Reasoning in Historiography
The paradigmatic change that narrativism brought about in the theory and philosophy of history and historiography, as already emphasized a number of times, is that it turned our attention to what the books of history as more or less autonomous entities are. After all, they are the main knowledge contributions of historiography. Since books are composed of a large number of sentences organized in a handful of chapters, and perhaps subchapters, it is only natural to ask what it is that binds all these elements together. There has to be something that ties together the chapters and, ultimately, the content in those chapters. What is it?
The short answer is that historiography produces texts. For this reason, the analysis of historiography as a text and the examination of textual qualities have taken centre stage in narrativist examinations. Ankersmit, much like other narrativists, claimed that no other type of philosophy has studied texts as objects of analysis in the philosophy of language, which is how he thinks they should be approached (e.g. Ankersmit et al., 2007). Further, as readers are well aware by now, narrativism has also attempted to define the theoretical entity that binds all the textual elements together more precisely. It has suggested that there are specific synthesizing structures that weave sentences and chapters into a comprehensive view of the past and that are more abstract than material sentences or texts on paper. The synthesizing structures have been given various technical names, such as 'story', 'plot', 'narrative', 'narratio', 'narrative substance', 'trope' and 'representation'.
In this chapter, attention is shifted to the structure of historical works. This means dealing with one of the key concepts of the narrativist philosophy of historiography, that is, holism, as identified earlier. Do the books of history or the views that they contain amount to indivisible wholes? Or should the internal structure of those books be understood in some other way?
It is worth repeating Ankersmit's main reason for regarding narratives or representations1 as holistic entities. In his philosophy, there is an absolutely essential distinction between higher-order entities, 'narrative' or 'representation', and lower-order entities - that is, all the 'singular' statements or 'facts' that are subordinated to the higher order. On one occasion he spells out this distinction clearly, if provocatively:
Theories of representation are, essentially, theories about how the whole of a historical text is related to the past that it is a representation of - and this is a problem that cannot be reduced to how a historical text's individual statements relate to the past ... one can quite well be (as I happen to be myself) an adherent of positivist and empiricist accounts of historical writing for what takes place in the historical text on the level of the statement while being, at the same time, and adherent of a theory of historical representation for the text as a whole. (Ankersmit 2005, xiv; my emphasis)
The argument is that if one forgets this distinction and assumes that a historical text amounts only to a set of lower-order statements, one loses the 'view' that the text as a whole articulates. Indeed, the 'view' is all those statements combined, which constitute or give birth to another kind of entity on a higher level, as it were, with emergent qualities and features of its own. The resulting narrative or representation is a new individual entity defined by all the statements that it contains. To put it another way, all the statements are analytically true, that is, true by virtue of meaning, of the narrative or representation.
Representation as a whole is related to a discussion concerning the meta-views of historiography initiated in the previous chapter, that is, to the question of what kind of a practice historiography is and what types of products it creates. The narrativists compare representations to artistic creations, initially to literary narratives, and later predominantly to visual artistic products, such as paintings. When art is the analogy for historiography, it is understandable that the outcomes of historiography are seen as holistic entities. A painting is typically what it is only as one, and not, for example, as two separated halves (when it would arguably be one or two new artistic creations) or as a thousand sliced pieces of the original. But if one were to shift the comparison to some other kind of practice, the problem would change or disappear. If historiography were viewed as a science, we would not conclude that its products are wholes, or at least not on the same grounds as when its products are compared to artistic creations. If the function of historiography were just to report 'facts', as might be the case when fitting it into a scientific framework, its product would arguably not be a whole but a list of factual statements instead. Similarly, if the role of historiography were seen as being to explain in the mode of a Hempelian covering-law scheme, its explanations would not amount to holistic entities. These are, of course, not the only options available. Towards the end of this chapter, my view on what historiography is and how its main scholarly products should be identified becomes clearer as I identify historiography as a rational and argumentative practice.2
In what follows I will first consider the central knowledge contributions of historiography from a practical angle. Can we identify something like a narrative or central thesis in works of history? What would concrete examples of this be? I will mention a number of historical examples. The virtue of these concrete analyses is that they make subsequent philosophical exploration of this topic more fruitful. I consider the kinds of entities that such central theses actually are. More precisely, I will examine whether they can be divided into smaller components or whether they are indeed wholes. What is the relation of the narrative or the 'central thesis' to the entire content of a book? Do all the possible statements that a text of history contains also define the historiographical thesis? The upshot of my view is that we need to distinguish the meaning of a historiographical thesis from the evidence for the thesis. This raises many challenging philosophical issues and also takes us to a more detailed examination of reasoning in historiography and the structure of its textual products. The structure will again be explicated with reference to concrete historical examples.