Resisting Representation ­ Replaying My Voice Mail [1]




David Lidov



I am a composer who was slightly diverted from composing for a number of years by a theoretical project. For most of the years when I studied problems in semiotics and, eventually, wrote and published my Elements of Semiotics (1999), I thought of my semiotics and my own composing as unconnected with each other except by their competition for time. Elements, after all, is not even a book about musical semiotics. It concerns general semiotics.[2]  Voice Mail, a suite, twenty‑five to thirty minutes long, of thirteen short pieces for solo piano is among the first compositions I wrote after Elements was published, and it is the first of my compositions on which I have brought semiotic contemplation to bear. If the music offers witness to any semiotic doctrine, it might be simply this: A musician needs no consistent or constant stance toward representation. A composer might write program music on Monday, "absolute" music on Tuesday, expressive music on Wednesday and abstract music on Thursday, and there is no reason why the same variability should not occur within one composition. It seems to me now that semiotic inconsistency, rather than merely a symptom of aesthetic theory, is and must be a symptom of the fine arts themselves. The karma of the fine arts is to continually renegotiate the fascination of perception itself and the subordination of perception to communication or practical use.


One of the objects of the present essay is to demonstrate this symptomology in a particular case. I think of Voice Mail as pretend program music, as asserting textural, topical and stylistic affiliation with suites of musical pieces focused on images and characters (less so stories). But Voice Mail unfolds a design which, in the end, resists its programmatic gambits. If it means something, it means something that the thirteen titles don't add to. Like much other music ­ sometimes I even think, like all music that is meant as foreground music, not background music ­ Voice Mail moves from image to abstraction. I tell my story as an instance of this general case. Fundamentally, music is always ready to signify: We identify somatically with its apparent movement and vocally with its apparent utterance. Resisting signification and abstracting from signification is not the same as an incapacity for signification.


The two objects of representation I will be most concerned to discuss in this essay are representations of tonality and representations of topics. I use the term "topic" technically in the sense that emerges in the musicological work of Ratner, Hatten, Monelle and others, which I will review sketchily below. I will describe syntactic relations and semantic elements of the composition in leisurely detail, especially the syntax because it turns out to illustrate the same final point, reserving my argument about their integration for a final section which also discusses performance.


Music hosts a three‑way competition among the material attractions of sound, the fascinations of structural elaboration, and engagement in reference, or, we might say, among sensation, syntax, and semantics. If we are careful, we must keep it in mind that the distinction between syntax and semantics can be problematic. Chomsky demonstrated in Aspects of Syntax that choosing to describe a linguistic regularity as syntactic or semantic may be arbitrary. In computer science the distinction can be formulated relative to a level of interpretation, in Elements, I treat it as relative to the delineation of a text and/or the presumption of an abstract grammar. Here, when I use those terms, a "common sense" or rough‑and‑ready understanding should be good enough.


Usually we think of tonality as a syntactic aspect of music, and the representations it requires are syntactic representations (for example, when a chord of the sixth degree or an inversion represents a tonic function). Voice Mail extensively represents tonality, but tonality does not provide very much of its syntax. Tonality is represented as a semantic content. The difference is not pure. Syntax may be roughly divided into two parts, patterns local to a work and grammars that govern a language or style. In tonal music, taking "tonal" in a strong sense, tonality functions as a grammar ­ a rule system. Tonality is not a grammar in Voice Mail, but tonal elements, oblivious of any tonal rules, do participate in patterning.



Thematic Forms


In the competition for attention, structure weighs in with a heavy display of thematic forms. They were one of my chief delights in composing Voice Mail, and I hope they add to the entertainment of listening to it. With the exception of numbers I, XII and XIII, each piece broadcasts a distinct pattern, and the exceptions, as exceptions, contribute to the same sport:



            I           Freely unfolding melody with (quasi Alberti) accompaniment.

II           Developing variation on a motive with climax and then a suggestion of


III         AA' with a quick turnaround in between.

IV         Theme and free transformations (melodic contours preserved).

V          A rigid two‑part invention.

VI         Theme with four strict variations.

VII        Song‑‑Development‑‑Song (short‑long‑short). The development progresses

      in rhythm, not in harmony.

VIII      AA'A" or three verses. Unlike III, where the model and its repetition differ

only in a few selected details, the repetitions here are systematic 'modal

            transpositions,' as explained below.

            IX         ABCB'. B' repeats the rhythm of B on a static chord

 X         Fugue by augmentation (ABA')

            XI         Imitation chorale prelude.

            XII        A parody of a well‑known piece, loyal measure for measure.

XIII       Potpourri of motives from previous numbers in free fantasia style.


These forms are aurally straightforward and blatantly different. Strangely, I can't bring to mind much repertoire in which the contrast of forms (not contrast of genre or contrast of style) is the main engine of a little suite. There is a big element of this in Agon, as part of much else. Berg amused himself and musicologists with a play of neo‑classical forms in opera, but surely it is secondary to the drama. Here it is in the center ring.



Harmonic Design


Charles Wuourinen refers admiringly to evocation of tonal idioms in Stravinsky's late style as "punning".[3] In Voice Mail, tonal idioms are too prevalent and sometimes too systematic to call puns, but the elements of tonality induced provide more color and imagery than logic. Features which evoke or refer to tonality include sonority types (triads, sevenths, etc.), voice leading relationships, imitations of harmonic progressions and, although with a very haphazard distribution, some dissonance resolutions.


Tonal grammar has an aural authority which is not annulled by simple fiat. Voice Mail is able to absorb extensive tonal reference only because its vocabulary is severely limited. Economy of vocabulary, if sufficient, can match tonality in its force of appeal to an intuitive sense of order. This is a simple principle and not a new one. A development passage limited in its motivic vocabulary to one cell and its variants can leap any harmonic distance in a single bound, and we follow happily. In Debussy's style repetition often serves to delimit a vocabulary so clearly that it becomes independent of its conventional harmonic interpretation. In both of those cases, vocabulary limitation enables music to escape tonal constraints or tonal interpretations. In Voice Mail, this common principle serves in the absence of any a priori tonal framework. The limitation is harmonic but not tonal. Voice Mail is made out of thirteen unordered pitch class sets that are (with one exception) never transposed or inverted or otherwise altered. The PC set organization is perceptually salient‑‑unusually so, I believe. Much of the time these sets, twelve pentads and one trichord, are either heard one at a time or kept separate from each other by rhythm and/or register. More often than not they occur complete and hardly ever with two or more PC's missing. The identity and perseverance of the PC sets is enhanced by motivic relations and registral limitations, but motives and registral assignments come and go. What stays and stays simple are the pitch class sets themselves. Tonal elements in Voice Mail may offer the comfort of familiarity, but those elements only rarely add up as complete structures. The PC sets are ultimately more prominent and establish a sense of order which becomes firmer and firmer as the music unfolds.


I feel half a poseur adapting this jargon, "PC sets". Given the simplicity of the method and the results, would it not be plainer and truer to say "modes"? Perhaps it is worth looking at the connotations of these terms. "Mode" should imply that the pitch collections have a tonic. Possibly the case here and there, but not consistently. On the other side, to speak of "PC sets" suggests much by association that is not legally implied: a stylistic tendency that makes living a long time in a small, recurring collection of pitch classes more the exception than the rule, a stylistic tendency that maintains a wary eye on octave duplication and registral compactness, a possible bad conscience if closure of the full chromatic aggregate is neglected  ­ in other words, a stylistic tendency that doesn't deploy "PC sets" the way we typically deploy scales and modes. But of course, these are only informal associations. Voice Mail  is not unique as a context where they do not apply.


The design which distributes the PC sets in Voice Mail (and which also contributes to their construction, but with considerable dilution) is a Fano Lattice or Fano Geometry. I employ this design, as others do various other sorts of pitch matrices, for the efficient, practical provisionment of a harmonic world which suites my aesthetic appetites ­ I make no Platonic claims about the essence of this particular structure. The sound world of Voice Mail is highly chromatic but not homogeneously so. Its harmonic world does not suggest symmetry. (Abstractly, the Fano lattice is all symmetries, but in my realization of this geometric structure, the symmetries are imperceptible.) Before getting to particulars of its realization in Voice Mail, I will describe the Fano lattice. To simplify the demonstration, I describe this lattice as a finite geometry and as if it were simply a design for constructing PC sets. My readers will understand that we could be talking about grouping pitches or durations or sandwich fillings or stock options, and, as I mentioned, in Voice Mail, the more important application is actually the distribution of PC sets, not their construction. (The world may not need to hear yet another design for PC sets, but I confess, having toyed with this thing occasionally for over twenty‑years, that I want to tell about it.)



The Fano Lattice


The Fano lattice can be defined as a geometry, a collection of points and lines. For the moment, we will think of the points as PC's and think of the lines as unordered sets. The design encompasses n2 + n + 1 points, that is 3, 7, 13, . . . points for n = 1, 2, 3, ... . The cases of 7 and 13 elements are all we need. A reasonable projection of the seven‑element geometry can be drawn in two dimensions, as in Figure 1, but the geometry is not really two‑dimensional. The lines appear to cross at several points on paper only because the picture flattens them onto its surface. One line would be in front, one behind. For this diagram, we must adopt the understanding that lines do not intersect except at the dots and also that straight line segments are equivalent to circles. Each line segment or circle corresponds to a set of three points (dots). Every dot lies on three lines.



  Figure 1: A 7-point (two dimensional) Fano Geometry ­ seven points grouped in seven sets.

                     In abstracto, this design is richly symmetrical, but symmetry is not a danger in its musical application to constructing PC sets. The design for seven has application to any seven‑note scale, but within twelve‑tone equal temperament, no seven‑PC scale is symmetrical. As I realize it with 12 pitch classes, the design for 13 comes out with a hole or a joker. No great danger of symmetry there, either. In Figure 2, I show three distributions of the diatonic collection that correspond with the seven‑element design. Apart from their diatonic transpositions, these and their three inversions are the only distinct designs of PC classes that realize this geometry within that scale.[4] The first of these is entirely trivial (in the case of a diatonic collection) and the second is at least partly trivial. The third may seem trivial to some readers, but I am fond of it.





 Figure 2: Collection of PC-sets in diatonic collection determined by Fano geometry.


The notes below are written as if we had 13 distinct PC's to work with ­ I don't; some people do! 


a)         Each PC of seven belongs to three sets (or with thirteen PCs, four sets.) 

b)         Each set out of seven includes three PCs (or with thirteen, four PC's.)   

c)         Any two sets have exactly one PC in common.   

d)    Every interval (not interval class) occurs just once.[5] 

e)    For the 7‑design, the complements of any two sets have two PCs in common (for 13,


f)          Every set has two PCs in common with each of the complements of the other sets (for

                        13, three)   

g)         For the 13‑design, the intersection of any two set complements is a subset of the union

     of the other two sets that share their one common PC. (Trivial with 7, a  compositional

resource with 13.)   

h)         A full lattice is uniquely determined when each element has  been inserted in one


i)          The homogeneous distribution of common tones determines that the union of any two

sets in the 7‑design is a five PC set, but the cardinality of a union of three sets is either

six or seven, depending on whether the three share one common tone or note. Parallel,

                        messier relations obtain for thirteen. 




My twelve‑note application of the thirteen‑note design, simply leaving a hole, provides nine mutually overlapping tetrads that also overlap four mutually disjoint trichords. (Associating the four three‑PC sets with an ordered row does not provide a simple key to ordering the thirteen sets. Proof withheld.) 


                   Figure 3 shows the pitch class set system of another composition Remembering Major Dreaming of Minor, for violin and piano. In this system, the PC's are arranged strictly to correspond to the Fano geometry with one "point" unassigned ­ hence the four trichords. 



Figure 3: PC set collection from a 13-point geometry (from Sonata for violin and piano). The four

instances of each PC are shown at the same pitch.


Figure 4 shows the distribution of PC sets in among the thirteen numbers of Voice Mail. One point in the geometry of thirteen corresponds to one set, not one PC. This distribution also corresponds strictly to a Fano geometry. There are four sets in each piece. Any two pieces have one set in common.


The design is compositionally and perceptually consequential. One consequence is that each piece has the potential to recall harmonic or melodic characteristics of any of the previous ones; the thirteenth number, though composed of four sets like the others, quotes motives from all the preceding pieces "at pitch". The recurrences of sets across the suite may or may not strike the attention of the listener ­ who knows what someone will notice! ­ but they are palpable enough to be strongly determinative of an overall harmonic character. There is a kind of resonance in memory that builds up from the long range repetitions of the never transposed sets and at least a sense of logic ­ if not closure ­ that emerges as less and less that is entirely new (harmonically) occurs. The harmony of the thirteen pieces surely sounds more rational than the harmony of any one or two of them; yet, the limitation of vocabulary in each single piece contributes to its own unity and identity, just as a rather uniform density of the harmonic vocabulary unifies pieces across the suite. In the composing out, recurrences of sets in different numbers are sometimes linked motivically. I don't think the fact that, by the end, every possible pair of two sets has occurred in just one piece is audibly salient, but this is a property of the syntax that ensures fresh harmonic combinations in each number. (As I order the pieces, the thirteenth set to appear enters in the seventh piece. But no matter how the pieces might be re‑ordered, the tenth piece would be the last one that might possibly introduce a new set.)


My fetishization of the Fano geometry responds to a number of personal tastes, for example, for small, unordered sets as "materials", for pitch relations that my memory can keep track of fairly easily. Non-transposition as a counterweight to chromaticism has interested me since I first noticed it Liszt's Tasso, where one of the principle themes is never transposed. I find the Fano design complex enough to hold my interest without transpositions or with few, and, when adopted with no or very limited transposition, the harmonic vocabulary it provides is quite manageable for memory. Above all, it is favorable to materials which defeat the symmetry of the equal-tempered twelve‑note scale in a chromatic context by recalling tonality.



Figure 4: Distribution of PC sets in Voice Mail.


Equal Temperament and the Representation of Tonality.


The sound of the twelve‑tone equal‑temperament heard for what it is and the sound of that same system when it is deployed to represent tonal relations are two different sounds. The sound of equal temperament displayed as such first enters music in passages of quick and casual modulation like complete, rapid cycles of fifths in Chopin (for example, in the E major Etude or the AbBallade.) It is a pretty sound. We hear it in all of Webern's serial music and in much music he influenced. It goes hand in hand with the principle of complementation. When Haydn offers us a slow movement in E major for a quartet in E flat, we are already poking our nose into that world, for the tonalities, placed side‑by‑side with no implication of movement between them, are truly supplementary, not conflicting. The opposite holds when Beethoven transposes a theme to the Neapolitan. Equal temperament takes some of the sting out of the opening of the Appasionata, but we can compensate because we know the acoustic tensions that tonality represents. When we are caught up in the tonal relations represented we somewhat discount the tiresome homogeneities of the chromatic scale, the rough consonances, compromised dissonances, and symmetries.


And make no mistake, those symmetries are both pretty and tiresome. Full equal tempered tuning for keyboard instruments took hold gradually in the nineteenth century. When flagrant modulation and then atonality became common place early in twentieth composition, some sharp‑eared people were already looking for a way out ­ Partch, Harrison, the quarter tone people and the overtone people. Others seek an exit (and I'm not convinced it works) in complexity. The way out that I like best was also the way in, via representation. In college, in theory class we used to sing:




Example 1: Schœnberg op. 37 ("This is really not D minor, it is a tone row.")


And really, it isn't D minor, but the surface invocation of D minor is vitally important in offsetting the tendency of the chromatic scale towards balance in symmetry.




Tonal References in Voice Mail


My recurrent engagements with the Fano design have been motivated in part by its facility in paraphrasing tonal effects, but Voice Mail, deploying pentads instead of tetrachords, greatly exaggerates the effect. How the pentads were individually constructed is almost peripheral. I did start with a Fano realization analogous to Figure 2. Ignoring the potential resource of complement intersections, which also happen to be pentads, 36 of them, I assigned a duplicate |A| to thirteenth PC position. (I use the bars to indicate the class, not a pitch.) That filled in three trichords. (The trichord which already included |A| didn't change.) I then arbitrarily added one more PC to each of the other twelve sets. It may seem silly to have announced all this theory and then to report a cooked collection, but the distribution of the sets is my excuse for the theory, not their construction. Yet, even the cooked sets retain the feature that any two have a common tone (now at least one common tone, not exactly one common tone.) And the cooked system still represents all PC intervals at least once, though not exactly once. There are eight |A|'s, not surprising if it sometimes sounds like a tonic, and five each of the other PC's. In Figure 4 the sets appear in the order of their first appearance. Some pieces introduce the one pitch (generally isolated) that is not included in any of their four sets. These exceptions take nothing from the audibility of the system and I doubt that much more extensive exceptions would need to. My brief for the system is not a defense of purity. Beyond a certain point, the purity of any musical procedure is only a labor saving device or a private amusement.


An anomaly of this particular system of sets is that one pentad type, the pentatonic scale, appears twice. Two of the sets, sets 10 and 11 in Figure 4 are pentatonic scales (B‑pentatonic and Abpentatonic.) It is also this one pentad that is once further transposed, to E‑pentatonic, in Number 12. 


Besides the two pentatonic pentads, three other sets lie wholly within diatonic scales. The fifth, suggesting A minor, the twelfth, suggesting Bbmajor and the sixth, suggesting G major do also, as does the first, the trichord. The second suggests E minor, and the eleventh can be accommodated in Bb minor. The five others are near misses, with at least one subset of four PC's compatible with a major or minor scale. This aspect of the system of sets is haphazard. Allusions to scales in the various sets rarely cooperate. The sets, individually, are deployed registrally and rhythmically in ways that make other harmonic allusions than those mentioned. Where sets or parts of sets overlap, still other harmonic associations are evoked. This said, what I would emphasize is that the tonally suggestive content of the sets is not repressed as it might be, for example, in a style that discouraged proximate octave duplication. The two numbers which come closest to observing a closed tonal syntax are Number 8, which I describe in some detail later on, and Number 12, which, built in parody technique, is a special case. One can readily hear No. 8 as "in" a blotchy G# minor and the following piece as a vagabond A flat major. Number 12 really has a tonic (or good imitation) with a functional IV and a "dominant prep" on the lowered 6th. Most other keys are local color patches, though sometimes persistent.


Note that the requirements of the set distribution dictate in this case that seven of the thirteen numbers must deploy a pentatonic scale, three with A flat, three with B plus one with both. The one with both is Number 4. At its beginning, number 4 quotes a traditional pentatonic tune in its entirety with a pentatonic accompaniment ­ a passage of some extent that is entirely diatonic. In Number 10 extended passages are confined to the "A‑minor" pentad. In both these cases the quasi‑tonal passage is interrupted by an interjection which makes no tonal sense whatsoever. These interjections sound familiar when heard in the context of the whole suite as they repeat prior harmonic material. I estimate that they sound arbitrary out of context, but the context is fixed in my mind.


It is because the presentation of the pentads is often so relaxed and redundant that the natural word for them in my head seems to be "mode" instead of "PC set". I will point out in a fuller analysis of Number 8 how another logic is becomes syntactically and perceptually predominant, ultimately resisting the representation of tonality. For a first preview of that logic, Example 2 shows entrances of the subject of Number 10, the fugue.




Example 2: Incipits of principal fugal entrances in No. 10 Per Augmentationem


The question of what tonality becomes when it reappears in post‑serial or elaborated minimalist styles has been in the air for a few decades, and I don't think we have a standard tool‑kit for describing it. Perhaps we have not adequately appreciated how diverse a phenomenon tonality was before 1900. My own quick remark above, that tonality in a strong sense of the term is a rule system, short changes the problem. Functional harmony entails rules about complete, ordered sequences of chords, but there are lots of other manifestations of tonality. Tone centricity is an element often mentioned. Fields of connotation are elements more often neglected. The Tristan chord is not enough to imply tonality by itself, but the opening measure and a half of the Tristan Prelude, played well, with suggestive dynamic phrasing, certainly is enough. References to vernacular styles associated with tonality can prejudice our hearing and interpretation of harmonies that, in themselves, offer minimal grounds for a tonal reading. Perhaps that is the case in Number 8, Tango. A hint of tango rhythms and tango gesture might be enough to encourage interpretations of the harmony in relation to minor and even to establish a positive valence for a suggestion of flat second degrees.


The full score is reproduced as Example 3. Figure 5 shows the first of the three, strictly parallel 17‑measure verses analyzed by lynes on the underlayed staves. Except at meas. 14 and 15, I think it is evident that the segmentation of the verse by PC sets largely corresponds with obvious Gestalt boundaries.


No need to catalogue all the elements of a potential tonality. To remark just a few, the four‑three resolution to G sharp minor at the opening is repeated in the last measure of the Coda (in Example 3, not 4). There is a full‑fledged V7‑I progression in B flat at measure 13. The parallel 6ths at measure 7 add by their connotation of popular styles to the impression there of E minor. The F natural with which E‑major elements are associated through out does not weaken the  centrality of E because the flat second degree, a  characteristic of some Spanish vernacular styles, was imported to the Argentine tango. If the first verse, mm. 1‑17, is taken in isolation, then I think the play of tonal suggestions, disorganized as it is, can  appear  the governing unity. But play through Example 4 which brings together measures 14‑16 with its corresponding transpositions in the second and third verse. Notice that the E‑major triad of measure 14 disappears through measure 31 and 48. The chromatic cluster which has replaced it in measure 48 needs some elegance of touch from the pianist to hold its place, but granted that delicacy, the notes sound "correct" because of the constancy of the set.  Note also that the one foreign body, the C, is invariant except that it moves up one octave. These are simple relations. The whole Tango proceeds by nearly systematic modal transposition of this sort. Since the structure of verses makes it easy to compare parallel passages, I leave it to any interested reader to extract other examples. I think it will be clear if the piece is studied as a whole that these modal transpositions (or rotations) establish an aural organization sufficiently forthright to dominate the more chaotic tonal references. The effect of this logic is stronger over the whole Tango than in a 17‑measure excerpt, and it is enhanced further by resonance with other numbers. Example 5 samples figures in PC set 2 from each of four numbers where it appears.



Example 3: "Tango" No. 8 from Voice Mail.








Figure 5: Voice Mail , No. 8 mm. 1 - 17 analyzed by lines.



Example 4:  Measures 14 - 16, as above with parallel figures (modal transpositions).





Example 5: Example of second pentad in three other numbers (see also Figure 5).


As the set class identities gain force across the suite, showing stability and accessibility, the tonal indications, though salient, are revealed as ephemeral. G‑sharp minor frames the Tango, but to see that this tonality has been represented without its really taking hold, I think it is sufficient to play the Tango (ideally omitting the six high register C's) substituting a tierce de Picardy at the end. By the time we get to the Coda, the 'dissonant' bass notes which respond to the triads in measure 52 and 53 sound like right notes ­ one recognizes the local closure when the pentad is completed. But a B sharp substituted in the last chord sounds like an intrusion from another universe.


Of course the flicker of competing triads has as much to do with representing "tango" as with representing "key". Genre stylization injects color. Like tonality, genre also encounters resistance from the rigidity of the pitch system.



Topics, Gestures, and Genres ­ Voice Mail as Program Music.


Voice Mail plays with programmatic titles and allusions to genres. In contemplating the game, I would hope we would recall the idea of imagination in early 19th century music. "Imagination" is not a synonym for "invention" or "creativity." Coolridge investigates imagination as the special faculty that produces and elaborates "image". Voice Mail offers grounds for imagining but often twists or undercuts its images. Such irony was not at all foreign to Romantic aesthetics, but I believe it was then a counter‑current, a reaction to the foregrounded ideal of the imagined image as a transcendent symbol. In Voice Mail, features that either resist or distort representation are quite pervasive. The harmonic techniques already described support this tendency. Early 19th century music depends on the hallucinatory power of smooth and sensuous tonality. The harmonic system of Voice Mail encourages various grating and abrupt transitions that make no straightforward contribution to the development of its initial images. The overall quality is not dream‑like, unless it be a dream recurrently interrupted, perhaps dreams of bucolic idylls littered with pop cans and Styrofoam. Such a metaphor illustrates an ambiguity inherent in the musical situation. Almost inevitably, the interruption itself or any other twist or distortion permits further interpretation as an elaboration or ornament, as what Hatten (1994) calls a trope. One might give up pastoral dreams because of the damn beer cans, or one might dream the cans, too.


                 One kind of resistance can be understood as a contest of topic and texture. Following Monelle (but abridging some subtleties) a topic is a type of musical figure that has developed an association with a unit of thought determined by cultural tradition‑‑moderately slow, minor mode dotted rhythms with funerals, for example. Adopting the notion of topic, we take advantage of a rare consensus in musical semiotic theory. The notion of "topic" from the older form "topos" re‑entered musicological discourse with Ratner's well‑known study, Classic Music.[6]  His argument that stylistic features were understood in the eighteenth century to carry semantic weight nourished a number of studies, among them, Hatten's Expressive Meaning in Beethoven, which accorded the notion of topic a rigorous reconciliation, till then lacking, with general theories of representation. Monelle's analysis, now available in his Sense of Music, shows decisively that Ratner's concept of topic belongs to our time, not to the Eighteenth century, and that it is a broadly applicable frame for investigating musical semantics, workable if the research is thorough, for any style of music. He also shows the kind of history homework you need to do to use the concept honestly. In the present case I hope to cash in on the illogical and indefensible privilege of composers to speak of their own music without doing that sort of proper homework. I will pretend you don't know the many reasons why a composer's interpretations of his or her own music should invite skepticism.


Titles and program notes should be part of the game in an account of the way the numbers of Voice Mail propose and extend or resist various topics,. It is quite possible that hearing Voice Mail in complete ignorance of its titles would be a better musical experience. My personal reactions regarding those alternatives with other music are inconsistent. In Voice Mail I have by now forfeited my choice, and I stick to the original agenda.



1. Invocation


To Mnemosyne and Amnemosyne, or to the goddesses of Memory and Memory Slips, if you will. The melody of Number 1 is a fake. Like a first assignment in a composition class, its notes are |A|, |G#|, and |B| throughout. I think it depends as a form on evoking associations with the melodic rhetoric of other styles of music. At the climax the melody is reduced to one note over pseudo arpeggios. The strain between the rhetoric and the repetitious, contrived pitch constraints are like brackets or scare quotes around the "melody" thematizing, with a little help from the title, a topical reference to memory, but if that signal is too subtle, the quotation ending the piece, from Schumann, Op. 78, ought to do the job.



2. "Midnight Skyscape"


"Midnight Skyscape" is representationally straightforward even if the texture has turned inside out before the end. The musical syntax is orderly. Any listener so inclined can make up a picture or a story, though I don't have one; the music does not resist. (Representationally straightforward does not mean representationally definite.) In combination with a mounting song and culminating fireworks, the title invites the pathetic fallacy: nature as a metaphor for soul.


3. "Pas de deux"


In "Pas de deux" the principle topic that suggested by the title and by the tune is "ballet", which the performer can realize in the somewhat static, seven‑quarter note measures (or phrases) that predominate with their recurrent foot (­ ­ _  ) a dactyl accentually or an anapest quantitatively. That is the same as the Sugar Plum Fairy's foot in Nutcracker, but much too fast, and accidents ensue. This representation is arbitrarily disrupted by a muddy outpouring, fortissimo, in the lowest register. In the written out repeat, the wound is repaired. A trope or a hole in the canvass? As I remember, the composer, when composing, only wished to fool with the sound and neutralize its semantic character, but by the time he came to write program notes, he had figured out a literary interpretation (which I will spare the reader).



4. "Traditional Song (Anglo‑American)" 


The fourth number, "Traditional Song (Anglo‑American)," which I mentioned above as the number that presents both of the pentatonic scales, is the one number I offer without hesitation to a listener who has no habit of attending to learned music. The folksong is more than a topic here, as it is fully quoted (below) and needs no quotation marks. There is no doubt in my mind what my representational intentions were, right from the start. I learned and loved the "Riddle Song," when I was a young child, and years later, not yet an adult, but then a camp counselor teaching it to campers, I suddenly understood it as a song about stillbirth.


I gave my love a cherry that had no stone;

                                    I gave my love a chicken that has no bone    

I gave my love a ring that has no end;     

I gave my love a baby with no cryin'       


How can there be a cherry that has no stone?    

How can there be a ... etc.       


A cherry when it's bloomin' it has no stone.    

A chicken when it's pippin, it has no bone.    

A ring when it's rollin' it has no end.    

A baby when it's sleepin' has no cryin'.   


The program notes set up the music to sound sad. Perhaps, without a program note, it would not sound sad.  (There is no way for me to know.)  I had no intention of opacity or irony; yet, I have a special reason to mention this representation in the present discussion of resistance.                   


The melody itself is presented with slight irregularities of rhythm that, to my ear, compensate for the rhythmic nuance words would afford in a sung performance (Parlando style,the topic of "speech".) As the number proceeds, phrases are increasingly separated by isolated and angular interjections (gestures, "speechlessness.") A tiny elf pulling at the composer's left ear, complained of the corny, maudlin result.  It was absolutely necessary to oppose representation, to return for an instant to pure sensory listening. My solution (not motivically irrelevant, by the way) was the measure shown in Example 6.



Example 6: Measure 32 of No. 4.


Surely, if one measure of music in isolation can mean nothing, this one should. But resistance to a force does not cancel it.  Resistance attests its object. We look at a cathedral and see gravity made visible. Sails let us see the wind. I can't say no one will find the song too sentimental or that everyone will find it sentimentally effective, but I think I can say that in this case, resistance enhances representation. In music and in persons, we sometime identify most with emotion held back. If you accept the logic of my demonstration, examples of the same principle in standard repertoire may come to mind.



5.‑ 6.


The fifth and sixth numbers, the two most suggestive of 20th century neo‑classicism in style, don't participate readily in the free‑ranging imagery of the others. Number five, the two‑part "Invention." is faithful to its genre, and this study in imitative counterpart may have no topics outside musical style history itself. The subject, fixed by rhythm and contour, is a type known from Bach, fugal subjects in the manner of a gig. The counterpoint opposed to it invokes the athletic pointillism of a more recent era. Until toward the end, where things speed up and fragment, the texture formed by the contrasting lines is merely an exaggerated example of a contrapuntal strategy already familiar from Bach and Scarlatti. The sixth piece, "Variations" makes no particular references that I know how to specify, even though I am privately aware of some sources. 



7. "Opera Song (Male Voice)"


"Opera Song (Male Voice)," on the other hand, is elaborately representational. The bass line is complicated enough to suggest both a persona and a developed rhetoric. The three sections of the aria are separated by a quiet figure blurred by pedal that I had thought of as changing stage lights, until another pianist told me they were actually the off‑stage choir. The representation goes awry, and I don't think you will find a story or a trope that can undo the semantic nonsense. The song's development proceeds through a series of rhythmic mutations that have more to do with kriti than bel canto. It isn't worth the ink it would take to figure out whether the result is semantic or anti‑semantic; suffice it to note that the question is on the table. (Further reference: as a piano's aria for bass voice, "Opera Song" renders homage to Chopin's C# minor Etude  ­  and of course, the composer bears the risk of unflattering comparison. But let it not be said that he ran away.)






8. "Tango"


The "Tango" was meant to sound like a tango and be played like a tango. But it is (another pianist pointed out) a cubist tango. The roughed up rhythms and disjointed phrasing might either enhance or endanger the reference. At a detailed level, its genre is not consistent for there are, I believe, elements of two distinct traditions, the sung tango and the danced tango. Tango is a loaded topic evoking images of sexuality, violence and, for the sung tango, complaint. (The typical lyrics are about the woman who did me in.) None of these themes are evoked consistently, but perhaps all are available to the pianist in fragments.



9. "Party Song"


"Party Song" is fast rock, perhaps instigated in my imagination by Chuck Berry, or perhaps instigated by memories of unsuccessful auditions by students hoping to be admitted to my department. I don't know a lot of rock. A young woman who hears me practice often and manages to ignore my music very consistently said, "Hey that sounds modern; I didn't know you were interested in contemporary music." I took this remark as proof positive of a successful representation. Nevertheless, this representation is, depending on your taste in semiotic theory, either resisted or troped by long and short unrock‑like silences and extra and missing half beats.



10. "Per Augmentationem"


I think of "Per Augmentationem," the fugue, as the most abstract number of the whole suite; yet, it is not like numbers five and six in this respect. The Invention and the Variations are both busy and more insistent on focusing attention on stylistic features. Number ten is slower and in any reasonable interpretation, the calmest of the pieces. I provides the performer with an opportunity to display a contemplative mood. The syntax of Voice Mail is largely patent, but the relations of augmentation here are perceptually liminal, a characteristic that seems to me to participate in evoking a more diffuse state of consciousness. There are both pro‑genre and anti‑genre elements. Anti: The action stops for an exposition of the subject in inharmonic similar motion at registral extremes‑‑a moment of contrapuntal nonsense. Pro: The counterpoint then resumes in a pianistically enriched texture that, despite ugly little blotches, is right out of the tradition of keyboard fugue.



11. "Per Organo Pleno"


Dense over‑pedaled chords represent the organ and a plodding melodic rhythm represents chorale style. But in the middle, the image turns inside out, pedal off, doubling off, meter awry. Once again, whether that reversal is an interruption of reference or a trope on it, I will not try to figure out. This is data enough. I must now explain in what sense none of this matters and in what sense I think it does.



The Semiotic Character of Voice Mail.  



There is no story, no argument, no plan organizing the collection of programmatic and generic references or at least none that I've noticed. There is a stylistic prise de position, but no integrating discourse. Like the tonal structures that are broached but rapidly discarded within several of the numbers, the fairly rich splatter of semantic pointers offers color and variety but contributes little to coherence. Furthermore, for all the talk of topics in classical music, the same balance often holds there. Topics bring life to an abstract design, provoking greater identification and involvement with its parts but not accounting for the sense.  The sense is made by syntactic design. In the present design, the Fano lattice space is gradually saturated. As that happens there is an irregular but still palpably directional enrichment of pianist means, (for example, richer pedal effects in the 11th and 13th pieces than elsewhere.) The strongest genre contrasts are in the second half. Music can be abstract in its overall conception and nevertheless very dependent on reference to sustain the vivacity of its character. Music lives as an abstract medium when it is abstracted from something. Pure design without the coloration of referential suggestions is sometimes of great interest to us, but over all, I think it forms a rather minor part of music, I would say even a rather minor part of abstract music.  The refusal to pin down ultimate references is not a flight from semiotics. What I advertise here is no more than Suzanne Langer declared when she defined a sign as "anything we use for abstraction."


Notation permitted European music to develop a wonderful division of responsibilities between composition and performance, but the basis of this division is not a European invention and does not depend on writing. Musical thought and musical structure combine categorical and quantitative elements. The choice of a chord or rhythmic figure is categorical insofar as the element selected can be recognized as an instance of an equivalence class. Finesse of intonation, dynamics and rubato (or swing or "feel") is quantitative.  These nuances may be similar, one to another, but they are never entirely equivalent. Fundamentally, anything an improviser might be able to decide about a performance before beginning to perform is abstract and categorical ­ these are compositional decisions. Composition lends itself to notation.  Quantitative decisions can be measured and graphed, but an analog graph is not, technically, a notation.[7]


The consequence of separating composition from performance is largely to alienate composition from the deepest, plainest and most prevalent determinants of musical meaning, the nuances of rhythm and color that convey gesture and tone of voice. Those are the responsibility of performance.


How I understand that responsibility can be shown with the opening three measure of the first number (Example 7). The diacritical markings (which are more graph than notation) are more elaborate here in than in most of the suite. A simple double turn is repeated with an indication of a swell. The swell is contrametrical, so if one is inclined toward meter ­ I am ­ one might want to do more. Consider the possibilities of accenting the first note or the last note, much or little, of either figure, the possibility of playing them the same way or playing them to contrast. Performance makes gesture, not composition. The gestural intentions of the performer establish a direction among diverging worlds. What's the right one?




Example 7: Measures 1-3 of No. 1.


But that is merely the pianist's right hand. More is at stake in the left hand in realizing meter, which will itself become more representational when it is firmly embodied and less representational when it is disembodied. I know my own taste in this regard. The gates of music opened for me so that there was no turning back when, as a child, I first encountered the rollicking bass octave ostinatos in musette of Böhm that J. S. Bach copied out in the notebook for Anna Magdelena. I never outgrew it. I still sometimes feel like a child reaching out for Mr. Böhm's hand, to catch his beat. But what does the score tell us of this with its sometimes persistent, sometimes changing, sometimes obvious, sometimes contradictory metrical structures? The pianist, Lidov, now plays parts of Voice Mail with a bouncing measure that the composer, Lidov, never had in mind. The semiotician, Lidov, finds this entirely as it should be.


When I was a young composer I heard much talk of a need to liberate music of the Western classical tradition from its dependence on the score. That discourse was salutary at the time, but the thesis was false. A better thesis would have been that performance needed to be liberated from the score.  The first half of the 20th century suffered the convergence of at least three tendencies that inhibited imaginative performance. In composition, expressionism gave rise to a style of notation that aimed, or seemed to, to be fully determinative of performance, as if it were the performance that was to be notated, not the composition. The progress of music historical studies in documenting the theories of style of various epochs as well as their performance practices undermined the assumption that composers' intentions could be best discovered by the performer's development of his or her personal intuition. And, to provide the new performance style authorities with the police services, we elaborated the art of sound recording, which ensures that we live with our mistakes and exaggerations.


                    The problem of constraint on performance has somewhat healed itself in recent decades because of experimental notations, renewed interest in improvisation, more subtle historical reconstructions, and even a fluidity of recording that robs the record/tape/CD of monumentality. My thought when composing Voice Mail was that we have a ways to go. We can see an older attitude in the questions and injunction with which Joseph Lhevinne concludes the penultimate chapter of his slender but very thoughtful book, Basic Principles in Pianoforte Playing.[8]  



1. Do you express the composer's thought and mood?

2. Do you express what you feel and wish?

3. Whatever it is, by all means express something! (Emphasis in the original, numbers added.)



This will not quite do for us. We know that the score is not sufficient evidence by itself of the composer's thought and mood. But what if the composer's mind had never settled the this matter to begin with? May I presume, as the composer in this instance, to assert that my intentions in composing, if there were any, should be irrelevant to the performer? I never renounced my own intentions; I simply used them up. I used them for choosing notes, knowing full well that the notes could not capture all the intentionality with which they were, fleetingly, invested ­ in my imagination. 


My technique in Voice Mail aims to arrange musical notes in such a manner that the performer will have specific incentives (topical clues, traces of gesture, hints of voice and rhetoric, evocations of memory) to assign intentions to the music and will have space to realize them. The page is neither blank nor fully colored in. Option 1, above, is rejected outright for most of Voice Mail. (I don't want to pretend that my stance is pure or radical. I certainly did mean, for example, that the tango should sound like a tango. What a tango is has as much to do with how you perform it as how you compose it. I hope the score provides grounds for the hesitations and violence that I love in tangos.)


But option 2 is also rejected. If it does not convey my mood, the score does suggest a way of making sense. The only acceptable manner to realize option 3, heartily endorsed, is by making sense of the score, as if it had consistent intentions. The pianist ought to play as the only one who, by special privilege, knows the composer's exact thoughts. One might posit that attitude for any music, but in many cases, there would be no specific grounds for it. Here, this disposition is implied by everything I have reported about structure and representation in Voice Mail. The irresponsible performance is the performance that does not aim at interpretation. The task of the composition was to suggest but ultimately to resist representation. The responsibility of performance is to overcome the resistance. 









Hatten, 1994. Musical Meaning in Beethoven: Markedness, Correlation, and Inter-retation.

Bloomington: Indiana University Press.


Noske, Fritz. 1977. The Signifier and the Signified: Studies in the operas of Mozart and

Verdi. The Hague:Nijhoff.


Ratner, Leonard. 1980. Classic Music: Expression, Form and Style. New York: Scribner.

[1] This paper will appear, with some revision, as the final chapter of the author's book, Is Language a Music? forthcoming from the Indiana University Press.

[2]  Semiotics is concerned with general problems of representation: how systems of reference are comparable and different; how reference is possible and how it is limited; how signs are biased and how they compete with their objects, etc. It has always seemed to me that a special semiotics, like the semiotics of music, is intellectually interesting only if it is clearly related to a general theory of semiotics. When I began to consider problems in musical semiotics, I got stuck because I did not find a general theory to which I could relate the problems of music. In consequence, I felt impelled to try my hand at general semiotics.

[3]  Lecture at the Toronto Royal Conservatory of Music, April, 2001.

[4] I first learned about this design in an introductory course in finite mathematics which I had the pleasure to attend in ca. 1978 with Morten Abrahmson. I wrote a paper for the course which enumerated the distinct ways of labeling the vertices of a seven point Fano geometry with the PC's of one diatonic scale. Sometime after that I permanently loaned the paper to a student.

[5] The all‑interval tetrachords provide similarly trivial realizations in a thirteen‑note scale.

[6] But see also Noske, 1977.

[7]  This paragraph is a crude summary. Notation is characterized technically in Nelson Goodman's Languages of Art. His system is exploited in the distinctions between performance and composition I develop in Elements of Semiotics.

[8]  Joseph Levhinne, Basic Principles in Pianoforte Playing, with a new foreword by Rosina Lhevinne. Dover, New York, 1972. First published serially in Etude and in full by Theo. Presser, Philadelphia, 1924.