The correspondence theory of truth and truth-makers
The correspondence theory of truth is arguably the most venerable and oldest theory of truth. By some accounts it goes back all the way to the roots of Western Philosophy, to Aristotle and Plato. Most epistemolo- gists, such as Nicholas Rescher (who nevertheless develops a coherence theory of truth as a 'criteriological theory') (Rescher 1973, 9), for example, agree that it is also the most intuitive theory for expressing the meaning of truth. David Armstrong writes that 'it is entirely natural to think that a proposition is true or false according as it corresponds or fails to correspond to an independent reality' (1997, 128). Further, Mandelbaum thinks that the correspondence theory is presupposed by all works of history (1938, 184).
We can understand the correspondence theory of truth as saying that a proposition or a statement is true if and only if the state of affairs stated prevail; in this way a true proposition or the statement corresponds to 'facts'. When I say that the statement 'there is a cup on the table' is true, it requires that there is indeed a cup on the table. But what does 'correspondence' mean more specifically in the correspondence theory of truth? Without going into all metaphysical intricacies, 'correspondence' could in general be understood as the intuitive idea that some factual elements in the world correspond to true propositions and thus make them true. In this sense, the correspondence theory seems naturally connected to the truth-maker and truth-bearer theory. Armstrong has suggested that 'the correspondence theory tells us that, since truths require a truth-maker, there is something in the world that corresponds to a true proposition. The correspondent and the truth-makers are the same thing' (1997, 128). He also claims that 'anybody who is attracted to the correspondence theory of truth should be drawn to the truth- maker' (1997, 14).
The truth-bearer and truth-maker vocabulary is very useful for expressing what is epistemologically and metaphysically at stake in trying to establish the truth of a historiographical thesis. The general idea is that the truth of a historiographical thesis requires that there is a relation between a truth-bearer and a truth-maker so that the latter makes the former true. Further, the notion of 'truth-maker' captures the idea that the truth of something depends on how things are in an independently given reality. In other words, truth-maker T is in some sense in the world, 'a portion of reality' (Armstrong 2004, 5-6), in virtue of which 'that T' is true. The intuition in Armstrong's words is that
It seems obvious that for every true contingent proposition there must be something in the world ... which makes the proposition true. For consider any true contingent proposition and imagine that it is false. We must automatically imagine some difference in the world. (Armstrong 1973, 11)
Now it can be understood why the idea that colligatory notions correspond to the past is problematic. The reason is not that the criterion of correspondence could not be applied to the actual practice of history, as many have argued (e.g. Goldstein 1976, 41). An epistemic problem does not mean that the correspondence theory could not capture what is at stake with truth-clauses in historiography. The problem is that there does not seem to be any one thing that would make a colligation true. Colligation is an arrangement, a construction, which does not have an independently given corresponding object. This is to say that even if a colligatory expression could be regarded as a potential truth-bearer, and I think it can, it does not have a truth-maker that would make it true.
However, perhaps I am advancing too quickly here. One might say that all the individual states of affairs combined make the thesis true. That is, if all the statements that are colligated under a colligatory expression are true in virtue of their truth-makers, then the colligation, a higher-order expression, is true. All these states of affairs together would function as a collective truth-maker for the colligation. In this way, it might be possible to retain the intuition that the truth depends on something external to it, yet avoid the problems associated with the correspondence theory. However, the problem with this suggestion is that even if all the statements were true and all the states of affairs described consequently prevailed, it would not be possible to derive a historiographical thesis without some additional elements; without some subject-sided imposition that links all the descriptive statements together. To express this another way, the links, the relations between phenomena or entities and their significance, are not objectively given unless one is prepared to accept that the reality of truth-makers admits logically complex facts in the early Russellian sense.1 I suggest that the metaphysical problems associated with this view are not insignificant. To see this, it is enough to consider what kind of 'complex fact' might make Clark's 'Sleepwalking' thesis concerning the First World War true.
But could one not see the lower-order statements used by Clark as directly referring and being true in the truth-functional way? Perhaps this is possible, but it does not change things since the problem is still that the truth of lower-order statements does not guarantee and, moreover, does not enable one to infer the truth of higher-order historiographical theses. The narrativist insight is in part based on this observation. And we have seen that especially Ankersmit underscored the qualitative difference and the absence of 'translation rules' between these two levels. Assume that all the statements that describe the route to war were true. This includes the statements describing the meeting between the Serbian Prime Minister Nikola Pasic and the Austrian Foreign Minister Leopold von Berchtold in 1913, British obliviousness of the proximity of war, the 'beehive' structure of Austrian decision making, miscom- munication between the Kaiser and German officials, the idiosyncratic behavior of French ambassador to Russia, etc. This set of statements would not allow one to infer the 'Sleepwalking' thesis regarding the First World War. Only when the thesis is suggested can one come to appreciate that all the states of affair are reasonably integrated under it.
What an advocate of a truth-maker solution needs is an entailment from lower-order entities to the higher-order entity of colligation so that if lower-order entities are true they entail the truth of the higher-order entity. This does not seem to be available. There are no such inferential relations from descriptive statements to their colligation, as the discussion of the empirical determination of colligatory notions has already showed. To see this yet more clearly, let C be a colligated expression, which colligates a large number of statements describing historical events, such as p, q, w, z, etc. The simplest expression of this state of affairs is to say that C creates a world in which it is the case that p, q, w, z . .. . n. The problem is to infer C from p, q, w, z . .. . n, which is exactly the original problem of colligation, as expressed by Whewell and others: how to derive general or other integrative concepts from a set of data describing particular states of affairs? Without some subject-sided imposition this does not seem feasible. It would, of course, be possible to go the other way round, from C to p, q, w, z ... .n; from the truth of a synthesized thesis C to the truths of statements that it entails. But this does not take one anywhere because the problem is that of how to establish that C is the case.
Where does this leave us with regard to historiographical theses? They are identified as conclusions of historiographical arguments displayed in the books of history; they are something for which the historian argues. A historiographical thesis is a statement in terms of its form and, therefore, does not raise similar problems with to regard to truth-functionality as a narrative, for example. That is, a narrative is typically considered to be some kind of holistic entity that does not allow decomposition into its component parts or the determination of the referring relations of those components. A statement, at least in principle, allows this. However, it is important to pay attention to what kinds of claims historiographical theses typically make. Let us consider the thesis that Europe and its great powers went to war like sleepwalkers. It should be evident that sleepwalking is a metaphor, and hence we should not expect it to have a truth- value. Instead, it is a suggestion for how the development towards the war should be seen. Further, describing the process as sleepwalking is a colligatory expression that subsumes a large number of lower-order statements, which describe various kinds of occurrences prior to the war.
The situation is still more problematic than this. First, it is not tenable to make a clear-cut distinction between a higher-order thesis and lower- order factual statements. Colligated statements themselves may contain colligatory expressions, thus making them non-referring as well. For example, the suggestion that Austrian decision-making was organized like a beehive with unclear hierarchical relations is a metaphor. Further, the meeting between the Serbian Prime Minister and the Austrian Foreign Minister does not appear to be a neutral description of a state of affairs, but conveys a sense or meaning too. It could perhaps be subcolligated in terms of something like the 'meeting of the deaf', because communication between the two repeatedly failed despite numerous opportunities. Second, it is conceivable that another historian would draw a different conclusion from the same set of statements, provided that the two somehow ended up with or were given the same set of evidence (as that set itself is naturally the result of many selections). The narrativists emphasized the importance of ordering 'singular statements' in a consecutive or other narratively fitting order. In the case of 'sleepwalking', the statements take no specific consecutive order, except that they all support the 'sleepwalking' hypothesis. It would be better to speak of the emphases and valuations of some descriptive statements, although even then inference from a lower-order set to a historiographical thesis is impossible.
If a historiographical thesis can thus be seen as a colligatory or/and containing colligatory expressions, it seems as if it amounts to a whole or is at least 'molecular' in Fodor and Lepore's sense (see Chapter 5). However, it is very important to make some qualifications. Colligations try to reach beyond the text to the past itself, even if they do not instantiate any reference there. One can say that a text makes the case for seeing the past in some specific fashion. Further, the text does not normally amount to one colligatory notion but, rather, specifies the meaning of the thesis and evidence for it. In other words, everything mentioned in a book does not necessarily fall under any one colligation; instead, different parts and statement play different roles. If the historian rebuts, for example, a potential objection to an interpretation or rehearses the possible issues that a reader should bear in mind, this does not make them part of the colligation. The structure of the text is different from the colligatory power and the function of the specific (including colliga- tory) expressions it contains.
Colligation should not be seen as defined by its lower-order elements. While it is true to say that the latter compose the former, the talk about definitions has a connotation of something like analytic meaning, whose identity requires the presence of one particular set. This is not the case with colligatory expressions. A historian may subsume the meeting between the Serbian Prime Minister and the Austrian Foreign Minister under the sleepwalking thesis in one case and not do so in another, and yet commit to the same colligation in both cases anyway. Or let me take the more mundane example of the colligation 'all TV-programs that I like'. It is possible that those programs have nothing obvious in common except the fact that I like them. This kind of colligation does not require that some specific programs are definitionally included. If a new program is launched on TV that I specifically like, and it thus ends under category 'all TV-programs that I like', the sense of the colligation has not changed. In this case, the TV-programs that I like compose the category but do not define what 'all TV-programs that I like' means. The same is true of the 'Renaissance', which may or may not subsume some specific paintings or thinkers and still retain largely the same sense. It may be the case that many of these categories are simply i nherited from previous historians and by-and-large retain their meaning and functioning despite the fact that the borders of the subsumption are porous.