98
K. M. A. BARNETT
in the light of Latin and Greek. Like Latin and Greek, all languages had to have clearly identifiable nouns, pronouns, adjectives, verbs, adverbs, prepositions, conjunctions and interjections; nouns, pronouns and adjectives had to be declined, with genders, numbers and cases; verbs had to be conjugated, with person, number, voice, mood and tense; even prepositions had to dance to the grammarians' tune, even though in Greek many of them were post-positions and even in Latin some of them were really suffixes, like the cum in TECUM.
All this had its utility, and I will suggest later that Cantonese too may with some advantage be studied in this way. But the grammarians' framework of an agglutinative hypotactical language with certain "parts of speech" and no others must first be loosened. We need, might I say, less a strait-jacket than a body-stocking, with considerable S-T-R-E-T-C-H.
The trouble, you see, with the classical grammarians' approach was that they left out so much. Having set up, for nouns, pronouns and adjectives, the landmarks of gender, number and case they were inhibited from even noticing that "gender" (which means only "kind", "description") need not be confined to masculine, feminine, neuter. "Number" need not be merely singular, dual, plural. And similarly with verbs.
The original basis of the Indo-European verb was voice (active, passive or reflexive) mood (infinitive, imperative, indicative, subjunctive, optative and various kinds of gerunds and participles) and aspect (perfective, imperfective, iterative, unique, inceptive). Tense, the indication of time, was a later development.
These forms were built up by prefixes, suffixes and infixes stuck on to the base or root; and the root itself could be modified (but not to the extent found in the Semitic family of languages) by reduplication, transposition, augment and vowel-changes.
There was probably also a structure of tones and stresses, but this is so inadequately covered by the various alphabets used in transcribing this family of languages that we know very little about how the patterns of words and sentences developed. But all these languages share the idea of syllables not separated in writing, of words written with a separating space, phrases and sentences,