Should Music Be Composed With the Aid Of Systems?

ca. 1959

Ivor Darreg

Science, engineering, and mathematics have become indispensable factors in today's world. It is not surprising, then, that attempts have been, and are being, made to apply modern technical methods to the composition of music, as well as other artistic endeavors.

Many musicians and composers, though, are puzzled: why can't the creative artist trust to inspiration and intuition? Can emotions actually be engineered? Doesn't a composer need talent? The acknowledged masters got along beautifully without systems; why can't we? How about improvisation? A whole string of questions of this sort comes to mind. On the other side, there are just as many: if an architect worked without system, the building would fall down, perhaps before it was completed. Sound recording and reproduction has made marvellous strides in the last few decades; why can't composers take advantage of this fact? Far more is now known about the physics of sound and the physiology and psychology of hearing,than was known in the time of Bach, Beethoven, or even Brahms. Why not a system based upon this knowledge? Since the Ancient Greeks, mathematical calculations for the sake of music have been made--why not see what can be done with modern mathematics? With all this and more to be said, one has every excuse for being puzzled.

Some composers, especially those who go in for so-called popular music, complain of getting "written out"--the fount of inspiration appears to have dried up, and they wonder if a system will help them. Other composers, forced to listen to background music and neighbors' radios and hi-fi sets whether they will or no, start wondering if all this exposure to hackneyed music as an involuntary captive audience will hamstring their originality and unconsciously compel them to timeworn patterns.

Much attention was attracted in the late 1940's and early 1950's by the system of the late Joseph Schillinger, appearing posthumously in two heavy volumes at a substantial price and accompanied by another book with the audacious title The Mathematical Basis of the Arts. Proponents of this system have indulge din plenty of name-dropping, but until recently no one has bothered to investigate the claim that the system is scientific. Dr. John Backus, a physicist, wrote an article in the November 1960 issue of the The Journal of Music Theory, entitled "Pseudo-Science In Music" and giving the Schillinger System a thorough debunking treatment.

We recommend the reading of this article, and agree wholeheartedly with many of its criticisms, but after a careful re-reading of Schillinger's books and investigation of the attitudes of several musicians toward the system, we do not feel it cane be dismissed entirely as quackery. Those composers who have had suitable training can pick out from the system many items of use to them, but anyone who accepts at face value Schilllinger's arrogant boasts that these books of his contain all that is needed to make a composer out of a diligent, but untalented and uninspired person, will be in for bitter disillusionment!

Perhaps the hardest thing for any artistic musician to take is Schillinger's graphic analysis (appealing to geometry, of all things) of the main theme of the third movement of Beethoven's Sonata Pathetique, and then his claim that the melody can be improved (?!) by going things to its "axis"--this annoyed Dr. Backus enough so that he categorically stated that "there is no 'axis of melody'" at all. The present writer has prepared a supplement to this article giving the graph in question and another graph revealing the context in which Beethoven set his melody: once this context is taken into consideration, you can see how presumptuous it would be to try improving on beethoven. For that matter, how would any composer as busy and as full of novel ideas for compositions as Schillinger said he was, have the time to go around trying to improve on the classics?

In the present writer's opinion, Schillinger must have had training in only certain branches of the vast field which is mathematics, and thus had attempted to twist such unlikely departments of that subject as geometry and the algebra of polynomials to the purposes of music, when many other divisions of the mathematical field would have been far more suitable. Or take the case of a sequence of numbers which Schillinger arbitrarily applies to semitones to produce and ascending series of notes: when the numbers become so large that the notes climb up too high above the staff, Schillinger arbitrarily deducts one or more octaves--in one place, this saves him from leaving the range of human hearing, but 5 octaves subtracted given a repetition of the same note. Wouldn't it have been more sensible to have selected a series of number that did not increase so rapidly, or that alternately increased and decreased, and thus avoid making fun of his own system?

In other places, he actually piles up a series of chords which goes three octaves or more above the audible range,and labelling such an impractical affair "a hypothetical case" does not make it any more valuable to composers of audible music. It is some relief, then, to be able to recommend the tables of scales and chords possible within the usual 12-tone tuning system, and some of the data on permutations and combinations. While these are not the only such tables in the world, they are as good as most. The harmonic combinations suggested in the system are suggestive of post-Wagnerian chromaticisim and of polytonality, but deliberately not too helpful to atonal writers. We find a 19th-century conservatism where we least expect it.

The system does not pay much attention to acoustical laws, nor to the intensive work done on theory of harmony by Helmholtz and others. The customary 12-tone equal temperament tuning is accepted without any explanation being given--no whys or wherefores--and the conventional musical notation is taken for granted without any explanation of the odd way Schillinger uses enharmonic equivalents. The graphs are alleged to show tonal relations more accurately than the conventional notation does, but they leave out information contained in the ordinary notes and markings, and terms like "axis" and "trajectory," denoting structural features that are not discernible by the ear or particularly relevant to music-as-played-and-heard, are bandied about, giving the impression that geometrical designs ought to be more important to composers than any sounds. Indeed, Schillinger had his pupils write music from numbers in telephone directories, stock-market charts, the silhouette of the buildings in the Manhattan skyline, etc. This is better than inspiration? Well, Schillinger said it was not only better, but was more scientific and efficient!

If this were the only available or possible system of musical composition, or systematic aid for composers, we would be in a sad way indeed--but fortunately there are other systems, past, present, and future.

The composers of so-called "atonal" or "12-ton" music often follow the tone-row system, pioneered by Arnold Schoenberg, and elaborated and discussed by a number of authors, so we do not need to set it forth in detail here. The main principle of a tone-row is that all 12 of the tones of the usual piano scale must be sounded before repeating any one of them. The number of such tone-rows possible is factorial 12 or 479,001,600, which should give composers plenty of resources! The basic idea of a series of permutations applied to music is at least as old as the custom of change-ringing, still widely practiced in England. Actual compositions based upon a tone-row or derivative method use more sophisticated techniques than merely sounding the notes over in succession. The value of this system can be left to the judgment of posterity, as it is still evolving.

Those who follow the dodecaphonic system will find themselves in plenty of company, so no one need be timid about trying out the system. Again, here is a system that does not lean very much on acoustical principles, but rather it results form accepting the keyboard, tuning, and tone-quality of the piano as found at the beginning of the 20th century, and imposing these piano-inspired and piano-determined concepts upon all the other instruments,and upon the singing voice. On paper, these 12-tone compositions look rational, orderly, and disciplined enough. Sometimes this is also evident in the way they sound when played, sometimes not. No composition system can guarantee success. The present writer has employed the atonal or dodecaphonic system only seldom in his compositions, and does not plan to make more than occasional use of it in the future. It has its uses, but these comprise only a small portion of all possible music yet to be written. Schoenberg himself very wisely cautioned up-and-coming composers not to adopt atonality unless they felt they had to.

In the author's opinion, the weakest points of the 12-tone system are: [1] its dependence upon certain characteristics of the piano, which means that if the piano ever loses its commanding position in the present-day musical world, so will the dodecaphonic system, sooner or later; [2] the need constantly to avoid and sidestep melodic and harmonic progressions which would establish a sense of tonic or key in the 18th- or 19th-century manner; [3] the very real difficulty of carrying one's orderly musical thoughts and ideas conceived according to this system over to the actual sounds that will be heard by one's listeners when or if the music gets played. It has been our experience that this type of music almost always sounds more chaotic than it looks.

The analytical method has been used to teach composition for a long time. Everyone in composing, except possibly the most rebellious of the self-taught, has come under its influence, and perhaps they have also, for they would have to find out what they were rebelling against before they could proclaim themselves as rebels.

The analytical method is necessarily retrospective: the compositions of recognized masters are dissected and from this and his own prejudices and theories the analyst constructs what he fondly hopes will be a blueprint for further masterpieces. The models to be dissected and tortured till they yield up their secrets may be from any period, Palestrina's to Hindemith's, but thoroughly obsolete and worked-out styles are preferred, such as "strict counterpoint." The average time-lag between the date of composition of the model and the present time is a century, and just how one's learning to compose reasonable facsimiles of Haydn symphonies or Bach chorales will make one more up-to-date and ready to go and compose something in tune with 1970 than someone else would be who undertook the analyze the music being written right now--well, it just doesn't make sense. It is fine to be conversant with the history of music, but the attitude of pretending, with delusional fixity, that it is still 1870, would never be tolerated outside the musical world. Once we even met a composer who dressed up to look like a reincarnation of Schubert, then later got tired of this and decided to become Beethoven--possibly on the ground that it would make is displays of temper a little more in character! "Tchaikovsky could well have written this!" was his proud boast on showing one of his compositions; "Anyone would surely believe that Debussy had composed this!" he said on showing another. Such is the result of the analytical system of composition--indeed, if all the imitation-Chopin ever written were laid end to end, it would doubtless reach to the moon.

Recently, the book Experimental Music by Hiller and Isaacson set down very clearly the possibilities of automating the analytical method. ELectronic computers can now be programmed to perform collating operations stylistic analyses, which means that "standard average Schumann" or "typical Ravel" could be turned out by the yard on demand. Let it be to Hiller & Isaacson's credit that they hastened to add that such a procedure would be inartistic in the extreme.

But no sooner had they written those lines than a phonograph recording company came out with an imitation-stereo recording of a Toscanini concert, while articles appeared in several magazines about a computer that "composed" popular song tunes, and just recently announcements have been published about a computer system for analyzing the style of a composer and then duplicating it. This proves that some people are not as scrupulous as were the authors of Experimental Music. Now that we may expect any day to find recordings of Beethoven's Eleventh and Thirteenth Symphonies in the music stores along with Mozart's brand-new Requiem for the Hydrogen Bomb Victims, the analytic road to imitating and emulating the great composers of the past will lose much of its glamour.

With your neck twisted around 180 degrees and your eyes and ears intent on what was done over 60 years ago to the exclusion of all else, you cannot hope to write something to help musical progress, no matter how thoroughly and conscientiously you study the greatest music of the past. Bear in mind that linguistic science has made seven-league strides since they got over the traditional philologists' approach of typing everything found in contemporary languages back to ancient times. When they studied present-day English and French and Spanish on their own terms as living, spoken communication media, a whole new vista opened. So with music, one would think.

Thus the analytical system of composing should be updated.

We now come to the stochastic system of musical composition: composing by chance. No less a figure than Mozart had a method for composing by throwing dice, and such stratagems as dealing out cards and spinning a wheel have been tried,and of course people have drawn slips of paper from a hat. Perhaps it is uncharitable to say that much contemporary music sounds as if the notes had been thoroughly mixed up and then drawn by lot, but we'll say it anyway.

Readers deeply interested int his subject are advised to read Hiller & Isaacson's Experimental Music, wherein the application of the stochastic compositions system to an electronic computer is well described. Note that analysis of musical style by computer is not the same thing as "composing" with a computer operating on random numbers. While the same machine could do both, its programming by a human being for these two tasks would be radically different, so that in effect one would have two different machines. Neither of these two computer functions has anything to do with performing the music automatically. The music stochastically composed may be performed by ordinary musicians on conventional instruments, and this has been done a number of times. Conversely, music may be composed without the use of any system or machine, but played by automatic instruments.

The branch of mathematics called permutations and combinations, already referred to in this article, can be applied to composition in several ways. For example, a basso ostinato of four notes may be run through factorial 4 (1 x 2 x 3 x 4, or 24) permutations before the notes come out in the same order they had when first played. A five-note theme has 120 permutations, while an 8-note theme would have 40,320 "changes that could be rung" upon it before coming back to the starting order. (It is understood, of course, that all five, or all eight, notes are different, although one of the notes could be replaced by a rest.)

A chord, with four different notes, such as the dominant seventh or the added-sixth chord, has 24 permutations, just as the 4-not theme did, and this does not count the variations possible by inserting an extra spacing of an octave or two. Rhythmic groups maybe permuted also, as for instance a triplet figure maybe placed on the different beats of a measure in turn.

Various tables of permutations exist, which may be of service to composers. There are also tables of random numbers available, which can be used to provide a completely chance element when that is desired. Any composer could do well to compile some tables of his won, in the manner of a score-card or business record, so that he may keep track of the basic patterns he carries over from one composition to another, varying them enough to avoid monotony, but keeping enough consistency to have a definite style--if he wants to be thus consistent, that is. The application of permutation and the use of chance processes can be a great help to the composer who feels he is in a rut.

Just as there are novelists who write the same story over with different characters and setting, so there are composers who write the same piece over ten or twelve times, changing little beyond the title. A system of pre-shuffling the basic theme or harmonic progression, or the rhythms, before starting the next piece, should be really helpful. This is also true of improvisations, though there may be serious practical difficulties to applying any system on the spur of the moment when one is confronted with both an instrument and an audience. This could be why duet improvisation works better than might be supposed: each partner in effect reshuffles the other's ideas, helping to prevent unthinking repetitions of the kind that occur in solo improvisations.

The experiments which have been carried out,and will go on being tried, to use a computer or similar device for "composing" are naturally disquieting to any composer or prospective student of composition. Reassurance is certainly in order. Composing by computer is actually composing at the second or third order. Human beings still take part, albeit indirectly. First someone feeds the rules of counterpoint and/or melody-writing and/or harmony and/or various other phases of musical practice into the "memory" (data-storage department) of the machine. Now these rules and principles were not the result of the blind grinding of some robot's gears; they were deduced by one human being (the musical analyst) from an examination of the compositions written by other human beings.

The very fact that these rules change with time and are not as stern and unyielding as are scientific laws, should reveal their human origin. The machine containing a coded equivalent of musical rules in then programmed by another human being to go through various mathematical manipulations of the stored data. Generally this consists in applying a series of random numbers, or numbers random for all practical purposes, to the stored rules, and rejecting automatically any results that do not conform to such rules. The real, human composer would break the rules now and then when he chose to, for he is a free agent to some extent at least. The machine cannot do this--at least not for the time being. Finally when the machine puts out the results, some human being selects only a part of the results, and consigns the remainder to the nearest wastebasket. Also, someone decides when the computer is to be shut off!

Thus we don't have any mechanized Grieg or robot Weber or Mendelssohn, but rather an automated Ebenezer Prout, and subejct to human veto--which most music students would consider an improvement over the original! Theoretically,the machine now does the drudgery, the clerical work, leaving the composer free for the higher-order tasks. Remember that the machine's work here is still retrospective, and that if the machine is fed the proper data, it can be strictly up-to-date in its methods, but cannot advance even one step beyond today's musical practices. The main advantage in using a machine is that all the rearrangements which would take a composer hours to try out, can be run through by the machine in a matter of minutes or less. The other feature, which may or may not be a real advantage, is that the machine will always observe all the rules supplied it, and the composer may forget to, since his attention and alertness may falter.

It is worth noting here that popular songs and dance arrangements, as well as what is euphemistically termed "background music" or even "neutral background" for films, TV programs, offices, stores, and the life, is subject to even stricter rules than the more serious type of music. Naturally enough, such music is more amenable to "systems;" and this may well explain the comparative success of the Schillinger system among Hollywood film-score composers and the like. Similarly, we may expect the application through machines of the stochastic system to have more success in the background-music field than in the serious music domain. The composer of such music as popular dance-tunes or background for something else has to to contend with stricter rules and more of them, and because of this he must make more trials and spend more time making those trials before something conforming to all those rules turns up. Inconsequence, he will feel more need for systems or machines or both.

The techniques of probability theory and statistics have been used to analyze the styles of various composer sand thus to enable their imitators to do a better job of imitating. A more legitimate use of this technique is by scholars seeking to establish the authenticity of spuriousness of a newly-found work said to be by such-and-such a composer.

it is quite possible that someone with the proper mathematical bent might take the new theory of games, a rapidly-growing branch of mathematics, and evolve some kind of musical composition system out of it. The fondness some musicians have had for chess, checkers, solitaire, tennis, and other games may show an intuitive grasp of this similarity between music and games. Again,there is a remote affinity between the fugue and other involved contrapuntal forms of music and the solving of puzzles or the working of chess problems.

The presnet writer, is exploring the possibilities in the quasi-mathematical fields of symbolic logic (sometimes considered a branch of formal logic) and structural lignuistics. He expects to be engaged in this for some time to come, and will of course find collaborators, as these are vast unexplored fields, and only a little has been attempted in the application of logic to music, while structural linguistics is too new for anyone to have thought of applying it to musical composition.

There is beauty in logic, whether this is generally realized or not. A musical composition, especially one of considerable length, is more apt to hold our attention and stay in our memory if it is logical in nature and form. The opposite of a logical composition is of course an incoherent, aimless one. harmony and melody exemplify relations among tones, while rhythm exemplifies relations of duration and intensity--thus the branch of symbolic logical called the calculus of relations seems a possible aid to composing music. We are thinking here in terms of aids, auxiliaries rather than systems, as it is our belief that a complete, exhaustively-developed system is too much of a restraint on any creative artists. But this should not deter him from seeking a little help from here and a little from there and the other place when it is available and appropriate to the problems of the moment.

The case for deriving aids to composers from structural linguistics is, if anything, even stronger than the case for logic. Language is a human phenomenon. People talk languages, and they do so according to unconsciously-acquired principles of grammar, phonetics, and syntax (not exactly in the schoolteacher's meanings of these terms, we hasten to say) and the newer breed of structural linguists discuss this under such headings as patterning and behaviour inspired by situations This is easy to conceded, without having to make up one's mind on the older question wether music is or is not a language. More than this: written musical notation is no more "phonetic" than is the spelling of most languages, so here again music and languages share a problem. Form what little we have said here, it should be obvious that a comparison of language patterns with musical patterns should give promising results.

There has been some research in the application of Gestalt psychology to music. Future research in this direction will no doubt be very fruitful. As an introduction to this idea, consider that a melody is much more than the sum of the notes which go to make it up. Consider further the various transformations a melody can undergo and still be recognizable as "the same tune." So, if anyone comes up with a theory of musical gestalten, it will doubtless be of value to composers.

Since music is a form of audible communication, and much music is recorded and/or transmitted over various types of communication systems, the mathematical theory of communication is relevant to music. The present "high-fidelity" craze evinces some awareness on the part of the public that music is a form of communication, but this is only a beginning. There are some interesting problems raised by information theory, quite relevant to music, particularly from the composer's standpoint. First of all, music is "highly redundant," in the terminology of information theory--almost all pieces ever written contain numerous repetitions, imitative passages, and restatements. The recognized musical forms depend upon pattern repetitions, recapitulations, and restatements of a theme in another key or slightly altered in configuration.

In this connection, we must no forget the profound influence exerted upon nearly all composers by their practicing regualarly on one or more instruments, which deeply engraves in their nervous systems both the musical patterns most congenial to the particular instruments practiced on, and the exercises, etudes, pieces, scales, and whatever material is habitually practiced on such instruments. Certainly the musical world at large is the poorer because of the countless instances in which composers have written scraps of Hanon and Czerny finger exercises and needlessly-multiplied scale and arpeggio figures in their compositions instead of putting in something more original, something properly the composer's own idea. The pity is that this has mostly been done entirely unconsciously! We scarcely need information theory in cases of this kind to tell us that such music does not convey as much information as it might. Right here we see another danger of composing by system: if the system is at all comprehensive in scope, it disciples will be in effect rewriting or arranging one composition over and over. Instead of communicating their own information to the audience, they will be communicating a more or less encoded version of information already contained in the composition system.

Artificial restraints of this kind necessarily deprive the composer of possible avenues of communication, and thus of expression. We must deplore the current vogue of chord organs, that is, new instruments patterned upon the accordion keyboard principle, which compels all music played on instruments of this pattern to be a solo in the treble range with an accompaniment in the bass limited to a standardized selection of chords. In popular music, the notation for chords, while not so limited as is the selection of accordion buttons, also operates to restrict the possibilities available to the composer. This solo-and-accompaniment system excludes a great body of music, for example Bach, Debussy, Schoenberg, and practically everything contrapuntal or outside the key-system. The solo-and-accompaniment system is seldom labeled with a big sign saying THIS IS A SYSTEM OF COMPOSITION, but it is being built right into instruments which are going into many homes, churches, and schools, so its influence will continue. Furthermore, it is descended from the figured-bass system of respectable antiquity, which means that jazz and dance music generally is much more conservative than the general public realizes.

The tonality- or key-system is fighting for its life right now, so little comment is needed here. Suffice it to say that those who choose to write music outside the key-system are often faced with terrific problems to keep their themes or harmonic progressions from appearing to be in some tonality.

The anthropological approach is almost a composition system: one takes the music of some exotic people and either tries to compose along their lines or more often adapts one or more of its features to standard Western European instruments. Those who attended silent movies in their childhood will remember the pieces that were played to give an Egyptian, Arabic, Chines, or other atmosphere, and it is likely that most of these atmospheres would be quite unrecognizable by natives of India, China, Arabia, or Iraq.

The harmonic series, or the overtone series as it is also called, has served as the acoustic basis for erecting many systems of harmony,and also is intimately involved in the design of musical instruments, chiefly as regards their tone-quality. For example, the tone-quality of the piano has come to be deliberately designed so as to eliminate those overtones not congenial to the usual 12-tone equally tempered scale. The Hammond organ went even further by deriving all its overtones from a series of tempered tone-generators. This gives it a peculiar insipid tone-quality which has not been imitated either in other instruments produced by the same company or in most instruments produced by other firms.

Scriabin and other composers have employed scales or chords based on the nearest piano-keyboard representations of the higher overtones, such as the eleventh and thirteenth harmonics, while other persons more intrepid have composed for scales with more than 12 tones to the octave, in order to represent the higher members of the harmonic series more accurately. Now that electronic and automatic instruments are coming into their own, experiments in this microtonal direction may become more successful. Already the new ability of controlling the overtone structure of all the notes on an instrument is creating a new interest in the harmonic series.

One of the oldest systems of compositions is that of writing counterpoint to a cantus firmus, a melody usually put in long notes, taken from a folk-song or the like, and given to the tenor part. This custom eventually went to absurd extremes, such that the cantus firmus became completely unrecognizable among the other parts written around it. modern jazz often uses a system which can be considered a manner of writing to a cantus firmus, with counterpoint of an improvisatory or pseudo-improvisatory nature taking place against a rigid, almost unvarying background.

The theme with variations form does not quite constitute a system, but it deserves mention here as bearing a logico-mathematical relationship to the cantus firmus method. This is especially true of the many sets of Variations on a Theme by Someone Else.

Most recent, perhaps, of the composition systems--though its results can be highly unsystematic in sound--is the tape-recorder-manipulated music, one variety of which is called musique concrete. Sounds are taken from ordinary instruments, from electronic instruments, and in musique concrete from street and machinery noises, as well as the cries of humans, animals, and birds--and these are recorded on tape, re-recorded, passed through electric filter circuits, played back at high or lower speeds, superimposed, and otherwise transmogrified. At one end of the spectrum the results are typical musical composition, while at the other end it is doubtful whether the montage of sounds can be called "music" in any normal meaning of the term. Let us be charitable and speak of manipulated sound-effects. One form of this manipulated music can be called composition at one remove (or two removes). In this method, already-existing music is re-recorded and recombined and even played backwards, so that often the original composer would not recognize his own brainchild.

The programming of an electronic computer (or similar machine) to play music is not necessarily a system, since this might be done with a conventional type of music. However, the fact that the composer would ordinarily have to take on the added responsibility of including everything in the piece that performers would ordinarily take care of for him, such as vibrato, deviations in pitch, tone-quality variations, nuances in rhythm, etc., may result in the creation of methods and systems for these purposes. Much the same can be said of preparing music for performance by an automatic instrument.

In this exposition we have probably left out a few composition systems, so we hope our readers will call any important omissions to our attention for possible later publication.

Now, to sum up, let us first go back to the remarks on information theory. Why composer music at all, unless one has something to say? Perhaps, for background or to score a film or animated cartoon or to accompany some other art-form or action. In such cases the form and content may be determined by the film or play-action or the movements of the cartoon figures or the time requirements. One composer, asked to fill in some time on the radio, complied by writing a piece called 3 minutes and 20 seconds! In circumstances like these, we may speak of extra-musical systems or structures imposed from without. This is not really the province of this article.

What our title question is intended to mean is, Does the composer who is writing music to be listened to, not just go in one ear and out the other, the composer who is expecting to hold his audience's attention for a reasonable length of time, have need of a "system of musical composition"? If the composer is going to conform to a system, he is not going to put out much information in the time it takes to hear his piece. That is, the course of the composition will be too predictable. If he has to lean on the crutch a system provides, then perhaps he should be an arranger or conductor or performer instead of a composer. Or perhaps his real creative talent lies outside the musical field: writing, painting, or whatever.

Just as a writer uses reference books and summaries, and a painter looks up reproductions of art work to get idea, so a composer may use various aids and tables such as we have discussed--but he ought to to let the systems do his composing for him, because then it won't be his--and fortunately there are ways of being inspired, even if the most widely-advertised system-maker did not approve of inspiration.

As composing is a process of selection, and there is now more from which to select, and there will be labor-saving devices to free composers from such drudgery as copying out orchestra parts of having to wait until rehearsal to discover various mistakes, the outlook for the future is most favorable.