This device of undercutting the rise from ^6 to ^7 is discussed in my JMT article and seems to be particularly characteristic of the later 18th century. To speculate: the conventions associated with the dominant Italian style (which we know much better nowadays thanks to important research on the partimenti, evidence of methods of instruction) were so strong that Haydn felt an obligation to observe them in some situations, rather than take full advantage of the rising cadence gesture. In any case, the leap downward from a subdominant to the leading tone is very expressive in and of itself. (Survey, p. 64)In the note, five compositions are mentioned. I have already written about three of them in the essay Ascending Cadence Gestures: A Historical Survey from the 16th to the Early 19th Century: (link).
Haydn, Piano Sonata in E-flat, Hob. XVI/52, II. Survey, pp. 76-78.The other two are Beethoven, String Quartet, op. 74, IV, and Corelli, Trio Sonata, op. 2, no. 8, Preludio. I'll discuss the latter first, because it affords an easy opportunity to sort some of the issues related to register.
Haydn, Piano Sonata in A-flat, Hob. XVI/43, Menuet. Comment in the note: "the large-scale structure is obscured somewhat by strong emphasis on ^3 in the Trio." Survey, pp. 74-76.
Haydn, String Quartet, op. 76, no. 2, II. Survey, pp. 78-83.
Register transfer in the rising line is worth some comment. Examples (a) - (e) apply octave or seventh registral changes to each successive tone of the rising line from ^5. In (a), the very common change of octave over a stable bass; in (b), the figure used by Bach in BWV 924; in (c), the registral variant I reference in note 32; in (d), the highly violinistic broken figures one frequently finds in Baroque music, where it is a 50-50 chance the final ^8 will be in the lower or upper octave; in (d'), a variant that applies the register change to a neighbor note -- this is a major-key version of the figure in the Corelli prelude to be discussed below; (e) is similar to (a), a simple octave embellishment of ^8.
My comment in note 32 is that "Very occasionally register transfer is applied to other tones [than ^6]: in Corelli, Trio Sonata, op. 2, no. 8, Preludio, the variant ^5-^6-^7-(^8-^7)-^8 has a dramatic octave-leap downward applied to the first ^8." As my parentheses suggest, the register change here is applied to a middleground neighbor note, not to an Urlinie tone.
The reading requires a line from ^5, which is certainly as plausible as one from ^3, even if we were to insist on a descending Urlinie form. In the closing cadence, the first violin takes the line steadily up but breaks at the dramatic #4 diminished chord to place its final notes an octave lower -- and below the persistent descent of the second violin. Here is another notation of the ending, emphasizing the parallel 10ths between bass and first violin and positioning the final notes in their "correct" octave. I just placed "correct" in scare quotes but it doesn't really need them -- the correct, simple, and proper voiceleading of all the parts above this harmony clearly demands that the first violin end in the fifth octave (its obligatory register, in other words).
In the next post I will examine the problem of the "descant" voice in Corelli, as presented in a book chapter by William Rothstein.
In the JMT article, note 32, I also mention Beethoven, String Quartet, op. 74, IV. A subsequent post will discuss that.
In the JMT article, note 32, I also mention Beethoven, String Quartet, op. 74, IV. A subsequent post will discuss that.
