Can You Seek The Answer To This Question?

Abstract

Plato articulates a deep perplexity about inquiry in ‘Meno's Paradox’—the claim that one can inquire neither into what one knows, nor into what one does not know. Although some commentators have wrestled with the paradox itself, many suppose that the paradox of inquiry is special to Plato, arising from peculiarities of the Socratic elenchus or of Platonic epistemology. But there is nothing peculiarly Platonic in this puzzle. For it arises, too, in classical Indian philosophical discussions, where it is formulated with great clarity, and analysed in a way that casts it in a new light. We present three treatments of the puzzle in Indian philosophy, as a way of refining and sharpening our understanding of the paradox, before turning to the most radical of the Indian philosophers to tackle it. The Indian philosophers who are optimistic that the paradox can be resolved appeal to the existence of prior beliefs, and to the resources embedded in language to explain how we can investigate, and so move from ignorance to knowledge. Highlighting this structural feature of inquiry, however, allows the pessimist philosopher to demonstrate that the paradox stands. The incoherence of inquiry is rooted in the very idea of aiming our desires at the unknown. Asking questions and giving answers rests on referential intentions targeting objects in a region of epistemic darkness, and so our ‘inquiry sceptic’ also finds structurally similar forms of incoherence in the pragmatics of interrogative discourse.

Notes

¹See Dominic Scott, who makes a similar observation [Scott 2006: 76–7, 83].

²Whether Socrates offers a reinterpretation, a correct interpretation, or a rather different version of Meno's questions is disputed. See M. M. McCabe [2009] for an insightful way of taking the two formulations to be distinct but complementary. Because the passages we will look at bear strong similarity with Socrates' formulation, this will be our primary focus in what follows.

³Contrast this interpretation with John McDowell's view that, as late as the Theaetetus, Plato himself believed there to be no other option but complete knowledge or complete ignorance—he ‘regard[s] the knowledge which is required for a thing to figure in one's judgement as an all-or-nothing matter’[McDowell 1973: 197]. Panagiotis Dimas [1996: 9–11] sees the Meno as an exercise in working out this distinction, but not one presupposed by Plato.

⁴Compare Vlastos' interpretation of Plato's solution in Vlastos [1994b].

⁵So Terence Irwin [1977: 139, 140] writes, ‘The slave and Socrates' other interlocutors discover the resources they need in their present beliefs … Plato's explicit distinction between knowledge and true belief disarms Meno's paradox and its attack on Socratic inquiry’.

⁶Of course, one might say this only proves that one needs more than beliefs to make progress in inquiry, for after all the elenchus is notorious for its inability to progress beyond aporia. Vlastos [1994b] and Irwin [1977: 69–70] both take the paradox, and its solution in the Meno, to be Plato's reflections on vindicating (or rejecting) Socratic Method in one way or another. Both, however, take pre-existing (true) beliefs to be sufficient to dissolve the paradox. Alexander Nehamas [1985] agrees that justifying elenchus is part of Plato's motivation—‘He uses the paradox not only in order to discuss serious epistemological issues, but also to resolve a number of dialectical difficulties to which Socrates’ practice had given rise' [ibid.: 8; cf. 14–16]; he disagrees, however, about the sufficiency of true beliefs in addressing the paradox [ibid.: 16–17].

⁷These will offer terms through which new information can be conceived and conveyed; cf. Michael Beaney's gloss: ‘Here, if ‘learning’ is to take place, some knowledge must indeed be presupposed, as Socrates himself argued when the general problem was first formulated’[Beaney 1996: 139].

⁸This is not, of course, to say that Plato does not offer any further resources for reflecting on the paradox. Plato remained deeply concerned with the paradox, throughout his epistemological writings. The point is only that he does not present an answer to the paradox, explicitly formulated as such—apart, perhaps, from ‘recollection’, which raises at least as many difficulties of its own.

⁹‘I do not insist on the truth of everything that I've said; only that we will be better and braver men if we search for what we know, than if we believe inquiry is impossible’[Meno 86b6–c1]. It is, of course, debated what exactly of 81a–86b Socrates intends to distance himself from; and there are those (Menn [2002: 221–2] is a recent case) who think that Plato at least meant recollection to be the literal solution to the paradox. If so, it is inadequate, for it explains neither de novo acquisition, nor the requisites for recovery now. For such reasons, Scott [2006: 80–14] takes recollection to be a solution, but only to the problem of discovery, recognition of something as the object sought, not to the problem of starting inquiry. Others take recollection to be a metaphor or analogy for some key aspect of inquiry—perhaps that it is rational to engage in inquiry [Dimas 1996: 27–30], perhaps that we have rational capacities and tendencies towards truth [Fine 2003: 62; 1988: 138–41], or perhaps that there is something besides exchanging words (namely, comprehension) that happens in the person and cannot be pointed to [Moravcsik 1971: 64–5].

¹⁰Other passages in which Plato discusses knowledge, judgement and method might be seen as implicit attempts to address the paradox; each of these brings with it interpretive difficulties.

¹¹‘… one can inquire even if one lacks all knowledge of the subject, for the slave has just done so. The slave can inquire, although he entirely lacks knowledge, because he has both true beliefs and the capacity for rational reflection and revision of his beliefs, and these are adequate for inquiry’[Fine 2003: 56].

¹²As Vlastos [1994b: 25] suggests.

¹³As argued by Dimas [1996: 9].

¹⁴‘All instruction given or received by way of argument proceeds from pre-existent knowledge … [sometimes from] knowledge of the particulars actually falling under the universal and therein already virtually known … [So that] before he was led on to recognition or before he actually drew a conclusion, we should perhaps say that in a manner he knew, in a manner not’ (Posterior Analytics I.1, 71a1–2, 18–19, 24–5; and at 71b5–9: ‘I imagine there is nothing to prevent a man in one sense knowing what he is learning, in another not knowing it. The strange thing would be, not if in some sense he knew what he was learning, but if he were to know it in that precise sense and manner in which he was learning it.’) Or consider also Moravcsik's claim [1971: 63, cf. 59] that ‘The recollection thesis’ solves the paradox because it ‘enables us to give a consistent and intelligible account of how the inquirer at the outset both knows and does not know what he is seeking’.

¹⁵Vlastos' focus on the importance of insight into logical relations suggests this [Vlastos 1994: 27; 1965: 156–7].

¹⁶See, for instance, Scott [2006: 77,83–7] and White [1974; 1976: 42–7]. For Gail Fine [2003: 60], agreed examples perform the work of initial specifications: ‘He does believe there is great dispute about the correct definition of virtue terms, and of course there is some dispute about particular moral cases. But there is also considerable agreement, enough agreement to secure the reference of the terms and so to ground inquiry’.

¹⁷‘The role of questioning in bringing this matching about is crucial: Plato's resolution of Meno's paradox is dialectical rather than logical’ writes Nehamas [1985: 24]. Presumably this means that he also takes the puzzle itself to be more a pragmatic difficulty than a logical one.

¹⁸As with most Indian sources, the text is difficult to date, and one must in any case distinguish between the dates of composition, compilation and redaction. Much of the sūtra literature is now thought to have achieved a relatively stable form by the first or second century CE, some strata of the texts invariably being of considerably greater antiquity.

¹⁹The term prasiddha is used several times in the texts we are discussing. A past participle from the verb sidh-‘to achieve, accomplish’ with the verbal prefix pra-, its common meaning is ‘renowned, famous, celebrated’ (German: ‘bekannt’). The sense is of something generally accepted or commonly agreed to be the case. The Sanskrit verb jñā‘to know, learn, find out, recognize’ is cognate with ‘know’; but the derived noun jñāna is used in philosophical Sanskrit with a meaning more akin to ‘belief’, the term employed for ‘knowledge’ being instead pramā.

²⁰‘dharma h prasiddho vā syād aprasiddho vā| sa cet prasiddho, na jijñāsyah| athāprasiddho, natarām | tad etad anarthakam dharmajijñāsāprakaranam |’[Śabara, Mīmāmsāsūtrabhāsya, inf. 1.1.1, 14,2–15,2].

²¹We have serious reservations in general about the ‘diffusion’ thesis, as defended most recently and elaborately by Thomas McEvilley [2002]. There are at least two alternatives to supposing an actual transmission of ideas: one is that the co-occurrence of similar ideas in Greece and India is due to their having a common origin in an Indo-European ‘protophilosophy’; the other is that the deepest philosophical problems are essentially ‘perennial’ or culture inspecific. For the first hypothesis, see the work of Nicholas J. Allen, for instance his review article [Allen 2005]; for the second see the work of B. K. Matilal, surveyed in the introduction to Matilal [2002]. See also Carmen Dragonetti and Fernando Tola [2004].

²²M. M. McCabe [2009] is a notable and rewarding exception. Dominic Scott argues that the paradox arises because of Plato's commitment to the principle of the ‘Priority of Definition’ and ‘Foreknowledge Principle’[2006: 84–7]. Gail Fine [2003] and Gregory Vlastos [1994a: 78–9] both likewise take the paradox to rest on the peculiarly Platonic requirement that one know what something is in order to know what it is like (Fine calls this ‘Priority of Knowledge What’—or PKW—in Fine [2003: 44–50]); Dimas [1996: 8–9] ties this feature even more closely to Plato by making formulation of the PKW principle in more tendentious terms the seed of the paradox. Irwin describes the paradox as one that arises in the context of Socratic inquiry, and must be disarmed if Socrates' idiosyncratic method of inquiry (including PKW and a distinctive Dialectical Requirement) is to be defended [Irwin 1977: 70, 138–9]. Moravcsik [1971] may be an exception to this, if he sees the unobservable ‘gap between the question and the response’[ibid.: 65] as an actual feature of the experience of inquiry, which anyone must explain in some way or other.

²³The Sanskrit vicārya‘having deliberated, considered, examined, discussed, investigated’, is a causative derivation from the verb vicar-‘to wander about, roam over’.

²⁴athavā arthavat dharmam prati hi vipratipannā bahuvidah| kecid anyam dharmam āhuh, kecid anyam | so ’yam avicārya pravartamānah kamcid eva upādadāno vihanyeta anartham ca rcchet | tasmād dharmo jijñāsitavyah| sa hi nihśreyasena purusam samyunakti iti pratijānīmahe |[Śabara, Mīmāmsāsūtrabhāsya, 16,3–16,6].

²⁵Irwin [1977: 138–9] also discusses the problem as one of ‘securing reference’, although he does not seem to think it is such a grave problem: ‘The answer to Meno's paradox[is as follows:] though the slave does not know, he has true beliefs about the questions discussed (85c6–7) … To inquire into x we need only enough true beliefs about x to fix the reference of the term ‘x’, so that when the inquiry is over, we can see we still refer to the same thing’[1977: 139]. Dimas argues, unconvincingly, that the matter of ‘securing reference’ should not be on the table at all: ‘We do not need to have any true substantive beliefs about x to be able to identify the reference of x. What we do need is the capacity to acquire such beliefs. Most importantly, however, it is false that we need to identify in any way at all the reference of something in order to be able to raise the question ‘What is ——?’ about it’[1996: 18].

²⁶According to Fine [2003: 59], Plato ‘seems to assume that some important true beliefs are better entrenched than are various false beliefs (or will seem more reasonable to us when we first consider them) so that, in cases of conflict, we tend, upon reflection, to reject the false beliefs'; recollection is brought in to explain not the possibility of inquiry but the curious fact that ‘in inquiring, we tend to favour true over false beliefs’[2003: 62]. See also Vlastos [1994b: 25–7].

²⁷Scott's ‘problem of discovery’ seems to turn on similar worries: ‘the assumptions included in the specification play a crucial role in determining the direction and outcome of the inquiry … Yet, unless you already know that the specification is correct, how can you know that this proposed answer is the right one, even if it happens to be?’[Scott 2006: 83]. Compare a similar worry in the Platonic context, raised by McCabe: ‘In order, then, for true beliefs to underpin inquiry, the status of those beliefs must itself be under scrutiny’[McCabe 2009:245].

²⁸prasiddhahśakyate jñātum prasiddhatvāt tu nesyate | aprasiddhas tv aśakyatvān natarām ityato ’bravīt || 124 ||[Kumārila, Ślokavārttika 57, 1–2].

²⁹svarūpādisu dharmasya dvividhā vipratipadyate | pūrvam pramānarūpābhyām pādenādyasya nirnayah|| 126 || sthite vedapramānatve punar vākyārthanirnaye | matir bahuvidām pumsām samśayān nopajāyate || 127 || kecid āhur asāv arthah, kecin nāsāv ayam tv iti | tannirnayārtham apy etat paramśāstram pranīyate || 128 ||[Ślokavārttika 57, 5–58, 4].

³⁰Vlastos is tempted by this sort of resolution to Meno's paradox when he writes that the paradox does not get off the ground ‘[s]o long as your inquiries concern exclusively moral questions … For moral terms hail from common speech, where their meaning is established long before you could undertake to encapsulate it in a Socratic definition’[Vlastos 1994a: 84].

³²tatpunar brahma prasiddham aprasiddham vā syāt | yadi prasiddham, na jijñāsitavyam | athāprasiddham, naiva śakyam jijñāsitum iti |[Śa[ndot]kara, Brahmasūtrabhāsya 78,1–79,2 (inf. 1.1.1)].

³¹It is Śa[ndot]kara's view that an inquiry into brahman has four prerequisites, which consist in the possession of intellectual virtues of discrimination, equanimity, detachment from ordinary pleasures, and desire for liberation. Rāmānuja, on the other hand, thinks that a study of Vedic ritual and Mīmāmsā is the prerequisite. See also Chari [2004: 152–7].

³³Dimas [1996: 22] makes much of this formulation, and the fact that it does not explicitly state the impossibility of inquiry. He cites this as support for his view that Meno's paradox is a worry about the rationality of inquiry, rather than the possibility of it: ‘Meno objects that it is impossible, not to conduct an inquiry, but to do so rationally, i.e. hoping reasonably to discover the answer. Let us call this the Rational Impossibility account of Meno's challenge and oppose it to the Conceptual Impossibility one’[1996: 19]. For reasons stated below, we do not believe his distinction between the Rational Impossibility and the Conceptual Impossibility accounts will go through.

³⁴See Ratnakīrti, Santānāntaradūsana 147–9.

³⁵The claim might be (1) that as a matter of convention, one does not normally conduct inquiries into what one knows, just as, as a matter of convention, one does not normally bathe when one is clean; or (2) that as a matter of psychological fact, when one thinks one knows, one does not inquire.

³⁶What holds for desire (for knowledge) holds also for intention (to seek it). Compare Donald Davidson [2005: 97]: ‘Donnellan explains that intentions are connected with expectations and that you cannot intend to accomplish something by a certain means unless you believe or expect that the means will, or at least could, lead to the desired outcome.’

³⁸ucyate—asti tāvad brahma nityaśuddhabuddham uktasvabhāvam sarvajñam sarvaśaktisamanvitam | brahmaśabdasya hi vyutpādyamānasya nityaśuddhatvādayo ’rthāh pratīyante | brhater dhātor arthānugamāt |[Brahmasūtrabhāsya 79,2–81,1].

³⁷Those who are impressed by it might point out that Socrates' first question about the slave, before he begins questioning him, is ‘Does he know Greek?’ There are obvious pragmatic reasons for this question. But Plato's choice to make it part of the conversation at all might suggest it has rather more importance. Perhaps acquaintance with a natural language gives one prior conceptual resources—a grasp of ‘same’, ‘different’, ‘larger’, ‘smaller’, in the case of the geometrical inquiry—which one needs in order to get inquiry off the ground. Inquiry implicitly draws on these conceptual frameworks built into our very acquisition of language. Compare the account of Vlastos [1965], which emphasizes how the slave must draw out inferences from concepts already grasped: ‘Reduced to its simplest terms … ‘recollection’ in the Meno is any enlargement of our knowledge which results from the perception of logical relationships… . [I]f the relations are intra-propositional … , then to ‘recollect’ is to gain insight into the logical structure of a concept, so that when faced with its correct definition one will see that the concepts mentioned are analytically connected’[1965: 156–7]. Vlastos, however, does not connect this to language acquisition; but see note 30, above.

³⁹For discussion of the Indian tradition of semantic analysis (the nirvacana tradition), see Eivind Kahrs [1998]. For a treatment of Plato's discussion of the same practice, see David Sedley [2003].

⁴⁰ Cf. White's preferred solution to Meno's paradox, in White [1976: 43–7], and in greater detail in White [1974]. According to White, an initial specification of the object sought suffices to dissolve the paradox. We will see reasons, especially in the discussion of Śrīharsa, to doubt the viability of this proposal.

⁴¹[sarvasyātmatvāc ca brahmāstitvaprasiddhah] | sarvo hy ātmāstitvam pratyeti, na nāham asmīti | yadi hi nātmāstitvaprasiddhah syāt, sarvo loko nāham asmīti pratīyāt |[Brahmasūtrabhāsya 81,1–2].

⁴²In fact, Śa[ndot]kara's monism permits him the thesis that there is only one object of inquiry, namely brahman. Vedāntadeśika, however, wonders if there can be inquiry into a single, undifferentiated, reality [Chari 2004: 94–5].

⁴³ātmā ca brahma |[Brahmasūtrabhāsya 81,2].

⁴⁶tadviśesam prati vipratipatteh| dehamātram caitanyaviśis tam ātmeti prākrtā janā lokāyatikāś ca pratipannāh| indriyāny eva cetanāny ātmety apare | mana ity anye | vijñānamātram ksanikam ity eke |śūnyam ity apare | asti dehādivyatiriktah samsārī kartā bhoktety apare | bhoktaiva kevalam na kartety eke | asti tadvyatirikta īśvarah sarvajñah sarvaśaktir iti kecit |ātmā sa bhoktur ity apare |[Brahmasūtrabhāsya 81,3–82,3].

⁴⁴yadi tarhi loke brahmātmatvena prasiddham asti, tato jñātam evety ajijñāsyatvam punar āpannam |[Brahmasūtrabhāsya 81,3].

⁴⁵As John Perry [1979] has pointed out, the pronoun ‘I’ is, in his phrase, ‘essentially indexical’, that is to say irreducible to any non-indexical expression.

⁴⁷The suggestion of a progression from worldly to refined theories of selfhood may remind one of Prajāpati's graded instruction to Indra in Chāndogya Upanisad 8.7–12.

⁴⁸It seems to be considerations of this sort that move Moravcsik to claim that ‘it is intuitively easy to see that one would not be bothered by the paradox except in cases of a priori inquiry’[1971: 56]. Only in the case of a priori inquiry, he thinks, will getting hold of the object itself be problematic.

⁴⁹The editor divides the text of Chapter 3 into eight paragraphs. In the first two paragraphs, Śrīharsa distinguishes four possible meaning of the word kim, ‘what?’, the last of which is interrogative. What we are discussing now is the argument in paragraphs 3–5, that it is impossible on dialogical grounds to use kim to ask a question. Paragraphs 6–8 press a second argument, that the prior knowledge needed to ground inquiry itself makes inquiry impossible. We discuss this second argument in §5.

⁵⁰This is one of a pair of rules Śrīharsa quotes as ślokas from earlier, unidentified, sources; the other is introduced below as I′. The principle is first appealed to on page 555, but only stated in this complete form later. The Sanskrit for both principles is: atra ca sa[ndot]grahaślokau “yathāvidham yam visayam nijasya praśnasya nirvakti paro yathoktyā| vācyas tathaivottaravādināpi tathaiva vācā sa tathāvidho ’rthah| praśnasya yah syād visayah sa vācyo vācānayaivaisa bhaven niruktah| idam tvayāpyāsthitam etayaiva girā svaprcchā visayasya vaktrā”||[Śrīharsa, Khan danakhan dakhādya 557,7–10].

⁵¹We might recall Grice's ‘conversational maxims’, in particular, the maxim of Relation [Grice 1975].

⁵²Even if a questioner's use of a particular term must be corrected, it can only be so if the respondent first fixes the object of his reply as the same as that of the question. Problems that arise in such cases—as when someone, looking at a dolphin, asks ‘What kind of fish is that?’—will be discussed in the following section.

⁵³See note 50 for Sanskrit text.

⁵⁴Compare G. E. R. Lloyd's worry, in the context of learning from other cultures, that ‘We may be at a loss to explain, in general terms, how such learning can occur, how new insights into underlying ontological questions can be gained. It may seem that it cannot happen, as if either other ideas will be reduced to our own, or they will remain forever unintelligible’[Lloyd 2004: 9]. He replies that 1) we do in fact learn in such circumstances, and 2) there is no difference between such learning circumstances and what we might think of as the usual ones children find themselves in. But neither reply seems very satisfactory: the first simply reiterates the paradox to be explained; the second makes all such learning haphazard—young children, we might think, to a large extent just assimilate whatever is ‘around’; their learning is not targeted. But such undirected activity is not inquiry at all. Śrīharsa's final tongue-in cheek invitation to his interlocutor to become his devotee seems to be recognizing this very point.

⁵⁵praśnārthāt khalu kimśabdāt kasyacit padārthasya jijñāsyamānatā pratīyate, sā ceha pramānapadasamabhivyāhārāt pramānavisayinī pratīyate. yadvisayaś ca praśnas taduttaravādinābhidheyam tad ayam praśna īśvarasadbhāve pramānasāmānyavisayas tadviśesavisayo vābhipretah ? ādyaś ced īśvarasadbhāve pramānam ityevottaram āpadyeta, yadvisayo hi praśnas tadabhidheyam, pramānasāmānyavisayaś ca praśnah tac ca pramānaśabdenābhihitam eva | atha dvitīyah tathāpīśvarasadbhāve pramānam ityevottaram āpadyeta | yathā praśnavākye pramānaśabdo viśesaparah tathottaravākye ||[Khan danakhan dakhādya 555, 6–11].

⁵⁶kāsāvasādhāranī pramāna-vyaktir iti praśnārthah, tatra tādrśyāh pramāna-vyakter abhidhānam uttaram yuktam naivamvidhāh pralāpāh|[Khan danakhan dakhādya 556, 10–12].

⁵⁷praśnārthāc ca kimśabdāj jijñāsāvisayatā 'rthasya pratīyate | jijñāsā ca jñātum icchā, icchā cājñāte na sambhavati atiprasa[ndot]gāt |[Khan danakhan dakhādya 557, 35–6].

⁵⁸In paragraphs 7 and 8, Śrīharsa considers and rejects two possible strategies for defending the possibility of inquiry without prior beliefs about the object of inquiry, one being that the question has a merely causal function, the other that it is asked only for the sake of argument. Neither defence seems promising.

⁵⁹Compare Plato, Philebus 35a–c, where either perception or memory must put us in touch with an object, if we are to conceive a desire for it.

⁶⁰Indicative: ‘in accordance with the facts, the true state, truth, reality’.

⁶¹tatra svajñānam icchākāranībhūtam vaktavyam tadyathābhūtārtham vā syād ayathābhūtārtham vā ? yathābhūtārtham cet tenaiva jñānena svakīyo visayah pramānam upasthāpyate, visaye pramānapravrttim antarena tadīyayathārthatvasya vaktum aśakyatvāt, [tenāpi pramānena svagocara īśvarasadbhāvah upāsthāpyata ityanāyāsenaiva siddhosmākam īśvarasiddhimanorathah|] athāyathārtham tattasminn ayathārthajñānavisaye yadyasmābhir apy ayathārtham eva jñānam utpādanīyam iti bhavatah prcchato vāñchitam tadā keya svādhīne ’arthe parāpeksā ? bhavānevāyathārthajñānotpādanakuśalo yathaikam tatra mithyājñānam ajījanat tathā’param apy utpādayatu | vayam punar yathārthajñānasyotpādayitāro mithyājñāne sarvathaivākrtinah kim iha niyujyemahi ? [Khan danakhan dakhādya 558,1–558,10].

⁶²Vlastos [1994a: 84], Irwin [1977: 138–9], Fine [2003: 60], and White [1976] all in one way or another take this point to be a key to resolving the paradox—as did Śabara and Śa[ndot]kara, above.

⁶³‘To be able to specify some sign (σημεĩν) whereby that thing can be differentiated from everything else’ as Theaetetus 208c has it. Socrates' objections to this possibility are similar to Śrīharsa's.

⁶⁴atha madīyasyāyathārthajñānasya yo visayah sa madīyayathārthajñānavisayo bhavatā kriyatām iti ? tvadīyam vāñchitam tadā vyāghātādīdrśyarthe bhavatah pravrttir evānupapannā, śuktikā rajatātmatvena mama yathārthajñānavisayo bhavatvityetadartham preksāvān kathañkāram prayateta ? yena rūpenāyathārthajñānavisayatvam tena rupena yathārthajñānavisayatve vyāghātāt ||[Khan danakhan dakhādya, 558,10–558,14].

⁶⁵Tyler Burge's famous ‘arthritis’ thought experiment [Burge 1979], one point of which is to show that thought can have an object even if the concepts in play are under-specific and even erroneous, therefore does not bear upon the present discussion. For although a person whose conception of arthritis is woefully inadequate may indeed be thinking of arthritis, they are not in a position to formulate an epistemic project that has as its goal the improvement of a conception acknowledged to be inadequate. One cannot rationally think to oneself: ‘I am minimally competent in my conception of X, and let me seek to improve it with an improved competence.’

⁶⁶A desire, Śrīharsa says, must have some object. But, by definition, inquiry is for the unknown. Being unknown, it cannot be picked out as an object, and so cannot be an object of desire. ‘It is therefore unlikely that Meno's paradox is resolved by appealing to them [true beliefs] in order to secure, from the very beginning of the inquiry, reference to the object which the inquiry concerns’, Nehamas argues [1985: 23]. ‘These true beliefs are recovered by the slave at the end of his examination by Socrates; they could not therefore play the identificatory role Irwin asks of them, and which requires them to be there consciously at its very beginning’[ibid.: 16].

⁶⁷Ga[ndot]geśa, in his seminal post-Śrīharsa work on epistemology, reaffirms the principle that inquiry requires a target. He says: ‘A cognition (jñāna) of an object [to which contrary alternatives are attributed] is a necessary condition for doubt. Otherwise, there could be neither regulation by the object, in a doubt, nor the possibility of an imbalance [in epistemic weight] between the alternatives.’ (dharmijñānam ca samśayahetuh| anyathā samśaye dharmi-niyamah kotyutkatatvañca na syāt) [Ga[ndot]geśa, Tattvacintāmani 199].

⁶⁸Beaney's tussle with this problem on behalf of Frege casts it in terms of problems with ‘analysis’, or definition [Beaney 1996: 138–50]. ‘The real problem arises, however, when we do have some grasp of the meaning of ‘A’[that which we want to inquire into or, here, analyse]. If the analysis is to be informative, must not ‘B’ possess a different meaning? …’[1996: 139]. Frege's own attempt at a solution, Beaney writes, involved attributing the ‘same content’ to starting- and end-point, but “‘split up” differently’[1996: 139], so that we can think of coming to learn something that was not known already. Unfortunately, this solution, like most, seems not to be able to avoid the paradox's reduplicating itself at another level.

⁶⁹White remarks in passing that ‘inquiries … begin … with the possibility of distinguishing between the ability to say what they are seeking on the one hand, and their successful completion on the other.’ He does not make much of the importance of distinguishing between the two, nor does he see the tension between this and his claim that ‘it is the specification that defines the inquiry’[White 1976: 45, 46].

⁷⁰Again, consideration of Beaney's discussion of Frege's treatment of the problem is instructive: ‘If the original sense drops out of consideration, then no problem can arise in attempting to analyse it … But if no such judgement can be made, either where the senses are not obviously the same, or where the established sign has no clear sense at all, then we simply replace the old sign with a new sign, defined in the way we want, and hence bypass the question as to whether the senses are the same.’ This is obviously unsatisfactory, if what we are trying to do is formulate a rational inquiry. As Beaney concedes, ‘It might seem, at best, to evade rather than answer the paradox’[1996: 147].

⁷¹There are, as Moravcsik [1971: 53–4] points out, different ways to acquire intellectual skills; although one might, with Plato, doubt whether this deserves to be called learning [Nehamas 1985: 10–11].

⁷²Compare Aristotle's suggested solution, in note 14, above.

⁷³And as Fine [2007: 344] observes, ‘It is difficult, not easy, to say what criteria an adequate specification for fixing any given target must satisfy’.

⁷⁴Many thanks to Peter Sahota, and particularly the two anonymous readers for the AJP.