Martin (1992: 581):
Because coding orientations are variably realised, ideology will never be a question of this or that but one of more or less; and because these coding orientations distribute discursive power unevenly, there will always be semiotic tension in the community. The variable realisation of ideology provides the dynamic openness through which this tension can be resolved — it is a necessary condition for the system to evolve.
 This continues the confusion of variability in the linguistic realisation (semantic style) of one coding orientation with the different linguistic realisations (semantic variation) of different coding orientations; see earlier critique here.
 This continues the misconstrual of coding orientation as ideology; see earlier clarification here.
 Bernstein's coding orientations do not distribute (the undefined) "discursive power" — unevenly or evenly — and so this is not a cause of (the undefined) "semiotic tension in the community". The codes are different orientations to language by different social groups. Halliday (1978: 106):
What Bernstein’s work suggests is that there may be differences in the relative orientation of different social groups towards the various functions of language in given contexts, and towards different areas of meaning that may be explored within a given function.
And these sub-cultural angles are functions of the social structure; Halliday (1978: 123):
This angle of vision is a function of the social structure. It reflects, in our society, the pattern of social hierarchy, and the resulting tensions between an egalitarian ideology and a hierarchical reality.
- it provides the dynamic openness through which semiotic tension can be resolved;
- it is a necessary condition for the system to evolve.
No evidence or argument is offered to support either of these bare assertions.