History and Theory, Vol. 16, No. 4, Beiheft 16: The Constitution of the Historical Past (Dec., 1977), 53-71.
TRUTH AND FACT IN HISTORY RECONSIDERED
W. H. WALSH
At the outset of my Introduction to Philosophy of History, first published in 1951, I listed a number of problems which, I said, should undoubtedly be recognized as genuine by philosophers who approach the discipline of history from a critical point of view. The implication was that these problems would constitute the main repertoire of critical or analytic philosophers of history. It may be of some interest to others besides myself to see to what extent this program, or prophecy, has been borne out by events. One of my main topics, the nature of historical explanation, was already a subject of lively controversy at the time I wrote and has been debated extensively ever since, so much so that it is now widely recognized as among the central questions of philosophy generally. Another, the question of the extent to which historians can be expected to offer objective accounts of what happened in the past, has attracted continuing if less vigorous interest, if only because of its connections with the wider question of the extent to which statements about human affairs can be value-free. A third problem I distinguished, about the logical level of historical discourse and the relation of history to other forms of knowledge, has also attracted some attention, though not perhaps in the form in which it originally interested me: it has turned, in effect, into the much-debated question of the relations of history and the social sciences. Only the fourth of my problem areas, which concerned truth and fact in history, has until now been largely neglected.'
What philosophical problems are there about truth and fact in history? At the time I wrote my book I thought that the main problem was posed by the fact that historians purport to tell us what happened in the past, and yet have no direct access to it. They cannot directly verify any conclusions they come to, nor can they claim them to be true because borne out by testimony which is unimpeachable. As Collingwood had made clear, historians recognize no such testimony: they treat statements from the most reliable of witnesses as no more than evidence, and for them all evidence
must establish its credentials if it is to be taken seriously. The problem of arriving at truth about the past thus becomes the problem of whether, on the basis of such evidence as we can now command, historical thought can arrive at a convincing account of what occurred. But if that is so, historical facts are not so much discovered as arrived at by processes of argument, and the question whether we can accept something as fact is the question whether we can fit it in with the other conclusions to which we have already committed ourselves, or can fit it in without disturbing those conclusions to an undue extent.
If we ask, as I did in what I now fear was a somewhat pedantic way, to what theory of truth historians are committed, the answer is not altogether easy to find. What has just been said suggests that it is not possible to apply the usual Correspondence account to the case of history without getting into very great difficulties. The past as it actually was is not open to our observation, and there is no reason to think that any remains we now have of it constitute in themselves what might be termed unvarnished transcripts of past reality. Historical conclusions must accord with the evidence; but evidence, too, is not something which is fixed, finished, and uncontrover-sial in its meaning and implications. Evidence has to be authenticated, and again evidence has to be assessed. Just because of this the ideas of those who explain truth as Coherence seem to correspond to historical practice much more readily. In history, far more than in perceptual situations, it is possible to argue that we accept or reject suggested truths not by confronting each of them directly with reality, but by thinking about their coherence with the rest of our beliefs. As already said, we try to reconstitute the past on the basis of the evidence we now have, but it is not a question of making our constructions conform to the latter as independent fact. Rather what we have to do is think out a hypothesis about the nature of past fact which will allow us to take the evidence for what it is and offer a connected account of it. Here we are concerned to make one set of judgments coherent with another, precisely the activity on which supporters of the Coherence theory lay such stress. It is not surprising in these circumstances that some writers have tried to explicate historical truth exclusively in terms of coherence, particularly Michael Oakeshott, who said that "if history is to be rescued from nonentity" "what really happened" must be replaced by "what the evidence obliges us to believe." But to say with Oakeshott that "the past in history" not only "rests upon" and "varies with" the present, but actually "is the present," is surely to go too far, or so at least I thought when I wrote this part of my book. It was one thing to say that historical evidence must be present to us now, another to claim that it refers to present time. The truth was obviously that the evidence with which historians deal refers not to the present but to the past. And if it was asked how we could possibly have any access to the past, seeing that we could not
observe it directly, the answer I offered was that we do so through memory. Memory is not a species of knowledge by acquaintance, since its deliverances are all subject to interpretation; it nevertheless gives us a sure link with the past, just as sensation gives us a sure link with external reality. Sensation needs to be elaborated in perceptual judgment, pure memory in memory judgment, and neither form of judgment is infallible. Yet each form of judgment has its foundation in fact; each contains a "given" element which makes it impossible to dismiss.
The problems which I thus discussed in an elementary way have been taken up and illuminated quite independently by Leon J. Goldstein in his book Historical Knowing.2 Goldstein improves on my discussion not only by treating the subject in far greater detail, but also by situating the general problem in an altogether bolder way. For him, questions about truth and fact in history are not just one of the topics on which philosophers of history may be expected to concentrate their attention; they are the key questions for a properly constituted philosophy of history. That branch of inquiry has as its business to explicate historical thinking as it actually is, not as abstract principles dragged in ab extra dictate that it should be. And when we ask what historians actually produce in the way of knowledge, the answer is that they produce largely-agreed accounts of what happened in the human past. They reconstitute, or as Goldstein prefers to say, constitute the events of the past on the basis of the evidence before them now. Goldstein's attitude to that evidence is very much the same as was mine:
he takes it as something which, as Collingwood said, is relative to the questions asked rather than as something which exists in a vacuum of its own. Historians have to establish what is evidence in any particular inquiry, as well as to decide what conclusions the evidence points to. One point which Goldstein stresses and which I and, I think, Collingwood overlooked is that historical thought is carried on largely on a corporate rather than an individual basis. The consensus or near consensus of historians is important both when it comes to saying what is to be explained and when we ask what explains it. The fact that we have here to do with a common activity carried out according to rules and standards which are widely accepted is a large part of Goldstein's case for his claim that, as he puts it in the first sentence of the book, "history is an epistemically licit discipline which deserves to be taken seriously on its own terms."
We should notice that the epistemically licit discipline here spoken of has, for Goldstein, strictly limited aims. It is not, it would seem, an explanatory discipline, nor one which offers interpretations by arranging or seeing a
pattern in the facts.3 Philosophers of history have given enormous attention to the question of historical explanation in recent years; they have also repeatedly discussed the question of historical relativism. Goldstein clearly believes that such discussions are misplaced,4 since they spring from a totally false conception of what historians really do. Philosophers judge history by the end product it puts before the world and thus confine themselves to the historical superstructure; what they should be doing is giving their attention to the infrastructure involved in the constitution of historical fact. In this area historians have managed to achieve what Goldstein calls (HK, xii) an "extraordinary amount of agreement," though that is not to say that they agree about everything. Indeed, it is Goldstein's thesis that philosophers should give particular attention to the cases over which they disagree. Studying such disagreement, as in his case of the Dead Sea Scrolls, enables us to see historical thought at work, and thus brings out the delicate ways in which evidence is adjusted to theory and theory to evidence in the hope that an acceptable account will emerge.
It seems to me unfortunate that Goldstein does not say more about the precise nature of the end results of historical constitution, and again about the relations between constituting the human past and other activities in which historians engage. Normally what he says is that historical thought constitutes historical events, which is unsatisfactory in itself, for are not historians also concerned with structures (such as institutions) and states of affairs'] To say that historians constitute historical facts would be altogether preferable to saying that they constitute historical events. But what in any case is an event in history? In many cases at least it is an action, and thus includes a causal component. When Brutus killed Caesar, a certain causal transaction took place, and when Wellington defeated Napoleon, another. But once we agree that the historian can tell us not just about bare happenings but about the operation of causes, the line between constituting the past and explaining it begins to seem shadowy. Insofar as historians aspire, as I argued in my book that they do, to construct not so much a. plain as a significant narrative of what happened in the past, they recognize that there is more to their discipline than Goldstein allows.
Goldstein may reply that my whole distinction between plain and significant narrative in history is untenable, and add that he has given good reasons for thinking that history is not typically or properly concerned to
construct narratives (see HK, ch.V). But even if we accepted these points, we could not allow him to get away with his whole account of what he calls "the presently practiced discipline of history" (HK, 93). That discipline, as he expounds it, seems to be essentially concerned with the question what conclusions about past fact can be drawn from present evidence; once that question is answered its real work is complete. But it is obvious from the merest inspection of what historians do that this opinion is mistaken. It may indeed be true that tackling factual problems of the kind Goldstein specifies is the most basic of historical activities; it may be, again, that it has been neglected by philosophers. But it is simply absurd to describe it as the only, or the only legitimate, historical activity. Historians are concerned to sum events as well as to set them out; they see it as one of their tasks to offer interpretations, in the hope that this will enable their readers to get the different facts into perspective; they feel the need to explain, not merely describe. Admittedly, these tasks appeal in different ways to different individuals: there are some who remain almost exclusively at the level of constitution of fact, others who concentrate on fresh ways of taking the facts as established, others again who look for new ways of explaining them. But to emphasize one at the expense of the others, and indeed to separate them sharply one from another (as if one could concentrate on constituting facts without risking any kind of interpretation) is certainly a mistake. Philosophers of history may have got their priorities wrong in paying so much attention to historical explanation and interpretation and so little to the problems involved in deciding what went on at all. They may have been too much impressed by the surface disagreements of historians (about why things happened and what they amounted to) and too little impressed by their underlying agreement (on what in fact occurred). But they were not wrong in arguing that historians generally claim to do more than constitute facts and in maintaining that their activities in this regard are of real philosophical interest. My own view is that Goldstein would have gained rather than lost by admitting these points. It might not have been so easy for him to say, as he does, that history is an epistemically licit discipline, since it is obvious that explanation and interpretation in history remain highly controversial. He could, even so, have made plain that historians explain and interpret as an essential part of their work without losing sight of what he regards as the all-important fact that there is something else they do and do with conspicuous success, namely find out what happened in the past.
For the purposes of this discussion, however, we must take Goldstein on his own terms and accept his thesis that history is essentially fact-constituting. Assuming that it is, what points of importance does Goldstein make about it? One point which Goldstein himself clearly thinks of great importance is that the world which historical thinking constitutes is a world
which exists in relation to the thought of the historian and not in its own right. The "real past," the past as it occurred whether or not we know of it now, has no part to play in live historical thought, if only because it is over and gone for ever. The actual personages and institutions of the real past are simply inaccessible to the historian working now. But this is not to say that historical knowledge is impossible. If the "real" past eludes us, the whole point of history is to constitute an "historical" past whose authenticity will be recognized by the community of historians. And the fact that historians do largely agree about the character of the human past shows that this aspiration is successfully carried out.
In his essay above P. H. Nowell-Smith argues that Goldstein has given a correct and illuminating account of the way in which historians reach conclusions about what in fact happened, but is quite mistaken about the import of their activities. Goldstein says that historians constitute past events; what they actually do is construct accounts of past events, which purport to be about those events as they really were. It is the historian's claim, according to Nowell-Smith, to be in possession of knowledge of what the past was really like; largely on the score of considerations advanced by Goldstein, Nowell-Smith is prepared to concede the claim. But Goldstein, as his reply to Nowell-Smith shows, is far from grateful for the concession. For one thing, he is suspicious of Nowell-Smith's whole approach to this and other questions in philosophy of history, complaining that he reaches conclusions by appeal to general logical or linguistic considerations instead of asking himself what goes on when historians are actually at work. Nowell-Smith says that Goldstein has good methodological points mixed with dubious philosophy; Goldstein thinks the whole distinction between philosophy and methodology begs the question against him. Apart from such issues, however, Goldstein is sure that Nowell-Smith's account cannot be correct, since the real past which it invokes as the object of historical reference is for historical purposes a cipher. The historian has no means of getting in touch with the real past, and it accordingly plays no part in his thought. To retain it is only to provide a standing incentive to embrace historical skepticism. But as Goldstein sees the matter, historical skepticism is ruled out by the very fact of the successful functioning of history as such.
In part VI of his paper Nowell-Smith tries to make his point against Goldstein by means of a distinction between reference and verification. A proposition refers to whatever it is supposed to be about, and this holds whether the proposition speaks of things present, past, or future. The propositions of history thus refer to past facts, and are true just insofar as they correspond to such facts. But though fact is in one sense the touchstone of all truth, it does not follow that we verify propositions of every sort by matching them against fact, and it is definitely false that we
do so in the case of historical propositions. Nowell-Smith is ready to allow that historical statements are verified along the lines that Goldstein describes. But he still insists that the formal referent of such a proposition is past fact, something which, he admits, can no longer be confronted and observed.
It is interesting to see Goldstein's response to this move. In part it consists in grumbles about "conceptual distinctions which are abstract and removed from real quests for knowledge" (above, 48), distinctions which have to do with logic but which may well be irrelevant to a philosophy which insists on "the primacy of knowing." But Goldstein also argues that the objects referred to in actual historical statements are specified sufficiently in the statements themselves. The logical distinction between "a description and its referent" (34) or between "an assertion and its ground" (35) is simply not applicable in this sort of case. The reason for this is the point already indicated about the vacuous nature of the concept of the real past. We could, says Goldstein near the end of his paper (51-52):
force a distinction between the referent of an historical assertion and the way in which it refers. The former would presumably refer to the real event in the real past and the latter would involve the considerations of technique and methodological procedure I have just been talking about . . . But in the end this will not work. To be sure, the distinction is intelligible; the two are conceptually distinct. Yet when we seek to put it into the cognitive framework within which historical knowledge is actually acquired, there is nothing we can do with it. The referent remains entirely empty of factual content: whatever can be said of it depends on the outcome of methodological procedures.
Thus so far as history is concerned we have to exchange the real past, what really happened, for the past as constituted in historical thought, what the evidence obliges us to believe, though Goldstein does not recall Oakeshott in this connection.
I am inclined to think that Goldstein has not reflected enough about what he calls "the real past," nor again about his stance vis-a-vis idealism, the philosophical position from which Nowell-Smith wants to save him. Goldstein writes that "it is hard to doubt that there was a real pastor to formulate such a doubt in intelligible language" (33-34), and throughout his book he speaks as if we had to accept the idea. Yet he nowhere inquires what content the idea has, nor whence it derives that content. Two considerations may have led him to neglect the question. First, his view that the problem of the reality of the past, for all the prominence given it by Arthur Danto in his Analytical Philosophy ofHistory, is in truth not a problem of philosophy of history at all, since the class of statements about the past by no means coincides with the class of historical statements (see HK, ch. I). Second, the feeling that in setting out his own ideas about history, he should stir up as few hornets' nests as possible and should therefore acquiesce in commonly accepted views on non-historical matters as far as he
can. It is this policy, I suspect, which led Goldstein to proclaim himself a "realist" as regards the philosophy of perception, and it may well have had a bearing on his being willing to grant that there is a real past even though it is of no relevance to history.
The trouble is that the arguments Goldstein himself produces for saying that, as far as history is concerned, the real past is a merely vacuous notion apply equally to the real past as it figures in everyday thought. As Goldstein says, it is difficult to deny the reality of the past, and yet we can never assure ourselves of that reality by confronting it face to face. Memory, even at its best, hardly gives us acquaintance with past events, and lacking such acquaintance we can claim no direct knowledge of the past as it was. But what content can we then give to the idea? The only available content is what we can come by if we reconstruct what happened now, using our memories and those of others but also drawing on any other means we have of making our account more reliable old letters, reports in newspapers, photographs, and the like. Goldstein often writes as if what the historian did in constituting or reconstituting the past were utterly different from, altogether more esoteric than, what we do in our daily lives in trying to piece together something that happened. I agree that it is more sophisticated and depends on the development of techniques which most of us do not normally command, but would say that it nevertheless only develops out of our common-sense thinking. Nor is it true, as Goldstein also seems to suggest, that the historical past differs from the real past because it could not, while the real past could, in principle be observed. History, as he points out, is often concerned with developments which last for more than a lifetime and are therefore not the sort of things which can be witnessed at all. But the "real" past, if we take that notion seriously, was not just a succession of events; it included processes which lasted through long periods of time (the onset, continuance, and retreat of Ice Ages, for example), and these too could hardly have been observed as present happenings can.
It looks as if Goldstein cannot have it both ways on the question of the real past: he cannot deny its reality for history while proclaiming, or at least acquiescing in, that reality for the purposes of everyday life. He must put his money on one or the other. If his final wish is to go along with common sense (and we must remember in this connection his point about the difficulty of even formulating an alternative), he should have no difficulty in extending the notion to history as well, and thus in accepting the main point Nowell-Smith makes against him. Of course he will still want to say that our detailed answers to the question "What really happened?", in real life as in history, will be come by through reflection on evidence available now; the content of the idea of the past as it actually was will be dependent on our efforts to reconstruct it. But this is something which Nowell-Smith has no wish to deny; the realism he supports is
explicitly distinguished from the theory which claims that "the touchstone of historical truth must be direct observation of or acquaintance with the objects concerned" (see section III of his essay). For Nowell-Smith to say that there was a real past is to say only that certain things happened or were the case; what those things were has to be found out by examining evidence. If Goldstein can accept this notion as significant in everyday life, there is no reason why he should not accept it in history. He may say that its part in historical thinking is formal at best. But we have seen that its part in everyday thinking is equally formal.
Alternatively, of course, Goldstein can rid himself of the idea of a real past in everyday life as well, on the ground that it is without meaningful content, a past-tense version of the Kantian thing-in-itself. If he had to make the choice I suspect he would prefer this option. This brings me to my second point, about Goldstein's attitude to idealism. There are a number of places both in the book and in the paper where Goldstein toys with the idea that it is not only in history that reality is constituted, but in other areas as well. "The primacy of knowing" is not a condition peculiar to history; it has relevance to the world of science and even to that of perception. We think we confront independent reality in perception, but the truth is that, as Goldstein puts it in one place, "there is more to being a delphinium or a garden than the brutely given. To be either is to realize a concept, and that immediately plunges us into a context of knowing" (see his footnote 9). History may involve more explicit thought than does perception, historical fact be less "brutely given" than perceptual fact, but the difference between them is only one of degree.5 At one point in the paper Goldstein mentions some reasons for thinking that perceptual realism is tenable even though historical realism is false, but adds that "it might be possible to argue . . . that an examination of the way in which claims to historical knowledge are made provides a purer case on the basis of which to construct a theory of knowledge than other more usual cases" (48). It is, apparently, a general feature of the knowing situation that we there have to do with "the development of claims to knowledge by means of systematic methodologies into which unmediated reality cannot obtrude itself" (47-48). "Unmediated reality," reality as it exists in independence of thought, is in fact always a dubious notion so far as Goldstein is concerned; and it could
5. A major point of Goldstein's is that there is nothing in history that corresponds to sensation; his remarks about memory in HK, 145-148 show that he would not allow memory to fill the bill. He also says that historical facts cannot force themselves on us as do perceptual facts. History is an affair of thought, and of thought alone. Yet Goldstein also recognizes that historical facts are constituted on the basis of data, in the form of evidence: history is not wholly free thought. And in passages like the one under discussion he not only suggests that perception involves thought, but hints that it too may be properly a matter of making and reconciling judgments.
be argued that he would have done better, and certainly made his overall position more consistent, to have got rid of it altogether.
But would not that have solved one difficulty only to land him in others? Nowell-Smith argues that the restricted idealism to which Goldstein implicitly commits himselfhis theory of historical constitutionis not coherent as it stands. One problem he finds in it arises out of Goldstein's language, with its suggestion that when the historian constitutes events he somehow makes reality. A glance at the brief passage about constitution in the introduction to Goldstein's book (HK, xx-xxi) shows that this is unfair. Goldstein borrows the term "constitution" from Husseri, and quotes a commentator on that author for the view that "for Husseri consciousness is a necessary but not a sufficient condition for reality; while it constitutes it, it does not create it." The same commentator says further that "although intentionality alone constitutes meaning and objects, sensations also have a role in determining their constitution . . . the role of a 'raw material' out of which the objective is formed." Just how this happens is unclear from the text, but it is obvious that when Goldstein speaks of constituting events he is not implying that they are conjured up out of thin air. On the contrary, they are constituted on the basis of evidence, and exist in relation to that evidence. To put the matter in a different terminology, they are products of judgment, facts to which we conclude when we bring our thought to bear on such relevant remains of the past as we can now assemble. I have already argued that Goldstein would have done better to speak of the constitution of facts rather than the constitution of events, since historians are not concerned with events alone. But even if he were unwilling to accept that point, he could still have phrased his view more effectively had he spoken of historians arriving at the conclusion that such and such things occurred. He would then have made events, with any other historical objects there may be, constituents of judgment, and thus preserved an essentially idealist position without exposing himself to the crude objections which his own terminology invites.
To switch from "constitution" to "judgment" would clearly not satisfy Nowell-Smith, who could still argue that actual events are one thing, judgments to the effect that they took place quite another. Goldstein has given his reasons for rejecting this common-sense dichotomy, at least as far as history is concerned, and I can say of them only what Hume said of the arguments of Berkeley, that they admit of no answer and produce no conviction, so deeply engrained is our belief in an actual past. But instead of discussing this issue further now I should like to turn to the question whether Goldstein's position is really incoherent, as Nowell-Smith says it is. In addition to the general objection mentioned above, Nowell-Smith argues that Goldstein is involved in fatal difficulties about the identity of the objects of historical thought and again about the status of the events
concluded to (i.e., constituted) in conflicting historical theories. I do not believe myself that these objections are unanswerable, though I am not wholly satisfied by Goldstein's own attempts to answer them. I should therefore like to sketch a position which I think Goldstein could hold which would not expose him to the objections.
The first point to make clear here is that one who says that historical events are constituents of historical judgment must understand the latter in an impersonal sense. Particular thinkers on particular occasions commit themselves to judgments about what is or (in the instance we are considering) was the case, but they do so not as this thinker or that. They claim that what they say has the status of truth, and in so doing presume to speak for intelligent persons generally. If it is true that Brutus killed Caesar, it is not just true for me who now make the judgment; it is true without distinction of persons. The subject of judgment, as I tried to make clear in my commentary on Kant, Kant's Criticism of Metaphysics, is the Kantian unity of apperception, and this is something which is not concrete but abstract. Insofar as each of us thinks rationally he conforms his actual thinking to the ideal thought of which the unity of apperception is the subject. To speak of judgment is thus to speak of a proceeding which is logical rather than psychological, and which is accordingly not peculiar to the personal history of any particular thinker. As Kant made clear, there is an important contrast in this respect between judgment and the association of ideas. Association is, or may well be, confined to individual minds, but judgment has to do with mind as such or, in Kantian terminology, "consciousness in general."6
If this is correct, we can draw a contrast between judgment proper and particular attempts at judgment. Whenever any particular person commits himself to a statement on a factual issue he aims at judgment proper, since he claims to say what is true. If he succeeds if, that is, he argues with-out preconceptions from true and adequate premises and by proper methods he can be said to have constituted fact (there is nothing surprising in this, since the notions of fact and true judgment are internally related). But of course he may not succeed, perhaps because he is not in possession of the requisite evidence, perhaps because he is wrong in the conclusions he draws, perhaps because he is led into error by particular prejudices or preconceptions of his own. In this case the "events" to whose existence he concludes will not be real events; they will not belong to the world of fact. They will exist only in the context and as the upshot of thinking which is erroneous or defective, and will hence be unable to lay
claim to objectivity. What exists in the objective sense, on this way of thinking, is not what is acknowledged or concluded to by any actual thinkers, but what would be acknowledged or concluded to by thought in its ideal form. If each of us were able to put off the old Adam and appear only as an impersonal unity of apperception, a representative of consciousness in general, and if he had adequate materials at his disposal, the content of his judgments would coincide with that of judgment proper and would represent what is objectively the case.
It is with these considerations in mind that we should consider Goldstein's example of the dispute about the Dead Sea Scrolls and Nowell-Smith's objections to his views. Let us notice first that there are many occasions, in some branches of history at least, on which historians are not really in a position to make judgments at all. They raise questions which cannot be rationally answered, for the simple reason that they lack the evidence on which an informed answer could be based. Much of the material handled in early Greek history, and even in parts of the history of fifth-century Greece, answers this description. Historians may speculate in circumstances of this kind, offering "theories" which cover such evidence as there is but can hardly be convincing just because it is so patently exiguous. That there should be a plurality of such theories in such conditions, each issuing in the postulation of this or that supposed fact or event, is not in the least surprising. There is no mystery in their existence, nor any paradox in the fact that they argue for different and incompatible "events," for the latter exist in the context of the theories alone, and the theories are put forward in conditions in which more prudent investigators would abstain from theorizing altogether.
I do not know whether the dispute about the Dead Sea Scrolls during the period with which Goldstein is concerned fell into this class of undecidable cases; if so, no more need be said of it. From what Goldstein says (41; cf. HK, 104, 223-224), it may have been at a more interesting stage, since it seems that one of the hypotheses then canvassed has subsequently been widely accepted by the community of competent scholars. Goldstein has the impression that the view that the Scrolls were produced by Essenes or "Essene-like sectarians" is now "overwhelmingly supported." Unfortunately he does not tell us how this change has come about. Presumably it was by the discovery of fresh evidence and/or the adducing of considerations whose relevance was not seen, or not fully appreciated, at the earlier stage. If this is correct (and one hopes it is), the thinking which issued in the conclusions that the Essenes produced the Scrolls has been radically changed. Previously there was only a sketch of an argument, beginning from inadequate premises and eking out inference proper with more or less well-judged guesses; it purported to constitute an event, but could hardly
sustain its credentials. Now (one presumes) the situation is very different: evidence is fuller, argument altogether more convincing, so much so that the community of scholars is persuaded that the postulated event was real.
What we have to ask is whether in a transition of this kind I am not of course concerned with the particular case of the Scrolls we are warranted in claiming that we have got to objective fact. Suppose that the community of scholars is now at one in accepting some thesis that was previously controversial: does that show that the thesis is definitively true? May it not be the case that scholars generally are in error over the point, as they have apparently been in error at some stages in the history of science? It would certainly take a rash man to deny this possibility. Nor, of course, does the theory we have put forward require us to deny it, since it distinguishes sharply between thinking as it goes on in individual minds and thinking as it should be. That there is wide and general agreement among competent scholars that such and such an event took place will not prove that it actually did. But I think Goldstein is right in believing that it is not irrelevant to the question. The point on which he rightly insists is that historians over many generations have worked out techniques of considerable sophistication which have been widely and successfully applied in answering particular questions about what occurred; to shut our eyes to this fact is to misrepresent the whole discipline of history. The presumption must be that, given the right conditions (above all the right evidence, but also of course the appropriate background knowledge), competent historians can be expected to get things right, with the result that what is widely accepted in the historical community must indeed have been the case. There may be special circumstances in which historians fall into collective error. But such circumstances will be rare; normally things will go right, and hence normally the past as constituted to the satisfaction of the community of historians can be taken to be the past as it objectively was.
If this seems unduly complacent, two points can be made. First, that the discussion has been about the bare constitution of the human past, the discovery of what in fact happened, rather than of the whole range of activities in which historians engage. What Goldstein has claimed is that there is a wide measure of agreement among historians on what conclusions are appropriate at the factual level; it is this feature of history which, in his view, philosophers have failed to appreciate. If we grant Goldstein this point, as I think we should, we shall not be claiming that all the problems of history have been solved. Even where there is agreement among the community of historians about what we might call history in the round (and such agreement tends to be local and short-lived), we need not think that final truth has been reached. Second, if agreement among those properly versed in the subject is to play the part in the constitution of historical fact which Goldstein wants, it will be necessary to define with
some care in what conditions it can be said to be reached. When most scholars think that the language of the Linear B tablets is Greek, but a small minority is still skeptical, can we say that the matter is decided? What if only one man holds out against the majority, but he happens to be a scholar of very good knowledge who has done important work in the field? If historians are asked the question "Was there an historical George Washington?," they will be unanimous in their answer. But if they are asked "Was there an historical King Arthur?," even though there is a substantial majority for one or the other view, it is not clear that we can properly speak of the general opinion of historians. And there may of course be circumstances in which historians are not entitled to any general opinion, just because they lack the resources necessary for arriving at a rational judgment.
Nowell-Smith argues that historians do not constitute the past; they construct accounts of the past, accounts which are true if they conform to the past as it actually was, false if they do not. On this view the past is what it is, irrespective of whether anyone now knows anything about it. Nowell-Smith believes, I suspect, that this position has only to be stated to be seen to be true; the fact that Goldstein hangs on to the notion of a "real" past for everyday purposes suggests that he may be right to take this line. But Nowell-Smith does not rely on it exclusively, but offers supplementary arguments designed to discomfit his opponent. Hence his long disquisitions on the rival events postulated by scholars in the Scrolls controversy and on the problem of the identity of historical characters. The object of my own discussion here has been to show that Goldstein need not be embarrassed by the first problem at any rate, provided that he is willing to think out his epistemology in more detail than appears in his book and his essay. He needs to face the fact, of which he is remarkably shy, that if you are an opponent of historical realism you must be a supporter of historical idealism; and he needs to discuss what idealism allows and what it precludes. The idealism Nowell-Smith attributes to him is of a half-baked variety which argues that whenever anything is thought of, it is real; Gold-stein's sketchy account of "constitution," and his talk about the part played in it by "subjectivity," do not do enough to discourage this belief. An unsympathetic critic can take him to hold that the historical past exists not in relation to thinking generally, but to particular bits of thinking and hence to particular thinkers; it is vital to Goldstein's case that this illusion be dispelled.7 The Kantian argument I have suggested, which connects the notion of established fact with that of warranted judgment, but allows that particular thinkers on particular occasions may fail to satisfy the require-
ments of judgment proper, is designed to provide for this result. It follows Goldstein in insisting on what he calls "the primacy of knowing" and in making fact in history a function of evidence, but explores the general cognitive situation in more detail and (one hopes) to greater effect. If we take the line here advocated, Nowell-Smith's problem about real and unreal events should disappear. And the question about the status of the events which are historically authenticated gets an answer: they are conclusions forced on us when we think as we should and with all the resources we need.
I agree that this theory does not solve all the problems called up by a view like Goldstein's. Confined as it is to historical thinking, it can hardly deal with Nowell-Smith's point about the identity of the Franklin Roosevelt we knew and didn't much like and the Franklin Roosevelt we read about in our history books. It can give some account of the real Roosevelt as he appears in history, since it offers a general way of distinguishing what is real and what is merely fanciful in that sphere. To work this out it would be necessary to investigate the ontology of history as Goldstein does not: to show that there is more to history than events, or even events and actions, and that continuants of many different kinds come into it as well. One could hardly be expected to produce an account of this kind in the course of a comment like this. But even if it were available and the general theory were vindicated, that would still not meet Nowell-Smith's objection, which arises out of the sharp contrast Goldstein draws between historical knowing and other forms of knowledge, including direct perceptual acquaintance. Goldstein tries to combine an idealist theory of history with a realist theory of perception, and Nowell-Smith rightly calls attention to the difficulties which ensue. I have already argued that Goldstein cannot have it both ways with regard to the reality of the past, and pointed out his tendency to flirt with an altogether more comprehensive form of idealism. My own view is that he cannot stick to his main theory and meet Nowell-Smith's point without developing such an idealism, thus making the Roosevelt we knew as well as the Roosevelt we read about essentially a constituent of judgment. That he should be driven to such a course may serve only to convince some of his readers that his central thesis is radically defective. But if it is to be saved at all, I think it must be along these lines.
It remains to mention a problem for this type of view to which I can see no easy solution. Formulating his position at the beginning of his essay, Goldstein declares that his is "a belief in the primacy of knowing, not in the primacy of reals that are independent of knowing." What is real must be defined in an epistemological context, at least as far as history is concerned; established historical fact is what ensues from authentic and au-
thenticated historical judgment. I am not concerned now with the main realist objection to this point of view. Instead, I want to explore briefly the paradox which arises in this apparent attempt to equate the objective historical past with what the best-informed historical opinion takes to be established fact. Does not history progress? And if it progresses must that not mean that what at one time is taken to be fact will cease to be so thought of at another? As new evidence comes to Ught, or new general considerations are seen to be relevant, what was previously thought of as adequate historical judgment turns out to be mistaken. Must there not always be the possibility of mistakes of this kind? And if there is, will not the facts of history necessarily have a shifting character which seems on the face of things wholly out of keeping with what we expect of facts?
The theory of judgment I have outlined allows for a distinction between the objective past and the past as constituted in actual historical thought, since it defines the former in terms of judgment proper. If qualified historians agree, that creates a presumption that the past has been correctly constituted, but only a presumption. But to invoke this distinction may be thought to give at most a formal solution to our problem. For the truth, as Goldstein repeatedly points out, is that the only past historians know about for professional purposes is that which is constituted in their inquiries. The real past of the idealist is only marginally more attractive to the historian than the real past of the realist; because of his engagement in judgment he is not precluded from reaching it, and yet he can do so only through fallible processes of thought. If we ask what we know about what happened, we have to answer that we know what the community of historians says we know. We could protect ourselves from the revulsion this conjures up by saying that the term "knowledge" should be reserved for the products of ideal judgment, with the result that actual historians produce no more than a series of more or less well-supported beliefs. But if we did so we should certainly break with ordinary historical conviction as well as with common usage: is it really only a well-supported belief that George Washington was the first President of the United States?
The trouble is that, as Goldstein keeps emphasizing, the facts of history hang together in chains if not in a single chain, lending each other mutual support but also communicating one to another any weakness that they happen to possess. It is not possible to stick to Washington as first President without committing yourself to many other assertions and denials;
claiming that the first is a matter of knowledge means claiming that these others are too. Mostly historians can take this in their stride: the picture of the past they build up is not arrived at in jigsaw fashion piece by piece, but rather emerges into view as something of a concatenated whole. But of course there are lots of things about Washington and his times which we should like to know but do not; future historians may succeed in filling
some of these gaps. At this stage we cannot assess how seriously their discoveries will affect what we say we know now. If we allow even the possibility that new historical discoveries will result in situations presenting themselves to the public in a substantially different light, the problem of the provisional character of historical conclusions becomes acute.
One expedient for someone in this position would be to take refuge in a Moore-like faith in the powers of the professional historian, arguing that some bits of the past have been constituted beyond serious doubt and will stand as fact without regard to future historical discoveries. That George Washington was the first President is a piece of historical knowledge, accepted by the present community of historians as such and carrying such conviction that it must be seen as established for all time. No doubt many working historians would like to take this view. But Goldstein cannot and does not. Criticizing Mandelbaum he writes (HK, 46) that "points are never so settled statements so irrevocably established that they serve as criteria for the truth of all subsequent statements. In history, as in all spheres of inquiry, what is established is always subject to revision in the light of what may subsequently turn up." Historical conclusions are a function of historical evidence, and are hence liable to change as this is "enriched." As has been said before, they come to the test in clusters, not singly; atomism in history is definitely false. For someone who thinks on these lines the let-out sketched in the first lines of this paragraph is simply not available.
What then is to be done? Goldstein himself tends to speak of the discipline of history as an ongoing enterprise which can be presumed to be successful unless there is a reason to think otherwise, and which has the power to correct its own errors and so can be taken as advancing toward final historical truth (cf. HK, 90). But it is not clear that he is justified in using this language in view of his theory of historical constitution, which links historical actuality very closely with historical thought, duly authenticated by the community of historians. The facts of history, as he presents them, are the conclusions to which the best informed historical opinion is prepared to commit itself at any one time. If they are, we must be prepared for the possibility that some of them will be unstable; they may, though they need not, turn out to possess the shifting character we spoke of above. We need, of course, in considering this matter to distinguish carefully between propositions the community of historians is strongly inclined to accept and those which it regards as matters of knowledge; additions to the ranks of the former will not signify, nor will deletions, except when they are transferred to the other class. But insofar as new constitutions take place and are duly accepted as authentic (become part of what Goldstein refers to as "historians' tradition"), it will be necessary to ask how the overall picture of the past as currently understood has been affected. It
may just have been added to, but it could be also that it has been altered, the new discoveries excluding what was previously said to be tme. And if it has, historical knowledge will not have held steady: warranted facts will have shifted, to however slight a degree.
There is a strong repugnance on the part of philosophers to allow the possibility of shifting facts. In the common view facts are what true propositions state, and propositions are true or false without regard to whether anyone knows about them and indeed without regard to whether anyone thinks. In contrasting the events of the past with the account we construct of them now Nowell-Smith clearly works within these assumptions. Gold-stein with his "primacy of knowledge" approach is committed to a different conception of fact, one which makes fact dependent on judgment in a way no realist could accept. If he is to defend his position convincingly, I suggest that he must recognize its implications and embrace the paradoxes it involves. He will then be forced to admit that what is knowledge at one time may not be knowledge at another, without being able to gloss this in the usual way by saying that certain things were formerly taken to be true when all along they were not. Truth, fact, and knowledge will lose their total independence and become relative, in part at least, to the conditions of particular cognitive situations: if what is true for me remains true for you, unless you show that I am mistaken, that will be only so far as we share fundamental ways of thinking and recognize the same initial data. These ideas have only to be mentioned in orthodox philosophical circles to provoke opposition, even derision. But it is arguable that behind such reactions lies more dogma than considered opinion. The problem of cultural relativism already threatens the notion of a single unchanging truth, and though what we have here is clearly something different (since historians of different generations would normally be thought of as working within the same cultural tradition), once absolute facts go, the way is open to a radically new conception of truth and fact generally. I agree that no such view has so far been worked out with any seriousness. But I suggest that it might be, and that Goldstein's arguments point toward it even if Goldstein himself does not examine the question.
To repeat a point made earlier: Goldstein could avoid these difficulties if he were to follow the line indicated above and correlate objective fact with judgment as such rather than with any actual judgments. I think myself that this would be the proper course for someone in his position to take. But it is hardly likely that Goldstein would take it, partly because of his suspicion of "Kantian" approaches to history (see HK, ch. VI), partly because he will have no truck with a real past of any sort, such pasts being "factually vacuous." In history, Goldstein tells us (51), "there is no referent except as it is constituted or constructed by means of the technique of historical investigation": past facts exist in a context of thinking. It follows that they
must either be thought to be established once and for all, or be subject to change in the manner described above. Goldstein cannot accept the first of these alternatives without allowing that at least some historical facts are "loose and separate" and without abandoning his own opposition to atomism, nor does he attempt to do so. He must in consequence be driven back on some form of the view we have outlined, correlating fact with warranted assertions and recognizing that those who issue such warrants change their minds from time to time. Or if he does not accept this position, he must state and defend a third alternative. Jonathan German in his review of Historical Knowing described the book as lacking in philosophical refinement.8 To say that the book is philosophically weak is in one way absurd: the extent to which it has already attracted interest and provoked discussion suggests that the author has identified a previously unrecognized, or barely recognized, problem and has handled it in a penetrating and provocative way. Many of us would be delighted to have achieved as much in philosophy as that. But Goldstein could have done even better had he been prepared to consider his conclusions more broadly and to make explicit the overall theory which underlies them. It is in failing to attempt that task, if at all, that the book is philosophically defective.
University of Edinburgh