31
First-Person Authority
Donald Davidson's discussion of the phenomenon of first-person authority has two main strands. He aims to explain the phenomenon: in particular, to explain the asymmetry between our knowledge of our own minds and our knowledge of other minds (see Davidson 1984). And he aims to demonstrate that first-person authority is consistent with other elements of his philosophy with which it might seem to be incompatible: most notably, his externalism about content (see Davidson 1987).
Our beliefs about our own present attitudes, Davidson thinks, are distinguished from our beliefs about others' attitudes in two ways. In the first place:
When a speaker avers that he has a belief, hope, desire or intention, there is a presumption that he is not mistaken, a presumption that does not attach to his ascriptions of similar mental states to others
(Davidson 1984: 3).
And second, “the self-attributer does not normally base his claims on evidence or observation”; attributions of attitudes to other people, by contrast, rest on the evidence of what they do and say.
That characterization of first-person authority focuses on a linguistic phenomenon: the asymmetry between statements about one's own attitudes and statements about other people's. And it is that phenomenon – first-person authority in speech – that Davidson's explanation directly addresses. But his “main aim,” he says, is “to throw light on an epistemic problem, not a linguistic one” (Davidson 1993a: 211). His assumption is that “if first person authority in speech can be explained, we will have done much, if not all, of what needs to be done to characterize and account for the epistemological facts” (Davidson 1984: 3).
Davidson explicitly restricts the scope of his explanation of first-person authority to the case of a subject's knowledge of her own current propositional attitudes: states like belief, desire, and intention. He does not attempt to explain our knowledge of our past or future propositional attitudes. Nor does his account extend to attitudes that are directed at nonpropositional entities: admiring Cézanne, fearing death, and so forth. Nor does he try to explain our knowledge of our own sensations. “What holds for the propositional attitudes,” he writes, “ought, it seems, to be relevant to sensations and the rest, but I do not explore the connections here” (Davidson 1984: 3). And even for the case of propositional attitudes, Davidson further restricts his account to deal with just one kind of case: our knowledge of our own present beliefs (see Davidson 1984: 4). So, even if the explanation he gives is completely successful for the case he discusses, it will offer only a partial explanation of first-person authority as a whole. (We will return to this point later.)
Before developing his own explanation of first-person authority, Davidson considers and rejects two kinds of explanation that are common in the literature. There is the view that what explains a person's superior knowledge of her own beliefs is her access to a special sort of evidence: in particular, the evidence afforded by a faculty of introspection. And there is the view, which Davidson associates with Strawson and Wittgenstein, that it is just “an essential feature of our use of certain mental predicates that we apply them to others on the basis of behavioral evidence but to ourselves without benefit of such aid” (Davidson 1987: 16). Despite their differences, Davidson argues, these two traditional approaches face the same basic objections: they do not explain the asymmetry between self-knowledge and knowledge of others, and they invite skepticism about whether the beliefs we ascribe to ourselves and the beliefs we ascribe to others are states of the same kind. We can take these points in turn.
Davidson agrees with the Strawson/Wittgenstein claim that it is a feature of our mental predicates that they can be ascribed to others on the basis of observed behaviour and to ourselves without such a basis. But to say this, he thinks, is simply to describe the asymmetry between first-person and third-person ascriptions of mental predicates; it does not explain it (see Davidson 1984: 8). And the introspectionist view, he thinks, does no better. It tries to explain the first-person/third-person asymmetry by appeal to the fact that we have direct introspective awareness of our own beliefs but are only indirectly aware of other people's beliefs. But to say that I am directly aware of my own beliefs, Davidson thinks, is to say no more than that I know them without needing to observe my own behaviour. As before, that is true: but it does not explain the phenomenon of first-person authority; it simply redescribes it.1
Why does Davidson think that the traditional explanations of first-person authority invite skepticism? His argument is this. Both views turn on the systematic difference between the basis on which the predicate “believes that” is ascribed to other people and the basis on which it is self-ascribed. That prompts the skeptical question, why we should think that the predicate means the same thing when applied to oneself and to others. And neither view says anything to address that question. So, Davidson thinks, they both leave it open that the property of belief one ascribes to oneself without evidence is different from the property of belief one ascribes to others on the basis of their behaviour. That is why both views “invite skepticism” (see Davidson 1984: 8,1987: 16–17).
Davidson's conclusion, then, is that an adequate account of first-person authority must improve on the traditional accounts in two ways: it must explain, and not simply describe, the asymmetry between knowledge of one's own beliefs and knowledge of other people's, and it must do so in a way that does not invite the kind of skepticism just discussed.
Suppose I utter the sentence “Wagner died happy,” thereby expressing a belief. If someone knows what I believe, Davidson thinks – what belief I express in uttering the sentence – then she knows two things: she knows that I hold the sentence true; and she knows what I mean by the sentence.2 Correspondingly, he thinks, we can regard the asymmetry between my knowledge of what I believe and your knowledge of what I believe as involving two asymmetries: there is an asymmetry between my knowledge that I hold the sentence “Wagner died happy” to be a true sentence and your knowledge that I hold it true; and there is an asymmetry between my knowledge of what the sentence means on my lips and your knowledge of what it means on my lips.3 It is natural to think that a full explanation of the asymmetry between knowledge of one's own beliefs and knowledge of others' beliefs would explain both these asymmetries.
Now Davidson's account does not address the first asymmetry at all. There is, he says, an asymmetry in our knowledge that I hold the sentence “Wagner died happy” to be true: your knowledge depends on observation or inference; mine does not. But he does not attempt to explain this asymmetry; nor does he appeal to it in his explanation of first-person authority (see Davidson 1984: 12).4 So his explanation of the asymmetry between my knowledge of my belief and your knowledge of my belief rests entirely on his account of the second asymmetry: the asymmetry between my knowledge of what my sentence “Wagner died happy” means and your knowledge of what it means. What explains this asymmetry?
Davidson's idea, in broadest outline, is that “there is a presumption – an unavoidable presumption built into the nature of interpretation – that the speaker usually knows what he means” (Davidson 1984: 14); there is no corresponding presumption that an interpreter usually knows what a speaker means. But what explains that presumption? What explains the fact that a speaker normally knows what her words mean without observation or inference? And what explains the asymmetry between the speaker's knowledge of the meanings of her words and the hearer's knowledge of their meanings? Davidson's answer to those questions involves two key ideas.
The first idea is that a speaker can state the meanings of her words in a specially error-free way that is not available to an interpreter:
The speaker, after bending whatever knowledge and craft he can to the task of saying what his words mean, cannot improve on the following sort of statement: “My utterance of ‘Wagner died happy’ is true if and only if Wagner died happy”. An interpreter has no reason to assume that this will be his best way of stating the truth conditions of the speaker's utterance.
(Davidson 1984: 13. For other occurrences of the same point, see Davidson 1989: 66, 1991: 197–198)
We can call this the disquotational point.5 Davidson's second key idea is that a speaker must in general be right about the meanings of her words because it is her own use of those words that determines what they mean; there is no similar reason why a hearer must in general be right about the meanings of a speaker's words:
The explanation [of first person authority] comes with the realization that what a person's words mean depends in the most basic cases on the kinds of objects and events that have caused the person to hold the words to be applicable; similarly for what the person's thoughts are about. An interpreter of another's words and thoughts must depend on scattered information, fortunate training, and imaginative surmise, in coming to understand the other. The agent herself is not in a position to wonder whether she is generally using her own words to apply to the right objects and events, since whatever she regularly does apply them to gives her words the meaning they have and her thoughts the contents they have.
(Davidson 1987: 37. For essentially the same idea, see Davidson 1984: 13–14)
It is clear that these two ideas are central to Davidson's explanation of our knowledge of the meanings of our own words. It is less clear how they are related to each other, or to Davidson's overarching claim that the presumption that a speaker usually knows what her words mean is “essential to” or “built into” the nature of interpretation (Davidson 1984: 12, 14).
Many commentators treat the disquotational point either as carrying the whole weight of Davidson's explanation of a speaker's knowledge of the meanings of her words, or as being one of a number of distinct explanations he offers.6 Few critics have put much stress on Davidson's idea that the meanings of a speaker's words are determined by her own use of those words.7 And some treat his claim about the presumption that is essential to interpretation as a further, and separate, argument. My own view is that the different strands in Davidson's discussion make up a single explanation of speakers' knowledge of the meanings of their words. The main substance of the explanation comes from the idea that the meanings of a speaker's words are determined by her own use of those words. That is what underpins the presumption that a speaker usually knows what her words mean. The disquotational point has an important role in Davidson's thinking; but it is not intended as a self-standing explanation of speakers' knowledge of meaning.
Before proceeding, it is worth addressing a potential objection to Davidson's strategy. In Davidson's view, if someone knows what belief I express when I say “Wagner died happy,” then she knows two things; she knows that I hold the sentence true, and she knows what I mean by the sentence. For analytical purposes, he thinks, it is helpful to see this structure in the phenomena. But Davidson's view does not imply that someone who knows that I believe that Wagner died happy reaches that knowledge by inferring it from two prior and independent pieces of knowledge: the knowledge that I hold the sentence “Wagner died happy” to be true, and the knowledge of what I mean by it. And he can perfectly well accept that we rarely if ever know our own beliefs via that sort of inference.8 (Compare. If someone knows that Sam is an actress, then she knows two facts: that Sam is in acting; and that Sam is female. It does not follow that she must reach the knowledge that Sam is an actress by inferring it from prior and independent knowledge of those two facts about Sam. If she is told that Sam is an actress, for instance, then she can learn both facts simultaneously.) So it would be a mistake to think that Davidson's strategy for explaining first-person authority builds in an assumption that our knowledge of our own beliefs is essentially inferential.9
Suppose I want to say what the sentence “The plane will be landing momentarily” means in my language. Davidson says that I “cannot improve on” (Davidson 1984: 13) a statement of meaning like this:
Or, more straightforwardly, this:
When I state the meaning of my sentence in that way, I am using a sentence of my language to state the meaning of that same sentence. So there is no chance that the sentence I use on the right-hand side of the words “means that” will mean something different from the sentence I mention on the left-hand side; it is the same sentence in the same language. When I state the meaning of one of my own sentences in this way, therefore, I cannot misrepresent its meaning. By contrast, Davidson points out, there is no way for me to state the meaning of someone else's words that is similarly immune from error. Suppose I say this:
What I say will only be true if the sentence “The plane will be landing momentarily” has the same meaning in your language as it has in mine. And there is no guarantee that it does. In my language, for example, the word “momentarily” means for a moment. But suppose the same word in your language means in a moment. In that case, my utterance of (3) will be false. The lesson is a general one: our judgments about the meanings of other people's words are vulnerable to a kind of error that cannot arise in disquotational judgments about the meanings of our own words. That is Davidson's point.
Some of Davidson's critics have interpreted him as holding that it is sufficient for knowing the meaning of a sentence that one can produce a sentence like (1) or (2). Against that view, they object that one can produce a sentence like (1) or (2), and know that it is a true sentence, without understanding the sentence embedded on its left-hand side.11 Suppose I do not know the meaning of the English sentence, “I dehort a postprandial perambulation.”12 I can still know that (4) is a true sentence of English:
So, it is said, Davidson's point about disquotation does not explain a speaker's knowledge of the meanings of her words.
The objection misunderstands Davidson's point. We need to distinguish two things: knowing that (4) is a true sentence; and knowing what it states – that is, knowing that the English sentence “I dehort a postprandial perambulation” means that I dehort a postprandial perambulation. Davidson would agree that someone can know that (4) is a true English sentence without understanding the embedded sentence “I dehort a postprandial perambulation” and, therefore, without knowing what (4) states. But he never suggests that the ability to produce a true disquotational sentence like (4), or the knowledge that it is a true sentence, is by itself sufficient for understanding the sentence embedded on its left hand side. His point about disquotation is just this. Suppose I do understand the sentence “I dehort a postprandial perambulation”; I know what it means. Then I can use that sentence to state its own meaning in a way that is proof against the kind of error to which I am vulnerable when I use my words to state the meanings of someone else's words.
Davidson's point is plainly correct. But the critic may now raise a new concern. As I have just explained it, the disquotational point simply takes it for granted that the speaker understands the sentence she uses to state its own meaning; it takes it for granted that she knows the meaning of the sentence. It does not explain that knowledge. Nor does it explain what ensures that a speaker generally knows the meanings of her own words. What, then, does explain these things? For Davidson's answer to those questions, we must turn to his second key idea: that the meanings of a speaker's words are determined by her own use of those words.
Davidson writes: “a person cannot generally misuse his own words, because it is that use which gives his words their meaning” (Davidson 1993b: 250). Or again: an agent “is not in a position to wonder whether she is generally using her own words to apply to the right objects and events, since whatever she regularly does apply them to gives her words the meaning they have” (Davidson 1987: 37). That claim is the key to his explanation of a speaker's knowledge of the meanings of her own words.
How, exactly, does a speaker's use of her words give those words their meaning? Davidson writes:
In the simplest and most basic cases, words and sentences derive their meanings from the objects and circumstances in whose presence they were learned. A sentence which one has been conditioned by the learning process to be caused to hold true by the presence of fires will (usually) be true when there is a fire present; a word one has been conditioned to hold applicable by the presence of snakes will refer to snakes. Of course very many words and sentences are not learned in this way; but it is those that are that anchor language to the world.
(Davidson 1988a: 44–45)
In these “simplest and most basic cases,” Davidson thinks, my words get their meanings from the objects and events that “normally” (Davidson 1991: 197), “regularly,” or “generally” (Davidson 1987: 37) cause me to hold those words to be applicable.13 My word “fire,” for instance, refers to fires because it is normally fires, and only fires, that cause me to hold the word “fire” to be applicable. If it were something other than fires that normally caused me to hold the word “fire” to be applicable, then the word would refer to that other kind of thing instead.14 Given that view of what determines the meanings of words, a speaker cannot generally misapply her own words; for (in the simplest and most basic cases, at least) what a speaker's words apply to is determined by whatever she generally does apply them to. (An analogy: a person cannot generally act out of character; for what her character is is determined by the way she generally acts.)
What exactly does this establish? The immediate point is that “nothing could count as someone regularly misapplying her own words” (Davidson 1987: 38). But what Davidson wanted to show seemed more ambitious than that. For he aimed to establish that a speaker “usually knows what he means” (Davidson 1984: 14). How do we get from the first of these things to the second: from the claim that a speaker cannot generally misapply her own words to the claim that she generally knows what her words mean? (In terms of our analogy: How do we get from the claim that a person cannot generally act out of character to the claim that a person generally knows what her character is?)
There is a broadly Wittgensteinian tradition in which knowing the meaning of a word just consists in the ability to use it correctly. If we hold this deflationary view, there will be no gap between the claim that a speaker cannot generally misapply her own words and the claim that she generally knows what her words mean. For, on the deflationary view, all there is to knowing what a word means is being able to use it correctly. A number of commentators have suggested that Davidson's explanation of a speaker's knowledge of the meanings of her words depends on this deflationary view.15 And Davidson says things that certainly encourage that interpretation. For example:
The best [a] speaker can do is to be interpretable, that is, to use a finite supply of distinguishable sounds applied consistently to objects and situations he believes are apparent to his hearer. Obviously the speaker may fail in this project from time to time; in that case we can say if we please that he does not know what his words mean.
(Davidson 1984: 13)
“We can say that” a speaker who fails to be interpretable does not know what his words mean. That does not entail that a speaker who is interpretable – who succeeds in applying his words consistently to objects and situations in his environment – does know what his words mean. Still less does it entail that all there is to knowing what one's words mean is applying them consistently. But Davidson's comment easily suggests that deflationary view. Or again:
What I claimed was that in any normal utterance there is a presumption that the speaker knows what he means, while there is no such presumption in our understanding of others. (The word “knows” here is not essential in making the point; it would do as well to say that the speaker is not misusing his own words, or that he means, in a literal sense, what he says.)
(Davidson 1993a: 211–112; see also Davidson 1993b: 249)
That passage suggests that the sense in which a competent speaker must know the meanings of her words is relatively thin. And, as before, that might encourage the deflationary view.
Nonetheless, it would be wrong to conclude that Davidson takes a speaker's knowledge of the meanings of her words to be simply equivalent to the capacity to use them correctly. A competent and reflective speaker, he insists, can say what her words mean. And when she does so, she “cannot improve on” statements like this: “My utterance of ‘Wagner died happy’ means that Wagner died happy.” In making a statement like that, the speaker is not simply exercising a first-order capacity to use her words appropriately; she is making a higher-order statement about her words. By the same token, she is not simply exhibiting a bit of practical knowledge: knowledge of how to use her words. She is expressing a piece of genuinely propositional knowledge: the knowledge that her words “Wagner died happy” mean that Wagner died happy. So for Davidson, I think, our knowledge of the meanings of our words is not merely a matter of being able to use them correctly; he thinks we have propositional knowledge of meaning.
However, even if knowing the meanings of one's words does not reduce to the ability generally to apply them correctly, the step from that ability to propositional knowledge of the meanings of one's words is relatively small. Suppose a speaker does use her words correctly. In order to have propositional knowledge of the meanings of those words, she needs only two things: first, the reflective abilities and basic linguistic concepts that are necessary for making higher-order judgments about the meanings of her words (“The word ‘red’ applies to red things,” “The word ‘red’ means that color”); and second, the ability to move from a first-order application of a word to a simple higher-order judgment about it. Now on Davidson's picture of language and thought, it is a condition on someone's being a competent language user that she is sufficiently reflective to have higher-order thoughts about her own beliefs; for to count as having beliefs at all, he thinks, she must grasp the distinction between how things are and how she believes them to be.16 And it is a condition on someone's being a competent language user that she is an interpreter of others' language, and that requires her to grasp at least some linguistic concepts.17 It follows that for Davidson, if someone is a genuine language user, she must be capable of making higher-order judgments about the meanings of her words and sentences. And if someone is capable of making such judgments at all it is, as Davidson insists, a straightforward matter for her to make judgments about her own meanings that are by and large true.18
To summarize: We can see Davidson's explanation of the fact that speakers generally know the meanings of their own words as having two parts. The first is the argument that a speaker cannot generally misapply her own words, because the meanings of those words are determined by the way she generally applies them. The second is the argument that the requirements for having propositional knowledge of the meanings of one's words are so basic and straightforward that they will be met by anyone who does generally apply her own words correctly.
On Davidson's view, as we have seen, what my words mean is a matter of what I am disposed to apply them to. So knowing what my words mean involves knowing what I am disposed to apply them to. Now it is not generally true that if I have a disposition, I can know that I have it without observation or inference. Selfishness, friendliness, and having perfect pitch, for instance, are dispositional properties. But I have no special, non-observational way of knowing whether I have any of those properties. To know whether I am selfish, or friendly, or have perfect pitch, I must observe my own reactions in appropriate situations. With linguistic dispositions, however, things are different. The meanings of my words are determined by my linguistic dispositions. But to know what my words mean, I do not generally need to observe my linguistic reactions: I do not normally need to find out what my words mean by seeing what prompts me to apply them in actual situations, or by contemplating imagined situations and considering what would prompt me to apply them.19 On the contrary, I know in advance what to apply my words to because I already know what they mean. That is a crucial feature of our relation to our own words. It is the feature that Wittgenstein identifies when he talks about grasping the meaning of a word in a flash: on the one hand, meaning something by a word is (or is related to) a disposition to use it in a particular way; on the other hand, when I come to understand a word, it does not take experience to know what I mean by it – I know it in a flash.20 But what accounts for this ability? How is it that we can know these dispositional facts about ourselves without observation?
There is nothing in Davidson's account of first-person authority to answer that question. As we have seen, Davidson's explanation appeals to two asymmetries between our knowledge of the meanings of our own words and our knowledge of the meanings of others' words: there is the fact that the meanings of my words are determined by my own use of those words, whereas they are not determined by other people's use of the same words; and there is the fact that, when I use one of my words to state its own meaning, I am immune to a kind of error to which I am vulnerable when I use a word of mine to state the meaning of someone else's word. But neither of those points explains the distinctive character of one's knowledge of the meanings of one's own words. For both points (or analogues of them) would apply equally well to dispositions that we had no distinctive, nonobservational way of knowing we possessed. Imagine someone – Jane – who has a series of blind dispositions to produce a range of sounds in regular ways in response to objects in her environment. Because these sounds are systematically correlated with particular kinds of environmental causes, they carry information about those causes: they have a “meaning,” in something like the sense of Grice's “natural meaning.”21 How can Jane know what her sounds “mean”? It is clear that her relation to the “meanings” of these sounds is fundamentally different from our relation to the meanings of our words. In the normal case, we know what our words mean without waiting to see how we use them. That is not true for Jane: in order to know what her sounds “mean,” she must observe her reactions and correlate each sound with its regular cause. Nonetheless, the two points that Davidson stresses (or analogues of those points) apply just as much in Jane's case as they do in the ordinary case. First, the “meanings” of Jane's sounds are determined by the objects that generally cause her to produce them; so there is no possibility of her generally producing those sounds in response to the wrong objects. Second, if Jane says “My sound ‘Φ’ ‘means’ Φ,” she cannot be mistaken; the “meanings” of the two instances of the sound “Φ” will necessarily be the same. But if Davidson's two points apply in the case of Jane's blind dispositions, where there is no distinctive first-person knowledge of meaning, they cannot explain what is most striking and most puzzling about the character of our knowledge of the meanings of our words: the fact that we generally know what our words mean without needing to observe how we use them.22
Of course, it may be that no philosophical account can explain this feature of our knowledge of the meanings of our words; maybe we must simply accept it as a basic fact. In that case, it will not be a special shortcoming of Davidson's account that it fails to explain it. I suspect that that may in fact be Davidson's own assessment of the situation; he may think that the best we can hope for is an account that explains certain limited aspects of first-person authority, and that we must accept that some of the phenomenon's most distinctive and central features are not susceptible of philosophical explanation at all. If that is Davidson's view, however, it is not a view that he explicitly articulates.
Davidson himself has expressed dissatisfaction with his explanation of first-person authority. He writes:
I am by no means satisfied with my explanation of first person authority. More remains to be said in defence of my position, and indeed it may be that there are aspects of first person authority that cannot even be approached following the line I have suggested. It is clear, too, that my treatment of the topic is incomplete, because I have said nothing about knowledge of our own sensations.
(Davidson 1993b: 248)
In a similar vein he says, of the account he offered in “First Person Authority,” that he “had the feeling that there had to be more to the story, something that would throw more light on the nature of self-knowledge” (Davidson 1999: 226). Critics have agreed with that assessment.
The most frequent criticism of Davidson's account is this. In order to explain a person's knowledge that she believes that Bamako is the capital of Mali, say, we must explain two aspects or dimensions of her knowledge: we must explain her knowledge that the proposition she believes is that Bamako is the capital of Mali (rather than, say, that N'Djamena is the capital of Chad), and we must explain her knowledge that she believes that proposition (rather than, say, wanting it to be true, or intending to make it true). But Davidson's account focuses exclusively on the first task: on explaining our knowledge of the contents of our beliefs. He says nothing at all to explain our knowledge that we believe those contents. So his explanation of our knowledge of our own beliefs is at best radically incomplete.23
In fairness to Davidson, it should be said that he openly acknowledges this feature of his account. As we saw, he says explicitly that he is not explaining a speaker's knowledge that she holds a sentence true, but only her knowledge of what the sentence means. Nonetheless, the underlying point of the criticism is perfectly correct. A full account of our knowledge of our own beliefs should explain not only our knowledge of their contents but also our knowledge that we believe those contents. Davidson's account says nothing at all about the second. So even if it is successful on its own terms, it accounts for only part of what we want an account of first-person authority to explain.24
In a similar way, critics have complained that Davidson's explanation deals directly only with our knowledge of our current beliefs. It does not explain our knowledge of our past beliefs; or our knowledge of our other propositional attitudes, like desire and intention; or our knowledge of attitudes that do not have propositional objects; or our knowledge of our own sensations. As we saw at the outset, Davidson himself is again quite explicit about these limits on the scope of his account. So pointing them out is not in itself an objection to anything he claims to have done. Nonetheless, critics have said, what we want from a philosophical account of first-person authority is an understanding of the phenomenon as a whole: an explanation that addresses every instance of first-person authority. By his own admission, Davidson's account does not give us that.
The point is a fair one, even if it only states what Davidson himself always acknowledged. On the other hand, we should not necessarily assume that first-person authority is a uniform phenomenon, susceptible of a uniform explanation. It seems plausible, on the contrary, that the source and extent of our authority about our own mental properties may be different in different cases. In that case, we should expect the explanation of first-person authority, too, to be different in different cases. And if that is right, then there is no objection in principle to the idea that there could be a successful explanation of one aspect of first-person authority that did not generalize to other aspects.
Many have thought that the idea that we have first-person authority with respect to the contents of our thoughts is incompatible with semantic externalism: the thesis that the meanings of words, and the contents of thoughts, are constitutively dependent on the nature of the things and kinds in one's environment. Against that challenge, Davidson argues not only that externalism does not threaten first-person authority, but that externalism is actually part of what explains it (for this latter claim, see Davidson 1987: 17–18, 37).
As we have seen, Davidson's own account of meaning and content has an important externalist element:
what a person's words mean depends in the most basic cases on the kinds of objects and events that have caused the person to hold the words to be applicable; similarly for what the person's thoughts are about.
(Davidson 1987: 37)
And we have seen how this externalist account of meaning plays a central part in Davidson's explanation of a speaker's knowledge of the meanings of her own words. Davidson has his own arguments for externalism. But he also uses Putnam's Twin Earth thought experiment to illustrate the view25:
[Putnam] asks us to imagine two people exactly alike physically and (therefore) alike with respect to all “narrow” psychological states. One of the two people, an inhabitant of Earth, has learned to use the word “water” by being shown water, reading and hearing about it, etc. The other, an inhabitant of Twin Earth, has learned to use the word “water” under conditions not observably different, but the substance to which she has been exposed is not water but a lookalike substance we may call “twater.” Under the circumstances, Putnam claims, the first speaker refers to water when she uses the word “water”; her twin refers to twater when she uses the word “water.”
(Davidson 1987: 20–21)
Davidson agrees. He continues:
How about the thoughts of these two speakers? The first says to herself, when facing a glass of water, “Here's a glass of water”; the second mutters exactly the same sounds to herself when facing a glass of twater. Each speaks the truth, since their words mean different things. And since each is sincere, it is natural to suppose that they believe different things, the first believing that there is a glass of water in front of her, the second believing that there is a glass of twater in front of her.
(Davidson 1987: 21)
But do our two speakers know what they believe? The challenge that externalism seems to pose to first-person authority is this. Consider the inhabitant of Earth. Her word “water” refers to water, and her thoughts are about water. But from the inside, as it were, there is no detectable difference between the way things actually are and the way they would be if she were an inhabitant of Twin Earth, her word “water” referred to twater, and her thoughts were about twater. So, it is tempting to conclude, she does not know that her word “water” refers to water and not twater; and she does not know that her thoughts are about water and not twater. It seems, in short, that she does not know the contents of her own thoughts.
Davidson disagrees. In his view, there is no reason to think that the speakers on Earth and Twin Earth do not know what their thoughts are about. Two ideas underlie, or inform, Davidson's view: one is that “what determines the contents of thoughts also determines what the thinker thinks the contents are” (Davidson 1988b: 664); the other is that knowing what one is thinking about does not require the ability to discriminate the object of one's thought from every possible alternative. We can expand on these points in turn.
Davidson summarizes the first idea like this:
An interpreter must discover, or correctly assume on the basis of indirect evidence, what the external factors are that determine the content of another's thought; but since these factors determine both the contents of one's thought and the contents of the thought one believes one has (these being one and the same thought), there is no room for error about the contents of one's own thought of the sort that can arise with respect to the thoughts of others.
(Davidson 1991: 197–198)
That argument is an application to the case of externalism of the same line of thought we discussed earlier in connection with disquotational knowledge of the meanings of one's words.26 Suppose I see someone looking at a glass of water and saying “Here's a glass of water.” I assume that her word “water” refers to water and that she is expressing the thought that there is a glass of water in front of her. I am probably right. But I could be wrong. For suppose she is an inhabitant of Twin Earth, recently transported to Earth without her (or my) knowledge. In that case, her word “water” will refer to twater, and she will be expressing the thought that there is a glass of twater in front of her; so my judgement that she is thinking that there is a glass of water in front of her will be false. But there can be no error of that kind when I ascribe a thought to myself. Suppose I say, “Here's a glass of water.” The thought I express is a thought about water. Then I say, “I have just expressed the thought that there's a glass of water here.” There is no possibility that the word “water” in my second utterance means something different from the word “water” in my first utterance. And there is no possibility that the associated thoughts are about different liquids. For the meaning of my word “water,” and the content of the associated thought, is determined by the same external factors on both occasions. I cannot, therefore, make the kind of mistake about the content of my own thought that I can make about the content of another person's.
In the literature, that point is now treated as the standard response to the threat to first-person authority that is allegedly posed by externalism.27 But some commentators find the response unsatisfying. They agree with Davidson that the contents of my higher-order, self-ascriptive thoughts will always match the contents of the first-order thoughts they are about. So they agree that my beliefs about the contents of my thoughts will always be true. But, they say, that does not suffice for me to know what my thoughts are about. For I cannot tell the difference between water and twater. So I cannot tell the difference between a situation in which I am thinking about water and a situation in which I am thinking about twater. I therefore lack the kind of discriminative knowledge that, according to the critics, is needed if I am to know that I am thinking about water. (For discussion of this criticism, and further references, see Brown 2009: 770–772.)
Davidson's response to the objection is straightforward: he denies that knowing the contents of one's thoughts requires the discriminative knowledge that the critic supposes. Part of the point of externalism is that, given appropriate facts about my environment and causal history, my thoughts can be thoughts about Fs rather than Gs even if I cannot discriminate Fs from Gs. We see that point in the Twin Earth thought experiment. And we see it in real-life cases too:
I may not know the difference between an echidna and a porcupine; as a result, I may call all echidnas I come across porcupines. Yet because of the environment in which I learned the word “porcupine,” my word “porcupine” refers to porcupines and not to echidnas.
(Davidson 1988a: 48–49)
Similarly, even if I cannot distinguish porcupines from echidnas, the belief I express when I say “That's a porcupine” is a belief about porcupines and not about echidnas. And by the same token, Davidson thinks, I can know that I believe that the creature in front of me is a porcupine, without needing to be able to distinguish porcupines from echidnas. Davidson does not spell out detailed reasons for that view.28 But we might defend it in the following way, which is consistent with the kind of externalism about knowledge that runs through Davidson's work.
To know that I am having the thought that a creature is a porcupine, I must be able to discriminate that thought from relevant alternative thoughts. And I can certainly do that; for I can discriminate the situation in which I am thinking that something is a porcupine from a situation in which I am thinking that it is an anteater, or a hedgehog, or a sea-urchin. It is true that I cannot discriminate porcupines from echidnas. So it is true that I cannot discriminate the actual situation (in which my word “porcupine” refers to porcupines, and the thoughts I express when I use the word are thoughts about porcupines) from a possible situation in which I learned the word “porcupine” in a different environment, the word refers to echidnas, and the associated thoughts are about echidnas. But that does not threaten my knowledge that I am thinking that the animal before me is a porcupine. For the possible situation just described (the situation in which my word “porcupine” refers to echidnas) is not a relevant alternative to my actual situation. Accordingly, I do not need to be able to discriminate my actual situation from that alternative in order to count as knowing that I am thinking that the animal is a porcupine.
A full defence of this view would plainly require more detail than that. (e.g., What determines which alternatives count as relevant alternatives?) But we have said enough to show why Davidson might take the view he does, and what makes that view a plausible one to take.
Notes
1 This line of argument is most clearly spelled out in Davidson (1993b: 249).
2 What Davidson actually says is this: “if you or I or anyone knows that I hold [the sentence “Wagner died happy”] to be true on this occasion of utterance, and she knows what I meant by this sentence on this occasion of utterance, then she knows what I believe – what belief I expressed” (Davidson 1984: 11). That states a sufficient condition for knowing what belief I expressed in uttering the sentence. But it is very plausible that Davidson endorses the corresponding necessary condition: if someone knows what belief I express when I say “Wagner died happy,” then she knows that I hold the sentence true, and she knows what I mean by the sentence.
3 As Davidson puts it in a later comment: “As I said in [‘First Person Authority’], if there is asymmetry with respect to the content of the attitude, there is also asymmetry with respect to the claim that the attitude exists” (Davidson 1993b: 250).
4 Davidson holds that it would be “circular” to explain the asymmetry between knowledge of one's own beliefs and knowledge of others' beliefs by simply assuming, without explaining, “an asymmetry in the assurance you and I have that I hold the sentence I have just uttered to be a true sentence” (Davidson 1984: 12). Since he has no explanation of this latter asymmetry to offer, he does not appeal to it in explaining the asymmetry between self-knowledge and knowledge of others.
5 The point of the term “disquotational” is that the right-hand side of a judgement like “My utterance of ‘Wagner died happy’ is true if and only if Wagner died happy” “disquotes” – that is, removes the quotation marks from – the sentence identified on its left-hand side.
6 See for example, Bernecker (1996: 132), Hacker (1997: 299), Lepore and Ludwig (2005: 353, 361–364), Nguyen (2005: 459–462), Thöle (1993: 245), and Wright (2001: 348–349).
7 An exception is Jacobsen (2009): who (rightly, in my view) makes this point central to Davidson's explanation. I take the same view in Child (2007). See also Lepore and Ludwig (2005: 364–369).
8 It seems possible for someone to realize that she has a particular belief by reflecting on her own linguistic behaviour; just as it is possible for someone to learn that she has a belief by reflecting on her nonlinguistic behaviour. But these are clearly abnormal cases.
9 For an objection of this form, see Lepore and Ludwig (2005: 260).
10 For Davidson's approval, in the right kind of context, of the formulation “s means that p,” see Davidson (1976: 175).
11 For objections of this form, see the references to Hacker (1997) and Thöle (1993) in note 6.
12 The example is from Hacker (1997).
13 In some of Davidson's formulations, including the passage just quoted in the text (Davidson 1988a: 44–45), the idea is that words get their meanings from the objects and events that normally caused me to hold the words applicable in the learning environment. Other formulations (including those in Davidson's main discussions of first-person authority: Davidson (1984: 13–14, 1987: 37–38) ) suggest that the meaning-determining causes are the causes that are normal in the speaker's current environment. The difference between these formulations is significant, and will affect our verdicts about particular cases. But for present purposes, we can gloss over it.
14 Of course, Davidson thinks, this is only true in the simplest and most basic cases; it is not true that “all words and sentences are this directly conditioned to what they are about.” Indeed, someone “can perfectly well learn to use the word [‘fire’] without ever seeing [fire]” (Davidson 1987: 29); she can learn the word by seeing pictures of fires, or by being told what fire is, and so on. What Davidson insists, however, is that for each person, there must be many words and sentences that are directly conditioned to what they are about. It is those words and sentences, he thinks, “that anchor language to the world” (Davidson 1988a: 45, 1991: 197).
15 See for example, Jacobsen (2009: 255), Lepore and Ludwig (2005: 355, 364–365), Nguyen (2005: 463), Smith (1998: 417), and Thöle (1993: 243–244).
16 For this claim see Davidson (1982).
17 See Davidson (1992).
18 For further discussion of disquotational knowledge of meaning, with competing proposals about what exactly is required for a speaker to qualify as having such knowledge, see Higginbotham (1992, 1998) and Richard (1992).
19 We do sometimes proceed in that way. But the cases in which we do so are precisely those where we are not sure what our words mean; they are the abnormal cases where we do have to observe our own use of our words in order to tell what we mean by them.
20 See for example, Wittgenstein (1958: §§139, 191, 197). Smith (1998: 427) also draws on Wittgenstein in characterizing what is missing from Davidson's account. For a different presentation of essentially the same point about the character of one's knowledge of one's own language, see Higginbotham (1998: 440).
21 See Grice (1989).
22 For further discussion of this objection, see Child (2007: 171–173).
23 For objections of this form, see Bernecker (1996: 136), Hacker (1997: 296), Lepore and Ludwig (2005: 359–360, 370–371), Nguyen (2005: 460), Thöle (1993: 242), and Wright (2001: 349).
24 For one suggestion about how we might use materials implicit in Davidson's own work to provide what he himself does not give – an account of how a person can know not only the content of her belief but also that she has that belief – see Jacobsen (2009).
25 For the original thought experiment, see Putnam (1975).
26 Here I differ from Lepore and Ludwig (2005: 369–371), who hold that the quoted passage (Davidson 1991: 197–198) suggests a new explanation of our knowledge of our thoughts, different from that given in Davidson (1984).
27 For an independent statement of essentially the same response, see Burge (1988). Despite the similarity of Burge's and Davidson's responses, Davidson holds that Burge's brand of externalism really is incompatible with first-person authority, in a way that his own form of externalism is not. For Davidson, the meanings of my words are determined by my own applications of those words; that fact, he thinks, plays an important part in explaining why I generally know what my words mean. By contrast, Davidson thinks, “if what we mean and think is determined by the linguistic habits of those around us in the way Burge believes they are, then first-person authority is very seriously compromised” (Davidson 1987: 26–27; see also Davidson 1991: 199). Davidson would acknowledge that even on Burge's view of what determines the meanings of our words, a disquotational statement of the meaning of one's words is bound to be true. But, as we have seen, he does not think that the ability to produce a true disquotational statement of the meanings of one's words suffices for knowing what they mean.
28 For more by way of argument, see Burge (1988) and Falvey and Owens (1994).
References
Bernecker, S. (1996). Davidson on first-person authority and externalism. Inquiry 39:121–139.
Brown, J. (2009). Semantic externalism and self-knowledge. In The Oxford Handbook of Philosophy of Mind, B.P. McLaughlin, A. Beckermann and S. Walter (eds). Oxford: Oxford University Press, pp. 767–780.
Burge, T. (1988). Individualism and self-knowledge. Journal of Philosophy 85:649–663.
Child, W. (2007). Davidson on first person authority and knowledge of meaning. Nous (Detroit, Mich.) 41:157–177.
Davidson, D. (1976). Reply to Foster. In Inquiries into Truth and Interpretation, Oxford: Oxford University Press (1984), pp. 171–179.
———. (1982). Rational animals. In Davidson (2001): 95–105.
———. (1984). First person authority. In Davidson (2001): 3–14.
———. (1987). Knowing one's own mind. In Davidson (2001): 15–38.
———. (1988a). The myth of the subjective. In Davidson (2001): 39–52.
———. (1988b). Reply to Burge. Journal of Philosophy 85:664–665.
———. (1989). What is present to the mind? In Davidson (2001): 53–67.
———. (1991). Epistemology externalized. In Davidson (2001): 193–204.
———. (1992). The second person. In Davidson (2001): 107–121.
———. (1993a). Reply to Eva Picardi. In Stoecker (ed.) (1993): 210–212.
———. (1993b). Reply to Bernhard Thöle. In Stoecker (ed.) (1993): 248–250.
———. (1999). The Philosophy of Donald Davidson, L. Hahn (ed.). Chicago: Open Court, pp. 225–227.
———. (2001). Subjective, Intersubjective, Objective. Oxford: Oxford University Press.
Falvey, K. and Owens, J. (1994). Externalism, self-knowledge, and skepticism. Philosophical Review 103:107–137.
Grice, H.P. (1989). Meaning. In Studies in the Way of Words, H.P. Grice (ed.). Cambridge, MA: Harvard University Press.
Hacker, P.M.S. (1997). Davidson on first-person authority. Philosophical Quarterly 47:285–304.
Higginbotham, J. (1992). Truth and understanding. Philosophical Studies 65:3–16.
———. (1998). On knowing one's own language. In Wright, Smith, and Macdonald (1998): 429–441.
Jacobsen, R. (2009). Davidson and first-person authority: parataxis and self-expression. Pacific Philosophical Quarterly 90:251–266.
Lepore, E. and Ludwig, K. (2005). Donald Davidson: Meaning, Truth, Language, and Reality. Oxford: Oxford University Press. (Chapter 20 of this book reproduces and slightly expands an earlier paper: Kirk Ludwig, “First-person knowledge and authority,” in Language, Mind, and Epistemology: On Donald Davidson's Philosophy, edited by Gerhard Preyer, Frank Siebelt, and Alexander Ulfig, 367–398, Dordrecht: Kluwer Academic Publishers, 1994.)
Nguyen, A.M. (2005). Davidson on first-person authority. Journal of Value Inquiry 38:457–472.
Putnam, H. (1975). The meaning of “meaning.” In Mind, Language and Reality, Cambridge, UK: Cambridge University Press, pp. 215–271.
Richard, M. (1992). Semantic competence and disquotational knowledge. Philosophical Studies 65:37–52.
Smith, B.C. (1998). On knowing one's own language. In Wright, Smith, and Macdonald (1998): 391–428.
Stoecker, R. (ed.) (1993). Reflecting Davidson: Donald Davidson Responding to an International Forum of Philosophers. Berlin: de Gruyter.
Thöle, B. (1993). The explanation of first person authority. In Stoecker (ed.) (1993): 213–247.
Wittgenstein, L. (1958). Philosophical Investigations, 2nd ed, G.E.M. Anscombe, R. Rhees and G.H. von Wright (eds), translated by G.E.M. Anscombe. Oxford: Blackwell.
Wright, C. (2001). The problem of self-knowledge (II). In Rails to Infinity: Essays on Themes from Wittgenstein's Philosophical Investigations. Cambridge, MA: Harvard University Press, p. 345.
Wright, C., Smith, B.C. and Macdonald, C. (eds) (1998). Knowing Our Own Minds. Oxford: Oxford University Press.