The following definitions are employed throughout the present work as consistently as seems necessary for purposes of conceptual clarity. These definitions are different in many cases from those employed by other authors, and even from the definitions that are normatively used in philosophy and other disciplines. They should be understood as our own definitions, relative to the present work; there is no suggestion that they are or should be the only definitions. Within each definition, words appearing in bold-face type are themselves defined elsewhere in the glossary. At the end of most entries there is a reference to a particular chapter or chapters in this book where the concept under consideration is more fully discussed.

Abstraction. The mental process involved in creating a category, by mentally grouping things together, and thinking of them collectively, on the basis of certain shared characteristics, while ignoring other characteristics that are not shared. (See Chapter 4.)

Acquired gestalt. As used in the present book, a gestalt is an immediate sensory image which we form upon examining a particular entity, and which indicates to us that the entity is a member of a particular type, without any process of conscious analysis. An acquired gestalt is one which we are able to form only after we have had considerable experience in handling the type in question, in contrast to an intuitive gestalt which we are able to form even before types have been formally defined. (See Chapter 4.)

The preceding section has placed the Tuxtlas potters in an ecological context. Information on the producer environment, coupled with data on pottery production and consumption, has exposed the dynamic nature of production-related activities. It is now time to consider how such a system might be approached from an archaeological perspective.

This chapter explores some of the fundamental issues involved in the identification and interpretation of pottery production. The first section considers how archaeologists establish the existence of production areas. Both epistemological and empirical issues are raised. This discussion argues that confusion in these two arenas presently hampers our inferential approach to ceramic production. Through an examination of these issues the stage is set for addressing models of production organization.

The second section focuses on these models and critically evaluates their utility. These approaches are shown to be overly static with only a limited applicability to many archaeological goals. This discussion underscores the need for a more rigorous approach to investigating production activities.

A frame of reference

After many years studying ceramics at the regional and inter-regional level, Mesoamerican archaeologists have recently begun investigating the actual location of ceramic production activities. This new emphasis on “microprovenience” studies (Rice 1981:219) is an outgrowth of research questions that seek a more detailed understanding of the producer's socioeconomic environment (e.g. Deal 1988:113; Feinman 1985; Santley et al. 1989; Stark 1985:158–160). Research into the scale of a production system is conducted more effectively when the actual production entities can be identified and their characteristics evaluated.

In its early days the Typological Debate was somewhat beclouded by the lack of an adequate conceptual vocabulary, a problem that has been common to all sciences in their formative stages. Commentators like Rouse (1939; 1960), Krieger (1944), Brew (1946: 44–66) and Taylor (1948: 113–51) had to use too few words to mean too many things, and it is not surprising that their ideas were not always expressed very precisely. They rarely distinguished between entities and concepts, between variables and attributes, between classifying and sorting, or between typology and taxonomy.

Lack of an adequate vocabulary is no longer the problem that it once was. In some respects we now have more conceptual terminology than we can reasonably use. Many typologists in the last thirty years have attempted to solve the stubborn problems of artifact classification, or to reduce the gap between theory and practice, by offering new conceptual formulations. Unfortunately these have often been expressed in the form of false dichotomies, as we will see later in the chapter, and their effect has at times been to further becloud the Typological Debate rather than to clarify it.

The problem today is not so much one of inadequate conceptual tools, as of tools unused or unsystematically used. In this chapter we will consider a number of persistent problems with regard to classification and typology whose basis is at least partly conceptual.

One of the most insightful early discussions of archaeological typology and its problems was “The use and abuse of taxonomy,” by J. O. Brew (1946: 44–66). (In most contexts “taxonomy” was to Brew what “typology” is to us.) Brew was not the first to recognize the theoretical problems involved in archaeological classification (cf. Kluckhohn 1939: 338; Rouse 1939: 9–35), but he was one of the first who correctly understood their nature. It was not (as Kluckhohn had implied) that artifact types had no theoretical significance, but that their significance was widely misinterpreted. Brew saw this as evidence that typological practice was being used in ways inappropriate to theory-building – hence the title of his chapter. Here we will argue, conversely, that theory has often been used in ways inappropriate to typology-building.

We have suggested at many points in this book that there really was not, and is not, anything seriously wrong with the practice of artifact classification. By the time that Brew wrote, however, it had largely achieved its original goal of erecting a basic time/space grid for the contextualization of North American archaeological materials. Precisely because this edifice was now largely complete, archaeologists were beginning to look around for new questions to ask. As they did so, they began to criticize the old typologies because they were not answering the new questions. As we will see a little later, this is one of the frequent consequences of paradigm shifts in the social sciences.

In Chapter 3 we made the point that a lexicon is a congeries of words, a classification is a set of classes, but a typology is a system of types. What we mean by this is that a typology has systemic or structural features that are not entirely dictated by the nature of the constituent types, just as a language has structural features that are not dictated by its lexicon (cf. Waugh 1976: 20). In this chapter we will consider first what are the structural features of a typology, and then how they come about and what (if any) purpose they serve. The features that we will specifically discuss are those of boundedness and comprehensiveness of the typology itself, and the mutual exclusiveness, consistency of definition, equivalence, equidistance, and independence of the constituent types.

Basic structural features

Boundedness. In any classification that is to be used for sorting entities (that is to say, a typology), it must be clear at the outset what is and is not to be sorted. Hence typologies (unlike some other classifications) have quite rigidly specified and inflexible boundaries. Moreover, because sorted material is often subjected to statistical manipulations, the outer boundaries of the system cannot be adjusted to include new material which was formerly excluded, without invalidating any statistical operations that were previously performed.

Typological boundaries may be defined on the basis of either internal or external criteria, or both (see Chapter 14).

In Part II we talked in somewhat theoretical terms about how type concepts are formulated and communicated. In this chapter we return to the same issues from a more pragmatic perspective, and with reference especially to the relationship between types and attributes.

Discussion in the last two chapters might suggest that typology-making involves a simple and logical progression from the specification of purpose to the selection of attributes to the formulation of types on the basis of selected attributes. This is indeed the procedure recommended in many programmatic statements (Dunnell 1971b: 70–6; Spaulding 1982). Real-life typologies, however, are seldom made entirely in this way or in any single and simple way. The relationship between attributes and types is, in most typologies, a complex and multi-dimensional one.

As we saw in Chapter 5, the formulation and use of type concepts involves a number of successive stages, or processes. Processes that will concern us in the present chapter are those of formulation, definition, designation, and description. We will see that the relationship between attributes and types is somewhat different in each of the four cases, obliging us to make a distinction between diagnostic (prescriptive) and descriptive attributes. In anticipation of the discussion we may say that attributes are prescriptive of types at one level of analysis, and descriptive at another (cf. Whallon and Brown 1982: xvi).

One of the more important advances stemming from the “essential polemics” (e.g. Dunnell 1986:25) of the last thirty years is the realization that archaeological materials are not self evident. Rather, it is the responsibility of the archaeologist to describe material patterns, identify meaningful variability, and interpret that variability. The current wealth of competing paradigms and programmatic statements is in one sense a tribute to archaeologists' increasing awareness of their role in generating inferences about the past (e.g. Watson 1986).

The theme of evaluating assumptions about the archaeological record has served as the centerpiece of the present work. If archaeologists wish to understand how ceramic production developed and integrated with prehispanic complex societies, research cannot simply focus on the archaeological record. Instead, studies dealing with contemporary production systems must be undertaken. Only in this way can both the behavior generating the pattern and the material consequences of that activity be controlled. It is the ability to document both cause and effect that enables archaeologists to establish the necessary relationships between pottery making and the material record.

For archaeologists interested in prehispanic ceramic production, the concept of actualistic research is nothing new. Anthropological accounts of contemporary potters have a long, established history (see Rice and Saffer 1982). The existing body of ethnoarchaeological information on pottery making is similarly impressive and continues to grow (e.g. Arnold 1987; Deal 1983; Hagstrum 1989; Kramer 1985). Furthermore, ethnographic accounts of ceramic production have recently been synthesized and presented in a systems perspective (Arnold 1985).

The authors of this book are familiar, in theory or in practice, with quite a number of typologies, and so presumably are most readers. Since we hope to reach a wide and varied audience, however, we cannot be sure that all or even most of our readers will be familiar with the same examples of the genre. Rather than illustrate our points with examples from a number of different typologies, we have chosen to describe in detail one particular typology, which will provide many of the illustrative examples we will discuss in subsequent chapters. The typology chosen is WYA's Classification of Medieval Nubian Pottery Wares (WYA 1986a), hereinafter called the Nubian Pottery Typology.

This seemingly egocentric procedure can be defended on several grounds. The typology is a fairly complex one, with enough internal variability to illustrate most of the points we want to make. It was developed for a specific purpose, and its utility in serving that purpose can be not only described but measured. Most importantly of all, one author (WYA) is familiar in detail not only with the typology but with its family history and secrets. This makes it possible to consider how the scheme first arose, and the hits and misses of its development – a privilege we are not often given in discussions of typologies (cf. Dunnell 1971b: 139; Gardin 1980: 81).

In the affluent 1960s and 1970s, archaeological procedures were not always closely scrutinized for their practicality or utility. However, this was an exceptional era in archaeology, one that gave researchers the unaccustomed luxury of experimenting with field and laboratory techniques whose practical value was often undetermined. In more normal times, archaeological research is usually carried out under conditions of severe financial and resource constraint. Considerations of practicality and cost-effectiveness can rarely be ignored, and must sometimes be paramount. We (or our funding agencies) are usually obliged to ask: Do our analytical procedures really tell us more than is intuitively obvious? Does increased rigor lead to measurably more useful or meaningful results? Could the same results be obtained by cheaper or less time-consuming means? Can we justify our procedures not only in terms of what we want to know, but why we want to know it? Above all, is our work likely to diminish the unknown, or merely (perhaps for the nth time), to reconfirm what is already known (see WYA 1960: 19)?

In this chapter we will be considering the issue of practical utility with specific reference to archaeological classifications and their use. In theory, two separate issues are involved here: those of practicality and of utility. Practicality refers to whether a procedure is inherently doable (feasible) and/or affordable. Utility refers to whether the results are of any use. Taken together, “practical utility” refers to whether the cost of any procedure, in time and money, is justified by the results.