Grieg, Lyric Pieces, Op. 68, part 2

The somewhat strange Valse mélancholique is no. 6, the last number, in Grieg's Op. 68. Sounding like a merger of Chopin's well-known melancholy waltz in A minor and the most repetitious of Tchaikovsky's ballet waltzes, the Valse mélancholique is also insistently dissonant, breaks up its harmonic progression oddly on more than one occasion, and includes a decidedly un-melancholic explosion in its structural cadence.

A rising line is a central figure in the opening. A four-introduction is followed by a thirty-two bar period that forms the waltz's first strain. The 16-bar antecedent phase, bars 5-20, brings a line very gradually up from D5 (doubled with D4) but it is broken after the F5 in bar 19 by a surprise harmonic break to a Neapolitan (G minor: an Ab6 chord) in bar 20, after which the consequent phase begins and plods along, becoming more and more insistent and finally pounding out the concluding notes of the rising line.

I hear the more active "alto" as thoroughly meshed with the octaves just discussed above, but it is entirely legitimate to hear the internal voice more or less independently. I have traced it in red below, an extra strand emerging after bar 28 being in green. Regardless of how it begins, however, the dramatic rising trajectory of the octaves takes over in the final bars.

The coda follows directly on the repetition of the fortissimo cadences in the reprise and takes a classic coda form with "reminiscences" over a tonic pedal point. Note that the octaves, having fulfilled their role, are gone, and we hear the alto's tune, still in the middle of the texture.

Lehar, The Merry Widow, Act 1 Ball-music

Franz Lehar, Die lustige Witwe / The Merry Widow (1905), Act 1 Ball-music. This single waltz strain is to be played offscreen under conversation, but also ad libitum, meaning it can be left out (the same music appears later in the score -- a recording I listened to recently included the first statement but dropped the second one). The design is a 32-bar period (common in waltzes after about 1860, but also found regularly later on in marches, one-steps, and rags). The first 16-bar phase is one of those awkward units where period or sentence might apply equally well, depending on how you take the identical rhythms but different melodic shapes.

A triadic interval frame is established quite clearly -- circled at the beginning below -- and is returned to at the end of the first 16-bar phase -- circled at the beginning of the second system. In the second 16-bar unit, the triad is shifted upward -- circled across the second and third systems -- and overlaps with a simple scalar progress through the octave, with a registral drop on ^3. See the next example.

Here is a schematic of the triad frame (as proto-background) and a parsing of the octave line in the final six bars.

Brahms, Liebeslieder-Walzer, op. 52, n3

The third number in Brahms's opus 52 is brief, a duo for tenor and bass, for whom the slightly obnoxious text offers that a fellow would have become a monk if it were not for the charms of women.

Here is the piece in a piano reduction (not by Brahms but obviously based very closely on his piano four-hands version). All in all, a simple rising line from ^5 in the first strain, repeated to end the second.

Here is the texted version with solo piano accompaniment. The tenor might be said to trace a line from ^3 downward, so long as one is willing to understand ^7 in the penultimate bar as a substitution for ^2--not unreasonable. It is, however, telling that Brahms—far from emphasizing the voice parts in his own piano versions of the piece—dropped them.

Post no. 300

It is four years to the day that I started this blog (introductory post), and this post is the 300th in the series. The blog has provided material for over twenty essays published on the Texas ScholarWorks platform (link to my author page).

The focus has been quite narrow throughout: ascending cadence gestures in traditional European tonal music. Though my work on these figures has its roots in Schenkerian analysis—an article I published thirty-one years ago ("The Ascending Urlinie," in Journal of Music Theory)—I use those particular constructions only some of the time, because of problems with the figure of the focal tone (Kopfton), which I see as defined too narrowly to be generally useful. I opt more often (1) for the model of proto-backgrounds (link; see also Neumeyer 2009), (2) for style historical methods in connection with the dominant ninth chord and what I call the “waltz ninth,” (3) for similar methods and speculative modes for improvisatory practices, especially in the later 18th and early 19th centuries, and (4) for historical narratives of music for social dance, of music for the stage through the nineteenth century, and of composition in general in the later decades of that century, both narratives continuing to be relevant in the first half of the 20th century, as well.

The great majority of the traditional tonal music preserved in scores and manuscript makes use of form-defining cadences in which the principal melody line descends the scale to end on the tonic note. A significant minority, however, follow an upward path to end on scale degree eight (^8), a “circling” path around ^8, or a “mirror” path down from ^8 to ^5 and returning. An early surprise in my work was to find a considerably larger than average percentage of rising lines in the country dances preserved in John Playford's Dancing Master (first edition 1651; link to essay), which fact suggested to me that the figures were relatively common in dance-performance practice, including improvisation. It must be remembered that music for social dance was predominantly music for the violin (secondarily, flute) and that the instrument’s fifths mapping made it as easy to rise from the middle of the scale to ^8 as it did to descend to ^1. 

I later found considerable corroborating evidence in Scottish and Irish dances and dance-songs, these coming largely from the later 18th century (link; link; link), and in the Germanophone Laendler, one of the waltz ancestors and especially closely associated with the violin, later in the 18th century with an ensemble of two violins and bass.

In the example below, the four open strings are depicted at the left in (a) and (b). From the open A-string, one moves with the greatest ease down the D-major scale, as in (a), or up the D-major scale, as in (b). In counterpoint, supposing for example, one violin improvises a descant to the other, the work is almost as easy, one of the simplest versions shown in (c). Another version with a bit more melodic complexity is shown in (d). Register play, in other words, offered a simple device to "do something different," specifically to do something different for an ending/articulation that didn’t run afoul of the traditional cadenza, where ^2 goes down to ^1 while ^7 goes up to ^8, making intervals of the major sixth and the octave (or minor third and unison). Not coincidentally, the alternative higher-register cadence offered a sound that was "bright" or "brilliant."

Additional corroboration of early practices came from 17th-century Germany and Austria; these included repertoires across the entire spectrum of genres, excluding only sacred choral music: link to essay.

The history and practice of rising cadence gestures quickly became more complicated in the early 19th century. Broadly, though, cadences can be heard as prominent expressive gestures and a turn toward less common cadences fits nicely with our familiar nostrums about Romantic rebellion against eighteenth-century conventions—and it aligns well with music theorists' recent revelations about a kind of shadow tonality of hexatonic relations that arise from the exploitation of chromatic mediants, early on especially by Schubert, whose Laendler and Deutscher ("German dance," the other ancestor of the waltz) made significant use of rising figures. This is music, incidentally, we know passed back and forth between music for dance and music for performance, not only in Schubert’s case but in the pragmatic circumstances of music publishing and (especially) domestic use. 

Strangely, perhaps, I have found so far that straightforward expressive motivations for rising gestures—exhilaration, release, etc.—seem to have been far less prevalent than generic, topical, or formal-design considerations. The few coincidences of text and cadence, predictably, were found in the 19th century, after the clichéd cadence figures derived from earlier Italian practice had been largely abandoned (or, at least, their authority undermined). Examples: Grieg, Morgenstimmung (the analogy of musical ascent and the rising sun; link); Schubert, "Die Nonne" (the religious-utopian; link); Strauss, jr., Die Fledermaus n2 (increasing energy, demand, insistence: "Hinaus!”; link); Wolf, "Trunken müssen wir alle sein!” (as in Strauss; link).

As I have already noted, the rising cadence gesture was part of the toolkit of the waltz, and from there it went directly to the polka by no later than 1840. When an aria or other song used a waltz topic, the gesture went along with it. The floodgates were opened in French comic opera by no later than 1834 (see my essay on Adolphe Adam's Le Châlet: link), and rising cadences remained a factor in the opera bouffe and operettas of Offenbach, Lecocq, and others before finding a niche in the American operetta (Herbert) and the musical (notably those by Richard Rodgers).

The formal figure of the rising gesture in the coda of an aria or instrumental movement became firmly established in the last quarter of the 18th century. In some pieces, though, the boundary line between the "structural cadence" and coda figures became blurry—a process already underway earlier in the ensemble finales of Galuppi, with their many repetitions of cadential phrases, a dramatic device adopted by many composers, notably Mozart and Rossini. The process accelerated in the 19th century, more vigorously and consequentially in music for the stage than in instrumental concert music.

An important outlier in all this is Beethoven, for whom transcendence, as a philosophical-religious category, could mean striving to move not just to the top of the voice leading, but outside or beyond it. For my only comments on this to date, see my post on Op119n7: link. My plan is to do more with this eventually, engaging work by David Lewin, Robert Fink, and, more recently, Malcolm Miller.

Emile Waldteufel, Les Patineurs (Skaters), Op. 183

The famous "Skaters Waltz" is  Les Patineurs by Emile Waldteufel (published in 1882). Given its main melody—the first strain of waltz n1—the notion of ascending cadences would seem far afield, but look at this simplified piano edition:

In a version for salon orchestra, the piano/conductor score (a two-stave score with abbreviated instrumental indications) also shows the rising scale figure from F#4 to A4. In the full orchestral score, this figure is taken by the second clarinet and second trumpet. The piano/conductor score, btw, shows the complete texture: there are no subsidiary parts or notes in different octaves.

Les Patineurs has an introduction, four waltzes, and a coda. The 16-bar main theme is anticipated in the introduction, appears twice in the first waltz (a small ternary design), and three times in the coda. The cadential counter-motive appears in all of these but the first (the introduction) and the last, grandioso statement.

Looking at it in Schenkerian terms, three possible hearings are readily imagined. The first is a rising line from ^5, where an Urlinie parallelism traces across the middleground in bars 2-10. The E4 (circled) in bar 11 is particularly interesting in that it clearly initiates a lower, descending line through the 4-3 figure over B2, but E4 is also an aural trace of the initial ^5 that prepares for the ascent that begins a bar later.

Alternatively, if we give full sway to hypermetric accent, that allows a hearing from ^3, with a neighbor in bar 5, a possible return to C# in bar 7, a consonant C# touched on at the end of bar 10 and a twice-dissonant C# at the end of bar 12 and in bar 13.

Combining the figures of both lines in a three-part Ursatz design produces the following, quite satisfactory reading:

The problem is that the counter-motive is barely audible—to my aging ears, effectively inaudible—in all the recorded performances I listened to. The second clarinet and trumpet are simply overwhelmed by the unison melody in the first violins, cellos, and first clarinet and trumpet. In the end, then, if one is using Schenkerian tools, I think the best reading is this:

Note: I've left out one intermediate step, where the "quarter note" G#5 would have a C#4 below it and a bass C#3 (to account in some fashion for bars 7-8).

All in all, it must be said that this famous theme has a remarkable collection of diatonic dissonances.

Waltz "Du und Du," op. 367

Die Fledermaus draws on a variety of musical resources, including—as we have seen—popular dances of its era: the polka, galop, and waltz.

Strauss published a set of waltzes derived from the operetta; it is named "Du und Du," after the choral section of Falke's "Brüderlein und Schwesterlein" in Act II.

The set consists of an introduction, three waltzes (not the standard five), and a lengthy coda. The introduction immediately quotes "Brüderlein und Schwesterlein":

but before too long inches its way toward "Ha, welch ein Fest," the climactic dance of the Act II Finale:

The first waltz, then, offers up the two strains of that waltz:

Waltz n2 gives us Rosalinde's rebuke to the jail warden Frank in Act I and Adele's rebuke to Eisenstein from Act II.

Finally, n3 brings us two strains from "Du und Du":

In typical fashion, the coda then recapitulates several strains from the waltzes, separated by dramatic transitions.

As a point of interest, Strauss published several other pieces with material derived from Die Fledermaus: a Fledermaus-Polka, Op. 362; a Fledermaus-Quadrille, Op. 363; a polka "An der Moldau," Op. 366; and a polka-mazurka "Glücklish ist, wer vergisst!", Op. 368.

Postscript to Strauss, Die Fledermaus n11, Act II finale

The five parts of the Act II finale are (1) Orlofsky's toast to champagne--which we discussed in the previous post; (2) a comic exchange between Frank and Eisenstein; (3) Falke's waltz-song "Brüderlein und Schwesterlein," which leads into the "Du und Du" waltz; (4) ballet (handled in various ways in different productions); and (5) Prince Orlofsky's call to dance "Stellt Euch zum Tanz" and the dance itself "Ha, welch ein Fest, welche Nacht voll Freud!"

In the last of these, Strauss shows one of the strong tendencies in his later waltzes: toward 32-bar units, either by "stretching out" double periods or sentences (making them 16 + 16 rather than 8 + 8), or--as here--by so closely linking two 16-bar strains that they make a single musical unit:

There are no ascending cadence gestures in this extended and exhilarating waltz, but it doesn't have the last word in the Act II finale, as the proceedings come to a halt on a resounding cadential dominant -- see the beginning of the example below -- and everyone offers up a last salute to champagne by repeating the music for the Prince's toast. In the process the choral sopranos mark out the essential elements of the voice leading for the tune that Rosalinde—along with Orlofsky, Adele, and her sister—sings in a register that makes its yodeling topic even more obvious than it was earlier.

The orchestra, then, goes loudly to it one more time, stretching ^6 over IV and ^7 over V to two bars each and then beating on ^8 for no less than nine bars.

Strauss, Die Fledermaus n11, Act II finale

The setting of Act II is the ball to which Eisenstein and Falke have stolen away. Thanks to Falke's plan for revenge on Eisenstein, however, pretty much everyone shows up, although in disguise; the only person who isn't in disguise is the host, Prince Orlofsky. Unlike the first act, the second has no rising cadence figures outside of the finale.

The Prince opens with a toast to champagne. The music is mapped out as three eight-bar strains (A, B, C below; C is repeated as C') with a twelve-bar insert. Strain B, in the dominant, functions as Caplin's contrasting middle, a "B-section" in traditional form terms, to which the insert adds a retransition. Strains A and C are distinct, but both use ascending lines in their cadences.

The two phrases of A are essentially the same, and they would form a simple wedge figure except that ^2 in the descent has to be imagined. The ^3 (as F#5) is clear enough as a focal tone, but all the attention after the first bar goes to ^5 and then its tra-la-la-ing ascent in the cadence. The voice is accompanied by lower orchestral voices in this segment, and one can find the requisite ^2, though as E4 not E5 and in the third horn and viola, which placement doesn't inspire confidence about the musically revelatory.

Strain C has the toast itself, and its tune is built much like the one in A, but with the lower line stretched out to a sixth and the force of the upper ^3 as focal tone much diminished. In the cadence it's more plausible, to my ears anyway, to hear F# moving up to its similarly expressed neighbor G than it is to pull out the sixteenth-note E4 for a descent. The priorities I am hinting at here become obvious in C', where the cadence brings more attention to ^6 and an over-reaching ascent in the Flute 1 part (marked and notes circled below).

Strauss, Die Fledermaus n5, Act I Finale, "Herr, was dächten Sie von mir"

Alfred and Rosalinde's duo in the "Trinklied" was the first musical section of the Act 1 finale. In the subsequent section the jailor Frank enters and a comic szena ensues in which Alfred continues to sing phrases of the "Trinklied" while Frank attaches his own meaning to them, in the process mistaking Alfred for Eisenstein. Rosalinde realizes what is happening and, because she is trying to get rid of Alfred, claims he is in fact Eisenstein. Her strophic song "Mein Herr, was dächten Sie von mir" elaborates on this ruse ("How could you imagine I would be here with anyone other than my husband? Etc.). See the opening below. Strauss often uses polkas for happy moods or congeniality, but sometimes for irony or, as here, for a series of comic double entendres. (The polka I am referring to is the original type from the 1840s, known in the second half of the 19th century as the polka française, or a slower tempo polka. The music of the polka schnell, in a fast tempo, is barely distinguishable from a galop.)

Note the very strong emphasis on ^6 (E5) as the ninth in a V9. Also note the tonic with add6 at the end.

Still putting emphasis on ^6, the second half of each strophe switches to a waltz, which consists of a double period (Caplin's 16-measure theme) that is repeated. In the first iteration the melody makes its way through an octave -- see the beam.

In the second iteration, Frank and Alfred join in, making for a bit of contrapuntal play. In the cadence, Rosalinde takes the melody back up to G5.

Here is a reduction of the voice leading for the final bars.

Postscript 2 to "Trinklied": parallel fifths

Postscript to "Trinklied": assuming a dominant-root in the bass, the "fall from the dominant" in the cadence traces a V9 chord, which -- in the major key -- poses the danger of parallel fifths if 9 in one voice descends to 8 (that is, ^6 goes to ^5 in the tonic resolution), and 5 in a second voice descends by step (or ^2 goes to ^1) below the first voice. The problem is easily seen in the figure from my previous post (below): E5 goes to D5, A4 goes to G4.

Musicians obviously found ways to deal with this, as the sound of the V9 chord is a particular feature in 19th century music of all but the most conservative sorts—and it began early, with Schubert's generation. In yesterday's post on The Blue Danube waltzes, we saw Johann Strauss, jr., use the simplest method: resolve 9 to 8 before the tonic chord. Thus one gets the expressive sound of the ninth chord without the voice-leading hassle.

The problem of parallels in certain progressions--and clever devices of detail to overcome them--is far older than the 19th century. In the 16th century, the 5-3 sequence was occasionally used -- see Ex. 1a. In the 17th and 18th centuries, the ubiquitous circle of 5ths sequence avoided parallel octaves and fifths while invoking the sound of a stream of perfect intervals--Examples 1b and 1c.

In Example 2a, I have condensed the "Trinklied" falling figure into a chord. At 2b is the older voice leading for viiø7-I; this assumes that both C5 and A4 resolve to B4, but that means one cannot have a descending cadence melody: ^2 goes to ^3, not ^1. The recommended textbook voice leading for V9 in four parts (Example 2c) conveniently leaves out the fifth (A5 here), erasing the problem of parallels, but still making V9 unusable in closing cadences. Example 2d translates 2c into a cadential form, but that means ^7-^8 in the uppermost voice. Finally, Example 2e overlays 2d on the "Trinklied" melodic figure, which reintroduces ^2 (A5) but doesn't sustain ^6 (E5) throughout the bar, thus barely escaping directly sounding parallels.

In general the orchestral parts reinforce this. For example, one of the horn parts holds D4 throughout the four measures of the cadence phrase. One of the woodwind parts even traces a melodic line through A5 but then is silent during the final bar's tonic chord!

There is an exception, though. While the second violins hold A3 and resolve it directly to G3—

—the second clarinet (which is in C) holds E4, clashing with the F#4 in the second violins' double-stop, and only touches on F#4 for an eighth note before dropping to D4. This is as "barely" as barely escaping fifths can get.

Strauss's generation was the last to try to observe the prohibition against parallel fifths. In the next generation, Debussy and his peers divided parallel intervals/chords and contrary motion into two sets of effects, both equally expressive and usable.

Postscript to "Trinklied": The Blue Danube

At the end of yesterday's post on the "Trinklied" (first section of the act 1 finale in Die Fledermaus), I showed several examples of the cadence figure I call the "fall from the dominant" in one of the early waltz sets by Johann Strauss, jr., Die jungen Wiener Walzer, op. 7 (1845). In previous work I have shown that this gesture is a characteristic one in the waltz repertoire, starting in the early 19th century Ländler.

Here are more examples from one of the most famous of Strauss's later Viennese waltzes, The Blue Danube (1867).

The simplest is in the first strain of n5: a rise to 9 over V, resolved to 8 (circled) before a drop to ^7 (the whole figure boxed).

Note that Strauss—unlike his father—has not singled out the cadence for a higher register: we hear that already in the second bar of the excerpt. The most dramatic instance of this "early" arrival is in the first strain of n1 (below), where a firm upper-register ^3 over I is reached before settling to the cadence, which includes another simple 9-8 over V (circled).

An even more dramatic version is in in the second strain of n5  (below), where the high register is reached over the cadential dominant 6/4, and the "fall" is a long scale figure that moves through an entire octave.

Here (below) is another that reaches its (literally) high moment over the cadential dominant 6/4. This is the second strain of n3. Arrivals of this sort are the most traditional of Strauss's cadential constructions, as a dramatic expressive arc toward the cadential 6/4 was a commonplace in the early 19th century (its most exaggerated expression being the orchestral chord that signals the beginning of the cadenza in a concerto movement). Note here that the figure over the dominant seventh (the second circle) is the one used in the "Trinklied" (the key, G major, is even the same).


Here are two examples where significant emphasis goes to the S or pre-dominant chord in the cadence. The first example below is from the first strain of n2; below that is the second strain of n1.

In the first strain of n4, S and the cadential 6/4 are nicely linked (below). Note that, as in the first strains of n1 and n5, the high register actually precedes the cadence by several bars.

 Finally, in the second strain of n4, several of the elements shown above are combined: the highest point is over the root-position tonic (fifth bar from the end below), the fall from the dominant is again stretched to a leisurely scale (boxed), but the "Ur-form" of 9-8 over the dominant with a drop to the tonic note nevertheless concludes (circled).