160 Michael Fortescue
identification and class-assignment or the like (as above), i.e. where one
alternatively could talk of a zero-marked derivational copula (something
actually quite widespread in the languages of the world; see Fortescue
1998: 62ff.). Likewise, a lexical item referring to an action or state (a pro-
totypical ‘verbal’ morpheme) generally has to be nominalized in order to
function as a nominal (e.g. by addition of the definite article, as above).
- Word classes and the relationship between predication and
lexicon in FG
The question I want to turn to now is this: how are the facts about Nootka
sketched above to be fitted into the FG model? First, the choice of the
‘most important/newsworthy’ element that is positioned first in the typical
clause (as in sentences 1–5) could in theory be made at the level of prag-
matic function assignment (wherever that may be). On the other hand,
Nootka ‘phrasal’ suffixes (as in 15–17), like semantically similar West
Greenlandic ones, can be treated by individual predicate formation rules,
and its ‘zero copula’ construction (as in 20) by term-predicate formation
rules in the manner of Dik (1989: 172) or, preferably, by non-verbal predi-
cate frames in the manner of Hengeveld (1992: 32). The lack of evidence
for the assignment of syntactic Subject status to particular arguments (as
mentioned in connection with sentences 13–14) would seem to be ac-
counted for by aligning Nootka with those numerous languages where FG
recognizes that Subject assignment simply does not occur. Furthermore,
Dik’s fundamental principle of treating both nominals and verbals as un-
derlying predicates would seem to be amply justified for this language,
where all lexical content morphemes are indeed predicates in the sense that
they can function – given an appropriate morphosyntactic ‘slot’ – as
verbals or nominals or modifiers of either. However, with the standard ver-
sion of FG, designed principally for non-polysynthetic languages, there are
problems with all of these suggestions.^13
Specifically as regards clause-initial ‘focused’ constituents, it appears
that the choice of first constituent is not simply a matter of Focus in Dik’s
original pragmatic function sense. The constituent chosen is either one dis-
playing general ‘newsworthiness’ (in the discourse context) or one required
by the lexical items and/or construction involved (as in 7–9). There has
been much discussion within FG in recent times as to whether Focus (and
other discourse-related functions) really belong in the model here at all or
should be part of a separate discourse grammar or module (as discussed