A Treatise on Social Theory. Vol. I: The Methodology of Social Theory 
by W.G. Runciman.
Cambridge, 350 pp., £25, March 1983, 0 521 24906 6
Show More
Show More

By the time he was 34, Thomas Macaulay had had a fellowship at Trinity, practised law for a year or two, sat in the Commons for four, and been appointed to a seat on the Supreme Council in India. On the boat to Calcutta, he wrote to Ellis, he had read the Iliad and the Odyssey, Petrarch, Ariosto, Tasso, Don Quixote, Gibbon on Rome, Sismondi on France, Mill on India, ‘the seven thick folios of Biographica Brittanica’ and ‘the 70 volumes of Voltaire’. Once there, he took to ‘passing the three or four hours before breakfast in reading Greek and Latin’, started to reflect on politics at home, and decided to write his History of England. ‘For what is it,’ he asked, ‘that a man who might, if he chose, rise and lie down at his own hour, engage in any study, enjoy any amusement, visit any place, travel to foreign countries, consents to make himself as much a prisoner as if he were within the rules of the Fleet – to be tethered for 11 months of the year within a circle of half a mile round Charing Cross?’

Macaulay’s was a 19th-century choice, and his three-volume History was a 19th-century ambition. Synoptic sweeps are now out of fashion, and the professionalisation of letters, which has made them so, stops men moving, as Macaulay was to continue to do, between worlds. W.G. Runciman is an exception. By the time he himself was 34, he too had had a fellowship at Trinity, where he holds one now, had read very widely and written three books, joined the family business, which he now runs, and despite being thus tethered to within a mile of Charing Cross, had decided to write a Treatise. This volume, the first of three, is on method; the second will be on societies in general; the third, which Runciman promises for the end of the decade, on England since 1900.

Macaulay, back in London and contemplating his History, wrote to Napier in 1841 that ‘I really do not think that there is in our literature so great a void as that which I am trying to supply.’ Runciman might not unreasonably say the same. Only one other private scholar, Perry Anderson, has attempted anything similar. But Macaulay wished also to entertain. ‘I shall not be satisfied,’ he confessed to Napier, ‘unless I produce something which shall for a few days supersede the last fashionable novel on the tables of young ladies.’ Runciman is more restrained. He takes what he calls ‘an impartially benevolent view’. To describe the Tory members of 1730, for instance, as Macaulay did, as ‘little more than rows of ponderous foxhunters, fat with Staffordshire and Devonshire ale, men who drank to the King across the Water and believed all the fundholders were Jews,’ is, Runciman says, to lapse into ‘too much conscious literary merit to be counted even under the most generous rubric’ – and Runciman’s own rubric stretches to history itself – ‘as social science’. Not that Runciman is a Tory or a philistine. If he lacks Macaulay’s high spirits and the urge to move young ladies, or indeed anyone else, he has that faith in time and reason which Macaulay once confessed to James Mill. He shares the Whiggish benevolence that united Macaulay, the Evangelical, and Mill, the Utilitarian, more than their one dispute divided them. And in addition to also displaying a command of large parts of human time and space, Runciman approvingly reviews Marquez, commends Henry James, discusses some ‘splendid stuff’ from Tom Wolfe and starts a chapter with Angela Carter on rooms by the hour in Japan. If this is social science, it is in style and scope not social science as we have come to know it.

As we have come to know it, this science has, in its more principled moments, consisted either in attempts to show that accounts of people can look like accounts of particles, or in assertions of the truth of one or another anthropocentic cosmology. It has been an exhausting oscillation between anxiety and aggression. To the anxious, Runciman reasonably retorts that physics is a much less orderly and well-grounded affair than once it seemed. But to the morally aggressive, he simply says that the moment will come ‘when our understanding of human behaviour has so far advanced that ... values become as irrelevant to ... substantive conclusions as the doctrines of the Papacy became to the conclusions of rival astronomers about the number of Jupiter’s moons.’ There is a fine line between these two replies, a high wire to be walked. And he attempts it.

The required balance, he sees, lies in conceding that our accounts cannot be without presupposition, but in denying that they must thereby rule out others. We must concede that we cannot come to the world as pure visitors, without some interest in it, some frame through which to see it, but we can deny that in so doing we must pre-empt any alternative view. There no doubt is, as Quine insists, an irremediable ‘inscrutability’ in others’ reference, an ‘indeterminacy’ in translating what they say, but there is equally no doubt, as Quine also insists, that this is a point in principle which will not often trouble us in practice. In his famous example, an object moving against a contrasting background may be referred to by one man as a rabbit, by another as a gavagai, and although we can never in principle know that the one is referring to what the other is, and thus never know that ‘gavagai’ exactly translates as ‘rabbit’, we may ‘sensibly impose’ that he is and it does. Equally, the more inexplicably absurd or exotic a belief seems to be, the more suspicious we can be of its translation. Runciman is deliberately cavalier about Quine and all the more principled points. This is an essay on method, he says, and not a contribution to philosophy. But his purpose is the same: to rescue firm reference and sound explanation.

He does so by distinguishing three kinds of ‘understanding’. The first is reportage. There is no reason, he claims, to believe that reports cannot be given which, while necessarily presupposing some frame, use one which does not pre-empt any explanation or evaluation of what it reports and with which, therefore, ‘any rival observer will be bound to agree.’ Or at least, if in any such report ‘one differentiation is highlighted instead of another,’ he who makes it can be asked why, ‘and the answer may well be open to further question by those with other presuppositions and purposes in mind.’ And such questions, Runciman suggests, can be settled by privileging actors’ accounts of what they believe and do, for they too must be parties to the agreement. ‘Agents always know what they are doing, but not why, except where they are mistaken about something about their action which is constitutive of it.’ If a befeathered Mr Morley said he was doing a rain-dance in Bournemouth in the dry summer of 1976, and he did, then he was. If blacks say that the Stanford-Binet test does not measure their intelligence, it does not. But if I say I am taking a salute (as I was once persuaded to do in a village in India) and have neither rank nor troops, I am mistaken. And if actors’ accounts are not at issue, or if one wants some general category of one’s own to report instances which are themselves reported by a variety of actors in a variety of natural languages, that category may at a pinch be put into inverted commas, ‘thereby signalling to the reader that the author’s purpose is denotation alone and that the chosen name can innocuously be replaced by another.’ Bloch claimed that there was feudalism in Japan. But in that country, the chain of vassalage came to an end before reaching the throne, multiple fealties were unknown, there were no vassal courts, and there was a distinct awareness of the unity of the society quite absent in 11th-century Europe. In Japan, there was at best ‘feudalism’. These are striking arguments, but, thus qualified, correct. The question is: do we really want such neutral reports?

We clearly do if we want to move to Runciman’s second kind of understanding, to explanation. Reports which rule out the possibility, say, that blacks are not less bright, or that there was not feudalism in Japan, clearly rule out other ways of accounting for ethnic differences in Stanford-Binet scores or for the complicated pattern of relations between the centre and the provinces in the Tokugawa Shogunate. More exactly, we want reports which are neutral between explanatory theories if we hold to the view, as Runciman does, that reports and explanations are in principle and should in practice be distinct. And if we agree that they are, and should be, then we will be committed, as he also is, to the view that there is no special problem of explanation in social science. Explanations, on this view, consist in teasing out, by means of the ‘quasi-experimental’ methods of judicious comparison and controlled counterfactualising, causal sequences which are in turn plausibly grounded in some acceptable theory. We can be suspicious of explanations of Rome’s eventual inability to defend its frontiers in terms of its simultaneous inability to maintain its numbers of slaves, Runciman elaborates in a moment of characteristically casual erudition, because the otherwise very similar later Han Empire succumbed in exactly the same way in 311 without having had any slaves at all. We can be much more confident of an explanation which points to the logistic difficulty of maintaining troops and moving them at high speed across great distances to meet attacks which by the time the troops arrived might well have succeeded. Rome and the later Han had got to the stage at which with existing resources they just could not be defended.

But in the most original and arresting chapter of the Treatise, and, as he says, quite against what he has written elsewhere, Runciman readily agrees that such austere reports and explanations are not the only kind of understanding we may wish to have. We may want a third kind too, Angela Carter’s kind as she reflects on the initial mystery of primly crisp bedrooms by the hour in Japan:

Ah! Now I understand!
This is not an illicit bedroom at all.
It is a safety-net in which the death-
defying somersault of love may be performed
with absolute propriety.

We may want, in Macaulay’s phrase, to ‘exhibit such parts of the truth as most nearly to produce the effects of the whole’. We may want to know what it is like. We may want, as Runciman puts it, to describe. To do so, Runciman agrees, we have to elide the line between fact and explanation, for we do so by introducing a theory, a theory, say, of the importance of clearly-defined institutional provisions for everything in Japan, which meets the test of authenticity, of being at least prospectively acceptable to the agents themselves, to the Japanese, and of representativeness, of being prospectively acceptable to all kinds and conditions of Japanese: the one gives the ‘truth’, the other the ‘whole’. And although rival explanations can remain rival explanations, and rival evaluations remain rival evaluations, ‘no two authentic, well-grounded descriptions are ever incompatible. They can simply be added together to make a third.’

This is very striking indeed, and here, difficulties do arise. It is not so much that one might be thereby committed to going on and on. The skilled will not. It is not even that social science ceases to be distinct from other kinds of description. Novelists don’t have to be authentic, and no autobiographer can be expected to represent anything but himself. It is rather that descriptions may in themselves be severely indeterminate and yet incompatible with each other. If a description, as Runciman concedes, cannot be assessed as just true or false, as a report can, and if ‘good descriptions, no less than good explanations, have to be grounded in a theory, that is, some underlying body of ideas which furnishes a reason for both readers of them and rival observers of what they describe to accept them,’ then it cannot simply be the case, as Runciman also says it is, that ‘describing is a form of utterance marked off equally, on the one hand from saying what happened and why, and on the other, from making the claim that it was either a good thing or a bad one.’ At least, it cannot be the case unless ‘reasons to accept’ and explanations, answers to the question ‘why’, are kept quite distinct, and unless ‘reasons to accept’ carry no evaluation. Runciman suggests that the two are indeed distinct, and he repeatedly insists that neither requires any evaluation at all.

The issue thus turns on what is a reason to accept. Reasons are dreadfully indeterminate things, true or false of nothing but themselves, and so merely acceptable or unacceptable. Accordingly, if a good description has to give reasons for things being as they are, and if these reasons are to be authentic, then they must be reasons which the agents in question would themselves, at least on reflection, accept. But if we accept them too, as we must, since to reject them would be to reject the description, then we accept them because they are theirs. And to accept them because they are theirs is to give them not just cognitive privilege, whatever, quite, that might mean, but moral privilege too. It’s to grant them some legitimacy. Descriptions pre-empt explanations and would seem to pre-empt some evaluations too. Further, if explanations in Runciman’s sense rest on presumptions of motive, as he says they do, and if motives and reasons are not sharply distinct, as he also says, then our grounds for accepting an explanation are in part, and ineradicably, grounds of a similar kind. Finally, since in many societies, if not in most, what is an acceptable reason for some is not an acceptable reason for others, there will almost certainly be some equally ineradicable incompatibility between accounts. It seems too optimistic to suppose that two such descriptions ‘can simply be added together to make a third’.

Runciman is far too clever not to see all this coming. But his answers to it are not convincing. In the first place, he says, quite rightly, that although there may be a suggestion of approval, or not, in our reports, explanations and descriptions of whatever it is that we’re concerned with, such approval is not strictly implied. It would not be at all inconsistent of Macaulay to have said that tubby foxhunters were an altogether invaluable antidote to fundholders, or of Angela Carter to say that discrete sex in tight little bright little rooms produces some sort of stain on the moral fabric. But the fact remains, as Charles Taylor once put it, that there is a ‘value-slope’ in our accounts which, if left to itself, without countervailing argument, does tend to commend or condemn. Second, and much more interestingly, Runciman also supposes that ‘sociologists, anthropologists and historians of all schools do share some sort of tacit, if unspecific commitment to benevolence,’ do all ‘desire that other people’s desires should be fulfilled unless there is some reason against it,’ are all persons of ‘sympathy and detachment’, so that even if their accounts do tilt in this way, it does not in practice matter, and it certainly does not pre-empt some countervailing reason again being given. This, though, in turn supposes that such a seemingly innocuous ‘welfarism’, in Amartya Sen’s word, what Runciman more loosely calls a ‘utilitarianism’, does not itself pre-empt countervailing arguments. But it does. For unless an agent takes a wildly eccentric view of himself as a series of separate selves, each defined by a separate desire, he may well regard at least some of his desires as constitutive of the self he is. If then they are overridden, so is he.

The values of science are not just the values of science. Neutrality, even if possible, may not always be desirable. Benevolence, in Runciman’s sense, may seem to some to tip into indifference. Impartiality is not impartial. Nevertheless, it’s the partisan point that has been the more forcefully put in the literature of the past twenty years or so, and if Runciman’s high-wire act is at times so dazzling as to make one wonder whether there is a wire at all, that is exactly why this book is so good.

Gibbon, for whose ‘fine 18th-century disdain’ for obscurantists and bigots Runciman betrays a commendable taste, believed that his captaincy in the Hampshire Grenadiers had helped his history. Macaulay, despite his desire in 1835 to be away from Charing Cross, came to believe that being Secretary at War in Melbourne’s ministry had helped his. Runciman wishes that he could believe that he is ‘a better sociologist for being at the same time a practising capitalist’. There can be little doubt that he is. It’s therefore all the more tantalising to have to wait until 1990 and Volume Three for the first authentic but detached account of what capitalism, in England, is. Or was.

Send Letters To:

The Editor
London Review of Books,
28 Little Russell Street
London, WC1A 2HN


Please include name, address, and a telephone number.

Read anywhere with the London Review of Books app, available now from the App Store for Apple devices, Google Play for Android devices and Amazon for your Kindle Fire.

Sign up to our newsletter

For highlights from the latest issue, our archive and the blog, as well as news, events and exclusive promotions.

Newsletter Preferences