we start out with trisyllabic rhythm, but have only two syllables in the second and in the third foot. It does not seem certain whether the missing syllable after dying is to have its place filled by a pause or by a prolongation of either or both of the syllables dy-ing — perhaps by all three means combined. In the same way the missing syllable after moon may have its place filled either by the prolongation of the oo, or by the pause indicated by the comma, or by both. But in other cases the pause occupies the entire syllable-moment; for examples, see under Pauses in pages 20-22 above. The whole matter was well summed up in Lanier's saying that "rhythm may be dependent on silences as well as on sounds.

Let us now try to gather what we have been considering into the form of definite statements regarding the place of the timeelement in our verse.

I. In the normal verse, accents appear at equal time-intervals. This, of course, does not preclude all manner of variations; the unit of measure is not the distance between the accents as they are found in each verse, but between the points where they belong in the typical metre.

2. There is a tendency toward the coincidence of long and accented, and of short and unaccented, syllables. This we have seen to be true in two different senses. In the first place, an accented syllable is likely to be lengthened for the same reason that it is accented - because of its relative importance in the place where it stands. In the second place, syllables noticeably long are avoided in those places in the verse where the accent does not fall, and are preferred where the stress is heavy.

3. In the reading of verse, the length of the syllables is varied artificially, so as to tend to preserve the equal time-intervals.

4. In like manner, pauses are introduced where syllables are short or wanting, to preserve these intervals.

It is quite possible that these laws might be stated more fully and definitely. In Anglo-Saxon verse the conditions were perhaps not so different from those of modern English as we are likely to think; there we know that the principal stresses of the

verse always fell on long syllables, and scholars like Sievers, by analyzing the remains of our early poetry, have formulated certain other laws as to the position and relations of the short syllables. If similar laws were to be formulated for our modern verse, we should probably find them no more perplexing than our ancestors would find those we have formulated for their verse. In every case the "law" is only an attempt to express what the ear has long known and obeyed. Mr. Goodell, in an article on "Quantity in English Verse," in the Proceedings of the American Philological Society for 1885, attempted to do for our verse what has just been suggested. He stated such laws as these:

"The thesis becomes a triseme if the next syllable bears the ictus. No syllable can be placed in this position which is incapable of prolongation."

"If the arsis is monosyllabic, a short vowel in the thesis followed by a single consonant is not lengthened by the ictus; the arsis is instead prolonged."

"With arsis monosyllabic, the strong tendency is to make the thesis short."

Perhaps these rules are on the right track; the terminology is somewhat difficult, and makes one hesitate to criticise carefully. But since, as we have seen, the terms "long" and "short," as applied to English syllables, have come to be so purely relative, since our syllabic quantities vary so much at the will of the reader, and since the whole matter of the reading of our verse is in good measure one of subjective interpretation, it seems very doubtful whether any statements more explicit than those already laid down would be found of practical service.

Finally, we come back to the question whether we shall use for English verse the classical terminology which has for so long been applied to it. Those who object to such terminology do so either on the ground that it implies that English accented and unaccented syllables are equivalent respectively to Latin long and short syllables, or on the still more fundamental ground that there is nothing in our verse which can properly be called a "foot." It is undoubtedly true that the use of terms

based on quantity has given rise to some confusion when applied to phenomena based on accent, yet the terms are now understood with as fair a degree of clearness as any terms relating to so disputed a subject as English verse; and it seems very doubtful whether it is not easier to explain them than to introduce new ones. Experiments in the latter direction have not been very successful. The latest writer on the subject objects, with considerable severity, to the classical nomenclature "hardly pressed and barbarously misapplied." Our current prosody, he says later, "ignores" the frequent occurrence of an accented syllable at the beginning of a line of Shakspere's verse, "turning it off with the statement that a trochaic foot may begin an iambic verse."" Yet when we reach the summary of the author's discussion of the subject, we find the same phenomenon "turned off" with this statement: "In rising rhythm a thought-moment may begin with a falling wave-group." One cannot avoid querying whether this interesting combination of words conveys any simpler and better idea to the normal English reader than the familiar statement that "a trochaic foot may begin an iambic verse." The case is instructive as to the danger of attempting a new terminology where one is already established, and of imagining that one has thereby made the discussion of the subject more scientific.

The second of the objections to the usual terminology, that there is no real foot in English verse, has already been considered. If there are no regular units of measure in our verse, then to attempt constantly to find such units, and to use terms that imply their existence, is certainly a mistake. But those are on the wrong track who would find the divisions of the verse in the natural phrase-divisions of English speech.*

In " arma

*This seems to be a part of the old effort to seek a grammatical rather than a musical origin for metre. On this subject the reader should see the brilliant discussion of Professor Gummere in The Beginnings of Poetry, from which a few paragraphs are quoted in Part Four.

virumque cano" the syllable vi- is far more closely connected with the syllable -rum, for all prose purposes, than with the preceding syllables; but in the verse the Romans thought of it as being in the same foot with arma; and later in the verse the last syllable of cano is rhythmically connected (over the barrier of a comma) with the first of Troja. Indeed, the Latin poets instinctively avoided the regular coincidence of metrical units with word or sentence units. Precisely the same thing is true of English verse. It has been suggested more than once that the great preponderance, among English dissyllables, of those accented on the first syllable, goes to explain our preference for iambic over trochaic measures; and that one reason why the rhythm of Hiawatha, for example, so soon wearies the ear, is because its metrical divisions and word divisions so frequently coincide. The fundamental principle of verse is that it sets up a new order of progress which constantly conflicts with, yet without destroying, the order of progress of common prose speech.

So the foot means, not a unit of measure for the words, but for the syllables viewed as rhythmical sound; and the attempt has already been made to show that it represents the timeinterval between the regularly recurring accents of the normal metre. When there are two syllables in the interval, it is convenient to call the foot an iambus, a trochee, a pyrrhic, or a spondee; when there are three, it is convenient to call the foot an anapest or a dactyl. According to this system, the number of feet in the metre will always depend on the number of regularly recurring accents, which of course is not the case in classical prosody. For the same reason, all exceptional feet can be named by one of the six terms indicated, except where (as in Swinburne's "Choriambics") some classical metre is deliberately imitated. There is no sufficient reason for speaking of the choriambus as occurring in Shakspere's verse, because where four syllables occur in such succession as to form a sort of choriambus, they will be found to fill the place of two ordinary feet, not of one; hence it would be irrational to combine them into one exceptional foot. But on this matter of convenience in the

terminology of verse, one cannot do better than to refer the reader to Mr. Mayor's Chapters on English Metre, where a refreshingly simple system is set forth, such as will not break down under any reasonable test.

There is one defect, it may be freely admitted, in these classical names of feet. They provide no place for the secondary accent. A foot made up of a fully accented plus a slightly accented syllable must be called either a spondee (the second syllable being thought of as approaching the stress of the first) or a trochee (the syllable being thought of as approaching no stress). The abundant use of secondary or compromised accents - and one might say, too, of secondary or compromised quantities is a Germanic characteristic, for which no classical terminology can provide. There is, theoretically, room for some new names of feet recognizing these ambiguous syllables. Yet since degrees of accent are purely relative, and no two readers would be sure of agreeing as to which syllables are fully stressed, and which are half-stressed, it is not likely that such additional terms would make our terminology any more exact for practical purposes. The present system does, in fact, represent a characteristic feature of modern English as distinguished from early English verse; namely, that our metres strive after a regular alternation of stress and no-stress, and that the ear imagines this alternation even where (if it were a matter of prose utterance) it can scarcely be said to exist.

It would be absurd to strive with any warmth for the classical system of terminology in English prosody. It is undoubtedly not an ideal system, nor such a one as we should adopt if we were naming everything anew; few existing terminologies are. The only object of the present defence of its carefully limited use is to show that it does stand for some fundamental facts in our verse, and to suggest that it is usually wiser to make the best of the vocabulary we have than to fly to one we know not of. The important thing, in any case, is not the question of terms, but the end that we should not lose hold of the musical rhythms of our verse, made up of delicately adjusted elements of accent and time.