Page  487 ï~~Music and technology Frank Pecquet Esthetic Institute, Paris-I University. 162 rue Saint Charles, 75015, Paris. c-mail: l(X)15,1143@compuscrvc.com ABSTRACT A methodological reflection on the objectives of computer music composition facilitates the transition from a logical use or a rationalized proceeding of technological tools to the domain of the esthetic. This text clarifies in three connected axes of thought the development of music composed with computers. The first one is the concept of material, i.e. the sonic technical phenomenon or "operative technical method", second the concept of techniquecompositional thought, the writing and the notation, third and last, the concept of new relationships between the composer and the sound. 1. Conception/Material A sensibilisation that is different to the sounds as to the structures which differentiate them, implies that composers have a distinct perception of the relationship between the intelligible world and the sensible world. It is evident that theoretical research undertaken by Webern, who died in 1945, illustrates that the intelligible world complements the sensible world. The true challenge raised by new technologies is seen by composers in the mastering of new tools that will allow them to structure this sensible and intelligible order. Integrating operative functions (mathematical laws) with sonic transformation processes (musical laws) and with the development representative of the solution (esthetic choices), proves that there is on an information theory level (in itself attesting to an objective reflection of a scientific tendency), a musical implication beyond the solution. In this context, solutions or tests generated by the machines have a determining significance: subject to a user's musical choice dictated by the computer, these solutions were conceived on a strict objective and scientific level. As a result, an intuitive know-how, originating from one's own sensibility, will determine the musical possibilities that exist. This subjective control of technological data provided by the machine, allows one to re-integrate the mental participation, the musical subjectivity necessary to the transformation of any concrete solutions in music. Even in the computer assisted composition, the initial computer theory develops a material without any definite musical trajectory. As a general rule, it is the selective process (subjective in itself) which determines the musical possibilities carried out in the solution provided by the machine, and not the solution in itself. At this stage, before any programmed computer calculations, the sensibility or intuitive representation of sonic phenomenon, corresponds to esthetic criteria which have no explicit connection to the technological experience. This re-conceptualization stage shows the separation of "abstract sensibility" from "concrete intelligibility", and the transition from science and esthetic, to objective and subjective. The second stage, transition from computer to musical writing, transfers concrete intelligibility (known as formalism or computer theory) into abstract intelligibility (known as esthetic conceptions), thus giving form to the sensible by way of the musical concept. The electronic musical material is endlessly extensible and might even seem purely excessive. In order to limit its size, the composer is responsible for managing both: the compelling logics dictated by the machine as if they were proceedings in composing and, in keeping with his esthetic intentions, the technological response to his choices. As we can see, there exists a deontological contradiction: in order to develop musical possibilities, the theory (by way of logical formalism) provides the solution, is subject to an esthetic arbitration in the selective phase which is attributed to the solution. In this strange and unguarded realm connections emerge between ideas and production performed according to a sacred ritual of meditation between the pencil, the paper and the musical instrument. The acoustic instrument has disappeared, and unless the composer uses the score editor, the pencil and the paper are still mediators between research and rendition of the final choices. This traditional proceeding materializing after a period of experimentation, plays a fundamental role in the creation of sounds and sequences. This proceeding consists of two types of solution: silent tests - numbered output, or sonic tests -play-back. Computer programming requires as well an intellectual practice of rationalization which may possibly alter the esthetic experience due to the theoretical rigidity that it implicates. In composing, it is never a matter of proving theoretical principles but of producing sonic sequences where coherence can only be measured by its intuitive representation. The distinction between research and creation is obviously limited in this case. The integration of the technological process with invention does not necessarily mean that the elaborated theoretical research for a given work is the conclusion of computer composition. Music remains the main objective. The project integrated with the esthetic creation and also added to a scientific research project (such as psycho acoustic and cybernetic) invariably and ideally still aims at the structure of the material. The purpose of this structure is to offer composers infinite possibilities in musical production. It is here that the concept of musical research becomes evident and at which computer composition aims, for the theoretical phase of composing ICMC Proceedings 1996 487 Pecquet

Page  488 ï~~has become a scientific cognitive act surpassing esthetic framework. From a writing point of view, composers find themselves in a delicate situation between scientific objectivity and esthetic subjectivity; a situation comparable to the one found between conception of a score, where the composer is guided by a precise musical thought (the esthetic research in question), and between production, the accomplishment of this thought as expressed in the transition from writing to notation. If the computer has a superior intelligence, conscious of its functions, then it can only respond to its own properties when being assisted in its objectives. Therefore in order to establish the relationship, it is the composer who assists the computer rather than the computer who assists the composer (unless he is supervised by an expert). The translation of solutions within musical language remains the pride of human spirit which can only interpret anew, a posteriori and according to its own uniqueness, the technological consequences of artificial intelligence. Staying within the strict limit of solutions generated by the computer, even if these latter are guided by a discerning analytical thought, does not guarantee the attainability of the musical level which is obviously generated by the adjustment of the sensible with the intelligible. 2. Intuition/Reasoning Because the composer oscillates invariably between logical calculation and the expression of his sensibility, there is between intuition and reasoning, a chronological break down of the musical operation, rather than an opposition to his nature. The relationship between intuition and reasoning is reciprocal. On the one hand, it represents the transition from compositional thought to computer practice; on the other, the transition from computer thought to compositional practice. Intuition is responsible for bringing us back to sensibility which is precisely what motivates us to make certain choices and to operate them. It provokes processes that are irrational since they depend on musical concerns not encountered by definition in information technologyl. This intuition immanent to the esthetic creation comes close to Kant's "empirical intuition", described as an aposteriori knowledge arising from a received sensation. This kind of intuition rejects determinism and formalism determined by logical thought (in scientific term), and bases its acknowledgement on empirical phenomenon that governs immediate cognition, such as, sensations, feelings, emotions, and integrating as well the socio-cultural dimension. In line with this expose, this intuition paradoxically represents everything that tends to reject the rational logic required by any computer system, especially in the programming stage. Cognitive mechanisms that play a role in the relationship between computer thought and compositional practice, put a dualistic tension in the methods of reasoning. Two kinds of antagonistic thoughts stand out: mental processes devoid of objectives (in logic term) which are defined in cognitive sciences as "reveries", and mental processes of a determinist type, having a precise objective and defined as "calculations". In this taxinomie of thought, some modes characterize creation, others formal logic. In cognitive science, inductive processes are distinguished from deductive processes as if they were mental processes of a specific semantic information. - Some (by induction) substitute intermediary conjunctive informations until adopting the disjunction process as a solution, and adding more information: i.e. taking intermediary steps in problem solving, different ways to go, for example from point A to point B; - Others (by deduction) proceed by a mutual process of elimination of incompatible propositions and only retain what conforms to them: by excluding compared inadequate solutions according to a pre-elected agreement; the point is to conform to the logic clarifyied in the program. When excluding intermediary situations, the transition from compositional thought to computer practice favors deductive mental processes. The decision connecting each mode of thought, composition and information technology, is made by going from one to the other: from an intuitive, representation of material, infered from a sensible knowledge of technically sonic phenomenon (subjectivity), to a disensibilisation or objectivity within computer code. In information technology, to introduce certain pitch in a given sequence for example, does not imply that one must solely indicate maximum and minimum ambitus, nor show a filtering mode, or interpolated curb between values, etc...; but also to deduce their consequences on perception by reducing to the maximum all intermediary situations. In other words, it is a matter of clarifying all implication in the composer's "reverie" in order to optimize the computer's job. "A program must reduce to the minimum the amount of time necessary for determining the validity of an inference, the longer the time, the more difficult it is to know if a program will lead to a decision or if it will continue endlessly". Whereas the transition from computer thought to compositional practice seems to favor the inductive mental processes, choices extend beyond the determinist framework of solutions that the machine proposes. In this case, it is a matter of establishing the subjectivity in the material by way of an auditive experimentation of the solutions. In composition, the deductive and essentially non determinist process is always the subjective issue in regards to solutions, earmarking the integration of calculation for a rational explaination, thus expressing in an often not very coherent way the analytical interpretations of the work. This explains why the best programs are those that allow a simulation of arbitrary choices all the way to their structures, thus making it possible for the user to intervene and to modify both the op~erative orientation (the operations for calculation) and their parameters. Because the decisions are made equally and almost independently in favor of formal logic, as if they were desired expressions connected to specific extralogic requisites, in certain compo~sitions, some ideas have a more intellectual nature than others. They will encourage the compo~ser to respond to a distinct complex problem by developing a theoretical system as paradigm of a given phenomenon. In other compositions on the contrary, it will be a question of staying closer to the emotional sensibility by favoring the associations of intuitive thought that are independent of a formal theory. Pecquet 488 ICMC Proceedings 1996

Page  489 ï~~The types or instrumental groups express this idea, as well as to equally reveal these requisites: requiem and opera obviously do not have the same compositional objectives. 'Thus the transition to information technology requires the composer to describe at first, logical principles while situating them respectively in a musical application. The uniqueness of the processes will depend inevitably on specific orientations that the computer takes: both the choice of theoretical devices (for example softiwares) and mathematical functions (computation and rules). It is therefore formalization which becomes the composingmethod that can connect thought to execution, and idea to its creation. If composition brings about a speculative type logic which is dictated precisely by organized rules of any writing system, one must therefore distinguish it from the logic that information practice dictates, i.e. operational logic rather than cognitive logic. In information technology, logic has first of all, to acccomplish precisely what it is being asked to do. It must describe exactly the steps leading to the solution, by calculating and reasoning that contribute to the efficacy of operations. The definition of sounds and structures that give them life is set and found in the descriptive stage2. In this area, mathematics becomes the essential metaphor in explaining the sonic phenomenon, precisely because, still today, communication with computers is almost impossible other than with operational languages. thus, in inlormation technology, the reasoning logic requires the user to describe relations or processes that must imply knowledge of mechanical principles that operate objects or bring about conditions. thus, in information technology certain operations are of prime necessity and the recursive definition is fundamental. As far as the founding principle of form is concerned, today the musical idea inspired from the rhetorical code (theme, motive, sequence...) becomes obliterated. Instead, the process inspired from logical formalism, though inconstant, is the way to the future. From this evolves a teleologic transition, from a syntactic area to a dynamic area. 'there is therefore a difference between the presentation of relations between object, mathematical operations, theory of numbers regarding the equivalence between objects and actions; and the representation of objects, known as conceptualism. On the one hand, semiologic operation shows the difference between concept and object by proposing an empirical interpretation of phenomenon; and on the other hand, mathematical operation shows their unique cohesion in the operative optic which makes them easily perceived. lundamentally, the one and the other do not seem foreign to music. [or the composer, musical idea which is already at the transition stage from writing to notation, directly defers to intuitive knowledge which demands an apriori representation of the sought alter object or of the desired process. Intuitive knowledge often comes to mind spontaneously when it is not provoked by methodological mechanisms of musical writing, such as the drawing up of a form, organization of parameters, scale of value, etc. On its own, computer practice requires an a posteriori knowledge of the object or of the process as well as a defined description. It is limited by its abilities and demonstrates the challenging ambivalence encountered in computer composition as it races the hazards of creation. The idiosyncratic nature obviously shows through in the structure of programs that are developed by and for composers. In as far as they are complex systems that include varied levels of intervention, such as programs, under-programs, functions, etc... these programs in themselves convey the spirit of their creator with their specific approach to composition. For example, the library "Esquisse" created by Tristan Murail, composer and head of "the spectral school", based on complex and dynamic sonic properties:. Conceived in "LisP" language (List Processor) and programmed on"Patch Work" (visual interface), this library includes models (or operative functions) that mainly respond to the demands of spectral composers or to other non-spectral composers, (since the spectrum is an integral part of sound). This library comprehends the following: a sub-library of harmonic sequences, programmed calculation of specific parts in a basic given sound, a model than can draw out parts according to a given order, the match for inferior harmonics, etc... Another sub-library for modulations, frequency modulation, modulation in links, etc...; a sublibrary for processing harmonics as well as their analysis. Henceforth, certain models such as the "fundamental virtual", based on the algorithm of lerhardt, have a direct musical impact that can influence the writing due to their response to pertinent harmonic criteria. This model allows one to reconstruct on a sonic aggregate a fundamental virtual, using analytical criteria that ascribe a perceptive importance to each part of the harmonic sequence3. At the same time, it seems that a developing technique is perfectly represented in the machine when analysing the models of the "Combine" sub-library that were developed in the same language found in "Esquisse" by Mikhail Malt for composer Brian Ferneyhough (the latter known for its leadership in "complex" music). This developing technique, founded on combinative calculations, is used in Ferneyhough's complex piece. Using the functions that were developed in "Combine" allows for an apparently close reproduction of solutions similar to those encountered in complex matrix of calculations that the composer uses. Such programs exemplify the cohesion of a system and of an idea, of a composer's style and personality. '[hey indicate as well that the differences between intuition and reasoning can be sensibly lessened when the program responds exactly to the composer's mode of thinking. 3. The Tool and Its Usage F or a long time, two levels of mastering the tool have been esssential: 1. the composition of algorithm in abstract code, 2. computer programming and the choice of input parameters. 'Today, these levels blend into one: a simplified program that includes musical components and composition, obviously lessening "the lack of sensible causality in computer material"4. These two levels are today accessible and complement themselves in the practice. In the programs, visual interface and graphics, the code acts as an intermediary, representing program and traditional notations. The first (the composition of algorithm) is homogeneous and only has an indirect relationship to music. It practically does not influence the ICMC Proceedings 1996 489 Pecquet

Page  490 ï~~musical sensibility, other than in its austere and distant presentation of a symbolic musical representation. This composition requires the user to remain at a constant level of abstraction and guarantees a certain integrity in logical thought vis-A-vis esthetic research. If the principle of visual connnections between different "operational units", which is a typical principle in graphic programming, is to secure a wider empirical freedom, it can also confuse the interior logic of thought during the programming phase in coding. Because with "play back", it is possible to correct the listening, this principle can therefore challenge the existence of certain musical principles known in the programming phase and denied to the hearing by a sensible correction of the ear. But, is this not a justified contradiction since it assists musical reason? For the purists, graphic programming is a short cut compared to programming in code, and it fails the mastery of logic that is required by the information system. This does not prove anything musically speaking, even if we must admit that this attitude leads to an arbitrary and approximate situation, most likely conflicting with theoretical work found before the solutions. However, the adequacy of scientific and artistic research is fortunately not the ideal belief of computer composition. It is not a matter of writing an instrumental piece by being satisfied to just unconsciously assign to instrumentalists logical instructions. The complete control of information logic does not mean mastering the emotions and thus creates this ambiguity. From one universe to another there is always an objective and subjective stake which can only be resolved by composers and only by means of an esthetic experience. What composers prefer best is to reduce the machine to only an instrument of music. However, with caution, the delicate borders between intuition and reasoning disappear as soon as the machine goes beyond the functional role and the determining logic has a chance to regain its rights. The same way a specific tool produces a given solution, a specific synthesis creates its own type of sound. The difficulty in integrating strategies of writing in the machine comes from the questionable relationship between two kinds of thought as mentioned above: calculation and creation or theoretical speculation and esthetic choice. In creation, the imaginative abilities, expressed by the transition and the translation of ideas to their completion, are encouraged by multiple causes which are often external or foreign to all determining principle. In addition, the auditive representation of technically sonic phenomenon (of material) resists the fact that a composer often does not have the technical control of his tools. To him, the technological tool seems often disagreeable because of its internal determining function mode. ln spite of the incompatibility between formalism which is dictated by the technological instrument, and the organizational sense that is required for compositional thought, there is an inevitable confrontation between one's own thought and the determining logic of the tool (use of a given machine, studying the manual or program training). As a consequence, the composer finally compromises between two connected attitudes in order to balance the implicit and explicit part of the relationship, to think in order to formalize, formalize in order to rethink and to validate his choices. Practicing the material has little in common with training on a given machine. In information language, mastering the programming must not exceed mastering compositional technique, since it might unbalance the relationship between intuition and reasoning and might increase the speculative work rather than creation. There is therefore a limit on how much to tolerate in this relationship between compositional thought and logical formalism. This limit changes the codification of music intelligible in object and the "decodification" of the object into music. It takes the course previously discussed between user and information tool. Programming embodies the abstract construction of the tool, the instrument in the machine. It simulates a compositional system without using continuous automatic inputs, dictated by the user. A simple operation which responds to a precise algorithm can amount to a unique command. Integrated to other algorithmic functions done in sequence, this operation is enough to produce ultrasophisticated solutions. This is not a matter of compositional short-cuts (as it is with truncated writing), rather it is about controling rules determined before the composition and applied automatically as if they were esthetic conditions given by the user. These "short-cuts" have obvious esthetic implications. Making certain steps automatic when composing, certain motions are inhibited, sensibly neutralized and others exacerbated. Very few softwares conceived by composers can meet compositional requirements. The large multifunctions model "Situation", conceived in the PatchWorkLisP language by Camillo Rueda for composer Antoine Bonnet, is an example indicating that the more a program meets the specific requirements of the composer, the more it will only interest a small number of users, though it is a bigger satisfaction for the conceiver. Adapting foreign programs restricts composers to acccept some programming rules that do not necessarily correspond to their needs, nor agree with a logical system. Research can only lead to creation by submitting implicitly and quietly to scientific objectives where its existence has been recognized. Creation is only the product of research and a further theory of found solutions. The idea of recognizing the sensible in each discipline, music and technology, allows the truth to reveal itself when those disciplines cross each other's path. References [Johnson-Laird, 1994] L'ordinateur et l'esprit. Published by Odile Jacob, translated from the English by Jacqueline Henry, Paris, 1994. [Agostino Di Scipio, 1994] "Formal Processes of Timbre Composition Challenging the Dualistic Paradigm of Computer Music" Proeedings of the 1994 International Computer Music Con~ference, Aarhus, Dannemark. [Terhard, Stoll and Seewann, 1982] "Algorithm for extraction of pitch and pitch salience from tonal signals", Journal of Acoustic Society of America 71 (3), March 1982. [IDuchez, 19911 "L'6volution scientifique de Ia notion de mat6riau musical ", published in Le Timbre A4taphore pour la Composition, edited by Christian Bourgeois, I.R.C.A.M., Paris: 1991. Pecquet 490 ICMC Proceedings 1996