Arnold Schoenberg · section 36 of 41

Guidelines

I. Secondary dominants can stand wherever the diatonic degree can stand, provided that the root progressions permit it. In keeping with their meaning, however, they are best where there is a root progression similar to the dominant function—that is, on the pattern V–I, V–IV and V–VI. For their purpose is above all to support this tendency toward a dominant-like step by means of the artificial leading tone. The root progression a third down is also often quite usable, though without achieving any significant enrichment. By the root progression a fourth down, chords are likewise strung together that exercise hardly any other than a passing function (that is, melodic, in the sense of a change of position, an exchange). The root progression a third up is almost entirely pointless and usually cannot be applied either, because in part it leads to chords not yet explained, a reason that also rules out some of the successions named before; as the example shows, where the seventh chords are treated at once (see 137A).

II. Secondary-dominant seventh chords can stand wherever secondary dominants are permissible and the treatment of the dissonance can be carried out. In general they achieve the purpose aimed at by secondary-dominant triads even better, because the seventh as a rule promotes the dominant-like steps by giving direction.

III. Artificial augmented triads are to be used on the pattern of the natural augmented triad (III in minor). This can, to be sure, precede and follow all degrees; its most important functions, however, are III–VI, III–I, III–IV and III–II, the last of which, because of the complicated voice leading, is seldom usable. Naturally the artificial ones can be followed by the same chords as the natural one: secondary dominants, artificial minor triads, artificially diminished triads, seventh chords with artificially diminished fifth, and artificial diminished seventh chords. Their main purpose, however, is to give direction to a connection by means of the artificial leading tone (see 137B).

IV. The artificial minor triad on the V degree will, as the examples show, be especially suited to bringing about the subdominant region (IV and II degrees). That it also introduces more neutral ones well—the unraised VI and the raised III—is shown, among other things, in 137C under a) by the case V–II etc., under c) by the case V–VII. The most usual forms, however, will be obtained if, in the case of its V–I and V–IV progressions, one understands it as though it were II of the subdominant key (F major), and in the case of the V–VI progression as though it were IV of a minor key (D minor) (see 137C under a, and corresponding things in the following examples).

V. Artificially diminished triads can be treated like the VII in major or, what is usually better, like the II in minor (almost exclusively as six-chords [137D]). For almost all purposes, however, preferable to them are those arising through the addition of the seventh:

VI. Seventh chords with diminished fifth [are to be preferred], because for the more modern ear the seventh still gives more direction than the diminished fifth. The one on the III degree, which stands close to the subdominant region (137E 1), fits most easily into the cadence. Its most usual forms are those in which it introduces the II degree (III–II, III–VI–II), where it is treated similarly to a II of minor (D minor). It is less convincing for introducing IV, because there it is used like a VII in major (VII–I!!). On I and V those with a minor seventh are for the time being unusable for us, while the other two (E 2b and 4b) are diminished seventh chords, about which more will be said in what follows. The one on the IV degree recalls a II or VI in minor (E or A minor) and a VII in major (G major). It fits best into the cadence when it is treated on the pattern of a VI from (A) minor, whereas on the pattern of a II in (E) minor it emphasizes the upper dominant too much, and on the pattern of a VII in major it is less convincing for reasons already mentioned. (VII–I!)

Following on from this, the new means give rise to two further remarks about chords already used earlier.

I. As the dominant in the cadence (that is, as the penultimate chord) we shall use only a major triad of the V degree (thus not the augmented one) whose major third does not arise through chromatic raising. A lowered seventh tone occurring before it (in C major, for example, B-flat) would, as a subdominant characteristic, first have to be resolved downward in the manner of a sixth tone in minor. Other things are indeed sometimes found in the literature, but we have no occasion to make use of them.

II. For good reasons we shall now impose on ourselves some restrictions in the use of certain diatonic secondary seventh chords. In the course of what follows the student will increasingly recognize that every chord, in keeping with its various tendencies, can have manifold functions; that it is therefore not unambiguous, and that its meaning is established only by its surroundings. If I now restrict the use of the secondary seventh chords to some extent, this springs from the endeavour to use them in the sense of their most important tendencies. For this one can set up the following principle: every chord will (if its surroundings do not prevent it) demand the same continuation as the chord (consisting of entirely different notes) to which it is equal in its intervals. Thus the seventh chords of the II, III and VI degrees in major are of exactly the same construction. Now since that of the II degree, as is well known, has a definite function (II–V–I, II–I⁶₄–V), the ear expects the same continuation from the identically built ones on III and VI (137F a and b), and there one sees at first glance that this continuation can no longer easily be referred to the key. This is by no means to say that only this continuation is possible, and as proof I myself give some examples from which it appears that the fate of the tonality is by no means sealed thereby. But in any case this use is less characteristic than the one after the stereotype of the II. And so we see that in example c) the secondary dominant on VI would have worked better in place of the III, in d) the one on I or III, in e) the one on II, while in f) there is the danger of a modulation to G major—which, to be sure, one can easily master. Since, then, these chords are questionable in their most characteristic successions and indecisive in the others, I recommend—and this also corresponds to their occurrence in the literature—the greatest economy. They will most likely still be of a certain (melodic) value as passing phenomena (passing seventh). Otherwise, however, depending on the direction in question, they are best replaced by the artificial chords on the same degrees that we have just become acquainted with: secondary-dominant seventh chords, artificial diminished ones, etc. [compare c) with g), d) with h)]. Of the remaining secondary seventh chords, that of the I and that of the IV are identically built. They too are best used in passing. But clearly, in almost all cases, the secondary-dominant seventh chord of I will have a more definite effect than the diatonic seventh chord [compare i) with k)]. The secondary seventh chord of the VII degree recalls that of the II in minor too much for one to like using it otherwise than as that chord (m and n). The student is easily able to apply exactly the same principles to the secondary seventh chords in minor, and accordingly he will there, too, restrict himself in the use of these chords to what is characteristic and suitably replace what is uncharacteristic.