Remarks on layering 287sions in an actual communicative situation. It is, in fact, a major element
driving this process.
Certainly, illocution is related in very profound ways to the conceptual
system, and even to certain qualificational dimensions there, i.e., probably
not accidentally, the high-level ones, belonging to the ‘attitudinal stack’
(see Section 4.2). For one thing, illocution shares with these qualifications
the matter of performativity. In fact, quite like illocution, speaker attitudes
are bound to the hic et nunc of the speaker.^16 Of course, what is performed
in qualification and in illocution are quite different, i.e. a conceptual opera-
tion of determining an attitude towards an SoA vs a linguistic act towards
an interlocutor. Nevertheless, the ‘performing entity’ is no doubt the same
in both (cf. Section 5 below).
For another thing, the sources of illocutionary force are obviously also
conceptual, and in this respect too, there are links with (the same)
qualificational categories. Information questions ensue from gaps or
unclarities in conceptual knowledge, i.e. essentially from epistemic
uncertainty, and the intention to fill or resolve them. And orders or requests
for action, or promises, probably result from an epistemic assessment of
the (non-)existence and a deontic assessment of the (un)desirability and/or
(non-)necessity of an SoA, and the intention to do something about that.
This may explain why illocutionary markers and deontic and epistemic
markers are often closely related. It also explains why performative episte-
mic expressions do not occur in questions.^17 Thus, a speaker can deal with
an informational gap in two ways, depending on what (s)he aims to
achieve. Either (s)he can communicate his/her own view, with an epistemic
assessment, of how it might be filled, leading to a declarative with an
epistemic expression. Or (s)he can ask the hearer to fill it, leading to a
question thematizing the gap or unclarity as such. But then, there is no use
for an epistemic expression. In fact, if there is such an expression in other
than a speech act-modifying function, then this is automatically part of the
thematized gap, hence descriptive (cf. note 16).
A second and even more complex issue is how to deal with conditional,
temporal or causal relations, relations of purpose, reason, motivation, etc.
in linguistic expressions. FG fully integrates these in the layered system, as
satellites at the different levels. But conceptually, these dimensions are ob-
viously quite different from ordinary qualifications of SoAs: they concern
relations between SoAs. Thus, even as satellites they usually take the form
of full subordinate clauses, although the same links exist between syntacti-
cally independent clauses. So these relations extend beyond the domain of
a single utterance or SoA. How such relations get expressed, then, is a dis-