1. What must a (musical) assertion entail?

Can music express a proposition? In Musical Meaning and Expression Stephen Davies gives a clear set of hurdles a musical assertion (i.e., one which conveys a simple proposition) must clear:

Assertions take truth values. If musical works, themes, or phrases are not appropriately to be assessed for truth, then music performs no assertoric function: either it does not pick out a subject, or it fails to characterize the subjects it picks out in a way analogous to predication (p. 12).

As he makes clear, Davies does not think music can do these things, or rather, it can't do both at the same time. Indeed, for him, this is the main reason why comparisons between music and language are inapt, since he prefaces the above remarks by saying that "assertion is a foundational function of language, a function that must be performed by anything worth the name" (p. 12). But what exactly is being asserted in the following linguistic cases(?):

"She is hopeful."

"She is here."

"Columbus."

"Over There."

Standing alone, none of these take truth values, since (in the case of the first two) we do not to whom "she" refers (not to mention where "here is"). And the last two are fragments, and thus would seem to fail utterly. But of course these are the kinds of utterances we often use to convey propositions. Context not only defines the token-reflexive aspects of an utterance; it also supplies missing parts of a proposition--since "Columbus" is the answer to the question "What is the capital of Ohio" and "Over There" the answer to the question "Name a famous song written by George M. Cohan." Thus while Davies is of course correct in listing the necessary requirements for making an assertion, not all linguistic utterances are complete assertions in and of themselves. Many require contextual help. This leads me to consider the following question and its corollary: If a semantically-rich context can make up for a semantically-impoverished utterance in the case of a natural language, then how might semantically-rich musical context make up for the semantic shortcomings of music alone(?). And if there are linguistic assertions which are context-dependent, perhaps they, rather than complete sentences, are better models for a consideration of how music may express a proposition or perform other quasi-linguistic acts.

To explore this question, I will present an analysis of the "Maria" motive from Leonard Bernstein's West Side Story. I will address each of Davies' requirements: how the motive is able to pick out a subject, how it characterizes that subject, and how we are able to assess the truth value of that characterization. At the end, I hope to have made a plausible case that in the narrow context of the musical (or its film version) the "Maria" motive conveys the proposition "Maria is hopeful."

From the outset, let me make it clear that I am not claiming that music can express propositions just like language can. Clearly, it cannot, and indeed, I am largely in agreement with those who have probed the semantic limits of what Kivy has aptly termed "music alone." My larger point, though, is that music alone is not the whole story, and so a more complete account of musical semantics must examine musical structures as they occur in other, semantically richer (and messier) contexts.

2. "Maria" refers to Maria.

It has been a long standing practice in opera, drama, and (more recently) film to have particular thematic material associated with particular persons, objects, or even ideas expressed by the text and/or dramatic action. Since Wagner these special themes have been known as "leitmotives." To establish this association the person, place or object a leitmotive designates must normally be onstage or otherwise in view when the leitmotive is first sounded . In West Side Story there are a number of leitmotives, but of course the most famous is this the "Maria" motive. The audience is introduced to this leitmotive in a number of ways. Following both musical and filmic conventions, the overture begins with introductory material, then presents the first "big tune" emblematic of the musical--"Tonight, Tonight." This is then followed by the theme for the female lead--"Maria." Following this convention, the audience knows the "musical name" for the prima donna (though at this point we don't necessarily know her spoken name). We next hear Maria's leitmotive in the dance scene, when Maria and Tony first meet. We hear the leitmotive again at precise moment Tony first looks at Maria. The scene immediately shifts to blackout, save for Tony and Maria (they "see only each other"), and in that transported state, they dance to Maria's theme presented as an airy cha-cha. Tellingly, Tony knows Maria's leitmotive (as evidenced by his dancing) before he knows her name; after their dance and brief conversation he overhears Chino say "come Maria," and she is led away from the gymnasium. Maria's leitmotive is of course the basis of Tony's following song.

The way in which we first learn of the association between a leitmotive and its referent, and witness its subsequent usage is analogous to how we understand proper names in linguistic contexts. In both the linguistic and musical cases there are particular conventions through which the link between a proper name or leitmotive and its referent is established. Fixing the referent of a proper name forges the first link in a causal chain of usage which subsequently extends the connection between name and referent.

How do proper names refer? Frege and Russell have both posited that names must have a sense which determines reference: "[a name] designates its bearer . . . because the definite description which expresses its sense denotes that bearer" (see Devitt and Sterelny, p. 39). Under the Frege/Russell model names are abbreviated descriptions. And this seems like a very good account of musical leitmotives--given that they cannot help be musical expressive (at least to some degree), then they function as an expressive description of their referent. So the "Maria" leitmotive does not refer to Maria, but simply to "the one with THOSE expressive characteristics" (more on what those expressive characteristics are in a moment).

Here is an example which counters this argument. Imagine a western, with the lead characters Jones and Smith. Jones is our optimistic hero, while Smith is his perpetually dour sidekick. Here are two plausible leitmotives, LM1 and LM2, composed by the author, which appropriately underscore the "characters" of Jones and Smith, respectively [click here for examples]. Now, at some point in our story, perhaps in order to raise money to save the farm, Jones goes into the saloon to gamble his fortune in a poker game. Time passes, and then we see him exit the saloon and walk up to Smith. We hear this LM1a. This is an "unhappy" version of LM1--its "unhappy" character isachieved by a shift to minor harmony, slower in tempo, and muted dynamics, but note that its essential pitch and rhythmic contour is preserved. To whom does LM1a refer? On a descriptive account, we might have a case of ambiguity, since descriptively it could now refer to Smith or Jones. But if one grants that LM1a unambiguously refers to Jones, then the referential ability of a leitmotive cannot be based on its descriptive content.

If this sounds like a Kripkean account of leitmotives, it's because it is. Kripke argued that proper names need not be yoked to any specific property or set of properties that their referents may have, and in this sense proper names are semantically empty or meaningless. In fact, leitmotivs are an excellent example of Kripkean rigidity, and so if one accepts Kripke's account of proper-name reference (given in given Naming and Necessity; see also Donellan, 1972), then not only is leitmotivic reference separate from its descriptive (read "expressive") content, but this separation may be exploited. And as is doubtless familiar, the expressive character of a leitmotive often changes radically over the course of a film or musical-dramatic work. Indeed, it must, for in a drama it is the emotional state of the characters that is the most mutable, and its that state that the leitmotive so often "describes." Leitmotives are thus musical proper names.

One might also object to a Kripkean account of leitmotives-as-proper-names under the claim that names are arbitrary, while leitmotivs are not. That is, there has to be some basis for using a particular leitmotive, with particular expressive properties with a given character--you presumably wouldn't use Maria's leitmotive for Darth Vader, or vice versa. While this is true, the premise that proper names in language are wholly arbitrary does not hold up under scrutiny. Names are not arbitrary, unless, perhaps, you are naming a car (e.g., "Elantra," "Stylus," "I30"). Proper names for persons have conventionalized associations with gender, ethnicity, and nationality (which is why one can be confused or taken aback upon learning someone's name, as in the classic "funny, you don't look Jewish" joke). Moreover, proper names, like all words, when spoken have expressive properties created by their particular phonological features (for a discussion of this "phonological poetics" see Reuven Tsur on What Makes Sound Patterns Expressive).

3. "Maria" is an expressive characterization of Maria.

How does the "Maria" motive characterize Maria? The leitmotive consists of three notes: Eb-A-Bb, with the Eb as a pick-up to the short, dissonant A on the downbeat followed by its upward resolution to Bb. The opening Eb-A is a (very) dissonant tritone, and the placement of the A on the downbeat enhances the sense of strain. The motive then rights itself with the Bb. The overall quality is one of stretching, perhaps almost missing, but then ultimately grasping a desired object. In most instances this figure is repeated several times. Now, consider Levinson's comments regarding the expressive qualities of another musical passage:

We hear these phrases as a reaching for something, for something higher. These successive leaps, the second an amplification of the first, go some way to account for the melody's quality of aspiration . . . The [following] figure, seems in context also suggestive of the poised bearing, the restrained carriage of one who calmly hopes in the face of tribulation (Music, Art, and Metaphysics, pp. 367-68).

This is Levinson's analysis of mm. 57-66 of Mendelssohn's Hebrides Overture (Op. 26), which he claims are expressive of hope. I would claim that Maria's motive embodies similar qualities of aspiration, though perhaps not as poised or restrained. "Maria" (the leitmotive) seems more an expression of outright yearning.

What transforms this yearning and stretching gesture to an expression of hope is our knowledge of the story in which it occurs, as well as knowing that it is associated with a particular character in that story. When we first hear the gesture in the overture, we are impressed with its sense of stretching and reaching, but we don't necessarily know that it is hopeful. But when we hear it again, we have since met Maria, and we know that she has aspirations and plans. Knowing this, we can hear the Eb-A-Bb as a gesture of aspiration, and of hope.

This account of expressive qualities of the "Maria" motive follows Kivy's contour theory of expression, and yields a characterization of Maria which fits the story (see Sound Sentiment, esp. 71-83). An even richer expressive account is possible when expressive conventions are taken into account. For the "Maria" motive is made with harmonically plangent materials. The opening interval is a tritone, the paradigmatic dissonance of Western tonal music. Moreover, this skip has the added problem of de-stabilizing the sense of the opening tonic. One of the reasons the tritone is the paradigmatic dissonance is that while every key contains some major thirds, some minor thirds, some seconds, and so forth between each of the seven notes of the scale, there is only one tritone (between the fourth and seventh notes of the scale). Tritones therefore require careful handling--the two characteristic resolutions of a tritone are a collapse inward to a 3d, or expansion to a 6th. Here we have neither. Maria's tritone is a skip from the tonic Eb to a chromatically-altered note (in the key of Eb we should have Ab, not A-natural). The first note is the fundamental note of the key; the next lives outside of it. To be sure, the leitmotive is undergirded with Eb major harmony, which will reinforce our sense of Eb-as-tonic. But the shift to Bb (the second "Maria," after "I just met a girl named . . . ") which ought to be an increase in harmonic tension instead creates the opposite effect, one of harmonic relaxation. The melody continues to mix the notes Ab and A-natural, and it turns to Eb for its first cadence. Subsequent phrases move more firmly into Bb. But it is only on at the end of the verse--"I'll never stop saying 'Maria'"--at that Eb-as-tonic is finally secure. To create this masterful harmonic ambiguity Bernstein made skillful use of the conventions of tonal syntax, turning them on their head, as it were.

This unconventional harmonic gesture is concomitant with an "improper" melodic gesture. One of the oldest rules of thumb in writing melodies is that leaps should be followed by stepwise motion in the opposite direction. Here, however, while we get the step, we do not get a change of direction. Thus the third note should have been a half-step descent, along the lines of the tritone collapse (which would have then established E, rather than Eb, as the key). So there is something more going on here than simply a reaching upward. We also have a gesture which attempts to break out of the conventional melodic and harmonic syntax of the tritone. Not only is this a gesture of "aspiration," but it is a gesture which strains many of the conventions of melodic and harmonic syntax. In other words, it is not just hope, but hope of a markedly unconventional kind. In the dramatic context, the music says something along the lines of "Maria hopes for something one would not normally hope for." She hopes for a better life than what is typical of a poor Puerto-Rican immigrant. Perhaps, even, in that moment in the dancehall, that she hopes for Tony.

Davies also notes that there is a Goodmanian question as to whether one should regard expressive qualities of music as instances of a particular emotion (a swatch of hope) versus denotations of particular emotions (expressions of hope). Davies says:

If composers deliberately employ expressive features in their works, it might be that they aim to make their works expressive, thereby intending to refer more widely to the emotions exemplified in those works. By making the work sad, the composer might intend to denote sadness (and not merely to present an instance of it) (Musical Meaning and Expression, p. 10).

I think this is exactly right. Furthermore, the context in which expressive gestures are presented allows the listener to distinguish between exemplifications versus instances of various emotions. In the case of a leitmotive, the context of the film score or opera tells us that the usage of these properties is denotative with respect to their embodiment in a leitmotive. This is because we understand the orchestra in an opera or the underscoring of a film as kind of musical narration which accompanies the dramatic action. The ways in which leitmotives (as well as other parts of a soundtrack) can fulfill such "cueing functions" are well known to film composers and film scholars. As Gorbman has noted:

Narrative film music "anchors" the image in meaning. It expresses moods and connotations, which, in conjunction with the images and other sounds, aid in interpreting narrative events and indicating moral/class/ethnic values of characters. Further, attributes of melody, instrumentation, and rhythm imitate or illustrate physical events on the screen (Unheard Melodies, p. 84).

4. "Maria is Hopeful" is true.

Though the "Maria" leitmotive is best known in the context of Tony's song, it remains Maria's throughout the musical, and so its expressive properties must therefore be understood as Maria's. If this is so, then I think one is warranted in making the following claims about its use:

1. The leitmotive "Maria" is "uttered" by the soundtrack

2. The "Maria" leitmotive refers to Maria.

3. At the same time, the "Maria" leitmotive has expressive properties which denote hope

4. THEREFORE, we are warranted in attributing those expressive properties to its referent. In words, "Maria is hopeful."

Is this assertion true? Yes (at least until the end of the show), but in practical terms, this is a trivial question, since leitmotivic characterizations are always true. This is because the "speaker" of these leitmotives is the composer, the omniscient musical narrator of the sound track to our movie or opera. When we listen to a storyteller we accept her account of events and description of characters are "true." Now one can imagine instances where the soundtrack might "lie" to us--advertise the presence of a character who is really absent, or incorrectly characterizing a character's emotional state. But this would be a radical failure of the normative discursive role of an omniscient narrator.

5. Conclusion

Ever since Hanslick, musical aestheticians have been confronted with all of the things that music cannot do: music cannot be representational; music cannot express emotions; music cannot express propositions; and so forth. And ever since Hanslick, studies of musical semantics have concentrated upon meaning in the context of abstract instrumental music, music apart from any sung text, literary program, or dramatic setting--in Kivy's apt term, "Music Alone." In his book of that title, as well as in his path breaking Sound Sentiment, Kivy has led the charge to reclaim a few things that music can do, and so we now grant that music alone can express at least some emotions (which ones depends on whom you ask). Yet Kivy, like Davies, remains skeptical regarding music's capacity to achieve anything like linguistic meaning:

Although the ear does, like the eye, have a strong tendency to interpret, its tendency is not to interpret sounds as representational or as natural phenomena but to interpret them as meaningful in the full linguistic sense. And since such a tendency is easily defeated by the stringent semantic requirements on successful linguistic interpretation (Music Alone, p. 9)

And

If the tendency of the ear to interpret musical sound as meaningful human utterance fails, as it must [ital.. mine], what linguistically is left? Take the meaning away from the utterance, and one thing you may still have (though not necessarily) is the utterance's emotional cast. I can, for example, sometimes tell by tone of voice that someone has said something to me angrily or sadly, even though what he or she said may have been lost on the wind (ibid.).

I think Kivy overstates his own ignorance here. Linguistically, in the fullest sense of the term, one has a lot left. If I am in mid-conversation, and hear an angry retort that is lost on the wind, in most cases I still have a pretty good idea what was said. Like Searle's Italian soldiers, my understanding of my interlocutor's illocutionary intent may not require any knowledge of her locution at all (see Speech Acts, p. 44). But of course I have a lot of knowledge: the prior conversational context provides one discursive thread; knowledge of the general subject of conversation another; knowledge of one's interlocutor (and her interests relevant to the subject at hand) yet another.

We are rarely confronted with "language alone." Most of utterances occur in semantically rich contexts--with defined social roles, conversational goals, and shared knowledge and presumptions. Consider too our understanding of the "meaning" of certain kinds of human behavior (i.e., all those behaviors which ground a contour theory of musical expression): are these behaviors intelligible in the absence of a social/discursive context? Again, I would say for the most part they are not; we don't have "expressive behavior alone" either.

And we are rarely confronted with music alone. For most listeners, "music" means "song" (indeed, as a college music teacher, one of the first tasks is to get my students to stop referring to every piece as a "song"). Music with a sung text or which accompanies a film is the normative condition for musical listening; music alone has always had a relatively small "market share" of the total audience for music. Often these days one hears (at least in academic and arts funding circles) that shrinking audiences for orchestral and chamber music are a sign of cultural decline, etc., etc. While this may be partially true, there may be another reason: listeners have always gravitated toward music in the context of songs, cantatas, motets, operas, pantomimes, tableaux vivants, and films, not because they don't understand symphonies and string quartets, but because it is those contexts that the music itself can be most meaningful.

Davies, Stephen. Musical Meaning and Expression. Cornell University Press, 1994.

Devitt, Michael, and Kim Sterelny. Language and Reality. MIT Press, 1987.

Donnellan, Keith S. "Proper Names and Identifying Descriptions." In Semantics of Natural Language, edited by Eds. Harman and Davidson, 356-79. Dordrecht-Holland: D. Reidel, 1972.

Gorbman, Claudia. Unheard Melodies: Narrative Film Music. Bloomington: Indiana University Press, 1987.

Kivy, Peter. Music Alone. Cornell University Press, 1990.

Kivy, Peter. Sound Sentiment. Philadelphia: Temple University Press, 1989.

Kripke, Saul. Naming and Necessity. 2nd ed. Harvard University Press, 1980.

Levinson, Jerrold. Music, Art, and Metaphysics. Cornell University Press, 1990.

Putnam, Hilary. "The Meaning of "Meaning"." In Minnesota Studies in the Philosophy of Science, Volume VII: Language, Mind, and Knowledge, edited by Keith Gunderson, 131-93. University of Minnesota Press, 1975.

Scruton, Roger. The Aesthetics of Music. Oxford University Press, 1997.

Searle, John R. Speech Acts. Cambridge University Press, 1969.

Tsur, Reuven. What Makes Sound Patterns Expressive. Duke University Press, 1992.