Martin (1992: 215):
The final internal comparative systems to be considered have to do with difference rather than similarity. Within this group [converse] relations in which two aspects of a message trade rôles can be contrasted with those in which rôle reversal of this kind is either not present, or not explicitly drawn to attention.
DIFFERENCE:CONVERSE[4:129] The most general and important point that has come out of this chapter is probably the close connection between data and theory. Until the data on quantitative variations on linguistic variables became available through the work of Labov, it was unnecessary to take seriously the need for quantitative statements in a linguistic theory,and conversely the lack of a place for such statements in linguistic theory prevented most linguists from bothering to look for the relevant data.
DIFFERENCE:OPPOSITION[4:130] The lack of a place for quantitative statements in linguistic theory prevented most linguists from bothering to look for the relevant data.On the other hand it could be argued that the relevant data could not be interpreted in any case since the tools for its collection and statistical analysis were not readily available.
In SFL theory, the expansion relation in both texts is adversative addition, a type of extension; its meaning is X and conversely Y (see Halliday & Matthiessen 2004: 405).
In the first text, this is manifested structurally as a paratactic relation between clauses (logical metafunction), whereas in the second text, on the other hand, it is manifested non-structurally as cohesive conjunction (textual metafunction).
That is, the contrast presented here as types of difference — a type of enhancement — is actually, in SFL theory, a contrast between metafunctional manifestations — logical vs textual — of adversative addition, a type of extension.