The solution, referred to by Hjelmslev (1961) as the essential genius of human language, involves combining sounds in sequence to distinguish meanings. This immediately provides a stratified model with two ranks in the phonology, phoneme (distinct sounds) and syllable (distinct combinations of sounds) and one in grammar, the word (distinct meanings). At this point a second language can be proposed, lexese, in which there is a one to one relation between meanings and words.
As far as we know, evolved systems on the model of lexese do not occur. Once again, their limitations are obvious. They would require humans to remember an inordinately large number of words; and they have the further important limitation that no-one could mean something they hadn't already heard. The logical way out of systems of this kind is once more a combinatorial one — adding on a grammar than organises sequences of words. This brings us to something more closely resembling human language, at least as it is modelled in Hjelmslevian terms — a two strata system, with a minimum of two ranks on each stratum and an non-biunique (and therefore in principle arbitrary) relation between sound and meaning. This kind of two strata system is outlined in Fig. 1.11.
Blogger Comments:
[1] To be clear, Hjelmslev (1961) is concerned with expression and content; see [6] below.
[2] To be clear, the word, as grammatical form, is not meaning. The word realises meaning.
[3] This is a false conclusion to draw. To be clear, the one-to-one relation in this scenario is between wording and sounding, not between meaning and wording.
[4] To be clear, a language like lexese could not occur, because it requires grammatical form to emerge prior to the meaning it realises.
[5] To be clear, this does not resemble human language, because it is model in which words have no meanings.
[6] This is misleading, because it is not true. Hjelmslev's two levels are content and expression, whereas as Martin's are grammar and phonology. Where Hjelmslev's model includes meaning (in content), Martin's model does not.
On the one hand, this reflects Martin's misunderstanding of stratification, mistaking every stratum as a stratum of meaning; see later posts. On the other hand, it serves a rhetorical purpose, since Martin will later claim that he is filling a gap in SFL Theory by proposing a stratum of discourse semantics above the stratum of grammar, ignoring the fact that SFL Theory has always had a stratum of semantics above the stratum of grammar; see e.g. the primary source of Martin's model: Halliday & Hasan (1976: 5).
No comments:
Post a Comment