Historical Manual of English Prosody. Saintsbury George
Its two laws.
18. This process of substitution is governed by two laws: one in a manner a priori, the other the result of experience only.
Confusion of base must be avoided.
19. Substitution must not take place in a batch of lines, or even (with rare exceptions) in a single line, to such an extent that the base of the metre can be mistaken.
(Of which the ear must judge.)
20. Even short of this result of confusion the ear must decide whether the substitution is allowable.
Certain substitutions are not eligible.
21. As a result of experience we find that the feet most suitable—if not alone suitable—as substitutes for the iamb—the commonest foot-staple—are the trochee, the anapæst, and the tribrach; that the dactyl substitutes well, if not too freely used, for the trochee.[27] These equivalences are reciprocal.
§ D. Pause
Variation of pause.
22. Next to equivalence, the most important and valuable engine in the constitution of English verses is the variation of the middle or internal pause.
Practically at discretion.
23. Except in very long lines—which always tend to pause themselves either at the middle or at two places more or less equidistant—there is no reason why the pause of an English line should not be at any syllable from the first to the penultimate, and none why it should or should not occur at the end of a line, couplet, or even stanza—though in the last-named case rather special reasons are required for its omission. Not every line need necessarily have any pause at all.
Blank verse specially dependent on pause.
24. The effect of blank verse depends more upon pause-variation than upon anything else; and by this variation, accompanied by stop or overrun ("enjambment") at the end of the line, verse-paragraphs are constituted, which can contain verse-clauses or sentences, in like manner brought into existence by pauses.
§ E. Line-Combination
Simple or complex.
25. Lines, composed as above of feet, can be used in English either continuously on the same or equivalent patterns, or in batches of two or more.
Rhymes necessary to couplet.
26. The batches of two almost necessarily require rhyme to indicate and isolate them, especially if the individual lines are of the same length. Other batches [stanzas] might, as far as any a priori objection goes, consist of unrhymed lines, symmetrically correspondent, or irregular [Pindaric].
Few instances of successful unrhymed stanza.
27. It is, however, found in practice, despite the examples of Campion, Collins, and one or two others, that rhymeless batching or stanza-making is very seldom successful. [28]
Unevenness of line in length.
28. There is neither a priori objection nor a posteriori inconvenience to be urged against the construction of stanzas or batches in lines of very uneven length.
Stanzas to be judged by the ear.
29. Every stanza-scheme must undergo, and is finally to be judged by, the test of the ear, and that only.
Origin of commonest line-combinations.
30. The commonest and oldest line-combinations—octosyllabic couplet, "common" or "ballad" measure, "long" and "short" measure, etc.—in some cases demonstrably, in all probably, result from the breaking up of the old long line ("fifteener" or "fourteener"), which itself came from the metricalising of the O.E. double stave.
§ F. Rhyme
Rhyme natural in English.
31. It is natural to English poetry—i.e. Middle and Modern English, or English poetry proper—to rhyme; and, except in the case of blank verse, no unrhymed measure for the last seven centuries has ever produced large quantities of uniformly satisfactory quality.
It must be "full,"
32. Rhyme in English must be "full," i.e. consonantal (on the vowel and following consonant or consonants), not merely assonantal (on the vowel only). Assonance by itself is insufficient.
and not identical.
33. It should not, according to modern usage, be identical—that is to say, the rhyming syllables should not consist of exactly the same vowels and consonants. But exceptions to this may be found in good poets, especially when the words are not the same.
General rule as to it.
34. Good rhyme has necessarily varied, at different times, with pronunciation; but a certain rough rule may be seen prevailing not uncommonly, that vowels in rhyme may take the value which they have in words other than those actually employed.[29]
Alliteration.
35. What is sometimes called "head-rhyme" (i.e. "alliteration") has now no place in English as rhyme at all, nor does it constitute either metre or stanza; but it is a permissible, and often a very considerable, ornament to verse.
Single, etc., rhyme.
36. Rhyme is either single (on the last syllable only), double (on the two last), or triple (on the three last). Beyond three the effect would be burlesque, and this is hard to keep out of triple rhyme, and even sometimes seems to menace the double.
Fullness of sound.
37. In serious poetry the fuller in sound the single rhyme is the better.
Internal rhyme permissible,
38. Rhyme is usually at the end of the line; but it may be "internal"; that is to say, syllables at one or even more than one place within the line may rhyme to the syllable at the end or to each other, and syllables within one line may rhyme to those at corresponding places within another.
but sometimes dangerous.
39. But this has a dangerous tendency to break the lines up.
§ G. Miscellaneous
Vowel-music.
40. The effect of English poetry at all times, but especially for the last hundred years, has been largely dependent on Vowel-music. This is by no means limited to the practice of what used to be called "making the sound suit the sense," though the two sometimes coincide. Vowel-music, not without occasional assistance from consonants, establishes a sort of accompaniment to the intelligible poetry—a prosodic setting.
"Fingering."
41.