Chapter 7. Conclusion

1This book began with the introduction of a dispute between historians which demonstrated the necessity, for the practice of historical writing, of making a choice between different available approaches to history. If historians were to provide knowledge about reality, then a justification would be needed for their choice of approach. Thus the practice of history has to assume a solution to certain theoretical problems in the philosophy of historical knowledge. What, then, would be proper justifications for our exemplified approaches, for the traditional or for the cliometric approach to history? We needed, first, to understand how historical knowledge was in principle achievable before we could understand what the different approaches were and how they might be justified.

2Empiricism initially set us the standard for rational justification, but its own irrationality became apparent when it appeared that it set a standard which could not be achieved, since it could not be applied without further background assumptions about what experience is experience of. These background assumptions could not themselves be directly derived from experience, but were, if known at all, known a priori. Metaphysical assumptions about experienced reality thus have to be made. Historians, therefore, have to describe and explain the historical world, given a priori non-empirical assumptions as to what that world is. It is these assumptions which frame the questions historians ask, and frame their procedures, descriptions and explanations. A historical approach, embodying such assumptions, requires a rational defence if historical knowledge is to be constructed on the basis of it.

3In chapters 4 through 6 we analysed and compared the traditional and cliometric approaches in order to recover the fundamental assumptions each made about historical reality. On the traditional approach, we have a metaphysics which permits at the minimum two kinds of entities: those in the natural world which are causally related, and those in the human world which are individual persons. The natural and human worlds co-exist and interact. Each kind of entity has its associated theory of explanation: the natural world is to be understood in terms of covering law theory; and the world of individual persons must be approached in terms of empathetic understanding.

4While we were able to provide an analysis of empathetic understanding in terms of a perceived cost-benefit analysis, we saw that the economic approach embodies a further dimension: we need to be able to explain people’s actions, not necessarily in terms of the beliefs and desires we actually empathize with, but in terms of the beliefs, desires and costs we deem them to have in the light of an independent economic theory. This economic theory is used to specify an economic “reality” involving cost-benefit calculations of “interest” which are independent of the calculations of the individuals involved. Our choice of historical approach thus depends on the decisions we make about what historical reality consists in and how it is to be understood. Historical knowledge requires that our choice be justified.

5We may recall from the exposition of philosophical issues concerning knowledge and reality in chapters 2 and 3 that we have offered two standards for justification for knowledge claims: rationalism and empiricism. Our choice of standard must itself be justified, for the justification of knowledge requires the justification of the foundation adopted. But how can the standards themselves be justified? For rationalism, while arguing for the permission of a priori metaphysical assumptions, has provided no justification for them, while empiricism has been shown to involve metaphysical assumptions which require justification on non-empirical grounds. Whether we adopt empiricism or rationalism, justification of our metaphysical assumptions is required, and yet neither standard shows us how such a justification may be provided.

6It may then appear to be an arbitrary choice, or a mere matter of unfounded opinion, which metaphysical assumptions we make. If this is so, then it will also be an arbitrary choice how we interpret our evidence and write our narratives. If it is an arbitrary choice, then it may seem that, in the end, Catherine Morland is right: history is sheer invention, because we are apparently writing about arbitrarily invented entities.

7If, therefore, we write our historical accounts partly on the basis of, for example, the following central belief: “human beings are the kind of entities which must be empathetically understandable,” then, if we seek historical knowledge, we face the problem of justifying how that belief can be known to be true. In chapter 2 the grounds of truth were explained at length; at the most elementary, we saw that beliefs (statements) were either analytic (“true by virtue of meaning”) or synthetic (‘true by virtue of fact’). To say “human beings are the kind of entities which must be empathetically understandable” is not to say something analytic, for if it were, it would be true by definition and empirically empty; moreover, how would you be able to tell that a person was a human being? You could not tell just by looking at them, on this supposed definition, because you cannot tell just by looking at some entity whether or not it is empathetically understandable. Furthermore, to say that “human beings are the kind of entities which must be empathetically understandable” is not to say something synthetic and necessarily true, for it might well be false, since alternative metaphysical views are available to us. Is it, then, to say something synthetic and known a posteriori, known on the basis of simple experience? We have seen that it cannot be so; there is no single experience which could warrant this belief, no single fact in virtue of which it can be experienced to be true.

8Yet need “true by virtue of fact” require a single independently experienced fact to make a synthetic statement true? In chapter 2 we noted an idealist theory of truth. “True by virtue of fact,” for the idealist, cannot mean “true by virtue of independent fact,” and this has to be understood in some more complex way. We noted that the idealist often speaks of synthetic truth as being assured by the place of a synthetic statement in a vast web of beliefs, the acceptability of which depends upon its internal coherence, consistency and comprehensiveness, and the whole of which expresses (mind-dependent) reality. It is a so-called “coherence” theory of truth of this kind which will be outlined in this chapter as the best solution yet available to our epistemological problem.

9The truth of the claims we make about reality depends partly on the experiences we have, and partly on the background “a priori” assumptions which we bring to experience. We have many experiences and make many assumptions. In other words, the truth of the claims we make about reality depends partly on our many experiences and partly on the truth of many other beliefs. We cannot separate, within our total set of claims to knowledge, certain particular beliefs which meet experience, “copy” experience and are checkable by experience independently of the rest of what we believe. Strictly speaking, because of the presence of kind-words, no statement is “true by virtue of fact,” where “fact” is understood as an atomistic Humean simple impression. Even to describe something as “blue,” it will be recalled, is to liken it to other things not now present to us.

10The American pragmatist philosopher W. V. Quine suggests that, when we try to make sense of the experienced world, there is room for adjustment regarding which statements we propose to hold true and which we propose to discard as false.1 The following example is one way of clarifying this point. Imagine an ornithologist, firmly convinced “a priori” that all swans are white, who travels to Australia and discovers a black one. There is a clear conflict between his knowledge claim and his experience. But must he discard as false his belief that all swans are white? Not necessarily. Two inconsistent beliefs are in the forefront of his mind at the relevant point: “All swans are white” and “This is a black swan.” These beliefs are inconsistent with each other. One of them has to go. But neither logic nor experience tells us which it should be. He can keep the belief that he has before him a black swan, and discard the belief that all swans are white. Equally, he can keep the belief that all swans are white (treat it as “a priori”), and discard the belief that this is a black swan. (He can, if he wishes, discard both beliefs, but he has, so far as the example goes, no reason or need to do so.) Is whiteness essential to swans, or not? To hold that it is not permits the ornithologist to deny that all swans are white, but the removal of colour as part of the criterion for being a swan may lead to the removal of colour as an important part of his taxonomy of birds, which might have far-reaching effects for his understanding of the rest of biology. On the other hand, one way in which the ornithologist might justify to himself keeping the belief that all swans are white, and discarding the belief that the object before him is a black swan, would be by holding that, since this thing before him is black, it cannot be a swan (since all swans are white). It must, then, be some other kind of bird. To make this move would again require consequential changes to his taxonomy of birds, changes which might be extremely far-reaching. Whichever belief the ornithologist decides to discard, it is a matter for human decision and not of particular experience, and it is a decision which has consequences for the rest of his beliefs to a lesser or greater extent.

11This example is merely one instance of a general point. No one of our beliefs is so fully and directly related to experience that we are forced to keep or amend just that one if experience requires it. Equally, what we choose to treat as “a priori” is not a matter independently given to us. It is open to us to amend our knowledge claims as we find pragmatically convenient, and there are in principle many ways of effecting any required change. Quine notes, “any statement can be held true come what may, if we make drastic enough adjustments elsewhere in the system.”2 The history of science seems to show the wide range of possibilities which are open to us here.3 Quine’s position, while empiricist (for it is still experience which provides the ultimate ground for what we can rationally claim to know), understands experience as holistically rather than (with Hume) atomistically related to our beliefs: that is, all our beliefs meet all of experience as a body, and not on a one-to-one piecemeal basis. It is this holistic empiricism which we adopt in this chapter.

12All the assertions which we make are warrantable on the basis of their coherence with the rest of our claims to knowledge, together with experience. If a certain so-called “metaphysical” set of beliefs, together with the other beliefs which we have, fails to make sense of experience for us, if it leaves us with unexpected experiences or unsolved puzzles, then this set will need to be revised in favour of another. But the revision is neither total nor arbitrary; it is a matter of pragmatic convenience for us all. Thus the central belief that “human beings are the kind of entities which must be empathetically understandable” is warranted, if it is warranted, in virtue of the crucial place it appears to occupy within the entire set of our beliefs about all of reality in all of its guises, a set of beliefs warranted in so far as it makes sense of all experience for us. Yet we have seen that puzzles remain, puzzles which economic theory may have to be added to remove. Such an addition can push empathetic understanding from its original central place. Yet this is not easy: the coherence which we require between all our beliefs is a massive demand. The constraints upon our choice of metaphysics are many, for convenience requires that we save as much understanding as we can.

13We have noted that in principle there are many ways of effecting the changes which puzzling experiences may demand of our system of beliefs. Even so, some beliefs may not be possible to hold, given what else we want to believe. For example, it may be (it is a central philosophical question) that a belief in Hume’s and Hempel’s mechanism means that we cannot make sense of persons engaging in meaningful activities, and that pragmatically we prefer to jettison the belief in Hume’s and Hempel’s mechanism rather than the belief in the possibility of meaningful activities. In principle we could make the opposite choice. Yet we saw in chapter 4 that there can be moral constraints upon our choice of metaphysics, which may require us not to regard ourselves or others as mechanically determined to act. We are not to escape responsibility. On the traditional view, if we regard others or even ourselves as machine-like, then we are morally at fault. Value, therefore, can enter into choice of belief, can enter into metaphysical choice.4 Indeed, the requirement mentioned above, that we “save as much understanding as we can,” may itself be interpreted as an evaluative demand which places merit on conservatism and simplicity.5 Does the room for evaluation and the many ways in which even metaphysical beliefs may be changed mean that, ultimately, what we believe is arbitrary, and thus knowledge unfounded? Our final task in this book is to examine the place of value in our decision what to claim to know.

14There is no doubt that we do ordinarily make a distinction between factual matters and evaluative matters. The central question for us is that of knowledge, however. Why should the introduction of evaluative considerations preclude the possibility of a claim to knowledge being well-founded? The only plausible reason is that evaluative claims are themselves mere matters of opinion, which cannot be rationally founded. Is this true for moral and other evaluative judgements? With our conditional commitment to empiricism, we have only one answer to questions about how statements are to be rationally supported: such knowledge must be derived from experience.

15It is a famous observation of David Hume’s that we cannot derive an “ought” from an “is.”6 More generally, factual statements cannot be used to prove moral statements, unless we assume other moral statements also. This distinction between fact and value is widely accepted among scientists and others brought up in the Humean empiricist tradition. Milton Friedman, for example, when introducing the example of the leaves (used in chapter 5), insisted on distinguishing normative questions about what ought to be done from factual questions about what is the case.

16If knowledge must be derived from experience as Hume requires, then one quick answer to the problem of the epistemology of evaluation is to say that evaluative judgements are neither true nor false, and are thus not the kind of thing that can be known. This is a familiar empiricist route, and Hume sometimes spoke as if this were his view: “morality,” he said, “is more properly felt than judg’d of.”7 It is sometimes said that moral judgements are expressions of taste, of approval or disapproval. If it is rationally arbitrary judgements like these which importantly go towards determining our choice of metaphysical belief, then it might well be appropriate to hold that knowledge is not achievable.

17But Hume also states that the “ought” relation is an affirmation, and thus that we do judge of it. On this view, which is surely closer to common sense, some of our beliefs are moral beliefs; we can reason with them, they can be true, and, given empiricism, if we know them, then we know them on the basis of experience.

18If moral affirmations share, as for the empiricist they must, the same kind of empirical warrant as do non-moral or factual claims, then the epistemological contrast between descriptive and evaluative statements must, if it exists, derive from a contrast between different kinds of empirical sources. Hume could properly assume this because his was an atomistic empiricism. On this view, we must distinguish between external experiences (which warrant the factual beliefs) and internal moral experiences (which warrant the evaluative beliefs). This Hume did.

19But if atomistic empiricism is false and is best replaced by the holistic empiricism now described, then we cannot make this move. The situation is then this. Moral beliefs are either true or false. Our claims to knowledge include moral claims. Moral beliefs are supported epistemologically just as any other statement is supported epistemologically. Since the entire system of beliefs is supported by experience as a whole, then the moral beliefs are similarly supported by experience as a whole. But although the moral beliefs are supported by experience, they are supported in a holistic way, and not supported by atomic moral experiences which are given to us already labelled as such. Thus the fact/value distinction is not epistemologically fundamental.

20There may therefore be moral or other evaluative constraints upon our metaphysical theorizing or choices of factual description, constraints which raise no additional epistemological problems. We have already noted that the requirement that we “save as much understanding as we can” may be interpreted as an evaluative demand which places merit on the convenience of conservatism and simplicity, and thus affects our choice of theory. It is sometimes held that we have a duty to be conservative of the existing belief system. Quine himself, who did not venture into moral philosophy, merely observed, correctly, that we have a natural tendency to preserve the existing system, and we may further note that new information has a cost.

21We may conclude that these conservative tendencies involve a pragmatic commitment to the view that what requires justification is not so much belief as rather change in belief. Knowledge is therefore not, as so often said, “justified true belief,”8 for to suggest this would be to hold that what requires justification is belief itself, rather than just change in belief. By contrast, knowledge should be understood as consisting in any system of beliefs which is the consequence of a justifiable series of variations upon an earlier set of beliefs accepted as true. We are not to ask by what title truth was originally claimed. The justification of truth is thus partly historical.

22Thus we see that evaluative constraints like appeals to morality or simplicity or conservatism may properly exist in our theorizing. When we “deem” a person to have a certain net preference, and construct our historical approach to his actions on the basis of this, we are introducing evaluative considerations. We may defend our evaluation by pointing to the obligation or commitment which a member of some firm or institution has undertaken, or we may refer to desires and beliefs which we think that, as a “rational” being, he “ought” to have. We saw in chapter 1 how the historical assessment of slavery may be affected by evaluative considerations. This need not be simple irrational bias, however. The truth in morality, as within the rest of our beliefs, is warranted in a revisable fashion by the totality of experience together with the rest of what we believe.

23We have to achieve an equilibrium in moral judgment and factual judgment, not just within the realm of the moral or just within the realm of the factual, but across the entire system including both.9 Moral matters can thus rightly constrain our choice of historical approach, our metaphysics and our choice of factual description. Perhaps on moral grounds we should reject a mechanistic or deterministic conception of people. Value thus does enter into our decision about the proper metaphysics for history, about the correct theory of human nature, but value is ultimately as well-founded as any factual matter, and our decision is not a rationally unfounded opinion. Value provides an input into our total claim to knowledge, but not in a way which should cause us special doubt. The system of knowledge itself is historically warranted, and historical research supplies an essential and large-scale input into and constraint upon our overall understanding, constraining fact and value and metaphysics.

24In conclusion, is answering the original question “was slavery profitable?” a way of achieving knowledge, so that, once answered, the issue is disposed of once and for all? It was the disagreement between historians which first gave us reason to suppose that perhaps historical knowledge was not possible. Is disagreement, in principle, avoidable? It seems not. For knowledge, on the approach adopted here, consists in any system of beliefs that is the consequence of a justifiable series of variations upon an earlier set of beliefs accepted as true. Knowledge is thus relativized to the earlier beliefs, and the best we can do is improve on what we now believe. But “what we now believe” is not unambiguously identifiable; moreover, there is more than one way of making an improvement, and pragmatic convenience is the sole constraint upon it.

25The most that could be claimed (although it is not claimed here) is that the cliometric approach is the best we have, and the necessarily holistic defence of, or attack on, this claim has to be undertaken in part by economists and historians, who are in the empirical front line, and in part by philosophers, who are professionally committed to a synoptic view of the presuppositions of the many disciplines and sciences. Philosophical and historical understanding are entwined. We never have access to some touchstone which will assure us that we have now achieved truth, and need never change our beliefs again. Thus historical knowledge should not be understood as being inconsistent with the possibility of disagreement, for that would set an impossible and useless standard. The issue is never “disposed of, once and for all.”