Answering the Subject: Real and Tonal

In a concerto, the ritornello waits for an episode before it comes back in a new key. In a fugue, the subject comes back at once, in a second voice: the answer (see From Concerto to Fugue).

① and ⑤. A subject starts on ① or on ⑤ (of 27 subjects in the Well-Tempered Clavier, 18 start on ① and 8 on ⑤), and the answer starts on the other one. The later entries use both, but not always in turn: the C major fugue of WTC I goes ① ⑤ ⑤ ①, and so does Handel’s first exercise.

The question on this page: is the answer an exact copy of the subject, and in which key? Most often the copy is in the key of ⑤, but it can be in the key of ①, or of ④ (BWV 872, below).

Often, yes. That is a real answer. Sometimes some notes near the start have to change. That is a tonal answer. Every example on this page is printed from a score (Bach’s, and Kirchhoff’s partimento), and every count comes from checking 27 fugues of the Well-Tempered Clavier one by one.

How to read the examples: the circled numbers are scale degrees; ♯ or ♭ in front of one means the note is raised or lowered. A red note in the answer is one that does not match the exact copy. Under the answer, the first notes are numbered in the home key, so you can see ① and ⑤ trade places; from “new key →” on, the answer is numbered in the key it is copied into, and its numbers are the subject’s own. The menu under each example switches to all home key, or all in the answer’s key.

Try a real answer first

Handel’s first fugue exercise works the same way: his subject on G (G held, A–G, F, B♭, A) is answered on D with every note a fifth higher (D held, E–D, C, F, E).

When the answer changes: what Bach does

Look at how the subject starts. In 27 fugues of the Well-Tempered Clavier:

The subject…Tonal answerReal answer
starts on ⑤80
leaps between ① and ⑤ in its first few notes (or runs up ① ③ ⑤)62
does neither110

Ebenezer Prout drew the same rule from Bach and Handel in 1891 (Fugue, §58):

Every subject in which there is no modulation to the dominant, either expressed or implied, may have a real answer, excepting, first, when it begins on the tonic and leaps to the dominant either direct or with the third of the scale as an intermediate note; and secondly, when it begins on the dominant.

In those cases, he says, “the tonic must be answered by the dominant, and the dominant by the tonic” (§86), and this “applies only to the beginning of the subject” (§90). In our numbers: at the start, ① answers ⑤ and ⑤ answers ①; the rest is copied exactly.

Why: two ways to hear it

The chord under the answer. When the answer comes in, the first voice is usually on ① or ③ (in 20 of the 27 fugues), so the answer starts over the tonic chord. Play the exact copy against that: its first notes belong to ⑤’s chord. The tonal answer fits.

The boundaries of the key. Johann Mattheson gave a three-part fugue rule in Der vollkommene Capellmeister (1739):

Do not overstep the boundaries of the key, below or above; do not begin the answer on a sound that goes against the key; otherwise make the intervals of the transposition as alike as can be done in good taste.

He ranks them: the first is the most important (though it allows exceptions), the second is necessary in regular fugues, and the third gives way most. The boundaries are ① and ⑤. The octave of a key splits at ⑤ into a fifth (① up to ⑤) and a fourth (⑤ up to ①). A subject that fills the fifth is answered in the fourth, and the other way round. That is why ⑤ answers ① and ① answers ⑤.

What it costs

Mattheson’s third part gives way: where the changed notes meet the copied ones, one interval changes size. A fifth becomes a fourth (A♭ major, above). A step becomes a third (G minor). Here a step becomes a repeated note:

In a long subject the change can be a single note:

When the new key is a fourth up: BWV 872 and Kirchhoff

Bach’s Fugue in C♯ major from Book II (BWV 872) is the fugue we will study on November 30. Its answer starts with ① ③ ① copied a fifth higher, swaps ⑤ for ①, and then goes on a fourth higher, not a fifth: there is a B♮ in it. Read in the key of F♯, the rest of the answer has exactly the subject’s numbers, ⑤ ④ ③ ② ①.

Your Kirchhoff partimento follows the same plan, which is one reason we use it:

Look at the fourth note after “new key”: D♯, marked ♯④. Copied exactly into A, it would be D♮. Kirchhoff writes both. The first answer, above, has D♯; when the bass answers in bar 6, it has D♮. Both readings are in the 1734 Amsterdam print, so neither is a misprint we can fix:

With D♮ the answer copies the subject exactly, a fourth up, just as BWV 872 does with its B♮. With D♯ it keeps E major’s leading tone and does not lean toward A. Play both answers. Which one would you write in your realization, and why?

Why a fourth? In both, the end of the subject’s head moves onto ⑤ and its chord (in BWV 872, G♯ then F♯; in Kirchhoff, B–F♯–G♯–A–G♯, mostly B, F♯ and A). The answer puts the same figure over ①. Prout states the principle (1891, §80): dominant harmony in the subject is answered by tonic harmony. It also keeps the answer inside Mattheson’s boundaries: Kirchhoff’s answer stays between B and E, where the exact copy would climb to F♯.

Bach changes four notes at the start of the B major fugue in Book I, the one subject in our table that neither starts on ⑤ nor leaps between ① and ⑤ and still gets a tonal answer:

Step by step
  1. Write the circled degrees under the subject.
  2. Does it start on ⑤, or leap from ① to ⑤? If not, copy it exactly, starting on ⑤.
  3. If it does: at the start, ① for ⑤ and ⑤ for ①. Copy the rest exactly.
  4. Play subject and answer one after the other.

For a long or decorated subject, use what you know from diminution: find the framework notes, answer those, then decorate the answer the way the subject is decorated. William Renwick proposes this method for fugal answers: reduce the subject, answer the reduction, then embellish (1991, p. 207).

A guide, not a law

The table has exceptions both ways. Two subjects that leap between ① and ⑤ get real answers (WTC I in E minor, which runs up ① ③ ⑤, and WTC II in F♯ major), and B major gets a tonal one without the leap. Prout himself adds that “even in these two cases a real answer is always possible” (§58). Mattheson says his first rule allows exceptions, and Prout found many real answers to subjects that run up the tonic chord, especially in Handel. When in doubt, try both and choose the one that sounds like the subject and stays in the key.

Prout’s third case, a subject that ends in the key of ⑤, needs an answer that brings the music back home: the part of the subject in the key of ⑤ is answered in the home key, turning back at the same point where the subject turned away (1891, §§116–117).

Practice

Write an answer to each subject on paper or in MuseScore. Then tick “Show Bach’s answer” and compare.

More, with listening examples (in German)

Fuge I – Themenbeantwortungen on Open Music Academy: a slider shows each answer change on the score, with recordings. It walks through Mattheson’s rule, the modes behind it, and two places where the rule is broken on purpose: a Frescobaldi canzona and Bach’s G minor fugue for solo violin.

Sources
  • Mattheson, Johann. Der vollkommene Capellmeister. Hamburg, 1739. The rule as quoted in “Fuge I – Themenbeantwortungen,” Open Music Academy (Ulrich Kaiser, Ilka Mestemacher, Theresa Ulbricht; CC BY-SA 4.0). Translation ours.
  • Prout, Ebenezer. Fugue. London: Augener, 1891. Chapters 3–4, “The Answer.”
  • Renwick, William. “Structural Patterns in Fugue Subjects and Fugal Expositions.” Music Theory Spectrum 13, no. 2 (1991): 197–218.
  • Scores and counts: J. S. Bach, The Well-Tempered Clavier, Humdrum edition, 27 fugues (Book I nos. 2, 4, 5, 8–11, 13, 14, 16, 17, 21–23; Book II nos. 2, 4, 7, 9, 11, 13, 14, 17–19, 22–24). G. Kirchhoff, Fugue No. 5, from the course file.

Next: the countersubject, and the Kirchhoff workshop.

Leave a Reply

Your email address will not be published. Required fields are marked *