Halliday & Matthiessen (1999: 568):
The folk notion of the “word” is really a conflation of two different abstractions, one lexical [lexical item] and one grammatical [word rank].
Halliday & Matthiessen (2004: 51):
Word classes were traditionally called ‘parts of speech’, through mistranslation of the Greek term meroi logou, which actually meant ‘parts of a sentence’. These began, with the Sophists, as functional concepts, rather close to Theme and Rheme; but they were progressively elaborated into, and replaced by, a scheme of word classes, defined by the kinds of inflection that different words underwent in Greek … .
Viewing Word Classes
Halliday & Matthiessen (2004: 38):
Word classes can be viewed ‘from above’ — that is, semantically: verbs typically refer to processes, nouns to entities and adjectives to qualities (of entities or of processes). They can also be viewed ‘from round about’, at their own level, in terms of the relations into which they enter: paradigmatic relations (the options that are open to them) and syntagmatic relations (the company they keep). On either of these two axes we can establish relationships of a lexical kind (collocations and sets) and of a grammatical kind (structures and systems).
Processes, Participants And Circumstances
Halliday & Matthiessen (2004: 176-7):
This tripartite interpretation of figures … is what lies behind the grammatical distinction of word classes into verbs, nouns and the rest, a pattern that in some form is probably universal among human languages.
From Process Differentiation To Word Classes
Halliday & Matthiessen (1999: 513):
… construing processes in this way clearly depends on generalising whole classes of phenomena; the grammar sets up classes of process, of participant, and of circumstance. There are various ways of doing this; one that is familiar in many languages is by means of a taxonomy of different kinds of word. The classes of word may be distinguished by their internal form, or by the way they are able to enter into larger constructions (or both).
Class & Function
Halliday & Matthiessen (2004: 52):
The class of an item indicates in a general way its potential range of grammatical functions. … But the class label does not show what part the item is playing in any actual structure. For that we have to indicate its function. The functional categories provide an interpretation of grammatical structure in terms of the overall meaning potential of the language … .
The Method Of Traditional Grammar
Halliday & Matthiessen (1999: 466):
A characteristic of work on grammatical semantics, where this has been based on linguistics or on natural language philosophy, is to move in at the lower ranks of the grammar rather than the higher ones, and to start with classes rather than with functions. This is a continuation of the method of traditional grammar, which (because it originated with the study of observable features of language) was word-oriented and leant heavily on word classes in its descriptive statements. We find this tendency in discussions of word classes and their semantic values — the issue of the proper interpretation of adjectives, the exploration of various verb types, and so on.