This page has been archived by the Computer Arts Society.

Problems with Goodman: Lydia Goehr and Richard Wollheim

<<Previous Up Next>>

Other theorists have raised doubts about the suitability of Goodman’s theory for music, so it may not be ideal for comprehending the digital nature of Computer Art. A computer image may well exist as notation in its basic form, but we judge it on its visual appearance and content. That is usually how the artist conceived it, even if they constructed it from co-ordinates and values. If Goodman’s theory is applied across the board, there is a danger of mistaking the “art” for the “code”, even if on one level the art is the code to some extent. Richard Wollheim considers this in relation to older media – the book and the operatic performance:

That there is a physical object that can be identified as Ulysses or Der Rosenkavalier is not a view that can long survive the demand that we should pick out or point to that object. There is, of course, the copy of Ulysses that is on my table before me now [or] there is the performance of Der Rosenkavalier that I will go to tonight, […] from which it would be tempting (but erroneous) to conclude that Ulysses is just my copy of it, Rosenkavalier just is tonight’s performance.[1]

Richard Wollheim used Charles Peirce’s idea of works as “types” and their individual copies and performances as “tokens”. On a simplistic level, the computer artwork as unexecuted idea may be considered the type, and all its instances as tokens, whether the digital form held in potential or the visual rendering of the same. Both score and performance are tokens. In this respect, Computer Art is closer to the performing arts than to visual or literary forms with its performative aspect.

However, as Wollheim also notes (see Hanfling, p80),[2] tokens have to have certain necessary, or invariant, properties to be of a given type, yet they also possess other non-necessary properties that are not specified in the original. Unlike musical notation, the computer’s execution of artforms is constrained by its code, but this is offset by the degrees of serendipity and sentience that can also be built into the program, or its physical execution.

This process bears a closer resemblance to musical composition when the picture is construed in mathematical terms, and to non-digital graphic art when it is not. By which I mean that a picture conceived of in terms of maths or process can be more readily expressed in notation or code, which is what many early computer artists were forced to do owing to the constraints of their machinery. But since the whole point of the Graphical User Interface was to open up the computer interface through gesture-based actions rather than command-line instructions, it became easier to approach the computer as a drawing surface than as an algorithmic system.

As Larry Cuba noted, the composition of a digital piece may itself be the result of a “dialogue” between artist and tools. Hanfling recognises that creators of musical or dramatic works may be so delighted by an expressive performer that they alter their work based on this experience: “… they may also feel that the new version is a better expression of what they ‘wanted to say’ than what they had themselves written.”[3]

This process is an essential component of some computer artists’ work, where indeterminacy and evolution are employed in the program; it also points towards a rationale for interactivity. The creator becomes the originator of a core idea which is built upon and changed over time, until its most recent iterations are wholly different from its origins. Yet it will demonstrably be part of the same continuous artistic process that ties it back to its conceptual origins, in the mind of its creator.

In this case, the artist’s role is akin to that of a director or a conductor, organising and specifying the roles of various forces to fulfil his vision. The computer artist steps back from the execution and combines his deep knowledge of the computer with an overarching concept to enable the computer to select a route through various possibilities.

The result is not a demonstration of plastic skill or draughtsmanship, but the presentation of an artistic vision. The computer artist directs the computer to make his vision concrete, summoning new forms out of the inscrutable code. There is an obvious parallel with the artists’ workshops of the Renaissance, where the apprentices would do much of the work based on cartoons and sketches, and the master applied his skill to the hands and face.

As yet, the computer itself cannot bring an overall vision to any work of art; only in the bounded universe created through Harold Cohen’s AARON program can a computer make aesthetic choices based on knowledge of results. The computer artist employs randomising processes in pursuit of his vision, by tailoring the starting point of these processes and setting them in motion. In this he may exercise more or less control according to temperament, but there is usually an element of indeterminacy which leaves room for discovery – whether it is found in Hébert’s choice of sand as a material, or the serendipitous results of genetic transformations in Latham’s work.

Moreover, the rendition of a computer artwork may involve unpredictable randomising factors that change its very structure every time it is manifested. In this sense, every manifestation is unique, though within parameters that make it obviously related to the body of work produced by the program.

Predictability is a typical feature of mechanical art which was valued by Moholy-Nagy and by early computer artists. A computer whose program results in seeming unpredictability goes beyond such expectations and challenges our simplistic perceptions of it as a purely mechanistic device.

Indeed, much Computer Art involves controllable unpredictability. This makes it less mechanical than simpler processes such as printing, where the brain can comprehend the operation of the hardware, and makes it closer, though not entirely comparable, to media such as watercolours and oils. Here, the important factor is not so much the execution of a particular operation as familiarity and fluency within the constraints of the medium.

The rendition of a Computer artwork may involve unpredictable randomising factors that change its very structure every time it is manifested. In this sense, every manifestation is unique, though within parameters that make it obviously related to the body of work produced by the program. The artist’s relation to the computer is more complex than has often been acknowledged because the computer’s responses, while programmed, can have such a large degree of latitude and inbuilt randomising factors that the artist cannot necessarily predict the outcome based on their internalised model of the system.[4] In this sense it goes against the expectations of earlier mechanical art devices. Robert Mallary mentions this opening up of possibilities in Artist and Computer (1976):

[…] in the course of debugging a program, the output might reveal a possibility that could not be foreseen when the program was first planned […] In fact, one of the delights of complex programs with many parameters is these surprising productions. [pp6-7]

Even the frustrations encountered when using the computer are an integral part of the computer-assisted artistic process and as such inform the art as much as the other constraints imposed by the interface and factors such as software design. Thus Computer Art is not the facile product of digital wizardry it is supposed to be, though in the hands of an expert – as in any other field – it is made to look easy. [QUOTE from MIT book about frustrations of digital practice.]

This is very reminiscent of the “discovery” of artistic direction in the very resistive properties of materials that may initially deflect the course of an artwork. As Anton Ehrenzweig describes, these frustrations may finally lead to a better piece of art:

The wood carver is delighted and not upset when the wood grain inside his block gradually reveals itself and by its obstinate twists compels him to modify his tentative ideas. ‘Accident’ is a relative term. […]A new idea will inevitably be modified through its impact on the resisting medium [p57]

There will always a be a slight element of mystery in the working of the software, requiring knowledge and skill to use to best advantage and thus rendering the computer less mechanical (and predictable) and more ‘artistic’ in the sense that the outcome of a particular operation cannot always be foreseen by the artist. However, the computer artist Tom Kemp, who has recently been programming a particle system in Java for creating drawings, says that most operations will work as intended and that most outcomes are relatively predictable. Whilst agreeing that the program will never develop entirely randomly, it would seem that some outcomes will surprise the programmer, leading to more modifications to either increase the effect or prevent it.

This is where Computer Art proper differs so fundamentally from all previous plastic arts, even those predicated on reproduction such as engraving. Hanfling recognises that “our consciousness of the unique identity of a work is an important element in our experience of it, and this is not so in the case of multiple arts, such as music.” This in turn derives from its physical existence as a piece of art and the knowledge that one is looking at the original instance of the work. For me, at least, this feeling may be heightened by any reproductions I have previously seen: they convey an element of familiarity which is often entirely undermined when seeing the work in its unique physical form.

Works of Computer Art by their very nature do not have a unique manifestation, at least in the sense of a finalised and complete unique work. So long as they are dynamic and malleable, they may be changed; so long as they are the result of pre-ordained programming they may be recreated. Yet they often possess a physical presence, as Hébert’s Ulysses demonstrates so well, and this physicality usually demands the viewer to be local as well.

It is true that much digital art is “portable”, in the sense that it can be displayed on most computers (dependent on compatibility of formats); but even Internet-based interactive art demands the user be present at workstation to view and manipulate the artwork. In the case of large-scale Computer artwork, which requires unique or expensive hardware to manifest itself, then the viewer has to be physically present to appreciate it fully. This is the price we pay for digital malleability: we need the computer as intermediary to display it.

Since Computer Art comprises a sequence of instructions which can be copied with no intervening loss of quality, the work itself, as an entity, is hard to identify. This is not the same as the original; rather it is the totality of artistic conception, intention, material, presentation and the final rendering. This question is especially salient with regard to music, but it would seem that Computer Art has affinities with music in this respect.

Music theorist Lydia Goehr argues that musical works are not concrete objects or “private ideas existing in the mind of a composer, a performer, or a listener”. Nor are they identified with a particular performance, which is time-limited and sequential. Rather, the work has a continuous temporal existence in which all its parts “exist simultaneously”, along with those expressive properties which are not necessarily set down on the score.

In a sense, the work exists through all its manifestations, including the purely conceptual ones amongst composers and listeners; yet it also exists in a space quite separate from them. All reifications of the work are interpretations, though I would say that each successive performance adds to the range of possible interpretations and thus expands the work itself. On the basis of this immaterial existence, Goehr asserts that even if all copies of a Beethoven Symphony are destroyed, the work may continue to exist.[5] In a similar (if slightly misleading) vein, the surviving photograph of a Computer artwork enables the survival of its whole conception, even though the hardware, software and the artist himself have disappeared and the photograph is their last record.

In the case of Edward Ihnatowicz’s Senster, this is very much the case. The Senster was a four metre high robotic sculpture constructed for the Philips Evoluon exhibition at Eindhoven in Holland, in 1972. Ihnatowicz applied considerable artistic and technical virtuosity to the Senster’s design, enabling the giant structure to move about on its base and home in on sources of sound. The surviving photographs show an imposing bare metal frame with a distinctive articulated neck and “head” with the sound sensors; it impresses with its scale and starkness. Yet it is only in surviving film footage that the Senster’s graceful movement – and the breadth of Ihnatowicz’s conception – becomes apparent. The sculpture seems to “look” towards each sound, possessed of some basic intelligence even though its code occupied barely 4K of memory! Indeed, knowing the elegance of the code is itself another realisation of Ihnatowicz’s abilities: he was able to leverage the primitive computers of the day to manoeuvre the sculpture around.

[Plate XXIII: The Senster at Philips’s Evoluon exhibition, Eindhoven, 1972]

Now that the sculpture has been put out to grass in a forlorn Dutch field, the computers are long-since dismantled and Ignatiwicz himself is sadly deceased, what remains of the Senster’s remarkable concept? Only the rusting frame, which in its inanimate state hardly represents the work which Ihnatowicz had in mind. In this sense, the photographs and films are more than snapshots or footage: they incorporate something of the original work in their very existence because quite apart from their status as visual records, they also show the Senster’s operation and the audience’s reaction. In this, they convey Ihnatowicz’s concept in more than merely visual ways.[6]

R.G. Collingwood considered whether works of art reside solely in the minds of their authors. He compares the creation of a work of art to the making of a bridge, where its original locus is the engineer’s plan in his mind, not its instances on paper as notes and sketches.  Similarly, the musical or artistic plan may be executed, but this is “the means by which the audience [can] reconstruct for themselves the imaginary tune that existed in the composer’s head” [Collingwood, quoted by Hanfling]. Thus the “art” remains in its originator’s purely mental conception and its execution should not be confused with its origins. Insofar as the work exists externally, it is through the interpretation by the players that it is brought into being, and in the audience’s interpretation that it grows to an approximation of its original concept.

With Computer Art, the extra facet added by material realisation, or films that record the operation of large-scale installations such as with the Senster, is that these renderings are not interpretative but aid interpretation of the original concept. Obviously, they also enable it to survive the dissolution of the computer that first generated the images and propagate the artist’s original concept, long after the dynamic computer-based artwork has ceased to function.

Goehr makes a distinction between the temporal structure of works, existing simultaneously, and that of their performances which unfolds over time. The temporal and spatial element in Jean-Pierre Hébert’s computer-driven sand pieces is their most important feature. Hébert contrasts the necessity of controlling the exact shape executed by the steel ball on the surface of the sand with the imperfections that occur when the sand creates an obstruction. Although the program encodes the instructions for creating an exact shape, the uncertainties inherent in the physical execution are part of the final, physical artwork; indeed, Hébert aims to include them through his choice of materials and printing process

[1] Wollheim Art and Its Objects, pp4-6

[2] Hanfling, Oswald, Philosophical Aesthetics, p80

[3] Hanfling, ibid, p87.

[4] Predictability is surely a typical feature of mechanical art (valued by Moholy-Nagy and indeed by early Computer artists). So a computer with programmed unpredictability goes beyond this stereotype and changes our simplistic perception of it as a mechanical device.

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

[6] Edward Ihnatowicz, “Cybernetic Art – A Personal Statement” 1985