davidson and derrida
Sam Wheeler III interviewed by Richard Marshall.
Sam Wheeler III is always brooding on Davidson and Derrida, on connecting deconstruction with analytic philosophy, on how to do this, on their approaches to metaphor, on the legitimacy of some of De Mann’s ideas, of how Davidson improves on Derrida and Wittgenstein, on Quine vs Davidson, on an objection to naturalism, on being a fan of physics, on exotic objects, on anti-metaphysical strands in Quine and Davidson, on objections to contemporary semantics, on Kripke and metaphysical realism, on how Davidson can be compatible with Kripke and Aristotle, on the logic of ought, and on never figuring out what literary theory was doing. This one goes to the depths.
3:AM: What made you become a philosopher?
Sam Wheeler III: I was a faculty kid at a small liberal arts college in Ohio, living on campus for much of my childhood. There was a summer theatre there, and I scrounged a job as tent-cleanup kid/ child actor. I listened to discussions about God, reality, Marx, Nietzsche, etc. I attended a Quaker Sunday school class run by the philosopher, Mahlon Hepp, where the readings were the Upanishads, the Bhagavad Gita, and other such texts.
I was always fascinated by puzzling arguments. For a confirmation class I took in order to meet girls, I did a report on “predestination” which I took to be the doctrine that, if God made the world with exceptionless physical laws, whatever happens would be God’s doing. I argued that the alternative, occasional randomness, would not make people responsible either. Since those options exhaust the possibilities, there is no genuine responsibility. I’ve only recently figured out what is wrong with that argument. When I got to college, I was going to be a philosophy major.
3:AM: You make an interesting and, at first glance, rather unlikely connection between analytic philosophy and deconstruction. The two go-to guys in this equation are Davidson and Derrida. And they have tended to be seen as belonging to two distinct traditions – Davidson as an Analytic, and Derrida as a Continental. Is this a distinction that you find useful? You suggest that, for example, when Derrida talks about ‘presence’ he’s talking about what Wilfred Sellars calls ‘the myth of the given’ and this suggests we can translate terms from the two traditions and find a common set of problems. And ‘deconstruction’ links a unifying set of issues from the two traditions. Is that right?
SW: Very briefly, I think of philosophical traditions as like literary or artistic traditions. Philosophers are trained in certain texts, respond to those, respond to responses to those texts, and so on. Different texts are added in the different traditions, and different topics become important. So divergence grows. At the moment, the latest texts that analytic and continental philosophers will both have read are those of Kant.
I got interested in Derrida by pure accident. I was trained as an analytic philosopher of language at Princeton in the 70s. I had been focused on problems generated by meaning, conditionals, vagueness and the like — problems on which of course I still primarily work. Anyway, sometime in the late 1970s, I was trying to catch up on literature by taking literature courses, something I had neglected as an undergraduate. At an APA, Dick Rorty was giving a talk on Derrida. I had never heard of Derrida, but when Dick said I should look into him, I did. Rorty is second only to Davidson as a mentor and influence. Once I started reading Derrida in a study group at Connecticut College, I found that the interpretive hypothesis that he is doing something like Quine and Davidson turned out to explain a lot and to be very productive.
There are parts of Derrida’s work that I am not equipped to understand. Glas, for instance, is opaque to me, at least now. Understanding another tradition is a lot of work, involving reading the texts the author presupposes, seeing what the issues are, and so forth. I do understand some of Derrida’s central views in the Philosophy of Language. If you read Husserl’s Logical Investigations and Saussure and get entrée into Derrida’s thought via “Signature, Event, Context” and his writings about philosophers you already know something about, you can get started. Getting into another tradition is not something I would recommend to an untenured analytic philosopher, given the focused specialisation now required in academia.
In spite of the differences, there is at least one similar progression of questioning. One issue with similar trajectories in both traditions is the nature of meaning. A (very rough) version of my narrative is the following: Husserl and Frege are working on the same issue from the same general position. They are both against psychologism as an explanation of why logical truths and mathematical truths are true. They both postulate some kind of abstract object which is meaning or carries meaning independently, by its very nature. Derrida, following Heidegger’s lead and Davidson, following Quine’s lead, attack these meanings.
Derrida’s first book was a critique of Husserl, arguing that Husserl, on his own account of the experience of time, could not construct presence. His next works develop consequences of there being no present-to-the-mind meanings. If there are no language-transcendent meanings, then the idea of Truth as the matching of meanings to reality is unhinged. There is no such thing as a fixity of meanings which grounds a single correct interpretation, and so on.
The progression from Carnap to Quine’s “Two Dogmas,” to Word and Object and then to Davidson seems to me an analogous trajectory. Davidson continues and purifies Quine’s project, dropping the idea of a given manifold and Quine’s scientism.
I read “Two Dogmas of Empiricism” as showing that the attempted denial of essences in the analytic tradition implicitly recognised meanings as one kind of entity which, by their very nature, determined their referents. Meanings as Fregean senses are quasi-linguistic items with Aristotelian essences. The kind of meaning that supports the analytic-synthetic distinction is an entity that means something in virtue of what it is. Carnap and other empiricists could only hope to ground meaning empirically with an unrealistic picture of how experience and belief interact. As Quine points out, Carnap’s Aufbau itself recognized the holistic relationship of experience to sentences. Quine’s argument against meanings is very like what Derrida calls “deconstruction.” Implicit in a view which seeks to overcome X is a commitment to X.
Derrida, responding to Husserl in somewhat the way Quine responds to Carnap, denies logoi just as Quine and Davidson reject Fregean senses. Derrida, being also a Heideggerian and French, places Husserlian noemata in a tradition that goes back to Plato—the distinction between logos and rhemata, between the genuine meaning (logos) of a word and other features of a word. Derrida questions this distinction. If we are skeptical about this distinction, then the critique of a discourse can focus on rhetorical features as well as what are considered ‘logical” features of the discourse. I have argued that Quine should agree. Since there are no meanings, the language cannot be divided into the truth-conditional (logical) and “other.” Quine did not pursue this line of thought. Derrida does, and his discussions of Plato illustrate what it would be like to really take there to be no clear “logical” core to a text.
Given their different traditions and backgrounds and interests, the consequences they explore are different. Derrida is interested in the topics and writers that are in his tradition. Derrida’s philosophical writing refers to a different set of background texts and different discussions, most of the time.
I never asked Derrida what he thought about Tarski or deflationism. His denial of “truth” seems to be based on the idea that, given Aristotle’s formulation of truth as correspondence, and given that there are no self-interpreting tokens, there cannot be “adequation of language and reality,” so that a consequence of denying meanings is a denial of truth.
I think that other possibilities were just not salient enough in his background of texts so that he could appreciate Davidson’s or Tarski’s or Quine’s conceptions of what ‘true” could mean beyond correspondence of a logos and Being. I think that the fact that Davidson takes truth to be primitive and fundamental, while Derrida takes the end of meaning to be the end of genuine truth does not undermine the fundamental agreement between the two.
Sellars’ Myth of the Given is actually very much a phenomenological position, rather than a view about how to think about language. Sellars in many ways is much more akin to the early Derrida’s critique of Husserl than either Quine’s or Davidson’s discussions. Davidson’s denial of the given is primarily an ontological view. There is a sequence of interesting essays to be written by someone on Husserl, Sellars and Derrida.
3:AM: So firstly, what are the similarities between Davidson and Derrida you find interesting? And how is Wittgenstein part of this ‘deconstruction’ mix you develop? You say it’s taken you longer to work out how Davidson and Derrida differ? Have you got it sorted now? What are the differences?
SW: The correspondences are between notions like indeterminacy of interpretation and Derrida’s “free play” and their shared denial that there is a given. Absent a language (the logoi, the language of thought, Fregean senses) whose words, by their very nature, refer to their referents, interpreting one system of signs in terms of another allows choices. That is, without a common thing that expressions from two systems express, there is nothing to go on but what people say when. So, there is slack.
Without a given domain of objects, another kind of fixity of meaning, “referring to the same object” is unavailable. Derrida in relation to Heidegger is somewhat akin to Davidson in relation to Quine, in dropping the remaining bit of realistic metaphysics.
Derrida and Davidson go in different directions. Davidson’s applications of his semantics and ontological views is largely on topics central to analytic philosophy, with some exceptions. Derrida, of course, primarily writes about topics in continental philosophy. He is interested in interpretations that go beyond the logical/truth-conditional. He takes the denial of logoi to mean that rhetorical features of a discourse can be just as important as what we would regard as “logical.” So, many of Derrida’s interpretations of, say Plato, rest on what we could regard as accidental features, for instance that “pharmakon” is both drug and poison.
The one figure I know they both thought about is J.L. Austin. Davidson spent several classes on Austin’s How to Do Things with Words in his 1967 class on philosophy of language. Derrida wrote Signature, Event, Context on the same text. They raise some of the same issues. Both Davidson and Derrida are suspicious of theories which ignore “marginal” cases. Davidson was particularly interested in arguing that what you can do with a sentence is not a good starting-point for semantics. Derrida and Davidson both point out that there is no way to fix a particular speech act to a sentence. Davidson’s notes that no sign can label something an assertion. Derrida makes a similar point about signatures.
Derrida may have read Davidson, since he did read my essay relating his views to indeterminacy of interpretation. I’m pretty sure Davidson never read Derrida. But Derrida would have very much admired Davidson’s essay on James Joyce, I think. Davidson would, if patient, have admired Derrida’s “White Mythology.”
3:AM: How is the analysis of metaphor and Davidson’s avoidance of the mysteries of Derridean Different and Differance important and how is De Man helpful?
SW: On metaphor: In my opinion, the two best essays in the twentieth century on metaphor are Davidson’s “What Metaphors Mean” and Derrida’s “White Mythology.” I am working on an essay which I hope will show that these two essays are complementary. There are some features of metaphor that Davidson chooses not to discuss that Derrida does discuss. Each could have learned from the other. In particular, I think Derrida’s reflections on free play could have convinced Davidson that dying metaphors reveal a kind of indeterminacy of interpretation different from any that he and Quine consider. From the other side, I think Davidson’s reflections in “On the Very Idea of a Conceptual Scheme,” could have convinced Derrida that differance, as pre-conceptual articulation which creates a domain to be sorted, can be dropped. I’ve argued that differance comes from Plato via Saussure, which is sort of interesting, given that in many other respects Plato and Derrida are antagonists. I would say that differance is a remnant of the given in Derrida’s thought.
For me, De Man is primarily someone else in a quite other tradition who thinks of metaphor as a rhetorical phenomenon. I wrote my one piece touching on De Man as kind of support for Davidson’s view. The theory in which this idea is imbedded seems to me to need sharpening. De Man makes large claims about self-referentiality, only some of which are plausible. I have argued that some of de Man’s ideas can be defended as legitimate applications of the great meta-theorems of twentieth century logic.
Wittgenstein, just to complete the question, seems to me to have affinities with both Davidson and Derrida. With Derrida, he thinks that philosophical accounts do not really work as theories. On the other hand, unlike Derrida, he thinks that the fact that a distinction cannot really be made precise is not a defect, but rather just the way language is. Derrida, on the other hand, seems to think that distinctions that cannot really be made to work in every case show something deeply wrong. Derrida takes distinctions that divide ambiguously not to be real distinctions.
Davidson is an improvement on both. With Wittgenstein, he takes seriously what people actually say and would say. He is much more appreciative of the later Wittgenstein than Quine is, for instance. On the other hand, Wittgenstein has no explanation of how we understand new sentences, which problem is the impetus of Davidson’s semantical program. Derrida, like Wittgenstein, does not appreciate that language, for all its vagaries and possibilities for innovation, has to be compositional. Davidson, in “A Nice Derangement of Epitaphs” and his essay on James Joyce, seems to me to get both aspects of language right.
3:AM: Quine vs Davidson is part of your analysis and you suggest that this Quinean idea of the given and Davidson’s arguments against it in his essay ‘The Very Idea of a Conceptual Scheme’ is a crucial debate. So what’s at stake and how does Quine argue that there is a given and that it’s no myth?
SW: What’s at stake is the reality of the intentional. Quine and Davidson agree with Chisholm that the intentional is not reducible to the physical. That’s what indeterminacy of translation/ interpretation essentially is. Quine and Davidson differ on what to make of that irreducibility. Quine takes it that in some sense, beliefs, desires, intentions, and the like are not real, and would be replaced in a completed physics. Thus, in particular, semantic notions do not really characterize reality.
Davidson takes the irreducibility of the intentional to the physical to be a reason to take truth to be primitive. Truth is the fundamental notion of the “intentional scheme,” by which I understand the inter-related system of predicates by which we think of ourselves and our actions. So, as Kemp points out in his recent book, Quine can be a pure deflationist about truth, whereas Davidson has to take truth to be a primitive notion.
The basic issue is naturalism. I take naturalism to be a monistic metaphysical realism, the view that reality is (close to) what the best physics says. For Quine, of course, there will be several equally good versions of the best physical theory at a time, but I take that not to essentially change the basic view that physics is the true arbiter of what is real. The attraction of naturalism is that it gives a unified account of everything by privileging one very successful way of thinking about the world.
My objection to naturalism is only that it is exclusivist. It seems obvious to me that many kinds of real objects, including the ones we care about most, make no sense whatsoever from the perspective of micro-particles. Only a metaphysical monist who thinks that somehow the reality of people and their beliefs is undermined by not fitting neatly into particle physics should think that people and micro-particles cannot be equally real. This kind of monism was the view I took in “Reference and Vagueness” the first “nihilist” essay on the sorites ever published. The issue bothered me for decades. I think Quine’s idea that there should be a single theory of everything into which all the kinds of things we think exist fits is a version of the Unity of Science project in which he grew up. I think it is very attractive but mistaken.
I am a fan of physics. My father was a physicist. As an anti-metaphysician, I am a firm believer in physics as the only worthwhile account of the physical. I think, for instance, that the nature of time is not something about which metaphysics, as opposed to physics, has anything useful to say. Likewise, whether reality could or could not be gunk, whether the physical world is hunks of matter, whether the world is one thing or many is unlikely to be advanced by intuition-mining and a priori argument. There is no intermediate science, called “metaphysics,” between the common sense conception of the world and physics. But that does not mean that the entities of physics are all the kinds of objects there are.
Davidson, Quine, and I agree that people coincide with physical systems and that there would not be people if there were no neurons and quarks. Where I disagree with Quine is whether that means that people are in some sense “really” physical systems. In the same way, feet coincide with meters, but that does not mean that feet are nothing but meters (nor vice-versa.) Davidson actually seems to me to be of two minds here. “Mental Events,” by making mental events strictly identical with physical events, in effect undermines his views in “On the Very Idea of a Conceptual Scheme.” I propose a fix in Chapter 4 of my new book, which takes these to be different but coinciding events with different modal properties.
On the particular textual issue of whether Davidson was right to lump Quine with the “sorters” of the given in “On the Very Idea of a Conceptual Scheme,” I don’t have any views I care much about. Quine denies that his appeal to stimulations is ontological, and he is certainly right to distinguish his view from the “sorting of a domain” view that relativisms suppose. In one domain, however, brain events and mental events, he does think that the brain events are what is really happening. If we ourselves are part of the world, then our evidence that we believe that it is Tuesday or that we are slightly hungry is based on stimulation. So, at least for the special field of brains and mental events, it seems that Quine takes there to be an ontological given. Stimulations so understood are events, not components of inferences.
Davidson criticises Quine’s term “posit” for the beings we accept on the basis of stimulation, according to Quine. I use the term “posit” to describe our articulation of the world by means of singular terms and predicates. It is not selecting a subset from a domain, as both Quine and Davidson take the term, but as coming up with a system of beings and features in order to think about a world that is not in itself articulated. Just as space requires the imposition of units in order to apply mathematics to it, so the world requires singular terms and predicates in order to be thought about. Different systems of predicates are successful for various things. Micro-particles are good beings to think in terms of for many projects; not so good for others.
My interpretation of both Quine and Davidson treats them as quasi-Kantians. A system of objects and properties is a requirement of thought. Singular terms and predicates are a requirement if we are to be able to make inferences that depend on sub-sentential structure, that is, to think. One could hold that by a happy accident the world itself is articulated into beings and properties. However, the objects and properties we naturally arrive at, organisms and medium-sized objects, bear little relation to what physics appears to tell us is the basic articulation of the world. So, a more likely explanation of the match between the world and thought is that beings and properties, that is singular terms and predicates, are something we impose on the world. The world itself has no intrinsic articulation into objects and properties.
This is the core of the view I call “relative essentialism,” which gets both Kripke and Davidson to be close to correct. The germ of the view is in Quine and Davidson, I would argue. From the perspective of this view, many metaphysical issues are empty.
3:AM: How is this different from Derrida’s response?
SW: I can’t translate the issue between Quine and Davidson into anything I’m confident Derrida thought. I’m not sure Derrida ever responded to any analytic philosophers other than Austin and Searle.
3:AM: Your new book is about continuing the Davidsonian anti-metaphysical program isn’t it? But Quine would argue that, for example in the area of Montague’s linguistic semantics, possible worlds and exotic functions fail the anti-metaphysical test. So has he and so the austere Davidsonian program either been betrayed or not even tried? Can you say something about what the new Davidsonian work needs to be doing? And why wasn’t his approach been taken up further?
SW: “Betrayal” is not the category I apply to these developments. Exotic objects don’t disqualify a person from being a Quinean. Consider David Lewis. He has perhaps the most exotic ontology around, but the arguments he gives for it are absolutely Quinean. Postulating possible worlds is justified by what you can explain with them. This is why Quine accepts numbers and sets, also pretty mysterious things. David was quite Quinean and, in a sense, anti-metaphysical in rejecting necessary relations among independent objects, primitive modality, and the like.
There are two “anti-metaphysical” strands in Quine and, in purified form, Davidson. One is the rejection of an intrinsic articulation of the world into beings. I talked about that above. The second is the theory of predication that Davidson formulates in his last book, Truth and Predication, together with his early remarks on semantics as logical form. His idea in 1967 was that a theorist should only assign a logical form to a construction if the construction could not be accounted for by information special to a predicate. His idea was an application of Quine’s idea that the mark of the logical is indifference to particular predicates.
I also think there are fewer real divergences between Montague students and Davidson than some think. The followers of Montague use possible worlds, but I think most of them would not take them to be serious posits, but rather heuristic devices to talk about how things would be, and the like. As linguists, they would probably defer to Stalnaker or other “ersatzists” for what these objects actually are. I argue in the new book that you can understand modals better by using Davidson’s “On Saying That,” essentially going back to Carnap’s idea in The Logical Syntax of Language. Given that you are going to try to construct possible worlds out of sets of propositions, you might as well do that directly. I think Carnap would have liked relative essentialism as an explication of his “Principle of Tolerance.”
The “Davidson program” was primarily a program in semantics with a goal very much akin to Montague’s. Both thinkers realised that English, for instance, had to be a formal language. As far as basic semantics goes, Davidson and Montague are both anti-metaphysical, and that continues. Heim and Kratzer’s Montagovian textbook, for instance, takes the truth conditions of “smokes” to be in effect equivalent to “`smokes’ is true of an object a if and only if a smokes. This is disquotation. There is no metaphysics of truth-makers.
“Truth-maker” accounts of predication, that there are pieces of the world corresponding to sentences, is alongside “being monism” the main source of metaphysical mysteries. Try to formulate McTaggart’s worry about “past,” “present” and “future” applying to an unchanging moment without properties as designata of predicates. Try to formulate a difficulty about how a single thing, Redness, can be present in numerically distinct objects without treating predication as requiring something beyond “’Red’ is true of Joe if and only if Joe is red.” It is sort of sad to lose these puzzles. It is as if the Guardian stopped publishing cryptic crosswords.
My objection to much of contemporary semantics is primarily that it mis-labels theories about predicates as parts of semantics. Other than that, I learn a lot from reading Montagovian accounts of constructions that interest me. For instance, Angelika Kratzer’s view on “if,” translated into Davidsonian, is close to mine.
Likewise, my objection to “logics” of this and that (deontic, temporal, etc.) is that they similarly mis-label theories about predicates as “logics” of subject matters. I agree with Kripke that logic does not have a useful plural. I agree with Lewis, for instance about “modal logic.” The “modal logics” can be treated as first-order theories just by positing possible worlds and proposing various R-relations. If we accept modal logic as a distinct kind of logic, we may as well have a “logic” of Labrador Retrievers which includes axioms such as “All Labrador Retrievers are dogs.”
I propose an adaptation of Davidson’s “On Saying That” as a simpler and more direct account of the modals, bypassing possible worlds. This yields a version of Carnap’s view that modality is about essentially linguistic items. A number of people have claimed that Davidson’s idea in that paper cannot work. Two of my students, Daniel Blair and Nilanjan Bhowmick, have shown, at least to my satisfaction, that the objections can be overcome. Whether Davidson’s original formulation is exactly right is less important than his way of responding to the Church translation objection to treating intensional contexts as linguistic.
Quine is wrong to think he can get along without modal notions. I argue that positing objects is implicitly positing identity and survival conditions, and that both Quine and Davidson are thus in some sense essentialists. However, unless one thinks that positing one kind of object as real somehow gets in the way of positing other overlapping objects as also real, there is no problem in saying that there are modal truths. But the modal truths depend on what the posit is. That is, for instance, the Battle of Stalingrad could have gone slightly differently but still be the same battle, whereas the space-time worm of events around the Volga at the time would be a different mereological sum if any events in it had been different. These are different coinciding objects with different modal features. They are both real.
There is no reason whatsoever to require a place in natural science as the criterion of real existence. We no longer believe that “Earth” is a fundamental component of reality, but we still believe that bulldozers move earth. There is such a thing as “earth” whether or not it has an account in terms of basic scientific concepts. Beliefs, Chevy Impalas, and nations are as real as quarks and black holes.
Davidson, being committed to the truth of most of what most people believe, has to have some account of things like “earth,” “has to,” “we hold these propositions to be self-evident,” “there are many properties he and I share,” and the like. He did not get around to addressing this kind of stuff. My recent book supplies accounts of properties, propositions, and facts that are close to what he would have provided if pressed.
It sometimes did not occur to Davidson that things he thought were obvious were not obvious to others. When he was working on his last book, on truth and predication, he told me that he had just not realised that many people did not see that there is no metaphysical issue about predication—that many people did not see as obvious that all you need to say about predication is that “is a frog” is true of Fred just in case Fred is a frog. I think there are similar reasons why he never said what “You must not turn left at red lights” and “You can’t get blood from a turnip” mean. His minimalist semantics, with no truth-makers and no semantically relevant notion of “analyticity” has lots of consequences that he never worked out. Some of the chapters of my book consist of applying the idea that “ `is a frog’ is true of Fred just in case Fred is a frog” to other puzzles, such as how it is possible for a moment to have the properties pastness, presentness and futurity without changing.
3:AM: How far have your views about Realism and metaphysics changed your mind about Davidson? Kripke’s arguments have been important haven’t they in suggesting metaphysical realism is ok – and you’re an Aristotelian of sorts aren’t you? So why isn’t Davidson’s program a dead end?
SW: I attended Kripke’s lectures at Princeton in 1970, and was immediately persuaded. Aristotle suddenly seemed so interesting that I spent three years studying Greek in order to be able to really understand him. For a decade or so in the seventies and eighties I was very attracted to metaphysical realism. I published a number of papers that presupposed realism, including some notorious “nihilist” essays on the sorites.
The sorites argument seemed to show that real scientific realism would have no room for people, tables, or Chevy Impalas as genuine objects. This seemed to be a crazy result, but a consequence of taking essentialism seriously and interpreting “essence” as grounded in natural law, as Aristotle did. Sometime in the mid-nineties it dawned on me that Davidson and Quine were themselves some kind of relative essentialists and that Kripke’s examples should not have been persuasive. The very stuff making up the table in the Woodrow Wilson School could have been arranged to in fact be quite dirty ice. Kripke selected “wood” as what the stuff itself was, but the same material was also a collection of micro-particles which happened to be arranged in wood configuration. I realised that the notion of “the thing itself” makes little sense apart from a description. This is wood, it is also an aggregate of micro-particles, and also a table. Relative to each of these characterisations, there are existence and extinction conditions, relative Aristotelian essences. Different changes make it cease to be a table, from those that make it cease to be wood, and those that make it cease to be an aggregate of particles. But there are modal truths about wood, tables, and aggregates of micro-particles as such. They all are real. Only a metaphysical monism could persuade a person otherwise. Aristotle’s difficulty about parts of substances, that would make corpuscles not be full beings, illustrates one reason not to be a monist. I should mention that Epicurus is probably the originator of “relative essentialism.”
I think Quine was blinded to the modal implications of his notion of a “posit.” My central thesis is that both Davidson and Kripke are right, more or less. Things have essences, there are necessary truths about things, but any thing is many different things. Chapters 2-4 of the new book make this claim and illustrate its consequences by arguing for a revision of Davidson’s view in “Mental Events.” Briefly, the correct view of the mental in relation to the physical should be called “innocuous dualism” rather than “anomalous monism.”
So, to summarise: I lost faith in the Davidson program for a while, seduced by Saul, but then found a way to make both philosophers more or less right.
3:AM: Are you saying the Davidson program can be made to be compatible with the Kripkean and Aristotelian intuitions?
SW: Yes, except for the “monistic” part of Kripke’s and Aristotle’s realisms. Aristotle in Metaphysics Z13 says that no substance contains another substance as a proper part. His reasons seem to be that he regards a substance as determining the “what it is” of the whole region, so that a human, for instance, cannot also be an aggregate of particles. These reasons are not compelling, and his view that medium-sized objects and organisms are scientifically basic was a reason early modern science abandoned Aristotle and preferred Lucretius.
Aristotle and Kripke are both committed to the notion that there is such a thing as “the entity itself,” having an essence that fills a whole region with it and nothing else. That is, the world divides into objects in a single way. I argue that this is wrong, and that dropping it allows different modal truths about overlapping, even coinciding objects.
The result is a kind of essentialism that accommodates Kripkean intuitions, while observing Davidson’s dictum that “sameness is always relative to a predicate.”
3:AM: You argue that what’s crucial for the Davidson program is his account of truth and predication. Why do you think a disquotational truth-definition is the only proper semantics – and why is this so crucial?
SW: The argument that this is the only proper semantics rests on doubts that there is anything for metaphysics to add to the sum of disquotation and real-world information. Consider “Joe is a frog.” There are a lot of consequences of this being true. If Joe is a frog, he must live in an environment where there is water, must be an amphibian, cannot have the same DNA as a dandelion, and so forth. There is a scientific specialty devoted to amphibians, of which frog-science is a part. But there is also the simple consequence: If “is a frog” is true of Joe, then Joe is a frog. So there is a very expansive set of truth-conditions, uncovered by the on-going projects of biologists, as well as the “simple aitia” of (in the homophonic case) disquotation.
A metaphysical account of predication, for instance, in terms of truth-makers, imagines that there is another, intermediate science between natural science and “`is a frog’ is true of Joe just in case Joe is a frog.” I think this imagining is pure fantasy and bullshit.
I don’t think contemporary linguistic semantics is bullshit. Much contemporary semantics consists of theories about the application-conditions of predicates. This work accurately theorises about the application-conditions of a variety of predicates. For instance, Barwise and Cooper illuminate the truth-conditions of the quantifiers, as does subsequent work. My only objection is really merely “semantic”—what this kind of semantics is doing should be categorised as “theory” rather than “semantics.” Quine’s notion of logic as indifference to predicates seems just right to me. A consequence is that, if, for instance, the quantifiers are predicates of pairs of sets, then the difference between when “all” holds between a pair of sets and when “most” does is explained by the truth-conditions of two different predicates. That is, it falls into the category “theory,” just as the information that “mother of” is not symmetric is not part of the semantics of “mother of” but part of the theory associated with this predicate.
The basic idea is that semantics is logical form, and that logical forms should be ascribed only when for some reason information about the truth-conditions of a predicate will not explain peoples’ ability to make inferences. The inference from “Joe is a tall man” to “Joe is a man” has to be formal, whereas the inference from “Joe is a man” to “Joe is an animal” is explained by a piece of general knowledge that humans are animals.
3:AM: Of all the different areas for the Davidsonian program you’re interested in moving forward the ethical theory is one that seems an unlikely area for Davidson – he didn’t do much about this area did he? So can you say how you think your Davidsonian program ought to handle ethics?
SW: Remember that Davidson characterised interpretation as treating agents as “believers in the true and lovers of the good.” So, “reasonable values” as well as “reasonable beliefs” was something he thought should be maximized in interpretation. He was always interested in, and used, decision-theory as a model of his view of interpretation. On a couple of occasions Davidson did discuss “ethical” concepts. He has an essay arguing that sentences using “good” have truth-values, which of course is the natural view for a philosopher proposing a truth-conditional semantics.
Buried in “How is Weakness of the Will Possible” is an account of ethical reasoning that treats it as essentially inductive, rather than deductive. So, what a person ought to do can be very different from what a person is obligated to do. If “ought” is like conditional probability, as he argues there, then new data can turn a very strong argument that A ought to do B into a very weak argument, just as new information can turn an argument that Joe is the murderer from strong to weak.
This extension of Davidson’s ideas about truth-conditions to ethical notions was the project of my dissertation (Princeton, 1970). Since then my thoughts on the topic have evolved. However, the basic idea that constraints on interpretation (maximise true belief, maximise correct evaluation) can provide the basis for judgments of what one ought to do is basically the same between 1969 and 2013. The fundamental idea is that the “logic” of “ought” is like the “logic” of conditional probability. That is also the logic of Carol Gilligan’s “care perspective.”
3:AM: If you were to summarise the situation regarding this new Davidsonian program you write about, where do you think we are at the moment and what do you think is at stake if it develops as you think it might?
SW: The situation at the moment, as far as I know, is that I’m working on various neo-Davidsonian projects. I think Kirk Ludwig, Ernie Lepore, and maybe some other folks out there I don’t know about, are doing similar things. I hope people will read my book, realise that they could come up with better Davidsonian accounts of, for instance facts, “can,” comparative adjectives, propositional attitudes, and so forth, and start publishing. Meanwhile some of these converts could attack various metaphysical views.
That’s fantasy. What I’m sure about is that I’m continuing to work on the project. Since publishing the book, I have come up with an improved theory of “ought” and other modals. I’ve found (I think) a way to think about counterfactual conditionals that de-mystifies them.
3:AM: ‘Theory’ has been treated with deep suspicion by many philosophers and scientists – does your Davidsonian position help defend Theory or are suspicions justified? Is this no longer a live issue or are there still things we can learn from your analysis?
SW: I sat in on a lot of “theory” courses at Yale, and spent part of a summer at the School for Criticism and Theory. My impression at the time was that most literary theory writing is very good at close reading, but not so good at theory construction. The insights seemed to be literary rather than theoretical insights. I remember being vastly impressed in Geoffrey Hartman classes, primarily because he was an exceptional reader and interpreter. I never did figure out what exactly literary theory was supposed to do.
I’ve come around to an intentionalism which basically agrees with Knapp and Michaels. As it happens, a paper Bill Tolhurst and I published in 1978 provides a kind of foundation for a version of intentionalism about the meaning of literary texts.
3:AM: And finally, are there five books you could recommend to readers here at 3:AM that would help us understand better your philosophical world?
SW: Besides my recent book, Neo-Davidsonian Metaphysics, Davidson’s Truth and Predication, Quine’s Word and Object, Aristotle’s Metaphysics, Kripke’s Naming and Necessity, Derrrida’s Speech and Phenomena.
ABOUT THE INTERVIEWER
Richard Marshall is still biding his time.
First published in 3:AM Magazine: Friday, April 25th, 2014.