Philosophy of the Performing Arts

performance i: improvisation and rehearsal 159
certain features must be present in correct performances of the work. But
any given performance of a work will always be thicker in its appreciable
audible qualities than the work of which it is a performance, because of the
role of interpretation in performance. This, we may recall from Chapter 6, is
one of the problems with the idea of performable works in dance: the notator
attempts to record an actual dance performance rather than prescribing in
advance what the dancers are to do. To constitute something as a musical
work through a process of improvisation, therefore, requires something that
renders determinate which features of the improvisation are to be constitu-
tive of the work. This could be either a prior or simultaneous decision on the
part of the performer, or, more likely, the existence of a set of public conven-
tions that the performer can rely upon to determine what the constitutive
features are. The act of composition in the Bach case must be seen as contem-
poraneous not with the act of improvisation but with the act of scoring. For
it is only at this point that the relevant decisions are made in a binding form.
Others have made similar points. Wolterstorff, for example, points out that to
compose a musical work we must select “that particular set of requirements
for correctness of occurrence to be found in the score,” something the impro-
viser is unlikely to do while in the very process of improvising (Wolterstorff
1980, 64). And Stephen Davies argues that, whether or not Bach wrote, in
the score for The Musical Offering , what he actually performed for Frederick,
he always had the option of prescribing, for that work, something other than
what he performed (S. Davies 2001, 15).
Do these arguments show that improvisational composition is impossible?
They certainly suggest that, if a performer is to be counted as composing
a performable work at the very same time that she is improvising, this is
because she is carrying out two distinct actions at that time – the act of
improvising and the act of composing. Thus the act of improvising per se is
not an act of composing, because the latter requires that some of the fea-
tures of the performed sequence of sounds be selected as prescriptive for
the performable work. However, we can imagine a situation in which what
serves as the act of improvising also counts as the act of composing. This will
be the case if the selection of certain features of the improvised sequence
of sounds as prescriptive is made not by the performer as she is playing but
by certain conventions in place at the time of performance. Such conven-
tions might be generally acknowledged in the musical community to which
the performer belongs, or they might result from a prior decision on her
part. The conventions would have to take the following form: to improvise a
sequence of sounds S is to prescribe features F(S) for instances of a work W.
Indeed, this is precisely what the Goodman argument that we examined in
Chapter 3 assumes is possible for classical performable works – that we can
unequivocally derive a score for a work from a correct performance. If such