Try to consider music from the viewpoint of a linguist. Western music is a both written and 'spoken'. The written language, as with most languages is a reasonable approximation to the spoken form. The written language is not universally applicable to all music as listening to any middle eastern music illustrates. For one thing, it has no standard provision for microtonal embellishments that interpolate the semitone. Although many complications can be written, microtones are not. As the written form of western music crosses human language barriers, written western music has a functional similarity to Chinese ideograms. The written form is practically a language unto itself. Its possibilities and rules are also its constraints. To write music is to limit what can be expressed.

To simplify then, consider the western written form of music as compared to the written form of a language. It is an alphabetic language based on the 12 well tempered pitches of an octave. In practice, the notation comfortably accommodates an extended alphabet of say 8 octaves. The number of octaves is in any practical case finite, since the range of human hearing is finite; it wouldn't make terribly much sense to write music that can't be heard.

The first major complication and difference between written music and written human language appears when one considers the formation of words. In all human languages, whether their written form is alphabetic, syllabic, ideographic, or any mixture of the above, the written form is linear. Perhaps, in a sequence, and therefore 1-dimensional is a more descriptive terminology. Meaningful groupings within alphabetic sequences are either understood by readers of the language, or there are built in "stop codons" as in DNA basepair sequences, to distinguish the end of a structure. Alphabetic languages generally use space to separate words; comma to separate phrases; and period, question mark and exclamation mark to separate sentences. The point that levels of structural units exist in human languages has been made, so I won't belabor it by going on about the other punctuation marks.

The written form of music is immediately seen, however, to be 2-dimensional, so it is over a 2-dimensional space that the extended alphabet is deployed. In this context, the simplicity of any form of stop codon is thwarted; and one asks the question: what are the words of music? The notion of word subdivision in most languages is a natural one; the idea is not so simple in many of the Uto-Aztecan languages. The grammatical structures that we take for granted in Indo-European languages are either completely absent in other languages or so vastly complicated as not to be recognizable. The structure of either a "noun" or "verb" in the Cherokee language can be complex enough to embrace an entire detailed sentence with merely a few syllables. Nouns can contain verbs and verbs can contain nouns.

The caution is therefore given that Indo-European grammatical structures are not necessarily appropriate containers for the "parts of speech" in a musical language. One should look, rather, for the units of meaning, in the language itself.

Let's try anyhow, just how well we can do, and perhaps more informatively where we fail.

Human Languages have both syntax and semantics. Syntax is the answer to the questions: how do the words go together, and what are the morphological rules for words? Semantics is the answer to the question: what does an expression of the language mean, or how is meaning assigned? In the 'or' part of the previous question, syntax often creeps into the answer. What does music mean? Upon being accosted by a member of the ladies auxilliary, with this question, Stravinsky is reported to have thought for a minute and said, Its meaning is its dramatic unfolding of its own architectural structure and internal internal relationships, or something very similar. I don't remember the reference. The important point being made is that music is essentially autoreferential, programme music notwithstanding. As one of my composition teachers, Raoul Pleskow expressed it, "It doesn't all have to sound like angels singing, but it *does* have to make sense." That sense is in it's internal structuring of autoreference; music speaks only of itself. Music is a language of forms, not unlike pure mathematics; human language is a language of symbols of things exterior to it, and only grudgingly gives reference to itself. It is about symbols, metaphors of symbols and symbols of metaphors. In this sense, it has no fundamental intrinsic meaning; it babbles incoherently. Human language is not built for autoreference.

It is curious that exactly one computer language LISP is, in fact, so structured that its data types contain the language itself so that any program written in LISP is also a data type of LISP. LISP is still the premiere artifical intelligence language, and music writing systems like "opus5" , Common Lisp Music [LINK] have been written in it.

Music can, of course, acquire extramusical symbolic meaning, but that is not its essence. Without its structural selfmeaning, it is nothing but "unorganized sound". It is precisely the relationships between it elements that make music what it is. Not to hear or listen to these relations is not to hear music.

Human Languages have both syntax and semantics. There is also the matter of 'style'. In artistic writing or speaking of a human language the personal choices of the creator manifests itself in a personal voice or style; so too in the writing of music.

What is the status tonality?
A fairly wide spread adjunct of Chomsky's deep grammar. It is the most primitive and deepest form of architectural cohesiveness and appears in any music that has the parameter of pitch.

What is the status harmony?
Harmony as determined by what period of musical history?
Harmony, in its most general sense has not to do with what sounds pretty and what doesn't (which is pretty much ephemeral and culture dependent), but to do with conventions concerning how pitches are combined both vertically and horizontally. There are concepts and concerns of harmony within atonal music, as there are within tonal music.

"The superannuated system of classical tonality, which has served as the basis for musical constructions of compelling interest, has had the authority of law among musicians for only a short period of time - a period much shorter than is usually imagined, extending only from the middle of the seventeenth century to the middle of the nineteenth. From the moment when chords no longer serve to fulfill merely the functions assigned to them by the interplay of tone but, instead, throw off all constraint to become new entities free of all ties - from that moment on one may say that the process is completed: the diatonic system has lived out its life cycle."

                               -- Igor Stravinsky  [Stravinsky 1942]

Music is then not a thing universal, but a collection of musical languages that might be generally categorized by syntactical structure as:

             Tonal
                  Horizontal
                  Vertical
                  Polyphonic

             Atonal
                  Horizontal
                  Vertical
                  Polyphonic

If there is a deep grammar of all music, then it must be founded in the physics of sound, and the physiology of the ear as well as the psychoneural implications of sound. Although the first two subjects have already been treated thoroughly in [Helmholtz 1877], and yet more research has also been done, I'm unaware of any serious work having been done in the psychoneural area. That this area of study is nonvacuous, is suggested by the use of droning, drumming and mantras as aids in producing altered states of consciousness.
What does a Bach fugue do?

Horizontal precedes Vertical in historical development. The Phrase, is the fundamental unit of meaning. Yet, Webern first created a musical language with such a low density of material and such a high density of relationships, that the pitch itself becomes the fundamental unit of meaning; the phrase is abstracted to pitch.

If music is an autoreferential language, how does it achieve its own reference? Parts refer to other parts by transformational relations, including the identity transformation yielding a repetition.

Events and activities - time slicing gives a sequence of complex events which might miss important cross relations. It is better to hear what is being expressed, i.e., all that is going on both vertically and horizontally. To "hear" music is not something that is innate, and not something that requires no effort. Nor can it be done without learning and practice, yet that *is* what is required.

To hear and understand music as it expresses itself, is at least as difficult as for one whose mother tongue is English to learn to hear, read and understand Tamil. Speaking it, is another matter.

   Suggestion of cognate "parts of speech",
   and grammatical structures:
   -----------------------------------------


   Agreement    - with respect to a tonal center
                  invariant of a transformation

   Gender       - Modality

   Number       - Number of repetitions in a sequence

   Cases        - 
	          Word order (no inflective *-fixes)
                  [Prefixes, infixes and postfixes exist but as
                   expanding or agglutinative transformations,
                   not as case inflections.]
                  Genitive (relationships by transformations)
                  Dative - (object towards which an activity moves)

   Tense        -
                  Past - back reference
                  Present - the radical, novel event
                  Future - anticipation or foreshadowing.

   Negation     - NONE

   Articles     - NONE

   Prepositions - NONE

   Conjunctions -
	          (of phrases) horizontal
                  by transformational similitude
                  by overlap
                  by polyphony vertical

   Verb          - Horizontal aspect of activity

   Noun          - Vertical aspect of activity

Music can also be analyzed as architecture where the parts have function and form only with regard to the whole. An architecture of integrity is of a whole when it procedes from unifying principles, one of which is the collection of its substantive materials. The materials are seen to generate the whole: the form of a chaconne has as unifying principles, its customary 3/4 signature, a tonality, and a recurring harmonic sequence upon which a sequence of variations is constructed constituting the chaconne.

A language of architecture shares with music, a decided multidimensionality of expression, including cognates of color, texture, smoothness and fracturedness of surface. As architecture is geometrical, so also then is music.

Everything, so far, in this essay suggests that the semantics of music as a language, like pure mathematics is equivalent to its syntax; structure is all. Yet, somehow we all know that this exclusive view that cannot be supported by the existing literature - or experience. A syntactical structure must be present for the forms to make sense, that is, to hold together. But the syntax does not determine the structure any more than the grammar of English automatically generated Hamlet, or King Lear. As one writes either prose or poetry and thinks in human language, semantics by metaphor is inescapable. It could be argued that some rather large percentage of our alleged knowledge, expressed in language is essentially metaphorical.

By the term metaphor, I do not mean the narrow technical distinction made in poetic analysis from simile, but mean to include the notion of simile, which is to say, pretty much as its Greek origin translates. The preposition "met\'a", has different meanings depending on the case with which it is used. With the genitive, it translates as "with, in company with, among". With the accusative, it translates in place or time as "behind, after, next". In agglutination it signifies a sharing "with" or "among", or regarding time or a quest, "after", or "change". The "phor" root we translate as "symbol" which is itself a metaphor since I see the root as being derived from "phoreus" meaning "bearer" or "carrier". At any rate, the linkage implied by a metaphor is a sharing so that one can act as a symbol for the other.

Music also has its metaphors and they are, loosely speaking, of two kinds: intrinsic and extrinsic.

Intrinsic metaphors link things solely within a musical context, they can be divided into two classes: interior and exterior. Interior intrinsic metaphors, repetitions, variations and transformations are part of the musical structure: devices of structural composition. External intrinsic metaphors abound in composers quoting themselves and other composers in music: "Variations on a theme of X"; Bach's stylistic metaphor in his Italian Concerto; Bach quoting Vivaldi; Wagner quoting the "Dresden Amen" in Parsifal; Max Reger quoting Bach quoting "BACH" and quoting Telemann; etc.

Extrinsic metaphors link things within a musical context to things outside of a musical context, and this is where the trouble begins. There are some wonderful examples of how some composers play with this device. Händel in the Messiah on the words "and He shall shake", writes a vocal line that actually shakes. Both Richard Strauss and Gustav Mahler are fond of imitating birds in their orchestral pieces, specific birds mind you that have their own mythosymbolic meanings. Many composers like birds; perhaps their musical ears cannot take them for granted and they wish to share their own delight at listening. Beethoven simulates a pretty good thunder storm in his sixth symphony; and this from someone who hated the idea of "programme music". Is this really programme music? It tells no story, but does elicit imagery that is familiar, and that familiarity it crucial to the understanding of the metaphor. Real programme music, has to tell a story.

Can one actually write programme music? Well - maybe - try some of the tone poems of Liszt, or Moussorgsky's "Pictures at an Exhibition". Without knowing what is supposed to be going on, can you guess? My guess is not. Music as programme music only works when the listener already knows the story or picture, and even then the listener needs clues, cues and signposts. Only opera, where the libretto leads us by the nose, presuming the appropriate homework has been done, is the ultimate programme music, where there is a distinct metaphorical link from the music to the story and the characters.

Can one actually write pure programme music? Perhaps, and I tread softly here: the B minor Mass of Bach, the German Requiem of Brahms, the Missa Solemnis of Beethoven, the Mass of Bernstein, various Requiem Masses written by composers knowing full well that they could never serve a liturgical function. "The Mass" has become as much a musical form as a passacaglia, and it comes intrinsically endowed with structure as well as a truckload of emotional metaphors. The obvious difference between a mass and a passacaglia, is that the passacaglia is simply a musical form whose origins in dance have been culturally lost.

Now we've gotten to the nuclear energy of music: extrinsic metaphors where the metaphorical links are emotions. As nuclear energy can be used constructively or destructively, so can these emotional links. They are dangerous, a fact understood quite clearly by Plato in The Republic. They are dangerous, precisely because they can and have been used to manipulate and destroy. The field is wide enough here to fill a myriad of volumes uniting music, metaphysics, ethics, physiology, psychology and neurology, with about two dozen other areas of knowledge.

All of these volumes will deal with the simple fact that music is not just structured sound, but a syaesthesic phenomenon that acts on human nervous systems directly through all senses in addition to having its cognitive appeal to the analytical musical mind. Statistically and historically, the analytical and structural aspects of music: that which concerns all of my music pages, are utterly irrelevant to the actual societal function of music, which is probably why I've written them.

Some other related links

1. Meaning Generation in Music Listening
   by Halljerd Aksnes
2. The Brain Opera - MIT