Tucker, Our Knowledge of the Past (Cambridge: CUP, 2004)
Tucker provides a welcome and robust rejection of scepticism about historical knowledge, and is moreover correct in thinking that a well-founded concept of the nature of historical knowledge is part of a civilized politics. There are obvious empirical examples that can be cited in defence of the possibility of historical knowledge, such as the successes of historians in correctly predicting the existence of evidence [Giesebrecht], and Tucker assembles a good range of examples in this respect. But he wants to go further by putting forward a genuinely philosophical argument in defence of the scientific nature of historiography, which he calls the '[representation] of past events...that generates probable knowledge of the past' .(2)
This argument is based on the similarity between history and other sciences including linguistics and evolutionary biology, all of which study information flows. In effect, Tucker wants to show that history is a science by demonstrating that it shares the characteristics of these other disciplines which are already admitted to be sciences. Naturally, this involves some consideration of the nature of science itself. Tucker opposes the old positivist conception of science which was essentially Newtonian; modern philosophy recognizes science 'as stochastic, probabilistic, with limited powers of prediction, and irreducible to another science' (211).
If Tucker is read as claiming that there is more ground in common between the human and natural sciences than has often been thought, this is unobjectionable. Both branches of science are rational, critical, sceptical, evidential, and inferential in their approach to knowledge. Perhaps, as Tucker claims, too much has been made of the difference between explanatory natural science, in which an effect is supposed to be related to its cause or causes in terms of a general law or laws, and understanding in the human sciences which focusses on the reasons why this event in particular was the outcome of previous states of affairs.
Nevertheless, it is also possible to feel at certain points that Tucker has in fact shown what makes history scientific, rather than what makes it history; that is, he has focussed on the common factors (which undeniably exist, and no doubt do need to be highlighted) between history and various other sciences, natural and human, to the neglect of the factors that enable us to pick out history as a distinctive discipline in its own right. In doing so, he also sometimes seems to unconsciously avail himself of the positivist approach that in other respects he is rightly keen to leave behind.
A key problem lies in the use made of the notion of 'information'. Despite repeatedly insisting that there are no 'ready to eat facts' (2, 12, 14, 43, 93, 138, 178), 'information' is treating as a kind of scientific constant. 'Historiographic information-bearing signals are considerably slower and composed often of words, rather than of light as in science, but these properties do not distinguish epistemically between historiography and science. Scientists understand the transmission of light; historians analyze the fidelity of textual information' (94).
The problem is that this entirely neglects the nature of the information involved; its qualitative aspect, if you like. It would be foolish to object to talking of historical evidence as containing information for historians; but it would be equally foolish to imagine that we mean the same thing by it as scientific information. When a biologist extracts DNA, or a linguist studies a sound recording, they are entitled to presuppose that their 'information' is present before them in a quantifiable form that exists entirely independently of their own subjectivity, because they are dealing with the results of non-intelligent processes; but this is never the case with evidence that must literally be read before it can be used, as is the case in history.
Notice how Tucker falls back on the language of realism; the 'fidelity' of textual information. To what is such information faithful? It cannot, by Tucker's own admission, be 'the facts'; there are none, in advance of historical inquiry (even if we admit, as he rightly does, the received knowledge of the tradition of historical investigation as part of what we mean by historical knowledge, this too comes under the heading of material that must be read or heard to be useful to the historian; it does not supply the role of a DNA sample). It cannot, also, be faithful to 'the past'; we have, again by his own admission, no way of knowing whether or not this is the case, nor even of knowing whether there is an independently existing past. Information can only be faithful to the events; but these are precisely what we piece together on the basis of the evidence, in hermeneutic fashion.
Tucker, however, does not like the hermeneutic tradition of Verstehen which would have rescued him from his frequent but unacknowledged retreats into a realist position. He makes the point, which must be conceded, that in many cases its proponents (like Dilthey and Collingwood) emphasized rational, conscious, and voluntary action, which leads to an untenably truncated concept of historical understanding (201). Much action is not rational, and much of history is not limited to the study of action; historical movements take place over centuries, even millennia, and historical actors may be blithely unaware of them. Demographic and economic trends, for instance, may operate without anyone alive at the time being aware of them.
Tucker is also right to say that the classic hermeneutic situation of a 'reader, a text, [and] the relationship between them' does not apply to historiography. In history we are dealing with 'a community of interpreters, their theories, and sets of documents'. (259-60) This is entirely true. But in the very next sentence there seems to be a tacit accommodation with at least a moderate type of hermeneutic tradition, when he concedes that 'The world is not a text, but it can be interpreted as texts'. (260) This is all that any plausible hermeneutic theory would require to defend a theory of history as providing an understanding of the meaning of past events, where their meaning involves some grasp of the reasons for their occurring as they did.
However, Tucker neglects to bring out the connection between hermeneutic theory and the study of meaning, preferring to concentrate exclusively on the admitted defects of the hermeneutic approach. This means that he cannot accommodate the qualitative difference in the nature of the information dealt with by history. It is true that history must be able to study long term trends in population, economy, and so forth as well as individual actions and decisions; but the ultimate focus is always on the difference that such trends made to the people who were alive at the time. An individual or a society does not have to be aware of an economic or demographic trend to experience, however unwittingly, its effects, and to respond to them. The focus in history, in other words, is on the changed life-experience that such trends produce, and here there is ineliminable reference to subjectivity and reflexivity of a sort absent from biology and linguistics.
There is however a commendably powerful sense of the seamless nature of events in historical thinking. For Tucker, historical explanations can be either determined or underdetermined. A determined historical proposition is one that is fully warranted by the evidence; an underdetermined one is one for which evidence is partially lacking, so that there may be several (though not indefinitely many) equally plausible explanations for the evidence as it stands. This is one major source, Tucker claims, of historical disagreement and of the division of historians into identifiable schools. But it is clear from this distinction that there is no period in history about which statements would be inherently underdetermined simply because of the kind of events they were dealing with; all historical knowledge may be determined if we possess adequate evidence, on Tucker's view. 'Justification comes form the evidence and the theories that connect the explanatory structure with the evidence, a distinct deductive, inductive, or other justification is redundant' (187).
In history, there are no mysteries from necessity, only from lack of sources. And the sources are all that is required. Always allowing for genuine cases of ambiguity and incoherence within the evidence itself, and allowing too that no historical explanation is ever completely certain, the evidence is, ultimately, all that historians need; they do not need additional covering laws to work out the relationship between the events they describe, because the events they describe are related to one another in terms of other events of which ultimately only a handful of interpretations are possible, one of which will usually turn out to be the right one. Or at least, not obviously wrong like the others.
To claim otherwise, in the face of the achievements of modern historiography, is to resort to outright scepticism. Tucker does not make as much as he might of the fact that in the twentieth century humanity achieved for the first time a completely secular, critical, inferential perspective on its own history as a universal whole, definitively superseding the kind of overview of world history offered by, for example, Hegel in the early nineteenth century. This knowledge of human history was achieved without the use of covering laws, nor have any been derived from it.
Of course, in some contexts, this may not be relevant. If one wants to show what history has in common with all other sciences, it can perhaps be disregarded. But in understanding what makes history distinctive, it is crucial, and Tucker's neglect of the point shows in the tension between his tacitly positivistic conception of neutral information on the one hand, and on the other, his insistence that there are no ready made facts of the sort on which positivists traditionally insisted. His recognition that history cannot be made to fit the model of the natural sciences seems tinged with regret; 'For now, fully scientific historiography is science fiction' (253).
This is perhaps because of the belief that Tucker has in the 'undistorted' communication that he believes takes place in a scientific community (179). There is a related belief that the interpretation of meanings of historical events can be entirely divorced from the acquisition of historical knowledge which is surely false given Tucker's own belief there are no prefabricated facts (215). These two claims are rather analogous to the belief in neutral information mentioned above. Of course, insofar as scientists can communicate in entirely formal or mathematical terms, there is some sense to the notion of undistorted communication; pure mathematicians, mathematical physicists, and logicians really are freed from the inherent ambiguities of natural language.
But these activities form only a part of what we call science. In fact, Tucker quickly retreats from this position to the view that 'within a sociohistorical community of discourse a fairly undistorted and meaningful communication frequently takes place' (179). This is unexceptional, but it is as true of the community of historians as it is of any (other) group of scientists, and really states little more than the trivial truth that we normally succeed in understanding one another most of the time.
The area in which Tucker is most genuinely innovative is in showing how Bayesian probability can provide a model of historical reasoning. A particular strength of his argument is its ability to distinguish Bayesian probabilism from the kind of probabilism advanced by Hempel, which still relied on the covering law approach. There is some real philosophical value in being able to state in formal (not of course quantitative) terms the grounds on which historians distinguish between the relative likelihood of various explanations for given evidence; it is not to be confused with the attempt, shown to be futile by Goldstein, of formalising historical statements themselves. For clearly stating the logic of historical judgments of probability, he deserves recognition as a truly pioneering thinker in the philosophy of history.
Ultimately, this Bayesian approach involves Tucker in a major claim about the organization of the sciences; he asserts that 'The sciences are divided neither into human sciences and natural sciences according to their subject matters, nor into ideographic and nomothetic sciences according to the purpose of their inquiries, but between sciences that examine the similar effects of common tokens of causes that preserve information about their common causes, and sciences that examine the similar effects of shared types of causes' (260). This claim is deliberately intended to overturn what Tucker regards as two major traditions of thinking about the organization of the sciences, the positivist one noted above in which the human sciences were always the poor relations of their natural scientific cousins, and the neo-Kantian tradition of Rickert, Simmel, and Dilthey that tried to establish the autonomy of the human sciences, including history, and grant them equal status with the natural sciences in doing so.
However, what Tucker does not rule out is that the sciences are divided into human sciences and natural sciences, and that within each of these broader groups there are sciences that examine the similar effects of common tokens of causes that preserve information about their common causes, and sciences that examine the similar effects of shared types of causes. What he has highlighted, arguably, is an important additional means of classifying forms of thought rather than a classification which entirely overturns those used hitherto. Within the human sciences, for example, it would seem that history comes into the former class, of ' sciences that examine the similar effects of common tokens of causes that preserve information about their common causes', and sociology falls into the latter group of 'sciences that examine the similar effects of shared types of causes'.
His revisionist position allows Tucker to claim that he does not adhere straightforwardly to any version of either realist (positivist) or neo-Kantian (constructionist) philosophy of history (256), although he admits that 'It is impossible to refute determined constructionism. It fits determined historiography as evidence just as well as a realist interpretation. Yet, though construction is simpler than realism, it makes fewer assumptions, realism is a better explanation of the historiographic Rankean paradigm, why a uniquely heterogenous and uncoerced large group of historians has come to agree on the theories and methods that define the historiographic community' (257)
But this objection can be dealt with by raising a point similar to Tucker's own caution that one should not confuse 'historiographic phenomenology and ideology; what historians may think of their enterprise and how they like to present their enterprise, with actual historiographic practice' (193). Just because the realist assumption is highly efficacious does not mean it is right (or wrong, for that matter); it is actually a naïve importation into historical theory from the ordinary lanaguage of practical experience, which requires realist assumptions. Ordinary language, however, is not the language in which history is written, as Tucker himself admits (178).
Tucker's account of constructionism is also flawed in other ways. He claims (without providing a specific example other than Goldstein) that constructionism 'denies that historiography refers to history, to the past'. In fact, Goldstein, if read carefully, does not deny this, exactly. And nor does Oakeshott, who makes pastness a category of historical understanding while maintaining a radical constructionist position. What both argue is that any historical talk about the past cannot be treated in the same way as commonsensical talk about the past, and in fact Tucker entirely agrees with this; constructionism is a strategy for preventing confusion of what Tucker calls 'therapeutic, nonscientific' (262) accounts of the past being allowed to hold sway unchallenged and thus potentially abused by demagogues.
In another attempt to undermine the constructionist case, Tucker mentions that 'Dummet argued that sentences about the past are not assertoric, they do not assert anything, because there are no clear truth conditions that would allow or disallow us to assert them' (255). Dummet, however, was not discussing the nature of historical knowledge when he made this remark. By Tucker's own admission, historiographic knowledge is theoretic knowledge. But even if Dummet had been discussing historical truth, how damaging would this charge be, if we have already established that history is capable of providing 'the best explanation of historiographic evidence' (254), as Tucker puts it? Why fixate on the notion of truth?
What is important in the kind of context that Dummet is discussing - the context in which what Tucker calls 'ordinary folks' (191) assert things about the past - is that their claims are efficacious. If I say to myself 'I left my wallet in the bedroom', what matters to me at that moment is not that this is true, but whether my wallet is in the bedroom when I walk in there to look for it. Similarly, if the historian can explain evidence in a way that rules out other possible explanations of it, what is to be gained by claiming that this explanation is either true or false when it has done what was required of it?
In fact, constructionism neither wants nor needs to abandon truth claims in the way that Tucker suggests, though there are, as we have just seen, reasons for thinking that it would not be as problematic as he believes even if it did. Nevertheless, all constructionism needs to assert is that there is a qualitative difference between historical explanations and narratives of the past, and other statements about the past. And this is tacitly admitted by Tucker in the distinctions he makes between historiographic and therapeutic accounts of the past, and between historians and 'ordinary folks'.
It is strange, then, that Tucker should still write that 'We understand history backward' (p. 229). Of course, there is a sense in which historians do understand history backward, because they have some idea of the outcome of the series of events they investigate. But there is a serious ambiguity in the notion of understanding history backward that needs addressing. Where the past is seen as significant or relevant, it is really the therapeutic past which is under consideration; this, by definition, has to be a kind of past, or view of the past (we need not worry too much about the distinction) in which its importance is as the preamble to our present.
By contrast, the theoretic or historical view of the past is of the past in terms of its past; the explanation of an event in terms of its antecedent causes. We do not understand why the French revolution occurred in terms of what came after it; we grasp the reasons why it happened in terms of what lead up to it. And it we ask 'what were the consequences of the French revolution', we must distinguish between asking that question in relation to those who experienced them in the nineteenth century, for instance, and asking it in relation to ourselves.
It is a hallmark of all theoretical thinking that our practical concerns must be temporarily shelved. This is as true for readers of historical works as it is for writers of them, and it is a precondition of the engagement of the historical imagination, which is the real source of the civilizing value that Tucker correctly detects in history. This critical species of imagination is an integral element of the educated mind, and it is one of the things that allows individuals to resist being moved to potentially self-destructive action by specious rhetoric.
Ultimately, Tucker is not really entitled, given his own tacit concessions to the hermeneutic and constructionist positions coupled with his illegimate tacit resort to a positivist and realist concept of 'information', to talk about “our knowledge of the past”, when we cannot know that there is any such thing. An indication of where he has gone wrong comes in the claim that 'evaluating how contingent history is can only be done empirically' (226). Contingency is here being contrasted with necessity. But history, as Oakeshott observed, is totally contingent. Oakeshott did not mean to say that history is accidental, for none of it is, in the sense of being composed chance or random occurrences.
Nor did Oakeshott mean to say that no necessary natural processes (sunrise, death, etc.) are at work in history, or even that, as Tucker alleges, 'universal or general statements are inapplicable to historiography because historical events are unique' (242). Historians are perfectly able to generalize; but the kind of generalizations they put forward never have the status of universal laws of the type associated with positivism.
What Oakeshott wished to draw attention to was that all of the historical past could have been otherwise, even if what might have been different would in many cases have been very similar to what did in fact occur, as Tucker rightly stresses in his reflections on the plot of It's A Wonderful Life (238). And it could have been otherwise because we are always dealing with the responses of people to events, and these, while they may have certain structural limitations (not all things are possible for all people at all times and in all places), are never foreordained. All history is equally contingent; which is quite consistent with holding that some events in history were more likely than others. Their likelihood, however, at least by itself, tells us nothing about
The contrast between contingency and necessity only arises in the context of the very debates over the scientific nature of history that Tucker is rightly trying to leave behind. It presupposes a certain metaphysical viewpoint; namely, that we inhabit a natural (perhaps 'external') world in which certain (necessary) processes with an identity entirely independent of consciousness are going on, but into which subjectivity introduces an element of randomness (contingency). The scientific, historical, and practical attitudes are all caught up in this continuum. This metaphysical viewpoint however is something that we simply cannot know the truth of. The most that the scientist can do is to presuppose that this is the case. Such a presupposition is perfectly legitimate, but not to be turned into an unqualified truth about the nature of reality.
But Tucker in fact sees that 'The epistemic question whether or not scientific historiography is possible is independent of the ontology of history' (213). And indeed, he does not need to avail himself of the notion of the past to make good his account of history as a form of probabilistic inference to the best explanation. The irony is that a more careful consideration of the neo-Kantian constructionist position that he claims to repudiate might have revealed to him how much closer he stands to it than he thinks. His claim that 'All historiography is hypothetical, unobserved, and should be the best explanation of observable evidence' (149) is, at the very least, entirely consistent with the constructionism of Goldstein and Oakeshott. And constructionism also contains the key to resolve some of the outstanding difficulties with what, by any standards, is one of the most important contributions to philosophy of history of recent years.