Thompson, however, contests that history is a science at all, and might therefore discount any comparison with other disciplines. ‘The attempt to designate history as a “science”,’ he argues, ‘has always been unhelpful and confusing’,20 because historical knowledge is of its nature provisional, incomplete and approximate. ‘The older, “amateurish”, notion of History as a disciplined “Humanity” was always more exact’.21 Now, to quarrel over terms would in itself be idle. But Thompson’s refusal of the title of a science to history in fact rests on a serious substantive misunderstanding of the nature of the sciences in general, which leads him to create a false extra-territoriality for history. For he goes on to assert: ‘In this sense it is true (we may agree here with Popper) that while historical knowledge must always fall short of positive proof (of the kinds appropriate to experimental science), false historical knowledge is generally subject to disproof.’22 The contrast postulated here is an imaginary one, however, which suggests a rather limited acquaintance with contemporary philosophy of science. For Popper, of course, has always maintained that conclusive verification of scientific hypotheses—in physical or any other branches of knowledge—is axiomatically impossible: the cornerstone of The Logic of Scientific Discovery was precisely his rejection of the ‘verification principle’ of logical positivism.23 In its stead, he put the falsification principle—that hypotheses were scientific only in so far as they could be falsified, by pertinent empirical testing. What Thompson thus takes to be the exceptional condition of history is, in fact, the normal status of all science. Provisionality, selectivity and falsifiability are constitutive of the nature of the scientific enterprise as such. Even lack of experimental controls is not confined to historiography: astronomy permits of no laboratory tests either. The most important recent philosophy of science, that of Lakatos, has shown the limits of even Popper’s account by demonstrating that a scientific theory can survive a number of falsifications, and must be judged by the long-run development or deterioration of its ‘research programme’, rather than by its immediate pattern of disconfirmations or failures.24 In other words, the prolonged ‘adjournment’ of the verdict on historical materialism, in Thompson’s memorable metaphor, is very close to a description of the ordinary circumstances of any scientific theory.
Thompson’s disclaimer of ‘scientific’ accuracy for history, on the other hand, proves to be a preamble to a much grander claim for it. For he goes on to write: ‘“History” must be put back upon her throne as the Queen of the humanities, even if she has sometimes proved to be rather deaf to some of her subjects (notably anthropology), and gullible towards favourite courtiers (such as econometrics). But, second, and to curb her imperialist pretensions, we should also observe that “History”, in so far as it is the most unitary and general of all human disciplines, must always be the least precise. Her knowledge will never be, in however many thousand years, anything more than approximate.’25 This is certainly a pleasing image. But is it a persuasive one? The answer must surely be no. In what sense is history ‘less precise’ than aesthetics or literary criticism? It is obvious enough that, if we wish to keep these terms, it is far more so. Why should history be incapable of ‘anything more than approximate’ knowledge? Do we suppose that the date of the October Revolution is subject to alteration in the next century? Exact and positive knowledge has never been beyond the powers of history: its vocation, as with its sister disciplines, is to extend it—although the process, as Lenin noted, will always be asymptotic to its object. Any real scrutiny of Thompson’s construction undoes it.
A central question remains, however. What defines the content of history’s ‘unitary and general’ supremacy over all other human disciplines? We arrive here at the final issue of Thompson’s opening discourse on method: what is the specific object of historical inquiry? The problem constitutes the classical conundrum of all theories of history. None has proved so intractable to generations of debate by historians and philosophers. Thompson’s initial answer to it is surprisingly simple. He equates history with the past. ‘ “Historical” is a generic definition: it defines very generally a common property of its object—appertaining to the past and not to the present or future’.26 At the same time, he contends that ‘the human past is not an aggregation of discrete histories but a unitary sum of human behaviour’.27 The logic of these propositions seems to be that history is the record of everything that has happened—a notoriously vacant conclusion to which virtually every previous thinker on the subject has given a fin de non recevoir. Carr’s criticism of it is famous.28 In fact, Thompson’s slippage towards it is an unpremeditated movement of thought, not his due and deliberate answer to the question—although it is not without significance for another theme of The Poverty of Theory, as we shall see. When he consciously addresses the problem in a later section, in response to the very sharp formulation of it by Althusser, he concedes that ‘if I get up from my desk (as I will do shortly), to take the darned dog for a walk, this is scarcely an “historical” event. So that what makes events historical must be defined in some other way.’ But in what way? It is striking that Thompson scarcely attempts even the most cursory tour of the problem. He merely writes: ‘Even when we have defined out innumerable events as of negligible interest to historical analysis, what we must analyze remains as a process of eventuation. Indeed, it is exactly the significance of the event to this process which affords the criterion for selection.’29 In a text of two hundred pages, two lines. What do these yield us? A tautology. A historical event is one that is significant to the process of historical eventuation. How do we know whether an event has such significance or not? How do we delimit the eventuation to which it is significant? The two sentences form a single, empty circle.
The reason for Thompson’s lapse here is probably that his polemical attention was so polarized by Althusser’s solution to the problem that he failed to notice how scant was his own. Curiously, his dislike for Althusser’s language is such that he here actually misreads what it is in fact saying. For Althusser does attempt a more substantive definition of the object of history: a historical fact is one ‘which causes a mutation in the existing structural relations’.30 Thompson’s comment is indignant: ‘Process turns out to be, not historical process at all (this wretched soul has been incarnated in the wrong body) but the structural articulation of social and economic formations … The soul of process must be arrested in its flight and thrust into the marble statue of structural immobilism.’31 In his ire at the phrase ‘structural relations’, Thompson has overlooked what is the hinge of the definition he is attacking, the term ‘mutation’. Althusser’s formula puts an impeccable emphasis on change, rather than on stability as Thompson imagines it to do. This is not to say that it furnishes a satisfactory solution to the problem. On the contrary, it is undoubtedly too restrictive. Did Marx’s death, for example, cause a mutation in existing structural relations? Scarcely. Yet it remains an eminently historical fact. The actual terrain tilled by the historian lies somewhere between a confinement to structural changes and an infinity of human behaviour. It is not a matter of reproach that neither Thompson nor Althusser should have resolved one of the oldest and most obdurate puzzles in the philosophy of history. But of the two, it must be said that it is the