[...] ʻmeaning potentialʼ, which is what the speaker/hearer 'can' (what he can mean, if you like), not what he knows.
- Halliday (1973), a pag.25

[...] the emphasis is here on the form of the language. This consists of a meaning potential, represented as a network of options, which are derived from a particular social function and are realized, in their turn, by structures whose elements relate directly to the meanings that are being expressed.
- Halliday (1973), a pag.29

The potential of language is a meaning potential. This meaning potential is the linguistic realization of the behaviour potential; ʻcan meanʼ is ʻcan doʼ when translated into language. The meaning potential is in turn realized in the language system as lexicogrammatical potential, which is what the speaker ʻcan sayʼ.
- Halliday (1973), a pag.51

Meaning potential is defined not in terms of the mind but in terms of the culture; not as what the speaker knows, but as what he can do – in the special sense of what he can do linguistically (what he ʻcan meanʼ, as we have expressed it). The distinction is important because ʻcan doʼ is of the same order of abstraction as ʻdoesʼ; the two are related simply as potential to actualized potential, and can be used to illuminate each other. But ʻknowsʼ is distinct and clearly insulated from ʻdoesʼ; the relation between the two is complex and oblique, and leads to the quest for a ʻtheory of performanceʼ to explain the ʻdoesʼ.
- Halliday (1973), a pag.52-53

The meaning potential is the range of 'significant' variation that is at the disposal of the speaker. The notion is not unlike Dell Hymes’ notion ʻcommunicative competenceʼ, except that Hymes defines this in terms of ʻcompetenceʼ in the Chomskyan sense of what the speaker knows, whereas we are talking of a potential linguistic sense of what he can mean – and avoiding the additional complication of a distinction between doing and knowing. This potential can then be represented as systematic options in meaning which may be varied in the degree of their specificity – in what has been called ʻdelicacyʼ. That is to say, the range of variation that is being treated as 'significant' will itself be variable, with either grosser or finer distinctions being drawn according to the type of problem that is being investigated.
- Halliday (1973), a pag.54

A network of socio-semantic options - the representation of what we have been calling the ʻmeaning potentialʼ – has implications in both directions; on the one hand as the realization of patterns of behaviour and, on the other hand, as realized by patterns of grammar. The concept of meaning potential thus provides a perspective on the nature of language.
- Halliday (1973), a pag.65

The term ʻmeaningʼ has traditionally been restricted to the input end of the language system: the ʻcontent planeʼ, in Hjelmslev’s term, and more specifically to the relations of the semantic interface, Hjelmslev’s ʻcontent substanceʼ. We will therefore use ʻmeaning potentialʼ just to refer to the semantic options (although we would regard it as an adequate designation for language as a whole).
- Halliday (1973), a pag.72