1 The (Gappy) Postnarrativist Philosophy of Historiography

The purpose of this section is to introduce Kuukkanen’s postnarrativist philosophy of historiography. In doing so, I’ll begin with a brief sketch of the defining innovation of this account – namely that works of historiography are best understood as arguments. After this, I’ll focus specifically on those features of Kuukkanen’s proposal that are most relevant for our purposes: (1) the relationship between historiographical theses and truth, and (2) its standards of historiographical evaluation.

The jumping-off point for postnarrativism, which Kuukkanen terms the “narrativist insight,” is the idea that a work of historiography is best characterized as a “content-synthesizing entity.”² Rather than simply reporting on particular facts of historical reality (events, persons, etc.), the role of the historiographer is an interpretive one, synthesizing these facts in order to produce broader historiographical knowledge. While postnarrativism agrees with narrativism up to this point, it breaks with its predecessor with regard to the nature of this synthesizing entity. Rather than a holistic narrative, Kuukkanen characterizes “the historical text as an informal argumentative entity,” the aim of which is “reasoning for theses of history.”³ Roughly, these arguments might be understood as being constituted by both (i) theses, the products of the historiographical synthesis, and (ii) descriptive statements about historical reality that provide evidentiary support for these theses.⁴ As the purpose of this paper is not to contest this general postnarrativist framework, I won’t say any more about it here.

Instead, this paper is concerned primarily with historiographical evaluation and truth under postnarrativism. In order to discuss this, I first want to highlight two key claims advanced by Kuukkanen. The first of these I’ll call “gappy colligation.” Roughly, the colligatory terms frequently employed in historiography are non-referring, meaning that any historiographical statements or theses containing these terms will be neither-true-nor-false. The second I’ll call “anti-truth-functionality”, the claim that we should not evaluate competing historiographical theses in terms of their truth. Let’s unpack each in turn, looking both at what exactly Kuukkanen has in mind with each claim and his motivations for making them.

(1) Gappy Colligation: “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.”⁵

As something of a preface to the actual discussion, it’s first important to understand that there is a key ambiguity in in Kuukkanen’s account with respect to the concept of truth he chooses to employ, insofar as he never settles on what he means by “truth.” On the one hand, he often specifies that his claims apply specifically to truth taken in the correspondence sense, for example in this alternate expression of what I’m calling his “gappy colligation” claim:

Higher-order historical knowledge cannot be true in the sense of correspondence.⁶

However, while he goes so far as to remark that “the correspondence theory is intuitively appealing as an expression of the meaning of truth,” he doesn’t actually commit to this as what he means by truth.⁷ However, on the other hand, not only does he not endorse some alternative conception of truth, but he indicates that he doesn’t think such conceptions would even have a claim to the word “truth,” remarking that he would then “suggest that different terminology be used.”⁸ Accordingly, going forward it is important that we don’t present a caricature of Kuukkanen or attribute a stronger position to him than he actually takes up. His claim is not strictly that colligatory expressions cannot be truth or false, but that they cannot be true or false in the correspondence sense.

With Kuukkanen’s approach to truth in mind, I now want to say a bit on what it means for colligatory expressions to be gappy, along with Kuukkanen’s motivation for thinking along these lines. At its core, this simply means that any expressions containing colligatory concepts (e.g. “The Renaissance”, “The Cold War”, etc.) cannot have truth-values. On Kuukkanen’s framework, this holds for both descriptive statements that happen to contain colligatory concepts and colligations themselves.⁹ As I will discuss in more detail below, it’s important to keep in mind that Kuukkanen is not saying here that we just aren’t in an epistemic position to assign truth-values in such cases, but that such statements fundamentally do not have them.

Let’s now consider an example each of a “higher-order” colligatory statement and a “lower-order” descriptive statement that happens to contain a colligatory expression, both taken from McNeill’s landmark work of historiography, The Rise of the West:

The twin movements of Renaissance and Reformation tore apart the fundamental fusion between Hellenic and Judaeo-Christian elements in the cultural heritage of Europe.¹⁰

In the eleventh century, at the height of the Hellenic literary renaissance, a notable naturalistic style of painting appeared, attesting to the abiding power of classic models.¹¹

The first statement is a straight-forward synthesizing historiographical thesis. It uses colligatory concepts as a means of subsuming a wealth of facts from historical reality under unified interpretive frameworks. Accordingly, per Kuukkanen, it cannot be true or false. However, something more complex is going on in the second. Clearly, among other things, this statement is making a descriptive claim about the increase in number of paintings of a particular style during a particular span of time. However, it uses a colligatory expression, “the Hellenic literary renaissance,” in doing so, which complicates things considerably. Taking McNeill at face value, he appears to be advancing at least one of the two following descriptive claims: (i) The height of the Hellenic literary renaissance was during the eleventh century; (ii) a notably naturalistic style of painting appeared during the height of the Hellenic literary renaissance. The upshot here is that, under gappy colligation, neither (i) nor (ii) can be true or false.

I now want to say a bit on Kuukkanen’s motivation for contending that statements containing colligatory concepts cannot be true or false. As indicated above, this is based on the idea that there is nothing in the past that could make them true. As Kuukkanen maintains that “colligatory concepts … do not refer to corresponding entities in historical reality,” he concludes that they cannot be true in the correspondence sense.¹² His argument for this ultimately hinges on the difficultly of specifying what the truth-makers for colligatory statements would be. On the one hand, “there does not seem to be any one thing that would make a colligation true.”¹³ However, it equally doesn’t seem like there is any combination of events or states of affairs that could do this either. Specifically, Kuukkanen maintains:

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 available.¹⁴

Here I simply want to highlight these two key elements motivating Kuukkanen’s commitment to gappy colligation: (i) The truth of colligatory statements requires they be entailed from a set of facts about historical reality, and (ii) it doesn’t seem like there will ever be any such set of facts. Criticism of this argument will be reserved for section four, when I return to it in the context of the disadvantages associated with gappy colligation.

Finally, it is important to note here that this argument is entirely incidental to the postnarrativist framework. That is, there is nothing about understanding historiography as a “practice that produces arguments in favor of theses about the past”¹⁵ that requires that these theses be gappy. As Kuukkanen remarks:

Argumentation theory teaches us that arguments have premises, which take, and in this case ought to take, one to a conclusion that is typically a statement expressing a certain state of affairs. That is, the components of a statement can in principle be understood as having truth-value in the historical world.¹⁶

This point will be crucial moving forward. As I will discuss in section four, we might rescind our commitment to gappy colligation, and thereby avoid the problems it invites, all while remaining within the postnarrativist framework.

Now I’d like to turn to the second key claim we find in Kuukkanen with respect to evaluation and truth.

(2) Anti-Truth-Functionality: “Synthesizing historical theses should normally not be evaluated truth-functionally.”¹⁷

Put simply, insofar as historical theses contain synthesizing elements like colligation, they shouldn’t be evaluated in terms of whether they are true or false. Given that (1) is our starting point, the motivation for this is obvious. If all these theses are neither-true-nor-false as the result of containing colligatory terms, they of course cannot be evaluated along truth-functional lines. As Kuukkanen puts it:

The appeal to truth is out of the question, provided that the meaning of “truth” is correspondence. We have to find another principled explanation for why some synthesizing historical expressions are to be prioritized over others.¹⁸

Here I won’t contest that (2) is a natural consequence of (1). Critically, the purpose of this paper is not to argue against (2). As I discuss in section four, I think this position is broadly correct. Instead, in the following two sections I will argue that Kuukkanen is fundamentally mistaken in committing to (1). Not only does it seem to be in tension with the very notion of historiographical knowledge itself, but it raises a number of concerns for the justification-based evaluative framework he proposes to replace truth-functionality.

2 Truth and Knowledge

In the next section, I will get to the heart of my critique of Kuukkanen’s account, detailing a problem associated with accepting (1) that derives from specifics of his justificatory account of historiographical evaluation. However, first I want to quickly point out a more general problem for postnarrativism associated with (1). This problem is remarkably simple: Knowledge is factive. That is, knowledge entails truth (in the “correspondence sense”). Proposition p constitutes knowledge only if p is true. Thus, to the extent that statements containing colligatory concepts cannot be true, they cannot constitute knowledge. This poses a fundamental problem for Kuukkanen, who takes the practice of historiography to be fundamentally knowledge-producing.

It is of course widely accepted that historiography, in some sense, produces knowledge about the past. This certainly holds for Kuukkanen, who takes this as a given throughout Postnarrativist Philosophy of Historiography. Although I’ve quoted the following sentence above, it bears repeating simply in virtue of its ability to capture the fundamental problem this creates for Kuukkanen’s postnarrativism:

Higher-order historical knowledge cannot be true in the sense of correspondence.¹⁹

In this single sentence, Kuukkanen (by all accounts inadvertently) highlights the inherent tension in his account due to his twin commitments to (1) and to the corollary of historical knowledge: that knowledge is factive. The view that knowledge is factive is all-but-universally accepted in epistemology. The most notable partial exception to factivity comes by way of Hazlett, but even he doesn’t actually argue against factivity for the epistemological concept of knowledge.²⁰ In short, as put by Davidson, “everyone agrees that what is known must be true.”²¹ Furthermore, it is worth noting here that the conception of truth being employed here is more or less what Kuukkanen calls the “correspondence sense,” the intuitive notion on which truth is some sort of correspondence with objective reality. Accordingly, in Kuukkanen’s parlance, we might express the factivity of knowledge as what is known must be true in the sense of correspondence.

One potential objection here is that perhaps Kuukkanen isn’t talking about knowledge in the normal epistemological sense, on which it is factive. Such an objection, however, would seem to ignore the actual language Kuukkanen uses in the discussion of historical knowledge. Consider the following:

The fundamental problem with narrativism is that it cannot provide an epistemologically or otherwise cognitively meaningful evaluative framework for [synthetic] higher-order historical knowledge.²²

Throughout Postnarrativist Philosophy of Historiography, Kuukkanen talks about historical knowledge in terms of cognitive evaluation, epistemic values, and, of course, justification. Elsewhere, he even talks about historiographical arguments as giving us “a reason to believe something about the past.”²³ None of this makes any sense if by “knowledge” he doesn’t mean the thing that epistemology has in mind. Moreover, this would then raise the question of what he means with “knowledge.” Certainly, he couldn’t mean practical knowledge (knowledge-how) or “knowledge” in some non-literal sense.²⁴ Given that he doesn’t lay out any complex, idiosyncratic account of knowledge, like he does for justification, the only way we might take meaning from his uses of “knowledge” is to assume that he means the same thing epistemologists do with that word. That is, knowledge is a thing that entails truth in the correspondence sense.

In short, the factivity of knowledge poses a fundamental problem for Kuukkanen’s account. Insofar as he assents to both (1) and the idea of colligatory historiographical knowledge, his account will be inconsistent. In section four, I’ll argue that the solution is simply to reject (1) and instead maintain that it is possible to assign all historiographical theses truth-values, regardless of how difficult that might be in practice. Before this, however, I first want to dig into the problem (1) causes for Kuukkanen’s specific theory of justification.

3 Truth and Justification

In this section, I discuss the second problem for Kuukkanen’s account that emerges from his commitment to (1). Put simply, the evaluative framework he proposes, on which historical theses are evaluated according to their degree of justification, is exceedingly difficult to make sense of under gappy colligation. In order to illustrate this, I’ll begin by saying a bit on the relationship between truth and justification, understood in the standard epistemological sense. The purpose of this is to illustrate the extent of the challenge posed by developing a theory of justification divorced from truth. After this, I will detail the problems gappy colligation causes for Kuukkanen’s theory of historiographical justification.

As discussed in the first section, a core component of Kuukkanen’s account is the rejection of truth-functional evaluation for historiographical claims, i.e. (2). In the place of truth-functionality, Kuukkanen proposes a justificatory theory of historiographical evaluation: Rather than evaluate historiographical theses on the extent to which they are true, we should instead evaluate historiographical theses on the extent to which they are justified. Before detailing Kuukkanen’s specific conception of justification, we might pause here and note that, especially from an epistemological perspective, this proposal appears exceedingly odd. On the one hand, Kuukkanen commits to gappy colligation and maintains that many historiographical claims are neither-true-nor-false. On the other hand, the epistemological concept of justification is so wrapped up with the concept of truth, so embedded within a framework devoid of truth-value gaps, it’s not at all clear what it even means for a neither-true-nor-false claim to be justified.

To stress this link between justification and truth a bit more, there are two points we might consider. The first is that in the standard epistemological picture, justification is truth-conducive. That is, part of what it means for beliefs to be justified is that justified beliefs are ceteris paribus more likely to be true than false beliefs. As put by BonJour:

One crucial part of the task of an adequate epistemological theory is to show that there is an appropriate connection between its proposed account of epistemic justification and the cognitive goal of truth. That is, it must somehow be shown that justification as conceived by the theory is truth-conducive, that one who seeks justified beliefs is at least likely to find true ones.²⁵

Put in a slightly different way, the aim of justification is truth:

Does epistemic justification aim at truth? The vast majority of epistemologists instinctively answer “Yes”; it’s the textbook response.²⁶

To be clear, this isn’t to say that justification entails truth, but only that getting at the truth is the purpose of justification. Of course, if we think about justification in such a way, it is difficult to imagine what it even means for a necessarily neither-true-nor-false claim to be justified. If justification is truth-conducive, then a justified thesis is likely to be true. Therefore, if we understand justification in the usual way, it seems incoherent to claim that necessarily neither-true-nor-false claims are justified.

The second point is that justification is usually understood in epistemology as a property of beliefs,²⁷ and Kuukkanen even indicates that he’s also thinking about justification at least somewhat along these lines.²⁸ However, as believing some proposition/claim/thesis is equivalent to believing that proposition/claim/thesis to be true, this means that, understood in this way, justifying some proposition/claim/thesis is equivalent to justifying the truth of some proposition/claim/thesis. As summarized by Lehrer and Cohen:

Of course, a person is justified in believing P only if the person is justified in believing that P is true. This connection is trivial, because to believe that P is just to believe that P is true.²⁹

Accordingly, the problem this poses for Kuukkanen is that understanding justification in the typical way means that a colligatory statement being justified entails that the truth of that colligatory statement is justified. For example,

“The twin movements of Renaissance and Reformation tore apart the fundamental fusion between Hellenic and Judaeo-Christian elements in the cultural heritage of Europe” is justified.

entails

It is true that “The twin movements of Renaissance and Reformation tore apart the fundamental fusion between Hellenic and Judaeo-Christian elements in the cultural heritage of Europe” is justified.

This is of course a problem, as Kuukkanen wants to commit to only the former sort of statements. Again, the key here is that, when justification is understood conventionally, it is not at all clear what it even means for a neither-true-nor-false claim to be justified. Kuukkanen, to his credit, acknowledges as much, specifically remarking:

It may seem that talk of justification without truth clashes with the entire tradition of epistemology because justification is typically seen to connect with truth, and the point of theories of justification is to explicate this link…. This may be so.³⁰

Because Kuukkanen takes this sort of highly heterodox stance, one might question whether there was any point to the preceding discussion about the standard epistemological concept of justification. To quell such a concern, I want to reiterate that the purpose was of course not to ding Kuukkanen’s theory of justification for being insufficiently orthodox, but rather to stress what such a theory is up against. As the epistemological concept of justification is inexorably linked to truth, Kuukkanen cannot rely on our epistemological intuitions to help fill in the meaning of justification. Moreover, there is a significant risk of back-sliding into the epistemological concept, especially in places of vague or underdeveloped writing. Further still, Kuukkanen still wants a theory of justification which explicitly handles “the evaluation of (synthetic) historical knowledge,” complete with an “epistemic dimension.”³¹ All this is to say that the task Kuukkanen sets out for himself is a remarkably daunting one.

In order to accomplish this task, Kuukkanen offers a pluralistic account of historiographical justification which is indeed rather unlike any theory of justification to be found in epistemology. In what he calls the “tripartite theory of justification of historiography,” justification is comprised of the following desiderata: exemplification, coherence, comprehensiveness, scope, and originality (which collectively constitute the “epistemic” dimension of justification); “success in argumentation” (which constitutes the “rhetorical dimension”); and “[success] in argumentative context” (which constitutes the “discursive dimension”).³² While I think it’s doubtful that “justification” is an appropriate label for this set of evaluative criteria, I don’t want to contest the appropriateness of evaluating historiography qua argument according to these desiderata. Accordingly, I have no interest in arguing against the evaluative standards set out by this theory and will in fact argue in support of something like these standards in §4. Rather, my intention with the remainder of this section is to simply show that this account is no better at handling neither-true-nor-false historiographical theses than standard justification. In short, gappy colligation poses a fundamental challenge for such an evaluative approach, and it’s not at all clear how this challenge might be met. In order to illustrate this challenge, let’s simply go through the desiderata one-by-one, with a mind for how they might be applied to claims lacking truth-values.

First there is exemplification, which is understood in the sense that the “descriptive content of a colligatory expression exemplify the historical data it subsumes.”³³ This, as Kuukkanen points out, is simply the “minimal truth-requirement” for his account.³⁴ The descriptive content of historiographical theses should exemplify the relevant data. Accordingly, this of course of no help in the evaluation of theses to the extent that they are neither-true-nor-false, and Kuukkanen seems to indicate as much when he exempts “colligatory statements” from this requirement.³⁵

Next, there is coherence, which here simply means that historiographical claims should be “maximally coherent.”³⁶ This marks Kuukkanen’s closest point of contact with an epistemological conception of justification, and it is therefore no surprise that it too is unable to help evaluate claims lacking truth-values. After all, it doesn’t appear that neither-true-nor-false claims will ever fail to cohere with any other claims. Therefore, coherence isn’t a useful concept for colligatory statements. Kuukkanen seems to indicate as much when he limits coherence to the evaluation of “descriptive content.”³⁷

Matters are a bit more complicated for comprehensiveness. Although again I think it rather clear that this concept is useless for evaluating claims that lack truth-values, in this case Kuukkanen maintains the opposite, and thus it requires a bit more scrutiny from my part. To begin with, comprehensiveness denotes the idea that, when evaluating historiographical claims, “the one that explains more of the evidence available is better.”³⁸ To illustrate this point, Kuukkanen considers two potential colligatory concepts that might be applied to Khrushchev’s time as the leader of the Soviet Union: the “Cold Snap”/ “Big Chill” and the “Thaw.”³⁹ The idea here is that the former concepts can explain less historical data than the latter one. In this manner, we might say that the Thaw is more comprehensive.

The problem with such a proposal comes into focus when we attempt to understand what a difference in comprehensiveness actually means. Kuukkanen characterizes a comparative lack of comprehensiveness for a given colligation in terms of consistency with other historical knowledge:

If a historian suggests that it would better to call Khrushchev’s period the “Cold Snap” or the “Big Chill,” the historian would be confronted with and be required to explain a vast amount of contradicting historical material.⁴⁰

This is of course a natural way to understand comprehensiveness, as something like a lack of external contradictoriness. However, not only is it hard to see much daylight between it and coherence, but this faces the same problem as coherence when it comes to the evaluation of gappy colligations. It’s not at all clear what it means for a colligatory thesis, which lacks a truth-value, to contradict anything else, be it another neither-true-nor-false thesis or a presumably true descriptive claim. Were historiographical theses only ever true or false, it would be a trivial matter to specify what it means for one to contradict historical material. Given Kuukkanen’s commitment to (1), however, it’s exceedingly difficult to make sense of the meaning of “contradictory material,” let alone why it should be obvious that “Cold Snap” contradicts anything. Accordingly, it too doesn’t seem capable in contributing to the evaluation of gappy colligations.

Next, I want to talk about evaluation along what Kuukkanen calls the “rhetorical dimension,” the desideratum that arguments should be successful. Although not presented this way, I think we might understand this as the core of postnarrativist evaluation. After all, if we take the end products of historiography to be arguments, then of course we should evaluate historiography in terms of the success of these arguments. The sort of success Kuukkanen has in mind is “a specific form of argumentative persuasion that relies on informal argumentative strategies and reasoning.”⁴¹ Immediately we might observe that, while not necessarily misguided, persuasiveness is a rather narrow take on what it means for an argument to be successful. Notably, this doesn’t seem to capture argumentative success in the sense that its conclusions follow from its premises and evidence used to support them. Perhaps this isn’t surprising, given Kuukkanen considers all the conclusions of historiographical arguments to be neither-true-nor-false, insofar as they are colligatory, and specifying what it means for such conclusions to follow from anything might be tricky. Nevertheless, at best, this desideratum seems dangerously thin in its ability to gauge argumentative success.

However, there is a greater problem in reducing the rhetorical dimension to persuasion, which is that it’s not clear what it means to be persuaded of a thesis that cannot be true or false. It is natural to think that being persuaded of some statement means being persuaded of the truth of that statement. “You’ve persuaded me of X, but I don’t think that X is true” even strikes the listener as something of a contradiction. The problem is, if we specify that we can’t mean persuasion in this sense, then it’s not at all clear what persuasion even means. This problem is compounded by the fact that Kuukkanen seems to mean persuasion in the natural truth-functional sense, maintaining that historiographical arguments provide reasons for belief:

There can be many kinds of reasons for accepting a thesis: reasoning from premises, a (narrative) description of the state of affairs, exemplification, statistics, and so on. The shared feature of all these is that they play a role in historiographical reasoning to support a thesis about the past. Alternatively expressed, they provide a reason to believe something about the past.⁴²

This is a natural way to think about arguments, as providing reasons for belief, but given Kuukkanen’s aims it is likely out of step. As discussed above, providing a reason to believe something about the past is equivalent to providing a reason to believe that something about the past is true. Therefore, no component of a historiographical argument can provide reason for believing neither-true-nor-false claims, which is to say that such claims cannot be supported by argument. This then means that these arguments will never be successful in supporting neither-true-nor-false claims, which in turn means that success of argument – even in the limited persuasiveness/reason-to-believe sense – is a useless evaluative standard to the extent that historiographical arguments are constituted by such claims.

We might say something similar about evaluation along the “discursive dimension,” the desiderata of success in argumentative context. Say we want to compare how two rival neither-true-nor-false historiographical theses are positioned in the same argumentative context. Understood naturally, to maintain that thesis 1 fits that context better seems to mean that thesis 1 is more supported by this external historiographical context, which again seems to naturally mean that the argumentative context gives us more reason to believe thesis 1. However, at this point we reach the same problem as with the “rhetorical dimension,” since having reason to believe a thesis entails having reason to believe that that thesis is true, we can never have reason to believe necessarily neither-true-nor-false theses. Moreover, to appeal to something like consistency with the argumentative context is similarly unhelpful for the same reasons discussed above. Thus, it’s not at all clear how to even make sense of this desideratum – let alone make use of it – for the evaluation of these necessarily neither-true-nor-false theses that populate Kuukkanen’s ontology.

This leaves us with two desiderata for the evaluation of the neither-true-nor-false components of historiographical arguments, scope and originality, both of which mean exactly what we might expect. While I don’t think these face problems like those of the other desiderata – that is, there doesn’t appear to be any obvious problems with their application to claims lacking truth-values – it is obvious that these are woefully insufficient for evaluating the sorts of claims that Kuukkanen has in mind. As discussed in the next section, Kuukkanen maintains that substantial, historiographically interesting claims are usually neither-true-nor-false. It shouldn’t be at all controversial that we need to evaluate these claims, like McNeill’s colligatory Renaissance/Reformation thesis, along more lines than simply their scope and originality. It is a trivial matter to imagine outlandish claims that dwarf the scope and originality of all serious historiographical theses.⁴³ If such claims are indeed neither-true-nor-false, we need evaluative standards that reflect their outlandishness.

In short, the evaluative standards Kuukkanen proposes simply cannot accommodate the sorts of neither-true-nor-false claims associated with his acceptance of gappy colligation. Not only does his account clash with the standard epistemological conception of justification, but most of the standards he cites are ill-defined and/or useless for the evaluation of claims without truth-values. The solution to this problem, which I discuss in the next section, is incredibly simple: drop gappy colligation. These ostensibly neither-true-nor-false claims cause far more trouble than they are worth, rendering otherwise helpful evaluative standards largely useless.

4 Towards a Two-Valued Postnarrativism

In this section, I want to lay the groundwork for the future development of a two-valued postnarrativist philosophy of historiography, that is, one in which colligatory theses and descriptive statements containing colligatory concepts can only either be true or false – no truth-value gaps (or gluts). In doing so, there are two related points I’ll make. First, considerations from the above discussion directly undermine Kuukkanen’s argument for gappy colligation. Second, the points Kuukkanen raises function better as narrower arguments specifically against truth-functionality. On this basis, we might reasonably conclude that a two-valued postnarrativism is preferable to the gappy one proposed by Kuukkanen.

To begin with, I want to focus further on Kuukkanen’s original claim that we cannot in principle assign truth-values to historiographical theses that contain non-referring terms. As alluded to in section one, Kuukkanen contends that it is not simply the case that we lack the epistemic standing to confidently assign truth-values in such cases, writing:

It is worth adding that this reason does not have to do with any epistemic problems in establishing that something is true…. What criteria correlate with truth and what theory of justification indicates truthfulness, does not help because the problem is principled: there is no fact of the matter and no truth-makers in the world to make a colligatory expression true even in the ideal epistemological situation.⁴⁴

However, considerations from section two allow us to reevaluate this claim in a new light. First, knowledge is factive, and historiographical theses often constitute knowledge. This alone suggests that historiographical theses are often true, and that gappy colligation is mistaken. Next, taking this a step further, we might note that there seems to be something profoundly mistaken in maintaining that simple descriptive claims containing colligatory concepts cannot be true or false. Consider the following from Cardwell:

The Cold War came to an abrupt and rather surprising end in 1991.⁴⁵

This statement seems obviously true. However, per Kuukkanen, “the Cold War” is non-referring and therefore there is nothing that could make such a statement true or false. Taking it a step further still, there doesn’t seem to be any intuitive problem with asserting the truth of more complex, synthesizing historiographical theses. Let’s say that we agree with McNeill’s thesis that the Renaissance and Reformation together precipitated a divorce between Hellenic and Judeo-Christian cultural elements, etc. We might express this by saying, “It is true that the twin movements of Renaissance and Reformation tore apart the fundamental fusion between Hellenic and Judaeo-Christian elements in the cultural heritage of Europe.” Similarly, if we disagree with McNeill’s thesis, we might instead say, “It is false that the twin movements of Renaissance and Reformation tore apart the fundamental fusion between Hellenic and Judaeo-Christian elements in the cultural heritage of Europe.” These statements are highly colligatory and synthetic. However, intuitively, it just doesn’t seem like we do anything wrong when we assign truth-values to them.

This leaves us with two conflicting considerations: gappy colligation vs. the factivity of knowledge and our intuitions about whether historiographical theses can be true or false. In order to resolve this conflict, we should now return to Kuukkanen’s original motivation for positing gappy colligation. Kuukkanen notes that there is no set of historical facts that would together entail the truth of a colligatory statement. While this may be is correct it is certainly not some unique problem for historiographical theses. The truth-maker theory famously struggles to handle similar cases, like universal truths, in which no conjunction of atomic statements will entail the truth of universal statements that quantify over them.⁴⁶ The key here is that we don’t need to commit to the truth-maker account in order to say historiographical theses are true, even in the “correspondence sense.” Not all truths have truth-makers! Given the problems such an account poses for postnarrativism, it’s not at all clear why we should insist that historiographical truths need obvious truth-makers. Taking the above considerations into account, it seems clear that committing to gappy colligation is a mistake.

The important question now is what rejecting Kuukkanen’s argument for gappy colligation would do for the broader postnarrativism framework. Immediately, of course, we would be left with a two-valued conception of colligation. In keeping with our intuitions about historiographical theses, we might say that they are either true or false. This would then avoid the problem with the factivity of knowledge and enable his theory of historiographical justification to evaluate such theses in terms of exemplification, coherence, etc. In short, this seems like a more promising route to take.

Moreover, we might note that Kuukkanen’s account still provides a compelling critique of truth-functionalism, giving us reason to retain (2). Even if we can assign truth-values to historiographical theses in principle, this doesn’t distract from Kuukkanen’s point that, practically, it often doesn’t seem we can do so. Kuukkanen’s work highlights that colligation is complex and often opaque, and firmly establishing the truth of actual colligations seems difficult even under the best circumstances. Something more obviously evaluable, like justificatory status, seems preferable here. As Kuukkanen establishes at length, his theory of justification is tailor-made to evaluate actual historiographical theses, and in this way it seems more promising than truth-functionality for the actual work of historiographical evaluation.

In short, I think a better version of postnarrativism is one that retains Kuukkanen’s evaluative framework but replaces gappy colligation with two-valued colligation. Such a theory works because all the key elements of postnarrativism are easiest to make sense of in relation to truth. This is most obvious for the epistemological concepts he employs: Knowledge entails truth, and justification is truth-conducive. Historiography should be evaluated on a cognitive basis. However, this also holds for the primary organizing concept of Kuukkanen’s theory: argument. Not only is there no reason why all the components of an argument cannot be restricted to having truth-values, but it is of course a natural way to think about arguments. After all, this is exactly what we do when we evaluate whether an argument is valid, sound, etc. As the cognitive evaluation of argument is clearly a welcoming framework for two-valued theses, we can understand that postnarrativism seems primed for a switch to two-valued colligation.

Beyond this, we might note that Kuukkanen’s postnarrativism might benefit from further development. First, I think we might question the wisdom of labelling his evaluative criteria “justification,” given how little they resemble any recognizable conception of justification. Next, and more significantly, there remains much to be said on the details of these criteria. This incompleteness is most obvious in the “rhetorical dimension” of evaluation, which, you’ll recall, currently defines the success of an argument primarily in terms of its persuasiveness. Finally, we might note that many epistemic virtues are conspicuously absent from his “epistemic dimension,” and we might question why this is. Is it really the case that we shouldn’t use safety, sensitivity, reliability, anti-luck considerations, etc. in the evaluation of historiographical arguments, or is their exclusion merely an oversight? Again, I think there is fertile ground for future development here.

In conclusion, I think it is clear that if postnarrativism is to constitute a successful philosophy of historiography, it needs to so with a two-valued conception of colligation. Kuukkanen’s commitment to gappy colligation poses significant problems, not only for his evaluative framework but for the very idea of historiographical knowledge itself. Moreover, postnarrativism as a whole seems so amenable to two-valued colligation that we might speculate its rejection is merely something of a narrativist holdover.