(1) from simply inserting an additional cadence, or from omitting one. Or

(2) it may be the consequence of some specific manipulation of the phrase-melody with a view to its extension or expansion, its development into a broader and more exhaustive exposition of its contents.

THE SMALL AND LARGE PHRASES.--If a cadence is inserted before it is properly due, it is almost certain to occur exactly _half-way_ along the line toward the expected (regular) cadence,--that is, in the _second_ measure. This is likely to be the case only when the tempo is so slow, or the measures of so large a denomination, that two of them are practically equal to four _ordinary_ measures. By way of distinction, such a two-measure phrase is called a Small phrase. For example:--

[Ill.u.s.tration: Example 37. Fragment of Mendelssohn.]

There is no reasonable doubt of the semicadence in the second measure, because enough pulses have been heard, up to that point, to represent the sum of an ordinary phrase. If this were written in 6-8 measure (as it might be), it would contain four measures. See, also, Song No. 22 of Mendelssohn,--9-8 measure, adagio tempo; the phrases are "Small"; note particularly the last two measures. The same is true in No. 17. About Schumann, op. 68, No. 43 (_Sylvesterlied_), there may be some doubt; but the measures, though of common denomination, contain so many tones, in moderate tempo, that the effect of a cadence is fairly complete in the second measure.

If, on the other hand, one of the regular cadences is omitted,--owing to the rapidity of the tempo, or a small denomination of measure,--the phrase will attain just double the ordinary length; that is, _eight_ measures. An eight-measure phrase is called a Large phrase. For ill.u.s.tration:--

[Ill.u.s.tration: Example 38. Fragment of Beethoven.]

There is not the slightest evidence of repose or interruption in the fourth measure, nor of a new beginning in the fifth, wherefore the cadence is not expected until four more measures have pa.s.sed by. The inferior points of repose in the upper parts, at the beginning of the 5th, 6th and 7th measures, serve only to establish melodic, or rather rhythmic, variety, and have no cadential force whatever. See Mendelssohn, Song No. 8; the first cadence appears to stand in the _eighth_ measure; the tempo is rapid and the measures are small; it is obviously a large phrase. The phrase which follows is regular, however; there _is_ a cadence in the twelfth measure, thus proving that Large phrases may appear in company with regular phrases, in the same composition. In other words, the omission of an expected cadence (or the insertion of an additional one) may be an _occasional_ occurrence,--not necessarily constant. See, again, No. 22 of the Songs Without Words; the first and second phrases are small; the third phrase, however (reaching from measure 6 to 9 without cadential interruption), is of regular dimensions.

THE PRINCIPLE OF EXTENSION.--The other cause of modified phrase-dimension is one of extreme importance, as touching upon the most vital process in musical composition, namely, that of _phrase-development_.

Setting aside all critical discussion with reference to the question, "What is good music?" and simply accepting those types of cla.s.sic composition universally acknowledged to be the best, as a defensible standard (to say the least), we find that such a page of music exhibits the pursuit of some leading thought (melodic motive or phrase), with precisely the same coherence and consistency, the same evidence of determined aim, as is displayed in the creation of a forcible essay, a masterly poem, an imposing architectural plan, or any other work of art that betrays intelligence and a definite, fixed, purpose. This is no more nor less than might be expected from the dominion of the law of Unity.

The equally inflexible demands of Variety are satisfied by presenting this self-same leading thought in ever new and changing aspects,--_not_ by exchanging the thought itself for a new one at each successive angle.

This latter faulty process would naturally lead to a conglomeration of impressions, baffling comprehension and jeopardizing real enjoyment.

In a cla.s.sic page of music we perceive that each successive unit grows, more or less directly, out of those which go before; not so directly, or with such narrow insistence as to produce the impression of sameness and monotony, but with such consistency of design as to impart a unified physiognomy to the whole. Hence, it will often be found that every melodic figure, during a certain section (if not the whole) of a composition, may be traced to one or another of the figures which characterized the first phrase, or the first two or three phrases, of the piece. This was emphasized by our reference, near the end of the first chapter, to the 8th Song Without Words of Mendelssohn. If the student, in a.n.a.lyzing the melody of that composition, will endeavor to penetrate some of the clever disguises employed by the composer (for the sake of Variety), he will find the whole piece reducible to a very few melodic figures, announced at or near the beginning. See also No. 45 (C major), No. 36, No. 26. Also Schumann, op. 68, No. 7, No. 8, No. 18, No. 23.

Also Beethoven, pianoforte sonata, op. 10, No. 2, last movement; op. 26, last movement.

In musical composition this process is known as thematic development, and it generally extends over the whole, or a greater part, of the piece.

Its operation on a smaller scale, with more limited reference to one phrase alone, effects the development of the phrase _by extension_.

The process of extension or expansion, by means of which the phrase usually a.s.sumes a somewhat irregular length, consists mainly in the varied repet.i.tion of the figures or motives that it contains; and the continuity of the whole, as extension of the _one phrase_, is maintained by suppressing the cadence--suspending all cadential interruption--during the lengthening process. For example:

[Ill.u.s.tration: Example 39. Fragment of Mendelssohn.]

These six measures result from a repet.i.tion (variated) of the third and fourth measures of the original--regular--four-measure phrase. A cadence is due in the fourth measure, but it is not permitted to a.s.sert itself; and if it did, its cadential force would be neutralized by the entirely obvious return to (repet.i.tion of) the motive just heard. Further:--

[Ill.u.s.tration: Example 40. Fragment of Mendelssohn.]

There is no cadence in the fourth measure,--the current of the melody obliterates it and hurries on, voicing the last measure again and again until it dies away in the tenth measure, where a cadence ends it. That it should be the _tenth_ measure is purely accidental; the number of measures is of little account in the act of extension; here, it was continued until a convenient place was found (with reference to chord and key) for the cadence. Further:--

[Ill.u.s.tration: Example 41. Fragment of Mendelssohn.]

Measures 1, 2, 3 and 8 const.i.tute the original regular four-measure phrase.

The following regular phrase (to be found in the last movement of Beethoven's pianoforte sonata, op. 28):--

[Ill.u.s.tration: Example 42. Fragment of Beethoven.]

is immediately followed by this lengthy and elaborate extension:--

[Ill.u.s.tration: Example 43. Fragment of Beethoven.]

[Ill.u.s.tration: Example 43 continued.]

The portion marked _b_ is a complete repet.i.tion, with quaint variation, of the original four-measure phrase, marked _a_ in Ex. 42; _c_ is a repet.i.tion of the last figure (just one measure) of the phrase, with the melodic parts inverted, or exchanged; _d_ and _e_ are a literal repet.i.tion of the two preceding measures--(_c_) and _c_; _f_ is another recurrence of (_c_), with still another inversion of the melodies; _g_ repeats _e_ an octave higher; and _h_ is nothing more or less than a curious repet.i.tion of _g_, in longer tones, and in reversed direction.

Distinct cadential interruption is carefully avoided after the original phrase has been announced, that is, throughout Ex. 43,--which is the significant proof (borne out by the manifest ident.i.ty of the _melodic_ members) that these measures form part and parcel of the original phrase, as extension or development of it, and _not_ a new phrase. The total length is sixteen measures, developed thus out of the original four.

For an exhaustive explanation of phrase-extension, with all the technical details, the student is referred to my h.o.m.oPHONIC FORMS, Chapter III.

Another method of extending a phrase consists in prefacing a measure or two of purely _introductory_ material; it is, therefore, rather antic.i.p.ation than prolongation, and is composed most commonly of the figure of the accompaniment, announced briefly before the actual phrase-melody begins.

This is shown very clearly in the first measure of the 22d Song Without Words; also in the first measure of No. 7, No. 31, No. 42, No. 40, and others; the first _two_ measures of No. 34, and No. 1; the first _three_ measures of No. 19, No. 26, and No. 37,--and needs no further ill.u.s.tration. It emphasizes the necessity of vigilance in defining the correct _starting-point_ of the first phrase; for a mistake at the beginning may interfere seriously with the locating of the cadences (according to our fundamental four-measure rule). For instance, in No.

42 the cadences do _not_ fall in the 4th, 8th, 12th measures--and so on--but in the 5th, 9th, 13th, 17th, from the very beginning of the piece.

When the introductory pa.s.sage is longer than _three_ measures, it probably const.i.tutes a complete phrase by itself, with its own cadence; in which case, of course, it must not be a.n.a.lyzed as "extension." For example, at the beginning of No. 29; still more apparently at the beginning of No. 28, No. 41, and others.

INHERENT IRREGULARITY.--Finally,--there exists another, third, condition, besides those mentioned at the head of this chapter, whereby a phrase may a.s.sume an irregular dimension; not by doubling or dividing its length (as in the large and small phrases) nor by the processes of extension,--but by an arbitrary and apparently incalculable act of _melodic liberty_,--by allowing the melody to choose its own time for the cadential interruption. This comparatively rare occurrence is ill.u.s.trated in Ex.

17, No. 1 (five-measure phrase), and Ex. 17, No. 2, second phrase (six measures long). It is true that in each of these cases the "extra"

measures might be accounted for as "extension by modified repet.i.tion,"--for instance, in No. 1 the _second_ measure might be called a reproduction (or extension) of the first measure. But cases will be encountered where a phrase of three, five, six, or seven measures will admit of no such a.n.a.lysis. In such instances the student is compelled to rely simply upon the evidence of _the cadence_. As was advised in the context of Ex. 17, he must endeavor to define the phrase by recognition of its "beginning" and "ending," as such; or by exercising his judgment of the "cadential impression." See also Ex. 48, second phrase (six measures).

See Schubert, pianoforte sonata No. 1 (A minor, op. 42) _Scherzo_-movement; first 28 measures, divided into 5 phrases,--as demonstrated by the melodic formation--of 5, 5, 5, 7 and 6 measures.

Also Schubert, _Impromptu_, op. 90, No. 3, measures 42 to 55 (phrases of 5, 5 and 4 measures.)

LESSON 6. a.n.a.lyze the following examples, locating the cadences and defining their value (as perfect or semicadence); and determining the nature of each irregular phrase (as small, large, or extended phrase):

Beethoven, pianoforte sonata, op. 22, second movement (_Adagio_), first 30 measures.

Beethoven, pianoforte sonata, op. 28, _Scherzo_-movement.

Beethoven, pianoforte sonata, op. 14, No. 3, _Menuetto_.

Mendelssohn, Songs Without Words: No. 4, first 5 measures.

No. 46, last 9 1/2 measures.

No. 42, last 15 measures.

No. 45, last 11 measures.

No. 12, last 12 measures.