Eric Hobsbawm was perhaps the most productive generator of historiographical theses of his generation. In a review of the posthumous collection of Hobwbawm's articles, Richard J. Evans suggested that one reason for his global appeal was the extraordinarily 'enormous, astonishing fertility of his historical imagination', which resulted in 'a whole shedload' of concepts: the 'General Crisis of the 17th Century', the 'dual revolution' (the French and Industrial revolutions, the formative events of modern times), the 'invention of tradition', 'primitive rebels', 'social banditry', the 'long 19th century' (1789-1914), the 'short 20th century' (1914-1989), etc. Evans' character assessment of Hobsbawm passes well as a characterization of an ideal historian: 'ability to see the big picture and devise a framing concept to sort out the diverse and unruly detail of history' (Evans 2013). At least some of these concepts also amount to the theses of entire books. The Age of Extremes: The Short Twentieth Century 1914-1991 (Hobsbawm 1995) springs to mind above all.
Another example of a historical thesis can be taken from E. P. Thompson's The Making of the English Working Class (1991). Thompson's book is a classic in labor and social history. What is the view for which Thompson argues in this monumental text of 958 pages? The main thesis of the book is that the English working class was born between 1780 and 1832, and further, that its birth was an active process, that is, that the working class was not made but that it made itself. As Thompson eloquently expresses this view at the beginning of the book: 'The working class did not rise like the sun at an appointed time. It was present at its own making' (Thompson 1991, 8). Thompson's book has great depth and contains a wealth of information on the period, some of which I will introduce and discuss below. But it is this thesis above that provides the topic and angle for our analysis, as none of the details is central in the book. Alternatively expressed, all the details are subservient to the main thesis, which thus synthesizes the book and its informational content.
Christopher Clark's book on the origins of the First World War The Sleepwalkers: How Europe Went to War in 1914 is another book of history that will be studied in depth. This is a recent book that hit the bookshelves in good time for the 100th anniversary of the Great War. What is the central view of this book? It is already included in the title, as often is the case in good historiography. The claim is thus that European decision-makers were like sleepwalkers who moved step by step towards a goal without being fully aware of or without fully understanding what the end station of this process would be.
It would be possible to go on mentioning examples of equivalent historical theses almost endlessly, but these are enough to give concreteness to this task. These represent first-class scholarship in historiography and that is the main reason for their inclusion here, but I could have chosen innumerable others. Like so much of Hobsbawm's work, Thompson's volume has also acquired the status of a classic. R. J. W. Evans (2014) calls Clark's volume in turn 'the most consistently subtle, perspicacious, and thought-provoking', referring to the wave of publications on the event of the 100th anniversary of the Great War. In what follows, I analyze how the central theses can be identified in these two books and what form they take.10
The narrativist suggestion is thus that one understands a literary historical thesis only in the situation in which one knows all the subclauses that define it or are somehow related to it. Ankersmit writes that 'none of the statements which constitute the text is ... irrelevant to the text's presentation of the past' (1995b, 225). In other words, it would be possible to understand the thesis only if one knew the entire content of the book, because it is assumed that the book forms an integrated whole. Alternatively, if a book is irreducibly one, then the understanding the book requires knowing what this one is, and this in turn requires knowing all the components that define the one.
Holism thus presents a very strict criterion for both the identity of narratives and for the understanding of them. Books contain thousands of sentences and they all constitute what a narrative is, that is, its identity. An omission of even one meaning-defining sentence changes the object, that is, meaning, to something else. This kind of semantic holism11 has many unintuitive consequences, such as that very few, if any, would understand the main thesis of Thompson's The Making of the English Working Class. How likely is it that even the best expert in social history is able to memorize or have some other kind of mental access to all the sentences on 958 pages in Thompson's volume? Even if one had read the book from cover to cover, it is hardly realistic to expect one to keep all the details in the mind.
If one now admits that it is not necessary to know all the sentences in order to understand Thompson's main thesis, then one either admits that all the sentences do not define the thesis or that we can draw a distinction between understanding and identity or meaning so that understanding a thesis does not require knowing its meaning. The first option would amount to abandoning holism and the second remains unclear without some further theory on the relationship between understanding and meaning.
It is therefore important to ask, what defines the meaning or identity of a specific historical thesis. Or simply, what constitutes a specific historical thesis? The assumption about the holistic nature of historical theses is best assessed by considering it on the level of a concrete historiographical example. Is the claim that Earl Fitzwilliam was removed from his Lord-Lieutenancy for protesting at the massacre of Peterloo in 1818 (Thompson 1991, 751), for example, a constitutive part of the meaning of Thompson's main historical thesis that the English working class was born more or less spontaneously between the years 1780-1832? Or is the claim that the publication The Political House that Jack Built sold 100,000 copies (Thompson 1991, 743) a part of its identity? Or that the London Correspondence Society called a great demonstration in Copenhagen Fields in Islington on 26 October 1795 (Thompson 1991, 157)? All these details are part of the 'representation' or 'narrative' if it is conceived of holistically. Is it the case that, if a person does not know one of these sentences, one does not understand what Thompson claims in his book? Let us assume that two extremely intelligent and competent readers know two respective sets of sentences that are almost complete, but that differ with respect to one sentence (that is missing in one's set or is replaced by another in the set). Should we in this situation assume that the theses they understand are different? If holism is correct, we should, no matter how insignificant the difference or the sentence is.
It is unreasonable to assume that all these details amount to meaning- constituting parts of the historical thesis in such a way that knowing them is required for understanding Thompson's main claim. Thompson could have dropped the mention of how many copies this particular underground journal sold and still be seen to make the same case for the birth of the English working class. Needless to say, it would be possible to mention literally hundreds of other minute 'facts' found in the almost 1000 pages that do not seem central to the main historiographical thesis.
In addition to the negative case, that no one (including the author) would understand historical theses and all would in practice understand a different thesis, it is possible to make a positive case (below) that one can as a matter of fact understand what the historian is saying without knowing all the details and claims that the book contains.12 These considerations suggest that we should make a distinction between identity or meaning on the one hand and evidence on the other. The difference can be put as follows. Ideally all the material in a work of history supports the main thesis (and in this sense the narrativists are correct that the whole literary work matters), but all the details mentioned do not define the meaning of the thesis, that is, what the thesis is. Minute details about minor agents and their actions and movements provide evidentiary support for the thesis but the understanding of the thesis does not require knowing all of them. The case is rather that the more of this kind of information there is, the stronger the evidence for the thesis. The bits and pieces should naturally be appropriately connected to provide effective support.
Distinguishing between meaning and evidence means rejecting semantic holism, as the distinction establishes a border between meaning-constituting and non-constituting parts. This kind of distinction is very traditional and familiar from the history of the analytic philosophy of history. It may appear to be a risky commitment because it resembles the analytic and synthetic distinction derided by some of the finest philosophers of the twentieth-century, such as Quine (1951). It is necessary to make clear that my intention is not to re-establish the analytic-synthetic distinction.13 The main issue is that that the distinction between meaning and evidence appears useful, and more, necessary, if we are to make sense of the structure of the studies of history; this is my suggestion for the 'logic' of historiography. The reasonable approach in this context is not to get bogged down in a debate on the nature of meaning as such but to consider what the understanding of (the meaning of) a thesis practically requires. It is worth remembering that we are dealing with something that emerges through the writing of history. With this point in mind one needs to direct attention to the question of what it takes to understand a historical thesis. The rationale for history writing is obviously to communicate something about the past to contemporaries, and this places the question of understanding in a central role. In what situation can the readers of a historical book be said to understand what the book is arguing for? The key problem is not what the 'meaning' is but when that 'meaning' can be said to be understood. And the answer is that it is understood when a sufficient number of appropriately related beliefs or claims, as well as the relations between them, are known. Consequently, what is not required for understanding (a thesis) is evidence, the role of which is to convince the reader that the thesis is tenable.
Even if there is no sharp boundary to be drawn between meaning and evidence it does not follow that there is no distinction. Although there is no sharp distinction between the bold and the non-bold either, this does not mean that all are bold or that no-one is bold. I am well aware that the philosophical problems of semantic holism and its alternatives are extremely challenging. One might say that Quine's argument that there is no principled way of drawing the analytic-synthetic distinction is still the chief reason to advocate semantic holism in some form. And although the problems with semantic holism are recognized, it is difficult to outline alternatives to it without endorsing equally problematic semantic atomism. Fodor and Lepore mention two attractive ways of reaching a middle ground. One could try to rely on the notion of meaning or belief similarity instead of identity. Or one could be a 'molecularist' who thinks that 'there are other beliefs that we must also share if we are to share the belief that P, but ... denies that all our other beliefs have to be shared' (1992, 31).14 However, they argue that in the first case, all notions of similarity must ultimately rely on the notion of identity and that molecularism does not hold without the adoption of the analytic-synthetic distinction. Now, as already stated above, the challenge that I take on in this book is not to sort out this interesting philosophical dispute but to find a credible and practically functioning way of using the notions of meaning and evidence in the context of historiography. Nevertheless, it would be possible to characterize my suggestion both in terms of molecularism and similarity about meaning content, as will be explained later.