NetNews Usenet Archive 1993 #3/NN_1993_3.iso/spool/comp/ai/5040

Internet Message Format | 1993-01-27 | 3.9 KB

Xref: sparky comp.ai:5040 rec.arts.books:26402 misc.writing:4292 rec.arts.int-fiction:1290
Path: sparky!uunet!gumby!yale!yale.edu!not-for-mail
From: [e-mail] (Dr. Disorder)
Newsgroups: comp.ai,rec.arts.books,misc.writing,rec.arts.int-fiction
Subject: Re: The language of music
Date: 27 Jan 1993 14:59:09 -0500
Organization: Yale University Science & Engineering UNIX(tm), New Haven, CT 06520-2158
Lines: 60
Message-ID: <[e-mail]>
References: <[e-mail]> <jakobs.728043779@utrurt> <[e-mail]>
NNTP-Posting-Host: minerva.cis.yale.edu

In article <[e-mail]> [e-mail] writes:
>In article <jakobs.728043779@utrurt>, [e-mail] (Oliver Jakobs) writes:
>> In <[e-mail]> [e-mail] (Eric Behrens) writes:
>>
>>>But music has only a dozen words in its vocabulary (notes of the scale) and
>>>the English language has tens of thousands.  I cannot believe that computer
>>>science is capable of replicating the intricate constructions of language.
>>
>> I'd say the analogy is not of the right order. I've heard that scales are
>> more like phonemes or letters, and that there are groups or patterns of
>> scales which could be regarded as ``words of music''. There has been done
>> some work trying to apply linguistic knowledge to musical ``texts''. I don't
>> remember the scientist's name, but I could look up some references if someone
>> was interested.
>>
>> Oliver
>>
>>
>> --
>> Oliver Jakobs, Dept. of Computational Linguistics, Trier University, Germany
>> |                Internet: jakobs@[e-mail]                       |
>> +--------------------------------------------------------------------------+
>
>    Indeed, I would consider notes to be much more synonymous to the
>letters of the musical vocabulary.  Patterns of such notes, then, could be
>the words of the musical language.  Finally, phrases (groups of music, usually
>in 4 or 8 bars) are the musical sentances.

[Stuff on Mozart's style deleted for space]

>    Although I think the definition for Mozart's style may be a bit more
>straightforward than those for the English language (I believe it would be
>slightly easier to know what 'makes sense' in the musical vocabulary), it is
>no small feat to copy the style of a composer.  Inductively, I don't know how
>hard it would be to copy the style of an author (not beyond the realm of
>possibility, though).

Perhaps this is the wrong place to take up this thread, but as a composer,
I just can't let this go by. IMHO, the second poster made a much more plausible
analogy (i.e, notes are closer to letters rather than words - although, in
certain styles, a single note can take on many different levels of meaning,
from the letter to the paragraphand beyond). However, I strongly disagree with
the statement that it would be easier to know what "makes sense" in musical
vocabulary. As with language, the basic rules for making combinations out
of primitive expressions can often contradict each other, with some kind of
difficult-to-precisely-characterize meta-system required to resolve the
lower-level conflict. It is my intuitive belief (based on a life-long deep
immersion in music, but with an admittedly paltry understanding of
natural-langauge modeling) that the rules for creating "proper" expressions
in the style of a specific composer (especially one as rich as Mozart) are
even more resistant to explicit definition than the rules for what
"makes sense" in the English language. I think that there are even more
decisions at every level of the musical process that involve aesthetic
judgements (of a variety that would be imposssible to describe unambigously).
As easy to trip up as it is, even ELIZA does a better job of
acheiving its goal than any computer-generated piece in the style of
Mozart that I've ever heard.
--
Practice random acts of randomness.........
        [e-mail]