Discourse structure, the generalized parallelism hypothesis, and FG 305He then suggests adding to the existing term layers a fourth one which
he calls the ‘modality layer’. This yields structure (10), assumed to under-
lie both Term and Proposition :
(10)
[Mod-Op Mod-Sat]
[Loc-Op [Quant-Op [Qual-Op [Nucleus] Qual-Sat] Quant-Sat] Loc-Sat]The next and final step in the development of the parallelism hypothesis
comes with Dik’s (1997b) proposal that Clause and Text (Discourse in his
terminology) can be said to display a strong similarity, both on the struc-
tural level and the relational level. The basic idea is that intra-clausal
layered structure as well as intra-clausal functional relations can be pro-
jected onto Text. According to this view, the structure which can be taken
as common to both Clause and Text may be represented as follows:
(11)
[Loc-Op [Quant-Op [Qual-Op [Nucleus] Qual-Sat] Quant-Sat] Loc-Sat]It should be noted, finally, that the works (Hengeveld 1997, Cuvalay
1997 and Moutaouakil 1998 among others) in which the suggestion is
made to deal with supra-clausal phenomena in terms of an expanding up-
ward-layering approach (where Text is conceived of as a supra-clausal, or
indeed a supra-sentential, layer rather than constituting a separate module)
can somehow also be regarded as contributions to the development and the
extension of GPH in the sense that they all postulate, though to different
degrees, a certain structural similarity between clause, sentence and text.
3.2. Archetypal Discourse Structure
Before it can be taken as underlying the various discourse categories dis-
cussed in the previous section, structure (11), requires, it seems to me,
some further refinements and enrichments on both the constituency and the
relational levels. Here are some suggestions.
According to the typology proposed in Dik (1997b: 384–386), greetings
[Ill-Op [Mod-Op Mod-Sat] Ill-Sat]