Advances in Spoken Discourse Analysis

56 Advances in spoken discourse analysis

abstract discussion of the nature of linguistic description. Thus for anyone
seeking, as we are, to describe a new kind of data following well-tried
linguistic principles, it is a perfect starting-point.
The first questions one asks of a linguistic description are what are the
descriptive units and how are they related to each other—as we have already
seen these are not questions that are easy to answer for the units proposed
by the Conversational Analysts. For any unit one must provide two kinds of
information: what position or function it has in the structure of other larger
units and what its own internal structure is.
Such information, about the interrelationships between units, can be presented
very simply in terms of a rank scale, whose basic assumption is that a unit
at a given rank—to take an example from grammar, word—is made up of
one or more units at the rank below, in this case morpheme, and combines
with other units at the same rank, that is other words, to make up one unit
at the rank above, group or phrase.
Organizing descriptive units into a rank scale can be part of the heuristic
process; as Labov observes (1972:121),

formalisation is a fruitful procedure even when it is wrong: it sharpens

our questions and promotes the search for answers.

It was their attempt to fit utterance into a rank scale which made Sinclair
et al. (1972) realize that it was not in fact a structural unit and if we try to
create a rank scale from the Conversational Analysts’ descriptive units discussed
above we get similarly enlightening results. One criterion for placing units
at a particular point on a rank scale is relative size and thus we would
expect the following:

sequence

pair

turn

However, in a rank scale, larger units are, by definition, related to smaller
ones in a ‘consists of relationship, and we can in no way pretend that the
Conversational Analysts’ sequence consists of one or more pairs; rather
both consist of two or more turns and thus we realize that structurally,
sequence and pair are varieties or classes of the same unit, with pair being
a label for one subclass of sequence just as transitive is a label for one kind
of clause.
Distinguishing analytic units is only a first step; a description must then
set out to isolate the different kinds or classes of unit at each rank, and
these classes must be distinguished in terms of their structure, the way in
which they are composed of particular units from the rank below in a particular
sequence. For example at the rank of clause one can distinguish four major
or primary classes declarative, imperative, interrogative and moodless according
to the occurrence and relationship between two elements of structure, Subject
and Predicator.