Enough is Enough: Austin on Knowing Guy Longworth1 November 9th 2016 1. Introduction. J. L. Austin’s main discussion of knowledge is in “Other Minds”. (Austin 1946; 1979.) The essay gives rise to numerous questions, both local and global. Ostensibly, the topic of the essay is knowledge of other minds. However, explicit discussion of that topic is postponed until its twenty seventh page. Apparently by way of preamble, the main body of the essay comprises an analysis of aspects of our ordinary treatment of expressions of knowledge, stippled with tantalizing pronouncements about knowing in general. One aim of my discussion is to address a global question about the function of Austin’s more general discussion of knowledge. How, if at all, does it further pursuit of our knowledge of other minds? More local questions arise about Austin’s pronouncements. I’ll also address a more local question. Austin invites us to consider a natural way of treating a claim to the effect that a goldfinch is present: If you have asked ‘How do you know it’s a goldfinch?’ then I may reply ‘From its behaviour’, ‘By its markings’, or, in more detail, ‘By its red head’, ‘From its eating thistles’.... You may object:... But that’s not enough: plenty of other birds have red heads. (Austin 1946: 154–5; 1979: 83) Reflecting on the extent to which one who knows is required to be in a position to address such challenges, he writes the following: (b) Enough is enough: it doesn’t mean everything. Enough means enough to show that (within reason, and for presents intents and purposes) it ‘can’t’ be anything else, there is no room for an alternative, competing, description of it. It does not mean, e.g., enough to show it isn’t a stuffed goldfinch. (Austin 1946: 156; 1979: 84) Austin’s pronouncement is both enticing and elusive: enticing, in that it seems to expose a significant limit on our obligations as knowers; elusive, in that Austin fails to resolve or vindicate the alleged limit. What does he mean to claim about knowing? Should we believe him? 1 I’m grateful for discussion and comments to audiences at Oxford, Porto, and Warwick, and to Bill Child, Thomas Crowther, Naomi Eilan, Elizabeth Fricker, Anil Gomes, Mark Kalderon, Hemdat Lerman, Ian Phillips, Johannes Roessler, Matthew Soteriou, and Charles Travis. 1 Philosophical positions rarely form in a vacuum, and so one way to enhance one’s understanding of a position is by attending to its roots. On the topic of knowledge, Austin’s avant-couriers were his teacher, H. A. Prichard, and Prichard’s teacher, John Cook Wilson, and so it is on their work that I shall focus. (Important influences who won’t be considered here include G. E. Moore, especially his 1905–6.) I’ll begin by sketching the shared core of Cook Wilson’s and Prichard’s views about knowledge—the epistemological component of Oxford Realism—before pointing to ways in which Austin’s position emerges naturally from theirs. I’ll suggest that Austin would have viewed his discussion of knowledge, not only as a preamble to a treatment of knowledge of other minds, but as a case study. And I’ll suggest that Austin’s pronouncement is best understood as a partial characterization of a necessary, but insufficient, condition on knowing. We’ll see that when Austin’s pronouncement is so understood, there is no reason to take him to be proposing that someone might know that a goldfinch is present without being in a position to know that it isn’t stuffed. Section 2. comprises a sketch of six core commitments of Oxford Realism. Section 3. explains how Austin’s general discussion of knowledge constitutes a case study of our knowledge of other minds. Sections 4. and 5. develop an interpretation of Austin’s project on which his pronouncement leaves intact that being in a position to know is closed under known entailment. 2. Oxford Realism. 2.1. Knowledge as primitive. The first and most basic commitment of Oxford Realism is that knowledge is primitive. Thus, Cook Wilson writes: Perhaps most fallacies in the theory of knowledge are reduced to the primary one of trying to explain the nature of knowing or apprehending. We cannot construct knowing—the act of apprehending—out of any elements. I remember quite early in my philosophic reflection having an instinctive aversion to the very expression ‘theory of knowledge’. (Cook Wilson 1926: 803) Prichard echoes: Knowledge is sui generis and therefore a ‘theory’ of it is impossible. Knowledge is simply knowledge, and any attempt to state it in terms of something else must end in describing something which is not knowledge. (Prichard 1909: 245) According to Cook Wilson and Prichard, knowledge is a distinct kind, and cannot be constructed out of elements distinct from knowledge. If the goal of a theory of knowledge is to say how knowledge is constructed, then the goal is unachievable; there can be no such theory. 2.2. Knowledge as akin to proof. The first commitment of Oxford Realism is bound up with a second: knowing amounts to, or is equivalent to, possession of proof: 2 In knowing, we can have nothing to do with the so-called ‘greater strength’ of the evidence on which the opinion is grounded; simply because we know that this ‘greater strength’ of evidence of A’s being B is compatible with A’s not being B at all. (Cook Wilson 1926: 100) The view is not that one who knows must possess a cogent derivation of what they know from premises that are distinct—and, perhaps, known. That is, the view is not that knowing is equivalent to possession of a proof. Rather, the view is that knowing is itself proof of what is known Where one knows, one has a conclusive guarantee of that which one knows. One’s standing with respect to that which one knows is incompatible with falsity. Thus, meeting a threshold condition on strength of evidence could not suffice for knowing if meeting that condition were consistent with falsity. As Prichard notes, Cook Wilson appears to have derived his model of knowing from reflection on mathematics: The point of departure of Cook Wilson's views lay in his unwavering conviction of the truth of mathematics. In mathematics we have, without real possibility of question, an instance of knowledge; we are certain, we know. (Prichard 1919: 302) The model of mathematics is useful in presenting a case in which it is plausible that we can possess absolute guarantees of truth. Furthermore, it presents a case in which we can possess such guarantees both with respect to basic and derived truths, and so a case in which the distinction between possession of proof and possession of a proof is operative. However, the model also points to a delicate issue about the notion of incompatibility that figures in Cook Wilson’s exposition. For suppose that incompatibility of a standing with falsity amounted to the impossibility of having that standing with respect to a proposition whilst the proposition is false. On the natural assumption that true mathematical propositions are necessarily true, it would be impossible to have any standing with respect to such propositions whilst those propositions were false. And in that case, incompatibility with falsity would fail to characterize a distinctive property of the standing of one who knows. Thus, the operative notion of incompatibility must be more demanding than the simple modal notion. 2.3. Knowledge as a state of mind. The third commitment of Oxford Realism is a partial corollary of the first two: knowing is a state (or frame, or condition) of mind. We can reconstruct a path to the commitment as follows. If one’s standing when one knows is to furnish one with a guarantee against falsity, then it must make a difference to one’s subjectivity. Thus, one’s standing must partly comprise a mental state. Suppose, then, that that mental state did not suffice for knowing. In that case, knowing would comprise that mental state together with whatever extra-mental elements provide a guarantee against falsity. But in that case, knowing would be constructible out of elements. Hence, since knowing is at least partly mental, and is not constructible out of elements, it must be wholly mental. That is, 3 knowing is a mental state, occupancy of which state is incompatible with falsity. (See also McDowell 1982; Williamson 2000.) 2.4. Knowing as distinct from believing. Connected with the third commitment is a fourth: knowing is not a form of believing. In particular, knowing is not believing whilst meeting further conditions. In fact, Cook Wilson makes the stronger claim that knowing excludes believing: Belief is not knowledge and the man who knows does not believe at all what he knows; he knows it. (Cook Wilson 1926: 100) Prichard brings together the previous four commitments in a way which again echoes Cook Wilson: Knowing is not something which differs from being convinced by a difference of degree of something such as a feeling of confidence, as being more convinced differs from being less convinced, or as a fast movement differs from a slow movement. Knowing and believing differ in kind as do desiring and feeling, or as do a red colour and a blue colour. (Prichard 1950: 87) Both Cook Wilson and Prichard are drawn to the stronger claim that knowing excludes believing due to their positive views about the distinctive nature of believing—in particular, that believing is a matter of holding something true on broadly evidential grounds whilst recognizing that one’s grounds fail to decide the issue. By contrast, some contemporary thinkers will be willing to adopt a more minimal conception of believing, or to accept the existence of a more general kind of state of mind that encompasses believing. For example, they will be willing to allow a conception on which believing, or some more general sort of state of mind, is a matter of holding something true in a way that is potentially responsive to evidence. However, the crucial claim here is that knowing isn’t itself a form of believing. And one could consistently endorse that claim whilst allowing that knowing doesn’t exclude believing, or even that knowing entails believing. I can see no grounds for thinking that Cook Wilson or Prichard would have denied that knowing was, or at least entailed occupancy of, a frame of mind of the more general sort. (See Williamson 2000: 41–48.) 2.5. The Accretion. The fifth commitment of Oxford Realism concerns our capacities to know which frames of mind we occupy. The seemingly implausible strength of this commitment, together with its seeming independence from other commitments, has led Charles Travis to label it the Accretion. (Travis 2005.) Cook Wilson presents the commitment in the following passage: [knowledge cannot be one of] two states of mind…the correct and the erroneous one…quite indistinguishable to the man himself. [For] as the man does not know in the erroneous state of mind, neither can he know in the other state (Cook Wilson 1926: 107) 4 The first thought contained here is that subjects must be in a position to distinguish any state of knowledge from other “erroneous” states of mind, at least in principle. The second is that if a state were not in the required sense distinguishable by its subject from “erroneous” states, then—since those other states are, by assumption, not states of knowledge—that state could not be a state of knowledge. Cook Wilson’s view may have been sponsored by an argument like the following. In order for a state to be a case of knowledge it must be different in kind from any “erroneous” state. Furthermore—and, perhaps, because the kinds in question are mental kinds—the required difference must have a subjective reflection: it must make a difference to how things are, or seem, from the subject’s perspective. If the difference between the target state and its “erroneous” ringers were blankly external to the way things are for the subject, then how things were subjectively for the subject of either kind of state would be compatible with their not knowing. And in that case, even if they occupied the target state, they wouldn’t know. A final step in the argument is required in order to connect the requirement that the difference between knowing and its “erroneous” ringers be reflected subjectively with the further requirement that the subject be in a position to distinguish the two states—that is, to tell the two states apart. What, more precisely, does Cook Wilson mean by claiming that subjects must be in a position to distinguish states of knowing from ringers? Prichard offers the following elaboration: We must recognize that when we know something we either do, or by reflecting can, know that our condition is one of knowing that thing, while when we believe something, we either do or can know that our condition is one of believing and not of knowing: so that we cannot mistake belief for knowledge or vice versa. (Prichard 1950: 88) Prichard’s elaboration of the Accretion invokes two conditions on knowing: (i) If one knows p, then one is in a position, at least in principle, to know by reflection that one knows p. (ii) If one believes p without knowing p, then one is in a position, at least in principle, to know by reflection that one believes p without knowing p. Condition (i) does not obviously entail condition (ii). It is consistent to hold that one might fail to know p whilst being unable to know by reflection that one failed to know p even if one also held that if one knew p, then one would be in a position to know that one knew p. To take one sort of example, one might reasonably hold that a severely drunk person can fail to know that they are drunk (and can even believe that they are sober) whilst at the same time being precluded by their drunkenness from knowing that they don’t know that they are drunk. And holding that would seem perfectly consistent with also holding that someone who is sober, and who knows that they are, would be in a 5 position to know that they know that. (See Williams 1978: 309–313; Soteriou 2016: 117–156.) Furthermore, when conditions (i) and (ii) are distinguished, it becomes apparent that meeting the former condition would be enough to render states of knowing distinguishable from their ringers. On at least one reasonable understanding, one can distinguish Fs from Gs just in case one can activate knowledge that presented Fs are not Gs. Similarly, one can distinguish Gs from Fs just in case one can activate knowledge that presented Gs are not Fs. Thus, distinguishability is asymmetrical. Very often the required capacities run in step: one can activate knowledge that a presented sheep isn’t a wolf when, and only when, one can activate knowledge that a presented wolf isn’t a sheep. In those cases, the asymmetry doesn’t matter. But in the sorts of cases we’ve just considered, the capacities can come apart: one can activate knowledge that a case of one’s sobriety isn’t a case of one’s severe drunkenness, even though one cannot activate knowledge that a case of one’s severe drunkenness isn’t a case of one’s sobriety. One can distinguish one’s knowing from one’s occupying ringer states if one can activate knowledge that one’s state is not a ringer state. And that can be so even if one cannot distinguish one’s occupying a ringer state from one’s knowing. Cook Wilson’s requirement that knowing be distinguishable by its subjects from ignorance can be implemented by condition (i); condition (ii) is needless. Even if we treat the Accretion as incorporating only condition (i), it seems implausibly demanding. For according to condition (i), that one knows is, in Timothy Williamson’s sense, a luminous condition: for every case a, if in a one knows, then in a one is in a position to know that one knows. And Williamson has offered powerful arguments that no condition which obtains only sometimes—and in particular no such condition that one knows—is luminous. (Williamson 2000: 93–123) 2.6. Being under the impression. The Accretion presents Oxford Realism with a difficulty. Given the luminosity of one’s epistemic position, how is it possible for one to make mistakes? More carefully, how is it possible for one to make mistakes that one cannot, by reflection, correct? Cook Wilson offers the following example: …we see at a little distance a person whom ‘we mistake for an acquaintance’ and without hesitation perform some act which it would be a liberty to take with anyone but an acquaintance, do something in fact which we rightly say we should not have done if we had ever suspected he was not an acquaintance. We did not act on an opinion that it was our friend; for, in forming an opinion, we are aware that the evidence is insufficient and, if we had thought that, we should never have done the act. It seems more like belief; but if we had consciously made it a matter of belief, we should have distinguished it from knowledge, and then again, ex hypothesi, we should not have done the act. Probably one answer offered would be that, though we didn’t know, we thought we knew. But this will not suffice. Apart from the criticism we have already passed on this phrase itself, if we really thought we knew, we must have reflected and must have thought the evidence conclusive, 6 whereas, ex hypothesi, any reflection shows it could not be conclusive. (Cook Wilson 1926: 109–10) Cook Wilson fails to detail what no Edwardian would have risked from a position of ignorance. His response to the challenge of explaining its performance is the invocation of a further species of attitude: being under the impression. This is the sixth commitment of Oxford Realism. Being under the impression is—like opinion or belief—a mode of holding something to be true. However, unlike belief or opinion, being under the impression need not be installed or sustained by reflection. Indeed, being under the impression is incompatible with, and so apt to be destroyed by, reflection. (Cook Wilson 1926: 108–113) Being under the impression falls outside the scope of the Accretion. One cannot know by reflection that one is under an impression, since reflection would release one from its hold. Thus, being under the impression can help to explain how subjects can hold things to be true in a way precluded by the activation of knowledge. It can explain how unreflective mistakes are possible. However, it’s natural to think that some mistakes can withstand reflection. If one were ensconced in a standard skeptical scenario, then one might be subject to mistakes both about how things were—e.g. that one had hands—and about one’s attitudes to how things were—e.g. that one knew that one had hands. And it’s natural to think that no amount of reflection would remedy one’s situation. There are two broad routes via which an attempt might be made to exploit the state of being under the impression in order to explain reflective error. The first would be to deny that genuinely reflective errors are possible. Genuine reflection would reveal that one doesn’t know that one has hands and, so, that one isn’t entitled to hold true that one has hands. The claim would be that we are commonly less reflective than we take ourselves to be—that is, than we are under the impression of being. The second route would be to extend the first by seeking to explain barriers to reflection. The claim would be that in certain circumstances genuine reflection cannot be undertaken. Severe drunkenness might block reflection. More delicately, being the subject of a skeptical scenario might prevent the activation of knowledge that would otherwise enable reflection. That completes my sketch of the six core commitments of Oxford Realism: (1) knowledge is primitive; (2) knowledge is akin to proof; (3) knowledge is a state of mind; (4) knowledge is distinct from belief; (5) the Accretion; (6) mistakes depend upon the unreflective state of being under the impression. (See also Marion 2000a, 2000b; Travis 2005; Travis and Kalderon 2013.) 3. Other Minds. There are numerous echoes of Oxford Realism in Austin’s work. Cook Wilson and Prichard both present commitment (1), knowledge as primitive, as precluding one form of theory of knowledge. Austin agrees: …there could be no general answer to the questions what is evidence for what, what is certain, what is doubtful, what needs or does not need evidence, can or can’t be verified. If the Theory of 7 Knowledge consists in finding grounds for such an answer, there is no such thing. (Austin 1962: 124) Commitment (2), knowledge as akin to proof, reverberates more widely. Thus, for example, Austin contrasts (inconclusive) evidence with what settles a question: The situation in which I would properly be said to have evidence for the statement that some animal is a pig is that, for example, in which the beast itself is not actually on view, but I can see plenty of pig-like marks on the ground outside its retreat. If I find a few buckets of pig-food, that’s a bit more evidence, and the noises and the smell may provide better evidence still. But if the animal then emerges and stands there plainly in view, there is no longer any question of collecting evidence; its coming into view doesn’t provide me with more evidence that it’s a pig, I can now just see that it is, the question is settled. (Austin 1962: 115; cp. 1946: 176–182; 1979: 105–111) Austin makes a closely related point here: …saying ‘I know’ is taking a new plunge. But it is not saying ‘I have performed a specially striking feat of cognition, superior, in the same scale as believing and being sure, even to being merely quite sure’: for there is nothing in that scale superior to being quite sure. (Austin 1946: 171; 1979: 99) It is natural to read Austin as appealing here to a distinction between the accumulation of grounds for sureness of belief and the achievement of a position that differs not merely in degree, but in kind. The distinction between evidence and proof feeds into the idea of knowledge as a state of mind distinct from belief (commitments (3) and (4)). At first blush, Austin might be read as rejecting the first idea in the following passage: If we like to say that ‘I believe’, and likewise ‘I am sure’ and ‘I am certain’, are descriptions of subjective mental or cognitive states or attitudes, or what not, then ‘I know’ is not that, or at least not merely that: it functions differently in talking. (Austin 1946: 150; 1979: 78–9) However, the key phrase is “or at least not merely that” and the main point is to distinguish knowing from believing. The claim is that ‘I know’ is not merely a description of a subjective mental state. That is either because knowing differs from believing in not being merely a subjective mental state, or because ‘I know’ has functions over and above describing the subjective mental state of knowing. (The latter idea figures in Austin’s infamous comparison of ‘I know’ with ‘I promise’. See Austin 1946: 169–175; 1979: 97–103.) There is evidence, then, if not proof, that Austin accepted at least the first four commitments of Oxford Realism. That supports a straightforward answer to our first question. A general discussion of knowledge might be useful preparation for a discussion of knowledge of 8 other minds. However, insofar as Austin views knowledge itself as a state of mind, such a discussion assumes a more central role. For the question how we can know about what other people know is viewed as a special case of the more general question of how we can know about other’s minds. Furthermore, there are forces internal to Oxford Realism that make it pressing to address questions about our knowledge of what other people know. One source of pressure is the idea that knowledge is primitive. It is apt to seem to follow from that idea that there are no independently specifiable criteria by which to discern whether or not someone knows. And in that case, we can’t account for our knowledge of whether or not someone knows by appeal to our knowledge of whether or not they satisfy such criteria. A second source of pressure arises from the Oxford Realist treatment of the idea that knowing is a state of mind and, in particular, the Accretion. For by making the reflective subject decisively authoritative about whether or not they know, the Accretion problematizes the idea that other people might have access to ways of determining whether or not the subject knows other than via recourse to the subject’s avowals of their own reflective view about whether or not they know. The first two sources of pressure sponsor a third: the threat of dogmatism. Suppose, first, that reflective subjects can be decisively authoritative about whether or not they know and, second, that there are no independently discernible conditions to which appeal can be made in attempting to show that subjects are wrong to take themselves to know. In that case, it would be hard to see how a challenge could be mounted to someone’s sincere and reflective claim to know. The situation would be akin to one in which a subject reflectively avows that they believe something. Typically, the most that one would achieve by challenging such an avowal would be to encourage the subject to reflect again. So questions about how we can know whether other people know arise naturally from engagement with Oxford Realism. (R. G. Collingwood’s central complaint about Oxford Realism was that it enabled dogmatism. (Collingwood 1939.) The threat was live: Cook Wilson thought he knew that non-Euclidean geometry is impossible. Cook Wilson 1926: 456, 561. See also Prichard 1950: 99–100.) As Austin in effect observes, the Oxford Realist denial that knowing can be reconstructed out of other materials—that is, their claim that there are no independently specifiable sufficient conditions for knowing—leaves open that there are independently specifiable necessary conditions on knowing. One such condition to which the Oxford Realists make explicit appeal is that if one knows, then one can’t be wrong. (See e.g. Cook Wilson 1926: 69; Prichard 1950: 88.) Opining falsely is incompatible with knowing. It is to that condition that Austin points in his comment to the effect that talk about knowing functions differently from talk about believing. Whilst falsity is no immediate threat to the claim that someone believes, it excludes the claim that they know. So, if it is belief’s compatibility with falsity which is supposed to ground the idea that belief is merely subjective, and which in turn grounds the idea that thinkers are decisively authoritative about what they believe, then that ground is unavailable with respect to knowing. Austin’s central project involves seeking to discern further necessary conditions on knowing, as revealed in our ordinary handling of 9 challenges to a thinker’s claims to know. A typical challenge might begin with the question how the thinker knows, raised with the aim of having the thinker reveal sources of their standing, with those revelations potentially giving rise to further challenges. Thus, for example, If you have asked ‘How do you know it’s a goldfinch?’ then I may reply ‘From its behaviour’, ‘By its markings’, or, in more detail, ‘By its red head’, ‘From its eating thistles’. (Austin 1946: 154; 1979: 83) Austin focuses on challenges to the presumption that the thinker possesses a standing equivalent to proof: You may object:... But that’s not enough: plenty of other birds have red heads. (Austin 1946: 155; 1979: 93) If the thinker’s standing were exhausted by their awareness of the bird’s red head, then—according to the challenger—their standing would be compatible with falsehood, and so could not amount to proof. Superficial appearances notwithstanding, we can see Austin as pursuing questions about our knowledge of other minds throughout “Other Minds”. In doing so, he seeks to discern necessary conditions on knowing. In the following section, we’ll consider some of the conditions that he discerns. 4. Reasons for Doubt. The question at issue is whether some particular thinker knows some particular fact. The answer is pursued by a challenger who seeks to determine whether the claimant meets necessary conditions on possessing proof: It is in the case of [this] objection that you would be more inclined to say right out ‘Then you don’t know’. Because it doesn’t prove it, it’s not enough to prove it. (Austin 1946: 155; 1979: 84) One natural reading is the following. It is a necessary condition on knowing that a goldfinch is present that one’s standing is equivalent to proof. It is a natural consequence of that condition that one who knew would be in a position to articulate that standing, so as to show that they had the required standing. And showing that one had the required standing would amount to providing proof. If the best that a thinker could do by way of articulating their standing didn’t amount to proof, then their incapacity would provide reason to think that they lacked proof and, so, failed to know. On the natural reading, it would be reasonable to expect that if a candidate knower were able successfully to navigate all appropriate challenges to their claim to know, then their doing so would constitute a conclusive defence of their claim. That is a perspective from which it is natural to read Austin’s further comments on the obligations attending the candidate: (a) If you say ‘That’s not enough’, then you must have in mind some more or less definite lack. ‘To be a goldfinch, besides having 10 a red head it must also have the characteristic eye-markings’: or ‘How do you know it isn’t a woodpecker? Woodpeckers have red heads too’. If there is no definite lack, which you are at least prepared to specify on being pressed, then it’s silly (outrageous) just to go on saying ‘That’s not enough’. (b) Enough is enough: it doesn’t mean everything. Enough means enough to show that (within reason, and for presents intents and purposes) it ‘can’t’ be anything else, there is no room for an alternative, competing, description of it. It does not mean, e.g., enough to show it isn’t a stuffed goldfinch. (Austin 1946: 156; 1979: 84) My focus in this section will be (a). (We’ll return to (b) in the next section.) The central thought in (a) seems to be that there can be reason to doubt that a candidate knows something only if there is a reason for doubting that the candidate’s standing amounts to proof that is available to the challenger—that is, a reason that is specifiable, and so known. On the face of it, this represents a significant restriction on the candidate’s obligations. They are required to be in a position to respond to a challenge only if that challenge isn’t silly (outrageous); and a challenge meets that condition only if it is backed by reasons that are specific and available. Austin expands on the apparent restriction: The doubt or question ‘But is it a real one?’ has always (must have) a special basis, there must be some ‘reason for suggesting’ that it isn’t real, in the sense of some specific way, or limited number of ways, in which it is suggested that this experience or item may be phoney. (Austin 1946: 159) According to the natural reading, where no challenge that is appropriately grounded in available reasons is forthcoming, the candidate knows. By way of comparison, consider a similar idea expressed by J. M. Hinton: …the apparent rigorism which runs ‘I need more than there is here, before I stop saying that an item is not known’ is at the same time the apparent laxism, ‘I need no more than there is here, to make me admit an item “not p” as epistemically possible—not known not to be the case’. (Hinton 1989: 232) Believing that something is known and believing that something is not known should be treated symmetrically. Insofar as one should believe something only if one has reasons for believing it, one should believe that something is known only if one has reasons for believing that it is known. But for the same reason, one should believe that something is not known only if one has reasons for believing that it is not known. Thus, one should believe that a candidate does not know that a goldfinch is present—that from their perspective it is epistemically possible that it’s not the case that a goldfinch is present—only if one has reasons for believing it. Doubt and belief are equally reason hungry. Hinton’s proposal that belief and doubt be treated symmetrically doesn’t impose any substantive restriction on the obligations attending 11 candidate knowers. In particular, it doesn’t insulate their claims to know from defeat by countermanding reasons on grounds of the unavailability of those reasons. To what extent does Austin’s proposal impose such a restriction? We can pursue that question by considering a development of Austin’s proposal due to Adam Leite: If one recognizes that there is no reason in favour of some possibility that would undermine one’s authority, competence, or reliability regarding a certain domain, then (other things being equal) one may reasonably believe things in that domain even if one lacks adequate independent grounds for believing that the possibility does not obtain, and one may reasonably dismiss as groundless the suggestion that it does obtain. (Leite 2011: 94) If one recognizes (and so knows) that there are no reasons for believing that not-p is epistemically possible, then there are no reasons for believing that not-p is epistemically possible; and in that case, one may reasonably believe p. Although Leite doesn’t present his proposal as covering the case of knowledge, a natural extension would be the following. If one recognizes that there is no reason in favour of any possibility the obtaining of which would undermine a claim to know, then one is in a position to dismiss as groundless any objections to that claim, and so to endorse the claim. Indeed, that would be a way of possessing proof. Leite’s proposal seems close to Hinton’s plea for symmetry and, to that extent, not to impose substantive restrictions on candidate knowers’ obligations. However, Leite’s applications of the proposal reveal that, as he understands it, it differs from Hinton’s. Leite seeks to apply his proposal in order to undercut forms of skepticism that are based on an alleged inability to know that one isn’t dreaming. Leite considers two types of case that might be thought to sponsor a threat to knowledge based on one’s current experience. The first type of case is that of a dream that is phenomenologically indistinguishable from wakeful experience. Here, the threat to knowledge is supposed to arise because one lacks positive reasons for believing that one isn’t currently suffering such a dream. If it were an epistemic possibility that one is currently suffering such a dream, then it is plausible that its being so would preclude one from exploiting one’s current experience in order to know things about one’s environment. Hence, the fact that one has—and as a matter of principle can have—no positive reasons for excluding that epistemic possibility seems to undercut one’s claim to know. However, the nature of such a dream not only rules out one’s acquiring positive reasons for thinking that one is not suffering one. It also rules out one’s acquiring reasons for thinking that one is suffering one. Thus, on reflection, one can recognize that one can never be apprised of reasons for believing that one is now suffering such a dream. And now, according to Leite, his proposal operates to deliver the result that the putative epistemic possibility can reasonably be dismissed as groundless. (Leite 2011) Thus, in serving to exclude what might otherwise have seemed a threatening possibility, Leite’s proposal is shown to impose a substantive restriction on the obligations of knowers. 12 Given the substantive restriction on the obligations of knowers, admissible challenges to claims to know must be grounded on facts or reasons that are themselves available, and so knowable. Thus, the only type of dream the possibility of which could pose a threat to one’s claim to know would be a dream which was phenomenologically distinguishable from one’s actual experience, and that is the second sort of case that Leite considers. In this case, he allows that we have some reasons for believing that we might now be suffering such a dream since we have, after all, suffered such dreams in the past. However, we can in principle establish that we are not suffering such a dream, by exploiting their phenomenological distinguishability from wakeful experience. We therefore have the resources to exclude the possibility. Unlike Leite’s initial proposal, his applications rely on restricting admissible challenges to those grounded on available reasons. That suggests that the transition from proposal to applications involved slippage. On closer inspection, that is what we find. This can be seen by comparing the following principles: 1. 2. 3. If one recognizes that there are no reasons that support a possibility, then the possibility may be dismissed. If one doesn’t recognize that there are reasons that support a possibility, then the possibility may be dismissed. If one recognizes that one can’t recognize that there are reasons that support a possibility, then the possibility may be dismissed. Leite’s initial proposal is a version of principle 1. However, his applications rely not on one’s recognizing that there are no reasons that support a possibility, but rather on one’s recognizing that one cannot recognize that there are such reasons. Thus, Leite’s applications rely on a version of principle 3., rather than 1. We can most clearly see the difference between those principles by reflecting on the difference between principles 1. and 2. 1., as we observed above, relies on one’s recognizing, and so knowing, that there are no reasons that support a possibility. Since knowing is factive, that principle delivers that there are no such reasons. By contrast, principle 2. relies only on one’s failure to recognize any reasons that support a possibility. That would entail that there are no such reasons only on the assumption that we are incorrigible with respect to the geography of reasons. Having noted the weakness of principle 2., by contrast with 1., it becomes obvious that principle 3. represents no significant advance. For again, one’s recognizing that one is debarred from recognizing reasons in support of a possibility leaves open that that is due to limitations on our ability to track whatever reasons there are, rather than to the absence of such reasons. Thus, principle 1. is quite plausible, but is of no assistance in dealing with, for example, phenomenologically indistinguishable dreams. In order to exploit that principle, one would have to recognize, and so know, that there are no reasons for believing that one is suffering such a dream. By contrast, principle 3. might sustain the dismissal of the possibility that one is suffering such a dream, but only at a cost to plausibility. For meeting its condition is consistent with getting things wrong, for reasons of which one is unaware even in principle. (That is 13 the central message of Barry Stroud’s use of Thompson Clarke’s example of the plane spotters. Stroud 1984: 67–74; Clarke 1972.) The upshot is that principles 2. and 3. impose no plausible restriction on the obligations of knowers, since one can adhere to the restriction and yet fail to know. And principle 1., whilst plausible, imposes no substantive restriction. With that information in hand, let’s return to consideration of Austin’s proposal. Is it a version of the insubstantial principle 1., or is it rather a version of the substantial but implausible principle 3.? It is neither. It can’t be the second because the gap between a thinker’s being unaware, even in principle, of reasons that would undermine their claim to know and, on the other hand, there being no such reasons, is too large and too obvious to have escaped Austin’s attention. And it can’t be the first because Austin builds into his proposal that operative reasons must be available. But if Austin’s proposal isn’t to be understood in either of those ways, how is it to be understood? At the outset, I suggested that Austin’s proposal seems to impose a substantive restriction on the range of challenges that a knower should be in a position to address. And I suggested that his proposal seems to allow that someone able to answer all such challenges would thereby count as knowing. If we adhere to both of those suggestions, we are forced to read Austin as proposing a restriction on what knowers should be in a position to do that is either insubstantial or implausible. Since Austin explicitly endorses the first suggestion, we should reject the second: it is no part of Austin’s proposal that being able to deal with all available reasons for doubt is a sufficient condition on knowing. John McDowell provides a useful model for the proposed reading of Austin. McDowell presents a view of the standing, and obligations, of knowers comprising three main claims. First, knowing p is a matter of possessing reasons that suffice for truth. (Compare Oxford Realism (2) and (3).) However, second, one can meet that condition without knowing specifiable, non-question-begging reasons that suffice for that truth. (Compare Oxford Realism (1).) For example, one’s possession of reasons for holding p might be constituted by one’s seeing p or one’s remembering p, where those standings could not be specifiable independently of appeal to one’s knowing p. Appeal to those standings would therefore beg the question against a challenge to one’s claim to know. One might possess reasons that don’t beg the question in that way—for example, facts about the way things look to one. But there is no general expectation that it will be possible to reconstruct sufficient conditions for truth on the basis of those reasons. What is the role of those reasons, if not to constitute one’s standing as a knower? McDowell’s third claim addresses that question. Knowing requires that one is doxastically responsible. That is, it requires that one is appropriately sensitive to non-question-begging reasons for or against the truth of what one knows. (McDowell 1994.) As I’ll explain, that necessary condition on knowing is central to the proposed reading of Austin. McDowell presents his model in the following passage: There is a completely cogent argument from the fact that someone, say, sees that things are thus and so to the conclusion that things are thus and so. But that argumentative transition 14 cannot serve to explain how it is that the person’s standing with respect to the fact that things are thus and so is epistemically satisfactory…. Genuinely mediated epistemic standings, on the conception I have in mind [—that is, the governing conception that McDowell rejects—], would have to consist in the cogency of an argument whose premises do not beg the relevant question of epistemic standing “If one’s takings of things to be thus and so are to be cases of knowledge, they must be sensitive to the requirements of doxastic responsibility. We could not conceive remembering that things are thus and so, say, as a standing in the space of reasons if a subject could count as being in that position even if he were not responsive to the rational force of independently available considerations—the material to which the governing conception appeals. But we can separate that point from the idea that one can reconstruct the epistemic satisfactoriness of the standing in terms of the rational force of those considerations. (McDowell 1994: 416) Austin endorses the proposed distribution of epistemological labor. In accord with Oxford Realism, he views knowing as possession of conclusive reasons. However, in order to have access to such reasons, one must be appropriately sensitive to independently specifiable, nonquestion-begging reasons. Crucially, that obligation is limited. Although one must be able to address appropriate challenges to one’s claim to know, such challenges must be based upon reasons that are both specific and available. As we saw, the limitations that Austin proposes are implausible if taken to exhaust the obligations on knowers. However, when construed as limitations on the demands of doxastic responsibility, where the latter is a necessary but not sufficient condition on knowing, they appear defensible. In the next section, the proposed reading of Austin is extended in order to encompass part (b) of the target passage. As we’ll see, this allows for a reading of that passage which is less dramatic, but more plausible, than an alternative reading due to Mark Kaplan (2011). 5. Enough is Enough. In the previous section, I suggested that we read Austin as aiming to characterize only a necessary condition on knowing, according to which knowing requires appropriate sensitivity to non-question-begging reasons for or against what one claims to know. The conditions on doxastic responsibility are weaker than the conditions on knowing, since knowing requires possession of conclusive reasons and, so, reasons that, in effect, beg the question against challenges. In this section, the proposed reading is applied to Austin’s claim, in (b), that enough is enough: that meeting the obligations that attend knowing that there is a goldfinch requires being in a position to do enough to show that “there is no room for an alternative, competing, description of it. It does not mean, e.g., enough to show it isn’t a stuffed goldfinch.” (Austin 1946: 156; 1979: 84) One might read the passage as suggesting that the description of something as stuffed needn’t compete with its description as a goldfinch, because it’s possible for something to be both. (Austin was 15 skeptical about the idea that claims about ordinary things carry a determinate range of entailments: 1946: 159–161; 1979: 88–89; 1962: 118– 124.) Another way of reading (b) would be as allowing that one might be in a position to show that something is a goldfinch without (yet) being in a position to show that it isn’t stuffed because one hadn’t realized that being a goldfinch entails not being stuffed. Kaplan proposes a more tantalizing reading of the passage. On Kaplan’s reading, Austin’s proposal is that one might be in a position to meet sufficient conditions on knowing that something is a goldfinch, and also that its being a goldfinch entails that it isn’t stuffed, and yet not be in a position to know that it isn’t stuffed. (Kaplan 2011) Kaplan’s reading can be developed via the sixth component of Oxford Realism, the appeal to a state of being under the impression. On this reading, knowing that there is a goldfinch, and that its being a goldfinch entails that it isn’t stuffed, would impose an obligation to hold true that it isn’t stuffed. However, one can meet that condition without knowing that it isn’t stuffed. For one might hold it true unreflectively in being under the impression that it isn’t stuffed. (Kaplan 2011: 60–72) On the assumption that Austin follows the Oxford Realists in identifying knowledge with possession of proof, Kaplan’s reading would have Austin allowing that one can have proof that there’s a goldfinch, and that that entails that it isn’t stuffed, without being able to convert those proofs into proof that it isn’t stuffed. But if one’s purported proofs leave it open whether it’s stuffed, then one might reasonably doubt that what one has are proofs. It’s imaginable that one might have done enough to show that something is a goldfinch, and that its being so entails that it isn’t stuffed, without having done anything further to show that it isn’t stuffed. But one might wonder why Austin would have accepted the demand that showing that something isn’t a stuffed goldfinch must proceed independently of showing that it is a goldfinch. Furthermore, and as noted by Leite, Kaplan’s reading is in tension with Austin’s explicit animadversions about assumptions: To say…that we are making assumptions and taking things for granted whenever we make an ordinary assertion, is of course to make ordinary assertions look somehow chancy.... (Austin 1962: 138; Leite 2011: 91, fn.24) Austin’s general target is the idea that one isn’t always in a position to know all the things that are entailed by things that one takes oneself to know. In that case, Austin suggests, ordinary claims to know would look somehow chancy, since those claims would be dependent on things about which one was strictly ignorant. One’s standing with respect to the things that one claims to know would appear no better than one’s standing with respect to the mere assumptions on which those claims to know depend. Austin’s response casts doubt on the conjunctive claim (i) that there are things entailed by things that one takes oneself to know and (ii) that one is confined to assuming those things. Since Kaplan’s reading relies upon Austin’s endorsing that conjunction, the reading is undermined. We’ve seen, in the previous section, that an alternative interpretation of Austin’s project is available. According to the alternative, Austin’s discussion of what a knower must be in a position 16 to show is concerned not with the totality of a knower’s reasons, but only with those that sponsor their doxastic responsibility. Knowing requires being appropriately sensitive to non-question-begging reasons for or against what one claims to know. In this case, it requires being appropriately sensitive to non-question-begging reasons for or against the claim that there is a goldfinch. However, one’s non-question-begging reasons needn’t furnish proof that there’s a goldfinch. So, although a proof that there is a goldfinch would be convertible, inter alia, into a proof that it isn’t stuffed, the reasons that figure in that proof are liable to beg the questions whether it’s a goldfinch and, thus, whether it’s stuffed. Alternatively, one’s non-question-begging reasons for holding that there is a goldfinch can leave open whether it’s a goldfinch—for example, by leaving open whether it’s stuffed. Thus, we can find passage (b) intelligible without reading Austin as allowing that someone could know that there is a goldfinch, and that that entails that it isn’t stuffed, whilst being incapable of knowing that it isn’t stuffed. 6. Conclusion. I’ve suggested that that we can better understand some otherwise puzzling aspects of “Other Minds” if we read that work against the background of Oxford Realism. First, we can discern evidence that Austin agreed with the Oxford Realists in viewing knowing as a mental state. In that way, we can see the whole of Austin’s essay as addressing its titular topic. Second, we can discern evidence that Austin agreed with the Oxford Realists in viewing knowing as primitive. In that way, we can see him as seeking to uncover necessary conditions on knowing—including, especially, the requirement of doxastic responsibility—without treating those conditions as elements in a conjunctive reconstruction of knowing. And that, in turn, makes space for a plausible reading of some of Austin’s tantalizing pronouncements. References. Austin, J. L. 1946. “Other Minds.” Proceedings of the Aristotelian Society, Suppl. Vol. 20: 148–187. Reprinted in his 1979. Austin, J. L. 1962. Sense and Sensibilia. Reconstructed from the manuscript notes by G. J. Warnock. Oxford: Oxford University Press. Austin, J. L. 1979. Philosophical Papers. 3rd edn., J. O. Urmson and G. J. Warnock eds. Oxford: Clarendon Press. Clarke, T. 1972. “The Legacy of Skepticism.” The Journal of Philosophy 69, 20: 754–769. Collingwood, R. G. 1939. An Autobiography. Oxford: Clarendon Press. Cook Wilson, J. 1926. Statement and Inference. A. S. L. Farquharson ed. Oxford: Clarendon Press. Gustafsson, M. and Sørli, R. 2011. The Philosophy of J. L. Austin. Oxford: Oxford University Press. Hinton, J. M. 1989. “Scepticism: Philosophical and Everyday.” Philosophy 64, 248: 219–243. Kaplan, M. 2011. “Tales of the Unknown: Austin and the Argument from Ignorance.” In M. Gustafsson and R. Sørli eds. 2011: 51–77. 17 Leite, A. 2011. “Austin, Dreams, and Scepticism.” In M. Gustafsson and R. Sørli eds. 2011: 78–113. Marion, M. 2000a. “Oxford Realism: Knowledge and Perception I.” British Journal for the History of Philosophy 8, 2: 299–338. Marion, M. 2000b. “Oxford Realism: Knowledge and Perception II.” British Journal for the History of Philosophy 8, 3: 485–519. McDowell, J. 1982. “Criteria, Defeasibility, and Knowledge.” Proceedings of the British Academy 68: 455–479. Reprinted in his 1998: 369–394. McDowell, J. 1994. “Knowledge by Hearsay.” In B. K. Malital and A. Chakrabarti eds. Knowing from Words: Western and Indian Philosophical Analyses of Understanding and Testimony. Dordrecht: Kluwer: 195–224. Reprinted in his 1998: 414–443. McDowell, J. 1998. Meaning, Knowledge, and Reality. Cambridge, MA: Harvard University Press. Moore, G. E. 1905–6. “The Nature and Reality of Objects of Perception.” Proceedings of the Aristotelian Society 6: 68–127. Prichard, H. A. 1919. “John Cook Wilson”. Mind 28, 111: 297–318. Prichard, H. A. 1968. Moral Obligation and Duty and Interest: Essays and Lectures. Oxford: Oxford University Press. Prichard, H. A. 1950. Knowledge and Perception: Essays and Lectures. Oxford: Clarendon Press. Soteriou, M. 2016. Disjunctivism. London: Routledge. Stroud, B. 1984. The Significance of Philosophical Scepticism. Oxford: Clarendon Press. Travis, C. 2005. “A Sense of Occasion.” The Philosophical Quarterly 55, 219: 286–314. Travis, C. and Kalderon, M. E. 2013. “Oxford Realism.” In M. Beaney ed. The Oxford Handbook of the History of Analytic Philosophy. Oxford: Oxford University Press: 489–517. Williams, B. 1978. Descartes: The Project of Pure Inquiry. London: Penguin. Williamson, T. 2000. Knowledge and Its Limits. Oxford: Oxford University Press. 18