This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
A NEW BAYESIAN CHRONOLOGY FOR POSTCLASSIC AND COLONIAL OCCUPA-TION AT XALTOCAN, MEXICO Lisa Overholtzer
Department of Anthropology, Wichita State University, 1845 Fairmount St., Wichita, KS 67260, USA.
ABSTRACT. This article proposes a new four-phase chronology for Postclassic and colonial occupation at Xaltocan, Mex-
ico, using Bayesian statistical modeling of a suite of 54 radiometric dates. Of these, 46 samples come from recent extensive
H[FDYDWLRQV�RI�VHDOHG��VWUDWL¿HG�KRXVHKROG�GHSRVLWV��IDFLOLWDWLQJ�LPSURYHG�XQGHUVWDQGLQJ�RI�VDPSOH�FRQWH[W�DQG�UHVXOWLQJ�LQ�a more accurate chronology. The timing of the adoption of major ceramic wares at the site and intrasite level is outlined and
contextualized within broad, regional phases. These results are particularly valuable given recent research that repudiates
D�XQLIRUP�FKURQRORJLFDO�IUDPHZRUN�IRU� WKH�%DVLQ�RI�0H[LFR�LQ�WKH�3RVWFODVVLF�SHULRG��DQG�LQGLFDWHV�LQVWHDG�D�VLJQL¿FDQW�amount of chronological overlap and regional variation in the use of distinct ceramic types.
INTRODUCTION
Despite the vast and detailed knowledge available for the Postclassic period in the Basin of Mexico,
otlatelco and Aztec I ceramic wares, “We are still a long way from understanding the spatial-tempo-
ral distribution of [these types] in central Mexico, much less the sociopolitical and socioeconomic
implications of this distribution.” This meager grasp is partially due to the apparent complexity of
these distributions, as well as the small number of existing radiocarbon determinations and a lack
RI�GHWDLOHG�FRQWH[WXDO�LQIRUPDWLRQ�IRU�WKRVH�GDWHV��7KLV�VLWXDWLRQ�LV�H[HPSOL¿HG�DW�WKH�VLWH�RI�;DO-tocan, capital of the pre-Aztec Otomí city-state, and subsequently subject to the successive Aztec
and Spanish colonial empires. With over 20 years of archaeological research at Xaltocan, the site
chronology is better understood than most, and yet the existing chronological framework relies on
lutionary trajectory between types with the exception of some colonial Aztec IV motifs that showed
REYLRXV�(XURSHDQ�LQÀXHQFH��
7KLV�UHODWLYH�FKURQRORJ\�ZDV�FRUUHODWHG�ZLWK�DEVROXWH�GDWHV��UH¿QHG��DQG�H[SDQGHG�WR�LQFOXGH�RWKHU�central Mexican regions associated with the Aztec Empire as a result of major regional survey and
H[FDYDWLRQ�SURMHFWV� EHJLQQLQJ� LQ� WKH�����V� �&KDUOWRQ�������3DUVRQV�������%ODQWRQ� DQG�3DUVRQV�������:KDOHQ�DQG�3DUVRQV�������6PLWK�DQG�'RHUVKXN�������(YDQV�DQG�)UHWHU�������+DUH�DQG�6PLWK�������1LFKROV�DQG�&KDUOWRQ�������3DUVRQV�HW�DO��������+RGJH�������%UXP¿HO�����E���6DQGHUV�HW�DO���������DUJXHG�WKDW�$]WHF�,�DQG�,,�FHUDPLFV�ZHUH�URXJKO\�FRQWHPSRUDQHRXV��UHJLRQDO�YDULDQWV�XVHG�LQ�WKH�%DVLQ�RI�0H[LFR�EHWZHHQ�$'������DQG�������$]WHF�,�FHUDPLFV�ZHUH�XVHG�SUHGRPLQDQWO\�LQ�the southern Basin, with the northern Basin island of Xaltocan standing out as a marked exception,
while Aztec II ceramics were popular in the northern Basin. In stratigraphic excavations at sites
ZKHUH�ERWK�$]WHF�,�DQG�,,�ZHUH�XVHG��VXFK�DV�&KDOFR�DQG�&XOKXDFDQ��2¶1HLOO�������6HMRXUQH��������:KDOHQ�DQG�3DUVRQV��������IRXQG�WKDW�$]WHF�,�SRWWHU\�ZDV�XVHG�H[FOXVLYHO\�IRU�VRPH�WLPH�EHIRUH�Aztec II was adopted, suggesting some degree of temporal variation. Aztec III ceramics were con-
VLGHUHG�WR�EH�WKH�KDOOPDUN�RI�WKH�SHULRG�$'�����±������6DQGHUV�HW�DO���������+RZHYHU��&KDUOWRQ��������������GHPRQVWUDWHG�WKDW�LQ�PDQ\�UXUDO�DUHDV��(XURSHDQ�LQÀXHQFHG�FHUDPLFV�GLG�QRW�DSSHDU�until approximately AD 1650, and Aztec III pottery continued to be produced in the colonial period
alongside Aztec IV pottery.
More recent Basin of Mexico chronologies have replaced a model of unilinear change with a more
logical overlap exists for several types, and chronological overlap varies geographically. Investi-
JDWLRQV�E\�%UXP¿HO�DQG�FROOHDJXHV��3DUVRQV�HW�DO��������%UXP¿HO�����D��SODFHG�$]WHF�,�SRWWHU\�earlier, beginning in the 10th century, and thus partially contemporaneous with the other regional
FHUDPLF� FRPSOH[HV��&R\RWODWHOFR� DQG�0D]DSDQ��7KHLU� UHVHDUFK� LQGLFDWHG� D� SHULRG� GXULQJ�ZKLFK�$]WHF�,�SRWWHU\�ZDV�XVHG�H[FOXVLYHO\�IROORZHG�E\�D�VLJQL¿FDQW�SHULRG�GXULQJ�ZKLFK�ERWK�$]WHF�,�DQG�,,�FHUDPLF�W\SHV�ZHUH�XVHG�DW�VRPH�VLWHV��OLNH�;DOWRFDQ���ZKLOH�RWKHUV��OLNH�WKRVH�LQ�WKH�VRXWKHDVWHUQ�%DVLQ��ZHUH�FKDUDFWHUL]HG�E\�FRQWLQXHG�XVH�RI�$]WHF�,�SRWWHU\�DQG�D�GHOD\�LQ�WKH�XVH�RI�$]WHF�,,��7KXV��DV�1LFKROV�DQG�&KDUOWRQ������������VXJJHVW��ZH�PXVW�PRYH�EH\RQG�SUREOHPDWLF�DVVXPSWLRQV�“of a uniform chronological framework for the entire basin during the Postclassic period.” This
conclusion presented challenges and exciting potential for research exploring spatial and temporal
variation in ceramic traditions related to Postclassic ethnicities and city-state confederations. It also
demonstrated the need for the development of detailed local chronologies within broad regional
Local chronologies for the site of Xaltocan, located on a human-made island in the northern Basin
RI�0H[LFR� �)LJXUH� ����ZHUH� ¿UVW� HVWDEOLVKHG� E\�%UXP¿HO� �����D���8VLQJ� VWUDWLJUDSKLF� HYLGHQFH��multidimensional scaling of ceramic variants, and 10 14&�GDWHV�IURP�WZHQW\�IRXU���î���P�WHVW�SLWV��2SHUDWLRQV��H[FDYDWHG�LQ�DUELWUDU\����FP�OHYHOV�LQ������DQG�������%UXP¿HO�SURSRVHG�IRXU�SKDVHV�of occupation at the site. These four phases were based on the following stratigraphic sequence of
diagnostic Aztec Black-on-Orange ceramics: pure Aztec I contexts lie underneath mixed Aztec I
and II deposits, which lie underneath Aztec II ceramics, which in turn lie beneath Aztec III and IV
FHUDPLFV��%UXP¿HO�����D����±����
A 14&�VDPSOH�IURP�WKH�YHJHWDWLRQ�XQGHUO\LQJ�WKH�KXPDQ�PDGH�LVODQG�SURYLGHG�DQ�HDUO\�ERXQGDU\�HVWLPDWH�IRU�VLWH�FRQVWUXFWLRQ��WKLV�GDWH�KDG�D�PHGLDQ�RI�FDO�$'������7KH�¿UVW�RFFXSDWLRQ�SKDVH�ZDV�DVVRFLDWHG�ZLWK�SXUH�GHSRVLWV�RI�$]WHF�,�%ODFN�RQ�2UDQJH�SRWWHU\�DQG�&KDOFR�3RO\FKURPH��3KDVH���had four 14&�GDWHV�ZLWK�PHGLDQV�RI�FDO�$'����������������DQG�����DQG�UHSUHVHQWV�RFFXSDWLRQ�GXULQJ�the Early Postclassic period.
Figure 1 Aztec I–IV Black-on-Orange diagnostic ceramic types: a. Aztec I bowl; b. Aztec II plate; c. Aztec III molcajete;
and d. Aztec IV molcajete.
����Bayesian Chronology for Occupation at Xaltocan, Mexico
%UXP¿HO¶V� UHVHDUFK� �����D�� LQGLFDWHG�D�SHULRG�RI�RYHUODS�EHWZHHQ�$]WHF� ,� DQG� ,,�SRWWHU\� W\SHV��FDOOHG�3KDVH����WKDW�FRQ¿UPHG�WKH�JUDGXDO�VKLIW�¿UVW�VXJJHVWHG�E\�:KDOHQ�DQG�3DUVRQV���������,Q�DGGLWLRQ� WR�PL[HG�$]WHF� ,� DQG� ,,�%ODFN�RQ�2UDQJH�SRWWHU\��3KDVH���ZDV�DVVRFLDWHG�ZLWK�&KDOFR�Polychrome and small amounts of redwares, particularly Black on Red, and with Brown Incised
ware. Two 14&�GDWHV�ZLWK�PHGLDQV�RI�FDO�$'������DQG������LQGLFDWHG�WKDW�3KDVH���UHSUHVHQWV�WKH�KHLJKW�RI�;DOWRFDQ¶V�SRZHU�GXULQJ�WKH�(DUO\±0LGGOH�3RVWFODVVLF��
Phase 4 was associated with Aztec III and IV Black-on-Orange pottery and redwares, especially
Black-on-Red. It had one 14&�GDWH�ZLWK�D�PHGLDQ�RI�FDO�$'�������EXW�WKH�PLGGHQ�WKDW�SURGXFHG�WKDW�VDPSOH�DOVR�\LHOGHG�D�FRORQLDO�VW\OH�¿JXULQH��VXJJHVWLQJ�WKDW�WKLV�SKDVH�UHSUHVHQWV�RFFXSDWLRQ�GXULQJ�WKH�/DWH�3RVWFODVVLF�DQG�FRORQLDO�SHULRG��3KDVHV���DQG���UHPDLQHG�SRRUO\�XQGHUVWRRG��VLQFH�3KDVH���was represented by only two dates and Phase 4 by a single date, and there was no chronological
separation between the 14&�GDWHV�IRU�3KDVHV���DQG����
METHODSSamples and Sample Context5HFHQW�H[WHQVLYH�H[FDYDWLRQV�RI�GRPHVWLF�VWUXFWXUHV�FODULI\�WKH�VLWH¶V�RFFXSDWLRQDO�KLVWRU\�DQG�SUR-
FHQWHUHG�RQ�3KDVHV���DQG����7KHVH�VWUDWL¿HG�GRPHVWLF�FRQWH[WV�SURGXFHG�QHZ�VWUDWLJUDSKLF�LQIRU-PDWLRQ�RQ�WKH�RUGHULQJ�RI�GHSRVLWV�FRQWDLQLQJ�WKH�GLDJQRVWLF�$]WHF�,±,9�FHUDPLFV�XVHG�WR�FRQVWUXFW�;DOWRFDQ¶V�FKURQRORJ\��7KH�H[FDYDWLRQV�DOVR�SURGXFHG����VDPSOHV�RI�FKDUFRDO�DQG�ERQH�WKDW�ZHUH�sent for radiometric analysis and which were available for Bayesian modeling. For this article, 14&�determinations were available from nearly all the excavations carried out at the site to date, and
%UXP¿HO�VXEPLWWHG����FDUERQ�VDPSOHV�IRU�14&�GDWLQJ�����$06�DQG���VWDQGDUG�UDGLRPHWULF�GDWHV��DW�%HWD�$QDO\WLF��,QF��)URP�2YHUKROW]HU¶V�H[FDYDWLRQV�����ERQH�DQG�FKDUFRDO�VDPSOHV�ZHUH�VXEPLWWHG�IRU�GDWLQJ�E\�DFFHOHUDWRU�PDVV� VSHFWURPHWU\� �$06��DW� WKH�8QLYHUVLW\�RI�$UL]RQD��%RQH�VDPSOHV�ZHUH�SURFHVVHG�XWLOL]LQJ�WKH�PRGL¿HG�/RQJLQ�PHWKRG��%URZQ�HW�DO���������$OO�ERQH�VDPSOHV�\LHOGHG�VXI¿FLHQW�FROODJHQ�DQG�DOO�FKDUFRDO�VDPSOHV�VXI¿FLHQW�FDUERQ�IRU�GDWLQJ��7ZR�GDWHV�IURP�WKH�PL[HG�$]WHF�,,±,,,�FRQVWUXFWLRQ�¿OO�ZHUH�H[FOXGHG�IURP�%D\HVLDQ�PRGHOLQJ��EHFDXVH�ZKLOH�WKH\�FRXOG�RI-fer insight into the timing of that particular construction, they could not be associated with a single
phase. The remaining dates were combined with the 10 14&�GDWHV�IURP�%UXP¿HO¶V�RULJLQDO�WHVW�SLW�H[FDYDWLRQV�WR�IRUP�WKH�EDVLV�IRU�WKH�FKURQRORJ\�EXLOW�XVLQJ�%D\HVLDQ�VWDWLVWLFDO�PRGHOLQJ��$SSHQ-
$OO�GDWHV�ZHUH�FDOLEUDWHG�XVLQJ�2[&DO��Y������%URQN�5DPVH\�����D�E��DQG�WKH�,QW&DO���FDOLEUDWLRQ�FXUYH��5HLPHU�HW�DO���������%HFDXVH�RI�D�VWURQJ�³ZLJJOH´�LQ�WKH�FDOLEUDWLRQ�FXUYH��WKH�FRORQLDO�SHULRG�dates demonstrate bimodal distributions and have two likely date ranges—one in the 16th and one
in the 17th centuries. Unfortunately, the material culture present in the middens does not allow us
WKRU¶V�GLVVHUWDWLRQ��2YHUKROW]HU��������WHVWHG�%UXP¿HO¶V�IRXU�SKDVH�FKURQRORJLFDO�IUDPHZRUN�XVLQJ�the newly available 14&�GHWHUPLQDWLRQV��7KLV�LQLWLDO�DWWHPSW�LQGLFDWHG�WKDW�D�VXEVWDQWLDO�UHYLVLRQ�RI�WKH�VLWH�FKURQRORJ\�ZDV�QHFHVVDU\��)LUVW��3KDVH���FRXOG�DQG�VKRXOG�EH�GLYLGHG�LQWR�SUH�+LVSDQLF�
Figure 3 Reconstructed island of Xaltocan, showing excavation units as black squares and radiomet-
ric sampling as stars within the squares. Radiometric samples from recent extensive excavations of
domestic contexts are denoted by white stars, while those from test pit excavations are shown in gray.
1081Bayesian Chronology for Occupation at Xaltocan, Mexico
and colonial contexts. Unfortunately, excavations have uncovered only two colonial-period houses
that directly overlie Aztec III occupation, and in those instances, the ground surface was the same.
+RZHYHU��XVLQJ�14&�GDWLQJ�DQG�WKH�SUHVHQFH�RI�(XURSH�IDXQDO�UHPDLQV�DQG�(XURSHDQ�VW\OH�PRWLIV�RQ�FHUDPLFV�DQG�¿JXULQHV��LW�ZDV�GHWHUPLQHG�WKDW�FRORQLDO�SHULRG�GHSRVLWV�IURP�WKH���WK�DQG���WK�centuries at Xaltocan can be distinguished by the presence of large quantities of Aztec IV pottery.
6HFRQG��QR�VWDWLVWLFDOO\�YLDEOH�PRGHO� LQ�ZKLFK�3KDVHV���DQG���ZHUH�DEXWWLQJ�DQG�QRW�FRPSOHWHO\�RYHUODSSLQJ�FRXOG�EH�FUHDWHG��OHDGLQJ�2YHUKROW]HU�WR�FRQFOXGH�WKDW�%UXP¿HO¶V�3KDVH���GLG�QRW�H[LVW��DQG� WKDW� VXFK�GHSRVLWV� OLNHO\� UHSUHVHQWHG�PL[HG�FRQWH[WV� �2YHUKROW]HU������������+RZHYHU�� WKLV�H[SODQDWLRQ�ZDV�XQVDWLVIDFWRU\��VLQFH�LW�GLG�QRW�DFFRXQW�IRU�RWKHU�HYLGHQFH�IRU�WKH�SKDVH��VSHFL¿FDOO\�the association of mixed Aztec I and II and pure Aztec II deposits with distinct redwares and other
WH[WV�SHUWDLQLQJ�WR�%UXP¿HO¶V�3KDVHV���DQG��²PLGGHQV�ZLWK�$]WHF�,�DQG�,,�SRWWHU\�DQG�PLGGHQV�with purely Aztec II pottery, respectively—with the same associated redwares and other decorated
ceramics. Moreover, these contexts produced 14&�GHWHUPLQDWLRQV�WKDW�RYHUODSSHG�HQWLUHO\��)XUWKHU�examination of the data revealed that Phase 2 contexts always represented continuous occupation
FDQ¶V�ODNHVKRUH��7KHVH�VHWWOHUV�MRLQHG�WKH�RULJLQDO�LQKDELWDQWV��ZKR�DGRSWHG�VRPH�$]WHF�,,�SRWWHU\�ZKLOH�VWLOO�XVLQJ�WKH�$]WHF�,�ZDUH�WKH\�KDG�FRQVXPHG�IRU�FHQWXULHV��7KXV��LW�ZDV�FOHDU�WKDW�%UXP¿HO¶V�Phase 2 was “real” in the sense that it represented an assemblage of materials used by a set of people
6HYHUDO�TXHVWLRQV�UHPDLQ�UHJDUGLQJ�WKLV�WUDQVLWLRQ��KRZHYHU��6RPH�FRQWH[WV�IURP�%UXP¿HO¶V�H[FD-vations showed Aztec II deposits overlying deposits containing mixed Aztec I and II. This suggests
that some residents may have stopped using Aztec I pottery before the arrival of Aztec III wares.
Further research, including additional extensive excavations of domestic contexts, is necessary to
determine the likely variable timing of this shift. For now, this study acknowledges and models the
basic chronology of a phase associated with the arrival of Aztec II pottery, and comments on the
timing of the use of Aztec I and II wares in the two households for which we currently have data.
6WUDWLJUDSKLF�LQIRUPDWLRQ�RQ�WKH�RUGHULQJ�RI�$]WHF�,±,9�GHSRVLWV�IURP�WKHVH�UHFHQW�H[FDYDWLRQV�DW�Xaltocan validates the new four-phase chronology proposed here. In all excavations at the site, a
similar stratigraphic sequence was evident: deposits containing Aztec III ceramics overlie contexts
SRWWHU\��LQ�WKHVH��$]WHF�,,,�DQG�,9�FHUDPLFV�DUH�VLPLODUO\�XELTXLWRXV��7KH�QHZ�PRGHO�SURSRVHG�KHUH�FRPELQHV�GDWHV�IURP�GHSRVLWV�WKDW�ZRXOG�KDYH�EHHQ�FDOOHG�3KDVHV���DQG���LQ�%UXP¿HO¶V�FKURQRORJ\��and separates her Phase 4 into two distinct phases. 14&�GDWHV�IURP�GHSRVLWV�DVVRFLDWHG�ZLWK�WKHVH�IRXU�phases—determined by the presence of Aztec I, Aztec I and II or pure Aztec II, Aztec III, or Aztec
III and IV ceramics—and prior knowledge of archaeological context were used to make Bayesian
statistical models.
)LJXUH����1RUWK�SURÀOH�RI�2S�*��*���VKRZLQJ�+DL�SKDVH�GRPHVWLF�DUFKLWHFWXUH�VXSHULPSRVHG�RQ�'HKH�SKDVH�GRPHVWLF�DUFKLWHFWXUH�HQFRXQWHUHG�LQ�WKH�2S�*�*��H[FDYDWLRQ�XQLWV��7KUHH�14C dates from this occupational
)LJXUH����(DVW�SURÀOH�RI�(VWH����VKRZLQJ�7ODOOL�DQG�,VOD�SKDVH�GRPHVWLF�DUFKLWHFWXUH�VXSHULPSRVHG�RQ�+DL�SKDVH�GRPHVWLF�architecture encountered on Structure 122. Fifteen 14C dates from this occupational sequence are included in the Bayesian sta-
����Bayesian Chronology for Occupation at Xaltocan, Mexico
7DEOH����5LP�VKHUG�FRXQWV�IURP�PLGGHQV�DQG�¿OO�GHSRVLWV�DVVRFLDWHG�ZLWK�WKH�'HKH��+DL��7ODOOL��DQG�Isla phase domestic architecture seen in Figures 4 and 5.
The chronological model proposed in this article is qualitatively, quantitatively, and methodologi-
cally different from its precursor. Qualitative sample context merits some attention, as the chrono-
logical model proposed for the later phases is based predominantly on samples from vertical and
horizontal household excavations from two areas of the site. The model it is designed to replace was
based entirely on samples from test pit excavations. Test pit samples, by virtue of being distributed
widely across the site, result in chronologies that are more likely to be representative of popula-
tion-wide trends. By contrast, single households might not have been occupied for the full range
of a given ceramic phase, or their occupants could have adopted the use of a diagnostic ceramic
type earlier or later than most residents. Thus, the use of 14&�VDPSOHV�IURP�IHZHU�ORFDWLRQV�DFURVV�the site may lead to sampling bias. In the case of the samples used for this article, the concordance
between individual dates from Aztec II contexts in two distinct parts of the site—centrally located
2S�=RF��DQG�2S�(VWH��RQ�WKH�VRXWKHDVWHUQ�HGJH�RI�WKH�LVODQG��VHH�)LJXUH���²VXJJHVWV�WKDW�VDPSOLQJ�bias is not an issue. Unfortunately, no other Aztec III or colonial period household contexts have
been excavated at Xaltocan, but the dates we do have correspond with published chronologies for
other sites. This concordance suggests that no severe sampling bias exists. Sampling bias issues
notwithstanding, samples from extensive household excavations facilitate a better understanding of
stratigraphy and sample context, and this understanding of context leads to more accurate and more
SUHFLVH�FKURQRORJLHV��7KH�DXWKRU�VXJJHVWV�WKDW�WKLV�IDFWRU�LV�UHVSRQVLEOH�IRU�WKH�VLJQL¿FDQW�FKDQJH�LQ�chronology proposed in this article.
Quantitatively speaking, continued research at the site has increased the available number of 14&�GHWHUPLQDWLRQV�IURP�WKH����XVHG�LQ�%UXP¿HO¶V�FKURQRORJ\�WR�D�VDPSOH�VL]H�RI��������RI�ZKLFK�ZHUH�LQFOXGHG�LQ�WKH�PRGHO��7KH�VLJQL¿FDQWO\�ODUJHU�QXPEHU�RI�GHWHUPLQDWLRQV�IDFLOLWDWHV�WKH�FRQVWUXF-WLRQ�RI�D�PXFK�WLJKWHU�FKURQRORJ\�DQG�DOORZV�XV�WR�LGHQWLI\�RXWOLHU�VDPSOHV�PRUH�HDVLO\��2I�WKH����RULJLQDO�GDWHV��¿YH�ZHUH�H[FOXGHG�DV�RXWOLHUV��UHVXOWLQJ�LQ�D�¿QDO�PRGHO�VDPSOH�VL]H�RI����14&�GDWHV�
1084 L Overholtzer
Methodologically, most previous chronological studies in central Mexico, including that of Brum-
¿HO��KDYH�IRFXVHG�RQ�FHUDPLF�VHULDWLRQ��,Q�RUGHU�WR�DFKLHYH�¿QHU�WHPSRUDO�UHVROXWLRQ�LQ�0RUHORV�DQG�WKH�7ROXFD�9DOOH\��0LFKDHO�6PLWK��6PLWK�������6PLWK�DQG�'RHUVKXN�������+DUH�DQG�6PLWK�������6PLWK�HW�DO��������DSSOLHG�PXOWLSOH�WHFKQLTXHV��LQFOXGLQJ�WKH�FDOFXODWLRQ�RI�(XFOLGLDQ�GLVWDQFHV�DQG�nonmetric multidimensional scaling. In addition, Smith and colleagues seriated whole excavated
contexts, rather than relying on the presence of diagnostic ceramic types such as Aztec Black-on-
ods to 14&�GDWHV�KDV�DOVR�VKRZQ�VLJQL¿FDQW�SURPLVH�IRU�WKH�FUHDWLRQ�RI�PRUH�SUHFLVH�SKDVH�EDVHG�chronologies.
The Bayesian approach to 14&�GDWLQJ��%XFN�HW�DO��������LV�D�ZD\�WR�FRPELQH�DUFKDHRORJLFDO�NQRZO-edge on the nature of the sample, archaeological context, and stratigraphy, called “prior informa-
tion” in Bayesian terminology, with explicit, probabilistic modeling of date estimates. The use of a priori knowledge in the interpretation of data is the fundamental difference between Bayesian and
FODVVLFDO�DSSURDFKHV�WR�VWDWLVWLFDO�LQIHUHQFH��%XFN�HW�DO������������)RU�H[DPSOH��DUFKDHRORJLVWV�FDQ�use the information that a set of samples comes from a stratigraphic sequence—sample A is older
WKDQ�VDPSOH�%��ZKLFK�LV�ROGHU�WKDQ�VDPSOH�&²WR�FUHDWH�D�PRUH�SUHFLVH�SUREDELOLVWLF�UDQJH�IRU�HDFK�GDWH��%HFDXVH�SULRU� LQIRUPDWLRQ�FDQ�JUHDWO\� LQÀXHQFH� WKH�UHVXOWLQJ�FKURQRORJ\��VFKRODUV�PXVW�EH�FDUHIXO�ZKHQ�WUDQVODWLQJ�WKHLU�DUFKDHRORJLFDO�NQRZOHGJH�LQWR�VWDWLVWLFDO�LQSXWV��%XFN�HW�DO������������Archaeologists should create multiple models, evaluate their robusticity, and propose the preferred
model. This combination of archaeological knowledge and probabilistic modeling results in better
Archaeologists have applied Bayesian theories to the interpretation of 14&�GDWHV�RI�VLQJOH�PRQX-
PHQWV��VXFK�DV�WKH�VWRQH�FLUFOHV�DW�6WRQHKHQJH��%D\OLVV�HW�DO���������VLQJOH�HYHQWV��VXFK�DV�WKH�GDWH�RI�WKH�HUXSWLRQ�RI�6DQWRULQL��)ULHGULFK�HW�DO���������DQG�FXOWXUDO�VHULHV�DQG�VHTXHQFHV�LQ�&KLQD�DQG�WKH�$HJHDQ�%URQ]H�$JH��/X�HW�DO��������0DQQLQJ�HW�DO���������:LWKLQ�FHQWUDO�0H[LFR��%D\HVLDQ�statistical modeling has been applied to traditional phase designations at the capital of Teotihuacan,
%D\HVLDQ�VWDWLVWLFDO�PRGHOLQJ�ZDV�SHUIRUPHG�XVLQJ�%&DO�DQG�2[&DO�Y������ERWK�RQOLQH�%D\HVLDQ�14&�FDOLEUDWLRQ�WRROV��%&DO��KWWS���EFDO�VKHI¿HOG�DF�XN��LV hosted by the Department of Probability and
6WDWLVWLFV�DW�WKH�8QLYHUVLW\�RI�6KHI¿HOG��%XFN�HW�DO���������ZKLOH�2[&DO�LV�KRVWHG�E\�WKH�5DGLRFDU-ERQ�$FFHOHUDWRU�8QLW�DW�2[IRUG�8QLYHUVLW\��%URQN�5DPVH\�����D���1HDUO\�LGHQWLFDO�UHVXOWV�ZHUH�REWDLQHG��FI��2YHUKROW]HU���������±����ZLWK�2[&DO�SURGXFLQJ�SRVWHULRU�FDOHQGDU�\HDU�SUREDELOLW\�GLVWULEXWLRQV�WKDW�ZHUH�VOLJKWO\�PRUH�FRQVHUYDWLYH��,Q�VRPH�FDVHV��WKH�GLVWULEXWLRQV�ZHUH�WKH�VDPH��in others, the date ranges were 10 yr longer. In all cases, the ranges produced with the two models
overlapped completely, and the medians for the beginnings and ends of phases were identical. This
DUWLFOH�UHSRUWV�WKH�PRUH�FRQVHUYDWLYH�2[&DO�HVWLPDWHV�DQG�QRWHV�WKDW�WKH�FRPSDUDEOH�UHVXOWV�LQGLFDWH�that the proposed model is robust. In accordance with stratigraphic evidence, all phases in the model
are abutting, that is, there are neither chronological gaps nor overlapping periods. In addition, 1521
was set as the absolute calendar date separating the Aztec III and Aztec III/IV phases. In this article,
1085Bayesian Chronology for Occupation at Xaltocan, Mexico
EHORZ������UHMHFWLRQ�ZDV�FRQVLGHUHG��%URQN�5DPVH\�����E���,Q�¿YH�RI�WKH�VL[�FDVHV��WKH�GDWHV�ZHUH�H[FOXGHG��,Q�WKH�VL[WK��VDPSOH�%HWD��������KDG�D�ERUGHUOLQH�DJUHHPHQW�LQGH[�RI������7KLV�VDPSOH�UHSUHVHQWV�WKH�¿UVW�DSSHDUDQFH�RI�$]WHF�,,�SRWWHU\�DORQJVLGH�$]WHF�,�LQ�RQH�KRXVHKROG��:KLOH�LW�LV�possible that the sample represents a case of old wood, it is also possible that this household may
have been one of the earliest adopters of Aztec II pottery and may have done so before the arrival of
Aztec II-using migrants. Thus, the sample was not excluded, but its low agreement was noted. After
H[FOXGLQJ�WKH�¿YH�RXWOLHUV��WKH�PRGHO�ZDV�UHUXQ�DQG�WKH�RYHUDOO�DJUHHPHQW�LQGH[�ZDV�FDOFXODWHG�WR�be 80%, indicating that the model is acceptable.
7KH� UHVXOWLQJ�PRGHO� �$SSHQGL[�Table S2��)LJXUHV��±����SURGXFHG� LQGLYLGXDO� SRVWHULRU� FDOHQGDU�year probability distributions with more precise date ranges than those of the previously accepted
FKURQRORJ\��$W��ı��WKH�DYHUDJH�UDQJH�RI�HUURU�IRU�DOO�FDOLEUDWHG�GDWHV�ZDV�UHGXFHG�E\�D�PRGHVW������More importantly, the model also created posterior calendar year probability distributions for phase
boundaries. The mean values were used to provide estimates of the beginning and end dates of the
ics, called the Hai �³ODQG´�LQ�2WRPt��SKDVH��GDWHV�WR�$'�����±������RFFXSDWLRQ�DIWHU�WKH�DUULYDO�RI�Aztec III ceramics, called the Tlalli��³ODQG´�LQ�1DKXDWO��SKDVH��GDWHV�WR�$'�����±������DQG�FRORQLDO�occupation, called the Isla��³LVODQG´�LQ�6SDQLVK��SKDVH��GDWHV�WR�$'�����±������
7DEOH����&RPSDULVRQ�RI�%UXP¿HO�DQG�UHYLVHG�FKURQRORJLHV�%UXP¿HO������E��PRGHO Revised model
Phase &HUDPLFV n
Range of
medians
�FDO�$'� Phase &HUDPLFV n
Range of
medians
�FDO�$'�
Phase
range
�FDO�$'�1 Aztec I 4 �����±��� Dehe Aztec I 24 �����±����� �����±����
2Aztec I &
II2 ����±����
+DLAztec I & II
or Aztec II
16 �����±���� ����±����� Aztec II 2 ����±����
4Aztec III &
IV1 1420
Tlalli Aztec III 11 �����±���� ����±����Isla Aztec III & IV 2 �����±���� ����±����
)LJXUH����3UREDELOLW\�GLVWULEXWLRQ�IRU�ERXQGDU\�WUDQVLWLRQ��6DPSOHV�FDOLEUDWHG�LQGLYLGXDOO\�DUH�VKRZQ�LQ�JUD\�OLQHV�DQG�with Bayesian model as black areas.
7KH�SKDVHV�ZHUH�JLYHQ�QDPHV�UDWKHU�WKDQ�QXPEHUV�WR�DYRLG�FRQIXVLRQ�ZLWK�%UXP¿HO¶V�3KDVH��±��designations. Otomí words were chosen for the periods when Xaltocan is thought to have been an
ually are shown in gray lines and with Bayesian model as black areas.
1087Bayesian Chronology for Occupation at Xaltocan, Mexico
2WRPt�SROLW\��D�1DKXDWO�ZRUG�IRU�WKH�SHULRG�GXULQJ�ZKLFK�ZH�VHH�D�GLVWDQFLQJ�IURP�2WRPt�HWKQLFLW\�and during which Xaltocan was subsequently incorporated into the Aztec Empire, and a Spanish
word for the colonial phase. All words chosen relate to the fact that Xaltocan was located on an
SOLHG�E\�%UXP¿HO¶V������D��LQWHUSUHWDWLRQ�RI�PHGLDQ�14&�GDWHV��,Q�DGGLWLRQ�WR�WKH�QHZ�GLVWLQFWLRQ�EHWZHHQ�SUH�+LVSDQLF�$]WHF�,,,�DQG�HDUO\�FRORQLDO�$]WHF�,,,�DQG�,9��WKH�$]WHF�,,�DQG�,,,�GDWH�UDQJHV�UHVXOWLQJ�IURP�DQDO\VLV�RI�QHZ�VDPSOHV�IURP�VHFXUH�KRXVHKROG�FRQWH[WV�DUH�VLJQL¿FDQWO\�HDUOLHU��7KH�large number of determinations used in the proposed Bayesian model and the excellent understand-
ing of the stratigraphic context of the samples lends credence to these revisions.
)LJXUH����3UREDELOLW\�GLVWULEXWLRQ�IRU�+DL�SKDVH�GHWHUPLQDWLRQV��6DPSOHV�FDOLEUDWHG�LQGLYLGXDOO\�DUH�VKRZQ�in gray lines and with Bayesian model as black areas.
1088 L Overholtzer
)LJXUH����3UREDELOLW\�GLVWULEXWLRQ�IRU�7ODOOL�SKDVH�GHWHUPLQDWLRQV��6DPSOHV�FDOLEUDWHG�LQGLYLGXDOO\�DUH�VKRZQ�LQ�JUD\�lines and with Bayesian model as black areas.
)LJXUH�����3UREDELOLW\�GLVWULEXWLRQ�IRU�,VOD�SKDVH�GHWHUPLQDWLRQV��6DPSOHV�FDOLEUDWHG�LQGLYLGXDOO\�DUH�VKRZQ�LQ�gray lines and with Bayesian model as black areas.
����Bayesian Chronology for Occupation at Xaltocan, Mexico
In addition, the resulting posterior calendar year probability distributions for several 14&�GDWHV�SUHV-ents preliminary evidence for the timing of the contemporaneous consumption of Aztec I and II
wares by some residents while others used Aztec II pottery. This sample, while admittedly small,
SURYLGHV�WKH�¿UVW�VXFK�GDWD�VHW�IRU�FHQWUDO�0H[LFR��([FDYDWLRQV�IURP�WZR�KRXVHV�RQ�WKH�HDVWHUQ�HGJH�of the island indicated that the island had been expanded and people using Aztec II pottery had set-
tled on the new lakeshore by AD 1240. Three 14&�GDWHV��$$�������$$�������DQG�$$�������IURP�WKH�HDUOLHVW�SXUH�$]WHF�,,�PLGGHQV�ZHUH�DOO�HVWLPDWHG�LQ�WKH�%D\HVLDQ�PRGHO�WR�FDO�$'�����±������Also by 1240, households located in the center of Xaltocan had adopted Aztec II alongside Aztec I
SRWWHU\��)RU�WKH�2S�=�KRXVHKROG�H[FDYDWHG�E\�'H�/XFLD�DQG�%UXP¿HO��'H�/XFLD��������UHSRUWV�WKDW�14&�GDWHV�SODFH�$]WHF�,�RFFXSDWLRQ�RI�WKH�KRXVH�LQ�WKH�PLG���WK�WKURXJK�PLG���WK�FHQWXULHV��7KH�consumption of Aztec II pottery alongside Aztec I in this household is represented by a single sam-
formal hearth associated with a sealed midden containing Aztec I and II rims. This determination
ZDV�FDOLEUDWHG�WR�KDYH�D�����OLNHOLKRRG�RI�GDWLQJ�WR�FDO�$'�����±������WKRXJK�XQIRUWXQDWHO\��WKLV�date has poor agreement in the model, as mentioned previously. The declining use of Aztec I in this
$]WHF�,�DQG�,,�KRXVHKROG�XVH�DW�;DOWRFDQ�FHUWDLQO\�EHJLQV�E\�WKH�PLG���WK�FHQWXU\��WKRXJK�ZKHQ�it ends remains unclear. More dates from additional household contexts are needed to assess the
potentially variable chronology of this consumption.
Figure 11 Comparison of cultural sequences in central Mexico, including new proposed chronology for Xaltocan
���� L Overholtzer
CONCLUSION
Precise understandings of past chronologies have always been a central concern of archaeologists.
1HZ�WRROV�VXFK�DV�%D\HVLDQ�VWDWLVWLFDO�PRGHOLQJ�RI�14&�GDWHV�SHUPLW�WKH�FRQVWUXFWLRQ�RI�FKURQRORJLHV�with more precise phases and individual 14&�GDWHV��,Q�WKLV�DUWLFOH��%D\HVLDQ�VWDWLVWLFDO�PRGHOLQJ�ZDV�applied to a suite of 54 14&�GDWHV�UDQJLQJ�IURP�FDO�$'����±�����IURP�KRXVHKROG�H[FDYDWLRQV�DW�Xaltocan. This modeling was compared to earlier chronologies at the site, and a new chronology
was proposed. The conclusions reached here are of broad relevance to central Mexican archaeolo-
J\�EHFDXVH�WKLV�DUWLFOH�SURYLGHV�DQ�H[WHQVLYH�UDGLRPHWULF�GDWD�VHW�IURP�VHDOHG��VWUDWL¿HG�GRPHVWLF�FRQWH[WV�VSDQQLQJ�WKH�(DUO\�3RVWFODVVLF�WR�(DUO\�&RORQLDO�SHULRGV��WKH�¿UVW�GDWD�VHW�RI�LWV�NLQG�LQ�WKH�Basin of Mexico.
The new chronology—composed of the Dehe, Hai, Tlalli, and Isla�SKDVHV²FODUL¿HV�WKH�WHPSRUDO�distribution of all four Postclassic and colonial indigenous Black-on-Orange ceramic types at Xalto-
U\��$]WHF�,,�SRWWHU\�ZDV�DGRSWHG�LQ�KRXVHKROGV�LQ�WKH�PLG���WK�FHQWXU\��DORQJVLGH�$]WHF�,�FHUDPLFV�in descendants of the original inhabitants, and exclusively by settlers from a place without a tradi-
WLRQ�RI�XVLQJ�$]WHF�,�FHUDPLFV��$]WHF�,,,�SRWWHU\�ZDV�FRQVXPHG�E\�$'�������VRPH����\U�EHIRUH�WKH�VLWH¶V�FRQTXHVW�DQG����\U�SULRU�WR�LQFRUSRUDWLRQ�LQWR�WKH�$]WHF�(PSLUH��7KH�7ODOOL�SKDVH��WKHUHIRUH��does not correspond only to the period of Aztec rule at Xaltocan, but rather also includes several
decades of pre-conquest life. Distinguishing between pre- and post-conquest practices at Xaltocan
PXVW�UHO\�RQ�RWKHU�PHWKRGV��VXFK�DV�%D\HVLDQ�VWDWLVWLFDO�PRGHOLQJ�RI�VSHFL¿F�VWUDWLJUDSKLF�VHTXHQF-HV��DV�ZDV�GRQH�IRU�SDUWLFXODU�KRXVHKROG�FRQWH[WV��2YHUKROW]HU��������RU�SHUKDSV�VHULDWLRQ�RI�ZKROH�contexts using multiple statistical techniques, as Smith has done for sites outside the Basin of Mexi-
FR��6PLWK�������6PLWK�DQG�'RHUVKXN�������+DUH�DQG�6PLWK�������6PLWK�HW�DO���������8QIRUWXQDWHO\��H[LVWLQJ�GDWD�DUH�LQVXI¿FLHQW�WR�FKDUDFWHUL]H�IXOO\�WKH�FKURQRORJ\�RI�$]WHF�,9�XVH�DW�;DOWRFDQ��EXW�preliminary evidence suggests that Aztec IV pottery was consumed in low frequencies during the
SUH�+LVSDQLF�SHULRG��DQG�HDUO\�FRORQLDO�UHVLGHQWV�XVHG�PDLQO\�$]WHF�,,,�DQG�,9�SRWWHU\�XQWLO�QHDUO\�the end of the 17th century, if not later. While this research makes progress in clarifying dates for
WKH�DUULYDO�RI�$]WHF�,±,9�SRWWHU\�DW�;DOWRFDQ��DQG�VXJJHVWV�WKDW�DGRSWLRQ�RI�WKHVH�ZDUHV�PD\�KDYH�varied by household or neighborhood, additional excavations of household contexts across the site
DUH�QHHGHG�WR�GH¿QH�WKH�QDWXUH�RI�WKDW�YDULDWLRQ��H�J��KRZ�ORQJ�GLIIHUHQW�KRXVHKROGV�FRQWLQXHG�WR�XVH�Aztec I pottery alongside the newly adopted Aztec II ware.
SORUHU�*UDQW�IURP�WKH�1DWLRQDO�*HRJUDSKLF�6RFLHW\��D�*UDQW�LQ�$LG�RI�5HVHDUFK�IURP�WKH�6LJPD�;L�)RXQGDWLRQ��D�1DWLRQDO�6FLHQFH�)RXQGDWLRQ�*UDGXDWH�5HVHDUFK�)HOORZVKLS��D�*UDGXDWH�5HVHDUFK�*UDQW�IURP�WKH�8QLYHUVLW\�5HVHDUFK�*UDQWV�&RPPLWWHH�DW�1RUWKZHVWHUQ�8QLYHUVLW\��DQG�D�5HVHDUFK�*UDQW�IURP�WKH�/H&URQ�)RVWHU�DQG�)ULHQGV�RI�$QWKURSRORJ\�DW�1RUWKZHVWHUQ�8QLYHUVLW\��-XDQ�-RHO�Viveros�6iQFKH]�GUHZ�)LJXUH����DOO�RWKHU�¿JXUHV�DUH�E\�WKH�DXWKRU��,�ZRXOG�OLNH�WR�WKDQN�(OL]DEHWK�%UXP¿HO��5RVHPDU\�-R\FH��&\QWKLD�5RELQ��DQG�0DU\�:HLVPDQWHO�IRU�WKHLU�IHHGEDFN�DQG�VXSSRUW��-�+HDWK�$QGHUVRQ��7UDYLV�%UXFH��DQG�$QG\�0XOOHQ�DOVR�SURYLGHG�KHOSIXO�VXJJHVWLRQV��WKRXJK�DOO�remaining errors are my own.
����Bayesian Chronology for Occupation at Xaltocan, Mexico
REFERENCES
%D\OLVV�$��%URQN�5DPVH\�&��0F&RUPDF�)*��������'DWLQJ�6WRQHKHQJH��Proceedings of the British Academy������²���
%UXPÀHO�(0������D��&HUDPLF�FKURQRORJ\�DW�;DO-tocan. In: Production and Power at Postclas-sic Xaltocan�� 0H[LFR� &LW\� DQG� 3LWWVEXUJK��,QVWLWXWR�1DFLRQDO�GH�$QWURSRORJtD�H�+LVWRULD�DQG�8QLYHUVLW\�RI�3LWWVEXUJK��S����²���
%UXPÀHO� (0�� ����E�� Production and Power at Postclassic Xaltocan��0H[LFR�&LW\� DQG�3LWWV-burgh: Instituto Nacional de Antropología e
+LVWRULD�DQG�8QLYHUVLW\�RI�3LWWVEXUJK�%UXPÀHO�(0��������Estrategias de las Unidades
ogy in western Morelos, Mexico. Latin American Antiquity ��������²����
6PLWK�0(��%RUHMV]D�$��+XVWHU�$��)UHGHULFN�&'��5RGUt-JXH]�/ySH]� ,��+HDWK�6PLWK�&�� ������$]WHF� SHULRG�houses and terraces at Calixtlahuaca: the changing
morphology of a Mesoamerican hilltop urban center.
Journal of Field Archaeology ���������²���7ROVWR\�3��������6XUIDFH�VXUYH\�RI� WKH�QRUWKHUQ�9DOOH\�
RI� 0H[LFR�� WKH� &ODVVLF� DQG� 3RVWFODVVLF� SHULRGV��Transactions of the American Philosophical Society