72 The Language of Fashion
not, because of this, permanently removed from the history or society
in which it operates).^14 Therefore semiological analysis, even when it
remains methodologically immanent to its object via its acceptance
of the pertinence principle, itself demands a sociological dimension;
however, we cannot then attach it back on to classical sociologies of
‘knowledge’, because the intelligible is not only—or perhaps ever—
directly part of the intellectual; but rather it has to be inserted into the
socio-logic, or sociology of classifications which Durkheim was already
thinking about; so semiology is research into where things are placed,
and not into the things themselves.
having accepted the principle of pertinence and having also selected
the particular sort of pertinence, the next step was to decide on the
corpus of materials on which it was going to be brought to bear.^15
at first, we wanted to analyse the semantics of real clothes that are
worn and not those of Fashion clothing.^16 This project was attractive
because, as we have just said, it allowed us to link up with an authentic
sociology of Fashion by studying its rhythms of renewal and its circuits
of dissemination. But without even having to make a choice between
a sociology of practices and a socio-logics of the intelligible, it was still
difficult to undertake an initial ‘exercise’ in semiological analysis on an
object that was probably full of meaning (real clothing is undeniably
meaningful: humans communicate via clothes, tell each other if they
are getting married, being buried, going hunting or to the beach, if they
are department store staff or intellectuals, if they are doing their military
service or painting), but in which finalities other than those of meaning
remained extremely active (protection, ornamentation, economics), and
which mixed, in a way which was tricky to discern, the individual act and
the social institution, or what is termed, in Saussurian language, parole
and langue. Since the object of the research was essentially the testing
out of a method and not directly the discovery of a sociological truth, it
was probably better to accept a simplification in all its openness, and to
reduce the corpus by restricting it to descriptions of a particular clothing
type, the sort covered in fashion publications and which we will call here
fashion clothing.
But as well, although we still wanted to make sure that we remained
in contact with the future tasks of a sociological study of real clothing,
we restricted our analysis here to Fashion clothing, constituting a