78 Carlos Inchaurralde
have, for instance, the form o-kaki ni naru ‘to become a writing’ instead of
kaku ‘to write’. These forms are very lexicalized, but they still involve dif-
ferences in cognitive construal that require a different form at the
representational level. In these cases, even if we do not accept changes at
this level, it is clear that they do require different forms at the expression
level, which justifies the input from both the cognitive component and the
interpersonal level. There are many expressions in kei-go that differ only
lexically. We have the honorific prefixes o- and go- (e.g. o-tomodachi ‘his
friend’ instead of tomodachi ‘friend’), alternative forms (e.g. achira instead
of asoko ‘there’), and special words for the different kinds of kei-go (e.g.
iku ‘go’ and kuru ‘come’, which both change into irassharu in sonkei-go
and into mairu in kenjoo-go and teinei-go).
Of course, the information about the relative social distance of both in-
terlocutors is part of the communicative settings. General knowledge about
different kinds of communicative settings (as well as of different text ‘gen-
res’ or conventional text structures), with their implications for different
performative acts – cf. Austin’s (1962) mention of ‘felicity conditions’ – or
for the use of different communicative functions, is also part of the long-
term information stored in the cognitive component.
2.3. Linguistic competence and knowledge of the world
Hengeveld refers to linguistic competence as being part of the long-term
knowledge available to the cognitive component. If by linguistic compe-
tence we mean knowledge of the phonology, grammar, and vocabulary of
a language, then obviously we would have to include here the knowledge
of all vocabulary items and all rules of the grammar, as well as phonologi-
cal rules. However, although the form and certain combinatorial
characteristics of vocabulary items can be considered to be linguistic
knowledge, the above-mentioned ‘lexicopaedic’ information seems to go
beyond this kind of knowledge. It seems more suitable to consider lexical
linguistic information to be specifically the kind of information that is
available in the lexicon for predicates in terms of form, syntactic category,
quantitative valency, qualitative valency, selection restrictions, meaning
postulates, and meaning definitions (Dik 1989: 68ff.). This kind of infor-
mation, stored in long-term memory, is useful for a metalinguistic usage
of language, and for monitoring language production, but vocabulary
choice seems to have some degree of automatization. The same can be
said of syntactic and phonological rules. There seems to be some evidence
of independent modular capacities that deal with syntax and phonology in