The origin of parts of speech systems
The parts of speech categories used in modern linguistic theory go back to the classical grammarians of Greek and Latin. Both Plato and Aristotle include embryonic classifications of words into parts of speech; the expression ‘parts of speech’, indeed, is first recorded in Aristotle. But the first systematic classification of parts of speech is due to the Stoics – philosophers who belonged to the school of Zeno of Citium, who was active round 300 BC. The early Stoic classification recognized just four parts of speech: nouns, verbs, articles (including pronouns) and conjunctive particles (including prepositions; Michael 1970: 48). Dionysius Thrax, the grammarian responsible for the first surviving work of Western grammar (see Kemp 1986), recognized twice the number, distinguishing nouns, verbs, participles (which roughly correspond to the -ing form of English verbs), articles, pro nouns, preposition, adverbs and conjunctions.
This eight-fold classification persisted in the influential system of the Latin grammarian Donatus (mid-fourth century AD), the teacher of St Jerome (the translator of the Bible into Latin). Donatus’ list of categories was the following: noun, pronoun, verb, adverb, participle, conjunction, preposition and interjection. Note that this list substitutes interjections for articles. This reflects a significant difference between Greek, the language for which Thrax’s grammar was designed, and Latin, Donatus’ concern. Unlike Greek, Latin lacks a word meaning ‘the’, making the ‘article’ category completely redundant for this language (neither Greek nor Latin has a word for ‘a(n)’). This difference aside, however, the two languages are sufficiently simi lar for the rest of the Greek word classes still to play a plausible role in the description of Latin. Donatus’ system forms the basis for the whole grammatical tradition of the Middle Ages, and, consequently, for the classification which we’re familiar with ourselves. That’s not to say that the system didn’t go unchallenged. Various scholars throughout the medieval period tinkered with the basic eight-way division found in Donatus (details in Michael 1970). But it was Donatus’ system that held sway. Smaragdus (a celebrated theologian active towards the start of the ninth century AD) followed Donatus’ view that there were only eight parts of speech, adding that ‘the whole church . . . holds that there are only eight, and I have no doubt that this view is divinely inspired’ (quoted in Michael 1970: 51). When grammarians first began to describe the grammar of European vernaculars, they based their classifications closely on these Latin systems, regardless of how appropriate they were to the language being described. It is as though principles from cricket were automatically used to describe any ball game, with terms like stumps, wicket, run, fielder and innings being applied indifferently to football, hockey, lawn bowls and tennis.
