kripke’s unfinished business
Scott Soames interviewed by Richard Marshall.
[Photo credit Philip Channing.]
Scott Soames thinks philosophy is majestic and multi-faceted, is a leading philosopher of language and writes about the history and development of the analytic tradition in philosophy. He thinks about the progress we’ve made in the study of language since Frege, about the central semantic features of language, about truth, about what Russell got wrong, about why propositions aren’t enough and why they aren’t just theoretical entities, about Davidson, about Cognitive Realism, about why his theories don’t distort very much, about the unfinished business of Saul Kripke and going beyond rigidity, about what went in to his books on 20th century analytic philosophy and why analytic philosophy is not a philosophical school, about the analytics in America and his theory of Legal Deferentialism. Fee Fi Fo Fum, this is one mileshigh mind-bomb!
3:AM: What made you become a philosopher?
Scott Soames: I never had serious intellectual interests before my undergraduate days at Stanford, which I entered convinced I would go into my father’s small family business across the street from the Pike Place Farmers’ Market in Seattle, where I worked in my teens. After a year at Stanford fulfilling requirements – freshman English, western civ, biology, calculus, and the like – I had lots of intellectual interests and I knew I wouldn’t be a businessman. In my sophomore year I gravitated toward the humanities, but quickly became dissatisfied with what seemed to me to be the self-indulgent and unserious ways in which the issues on which they touched were discussed in class. Philosophy was different. Though I often had difficulty making sense of the grand systems of great historical figures, or appreciating some of their metaphysical problems, I was at home with reason and argument. In my first course – a historical introduction – I wrote a term paper on the difference between Hume and Kant on morality. When it was returned with my professor’s comment “you have the true makings of an able philosopher,” I was encouraged. My next course, “Good and Evil,” was an eclectic survey of ethical thought from a wide variety of sources. I found it so interesting that I attended the Saturday morning discussion section let by the professor.
In addition to taking valuable courses from professors nobody remembers today, I also took courses from Donald Davidson, Patrick Suppes, and Dagfinn Follesdal – all of whom I would know personally decades later. After completing Dagfinn’s course on Brentano, Bolzano, Husserl, and Quine, I took a reading course from him in which I read nearly all of Husserl. I wrote an essay every week, which he spent an hour talking to me about. During one session, he gave me one of the two best pieces of advice about writing philosophy I ever received. He said “Mr. Soames, you should write so that if you make a mistake, anyone who knows the subject will immediately be able to identify it.” The other piece of good advice, later given by Judy Thomson, was “Don’t be afraid of mistakes; if you never make mistakes, you’ll never be a success.” Not to worry. My critics have Judy and Dagfinn to thank.
Though I only had one two-minute conversation with Pat Suppes when I was a student, he also had an impact on me. In 1970-71, I took his computerized logic course – no lectures, no book, no discussions. Pat had his own building that housed what must have been a giant computer. I saw him only on the first day, when he told me how the course worked and where the terminals were – no screens, just rolls of paper and typewriters on which the (hidden) computer gave us lessons, and we answered questions. The final exam was to find the smallest axiomatization of 23 statements and do the proofs. I loved it.
Although I had no clue about the philosophy of language, Davidson was the big philosopher on campus, so, on the first day, I wandered into his class on the subject. The fact that half the people in the room were graduate students was a little intimidating, but what really put me off was his introductory discourse on why truth was the key to meaning. It seems this guy “Tarski” had discovered that the sentence ‘Snow is white’ is true iff snow is white. From this we were supposed to see that if we could frame axioms from which we could derive a similar theorem for every English sentence (as it seems this fellow Tarski had done for logical languages) we would have a theory of meaning for English and an answer to the question “What is meaning?” That made no sense to me, so I never went back. But I was a student in his large lecture course on ethics. Entering the room accompanied by an entourage of grad students and Assistant Professors, Davidson would dazzle us by speaking brilliantly for an hour without notes. The first half of the course was his own remix of chapter one of Principia Ethica, with a dose of Ayer and Stevenson to explain why the Open Question had to remain open. From this we were supposed to conclude there can be no true ethical theory – which we found shocking but hard to answer. The other half of the course — on actions, reasons, and causes – was, I think, sounder, but not so exciting.
Despite majoring in philosophy, and liking it, I had no idea, as graduation approached, what I would do. The Viet Nam war was raging and I was anti-war – demonstrations, a few weeks in jail, and a decision to make about the draft. I devoured every book on Vietnam, China, Asia, economic development, and political theory that I could find. I decided that though I would not to serve in the army, I would also neither accept a draft deferment nor leave the country. Instead, I wrote a manifesto to my draft board applying for conscientious objector status (and alternative non-military service) on non-traditional grounds. Although I was anti that war I wasn’t anti-war “in any form,” and although my objection was moral, it was not based on “religious training and belief.” So I was drafted, refused induction, and waited two years to come to trial.
At some point after graduating, when I needed a job, I read John Holt’s How Children Fail – which was a heartfelt, but I now think misguided, plea for “progressive education” as a way of rescuing poor children who were failing in public school. I found a federal project scheduled to begin in Mountain View, California aimed at diagnosing and overcoming the reading problems of first, second, and third grade Mexican-American children. Despite having had no teaching experience and no education courses, I convinced them to hire me as a teacher. Being a federal program it was top-heavy – a director, assistant director, materials specialist, two psychologists, a social worker, two teachers and two teacher aides. Not surprisingly, it didn’t live up to the promises made to secure its funding, and it folded after two years. But it worked for me and some of my students. I spent two rewarding years trying everything possible to teach them, sometimes succeeding and sometimes not. No teaching I have ever done has been so challenging.
It was then that I developed my first interest in language. When I was teaching Mexican-American kids, the “experts” in my program speculated that my students were linguistically and conceptually underdeveloped because they came from poor families with uneducated parents who didn’t verbally interact with them very well. That – rather than failing to crack the code in going from spoken to written English – was the supposed source of their problems. Skeptical about this, I read some developmental psychology and some psychology of language, which led me to linguistics and Noam Chomsky. That, in turn, kindled a theoretical interest in language that would later lead me to the philosophy of language. But that was still a little way off.
During my final year teaching in the federal program, I was tried for refusing induction into the army, but, to my surprise, I wasn’t convicted because my two lawyers, Norman Leonard and Richard Gladstein (both of ILWU boss, Harry Bridges, fame) convinced Judge Oliver J. Carter (of Patty Hearst fame) to read my manifesto to the draft board. Although he could see I didn’t fit the legal definition of a conscientious objector, and although he was one of San Francisco’s toughest judges on draft resistors, somehow he was impressed with what I wrote. Believing the draft board owed me a word of explanation, rather than summarily dismissing my C.O. application and turning down my request for an interview, he sent the case back to them for review. When they repeated their earlier behavior, I was free. I came away impressed by the system that had given me a fair hearing, by the intelligence and humanity of Carter, Leonard, and Gladstein, and by the generosity of the latter pair for working on my behalf for what was, for them, the ridiculously modest fee of $900, which was all I had.
With my future no longer under threat, I was free to do as I wished when my federal project folded. I decided to teach young children in the Claremont Colleges experimental elementary school, enroll in their School of Education and take philosophy courses on the side. But it was not to be. In less than a week I decided that neither the intellectual nor the early educational environment there was promising. So I moved back to the mountains above Stanford, enrolled part-time in philosophy as a non-matriculated student, and supported myself by teaching half-time in a middle school in San Mateo, helping kids with learning problems. That’s when I discovered the philosophy of language. I pursued my new interest first by taking a philosophy of language course at Stanford from Julius Moravcsik (who was to remain my friend until he died several decades later), and then by taking graduate seminars offered by David Kaplan and Barbara Partee, who flew in once a week from UCLA. Those were spectacular. I was hooked, and the next year I started graduate school at MIT – there to work with and be inspired by Dick Cartwright, George Boolos, James and Judy Thomson, Jerry Katz, Sylvain Bromberger, Noam Chomsky and David Perlmutter.
So why did I go into philosophy? Because I found it to be a majestic, multi-faceted subject, fascinating both in its own right and for the perspectives it offers on a wide range of thought and experience, because I had the good fortune of being taught by exciting and accomplished teachers, because I began to feel I was made for it, and because I judged that nowhere else could I interact with such brilliant and interesting people. All true.
3:AM: As a philosophy of language philosopher one of the things you’re interested in is meaning. I guess the obvious question is why philosophers of language find this an important topic. Some might suggest that getting at the truth is a real issue. But wondering about how we mean seems less pressing. So what do you think is at stake in this area of philosophy?
SS: What is a stake is whether the progress we have made in the study of language since Frege can be extended to a genuine science of language and information. We are not there yet, but if philosophers continue to make foundational contributions, eventually we will be. The central semantic feature of language is that agents use sentences to represent things in the world as being certain ways. This is the basis of speaker meaning, which, in turn, is related to linguistic meaning. Without meaning in these senses, the ordinary notion of truth — which is conceptually tied to these senses of meaning — loses its utility for empirical theorizing about language and mind, and becomes little more than a vehicle for Tarskian semantic ascent. The reason we want a theory of meaning is that it is a necessary component of an account of what it is to understand a sentence of a language, of what is asserted by utterances of such sentences, of what is believed when one understands the sentence uttered and believes what it is used to assert, and of what must is true if what one asserts or otherwise expresses using a sentence is to be true.
3:AM: You argue that a theory of meaning needs propositions but also needs to account for the cognitive stance of a person towards a proposition and also how propositions manage to represent the world and have truth conditions. Is that right? So firstly, can you explain why you find Russell’s attempts to account for them not right?
SS: Yes, a theory of meaning for a language L needs propositions that represent the world and so have truth conditions. Yes, it also needs an account of cognitive stances agents take to propositions – if L has sentences – e.g. attitude ascriptions – that predicate properties of propositions. My answer to your question about Russell is explained in chapter 9 of my new The Analytic Tradition in Philosophy, Vol. 1. Here is the gist of it. Between 1900 and 1910 Russell believed in propositions constituted by objects and properties, but he couldn’t explain their “unity”. Just as sentences aren’t collections of unrelated expressions, but have a structural unity that distinguishes them from mere lists and allows us to use them to represent the world truly or falsely, so propositions aren’t collections of unrelated meanings of the words used to express them, but have a unity that endows them with truth conditions that mere aggregations of their parts don’t have. Russell struggled unsuccessfully to explain this unity until he rejected propositions in 1910 in favor of his multiple relation theory of judgment. Although that theory was disastrous, the insight behind it was brilliantly correct. The intentionality of agents can’t be derived from the supposed sui generis intentionality of propositions to which agents bear attitudes. Instead, the unity that brings together Desdemona and being unfaithful in Othello’s belief that Desdemona was unfaithful is provided by the sui generis fact that the agent predicates being unfaithful of Desdemona. What Russell failed to see was how this insight can be used to reconstruct genuinely unified (i.e. representational) propositions by deriving the intentionality of propositions from the intentionality of agents who entertain them.
3:AM: Why do you think it is not enough just to have propositions, but we need the cognitive stance and an explanation of how they have truth conditions?
SS: Let’s take a few examples. To entertain the proposition that B is red is to predicate redness of B, which is to perceive or conceive of B as red. Other propositional attitudes are based on this. To judge that B is red is to perform the predication in an affirmative manner, which involves accepting it as a basis for possible action. This doesn’t involve predicating any property of the proposition, or making it an object of cognition; it does involve a predication that itself involves forming, or activating already formed, dispositions to act, both cognitively and behaviorally, towards B in ways conditioned by one’s attitudes toward red things. So, to judge that B is red is for one’s predicating redness of B to include one’s forming or activating certain dispositions. To believe that B is red is to be disposed to judge that it is. To know that B is red is, roughly, for B to be red, to believe that B is red, and to be safe, or to be cognitively justified, in so believing. To assert that B is red is to commit oneself, by uttering something, to knowing that B is red. These attitudes – judgment, belief, knowledge, and assertion – aim at truth. But the story is the same for other attitudes — e.g., doubting, denying, disproving, and imagining – that don’t aim at truth.
The objects of these attitudes – the things doubted, denied, disproved, or imagined – may be true or false, just as the things judged, believed, known, or asserted may be. The very same proposition that is believed by one person may be doubted, denied, disproved or merely considered by another. This suggests that the intentionality of both sets of attitudes is due to something common to all the attitudes, and hence that it is not any of these attitudes but their objects that, in the first instance, represent things as being certain ways. Because of this, propositions have truth conditions independently of agents’ further affirmative, negative, or non-committal stances toward them. Cognitive stances are needed to distinguish these different attitudes from one another.
3:AM: Why do you think Davidson’s account of truth conditional meaning is inadequate?
SS: I have spilled a lot of ink on this, so I will only summarize. We begin with John Foster’s demonstration (“Meaning and Truth Theory,”1976) that Davidsonian truth theories don’t — and can’t — satisfy Davidson’s own original criterion for calling it a theory of meaning – namely that knowledge of what it states is sufficient for understanding the language. We continue with my arguments (“Truth, Meaning and Understanding,” 1992) that Davidsonian theories also don’t satisfy his revised criterion, proposed in response to Foster, and my further arguments presented in “Truth and Meaning – in Perspective,” 2009, and reprised in What is Meaning? , that no proposed rescue has worked, and it appears that none will. Roughly put, a semantic theory should be the central component of a theory that allows us to interpret the cognitive and communicative activities of speakers. Davidsonian theories are too impoverished to play this role, even for purely extensional object languages. Things get worse when intensional and hyperintensional constructions are added.
3:AM: Why don’t you think propositions are just theoretical entities?
SS: I tested and rejected a version of that idea in chapter 5 of What is Meaning? The idea was that propositions are theoretical fictions used to track and compare the cognitive states of agents who predicate properties of things. Just as in physical theory we use numbers and other abstract objects to talk about theoretically significant relations that physical magnitudes bear to one another, so, it might seem, in semantic and psychological theories we use abstract “propositions” as mere instruments to talk about theoretically significant relations that representational cognitive states of agents bear to one another and to the world. Although this idea could in principle be made to work for a restricted range of cases, the rocks on which it founders are cognitive states corresponding to complex propositions in which properties are predicated of other propositions. Here the cognitive states being tracked are those in which agents predicate properties of propositions. They cannot do this, if propositions are theoretical fictions.
There is another, more common way, in which propositions are treated as theoretical conveniences. Consider a version of possible worlds semantics that takes truth at w to be a two-place primitive, not definable in terms of our ordinary monadic notion of truth, while identifying propositions with functions from world-states to the truth values of sentences at those states. Stalnaker accepted this picture in Inquiry (1984), saying that truth and falsity can be identified with 1 and 0, because it doesn’t matter what they are so long as they are distinct. On this picture, propositions are nothing more than (supposedly useful) theoretically fictions. As I argue in New Thinking about Propositions (with Jeff King and Jeff Speaks), this is pretty much how one has to think of them within a possible worlds framework, if one doesn’t use a distinct and independent conception of propositions together with our ordinary notion of truth in defining truth at w. If one takes this route, one’s possible worlds “semantic” theory becomes a mere formal device for sorting sentences into classes which, by itself, tells one nothing about the meanings of sentences, or their cognitive and communicate use by speakers. Thus, the whole empirical content of such a “semantics” must come from extra-semantic theses about how sentences are used. This has never been successfully done, in part because it has seldom been attempted.
3:AM: You argue that propositions are best understood in terms of the Cognitive Realist theory. Can you say how this works?
SS: The theory starts from two premises: (i) Agents represent things as being certain ways when they perceive, visualize, imagine, or otherwise think of them as being those ways. (ii) Propositions represent things as being various ways, and so have truth conditions, because of the relations they bear to agents who entertain them. Next we ask What must propositions and entertaining be in order to guarantee that one who entertains the proposition that a given piece of fruit B is sour, thereby represents B as sour? My answer is that propositions are repeatable, purely representational, cognitive acts or operations; to entertain one is to perform it. When I perceive or think of B as sour, I perform the act predicating sourness of B, which is to represent B as sour. The act represents B as sour in a sense similar to that in which acts can be insulting or irresponsible. An act is insulting when for one to perform it is for one to insult someone; it is irresponsible when to perform it is to neglect one’s responsibilities. A cognitive act represents B as sour when for one to perform it is for one to represent B as sour. The minimal, purely representational act of doing so is the proposition that B is sour, which is true iff to perform it is to represent B as it really is.
Familiar propositional attitudes are defined as the cognitive stances taken toward propositions, but not all attitudes have propositions as objects.
The object of the attitude reported by “John asked whether B is sour” is a question. Whereas a proposition represents the world as being a single way, and so has truth conditions, questions are cognitive acts that represent the world as being several ways, and so don’t have truth conditions. The question whether B is sour is the act Q of (i) entertaining the proposition that B is sour, and (ii) applying an operation to it that results in representing the world in two ways – as being such that B is sour and as being such that B isn’t sour (which isn’t to represent the world as being such that B is sour and B isn’t sour). To ask Q is to request a judgment as to which of the two ways the world is.
This is the basis of a naturalistic epistemology of propositions. Since believing p doesn’t require cognizing p, any organism that can perceive or think of its constituents as being certain ways can believe p, whether or not it can predicate properties of propositions. Knowing things about propositions requires the further ability to distinguish one’s cognitive acts from one another. One who can do this can ascribe attitudes to oneself and others, and predicate properties of propositions. Focusing on their cognitions, such agents identify distinct propositions as different thought or perception types, which leads them to conceive of truth as a form of accuracy in representation.
Here, I have mentioned only simple propositions, which predicate properties of objects. Complex propositions involve additional cognitive operations. But the basic idea is the same. In speaking of propositions as acts, I don’t mean they are always intentional or conscious. They aren’t. But they are doings in which things are cognized as being one way or another. How a proposition represents things is read off the cognitive doings with which it is identified. From this we derive its truth conditions. P is true at world-state w iff were w instantiated, things would be as p represents them – where what p represents is what any conceivable agent who entertains p would represent. Since this doesn’t vary from state to state, p’s truth conditions don’t either. No one has to entertain p for p to be true.
This conception has the foundational advantage of explaining how an organism without the concept of a proposition or the ability to cognize one can know or believe them. It also explains how sophisticated agents can acquire the concept, and come to know things about propositions by monitoring their own cognitions. Further progress comes from solving the problem of propositional unity and from the account what it is for a proposition to be the meaning of a sentence. For S to mean p in L is for speakers of L to use S to perform p. One who understands the sentence ‘The Union Jack is a symbol of Britain’ uses the name to pick out the flag and the verb phrase to predicate being a symbol of Britain of it. Since to do this is to perform the act that is the proposition expressed, one’s use of the sentence is one’s entertaining the proposition. Since no other cognition is needed, understanding what S means in L, when in fact s means p, doesn’t require knowing that S stands in some theoretical relation R to p and L, or having any concept of a proposition or a language.
3:AM: Two arguments against your theory have been that it distorts the normal way we talk about propositions and also that propositions become so complex no one could ever entertain them. What’s your response.
SS: The theory doesn’t distort the normal way in which we talk about propositions very much. Ordinary folks have no more idea of what the proposition that snow is white is than they do of what the number 7 is. They don’t talk about the kinds of things that propositions are, except to say that they are things we believe, assert, deny, etc., as well as being things that may be true or false, either contingently or necessarily. When one gets down to the nitty gritty of what, exactly, is asserted or believed, when, where, why, and by whom, the cognitive theory gives us tons of good results that we need in semantic, pragmatic, and psychological theories that no other theory of propositions provides. I have in mind results about de se attitudes, sentences with articulated terms, temporal cognition, perceptual cognition, essentially linguistic cognition, and cognition of propositions involving recognition of the recurrence of different items. The case for the cognitive theory is (i) that it delivers the goods, and (ii) that it has the foundational advantages for the epistemology of propositions I talked about in answering the previous question.
But there is one way in which it might be said to distort what we otherwise would have thought. It does sound very strange to be told that a proposition is something we do, so strange that one might be tempted to think it is a category mistake. Why does it seem so strange? If one asks oneself, pretheoretically, “What is a proposition?” one naturally starts with examples. What are the propositions that 7 is a prime number, that Seattle is sunny, or that there is a red dot on the wall in front of me? In bringing these examples to mind, one thinks of 7, Seattle, and the wall as being certain ways. In some cases one may also conjure up images, but even then one knows that the proposition entertained isn’t the image in one’s mind, but something more general. Still, there is something in the visual model to which we wrongly tend to cling. Just as seeing a wall with a red dot is a phenomenally robust form of being aware of it, so, we are incautiously inclined to think, visualizing the wall is a phenomenally poorer form of awareness of a mental image, while entertaining a proposition about the wall is a minimal, or even phenomenally empty, form of awareness of something that represents the wall as being a certain way. It isn’t. To entertain a proposition is not to be aware of it, nor is to believe it to affirm something of which we are aware.
We are wrongly encouraged to think otherwise by the parallels, (i) and (ii), between our talk about perception and our talk about propositional attitudes.
(i) Just as ‘see’ is a two-place predicate relating agents to things seen, so ‘believe’, ‘assert’, and ‘know’ are two-place predicates relating them to propositions believed, asserted, and known.
(ii) Just as standing in the relation expressed by ‘see’ requires agents to be aware of things involved in the perception, so standing in the relations expressed by ‘believe’, ‘assert’, and ‘know’ requires them to be aware of things involved in the attitudes.
The natural, but readily explainable, error is to jump from (i) and (ii) to (iii), when in fact it is (iv) rather than (iii) that is true.
(iii) Just as standing in the relation expressed by ‘see’ requires one to be aware of the things seen, so standing in the relation expressed by ‘believe’, ‘assert’, or ‘know’ requires one to be aware of the propositions believed, asserted, or known.
(iv) Although standing in the relation expressed by ‘see’ to an object o that is seen requires one to be aware of o, standing in the relation expressed by ‘believe’, ‘assert’, or ‘know’ to a proposition p merely requires one to be aware of the things p represents and the ways they are represented.
The error of opting for (iii) rather than (iv) is one main source of the mistaken idea that propositions can’t be cognitive acts. It’s obvious that one can perform an act without making it the object of one’s awareness, and also that forming cognitive and behavioral dispositions when performing an act needn’t involve thinking of it. Since this is what entertaining and believing cognitive propositions amount to, an objector’s commitment to (iii) leads him to conclude that the cognitive conception is absurd. But it isn’t absurd; (iii) is false.
Think of what happens when an agent entertains the proposition that Seattle is sunny, and then thinks, that is widely disbelieved. In so doing he focuses on the concrete cognitive event in which he considers Seattle as being sunny. This token event (which lasted a second in his own mind) isn’t what he takes to be widely disbelieved. Rather, he would say, that type of thing is what is widely disbelieved. What type? Is it the event type in which one predicates being sunny of Seattle, is it the act type performance of which results in an event of that type, or is it some other type? It is, I think, whatever type best plays the proposition role in our theories. But this philosophical answer isn’t one the agent would give. All the agent can say about the type (the proposition) is that it is the thought he just had – which is true enough provided he doesn’t succumb to the seeing-in-the-mind’s-eye temptation of taking the thought to be something on which he focused when whispering “Seattle is sunny” to himself. If he does succumb, he may object if he is told that the thought he was entertaining (and so, he wrongly thinks, focusing on) was his act of thinking (i.e. focusing on) it. But he will be wrong. The cognitive act p (predicating being sunny of Seattle) is identical with the act of entertaining p – since, in general any act A is identical with the act performing A.
Perhaps the second objection in your question eludes me. Are some propositions too complex for humans to entertain? Of course. But they are not too complex for any conceivable – i.e. epistemically possible — agent to entertain. That’s all I need. If there is a serious worry about this, I don’t see it.
3:AM: You’ve found unfinished business in Kripke. Kripke is taken to have overthrown Russell’s theory of descriptions for names, which said that descriptive information that people connect to a proper name determines the referent. Can you say a little about Kripke’s arguments and the idea of a rigid designator and why you agree with them? You do think that some names are partly descriptive though don’t you? Doesn’t that complicate your relationship to the Kripkean project? Isn’t there a small toe-hold for descriptivism after all?
SS: That is a complicated bunch of questions which I will do my best to answer. First, my relationship with the Kripkean project is complicated. My aim has been to use it as a platform on which to build a larger structure. When David Lewis and I co-taught a year-long graduate seminar at Princeton “Naming and Necessity and its Aftermath,” in 1999-2000, he announced on the first day that if Naming and Necessity was revolutionary, then he, David, was a counter-revolutionary, whereas I announced that it was revolutionary, and that I and others wanted to take it further. As I see it, Saul clearly established that proper names are rigid designators that are not synonymous with (unrigidified) singular definite descriptions (or clusters of such). He also recognized the necessary aposteriori and the contingent apriori, while laying the basis for systematically distinguishing epistemic possibility from metaphysical possibility. These were achievements that changed the direction of philosophy. But he didn’t offer a positive thesis about what names mean, or what they contribute to the contents of statements made using them, and so provided no solution to Frege’s puzzle. In addition, his argument in Naming and Necessity for the aposteriority of the statement that Hesperus is Phosphorus was faulty and his related explanation of the aposteriority of the statement that heat is molecular motion, and similar examples, was badly in need of correction (leading, as it did, to the disastrous two-dimensionalist attempts to trivialize the necessary aposteriori). Related to this, his failure to recognize epistemically possible but metaphysically impossible world-states led him to mischaracterize the role of evidence in explaining the aposteriority of his celebrated examples. This, in turn, prevented him from clearly articulating the true relationship between conceivability and (metaphysical) possibility that his own work made manifest – namely, that conceivability plus knowledge of actuality is our guide to genuine possibility. Finally, these last two failures undermined his argument against the identity of pain with C-fiber stimulation. I have tried — in Beyond Rigidity 2002, Reference and Description 2005, “Naming and Asserting 2005,” “The Philosophical Significance of the Kripkean Necessary Aposteriori” 2006, “Actually” 2007, and “Kripke on Epistemic and Metaphysical Possibility” 2011 — to correct these shortcomings while preserving Saul’s core insights.
Now to your specific questions. There was an early attempt by Dummett and his followers to block the conclusion that names are rigid – and so not synonymous with nonrigid descriptions – by arguing that they are descriptions that are required to take wide scope over modal operators. The basic idea was that we believe that ‘Aristotle’ is a rigid designator because, as Kripke showed, (i) is true.
(i) There is a unique x such that for all possible world-states w, the statement that Aristotle was a philosopher is true at w iff x was a philosopher at w; ditto for other statements made using ‘Aristotle’.
Thesis (i) contains an occurrence of ‘Aristotle’ embedded under a modal quantifier. Now replace the name with a nonrigid description ‘the so-and-so’ that is required to take wide scope over modal notions. Substitution leads to (ii); adjusting for wide scope gives us (iii), which simulates rigidity.
(ii) There is an individual x such that for all possible world-states w, the statement that the so-and-so was a philosopher is true at w iff, at w, x was a philosopher; similarly for other statements made using ‘Aristotle’.
(iii) There is a unique y that is so-and-so and an x such that for all possible world-states w, the statement that y was a philosopher is true at w iff x was a philosopher at w; ditto for other statements made using ‘Aristotle’.
So, the thought went, since (i) is true iff (iii) is true, we can capture everything Kripke established while taking names to be synonymous with nonrigid descriptions that are required to take wide scope over modal notions — but not over epistemic notions (preserving Frege’s solution to Frege’s puzzle). In chapter 2 of Beyond Rigidity, I showed why this won’t work. We need rigidity, not its simulation. I also showed that the modal argument against the synonymy of names with descriptions can’t be blocked by rigidifying the descriptions; names are never synonymous with rigidified descriptions. Nor does it help to appeal to descriptions that are too subtle or complex for speakers to articulate.
Your question about reference determination is different. Names might be rigid while still having their reference determined by nonrigid descriptions. Kripke offered a battery of good arguments against this. Descriptivists countered by claiming that the arguments merely eliminated reference determination by the wrong descriptions, arguing that the right descriptions must exist. Why must they exist? There were two main thoughts. One piggybacked on the claim by Kripkean causal-chain theorists that the referent of a name n is the object that initiated the (relevant) causal chain leading to one’s current use of n. If D incorporates the details of this chain of transmission, then D fixes the referent of the name; so Kripke is wrong. Not so fast. The problem is that no one, certainly not ordinary speakers, associates any such D with the name. They may, to be sure, know that the referent of their use of n is inherited from earlier uses of n, without being able to say how it is inherited or from which uses. Thus, they may associate n with ‘the referent of whatever earlier uses from which my present use somehow inherited its referent’. But satisfaction of this description by o presupposes some antecedent fact by which the inheritance of o as referent takes place. Thus the fact that o does satisfy this description can’t be the constitutive fact that fixes the reference of one’s use of n. The other thought, defended by Frank Jackson in “Reference and Description Revisited,” 1998 is that Kripke’s own methodology presupposes that speakers must implicitly have the material for constructing a reference-fixing description at their disposal. In arguing that ‘the author of the completeness theorems’ can’t fix the referent of ‘Gödel’, Kripke assumes we can intuitively see, when presented with descriptions of certain world states, that the reference of the two expressions comes apart. From this Jackson concludes, I believe quite wrongly, that speakers are in possession of descriptive information that determine the referents of the names with which they are familiar at every possible scenario of use. Since I have gone on long enough, I won’t explain here why Jackson’s claim is incorrect. Interested readers should consult my “Reply to Jackson” in “The Substance and Significance of the Dispute over Two Dimensionalism”.
Are partially descriptive names toe-holds for descriptivism? In Beyond Rigidity I argued that the semantic content of e.g., ‘Princeton University’, is the same as the semantic content of ‘the x: x = y and x is a university (at t)’ relative to an assignment to ‘y’ of the object o at the end of the historical chain of uses leading to my present use of the name. The ‘at t’ part was added in my “Reply to McKinsey” 2006 to avoid some counterexamples. Since developing my new theory of propositions, I have come to doubt that partially descriptive names really do have semantic contents that include descriptive material. I have come to think that everything I previously wished to capture with such contents – including the fact that what I assert when I utter ‘I once taught at Princeton University’ includes the information that I taught at a university — can be captured (without the attendant problems) by the notion of essentially linguistic cognition I develop in chapter 4 of Rethinking Language, Mind, and Meaning, forthcoming, which I briefly explain in “Cognitive Propositions”.
Descriptivism does play an important role in what is asserted by uses of sentences containing names – whether they be partially descriptive or not. But the descriptive material in question is not semantically encoded. Semantics is thoroughly Millian, but assertion is partially descriptive, and part of what is needed to solve Frege’s puzzle. The gap that I see between the semantic content of S and the assertive content of an utterance of S has grown over the past decade. Interested readers should see: “The Gap Between Meaning and Assertion: Why what we literally say often differs from what our words literally mean,” 2009, plus the two items — “Cognitive Propositions” and Rethinking Language, Mind, and Meaning – just mentioned.
3:AM: Kripke wanted to extend rigid designation to some general terms. Is this the unfinished agenda you talk about and is he right to do this?
SS: Yes, one of the unfinished items on Kripke’s agenda – over and above specifying the semantic contents of names – was to explain precisely how his main conclusions about names carry over to general terms, including common nouns, compound nouns, and some adjectives (examples of which are given in Naming and Necessity). Although I was right in thinking this was needed, I erred in Beyond Rigidity by conflating simple general terms like ‘red’ and ‘water’ with predicates ‘is red’ and ‘is water’ formed by combining them with the copula. (See my reply to Linsky, who pointed out this problem, in my contribution to the symposium on Beyond Rigidity in Philosophical Studies, 2006.) Kripke’s general terms are rigid, in essentially the same sense that names are, but this turns out to be of minor significance, since nearly all simple general terms, including ‘doctor’, ‘bachelor’, and the like are also rigid. This means that something other than the rigidity of ‘electricity’, ‘water’, and ‘heat’ explains the necessary aposteriority of Kripke’s key examples, including ‘Lightning is electricity’, ‘Water is a substance molecules of which are composed of two hydrogen atoms and one oxygen atom’, and ‘Heat is molecular motion’, and ‘For all x and y, x is hotter than y iff the mean molecular kinetic energy of x is greater than that of y’. Thus, we need an account of what is responsible for their necessity and aposteriority. Chapters 9 and 10 of Beyond Rigidity remain part of the story, which has is taken further in my “What are Natural Kinds?”.
3:AM: You have also written a huge two volume book on 20th century analytic philosophy. Can you summarize for us the criteria for who got in and who didn’t? I guess one of the reasons for the question is that there has been debate about whether there really is an analytic tradition in philosophy, and whether there is an analytic/continental divide that is anything more than a sociological divide rather than a philosophical one.
SS: The criteria for who got in those volumes were (i) who was important enough to be included among the relatively small number that could be studied by Princeton undergraduates in two 12-week semesters of a course designed to cover most of the 20th century, (ii) who could students, some of whom had no formal training in logic, be profitably introduced to in the short amount of time available for any single figure, and (iii) who could they become interested in without too much stage setting. In my opinion, the three most important analytic philosophers who were excluded by these criteria were Frege, Carnap, and Rawls. These deficiencies will be made up by my much more extensive and advanced five-volume work The Analytic Tradition. The contents of those volumes are not limited by (i) – (iii). In addition to discussing substantial new material on most of the figures covered in the two original volumes, the new volumes will feature extensive coverage of figures such as Carnap, Gödel, Tarski, Chomsky, Kaplan, David Lewis, Stalnaker, Chalmers, Jackson, Putnam, Fodor, Dennett, Rawls, Nozick, H. L. A. Hart, Dworkin, and Raz, with less coverage of many others. The first volume is due out in mid March this year, (700 pages on Frege, Moore, and Russell alone, including the early days of Moore and Russell)
Is there really an analytic tradition in philosophy?. Of course there is an analytic tradition in philosophy, but analytic philosophy is not a philosophical school. There is no set of philosophical doctrines that all, or even the great majority of analytic philosophers adhere to, and there is no restricted set of common goals or interests. There is also considerable overlap between continental philosophers like Brentano, Husserl, Gadamer, Levinas, and Habermas, on the one hand, and various collections of analytic philosophers on the other. But that doesn’t mean that the divide is merely sociological. The analytic and continental traditions are paths of historical influence that have led to family resemblances among their members, even though some members of each tradition resemble some members of the other more strongly than they resemble various members of their own tradition.
3:AM: Does your new book add to the picture presented in the two earlier volumes by showing the growth of the analytic out of American Pragmatism? Can you say something about this and whether Pragmatism has not been rejected or whether much of the American Analytic scene is still Pragmatist at heart.
3:AM: Well, Volume 1 of The Analytic Tradition, does discuss why Moore and Russell rejected American Pragmatism. But it doesn’t grapple with the main pragmatists in their own right. The story of the impact of Pragmatism on the development of analytic philosophy in America is (briefly sketched) in the first essay of my new collection of essays, Analytic Philosophy in America, due out at the end of March of this year. In my view the most important American Pragmatists were Peirce, James, Dewy, and C.I. Lewis – of whom the first and the last had the most enduring influence. I don’t think there is anything really dead in Peirce and Lewis; they repay close reading today. The others less so.
3:AM: The final essays in Analytic Philosophy in America advance an originalist theory of interpretation applied to U.S. constitutional rules about due process. Can you say something about your approach here?
SS: Yes, I outline a theory of legal interpretation I call “Deferentialism”, which can be taken to be a version of originalism, though I hope it is an improved version. Its main features are:
(i) The legal content of a written statute or a specific provision of a written constitution cannot be identified with either the semantic content of the relevant text or the legal or political rationale that provided the purpose of its passage, but it can be identified with what was asserted or stipulated by the relevant lawmakers or ratifiers in passing or approving it.
(ii) In applying this content to the facts of a particular case, the legal duty of the judicial authority is to reach the verdict determined by the application of the asserted or stipulated content, unless (a) the content is vague and, as a result, the conjunction of that content and the facts of the case doesn’t determine a single definite verdict, or (b) the conjunction of the content with other surrounding law and the facts of the case determines inconsistent verdicts, or (c) the content together with the facts of the case is inconsistent with the rationale of the relevant statute or constitutional provision, which is the stated purpose that proponents of the law or provision advanced in justifying its passage.
(iii) In situations conforming to (ii), the duty of the judicial authority is to make new law by making the minimum change in the content of existing law or laws that maximizes the fulfillment of the stated rationale for the law or laws.
Application of (i) – (iii) requires historical research to determine the original asserted or stipulated content plus the original rationale. Point (iii) requires creativity in crafting new legal content, which authorizes limited legislative discretion to courts, guided by deference to the rationale of the original lawmakers. Provision (c) is especially relevant to the interpretation of constitutional provisions, which are often brief, overly general statements of guiding principles with relatively clear and value-laden rationales that guide the process of future adjustments in light of unanticipated facts.
I argue that this conception of the role of the judiciary is both normatively superior to its competitors and descriptively more accurate than those competitors in specifying the legal responsibilities of the judiciary in the federal and state governments of the United States. The chief basis of the claim of descriptive adequacy is the doctrine of the separation of powers, which is firmly rooted in American tradition and embedded the national and state constitutions.
In the last of the articles referred to in the question, I apply my theory to the historical line of cases involving the Due Process Clause of the 5th and 14th Amendments to the American Constitution. Following the excellent historical and legal analysis of Chapman and McConnell in “Due Process as Separation of Powers,” , I argue that the recent line of cases from Griswold, Roe, and Casey to Lawrence were wrongly decided in a way that mirrors the wrongly decided case of Lochner in 1905.
3:AM: How much is lost and subject to misunderstanding if philosophers don’t pay attention to the fine-grained nuances of the different cultural influence of analytic philosophers – is there anything in the suspicion that there is a very distinct Australian, American, British, and European approach to analytic philosophy?
SS: I don’t see substantially distinct Australian, American, British, and European approaches to analytic philosophy, nor do I see broad cultural differences having much effect on the differences that do exist. Of course, a place like Australia, with far fewer philosophers than the UK or the US, will display less diversity of opinion and look more cohesive. But leaders in different sub-fields of philosophy who work in any of these places interact with leading figures everywhere. America leads because of sheer numbers, but, by virtue of the excellence of their leading figures, Britain and Australia punch far above their numerical weight.
3:AM: Are there five books you could recommend to readers of 3:AM to take them further into your philosophical world?
SS: I will name six. Kripke, Naming and Necessity; Frege, The Foundations of Arithmetic; Russell, The Problems of Philosophy; Austin, Sense and Sensibilia; Kaplan, Themes from Kaplan, Chomsky, Syntactic Structures.
Although I regard the first two as the most important books in analytic philosophy, importance isn’t the main criterion used to make the list. The main criterion is that each is, in its way, both theoretically interesting and delightful to read. Reading Kripke, Russell, and Kaplan is nearly always a joy. (Themes from Kaplan made the list because Kaplan’s two long articles are, together, a book-length delight.) Although Frege, Austin, and Chomsky can all be turgid, they aren’t in these sparkling little works. Sadly, the production of philosophical prose that is both delightful and philosophically perspicuous is too rare and too-little valued today. As Timothy Williamson said, referring to a different philosophical virtue, “We must do better!”
ABOUT THE INTERVIEWER
Richard Marshall is still biding his time.
First published in 3:AM Magazine: Friday, March 28th, 2014.