H. P. GRICE E J. L. SPERANZA: LA CONVERSAZIONE -- I VERBALI: RAIMONDI

 G.: Raimondi has landed us in Examinations Hall with Persian morpho-syntactic categories, which is more than I ever expected to owe the Medici press. S.: If the Hall can swallow it. G.: The Hall has swallowed worse. It once swallowed my views on meaning. S.: And expelled them in Schools. G.: Precisely. But Persian is another matter. Austin once called it useless, which was merely his way of saying he had not learnt it. S.: Or that it would not help him distinguish “voluntary” from “involuntary” before luncheon. G.: Yes. Still, the deep berths of lingo matter. Persian is Indo-European, or Indo-Germanic if one wants to irritate a modernist, and therefore a cousin to the Yorkshire tongue, though with better carpets. S.: Yorkshire will not thank you. G.: Yorkshire rarely does. But I mean the family resemblance seriously enough. One cannot talk categories in language without asking whether one is merely rediscovering old kinships under new labels. S.: Which is what Raimondi would have liked. G.: Raimondi liked many things too infrastructural for Oxford. Typefaces, grammars, Oriental corpora, channels of transmission — the whole man is a warehouse with a theological licence. S.: You say that fondly. G.: Only because he gives us something better than another dictionary of philosophical jargon. He gives us scripts, grammars, and the possibility of saying “my mother is quick” in several civilisations. S.: Let us begin there, then. “My mother is quick.” G.: In English, yes. In Latin one might say mea mater celer est or mea mater velox est, depending how one wishes to sound. In Greek, perhaps ἡ μήτηρ μου ταχεῖά ἐστιν, though one should check the adjective if one wants real speed rather than English bluff. S.: And in Persian? G.: Modern Persian would give us مادرم تند است or مادرم سریع است, though the first leans to temperament and the second to speed in the newer register. Better perhaps مادرم تیز است in some contexts, but then one is already disputing the semantics of “quick.” S.: Which is why categories matter. G.: Exactly. The sentence looks simple in English only because English is badly behaved in a manner we have learnt to tolerate. S.: And Proto-Indo-European? G.: That is where one begins to enjoy oneself irresponsibly. Something like méh₂tēr h₁ésti kʷékʷlos would be absurd, because I have just smuggled in wheel-like speed. Better simply to confess that the exact PIE for “my mother is quick” is a reconstruction too far for a class before eleven. S.: So you will not say mea mater est quicka? G.: Only as a joke, and a good one. “Mea mater est quicka” is what happens when Indo-European piety meets schoolboy forgery. S.: Then you admit there is pleasure in family resemblance. G.: Of course. Latin mater, English mother, Persian mādar, all cousins in one great domestic conspiracy. And “quick,” if not cognate, still belongs to the sort of semantic field that makes one ask how languages carve attribute from action. S.: Which is where morpho-syntactic categories appear. G.: Yes. Raimondi’s thought, or one way of using him, is that one cannot speak of language philosophically without looking at the deep joints: noun, proper noun, substantive noun, verb, adjective, relation, agent, patient, action, all that old Aristotle-and-grammar business. S.: NOMEN, NOMEN PROPRIUM, NOMEN SUBSTANTIVUM. G.: Exactly. The old schoolroom parade. Better than half the newer taxonomies, because at least it knew it was half logic and half declension. S.: Does it apply to Persian? G.: And to Persian cats, if they could speak — though we might not understand them, which is a useful check on philosophy. S.: Let us try the cat. G.: Very well. “The Persian cat sat on the mat.” English, embarrassingly plain. Latin: cattus Persicus in storea sedit, though the Roman would complain about cattus and storea as if they were both latecomers. Greek would need some schoolroom barbarism because the cat itself is already a traveller. Persian: گربهٔ ایرانی روی حصیر نشست. S.: And Hebrew? G.: Something like הַחָתוּל הַפָּרְסִי יָשַׁב עַל הַמַּחְצֶלֶת. Which at once reminds us that Semitic patterns do not line up politely with our Indo-European smugness. S.: So Hebrew and Persian do not compare? G.: They compare magnificently, but not by inheritance of the same sort. The point is not to rank them but to see where categories match, fail, shift, or overperform. S.: For example? G.: The relation between noun and adjective, possession, definiteness, copula, and word order. English says “my mother is quick” with shameless simplicity. Persian often omits the explicit copula in some contexts, though not here in formal prose. Hebrew treats predication differently again. One sees at once that “being” is a problem of grammar before it becomes a problem of ontology. S.: There is your Aristotle again. G.: He never quite leaves. The Categories are really about what can be said in one breath of a subject, or at least that is one decent way of reading them. Substance, quantity, quality, relation, place, time, position, state, action, passion. S.: Which Cicero turns into quantity, quality, relatio, and the rest. G.: Yes, and thereby Rome at least had the grace to make them pronounceable. Then Kant arrives and multiplies necessity beyond appetite into quantity, quality, relation, modality, each with three subforms, to make twelve. S.: Three by four: twelve. The theological look of arithmetic. G.: Exactly. Quantity, quality, relation, modality, each triad behaving as if Königsberg had annexed the schoolroom. S.: Give me Persian any day. G.: My point exactly. If one wants to know whether morpho-syntactic categories are real, look not only at Kant’s table but at Persian, Hebrew, Latin, Greek, and English in use. S.: And what does Raimondi actually care about? G.: Transmission, first. Typefaces, corpora, Arabic Euclid, Avicenna, grammars of Hebrew and Chaldean, Persian manuscripts. He is a philosopher only in the older, larger sense in which making meaning portable is already a philosophical labour. S.: A macro-Gricean, then. G.: That is too flattering. Better to say that he constructs the conditions under which any micro-Gricean act of understanding can scale beyond a room. S.: Which is not nothing. G.: Certainly not. Without presses, grammars, copied scripts, and men patient enough to compare lexical habits, there is no Hall for us to preen in. S.: You are in a charitable mood. G.: Persian does that. It obliges one to acknowledge civilisation outside Oxford. S.: Let us return to categories. How many do you want? G.: Fewer than Kant and more than a railway timetable. Let us say the old grammatical-philosophical essentials: noun, proper noun, substantive, adjective, verb, pronoun, relation-marker, quantifier, modality-marker, perhaps copula if one insists, and then the inferential devices that tie propositions together. S.: “If,” for example. G.: Ah yes. The beloved “if.” Persian اگر, Latin si, Greek εἰ, English if. One sees at once both kinship and divergence. The connective is not merely logical; it is historical, social, and deeply irritating. S.: To Strawson too. G.: Of course. Strawson would say the horseshoe is not the vernacular if, and he would be right in his superior way. S.: Then how does Persian help? G.: It reminds one that a language can parcel agency, possession, predication, and relation with a surface economy that makes English look both elegant and careless. Persian’s ezafe alone should teach humility to analysts. S.: Explain. G.: The ezafe marks noun-adjective and noun-genitive relations with a little linking vowel and a great deal of civilising labour. It ties words in ways that our old school categories must notice or become ridiculous. S.: So “the Persian cat” is not merely noun plus adjective. G.: Exactly. It is often a linked construction: گربهٔ ایرانی — cat-e Persian. The relation is audible and graphic in a way English usually leaves to adjacency. S.: Then Aristotle’s categories are not enough. G.: Not enough, no, but still useful as the first theatre. One begins with substance and quality, perhaps, then learns that languages implement those distinctions through very different morpho-syntactic arrangements. S.: Which sounds less metaphysical than grammatical. G.: As it should. One of the oldest mistakes in philosophy is to mistake a grammatical convenience for a metaphysical revelation. S.: Such as “is.” G.: Such as “is,” yes, and “has,” and perhaps “quick.” S.: You are enjoying that word too much. G.: Because it behaves like a small Anglo-Saxon tyrant. “My mother is quick” can mean swift, lively, perceptive, alive, pregnant in older registers, all under one curt monosyllable. S.: Persian will not let you get away with that. G.: Precisely. Nor Latin, if properly taught. Velox, celer, acer, vividus — one must choose the shade. Grammar and lexicon together prevent philosophical laziness. S.: Then your class in the Hall is really a rebuke to English. G.: Every decent class is. S.: You said Persian was possibly Indo-European like Yorkshire. G.: I said it to provoke, but yes, in the large family sense. Persian is Indo-Iranian within Indo-European. Yorkshire is merely English with weather. Yet the comparison is useful because it reminds one that kinship in roots does not entail identity in categories. S.: So mother, mater, mādar, μήτηρ, all cousins, but the categories that frame them may differ. G.: Exactly. In one language possession is straightforwardly pronominal, in another enclitic, in another attached, in another a genitive relation. The old philosophers who thought grammar transparently mirrored the world deserved some Persian. S.: Or some Hebrew. G.: Hebrew too, though Semitic gives a different lesson. Root-pattern morphology alone is enough to unsettle any complacent Greek notion that the noun and verb parade one by one in tidy robes. S.: Yet Aristotle was not entirely wrong. G.: Certainly not. He saw that there are recurrent ways in which language says being, quantity, quality, relation, and so on. He was wrong only where later people treated his list as final and universal. S.: Which Kant then multiplies from XXX to 121212. G.: Yes, with four headings and three under each: quantity, quality, relation, modality. It is all very German and very overupholstered. S.: Give me Persian any day. G.: That is becoming our refrain. S.: It deserves to. But tell me again how Raimondi enters. Is he not mainly Arabic and Syriac and Hebrew? G.: Yes, and that is exactly why Persian becomes amusing in his orbit. Once you direct the Medici oriental press, produce Arabic Euclid, and print philosophical grammars of Hebrew and Chaldean, the leap to Persian morpho-syntax no longer looks like pure whim. S.: So the Persian cat in the passage is not merely zoological. G.: No, it is philological allegory in fur. S.: Very good. G.: Thank you. Persian, like the cat, will not be learnt by decree. It must be coaxed, and grammar is the first saucer of milk. S.: Austin said it was useless. G.: Austin thought many things useless if they did not sharpen an English distinction by tea-time. S.: He might have liked the categories, though. G.: Only if they stayed close to examples. “The fat cat sat on the mat” would probably have pleased him more than “morpho-syntactic categories.” S.: Then let us do the fat cat. English: “The fat cat sat on the mat.” G.: Latin: cattus pinguis in storea sedit, though again one feels the lateness of both cat and mat. Greek: something equally schoolroom and suspicious, perhaps αἴλουρος πίων ἐκάθισεν ἐπὶ τοῦ τάπητος, though I should not stake my dignity on the exact cat. S.: Hebrew? G.: הַחָתוּל הַשָּׁמֵן יָשַׁב עַל הַמַּחְצֶלֶת. Persian: گربهٔ چاق روی حصیر نشست. S.: And what do we learn? G.: That noun, adjective, article, and relation are not implemented uniformly; that predication and attributive linkage differ; that the so-called same sentence is not the same act of grammatical architecture in each language. S.: Which suggests that the categories are partly abstract and partly language-specific. G.: Precisely. A philosopher who ignores either side will deserve his students. S.: Then where does “meaning” come back in? G.: Everywhere. If we are to speak of “mean” triadically — sss means mmm for iii — we must know what counts as the sign-token sss in a given language, how it is morphologically formed, what relations it encodes, what it leaves to context, and what an interpreter may reasonably recover. S.: So the meta-language is crucial. G.: Exactly. One cannot discuss “meaning” in the abstract while pretending that signs arrive already individuated and relations already fixed. The meta-language tells us what sort of thing a sign is in the object language. S.: Raimondi would approve that. G.: He might, though he would probably insist on a better script and more manuscripts before trusting us with the Hall. S.: A fair condition. G.: Entirely. S.: Then perhaps the true class title is not “Persian morpho-syntactic categories” but “What language must already be like for meaning to happen.” G.: Too long for the board, but exactly right. S.: And your answer? G.: That language must already distinguish enough — beings, actions, agents, relations, quantities, modalities, and all the rest — for a speaker to place something before an interpreter in a way recoverable as meaningful. S.: Recoverable, not merely decodable. G.: Exactly. Even Persian, for all its elegance, does not abolish interpretation. It only disciplines different parts of it differently. S.: So the Hall will hear that Persian, like English, like Latin, like Greek, like Hebrew, signifies through grammar, but not identically. G.: Yes. And they will hear that “category” is dangerous if one forgets whether one means Aristotle’s, Kant’s, a school grammar’s, or an actual language’s working joints. S.: You should say that more slowly. G.: In the Hall I say everything more slowly. It gives the illusion of depth. S.: Which is half teaching. G.: The respectable half. S.: And the other half? G.: Surviving questions. S.: Dry enough? G.: Sufficiently Persian, with one eye on Aristotle and the other on the cat.