Process and pattern interpretations 175
from language to language than we often imagine. The reason Nootka ap-
pears to be close to the extreme here is the high degree to which purely
discourse-pragmatic factors directly determine coding choices (by contrast,
English employs to a greater degree subtle, analogue differences of intona-
tion to articulate such matters). Mackenzie (1998) follows Dik’s cue in his
recent exploration of the ontogenetic expansion of the ‘holophrase’, taken
as the essential item under Focus in an information ‘chunk’ that initially
fills the P1 position. This meshes nicely with Hannay’s emphasis on P1 as
the locus of initial message management choices. If English itself could
then be seen as in some sense also the phylogenetic result of a long histori-
cal development where P1 became gradually freed up for other items than
those under Focus (but with the initial situation remaining in adult holo-
phrases), one might want to say that Nootka represents a language that is
still much closer to its holophrastic roots (not for nothing is it also ‘polysyn-
thetic’). P1 is still reserved for Focus (in some sense) in Nootka in virtually
all its utterances types – here is positioned the one item of an information
unit that cannot just be dropped. This contrasts, by the way, with equally
polysynthetic West Greenlandic (which is basically SOV rather than VSO
like Nootka), where there is no P1 with that function – in that language Fo-
cus often adheres to final position in the clause (except in its contrastive
focal construction).
- Note that I am discussing a level of representation corresponding to (poten-
tially) conscious experience. This I see as essentially a matter of inferential
processes (whether on-line or automatized), both lexically-cued and gram-
mar-cued (see Givón 1995: 364ff.). These may well in turn involve basic
operations such as ‘search and retrieve’, ‘establish new node’, ‘connect
node’, etc., of the kind Givón proposes – or indeed Nuyts’s (1992: 272ff.)
sentencing (i.e. the breaking up of information into clause-sized units) and
the like – but this is surely at a level of assumed cognitive activity below
what is accessible to conscious experience. Moreover, they are not relevant
to minimalizing the FG model as such. Nuyts’s cognitive component actu-
ally consists of declarative rather than procedural knowledge (it is a
situational network consisting of representations, i.e. – like the FG compo-
nent– represents pattern), the claim being that such knowledge is relevant to
both logic (process) and to conceptual knowledge (pattern). The processes
meant to implement it are only vaguely adumbrated, however. The kind of
basic processes I am concerned with are more like Givón’s referent pointing
(a matter of attracting attention), easily understandable as complex concres-
cences and describable in prehensional terms (in this case involving an
indicative prehension).
- In general, interpreting the vertical axis of Hengeveld’s revised model as
universal process does not seem warranted. As regards Nootka, for exam-
