Jerrold Katz, when he passed away early in 2002, had just signed the contract for the publication of his last book, Sense, Reference, and Philosophy. Late in 2003 the book finally came out (actually copyrighted "2004"). If the errors of linguistic/analytic philosophy (what Katz calls "L/AP") were like much of 20th century philosophy falling into a well, with Jerrold Katz we have finally climbed back out and now again stand in the open air and wind and sunlight -- like Robert Duvall having climbed out of George Lucas's underground dystopia at the end of THX 1138 (1971). Indeed, that is the image with which I would like to associate Jerrold Katz.
...the closest thing the English-speaking world has to a broadly accepted view of the nature of philosophy is L/AP. [p.187]
This is somewhat chilling when Allan Bloom, as far back as 1987 in The Closing of the American Mind [Simon and Schuster], wrote:
Positivism and ordinary language analysis have long dominated, although they are on the decline and evidently being replaced by nothing. These are simply methods of a sort, and they repel students who come with the humanizing questions. Professors of these schools simply would not and could not talk about anything important, and they themselves do not represent a philosophic life for the students. [p.378, boldface added]
Whether Katz or Bloom are right, neither the living presence nor the mere inertial continuation of classic L/AP speaks well for the state of academic philosophy. What was the worst about all this stuff was the aim of much of it to justify why the philosophers involved were no longer seriously interested in metaphysics or ethics -- the truths of Being and Value, Bloom's "humanizing questions." If metaphysics and ethics are either meaningless or just not matters of knowledge, then philosophy doesn't have to worry about them.
In his late period [I might say, even in his late period, ed.], Wittgenstein, like Carnap, continued to pursue his former positivist aim of showing that metaphysical sentences are nonsense. [Katz, p.190]
The result was that no ordinary sensible person would find anything of interest in philosophy; and students looking for discussion of the meaning of life went elsewhere. Although the L/AP generally revered science, the gnawing skepticism and nihilism of their principles was ultimately turned against science itself by the deconstructionists and post-modernists. Dismissing science as Euro-centric oppression is now conventional wisdom in English and history departments, with Wittgenstein often quoted in support.
In The Metaphysics of Meaning (1990) Katz served notice that the naturalistic theories of people like Wittgenstein and Quine had failed and that it was going to be necessary to do some actual metaphysics. In Katz's Realistic Rationalism (1997) we learned that the metaphysics was going to be realistic, with a rationalistic epistemology. In Sense, Reference, and Philosophy we do not get a lot more about just what that metaphysics is going to be, but the case is further consolidated against L/AP and the directions for development clarified. Like Socrates, Katz shows people their ignorance, with the implication that it is time for us to learn better. My own view, of course, is that the indicated metaphysics should be Kant-Friesian. I never noticed that Katz had any particular objection to the program or methods of the Rationalists, whose history I regard as a reductio ad absurdum of their project. I think he mistakenly regarded Kant as a conceptualist on universals and accepted the conventional wisdom that he was wrong about geometry. I don't think this was promising for the metaphysical doctrine Katz might have tried to developed. But that is another question.
From the wealth of Katz's book, let me focus for the moment on three of Katz's criticisms.
Frege's concept of a logically perfect language has survived in contemporary analytic philosophy in the form of regimentation. [p.202]
"Regimentation" means paraphrasing or restating sentences in forms required by the impoverished grammar of symbolic logic or another artificial language -- and then claiming that this is the "real" semantics of the natural sentences. Thus, even though almost no one today would disparage natural language and claim that a "logically perfect language" can and should be created, nevertheless the spillover from Frege's project influences contemporary thinking. All of this involves a clinging to the notion that a proper understanding of language means that substantive truths can be established about the world, and that metaphysics and ethics can be ejected from possible sense or knowledge. The conflation of language and theory, in the end, means that a proper understanding of language can supposedly be used to reject metaphysics. This was not Frege's project, but it became the principal attraction of the method in the analytic tradition.
Separating language from theory provides a tool for resolving many traditional problems. One of these is what Katz calls the "Epimenidean dilemma" [pp.93-102]. Epimenides of Crete is supposed to have said, "All Cretans are liars." Since Epimenides was a Cretan, if he was speaking the truth, then he would have to be a liar, asserting a falsehood, which means that not all Cretans would be liars. But if not all Cretans are liars, then what Epimenides says is false, which means he is lying, which means that what he says can be true. This is a version of the "Liar" paradox discussed in Hellenistic philosophy, which reputedly led to the death of one frustrated philosopher. In modern philosophy, it leads to what Katz calls the "Epimenidean dilemma," which is that:
...we cannot say both that natural languages are in some sense universal or unlimited in expressive power and that they are consistent. [p.93]
If natural languages are universal in expressive power, then Liar paradoxes are part of them, introducing a contradiction and an inconsistency, just as Russell's Paradox introduced a contradiction into Frege's program of developing arithmetic out of logic (or analytically, by his definition, out of definitions and logic). This paradox can be defused either by rejecting Frege's definition of analyticity, as discussed below, or simply by separating language from theory. A Liar paradox introduces a contradiction into language only if it is part of language as a system of knowledge or a logical development of theorems from axioms. But if language is not an axiomatic system and is neutral on matters of substantive knowledge, then a Liar paradox introduces no contradiction. What system of knowledge is a Liar paradox part of? Well, if we are interested in Cretans, it is simply not true that all Cretans are liars. A simpler form of Liar is just the sentence, "This sentence is false" (the "Eubulidean" Liar, after Eubulides of Miletus). If it is false, then what it says is true. If it is true, then what it says is false. What system of knowledge is this part of? None. It is only about itself, but, unlike an analytic sentence, it cannot make itself true. Comparison can be made with the sentence, "This sentence is true." This sentence is not analytic either, but since it is not about anything else, there is nothing else to make it true or false. There is no epistemic procedure, except reductio ad absurdum, to assign it a truth value. The same can be said of the Liar sentence. It is then irrelevant that the Liar implies a contradiction on the (reductio) hypothesis of either truth or falsehood. Although much ink has been spilled over the centuries about whether the Liar is false or neither true nor false, it really doesn't matter. If a Liar paradox is distinct from language itself, then it really doesn't matter, and Epimenidean dilemma is resolved. Language can be universally expressive and imply no contradiction.
Mill's and Wittgenstein's examples show that the cluster of properties that are most naturally taken to pick out the bearer of a name in actual situations do not pick it out in some counterfactual situations. [p.21]
One of the most interesting of these problems is that of "false description," when one is mistaken about the sense of a word but is still in some obvious way referring to particular individuals. Thus, one of my students may write, "Socrates believed in reincarnation," confusing the ficticious Socrates of Plato's Meno for the historical Socrates, about whom the question was asked. Unless the issue is the actual personal refered to as "Socrates," it is wrong for me to correct the statement with, "No, Plato, not Socrates, believed in reincarnation." The referent of "Socrates" is thus independent of the sense of "Socrates" and indeed may be used, foundationally, to correct the sense (assuming we have access to Socrates, which is a larger epistemic issue).
For most of L/AP, the counterexamples to the principle of sense determining reference were used to deny that there were intensional senses -- often then used as evidence that meaning is extensional, i.e. consisting of the individuals to which the terms refered. Katz, however, properly argues that they only refute Frege's theory of senses, specifically the part of the theory that holds that sense determines reference. There are intensional senses; but sense does not determine reference. It's simple. But, unfortunately, it provides no traction for the project of the tradition to reject metaphysics. Thus, Katz introduces an "autonomous" theory of sense, that issues and truths involving sense are independent of issues and truths involving reference. This enables him to shed one mistake after another that was put forward by Wittgenstein or Quine.
In the theory of reference itself, there are two broad philosophical issues, one about names, the other about universals. In these pages, names and universals have been already discussed separately. I do not believe that Katz achieves the simplicity of explanation that he could in these matters. The independent meaning of the reference of a name is that it denotes a particular individual. With universals, the metaphysical question is most importantly about natural kinds. This comes up in Katz with one complication concerning false descriptions of natural kinds. Hillary Putnam proposed [p.21] the thought experiment that the creatures we have always refered to as "cats" might turn out to actually be "Martian robots that look and act like real cats." The question then is, if we have a false description of actual cats, what was the actual reference of the term "cat"? One possible answer is that, as the term "Socrates" refers to a particular individual, "cat" refers to a class or a set of individuals.
Frege himself, however, held that the reference of general terms is a "concept." The metaphysics of this was unexplored. What we would expect in a realistic metaphysics is that the "concept," as something objective and external, would be the essence of the thing. The essence is what makes the thing what it is, which enables us to understand the kind of thing, rather than just listing individuals that, for whatever reason, have been grouped in the class (the reason, of course, is the essence). False descriptions mean that we get the essence wrong, but as long as we are talking about the same thing, i.e. the same essence for the same kind of thing, we have a chance to correct our understanding. This is why a philosopher like John Locke, often cited by Katz, distinguished between nominal (what we are aware of) and real essence (what is really there). A false description that determines reference would mean that our reference would be different than what we thought. If cats turned out to be Martian robots, that would be rather surprising, but not so different than many things that have happened in science. From Heraclitus to Phlogiston, people thought that fire was a substance, a material, but then it turned out that it was nothing of the sort. Fire is a process, rapid oxydization, that can happen in certain circumstances to things that actually are materials. Thus, the essence of fire is now understood to be very different that what people thought for more than 2000 years. But there is no doubt that Heraclitus and Lavoisier were talking about the same thing.
If the content of concepts is determined on the basis of laws of logic, then there can be no concepts. The point goes back to Wittgenstein, who observed that, as a determinant of the content of concepts, P -> (P v Q) commits us to saying that the sense of a sentence S includes the sense of the sentence S or S' where S' is any sentence whatever... [p.30]
The logical law of "addition," where a disjunction can be created with a given sentence and any other sentence, means that the sense of any sentence implies the disjunction of all sentences:
Hence, the content of every sentence is the same as that of every other sentence: all sentences are synonymous. [p.30]
This blows away Frege's theory, but the point, naturally, was used by Wittgenstein to deny intensional senses altogether (in favor of his "usage," behavioristic, theory of meaning). Katz simply points out that it recoils only on Frege's, not on Kant's, conception of analyticity. Katz characterizes Kant's theory of analyticity, not as logical, but as "mereological," i.e. not as logical implication but as relations of part (Greek méros) to whole. The meaning of the predicate of an analytic proposition for Kant was already contained in the subject, was part of the subject. This is the "beams in the house" containment rather than Frege's "plant in the seed" containment. In Katz's autonomous theory of sense, this is absolutely right. Frege's theory had a strong appeal for people like the later Positivists, because they could take propositions that Kant regarded as synthetic a apriori, and if these could be logically derived from certain definitions, then it could be claimed that they were simply analytic rather than synthetic. Once Wittgenstein and Quine were through attacking Fregean analyticity, however, it looked like propositions were not much of anything -- determined by Wittgenstein's language games or other naturalistic, and basically non-cognitive, constructions.
Where I retain some sympathy for Frege is the result of a negative definition of analyticity that can be constructed from Kant's treatment. Thus in Kant, as in Hume, an analytic proposition is one that makes itself true by virtue of its meaning, i.e. it does not need a ground external to the sentence to make it true. This works pretty obviously in the mereological relationships of Kant's paradigmatic propositions, but it is also true for logical truths, as Frege may have realized, like P or not-P. Such a logical truth (the principle of the Excluded Middle) is analytic in terms of the negative definition, without involving a mereological relationship. Indeed, it seems rather like Kant's own example of 5 + 7 = 12 as a synthetic proposition (which is what it will be, if we reject Frege's conception of analyticity). Katz simply does not address himself to this issue. However, I believe that Katz does provide us with the tools to understand this by his treatment of "color incompatibility" [pp.151-165]. Thus, "Red is not blue" [p.155] is analytic, but the predicate, with the sense "blue," is not contained in the subject. Katz's treatment seems to me awkward, and I think he missed a significant way to both generalize and simplify the issue. Color incompatibility is a messy issue in part because they are many colors. However, "Good is not evil" is analytic in much the same way as "Red is not blue," but it is more obviously part of the discussion of opposites that goes back to the beginning of Greek philosophy.
Katz's theory is that, as red is a color, the sense of color contains the sense that colors are "members of an antonymous n-tuple," such that each color negates all others. The antonymous structure, while not obviously part of the meaning of "red," comes out in the sentences formed with it: "..the meaning of a word in isolation and the meaning of a word in a sentence" [p.164] can be different. Whether that is precisely the way to look at it, I am not sure, which is where opposites come in. Opposites are not "members of an antonymous n-tuple" but simply members of an antonymous double, where the duality is a more explicit and recognizable form of contradiction. Thus, in the Pythagorean table of opposites we find doubles like "light and darkness," "male and female" whose antonymity is buried in their sense, but also a double like "limited and unlimited," whose members only differ in the prefix "un-" (a-, the "alpha privative" in Greek), whose explicit sense is "opposite," "negation," or "privation." Many opposites might be considered members of an antonymous triple rather than double (male, female, and hermaphrodite; good, bad, indifferent), but opposites involving a privative prefix clearly are a double in which only a logical property, negation, supervenes on a single sense. Rather than "members of an antonymous n-tuple," what we can simply say of opposities is that part of their meaning is that they are, indeed, opposites. In "Red is not blue," part of the meaning of "red" is that it is an opposite to other colors. Since "blue" is understood to be a color, and an opposite, the syntactic negation in the sentence reflects the privative (mereological) meaning in the terms.
This brings us back to P or not-P. Every sentence is opposite in truth to its negation. We only have to change Katz's "..the meaning of a word in isolation and the meaning of a word in a sentence" to "..the meaning of a sentence in isolation and the meaning of a sentence in a compound sentence." This is not going to be different in principle -- both are "compositional" senses in Katz's terminology. The disjunction of P or not-P will work the same way as a sentence like, "The function will be either differentiable or non-differentiable." Indeed, P or not-P can be restated as "P will be either true or false" -- where the Excluded Middle means precisely, "'False' means 'not true'."
Katz treats of many other issues, but I will leave these to suffice for the moment. A remaining puzzle is Katz's dedictation of the book, in part to Quine himself. Katz has great respect for Quine because of the importance of his contribution to the debate and the inspiration of his example. However, if Katz is right, Quine has done as much damage to philosophy as good, and his entire approach, indeed his entire purpose for philosophy, is profoundly flawed and misdirected -- paradigmatic of the sterility and emptiness, let alone esotericism, of analytic philosophy. Something of the same problem occurs in the following quote:
To be sure, Locke's theory of ideas, Kant's theory of concepts and judgments, and Moore's account of analysis falls short of what modern standards of rigor demand in the way of philosophical clarity and precision. [p.199]
Exactly what thanks should philosophy have for "modern standards of rigor...in the way of philosophical clarity and precision"? The result of the L/AP tradition has largely been disastrous. Katz himself represents a kind of redemption of this, but then (1) Katz's views are not generally accepted, and (2) the amount of damage done, and perpetuated by the majority of the academic tradition, is a net loss for philosophy, let alone a waste of time (and public money in research institutions). Katz himself perpetuates one of the mistakes in Frege that he fails to identify. Frege, fulfilling a dream going back at least to Leibniz, wanted to turn logic into mathematics. The result of this was the notion that a formula is superior to a natural language sentence. Relying on formulas, however, opens one to the Sin of Galileo, in the form that excessive abstraction enables one to ignore important aspects of the meaning of an issue. Indeed, it means getting lost in one's own jargon, and then forgetting what the point was -- all the easier dealing with a desiccated formula rather than with a proper, semantically full thought. There is, indeed, little of this in Katz, but he does resort to formalistic diagrams in a few places, including the section about color incompatibility. If his thought was that this explains or accomplishes something that cannot be done in natural language, (1) I would be suprised, and (2) he would be wrong. But excessive abstraction and gratuitious symbolism are among the greatest evils of analytic philosophy -- one of the major reasons why the Positivists and people like Quine have been able to perpetuate grotesque or absurd errors for so long. It puts on a great show of "rigor," "clarity," and "precision," when the results may actually be incoherent or farcical (my unqualified judgment about Logical Positivism). Katz is better than this and rises above it, but he may not have realized just how bad it was. If he had gone on to work out the metaphysics that would be conformable to his theory, it may have become more obvious. Since he didn't, he does not seem to have freed himself sufficiently from the milieu of his origin. Katz is standing in the open air and wind and sunlight, looking towards a hopeful future for philosophy, but that is only the beginning.
The Metaphysics of Meaning, Jerrold J. Katz