The Myth of Factive Verbs

Abstract

At least since the days of 'ordinary language philosophy', epistemologists (some more than others) have been interested in knowledge attributions—the meaning or meanings of 'knows', the use or uses of sentences of the form 'S knows p', and so on. And there has more recently been renewed interest in how 'knows' is used.1 Theorizing about knowledge on the basis of how 'knows' is used is, at least in principle, quite different from the epistemological tradition exemplified by the 'post-Gettier' literature on the analysis of knowledge. In this tradition, necessary and sufficient conditions for knowledge are proposed, and tested against potential counterexamples. Imaginary situations are described, and intuitions (which are presumed correct) are elicited concerning whether or not a character in the story knows a certain proposition.2 Contrast this with the linguistic method: imaginary situations are described, and intuitions (which are presumed correct) are elicited concerning whether or not a character in the story said something acceptable.3 On the basis of this, a theory of the meaning of 'knows' is developed. But note that the result of this linguistic method might be a conclusion of the same kind as the result of the post-Gettier method, for one theory of the meaning of a word is simply a definition of it, in the form of necessary and sufficient conditions for its correct (i.e. true) application. So we can, at least in principle, distinguish between two competing methods of theorizing in epistemology—one based on intuitions about knowledge, and the other based on intuitions about language.4 Alternatively, a range of 'value based' methods have been suggested. Edward Craig proposes a genealogical approach, on which an inquiry into the sources of the concept of knowledge is the starting point for epistemology.5 Jonathan Kvanvig argues that no theory of knowledge can ignore the question of the value of knowledge.6 And Sally Haslanger has suggested an "ameliorative" approach to conceptual analysis, which can be applied to the concept of knowledge, on which the theorist identifies "what legitimate purposes we might have (if any) in categorizing [e.g. beliefs in terms of knowledge] and to develop concepts that would help us achieve these ends."7,8, 7,8 Nothing speaks against using these various methods in tandem. Kvanvig, for example, favors a combination of the 'post-Gettier' method with a 'value based' method, where "the presumption in favor of the value of knowledge is strong enough that it gives reason to abandon even a counterexample-free account of the nature of knowledge if that account leaves no way open for defending the value of knowledge."9 But the existence of a plurality of approaches to the question 'What is knowledge?' should give us pause—have we any reason to think that there is such a thing as the analysis of knowledge? What would make it the analysis of knowledge, as opposed to some competing definition? Do we really expect that a single definition of knowledge might be (even relatively) counterexample free, (even relatively) suitable as the ordinary meaning of 'knows', and (even relatively) successful vis-à-vis the 'value based' criteria proposed by Craig, Kvanvig, and Haslanger? At this stage one begins to wonder what the point of giving a definition of knowledge would be. In any case, it is within this context that I propose to criticize the following thesis: (Factivity) Certain two-place predicates, including 'knows', 'learns', 'remembers', and 'realizes', which denote relations between persons and propositions, are factive in this sense: an utterance of 'S knows p' is true only if p, an utterance of 'S learned p' is true only if p, and so on. Standard analyses of knowledge include a truth condition (S knows p only if p is true) and this element of the analysis is often justified (nowadays, anyway) by appeal to Factivity—a decidedly linguistic thesis. If I'm right in my criticism of Factivity, then epistemologists will have to look elsewhere for support for the truth condition. But more importantly, if I'm right, then epistemologists may have reason to stop looking at linguistic phenomena altogether—at least if they want to keep working on anything like the standard analysis of knowledge. The concept of knowledge that epistemologists have been interested in since the Meno is a factive concept (in the sense that nothing false can be known). But, if I'm right, the concept of knowledge that serves as the meaning of 'knows' in ordinary talk isn't. This is strong prima facie evidence that traditional epistemology shouldn't be especially interested in the concept of knowledge that serves as the meaning of 'knows' in ordinary talk. The same could be said about the thesis of 'pragmatic encroachment', according to which (roughly) the importance of a belief (for the believer) may determine whether or not it is (or counts as) a case of knowledge.10 If this is supported by consideration of ordinary language, so much the worse for supporting epistemological claims by appeal to ordinary language. For the epistemologist is interested in an epistemic concept of knowledge, if she is interested in a concept of knowledge at all.11 What I'm claiming is that epistemologists have every right to insist that knowledge (as they understand it) is factive—but the price to pay for this (which many will be happy to incur) is to give up the linguistic method described above. I'm suggesting, in other words, a divorce for the linguistic theory of knowledge attributions and traditional epistemology.12 I've called factive verbs a myth, and I mean something by that. Compare a myth (which, though false, captures something true) with a dogma (which is just plain wrong). Factivity captures something true, and it is a large part of my task here to explain the true content of this 'myth'—the linguistic phenomena that the 'myth' is useful in capturing. In §2 I describe some intuitively acceptable utterances that, if (possibly) true, are counterexamples to Factivity. In §3 I present and criticize two arguments for Factivity, and one for the related claim that nothing false can be known. The appeal of these arguments partly explains the appeal of Factivity. In §4 I focus on 'knows' in particular, and defend the claim that a non-factive concept of knowledge is the best candidate to serve as the meaning of 'knows' in ordinary talk. Explananda for a theory of knowledge attributions are proposed. In §4.1 I present my preferred non-factive theory of knowledge attributions. Crucial here will be a Gricean explanation of fact that one who utters 'S knows p' typically implies that p is true, which provides further explanation of the appeal of Factivity, and additionally accounts for its status as a 'myth'. §4.2 discusses a traditional theory of knowledge attributions, and §4.3 discusses an alternative both to my theory and the traditional theory. The utterance of the following sentences does not strike ordinary people as deviant, improper, unacceptable, necessarily false, etc.: Everyone knew that stress caused ulcers, before two Australian doctors in the early 80s proved that ulcers are actually caused by bacterial infection.13 He figures anything big enough to sink the ship they're going to see in time to turn. But the ship's too big, with too small a rudder … it can't corner worth shit. Everything he knows is wrong.14 In school we learned that World War I was a war to "make the world safe for democracy," when it was really a war to make the world safe for the Western imperial powers.15 I had trouble breathing, sharp pains in my side, several broken ribs and a partially collapsed lung, and I was in the middle of nowhere without any real rescue assets—it was then that I realized I was going to die out there.16 Talk of false memories is, I assume, familiar enough. Since these uses of 'knows', 'learns', 'remembers', and 'realizes' are unexceptional, and do not strike ordinary people as deviant, I claim that we need some theory of these words on which (1)–(4) are acceptable. I contend that the best working hypothesis is that utterances of these sentences could be true—so we have a prima facie threat to Factivity. Ultimately the matter should be decided by consideration of competing theories of 'knows', 'learns', 'remembers', and 'realizes'; §4 examines 'knows', where I'll argue that the best theory of knowledge attributions says that (1) and (2) could be true. But some objections to this idea can be dealt with immediately. Objection: The uses of factive verbs you appeal to are cases of loose talk. Reply 1: On some views of loose talk, 'loose' utterances are literally true, but not 'strictly' true, not as accurate as they could be, etc. For example, Peter van Inwagen argues that an ordinary utterance of 'There are two very valuable chairs in the next room' could be true (though it is 'loose talk'), even though 'strictly speaking' there are no chairs.17 If this is the right model of 'loose talk', then the objection fails to eliminate the prima facie threat to Factivity. Reply 2: A paradigm case of loose talk is an utterance of 'It's two-thirty' when it is, exactly, 2:29. The 'loose' utterance is 'close' to be being true, it 'approximates' the truth, etc. The uses of 'knows', 'learns', 'remembers', and 'realizes' in the examples above are nothing like that.18 (It is not as if, for example, Zinn thinks that World War I was close to being, but not precisely or exactly, a war to make the world safe for democracy.) Objection: People's resistance to the truth condition on knowledge is the result of the popularity of relativism, as those "infused with relativistic ideas" will maintain that it was once known that the earth was flat.19 Reply: According to at least one common (and, it seems to me, mistaken) relativism, it was not only known that the earth was flat, it was also true that the earth was flat. The person influenced by this relativism, however, has no quarrel with the claim that 'knows' is factive. It seems to me, furthermore, that one who claims that falsehoods can be known is in a unique position to reject relativism while accommodating some liberal and pluralistic intuitions. Suppose it's right, as we learn from anthropologists and sociologists, that just as there is Western scientific knowledge, there is also tribal knowledge, ancient Greek knowledge (perhaps this includes the knowledge that the earth is flat), and so on; suppose we agree that there are a plenum of incompatible 'epistemologies', each equally legitimate, each just as much a body of knowledge as the others. If nothing false can be known, it follows from this that there are a plurality of incompatible truths—and this is an unlovely relativism that violates the law of non-contradiction. If, on the other hand, falsehoods can be known, then relativism about knowledge gives us no motivation to slip towards relativism about truth.20 Objection: The uses of factive verbs you appeal to are cases in which said verbs are used with a different sense than they are often used. 'Knows' and other factive verbs are polysemous, in some contexts they take a factive sense; in others they take a non-factive sense. Reply: Following Grice, I take the positing of polysemy to be a vice, ceteris paribus, for a linguistic theory.21 So while polysemy should not be ruled out, we should first seek a theory on which 'knows' is univocal. After considering three such theories in §4 I'll return to the question of whether 'knows' is polysemous in §4.3. An appealing thought, for many defenders of Factivity, is the idea that (1)–(4) are false, but pragmatically appropriate because they are instances of metaphorical 'projection' into the perspective of the 'knower', 'learner', or 'realizer'. I reserve discussion of this proposal for §4.2. We can distinguish two claims. Factivity, on the one hand, is a claim about the truth conditions of certain utterances. The claim that knowledge is factive (i.e. that nothing false can be known), on the other hand, is a claim about the necessary conditions on knowing. Above I proposed a divorce for the theory of knowledge attributions and epistemology. The terms of this divorce would allow us to say that knowledge (as traditionally understood) is factive, but at the same time say that Factivity is false (and hence that knowledge, understood as that which 'knows' refers to, isn't factive). In this section I present three unsound arguments. The first two are arguments for Factivity, as stated; the first is based on syntax, and the second is based on the appearance of contradiction in an utterance of 'I know p, but not-p'. The third is an argument for the claim that nothing false can be known. In criticizing these arguments, I shall partly explain the appeal of Factivity. For these arguments are compelling at first glance, though unsound. That Factivity has appeal is uncontroversial. It has almost universal appeal for philosophers, and many non-philosophers find it intuitively compelling as well. This is something that the critic of Factivity must explain. But it should not be ignored that, this intuition notwithstanding, most people do not find the claim that nothing false can be known to be obvious. Now the fact that it isn't obvious is not a significant mark against the claim; some truths of this sort aren't obvious to most people. But as far as common sense goes, the best the defender of Factivity can say is that common sense doesn't have an opinion about Factivity. At worst, common sense is downright hostile to it. According to Zeno Vendler, philosophers, and not linguists, first claimed that that 'knows' and its kin are factive.22 It's often said (e.g. by me, above) that the origin of the idea that truth is necessary for knowledge is the Meno. The contrast set up there is between knowledge and true belief; the idea that falls out of introducing the concept of knowledge this way is that knowledge is to be defined as true belief plus something else. One thing to note, though, is that it does not follow from the fact that knowledge is more valuable than true belief (insisted upon by Socrates) that truth (or belief) is necessary for knowledge, that truth (or belief) is a 'component' of knowledge. Gold is more valuable than lead; it doesn't follow from this that lead is a component of gold. More important is the fact that the contrast set up in the Meno is not natural, from an ordinary language point of view. The natural contrast is between knowledge and ignorance. And there seem to be two kinds of cases in which knowledge is denied, and thus ignorance implied: cases in which the truth of some proposition is presupposed ("Anne doesn't know that the party was postponed," for example, when the party has been postponed, and Anne doesn't believe that it has been) and cases in which the truth of some proposition is in question, though someone believes it ("John doesn't know that this is the right road, because he didn't check the map," when it's unclear whether this is the right road).23 It is very rare, however, to find people using 'knows' the way that Plato does in the Meno– to deny knowledge to a true believer. This, I submit, is rare outside of the context of philosophy, though the idea that knowledge is true belief plus something is natural in the context of philosophical skepticism, where worries about the epistemic status of our (admittedly correct) beliefs occupy us.24 In any case, let us consider the three aforementioned arguments: The Argument from Syntax: The factive verbs are members of a class of expressions with certain syntactic features in common, which we'll call the syntactically factive expressions. First, syntactically factive expressions can always be followed by 'the fact that …', while others cannot. Compare: I remember the fact that I opened the door. * I believe the fact that I opened the door. Second, syntactically factive expressions are always able to be followed by gerunds, whereas others are not. Compare: I remember having opened the door. * I believe having opened the door. Third, syntactically factive expressions, by contrast with others, cannot be followed by infinitives. Compare: * I realize Martin to have opened the door. I believe Martin to have opened the door. Given this definition, it is maintained that the class of syntactically factive expressions are factive, in the sense defined at the outset (call this being semantically factive).25 Reply 1: That a class of expressions has certain syntactic features in common is not a good reason to conclude that they have any particular semantic feature in common. (Perhaps it is a good reason to conclude that they will have some semantic features in common.) So even if the verbs on our list are syntactically akin to one this does not a reason to conclude that they are semantically factive. Reply 2: The verbs on our list aren't syntactically factive. you may have we had to be in verbs from our list to these because the verbs on our list do not three * I know the fact that I opened the door. * I learned the fact that I opened the * I realize the fact that I opened the door. * I know having opened the door. * I learned having opened the door. * I realized having opened the door. I know Martin to have opened the door. I remember Martin to have opened the door. In other words, as and note, 'knows' and others are syntactically though, that they are semantically factive in of But since we are looking for to suppose that these expressions are factive, this is in with my that gives us no reason to say that 'knows' and others are factive. The Argument from who says 'I know p, but 'knows' is factive. for the other factive know p, but is not but an utterance of it is to know that p is to believe that p, and 'I believe p, but is It is to take a the utterance of which would be for a An utterance of 'I know p, but is always improper, but the is not a that, just as 'I p, but is not is 'I knew p, but as in the case of the It may be that knows p, but also to be a In §4 I some concerning the of our use of and there I maintain that one who utters 'S knows p' typically implies that p is true. This may explains 'S knows p, but can But the case above that sentences of this form are We should also consider the fact that find that 'I remember p, but and 'I learned p, but The Argument from knows p only if belief that p be false belief is one that be So knowledge knows p only if belief that p is Reply: is I reject it for the familiar and as it as part of my criticism of Factivity is not out of even though it has if is right the truth condition follows from then the appeal of the truth condition can be by the appeal of that might resistance from those who take it as the of the truth that some might be to and from the truth condition to At we can conclude that the intuition that knowledge truth is up with the intuition that knowledge must be be of any claim of for one intuition the I only that the that knowledge truth may be appealing (even though it's because it is a of the appealing thesis of Suppose we a concept of knowledge that would be the best candidate to serve as the meaning of 'knows' in ordinary talk, in other words, the most suitable concept to use in giving a and for our ordinary knowledge What sort of concept would we In this section I argue that a non-factive concept is to its when it to giving a and for our ordinary knowledge attributions. The method here that used in theories will be described, and a on the basis of and The will be intuitively acceptable knowledge attributions, with This is especially important for the of a non-factive For it is that, at least one who utters 'S knows p' implies that p is true. In Factivity, I the claim that this is by an from 'S knows p' to the truth of So I must explain the some other I'll consider three three agree that 'knows' is and that a can be for The theories are in terms of semantic

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 187 citations exist upstream.