Stephen Davies - Themes in The Philosophy of Music

The Ontology of 4
Musical Works and the

Authenticity of their
Performances1
The authentic-performance movement is a phenomenon of the last fifty years.

Once one could rarely find a recording of baroque music played on the
original instruments and such performances were often lame and faltering
because of the players' unfamiliarity with the instruments and with the
appropriate performance practices. Now, it is difficult to find a recording of
such music not played in the 'authentic manner' and different orchestras vie
with each other in the vibrancy of their interpretations.
Not surprisingly, the authentic-performance movement has raised a hue
and cry among performers and musicologists, since it challenges entrenched
traditions of performance. Very recently some philosophers have also turned
their attention to the subject of authentic performance. The ontology of
musical works has also attracted the interest of philosophers in recent
decades.
There is an important connection between any theory of the ontology of
musical works and a specification of the characteristics that must be exhibited
in an authentic performance of a musical work, though this connection has
not received much comment in the literature (but see Levinson 1987). If an
authentic performance is (at least) an accurate performance of a work, then
theories of musical ontology should tell us the type and range of properties
1
First published in Nous, 25 (1991), 21-41.
Ontology of Musical Works and Authenticity of their Performances 61
that must be produced in an authentic performance of a work. In the main
part of this paper I hope to characterize the debate about the ontology of
musical works in a way that draws out this connection.
I
Opponents or critics of the authentic-performance movement have made
points such as the following against the use of musical instruments and
performance practices from the work's historical period: (a) The attempt to
produce authentic musical sounds and styles has often resulted in dull, lifeless
performances. Many features, other than literal-minded, mechanical accur-
acy, make for good performances. The use of unfamiliar instruments and
styles can inhibit fluency and spontaneity, which are hallmarks of vital
musical performances (Kivy 1988c; Taruskin 1988). (b) Even if we could
reproduce the sounds of the work as these might have been heard at the
time of its composition, authenticity, as the recreation of the experience of the
work shared by the composer and his or her contemporary audience, is
unattainable, because we cannot reproduce the physical, social, cultural,
and historical context of the composer's time (Dipert 19806; Young 1988).
The way we hear music has been affected by the changing history of music;
we cannot bridge the gap separating us from the past. Our understanding of
the work may be better than that of the composer and of her contemporaries,
because we, unlike them, can place the work within the historical tradition
binding it to its future, as well as to its past. To sum up: The type of
authenticity so many performers take as their goal is impractical (indeed,
impossible) to achieve and undesirable.
What is it that explains the appeal and success of the authenticity move-
ment? Some authors deconstruct the notion and thereby discover (lo!) that the
movement is a modern one, offering the attraction of novelty (Taruskin 1988;
Tomlinson 1988). (The appeal to deconstruction is often used unselfcon-
sciously, with neither a suggestion that deconstruction is itself a new and
fashionable theory, nor a hint, whether of glee or embarrassment, that one
might deconstruct the theory's own foundations.) These writers see 'authen-
tic' performance as a modern style of performance (no less reconstitutive of its
object than have been other styles of performance) that claims for itself an
illegitimate superiority through its invocation of the imprimatur of the
composer.
It is sometimes said that performers should strive for a different type of
authenticity—for the compelling vibrancy that brings life to a work (Kivy
62 Themes in the Philosophy of Music
1988c). (Such a performance might follow from a mastery of old instruments
and a sympathetic commitment to works from a certain period, but in that
case the 'authentic' approach is justified as the means to an end, and neither
as an end in itself nor as a means superior to a variety of alternative
approaches.) There is no single, ideal performance of any work—performing
must be creative if it is to be convincing. Performers might allow themselves
to be advised by composers' intentions, where these are known, but they
should not sacrifice their creative autonomy to the fixed will of the composer,
for without the exercise of that autonomy performance reduces to the bare
transmission of characterless notes.
Against the view outlined above one might argue as follows: Our aesthetic
interest in music in general and our favoring music above sounds that occur
naturally as well as above sounds specified and produced by us for other
reasons presuppose that composers, more often that not, succeed in writing
artistically interesting works, and that they do so not by chance but by design.
The musical work is known to us through its performances. The composer
needs the services of the performer if her ideas are to be publicly presented,
and the composer relies on the performer to respect what has been specified of
the work in question if the work as specified is to reach an audience of non-
performers. For this reason, the so-called 'intentional fallacy' is no fallacy at
all in so far as it applies to the performer. The performer can be intending to
perform the work in question only when intending to perform what is consti-
tuted as the work by its composer. The hearer might dismiss the composer's
intentions as worthless and refuse to allow his response to the work to be
ruled by those intentions, but the performer, in order to perform the work in
question (and not to improvise or fantasi/e on that work instead), must be
dedicated to preserving those of the composer's intentions that are determin-
atively expressed and that identify the work as the individual it is.2
From this it does not follow (as is so often implied) that the performer's
creativity is compromised by her pursuit of authenticity in performance
(Davies 1987, 1988a). This would follow only if the composer's specification
exhaustively determined every aspect of the work, so that a performer would
have to do no more than copy the work from a recipe supplied by the
composer. Plainly this is not the case where performers work from notations.3
2
Not everything recorded in the score need be determinative. Conventions of performance
practice and score reading, as well as composers' intentions, set the standards for determinativeness.
For a fuller account of this issue see Davies 1987.
3
Some contemporary composers and transcribers have attempted to make notations as specific
and detailed as possible—see Bartok's transcriptions of Hungarian folk music, where many supple-
What is specified by the composer underdetermines the sound of any accurate
performance of the work. If'authentic' means 'accurate', then many different-
sounding performances could be equally and thoroughly authentic. More-
over, because the performer's contribution to the work's realization is by no
means fully determined, authenticity and creativity in performance will be
complementary rather than exclusive. If one cannot perform the work at all
except by exercising one's creative skills as an interpreter and reali/er of the
material provided by the composer, then one cannot perform the work
authentically except by being creative.4
The fact that performance is creative explains the reluctance nowadays to
talk of the performer as owing a moral duty to the composer (see Dipert 19806
and cf. Kivy 19886) or (though this is rarely considered) to the audience that
relies on performers for access to the composer's work. Talk of performers'
duties as correlative with composers' or audiences' rights, whether the duties
be 'moral' or not, seems inappropriately restrictive, given the creative free-
dom that is essential to the performers' fulfilling their role. But whatever
difficulties there may be in the terminology, still there is an important notion
such talk aims to capture. Where musical works exist and where audiences
attend performances in order to hear those works, the first aim of the activity
of performance is to deliver the work in question to the audience (and a
crucial further aim is to do so well). To meet these aims the performer must
exercise her creative talents within bounds prescribed both by the composer
and by the wider conventions of the composers' day that governed the per-
formance of works of the type in question. Performers and audiences come
together on the basis of an understanding of the point of the activity in which
they are jointly involved. Players who are not prepared to direct their talents
to the delivery of the work are unilaterally rejecting the enterprise in terms of
which they have come together with the audience. If the musicians are
professional and the conventional or contractual circumstances make the
purpose of their employment clear, then a failure to focus their efforts in
the appropriate way might well involve the dereliction of a moral duty. If the
musicians are amateurs, then, still, the activity would be misrepresented as a
performance of a given work unless it were a part of the performers' intentions
mentary notational symbols are used. Simply, I doubt that any written notation can fully specify
every aspect of the sound of an accurate performance of it. The standard musical notation certainly
does not.
4
Admittedly the notion of creativity appealed to here is minimal and is consistent with thought-
less, even mechanical, playing. The creative element in most performances may often go beyond this
minimal level, though this is not always the case.
to do what is necessary, given their capabilities, to deliver the work. Within
the tradition of 'classical' music, to aim at music making is usually to aim at
performing particular musical works, and to aim at this is willingly to accept
constraints on the exercise of one's freedom.
Is the above equation of authenticity in performance with accuracy in the
presentation of the individual musical work justified? Authenticity is a rela-
tive notion, so we might always ask: authentic with respect to what? Perform-
ances of musical works might be authentic (or not) with respect to many
possible factors—the dress of the musicians, the physical environment within
which the performance takes place, the size of the audience, the price of
admission, the work being played. Clearly the pursuit of authenticity in
performance is selective, and it is so with a very particular purpose in mind.
Where music making takes place more or less in the absence of particular
musical works, authenticity is concerned with styles of playing. Where an
interest in authenticity follows from a concern to present an authentic per-
formance of a musical work, authenticity is aimed at delivering what consti-
tutes the work as the individual it is. An authentic performance of a work
might aim to be authentic in further respects; for example, it might (also) aim
to re-create the physical environment of the work's first performance (e.g.
where a film is being made of the composer's life). But this further kind of
authenticity, the authenticity that goes beyond delivering the work itself, is
not required in the standard concert setting. What we require from an
authentic performance of the work is a performance that is accurate in the
sense that it truly represents that in virtue of which the work is the individual
it is. It is for this reason that I have equated authenticity with accuracy. An
interest in the work being performed is primary in that it gives point to the
activity of performing musical works (as opposed to music making in
the absence of musical works, or for the sake of historical reconstruction,
etc.). So it is that the notion of authenticity in the performance of particular
works is centrally and importantly an interest in accuracy in performance.
Now, though, having allowed that the goal from which performance takes
its first aim is that of faithfulness to the composer's determinatively expressed
musical ideas, it is only fair to concede that performance serves a variety of
goals. When a work is familiar and often performed, the attempt to approach
the work in a fresh and unusual manner might become desirable. After all, the
composer presumably intends that performances of her works be interesting,
as well as faithful, and what an audience will find interesting in a performance
depends on what they already know of the work. That is to say, where the first
goal of performance already has been realized in other performances (and the
audience is familiar with the work in question), other purposes of perform-
ance rightly come into greater prominence (see Levinson 1987).
What of the point, often raised, about the inaccessibility of the significance
of past events, given the number of beliefs and experiences that are no longer
common to us and our musical predecessors? The issues raised by this ques-
tion are subtle and complex, but I shall comment only briefly. Such an
objection to the project of authenticity will work only if there are insurmount-
able differences between us and our predecessors and only if those differences
are such as to affect totally the experience of the musical work qua the work
that it is. No doubt there are many respects in which our experiences of music
are bound to differ from those of its contemporary listeners, and no doubt
many of those differences are ineradicable, but to allow this is to be far from
having to accept that authenticity in performance is undesirable and unattain-
able, for it is not obvious that interpersonal judgments of authenticity are
rendered impossible by just any disparities in the experiences of different
listeners (Davies 19886).
II
Different theories of the ontological character of the musical work describe it
as variously thin or rich in properties. At its most spare, the musical work is
said to be a sound structure of (timbre-less) rhythmically articulated notes, or
a relationship between notes, or some combination of these two (Goodman
1968; Webster 1974; Cox 1986). (To use Webster's example, on the view that
the work is a set of pure pitch relationships one would be performing the Bach
E major Violin Concerto just so long as one preserved the appropriate note
relationships, which one might do by playing the piece in B major with piano
and sousaphone.) At the other end of the spectrum is the view that it is
essential to the musical work's being the piece it is that it possess a sound
structure with tempo, timbre, etc. that must be produced by the playing of
certain types of instruments and that must have been composed by a particu-
lar individual at a particular time and place (Levinson 1980; see also Walton
19886). (For example, on this view one would be performing the Bach E
major Violin Concerto accurately only if a violin and orchestra such as is
specified in Bach's score were used to produce the sound of the work and only
if a causal thread might be traced between what one was doing and Bach's
having composed that work at a particular time and place.) Between these
poles, alternative views are possible—for example, that the work is a sound
structure with a certain tempo and timbre but that the means by which such a
sound structure is produced are not part of the identifying features of the work
as such (see Kivy 1988a). Whether or not the work is taken to include
expressive properties depends on what one takes to be the musical substrate
of such properties. If pure note relationships can be expressive, expressiveness
might be a property of the work according to the thin characterization; if they
are not, it will be performances or interpretations, rather than the work itself,
that are expressive (see Pearce 1988). Alternatively, if expressiveness derives
as much from the manner by which sounds are produced as from the sounds
themselves, only the thick characterization of the work could include expres-
siveness among the work's properties (Levinson 1990a).
The dispute between theorists with different views on the ontology of the
musical work takes a number of forms—whether an analysis of the concept
should be ruled by ordinary language, as opposed to an analytical perspicuity
being necessary to penetrate the confusions of such ordinary language (Good-
man 1968); whether two composers who independently produce specifica-
tions that would be interpreted as generating identical sound structures
(produced by identical performance means) have composed one or two
works (Levinson 1980; Anderson 1982); whether any work might have been
composed by a different composer, or at a different place and time; whether a
musical event that aims to preserve no more than the thinly characterized
sound structure of a work is really a performance of the work; and so forth.
(For the most part discussion has centered on musical works of the type
written from c. 1650-1940. Surprisingly little has been said about works such
as Cage's 4' 33" or pieces in which chance plays an important function; but
see Ziff 1973, with responses in Sircello 1973 and Walton 1973; see also
Tormey 1974 and Cavell 1976.)
One part of the current debate concerns whether musical works are dis-
covered or created. If a work exists (between the times of its performances) as
the possibility of its production, then it also exists prior to its composition just
so long as it is logically possible that it might be instanced prior to the time of
its composition. And if it exists prior to the time of its composition, then the
composer must discover, rather than create, the work. Thus, if the musical
work is characterized as a thin sound structure (and anything that reproduces
that sound structure is an instance of the work, if not a performance of it),
then the work might be instanced prior to its composition and must exist for
all time, since it might be instanced at any time (Wolterstorff 1975; Cox
1985). Partly in reaction to such a view, the thicker characterization of the
work, as necessarily including a performance means and as necessarily being
indexed to a person, time, and place, rejects the claim that the work exists
eternally (Levinson 1980). This view ties the work into the world of time and
space, so allowing that the work is created and not discovered—and, hence, it
rejects the idealism of the alternative view. That is, the argument specifies that
an instance of the work must be a performance of it and that performances of
the work become possible only from a particular time. The reply to this
argument might take different forms: Simply, one could reject the inclusion
of performance means within the account of the work's ontology and thereby
allow for the possibility of instances of the work that are not performances of
it (Kivy 1988a); or one could prize the work free of the world by arguing that it
might have always been composed by another person, at another place, or at
another time, and so might have always been performed at some time before
its actual but contingent time of composition; or one might argue that the
work still exists eternally as a possibility prior to its composition, even if (as a
contingent fact) that possibility could be realized in this world only with the
birth of (for example) Beethoven, with the realization of the possibility of
the instrument we call a piano, with the realization of a particular cultural and
musical context, and so on (Kivy 1987; Walhout 1986). And, to complicate
what is already a complex issue, one might argue about the difference, if any,
between creation and discovery and about what is supposed to hang on that
difference (see esp. Kivy 1987).
Ill
The connection between a work and its instances has been characterized on
the model of a class to its members (Goodman 1968), a kind to its instances
(Wolterstorff 1975), and a type to its tokens (Wollheim 1980). The difference
between these analogies is not always as clear as it might be, but might come
to this: A class is the collection of its instances and does not usually share
many properties with its members; a kind stands as a concept the prepos-
itional content of which (subjunctive conditional) specifies the nature of its
instances without its being a collection of those instances; a type is an abstract
individual that possesses and shares the definitive properties of its tokens.
If one thinks there may be such a thing as an imperfect performance of a
musical work—something that misrepresents some characteristic of the work
though remaining recognizable as an instance of the work—then one might
introduce the suggestion that the relationship between the work and its
instances is normative rather than descriptive (Wolterstorff 1975; Anderson
1985). This view tends to be associated with the account of musical works as
kinds, but I can see no reason why the alternative views might not avail
themselves of some such notion. Thus, the class that constitutes the work
might be the subset of those of its performances that are correct in all relevant
respects; the norm kind might specify the properties that a correct perform-
ance should have; the type may be betokened by more or less well-formed
tokens.
On any view, the work determines or exemplifies the properties its in-
stances must display in order that they be instances of it; ontologically
speaking, it is the nature of the work that determines those properties of its
instances by virtue of which they are its instances. The epistemic process goes
in reverse, however. We come to know the work through its performances.
We abstract the work from its instances, stripping away from its performances
those of their properties that are artistically irrelevant, and then stripping
away those artistically relevant properties that are properties of the perform-
ance but not properties of the work, thereby exposing the work and its
properties.
Even if one does not know what properties Beethoven's Fifth Symphony
has except by recovering them from performances of the work, one could not
recover the work without the aid of some theory about the ontological status of
musical works (or of musical works of this type). One can distinguish the
irrelevant from the relevant properties only in terms of a theory establishing
criteria for relevance. Theories of musical ontology are a priori in this sense:
our acquaintance with musical works is indirect, mediated, and we can
separate the message from the medium only in view of a conception of what
it is that distinguishes the two. Unfortunately, the range of theories presented
in the literature suggests there is little agreement at the level of the intuitions
grounding the relevant a priori judgments.
I offer just one example by way of illustration: R. A. Sharpe (1979) has
denied that performances stand to musical works as tokens stand to types; if
they are tokens of anything, he concludes, they are tokens of interpretations.
Sharpe arrives at this conclusion by suggesting that it is a feature of the tokens
of any given type that their equivalent parts may be interchanged without
their status as tokens being impaired—a part of a linen flag might be replaced
by an equivalent part of a plastic flag and one would still have a flag, he
suggests. But, so continues the argument, parts of different interpretations of a
musical work are not similarly interchangeable. What are interchangeable,
instead, are parts of performances that are interpretationally consistent the
one with the other.
One might challenge this argument on a number of grounds, (a) One could
begin by pointing out that there is no bar to a single item's being at the one
time a token of more than one type, so one cannot show that performances
are not tokens of musical works by showing that they are tokens of interpret-
ations of musical works, (b) Or, pointing out that internal interpretative
consistency is not a necessary condition for something's being a performance
of a given work, one might suggest that Sharpe is wrong in denying
the possibility of substitution between different interpretations (Kivy 1983).
After all, an internally inconsistent performance is often played without any
substitution having taken place! (c) Or, one might question the claim that
intersubstitutability of parts is a definitive test of a common betokening
function. This final criticism ties the objection to the point made above—
what one takes to be a token (or class member, or kind instance) depends on
one's view as to the nature of the type in question (Dipert 1980a). Whether
the, the and the, all are tokens of the same type depends on what one takes the
type to be—they all are tokens of the definite article but they are not all tokens
of a single typeface. Sharpe's objection to the type/token account of the
work/performance relation reveals an implicit commitment to a theory
about the nature of the musical work (as well as of the type/token relation).
IV
Already I have emphasized that one could abstract the work from its authen-
tic performances only in the light of a theory about the nature of musical
works. Granting that, how does one do it? A crude but tempting answer,
perhaps, is this: Find the lowest denominator common to all authentic
(accurate) performances of the work, discard those common factors that,
according to one's theory, are not relevant to its identity—that all perform-
ances took place in the evening, for example—and what one has left is the
work. Reflection suggests that this approach is mistaken, however. If every
element of the work were determinatively fixed, presumably some such
procedure might succeed, but if the work contains elements that are variable,
with only the limits of possible variation fixed, then the lowest common
denominators underspecify the work. For example, where the work contains
a figured bass, the only elements common to accurate performances of the
work might be the melody, the bass-line, and a harmonic structure between
the two. But that does not mean that the realization of the figured bass is not
part of what gives the work its identity. Even if different realizations of the
figured bass are possible, so that different (but equally accurate) performances
of a given work contain different realizations of its figured bass, an essential
part of the work might be the fact that its middle parts be realized
in accordance with quite definite sets of conventions. This suggests that in
determining the identifying features of any particular work we need to look
beyond the level of common factors and include variable elements where
there is a pattern to their variation from performance to performance.
Those who would concentrate their attention exclusively on the musical
parameters (pitch, rhythm, texture, instrumentation, timbre, etc.) common to
a work's accurate performance are likely to favor a thin characterization of
the ontology of the work, because, at that level, the common factors may not
go far beyond the notes and the relationships between them. And such
theorists are likely to regard conventions allowing for variations in perform-
ance as matters of musical style, where style is a characteristic of schools and
movements rather than part of what gives any particular work its identity. On
the other hand, those who favor the thicker characterization of the musical
work are likely to regard the style of the work's proper playing as essential to
its being the work that it is. Accordingly, they will tend to include in their
account of the work's ontology, as well as the lowest common denominators,
the patterns and limits of allowable variation. Where such patterns are
common to a number of works (as they are likely to be, given that the
conventions are not usually codified), they constitute a style.
The emphasis I have placed both on conventions of performance and on
conventions for the transmission of the composer's work-determinative in-
tentions to the musician who will execute the composer's work will strike
some people as too insecure a basis for an account of musical ontology.
Artistic conventions are not more than rules of thumb, and the history of
art just is the history of the overthrow and alteration of such conventions.
How could the conventions secure the work unless we have a check on what
they are and how they are being used? And how, without something such as a
score, could we draw the crucial distinction between the composer's making
a mistake in accidentally breaking some convention and the composer's
deliberately altering some convention? The reply to such questions is two-
pronged: (1) It is not the case that wherever music puts aside or minimizes the
role of notation we get styles of music making without thereby getting
(performances of) musical works. In the absence of highly developed systems
of notation, there is a tendency for musical works to become simpler and for
improvisational and performance skills to become more important for their
own sake, but this tendency is neither necessary nor universal. Javanese
notation is far less detailed and complex than is orthodox Western notation
(and the general run of Javanese musicians never have occasion to refer to this
notation), but there are long and intricate individually named works for the
gamelan orchestra. Many of these works have been in the repertory for
hundreds of years. The survival of long and complex musical works largely
in the absence of a notation is made possible by the fact that the conventions
of performance are complex, stable, widely understood, and generative in
nature (in that the widely differing parts for various instruments each can be
derived, in terms of the convention appropriate for that instrument, from the
work's melodic foundation). The first point then is this: There can be a
tradition of performing and preserving individual musical works, some of
which may be complex and prolonged, in the absence of a complex musical
notation. (2) Where a complex musical notation exists, the manner in which
it should be read is governed by conventions that may be invisible only
because they are so familiar to those at home with the notation. As well as
conventions for reading the score, there are conventions for going beyond
what is given in the score—decoration, double-dotting, a preference for
stopped rather than open strings unless the contrary is directly indicated,
fingerings, the method for realizing a figured bass, etc. etc. Whether some-
thing is recorded in the score depends on how well known and widespread
various of the conventions are—the composer does not always spell out the
limits to the performer's freedom, since those limits are established already
within the musical culture, period, and style. Because it is contingent whether
or not some particular part of the work (or of the manner of its performance)
is recorded in the notation, I believe there is no reason for insisting on a sharp
division between the score and the conventions controlling performance
practice with respect to such scores, no reason for confining the work to
what is notated and dismissing the rest as a matter of style that could play no
essential part in shaping the identity of the individual work.5
How are mistakes in composition to be distinguished from innovations? The
existence of a score guarantees nothing. What is printed in the score might be a
mistake (type-setting error, copying error); or, even if the score correctly
records what the composer wrote, what the composer wrote might contain
an error (for example, a slip of the pen, such as a failure to cancel an accidental
with a natural within the same bar). So, again, how can we separate com-
posers' innovations from errors, given that the standards of correctness are set
5
Goodman (1968) does make such a division because he argues that if the score specifies the
work univocally and recoverably, it can do so only if it meets various syntactic requirements (that
would not be met by conventions of the type I mention). He is happy to depart from ordinary usage
in denying that performances differing by a single note cannot both instance the same work; to do
otherwise, he thinks, is to undermine the notion of the transitivity of identity. On similar grounds he
denies that the verbal language of tempo (allegro molto, etc.) is notational. For a discussion see
Boretz 1970; Goodman 1970; Kulenkampff 1981.
only by mutable conventions? The answer: By seeing whether composers
repeat themselves, correct themselves when their attention is drawn to the
matter, teach their students the same procedures, describe the matter in their
theoretical treatises, and so on and so forth.6 Obviously, careful attention must
be paid to the background of musical practice against which the composer and
performers work and, in particular, to those conventions that stand fast for the
work (or type of work, or musical period) in question. (An interesting case is
that of the Javanese gamelan orchestra, where 'wrong notes' are standardly
played so that the gods will not be offended by the pride displayed by humans
who act as if they believe that they might attain perfection. Being a musically
fastidious people, there are conventions within the performance tradition that
govern what wrong notes will be played and the instrument that will play
them. In this case an authentic performance would have to contain 'wrong
notes'.)
Could the account offered above explain radical rejections of or alterations
in musical conventions, such as were involved in the overthrow of the modal
system, or in the move to twelve-tone technique? Yes, and for two reasons. (1)
Despite what is often said, it is not the case that whole systems of conventions
are overthrown at a single bound. Systems of conventions are eroded (and
restructured) rather than being dumped holus-bolus. The discontinuities are
very marked to those who stand near. With the passage of time and the wider
perspective that is thereby created, we often become increasingly aware of the
continuities that had always tied the new movement to the heritage against
which it reacted. (2) The accumulation of small changes can, in time, produce
wholesale alterations. Moreover, even small changes might make possible
spectacular aspect-shifts, so that the tiniest innovations could turn one's
musical world on its ear.
It is common to suppose that acquaintance with artworks comes from first-
hand experience and that, in the musical case, the experience (for the audi-
ence, if not the composer) will be an experience of performances of the work.
But if a person can become acquainted with a musical work solely by reading
its score, the second part of the conjunction is false. And if a few people can
experience a musical work solely from reading its score, perhaps they can
experience it in the following case also: Several people call out the pitch
6
When Jane Torville and Christopher Dean competed in the Olympic Games in Calgary in 1984
they appeared to make a mistake and the shocked crowd gasped. The dance, which followed a
prescribed pattern, was twice repeated. On the first repeat, when the questionable pattern of
movement recurred, some people gasped again. On the second repetition no one gasped, because
everyone recognized that they were seeing an innovative and risky step, not a mistake.
names of notes and their relative durations, the whole being preceded by an
announcement that the work is to be played on the piano (Carrier 1983).
Under such circumstances what one has, I think, is a 'performance' of the
score rather than a performance of the work. Nevertheless, it may be possible
for some people to recover the musical work specified by the score from such
a 'performance'.
V
An inauthentic performance is a performance that misrepresents the work of
which it is a performance while remaining recognizable as a performance
of the given work, despite its inaccuracies. The possibility of inauthentic
performance presupposes the possibility that mistakes in performance can be
recognized as such. I explained above how this might happen: (a) An audience
familiar with a work might recognize the way one performance of it differs
from others, Or (b), an audience familiar with performance conventions
appropriate to the work in question might recognize that those conventions
have been violated and might also come to know that this was not intended by
the composer and, hence, that the violation was an accident of the work's
performance. Sometimes, of course, one might suspect that an error has
occurred but not know if it is an error made by the performer, the printer of
the score, the composer, or if in fact it is an error at all.
How does one recover the work from its inauthentic performances? To do
so, it must be possible not only to detect errors as such but also to determine
what would have been correct. Very often this is possible. Most people, I am
sure, can tell in some contexts not only that a note has been sung wrongly but
also what note should have been sung instead. One might make the general
point as follows: Musical works are very complex. One kind of atomic unit of
musical content—the unit an alteration in which might make a difference to
the musical sense of any given passage—is the pitched tone.7 (I allow that
7 It is my intention here to indicate a musical unit equivalent in status to that of the phoneme in
language. The danger of such an approach, of course, is the temptation to draw too close a parallel
between music and language—to describe music as a semantic system generating 'sentence-like'
units of meaning from the combination of 'word-like' units according to rules of musical syntax.
Like so many others, I reject the view that music is a semantic system, which is one reason why I call
the atomic units 'phoneme-like' rather than 'word-like'. As I use the analogy, units of musical
meaning are combined to form patterns with musical significance; the patterns are significant in that
someone who understands music must recognize and appreciate such patterns. The crucial disana-
logy lies in the fact that those musical patterns no more have semantic content than do the marks left
by a snake as it travels across the sand.
duration, timbre, and dynamics each might also have their atomic units
of musical content.) The level at which musical significance arises is that of
themes, motives, ostinatos, chords, etc. Many atomic units contribute to the
creation of any unit of musical significance. It is always possible that some
atomic units be mischaracteri/ed without this resulting in the destruction of
the molecular level of organization at which musical significance begins to
operate. (Indeed, were this not the case, one could never have the same theme
in both a minor-key and a major-key version, or in an embellished version,
etc.). Because there are conventions for the generation of units with musical
significance from the atomic units, it is possible not only to tell when an
atomic unit has been mischaracteri/ed, but also to tell, within the wider
context of musical significance, what that unit should have been. But, having
said all this, it is obvious that the business of recovering a work from its
inauthentic performances is likely to be less secure than that of recovering it
from its authentic performances.
Could one recover Bach's E major Violin Concerto from a performance
played by piano and sousaphone, given that all the notes were played to
tempo and given that one was told that the work was written for the baroque
violin? Some people might do so. But notice that it is not a requirement of
something's being an inauthentic performance that the work be entirely recov-
erable from the performance. What is required is that sufficient is recoverable
to disambiguate the performed work from others. That is, a performance may
be recognizable as a performance of a particular work even if it is not possible
to recover all of the work from the performance. So, even if one could not
easily appreciate how Bach's work would sound for the violin if one heard it
played on a sousaphone, it does not follow from that fact that it is not a
performance of the work that one is hearing.
VI
At least one of the connections between musical ontology and musical
authenticity should be obvious by now: If an authentic performance is an
accurate performance, what is to count as an authentic performance depends
on presuppositions about the ontology of musical works, since the ontology
determines what it is that constitutes the work as the individual it is and an
accurate performance is a performance that reproduces all that is constitutive
of the work's individuality. That is to say, not only is the notion of a musical
work an artifact of theory, so too are the notions of a performance and of an
authentic performance.
Notwithstanding what I have just written, to the extent that the theorists of
musical ontology aim to characterize our intuitive notion of the musical
work, it should be possible to test their theories against our intuitions and
against the terms in which we identify and discuss musical works and their
performances. Some philosophers might regard ordinary language as hope-
lessly sloppy and treat philosophical analysis as prescriptive rather than de-
scriptive. In such cases there is always a difficulty in our accepting that we
should abandon our intuitions and ways of talking for the sake of a philosoph-
ical theory. Most philosophers, though, do take themselves to be analyzing our
present concepts and if it turns out that our concepts are sloppy and obscure,
then it will be part of the philosopher's job to map the limits of those obscurities
and point to the source of our conceptual sloppiness. The majority of the
philosophers who discuss musical ontology take themselves to be performing
some such descriptive role and it is possible (therefore) to test their views
against our shared intuitions—in theory.
Nevertheless, in practice that test seems not to get us anywhere fast—as
I indicated at the outset, there are no widely shared intuitions about the
nature of musical works entrenched firmly enough that the philosophical
debate about musical ontology can be easily resolved. However, in view of
the connection between ontology and authentic performance for which I have
argued, it is possible perhaps to reconsider the debate about musical ontology
in terms of our intuitions about musical authenticity, so long as those intu-
itions are firmly based. The issue is not one about whether authenticity in
performance is desirable; neither is it one about whether we can experience
authentically performed music as it was experienced by its composer's con-
temporaries. Rather, the issue is whether we are agreed on what is involved in
aiming at authenticity. If we are, it might be possible to draw inferences from
the agreed facts that reflect on the debate about musical ontology.
Are we so agreed? It seems to me that there is a considerable measure of
agreement about what is involved in aiming at authentic performance for
some kinds of music, at least. Performers have consistently tried to achieve
authenticity by the use of the instruments for which the composer wrote, by
the adoption of styles of playing and by the adoption of the performance
practices for reading and interpreting notations that held at the time of
composition, and so on. If we consider the kind of ontology presupposed by
such a view of authenticity, it appears we must favor a thicker rather than a
thinner characterization of the nature of the musical work. If the use of the
appropriate performance means is important not simply because other means
of producing the appropriate sounds are not available, but in the fuller sense
that the use of those means is integrally and inescapably a part of the process
by which authentic performance is achieved, then musical works are not
viewed as pure sound structures.
The point I have made above is an important one, I believe, in its implica-
tions for the correctness or otherwise of different accounts of musical ontol-
ogy. But, by focusing on a narrow area of performance and musical history,
its significance might be easily overestimated. If we take a wider perspective
and consider the way an interest in music might range from a concern with
music making in the absence of musical works, through an interest in music
making with musical works minimally important and the stress on the impro-
visational skills of the performer, and finally to music making with a complex,
more or less determinate musical work as the primary object of interest and
the music making the means by which that work is presented, what emerges,
I think, is the realization that the notion of authentic performance has no
single, fixed essence. The more it is that the musical work drops out of
account—for example, because it exists as no more than a cipher the per-
formers must expand and develop in the creation of a performance—the more
it will be the case that authenticity in performance is concerned with faithful-
ness to styles of playing rather than to the work itself. The more it is that the
musical work is sufficiently complex and stable to become the focus of
attention—for example, because it is recorded by means of a sophisticated
notation or because conventions for performances are sufficiently complex
and detailed to allow for the preservation of the individuality of long pieces—
the more it will be the case that authenticity in performance is concerned with
faithfulness to a determinative text. Moreover, because musical conventions
are mutable, as are complex systems of notation, what it is that can be
determined by the composer and the conventions as the text of a musical
work will be relative to the time of the work's composition. Accordingly,
what it is that can be required in the name of faithfulness from a performance
of a given work will depend very much on the work's period.
Given this wider perspective, what emerges, I suggest, is the idea that the
criteria for authenticity in musical performances are variable. In some cases it
is essential, if authenticity is to be achieved, that particular types of instru-
ments are used, because the use of those instruments is specified by the
composer and the performance practice of the day treats such specifications
as determinative. In other cases, a variety of instrumental ensembles might be
employed in different, equally authentic performances, because the conven-
tions of the day allowed the composer to determine nothing more definite
than a range of possibilities. By the mid-nineteenth century, notated phrasing
and dynamics are determinative and must be observed in an authentic per-
formance; in the mid-eighteenth century these musical parameters are variable
within wide limits and notations of them had the status only of recommenda-
tions. In some extreme cases, the limits of choice permitted to the performer
are perhaps so wide that the work appears to be no more than a pure, timbre-
less sound structure that might be realized authentically on a synthesi/er, since
not even an historical limitation on the work's instrumentation would be
recognized by its creator and the musical culture within which that composer
worked.
If what I have said on the basis of this wider perspective is correct, it might
appear to be appropriate to draw the paradoxical conclusion that all the
theories of musical ontology I have mentioned are correct and that none
of them is. More carefully, the moral to draw perhaps is this: The totality of
musical works from culture to culture and from time to time do not have any
single ontological character. Some musical works are thick with properties,
others are thinner—some works include the performance means as part of
their essential nature, and much more besides, whereas others are more or
less pure sound structures.
Part Two
Performance
Authenticity in 5
Musical Performance1
In this chapter I discuss musical performances and their authenticity with

respect to the independently identifiable musical pieces of which they are
performances.2
The adjective 'authentic' has a number of meanings that no doubt are
related. But I am not here interested in the unity of the concept, nor in the
relative primacy of these different meanings. Nor shall I discuss one familiar
notion of musical authenticity—that in which a performance is authentic
with respect to a style or genre. My limited interest is in the authenticity of
musical performances as performances of particular compositions (which
are independently identified with event specifications that, in the case of
the Western cultural tradition on which I shall concentrate, take the form
of musical scores). That is, if I talk of the authenticity of a performance of
Beethoven's Fifth Symphony, I am interested in its authenticity as a member
of the class of performances recognizable as performances of Beethoven's
Fifth Symphony and not with it as a member of other classes of per-
formances to which it may also belong, such as that of the nineteenth-century
symphony.
The view for which I argue characterizes authenticity in musical perform-
ance as follows: A performance that aims to reali/e the composer's score
faithfully in sound may be judged for authenticity. A performance of X is
1
First published in British Journal of Aesthetics, 27 (1987), 39-50. Reprinted in Alex Neil and
Aaron Ridley (eds.), Arguing About Art: Contemporary Philosophical Debates, New York: McGraw-Hill,
1994, 62-73, and in Patrick Maynard and Susan Feagin (eds.), Aesthetics: Oxford Reader Series,
Oxford: Oxford University Press, 1997, 228-34.
2
Though it might be argued, for example, that rehearsals are not performances, this is a subtlety
I ignore.
more authentic the more faithful it is to the intentions publicly expressed in
the score by the composer (where those intentions are determinative and not
merely recommendatory of performance practice). Because the composer's
score underdetermines the sound of a faithful performance, the authenticity
of any particular performance is judged against (the appropriate members of)
a set of ideally faithful performances. As a commendatory term, 'authentic' is
used to acknowledge the creative role of the performer in faithfully realizing
the composer's specifications.
The chapter is divided into six sections. The first four concentrate on the
aim of faithfulness in securing authenticity; as well as an attempt to define
authenticity, these sections contain a characterization of what is involved
in faithfully realizing a composer's intentions. In the penultimate section
I discuss why authenticity in musical performance is value-conferring. In
the final section I emphasize the creative nature of the performer's role.
I
In this first section I argue that the pursuit of authenticity involves the attempt
to produce musical sounds as opposed to the social milieu within which those
sounds were originally created.
Over the past fifty years there has been a growing interest in authenticity in
musical performance. The same period has also seen a developing interest
in the performance of premodern music. These parallel developments are
probably related. Where modern music is written for modern instruments and
notated in the standard fashion, a high degree of authenticity will be achieved
in performance by a competent musician. But the more foreign the styles of
performance and the more unfamiliar the instruments employed, the harder
will it be for musicians to produce authentic performances without the benefit
of scholarly advice and instruction.
A moment's reflection shows that the pursuit of authenticity in musical
performance has been highly selective. The price of admission, the dress of
the audience, the method by which the program is printed—each of these and
much else in the context of music's performance is decidedly modern. The
search for musical authenticity takes a very particular direction. A highly
authentic performance is likely to be one using instruments contemporary to
the period of composition (or replicas of such instruments) in its performance,
involving an interpretation of the score in the light of stylistic practices and
performance conventions of the time when the work was composed,
Authenticity in Musical Performance 83
employing ensembles of the same size and disposition as accord with the
composer's specification, and so forth.
The selectivity displayed in the search for authenticity in musical perform-
ance has been systematic in a way suggesting that the quest may be character-
ized as aiming at the production of a particular sound rather than at th/
production of, for example, the social ambience within which the music
would or could be presented by the composer's contemporaries. This point
is effectively illustrated as follows: Orchestral music composed in the latter
half of the eighteenth century might have been standardly performed in wood-
paneled rooms. Nowadays such works would be performed in concert halls.
Modern concert halls are designed with modifiable acoustics, the adjustments
being made by the use of baffles, etc. In performing music of the period in
question, the acoustics of the concert hall would be set with a reverberation
period such as one might find in a wood-paneled room containing a small
audience. Though the music now is performed in a large hall in front of a large
audience, the acoustic properties of the modern building are so arranged that
they duplicate the acoustic properties of the sort of room where the music
would have been performed in the composer's day. Though one might prefer
the intimacy of music performed in salons, I take it that it will be accepted that
the use of concert halls that reproduce the acoustic properties of wood-paneled
rooms would be considered not merely as an adequate compromise between
the demands of authenticity and, say, economic considerations but, instead,
would be accepted as a full-blooded attempt at authentic performance.3 That
modern acoustic technology might serve the aim of authenticity in this way
suggests strongly that musical authenticity aims at the creation of a particular
sound and not at the production of a particular visual, social, or other effect.
Some performances are less authentic for being given in buildings other
than that for which the work was written, but this is true only of performances
of works written with an ear to the unique acoustic properties of a particular
building. That is, it is true of performances of Stravinsky's Canticum Sacrum
and of many works by Andrea and Giovanni Gabrieli, which were written for
San Marco in Venice, and it is not true of Verdi's Aida, which was written
for the opera house in Cairo, because, whereas the acoustics of the opera
house in Cairo are not distinctively different from those of other opera houses,
3
As implied here, the desirability of musical authenticity may sometimes be outweighed by other
factors—musical, pragmatic, or even moral. (I assume that arguments against the use of trained
castrati in opera seria are of the latter kind.) Of course, where the choice is between no performance at
all and a less than ideally authentic performance, the latter may be preferable.
the acoustics of San Marco are unlike those of other buildings. These
examples do not count against the point that a concern with the authenticity
of a performance is a concern with its sound.
II
In this second section I suggest that one might best hope to make a perform-
ance authentic by recreating the musical sound of a performance that might
have been heard by the composer's contemporaries. (Why this is a formula
for success is a matter considered in the next section.) I argue also that the
sound to which an authentic performance aspires is that of a possible, rather
than any actual, performance; that is, authenticity in musical performance is
judged against an ideal.
So far, I have said a performance is more or less authentic in a way that
depends on its sound. One might ask—the sound of what? A musical work is
comprised of notes and relationships between them, so an authentic perform-
ance of a given work must be a performance that concerns itself with pro-
ducing the notes that constitute the work. The sound of an authentic
performance will be the sound of those notes.
But it is not easy to specify the set of notes that constitute a given work (see
Ziff 1973, Sircello 1973, and Walton 1973). The notes recorded in the score
are often not the notes the performer should play; there are conventions
frequently known both to composers and performers governing ways the
written notes are to be modified (for example by accidentals or embellish-
ment). So, an interest in discrepancies between what is written and what is
conventionally played is of practical and not merely scholarly significance.
Debates about the problems ofmusicaficta in music written pre-1600 strongly
reflect a desire to achieve authentic performances of the music in question.
Even where the conventions by which the score should be read are known, it
is not always a straightforward matter to say which notes should be played.
Consider music written at about the end of the seventeenth century, when
pitches were as much as a minor third lower than now. The modern performer
might play the work at the modern pitch level, but vocal and wind parts would
then sound strained even if sung or played brilliantly and correctly.4 Or, the
performer might tune down modern instruments, as a result of which their
tone will suffer, or transpose orchestral parts, in which case the sound is
4
Competent musicians do not usually stumble over fast passages, lose the tempo, or produce
gross tonal contrasts but, despite this, hard music sounds hard to play (Mark 1980).
affected by alterations in fingerings and embouchure, by changes in register, by
shifts to harmonics, etc. In view of such difficulties it is understandable that
performers have turned to the use of instruments from the period of compos-
ition, or to replicas of such instruments, so that vocal and instrumental parts
'lie' comfortably to the voice and hands. The use of such instruments is
ultimately justified by the resulting sound of the performance.
However, despite the use of instruments and the appeal to musical conven-
tions from the time of composition, clearly it is inadequate to characterize
authenticity in musical performance in terms of the sound heard by the
composer's contemporaries. His contemporaries could perform the work in
question in ways that were relatively inauthentic.5 Typically, this would occur
where the performance contained wrong notes or where the composer's
specifications were misrepresented in some other way. The musicians who
sight-read the overture to Don Giovanni from orchestral parts on which the ink
was still wet probably gave a performance that was not as authentic as it could
have been. Since the performances heard by the composer's contemporaries
often were less authentic than was possible, authenticity in musical perform-
ance cannot be defined in terms of the sounds actually heard by the composer's
contemporaries. This suggests that, in striving for authenticity, the performer
aims at an ideal sound rather than at the sound of some actual, former
performance.
Ill
In this third section I consider the relevance of the composer's intentions in an
assessment of the authenticity of a performance of the composer's work.
I suggest that only those intentions that are accepted by convention as deter-
minative are relevant to judgments of authenticity; other of the composer's
intentions or wishes might be ignored in an ideally authentic performance.
Because the composer's determinative intentions underdetermine the sound
of an ideally authentic performance of her work, there is a set of ideal
5
It might be objected to what I have said that judgments of authenticity apply only to perform-
ances that are historically removed from the period of composition, or culturally removed from the
place or style of composition, or in some other way distanced from the composition. On my view,
judgments of authenticity tend to reduce to judgments of accuracy. But this does not mean that a
performance by the composer's contemporaries (for whom the score is 'transparent' to the conven-
tions by which it should be read) is not distanced from the work in a way that leaves room for
judgments of authenticity. Performance involves a creative element that is integral and not merely
appended to the faithfulness of the performance. This creative element distances any particular
performance from the work of which it is a performance.
performances (and not any single ideal performance) in terms of which the
relative authenticity of actual performances is judged.
There are conventions in terms of which musical scores are to be read.
The composer is able to express her intentions in a musical notation only
because the conventions for realizing in sound that notation are known both
to the composer and to the performer of the day. Those conventions provide
not only a vehicle for but also a limitation on the intentions that may be
expressed in the score. Not all of the intentions that the conventions allow to
be expressed are determinative of what can be required in the name of
authenticity. Non-determinative intentions (as expressed in the score or in
other ways) have the status of recommendations. I take it that exact metro-
nome indications are non-determinative, in that tempo may be varied to suit
the performance conditions. Both the composer and the performing musician
who is her contemporary are usually familiar with the conventions and know
which of the expressed intentions are determinative and which are not deter-
minative of that at which an authentic performance must aim.
The conventions by which musical scores are to be read change over time
in ways affecting what the composer may determine with respect to the
performer's attempt to produce an authentic performance. Phrasing was not
notationally determined in the early seventeenth century but was notationally
determined by the nineteenth century. At some time, before the convention
was established, composers notated phrasings that would have been rightly
understood as recommendations for, rather than as determinative of, what
should be played. At that time, the composer's indications of phrasing might
be disregarded without any diminution in the authenticity of the performance
(though the performance may have been less good as a result on other
grounds). (These changes in convention sometimes arise from composers'
rebelling against the existing conventions, but such rebellions reject only a
few conventions at any one time and do so against a wider background of
accepted conventions.) Because conventions of determinativeness change
through time, the conventions appropriate to the authentic performance of a
score are those with which the composer would have taken musicians of
the day to be familiar. It is this fact that explains what I have emphasi/ed in
the previous section—that an attempt at an authentic performance is likely to
be successful by aiming to re-create the sound of an accurate performance
by the composer's contemporaries.6
6
The claim that the conventions of score reading and/or performance practice establish which of
the composer's publicly expressed intentions are determinative may be defeated where there are
Sometimes it is possible to infer from what is written in the score that the
composer would have preferred to write something else had the instruments
or the performers been capable of accommodating her intentions. For
example, a sequential pattern might be interrupted by an octave transposition
where a continuation of the pattern would have exceeded the singer's or the
instrument's range. In these cases, it is appropriate to talk of the composer's
wishes (rather than intentions). Sometimes nowadays, with the wider range
of some instruments and the greater proficiency of many musicians, these
wishes could be realized and there would be a musical point to doing so.
However, such wishes have no more a bearing on the authenticity of a
performance than do the composer's non-determinative intentions. Both the
work and the performance may be better for the modification, but not because
the alteration makes the performance more authentic. If it were accepted that
mere wishes could set the standards of authenticity, it would be accepted also
that many works could not have been performed authentically by the com-
poser's contemporaries and some could not be performed authentically at all.
Clearly, in taking the line I have, I must deny that authenticity in musical
performance is judged against the sound of some particular performance that
was envisaged by the composer. I have said that not all of the composer's
expressed intentions are determinative of what must be accurately rendered in
an ideally authentic performance, in which case I must also hold that the
sound of an ideally authentic performance is underdetermined by the inten-
tions in terms of which its authenticity is judged. The way we talk of
authenticity favors my view, I claim, rather than the view that authenticity
is measured against the sound of a performance that the composer had in
mind. First, in reaching judgments about the authenticity of performances,
we do not seem to face the epistemological difficulties that would inevitably
arise if the standard for authenticity was a sound that may never have been
realized. Second, rather than taking composers' performances as definitive
models that performers are obliged to copy slavishly, we take them to be
grounds for believing that the composer was not familiar with the conventions or that the composer
believed that the musicians who would perform the piece were not familiar with all the relevant
conventions. These double-take and triple-take situations are unusual. An example: If the composer
had only ever heard violins with a thin and reedy tone and by the indication 'violin' on the score
meant to designate instruments of that type, then the fact that Guaneri's violins were extant at the
time would not license their use in performances of the composer's works in the name of authenti-
city, not even if the composer had wished that the instruments she knew as violins had a richer,
fruitier tone. (To avoid such problem cases I should relativize all claims about the role of the relevant
conventions to the composer's knowledge of those conventions and beliefs about the performers'
knowledge of those conventions.)
revealing of what we expect to be an interesting interpretation. In a perform-
ance, the composer may make her intentions as regards the sound of a
performance more explicit than could be done in the score, but what is
made explicit is not thereby made definitive. Other performers are left with
the job of interpreting the score for themselves.7 Third, we would not (as we
do) accept that different-sounding performances of a single work might be
equally and ideally authentic if authenticity were judged against the sound
of a particular performance imagined by the composer. It is (a member of) a set
of ideal performances against which the authenticity of an actual performance
is judged.
This last point deserves emphasis. Because an ideally authentic perform-
ance faithfully preserves the composer's determinative intentions and because
those intentions underdetermine the sound of a faithful performance, differ-
ent-sounding performances may be equally and ideally authentic. For
example, many combinations of vocal and instrumental resources are com-
patible with what is determinative in the score of Guillaume de Machaut's
Messe de Nostre Dame. Even if the composer wrote for a particular combination
of singers and instruments (such as were assembled for the coronation of
Charles V in 1364, perhaps) the conventions of the day allow that performances
by quite different combinations would be no less authentic. As long as two
performances are faithful to the score and are consistent with the performance
practices in terms of which it is to be rendered, they may be equally authentic
while sounding different. Compare, for example, performances of Beethoven's
symphonies as conducted by Klemperer and Toscanini, both of whom have
been praised as interpreters of the works. Klemperer tends to take the pieces at
the slowest tempo consistent with Beethoven's instructions and he emphasi/es
the structural qualities of the music so that, for example, climaxes at relatively
weak structural points receive less weight than do those in structurally import-
ant places, even where the dynamics indicated in the score are the same in
both places. Toscanini takes the works at a brisk tempo and concentrates on
the drama or beauty of each individual passage, investing every note and
phrase with its full potential of power. Without Klemperer's staid approach,
the grandeur and architectonic qualities of Beethoven's music could not be
presented. Without Toscanini's volatile approach, the dynamism and verve
of Beethoven's music could not be appreciated. So, the ideally authentic
7
A pertinent discussion of musical authenticity and the relevance of composers' intentions may
be found in Taruskin 1982. The status of the composer's intentions is interestingly discussed in
Dipert 19806. The philosophical literature on the subject of artist's intentions is immense. Two of my
own papers bear on the topic—see Davies 1982 and 1983a.
performance has no particular sound because it is no particular performance.
Rather, the standard against which the authenticity of performances of a work
is judged is comprised of a set of performances each of which is faithful to the
composer's determinative intentions.
In view of the above I offer the following account: A performance will be
more authentic if it successfully (re-)creates the sound of a contemporary
performance of the work in question such as could be given by good musi-
cians playing good instruments under good conditions (of rehearsal time,
etc.), where 'good' is relativi/ed to the best of what was known by the
composer to be available at the time, whether or not those resources were
available for the composer's use.
IV
In this fourth section I analy/e musical performance as involving both certain
intentions on the part of the performer and a relationship of invariance
between the composer's sound specification and the performer's realization
of that score. Performing is briefly contrasted with improvising and fantasi/-
ing. The point of authenticity is said to be the faithful realization of the
composer's score in sound.
The notion of performance must be analy/ed in terms of the performer's
intentions. If the production of some set of sounds is a performance of X, then
it must be the intention of the producer of the sounds to generate a sound
faithful to an ^-specification. However, the intention to perform X is defeas-
ible; where the sound produced is not recognizable as a realization of the
^-specification the attempt at performance has failed. The notion of authen-
ticity operates within the range set on the one hand by performances that are
barely recognizable as such and on the other hand by performances that
are ideally accurate. The closer a performance, recognizable as such, comes
to the sound of an ideal performance of the work in question, the more
authentic is that performance.8
I have suggested that there must be, as well as the appropriate intentions,
an invariant relationship between the composer's specification and a
performance of that specification as a necessary condition of the success of
8
It is controversial, I realize, to regard a barely recognizable performance as authentic. Of course,
the level of authenticity expected in a competent performance is far higher than the minimum at
which a performance is barely recognizable as such. A minimally recognizable performance is
inauthentic when authenticity is relativized to a standard of acceptability at the level of a competent
performance.
the attempt at performance. There must be some common factor (or tolerance
across a range of features) necessary for a performance's being a performance
of X rather than of 7, and necessary for different-sounding performances all to
be performances of the same X. Now, clearly the standard by which an
attempted performance is minimally recognizable as such falls far short of a
standard that identifies the work with the totality of notes constituting it. By
this standard only a perfectly accurate performance could count as a perform-
ance of the work in question, yet we all know that the school orchestra may
play wrong notes, play out of tune, and fail to play together while performing
what is unmistakably Beethoven's Fifth Symphony. It is because musical
works are comprised of large numbers of notes, not all of which contribute
equally to the overall effect, that the identity of the work survives the per-
formance of wrong notes. So, what is invariant between performances of the
same work is patterns of notes (or aspects, gestalts, emergent properties,
functions, of notes) plus a tolerance for deviation from these patterns. Musical
works are so complex that there are patterns of notes within patterns of notes
and these various patterns may remain recognizable despite changes in or
omissions of individual notes. The standard of adequacy that must be met in a
successful attempt to perform the composer's score need not be one that
requires a high degree of accuracy.9 It is within the gap between a set of
ideally faithful interpretations of a work and of barely recognizable perform-
ances of that work that the notion of authenticity operates. A performance is
the more authentic the further beyond the minimum standard of adequacy it
falls. The more faithful is a musical performance to the work's specification
the more authentic is that performance.
The difference between a performance ofX, an improvisation on X, and an X-
inspired fantasia lies in the musician's intentions, the aim being to realize a
higher level of invariance with respect to the work's specification in perform-
ance than in improvisation and in improvisation than in fantasizing. Whereas
authenticity is appropriately predicated of performances of particular works,
it is not appropriately predicated of improvisations or fantasias inspired by
particular works; that is, authenticity applies only where there is intended to
be more rather than less invariance between the specification of the work and
9
The same kind of point may be made with respect to other musical parameters. A performance
on the piano of J. S. Bach's Concerto in D minor for Harpsichord, BWV 1052, is a performance of it,
despite the change of instrument, and not the performance of a transcription of Bach's work.
Conventions in Bach's time allowed quite free interchange between keyboard instruments and, in
view of this, merely changing the solo instrument does not transform the work enough for the
performance to count as that of a transcription. (One does not transcribe a musical work merely by
altering a word in its title, which, in effect, is what happens here.)
its rendition in sound. This suggests that the notion of authenticity applies
where a 'text' (usually a written score in literate music cultures and a model
rendition in oral music cultures) is interpreted by a mediator who stands
between the composer and his audience, and where the point of the interpret-
ation is to render faithfully to the audience what is determined of the sound of
the performance in the work's specification. A concern with the authenticity
of performances of particular works ultimately takes its interest from a more
fundamental concern with the authority of authorship.
A shift of focus to music that is primarily improvisational (i.e. most ja/7, a
substantial amount of non-Western music, and some recent 'classical' music)
helps to bring out the point. In such music, where the composer creates a
cipher lending itself to improvisational manipulation, we are more likely to be
concerned with the authenticity of the style of the performance of any given
work than with its authenticity as a performance of that particular work.
The less the sound of the performance is determined in a faithful realization
of the composer's specification, the less we are concerned with the type of
authenticity in performance I have been discussing (and the more the
musicians are rated above composers). The less the composer has a hand in
the final outcome, the less is a concern with musical authenticity a concern
with the authority of authorship.
V
In this fifth section I consider the way authenticity in musical performance is
valued. I suggest that though such authenticity would not be valued were it
not a means to an independently valued end—the end of presenting the
composer's interesting musical ideas—nevertheless, authenticity in musical
performance is not valued as a means to this end.
Beyond the level of an acceptably competent performance, authenticity is
value-conferring. A musical performance is better for its being more authentic
(other things being equal). Because we have an aesthetic concern with the
musical interest of the composer's ideas, and because those musical ideas
must be mediated by performance, we value authenticity in performance for
the degree of faithfulness with which the performance reali/es the composer's
musical conception as recorded in the score. I am not maintaining that
authenticity in performance takes its value from the worth of the musical
content contributed by the composer. Rather, my point is this: Were it not for
the fact that composers set out to write aesthetically rewarding works, and
were it not for the fact that they are usually successful in this, we would not
value authenticity in musical performances as we do. But, in any particular
instance, authenticity in performance is valued independently and irrespect-
ive of the aesthetic value of the work itself. A performance is better for a
higher degree of authenticity (other things being equal) whatever the merits of
the composition itself. A performance praiseworthy for its authenticity may
make evident that the composer wrote a work with little musical interest or
merit. It is the creative skill required of the performer in faithfully interpreting
the composer's score that is valued in praising the authenticity of perform-
ances of that score.10
Of course, authenticity is not the only quality for which a performance
might be valued. Where a relatively inauthentic performance is highly valued,
it is valued in spite of its inauthenticity. Thus, Schnabel's recorded perform-
ances of the Beethoven sonatas are well regarded despite the wrong notes they
contain.
VI
In this final section I emphasi/e how creative is the role of the performer in
faithfully realizing the composer's specification. In developing the point, a
contrast is drawn between performing and copying.
The performer transforms the notes-as-written into the notes-as-sounds. In
talking casually of the notes of a piece, and thereby obscuring this distinction,
one might easily lose sight of the creativity of the role enacted by the
performer in faithfully converting the one into the other. The sounded notes
created by the performer go far beyond the bare peg that the composer
provides and on which the musicians hang their art. An authentic perform-
ance concerns itself with the production of the notes that constitute the piece
and that the composer specified, but the notes-as-sounds produced by the
performer involve subtleties of attack, decay, dynamics, tone, and so on
that cannot be captured in any notation composers are likely to use. The
written notes and the way they are played come together inseparably in
the notes-as-sounds, and it is in no way to undervalue the role of the composer
as the specifier of the notes-as-written to acknowledge that the musician
10
Indulging in some armchair sociobiology: It is perhaps not surprising in a social species such as
ours—which is concerned with successful communication and for which there can be no guarantee
that any particular attempt at communication will not fail—that what facilitates communication
becomes valued for its own sake and apart from the worth of the contents it helps to communicate.
(Not that I think that music can be usefully compared to a language with respect to its meaning—see
Davies 19836.)
brings something original to the notes-as-written in rendering them into
sound (Harrison 1978). The creative role of the performer, rather than involv-
ing a departure from the concern to realize faithfully the composer's
intentions, is integral to the execution of that concern.
What is more, rather than consisting of mere aggregations of notes, music
is comprised of themes, chords, subjects, answers, sequences, recapitulations,
developments, motifs, accompaniments, and so forth. These are gestalts (or
aspects, etc.) and not mere successions of notes. Because their articulation in
sound owes as much or more to the performer as to the composer, it can be
seen how extensive and important is the creative role of the performer.
One way of bringing out the creative role of the performer as a necessary
intermediary between the composer and the audience is by contrasting per-
forming and copying. Copying need not be intentional; copying may be a
mechanical process performed by a machine. And where copying is inten-
tional, the aim of faithfulness is to be contrasted with that of creativity. By
contrast, performance is always intentional, because the performer must bring
more than is supplied by the composer to a performance that is faithful to the
composer's ideas. Performing must go beyond what is given by the composer
in order to present that accurately. But nothing not present in the original need
be brought to copying. A machine might copy a performance (for example, by
recording it on tape), but performing is done only by agents. 1 1 And copies a
authentic only in the sense contrasted with forgery or fakery, whereas per-
formances are authentic in the sense that has here been under discussion.
Authenticity is an attribute acknowledging the way the interpretation of a
musical score is both necessary in the presentation of the music-as-sounds
and is also inherently creative. Authenticity, as a praiseworthy attribute,
acknowledges the ineliminability of the performer's contribution to the
sound of the performance.
1
' I do not deny that copying by hand an illuminated manuscript might require patience, skill, etc.
in a way that suggests that copying is anything but mechanical in this instance. Nor do I wish to deny
that there are imaginable cases in which computers are programmed to produce sounds where we
would be tempted to say a machine performs. (Just as there are cases in which the musician performs
on a violin without our saying the violin performs, so there are cases in which musicians perform on
computers— but the example to be imagined is not of this type.) But if there were such computers,
talk of them as machines would begin to look inapposite; at such a point one begins thinking in terms
of intelligent or agent-like 'machines'.

Stephen Davies - Themes in The Philosophy of Music

Uploaded by

Copyright:

Available Formats

Stephen Davies - Themes in The Philosophy of Music

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Stephen Davies - Themes in The Philosophy of Music

Uploaded by

Copyright:

Available Formats

The Ontology of 4

Musical Works and the

The authentic-performance movement is a phenomenon of the last fifty years.

In this chapter I discuss musical performances and their authenticity with

You might also like