There are _scale_ hierarchies, in which each level is a
_constituent_ or _part_ of the next higher/larger level (cell,
tissue, organ, organism, ecosystem).
There are _specification_ (aka inclusion, inheritance)
hierarchies in which each level is a kind or subtype of the next
higher/ more general level (squares, rectangles, quadrilaterals,
polygons).
There are _realization_ (aka type-token, instantiation)
hierarchies in which each level is a specific way of coming into
being of the manifold potential by which the next higher level is
characterized (top/down: meaning, wording (vs. depiction),
sounding (vs. writing)).
Consider the weather/climate relation. Weather is what happens on
some occasion, it is local, instantial, individual if not unique.
There is local weather, now, and global weather, now (scale
hierarchy). There are constructed similarities between this
weather-instance and that one, and so types of weather: sunny,
rainy, windy, etc. And, on some particular scales of spatial
extension, we can construct averages or tendencies over some
scales of time that are usefully constant or predictable (with
fluctuations from the expected norms): climates, which can be
microclimates (Central Park), larger scale (NE USA), and even
bigger or global (Ice Age).
For language, we can start with texts or utterances as the
instances, and look at their geographical distribution (scale).
We can devise text-types, like genres, which also have
distributions in space and time. We can imagine a genre as a
realization of some social meaning-disposition (Jim Martin has
called these 'ideologies', though one could want a better term, I
think). We can also look for average tendencies in texts at
various internal scales (long texts, short texts like sentences,
clauses, nominal groups, words), and perhaps find useful
constancies, which we can call grammars, lexicons, etc. If we
include all scales of texts, the existence of such
generalizations gives rise to the notion of a 'language', though
if we restore focus to spatial and temporal scales, we dissolve
the language again into regional and historical dialects, down to
idiolects. Dialectologists look across spatial scales, but at
small internal-scale units (word-variants, sound-variants) and
find only a topology of geographical and temporal contours of
frequency of use of each feature (which do not necessarily
correlate with one another to define a uniform dialect).
Suppose we do this now for meaningful human behaviors, for
actions-in-activities, for social practices, taking a doing on an
occasion as the analogue of utterance or text (actually an
utterance is an instance of this, and a text is a trace or
product of such a doing-with-language, if not something more
abstract -- meaning-text vs. object-text). I am not going to do
this here; people can think about it. Somewhere in such a series
of constructions, I think, we will find what makes us take the
notion of individual agent as useful, along with the conditions
for its construction, and the limits of its usefulness. JAY.
------------
JAY LEMKE.
City University of New York.
BITNET: JLLBC who-is-at CUNYVM
INTERNET: JLLBC who-is-at CUNYVM.CUNY.EDU