Remarks on layering 285of construction is probably as superfluous in syntax as is the semantic
hierarchy of qualifications itself. For it follows directly from the layered
system at the conceptual level. Obviously, if low-level qualifications fur-
ther specify internal aspects of the SoA, they will naturally affect the
relation between the participants in the SoA. Since this relation usually gets
expressed as the predicate in a linguistic expression, they appear to be
predicate modifiers. If medium-level qualifications situate the SoA as a
whole, they affect not only the relation between the participants, but also
the participants themselves. So in linguistic expression they will usually
appear to affect the predicate plus the core participants, plus the expres-
sions of the low-level qualifications, of course. Finally, high-level
committing qualifications affect whatever there is about the SoA, including
its external situation, so it is only normal that their expressions will usually
also affect the situating operators and satellites.
Note that I am relativizing these statements with the labels ‘usually’ or
‘normally’. In fact, the situation need not be as described. Corpus data
(Nuyts 2001a: Section 2.4, 3.4 and 4.4) show that a qualification such as
epistemic modality, for example, sometimes only affects a subpart of a lin-
guistic expression, such as a single constituent or a parenthetical insertion.
Moreover, it has been observed that most or all qualificational dimensions
which can operate at the sentence level can also occur within the noun
phrase (the term operators in FG – cf. Rijkhoff 1990).^13 Our present ap-
proach allows for a straightforward explanation. Depending on how the
speaker, in view of the communicative circumstances, decides to organize
the chunk of information he aims to express, a relation in conceptualization
that is qualified somehow (e.g. epistemically) can be linguistically coded,
not in the main pattern of an utterance, but in some subdomain (a noun, a
constituent, a parenthetical, etc.). In such cases, then, the qualificational
expression can hardly be considered to be attached to – in the epistemic
case – the predication in utterance syntax. This observation offers another
good reason to consider the extension of scope over domains of the SoA to
be a matter of the layered system in conceptualization, and not in syntax.
4.4. Qualifications and dimensions of interaction and discourse
In spite of certain correspondences between the FG account of layering and
the present approach, many aspects of the organization in the FG system
are entirely absent here. I have argued elsewhere (cf. Nuyts 1992a: 196f.,
1992b) why the distinction in FG between an interpersonal and a represen-