DRAFT Not to be quoted or referred to.

The History of Ideas as Philosophy and History

When the Isaiah Berlin Professorship in the History of Ideas was being established at Oxford one of the questions to come up for discussion was what should be the name of the post. Bernard Williams (of whom more later) apparently protested: You cant give it that name! No one wants to be called a historian of ideas. Williams view did not prevail but he had a point.1 There is no doubt that the history of ideas is disparaged on two sides nowadays not philosophical enough for the philosophers, not historical enough for the historians. Speaking personally, I feel honoured to carry a label that goes back through Berlin himself to Lovejoy and Carl Becker, but even those of us who do not reject the name must admit that the history of ideas is, at least, methodologically puzzling.2 Just what are ideas anyway? Anyone with a little philosophical or historical awareness knows that the use of idea as a generic way of referring to the contents of the mind is actually something quite historically specific. The seventeenth-century idea idea, as Quine called it, was part of a constructive intellectual project that developed in England as a philosophical parallel to the new physics, a project whose presuppositions analytical atomism, mental self-transparency seem nowadays (to say the least) deeply
1

This essay is based on the first of my lectures as Isaiah Berlin Professor of the History of Ideas, delivered at Oxford in January and February 2010. 2 Many years ago, when the prospects for academic employment were even worse than they are now (hard to believe, I know) I was interviewed on behalf of an American university by a well-known British philosopher. Although I didnt get the job I did hear something of what he had to say about me. Though I had, he thought, a number of positive qualities, I was, in the end, he regretted to say, not really a philosopher but a historian of ideas. This essay is based on the first of my lectures as Isaiah Berlin Professor of the History of Ideas, delivered at Oxford in January and February 2010.

problematic. Crossing the Channel to Germany, home of das historische Bewusstsein, things are a little different. In Germany, the word Idee retains its Platonic resonance. Thus, when it came to translating Locke into German, the word chosen for idea was Vorstellung a good choice in many ways, suggesting as it does whatever items are placed before the mind. The word was adopted by Kant and translated back into English (again, quite reasonably) as representation thus complicating even further the understanding of Kant in the English-speaking world. Hence the preferred term in Germany is Begriffsgeschichte conceptual history. Although the achievements of Begriffsgeschichte are impressive one thinks of the Archiv fr Begriffsgeschichte, Ritters Historisches Wrterbuch der Philosophie, Brunner, Konze and Kosellecks Geschichtliche Grundbegriffe, to say nothing of such towering individual figures as Hans Blumenberg the label itself is equally historical and, arguably, no less problematic. Concept (as used to refer to an item that structures our judgement) is another philosophical term of art in this case the result of the need for an English equivalent to the Kantian Begriff, itself part of Kants triadic sub-division of the mind and its representations into sense, understanding (the site of concepts) and reason.3 Indeed, Ian Hacking has argued in his wonderful little book, Why Does Language Matter to Philosophy?,4 that the transition from a discourse of ideas to one of concepts marks an epochal shift in our understanding of thought and language (a shift that, he thinks, was followed in the twentieth century by another, from concepts to sentences). Whether one is
3

It seems to have been common enough to describe thinking as conceiving and the products of thought as conceptions but the modern idea of concepts as the organizing principles of judgements appears to have been introduced into philosophical discourse by Sir William Hamilton and Coleridge and was popularized by William James 4 Ian Hacking, Why Does Language Matter to Philosophy? (Cambridge: Cambridge University Press, 1975)

convinced by Hackings claim or not (I am not) it is clear that any defence of the history of ideas must include methodological reflection on its subject-matter. That is one task that I propose to take up in this essay. In the course of it, I want to develop a conception of the history of ideas that is, I shall argue, both philosophical and genuinely historical. My views place me at odds, both theoretically and practically, with the justly celebrated historian of ideas who has done most to bring the methodological issues at stake in the discipline to reflective awareness I mean, of course, Quentin Skinner and I shall take the opportunity to articulate the differences between my views and his. Let me make it clear at the outset, however, that there is one point on which he and I are very much agreed. Although what I am advocating is a conception of the history of ideas that is intended to be both philosophical and historical, I do not mean to say that the historical and the philosophical impulses are always one and the same. It is a fundamental part of the activity of the historian that she brings to awareness the simple fact of difference between ourselves and the lives of those whom she studies. To insist on this and that this awareness has intrinsic value irrespective of any wider use we might make of our historical understanding has been, I believe, a driving force behind Skinners long and immensely productive scholarly career. And on this point he is, I think, absolutely right. The philosopher, on the other hand, wants to make progress with our (or, at least, her) problems. Whether or not (better: to what extent) those problems are shared with the authors of the past is an open question. Once again, I have absolutely no quarrel with this. However, I believe that there is an important region where the two impulses are complementary. The historical awareness of difference can help us in the

task of articulating and reflecting on our own concerns even where those concerns are not shared with the past and it is this view that I shall defend in this essay.

The History of Philosophy and the History of Ideas

I shall start by returning to the presumed contrast between philosophy and the history of ideas. It evidently raises one puzzle: where does that leave the history of philosophy? Which side of the line does it lie on? For some analytical philosophers the history of philosophy is not part of philosophy in the strict sense and the fact that historians of philosophy are to be found sharing space with philosophers in the academy is, in the end, an accident of institutional design and nomenclature. But, more interestingly, there are others who, while holding to the contrast between philosophy and the history of ideas, nevertheless place the history of philosophy within philosophy. Here, for example, is how Bernard Williams opens his admirable book on Descartes: This is a study in the history of philosophy rather than in the history of ideas. I use these labels to mark the distinction that the history of ideas is history before it is philosophy, while with the history of philosophy it is the other way round.5 What, then, is that distinction for Williams? For the history of ideas, he writes, the question about a work what does it mean? is centrally the question what did it mean?, and the pursuit of that question moves horizontally in time from the work, as well as backwards, to establish the expectations, conventions, familiarities, in terms of which the author could have succeeded in conveying a meaning. The history of philosophy, by contrast, Williams says ... has to constitute its object, the work, in genuinely historical terms, yet there is a cut-off point where authenticity is replaced by the aim of articulating

Bernard Williams, Descartes: the Project of Pure Enquiry (Harmondsworth: Penguin, 1978), p.9

philosophical ideas.6 Williams takes his own task to be, as he puts it, the rational reconstruction of Descartess thought, where the rationality of the construction is essentially and undisguisedly conceived in contemporary style.7 It seems, then, as though Williams envisages two, or, possibly, even three, kinds of meaning for historical texts. There is, first, the meaning that the text once had (the province of the historian of ideas), second, the meaning that it now has and, perhaps also beyond that (it isnt clear to me whether Williams thinks that this is the same thing as that second kind of meaning) the meaning that the text would have if it were rationally reconstructed, the latter two kinds of meaning being the concerns of the historian of philosophy. Is this plausible? It all depends (as philosophers still like to say) on what you mean by meaning, but on an obvious understanding I think not. Imagine that a seventeenth-century Frenchman were to say Paris is north of Lyons (or, more likely Paris se trouve au nord de Lyon). Would the meaning of that utterance have been different then from what it is in the early twentyfirst century? If knowing the meaning of a sentence is a matter of knowing its truthconditions, surely not. We know how to settle its truth-value in just the way that the original speaker would have done. To grasp the meaning of Paris se trouve au nord de Lyon we simply need to understand that it refers to two cities and that it is true if the former is indeed to the north of the latter. Obviously, when Descartes says sum res cogitans the sense and reference of the terms are far, far harder to establish than Paris is north of Lyons, but why should anyone think that just because of this complexity history (or rational reconstruction) will actually change the meaning of what he says?

6 7

Williams, Descartes, pp. 9-10 Williams, Descartes, p. 10

Williams gives an analogy which, or so he claims, will help us to grasp his point. The approach to texts taken by the history of ideas is, he writes, the equivalent of playing seventeenth-century scores on seventeenth-century instruments according to seventeenth-century practice8, while, in the case of the history of philosophy the musical analogy is, as an ideal, Stravinskys Pulcinella, in which the melodic line is Pergolesis, the harmony and orchestration Stravinskys.9 Yet this analogy strikes me as deeply unpersuasive indeed, strongly misleading. A musical work has its identity in its score (that is what makes it the work that it is) but an essential part of the value of a piece of music lies in the experience that we have of it as performed. For that reason one might perhaps say that the experience that we have of the work as performed constitutes the meaning of a piece of music. But this is obviously not meaning in the way in which we talk of the truth-conditions of sentences being determined by the meanings of the terms that they contain. We might say, in Williamss spirit, that the meaning (for us) of a Beethoven sonata is different if performed on a contemporary fortepiano or on a modern concert grand (and, indeed, that its meaning as performed on the fortepiano now is different for us than it would have been for Beethovens original audience even if it were performed on precisely the same instrument). But is it really plausible to say the same thing about, for example, Kants Transcendental Deduction? Like Williams, I think that rational reconstruction has a central role to play in our understanding of philosophical texts. But the account that Williams gives of what rational reconstruction amounts to is to my mind wholly unconvincing. To explain how it works, Williams gives us the Stravinsky/Pergolesi contrast. Now to my ears, at least,

8 9

Williams, Descartes, p. 9 Williams, Descartes, p. 10

Pulcinella is a work by Stravinsky, even if the melody is taken from elsewhere. An essential part of Pulcinellas identity as a piece of music is the parodic context with which Stravinsky surrounds the melody. Should we see the rational reconstruction of historical texts in the same way? Would Williamss book have been better titled: Rhapsody on a Theme by Descartes? Of course, when we interpret a philosophical text we are engaged in recovering its meaning but we must not be misled by that phrase into a misunderstanding of what that involves. Here is a very crude view of meaning. What we have in front of us is a text so many black marks on pieces of paper. Their meaning is what brings them to life for us just as when we see a drawing of a few lines and circles as a face or grasp a series of sounds as a melody. On this view, meaning is something ineliminably experiential it is something that you or I grasp and the question then naturally arises: is meaning-for-you the same as meaning-for-me or: is meaning-for-us-now the same as meaning-for-themthen? I have no doubt that Williams (one of the most acute philosophers of the twentieth century) would be appalled at having such a caricature view of meaning attributed to him. Nevertheless, his (as it seems to me) misguided analogy with the playing of music on authentic instruments encourages just this kind of subjective view and hence leads to the mistaken apparent alternative that we must either aim to recapture the meaning of a text as experienced at the time (the obviously impossible task of the historian ideas) or construct its meaning for us (at the price, he claims, of replacing authenticity with another, philosophical ideal) Against this, I want to insist that when we ask what is the meaning of the text? there is only a single question here: namely, how should we understand the text? But that

does not imply that we must be unhistorical and anachronistic in our account of it. On the contrary, there are ways to distinguish between an understanding on our part that is historically accurate and one that is not. But a historically accurate interpretation is not one that leads us to retune our instruments so as to re-experience the text as it would have been experienced by its original readers (or even its author): it is simply one that enables us to understand it correctly. I take my slogan from the eighteenth-century theologian and historian, Chladenius: An interpretation is, then, nothing other than teaching someone the concepts which are necessary to learn to understand or to fully understand a speech or a written work.

II

The Language-Chess Analogy

If it is misleading to represent the contrast between what a text once meant and what it means now as if it were a matter of the different subjective experiences of two different audiences, it might sound as if I am proposing instead that we should direct interpretation at a single, objective entity, the meaning tout court. Of course, the idea that behind the specific texts and utterances of language there hovers a kind of invisible entity that animates them is easy to lampoon. But there is another, more plausible way of thinking about meaning as subject-independent, a view that was, I think, a largely unquestioned orthodoxy in much of twentieth-century analytical philosophy. On this view, language constitutes a public domain that is conceptually prior to and independent of the state of mind of individual speakers not because it places a peculiar, Platonic object before our minds but because the use of language is a shared practice governed by public rules.10

10

Michael Dummett, indeed, identifies this understanding of language with analytical philosophy as such: Only with Frege was the proper object of philosophy established: namely, that the goal of philosophy is

This conception of language goes together with what I call the language-chess analogy.11 In a game of chess we can give an account of what is being done at any stage of a game (what move is being made) without giving an account of the particular state of mind of either the participant or the observer. This is because we know the rules of the game (and can presume that the players also know and are following the same rules). Moreover, we can make a sharp distinction between the move itself and what follows from it. A game of chess can be thought of as a huge, descending tree-structure of possibilities. To make a move is to select among one of those possibilities and so close off a whole ramifying set of branches (and twigs and twiglets). These we may think of as the consequences both immediate and increasingly remote of the move. Each move has a determinate set of such consequences, even though we all (even the most gifted players) have only limited insight into what those consequences are. To the extent that the same is true of language in general, an apparently very plausible view of interpretation presents itself. The interpretation of philosophical texts, we might say, should start with establishing their meaning in the way that the observer of the chess game might set out the sequence of the moves. Only after that had been done would one go on to such further questions as what the logical consequences of this or that
the analysis of the structure of thought; secondly, that the study of thought is to be sharply distinguished from the study of the psychological process of thinking; and, finally, that the only proper method for analysing thought consists in the analysis of language. Michael Dummett, Can Analytical Philosophy be Systematic, and Ought it to Be?, in Truth and other Enigmas (London: Duckworth, 1978), pp. 437-58, p. 458. 11 See my Hegel's Dialectic and its Criticism (Cambridge: Cambridge University Press, 1982), Ch. 1. By goes together with I mean encourages and is encouraged by. Logically, the language-chess analogy entails the view of language as public and prior to individual speakers but the entailment does not go in both directions. Indeed, the view of language as prior to individual speakers that dominated in Frenchlanguage philosophy (based on the Saussure-inspired distinction between langue and parole) differs in some important respects from the language-chess analogy. In particular, the Saussurean vision (particular in the hands of many of its later advocates) is holistic, where the language-chess analogy is not.

utterance might be (the equivalent to the analysis of a position in chess) or ask why the author made the utterance that she or he did (the question of why the player made this or that move). In this way we would also have a plausible division of labour between the philosopher (who looks at the text sub specie aeternitas) and the historian (whose task it is to connect the text to the agent and, through that to its wider social context). Each presupposes a common requirement: the establishment of basic meaning. So it is important to spell out the reasons why the language-chess analogy does not do justice to what is at stake in the interpretation of philosophical texts. (1) The first reason is that philosophy does not have a fixed and agreed set of rules to

oversee its practice in the way that chess does. On the contrary, philosophy is essentially a reflexive discipline that is, its method is part of its subject-matter and (as Stanley Cavell puts it) there is no useful distinction to be made between philosophy and metaphilosophy. Philosophers do not just dispute this or that substantive issue but, more fundamentally, how to address philosophical questions at all. They divide radically, for instance, about whether there is a mode of discourse that is distinctive to philosophy or whether philosophy simply uses forms and procedures that are equally at home in other domains.12 So it isnt just a matter of picking out the arguments; it is also a matter of identifying what the structure of argument is supposed to be. (2) Secondly, we typically think that the terms of our language are set independent of

and prior to individual speakers it is not you or me who decided that Paris is Paris or Lyons Lyon and the language-chess analogy re-inforces that understanding. But
12

The idea that philosophy has a distinctive method that separates it from the discourse of the sciences is characteristic of much so-called Continental philosophy. Hegel, Husserl and Heidegger are all obvious examples. But it is also an issue within analytical philosophy (one way of reading the philosophy of Wittgenstein is as a continuing protest against the extension into philosophy of methods appropriate in other spheres of intellectual activity).

10

philosophical terms are very often terms of art, coined by and defined with greater or lesser degrees of explicitness as the case may be, by particular speakers. This is obviously so in some cases (think of Kants use of the term transcendental) but may be true too when the word is not originated by a particular author. So, although we may ask What does justice mean?, we also need to ask What does Rawls mean by justice? And this latter is not a psychological question, a question about Rawlss state of mind, but a question about his textual practice: how do we find it used in his texts? (3) Another striking contrast between the language of philosophical texts and chess

concerns the extent to which we grasp the meaning of philosophical terms. In chess, we know the powers of each piece (anyone who doesnt know that the king can castle simply doesnt know to play the game). On the other hand, the truth (if it is one) that justice requires equality is a conceptual one, part of the meaning of the term. Yet such conceptual truths (the meanings of philosophical terms) are not fully available to all competent speakers in the way that every competent chess player grasps the rules governing the pieces they are a matter of the deepest controversy in a way that the powers of a chess piece are not.13 (4) A related important feature of philosophical language is that it is characteristically

textual in the following sense. For simple empirical terms, our understanding is a mixture of linguistic knowledge (knowing, for example, of an adjective that it can be used to characterize a substantive within a judgement, that kind of thing) and perceptual

13

You might think that there is a parallel in the case of chess. There are certain things that we need to know to make a move (the basic powers of the piece) and certain things that are true a priori which we may or may not know (that moving that piece will lead to checkmate in six moves). But the parallel here is a false one, for the conceptual truths of philosophy are not the kind of indirect consequences that players foresee to a greater or lesser degree in chess. If they are true they license inferences now, at once the question is whether they are true or not.

11

recognition being able to pick out the things that fall under the term (picking out red things as red and so on). Something similar is, I think, true of philosophical concepts. But here the recognitional element is not a matter of picking out objects by perception so much as identifying the beliefs that instantiate the concepts, and those beliefs are, in turn, governing patterns which govern the structure of texts. So someone who believes in legal positivism will show that in the things that she says about law that law and morality are different things, for example; that we can identify the law without using normative concepts, and so on. Perhaps she will say explicitly that she is a positivist; perhaps not, and it will be up to the interpreter to show the presence of that belief by the pattern that is to be found in what she says elsewhere. In any case, it is the pattern of the text that is the final arbiter.14 (5) So far, I havent questioned an apparent parallel between chess and philosophical

texts. Just as a game of chess is an ordered sequence of individual moves made according to rules, so a philosophical text, we might think, is (at least roughly) an ordered sequence of sentences operating according to (more or less explicit and more or less agreed) rules as steps in an argument. Of course, some bits of philosophy look more like sequential arguments than others (Aquinass Five Ways of Proving That God Exists more so than Heideggers Being and Time, to state the obvious) but, even where philosophy aspires to the coercive condition of a deductive argument, there is an important difference

14

Putting these two ideas together, we can see how close the business of textual interpretation and the substantive questions of philosophy can come. Take a question like: What is transcendental idealism? It would be absurd to think that we could answer such a question without looking at what Kant means by transcendental idealism. But we want to know too: What does transcendental idealism entail? Is it, for example (as Kant claims) equivalent to empirical realism. And at this point the question is not (just): is transcendental idealism (as Kant construes it) equivalent to empirical realism in Kants sense. It is a question about transcendental idealism and empirical realism as such perhaps what Kant means by empirical realism is inadequate, for example.

12

between philosophical argument and the rigorous, step-by-step procedure of a chess game. Philosophical arguments are inevitably, even in those philosophers who aim for the greatest degree of explicitness to a high degree enthymematic: that is, they must rely on premises that are not explicitly stated. It is this fact that requires us to practise (as Williams and I agree) rational reconstruction. But rational reconstruction here does not involve treating the text as a musical score and adding to it an anachronistic orchestration, as Williams analogy suggested. A better analogy would be to see the text as a machine or a computer program and the interpretation that we give of it a way of opening up the text and showing the mechanism by which it works (or why it fails to). What premises need to be added to make what seems to be an invalid argument valid? Can we reasonably ascribe these beliefs to the author? As interpreters, we are engaged in reverse engineering our conception of what it would take for the text to work may or may not correspond to the authors (and often it is extremely hard to tell whether it does or not). We cant, however, simply ascribe whatever beliefs would be necessary to make the argument valid even the greatest philosophers make mistakes!15 Rather than pursuing this further in the abstract, I want to illustrate what I have in mind by a very familiar example.

15

Beliefs, then, in my view, are not entities that could be discovered if only we could see into the minds depths so much as entities that we ascribe to agents as ways of making intelligible patterns that we find in their actions. As Peirce puts it, for example, [The] essence of belief is the establishment of a habit; and different beliefs are distinguished by the different modes of action to which they give rise. How to Make Our Ideas Clear, Works, V, p.398 If we understand here that action in the case where the agent is an author involves involves the production of texts, then it seems to me that the parallel works very well. Just as our habit of keeping our hand away from the gas flame shows that we have certain beliefs about heat and combustion, so a philosopher could show (say) that she believes in the distinction between facts and values by the way that she organizes her texts.

13

The fourth chapter of John Stuart Mills Utilitarianism has as its title Of What Sort of Proof the Principle of Utility is Susceptible and it contains the following famous (indeed, notorious) passage: The only proof capable of being given that an object is visible, is that people actually see it. The only proof that a sound is audible, is that people hear it: and so of the other sources of our experience. In like manner, I apprehend, the sole evidence it is possible to produce that anything is desirable, is that people do actually desire it.... No reason can be given why the general happiness is desirable, except that each person, so far as he believes it to be attainable, desires his own happiness. This, however, being a fact, we have not only all the proof which the case admits of, but all which it is possible to require, that happiness is a good.16 How to read this passage? As every undergraduate who has completed the first year of Oxford PPE knows, on the face of it, it contains (to borrow a phrase) a non sequitur of numbing grossness. Yes, indeed, the fact that something is seen is evidence sufficient to establish that something is visible in the sense that that thing can be seen. But to be desirable in the way that something that is ethically good is desirable is to say that that thing ought to be desired. And this is not the same as saying merely that it is capable of being desired. As G.E. Moore put the criticism in another almost equally celebrated passage: Mill has made as nave and artless a use of the naturalistic fallacy as anybody could desire. This fallacy, he continues, is so obvious, that it is quite wonderful how Mill failed to see it.17 Is there a way to avoid the fallacy? Consider now the following argument: (1) Whatever is desirable (in the sense that it ought) to be desired is desirable

in the sense that it can be desired (2)

16 17

Whatever can be desired is desired

[[ref.]] G.E. Moore, Principia Ethica, Sect. 40

14

(3)

Only happiness is desired

Therefore: (4) Only happiness is desirable in the sense that it ought to be desired (strictly:

if anything ought to be desired, happiness is the only thing it can be). Here then we have what we might call a rational reconstruction of Mills argument. Is it a good argument? Well, it appears to me to be valid and at least it doesnt rest on the simple confusion between something that is capable of being desired and something that ought to be desired that Moore takes to be its obvious reading. The soundness of the argument, however, must depend on the plausibility of the premises. Yet is it Mills argument? This is more difficult to decide. Certainly, Mill spends a good deal of effort in Utilitarianism arguing for what looks very much like proposition (3). And a way of reading the passage about the evidence for visibility consisting in being seen might be that we should take it as intended to be an argument in support of (2) (roughly speaking, it would say to the reader: we are looking for the class of visible things; perhaps there are visible things that are not seen, but the only evidence we have for what things are visible is what we see; the best inference from the evidence that we have is that the class of visible things and the class of seen things is co-extensive). On the other hand, there seems to be nothing in Mills text that looks like an argument for (1). But might we not attribute it to Mill as something that he takes for granted, a belief that would complete an enthymematic argument? The reconstruction is, you might say, a friendly amendment: if it is not exactly Mills argument as he intended it to work, at least it gives his position a better chance of being accepted than Moores allegedly obvious confusion of meaning.

15

So, whether or not it is in any psychological sense Mills argument, he ought, you might think, to welcome it. Why stop there? A large number of works written in English over the last halfcentury interpret past philosophers by reconstructing their arguments, not by adding premises that it would be plausible to attribute to them but by removing ones that the commentator believes to be dubious and substituting ones that he or she thinks more plausible. P.F Strawsons The Bounds of Sense (in many ways the inspiration for this style of history of philosophy) for example, gives an interpretation of the Critique of Pure Reason that involves both addition and subtraction.18 Kants arguments in the Transcendental Analytic (particularly the Transcendental Deduction and Refutation of Idealism) can be defended, Strawson claims, provided that we purge them of their association with the imaginary discipline of transcendental psychology19 and replace it with arguments that draw on the idea of a necessary connection between basic concepts. Certainly, it can be illuminating to discover of an argument, not just why it works but, equally, why it doesnt. Yet the objections are obvious. Above all, how confident can we be that the interpreters judgement of what are good beliefs to hold are better than the original authors? Strawsons interpretation is a case in point. Even conceding that there are deep problems with Kants theory of synthesis, it is hardly beyond dispute that Strawsons alternative is superior. How are concepts connected except by their meanings and dont those yield analytical truths, not the kind of synthetic a priori propositions that Strawson (and Kant) want? A further consequence of this reconstructive approach is that, as we move away from the texts as originally presented, we lose context the gain

18 19

P.F Strawson, The Bounds of Sense (London: Methuen, 1966) [[??]]

16

in current relevance of seeing an author of the past purged of elements that the modern interpreter takes to be indefensible risks disconnecting us from the way that the texts were read and understood by the authors contemporaries. If you think, like Allen Wood, for example, that [noumenal freedom] should be ... quarantined from Kantian ethics just as strictly as if it carried the plague20 then you will probably find it difficult to understand why the theme of noumenal freedom should have been so centrally important for Kants German Idealist successors. But this, of course, is to make a historical objection which will not necessarily have force with all interpreters. Derek Parfit is supposed to have said that there are two types of readers for the philosophical texts of the past: archaeologists and grave-robbers. While the archaeologist will worry whether it is appropriate to attribute the desired belief to the author under examination, the determined grave-robber will ascribe whatever she thinks works to the argument and throw away what doesnt, regardless of historical context. If I am to make my case, I need to show that something like archaeology itself has a philosophical contribution to make.

III

Philosophical Problems

At this point, I want to re-visit an assumption of the last section: that our interest in studying the texts of the past is to reconstruct the arguments that they contain (understanding by argument a sequence of deductively valid steps from premises to conclusion). That we can do so for some of the central texts of the past is what I tried to show with my example of Mills proof (and I think that there are many other possible examples). Yet not everything that we find in the canon of past philosophy fits so easily
20

Kantian Ethics (Cambridge: Cambridge U.P., 2008), p. 138

17

into this model. Either the philosophers in question do not deal in arguments at all (Heidegger, Wittgenstein?, Nietzsche?) or, if they do, what they consider to be argument seems so remote from the mundane idea of argument as a sequence of deductive steps that it is hard to see how the gap between the two conceptions can be bridged, let alone adjudicated (Hegel). This picture of competing but quite heterogeneous philosophical discourses becomes less so, however, if we think of the object of our interpretation not as the reconstruction of arguments so much as the identification of philosophical problems. Philosophical problems, I am claiming, emerge when a number of commitments call them background beliefs, intuitions, received common sense, ways of seeing the world, or whatever that are individually felt to be compelling lead, when brought together, to apparent conflict. That philosophy has been centrally concerned with the response to such conflicts since at least the time of Plato seems to me to be beyond dispute. The different strategies of response to give up one or another belief; to claim that the beliefs are ambiguous and, when properly re-interpreted, may be seen to be consistent; to argue that the conflict is a product of further, unstated but questionable, premises; to claim, indeed, that we may just have to live with the conflict are legion and form the central core of philosophical debate in many of the subjects fields. To shift the focus onto problems in this way is not to abandon the connection between philosophy and argument altogether. On the contrary, the dilemmas the apparent conflicts that are central to philosophical problems on this understanding are highly susceptible to being presented as arguments: it is precisely because we feel the force of such arguments that we see them as identifying philosophical problems. Yet the

18

response to philsophical problems need not be another argument: it may be, but, then again, it may not. One way of reading Kants Copernican revolution is to interpret him as resolving an inconsistency by changing one of the (hitherto unchallenged) premises in order to open the possibility what he believes will be a successful argument against scepticism about the existence of the external world where previous ones had failed.21 Yet that is not the only kind of response to philosophical dilemmas. One that is particularly characteristic of post-Kantian Continental philosophy depends on the thought that the apparent conflicts characteristic of philosophical problems are the result of a way of thinking that is unreasonably and inappropriately limited. Thus, for Hegel, it is necessary to move beyond Vorstellung (representation) and the understanding to a speculative form of discourse, at which level the apparently compelling conflicts will fall away.22 I want now to give an example of how the rational reconstruction of

philosophical problems can illuminate our understanding of the texts of the past. The eighteenth century saw an ongoing dispute amongst those who studied (what we now call) organic nature regarding the nature of organic growth the ways in which it comes about that the seed transforms itself into the plant, the embryo into the adult animal between those who were called at the time evolutionists and epigenetists. The problem, it seems to me, was not merely empirical but conceptual: if growth was to be

21

Hitherto it has been assumed that all our knowledge must conform to objects. But all attempts to extend our knowledge of objects by establishing something in regard to them a priori by means of concepts have... ended in failure. ...We must therefore make trial whether we may not have more success in the tasks of metaphysics, if we suppose that objects must conform to our knowledge... We should then be proceeding precisely on the lines of Copernicus' primary hypothesis. (Critique of Pure Reason, B xvi,) 22 It is in this spirit, I think, that we should read Heideggers famous rejection of Kants attempt to give a proof of the existence of the external world. The scandal of philosophy is not that this proof [of the existence of the external world] has yet to be given, but that such proofs are expected and attempted again and again. Being and Time, (Oxford: Basil Blackwell, 1967) p. 249

19

understood as development, then it contained an apparently insoluble antinomy. I shall call this dilemma the problem of development. Somewhat oversimplified, it can be presented as follows: (1) (2) (3a) (3b) If a final state (FS) is a development from an initial state (IS) then there will be some features of FS by which it differs from IS. But then what relationship do those aspects of FS bear to IS? Are these apparently novel elements to be held to be somehow already present, implicitly, but not apparently, in IS? If so, then they are not really something new. On the other hand, if the novel features of FS are not to be found already in IS, then they must be something that is in relation to that antecedent state arbitrary, and, hence, unexplained.

Thus, (4) it seems that either FS is inexplicable in relation to IS or else it is not really a development from IS at all.

Whose argument is being reconstructed here, you may ask? While the account of Mills proof could be checked against the way that the text of Utilitarianism works, whose text is this supposed to explain? Perhaps there may be somewhere in the mass of eighteenthcentury texts on embryology and growth Stahl, Haller, Blumenbach, Maupertuis, and so on someone who puts the issue in terms that are recognizably similar. But even if there is not, my claim is that this is a useful representation of an issue that the thinkers of that era faced. In particular, it helps to explain why biology was felt by thinkers of the late eighteenth century to be both deeply problematic and, at the same time, philosophically instructive. Kant, of course, believed that the problems represented by biological growth and development were objectively insoluble: there could never be a Newton of a blade of grass and we must settle for a subjective, not objective, explanation of the indubitable

20

fact of organic nature. On the other hand, the Idealist philosophers of nature Schelling and Hegel, in particular, but also their precursors such as Herder and Goethe wrote about organic development in a peculiarly fervent and excited way. Here, for example, is a passage from Herders Ideas for the Philosophy of the History of Mankind: Bearing in mind these transformations, these living operations in the egg of the bird or in the womb of the mammal, I feel we speak imprecisely if we talk of seeds that are merely evolving, or of an epigenesis by which the members are superadded externally. It is Bildung (genesis), an effect of growing, inward Krfte, brought together in a mass by Nature in order that they might manifest themselves.23 For Herder, as for Schelling and for Hegel (or so I claim) nothing less than a new metaphysics a new conception of the ultimate nature of reality and a radically new conception of philosophical explanation could do justice to the requirement of explaining organic development without falling prey to one or other horn of the dilemma. And this is a philosophical requirement that seemed, I believe, as obvious to the thinkers of the time as it may seem strange and misguided to us now. My aim in the reconstruction of what I have called the problem of development is not to defend the course that they took, but to show its reasonableness in the intellectual context in which they found themselves to help us to enter their world. The question in this case is not: is this a belief that it is right to attribute to the author as an unstated part of an otherwise incomplete argument?, so much as: is this a problem to which it is reasonable to see the text as responding? Such attempts at rational reconstruction are, of course, more

23

Siehet man diese Wandlungen, diese lebendigen Wirkungen sowohl im Ei des Vogels als im Mutterleibe des Tiers, das Lebendige gebret, so, dnkt mich, spricht man uneigentlich, wenn man von Keimen, die nur entwickelt wrden, oder von einer Epigenesis redet, nach der die Glieder von auen zuwchsen. Bildung (genesis) ist's, eine Wirkung innerer Krfte, denen die Natur eine Masse vorbereitet hatte, die sie sich zubilden, in der sie sich sichtbar machen sollten. Ideen zur Philosophie der Geschichte der Menschheit, Werke, 6 (Frankfurt: Deutscher Klassiker Verlag, 1989) Bk V, Ch. 2, p. 172. 21

ambitious and harder to justify than more conventional and limited reconstructions. They take us beyond the following through of the consequences of an agreed set of rules to which we might restrict ourselves if we took the language-chess analogy too much to heart. What governs and what sets limits on what we may ascribe to the authors we interpret? But rational reconstructions of underlying problems are not, I think, illegitimate in principle and they are, I believe, of historical value in illuminating what is distinctive about the discourse of the past.

IV

Skinner: the Principle of Attribution

So far I have talked about the interpretation of historical texts principally from a philosophers perspective. I now want to say some things about the same subject as seen by the historian of ideas. In particular, I want to take issue with some claims made by Quentin Skinner in his methodological writings, in particular in his now-classic essay, Meaning and Understanding in the History of Ideas.24 In that essay Skinner sets out to demolish an antithesis which has been prevalent in theoretical thinking about the history of ideas between interpretation which is directed towards the text itself and that which seeks to understand texts by placing them in context. The main thrust of his argument may usefully be divided into two. There is, first, the claim that beliefs may not properly be ascribed to an author except in so far as they correspond to the authors intentions (I shall call this the principle of attribution) and, second, a positive account, based on the Austinian theory of speech-acts, which

24

History and Theory, 8 (1969), 3--53.

22

Skinner proposes regarding the nature of such intentions. I have reservations about both aspects of his position. The principle of attribution prohibits us from identifying the doctrines contained in a text in ways that conflict with how the author himself might have understood them -and, a fortiori, from taking a perspective which represents texts from quite different contexts and epochs in terms of the development of a single doctrine.25 To do so, Skinner argues, would be to conflate an account which might be true of the historical significance of the works... with an account of what they were doing which could not in principle be true (p. 23). He continues: The danger here is not merely that of seeing far too readily the modern elements which the commentator has thus programmed himself to find; there is also the danger that such interpretations may part company with anything that could in principle be a plausible account of what Machievellis political writings were meant to achieve or intended to mean....The surest symptom, in short, of this mythology of prolepsis is that the discussions which it governs are open to the crudest type of criticism that can be levelled against any teleological explanation: the action has to await the future to await its meaning. (p. 24)

25

In the original version of the essay, Skinner takes a position which is clearly opposed to any attempt at all to look at doctrines from a trans-historical perspective. As he writes: The characteristic point of departure in such histories is to set out an ideal type of the given doctrine -- whether it is the doctrine of equality, progress, Machievellism, the social contract, the great chain of being, the separation of powers, and so on. The particular danger with this approach is that the doctrine so readily becomes hypostasized into an entity... The fact that ideas presuppose agents is very readily discounted, as the ideas get up and do battle on their own behalf. (Meaning and Understanding, pp. 10-11) Skinner now concedes he may have expressed himself incautiously here: My way of putting the point appeared to deny the obvious fact that western traditions of philosophy have contained long continuities, and that these have been reflected in the stable employment of a number of key concepts and modes of argument; an error the more regrettable since it created the impression that my precepts and my own historical practice are out of line with each other (Meaning and Context, p.283). (One of Skinners essays on Machievelli carries, for example, the apparently proleptic title: The Idea of Negative Liberty: Philosophical and Historical Perspectives (in R.Rorty, J.Schneewind, Q. Skinner (eds.) Philosophy in History (Cambridge: Cambridge U.P., 1984) pp. 193-224)). Nevertheless, one may wonder whether this is good enough. Does his concession not raise an acute problem of consistency for Skinner: it is not clear how a trans-historical perspective could be reconciled with the principle of attribution, which he continues to endorse?

23

It may seem that Skinners criticism is making somewhat elastic use of terms. Perhaps it is true that an interpretation of Machievelli from a modern perspective cannot in principle be a plausible account of what Machievellis political writings were meant to achieve or intended to mean. But that is to beg the question: why must any interpretation be an account of what Machievellis writings were meant to achieve or intended to mean? Why not just describe what they in fact achieved or meant? But the principle of attribution, if granted, does indeed disallow anachronistic interpretation. Skinner describes the principle as follows: The relevant logical consideration is that no agent can eventually be said to have meant or done something which he could never be brought to accept as a correct description of what he had meant or done. This special authority of an agent over his intentions does not exclude, of course, the possibility that an observer might be in a position to give a fuller or more convincing account of the agents behaviour than he could give himself... But it does exclude the possibility that an acceptable account of an agents behaviour could ever survive the demonstration that it was itself dependent on the use of criteria of description and classification not available to the agent himself. For if a given statement or other action has been performed by an agent at will, and has a meaning for him, it follows that any plausible account of what the agent meant must necessarily fall under, and make use of, the range of description which the agent himself could at least in principle have applied to describe and clarify what he was doing. Otherwise the resulting account, however compelling, cannot be an account of his statement or action. (pp. 28--29) The suggestion that any account of the meaning of a text should be directed towards the authors intentions (what it was that he was trying to do in saying what he said) and hence that it must restrict itself to a range of description which the agent himself could at least in principle have applied to describe and clarify what he was doing may appear evident at first sight. But it is, I suggest, less plausible than it seems. When we ask: What was the meaning of that statement?, it may be because we are puzzled about the intentions behind it (What do you mean `Itll end in tears?) but it

24

may be because we want to know about the concepts involved and their implications. This is not to say that no questions of intention arise. Philosophical texts are extremely complex -- terms are often left undefined (or defined in more than one way) and arguments are presented without the premises which would make them persuasive. Frequently, too, there is the question of whether a particular view is being advocated or whether it is merely being presented hypothetically -- in order, perhaps, to illustrate its weaknesses. For all of these reasons it is quite right of Skinner to deny that the text itself forms the self-sufficient object of inquiry and understanding (p. 4, my emphasis): many hypotheses are necessary which go beyond what one would call narrowly the text itself.26 Texts are deliberately produced artifacts and any theory of interpretation which neglects that fact or treats it as being of minor significance is bound to go wrong. But it does not follow from this that interpretation must match the authors intentions in the sense of directing itself towards what the author may be said to have meant by saying what he said (p. 31) and characterizing it in terms which the agent himself could at least in principle have applied.27 The suggestion that we can describe a pattern accurately only by using an agents own terminology -- his own criteria of description
26

Even this is too simple. Perhaps there are philosophical texts in which it would be wrong to imagine that we could ever properly say whether a particular viewpoint was being advocated, attacked or merely examined -- and not just because we are not in a position to tell: that it was actually important for the author of the text to refuse to allow himself to be determined as doing any one to the exclusion of the others (readers of Wittgenstein or Kierkegaard will appreciate this suspicion). 27 One might think that this quotation contains a way in which my position and Skinners could be reconciled: why not say (in concession to Skinner) that an author could be said to mean what an apparently anachronistic interpretation would have her mean because she might in principle have been brought to accept a characterisation, given in our terms, of the sort that I am now advocating? The trouble is that the notion of attributability in principle (like the logical positivists verifiability in principle) has here become impossibly vague and elastic. Since we can, without further external evidence, be brought to accept the truths of mathematics (this, as Michael Dummett has pointed out, is a plausible account of just what we mean by the idea of mathematical reasoning) then it follows by parity of reasoning -- absurdly -- that we could already be said to mean all the mathematical truths that are in principle available to us.

25

and classification, as Skinner has it -- appears to presume that what we are after is some kind of reconstruction of the contents of the authors mind. Yet to perform an action consciously and deliberately is not the same as to perform it with a fully reflective awarenesss of what it involves. We all of us, in speaking our language, deliberately follow complicated and systematic rules of grammar, but how many of us have anything that resembles an explicit statement of those rules when we use them? Our use of concepts -- even philosophical concepts -- is like this, surely: we use concepts largely without knowing (reflectively) their implications. As interpreters, the descriptions of the patterns we find in a text should be accurate -- of course. But the state of mind of the person who produced the text does not provide a privileged way of describing the patterns which we find in it, nor should we assume that the person who produced such a pattern should be the last authority on its implications. The fact that our description of patterns goes beyond translating the authors intentions is only natural when we consider the following: as we become separated from an author (in time or by tradition) so the need increases for an explanation of terms which goes beyond the degree of reflexiveness of the author himself. The disagreement between Skinner and myself might, then, be represented as follows. Skinner is concerned to emphasize that the production of texts is a kind of action and that, to do justice to this, interpretation must orient itself towards those intentions of which the text is a realization and to apply only that range of descriptions which was in principle available to the agent. While I agree with the first point, I am not prepared to concede that it entails the second. In my view, an account of a particular authors use of a term should correspond to his intentions in the sense that our account of the pattern to be

26

found in a text must be projectible -- that it is a constraint on an interpretation of a term that it should allow one to trace the pattern identified through the range of occasions on which the term is used -- but it is not a condition on interpretation that any pattern legitimately identified by the interpreter should have been anticipated (and characterised in such terms) by the author.

Speech Acts

For Skinner, the connection between the understanding of textual production as a form of action and the orientation of interpretation to intention follows from his positive account of the nature of textual intentions, drawn from the theory of speech-acts, and it is to this that I now turn. According to speech-act theory, uses of language are actions in the sense that they represent the actualisation of some conventional possibility given by the currently existing framework of the language.28 In an article published in 1970, Skinner makes it clear that whatever can be meant can be said and that in any culture or historical period: It must be a necessary condition of... communication that the form of Ss utterance should fall within the conventions and limits acceptable at t1 as applying to those particular forms of communication...29 The problem of historical understanding becomes, then, that of grasping alien conventions -- hence Skinners repeated emphasis that it is the task of the intellectual historian to guard against the loss of our sense of the distinctiveness of the past, rather than to establish transhistorical truths or standpoints of

28

Thus, in arguing that interpretation should be an account of what someone meant to mean, Skinner is not defending the Cartesian idea that the intention in question consists in an ultimately purely subjective mental state. On the contrary, the relevant notion of intention is a fundamentally public matter: intentions involve the conventions constitutive of the (illocutionary) force which speech-acts may have. 29 Q. Skinner, Convention and the Understanding of Speech Acts, Philosophical Quarterly 20 (1970). pp. 118-38, p.135

27

criticism. To deny the importance of intentions is, Skinner argues, to reduce interpretation to the narrow domains of sense and reference and to ignore the important -in political philosophy, vital -- dimension of force altogether.30 As Skinner puts it: To apply any word to the world, we need to have a clear grasp of both its sense and its reference. But in the case of appraisive terms a further element of understanding is also required... For example, no one can be said to have grasped the correct application of the word courageous if they remain unaware that it is standardly used to commend, to express approval, and especially to express (and solicit) admiration for any action it is used to describe.31 If it is true that it is necessary to make allowance for force, as well as sense and reference, and if it is true too that the kind of force at stake here is the illocutionary force identified by the theory of speech-acts, then it follows that to neglect intentions of that kind is to reduce ourselves to a bloodless notion of texts as, at best, abstract representations of the state of things, and to ignore the diverse range of attitudes and assumptions which are implied by and articulated through them. To this I think that it is possible to make two kinds of critical reply. One might imagine, first, a rather deflationary response to the idea that interpreters should take into account the speech-acts realised in texts. The point might be conceded, but the question raised whether the range of basic speech-acts embodied in the texts of philosophers and political theorists are of any great historical diversity or interest. Second, one might ask whether, in order to take into account the force involved in certain terms, it is necessary to have regard to intentions to any greater extent than in the case of those terms sense or reference. I shall consider these points in order.
30

[Speech-act] theory reminds us that, if we wish to understand any serious utterance, we need to grasp something over and above the sense and reference of the terms used to express it, Meaning and Context, p.260 31 Language and Social Change, in Meaning and Context, pp. 119 - 34, p.122

28

Austins purpose in How to Do Things with Words was to argue (or, as he would put it, remind us) that not all apparently descriptive or assertoric uses of language were such. But, however one takes the details of his argument, there is no reason to suppose that these are not the central kinds of utterance to be found in works of philosophy: for the most part, surely, philosophers are asserting, arguing and maintaining (with a bit of suggesting, doubting, or proposing thrown in). With very few exceptions, it seems that these intentions are standard and our grasp of them remarkably stable -- it requires rather little in the way of nuanced historical understanding to see that the authors of the past are arguing this or that view, although a great deal to get at what the view in question might be. One reason why this should be disputed might be if no distinction is being made between what the basic speech-act realised in an utterance is -- a rather narrow range of possibilities, as it seems to me -- and the much richer variety of things that may be being done with, through or by means of it. We certainly need to grasp the former in order to be said to understand the text at all -- who could be said to understand a language if she didnt know the difference between an assertion and a command? The latter, on the other hand, would seem to be less indispensable. Let us take an example discussed by Skinner himself. Locke, in paragraph 93 of the Second Treatise compares the thinking which would have it that the imperfections of the state of nature justify the creation of an absolute ruler with the view that Men are so foolish that they take care to avoid what Mischiefs may be done to them by Pole-Cats, or Foxes, but are content, nay think it Safety, to be devoured by Lions32. It is often claimed

32

John Locke, Two Treatises of Government, edited by P. Laslett (Cambridge: C.U.P., 1967), p.346

29

that this passage is aimed against Hobbes. However, as Lockes editor, Laslett, and, following him, Skinner argue, this is a claim which is very much open to doubt for it is not documented that Locke himself had even read Hobbes.33 But, of course, it is easy to see that, if Locke had read Hobbes, he may, in the circumstances of the time, have had good reasons to conceal the fact. Furthermore, it is still possible that the passage was aimed at Hobbes, despite the fact that Locke had not read his work at all.34 Or, then again, it is possible that the passage was aimed not at Hobbes directly but at a family of positions known to Locke of which he may -- perhaps on hearsay grounds alone -- have taken Hobbess to be an example. The task of deciding between these alternatives is, I should say, a typical example of the kind of fearsomely difficult question of intention which arises in the history of ideas. But in such cases a gap emerges between that which must be grasped in order to be said to understand the text itself at all and those questions of intention which advance our understanding of what the author is doing in a text more broadly. It is not true that we need to know whether Locke was attacking Hobbes in order to understand the structure of Lockes counterargument to the absolutist. To think that attacking Hobbes would be a speech-act which could be said to form part of the basic meaning of the text would be to inflate that theory far beyond its original (as it seems to me rather modest) usefulness for the practice of interpretation. On the other hand, who could deny that it is a necessary part of our understanding of the term itself -- not some wider intention realised in or by means of the utterance of which it forms part -- that we should grasp that a term like courageous carries
33 34

Meaning and Context, pp.46-47 Of how many contemporary attacks on deconstruction might one not suspect something similar?

30

commendatory force? This brings me to my second critical reply to Skinner. It is not clear to me that to concede the above point must lead us back to anything like the principle of attribution; there seems to be no more reason why knowledge of the authors intentions should be involved in grasping that part of the meaning of courageous that relates to its force than in understanding its sense or reference.35

VI

Intentions

I have taken issue with Skinners insistence that what the text is meant to mean sets the limits to a properly historical form of textual interpretation. I agree with him, however, that attention to the authors intentions is essential to a complete historical understanding of a text. Let us concede, for the sake of argument, that Kants theoretical philosophy would be more cogent without the theory of synthesis or that his moral philosophy can be detached from Kants insistence on noumenal freedom. Such interpretations (if accepted) will, I believe, have genuine historical value they show us something (albeit something negative) about the way that Kants texts work. Nevertheless, they invite but do not answer a pressing further question: why then does Kant himself adhere so tenaciously to these elements of his theory? To answer this, we cannot remain at the level of the text itself but must form hypotheses about the way in which Kant himself thought about his text. If (as Skinner and I agree) an authors intentions are not a self-transparent, purely subjective mental state, but if (pace Skinner) they are not to be identified as the

35

The issue is complicated, not least by the fact that sense itself, on the standard acccount, can only be individuated -- as the grasp of the contribution made by a term -- in relation to the expression, and thus the speech-act, of which it forms part.

31

enactment of one or more of a range of conventional speech-acts publicly available to the author, how should we best understand them? Here are some suggestions.36 You look out of your window and notice a dog chasing a cat in the park across the street. The cat reaches a tree, claws its way up and disappears into the branches. The dog leans its paws on the trunk, barks for a few minutes before settling down at the foot of the tree, looking upwards. We can say quite a lot about this rather mundane set of events: that the dog believes that the cat is in the tree; that it is waiting for the cat; that it intends to attack it. We can say this with some confidence even without entering the mental world of the dog. We dont need to know what it is like to be a dog (and so, a fortiori, what it is like for a dog to have beliefs and intentions) in order to know (or, at least, have very good reason to believe) that the dog has such beliefs and intentions. Such pieces of mind-reading are very straightforward for most of us most of the time (fortunately so, since we rely on this kind of ability every time we cross a road). The performance becomes puzzling, however, when we think that it requires somehow getting into the mind of the agent whom we are interpreting (the Cartesian mistake) or reconstructing the conventions that underlie the speech-act that they are performing. Note that the way that we characterize the dogs intentions can be accurate without it being a recovery or translation of the dogs language for the most obvious reason (the dog doesnt have a language).37

36

These suggestions are tentative the topic is a very large one and what I have to say can, at best, cover a small part. I would like to think that what I have to say agrees with the to my mind extremely subtle and illuminating discussion by Stanley Cavell in A Matter of Meaning It in Must We Mean What We Say? (Cambridge: Cambridge U.P., 1976), pp. 213-37 37 Hence I reject and not just for dogs Skinners constraint that if a given statement or other action has been performed by an agent at will, and has a meaning for him, it follows that any plausible account of what the agent meant must necessarily fall under, and make use of, the range of description which the agent himself could at least in principle have applied to describe and clarify what he was doing. (Meaning and Understanding in the History of Ideas, pp. 28-29)

32

So what then constrains our interpretation? Let me give another example. You see someone busy in their kitchen. They live alone but there are two sets of cutlery on the table. Some of their intentions seem obvious. They are preparing a meal for a guest, you infer. Indeed, some intentions are so obvious that stating them seems redundant. You can assume that it is their intention to prepare food that is nutritious and hygienic. No doubt, they did not reflect on that fact when they started to cook, but that, surely, is the best explanation of why they are washing fruit, peeling vegetables, boiling potatoes, and so on. You could say that there are some standing intentions that are part of what middle-class Westerners do when they go about preparing a meal. Nevertheless, there may also be more specific intentions that are less obvious. Why is no meat being prepared? Is the guest a vegetarian? Someone who keeps kosher? Or perhaps its just that the host wants to try an interesting bean recipe. If youre interested, you will ask them. They have a lot of information that will help you interpret what they are doing. But that is not to say that they have some essential final authority over their intentions. Let us suppose, for example, that the evening goes well, the meal is delicious, the guest relaxes and, well one thing leads to another. Is this what the host intended? Perhaps you will say that this was too obvious to miss (why else the flowers and scented candles?) but perhaps not or perhaps one just cant tell. When it comes to our own intentions we are all self-interpreters. Of course, we have a lot of data about ourselves that others dont (and we tend to care a lot although that must be balanced by the fact that we have less distance on ourselves to see what we are doing). There is a spectrum between things that we couldnt but know we intended (to invite a guest), things that we undoubtedly

33

intended even if we did not reflect upon (to prepare a nutritious meal) and things about which we may be quite genuinely unsure (to establish a romantic atmosphere). Here, then, are some conclusions. (1) We ascribe intentions as we do beliefs because they enable us to make sense of behaviour. (2) The language in which we do this does not have to reconstruct or translate the language of the agent to whom the intentions are ascribed. The agent may have had no reflective thought on the matter (intending to cook a meal that is nutritious) or, indeed, not be capable of language at all (the dog). (3) Agents, certainly, have a great deal of information about their own behaviour but they do not have some privileged, final authority over its interpretation. Drawing on these conclusions, how should we apply them to the understanding of that very particular kind of practice, the production of philosophical texts? The first thing to say is that the attribution of intentions to authors is similar in principle to the ascription of beliefs in the reconstruction of arguments discussed earlier: it is what makes sense of the pattern that we find in texts. Yet, where the ascription of beliefs may be relatively straightforward (what may we reasonably add to complete arguments that are enthymematic?) the attribution of intentions is, of course, much less so. You may think, like Allen Wood and many others, that Kants moral theory would be much better if we did not see it as centred upon a conception of noumenal freedom. Let us say that you are right. Does that settle the matter about what Kant himself intended? In my view, certainly not. Not only are there many texts of Kants38 which, to my mind, make it quite clear that Kant actually canvases something like a compatibilist view of
38

Which I would be happy to show you if there were time

34

freedom (by which it is sufficient for freedom merely to act on the conviction that we are free) and finds it insufficient for his purposes. Furthermore, there are compelling reasons, I think, why Kant hangs on so determinedly to his more ambitious and demanding understanding of what freedom amounts to. Put briefly, a compatibilist view of freedom might be sufficient to sustain our moral practice, but it could not sustain the belief that we were acting in a way that would merit approval (or punishment) by an omnisicient divine judge. This is certainly not a requirement that most of Kants modern admirers (the majority of whom are, I think, secular liberal humanists) place on moral philosophy, but that does not mean that it is not something that was deeply important to Kant. But Kant, you might object, does not say as much explicitly. Well, actually, there are a number of places where he does39 but, even were the evidence sparser than it is, we should not reject the possibility out of hand. The idea that we should act in a way that we can be held responsible by a just God may be like my cook who wants to prepare a nutritious meal an intention too obvious, too fundamental, too much taken for granted at that time to be spelled out. It is at this point, I think, that the historical sense is indispensable: we need to be able to articulate those commitments that underlie the texts of the past even where such commitments would be vigorously rejected by modern readers.

VII

Iconoclasm

It is now time to draw things together, the more so since you would be forgiven for thinking that I have let them drift apart. I announced that I would put forward a conception of the history of ideas that would be both of philosophical value and
39

See, for example, the remarkable discussion of divine judgement at the end of the Metaphysics of Morals where Kant concludes that it is from the necessity of punishment that the inference to a future life is drawn. (Ak. 6:490)

35

genuinely historical. Yet you might think (if you were not feeling generous) that what I have offered so far is neither. I have, for instance, given an argumentative reconstruction of what I called the problem of development and claimed that this is a kind of background account that helps us understand a range of texts in the late eighteenth and early nineteenth centuries in which the notion of organism (as opposed to mechanism) and the idea of organic development play a prominent role. But this is not contemporary philosophy I am not claiming that the problem of development is one that has force for us.40 Nor is it the kind of history that the historian will recognize its more of a grand claim about a Weltanschauung than the sort of thing that can be checked against archives, databases or other empirical sources. At the very least, the apparent analytical cogency with which I presented it presupposes a considerable distance from the immediate texts. I claimed too that Kants commitment to the idea of noumenal freedom is deeply rooted in his thought because it is essential to being able to see human agents as properly held to account before a just and omniscient God. Yet so much the worse for Kant, Kants modern advocates may reply what we need to do is to reconstruct Kants thought in such a way that it moves away from such untenable commitments. To some extent, I have to plead guilty: the objections to this kind of history of ideas are serious. However, I want to argue that they are worth facing for both philosophical and historical reasons. To make my case, let me revisit two of the themes I raised earlier in this essay. In discussing the reconstruction of arguments from the past, I introduced a contrast (originally made by Derek Parfit) between archaeologists and grave-robbers. I want now to question that analogy. Vivid though it is, it fails to bring
40

Although it appears that Thomas Nagel subscribes to a rather similar argument. See his Panpsychism, in Mortal Questions (Cambridge: C.U.P., 1979), pp. 181-95

36

out the way that we are all affected by the past, whether we like it or not. The beliefs that have come down to us through tradition form the environment within which we I mean, immediately, we philosophers, but, in fact, I believe, we much more generally live and think. A more accurate analogy might be with people who occupy a very ancient, in places ramshackle city. Some parts, we might say, we can simply preserve as ancient monuments, but others we have to live and work in. There the question arises: are they structurally sound or do they need re-building? And there is always an alternative, of course. Instead of preserving or restoring, would we do better to knock them down? The contrast between archaeologists and grave-robbers suggests just two ways in which we may value the past: because we value knowledge of it for its own sake or for what it may add to our lives now. But there is another possibility: to seek knowledge of the past in order to emancipate ourselves from it. Instead of just archaeologists and grave-robbers, might there not also be iconoclasts? This is not just a logical possibility, in my opinion, but actually one of the most compelling reasons to study the history of ideas. Much of my own work fits this model. In Hegels Dialectic and its Criticism41, for example, I argued that, while the terms of dialectical philosophy (particularly concepts such as reflection and determinate negation) are coherent in the wider context of Hegels speculative idealism, their appropriation by the Critical Theorists of the Frankfurt School takes them out of their sustaining element. In On Voluntary Servitude42 I tried to make a similar case with respect to the organic theory of society as a self-maintaining system. In each case,

41 42

Cambridge: Cambridge U.P., 1982 Cambridge: Polity Press, 1996

37

situating concepts historically is meant to show their dependence on a particular theoretical context for their justification. Bearing this idea in mind, let us return to what I said earlier about philosophical problems. Philosophical problems arise, I suggested, when commitments that are felt to be compelling lead to conflict. When I say commitments I mean to include many things that may be far from explicit background beliefs, intuitions, received common sense, ways of seeing the world, practices, pictures or forms of life. In order to leave the domain as open as possible, let me refer to them as doxa.43 What is clear about such philosophically-generated conflicts is that there is generally no single mandatory way to respond to them. Perhaps there are cases when merely to reveal one of the doxa is enough: it is evidently untenable and all reasonable people will reject it. If so, I cannot think of an example. More commonly, we will find some doxa more plausible, others less so and different reasonable people will differ about the different weight to give them. This is why philosophy is not, ultimately, a matter of proof so much as of persuasion. It is not that arguments do not play a role but that, since the arguments of philosophy are so frequently demonstrations of inconsistency, they leave open the question what to do about them. History can play an important role both in discovery and in persuasion. If doxa are things that we take for granted, then, obviously, one way to recognize our commitment to them is to be confronted with the texts of thinkers who dont. The effort of entering their set of beliefs and assumptions forces us to articulate our own. And this may lead us to question our own doxa. It is true that showing that a belief has a certain origin (that is, a certain causal history and/or motivation) does not refute it, but it is

43

Noting that I intend to endorse thereby neither the Platonic contrast of doxa with episteme nor Bourdieuan functionalist sociology.

38

certainly reasonable for that to affect the way that we look at it. In short, the history of ideas as I advocate it connects with philosophy because it is central to philosophy to articulate and call into question what is taken for granted.44 To give an example, let us return to the language-chess analogy. I introduced the language-chess analogy as a way of making clear by contrast certain features of the interpretation of philosophical texts that I am advocating. The view that language is a shared practice governed by public rules and that it therefore constitutes a public domain that is conceptually prior to and independent of the state of mind of individual speakers, was, I said, a largely unquestioned orthodoxy in much of twentieth-century analytical philosophy. Even if I were wrong about this, my description of the languagechess analogy could still have a certain heuristic value in helping to clarify what is involved in the interpretation of philosophical texts, but if, as I believe, the languagechess analogy embodies something deeper, then the argument has force in drawing attention to a widespread, yet, on reflection, questionable set of assumptions. As I noted, the idea that language is something essentially public has been identified as the essential transformative idea of the Fregean revolution. Moreover, it is noteworthy that two other accounts of thought and language came onto the European scene around that time which also asserted the priority of public meaning over private acts of thinking Saussures distinction between langue and parole and the Husserlian phenomenological conception of intentionality and noema. It seems plausible to think that these were simultaneous reactions to the powerful strand of psychologism that ran through philosophy in Britain and Germany in particular in the later nineteenth century. But if the idea of language as public and prior (in some sense) to the acts of thought of
44

Isnt it remarkable that this is as true today as it was for Plato?

39

individual thinkers was widespread in the early years of the twentieth century, the form it took in the post-war years especially in Britain was quite distinctive. Here, for example, is a quotation from Stuart Hampshires Thought and Action (first published in 1959): Reality and experience cannot be thought about unless we have rules that correlate particular groups of signs with particular recurrent elements in reality and experience, in such a way that any familiar use of a particular group of signs will be taken as a reference to some particular element in experience.45 Characteristically, Hampshires claim about the pervasively rule-governed nature of experience is made at the outset of his book (in fact, in the first paragraph). It is not something that Hampshire sees the need to explore or defend; it is assumed. Gilbert Ryle, William Alston, Jonathan Bennett, Richard Hare and John Mackie are other authors who, likewise, invoke a similar thesis more or less in passing. The idea of the rule-governed nature of meaning is presented unusually explicitly, however, in John Searles Speech Acts, a development of his 1959 doctoral thesis, written under the supervision of Austin and Strawson. Searles book thus usefully articulates many of the ideas and assumptions that motivated philosophical work in Oxford in the nineteen-fifties and early sixties.46 According to Searle, the main hypothesis of his book is that speaking a language is engaging in a (highly complex) rule-governed form of behavior, underpinned by a (system of) rule(s) of the form `X counts as Y in context C.47

45 46

S. Hampshire, Thought and Action, Second Edition (London: Chatto and Windus, 1982), pp. 12-13 Speech Acts (Cambridge: C.U.P., 1969) As the reviewer for Mind, quoted on the back cover of the paper edition, notes: The brilliant but programmatic insights of Austins How To Do Things With Words are systematically developed and integrated with the more recent work of philosophers such as Grice, Rawls and Searle himself... 47 Speech Acts, pp. 12, 52

40

Two features of Searles position are worth noting. The first is what he terms the principle of expressibility: that whatever can be meant can be said.
48

In other words,

language is first and foremost a system of public rules; the private use of language and innovative or original development must be in some way parasitic on that public existence. The second is that there is no sharp line of distinction to be drawn between language and action. It is their common rule-governed character that connects the two: ...a theory of language is part of a theory of action, simply because speaking is a rulegoverned form of behavior.49 These two claims represent language (and action in general) as held within an invisible network of normative commitments. They were widely accepted; indeed, I believe that they form the implicit background to Skinners version of speech-act theory. We have here then, I suggest, a clear example of philosophical doxa a view taken for granted rather than argued for. In digging out such doxa and holding them up for scrutiny and objection the history of ideas and philosophy, understood as critical engagement with unquestioned assumptions about thought and language, come together.50 The study of doxa and the attempt to address philosophical problems connect because doxa are (very often) part of philosophical problems.51

48

Speech Acts, p. 19 Speech Acts, p. 17, my emphasis. 50 The view of language as rule-governed action is relatively easy to criticize now because the tide of Anglo-American philosophy has turned so strongly against it. I have examined some of the criticisms in The Role of Rules (International Journal of Philosophical Studies, 9 (2001), pp. 369-84). Put briefly, the idea was challenged both by the pragmatist strand in analytical philosophy represented by Quine and his pupils, and (more surprisingly) by a new reading of Wittgenstein. (I say surprising since Wittgenstein had been widely believed to be the chief advocate of the rule-governed conception of language.) Of course, though the view was dominant in the U.K. and widespread in the United States, it was far removed from the view of language to be found in much twentieth-century Continental philosophy in the writings of Adorno or Gadamer or Derrida, for example. 51 Another term for doxa might be Daniel Dennetts bugbears. (See Elbow Room: the Varieties of Free Will Worth Wanting (Oxford: Oxford U.P., 1984), Ch. 1, Dont Feed the Bugbears). Dennetts treatment of free will is entirely unhistorical but his philosophical approach the attempt to defend his view by loosening the hold of certain unreflectively accepted images and beliefs that apparently entail the contrary fits very well with what I am here advocating undertaking by means of history.
49

41

In the course of this essay I have engaged with Quentin Skinners account of how we should interpret the texts of the past. But, of course, historians want to do more than interpret texts. They want to ask why they came about and what effects they had. On an extreme view, ideas and ideals are mere epiphenomena; they are, as Namier said flapdoodle. This view is often ascribed to Marxists evidently, wrongly. Marxists believe that ideas and ideals play an essential causal role in maintaining unequal structures of power; that is what the theory of ideology is all about. Skinner agrees that political discourse plays an essential legitimating role. Thus, he says, the entrepreneurs of early modern Europe had a recognisable motive for wanting to pursue their ventures unhindered They needed as a matter of some ideological urgency to legitimise what they were doing to those expressing such comprehensive doubts about the morality of their lives.52 Like the Marxists, Skinner thinks that political discourse is legitimating in its effects but instrumental in its origins.53 In other words: Instrumentality-InLegitimation-Out. There is not the space here to take issue with why I think IILO views of the Skinner/Marx/post-modern (though not, I think, pace Skinner, Nietzschean) sort are deeply problematic (that they are extremely unflattering to the sincere selfunderstanding of philosophers is evident though that is hardly a conclusive argument).54

52

Moral Principles and Social Change, in Visions of Politics, Vol. 1 (Cambridge: Cambridge U.P., 2002), pp 145-57, p.147 53 Thus he writes in relation to the work of Reinhart Koselleck: Koselleck and I both assume that we need to treat our normative concepts less as statements about the world than as tools and weapons of ideological debate. Both of us have perhaps been influenced by Foucaults Nietzschean contention that the history which bears and determines us has the form of a war. (Retrospect: Studying Rhetoric and Conceptual Change, in Visions of Politics, Vol. 1 (Cambridge: Cambridge U.P., 2002), pp 175-87, p.177) 54 Many of the difficulties are raised in On Voluntary Servitude

42

In general, though, IILO views seem forced to depict the history of philosophy as a series of cunning tricks a ruse or strategem played on the gullible.55 My own preference is for what I will call a LILO approach Legitimation-InLegitimation-Out. Legitimation I take here in the very broadest sense ultimately rooted in human beings standing need to find ways to make sense of their lives in the face of (or, at least, finding ways of accepting) the unpalatable but inescapable facts of death and suffering. How those possibilities of legitimation have played out how they have become exhausted and new ones developed forms a central part of the kind of history of ideas that I aspire to (which is not, I hope, to deny that questions of economic interest, institutional power and, indeed, individual psychology also play a role in the story). In so doing, I would like to think of myself in a tradition that can trace itself back to Hegel, and includes Nietzsche (as I would argue), Weber, the Frankfurt School (particularly Die Dialektik der Aufklrung) and the extraordinary Hans Blumenberg.56 Thus we can (or so I would argue) see the tenacity Kants adhesion to the doxa of the sovereign individual acting freely outside the causal order as explained by the need to see the individual as apt to be judged justly by an omniscient divine judge. And this, in turn, is (or so again I would argue) a part of Kants very novel post-Lisbon Earthquake response to the problem of theodicy, a problem which (as Blumenberg has consistently shown) is endemic in monotheistic religion, the most central (and now questionable) of

55

It is not clear whether those who perpetrate the stratagems are themselves gullible in Skinners view. In Marxs view, famously, not: the ideologists are themselves subject to ideological illusion hence their ability to propagate ideology all the more convincingly. For Foucault, individual thinkers appear to be mere surface effects, carried along on the shifting tides of episteme or (later) some wholly unspecified conception of power. I am not sure where Skinner stands. 56 Yes, and Carl Becker, Lovejoy and M.H. Abrams too

43

the resources that Western human beings have used to make their place in the world intelligible to themselves. But that is a story for another day.57

Michael Rosen Department of Government, Harvard University mrosen@gov.harvard.edu

57

The second Berlin Lecture was entitled The Idealist Theory of History Defended (Sort of) and continues in that direction.

44