Process and pattern interpretations 169comes out in the wash’ in the expression rules anyway, as long as the rele-
vant functions are marked somewhere. In Hengeveld’s revised model one
would presumably start with Focus choices made at the interpersonal level
(where the factors affecting its assignment are obviously ‘at home’, with
direct access to the communicative context), but it is also possible that dif-
ferent kinds of Focus are relevant to distinct levels. It may turn out that
both contrastive constituent focus and the kind of ‘newsworthiness’ focus
leading to choice of first constituent in Nootka are determined from the
outset at the ‘interpersonal level’, whereas the actual articulation of the ut-
terance into successive predicate and referential phrases realizing these
choices must be accomplished (as process proceeds generally ‘from left to
right’) by reference to pattern possibilities belonging to the other two levels
(for example activating the contrastive focus template or summoning and
positioning the ‘telic’ affix -a).
This may be a disappointing conclusion for those who would like a de-
finitive answer to the question of ‘where pragmatics belongs’ in the FG
model. But perhaps we should not be so fixated upon the interrelation be-
tween grammar and discourse after all, but think rather in terms of Pattern
versus Process, a distinction that cuts across this divide, since Pattern in-
cludes the more ritualized/grammaticalized aspects of pragmatics as well
as core clausal grammar. It is not yet clear how exactly Hengeveld’s new
‘architecture’ is to be understood in this respect. One possibility for ac-
commodating both perspectives in one model (which appears to be
Hengeveld’s intention) would be, as hinted at above, to interpret it as pure
Pattern on the vertical axis, but as Process on the horizontal axis (‘left-to-
right’ as he puts it). This at least avoids the problem of psychologically
suspect ‘percolation’ processes – as long as all relevant triggering factors
are potentially there ‘in the top left-hand corner’ from the start.^22
Whatever the way forward, I suggest that the goal must be to bring the
Pattern and Process perspectives on FG into line in a relation of comple-
mentarity which does not blur over the essential distinction between them.
There is still much mileage to be derived from applying standard FG to the
abstraction of generalizations across linguistic pattern at all levels of its
layered template (including grammaticalized higher-level ones), thus
achieving still greater typological adequacy. Elaborating a specifically
Process variety of FG does not, it seems to me, require further theoretical
or formal sophistication – on the contrary, we need to simplify what we al-
ready have in the Pattern model. Meanwhile, let us at least avoid
ontologically suspect hybrid models. When talking about Pattern let us do
just that, but when talking about Process let us make it quite clear that we