Postmodernism in historiography, as seen for example in the writings of Keith Jenkins and Alun Munslow, is one descendant of 1970s and 1980s narrativism (e.g. Jenkins 2003a, 7-8). It should be remarked that I am not using 'postmodernism' here as any kind of abusive term but as a description of an intellectual orientation to which these authors are explicitly committed. Hayden White has undoubtedly been its main influence, but also Ankersmit's early writings have contributed to its formation. Although the argumentation and conclusions differ to some extent, it is fair to say that these two intellectual schools, postmodernism and narrativism, share many assumptions. In what follows, I will analyze the postmodernist philosophical position. This is important in order to understand the relation in which postmodernism stands to narrativism, but also in order to situate my project with respect to the former. Keith Jenkins' key writings are my foremost guide to postmodernism in historiography.
Postmodernism in historiography accepts and endorses the fundamental distinction between lower-order and higher-order entities of knowledge, as outlined by the narrativist philosophy of historiography. That is, it commits to the distinction between 'facts' and 'narratives' and to the difference in epistemic status between them. In a preface to Jenkins' Re-thinking History, Alun Munslow writes that Jenkins takes 'his cue' from White and Ankersmit that history is first and foremost a literary narrative about the past, a literary composition of data into a narrative where the historian creates a meaning for the past (Munslow 2003, xii). Indeed the idea of there being data, which are used to create meaning is central both to narrativists and to programmatic postmodernists, which leads both to historiographical constructivism. Jenkins writes:
The historian can then begin to organise all these elements in new (and various) ways - looking to that longed for 'original thesis' ... Here the historian literally re-produces the traces of the past in a new category and this act of trans-formation - the past into history - is his/ her basic job. (2008b, 27)
Occasionally postmodernists make the same problematic commitment to historical 'facts' as the narrativists rather than to a less loaded notion of 'data', which has sometimes led to a critique of their implied positivism.1 For example, Jenkins indicates that the claims that the First World
War happened between 1914 and 1918 and that Margaret Thatcher came to power in 1979 are historical facts. Ultimately 'facts' are deemed insufficient for the construction of proper historical knowledge, because 'such facts, though important, are "true" but trite within the large issues historians consider' (Jenkins 2008b, 40). And 'large issues' refers to the historian's interpretative task and to the historian's attempt to make the past meaningful, a job which requires the historian to consider the weight, position, combination and significance of the 'facts'.
Ankersmit (1983) declared that there are no translation rules that would tell us how to translate the past or the traces of the past into a narrative, and Jenkins is certainly in agreement about this (Jenkins 2003, 5). In Jenkins' language, the narratives or theses that input meaning into the past are the historian's referents (Ankersmit would talk about 'presenteds' or 'aspects'). However, the historian's 'referent' is not out there but 'the product of their inference'. Jenkins plays down the importance of 'facts' in inferences and attributes a significant role to personal and professional interests and concerns (2008a, 66). The idea that historiographical interpretations are inferences is promising because it suggests that interpretation is explicable and rule-bound. Nevertheless, Jenkins agrees with Ankersmit that there are no logical rules that would regulate historical construction. Jenkins even hints that the presumption of, or perhaps the illusion that there exist, rules or methodologies to determine historiographical interpretations is harmful. He thinks that adherence to rules and methodology precludes choices and responsibility, and would thus ultimately be unethical (Pihlainen 2013, 244).
Jenkins makes clear, just like the early Ankersmit did, that the 'empir- ical/epistemological element' can operate on the level of singular statements but not on the level of narrative historical representations because these are of an 'aesthetic kind' (Jenkins 2008a, 69). But what is Jenkins' fundamental reason for saying this? His talk of narrative representations as aesthetic kinds and figures (Jenkins 2008, 69) would perhaps suggest that the problem has to do with the lack of isomorphism between the past and these kinds of entities. That is, he would perhaps mean that a qualitative difference makes it impossible to match them. In actuality, the onus of Jenkins' argumentation is on epistemological problems. First, Jenkins confesses to being a minimal, or 'fig-leaf' realist to use Michael Devitt's language (Devitt 1997, 23), in the sense that he accepts that there is mind-independent 'material stuff' out there, even if we may never be able to describe it accurately (Jenkins 2008a, 60). The fundamental problem of the 'empirically/epistemologically' driven 'nonradical historian' is that his or her aim to 'establish assured historical knowledge ... cannot ever be met' (Jenkins 2008a, 64), implying that the difficulty is of the epistemic practical kind. More revealingly, Jenkins writes that '[w]e shall never know what History/history "really is" - that will remain a secret like the name and the face of God'.
The problem thus is not that there is no 'what history/History really is' but that we can never know it. Jenkins describes our perspective as anthropomorphically limited (Jenkins 2008a, 60). Here emerges a crucial difference to my analysis of the situation, by which I may ultimately come to appear more radical than the 'radical historian' of the postmodernist. In my analysis, not even God could know what history 'really is', because there is 'no real' history in the sense that the past would have an inherent and given shape. I agree with Jenkins that 'the truth-full reconstruction of the past', at least on the synthetizing level, 'is ... an impossible "myth"' (Jenkins 2008a, 63). Equally, Jenkins is on the right track when he states that 'History is about something that never did happen in the way in which it comes to be represented' because 'representations' are constructions (Jenkins 2008a, 67; my emphasis). My technical reason for this conclusion is that there are no truth-makers for the integrative theses on history. But I rush to point out that this does not mean, and here I again disagree with Jenkins, that all historians' construction are equally 'arbitrary' (Jenkins 2008a, 64). Jenkins seems to follow White in that there are no epistemological and empirical grounds to choose one interpretation over others.2
Postmodernism and narrativism (specifically in its earlier formulations) infer from the correct conclusion that there are no uniquely, absolutely, correct historiographical interpretations the erroneous one that no interpretation is cognitively more justified than another. First, 'constructed' does not automatically mean 'unreal'. The semantics of 'real' needs its own treatment (in Chapter 10) since it is frequently used in argumentation both by postmodernists and realists. Second, neither does 'constructed' mean 'unjustified', as there is room for many kinds of comparative and rational evaluations.
The mistake that Jenkins makes is that he equates historiographical interpretation with 'meaning', and 'meaning' with values. This assumed, Jenkins then relies on a traditional Humean principle that it is impossible to derive value judgments from facts (or 'ought' from 'is'), and concludes that one cannot infer historical 'meaning' from historical 'facts'. The conclusion is thus that, because the past has no intrinsic value, history can be '(logically) anything you want it to be (the fact- value distinction allows this ... )' (Jenkins 2008b, 13; similarly 2003, 43). In other words, because we are free to choose our values, we are also free to choose the meanings of history (cf. Pihlainen 2003, 243). Further, Munslow's explication of 'meaning' is misleading, if meant as a general 'meaning' notion in historiography, since he seems to conflate the 'meaning of the past' with the meaning of a past text (e.g. Munslow 2007, 100-101). Whatever 'meaning' the past can take, it can arguably be a meaning of something non-textual too. Martin proposes another way of understanding 'meaning' in the context of historiography, which reduces it to a cognitive notion. For Martin, to provide an answer to what meaning or historical importance is comes down to showing a coherent and intelligible pattern, such as making a case for the importance of the defeat of the Spanish Armada (e.g. European nations concluded that the God is not on the side of Spain and that religious unity was not to be reimposed by force) and, in this way, justifying the interest in the episode (Martin 1993, 44-47).
Perhaps the most important difference between my analysis of historiography and that of the postmodernists deals with how the central theses of historical works are understood. The assimilation of 'meaning' with values is problematic not only because it puts them beyond any kind of cognitive judgment but also because it threatens to exclude them from rational evaluations altogether. Jenkins states that 'history remains inevitably a personal construct' (Jenkins 2008b, 14) and it is personal tastes that make us choose one approach over another. Jenkins rhetorically asks: 'is it not likely that in the end one chooses, say Thompson, because one just likes what Thompson does with his method?' (Jenkins 2008b, 18). Munslow writes that 'It is the function of the reader to determine for herself or himself why some views of the past are plausible, satisfactory and convincing and others are not' (Munslow 2007, 116). Now, if the whole point is to make emancipatory, material differences to and within the present, and if interpretations are subjective choices entirely, it would seem that there is no room for other evaluative criteria than those that matter to the individual, no matter how ill-informed the interpretations may appear to be for others. In other words, there would be no role for rational judgments and corrections that go beyond the individual's tastes and preferences.
Although Jenkins' propagates 'antirepresentationalism', he nevertheless succumbs to the same kinds of problems with 'representationalist' vocabulary as Ankersmit.3 Jenkins' 'antirepresentationalism' boils down to the view that 'historians' representations ... are always failed representations,' and thus 'fictive' in his parlance (Jenkins 2008a, 65; 68; similarly Jenkins 2003, 5). One wonders why he talks about 'representations' since the term connotes that one is re-presenting in historiographical language something that exists there prior to any constructions (just like it is with the term 're-construction'), as discussed earlier in the book. And this is not just a harmless linguistic entailment but a commitment that channels argumentation into a particular direction and exposes it to certain problems, ultimately even resulting in discursive incoherence. In other words, the talk of 'representations' is vulnerable to all the Cartesian epistemic and skeptical problems of getting it right and having 'access to the actualities of "stuff"' (Jenkins 2008, 60) that is there, but which remains beyond our epistemic reach. But if interpretations do not have references, then why get stuck with this kind of discourse? Why not to say, as Pihlainen expresses it, that 'the past need not be represented at all' (Pihlainen 2013, 239; my emphasis).
I now return to the question of rules and methodology. The first thing to say is that, on the sociological level, it would simply be wrong to say that historiographical discourse is arbitrary since the communities of historians certainly evaluate and control what is accepted. While it is reasonable to suggest that one cannot achieve 'closure' in interpretation, this is not a case of 'interminable openness' (Jenkins 2008a, 65). Now, Jenkins of course means 'openness' in an epistemological sense, and not in general, since power and political interests still govern and restrain historiographical discourse. Like any 'knowledge', interpretation in historiography is used to legitimate power and material interests (e.g. Jenkins 2008b, 31). This provides a principled grounding for the evaluation of different 'histories'. That is, contemporary moral considerations take priority: 'radical historian don't [sic] work on behalf of the people who lived in the past: they work for us' (Jenkins 2008a, 64). According to Jenkins, all historiographical accounts should be directed towards emancipation and liberation, or they should make emancipatory and material differences in the present (Jenkins 2008a, 71; 2008b, 81).
His view that the existence of rules is incompatible with choice and ethical responsibility is an overreaction. It would be correct only if rules are of an algorithmic kind, strictly determining the outcome. In this situation, the only 'choice' left for the individual would be the correct following of rules. But there is currently a consensus among scholars in science studies and the history of science that this kind of strong rationality does not apply anywhere in the sciences, not even in the physical natural sciences. We might say that the key lesson of several decades of historical philosophy of science, the historiography of science and sociological science studies is that a multitude of cognitive, observational, personal and various kinds of social factors have a role to play in theory decisions, and none of them alone determines the outcome. This is to say that Jenkins could have been more circumspect and expressed his position through the thesis of underdetermination of theory by data, which is in effect expressed in the following: 'clearly there are all kinds of limits controlling the knowledge claims that historians can make' and 'sources may prevent just anything at all from being said, nevertheless the same events/sources do not entail that one and only one reading has to follow' (Jenkins 2008b, 12, 15).4 Elsewhere Jenkins indeed admits that '"the past" ... is so very obviously underdetermining in relation to its endless appropriations' (although he also jumps to the conclusion that the past can be 'read at will') (Jenkins 2003, 10).
A still more important a point is that ethical responsibility itself requires some kind of rule-boundedness. Ethical choices cannot be random. Ethical responsibility arguably implies the existence of or commitment to ethical principles or maxims in some form, which instruct (but do not determine) about right and wrong behavior in this or that situation. Further, choices cannot be totally 'free' either, unless freedom refers to freedom from physical coercion. 'Freedom' is arguably not synonymous with 'arbitrary'.
Despite Jenkins' inclination towards individualistic subjectivism with regard to evaluative judgments in historiography, this position would not sit well with his other commitments. It is possible to maintain that aesthetic, moral and ethical considerations entail their own rational standards. And as just discussed above, social and ethical responsibility imply the existence of some kinds of behavior-guiding rules and principles, which entails that they are inter-subjectively applicable. It is, as Rescher writes, that 'the idea of rationality is in principle inapplicable where one is at liberty to make up one's rules as one goes along' (1988, 158). Specifically, the 'social' aspect of historiographical interpretation takes one beyond individuality, and therefore entails normative requirements. Further, it is far from clear that we should accept the assimilation of historiographical interpretation into value judgments, if one accepts that value-laden interpretations can be subsumed under rational (aesthetic, ethical or moral) considerations. A weakness in theoretical discussions on historiography is often the lack of specificity and of concrete examples. We need to ask whether historiographical theses are like values.
Thompson's thesis concerning the origins of the English working-class or Clark's 'sleepwalking' interpretation of the causes of the First World War are ultimately cognitive claims; they are suggestions regarding what the past was like, how it should be seen. To say this does not mean denying that they relate to certain political discourses, as we have seen;
indeed, argumentation takes place in a politicized context. Nevertheless, they are knowledge claims which, however, cannot be evaluated truth- functionally, as I have argued, but which may be evaluable by other cognitive criteria. Jenkins' approach closes down this option before alternative rational and cognitive criteria have even been considered, which is ironic because a cognitive approach may be closer to his position than Jenkins has noticed. Consider how he suggests that inferences in historiography are 'always arguments' (Jenkins 2008b, 67). Jenkins hits the nail on the head when he writes that 'arguments are never true of false; arguments can only be valid or invalid ... All histories [are] always neither rigorously true nor rigorously false; at best it can have to recommend it "a certain appearance in its favour"' (Jenkins 2008a, 67-68). This is indeed so, but argumentation is a rational practice and arguments can be evaluated by considering how strong is the evidence that they provide for the main thesis. With this move we are firmly in the domain of rationality and cognitive assessment.