Semantics of the 'real'
Someone might accuse me now of backing away from my own proposal in Chapter 4 to reject representationalism in favor of non-representation- alism. The section above employed the conceptual pair subject-object. Let me explain how this accusation is harmless or misdirected. I said that I do not have a problem with representationalism as such, as Rorty for example has. I am not a global anti-representationalist. It is not a problem for me to accept that some expression may have a clearly definable object to which it refers. The target of my criticism was the assumption that synthesizing historical theses must be seen as 're-presentations', that texts on the whole must be seen as referring to some (abstract) objects. In my view, this mystifies the central insight, which is that historical theses are performative argumentative acts. But, of course, if no parts of argumentative speech acts refer or deal with object-sided entities at all, they would not have much informational value with regard to the past.
Now there is another question that may concern the reader. Are the historical phenomena argued for 'real'? Is there 'really' something like 'sleepwalking' that preceded the Great War or a 'Thaw' after Khrushchev? I have already made clear that these should not be taken as referring expressions and thus not as entities in the past to be referred to by the historian. In this sense, they are not real but colligated synthesizing expressions. There is, however, more to be said about this. The talk of 'reality' and the claims about what is 'real' imply something more. For this reason, it is worth considering the semantics of the 'real'; to consider what is meant by reality-talk. Further, I suggested that a central tenet of the narrativist philosophy of historiography is constructivism, and subsequently agreed about the constructivist nature of historiography (but rejected two other tenets, representationalism and holism). Constructivism relates to the theme of the 'real'. The thought of a constructed entity may suggest unreality. If an entity is constructed, can be it 'real'? In what follows, I will examine briefly the meaning of constructivism as well.
It is often assumed that if something is constructed, it is not real. And in some cases, to claim that a certain X is constructed is bound to prompt accusations of moral irresponsibility and nihilism. Kukla writes that our moral sensibilities are outraged by the suggestion that the destruction of Hiroshima was constructed and not 'real' in the sense that it would exist in an independent realm. This conveys the message that constructor-independent events are 'morally weightier' than constructed ones, although it is not clear why this should be so (Kukla 2001, 49). Indeed, the argument regarding the 'unreality' of some object is often an attempt to discredit the opponent's position on moral grounds; it is to claim that the constructivist position is somehow irresponsible.
Overall, it is much more difficult to understand what the meaning of the 'real' is than may initially seem. Those who insist on the 'reality' of something typically attach some extra layer of significance or weight to an event or object that is 'real'. The danger, and too often also the outcome of reality talk, is a situation in which the meaning of the 'real' is left entirely vacuous. This is the situation that Arthur Fine aptly parodied. The default position of both those who argue for the 'really real' and those who are skeptical of the value of this locution is the acceptance of the results of science or historiography, provided that the results have emerged out of normal disciplinary deliberation and justification. What can the reality talk of the realist add to this? Fine writes:
What the realist adds on is a desk-thumping, foot-stamping shout of 'Really!' So when the realist and antirealist agree, say, that there 'really' are electrons and that they really carry a unit negative charge and they really have a small mass (of about 9.1 x 1028 grams), what the realist wants to add is the emphasis that all this is really so. 'There really are electrons, really!'(Fine 1996, 129)
Of course, much more may be implied by reality-talk than deskthumping and a shout. Fine thinks that the realist wishes to link the 'real' to truth or existence. The 'full-blown version' includes truth as a correspondence with the world and the 'surrogate' version implies truth as approximate truth, that is, as near-correspondence (Fine 1996, 129). The idea seems to be that 'real' objects have some kind of robust and tangible existence in an independent realm about which one can make true and false assertions.
However, I don't think that the existence of objects in an independent realm is an issue of contestation, although the label 'unreal' is often applied to constructed objects such as fiction. Many artifacts in our daily lives are indisputably constructed, like smart phones, for example, but they are real nevertheless. No one would want to say that his or her most recent iPhone is not real just because someone has designed and manufactured it. Indeed, Paul Boghossian's (2001) 'metaphysical social constructivism' is a claim according to which something is real but of our own construction. Further, Latour has often been rightly or wrongly seen as a pernicious (social) constructivist, but his point is not to cast doubt on the reality of objects: 'When we [actor network scholars] say that a fact is constructed, we simply mean that we account for the solid objective reality' (Latour 2005, 91). Constructed material objects would seem to be able to function as truth-makers once they are constructed. One can make, in principle, objectively true and false statements of a constructed iPhone. Finally, talk of the 'real' may also connect to various philosophical doctrines of what fundamentally exists. For the materialist, only material objects would be 'really real' and for the idealist only mental or spiritual entities; whereas the physicalist would accept all physical phenomena as a truly 'real' substratum of the world, etc.
Provided that there is no unanimously accepted or even generally approved definition of the 'real', there are a number of options from which we have to choose. The first possibility is to endorse Fine's treatment of the 'realist' and assume that appeals to the 'real' or the 'really real' are rhetorical tricks designed to discredit the opponent's argument without any further substance. The second option would be to understand the 'real' as something the existence of which is not doubted (but could be doubted) in the fashion of Latour. This typically applies to all our everyday material objects but also, for example, to the corroborated results of science. The third possibility is to think that the 'real' is simply synonymous with 'unconstructed'. The 'real' would be something like independence from human construction (but more on constructivism below).
Unfortunately, all these options have their disadvantages. To always dismiss talk of the 'real' as a rhetorical trick would probably be unfair in many cases since something potentially more substantial may be implied. Further, to tie the definition of the 'real' to a collective endorsement would have the unintuitive consequence that something that was once real may not be so any more just because people changed some of their beliefs. In this view, witches were real in the Middle Ages but not anymore. Finally, as already remarked, it does not seem correct to say that something that is dependent on our construction is not real because of that dependence. The context of everyday objects was already mentioned but one can also find examples from the domain of natural sciences. The periodic table contains over 20 elements that are not found in nature and require synthesizing by humans. The reality of artificially derived elements would dawn on any human being if he or she came into contact with them as they tend to be radioactive. It is reasonable to assume that the unconstructed is real, but, to repeat, the same applies to the constructed too.
Provided that it is not possible to find unanimous agreement regarding what 'real' claims mean, and that they are often ways of claiming epistemic authority and credibility for the objects deemed 'real', the most appropriate course of action is to pay attention to a historiography-specific problem, that is, to the standing of colligatory and synthesizing objects in this regard. Colligatory and other synthesizing expressions organize data into historically meaningful patterns, but it was argued earlier that there is no privileged organization or colligation available. James McAllister has analyzed the relation between data and patterns interestingly. Although his analysis applies to data in the physical sciences above all, it also provides a way for understanding 'real'-locutions in historiography.
McAllister understands a 'data set' as the outcome of individual observations and experiments. A 'data set' is composed of two components: a pattern and a certain level of noise. 'Noise' equals those data points that go beyond the pattern exhibited in any specific case. McAllister argues that 'any given data set can be described as the sum of any one of infinitely many distinct patterns and a corresponding incidence of noise' (McAllister 1997, 219; my emphasis). For example, data on economic activity can be seen to exhibit patterns corresponding to economic cycles of various durations. If one wishes to have a zero 'noise' level, one has to reproduce exactly the entire set of data, which is something that no theory in science (and no interpretation in historiography) does. McAllister's (1997) main point is that the notion of investigator-independent phenomena as the objective features of the world cannot be thought to provide us a pattern in data. Instead, 'the term "phenomenon" is ... a label that investigators apply to whichever patterns in data sets they wish so to designate. Thus ... which patterns count as those corresponding to phenomena is entirely a matter of stipulation by investigators' (224). McAllister writes that while no scientists try to explain all the patterns that a data set exhibits, the pattern chosen is 'not dictated by the world: it is open to investigators to decide' (1997, 227).
McAllister (2010) expands his considerations of the relationship between data and patterns to the question of what kinds of structures it is possible to find in empirical data. He notes that it is common to assume that one can divide empirical data into a relatively simple component that corresponds to structures in the world and a more complex one that does not. The former he calls 'physically significant patterns'. The division into two types of patterns is used to identify the patterns that exist in the world. McAllister, however, argues that it is impossible to physically distinguish significant patterns from insignificant ones and suggests that because of this we ought to consider the possibility that no such separation exists (2010, 809). These considerations lead him to his radical thesis that the world is 'radically polymorphous' - that is, it contains all possible structures:
Distinct empirical data constitute evidence of, and in this sense, corresponds to, distinct structures in the world. Coupled with the finding that, mathematically, empirical data show all possible patterns, this suggests that empirical data constitute evidence for the proposition that the world contains all possible structures. (McAllister 2010, 810)
Finally, McAllister notes that the fact that one pattern is picked up does not show that alternative patterns do not exist in data. He concludes by disputing any allegations of relativism: 'all structures are objectively real for all observers' (McAllister 2010, 811).
One might say that McAllister's considerations amount to a form of the radical underdetermination thesis when theories are seen to individuate a pattern in data. That is, all theories are radically underdetermined by any set of data. In other words, no amount of or quality of data alone is sufficient for inferring the uniquely correct pattern. I suggest that this is essentially the case also in historiography even if the data in historiography are typically not the outcome of observations and experiments, but source material of various kinds, including masses of text, pictures and other artifacts. This data may make many interpretations appear unreasonable, but there are no grounds to claim that historical sources contain naturally given patterns. Although I would maintain that infinite avenues for possible interpretations exist within any set of data, I would not claim that data in historiography contain all possible structures. Phenomenological narrativists might note, correctly, at this point that much of historiography is not lifted out of source materials, but that narratives and stories exist prior to the 'configuration of the field' by the historian. However, this does not make some 'narratives' more real or natural than others with regard to historical phenomena or historical reality, but only pushes the problem one step further. Similarly, McAllister notes that reliance on prior stipulations regarding a data set assumes that investigators knew the patterns at an earlier point in time (2007, 222; 2010, 814). The phenomenological narrativist could of course argue that only the existing narratives are the target of historiographical investigations, but this would arguably limit the scope of historical investigation unacceptably and yet remain unmotivated philosophically.
Pihlainen asks, 'What may legitimately be inferred from these individual facts? And the answer, again: Nothing, really' (Pihlainen 2013b, 512). Pihlainen has a point there. But, following McAllister, we can turn the answer the other way round: 'Quite a lot. Actually, too much'. McAllister's stipulation that all patterns inferred from empirical data are real is a very interesting and novel suggestion in the dispute between realists and anti-realists. The narrativist philosophers of historiography, for example, are anti-realists in that they think that none of the configurations of historical data and narratives show us the real structure of the historical world. White famously claimed that evidence does not at all constrain the choice of tropes, which in effect means that all tropes are epistemically equally unjustified. By contrast, McAllister does not thus dispute the reality of patterns but only the claim that 'the world has a unique structure and [that] there is a single true theory of the world' (McAllister 2010, 2013). McAllister's idea can be used to turn the 'problem of unreality' of all historical interpretation into the claim of their reality as long as it is remembered that there is no uniquely correct interpretation or colligation of the historical world.7 Even better, there are two available options that at first seemed to be contradictory, but can now be seen to represent two sides of the same coin: given that there is no inherent or privileged organization, a colligated or otherwise synthesized entity may be seen as either real or unreal, from the standpoint of one's favored philosophical framework, as long as the colligation is epistemically (and more generally cognitively) warranted. The possibility of finding other kinds of colligations does not make it or its alternatives unreal since they can all be seen as being warranted. Beyond the cognitive warrant the 'real' talk is little more than an attempt to attribute epistemic authority and credibility for an entity. There may of course be many other criteria for choosing between cognitively warranted colligations of historical phenomena.
This is my solution to the problem of the claims about what is 'real' and it is up to the reader to choose realistic or anti-realistic language in this matter. My suggestion is to opt for the side of 'real' talk. A consequence is that although the Holocaust, for example, as a colligated historiographical phenomenon is not carved as a given in the 'joints of the past', it has a strong cognitive warrant and epistemic authority as a consequence of which its reality should not be doubted. The reality claim, insofar as it goes beyond the cognitive warrant, says that all the atrocities of the Nazis should be considered as one, colligated under one comprehensive concept, and seen as one phenomenon from the perspective of this colligation.