Recently, the use of accelerator mass spectrometry (AMS) dating has created a minor revolution in archaeology, particularly for dating the human use and manipulation of plant foods. It is not my intention to dispute the concrete aspects of this methodology, but rather to call my colleagues' attention to something that seems very dangerous and that, if not clarified from the start, could give rise to serious problems in the future. I am referring to the practice of relying solely on absolute dates, whatever the specific method employed, to fix in time a historical event. In the case of the use of AMS dating, a revision of the previously established antiquity of a series of cultivated plants in the New World has been postulated based on new AMS dates (see J. F. Gayle, 1994, Are the First American Farmers Getting Younger? Current Anthropology 35:305-309), while similar changes to our understanding of the prehistory of plant use in the Old World have also occurred (see P. Rowley-Conwy, 1995, Making First Farmers Younger: The West European Evidence. Current Anthropology 36:346-353).
I absolutely support, as I believe every person of science must, not only the utility but also the necessity of applying new technologies that are continuously being developed. Without such technology science cannot advance. But archaeologists cannot forget the basic scientific norms of the discipline that should govern their conduct. An important aspect is that a new scientific method should be used when possible but only within the parameters of archaeology and not as an end in itself.
In the question of dating methods, such as AMS, results cannot be automatically accepted as valid, and if these results do not agree with those obtained by another methodology, it cannot be assumed sic et simpliciter that the previous work was mistaken without first completing a careful analysis of the facts. The apparent advantages of AMS dating have caused us to forget the framework within which that methodology should be considered. When this happens, it becomes possible to declare that "The single most important development in the study of early agriculture in recent years has been the advent of the radiocarbon accelerator. Individual bones and plant macrofossils may be dated, and therefore stratigraphic associations need no longer be taken on trust." (P. Rowley-Conwy, 1995, my emphasis). If this were an isolated sentence it would hardly merit our attention. But I have heard the same said by several colleagues. This kind of statement is worrisome because it ignores all the fundamental principles of archaeology. The essential value of relative chronology, context, and within context, associations, which are so important to the archaeologist, cannot be ignored. As Wheeler (1961, Arqueología de Campo. Fondo de Cultura Economica: Mexico City, p. 51) asserted, "A chronologist is not an archaeologist." Archaeologists are much more than that; their responsibility is complete historical reconstruction.
Generalizations are always dangerous, and the advantages that the AMS method offers should not lead to the reevaluation of every relative temporal sequence that has been established with respect to the use and cultivation of plants. The problem, in my judgment, must be seen from another angle. Those sites that have been problematic in terms of establishing stratigraphy as well as those in which associations have not been clear enough must be distinguished from those in which the associations are strong and the data are reliable. The former should be submitted (if it is possible to do so) to careful examination. It is not enough to date a sample and, if the date does not agree with what was obtained earlier, to discount all previous work. There is no doubt that a discrepancy in dating, especially if it is very large, should send out an alarm, but one must also examine the whole context, comparing it with the assemblage of which it forms a part, if possible, extending the excavation, and if the discrepancy is confirmed, pointing out the errors and proposing a revision. But we must always be aware that if archaeological material has been well excavated, this material should be described by at least three variable dimensions: space, time, and form (J. Deetz, 1967, Invitation to Archaeology. Natural History Press: New York, p. 9). Time is only one of the variables. And we must not forget what D. P. Dymond (1974, Archeology and History: A Plea for Reconciliation. Thames and Hudson: London, p. 31) saw with great clarity: that an object, in this case the remains of a plant, that is found in an archaeological context has at least three chronologies: a date of origin or creation; another date related to when the object was used; and, finally, a third date that records its deposition at the location where the archaeologist finds it. In order to understand this third date we must carefully verify if the object was deposited by accident or on purpose.
If a site has been well studied and all its data are in agreement, and there are no discrepancies with contemporary data from other sites, then a reevaluation is unnecessary. Neither does it matter what type of sample was used to date a specific context, as long as it has been done correctly. That date is valid for the whole assemblage.
The other facet of this issue, which for me is at least as important, is the way in which many archaeologists have employed the 14C method (and, clearly, the problem is the same for all dating techniques). A critical problem is that some archaeologists have removed samples without the necessary control, including the use of assemblages that were not excavated (e.g., F. Engel, 1966, Geo-grafia Humana Prehistorica y Agricultura Precolombina de la Quebrada de Chilca. Universidad Agraria: Lima; H. Trimborn, 1979, El Reino de Lambayeque en el Antiguo Peru. Anthropos Institut: St. Agustin, Peru). In these cases, the resulting dates have been regarded as sufficient to define a "cultural affiliation" or a "cultural state." A deterministic value attributed to the date of a sample was automatically used to define an alleged cultural context without first investigating whether all of the components were in agreement.
I am in no position to judge the AMS method and establish up to what point it provides dates that correspond with those provided by 14C. So, as the specialists have indicated to us, in order to see the comparative effects with other methods that give us dates based on an absolute calendar, we must submit the results of 14C testing to certain "calibrations," which will tell us what protocol to use with dates obtained by AMS.
J. H. Rowe, referring to radiocarbon dates, wrote a great truth that should not be forgotten and is valid for using dating methods: "...consistency among...measurements can only be determined with reference to archaeological associations and the system of relative chronology which has been set up for the area." He also notes that "to compare measurements made on samples from different associations it is necessary to refer them to a system of relative chronology" (J. H. Rowe, 1966, An Interpretation of Radiocarbon Measurements on Archaeological Samples from Peru. In Proceedings of the Sixth International Conference, Radiocarbon and Tritium Dating, pp.187-198. U.S. Atomic Energy Commission: Oak Ridge).
At this moment there are several questionable assemblages in the Americas, and their botanical components have undergone or are at present undergoing dating using the AMS method. I want to make it very clear that it has not been my intention here to refer to any one of them in particular. Neither is the spirit of this paper a polemic one. Nothing could be further from my intentions. I only wish to put forth this concern in order to remind us that archaeology is a science that has very clear rules and methodologies that are sometimes forgotten in our enthusiasm for new and exciting scientific advances. We must not forget that if we want to practice a scientific archaeology we must understand that methods of dating are in the service of our science and not the other way around. In the final analysis, only the archaeologist must judge whether the results are valid. Woolley always reminded us that the value of an exact chronology consists in its ability to put the objects we excavate--objects that a human community has left to us--in their proper relation to the world of which they are a part in such a way that isolated evidence may find its place in the ordered history of humanity (L. Woolley, 1957. Il Mestiere Dellarcheologo. Giulio Einaudi Editore: Torino, Italy, p. 167).
Laboratorio de Prehistoria
Departamento de Biologia
Universidad Peruana Cayetano Heredia
Lima 100, Peru