1. Interpreting dead philosophers. Those of us who study the history of philosophy spend our time trying to understand texts written mostly in languages other than English by people long dead. Our primary aim, whose successful achievement is presupposed by any other aims we may have, is to determine what the text means, or what the author means, or meant (I take these all to be the same).1
This is often difficult to do. The writings of Kant, for example, often challenge our ability to understand them. This can happen at the level of a single term. (What does Kant mean by “synthesis” or “determination of the will” or “transcendental principle of judgment”?) Or there can be questions about specific assertions (that matter is an appearance rather than a “thing in itself”, or that the moral law is a fact of reason). Or it can happen when we try to understand the general structure of his system. (How does judgment mediate between understanding and reason?)
But the kinds of questions that I want to ask do not arise only in the case of some philosophers, and the fact that we have to raise them cannot be blamed merely on the regrettable unclarity with which some philosophers write. The texts of Kant and Hegel are famously obscure, but the meaning of even apparently lucid writers such as Descartes and Hume is something that begins to elude us when we ask questions about their views. Descartes says that the mind and body are two distinct substances, which together constitute one thing, the human being. But exactly how do they do so? Hume reasons at length about our idea of causal power or necessary connection, basing his reasonings on the thesis that we have no ideas that are not copied from impressions. But does Hume mean to say that we have an impression of causal power or doesn’t he?
Asking difficult questions about what philosophers mean in their writings turns out to be an important part of what it is to read a text in the history of philosophy, or at least to read it philosophically. And trying to decide what a philosopher means will also lead us into controversies that often seem to be about philosophy as much as they are about what an author thought or meant. But how can questions about what someone means be philosophical questions? How can controversies about what a text means be philosophical controversies?
There have long been disputes, for example, about whether Aristotle regarded form or matter as the principle of individuation of substances.2 Again, some think that in the famous discussion of the piece of wax in the second Meditation, Descartes was trying to establish that only the properties dealt with by mathematics belong truly and permanently to matter, while others think his aim was the more modest one of identifying which properties are necessarily involved in our concept of body insofar as this concept is a distinct one.3 One set of interpreters holds that Hume intended his philosophy to curb the pretensions of metaphysics and thwart the enthusiasm of religious zealotry by casting skeptical doubt over all human knowledge and belief; others say that far from trying to discredit human knowledge, Hume was trying to lay a new foundation for it on the basis of a comprehensive empirical science of human nature.4 Kant scholars ask whether noumena or things in themselves are entities distinct from their appearances and causing them, or whether things in themselves are the very same entities as appearances, distinguished from them only by the ways in which they are considered or referred to.5 There is a dispute about whether Marx condemned capitalism for distributive injustice or held a deflationary account of justice according to which capitalist exploitation is just but no less objectionable for being just.6
When we ask these questions and try to settle these disputes about the meaning of a philosopher or philosophical text, what exactly is it that we are trying to find out? And what kinds of arguments and evidence are relevant?7
2. Why study the history of philosophy? But perhaps some will want to ask a prior question. Why does it matter precisely what long dead philosophers, or their texts, really mean? It might be argued that from a historical point of view, all we really have is what the texts say, and what others have said about them. Endless philosophical disputations about precisely what the texts mean is of little use to those who are interested, as historians should exclusively be, in wie es eigentlich gewesen ist. Philosophers might argue that the only job of philosophy is to concern itself with questions about what material objects really are, or what makes a thing the thing it is and different from other things, or whether we can ever know reality as it truly is, or whether capitalist wage bargains are unjust. They might object that we make no real progress in answering these questions by studying the opinions on them held by Aristotle, Descartes, Hume, Kant or Marx -- especially if these opinions are so obscurely expressed that even the experts, with all their erudition and fine-grained analysis, still cannot agree on what they are.
My aim here is not to defend what I do as a historian of philosophy, but it might help if I at least sketch the reply I would try to make to these objections, since what we are looking for as the meaning of a philosophical text will be conditioned by what we are trying to accomplish in undertaking this kind of inquiry. I do not think that the philosophical importance of studying the history of philosophy can be demonstrated a priori by some rigorous argument. It can be appreciated only by those who engage in philosophical inquiry, and have studied enough of the history of philosophy to experience for themselves, in a variety of ways, how indispensably it contributes to that inquiry. But I will try to offer some general considerations that might summarize the results of such experiences for a scholar who has had them.
To the objections of historians I would be conciliatory, at least to a degree. To the extent that historiography is interested only in the historical influence of what philosophers wrote, rather than the significance of what they actually meant, it can afford to ignore subtle interpretive inquiries. But I would also point out that it is extremely hard for a historian to keep away from questions about what philosophers mean, since these questions arise as soon as they try to explain the influence of a text in terms of its intellectual content. It is also very easy to underestimate the danger of being satisfied with what is supposed to be obvious about this.8 There is also an unfortunate tendency on the part of some (to which vulgar Marxism has contributed) simply to identify the meaning of what philosophers said with the role their ideas have played in social or political struggles or with some set of historical consequences for which the philosopher’s ideas are commonly held responsible. The element of truth in this is that texts and ideas, like people and their actions, always have a historical fate they cannot escape. But when we reduce the meaning of a text merely to that fate (or, more often to some conspicuously lurid aspect of it), this does not tell us what the text means, but only gets in the way of understanding that.9
Philosophers’ objections to studying the history of philosophy are more fundamentally mistaken and more pernicious. Fortunately, in the last generation their credibility has declined sharply in American philosophy. G. E. Moore once confessed that it was not life or the sciences that suggested philosophical problems to him, but rather the things other philosophers had said about them.10 In the mid-twentieth century, many philosophers in the tradition from which Moore came would probably have understood this remark as meaning that philosophical problems are entirely artificial inventions, of interest only to the peculiar sort of diseased or befuddled mind that might think them up. But I think Moore’s point was really quite insightful, and therefore entirely different from this. Moore’s remark was his way of acknowledging a fundamental truth about virtually all philosophical questions, namely, that they are inherited from the thoughts of earlier philosophers. All such questions have been created and shaped through a long historical process in which philosophers have taken over the thoughts of earlier philosophers, criticizing and modifying them.
This means there is something fundamentally self-deceptive in the view of those who disdain the history of philosophy on the ground that they are “interested only in solving the problems themselves, not in endlessly rehashing the failed attempts of others to solve them.”11 For solving a philosophical problem is not like solving a problem in engineering, where the only issue is whether the solution enables you to do something in the future that you couldn’t do in the past. Above all, solving a philosophical problem means coming to understand the problem. Since these problems are always products of a history, you can’t fully understand them unless you understand their origins.
Sometimes it may look as though you can do this well enough merely by studying the thought of the previous generation of philosophers (the ones who taught you philosophy). After all, problems in mathematics are also inherited, but mathematicians do not need to engage in deep study of the history of mathematics. One thing we historians of philosophy learn to our chagrin is that most of the philosophers whose works we study with such care were not especially well-informed or accurate interpreters of their predecessors.12 Yet from the fact that philosophers have been extremely successful without knowing much history of philosophy, it does not follow that ignorance of the history of philosophy is not harmful to them as philosophers. (Beethoven and Smetana wrote great and original music after they were completely deaf. It does not follow that being deaf is not a serious drawback to composing.) Philosophical problems relate to more aspects of human life and experience than mathematical problems. There are many more things that might count as a solution to them, and no solution to a real philosophical problem is ever going to be as elegant, perfect or certain as a mathematical proof. Truly understanding philosophical problems therefore requires taking a wide view, which means, historically, a relatively long view.13
The Bible tells us that there is no new thing under the sun.14 Like much that is in the Bible, this is no doubt poetic hyperbole and not meant literally. But in philosophy a fertile source of the new is the re-emergence after a time, often in the form of a re-interpretation, of ideas and viewpoints that have for a while been unknown or else despised and neglected as dead, profitless and false. Some of the greatest movements in the history of philosophy have been sparked by the rediscovery and revitalization of old ideas: of Aristotle by Averroes and the Western scholastics of the high middle ages; of Sextus Empiricus by Montaigne, Gassendi and Descartes; of Spinoza by the German idealists. Or sometimes ideas that are not necessarily despised contribute to what is new by being reappropriated. Think of the diverse ways in which recent philosophy has been impacted by successive waves of the rediscovery of Kant (by Cohen and Cassirer, Strawson and Putnam, Rawls, Apel and Habermas), or of Hegel (by Sartre, Taylor, MacIntyre, Hösle, McDowell and Brandom), or even of Dewey (by Quine and Rorty).
As these examples illustrate, however, there is no Nietzschean eternal recurrence in philosophy; what is old never returns precisely as it was, and often the heritage of a past philosopher or past idea can become a bone of contention. This makes it a matter of far more than antiquarian interest whether past philosophers are being correctly understood and whether revisions and modifications of their views are well-motivated or merely the result of misreadings and distortions, blinkered through the influence of intervening prejudices. Deciding such questions is therefore not merely a matter of intellectual heraldry, but is essential to the proper philosophical assessment of theses, arguments and theories. Likewise, it matters for philosophical purposes (and is not of ‘merely historical’ interest) whether, for instance, as Myles Burnyeat has argued, our modern understanding of skepticism has been based on fundamental misperceptions about what ancient skeptics were up to and how they saw the world.15 One of the greatest services we historians of philosophy can render to philosophy is therefore to prevent the effacement of earlier views, and especially to keep alive what our age is likely to regard as “weird”, “foreign”, “outdated”, “no longer to be taken seriously” -- that is, what is incapable of easy assimilation into the prejudices and fashions of our own time. For precisely that (or at any rate some now unidentifiable and inscrutable part of it)is always the source of virtually every philosophical thing that is new under the sun.
3. Thinking dead people’s thoughts.When we interpret a text in the history of philosophy, a surprisingly varied set of considerations come into play. To begin with, to do it right we need to understand the language in which the text is written.16 We need to know what other philosophers had thought and were thinking at the time.17 Sometimes we have to be aware of how the philosophical questions addressed by the text had been shaped by political, religious or other kinds of social forces.18 Also of vital importance is philosophical expertise – the ability to formulate ideas clearly and precisely, to construct and evaluate arguments, even to build philosophical theories and systems for ourselves.
For this reason, the interpretation of texts in the history of philosophy raises a specific set of problems that might be thought to differ from the problems of interpreting documents in other fields of the humanities. In literary texts, for example, the author often does not address the reader directly, but speaks through other characters; even the persona of a narrator in a novel or of the ‘speaker’ in a poem may be a carefully crafted fiction, quite distinct from the person of the author. But problems of that kind arise in philosophical texts too – in the dialogues of Plato, Diderot or Hume, for example, or the pseudonymous writings of Kierkegaard, the aphorisms of Pascal, Nietzsche or Wittgenstein, or in philosophical novels such as those by Dostoyevsky or Sartre.
This kind of problem arises even in such a basic philosophical text as the first sentence of Descartes’ Discourse on Method: “Good sense is the best apportioned thing in the world: for each thinks he has been so well provided with it that even those who are hard to content in all other things are not accustomed to desire more of it than they have.”19 Descartes’ argument here is surely intended ironically; it is a self-conscious joke. What, then, are we to make of the fact that he goes on to treat the thesis that good sense is equally distributed as though it had been adequately demonstrated? Such features of philosophical texts are like the analogous features of poems, novels and plays; they add to the richness of a text, but also make it more difficult to interpret. I think many of the things I am going to say about interpreting texts in the history of philosophy might well carry over into the interpretation of literary texts or other works of art, or even to the interpretation of such things as the aims and intentions of historical agents. But I will not argue for any particular extensions of what I say to other kinds of interpretation.
What is the meaning of a philosophical text? R. G. Collingwood is well-known for advancing the thesis that the proper method of all history (including the history of philosophy) is that of re-thinking in one’s own mind the thoughts of people who lived in the past.20
There is a lot in Collingwood’s approach that I agree with. One of Collingwood’s aims was to rescue important figures in the history of philosophy from what he thought were the shallow, arrogant and shortsighted criticisms of his analytical contemporaries. He wanted them to see how difficult it was to be sure they had gotten the questions and aims of past philosophers right when they accused them of failed theories and bad arguments. He urged them to try to rethink the thoughts of the past so that they would not dismiss the thoughts of Plato, or Descartes, or Kant by taking them to express whatever simplistic (and usually erroneous) ideas, drawn from the contemporary analytical fashions, were suggested to their impatient and blinkered minds by a casual reading of the historical texts, thus turning the study of the history of philosophy into little more than a contemptuous survey of the stupid errors supposedly committed by famous dead men.
It is easy for me to sympathize with Collingwood’s aims here. If the only points he was trying to make were those mentioned above, I would wholeheartedly agree with him. But Collingwood went further. He ended up maintaining that the theories of philosophers in different ages were incommensurable, because they were attempts to answer different questions.21 That merely invites the thought I have just been inveighing against, that the history of philosophy is bound to be pretty irrelevant to the philosophy we do today. For the same reason, Collingwood’s account makes it hard to explain why not only historical empathy but also philosophical skill is needed in interpreting a text in the history of philosophy. It even seems directly to rule out something that good historians of philosophy regard as essential to interpreting texts, namely, the use of concepts and theories that have been developed since the text was composed and therefore could not possibly have been part of their author’s actual thought processes.
Let me develop Collingwood’s idea in a way that may be a caricature of it, but nevertheless succeeds in bringing out more clearly some of the problems I want to discuss. We might think of the meaning of a text as a certain inner mental process that was taking place in the mind of the author as the text was being written. The text is the author’s attempt to put down words that will enable the reader of the text to duplicate the succession of those thought-types in the reader’s own mind. Following this picture, my task as an interpreter of the text will be to complement the author’s efforts by bringing before my own mind, as far as possible, exactly the sequence of mental process-types that were in the author’s mind as the text was being written. For the sake of convenience, let me call this (possibly caricatured) version of Collingwood’s claims the “Collingwood picture”.
One problem raised by the Collingwood picture is whether it is even possible to re-think the same thought-types as people who lived in the past. How should we set about doing this? Even worse, how can we ever know whether we have done it? That way lies skeptical historicism. But the more serious problems arise even if you suppose, as Collingwood apparently did, that we can think the same thoughts as people in the past. I think there are good reasons for doubting whether successfully doing this would really constitute either what we do mean or what we ought to mean by interpreting and understanding a philosophical text. Reproducing the thought-types available to philosophers in past centuries would not insure that we are getting them right. In fact, limiting ourselves to the thought-types they had available to them would often cut us off from our best route of access to their meaning, and make it impossible for the study of past philosophers to make a contribution to ongoing philosophical inquiry.
It may help at this point to look at an example. In Chapter Seven of his recent book Representation and the Mind-Body Problem inSpinoza, Michael Della Rocca is trying to understand how three claims found in Spinoza can be consistent:22
Every mode of extension is caused by another mode of extension (cf. Spinoza, Ethics 2p7). 23
No mode of substance conceived under the attribute of thought can cause any mode of substance conceived under the attribute of extension, or vice versa (cf. Ethics 3p2).
“The mind and the body are one and the same thing, conceived now under the attribute of thought, now under the attribute of extension” (cf. Ethics 3p2s).
Della Rocca is responding specifically to R. J. Delahunty’s argument that (3) is inconsistent with (1) and (2). For Delahunty claims that the following form of argument is valid.
(i) Mode of extension A causes mode of extension B.
(ii) Mode of extension A = mode of thought 1.
(iii) Therefore, mode of thought 1 causes mode of extension B.24
Della Rocca, however, argues that, according to Spinoza, claims (1)-(3) are consistent. He does so by appealing to W. V. O. Quine’s notion of referential opacity.25 A context is referentially opaque when a substitution of one co-referring term for another yields an invalid inference. For example,
John knows that Jim is sitting next to him at the bar.
Jim = the serial killer.
Therefore, John knows that the serial killer is sitting next to him at the bar.
This inference is invalid because contexts like “A believes that …” and “A knows that…” are intensional. That is, what is true of a subject in those contexts depends not only on the identity of the subject but also on how the subject is referred to or represented. What John knows or believes about Jim depends on how Jim has been presented to him. If Jim has been presented to John as “Jim” but not as “the serial killer,” then the fact that ‘the serial killer’ and ‘Jim’ refer to the same person does not entitle us to substitute one expression for the other when we are talking about John’s knowledge or belief. Della Rocca argues that for Spinoza, causal contexts are also intensional, therefore referentially opaque. Granted that thesis, Delahunty’s argument (i)-(iii) would be invalid, and (3) would be consistent with (1) and (2).
I have spoken of the thesis that causal contexts are referentially opaque as ‘Spinoza’s thesis’, but of course Spinoza never said any such thing. In fact, Spinoza never couldhave said or even thought it, since the term and even the concept ‘referential opacity’ was devised by Quine in the second half of the twentieth century and therefore was not available to Spinoza.
If, following the Collingwood picture, we identify the meaning of what Spinoza wrote with some thought-processes actually going on in Spinoza’s mind sometime in the third quarter of the seventeenth century, then we must dismiss Della Rocca’s interpretation as mistaken solely on that ground. Given the Collingwood picture, Spinoza can no more have subscribed to the thesis that causal contexts are referentially opaque than Aristotle can have uttered an English sentence, such as ‘All men by nature desire to know’. For just as modern English did not exist in Aristotle’s day, and hence was not available to him to speak, so the concept of referential opacity did not exist in Spinoza’s day, and hence was not available to him to think.
Yet Della Rocca’s interpretation of Spinoza seems to me correct. Not only does it provide a simple and straightforward solution to the problem of reconciling (1)-(3), but Della Rocca also shows convincingly that it dovetails with other doctrines about causality and representational content which it is reasonable to attribute to Spinoza, such as his even more famous claims of substance identity and substance monism, and his belief in the mind-relativity of content (a thesis which Della Rocca also states in terms not historically available to Spinoza). Della Rocca also finds some direct textual support for his interpretation in Ethics 2p6, where Spinoza says that God causes given modes insofar as he is “considered through the attribute of which they are modes.”26 But that Della Rocca is right is not essential to the point I am making. For if it is even possible that Della Rocca’s interpretation is correct, then it must be possible that what Spinoza means in the Ethics can be properly understood only in terms of concepts not available to Spinoza, which therefore could not possibly have belonged to the thought processes passing through his mind when he wrote the Ethics.
The point I am trying to make here, to put it with a sharpness approaching paradox, is that people can mean things they can’t think, and therefore that they must be able to express thoughts they can’t have. Accordingly, my criticism of the Collingwood picture could be put this way: In discerning the meaning of a text, we are interested in the thoughts the author expressed, but not necessarily in the thoughts the author had.
We might avoid these paradoxes by stipulating that whatever someone means in what they say or write is eo ipso something that they think, or that there is some sense of ‘think’ in which it is necessarily true that a whatever a person means is something the person thinks. I have no strong objection to using “think” in that sense, as long as we realize that such a stipulation would no way save the Collingwood picture from my objections. On the contrary, it would be a fundamental rejection of that picture, since the picture holds that we get at the meaning of what the author wrote only through rethinking in our minds what the author thought (which is depicted as there prior to and independently of our process of interpretation), whereas this stipulation would have us get at what the author thought only through deciding what the author meant, and treats what we count as the author’s thoughts as derivative from what we count as the meaning of what the author wrote.
The study of Kant is a fruitful source of similar examples, because even more than Spinoza, his writings have been subject to a long history of reading and interpreting, and have interacted with the ongoing reflection of each subsequent generation of philosophers. Consider the claims that Kant held a functionalist conception of mental activity or was a constructivist in moral theory, or anticipated Marxian materialism in his philosophy of history.27 Functionalism, constructivism and historical materialism are all positions formulated only well after Kant’s death. It is even arguable that in the actual historical sequence of events, it became possible to formulate all three positions only because Kant wrote what he did, and because other philosophers then reflected further on his thoughts in creative ways that Kant could not possibly have known about.
If the meaning of what Kant himself wrote is restricted to the mental processes that might actually have passed through his mind, then on that ground alone we can dismiss out of hand the notion that he could have subscribed to these positions in virtue of the meaning of what he actually wrote. But whether we think the above interpretive claims are true or false, they certainly cannot be dismissed merely on these grounds. There are any number of real questions in the history of philosophy which take the form of asking whether Kant or some other philosopher) himself belongs to a certain tradition of thinking that was subsequent to him and was based on certain ways of appropriating his thought. The Collingwood picture, taken literally, would seem to commit us a priori to a negative answer to every question of this form, and thus to rejecting a priori a lot of what makes inquiry into the history of philosophy interesting and worthwhile.
This also shows how the Collingwood picture makes it impossible to understand an important aspect of historical development in philosophy. Ideas seldom spring from human minds Athena-like, fully mature and magnificently armored with cogent articulation and argumentative defense. They usually develop gradually, first anticipated, then adumbrated, and only later, after a long development is it possible adequately to articulate and defend them. It is part of what Hegel meant by saying that the Owl of Minerva spreads its wings only at dusk that this is a process that never really ends, so that our access to philosophical thoughts, as to the meaning of all other creations of spirit, is necessarily limited by history and hence reaches its uttermost boundary for us in the present.28
It follows that an idea may belong to the meaning of a text without its even being possible for it to have been part of the author’s actual thought-processes. This is also an important reason why Collingwood was wrong to say that philosophers in different ages are always addressing different questions.29 As we have seen, it is only because this thesis of Collingwood’s is false that the history of philosophy can be of interest to philosophers at all. And it is only because the history of philosophy is necessarily of interest to philosophers in the way that it is that philosophical questions themselves are the kind of questions they are. So the Collingwood picture gives us not only a false image of what interpretation is, but also presupposes a false image of what philosophy is – ironically, given many of Collingwood’s intentions, an image that would deprive philosophy of any philosophically meaningful history.