eliot@phoenix.Princeton.EDU (Eliot Handelman) (04/13/90)
In article <134123@sun.Eng.Sun.COM> music-research%bartok@sun.UUCP writes:
;Date: Mon, 2 Apr 90 10:51:11 EDT
;From: "David M. Chess" <CHESS@com.ibm.almaden>
;Subject: Only amateurs?
;To: music-research@com.sun.eng.bartok
;Message-ID: <9004021451.AA17888@Sun.COM>
;
;E. Handleman writes, in passing,
;
;> ...the kinds of things that amateurs find exciting, like generating
;> tonal music, or music "in the style of," or ways of making one piece
;> sound sort of like another piece, or trying to make music sound
;> "authentic" or "expressive," or arranging chords so that your tune
;> comes out moderately "happy" or "soul searching." I don't say this
;> sort of thing shouldn't be done -- I'm only saying that I don't care
;> about this kind of work at all.
;
;But you're also saying that "amateurs" find it exciting. Do you
;mean to say that professionals don't? Professional whats? *8)
;I admit that I am an amateur, and that I find at least some of that
;sort of thing exciting (I think there's still lots of good tonal
;music to be written, for instance), but might I venture that your
;reference to "amateurs" was more of a passing slap at stuff that
;doesn't interest you than it was a considered, rational, statement?
;Or is it really true that Professionals have all abandoned tonality?
;I suspect not! (Pls don't take this as a flame or an attack; I've
;been enjoying the discussion greatly, but just couldn't bring myself
;to let this get by without comment...)
;
;Dave Chess
;IBM T. J. Watson Research
You're merely substantiating my point, Dave, as does some of the
recent discussion here concerning fugues and fractal music. (Fractal
music, incidentally, is another amateur preoccupation, as are most
attempts to generate music algorithmically.) The statement "there's
a lot of good tonal music to be written" is at least 50 years old and
hasn't been demonstrated for much longer. Not to say that no good music
has been written since that time, or even good music that's triadic --
the word "tonal" doesn't (and has never) referred to music, but rather
to a theory of what holds a certain type of music together, and face it
Dave, there's just very little evidence that the theory of tonality
DOES accomodate the music that it's describing as tonal. Amateurs don't
understand that point, and figure that by generating i-iv-v cadences
via the computer they're on the road to "tonal composition," that this
is the first step towards producing the next Beethoven symphony etc.
The more you know about Beethoven Symphonies (for example) the less
likely that proposition becomes.
Now, as to professional interest: I see the NSF financing a certain amount
of work being produced by amateurs, especially neural nets that produce
"melodies" and that sort of thing. I also don't see that the NEA or other
foundations to whom musicians can turn -- generally "peer panel" foundations
-- are lending too much support to neural net research that generates melodies.
For that matter, very few professional journals appear to even be
publishing that sort of work, with a couple of exceptions. Looking over
the proceedings of the AI/Music conference, none of the papers written by
people in music departments have anything to do with tonal music (except
for Cope's paper); those that do come out of psych/cs departments.
-e.smoliar@vaxa.isi.edu (Stephen Smoliar) (04/13/90)
In article <15312@phoenix.Princeton.EDU> eliot@phoenix.Princeton.EDU (Eliot Handelman) writes: > I see the NSF financing a certain amount >of work being produced by amateurs, especially neural nets that produce >"melodies" and that sort of thing. Eliot, I think if we "look at the record," we might find this statement to be at least slightly misleading. In particular, let us concentrate of three neural net researchers and their sources of support: 1. The work reported in Peter Todd's "A Connectionist Approach to Algorithmic Composition" (COMPUTER MUSIC JOURNAL, Winter 1989) was supported by a National Science Foundation Graduate Fellowship. In other words, it was part of his research as a graduate student. As far as the NSF is concerned, graduate students are not professionals. They are students. The funds in the fellowship exist to enable them to be students without having to support themselves by other means, such as driving cabs. The primary concern at NSF is that you be a good student while you are receiving their money, which ultimately boils down to devoting most of your waking hours to writing a good thesis. Therefore, Peter Todd is probably an amateur, by your standards; but NSF was supporting him because he was a student, not because he showed any promise of providing musicians with any deep insights. 2. Todd's mentor, Jamshed Bharucha, has also been supported by the National Science Foundation. My assumption is that this money came from their psychology division (given that he is a psychology professor). In order to get this money, he probably had to write a proposal in which he put forth one or more psychological problems whose validity and interest would be recognized by his peers. Then he had to offer an experimental plan for investigating those problems which would be judged for its methodological soundness. Probably, he also said something about how the results of this work would lead to further investigations. The bottom line is that in order to get such a proposal approved, he had to demonstrate that he was a good PROFESSIONAL PSYCHOLOGIST. As a professional psychologist, he runs the risk of making assumptions with regard to both problem formulation and methodology which a professional musician would write off as unfounded. However, the National Science Foundation reviews proposals written by psychologists on the advice of OTHER psychologists. Unfortunately, we cannot expect the poor officer who has to deal with a mountain of proposals to know enough to assemble a review board for such work which will allow for input from musicians. (Chances are, he has enough trouble getting PSYCHOLOGISTS to submit their reviews on time!) The conclusion is that NSF supports him for being a good psychologist who is going to make valuable contributions to his community of fellow psychologists. Musicians who would question the value of this work really do not have any voice in the matter, for better or worse. 3. Teuvo Kohonen is in Finland. I do not think NSF supports ANY work in Finland. Certainly, his paper in the 1989 International Joint Conference on Neural Networks proceedings does not list ANY support for his music work. My guess is that he does all this "on the side" while being supported for his more scientific efforts with neural networks. Therefore, we certainly have ground for calling him an amateur; and he probably would not disagree. There is certainly nothing wrong with his present results of his personal investigation to an audience of other neural network researchers. My guess is that he would probably not consider presenting it to a meeting of, for example, the Society of Music Theory. > I also don't see that the NEA or other >foundations to whom musicians can turn -- generally "peer panel" foundations >-- are lending too much support to neural net research that generates >melodies. Unless I'm mistaken, the NEA does not support ANY research. They have so little money to give out that they devote ALL of it to supporting artists to make more art. I'm not even sure to what extent the National Endowment for the HUMANITIES supports research. I think they are more interested in general promotion of the humanities than in any speculations of basic research. Again, they are pretty short on funds. >For that matter, very few professional journals appear to even be >publishing that sort of work, with a couple of exceptions. Looking over >the proceedings of the AI/Music conference, none of the papers written by >people in music departments have anything to do with tonal music (except >for Cope's paper); those that do come out of psych/cs departments. > Professional journals basically have the same peer review problem that NSF has. Because there is such a gross over-production of scholars who must get their work in print in order to survive, we are faced with a glut of journals, each of which has a relatively narrow scope for what it publishes. For example, MUSIC PERCEPTION has more than enough material of the sort that Bharucha publishes to worry very much about expanding its horizons. It is true that someone with the credentials of a David Lewin can come along at get published there, but it is unreasonable to expect that one paper by a music theorist will tilt the editorial balance towards the concerns of musicians. Unfortunately, about the only professional journal which has tried to deal with questions of cognition from a musical point of view on a regular basis seems to be INTERFACE; and my personal opinion is that the editorial quality of this journal is very sloppy. The bottom line is that there are still too few of us trying to ask serious questions to have much of an impact on either the scholarship or the practice of music. Frank Zappa is probably right. If you are serious about music, you should go out an get a Real Estate License (or follow Ives into insurance or whatever). ========================================================================= USPS: Stephen Smoliar USC Information Sciences Institute 4676 Admiralty Way Suite 1001 Marina del Rey, California 90292-6695 Internet: smoliar@vaxa.isi.edu "Only a schoolteacher innocent of how literature is made could have written such a line."--Gore Vidal
eliot@phoenix.Princeton.EDU (Eliot Handelman) (04/15/90)
In article <12884@venera.isi.edu> smoliar@vaxa.isi.edu (Stephen Smoliar) writes: ;In article <15312@phoenix.Princeton.EDU> eliot@phoenix.Princeton.EDU (Eliot ;Handelman) writes: ;> I see the NSF financing a certain amount ;>of work being produced by amateurs, especially neural nets that produce ;>"melodies" and that sort of thing. ; ;Eliot, I think if we "look at the record," we might find this statement to be ;at least slightly misleading. Let me clarify. My remarks were intended to apply exclusively to the production of music, or research that is meant to culminate in compositional applications. My point was that there's not much common ground between those things that people in the field of music find interesting and the psychology department notion of just what a compositional application worth advertising might look like. Someone asked whether this wasn't just my own view of the matter, or whether my claim that non-music professional interest indeed did not intersect with professional musician's concerns could be substantiated. (It's obvious that I'm reading German these days, isn't it?) I said the simplest way to verify this was to see who was funding who's research. The NEA, as you point out, doesn't fund research, most probably because research into compositional applications of neural nets is not considered a germane topic among those who decide who gets the grant. Or even more probably, those who have done their homework and are aware of the history of music over the past 40 years know that hundreds and thousands of composers of avant-garde music each invented their own compositional systems, the least of which is probably more intriguing than the sorts of developments that the NSF might think worthwhile bringing to the attention of the research community. ; Therefore, ; Peter Todd is probably an amateur, by your standards; but NSF ; was supporting him because he was a student, not because he ; showed any promise of providing musicians with any deep insights. I'll wager that the NSF would have been less quick to support him if they had a general research category called "interesting ways of composing music." Not that I'm suggesting that they should. ; The conclusion is that NSF supports him ; for being a good psychologist who is going to make valuable ; contributions to his community of fellow psychologists. ; Musicians who would question the value of this work really ; do not have any voice in the matter, for better or worse. But this is a clear case where a music theorist should have been consulted. ; 3. Teuvo Kohonen is in Finland. ; not disagree. There is certainly nothing wrong with his ; present results of his personal investigation to an audience ; of other neural network researchers. My guess is that he ; would probably not consider presenting it to a meeting of, ; for example, the Society of Music Theory. That paper I haven't read yet. I'll return to it when I do. ;Unfortunately, about the only professional journal which has tried ;to deal with questions of cognition from a musical point of view on ;a regular basis seems to be INTERFACE; and my personal opinion is ;that the editorial quality of this journal is very sloppy. Here the problem is inverted -- I personally would like to see the distance between the research and musical communities (and those factions enthusiastic about the accomplishments of the other side) close in on each other: music cognition is generally a sloppy field, not much less sloppy than composing NN resarchers. Unfortunately the misunderstandings on either side appear to run pretty deep.