62 performance and the classical paradigm
uniquely determined by each correct performance of W , then verbal speci-
fications of tempo cannot be viewed as part of the score. As a consequence,
they cannot play a part in defining the work, which is the first thing a score
has to do. This is precisely what Goodman concludes (1976, 184–185). And
what applies to verbal indications of tempo presumably also applies to verbal
expressions of mood, such as Sibelius’s instruction “ lugubre .” In ruling that
verbal instructions on a score fail to meet the requirements for notationality
and therefore cannot play a part in defining a work, Goodman commits him-
self to the correctness of performances that we might regard as bizarre – for
example, a performance of the Second Symphony in which the standard tem-
poral and expressive features of the work are freely interchanged between
and within the movements, and where, partly as a consequence, “melancholy”
and “sunny” passages exchange their expressive character.
But Goodman not only counts such bizarre examples as correct perform-
ances of a work. He also maintains that any performance that in the slightest
way departs from what is notated in the score of a work W is not a perform-
ance of W at all, however exemplary it may be in other respects. Again, it is the
requirement that we be able to uniquely identify the score of W , given a correct
performance of W , that leads to this conclusion. Goodman argues as follows:
The innocent-seeming principle that performances differing by just one note
are instances of the same work risks the consequence – in view of the transi-
tivity of identity – that all performances whatever are of the same work. If we
allow the least deviation, all assurance of work-preservation and score-
preservation is lost, for by a series of one-note errors of omission, addition,
and modification, we can go all the way from Beethoven’s Fifth Symphony to
Three Blind Mice. (1976, 187)^16
The conclusion we are intended to draw from this argument – that there
cannot be performances of a performable work that depart even in the slight-
est way from what the work (through the score) prescribes – directly contra-
dicts our assumption thus far that performable works are things that by their
very nature allow of both correct and incorrect performances. It also, thereby,
contradicts the idea that works are, or contain essentially, norm-types, since a
norm-type just is a type that can have both correct and incorrect examples.
4 Answering the Goodman Argument
Why does Goodman insist that, for a score to fulfill what he terms its “pri-
mary function,” it must not only define a work – enable us to determine
whether or not something is a correct performance of the work – but also