It may be argued that colligatory classification is characteristic to historiography in general, that is, also in cases in which there is no obviously identifiable colligatory term or corresponding colligatory concept. It is worth quoting Goldstein at length on this. The question he ponders is whether one could see the contested conclusion and event that Norsemen reached North-America in pre-Columbian times, in 1362, as reflecting the 'natural order' of the past:
The pieces of evidence which, at the close of the inquiry, we see belonging together - the stone, artifacts found in Minnesota and in Scandinavia, all texts bearing on the interpretation of the linguistic material, the documents bearing on the Paul Knutson expedition and the dissatisfaction with King Magnus Erikson in Norway - are brought together not because they naturally belong together, that any suitably trained scholar could see that they belong together, but by the nature of the investigation as it is pursued to its proper conclusion. Should some scholar find reason to dispute the conclusions to which [Hjalmar] Holland [in Norse Discoveries and Explorations in America, 986-1362] comes, we would most likely find in his work a somewhat different ordering of evidence: presumably some of Holland's evidence would be grouped with other evidence not deemed by Holland to be relevant to his purpose, and others of it in other ways. That is, there would likely not be some natural ordering of the data to which all sides of the dispute might appeal for impartial judgment, but, rather the dispute of the scholars and the rivalry of constituted historical events would involve, as part of the very nature of the dispute, disagreement over the arrangements of the evidence. (Goldstein 1976, 59; similarly 131)
Cebik expressed colligatory organizing as follows: 'The colligation of events (and/or conditions) x, y, and z as a Q allows one to see x, y and z as one could not see them before, that is, logically prior to the colligation. Colligation adds something but not new empirical information. Rather, it adds ... a conceptual framework, a kind of discourse' (1969, 45; cf. Dray 1959, 406). The considerations above convey the message that the organizing principles of colligatory concepts are not 'object-sided,' not 'natural,' which suggests that they must 'subject-sided,' imposed by the historian. Perhaps we might say that colligatory concepts form 'nominal categories' in the sense that the concept (or associated term) is not much more than a nametag attached to objects. This would imply that historiography and colligatory concepts entail nominalism, an idea that has indeed been put forward by Ankersmit on various occasions.4
The debate between nominalism and its opposite, universalism (and essentialism, as it is sometimes seen), has a long history in philosophy and a parallel (but much younger one) in the philosophy of science. The classical nominalist commits to the view that the world is a world of individuals or particulars only, and therefore, denies the existence of the universals (and in some versions the existence of abstract objects) that the universalist thinks are needed to explain our talk of kinds of objects and properties. The universalist by contrast claims, for example, that red things are red by virtue of their being instantiations of a universal 'redness' and gold things gold due to their being instantiations of the natural kind 'gold'. According to the nominalist, the classification and use of (natural) kind terms do not require invoking any other entities beyond the individuals that fall in the classes and kind categories.5
It appears that there is nothing 'real' or 'natural' in the ontological sense in how the historian organizes historical phenomena into more general categories like the 'Renaissance' or the 'Cold War.' Is historiography thus nominalist due to its colligatory language? This is a correct conclusion insofar as 'nominalism' means nominal postulations, that is, that colligatory arrangements are not natural, given or provided by the object (the past). However, the discussion of nominalism and realism often focuses on kinds, a central aim being to decipher what concepts are natural kinds, and on what grounds. The problem with regard to colliga- tory concepts is that it is far from clear that it would even be correct to see colligatory concepts as being any kind of kind concepts? Should we say that the 'Thaw' is a category or a set whose extension covers all 'thaw'-like objects? Alternatively expressed, are the objects subsumed under the 'Thaw' kinds of the thaw?
If some of the examples of colligatory concepts are considered, a crucial difference to kind concepts emerges. First, colligatory concepts are not taxonomic, while kind concepts are. A certain individual animal is a
German Shepherd (and a dog) because it is a kind of 'German Shepherd' (and 'dog'), on the basis that it shares some, perhaps essential, features with other kinds in that category. And the same taxonomic principle applies to 'planets'. Although 'planet' is a nominal kind, and thus has no natural essence, planets are nevertheless kinds of planets. They all share the feature of traversing around the sun. German Shepherds are kinds of dogs and the Earth and Mars are kinds of planets, but a certain painting and a book are not kinds of the Renaissance. It is not possible to create taxonomic, genus-species, categories of the kinds of the 'Renaissance' in the way that taxonomies of the kinds of dogs, mammals, animals, etc. are created.6 This would make sense only in second-order cases, in which we classify colligatory concepts themselves, such as different concepts of the 'revolution': the 'Bolshevik revolution', the 'French revolution' and the 'English revolution'. In that case, different 'revolutions' may perhaps be expected share some common features and differ in some other respects. However, one has to rely on the ready-made and understood category 'revolution'.
There is also another difference to kind concepts, the importance of which cannot be exaggerated. The point is that colligatory concepts are not general concepts, but individuals in themselves, which regardless organize and subsume other individuals (events, objects, people) under them,7 as in the case of 'Renaissance' paintings, sculptures, practices, scholars, etc.8 Tucker is incorrect in suggesting that there is no difference between the use of colligatory notions and theoretical concepts (Tucker 2004, 138). The 'Renaissance' and the 'Cold War' name two unique periods in history, and thus have clear and restricted temporal and spatial references, while theoretical concepts apply to a large set of phenomena, which is perhaps even infinite in some cases. In the first case, the suggestion is that there was a period called the 'Renaissance' in history, with some beginning and end (1400-1700), within a certain geographical area (Europe), and which was manifested in various ways in cultural products, practices and thinking. As such, it was unique; there is no other 'renaissance' in history (secondary meanings such as 'neoclassicism' refer to separate unique events). This is philosophically peculiar, as general concepts are normally assumed to do such organizing.
McCullagh has nevertheless argued that colligatory concepts could be seen as general, not as particular, which indicates that all subsumed objects have some common features.9 His examples are 'revolution', which entails that some form of radical change occurred in the historical phenomena, and the 'Renaissance', 'which refers to a collection of events inspired and directed by a set of ideas and values ... of a general kind' (1978, 272). McCullagh is correct in claiming that 'revolution' seen in this way is general, implying that the 'French revolution', the 'English revolution' and the 'Bolshevik revolution' all have something in common. Yet, as discussed above, these are 'second-order' categorizations, that is, the categorizations of historian's language, once that discourse is first in place. One should indeed expect that the phenomena that all are called 'revolutions' should be somehow similar, and the natural expectation is that they all designate fundamental changes of some sort. However, while the term 'revolution' is general, each of these revolutions is specific. The essential question is whether 'revolutionary change' can be inherent in the events themselves that are colligated to form a specific whole, and revolution. Let Cebik provide an answer to this. He argued that x, y and z, which are colligated under Q, lack a common feature: 'painting, inventing, sculpting, writing, et al., in no way equal a renaissance, nor do any of the actions have a discernible feature we might term "renaissant'" (Cebik 1969, 46-47). Indeed, what would such an inherent feature of 'renaissant' or 'revolution' be in all the parts of a colligated whole? I believe that to claim that the subsumed historical phenomena all share such a common feature in the case of a specific revolution would commit one to some kind of teleological conception of history where the parts with a specific inherent feature determined and pre-figured development towards a telos. In Finnish historiography, seeing all the pre-1917 governmental events (such as Finland's autonomy within the Russian empire, its own money, its own postal system, etc.) as a preparation for independence, in this sense as sharing an 'independence feature' realized on December 6, 1917, has been aptly called 'key-hole' historiography (Jussila 2004, 15) because the past is perceived from a narrow retrospective perspective.
McCullagh is thus concerned with the use of certain general concepts, already colligated, and not with the way in which colligatory concepts are constructed and applied to historical data in the first place. If one thinks about historiographical language, it is of course true that common nouns, proper names and many other types of expressions are used. However, the issue at stake is how a historian constitutes and justifies colligations and what the relation of a colligation to the historical data is. Although the postulation of revolution implies some kind of change, and more specifically, that the events colligated amount to a revolution, it does not follow that there was some kind of 'revolution property' inherent in and shared by each event. Colligation is a synthesization by the subject-sided historian and the result of the historian's reasoning.
Ryan Shaw aptly says that there is no template that historians could use to determine whether a certain set of phenomena should be called a revolution or something else (2013, 1094), although, when the judgment is made, the historian postulates that those phenomena colligated under 'revolution' amount to some form of radical change. The 'magic' of colligation is exactly that it enables one to put together a diverse set of events under one concept.
It is instructive to pay attention to Walsh's talk about 'parts' and 'wholes' when he analyzes the relationship between colligatory concepts and the events they colligate. It follows that the events and phenomena that a given colligatory concept organizes are not instances or members of it in the conventional kind-category sense. It is more appropriate to talk about the entities (events, phenomena, objects) as constituting the 'Renaissance' in this particular interpretation. Insofar as there is the 'Renaissance' they form it. On one occasion, Walsh compares colligatory concepts to Hegel's idea of the concrete universal, understood as the thought of something as 'a unity in diversity' and a 'complex particular' (Walsh 1974, 143-144). Both are excellent expressions.
An important qualification is required before moving on. It is reasonable to say that although members of a colligatory concept category are not kinds of that colligatory concept, they need to exemplify it or its sense. There is thus this one feature or principle that has to apply to all subsumed entities in order for the colligatory category to be meaningful. This is where the hand of the historian and her valuation is felt, as the historian chooses what to illuminate. The highlighted aspect forms an invisible thread that keeps the entities of the category together, even though the feature or the category itself is in no sense 'natural'. It is possible to view all objects forming this new holistic entity as a Venn diagram or a circle drawn around objects. The objects themselves may not have anything else in common beyond the sense imposed on them, as we notice if we consider the entities subsumed under the 'Thaw'. They are very dissimilar, which explains why the notion of family resemblance did not help either. Colligations are not based on similarity and dissimilarity postulations in any obvious sense.
Colligatory concepts provide an entirely new approach in comparison to traditional theorizing on concepts and kinds in the philosophy of science. This approach, if not unique to historiography, is in any case characteristic of it. The term nominalism in the classical sense is not directly applicable in the context of colligatory concepts. The thought behind the postulation of colligatory concepts is not to deny the existence of universals. It is true that historiography deals with individuals and is therefore compatible with nominalism, and also that colligations are nominal postulations without an assumption of natural essential qualities, but the main point is not to argue for nominalism against univer- salism. With colligatory concepts we are dealing with something that does not have a direct predecessor in the debates of the philosophy of science (with the exception of Whewell).
Another qualification to be added is that even if colligatory concepts cannot be true of historical reality, this does not mean that some other statements could not. This only means that historiographical language and its most interesting part cannot be true in the sense of correspondence. Secondly, there are other conceptions of truth that might be employed instead of the correspondence notion. I will discuss this theme more specifically in Chapter 8, but two points are worth mentioning already. If one swaps the correspondence notion of truth to, say, an epistemic concept, one changes the subject, and this does not alter the point that the language of historiography and the historical reality are incongruent. Moreover, my view is that it is more fruitful to speak about epistemic authority than truth in this context since the idea of correspondence is only one way of attributing epistemic authority to a claim or a view.