It is now time to draw some conclusions from the above discussion. The best way to do so is to first provide a definitional summary of what colligatory concepts are. Colligatory concepts: (1) organise lower-order data into higher-order wholes; (2) categorize without any necessary shared features or resemblance among sub-ordinated entities; and (3) are particular, that is, deal with phenomena restricted to a specific time and place.10
It was concluded that colligatory concepts cannot refer directly to the historical world since they do not have corresponding counterparts there. And 'counterpart' should here be understood widely, as covering objects, entities, processes, structures and tendencies. In other words, colligatory concepts do not 're-present' any given aspect of the past or refer to unique individual entities. One way to express this conclusion is to say that if we identify colligatory concepts as potential truth-bearers, they do not have truth-makers in the historical past, and cannot therefore be true or false. Further, we might state that there cannot then be a congruence or isomorphism between colligatory concepts and historical reality. On the level of colligatory concepts, anti-realism rules. In Chapter 8, I will continue the discussion of the problems of truth and correspondence in a slightly more formal way.
Before moving on, it is worth noting in passing, however, that I do not accept the idea shared by Ankersmit and Hayden White that there is a qualitative difference between singular truth-functional statements and narrative non-truth-functional ones. The full explanation for this has to wait until next chapter, but the central reason is that my analysis of what historiography fundamentally is differs from theirs. In my view, the presentations of history do not form holistic units, although I do accept that whole books cannot be reduced to a set of singular statements either. My suggestion is that there is no clear demarcation line but instead a sliding scale, according to which all works of history can be located somewhere on an axis of the subject-sidedness and object- sidedness. Actual historiographies typically contain elements from both sides and the 'objectivity' of historical knowledge depends on the specific combination of these kinds of elements. For example, colliga- tory concepts and colligation entail constructivism and anti-realism with respect to historical knowledge, and this is what they share with nominal kinds. But, as noted earlier, this does not necessarily apply to lower-level language and expressions. I am thus not arguing for a full- blooded semantic anti-realism in historiography either. Non-colligatory expressions are subject to the same semantic problems and possibilities (in terms of their reference, predicates, vagueness and truth-values) as any 'literal' and straightforward statements about the world. This is an issue that needs returning to later in the following chapter.
This takes me back to the issue addressed at the beginning of this section: do we need colligatory concepts? Or rather could we do without colligatory concepts? I already argued that colligatory expressions amount to the most interesting and useful type of historiographical language. It is possible to find agreement regarding this even among scholars who adopt an otherwise very different approach. Arthur Marwick, who understands historiography as a 'purely empirical' (2001, 4) discipline and firmly believes that it produces cumulative knowledge, nevertheless arrives at the following conclusion: '"Periodisation," the breaking up of the past into manageable epochs or periods, is simply an analytical device: the periodization that is useful for political history may well differ from that useful for economic history, and once again from the periodization that is useful for social and cultural history.' He adds that it would be 'ridiculous' to treat these 'analytical devices' as having 'some ineluctable materiality' or 'inherent reality' of their own (Marwick 2001, 9-10; 53; similarly 207). I agree. Despite (or perhaps due to) their constructive nature, historiography purified of colligatory expressions would be much poorer, and much less expressive. It would resemble a chronology of low-level observational statements, which arguably could not be said to fit our idea of proper history writing.