The justificatory role of epistemic values
It is important to realise that empirical equivalence in the form of the underdetermination thesis does not mean evidential equivalence, as there might well be other rational, non-empirical means to choose between alternative constructions. Nor is it the case that the historian's proposal is either undisputably justified or unjustified (cf. Cebik 1969, 54). I suggest that justification in historiography is something more subtle than such polarity entails. Next, I will outline a theory of justification in historiography, which is modeled on the justification of colligatory expressions and suggests some ways in which one could make rational choices between them.
Above we considered the construction of colligatory concepts from a logical point of view. In other words, the question was that of whether a colligatory concept can be uniquely justified with respect to historical evidence, and the conclusion was that this is not possible. Whether this tells us anything about the actual construction of colligatory concepts is another issue. More specifically, do historians actually construct colliga- tory concepts on the basis of 'historical data'? Or are they constructed in some other way?
This contrast between the logical point of view and actual construction resembles the classical philosophical distinction between the context of justification and the context of discovery. Popper famously remarked that mysteriously appearing or dreamt ideas - such as Kekule's dream of snakes biting their tails as representing the structure of benzene - may well turn out to be correct despite their odd context of origin. In other words, it is one thing to ask whether a construction is justified with regard to evidence and quite another to ask how someone arrived at such a construal.4
White, Ankersmit and the early narrativists all suggest that synthe- tizing expressions, tropes, narratives substances, colligatory concepts, etc., are imposed on empirical data. The temporal order runs from construction to application to data. And they have good reasons to say this. When one considers how colligations or other synthesizing expressions emerge, one notices that the marching order is typically from the historian and historiographical discourse to source material, and not the other way round. Michael Oakeshott observed that 'History [that is, historiography] ... begins not with a collection of isolated particles of data, nor with a universal doubt, nor with a blank and empty consciousness, but with a homogenous world of ideas . .. And the work of the historian consists in the transformation of this world ... in the pursuit of coherence' (Oakeshott 1966, 98). Above, I suggested that the concept of 'Thaw' must have been constructed against the idea of Stalin's 'freeze' because otherwise 'thawing' would not make sense. And how was the concept the of 'Cold War' born? Was the term coined and applied by a single historian on the basis of historical evidence acquired while in the archives? This is not the case. The term 'Cold War' appeared in a newspaper article by George Orwell (1945) at the end of the Second World War, and thus before the beginning of the Cold War, which is often viewed as having commenced in 1947. Walter Lipmann's book The Cold War (1947) made the term more widely known, after which it spread to general political and historiographical discourse and came to nominate the whole historical period. This shows that the choices of language and interpretation take place within a framework of the existing social situation and discourse. Further, the concept of 'Renaissance' was by no means inferred from the historical data, but constructed by Michelet (1855), after which the term has become part of parlance of historiography.
It would not be wrong to suggest that historiography resembles a priori plotting of the past in one specific sense: integrating expressions are not outcomes of empirical investigations of source material but precede their application to that material and are thus prior to empirical work. Would this mean that historiography loses its status as an empirical discipline that studies what actually happened in the past? This is a worthwhile query. The problem is that if a colligatory concept implied by the historian or inferred from historical discourse bears no relationship to historical data, historiography really begins to look like the feared random figments of imagination. On the other hand, one might note that the context of creation ('discovery') does not necessarily have bearing on the context of justification. In the remainder of this chapter I wish to argue that a priori type of plotting and empirical justification are in no way incompatible.
This is a good moment to return to Ilya Ehrenburg's novel The Thaw. Given that it is the birth context of the colligatory concept with the same name, it is clear that the concept did not emerge as a result of direct consultation of historical evidence. Its context of origin is in literature. The 'Thaw' illustrates well the a priori nature of colligatory concepts with regard to historical evidence. It really is imposed on data and used to colligate data from above. Nevertheless, it has been successfully applied in factually based historiographical discourse. One might even say that the 'Thaw' has become an integral part of the general historiographical talk concerning Khrushchev's post-Stalin period and that it is considered to be a meaningful expression on empirical grounds. What, then, can justify its application to empirical material? And, more generally, how can we decide how appropriate and fitting a chosen concept is? I propose that the five criteria and conditions below can be used in deciding the issue.