Page 524 - Encyclopedia of Philosophy of Language

Key Figures
genike for the genitive, etc., or perhaps more likely the names used by Aristotle (to toutou 'the of-this,' i.e., 'genitive'; to toutOi 'the for- this,' i.e., 'dative,' etc.).
(d) The traditional order of the cases (nom, gen, dat, ace, voc) may have existed in the fifth century, though a clear example does not appear until Cleochares (ca. 300 BC), as quoted in pseudo-Herodian's Peri schSmatOn, pro- duces an instance of poluptdton, the use of a particular noun or pronoun (in this case 'Demosthenes') in the same position in five suc- cessive phrases, in a different case each time, with the cases in the traditional order. The tra- ditional order of genders (masc, fern, neut) appears already in Protagoras (quoted above) and several times in Aristotle, but the order of numbers (singular, dual, plural), though per- haps natural (since the klSsis or basic form is always nom, sing, masc) is not attested early.
(e) The notion of dual patterning (Hockett) or double articulation (Martinet), that, essen- tially, the rules for correctness of sequences of vowels and consonants are unrelated to those for sequences of nouns, verbs, etc., appears first in Democritus (565, as cited in Isidore's Etym. 13.2.4. 108) and then in Plato (Theaet. 202-04, Cratylus 424-25, 431-32), though until Apol- lonius Dyscolus (Synt. 1.1-2) it often appears as a merely hierarchical arrangement: letters make syllables, syllables make words, words make sentences, and (in Plato's Cratylus 432A- C) sentences make discourses. But Apollonius clearly makes letters and syllables concrete, perceptible entities (Stoic aisthsta), whilewords and sentences are abstract, thinkable entities (noSta), so the Stoics should be given credit for the correct form of 'dual patterning.'
The amount of phonology current in the fifth cen- tury BCis probably reflected in Plato, who (in Crat. 431-32) assigns the topic to the science of grammatiks, and in Aristotle's Poetics (ch. 20, 1456b 25-31). Here (and in Plato's Theaet.), the distinction between vow- els and consonants is clear, and among consonants there is a distinction between psophoi (noises: pre- sumably s, ksi, and psf) and phonai (voices: m, n, r, I) and consonants which are aphdna (all the stops— ptkbdg, phi, theta,chi); Aristotle's sumphOna (latinized as 'consonants' but much narrower here) include both s and the liquids and nasals, i.e., our 'continuants.' Though words like 'labial' and 'dental' do not appear until much later, there is a curious inscription (IG n.5.4321) of about 350 BC which describes some sort of special notation (interpreted by some as shorthand) for just those two classes. There is no evidence of categories like 'aspirated,' 'voiced,' or 'voiceless' until much later, though the relation between [h] (for initial
aspiration) and the aspirated stops must have been learned before students could spell correctly.
Another recurrent theme of Greek grammatical thought is the listing of sentence types, i.e., illo- cutionary forces (in the sense of John Austin) or (in a few instances) moods, which begins in Protagoras, as mentioned above, and culminates with the lists discussed in Aristotle and the Stoics.
Finally, the notion to which Saussure gave the name 'arbitrariness of the sign should be considered,' i.e., whether or not there is some rational basis for the association of any given phonological sequence (of a word or morpheme) with a particular meaning or function. As conceived by the Greeks, the question was this: is the sound-meaning relation of all or some Greek words inevitable and natural? This is the main topic of discussion in Plato's Cratylus, but Aristotle, the Stoics, and Epicurus all also discussed the ques- tion, generally agreeing that all words are now arbi- trary, though some suggest that there was once a time when they were not. Democritus (as quoted in Proclus' commentary on the Cratylus 16) offered four argu- ments (with four specially coined names) in favor of arbitrariness: (a) 'homonymy' or 'polysemy,' i.e., the same sequence of phonemes may be associated with two or more unrelated meanings; (b) 'polyonymy' or 'isorrophy,' i.e., the existence of synonyms; (c) 'metonymy,' i.e., the fact that words and meanings
change; (d) 'nonymy,' i.e., the nonexistence of single words for simple or familiar ideas. Elsewhere (in Diodorus Siculus 1.8.3) Democritus argues for the 'polygenesis' of language—in other words, that there was no single proto-human language.
But a stronger claim than that of nonarbitrariness is the one implied by the famous experiment of Psam- metichus as narrated by Herodotus (2.1-2), not just that some Greek words have a natural origin, but that some existing language really is the single proto- human tongue. This entails monogenesis, of course, plus the notion that only some languages undergo linguistic change. Psammetichus arranged for a child to be raised in such a way that it never heard anyone speak. When, at last, the child spontaneously uttered a word, it was bekos, which Herodotus tells us is Phrygian for 'bread.' This belief in a single original language comes into later European thought from the biblical tale of the Tower of Babel.
Plato thus gives a good idea of what might be called 'normal linguistics,' the kind of grammar that every man knew who had been to school as a boy in the fifth and early fourth centuries BC. This information can be supplemented by what can be read in Aristotle's Poetics, and one or two other places.
See also: Aristotle and the Stoics.
Bibliography
Ax W 1986 Quadripartite ratio: Bemerkungen zur Ges-
502