Identity and Synonymy

It is commonly supposed that “water is H2O” is both known to be true and synthetic (hence a posteriori). I think this is not so. The reason is not difficult to see: if the sentence is known to be true, then speakers will associate the same descriptions with each term (following Leibniz’s law), so the sentence will be analytic for them.  Consider a group of scientists working in the same lab and knowing the identity of water and H2O: when they speak to each other they use the two terms interchangeably, knowing that the denoted things are identical. For them, the terms are synonymous and the sentence “water is H2O” is analytic. It is always possible to substitute these terms in belief contexts, given that everyone knows the truth of the identity statement (and knows that others also know it). Nobody thinks there is any more to one of the denoted things than the other: there is complete ontological parity, utter indiscernibility. They might even use “hydrox” as a more vernacular term than “H2O”, so producing the identity statement “water is hydrox”, which is also analytic (everyone assents to this sentence). Before the discovery of the identity people would suppose the two terms had distinct references, but afterwards the difference of meaning disappears. It is the same with Hesperus and Phosphorous: once the identity is discovered and generally accepted “Hesperus is Phosphorous” is analytic, since whatever is believed true of one is believed true of the other. This means that being synthetic is not compatible with being known to be true: no identity statement can be both. Of course, a false identity statement can be synthetic, given that the two terms have distinct references. All this seems clear enough, but matters turn murkier when we come to statements like “pain is C-fiber firing”. Here we are encouraged to believe that the statement is both true and synthetic—informative, a posteriori. It could have those attributes and be false, but it is said to be both true and synthetic; it is “empirically true”. This combination is supposed to be exemplified in such statements as “water is H2O”, thus providing a precedent for the claim of identity between pain and C-fiber firing. But I just argued that the former case is not an example of a synthetic identity statement. So, if the latter case is synthetic, as it seems to be, then this can be so only because it is not true (and known to be true). The explanation of its appearing not to be analytic is simply that it is not true—the two terms refer to distinct properties. If it were known to be true, then speakers would associate the same descriptions with each term, thus rendering the statement analytic. This doesn’t happen because speakers can see that the two terms have different denotations, so the statement can’t be analytic. The case is essentially like “creature with a heart” and “creature with a kidney”: the two predicates are coextensive (the properties denoted are correlated), but the properties are numerically distinct. Similarly, “pain” and “C-fiber firing” are coextensive (the properties are correlated), but the properties themselves are numerically distinct. This is the reason the statement is not analytic: there is more to pain than C-fiber firing, as any competent person can see. That is, the identity theory is false. And its characteristic formulation has no precedent in the sciences—no identity statement can be both known to be true and synthetic. Therefore, we cannot be lulled into accepting the truth of “pain is C-fiber firing” by alleged cases of known synthetic identity in the (respectable) sciences. If that sentence were an accepted identity sentence, then it would have to be analytic; but it isn’t, so it must be false. The explanation of its apparent non-analytic status must be that the properties are distinct, recognizably so. In other words, the identity could only be known to be true by being a priori—as in the case of “water is H2O”. Once that statement is known (or believed) to be true it becomes analytic, as it is for my group of savvy scientists; but this can’t happen with “pain is C-fiber firing” for the simple reason that the alleged identity claim is patently untrue. Ifpeople came to accept it as true, then it would be analytic for them; but it clearly isn’t analytic for anybody, so it cannot be true. There is no obstacle to water being recognized as identical to H2O, since there is demonstrably nothing more to water than H2O; but there is an obstacle to recognizing that pain is identical to C-fiber firing, namely that there clearly is more to pain than C-fiber firing. If that were not so, then the identity statement could be happily analytic—but that it can never be. The sentence is robustly synthetic, permanently informative, indelibly a posteriori: but this is incompatible with the truth of the identity statement. For, if the statement were true, it would have the status I attributed to “water is H2O”—an analytic truth. The fact is that there is no meaningful analogy between “water is H2O” and “pain is C-fiber firing”. Accordingly, the only way to make good on materialism is to provide an a priori analysis of the concept of pain—just the kind of thing attempted by the analytical behaviorists and functionalists. The idea of an empirically true central state materialism is a chimera born of a faulty analysis of so-called theoretical identifications in science. The mind-body problem requires an a priori analysis of the mental. No adequate solution can take the form of an a posteriori synthetic reduction of the mental to the physical. None of the standard alleged analogues have the properties required: “heat is molecular motion”, “light is a stream of photons”, “genes are strands of DNA”, etc. These all involve post-discovery synonymy, produced by the operation of Leibniz’s law. Once you discover that Hesperus is Phosphorous the names “Hesperus” and “Phosphorous” become synonyms for you, intersubstitutable everywhere salva veritate. There is no such thing as a known synthetic identity.[1]

[1] It might be said that this is too strong, because some identity statements contain definite descriptions that enable them to be both known and synthetic (“Benjamin Franklin was the inventor of bifocals”). But this is to overlook the point that such descriptions will typically contain terms that refer to distinct properties or objects. In cases in which this isn’t so we get full synonymy, as in “the bachelor at the back is identical to the unmarried man at the back”. It is really always reference that ultimately determines whether words are synonymous or not.

Am I Certain That I Think?

Descartes wanted to build human knowledge on a foundation of certainty. He thought the Cogito provided an instance of certainty, and many have agreed (Montaigne was there before him). Critics have argued that the conclusion of the Cogito doesn’t follow from the premise (the Lichtenberg objection). However, the premise itself is seldom questioned; it at least is certain, whether or not it entails the existence of a thinker. But is it certain? First, we need to ask what it means: when I say “I think” what proposition am I asserting? Does it mean the same as “I am a thinker”? And does that imply “I am a rational thinker”? Am I asserting that I think rational thoughts?  But there are two problems with this interpretation: (a) how can I be certain that my thoughts are rational and (b) how can I be certain that I have thoughts plural? What if there is a brain disease that renders my thought processes irrational while giving me the illusion that they are rational? I can’t rule this out with certainty. And how can I be certain that I have thoughts in the plural given that this will require a series of thoughts over time? I can’t be certain I have had thoughts in the past since memory is fallible, and the existence of future thoughts is even less certain. So, should “I think” be read as “I have a thought now, which may or may not be rational?” But is it possible to have only one thought and it be completely irrational? Would that even be a thought? Thoughts occur in series of linked thoughts (“temporal holism”), and their title to being thoughts surely depends on some degree of conformity to rationality, however minimal. But these requirements are not open to certainty. In addition, thoughts in the ordinary sense have intentionality, but is that open to certainty? Couldn’t I be having thoughts about nothing—how can I be certain that my thoughts are aboutthings? What thought is it that I am supposed to be certain I am having? Is it the thought that it is (say) raining in Miami now? But how can I be certain my thought is about Miami and rain—what if I am a brain in a vat and content externalism is true? Is it the thought that I think (“I think that I think, therefore I am”)? But then we are back with the problem about rationality and thought plurality. It looks as if we need to retreat inwards a step: “I seem to think, therefore I am”. We thus cancel the proposition that I really think and maintain more modestly that I appear to myself to think. This falls far short of the original Cogito in that it concedes that I cannot be certain that I think: that proposition cannot be the premise the Cogito works with. We can gloss “I seem to think” as “I may be thinking now but maybe I am not—it only seems to me that I am”. Is this strong enough to get the conclusion we want? It seems like a very flimsy foundation for erecting a tower of certain knowledge. And if we are to retreat to a “seems” version of the premise, why not say simply “I seem to exist, therefore I exist”? For, if we can’t appeal to an unqualified “I think” but must restrict ourselves to “it seems”, then we may as well do it straightforwardly and honestly—”I seem to myself to exist, therefore I must exist”. But that looks like a straight non-sequitur: how can my existence follow from the mere appearance that I exist? Don’t we need the existence of something real to get anywhere—as in the existence of real thoughts? How can seemingentail being? Certainly, this revised Cogito looks weird and fishy compared to the classic version—hardly self-evidently true. This impression is confirmed by raising another skeptical possibility, namely that we can’t get existence out of seeming existence—out of the mere impression of existence. Nor can we get it out of the mere impression of thoughts as opposed to real thoughts. Suppose you are Meinong and believe in different kinds of being, subsistence and existence: maybe you can derive self-existence from genuine thought-existence, but how can you derive it from mere appearances of existence? Surely subsistence could seem like existence, so it may be that the self you derive from “I seem to exist/think” is just a subsistent entity not an existing one (like a hallucinatory golden mountain that looks real). In order to rule this out you would need to refute Meinong by claiming that all subsisting objects are existing objects—or else the Meinongian will insist that the conclusion could only be that I have some sort of being. In other words, the seeming premise only licenses an inference to “The subject of this seeming has being of some sort”. To be more specific, how can I be certain I am not a fictional character? It can seem to a fictional character that he thinks and exists, but he is not thereby a thinking existing being like you and me (as we believe, at any rate).[1] Or if you hold that fictional characters really have existence, then the new Cogito has to conclude “Either I exist in the real world or I am a character in fiction”—not quite what Descartes was hoping for. The problem is that seeming is too weak to give (real) being, but we can’t justify the original premise that asserts certainty about the existence of thoughts (thinking). Can we assert the existence of real seemings? If so, we might be able to move to real subjects of seemings. But of course, there can be fictional seemings: it can seem to a fictional character that things are thus-and-so. I therefore need a way of showing that my seemings are real, but all I am entitled to is that they seem real. We are now at an argumentative stalemate, whereas the original Cogito at least purports to start from a premise about incontestable reality, viz. the known reality of thoughts and thinking. The idea was that we can know with certainty that thinking really exists and then move from this to the existence of the self. But it now turns out that only a “seems” premise is acceptable, which undercuts the move to an “is” conclusion (“I exist”). So, Descartes needs his version of the Cogito and not the weakened “seems” version: if the argument were equally good under the “seems” interpretation, he could have simply said “I seem to exist, therefore I exist”—which looks like a blatant non-sequitur.[2] But the original version is open to skeptical doubt about the premise itself, not merely about the validity of the inference. The lesson is that trying to base human knowledge on certainty is hopelessly quixotic and should not be attempted (unless we stick with the ordinary man’s use of “certain” and forget about philosophical skepticism). Or to put it differently, an epistemology of science should not be in the business of refuting skepticism. It will only take a sound beating if it chooses to take on the skeptic (like Don Quixote himself fighting foes far more powerful than him[3]). Descartes clearly had problems with his attempt to prove the external world (what with the ontological argument and God’s supposed non-deceptive nature), but in fact the problems begin much further back with his attempt to find at least one indubitable proposition. It is the whole Cartesian project that is at fault. Certainty is a false god.

[1] This is exactly the situation of two characters in Sophie’s World by Jostein Gaarder (1994): they think they are real but discover that they are characters in someone else’s work of fiction. Fictional characters cannot use the Cogito to prove they are real.

[2] There is also the problem that in order to know that it seems to me that I exist, or that I seem to have thoughts, my seeming must have intentional content; but how can I be certain that it does, given that this doesn’t depend solely on my inner state? The skeptic will argue that my seeming might not have the content I take it to have, if I am a brain in a vat or sufficiently mentally deranged or confused. Episodes of seeming can be as subject to skepticism as thoughts are with respect to their content. Things are not as simple as Descartes supposed.

[3] Don Quixote de la Mancha cuts quite a philosophical figure: even after a series of terrible drubbings he refuses to abandon his absurd delusions, despite the entreaties of Sancho Panza (the voice of common sense). He is the living embodiment of human error—of life inside Plato’s cave. Descartes is quixotic in the precise sense that he keeps fighting for his misguided ideals even when he is clearly outgunned by the evil demon; he is too entranced by the romance of complete certainty. There is a nobility to this, but it’s none too realistic.

A Philosophy of Seeming

In “Seeming” I introduced seeming as a sui generis psychological natural kind. Here I will explore its uses in philosophical thought—the kind of impact it would have on philosophy were it to be taken seriously. I won’t repeat what I said in the earlier paper (this one should be read in conjunction with that) but the basic idea is clear enough: a seeming is an impression that something is the case—an impression of being. It is not to be assimilated to a belief, a sensation, an experience, a state of consciousness, a piece of knowledge, or a mental content. Roughly, it is an invitation to assent—a kind of intimation or suggestion (which may be rejected). It is not to be confined to the senses: although the senses do deliver events or episodes of seeming, they are not the only sources of seeming. There are also memory seemings, moral seemings, modal seemings, logical seemings, mathematical seemings, probabilistic seemings, introspective seemings, semantic seemings, other-minds seemings, existential seemings, negative seemings, and artistic seemings. Seeming is everywhere; it is hard to find a place where it isn’t. There don’t appear to be future seemings or unconscious seemings or self-reflexive seemings, but otherwise they are ubiquitous. They generally function epistemically: they provide reasons for beliefs or actions or emotions. They have intentionality. They interact with inference to produce knowledge. They can be episodic or dispositional. The concept of seeming is not a family resemblance concept (if such there could be); it is a genuine psychological natural kind with a real essence and a causal profile. Indeed, it is a biological natural kind, since psychology is really a branch of biology (the mind being a biological organ or system or faculty). There are many modalities of seeming, to be sure, but they are all united in a single psychological type, viz. impression of being. If it seems to you that something is so, you are under the impression that things are a certain way—the world strikes you as being thus-and-so (though you may reject this impression in the light of other impressions). Seemings are generally fallible, defeasible. Seemings are ontologically committed, but we can suspend these commitments in thought and speech. There are seemings of sense and seemings of reason; there are even mystical seemings and supernatural seemings. They have a distinctive phenomenology, a characteristic linguistic expression, and a logic of their own (referentially opaque, open to quantifying in). Presumably, animals have them too, and human infants; no doubt our species inherited them from earlier species (there were dinosaur seemings). They have a developmental psychology that might be empirically investigated. There is a neural basis for them. There must be a cognitive psychology underlying them (computational seeming algorithms, etc.). Some seemings will be alien to us because not shared by us (bats, sharks). And some will be intellectually beyond us, as ours are intellectually beyond simpler creatures (bats, sharks). Seeming science (“seemiology”) lies just over the horizon, I am convinced, poised to attract funds. However, let me now raise a few philosophical questions about this notion. First, what is the connection between seeming and concepts? Might it be true that all concepts require a basis in seeming? What about concepts belonging to the “absolute conception”, i.e., maximally objective concepts? Evidently, concepts of shape and color are tied to acts of seeming: to have these concepts is to be capable of its seeming to one that they are instantiated. But is this true of the concepts of mathematical physics or the concepts of morality or law or depth psychology? Has it ever seemed to anyone that General Relativity holds of space, or the Copenhagen interpretation of quantum mechanics? I think it is doubtful that these concepts have a basis in seeming, though the evidential warrant for the theories in which they occur may well involve acts of seeming. Do we have any idea of how curved space-time might seem? Our theoretical concepts can transcend our capacity to be subjects of seeming, even very abstract seeming. In other words, the View from Nowhere (the “absolute conception”) is not tied to seeming; so not all concepts are rooted in the capacity to host seemings. Perhaps many of our moral and legal concepts are (these are not particularly theoretical), so that they are linked to more primitive sorts of reactive psychological capacity; they don’t float free of our ability to be struck by the world. But concepts like infinity (of space, time, and number) are independent of our ability to entertain episodes of seeming—no one has ever had an impression of an infinite object or collection of objects. Second, acceptance of the psychological reality of seeming allows for a natural description of what is going on in cases of seeing-as. In the duck-rabbit drawing, for example, we can say that it both seems to the subject that the picture has changed and that it has not changed: both impressions exist simultaneously. The lines don’t seem to change their shape and orientation, but it comes to seem as if a different animal is depicted. So, we don’t have to limit ourselves to the concepts of sensation and belief; we can employ the richer terminology of seeming to capture the phenomenon. We now have a further resource, drawn from folk psychology, to work with. Third, we also have a new way to formulate epistemological issues: does all knowledge rest on a foundation of seeming? Is everything we know derivable by inference from premises about episodes of seeming (not sense data as traditionally conceived, or physical stimuli). This gives foundationalism a new freedom, liberating it from dogmas of the given or the universal reach of the human senses. Seemings have the power to justify, and they crop up in every domain of knowledge, perceptual and other. We can thus retain the attractions of foundationalism without its dubious baggage (the senses by themselves never justify, but neither is it true that only beliefs can justify). The knowing mind just has to be conceived more broadly, with seeming occupying a central epistemological place. Among other things, this will enable us to provide a uniform epistemology for ethics and physics: both rest on seeming—of value and physical fact, respectively. Reasons are all of essentially the same kind (the natural psychological kind seeming). Seeming thus provides a unifying epistemological framework. Fourth, we can supplement the old dispute between empiricism and rationalism with a new question—whether all knowledge rests on seeming or not. Empiricism and rationalism both base knowledge on seeming (sensory or ratiocinative), so they are not as far apart as we might have supposed; but now we have the further question of whether any knowledge is, or could be, based on nothing impression-like at all. Granted, in most cases knowledge is based on reasons deriving from events of seeming, but might there be cases in which not even this is true? These would be cases in which our reasons don’t advert to anything like a seeming but simply go directly to the heart of the thing in question—as it might be, direct insight into numbers. The numbers don’t seem a certain way to our intellect, but we nevertheless know their properties. We bypass seeming. I confess I don’t know what to say about this, but the issue is interesting and prima facie real. There is no natural name for the doctrine I am envisaging but it exists in logical space. We could call it “impressionism” but that has other connotations, and “seeming-ism” is too coy. Maybe some Latin term could be pressed into service. For now, I will just call it “S-epistemology” and stipulate that this is to mean “the doctrine that all knowledge, a prioriand a posteriori, is based on episodes of seeming, sensory or ratiocinative”. Then we can ask whether S-epistemology is true or not: this question is orthogonal to the old empiricism-rationalism debate, and arises out of the project of reconfiguring epistemology around the notion of seeming (construed as neutral between sense and reason). My guess is that S-epistemology is true across the board, but it is a substantive question whether this is so. No doubt there are further issues that can be raised under the new dispensation, but let that suffice for now.[1] Certainly, a lot of philosophy takes on a new aspect once the generalized concept of seeming is allowed to penetrate its precincts. This forces us to rethink our conception of the mind and that cannot but affect the way many philosophical issues are approached. The powers of the mind have been significantly expanded.[2]

[1] I will simply list a few issues that I have not discussed. Are Fregean senses (modes of presentation) explicable in terms of seeming? How does an epistemology of seeming bear on skepticism: does it make skepticism stronger or weaker or leave it the same? Are certainty and seeming compatible: if we are certain that p, can it also seem to us that p? Does intending involve seeming, or does the lack of future-seemings rule this out? Is seeming subjective or objective or both? Do we have impressions of nothingness? Could it seem that things are contradictory? Are Hume’s “impressions” the same as impressions as here characterized? Are seemings conceptual? Are there higher-order seemings? How are the concepts of seeming and evidence related? Could there be perception without seeming? Does pain involve seeming—if so, of what? Could knowledge be analyzed as seeming plus being? Do we have an impression of the self? Could we ever have an impression of our whole nature?

[2] I mean that we now have a general faculty of seeming, applicable across many domains, and capable of providing reasons in each domain. It has the generality of reason but it isn’t a kind of reason. It is a sui generis mental faculty not reducible to the traditional categories, lying somewhere between sensation and belief. Seeming is a vital and irreducible component of mental life.

Naming, Necessity, and Mind

I propose to offer an interpretation of Kripke’s Naming and Necessity that has not (I believe) been offered before. I do not say that this interpretation consciously occurred to the author of that work—in fact, I think it didn’t. But I do claim that it illuminates what is going on argumentatively in the text, and it should have occurred to Kripke (though it didn’t occur to me until very recently). Nothing in the text goes against it as far as I can see. It is an interpretation that was suggested to me by reading the text but not one that Kripke intended; I believe, however, that he would have welcomed it. It really systematizes what he was arguing by unifying several strands of argument taken as separate. Let’s think of it as produced by a philosopher named “McKripke” (I should say that I am very inclined to accept the interpretation as correct philosophy). There are three main topics: the nature of naming, the nature of necessity, and the nature of mind. In each case a certain dialectic is set up exhibiting a common form, which I will describe as follows: Is naming reducible to describing? Is necessity reducible to certainty? Is the mind reducible to the body? We begin with a certain chunk of discourse and we ask whether that chunk can be analyzed by another chunk, thereby reducing the subject matter of the first chunk to the subject matter of the second chunk. Thus, we have certain words called “proper names” and we speak of them as “naming” or “denoting” or “labeling” certain entities: the question then is whether this discourse can be reduced to discourse concerning what we call “descriptions” that are said to “describe” or “characterize” or “attribute properties to” certain entities. That is, we ask whether a description theory of names is true. Similarly, we speak of necessity, saying things like “This table is necessarily made of wood”, and we ask whether such talk can be construed as expressing propositions about what we know with certainty—as in “It couldn’t turn out that this table is not made of wood” or “This table is certainly made of wood”. That is, we ask whether an epistemic theory of necessity is true. Thirdly, and similarly again, we start off using words like “pain” and “belief” and we ask whether such talk can be understood as referring to (perhaps meaning) the same as words referring to the body and brain. That is, we ask whether a materialist theory of the mind is true. Some philosophers have maintained each of these reductions, explicitly or implicitly. Thus, Frege and Russell on names, the positivists on necessity (via the notion of analyticity), and many philosophers committed to a physicalist picture of the world. Kripke opposes each of these reductions (analyses, explanations): he thinks naming is not reducible to describing, necessity is not reducible to certainty (epistemic necessity), and the mind is not reducible to the brain (including its functional properties). In each case, he mounts a three-pronged attack: semantic, metaphysical, and epistemological. In the case of names, he argues, first, that sentences of the form “NN is the F” are not (generally) analytic, e.g., “Plato was the teacher of Aristotle”: this is the semantic argument. He then argues that the reference of a name could have been other than the description entails, e.g., Plato might never have taught Aristotle (in some possible world he didn’t): this is the modal (metaphysical) argument. Third, he argues that knowledge of the description commonly associated with the name (or associated with it only by the individual speaker) is not an infallible guide to what its actual reference is, as in the Godel-Schmidt case: this is the epistemological argument. So, the attempted reduction to descriptive reference doesn’t work: names are not disguised descriptions, the naming relation is not a species of the describing relation, naming is a sui generis type of referring. He then goes on to make some suggestions about how naming actually works, introducing causal-historical chains etc. The lesson is that referring takes two fundamentally different forms, naming and describing, neither being reducible to the other; so, names have a different kind of meaning from descriptions. We might call this strategy of argument the “SME strategy”—semantic, modal (metaphysical), and epistemological. It purports to show that naming cannot be analyzed as describing. This is all very familiar and involves little in the way of creative interpretation on my part; I’m just recapitulating the text, more or less. The next stage of interpretation, however, requires some bolder moves, some hermeneutic impositions. Kripke contends that there are two notions of necessity that must on no account be confused, which he calls epistemic and metaphysical necessity; in particular, we must not try to reduce the latter to the former. He gives several examples of metaphysical necessity and argues (convincingly) that they are not reducible to, or explicable in terms of, epistemic necessity. I will put this simply as the claim that metaphysical necessity is not the same as certainty (infallibility, incorrigibility). It is a distinctive type of necessity, over and above epistemic necessity, aptly labeled “metaphysical”. Here again we may apply the SME strategy, though Kripke does not explicitly do so. First, the sentence “Necessity is certainty” is not analytic, so the predicate term cannot analyze the subject term; it expresses a synthetic proposition. Second, there are cases of necessity without certainty and certainty without necessity, so the conditions are neither necessary nor sufficient (metaphysically speaking)—as with “This table is made of wood” and “I am in pain”. Third (and now we add something to N & N), knowledge of certainty does not add up to knowledge of necessity: you can know that something is certain without knowing it is necessary (even when it is), e.g., a mathematical truth, and you can have the former concept without having the latter (you might be modally blind or otherwise incompetent). Knowledge of one property does not give you knowledge of the other property, so they can’t be the same property. There is no a priori entailment from one to the other—because the relevant terms don’t mean the same thing. Just as knowledge of an associated description doesn’t give you knowledge of reference in the case of names, so knowledge of what is certain does not give you knowledge of what is necessary. In sum, the concept of necessity (metaphysical) is not the same concept as the concept of certainty—as the three prongs of SME demonstrate. Thus, no reduction is possible; we are dealing with an irreducible type of fact. A certain modal dualism is therefore in order, as a certain referential dualism is in order for names and descriptions. That is Kripke’s basic message, though he doesn’t put it this way: the arguments are logically similar, structurally parallel. This prepares us for the third main topic of N & N: the mind-body problem. And now it is easier to see the wood for the trees—the argumentative pattern stands out more clearly. The SME strategy is written all over it: first, we have the point about the meaning of (e.g.) “pain”; then the modal counterexamples; then a version of the knowledge argument. Suppose someone claims that “pain” can be analyzed by a bodily or cerebral predicate (let’s say “C-fiber firing” to stick with tradition): the obvious initial reply is that “pain is C-fiber firing” is not analytic, so “pain” can’t mean the same as “C-fiber firing”. Then, second, we have the possibility of pain without C-fiber firing and C-fiber firing without pain (i.e., possible worlds where these things are so). These possibilities rule out an identity theory, given that identity is a necessary relation. Third, we have a knowledge argument: you can know that someone’s C-fibers are firing without knowing they are in pain; there is no a priori link between the two. Kripke doesn’t provide such an argument in N & N, but we know that he provided one later[1] and may have had the idea at the time of the earlier work. Thus, dualism of mind and body is affirmed, as opposed to reductive materialism: the mind is as distinct from the body as naming is distinct from describing, and necessity is distinct from certainty. The same argumentative resources are deployed in all three areas: the meanings aren’t the same, the connection (correlation) is merely contingent (if it exists at all), and knowledge of the one doesn’t give knowledge of the other.[2] Thus, no reduction is possible and duality is the outcome. We can now note various things about this interpretation of Kripke’s treatment of these topics, which may serve to underline the basic structural commonality. One: the topics themselves are not closely related; none is a special case of the others. Indeed, they are quite disparate; yet they each exemplify a common pattern of argument, a certain kind of philosophical method (a methodology). Two: Kripke is not averse to all forms of reduction; he offers some himself. He accepts reduction of natural kinds (“water is H2O”); he provides a quasi-reduction of the naming relation in the shape of causal chains; and he has no qualms about accepting Russell’s theory of descriptions (descriptions are reducible to quantifiers plus identity). Three: he says little positively about each of the areas he defends against reduction; his claims are largely negative. No full analysis of naming, no reply to Frege arguments against denotational theories, no effort to dispel disquiet about what such non-descriptive reference might consist in (it seems rather magical). Likewise, nothing much about the nature, provenance, and problems of metaphysical necessity (for example, how exactly is it known?). Also, no attempt to deal with objections to mind-body dualism or the supervenience of the mental on the physical (does he deny this?). Four (and connected): aren’t each of these topics puzzling, mysterious even? How does the mind manage to reach out to objects without descriptive or conceptual mediation? Isn’t metaphysical necessity rather spooky, inexplicable, contra-empirical? What is it? And isn’t the idea of a separate mental substance rebarbative and contrary to science? In each area there are intimations of the “queer”, the “occult”, the “non-natural”? Yet we don’t find Kripke agonizing over these questions, or even raising them. He seems oddly cavalier. Interpretatively, this provides insight into his philosophical predilections and potential vulnerabilities. Six: he doesn’t tie the topics together but treats them seriatim and independently, perhaps not recognizing the commonality of philosophical moves. Seeing them together gives a stronger sense of the philosophical geography (even geology): we see their place on the map and their deeper underpinnings. Even such diverse problems display a common form, a parallel dynamic—as we are driven this way and that. There is a reason why philosophers favor such quixotic reductions—because the alternative strikes them as worse. The primitive is apt to produce the problematic—unanalyzable denotation, modal metaphysical excess, mysterious mind-stuff. Kripke was too good a philosopher not to be aware of these intellectual dynamics, these conceptual pushes and pulls, but he did little to articulate them or stress their unavoidability.[3] In any case, the indicated interpretation of N & N is along the lines I have suggested: a common argumentative thread hidden beneath the surface, shaping the dialectic, determining the conclusions. He could have called the work Naming, Necessity, and Mind: Some Common Themes. The book is more unified than might appear.[4]

[1] See Adriana Renero’s work on Kripke’s 1979 lectures on philosophy of mind.

[2] It may be useful to give an example of a successful analysis or conceptual reduction–Bernard Suits’s definition of games in terms of freely chosen obstacles to a given goal (see his The Grasshopper, 1978, and my Truth by Analysis, 2012). First, if correct, the analysis will produce an analytically true sentence along the lines of “A game is a voluntary attempt to overcome unnecessary obstacles”. Second, there will be no possible worlds in which the analysis fails. Third, knowledge of the definiens will suffice for knowledge of the definiendum. A good case can be made that Suits’s definition meets these conditions.

[3] Perhaps the most profound and intellectually revealing thing he ever said was, “I regard the mind-body problem as wide open and extremely confusing” (footnote 77, p.155, the very last words of Naming and Necessity). He realized how vexatiously difficult philosophical problems are, but also how intriguing.

[4] Kripke tried to impose argumentative unity on Wittgenstein in Wittgenstein on Rules and Private Language, with mixed results. I am suggesting that his own work displays more unity than appears on the surface: there is a pattern to the failures he discerns, despite the variety of subject matter.