This page has been archived by the Computer Arts Society.

Notation and Performance

<<Previous Up Next>>

For Goodman, musical notation depends on understanding and knowledge to execute it musically, each performance being a “genuine instance” of the work. Yet performance is more than a matter of “unpacking” the concentrated data in the binary score: the performer also adds any contextual data already possessed or conjectured, which (for example) makes the supposedly authentic reconstruction of early music doubly conjectural.

Thus, as Goehr implies, the score acts as a mnemonic aid rather than a straightforward instruction. The performers provide more than technique and skill: they actively interpret and develop the music, occasionally improvising beyond the score; they invest it with emotion. True, the listener judges their performance on its accuracy, but the music has to have character and interest; insightful listeners recognise a “mechanical” performance.

Goehr considers musical works to be a totality of compositional idea, performance and appreciation. This was intended as a corrective to Goodman’s view that the music itself was comprised of straightforward “allographic” musical scores. It would seem that the overarching musical work is both the sum of its constituent parts and an entity in its own right. This has at least some application to multiple formats and creative inputs of Computer Art, particularly interactive pieces where the final form is dependent on the viewer’s interaction. Goehr holds that musical works “cannot, in any straightforward sense, be physical, mental, or ideal objects.” She also points out that there is a temporal difference between a performance in real-time and the simultaneous existence of all parts of a work.[1]

Goehr’s argument favours the existence of a work of music that may be represented by but not contained within its notation. In other words, it has an existence independent of the score and its material form, which is only released in performance. This is my favoured position for understanding physically-realised Computer Art: the object is constructed through the exercise of the code, but is not wholly contained within it. Rather, the artist’s concept of the work contains both code and physical artefact even if he does not know the exact outcome of the code’s realisation. Again, the origins of the artistic act are important and merely copying the code will not give one possession of the work of Computer Art. Indeed, in Hébert’s case the action of the ball on the sand, or his plotter pens on paper, is dependent on physical factors not present in the original code. Thus each realisation would result in a wholly new artefact.

Goehr criticises Goodman for only dealing with the work as a performance and score, without any reference to its abstract existence outside these instances. In order to establish a score, Goodman distinguishes between the work’s constitutive and contingent properties: the former must be rendered into a notation system, whilst the latter are elements of a performance.[2] Goehr questions whether the identity of a musical work can be explained through notation alone; she implies that neither notated score nor performance is the actual work:

[Goodman] can allow either that there is a degree of vagueness in both the work and its performances, or that they are both equally and perfectly determinate as regards their constitutive properties. He opts for the latter.[3]

The overall concept can be related to ephemeral Computer Art, even though the computer’s “performance” is a reconstruction of visual material from image data, quite unlike the expressive and interpretative role of a performer. Exactitude of the computer’s “performance” is expected, because any glitches are usually seen as flaws that need to be eradicated; they generally cause the program to crash. [But see Hébert’s work] Thus the computer’s rendering of the data into visual form has to be transparent. Yet the same question can be posed, with regard to Goodman’s theory: Does it follow that the identity of the computational work of art can be explained in terms of its underlying data?

Music might be regarded as the auditory product of an intangible, intellectual process of composition and rendition. The recorded form of the music (i.e. the score as opposed to actual sound recordings) can be saved and performed again. The score is the tangible part of the musical work, but only when performed does it become music as such. Until then, it encodes the bare sonic and temporal details which are the basis for making the music; it preserves its digital form but not its auditory matter.

On the other hand, a sound recording may be analogue or digital, but it is a record of a specific performance at a certain time. It captures the auditory and interpretative elements missing from the score, and the factors embodied in that particular performance. By contrast, a MIDI file has a similar relation to the musical work as sheet music, since it is a set of instructions to the computer which can be decoded as instrumental parts.A broader analogy may be drawn with the digital tracks on a CD, which mean nothing until reconstructed into sound by the laser and broadcast to the listener’s ear. I would say (following Goehr) that the data on the CD is not the work held in a potential state, but rather the components of it that must be reassembled. It is in this reassembly that the work is to be found.

Here again the difference between digital data and musical notation occurs in the exactitude of its replay: the CD player is expected to be exact and any deviation is noted, because it is rebuilding sounds rather than decoding music; a performer reading that music and creating sound is understood to follow the score as closely as possible, but not absolutely and mechanically exactly. Of course, a synthesizer playing music from MIDI data actually reconstructs the notes and “performs” them in a way directly comparable to a musician. MIDI notates the actions that produce the resultant sounds; audio formats like Windows Media capture the sounds themselves.

For digital images, it is the difference between generating a 3D scene from algorithms, and simply displaying a digitised photograph: the difference rests in the amount of rebuilding – indeed, “interpretation” – the computer has to perform.

Between conceiving a work of art (whether or not the artist has an image in mind) and its final presentation an artist would traditionally work with a material to realise the idea. But when using a computer, the working and the material are contained within a medium that is separated from the artist at several removes. In this case, the artist either works via an interface that allows for the transformation of apparently visual elements which are digital processes; or instructs the computer directly via algorithms, in which case the artist has themselves processed their work through a series of stages.

At every stage, the medium is not in itself a visual form, unless one constructs an image from the very instructions themselves (as happens in ASCII art). Otherwise, there is no visual relationship between the material and its eventual form. Instead, if seen as a set of instructions, it can usefully be regarded as a score.

Yet this is not a score, or as Goodman would have it, an allographic artform. True, one can copy these instructions exactly and submit them to another computer in the hope of recreating the image. This will only happen if software and hardware happen to match the original configuration, or fall within the bounds of its functioning.

The issue of programs intended to create images autonomously further complicates the analogy of a score, since these are obviously not performances intended to follow a pre-designed path exactly, but to depart from it and produce new forms. In a sense, it is a form of improvisation (to borrow another musical analogy), but one that is inherently rule-based.

It would be unwise to draw further parallels with Computer Art, because the musical work involves a central connection between the compositional, sonic and tonal character of performance. The visual object, by contrast, always has a particular presence, no matter how fleeting its appearance, and its form can be shaped in direct ways as opposed to the indirect shaping of sound by the composer operating through the score. The artist’s control over shape, colour and other elements, even when operating at one or two removes via his assistants or by programming, still seems (to me) far more direct than a composer setting down his piece in notational form, to be reconstructed much later in performance.

Perhaps the comparison between artist-programmers and composers breaks down at this point. Nicholas Cook considers the composer to be working with “notes, motifs and forms”, not sounds in themselves; the work only presents itself as a collection of sounds in terms of these structures. The formal relations of structural components are what the composer becomes involved in, even if he is striving for a particular harmony in the finished work. [4]

It might be claimed artist-programmers manipulate commands in order to arrive at a visual form. In the case of those working with emergent systems, even this final form is in itself not directly predictable. However David Gelernter cautions both against confusing the musical score for the music, and confusing the written form of the program or its mathematical content for the program as a whole. The program might contain equations and have the form of a document, but in itself it is a machine; just as a musical stave contains symbols and one could compose music solely in terms of symbolic manipulation:

You can see what good counterpoint looks like on paper […] But if you get carried away, and start asserting that “music is the mechanical manipulation of symbols on staff paper”, “programming is mathematics”, you have committed intellectual suicide. You’ve mistaken the means for the end.[5]

Many examples of Computer Art exist as concrete, physical objects; yet this form is complicated by their existence as data, which encodes their overall form in a format that gives no clue to their visual appearance. At least musical scores give some visual indication of rising and falling phrases and the length of notes; and most importantly, they indicate the temporal structure of the music. Data, by contrast, is an inert list of instructions that may make linguistic sense when printed, but only when decoded and “performed” – in the broadest sense – by the processor and displayed through the correct hardware, can it make the visual aspects apparent. This even applies to Computer Art which is generated by the machine itself, where the artist only supplies the parameters rather than an exact description of the final form.

Gelernter’s warning makes me doubt that Computer Art is simply one of Goodman’s “allographic artforms”. It may well exist in its basic form as notation, but we judge it on its visual appearance and content, and that is also how the artist conceived it, even if they were building up from co-ordinates and values. So mistaking the art for the code is a problem, even if on one level, the art is the code.

Perhaps Goodman confuses the work of art with its record [transcription] and performance [in the case of music]. Certainly, Goehr’s concept of the “work” encompasses the composer’s concept, its musical realisation in the form of the score, and the subsequent rendering into audible form.

Although Mitchell only considers the digital photograph in his discussion about its lack of an “original”, this idea of the image as a score was also raised by the algorithmic artist Roman Verostko. He initially believed that his Hodos software produced something akin to a “score” for visual art, in that it provides instructions through which the computer can “improvise an original work every time.”  However, he later decided the Hodos image creation was more akin to a virtuoso performance by the computer.[6]

[Plate XXVI: Roman Verostko, Epigenetic painting.]

Verostko’s works originate with his Hodos program, but the artwork itself consists of the print, made by Verostko’s modified plotter. The physical production and material existence of the work place it outside the computer, yet it results from the directions Verostko gives his software. In this sense, it could not have resulted in this form without the computer’s involvement.

[1] The Imaginary Museum of Musical Works (Oxford, 1997) Lydia Goehr pp2-3

[2] Goehr, ibid, p21

[3] Goehr, ibid, p21

[4] Nicholas Cook, Music, Imagination and Culture (Oxford, 1992), p212

[5] Mirror Worlds, David Gelernter, p41.

[6] “Epigenetic Painting Software As Genotype, A New Dimension of Art” by Roman Verostko, 1988