II.1 Classification: the Language of Observation
p. 21-24
Texte intégral
Introduction
1All scientific activities require that we organise the infinite variation of unfamiliar things so that we can make sense of them. The humble pot sherd is no exception to this. The remains of the past must be sorted and organised into archaeological units—which we variously label “type”, “class”, “series”, “group”, etc.—so that they can be used for study. This is achieved by means of classification. Its role is therefore central to archaeology. These are also the means by which we communicate our ideas to fellow archaeologists, scientists and members of the public. Classification is our language of observation.
2This is an enormously complex subject, and includes not just methodological challenges but matters of philosophy and metaphysics. The intention here is to introduce some basic aspects of the structure of classification (fig. 1) and highlight some of the consequences that result from different approaches and kinds of units. The most important discussion on this subject is Dunnell (1971), as it is the singular review of the form that classification has actually taken in archaeology, which he compared to that used in other sciences. A useful companion for discussion is Adams and Adams (1991), who looked at its practical application in more detail. The study of Dunnell will largely form the basis for the outline of classification described below. Many aspects of this structure may seem unfamiliar, even to experienced ceramic specialists, as most have been content to concern themselves only with practical problems, rather than how or why classification works. We hope that this brief outline will act as a primer and encourage further exploration of these issues in order to develop a better understanding of this important subject.
Figure 1. Structure of archaeological classification and the relationship between 1) empirical units (i.e. real things) and theoretical units (i.e. concepts), and 2) the processes of grouping, typology, and classification, and the kind of units that result from these

Based on Dunnell 1971; with modifications after: ADAMS, Adams 1991; O’Brien; Lyman 2000; Rice 1987; Bortolini 2017.
The Structure of Classification
Terminology
3All fields of scientific practice carry the weight of specialised terminology. Sadly, archaeology maintains a terrible history of using quite useful terms in all sorts of inconsistent ways and usually without definition. This is typified by the use of “typology” and “classification” as meaning the same thing, when they do not. As we shall see below, typology is a special kind of classification. The term “type” has also been abused, having been applied to units of different kinds (how they are made), and different sorts (what they comprise), ranging from form to decoration to technology. The kind of unit cannot therefore be determined by its label. Moreover, different regional or period studies in archaeology also tend to employ different terms, as used by the distinct traditions of archaeology in each country. Thus, it is important to be aware of these differences.
Different Kinds of Units
4In differentiating between sort of unit (what it is made of), the student of pottery must be aware of what it comprises (parts of pots, whole pots, etc.) before comparing it to other units, as well as what kind of unit (how it was made) it is. Two of the most important aspects to consider are “rank” and “definition”.
5Rank is important for deciding if units are of the same sort or not. This largely depends on their scale, for example if they are based on part of an object (e.g. handles, rim, etc.), a whole object (e.g. shape, volume, etc.), or an assemblage, with each representing different levels of inclusiveness. This can affect such things as temporal or spatial distribution patterns (fig. 2) or the degree to which they measure variation (fig. 3).
Figure 2. Example of the different distributional patterns in time and space that result from using different criteria in the construction of archaeological units, based on a typical pattern produced by stylistic and functional criteria

After Allen 1996, fig. 1.
Figure 3. The different distributional and relational patterns typically produced by type units and class units in time and space

Types, based on “extensional” definitions, result in patterns where time is discontinuous and change between units is transformational, whilst classes, based on “intensional” definitions, results in patterns where time is continuous and change measures variation (representing the differences between essentialist (change as transformation) and materialist (change as constant variation) perspectives respectively (after O’Brien, Lyman 2000).
6The procedure for definition (how it was made) affects the kind of unit. There are two sorts of definition, “extensional” and “intensional”. Extensional definition derives criteria for grouping things together from the objects themselves. This results in units that are quite internally homogenous or similar. However, the differences between groupings are not necessarily based on consistent criteria, as they were selected for defining internal similarity rather than external difference. This can be described as an inductive or bottom-up approach. Intensional definition is based instead on specific criteria chosen by the archaeologist to address a particular analytical question. As the units are based only on these criteria, they can be quite internally heterogeneous or varied, as the definition ignores all other variation. This can be thought of as a deductive or top-down approach.
Different Kinds of Classification
7These differences provide an important basis for distinguishing further structural differences within classification, to which we apply the terms “typological classification” and “classification”.
8Typological classification derives from the initial procedure of sorting things into groups by the selection of multiple related criteria (e.g. shape, decoration, attachments, etc.). These provide the basis for the group. Typological classification therefore comprises a particular kind of classification, concerned with—and derived from—these groupings and the classes that result from them. These kinds of classes, called types, are extensional units, as described above. Because the process is largely intuitive, with different groups often based on different criteria, definitions for types often take the form of written descriptions. Typological classification therefore derives from the field of empirical things: pottery sherds and vessels. As a consequence, the purpose of typology is closely related to the sorting of pottery from excavation or survey, and the production of catalogues. Its empirical basis means it has limitations in the field of analysis and explanation.
9Classification, unlike typology, does not deal with empirical things. Here, we must imagine that our sherds and vessels have been packed away. Thus, emphasis is on analysis rather than sorting. Classification is concerned exclusively with concepts or theoretical units, called classes. Their criteria are defined by the archaeologist to address a specific problem. If a type is supposed to mimic something real, a class is instead a measuring device, like a centimetre or gram.
10Much of the structure of classification and its output is defined by these two distinct (but related) fields: empirical and theoretical. Groupings are empirical units, comprising much of the initial or in-field ordering of archaeological data. Here a unit comprises something real. As we address other concerns, emphasis will shift to theoretical units, where the archaeologist has left the pottery itself behind and begins the critical work of the analyst. Here types, initially based on groupings, can often transition from empirical to theoretical units, or are replaced by more suitable classes. This procedure can be traced in fig. 1, moving from the top-right to bottom-left corner.
Pragmatic and Analytical Approaches
11The above structure is embodied in all forms of ceramic analysis, with the differences between 1) levels and rank, 2) extensional and intensional, 3) grouping, typology and classification, and 4) empirical and theoretical, having tangible effects on everything from the comparison of medieval ceramic industries, the reconstruction of style complexes, to chaîne opératoire for pottery production.
12These differences should not be thought of as mutually exclusive. Selecting between them is a matter of purpose. If you are in the field, faced with sorting pottery resulting from an excavation, or museum, confronted with an unsorted collection, you are likely to employ extensional grouping methods to create units. Faced with a specific research question, you are more likely to employ well-designed intensional classes. Types, for example, are remarkably useful for dividing up time and space, and have traditionally provided the basis for constructing chronologies and delineating spatial divisions, such as archaeological cultures, exchange patterns or ceramic industries. Intensional units are better for defining variation (fig. 3). As a consequence, types are good for defining limits or borders between distributions, whereas classes are good for studying transmission and innovation. Such units are complimentary, serving different problems. No particular approach, kind or scale of unit, is necessarily better or worse in their own right. The tasks with which we are faced and the use to which we will put them is what determines which is right and which is wrong.
Auteur
PhD, Deutsches Archäologisches Institut, Berlin, Germany.

Le texte seul est utilisable sous licence Creative Commons - Attribution - Partage dans les Mêmes Conditions 4.0 International - CC BY-SA 4.0. Les autres éléments (illustrations, fichiers annexes importés) sont « Tous droits réservés », sauf mention contraire.
Rural-Urban Dynamics in the East African Mountains
Sylvain Racaud, Bob R. Nakileza, François Bart et al. (dir.)
2016
Kenya’s Past as Prologue
Voters, Violence and the 2013 General Election
Christian Thibon, Marie-Aude Fouéré, Mildred Ndeda et al. (dir.)
2014
Music and Dance in Eastern Africa
Current Research in Humanities and Social Sciences
Kahithe Kiiru et Maina wa Mũtonya (dir.)
2018
Where Women Are
Gender & The 2017 Kenyan Elections
Nanjala Nyabola et Marie-Emmanuelle Pommerolle (dir.)
2018
Le Kenya en marche, 2000-2020
Marie-Aude Fouéré, Marie-Emmanuelle Pommerolle et Christian Thibon (dir.)
2020
Kenya in Motion 2000-2020
Marie-Aude Fouéré, Marie-Emmanuelle Pommerolle et Christian Thibon (dir.)
2021