Integrating qualitative and social science factors in archaeological modelling aims to contribute to bridging the gap between the subdiscipline of archaeological computational modelling and wider archaeology. Agent-based modelling (ABM: a modelling approach based on simulating many autonomous agents to observe the aggregate results of their behaviour) largely acts as the lens through which the overarching problem of the incorporation of social and qualitative factors into computational models is investigated. The details may therefore be unique to ABM, but wider conceptual considerations apply to all computational modelling.
In their introduction, Saqalli and Vander Linden pose the question of why there remains a distance between computational model makers and archaeologists more broadly, which is most keenly felt in the consideration of social and qualitative factors. Such elements are notoriously difficult to integrate into computational models. In their two introductory chapters, the editors and their co-authors highlight what they see as the two main causes for this. These are: diverging attitudes to, and understanding of, simplicity, and differing solutions to the flawed and fragmentary nature of archaeological data.
Archaeological data are by their very nature incomplete. Archaeology has a long history of trying to understand and account for the elements that are missing from the material record. As the editors point out, modelling offers a different way of dealing with the lacunae in our data. By simulating different possible pasts, we can explore, in a repeatable way, how our interpretations work under different circumstances.
Oversimplification is a criticism that is often levelled at computational models. To modellers, however, simplicity is a vital aspect of making a problem manageable. One of the major arguments of this volume is that when experts from the different disciplines that feed into a particular model each sacrifice some disciplinary specificity or detail, the whole of the model can become more than the sum of its parts. Experts who are not modellers may, perhaps justifiably, feel that reducing the complexity known to have existed in the past leads to a loss of precision in our reconstructions of the past. The editors retort that there is an important difference between precision and accuracy, and that by sacrificing some precision a model might achieve a more accurate representation of the past albeit with less detailed brush strokes. It is important to bear in mind that no model aims to be correct. As the editors stress, computational models are thought experiments. Furthermore, they note that the quantification required for computational modelling requires explicit discussion between experts to weigh up such qualitative factors, a valuable exercise in itself.
The fact that, in order to create computational models, there is no choice but to cooperate between disciplines is perhaps their most promising aspect. Modelling not only facilitates interdisciplinary cooperation and communication, but also enforces this. Domain experts and model creators need to communicate, cooperate and also sacrifice some of their disciplinary specificities to make the endeavour a success. Judicious sacrifices by all participants clear the way for true integration, offering the potential to improve the end result. Interdisciplinary communication is difficult, however, and fraught with misunderstandings. This is acknowledged in the volume. In fact, those chapters that discuss the conversations and (mis)communications that occurred in the design process are the strongest in furthering the rapprochement between modellers and non-modellers that the volume aims for, precisely because they shed light on the difficulties and compromises that characterise the process of generating the model.
Another important realisation is that the value of computational models lies as much in the creation process as it does in the final product. As the modelling process forces an exploration of the implications of assumptions, and demands that these are weighed explicitly against other factors feeding into the model, it is in this process that greater understanding is achieved. This point is raised in abstracto in many of the chapters, but is most fully explored in those that give the reader an insight into the process of creation, rather than only the product (Chapter 5: Le Néchet et al.; Chapter 8: Carrer et al.). Certainly, for those who are not familiar with the modelling process, these are the chapters that shed most light on the practice of model-making.
Individual chapters present different case studies, but some general groupings can be discerned. Chapters 3 (Barceló et al.), 4 (Bentley and O'Brien) and 6 (Read) showcase the use of modelling for thinking through theoretical issues in the absence of (sufficient) data. They highlight the model-as-a-thought-experiment argument. Chapters 5 (Le Néchet et al.) and 8 (Carrer et al.) stand out because they demonstrate the importance of the modelling process by explicitly discussing both the conceptual and the disciplinary processes of model creation. Chapter 7 (Roux) highlights the importance of choosing the right variables to model. The volume closes with a concluding chapter by the editors in which they reiterate the goals of bridging the gap between the modelling community and the wider archaeological community through (interdisciplinary) communication.
Given the admirable aim of bringing together modellers and archaeologists and the illuminating insights into the process that some of the chapters offer, it is a pity that the fluidity of writing was not given more attention during the editing process. Some of the text makes the reader work hard to understand the meaning and may alienate the non-modeller audience. The volume is most likely to appeal to archaeological modellers who will find interesting case studies with a somewhat wider disciplinary outlook. The aim of generating mutual understanding is, however, commendable, and those non-modellers who venture to pick up a book about modelling and persevere will be afforded valuable insights in the process and merits of striving to integrate qualitative and social science factors into archaeological models.
