Richard Marshall interviews Robert Stalnaker.

[Photo: Steve Pike]

Robert Stalnaker is the grandmaster flash of contemporary metaphysicals. He thinks that a language-first approach to philosophy is ludicrous, Paul Grice an inspiration and Saul Kripke very important to his early thinking. He broods on issues about internalist and externalist doctrines and approaches, on our knowledge of the external world, about the nature of phenomenal knowedge, about the view from nowhere, the opacity of transparency, contextualism, relativism, possible worlds, the entanglement of semantics with metaphysics, haecceitism and the beauty of metaphysical theories, amongst other things. He is currently on a phased retirement at MIT and becoming a Visiting Professor at Columbia. He is simply a modern daddy of the mac!Chillin'!

3:AM:What made you become a philosopher? You’ve been a big player in many fields of the philosophical landscape which could mean that, like a fox, you have many independent questions – or it could mean that in order to get an answer to one big question you need to approach it from different angles, like a hedgehog. So are you a foxy or hedgehoggy philosopher?

Robert Stalnaker:That’s an interesting question. I certainly started off in a foxy way, reacting to things I read about different issues. First (in high school) I got interested in questions about the nature of mathematics from readings I happened upon and grasped only dimly. Mathematical Platonism seemed to me a crazy view, but a couple of papers by Ernest Nagel and Carl Hempel (who was later my graduate school teacher) struck me at the time as the unalloyed truth. When I got to college and took a yearlong history of philosophy class in my freshman year, a new intellectual world opened up, and my enthusiasm for the logical positivist picture dimmed. My teacher was an advocate for each philosopher we read, and I enjoyed getting into the different ways of thinking. St. Augustine was my favorite, even though I had and have no inclination to take religion seriously. Spinoza was another favorite. I decided that the positivist stuff that had impressed me in high school was narrow-minded and superficial. As I read more philosophy later in college, I went back and forth in my thinking about more expansive and more restrictive approaches – an ambivalence that I still have.

I got interested in the philosophy of history in college, under the influence of a wonderful teacher, Louis Mink, and that later became the area of my dissertation. But in the end I turned to more general issues in the philosophy of language and mind. Saul Kripke, who visited Princeton and taught a seminar in my last year of graduate school, was a big influence. The seminar was officially about Wittgenstein on rule following, but this was mixed with the presentation of ideas that later became Naming and Necessity, and I was convinced, both by many of the substantive points, and by the power of Kripke’s way of thinking. I had some thoughts about counterfactual conditionals as I was finishing my dissertation and the ideas about possible worlds that Kripke was presenting helped me to see how to make them more precise. My first job after graduate school was at Yale, where the faculty of a large and dysfunctional department consisted mostly of an assortment of phenomenologists, Whiteheadean metaphysicians, eccentrics of other persuasions, and logicians. I gravitated to the logicians, and collaborated with Rich Thomason in developing a semantics for conditionals.

So my way into philosophy was foxy, but a general picture inevitably emerges when one thinks about a lot of philosophical problem. One finds oneself reflecting on how it all hangs together, noticing recurrent themes and ways that one’s views reinforce each other. I like, and have quoted, a remark that David Lewis made in the introduction to his first collection of papers: “I should have liked to be a piecemeal, unsystematic philosopher, offering independent proposals on a variety of topics. It was not to be.” He tried to be a fox, but couldn’t help turning into a hedgehog. I suspect that the best hedgehogs – those that are not blinded by their big idea – come to their systematic theory in something like this way, and that the best foxes see the interconnections between their different projects, and the general ideas that motivate them, even if they resist building them into one grand system.

3:AM:By not assuming the priority of linguistic over mental matters, where a philosophy of thought is best approached through philosophy of language, you were taken to be a heretic of analytic philosophy by Michael Dummett. (There were quite a few of you!) Dummett didn’t think it actually made all that much difference in the end to what you did, but you disagreed with his evaluation of the situation. So why was Dummett wrong?

RS:When I first read Dummett’s remark (that philosophers who reject the priority of language over thought are “overturning the fundamental axiom of all analytic philosophy and hence have ceased to be analytic philosophers”), I thought it must be at least half a joke, since the idea that a philosophical tradition as diverse as the one labeled “analytic philosophy” should have a defining dogma – a litmus test for membership – is ludicrous on its face. But it is true that the “language-first” approach to the explanation of speech and thought has been a recurrent theme in analytic philosophy, and it is a good question whether and why it matters. I think the linguistic picture has distorted the discussion in many different ways. Let me mention just three:

First, the most central problem about speech and thought is the problem of intentionality – the problem of explaining how our words and thoughts, and more generally our representational states can be about the world. It is not just human beings who represent the world – who take in information and use it to guide their behavior. Perception and memory are capacities of primitive animals as well as humans, and I think the best way to understand the sophisticated kind of representation that is manifested in human inquiry, deliberation and communication is to see it as a special case of a more general phenomenon, and continuous with it. The “language-first” approach takes rational, articulate, sophisticated adult human beings as the paradigm representors, with the representational capacities of the rest understood by analogy with them. I think it is more fruitful to start with the simplest cases, and understand human cognitive capacities as more complex systems of essentially the same kind. By starting with language, we over-intellectualize perception and cognition, which I think results in a distorted picture, not only of the cognitive capacities of nonhuman animals, but also of human knowledge and communication.

Second, even if our main concern is to understand human speech, and the more theoretical aspects of thought that essentially involve language, I think it is important to start with an account of representation that is independent of language. Our languages are all highly context-dependent in the sense that the content of what we say depends on the contexts in which we say it, and speech contexts are best explained in terms of the attitudes, intentions and expectations of the parties to the conversation. If this is right, then linguistic content depends on the contents of the states of mind that define the contexts in which linguistic content is expressed. The language-first approach presupposes that we have a language – perhaps a language of thought – that has an autonomous semantics – a semantics that can explain the relation between the expressions of the language and their content or meaning independently of context. I am doubtful that there could be such a language.

Third, speech comes in units – sentences – with complex compositional structure that explains how they have the content that they have. A linguistic model of thought tends to assume that a belief state is appropriately modeled by a set of sentences with this kind of structure, mental sentences or mental representations of natural language sentences that describes the way the world is according to the believer. But the idea that mental representation should take this form is a speculative hypothesis about cognitive architecture, and one that I think has little support or plausibility. It seems reasonable to think that we represent the world in a diverse mix of ways – with images and maps and procedural organization, as well as with mental structures that are sentence-like. Psychology and neuroscience will eventually tell us more about the mechanisms of mental representation, but we can and should separate questions about what it is to represent, and about how we represent the world to be from questions about the means that our brains use to do it. The linguistic picture of thought tends to blur this line.

3:AM:Paul Grice was influential in your rejection of the priority of language because he made you see that speech was action that required explanation in terms of beliefs and purposes wasn’t he? Who else was helpful to you as you set off?

RS:Right – Grice was a major early influence on my thinking about intentionality and the structure of discourse. The influence continues, as will be evident in my forthcoming book on context. It was not just his emphasis on the fact that speech is a kind of action that was distinctive to his approach – that emphasis was common to Grice and others such as J. L. Austin in the British “ordinary language” tradition. The central issue that divided Grice from Austin was the question whether communicative action essentially involved a conventional, rule-governed practice. Grice’s aim was to characterize communicative action, in general, in terms of beliefs and intentions that were independent of any conventional practice with constitutive rules. Grice’s motivation, which I share, was to provide for the conceptual separation of an account of the functions of language from the means that languages use to serve those functions. For this, one needs an account of what it is for a speaker to mean something (to perform a communicative action) that explains this notion independently of the rules a device (a language) whose function is to facilitate the performance of that kind of action. My forays into pragmatics have all been attempts to develop this kind of Gricean program.

As mentioned earlier, Kripke was very important early influence on my thinking, both about intentionality and about metaphysical issues. Another early influence was the work of Keith Donnellan. My first paper on pragmatics was in large part an attempt to provide a framework for representing his distinction between referential and attributive definite descriptions.

3:AM:You’ve been a big player in philosophy of mind and have made significant arguments against the internalist view. Can you set out what you think the debate between internalists and externalists is about, why its important and why non-philosophers ought to be interested in getting the answer right (if you do)?

RS:There are different internalist and externalist doctrines and approaches that play a role in ethics and epistemology as well as the philosophy of mind, but there is a common theme that I think is important, and that I tried to articulate in the first chapter of my book, Our knowledge of the Internal World. The externalist approach is to try to understand ourselves as thinkers, experiencers and agents from the outside, as things in the world that we find ourselves in. Even if our main concern is with questions that arise from the first-person perspective (what shall I believe? what do I know? how should I act? what must I do?), it is fruitful to ask, from a third-person perspective, what it is to be a creature with a first-person perspective, to be someone who has the capacity for thought, experience, knowledge and rational action. Externalism in this broad sense is a strategy for formulating philosophical questions, and not a philosophical thesis, but it is a strategy that helps to motivate specific theses in epistemology, ethics and the philosophy of mind to which the label “externalism” has been applied. For example if you start by asking from an external standpoint what it is for a thing in the world to be a knower, it is natural to answer that it is (at least) to be a creature that has the capacity to take in information about the world – to be in states that tend to be sensitive in specific ways to the state of its environment. The picture suggests an explanation of the content of thought (or more generally of the representational states that are, when things go right, states of knowledge) in terms of the features of the environment that those states tend to be sensitive to. This conception of content is one that conforms to the more specific externalist thesis that was summed up in Hilary Putnam’s notorious slogan, “meanings just ain’t in the head.” More soberly, the thesis is that the content of thought is determined, not just by the intrinsic states of the subject, but also in part by the features of the subject’s environment. And this general picture of knowledge and content suggests that there is no way to factor propositional content (the subject’s conception of the world) into a purely internal component (the narrow content of the thought) and an environmental component.

Why does this kind of externalism about intentional content matter? Because it is incompatible with a seductive foundationalist picture according to which we can form, a priori and independently of any presuppositions about contingent fact, a conception of the space of possibilities in which we locate our world. It would be nice if we could separate the question, what are the alternative possibilities, from the question, which of them is actual, but I think it is a utopian fantasy to think that we can do so. The externalist picture suggests that we can articulate and understand the possibilities that we distinguish between in thought only with the materials – the things, events, properties and relations – that we find in the actual world, and we can’t know what those materials are without doing empirical inquiry. If this is right, then even the most basic inquiry will have factual presuppositions. We have to muddle through, bootstrapping our way to a conception of what the world is like.

3:AM:You draw an analogy between knowledge of phenomenal experiences (eg seeing red) and self locating knowledge to develop your version of externalism don’t you? Could you sketch out the basic contours of your theory?

RS:A number of philosophers have been struck by an analogy between Frank Jackson’sknowledge argument against materialism and John Perry’sargument that indexical or self-locating knowledge is irreducible to knowledge of impersonal fact about what the world is like in itself. In both cases, it was argued that one could know all the facts of a certain kind (physical facts in the case of the knowledge argument, impersonal facts in the case of the argument for the essential indexical.) while not knowing facts of a different kind. But there was also a striking disanalogy between the two arguments: Jackson’s argument draws a metaphysical conclusion (that a materialist metaphysics leaves out certain facts about the objective world), while Perry’s argument concluded that a certain kind of knowledge didn’t concern the objective world at all, but was about the knower’s perspective on the world. The hope was that the analogy could help us to see how to reconcile materialism with the knowledge argument by seeing phenomenal knowledge – knowledge of “what it is like” – as a kind of knowledge of one’s perspective on the world, and not of what the world is like in itself. But reflecting on how to do this required, I thought, rethinking how to understand perspectival or self-locating knowledge. I see the perspectival feature of a state of belief, not as a feature built into the content of thought, but as a way of connecting the content of a thought to the thinker who is thinking it. It is a crucial feature of the representational content of thought and speech that it be public and communicable (a feature that Frege emphasized). When I tell you who I am (say by introducing myself at a party), if you understand me, then the information you acquire is the information I express. The information is self-locating for me, but not for you. The problem of self-locating communication (really, of communication generally) is the problem of calibrating the contents of the thoughts and speech acts of different subjects – of seeing how each is locating him or herself, and the other, in an objective world.

So what does all this have to do with phenomenal experience? It is not that getting clear about the relation between an objective conception of the world and a subject’s perspective on it automatically provides some solution to the puzzle that Frank Jackson posed. The starting point for the connection is just the idea that phenomenal experience is essentially perspectival. For the materialist, the state I am in when I am having a particular kind of experience – seeing red, say – is a physiological state, but knowing what it is like to see red requires a first-personal way of identifying that state, just as knowing that it is now noon requires a first-personal way of identifying a certain time.

3:AM:Are you saying that phenomenal knowledge and intentionality – all our mental goings on – don’t give us more information about the world as such, but just tell us where we are?

RS:No, though I admit that what I have just said may seem to suggest that. While the information expressed by saying “seeing red is like this” is not detachable from the context in which it is expressed or thought, it is (I want to say) a piece of information about what the world is like, which means that we understand the content of the thought by the way it distinguishes between alternative possibilities. The analogy with simple self-locating thoughts is intended to make it less mysterious how a thought can be both “information about the world as such” and also essentially about where we are in the world. When Dorothy says “We’re not in Kansas anymore”, she is contrasting two objective possibilities, and ruling one of them out. What is ruled out is a possibilityin which a particular person who was in Kansas, and is now having certain particular experiences in still in Kansas. The thought is self-locating because the person who is still in Kansas in the possible situation being ruled out is the same as the one who is thinking the thought.

3:AM:Is phenomenal knowledge just a species of self-locating knowledge, or is knowledge that is self locating the essence of intentionality?

RS:I want to agree that phenomenal knowledge is a species of self-locating knowledge in the sense that phenomenal knowledge is essentially first-personal – it is knowledge of what it is like for me to have a certain kind of experience. But I am also saying that in calling knowledge self-locating, or first personal, I am not talking about the content of the knowledge, but describing a distinctive way in which a knower is related to the content of what she knows. I also want to say that we can’t make sense of any kind of knowledge without locating the knower both in the actual world, and in the worlds that are compatible with the way the world is taken to be. This is because all knowledge is about the knower’s relation to other things in his or her world. In that sense, I can agree that self-locating is part of the essence of intentionality.

3:AM:Is your externalist argument intended to defend it against charges from the internalist view that to be an externalist about intentions is to set up an objectivist stance, a view from nowhere?

RS:“The view from nowhere” is Tom Nagel’s phrase, but the idea is an old one that has taken many forms: There is the God’s eye view, which sees the world from outside of space and time. There is the omniscient narrator who tells a story from no particular perspective. There is something real that these metaphors are trying to get at, but they do it by mixing two incompatible ideas: first, there is the idea that we can make sense of a conception of the way the world is or might be in itself, independently of anyone representing the world as being that way; second, there is the idea that every conception of the way the world is needs a conceiver – someone who is both the one doing the representing, and also a part of what is being represented. One combines the two ideas by imagining an observer who part of the story, but who also stands outside of it, seeing the world it describes from a perspectiveless perspective. We can agree that all stories that are told (both fiction and nonfiction) are told by someone who tells the story from some point of view, but the teller of the tale need not him or herself be a character in the story, or a person in the world being described. Even when one is trying to tell the truth about the actual world – the world that one inhabits – the content of the story one telling need not identify a perspective from which it is being told. Stories (whether fact or fiction) must have authors, and they may in addition have narrators (omniscient or otherwise, more or less intrusive), but they need not. An author who is not herself in the story may just tell us what happens in it. One might think of “the view from nowhere” just as the perspective of a teller of a tale who is not a character in the tale, or who does not identify herself, in the story, as the teller of the story. But what is misleading about the image is the idea that this absence of a perspective, as a part of the content of a representation, is a distinctive kind of perspective.

I think what you mean by my “externalist argument” is the argument that one can approach questions about the world and our place in it (including questions about our perspective on it) in this way – by asking what a world must be like in itself in order for creatures like us to exist and have the distinctive capacities for experience and thought that we have. If all that is meant by viewing the world “from nowhere” is conceiving of it in this way, then I see nothing objectionable about it. We can study perspectives as features of the world as it is in itself, but they will all be perspectives from somewhere.

3:AM: Contextualism and perspectivalism are key ideas you use to draw the important distinction between internalism and externalism in terms of how to understand transparency, which is the term of art used by philosophers to talk about the supposed direct contact with our own minds. You want to remove transparency as a true picture of intentionality don’t you?

RS:As you say, “transparency” is a term of art, but the fact that it is a technical term does not mean that it is clear what it means. You are right that I want to reject the idea of some kind of direct contact that we have with the contents of our own minds, but it would seem paradoxical to deny that our thoughts are transparent in this sense: that when we entertain a thought, or form a belief, we know what the content of that thought is. This truism needs interpretation, but we need to make room for a sense in which it is true. I agree with philosophers such as Donald Davidson and Tyler Burge that the thesis that we necessarily know what we are thinking when we think it can be clarified and reconciled with an externalist account of content, and with a rejection of intentional foundationalism. There are interesting problem cases, and the reconciliation brings out the ways in which the ascription of content is dependent on context, or so I argue.

3:AM:Contextualismis a given a big role not just in theories of phenomenal knowledge but in terms of meaning generally isn’t it? Does your view of why context matters to meaning amount to saying not just that context is useful but actually indispensible?

RS:That is exactly right. The general picture I have been sketching implies that context-dependence is not just a convenient feature of the languages we happen to speak, but is essential to both speech and thought. Knowledge is essentially contrastive; it is a capacity to distinguish the way the world is from certain alternative ways that it might have been, and we can inquire and debate about what the world is like only in a context that presupposes that the world provides us with certain resources for characterizing the possibilities we are distinguishing between. Some kinds of context dependence are more superficial. In these cases, context mediates between linguistic expressions and the propositions they are used to express, and the propositions themselves could be expressed in different contexts, using different mean. Alice says, pointing at Michael Jordan, “He is a great player”. She can say what she says with that sentence only because of certain ephemeral features of her local context, but what she said can be detached from the context and said or reported in a different way. “Alice said that Michael Jordan was a great player.” But the propositions themselves are ways of distinguishing between possibilities, and we can’t identify the possibilities without drawing on what the actual world offers us. We aim, in our conceptual and empirical inquiries, to move from the local contexts provided by our immediate environment and idiosyncratic knowledge of it to larger and more robust contexts, but we never find a platform outside of the world from which to describe it.

3:AM:Why doesn’t this threaten to be a kind of relativist stance?

RS:It does threaten to lead to a kind of relativism, or anti-realism. The picture suggests that learning about – even thinking about – the real world is hard, and more disorderly than one might hope. The challenge to reconcile realism with radical contextualism is real and important, but I think it can be met. In any case, the right response is not to opt for an indefensible picture of our conceptual and epistemic connection to the world.

There is a philosophical problem that needs to be addressed, but the threat is not just an abstract philosophical concern. The contextualist picture also points to a practical threat that is worth worrying about. The contexts in which discourse and inquiry take place can be, and are, manipulated in ways that distort the outcome. If, as I believe, we can make sense of rational discourse, deliberation and inquiry, only in a given context which involves substantive presuppositions, we face a daunting challenge when the contexts we find ourselves in are skewed – when the basic presuppositions that define the context are false. When disagreements are deep, or when one judges that our whole way of looking at things is radically mistaken, we need to find our way into a new context, and there may be no neutral way to do so. But we have rich and diverse resources for talking and thinking about the world and for deciding what we must do, and even if there is no absolutely neutral set of rules governing rational activity, and no safe platform where we are guaranteed to find common ground on which to settle our disagreements and find the truth, with good will we can usually find a way to get to a place where we can understand each other, and engage in what we can agree is rational debate.

The more philosophical challenge, for a realist, is to explain the resources we use to do this, and to justify the claim the methods we use are ways of discovering objective truths.

3:AM:Is it in relation to context that possible world theorybecomes important here? Why are possible worlds required and are these possible worlds to be understood in terms of the philosophy of language, or metaphysics? I guess this asks whether the possible worlds theory here is about a metaphysical theory distinguishing possibilities or a semantic theory about how we represent those possibilities. Can we disentangle metaphysics from semantics?

RS:Yes, I do think that possible world theory is a helpful tool for characterizing context, and for answering the challenge to realism. I talked a lot earlier about the relation between a perspective on the world and a conception of the world as it is in itself, arguing for a way of understanding propositional content in terms of alternative ways that a world might be in itself. Our contexts are defined by our common perspective on the world – by our shared presuppositions – but I want to represent the content of our presuppositions by sets of possible worlds – ways a world might be in itself. Even our most local contexts – alternative states of our immediate environment, and of ourselves – are ways of distinguishing objective possibilities. We use ourselves and our local environment to identify and to fix reference to the relevant objective possibilities that we are distinguishing between. The point of all this fuss about perspectives and worlds as they are in themselves is motivated by the concern to clarify and defend the claim that the contextualist picture I sketch is compatible with realism. But the defense will work only if we have a conception of a possible world as something that is independent of the mental activity and linguistic practices that we use the possible worlds theory to help explain.

So does possible worlds theory belong to semantics or metaphysics? If the question is about what kind of thing possible worlds are, then we should locate the theory on the metaphysical side. Possible worlds, or possible states of the world, are properties that a total universe might have. But much of the interest of the theory is in its role in explaining thought and speech. Semantics is concerned with explaining the relation between linguistic expressions and what they are use to say. We can do this only if we have some apparatus for talking, in an abstract and general way, about what language might be used to say, and the role of possible worlds theory is to help with this task. But when we ask specific questions about the structure of the space of possibilities, we have to answer the questions in a language whose semantics may be contentious, and this is why metaphysics and semantics are difficult to disentangle.

3:AM:You think metaphysical possible worlds are real but abstract objects don’t you? I think this makes you a species of actualist realist doesn’t it? Other philosophers disagree, not just about the way you characterize possible worlds, such as Plantinga and Peacocke, but even with the idea that there’s anything other than what actually exists or that anything exists that might not have done – such as Wittgenstein, David Lewis and Tim Williamson.

RS:You’re right that my view is a kind of actualist realism, but I don’t think that properly understood, this is such a controversial doctrine. Plantinga certainly counts as an actualist realist, and while Peacocke may be skeptical about the fruitfulness of the possible worlds framework, he is not going to disagree that there really are different ways that the world might be, or might have been. All actualists agree that there is nothing other than what actually exists. The disagreements are about how to account for apparent facts (such as that there might have been things other than those there are) that seem to suggest otherwise. David Lewis – the non-actualist – is the outlier here – the only one of those you mention who would say that it is literally true that there exist things other than those that actually exist. But ironically, it was the austerity of Lewis’s metaphysical resources – his nominalism about properties - that led to his modal realism.

3:AM:What’s at stake in this work that non-philosophers as well as those in the business can grasp?

RS:Metaphysical disputes about individual essences and merely possible things certainly have the feel of scholastic exercises that are irrelevant to real life issues. It is often hard to see what is at stake, and I do think it is important to reflect both on general questions about what one is doing in doing metaphysics, and on why it matters how specific metaphysical questions are answered. The positivists, writing in the first half of the last century, argued that metaphysics was an illegitimate enterprise that arose from equivocation between semantic and empirical questions. In my view, there is a grain of truth in this critique of metaphysics, but also a serious mistake. The grain of truth is that metaphysics involves a mix of semantic and factual issues, and that much of the dialectic of metaphysical argument involves clarifying the language with which we debate fundamental questions about what the world is like, and diagnosing equivocations in the use of that language. The serious mistake is to think that we can separate, at the start of our inquiry, the semantic questions from the substantive empirical question – that we can first clear the ground and establish, by a priori stipulation, a perspicuous and unambiguous language for describing all of the alternative possibilities for the way the world might be, and once we have done that, carry out the task of empirical science. This is the utopian intentional foundationalism that I was going on about earlier in our discussion. Once we reject this fantasy, we can agree with the positivists that it is the essential nature of metaphysics to mix semantic with empirical issues, and an important task of philosophy to separate them out. But one must also acknowledge that the task of separating the semantic from the substantive can’t be done without making substantive commitments about what the world is like, and for this reason, metaphysics is inescapable. Even if we agree with Carnap’s aim – the elimination of metaphysics – we can get there only by doing metaphysics.

Now with some classical metaphysical issues such as the problem of free will, the nature of consciousness and thought, the analysis of causation and chance, it is easy to see some connection with the concerns of those outside of philosophy. The more abstract questions of modal metaphysics are harder to motivate on their own, but these issues are connected with the others. All of the more accessible metaphysical problems I just mentioned essentially involve the notion of possibility, and getting clear about just what possibilities are is crucial for addressing those questions.

3:AM:You are arguing for a very subtle version of possible worlds – you want to avoid extravagant metaphysical commitments – the kind that I suppose give room for Plantinga and Van Inwagen finding a metaphysics that doesn’t rule out God – but you seem to take seriously that your theory may be construed as a realist one. How can ‘we affirm the existence of possible worlds’ whilst avoiding ‘commitment to the existence of nonactual objects’, as you say in your latest book?

RS:I’ve argued, following Quine, that there is no absolute, once and for all neutral framework in which to theorize, but (also following Quine) that we do aim to find a framework that is neutral between alternative answers to questions at issue. One theorizes at two different levels: first we strive to make questions clear without begging them, and this requires some relatively neutral apparatus for formulating them. Second, one uses the framework to state and defend one’s own answers to those questions. In my book Mere Possibilities, my main aim was to develop a framework for talking about possibilities that was as neutral as possible about the character of those possibilities. I argued that (despite some arguments to the contrary) the framework itself is not committed to individual essences, the necessary existence of all that actually exists, or other theses that I (contentiously, obviously) described as extravagant metaphysical commitments. The framework is compatible with these commitments, and I don’t offer arguments against them. What I do offer is just arguments that the abstract framework of possibility, interpreted realistically, does not force us to accept these commitments.

At first pass, it is easy to reconcile an affirmation of the existence of possible worlds with a denial of the commitment to nonactual objects. “Possible world” is a colorful label for a kind of property that a universe might have, and merely possible worlds are uninstantiated properties of this kind. The properties themselves exist, and actually exist. But since the properties are uninstantiated, there do not exist any universes that have those properties. Possible worlds are like the possible sisters that sisterless people have. All there really is in such a case is the possibility that the person have a sister, and not a possible sister that the person has. (I say “at first pass” since one needs to tell a longer story to reconcile possible worlds semantics, with its domains of individuals, with the thesis that there might have been things that don’t actually exist. I try to tell the longer story in the book.)

The general framework, as I understand it, certainly does not rule God either in or out, and I assume that Plantinga and van Inwagen would both agree that theological questions are not settled, one way or the other, by our general theory of the nature of modality. Some “extravagant commitments”, I would argue, are motivated only by bad modal arguments of the kind I try to diagnose, and I think the motivations for those commitments will disappear if we succeed in defusing the bad arguments. But even though there are some bad modal arguments for the existence of God, I certainly don’t think theists’ belief in God depends much on such arguments. My own metaphysical views are thoroughly secular and naturalistic, and so contrast with those of Plantinga and van Inwagen, but I think this disagreement is independent of any disagreement about the commitments of a modal framework.

3:AM:You ask what haecceitism is and whether it is true. So what is it, and is it?

RS:Those are fair questions, since I ask them in the book, so let me start by giving the short answers before taking them back – or at least qualifying them. Haecceitism is the doctrine that individuals are not definable, or supervenient on, their qualitative character. So according to haecceitism, it is metaphysically possible that there be an individual qualitatively exactly like me, but not me. I think haecceitism is true. Or at least (which may not be quite the same thing), I think that the anti-haecceitist metaphysical doctrine is false. The reason for hesitation is that the characterization of the doctrine depends on the distinction between the qualitative and non-qualitative character of an individual, and on the notion of a purely qualitative description of the world, or of a possible world. I am skeptical about the distinction, and so skeptical about a reductionist thesis that depends on it. So it is not that I accept haecceitism because I think individuals are something over and above their qualitative properties. The point instead is that I am inclined to be skeptical about the distinction that the anti-haecceitist doctrine presupposes. One thing I argue in my chapter about haecceitism is that one can’t give a metaphysically neutral definition of the thesis, and then argue about whether it is true or not. One can give a precise articulation of the thesis only by using certain conceptual and metaphysical resources that are themselves contentious, so the process of explaining what the thesis means and the process of defending or attacking it cannot easily be separated. I think this is a general feature of the dialectic of metaphysical debate – it is part of what makes metaphysics hard (as well as part of what makes it interesting).

3:AM:Kripke’s ‘Naming and Necessity’ is clearly important in setting an agenda for philosophers handling Frege problems about how references get fixed and so on, but your way of responding to Kripke is not straightforward is it? How would you characterize Kripke’s influence on what you’ve been doing, and has his influence diminished or changed in significant ways as you’ve developed your thoughts? I guess this is partly about how much over the years since Kripke metaphysics and philosophy of language has changed?

RS:As I said earlier, Kripke was a big influence on me at the beginning of my career, and while a lot has happened in metaphysics and the philosophy of language since his early work, I think his insights have stood the test of time, and still have a lot to teach us. The internalist and descriptivist doctrines the Kripke criticized have proved to have a tenacious hold on our way of thinking about semantics and metaphysics, and new and more sophisticated forms of these doctrines that take account of some of Kripke’s points have kept them alive. As David Lewis wrote in 1984, “description theories of reference are supposed to have been well and truly refuted,” but “I think not: we have learned enough from our attackers to withstand their attacks.” Lewis’s development and defense of an internalist account of intentionality is the most explicit and well developed account of this kind, but I don’t think it works, and some of the reasons why it is on the wrong track are implicit in Kripke’s early work.

3:AM:How do you work out how to weigh up the beauty of a metaphysical theory with ugly facts that it can’t accommodate? Do you think that the balance between theoretical virtue and what there is seems to require an appeal to hunches or intuitions rather than facts in the end, and that threatens the whole project of metaphysics?

RS:I don’t entirely understand how a grand metaphysical theory can be not only beautiful, but clarifying, despite the fact that almost all such theories have consequences that are difficult to take seriously. But it seems that by tying ideas together, seeing interconnections, and following their consequences wherever they lead (as grand metaphysical theories do), one can illuminate a conceptual landscape, and discover presuppositions that are hidden in a way of thinking. Even if the end result is to reject the presuppositions, and the basic principles of the theory, it can be an advance. David Lewis’s modal realist theory, combining an austere nominalism with a rich ontological plenitude of parallel universes, is our best modern example of a grand and beautiful metaphysical theory. Many of us are dazzled by Lewis’s system, even if few of us can take it seriously as something we might believe. But this system, like its predecessors (for example, Leibniz’s metaphysical system, with its windowless monads) had a wealth of spinoffs: arguments, analyses, bits of apparatus that can be detached from the grand theory and appropriated for one’s own philosophical use (and for uses outside of philosophy). It may not seem fair, but I think the best way to achieve the balance you ask about is to admire and appreciate the grand theory, and then loot it for what you find of value in it.

3:AM:And finally, for the budding metaphysicians here at 3am are there five books (other than your own) that you could recommend that would help us further understand your philosophical issues?

RS:I’ll mention three classics that you can predict from what I have said already, and then two more recent books that I found particularly illuminating. First, everyone interested in issues at the intersection of metaphysics and the philosophy of language should read Kripke’sNaming and Necessity, and if you already have, you should read it again, since it is one of those books one keeps finding new things in. Second, David Lewis’sOn the Plurality of Worldsis the best place to get a sense for his metaphysical system, and for his philosophical style, which is crystal clear, direct and unaffected. Third, Paul Grice’sStudies in the Way of Wordsis a collection of most of his published papers, and includes the William James lectures and a long and fascinating set of retrospective remarks. One can see a lively and subtle philosophical mind at work. Two more recent books: Timothy Williamson’s The Philosophy of Philosophy is a sharp and highly original set of reflections on philosophical method, but with plenty of substantive ideas and arguments as well, for example about the analytic/synthetic distinction, about knowledge and intentionality, about modal epistemology. Finally, I recommend Tyler Burge’smasterful treatise, The Origins of Objectivity. This is interdisciplinary philosophy at its best, drawing on a wealth of empirical research about human and animal perception, but keeping its eye on the philosophical issues, and with an unerring sense for the relevance of the empirical results to those issues.

Richard Marshallis still biding his time.