Comment clauses and FDG 203` I would like to proceed to a revision also in the representational level,
which can be seen as the reverse of Hengeveld’s ‘downward layering’ in-
novation of the interpersonal level: just as the interpersonal level in
Hengeveld’s model is extended downwards so that C, T and R overlap with
items on the representational layer, I think representation should cover all
linguistically represented meaning (including that which codes interper-
sonal choices).
One example of why this would make sense is the analysis of the sen-
tence I apologize as on the one hand a performative utterance or on the
other hand as the answer to a question such as What do you do when you
realize you’ve hurt someone? A linguistic model that aims to specify rela-
tions between grammar and discourse should provide a level where it is
specified that the sentence codes the same represented content in both
cases – as well as indicate that under appropriate contextual conditions, the
sentence acquires the status of a direct coding of an interpersonal act rather
than of informative content. An illustration might be:
The performative reading:
Interpersonal Level Act: apology:
S apologize towards H
(No C)
Layer of linguistically represented
contentA-coding:
Decl (pres (apologize (S))The informative reading:
Interpersonal level Act: Informative:
S tell H C
C= T, R
Layer of linguistically represented
contentC-coding:
Decl (Pres (apologize (S))Figure 1. Performative and informative readings of I apologize.
The details of this representation are very preliminary, of course. The
point is to bring out even more clearly than in Hengeveld's version what I
see as a major advantage of the three-tier system, namely the possibility of
matching structures at different levels of analysis in order to stress the
similarities and differences (a perspective that has recently been in the