Part 5 (1/2)
A second complaint of Cohen's cuts deeper. He claims that Frankfurt's account of bulls.h.i.+t is incomplete. It misses out on an important, and different, kind of bulls.h.i.+t. In focusing ”on one kind of bulls.h.i.+t only,” writes Cohen, Frankfurt did not address another, equally interesting, and academically more significant, kind. Bulls.h.i.+t as insincere talk or writing is indeed what it is because it is the product of something like bluffing, but talking nonsense is what it is because of the character of its output, and nonsense is not nonsense because of features of the nonsense-talker's mental state. (pp. 12122) On Frankfurt's concept of bulls.h.i.+t, the bull, to borrow Cohen's expression, ”wears the trousers”; bulls.h.i.+t is whatever we get from the bull. What we need, according to Cohen, is a bulls.h.i.+t- (rather than a bull-) centered account of bulls.h.i.+t-an account of bulls.h.i.+t or nonsense independent of facts about the person serving it up (such as, for example, her mental state). And Cohen delivers an admittedly preliminary account of bulls.h.i.+t in this sense, one that emphasizes as a sufficient condition of bulls.h.i.+t its ”unclarifiable unclarity” (p. 131). Something, a sentence for example, is unclarifiable ”if and only if it cannot be made clear.” It's disappointing that Cohen declines to say what 'clear' means -and, indeed, he lets on that he doesn't think it's even ”possible to [say what 'clear' means], in an illuminating way” (p. 131). But it's an ironic disappointment, so at least we have that.
Frankfurt and Cohen each have some ideas about bulls.h.i.+t, then, and, not surprisingly, they are at odds. My own idea about bulls.h.i.+t consists, I hesitate to divulge, in adding yet more ideas about bulls.h.i.+t to the mix, in the hope of resolving what appear to be irresolvable differences between Frankfurt's idea and Cohen's. This strategy of simplifying the conceptual stew by adding more things to it strikes many (my students especially) as a bit perverse. But it is in fact a reliable (and, partly for that reason, quite popular) way to actually make progress in these sorts of matters, especially when the ideas added to the mix have garnered decent reviews on the intellectual stage; when they have, that is, a respectable intellectual ancestry. Determining this strategy's success in the particular instance of this paper I leave as an exercise for the reader.
The ideas I'll be bringing to bear on the schism between Cohen and Frankfurt are the ideas a.s.sociated with logical positivism-the premier, pa.s.sionate, remarkably successful, and altogether thoroughly entertaining anti-bulls.h.i.+t philosophical program of the 1920s and 1930s (its end met, tragically but not surprisingly, at the hands of two of the twentieth century's premier bulls.h.i.+t programs, European fascism and the Red Scare106). There are clear parallels between logical positivism and contemporary anti-bulls.h.i.+t programs,107 and, in fact, I'm surprised that so far so few, Frankfurt included, have either noticed the parallel or drawn upon it to add to the current discussion.108 But no matter; perhaps this book, and even this essay, is a start.
The logical positivists were not shy about bringing the hammer down on bulls.h.i.+t, but as often they described what they were up to in the appropriately positive terms of promoting unity among all the domains of genuine knowledge. Their idea was that the unity of science meant making clear the connections various domains of knowledge bore to one another, and that that led to eradication of the hidden depths and dark recesses that could serve as ma.s.sive underground bulls.h.i.+t bunkers. ”Unity of science!” they sang, warbling 'science' in its very general, distinctly German, sense. So using the logical positivist's own ideas about bulls.h.i.+t, its source, and its eradication to unify Frankfurt and Cohen's accounts, as I intend, invokes, it seems to me, good anti-metaphysical karma.
No Bulls.h.i.+t, Please, We're Austrian.
Now the logical positivists railed against metaphysics rather than bulls.h.i.+t. If you happen to be worried that the positivist's metaphysics isn't Frankfurt's (or Cohen's) bulls.h.i.+t, then I aim to allay your worries by showing how the logical positivist's metaphysics unites Frankfurt's, and Cohen's, bulls.h.i.+t. The most famous instance of positivistic railing against metaphysics, I believe, is Rudolf Carnap's 1932 ”Overcoming Metaphysics through the Logical a.n.a.lysis of Language.”109 Philosophers know this essay as the one in which the young Rudolf Carnap takes the then far-better established German philosopher Martin Heidegger to task for writing, on p. 34 of his 1929 Was Ist Metaphysik?, ”Das Nichts selbst nichtet,” that is, 'The nothing noths'.110 Here was a choice bit of metaphysics, Carnap noted, and just the sort of thing that we ought to overcome (rather than ponder, examine, debate, or refute) by sober, and rather elementary, logical a.n.a.lysis of the sentence itself. Carnap's essay is hardly the only instance of a logical positivist a.s.sault upon metaphysics, but (recall my ridiculous deadline) it's the only one I'll consider here. There is on the ontological horizon a succession of interesting articles and books with t.i.tles like ”Neurath on Bulls.h.i.+t,” ”Reichenbach on Bulls.h.i.+t,” ”Quine on Bulls.h.i.+t,” and so on, and someone (other than myself) ought to write them.
Carnap's target, and the target of the logical positivists generally, was meaningless utterances, but not just meaningless utterances. Carnap is interested in the much more interesting topic of meaningless utterances that can be, and often are, presented as and widely understood to be meaningful-their utter-ers might, for example, present them (falsely) as though they had a meaning, or the people who read or hear such utterances might believe (again, falsely) that they have a meaning. Such utterances are pseudo-sentences, and Carnap's claim is that metaphysics consists precisely of such pseudo-sentences. Spelling this out means, first, giving an account of what it is for an utterance to have meaning (thereby identifying what it is for it to be meaningless as well) and, second, explaining how it is that meaningless utterances could ever be confused with meaningful ones-how, we might say, metaphysics happens. Perhaps it's not hard, having said even just this little, to see how Carnap's approach to metaphysics might incorporate both Cohen's bulls.h.i.+t-centered notion of ”unclarifiable unclarity” and Frankfurt's idea that bulls.h.i.+t is a certain intention, characterized by the disregard for the meaning of what one says (and, by that fact, for the truth). Both are afoot in the metaphysics Carnap begs us to overcome.
Examples work wonders for Carnap. He invites us to imagine an encounter with someone using the word 'teavy', a new word, or at least a word new to us: In order to learn the meaning of this word, we ask him about its criterion of application: how is one to ascertain in a concrete case whether a thing is teavy or not? Let us suppose to begin with that we get no answer from him: there are no empirical signs of teavy-ness, he says. In that case we would deny the legitimacy of using this word. If the person who uses the word says that all the same there are things which are teavy and there are things which are not teavy, only it remains for the weak, finite, intellect of man an eternal secret which things are teavy and which are not, we shall regard this as empty verbiage. (p. 64) The meaninglessness of 'teavy' stems, Carnap holds, from the fact that what Carnap calls the term's elementary sentences-sentences with the form ”x is teavy” (such as 'This world is teavy' or 'My brother is teavy')-cannot be deduced from other sentences. At the time, this meant for Carnap that the elementary sentences could not be verified.111 Utterances, as opposed to terms, can be meaningless as well, even when all the terms they contain are meaningful. There are trivial cases (Carnap offers 'Caesar is and'), but also cases like 'Caesar is a prime number', which might at first glance be taken as meaningful. This latter category contains pseudo-statements, things that aren't statements but might initially be taken to be (in Carnap's world, margarine, which isn't b.u.t.ter but can be pa.s.sed off as b.u.t.ter, would be pseudo-b.u.t.ter). Meaningfulness for utter-ances, as for words, amounts to a certain disconnectedness to other claims: we cannot, in principle, bring evidence to bear on the meaningless expression, either for it or against it. Carnap offers as examples of these Heidegger's 'We find the nothing' and 'We know the nothing'; 'The nothing noths' wins special Carnapian exasperation points for being not just a meaningless arrangement of terms, but for having among its terms a meaningless one to boot, the pseudo-verb 'to nothing' (p. 71).
A Little Carnap in Everyone.
Carnap's notion of meaninglessness, his main diagnostic tool in his battle against metaphysics, is a much more precise rendering of ”unclarifiable unclarity,” Cohen's main diagnostic tool in his battle against bulls.h.i.+t. This is why I am genuinely surprised that Cohen doesn't mention Carnap or, for that matter, any of the logical positivists. I believe the parallel is confirmed by a careful, sustained, reading of Cohen's paper alongside Carnap's; but, really, that's the sort of thing one ought to do in private.
But I will offer a consilience that bolsters my claim. Cohen offers nonsense as an example of bulls.h.i.+t, and by 'nonsense' he means not merely unclear discourse but discourse that can't be made clear: the mark of such unclarity is that ”any apparent success in rendering it un.o.bscure creates something that isn't recognizable as a version of what was said” (p. 130). This manner of identifying the unclarifiable by its disconnectedness to other statements or texts is just Carnap's strategy for isolating the meaningless. For Carnap, it is not as though there are antecedently meaningful sentences, and connecting a new sentence to one of these somehow infects the former with the latter's meaning. The idea, rather, is that something is meaningful just in bearing the right (presumably, logical) relation to other a.s.sertions or texts. And it's the same for Cohen: it's not as though there are clear sentences out there, the clarity of which seeps into other sentences if we position the latter correctly. Clarity is a matter of bearing the right relation to other claims. So it's a demonstration of profound unclarity if, in trying to show a sentence's connection to others, you inevitably mangle the claim with which you began into ”something that isn't recognizable as a version of what was said.” This shows that there was no such connection to begin with.
On Carnap's account, then, our language holds, for us, its users, a danger. For in allowing for the formulation of nonsense words and meaningless expressions it allows us to lapse into bulls.h.i.+t. Carnap frequently mentions the possibility of being ”misled,” or ”seduced,” by our language, and he means misled or seduced into metaphysics. But that is not the only danger. Because our language allows for the formulation of pseudo-words and pseudosentences, it is a powerful and effective tool that can be exploited by those whose aims are served by misdirection or the obfuscation of truth short of lying, that is, by bulls.h.i.+tters in the Frankfurtian sense. By the very fact that they present meaningless statements as meaningful they express their disregard for the truth, and by the fact that they utter meaningless statements they can't be lying; they are, after all, saying nothing.
In this regard, consider these two more pa.s.sages from Carnap's essay, each of which emphasizes the intention of the metaphysician-the bulls.h.i.+tter, as I read it. The first invites us again to imagine a new term, 'toovy' this time, which, in contrast to 'teavy', is meaningful: Let the sentence ”this thing is toovy” be true if and only if the thing is quadrangular.... Then we will say: the word ”toovy” is synonymous with the word ”quadrangular.” And we will not allow its users to tell us that nevertheless they ”intended” something else by it than ”quadrangular”; that though every quadrangular thing is also toovy and conversely, this is only because quadrangularity is the visible manifestation of toovyness, but that the latter itself is a hidden, not itself observable property. We would reply that after the criterion of application had been fixed, the synonymy of ”toovy” and ”quadrangular” is likewise fixed, and that we are no further at liberty to ”intend” this or that by the word. (p. 64) For Carnap, the example is intended to show that the meaning of a term is exhausted by the deductive relations.h.i.+ps the term's elementary sentences bear to other sentences; anyone who claims for a term a meaning not captured by those deductive relations.h.i.+ps cannot be offering us a meaning at all. But the example tells us as well about the intentions Carnap clearly thinks are tangled up with metaphysics. We too would dismiss anyone who continued to profess additional meaning for 'toovy' after its synonymy with 'quadrangular' had been laid bare. What's her problem?!? Absent appeals to absurdity, comedy, or idiocy (three well-known bulls.h.i.+t-defeaters deserving of much more philosophical attention), such flagrant disregard for meaning can be explained only by concluding that it was never the toovy-talker's intention to convey information to us in the first place, or even steer us away from some information (to, that is, lie). She must have wanted to accomplish some other end for which uttering such pseudostatements would be of use. 'Toovy' was a meaningful term all along, but in her disregard for the term's meaning the toovy-talker was engaged in metaphysics . . . that is, bulls.h.i.+tting.112 The second pa.s.sage I have in mind, in which Carnap comments on the intentions behind metaphysics, comes, interestingly enough, in the context of Carnap's answer to the question of why there is so much metaphysics, and why we seem to put up with it-confirmation, incidentally, of my view that Carnap and Frankfurt are talking about the same thing. ”How could it be,” Carnap asks, ”that so many men in all ages and nations, among them eminent minds, spent so much energy, nay veritable fervor, on metaphysics if the latter consisted of nothing but mere words, nonsensically juxtaposed?” (p. 78). How, indeed? What's with all this metaphysics?
Carnap's answer is that metaphysics is a consequence of a desire to express some ”general att.i.tude towards life” (Lebenseinstellung) combined with a mistaken impression that an att.i.tude (towards life or anything else) is a state of affairs, that is, the kind of thing that can be expressed by a declarative sentence. An att.i.tude towards life can be expressed, but only in art, poetry, or music; to attempt its expression in a.s.sertions, as though the att.i.tude were not an att.i.tude but a state of affairs, is futile.113 ”Thus in the case of metaphysics,” Carnap writes, ”we find this situation:”
Through the form of its works it pretends to be something that it is not. The form in question is that of a system of statements which are apparently related as premises and conclusions, that is, the form of a theory. In this way the fiction of theoretical content is generated, whereas, as we have seen, there is no such content. It is not only the reader, but the metaphysician himself who suffers from the illusion that the metaphysical statements say something, describe states of affairs. The metaphysician believes that he travels in territory in which truth and falsehood are at stake. In reality, however, he has not a.s.serted anything, but only expressed something, like an artist. (p. 79) On its face, this diagnosis renders metaphysics rather benign. That is a strength, of course, if the question is why we tolerate it. And as far as the parallel with bulls.h.i.+t goes, it gives us another answer to the question of why bulls.h.i.+t is both ubiquitous and tolerated. To wit: bulls.h.i.+t arises when people have something they want to get across and are confused, perhaps but not always culpably so, about what tools are appropriate to that task.
But alongside these somewhat contented observations about metaphysics, and bulls.h.i.+t, in our life, there is of course a critical current in Carnap, and Frankfurt, and we can't afford to miss it. Carnap describes the case in which the metaphysician as well as his audience is under the illusion that his utterances make sense, but there are, as Carnap was more than aware, cases in which the metaphysician, but not the audience, is under no such illusion. After all, metaphysics can only ”pretend to be something that it is not” if behind the metaphysical utterance is a metaphysician pretending to say something, knowing at the same time that he is not. This is bulls.h.i.+t, Frankfurt-style, pure and simple. It's more egregious, of course, to the extent that the metaphysician-bulls.h.i.+tter propagates the illusion in his followers even after we've called him on his insolence regarding the meaninglessness of his utterances or, in Frankfurt's phrase, regarding the truth-value of his claims. Heidegger was a metaphysician before Carnap penned ”Overcoming of Metaphysics through Logical a.n.a.lysis of Language,” but his metaphysical bulls.h.i.+t was more offensive after Carnap called him on it. Ditto, mutatis mutandis, for modern-day Frankfurt bulls.h.i.+tters. On Bulls.h.i.+t's placement on the New York Times bestseller list not only sold a pile of books; it raised the moral stakes on people who don't care about the truth of what they say.
The Unity of Bulls.h.i.+t.
Recall that Cohen employed a rather useful metaphor to distinguish his view from Frankfurt's: Frankfurt's account of bulls.h.i.+t focuses upon the bull, and his, Cohen's, starts from the bulls.h.i.+t. In light of our discussion of Carnap's anti-metaphysical program and my promise to unify Frankfurt's and Cohen's account, it will pay to return to the metaphor.
There is a certainly a distinction to be drawn between a bull and bulls.h.i.+t, and between bulls.h.i.+t and the bulls.h.i.+tter. But, my local veterinarians a.s.sure me, a bull that doesn't s.h.i.+t is no bull, at least not for long. And it takes, as David Hume might put it, no nice metaphysical head to realize that we get bulls.h.i.+t from a bull, not necessarily of course, but in fact, in this world and all the close possible ones. Cohen's metaphor not only serves his purpose, but it ought to remind us that the two sides of bulls.h.i.+t, Frankfurt's and Cohen's, are two sides of one thing. There are in this world those whose have ends that are served by a misuse of language, and whose desires trump or even eradicate any concern they might have had for the meaning or the truth of what they say. These are Frankfurtian bulls.h.i.+tters, and so be it. But we also have a tool, a language, that is amenable, perhaps even suited for, just the sort of misdeeds bulls.h.i.+tters have in mind. All of us have, or can, fas.h.i.+on Cohen-style bulls.h.i.+t on demand, and so be that. Combine the two and you have fodder for books like this one and Laura Penny's, and for that matter for the many, many, conversations held today that included the phrase ”This is such bulls.h.i.+t.” Carnap, of course, didn't know from Frankfurt and Cohen in 1932, but he knew bulls.h.i.+t. His approach to it gave us the intellectual goods on offer today from Frankfurt and Cohen, and then some.
Ah, but what to do? Is it any solace to have one account of bulls.h.i.+t over two, if our aim at the start was, implicitly at least, to get rid of the bulls.h.i.+t? The question of how to respond to bulls.h.i.+t is more pressing, and depressing, when we realize not just that the bulls.h.i.+t tide is rising, with no recess in sight, but that all those enthusiastic bulls.h.i.+t-eradication programs of yore, logical positivism included, have that rather embarra.s.sing odor of ambition-c.u.m-failure. In this context, the very end of Laura Penny's book might look at a first glance like the quintessential twenty-first century post-whatever reply to bulls.h.i.+t: a none-too-hearty ”Oh, well.”
But that's just a first glance. Here's a suggestion that may sound less antique and more plausible the more our intolerance for bulls.h.i.+t and its perpetrators grows. Our a.n.a.lysis of bulls.h.i.+t as one part tool (a language amenable to misuse) and one part intention (to put something over by means of that tool) invites a strategy oddly familiar to advocates of gun control: control the gun. In this case, of course, it's control the language, the tool that bulls.h.i.+tter's employ. No one, not Frankfurt, Carnap, Cohen, or Penny, suggests that we will eradicate from our midst those with intent to bulls.h.i.+t; indeed, sometimes that very enemy is us. Bulls.h.i.+tters are inevitable. But we can take in hand the tool the Frankfurtian bulls.h.i.+tter turns to, and needs: our language. This taking in hand need not be the fas.h.i.+oning of the ideal language that Carnap and many (though, notably, not all) of the Vienna Circle imagined, or even the conceptual clean up Cohen calls for.
Again, I'm moved by the last line of Penny's book, and I don't mean the choice verb. The language we use, English or whatever, is ours, together, and each of us bears responsibility for its misuse and abuse in our presence. When one of us misuses it, placing it in the employ of bulls.h.i.+t, it is no prissy matter of grammar or style that is at stake but a common temple being defaced. Your task, gentle reader, is to stand at the door of the temple. This can mean writing a book, or an essay, or a letter, or a blog, but it will also and more often mean holding a sign, raising a hand, casting a vote, or interrupting a conversation.
It's a daily, mundane, thankless, and unending task, but it will be the way out when the alternative becomes too much bulls.h.i.+t to bear. Take some heart that you will be joined, in spirit if not in corpus, with Carnap's robust colleague, Otto Neurath, who impressed upon his wide audience that the shape of the world around is the result not of reasons beyond our control-pseudorationalism, Neurath called this idea-but of our own choices.114 It really is up to us.
10.
Raising the Tone: Definition, Bulls.h.i.+t, and the Definition of Bulls.h.i.+t.
ANDREW ABERDEIN.
I always love that kind of argument. The contrary of a thing isn't the contrary; oh, dear me, no! It's the thing itself, but as it truly is. Ask any die-hard what conservatism is; he'll tell you that it's true socialism. And the brewers' trade papers: they're full of articles about the beauty of true temperance. Ordinary temperance is just gross refusal to drink; but true temperance, true temperance is something much more refined. True temperance is a bottle of claret with each meal and three double whiskies after dinner.115 Bulls.h.i.+t is not the only sort of deceptive talk. Spurious definitions, such as those quoted above, are another important variety of bad reasoning. This paper will describe some of these problematic tactics, and show how Harry Frankfurt's treatment of bulls.h.i.+t may be extended to a.n.a.lyze their underlying causes. Finally, I will deploy this new account of definition to a.s.sess whether Frankfurt's definition of bulls.h.i.+t is itself legitimate.
Semantic Negligence.
Frankfurt's princ.i.p.al contribution to the study of bulls.h.i.+t is the distinction he draws between the bulls.h.i.+tter and the liar. Whereas the liar represents as true something he believes to be false, the bulls.h.i.+tter represents something as true when he neither knows nor cares whether it is true or false (On Bulls.h.i.+t, p. 55). As Frankfurt amply demonstrates, this indifference is much of what we find most objectionable about bulls.h.i.+t. The liar has a vested interest in the inst.i.tution of truth-telling, albeit a parasitical one: he hopes that his falsehoods will be accepted as true. The bulls.h.i.+tter may also hope to be believed, but he himself is not much bothered whether what he says is true, hence his disregard for the truth is of a deeper and potentially more pernicious character.
Our outrage is conditioned on our being the objects of a deception. When we know what the bulls.h.i.+tter is up to we can be much more indulgent. As the comic novelist Terry Pratchett observes of two of his characters, ”they believed in bulls.h.i.+t and were the type to admire it when it was delivered with panache. There's a kind of big, outdoor sort of man who's got no patience at all with prevaricators and fibbers, but will applaud any man who can tell an outrageous whopper with a gleam in his eye.”116 The gleam in the eye is essential here: it is this complicity between bulls.h.i.+tter and audience which const.i.tutes the ”bull session” (On Bulls.h.i.+t, p. 34). Only when it escapes from the bull session and masquerades as regular a.s.sertion is bulls.h.i.+t deceptive; however, the insidious nature of this deception degrades the commitment to truth upon which public discourse depends.
One way of characterizing Frankfurt's innovation is as the introduction of a new category of linguistic misbehaviour, which we might call 'semantic negligence'. It is this concept which enables him to distinguish the bulls.h.i.+tter from the liar. In British and American common law, a civil claim for negligence arises when the defendant has a duty of care to the plaintiff which he neglects to exercise, thereby harming the plaintiff. Here the deceptive bulls.h.i.+tter has a duty to tell the truth; neglecting this duty harms his audience if they come to believe his false statements. His indifference as to the truth value of his statements, that is whether they are true or false, a meaning-related or semantic property, may thus be termed semantic negligence. Lying involves a higher degree of culpability, since the liar convinces his audience of falsehoods intentionally, not just foresee-ably. Frankfurt's insight is that conventional accounts of deception provide no middle ground between this higher level of culpability and complete innocence, and therefore no room for many familiar forms of deceit, such as bulls.h.i.+t. My contention is that semantic negligence may arise with respect to features of meaning other than truth value, and as such may be used to disentangle a wide variety of deceptive dialectical practices. Furthermore, semantic negligence is itself a matter of degree. The legal understanding of negligence acknowledges that the a.s.sociated culpability can range from inadvertence to willful blindness. We may generalize Frankfurt's position further by recognizing that some instances of semantic negligence are worse than others. In a.s.sessing the gravity of semantic negligence we should ask questions such as 'How foreseeable was it that deception would arise?' and 'How much at fault is the speaker in not foreseeing this?'.
A Caricature History of Semantics.
My argument will draw on themes from the philosophy of language, chiefly the pioneering German logician Gottlob Frege's disambiguation of the naive understanding of 'meaning'. In what may be considered the primal moment of a.n.a.lytic philosophy, Frege drew a threefold distinction between Sinn, Bedeutung, and Farbung, or sense, reference and tone. The sense of a term is what we understand if we understand what the word means. The reference, however, is the thing which the word picks out. Hence, as Frege explains, ”a proper name (word, sign, combination of signs, expression) expresses its sense, [but] stands for or designates its reference. By employing a sign we express its sense and designate its reference.”117 For example, the sense of 'the longest river in the world' is just what we understand by the words in this phrase. Clearly, having that understanding does not depend on knowing what the reference is (the River Nile, all four thousand miles of it), let alone on having seen the river in question. The last of Frege's three divisions, tone, is the least familiar: it may be defined as that aspect of the meaning of an expression that is irrelevant to the truth value of any sentence in which it may occur. In languages with large vocabularies, like English, it is often possible to restate a phrase using different words, but preserving both sense and reference. Continuing with the earlier example, consider 'Earth's lengthiest natural watercourse.' The change here is one of tone.
Frege's distinction between sense and reference was not entirely original. Many earlier philosophers, perhaps as early as Aristotle, drew similar distinctions between these aspects of the meaning of a word or expression. In this context the terminology 'intension' and 'extension' is often used instead of sense and reference respectively. With proper nouns, and definite descriptions, like the example in the last paragraph, the terminology coincides exactly. With other sorts of noun, ”concept nouns” as Frege calls them, sense and intension have the same meaning, but the reference is to the concept under which the members of the extension fall. The value of distinguis.h.i.+ng between the reference and extension of a concept noun is most apparent when talking about short-lived or rapidly propagating things. Expressions such as 'snowflake', 'mayfly', or 'web page' have constantly changing extensions, but more or less fixed references. By concentrating on reference rather than extension, we can disregard superficial changes of this kind. Frege's approach was innovative in several respects, most of which go beyond the scope of this article, and has had a profound influence on subsequent philosophy. A crucial insight of Frege's is that sense cannot be reduced to reference: different terms can have the same reference, but different senses. In his well-known example, 'the evening star' and 'the morning star' both refer to the same object, the planet Venus, although the senses of these phrases are clearly distinct. Indeed, it was a genuine scientific discovery in the ancient world when it was realized that these two familiar sights were one and the same. Without the distinction between sense and reference we would be unable to describe this discovery.
Frege's formalizing project required the suppression of tone: ”separating a thought from its trappings” as he puts it.118 Tone is the part of meaning from which we must abstract before logical a.n.a.lysis can begin. This abstraction is essential to the representation of inference in terms of logical form-that is, formal logic. For example, 'and' and 'but' are formalized in the same way, despite their difference in tone. (Consider 'He is a patriot and supports the government' versus 'He is a patriot but supports the government'.) This is entirely appropriate for the logic of mathematics, which was Frege's primary concern, since tone is seldom of significance in mathematical reasoning.
What is nuanced in the master can become dogmatic in the pupils. Many of Frege's successors sought to extend tone-free logical a.n.a.lysis to natural language. Amongst more popular writers this idealism could become extremism. Consider, for example, the psychologist Robert Thouless's claim that ”We must look forward to the day when the thinking about political and international affairs will be as unemotional and as scientific as that about the properties of numbers or the atomic weights of elements.” Whereas many logicians attempt to treat the terms of natural language as though they were tonally neutral, Thouless hopes to eliminate altogether ”Such words as 'progress', 'liberty', 'democratic', 'totalitarian', 'reactionary', 'liberal', 'freedom', . . . .”119 This Orwellian scenario exhibits the limitations of Frege's program. Although enormously successful in the formalization of technical language, and an inescapable foundation for any study of natural usage, it has little to say about tonal properties which play a substantial part in ordinary discourse. Thouless's procrustean fantasy of excising from our language what our logic cannot a.n.a.lyze is a desperate remedy diametrically opposed to the real solution: taking tone seriously. Further progress in the study of natural argumentation will require us to rehabilitate this repressed element. We shall see that this project is foreshadowed in the Yale ethicist Charles Stevenson's account of what he called persuasive definition.
Persuasive Definition.
As introduced by Stevenson, a persuasive definition (PD) of a term ”purport[s] . . . to alter the descriptive meaning of the term . . . but . . . not make any substantial change in the term's emotive meaning.” 120 Although he coined the terminology, Stevenson was not the first person to spot this phenomenon. Indeed he quotes the memorable attack on PD from Aldous Huxley's Eyeless in Gaza with which we began. Stevenson also introduced the converse stratagem, persuasive quasi-definition (PQD), in which the emotive meaning of a term is altered without changing the descriptive meaning. When PD is discussed in logic textbooks it is usually treated as though it were invariably fallacious.121 However, this betrays the hostility to tone we diagnosed in the last section. As Stevenson recognized, many cases of PD are much less objectionable: the difficulty is in drawing a principled distinction between harmless and malign instances of PD. Stevenson's account of PD is couched in unfamiliar terms: 'descriptive' and 'emotive' meaning. These reflect his understanding of the meaning of an expression as a dispositional property of that expression, representing its potential to cause a psychological response in its hearer or utterer (p. 54). Descriptive and emotive meanings are then distinguished as provoking cognitive or emotive psychological responses respectively. Few if any modern philosophers would find this account even remotely congenial. Detailed criticism would be out of place here, although we can observe that the account is closely related to the emotivist theory of ethics, sometimes called the Boo-Hurrah Theory, on which ethical terms, such as 'good', are merely expressions of an emotional att.i.tude. That Stevenson's ethical and semantic theories have fallen out of fas.h.i.+on may explain the comparative neglect of PD. However, we shall see that this concept is independent of the theoretical context in which Stevenson articulated it.
Specifically, Stevenson's definition of PD may be restated in Fregean terms as changing the sense or reference of a term, while representing the tone as unchanged. Replacing the slippery distinction between emotive and descriptive meaning with that between sense, reference and tone has several advantages, besides the rescue of PD from its theoretically suspect origins. Firstly, tone is not just emotive. It can also, for example, be jargon-laden (with any number of different jargons), bureaucratic, politically correct, affectionate, poetic, boorish, metropolitan, circ.u.mspect, dated, or many other things. Secondly, a threefold distinction provides for a more fine-grained a.n.a.lysis of dubious definition-like activity than the simple binary of PD and PQD. Table 1 distinguishes the sixteen different possibilities that can arise from changing () or keeping fixed (-) the sense, reference and tone of a term, as well as the term itself.
Table 1 Options for Change.
We can also begin to see how the concept of semantic negligence which we derived from Frankfurt's discussion of bulls.h.i.+t may be used to distinguish good from bad PD. The persuasive definer represents the tone of his redefined term as unchanged: this may or may not be negligent of him. He might be justified in believing the tone will not change, making his usage un.o.bjectionable. He might realize that the tone will be dramatically affected by the redefinition, in which case he is unlikely to expect his move to be accepted. Or he may be negligent as to whether the tone is faithfully preserved. This strategy is not overtly deceptive, since the tone could be unchanged. Rather, the speaker's lack of control over the tone, and indifference as to its eventual disposition, makes his utterance semantically negligent. In this respect it is a.n.a.logous to bulls.h.i.+t, not lying.
In our discussion of semantic negligence we suggested that different degrees of negligence are possible, depending on the risk of deception occurring and how much at fault the speaker is in not foreseeing that deception would result. Aphoristic definitions, such as ”By 'work' I mean action done for the divine”122 and perhaps Huxley's ”Conservatism is true socialism,” are usually so surprising or paradoxical that they are unlikely to be truly deceptive. Many other definitions are inseparable from the theories which produce them: as Stevenson observes, ”To chose a definition is to plead a cause” (p. 210). There's no reason to accept such definitions unless one is convinced by the arguments with which the theory is defended. This situation is common in scientific contexts, where it is typically unproblematic: good practice requires the definer to make the theoretical indebtedness of his definitions explicit. Definers in natural language are likely to be less scrupulous, hence their interlocutors may be misled into endorsing the conclusions of arguments they would not judge sound, were they to be given a fair opportunity to appraise them. The resulting deception may be deliberate, but is just as likely to be inadvertent: it is easy to confuse oneself as well as others with this sort of definition. Such behaviour is less culpable than outright deceit, just as bulls.h.i.+t is less blameworthy than lying, but as with bulls.h.i.+t, it is also peculiarly pernicious since it degrades the standards of discourse.