Framing the Question: Aims and Background

The reliability of historical conclusions offered by him [the excavator] as the result of his work, is in the final analysis dependent on the thoroughness of his field technique and the completeness and practicality of his system of recording what he finds. Increasingly in the future, students of archaeo-logical reports will desire information on the methods employed. (William Badè 1934: 8–9)

To what degree are the results of each and every of our research projects dependent upon the perspective and methods of the modern scholar? This question is important for every re-searcher, but is especially vital for archaeologists, who often destroy the evidence, at least partially, during the course of their work. Making the choice about what to study in detail and what to discard, either wholly or by leaving it out of the analyzed material, has consequences for the results of any study. Large amounts of finds are typical of archaeological projects in the Near East or Mediterranean, for a variety of time periods, and for Medieval or later periods this holds true over most of the world. In these traditions, the selection-making process during and after excavations is an integral part of the study, which is however rarely discussed in detail in publications. Thus, the readers of archaeological reports are often left without im-portant information affecting the results.

The aim of this thesis is to describe and reconstruct the process of archaeological artifact study, from the selection of the finds to their presentation in publications, as well as the in-terpretations of the material. There is a chain of selections that is made during the study of ceramics. This chain begins during the field

work, and passes through several points before ending up with the publications. The pottery de-scription is traditionally written as a typology, often supplemented by type frequencies. The typology can be regarded as the end product of the artifact study. The selections that are made during this process affect the nature of the pot-tery assemblages. Defining the research-based differences in the pottery assemblages thus helps to evaluate the reliability of the interpre-tations that are made using the pottery material as evidence. The validity of the tools for present-ing the archaeological artifacts is essential for all subsequent research, be it historical or social.

Herein I will discuss practices of selection mak-ing, methods of presenting the pottery accord-ing to a typology, and other analytical tools used in artifact study.

Fig. 1.1 Tel Kinrot (ancient Kinneret) and its surroundings, drawn by Ronja Kratz &Stefan Münger © KRP

I will use two pottery assemblages from one site,Tel Kinrot(Fig. 1.1) in modern northern Israel, as a case study. I will examine the ways that the finds have been selected and analyzed, both during and after the excavations, using two different strategies (intuitive selection and sys-tematic sampling, see below). There have been three large scale excavation projects at the site. The first was conducted on the upper mound in the 1980’s, and was published in 1990 by Volkmar Fritz (1938–2007). The second project was conducted by Fritz in 1994–2001, on the eastern slope of the mound. The third, theKinneret Regional Project (KRP), was co-directed byStefan Münger, Juha Pakkala andJürgen Zangenberg, with excavations taking place from 2003–2008. The KRP continued the work on the eastern slope. I worked as finds registrar for the KRP from 2003–2008. This study focuses on the material from the two latter projects, as they were both carried out on the same area of the site and focused on the Early Iron Age, while the excavations in the 1980’s were focused on the Late Iron Age.

This study has a methodological interest. I first faced the difference between the pottery as-semblages of the two latter excavation projects when I was preparing the typological descrip-tions of the pottery from the excavadescrip-tions on the slope of Tel Kinrot for the final site report. I had started the classificatory work with well-preserved vessels from the excavation project directed by Fritz (1994–2001). First, I worked with drawings and made short definitions or labels for the vessel groups. Before starting a more detailed description, I also thoroughly an-alyzed the ceramic material, consisting mainly of shards found during the KRP. I found it hard to integrate the shard material from the KRP (2003–2008) with the preliminary descriptions based on drawings of well-preserved vessels. The assemblages were different both in their number and in their state of preservation. The earlier excavated assemblage included 705 registered vessels/vessel fragments (drawn and described on find cards). The majority of them can be described as well preserved. The drawn fragments were rather easy to identify and classify. In addition, a selection of shards from clean loci was kept “as a proof for the dating”

(personal communication, Münger 18.12.2012). However, these shards were not analyzed, nor was it planned that they would be published. I was able to have a look at the material from Fritz’s excavations in the German Protestant Institute for Archaeology in Jerusalem in December 2003 and April 2004. This material (excavated from 1994–2001) has since 2005 been stored by the Israel Antiquities Authority at Beth-Shemesh. It includes, by a rough esti-mation, a few thousand shards in addition to the well preserved items that were recorded on find cards. The selected shards were numbered and stored in cardboard boxes. As they were not further analyzed, they are of little value for the academic audience. The assemblage re-trieved since 2003 includes 3817 ceramic items (over 3700 vessels or vessel fragments), most of which were rim shards. All items were registered in a digital database. The quantitative as well as qualitative asymmetry between the pottery assemblages inevitably drew my attention.

The difference was related to the different retrieval strategies for the pottery finds, which determined what was kept and analyzed. These decisions, made early in the research process, have had a strong impact on the ways that the material can be analyzed, what kinds of results can be achieved, and how reliable the conclusions are.

The question

With the help of a detailed case study of the Tel Kinrot excavations, I will explore how the chosen retrieval strategies and analytic methods used in artifact studies affect the results of subsequent research. The retrieval strategy has two parts: the process of deciding what is kept and what is discarded in the first place, and then how the kept material is further analyzed. In addition, the way in which the material is presented to the audience has consequences for its future utility and any following interpretations. This holds both as regards the readers’ ability to follow the initial interpretations, as well as the usefulness of the material for subsequent research. The retrieval strategy used reflects the research tradition and goals of the scholars involved. The strategy of keeping only some material in the first place naturally has conse-quences for what kind of analyses can be executed. Thus, the type of questions that can be posed about the material are partly a result of the selected retrieval strategy. I will compare two pottery assemblages that were formed by two rather common strategies in Near Eastern archaeology: informal selection and systematic sampling. Ininformal selection a rather intui-tive selection of “important and indicaintui-tive” material is selected to represent the collected ma-terial (Orton 2000: 2). Such a strategy was used during Fritz’s excavations (1994–2001). Sys-tematic sampling is more formalized, i.e. it is an explicitly defined procedure. In the case of the Tel Kinrot excavation, this method of sampling included keeping and analyzing all rim parts, regardless of how important they were considered. In addition, an intuitive selection of other shards was kept. Systematic sampling was used in two excavation areas during the KRP (2003–2008). It is important to acknowledge that both strategies are selections. It would be a gross over-simplification to say that the informal selection process was bad and the systematic sampling was good – they served different purposes and had different impacts on the nature of the assemblages derived from them, as well as how the material could be later analyzed (Orton 2000: 2–3). Only preliminary articles have been published so far from the excavations on the slope, and this study is based on unpublished materials formed by the two differing strategies. I have made both the typology and the statistical analyses myself. Because each ancient site has a unique assemblage of artifacts, no earlier typology as such is wholly appli-cable, but requires modifications in order to faithfully describe the material. At the same time, comparability to other sites demands that the typology shares its vocabulary with and relates to the existing typologies. As there is a large amount of shards derived from the systematic retrieval, I used statistical methods in order to gain an overview of and insights into the ma-terial.

The differences in the information yielded by the artefact study relate to the different selec-tion strategies, recording methods, and methods of analysis. What kind of informaselec-tion is pro-duced, and to what extent is one able to expand the research, if one collects and records all rim parts of pottery, instead of working with selected well-preserved vessels? The amount of items is manifold, but the information gained is fragmented and sometimes difficult to inter-pret. The nature of the larger material assemblage also affects the descriptive work with pot-tery, which is commonly written as a typology. Pottery typologies in Israel-Palestine form a rather fixed literary genre (section 2.5). Typological methods have been discussed actively in

research (e.g. Klejn 1982; Adams & Adams 1991; Welinder 1991; Langin-Hooper 2011, 2013;

Gnecco & Langebaeck eds. 2014). However, the relation between the field work and selection of finds to the typological presentations has been largely ignored. It is unclear how the addi-tion of shard material affects typology building. In most excavaaddi-tion reports published in Israel, the shard material is included in frequency tables calculated for different vessel types (e.g.

Arie 2006, Panitz-Cohen 2009). Even though the whole form of the vessel cannot be directly observed from the shards, there are many features that can be readily observed from the fragments (e.g. tempering, forming technique, surface treatment). The information concern-ing such features could be integrated into the typologies, even though they are traditionally very morphological (Schaub 1996: 232).

The nature of a pottery assemblage enables certain questions, while it may preclude others.

Quantitative aspects of artifact study are largely excluded when an informal retrieval strategy is used. Certain features of the material might reveal more of the formation processes of the contexts, or be vital for distinguishing functional activity areas. Others might help in interpret-ing the mixed contexts so common at sites with a long history of settlement. It is of interest to examine whether the more systematic retrieval of the pottery finds brings about similar advantages in different contexts. In Israel, it has been customary to keep more material from so called “clean loci”, consisting of homogeneous material in the chronological sense (Sharon 1995, Davis 2004: 73–74). Similarly, artefacts from well-defined contexts are studied more extensively, while the chronologically mixed contexts are excluded from analyses (e.g. Arie 2006). This is a significant choice. In the study of well stratified contexts the increased amount of material enables subtle chronological trends to arise. The intensive study of materials from mixed contexts might yield information helpful for the interpretation of their formation (e.g.

constructional fill vs. accumulation). What are the benefits of more intensive and detailed re-cording, and what are the costs of pursuing it? A discussion of such themes should be benefi-cial for planning any excavation project. This study is first and foremost a case study of Tel Kinrot, but the results can be projected onto the wider perspectives of archaeology in Israel-Palestine.

How the question is approached: the source material

I focus on artifact studies, from the selection of the pottery finds to the analyses, publication, and interpretations achieved by this process. The nature and interpretations of archaeological material are always affected by the contexts of the finds. The research processes at different sites are difficult to compare with each other, because there are too many other factors re-lated to the nature of the site that blur the picture. Exploring the differences in the materials from one site is better able to produce insights into the differences brought about by selection strategies, recording methods, and analytical tools. Therefore, a case study is a valid method to determine the impact of the retrieval strategy and recording system. The primary material under study is from the Tel Kinrot excavations of 1994–2001 and 2003–2008, with their re-spective pottery assemblages, their documentation of the pottery, and their related publica-tions. These excavations have focused on the Early Iron Age remains on the eastern slope of

the tel. The excavated areas of the two projects are adjacent to each other, and in several cases the same architectural units have been the subject of study in these two projects. There-fore, the formation processes, both natural and anthropogenous, can be assumed to be very similar. This situation, combined with the introduction of fundamental changes in the field work concerning the documentation of loci and the retrieval strategy of the pottery from 2003 on, enables one to assign the differences in the pottery assemblages to the research processes themselves with minimal confusing factors.

I had already seen, during my first experience in field archaeology in 1999, how ceramics played a central role in the practical work of archaeologists. As an undergraduate student I witnessed the professionals engaged in lively discussion around a table with pottery finds spread over it. I chose pottery as the material for my own study because it holds a central position in archaeological reasoning in general, and the ample amount of finds allows general trends to be discerned. Pottery has traditionally been the focus of material study, for most periods from the Early Bronze Age on. The pottery typology in Israel-Palestine has long been the main dating tool for settlements. The typology is traditionally based on whole forms, with a special emphasis on rim forms and decoration. I will investigate whether a study on shard material can affect, refine, or even change the typology of the region. Constructing the typol-ogy of the Early Iron Age pottery from Tel Kinrot is only one part of the study – actually, it is the first layer of the work. It has brought about an insider’s view of the process. The typology may be considered as the empirical part of the study: a trial of typological classification with quantitative aspects included, and testing my initially intuitive typology against the quantita-tive data collected. I reflect on the process of typology writing, and combining quantitaquantita-tive analyses with it, in chapter 5.

Selective retrieval has been and still is a common way of working with finds in the Near Eastern context. Shards that are considered indicative for dating or other cultural issues are kept, while “un-indicative” shards are discarded (Ben-Shlomo & Van Beek 2014: 13). As for dating, the criterion of contemporaneity with the excavated habitation layers and the selected pot-tery items is problematic. In simplified form, the process can be as follows: shards that look earlier according to the adopted typology are consideredresidual,and shards that typologi-cally better fit groups from later periods are consideredintrusive. This way of working easily leads to circular reasoning, especially if combined with a belief in exact typological dating.

Especially between periods that are close to each other, such a method should be avoided.

Such identifications should always be supported by other indications that an item derives from earlier or later periods (e.g. being worn, or from a context with pits). The typological dating of artifacts, especially pottery, has been the most common way of dating archaeological deposits in Israel-Palestine. The pottery typologies tend to aim at a high resolution for dating, with intervals of ca. 50 years (see chapter 2.3). If selective retrieval with a focus on contemporane-ity is combined with a belief in the high resolution of dating from pottery typologies, the work easily creates assemblages that are too clean and homogeneous, which are actually much more of a scholarly construction and do not reflect the nature of the material in any unbiased

way. Especially in the case of periods that are chronologically close, the fluidity of the materi-als has to be acknowledged. However, the fluidity is not constant either, but varies from pe-riod to pepe-riod. At a multi-pepe-riod site, there are many processes that will move finds that orig-inate from an earlier layer to later layers, and finds from later layers to the earlier ones. This phenomenon should be validated by several criteria.

The excavation projects at Tel Kinrot are rooted in the culture-historical tradition, which has determined the role of their ceramic assemblages. Pottery studies have focused on chrono-logical issues and questions of cultural relations, resulting in a special interest on well pre-served vessels and features considered to be good chronological markers, such as their deco-ration. The publications have likewise focused on well preserved items considered to be good chronological and cultural markers, and providing good illustrations (e.g. Fritz & Münger 2002;

Münger et al. 2011; Münger 2013). Chronology and ethnicity have perhaps been the most discussed themes in the archaeology of Israel-Palestine since its beginning.

Aims

Intensive retrieval increases the information available to archaeologists. The potentialities created by systematic retrieval will be evaluated considering the amount of work and the costs it inevitably entails. When an excavation project is planned, the resources required for the analyses of a larger body of finds are often difficult to estimate. I hope to present some tools for the evaluation of the usefulness of the more intensive retrieval and analysis of ceramics.

When the artifact analysis is well planned, one can also avoid recording features that do not vary in the assemblage, or features that correlate with each other so strongly that only one of several measurements (e.g. clay colors of different surfaces or only one recorded color) would be needed. As the same rules cannot be applied to different sites, the importance of a pilot study on the artifacts together with trial excavations should be promoted.

This study demonstrates how the selection strategy and analytical tools used on ceramic finds affect the artifact study. The impact can be seen in the amount of the studied items, as well as in their nature. These aspects affect the possible interpretations that can be justified with the material, and the reliability of those interpretations. All of these factors relate to the re-search interests that the methods should serve. This study aims at making the interpretative process transparent, and enabling the audience to better evaluate the results. Such transpar-ency for site reports in general would be beneficial. The publications, both articles and final reports, can be regarded as “the end-product” of archaeological research. The documentation

This study demonstrates how the selection strategy and analytical tools used on ceramic finds affect the artifact study. The impact can be seen in the amount of the studied items, as well as in their nature. These aspects affect the possible interpretations that can be justified with the material, and the reliability of those interpretations. All of these factors relate to the re-search interests that the methods should serve. This study aims at making the interpretative process transparent, and enabling the audience to better evaluate the results. Such transpar-ency for site reports in general would be beneficial. The publications, both articles and final reports, can be regarded as “the end-product” of archaeological research. The documentation