H. BEN- . 1983a. "Sinodontyand Sundadonty:A dentalanthropological NET, W. C. KNOWLER, D. J. PETTITT, W. BUTLER, R. BAIRD, L. view of Mongoloid microevolution,origin, and dispersalinto the DOWD-REA, T. A. BURCH, H. G. MORSE, and C. G. SMITH. 1985. PacificBasin, Siberia, and the Americas,"in Late Pleistoceneand Gm allotypesin Native Americans:Evidence forthree distinct mi- earlyHolocene culturalconnections of Asia and America.Edited by grationsacross the Beringland bridge.American Journal of Physi- R. S. Vasilievsky,pp. 72-76. Novosibirsk:U.S.S.R. Academyof cal Anthropology66:1-19. Sciences,Siberian Branch. WORMINGTON, H. M. 1983. "Early man in the New World: 1970- . 1983b."Dental evidencefor the peoplingof theAmericas," in 1980," in Early man in the New World. Edited by R. Shutler,pp. Early man in theNew World.Edited by R. Shutler,Jr., pp. 147-57. 191-95. BeverlyHills: Sage. BeverlyHills: Sage. ZEGURA,S. L. 1984. The initialpeopling of the Americas:An over- . 1985. "The dentalsearch for Native Americanorigins," in Out view from the perspectiveof physical anthropology.Acta An- ofAsia: Peoplingthe Americas and thePacific. Editedby R. L. Kirk thropogenetica8:1-21.

ResearchGrants complete.They providea monthlystipend, housing, a private study,and the time,space, and quiet needed forcreative re- * At leastfour resident fellowships for the 1987-88 yearwill be search. Residencyis usuallyfor 11 monthsand beginsin Sep- awarded to scholarsin anthropologyand relateddisciplines by tember1987. The applicationdeadline is December 1, 1986, the School of AmericanResearch in Santa Fe, New Mexico. and recipientswill be chosen by the middleof February.For The fellowshipsare supportedby the WeatherheadFounda- furtherinformation write: School ofAmerican Research, Resi- tionand the National Endowmentfor the Humanitiesand are dentScholar Program, P.O. Box 2188, Santa Fe, N.M. 87504, open to Ph.D.'s and doctoralcandidates whose fieldworkis U.S.A.

Vol. 27 No. 5 December1986 497

This content downloaded from 128.143.023.241 on August 23, 2016 15:09:18 PM All use subject to University of Chicago Press Terms and Conditions (http://www.journals.uchicago.edu/t-and-c).