Document Type


Date of Award



Engelbert site (N.Y.), Indigenous People, New York, Dwellings

Degree Name

Doctor of Philosophy (PhD)



First Advisor

Fred T. Plog III

Second Advisor

Hans Hoffman

Third Advisor

Eugene L. Sterud


In the last decade archeologists have been taking a fresh look at their data. Among other things, they have discovered new units of analysis. One phenomenon that has heretofore been overlooked as a unit of analysis is the pit. In the pages which follow, pits are treated as artifacts. They are described, classified, and viewed in time and space as any other class of artifact. The main focus of the essay is typological. Since pits are empirical objects, they have attributes which vary in time and space, and perhaps in function as well. If this variation is patterned they can be grouped into types. Some of these groups are perhaps meaningful for understanding the past.

Archeologists in the Northeast have assumed since the 1930s that pits and other facilities (non-portable containers) represent different types. They have traditionally classified pits according to an intuitive functional typology which relies chiefly on contents, and which is more or less shared among archeologists of the region. This pseudo-functional typology is loosely tied to ethnographic data by simple analogy (e.g., Indians used large pits for storage; therefore, all large pits are storage pits). A number of other pit typologies based on size and shape attributes have recently been added to the literature. Generally these typologies are specific to individual site reports, but some archeologists have gone so far as to suggest that certain sizes and shapes of pit are characteristic of certain cultures.

The use of typologies assumes that pit types are significant and useful. Neither assumption has been tested, chiefly because the assumptions are implicit, but to some extent because sample sizes have usually been so small. A large sample of pit data from the Engelbert site presents a unique opportunity to apply rigorous techniques to the problem of defining types and testing the correlations of types with other kinds of data useful for interpretations, such as ethnographic data and time-space distributions.

Because formal typological methods have never been applied to pits, the appropriate attributes have not been identified. The intuitive typology that Northeastern archeologists have been using recognizes only contents and size and shape as attributes. Pits have many other attributes, however, such as stratigraphy, which may carry information about past cultures. Part of the problem in attempting to form a typology, then, is in selecting and defining attributes which seem to be promising for discriminating between types. Another part of the problem is finding and applying appropriate typological techniques.

Much of the essay which follows is devoted to the problem of identifying appropriate attributes. This is approached from 2 perspectives: (1) The ethnographic data, in order to extract attributes useful for constructing arguments of analogy in interpretations of archeological pits; all possibly relevant attributes were considered. (2) The archeological data on pits from the Engelbert site, in order to extract attributes which are useful for defining units for comparison of artifactual and time-space data; this analysis focused on attributes of form as the most likely to reflect original purpose of construction and stylistic variation.

The second part of the problem, selecting typological methods, involves finding a combination of techniques for multivariate analysis of both size-shape (interval) and stratigraphic (nominal) variables. Among the techniques available to the writer the most appropriate ones seemed to be cluster analysis and discriminant function analysis for interval variables and association analysis for the nominal ones. The analysis of interval variables produced 5 well-discriminated types of pits which appear to be correlated with some of the ethnographic attributes, and to be non-randomly distributed in time and space. Because association analysis for archeological data is in its formative stages there were some problems which prevented a satisfactory conclusion to the analysis of nominal data. The partial results suggest that it is worthwhile pursuing this avenue of research.

Overall, the typological analysis promises some fruitful results for pit studies. The results of even this limited study indicate that pits can profitably be used as units of comparison for temporal, spatial, and functional distribution analysis. For multicomponent sites where strata, houses, or other provenience units are lacking, pits provide a unit for intrasite analysis where no other exists.