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

 G.: Let us begin with a Lycian phrase, since Meriggi would complain if we entered by Latin only. Let us say simply: xñtabura ebeñnẽ prñnawa. A tomb for a householder, if one permits oneself the usual funerary dryness. S.: Dry enough already. And you choose Lycian because a declension is less likely to lie when carved in stone. G.: Quite. Living languages equivocate with charm; dead ones leave endings behind as evidence. S.: Then your point is that the endings are not mere grammar-school upholstery. G.: Exactly. Morpho-syntactic categories are public traces of deeper categories by which a community carves experience. Aristotle would call them ontological in one key, Kant categories of experience in another. A case-ending is a small metaphysics at the end of a word. S.: That is rather grand for a suffix. G.: Suffixes have long been made to do civilisation’s smaller work. S.: Let us start with the nominative, before you become transcendental. G.: Sensible. In Lycian, as in Greek and Latin, the nominative marks what a clause presents as subject or topic-bearer. Consider, by analogy, Latin: Lupus currit in agro. The wolf runs in the field. Lupus is nominative because the sentence is, so to speak, about the wolf as doing. S.: And the Lycian parallel? G.: Something like xñtawati prñnawa, if one imagines a householder acting or standing as the clause’s bearer. The exact lexicon matters less than the function. The nominative is the form under which a thing is presented as the one of which something is said. S.: Aristotle would say substance first, predicate later. G.: Yes, though grammar is never wholly obedient to metaphysics. Still, nominative feels like the case of presentation, the public front door of the noun. S.: So in Greek too, ho lukos trechei. The wolf runs. Nominative as the one doing the running. G.: Exactly. And Anglo-Saxon preserves enough of this to remind modern English that it once had manners. Se wulf rint. Subjecthood was not always entrusted to word order alone. S.: Then nominative corresponds to the category of subject of predication. G.: Or, if one is feeling Kantian, to the form under which an object of experience is taken as unified for predication. S.: You are warming up unpleasantly. G.: Meriggi would expect no less. Now the accusative. Latin: Agricola lupum videt. The farmer sees the wolf. Lupum is accusative, the object toward which the seeing is directed. S.: So the accusative is the case of directedness. G.: Very nearly. The case of affected or goalward objecthood in the transitive scene. Greek does the same: ho georgos ton lukon horai. And in a Lycian analogy one would look for the noun-form used where the action terminates upon or is borne toward another participant. S.: So if Lycian marks a direct object distinctly, the community has chosen to make explicit not merely who acts, but upon whom action falls. G.: Precisely. The nominative says, as it were, here is the bearer; the accusative says, here is the borne-upon. A very old ontology of asymmetry. S.: Anglo-Saxon still has that, in a way. Se mann geseah þone wulf. G.: Yes. And the demonstrative helps preserve what later English lets word order carry with weary diligence. S.: The genitive next, I assume, before you accuse the cases of disorder. G.: Quite. Latin: Liber pueri in mensa est. The boy’s book is on the table. Pueri is genitive. Greek: to biblion tou paidos epi tes trapezes estin. The relation is not merely possession but belonging, source, dependence, specification. S.: So the genitive marks what is of another. G.: Exactly. It encodes derivation, belonging, partition, kindred dependence. In a Lycian funerary phrase, one often suspects the genitival atmosphere even when the exact ending is disputed: tomb of so-and-so, child of so-and-so, house of such a lineage. S.: Aristotle again would call this relation rather than substance. G.: Yes, though Roman schoolmasters made it look almost domestic. But belonging is an ontological category too. The genitive tells us that things are not merely there; they are of something. S.: And Anglo-Saxon had that with ease. Hus þæs cyninges stent. The king’s house stands. G.: Very good. A respectable Germanic genitive before modern English began scattering apostrophes like confetti. S.: Dative now. G.: Naturally. Latin: Magister puero librum dat. The teacher gives the boy a book. Puero is dative. Greek: ho didaskalos tō paidi biblion didōsi. The dative marks recipient, advantage, reference, orientation toward a beneficiary or addressee. S.: So the dative is the case of “to” or “for,” though not always reducible to prepositions. G.: Precisely. It encodes the structure of directed giving, saying, showing, helping. A category of relation in which something is not merely done, but done to or for someone. S.: Which sounds rather social. G.: Many cases are. Grammar remembers that life is not a sequence of isolated substances but of directed dealings. S.: And Lycian? G.: One looks for the oblique forms serving recipient or target functions. The nomenclature may differ; the practical intelligibility does not. If a Lycian text says, in effect, “this tomb to X,” we are already in dative country, even if Meriggi and his predecessors quarrel over the exact map. S.: Anglo-Saxon again: Se lareow sealde þæm cilde boc. There it stands. G.: Yes. Enough survives there to embarrass modern English into modesty. S.: Now the ablative, your favourite grievance. G.: Not favourite. Merely unavoidable. Latin: Miles gladio pugnat. The soldier fights with a sword. Or: Puella ex urbe venit. The girl comes from the city. The ablative is an untidy Roman empire of separation, instrument, source, circumstance. S.: In other words, a bureaucratic success and a philosophical nuisance. G.: Exactly. Greek lets the dative and prepositions do much of what Latin heaps upon the ablative. This is why I have always thought the Latin ablative too successful for its own conceptual hygiene. S.: Yet you use it. G.: As one uses the Foreign Office. Reluctantly but repeatedly. If Lycian has oblique cases covering instrument, source, or circumstance, one should compare them to Latin ablative functions without pretending exact identity. One analogy at a time. For instrument, Latin: Scriba calamo scribit. The scribe writes with a reed-pen. If Lycian uses a marked oblique to convey “with” or “by means of,” the structural analogy is sound. S.: So the ontological category there is mediation by means. G.: Or source, depending on the sentence. The ablative is really several categories in one administrative overcoat. S.: You do not sound reconciled. G.: I am not. Which brings us to the locative, that furtive survival. S.: Ah yes, Romae. G.: Precisely. Latin pretends not to have a locative while quietly retaining it where place-names refuse instruction. Romae sum. I am at Rome. Domi maneo. I remain at home. The textbooks bury the locative because it interferes with tidy declensional propaganda. S.: Greek uses prepositions instead, mostly. G.: Yes, though one feels the older local functions behind them. The question for Lycian is whether one should posit a distinct locative value or treat place-relations through other obliques and particles. I am not eager to multiply locatives merely because comparative philologists miss them. S.: So your dislike is not of place, but of exuberant case-inflation. G.: Exactly. One should not invent a case every time a noun loiters somewhere. Still, if Lycian has forms regularly marking “in” or “at” in a sufficiently morphological way, one must take the evidence seriously. Meriggi, being Meriggi, would enjoy the quarrel. S.: And the full sentence for Latin locative? G.: Romae poeta habitat. The poet lives at Rome. One cognate comparison with Greek is harder because Greek will prefer en Rōmēi or en polei with the preposition doing the overt work. S.: So Latin here preserves a bit of Indo-European local dignity. G.: A bit, yes. Enough to annoy simplifiers. S.: What of the vocative? You have not greeted anyone yet. G.: True. Latin: Serve, veni huc. Slave, come here. Or more politely, Marce, audi. The vocative marks direct address. Greek likewise: ō Marke. It is not a syntactic participant in the same way, but a pragmatic one. A case whose ontology lies less in being than in interpersonal summons. S.: So the vocative is almost the case of second-person encounter. G.: Very good. The case of turning toward another. If Lycian inscriptions name someone in address, one should watch whether the form differs. Though funerary Lycian is not exactly the agora. S.: No one in a tomb answers briskly. G.: Not in the better inscriptions. Now the dual. S.: At last. Greek had it and used it just enough to rebuke later laziness. G.: Precisely. Dyo anthrōpō, the two men, and so on. The dual marks that two is not merely a small many but a specially structured pair. Aristotle would see a numerical distinction; Kant perhaps a peculiarity in the manifold of counting; common sense sees that two shoes are not yet a crowd. S.: Anglo-Saxon pronouns also kept a dual. Wit for “we two,” git for “you two.” G.: Yes, and that fact alone should cure Englishmen of supposing that their language was always philosophically negligent about number. S.: Latin lost the dual, mostly. G.: Or fossilised it into embarrassment. Ambō, duo in certain pair-bound uses, but no living nominal dual system. Latin treats two as the first plural, which is efficient but metaphysically unimaginative. S.: And Lycian? G.: If Lycian lacks a productive dual, it joins Latin and later practicality. If it preserves traces, one asks whether the community still hears twoness as a distinct category of experience rather than a mere arithmetic threshold. S.: You make grammar sound like anthropology. G.: It is anthropology in inflectional dress. S.: Let us go back and compare each case with one cognate example more explicitly, since Meriggi deserves schoolroom exactness. G.: Very well. Nominative. Latin: Pater venit. The father comes. Greek: patēr erchetai. Both preserve the Indo-European root in comparable shape and the nominative as subject-presentation. S.: Lycian parallel would then seek the subject form of a comparable kinship noun. G.: Exactly. Genitive. Latin: Domus patris magna est. The father’s house is large. Greek: hē oikia tou patros megalē estin. The relation of belonging is marked morphologically in both. S.: Dative. Latin: Filius patri aquam portat. The son carries water to the father. Greek: huios tō patri hydōr pherei. G.: Good. Accusative. Latin: Mater filium vocat. The mother calls the son. Greek: mētēr ton huion kalei. S.: Ablative or instrumental analogue. Latin: Mater manu puerum ducit. The mother leads the boy by the hand. Greek often with dative or preposition: tē cheiri agei to paidion. G.: Precisely. And one should not force Greek to produce a Latin ablative when it prefers a different distribution of labour. S.: Locative. Latin: Corinthi mercator habitat. The merchant lives at Corinth. Greek: en Korinthōi ho emporos oikei. G.: Yes, and the contrast is pedagogically perfect. Latin preserves a case remnant; Greek lets the preposition do the visible work. S.: Vocative. Latin: Pater, audi me. Father, hear me. Greek: ō pater, akoue mou. G.: Just so. The cognates help because they show not only similar lexemes but similar relational needs. S.: And neuter? G.: Ah yes, the neuter, that old delight. Latin: Donum in mensa iacet. The gift lies on the table. Greek: to dōron epi tēs trapezēs keitai. Nominative and accusative neuter coincide. The language here marks not a third kind of thing in nature, but a third pattern of agreement and a special economy between subject and object forms. S.: So the neuter is less ontology in the naturalist sense than grammar’s own quiet classification. G.: Exactly. Yet even there one may ask whether the language treats certain conceptual regions as less agentive, less person-like, more thing-like. One must be careful, but not blind. S.: Anglo-Saxon preserves the neuter too. Þæt hus stent. The house stands. G.: Yes, and modern English inherits the wreckage through “it,” while pretending not to have gender at all unless scandalised into it. S.: What about plural? G.: The plural repeats the categories under the pressure of multiplicity. Latin: Patres veniunt. The fathers come. Greek: hoi pateres erchontai. Nominative plural marks many bearers of predication. Genitive plural: Domus patrum magnae sunt. The houses of the fathers are large. Dative plural: Filii patribus dona dant. Sons give gifts to the fathers. And so on. S.: So number does not abolish the case-relations; it scales them. G.: Precisely. The plural is not philosophical novelty so much as the discipline of repetition. S.: Unless one comes to the dual, where two insists on being special. G.: Quite. Human beings are oddly sentimental about two. S.: Let me ask the larger question. Are you really saying that nominative, accusative, genitive, dative, and the rest reflect ontological categories in Aristotle’s sense? G.: Not crudely. I am saying that a language’s morpho-syntactic choices publicise what relations it chooses to mark as basic for the handling of experience: bearer of predication, affected object, belonging, directed recipient, instrument, source, place, address. These are not identical with Aristotle’s categories, but they are not unrelated to the same pressure to sort the forms of being and saying. S.: And the Kantian version? G.: That case-systems may be read as historical codings of recurrent relational forms by which experience is organised for judgment: subject, object, possession, destination, location, means. Grammar is not transcendental deduction, but it is sedimented reason. S.: Meriggi would have liked “sedimented reason.” G.: He would probably have declined it into Lycian. S.: You are not wholly hostile to him, then. G.: Not at all. Meriggi’s title, La declinazione del lizio, is charmingly schoolmasterly. It announces that one may begin from endings and arrive at civilisation. S.: Then let us consider the local difficulty again. If Lycian shows an oblique used in place-relations, how do we decide whether to call it locative rather than dative, ablative, or merely “oblique”? G.: By discipline, distribution, and restraint. One asks whether there is a distinct morphological pattern regularly associated with static location rather than recipienthood, source, or general obliqueness. If not, one resists locative enthusiasm. S.: In other words, no honorary cases. G.: Exactly. Grammarians are too ready to found principalities on small phonetic evidence. S.: Yet one should also not flatten real distinctions. G.: Quite. A dead language deserves justice, not economy alone. S.: This is beginning to sound Roman. G.: Most good philology does. Now let us make the analogies explicit in one sequence, if only to satisfy the classroom in Meriggi. S.: Proceed. G.: Nominative: Latin Pater filium amat. The father loves the son. Pater is the bearer. Greek Patēr ton huion philei. Lycian, mutatis mutandis, would mark the father in subject position. S.: Accusative: the son is the loved-upon. G.: Precisely. Genitive: Liber patris iacet. The father’s book lies there. Greek To biblion tou patros keitai. Lycian would use the form marking belonging or descent. S.: Dative: Mater filio panem dat. The mother gives bread to the son. Greek Mētēr tō huiōi arton didōsi. G.: Ablative or instrumental analogue: Faber malleo laborat. The craftsman works with a hammer. Greek prefers a different device, but the category of means is there. S.: Locative: Domi puer manet. The boy remains at home. G.: Exactly, and one sees how Latin hides a case inside an adverbial survivor. Greek: en oikōi to paidion menei. S.: Vocative: Fili, veni. Son, come. G.: Very good. One could almost build a civilisation from imperatives and case-endings. S.: Many did. G.: Quite. Now, one more thing. The Greek dual and the Latin lack of a productive dual show that number itself is culturally and grammatically negotiable. The world may contain pairs, but a language need not inflect for them. S.: So categories of experience are not simply read off nature; they are chosen for public marking. G.: Exactly. There is the Kantian note. Experience may be universally structured in some deep sense, but languages differ over what they force speakers to make explicit. S.: And that is why declension matters philosophically. G.: Entirely. Declinatio is not merely a list of endings. It is a map of what a linguistic community chooses to distinguish openly between nouns and the world. S.: Then Meriggi, by marching Lycian beside Latin and Greek, is asking whether a less familiar Indo-European language marks the same public relations or redraws them. G.: Just so. And that is why one should begin with a Lycian phrase rather than apologise for it. S.: We began with xñtabura ebeñnẽ prñnawa. Have we ended anywhere worthy of it? G.: We have ended where all good declensions end: with the suspicion that endings are not trivial. S.: And your final word on the locative? G.: That Rome kept it like an eccentric aunt and grammarians should neither disown her nor move her into every spare room. S.: Dry enough? G.: Sufficiently Lydio-Lycian, with a Roman ablative of annoyance.