It is true that books of history typically imply an underlying time-dimension, but it does not follow that such a book necessarily has to be organized chronologically. Often books of history have to contain and mention a series of important dates and events, such as Thompson's description of the Luddites, but no chronological narrative or a set of events is self-evident or dictated by the past itself. To repeat, they are argumentative choices made by the historian and it is my suggestion that the choice is made in light of the main thesis and its defense. Naturally, books of history come in many forms and with varying degrees of perfection. Some are more explicitly reasoning and argumentative than others and some are more successful than others in tying and justifying different elements together. The fact that in some cases it might be difficult to decipher the structure and the links between different parts is not a counter argument to the suggestion that the main rationale is to reason in favor of and make argumentative cases for historiographical theses. The problematic cases may be samples of bad implementation or the books in which the ratio of 'narrative' or 'reasoning' parts may be have been stretched to one extreme.
A central lesson here is that premises or grounds in historiography are more diverse than those of a standard understanding of these concepts in courses of logic and reasoning. Historiography is about argumentation in a looser sense than that of a clear set of premises and conclusions. It is about proving or giving reasons to accept certain general points or theses. It is about establishing points about the past. There can be many kinds of reasons to accept a thesis: reasoning from premises, a (narrative) description of the state of affairs, exemplification, statistics, etc.
It is time to take stock of what has been said in this book so far. I have argued that the narrativist insight that the written accounts of history contain some kind of unifying structures is correct. However, my view of these structures is crucially different in two respects. First, the central theses that synthesize a study of history do not amount to indivisible wholes. Most importantly, the structural analysis of historiography does not support the implication that all the statements in a work of history define the thesis. In addition, holism would make the standard picture of understanding and language-learning impossible, implying that understanding, sharing thoughts and language-learning require a total overlap of beliefs and meanings. I consequently rejected holism and suggested that it is necessary to distinguish the meaning of a thesis from the evidence for it. Nevertheless, this is not to say that the whole of a book of historiography does not matter. Sometimes holism is defended by throwing the ball back to the critic of holism by asking the question, why would anyone either write or read entire historical volumes, if the point argued for can be expressed in condensed form (e.g. Ankersmit 1990, 286-287; cf. Mink 1970)? The answer is that the central point is indeed normally understandable without taking all the elements of the book into account, but this does not mean that the non-meaning- constituting elements have no role to play. They have an evidential role and that is why the whole books and texts matter and, further, why the whole book is the primary cognitive unit in historiography. Many of the elements are there to support the main thesis in some way and therefore, if one wishes to know what the thesis is and what the grounds are to support it, one has to read the entire book. Second, the fundamental organizing principle of written historical accounts is not narrativity but the argumentative support provided for the central thesis. In my view, the rationale for writing books of history is to persuade readers to accept the view that is put forward in the book, not to merely report historical events in a chronological sequence. In other words, historiography is reasoning for a point or a point of view. This thought takes us once again to the question on the nature of historiography.
Rejecting holism means in effect abandoning the suggestion that historiography creates products akin to artistic artefacts. The proposal that the main rationale for historiography is to provide argumentative support for the main thesis entails that historiography is a form of rational practice. Some might see this as a suggestion that historiography is a scientific activity. In some ways this may be a correct impression, although it is too early to form a definitive judgment as there are important differences too. In any case, it is clear that the framework of rationality is different from the narrative-descriptive account that, if understood in the realistic mode, could form a ground to claim that historiography is a science of the particular, that is, a subject that provides as accurate descriptions of particular events located in time and space as possible. However, in order to acquire a more satisfactory answer to what historiography is, it is necessary to consider how historical theses can be evaluated. How to decide whether a rational-argumentative practice is successful? Traditionally, an argument is taken to be composed of several connected claims for a conclusion or conclusions that seem to allow truth-functional evaluation. Is this the case also in historiography? Can the central theses in historiography be true if we operate in the argumentative framework?