LIST OF EDITIONS REFERRED TO

The quotations of Old English poetry are taken from Grein-Wülker, Bibliothek der Angelsächsischen Poesie, Strassburg, 1883–94. For the Middle English poets the editions used have been specified in the text. Most of the poets of the Modern English period down to the eighteenth century are quoted from the collection of R. Anderson, The Works of the British Poets, Edinburgh, 1795 (15 vols.), which is cited (under the title Poets) by volume and page. The remaining Modern English poets are quoted (except when some other edition is specified) from the editions mentioned in the following list.

ERRATA

BOOK I. THE LINE

PART I. THE NATIVE METRE

CHAPTER I
GENERAL INTRODUCTION TO THE SCIENCE OF
METRE AND THE STRUCTURE OF VERSE

§ 1. The study of English Metre is an integral part of English Philology. It is indispensable to the investigator of the history of the language, since it supplies sometimes the only (or at all events the surest) means of restoring the older pronunciation of word-stems, and of inflexional terminations. In many cases, indeed, the very existence of such terminations can be proved only by the ascertained requirements of metre. As an aid to the study of English literature in its aesthetic aspects the science of metre is no less important. It exhibits the gradual development of the artistic forms of poetical composition, explains the conditions under which they took their rise, and by formulating the laws of their structure affords valuable help in the textual criticism of poems which have been transmitted in a corrupt or imperfect condition.

§ 2. The object of the science of metre is to describe and analyse the various rhythmical forms of speech that are characteristic of poetry in contradistinction to prose.

Poetry is one of the fine arts, and the fine arts admit of a division into Plastic and Rhythmic; the Plastic arts comprehending Sculpture, Architecture, and Painting, the Rhythmic arts, on the other hand, comprehending Dancing, Music, and Poetry. The chief points of difference between these classes are as follows. In the first place, the productions of the Plastic arts can be enjoyed by the beholder directly on their completion by the artist without the interposition of any help, while those of the Rhythmic arts demand, after the original creative artist has done his work, the services of a second or executive artist, who is usually termed the performer, in order that these productions may be fully enjoyed by the spectator or hearer. A piece of music requires a singer or player, a pantomime a dancer, and poetry a reciter or actor. In early times the function of executive artist was commonly discharged by the creative artist himself. In the second place, the Plastic arts have no concern with the relations of time; a work of painting or sculpture presents to the beholder an unchanging object or represents a single moment of action. The Rhythmic arts, on the other hand, are, in their very essence, connected with temporal succession. Dancing implies a succession of movements of the human body, Music a succession of inarticulate sounds, Poetry a succession of articulate sounds or words and syllables. The Plastic arts, therefore, may be called the arts of space and rest, and the Rhythmic arts the arts of time and movement. In this definition, it must be remembered, the intrinsic quality of the movements in each of these rhythmical arts is left out of account; in the case of poetry, for instance, it does not take into consideration the choice and position of the words, nor the thought expressed by them; it is restricted to the external characteristic which these arts have in common.

§ 3. This common characteristic, however, requires to be defined somewhat more precisely. It is not merely succession of movements, but succession of different kinds of movement in a definite and recurring order. In the dance, the measure, or succession and alternation of quick and slow movements in regular and fixed order, is the essential point. This is also the foundation of music and poetry. But another elementary principle enters into these two arts. They are not founded, as dancing is, upon mere silent movements, but on movements of audible sounds, whether inarticulate, as in music, or articulate, as in poetry. These sounds are not all on a level in respect of their audibility, but vary in intensity: broadly speaking, they may be said to be either loud or soft. There is, it is true, something analogous to this in the movements of the dance; the steps differ in degree of intensity or force. Dancing indeed may be looked upon as the typical form and source of all rhythmic movement. Scherer brings this point out very well.[1] He says: ‘Rhythm is produced by regular movements of the body. Walking becomes dancing by a definite relation of the steps to one another—of long and short in time or fast and slow in motion. A regular rhythm has never been reached by races among which irregular jumping, instead of walking, has been the original form of the dance. Each pair of steps forms a unity, and a repetition begins with the third step. This unity is the bar or measure. The physical difference between the comparative strength of the right foot and the weakness of the left foot is the origin of the distinction between elevation and depression, i.e. between relatively loud and soft, the “good” and the “bad” part of the measure.’

Westphal[2] gives a similar explanation: ‘That the stamp of the foot or the clap of the hands in beating time coincides with the strong part of the measure, and the raising of the foot or hand coincides with the weak part of it, originates, without doubt, in the ancient orchestic.’ At the strong part of the bar the dancer puts his foot to the ground and raises it at the weak part. This is the meaning and original Greek usage of the terms ‘arsis’ and ‘thesis’, which are nowadays used in an exactly opposite sense. Arsis in its ancient signification meant the raising of the foot or hand, to indicate the weak part of the measure; thesis was the putting down of the foot, or the stamp, to mark the strong part of the measure. Now, however, it is almost the universal custom to use arsis to indicate the syllable uttered with a raised or loud voice, and thesis to indicate the syllable uttered with lower or soft voice. From the practice of beating time the term ictus is also borrowed; it is commonly used to designate the increase of voice which occurs at the strong, or so-called rhythmical accent.

All rhythm therefore in our dancing, poetry, and music, comes to us from ancient times, and is of the same nature in these three arts: it is regular order in the succession of different kinds of motion.

§ 4. The distinction between prose and poetry in their external aspects may be stated thus: in prose the words follow each other in an order determined entirely, or almost entirely, by the sense, while in poetry the order is largely determined by fixed and regular rhythmic schemes.

Even in prose a certain influence of rhythmical order may be sometimes observable, and where this is marked we have what is called rhythmical or artistic prose. But in such prose the rhythmic order must be so loosely constructed that it does not at once obtrude itself on the ear, or recur regularly as it does in poetry. Wherever we have intelligible words following each other in groups marked by a rhythmical order which is at once recognizable as intentionally chosen with a view to symmetry, there we may be said to have poetry, at least on its formal side. Poetical rhythm may accordingly be defined as a special symmetry, easily recognizable as such, in the succession of syllables of differing phonetic quality, which convey a sense, and are so arranged as to be uttered in divisions of time which are symmetrical in their relation to one another.

§ 5. At this point we have to note that there are two kinds of phonetic difference between syllables, either of which may serve as a foundation for rhythm. In the first place, syllables differ in respect of their quantity; they are either ‘long’ or ‘short’, according to the length of time required to pronounce them. In the second place, they differ in respect of the greater or less degree of force or stress with which they are uttered; or, as it is commonly expressed, in respect of their accent.

All the poetic rhythms of the Indogermanic or Aryan languages are based on one or other of these phonetic qualities of syllables, one group observing mainly the quantitative, and the other the accentual principle. Sanskrit, Greek, and Roman poetry is regulated by the principle of the quantity of the syllable, while the Teutonic nations follow the principle of stress or accent.[3] With the Greeks, Romans, and Hindoos the natural quantity of the syllables is made the basis of the rhythmic measures, the rhythmical ictus being fixed without regard to the word-accent. Among the Teutonic nations, on the other hand, the rhythmical ictus coincides normally with the word-accent, and the order in which long and short syllables succeed each other is (with certain exceptions in the early stages of the language) left to be determined by the poet’s sense of harmony or euphony.

§ 6. Before going further it will be well to define exactly the meaning of the word accent, and to give an account of its different uses. Accent is generally defined as ‘the stronger emphasis put on a syllable, the stress laid on it’, or, as Sweet[4] puts it, ‘the comparative force with which the separate syllables of a sound-group are pronounced.’ According to Brücke[5] it is produced by increasing the pressure of the breath. The stronger the pressure with which the air passes from the lungs through the glottis, the louder will be the tone of voice, the louder will be the sound of the consonants which the stream of air produces in the cavity of the mouth. This increase of tone and sound is what is called ‘accent’. Brücke seems to use tone and sound as almost synonymous, but in metric we must distinguish between them. Sound (sonus) is the more general, tone (τόνος) the more specific expression. Sound, in this general sense, may have a stronger or weaker tone. This strengthening of the tone is usually, not invariably, accompanied by a rise in the pitch of the voice, just as the weakening of the tone is accompanied by a lowering of the pitch. In the Teutonic languages these variations of stress or accent serve to bring into prominence the relative importance logically of the various syllables of which words are composed. As an almost invariable rule, the accent falls in these languages on the root-syllable, which determines the sense of the word, and not on the formative elements which modify that sense. This accent is an expiratory or stress accent.

It must be noted that we cannot, using the term in this sense, speak of the accent of a monosyllabic word when isolated, but only of its sound; nor can we use the word accent with reference to two or more syllables in juxtaposition, when they are all uttered with precisely the same force of voice. The term is significant only in relation to a variation in the audible stress with which the different syllables of a word or a sentence are spoken. This variation of stress affects monosyllables only in connected speech, where they receive an accentuation relative to the other words of the sentence. An absolute uniformity of stress in a sentence is unnatural, though the amount of variation in stress differs greatly in different languages. ‘The distinctions of stress in some languages are less marked than in others. Thus in French the syllables are all pronounced with a nearly uniform stress, the strong syllables rising only a little above the general level, its occurrence being also uncertain and fluctuating. This makes Frenchmen unable without systematic training to master the accentuation of foreign languages.’[6] English and the other Teutonic languages, on the other hand, show a marked tendency to alternate weak and strong stress.

§ 7. With regard to the function which it discharges in connected speech, we may classify accent or stress under four different categories. First comes what may be called the syntactical accent, which marks the logical importance of a word in relation to other words of the sentence. In a sentence like ‘the birds are singing’, the substantive ‘birds’ has, as denoting the subject of the sentence, the strongest accent; next in logical or syntactical importance comes the word ‘singing’, denoting an activity of the subject, and this has a comparatively strong accent; the auxiliary ‘are’ being a word of minor importance is uttered with very little force of voice; the article ‘the’, being the least emphatic or significant, is uttered accordingly with the slightest perceptible stress of all.

Secondly, we have the rhetorical accent, or as it might be called, the subjective accent, inasmuch as it depends upon the emphasis which the speaker wishes to give to that particular word of the sentence which he desires to bring prominently before the hearer. Thus in the sentence, ‘you have done this,’ the rhetorical accent may fall on any of the four words which the speaker desires to bring into prominence, e.g. ‘yóu (and no one else) have done this,’ or ‘you háve done this (though you deny it), or you have dóne this’ (you have not left it undone), or, finally, ‘you have done thís’ (and not what you were told). This kind of accent could also be termed the emphatic accent.

Thirdly, we have the rhythmical accent, which properly speaking belongs to poetry only, and often gives a word or syllable an amount of stress which it would not naturally have in prose, as, for instance, in the following line of Hamlet (iii. iii. 27)—

the unimportant word ‘to’ receives a stronger accent, due to the influence of the rhythm, than it would have in prose. Similarly in the following line of Chaucer’s Troilus and Cryseide, l. 1816—

the inflexional syllable es was certainly not ordinarily pronounced with so much stress as it must have here under the influence of the accent as determined by the rhythm of the line. Or again the word ‘writyng’, in the following couplet of Chaucer’s Canterbury Tales (Prol. 325–6)—

was certainly not pronounced in ordinary speech with the same stress on the last syllable as is here demanded both by the rhythm and rhyme.

As a rule, however, the rhythmical accent in English coincides with the fourth kind of accent, the etymological or word-accent, which we now have to deal with, and in greater detail.

Just as the different words of a sentence are pronounced, as we have seen, with varying degrees of stress, so similarly the different syllables of a single word are uttered with a varying intensity of the force of the breath. One of the syllables of the individual word is always marked off from the rest by a greater force of tone, and these others are again differentiated from each other by subordinate gradations of intensity of utterance, which may sometimes be so weak as to lead to a certain amount of indistinctness, especially in English. In the Teutonic languages, the root-syllable, as the most important element of the word, and that which conveys the meaning, always bears the chief accent, the other syllables bearing accents which are subordinate to this chief accent. As the etymology of a word is always closely associated with the form of the root-syllable, this syllabic accent may be called the etymological accent. It naturally happens that this syllabic accent coincides very often with the syntactical accent, as the syntactical stress must be laid on the syllable which has the etymological accent.

The degrees of stress on the various syllables may be as many in number as the number of the syllables of the word in question. It is sufficient, however, for purposes of metre and historical grammar, to distinguish only four degrees of accent in polysyllabic words. These four degrees of syllabic and etymological accent are as follows: 1. the chief accent (Hochton, Hauptton); 2. the subsidiary accent (Tiefton, Nebenton); 3. the absence of accent, or the unaccented degree (Tonlosigkeit); 4. the mute degree, or absence of sound (Stummheit). These last three varieties of accent arise from the nature of the Teutonic accent, which is, it must always be remembered, a stress-accent in which the volume of breath is expended mainly on the first or chief syllable. The full meaning of these terms can most easily be explained and understood by means of examples chosen either from English or German, whose accentual basis is essentially the same. In the word, wonderful, the first syllable has the chief accent (1), the last has the subsidiary accent (2), and the middle syllable is unaccented (3). The fourth or mute degree may be seen in such a word as wondrous, shortened from wonderous. This fuller form may still be used, for metrical purposes, as a trisyllable in which the first syllable has the chief accent, the last the subsidiary accent, and the middle syllable is unaccented, though audible. The usual pronunciation is, however, in agreement with the usual spelling, disyllabic, and is wondrous; in other words, the vowel e which originally formed the middle syllable, has been dropped altogether in speech as in writing. From the point of view of the accent, it has passed from the unaccented state to the state of muteness; but may be restored to the unaccented, though audible, state, wherever emphasis or metre requires the full syllable. We have the line: ‘And it grew wondrous cold,’ for which we might have ‘The cold grew wonderous’. In other cases the vowel is retained in writing but is often dropped in colloquial pronunciation, or for metrical convenience. Thus, in Shakespeare, we find sometimes the full form—

and sometimes the curtailed form—

This passing of an unaccented syllable into complete muteness is very frequent in English, as compared with other cognate languages. It has led, in the historical development of the language, to a gradual weakening, and finally, in many instances, to a total loss of the inflexional endings. Very frequently, an inflexional vowel that has become mute is retained in the current spelling; thus in the verbal forms gives, lives, the e of the termination, though no longer pronounced, is still retained in writing. Sometimes, in poetical texts, it is omitted, but its position is indicated by an apostrophe, as in the spellings robb’d, belov’d. In many words, on the other hand, the silent vowel has ceased to be written, as in grown, sworn, of which the original forms were growen, sworen

§ 8. Written marks to indicate the position of the accent were employed in early German poetry as early as the first half of the ninth century, when they were introduced, it is supposed, by Hrabanus Maurus of Fulda and his pupil Otfrid. The similar marks that are found in certain Early English MSS., as the Ormulum, are usually signs of vowel-quantity. They may possibly have sometimes been intended to denote stress, but their use for this purpose is so irregular and uncertain that they give little help towards determining the varying degrees of accent in words during the earliest stages of the language. For this purpose we must look for other and less ambiguous means, and these are found (in the case of Old English words and forms) first, in the alliteration, secondly, in comparison with related words of the other Teutonic languages, and, thirdly, in the development in the later stages of English itself. After the Norman Conquest, the introduction of rhyme, and of new forms of metre imitated from the French and mediaeval Latin poetry, affords further help in investigating the different degrees of syllabic accent in Middle English words. None of these means, however, can be considered as yielding results of absolute certainty, chiefly because during this period the accentuation of the language was passing through a stage of transition or compromise between the radically different principles which characterize the Romanic and Teutonic families of languages. This will be explained more fully in a subsequent chapter.

Notwithstanding this period of fluctuation the fundamental law of accentuation remained unaltered, namely, that the chief accent falls on the root of the word, which is in most cases the first syllable. For purposes of notation the acute (´) will be used in this work to denote the chief accent, the grave (`) the subsidiary accent of the single word; to indicate the rhythmical or metrical accent the acute alone will be sufficient.

§ 9. In English poetry, as in the poetry of the other Teutonic nations, the rhythmical accent coincides normally with the syllabic or etymological accent, and this, therefore, determines and regulates the rhythm. In the oldest form of Teutonic poetry, the original alliterative line, the rhythm is indicated by a definite number of strongly accented syllables, accompanied by a less definite number of syllables which do not bear the same emphatic stress. This principle of versification prevails not only in Old English and Old and Middle High German poetry, but also, to a certain extent, in the period of Middle English, where, in the same manner, the number of beats or accented syllables indicates the number of ‘feet’ or metrical units, and a single strongly accented syllable can by itself constitute a ‘foot’. This practice is a feature which distinguishes early English and German poetry, not only from the classical poetry, in which a foot or measure must consist of at least two syllables, but also from that of the Romanic, modern German, and modern English languages, which has been influenced by classical example, and in which, accordingly, a foot must contain one accented and at least one unaccented syllable following one another in a regular order. The classical terms ‘foot’ and ‘measure’ have, in their strict sense, relation to the quantity of the syllables, and can therefore be applied to the modern metres only by analogy. In poetry which is based on the principle of accent or stress, the proper term is bar (in German Takt). The general resemblances between modern accentual and ancient quantitative metres are, however, so strong, that it is hardly desirable to discontinue the application of old and generally understood technical terms of the classical versification to modern metres, provided the fundamental distinction between quantity and accent is always borne in mind.

Setting aside for the present the old Teutonic alliterative line, in which a ‘bar’ might permissibly consist of a single syllable, we may retain the names of the feet of the classical quantitative versification for the ‘bars’ of modern versification, using them in modified senses. A group consisting of one unaccented followed by an accented syllable may be called an iambus; one accented followed by an unaccented syllable a trochee; two unaccented syllables followed by an accented syllable an anapaest; one accented syllable followed by two unaccented syllables a dactyl. These four measures might also be described according to the length of the intervals separating the accents, and according as the rhythm is ascending (passing from an unaccented to an accented syllable) or descending (passing from an accented to an unaccented syllable). We should then have the terms, (1) ascending disyllabic (iambus), (2) descending disyllabic (trochee), (3) ascending trisyllabic (anapaest), and (4) descending trisyllabic (dactyl).[7] But we may agree with Prof. Mayor that ‘it is certainly more convenient to speak of iambic than of ascending disyllabic’.[8] It is, however, only in the case of these four feet or measures that it is desirable to adhere to the terminology of the ancient metres, and as a matter of fact iambus, trochee, anapaest, and dactyl are the only names of classical feet that are commonly recognized in English prosody.[9] As to the employment in the treatment of English metre of less familiar technical terms derived from classical prosody, we agree with Prof. Mayor, when he says: ‘I can sympathize with Mr. Ellis in his objection to the classicists who would force upon us such terms as choriambic and proceleusmatic to explain the rhythm of Milton. I do not deny that the effect of his rhythm might sometimes be represented by such terms; but if we seriously adopt them to explain his metre, we are attempting an impossibility, to express in technical language the infinite variety of measured sound which a genius like Milton could draw out of the little five-stringed instrument on which he chose to play.’ The use of these and other classical terms is justifiable only when we have to deal with professed imitations of ancient forms of verse in English.

Whatever names may be chosen to denote the metrical forms, the measure or foot always remains the unity which is the basis of all modern metrical systems, and of all investigation into metre. For a line or verse is built up by the succession of a limited number of feet or measures, equal or unequal. With regard to the limit of the number of feet permissible in a line or verse, no fixed rule can be laid down. In no case must a line contain more feet than the ear may without difficulty apprehend as a rhythmic whole; or, if the number of feet is too great for this, the line must be divided by a pause or break (caesura) into two or more parts which we may then call rhythmical sections. This break is a characteristic mark of the typical Old English alliterative line, which is made up of two rhythmical sections. The structure of this verse was at one time obscured through the practice of printing each of these sections by itself as a short line; but Grimm’s example is now universally followed, and the two sections are printed as parts of one long line.[10] Before entering into a detailed consideration of the alliterative long line, it will be needful to make a few general remarks on rhyme and its different species.

§ 10. Modern metre is not only differentiated from metre of the classical languages by the principle of accent as opposed to quantity; it has added a new metrical principle foreign to the ancient systems. This principle is Rhyme. Instances of what looks like rhyme are found in the classical poets from Homer onwards, but they are sporadic, and are probably due to accident.[11]

Rhyme was not in use as an accessory to metre in Latin till the quantitative principle had given way to the accentual principle in the later hymns of the Church, and it has passed thence into all European systems of metre.

In our poetry it serves a twofold purpose: it is used either simply as an ornament, or as a tie to connect single lines into the larger metrical unity of stanza or stave, by the recurrence of similar sounds at various intervals.

In its widest sense rhyme is an agreement or consonance of sounds in syllables or words, and falls into several subdivisions, according to the extent and position of this agreement. As to its position, this consonance may occur in the beginning of a syllable or word, or in the middle, or in both middle and end at the same time. As to its extent, it may comprehend one or two or more syllables. Out of these various possibilities of likeness or consonance there arise three chief kinds of rhyme in this wide sense, alliteration, assonance, and end-rhyme, or rhyme simply in the more limited and usual acceptation of the word.

§ 11. This last, end-rhyme, or full-rhyme, or rhyme proper, consists in a perfect agreement or consonance of syllables or words except in their initial sounds, which as a rule are different. Generally speaking, the agreement of sounds falls on the last accented syllable of a word, or on the last accented syllable and a following unaccented syllable or syllables. End-rhyme or full-rhyme seems to have arisen independently and without historical connexion in several nations, but as far as our present purpose goes we may confine ourselves to its development in Europe among the nations of Romanic speech at the beginning of the Middle Ages. Its adoption into all modern literature is due to the extensive use made of it in the hymns of the Church. Full-rhyme or end-rhyme therefore is a characteristic of modern European poetry, and though it cannot be denied that unrhymed verse, or blank verse, is much used in English poetry, the fact remains that this metre is an exotic product of the Renaissance, and has never become thoroughly popular. Its use is limited to certain kinds of poetic composition, whereas rhyme prevails over the wider part of the realm of modern poetry.

§ 12. The second kind of rhyme (taking the word in its broader sense), namely, vocalic assonance, is of minor importance in the treatment of English metre. It consists in a similarity between the vowel-sounds only of different words; the surrounding consonants do not count. The following groups of words are assonant together: give, thick, fish, win; sell, step, net; thorn, storm, horse. This kind of rhyme was very popular among the Romanic nations, and among them alone. Its first beginnings are found in the Latin ecclesiastical hymns, and these soon developed into real or full-rhyme.[12] It passed thence into Provençal, Old French, and Spanish poetry, and has continued in use in the last named. It is very rarely found in English verse, it has in fact never been used deliberately, as far as we know, except in certain recent experiments in metre. Where it does seem to occur it is safest to look upon it as imperfect rhyme only. Instances are found in the Early English metrical romances, Lives of Saints, and popular ballad poetry, where the technique of the metre is not of a high order; examples such as flete, wepe; brake, gate; slepe, ymete from King Horn might be looked on as assonances, but were probably intended for real rhymes. The consistent use of the full-rhyme being difficult, the poets, in such instances as these, contented themselves with the simpler harmony between the vowels alone, which represents a transition stage between the older rhymeless alliterative verse, and the newer Romanic metres with real and complete rhyme. Another possible form of assonance, in which the consonants alone agree while the vowels may differ, might be called consonantal assonance as distinguished from vocalic assonance, or assonance simply. This form of assonance is not found in English poetry, though it is employed in Celtic and Icelandic metres.[13]

§ 13. The third species of rhyme, to use the word still in its widest sense, is known as alliteration (German Stabreim or Anreim). It is common to all Teutonic nations, and is found fully developed in the oldest poetical monuments of Old Norse, Old High German, Old Saxon, and Old English. Even in classical poetry, especially in the remains of archaic Latin, it is not unfrequently met with, but serves only as a means for giving to combinations of words a rhetorical emphasis, and is not a formal principle of the metre bound by strict rules, as it is in Teutonic poetry. Alliteration consists in a consonance or agreement of the sounds at the beginning of a word or syllable, as in love and liking, house and home, woe and weal. The alliteration of vowels and diphthongs has this peculiarity that the agreement need not be exact as in ‛apt alliteration’s artful aid’, but can exist, at least in the oldest stages of the language, between all vowels indiscriminately. Thus in the oldest English not only were ellen and ende, ǣnig and ǣr, ēac and ēage alliterations, but age and īdel, ǣnig and ellen, eallum and æðelingum were employed in the strictest forms of verse as words which perfectly alliterated with each other.

This apparent confusion of vowel-sounds so different in their quantity and quality is probably to be explained by the fact that originally in English, as now in German, all the vowels were preceded by a ‘glottal catch’ which is the real alliterating sound.[14] The harmony or consonance of the unlike vowels is hardly perceptible in Modern English and does not count as alliteration.

The most general law of the normal alliterative line is that three or at least two of the four strongly accented syllables which occur in every long line (two in each section) must begin with an alliterative letter, for example, in the following Old English lines:

or in early Modern English:

The history of the primitive alliterative line follows very different lines of development in the various Teutonic nations. In Old High German, after a period in which the strict laws of the verse were largely neglected, it was abandoned in favour of rhyme by Otfrid (circa 868). In Old English it kept its place as the only form of verse for all classes of poetical composition, and continued in use, even after the introduction of Romanic forms of metre, during the Middle English period, and did not totally die out till the beginning of the seventeenth century. The partial revival of it is due to the increased interest in Old English studies, but has been confined largely to translations. As an occasional rhetorical or stylistic ornament of both rhymed and unrhymed verse, alliteration has always been made use of by English poets.

CHAPTER II
THE ALLITERATIVE VERSE IN OLD ENGLISH

§ 14. General remarks. It is highly probable that alliteration was the earliest kind of poetic form employed by the English people. There is no trace in the extant monuments of the language of any more primitive or simpler system. A predilection for alliteration existed even in prose, as in the names of heroes and families like Scyld and Sceaf, Hengist and Horsa, Finn and Folcwald, pairs that alliterate in the same way as the family names of other Teutonic nations: the names of the three sons of Mannus, Ingo, Isto, Irmino, conform to this type.[15] The earliest monuments of Old English poetry, as the fragmentary hymn of Cædmon in the More MS. (Cambridge) and the inscription on the Ruthwell Cross, are composed in the long alliterative line. The great body of Old English verse is in this metre, the only exceptions being the ‘Rhyming Poem’ (in the Exeter Book),[16] and a few other late pieces, in which alliteration and rhyme are combined. This Old English poetry, therefore, together with the Old Norse and Old Saxon remains (the Heliand with 5,985 lines, and the recently discovered fragment of the Old Saxon Genesis, edited by Zangemeister and Braune, 1894, with 335 lines), affords ample and trustworthy material for determining the laws of the alliterative verse as used by the Teutonic nations. In comparison with these the remains of Old High German alliterative verse are both scanty and lax in structure.

§ 15. Theories on the metrical form of the alliterative line. Notwithstanding their comparative scantiness, the Old High German fragments (Hildebrandslied, Wessobrunner Gebet, Muspilli and two magical formulae, with a total of some 110 lines) formed the basis of the earliest theories of the laws of the accentuation and general character of the original alliterative line. They were assumed to have preserved the features of the primitive metre, and conclusions were drawn from them as to the typical form of the verse. When examined closely, the Old High German remains (and this is true also of the longer monuments in Old Saxon) are found to differ widely from Old Norse and Old English verse in one respect. While the general and dominating features of the line remain the same, the Old High German and Old Saxon lines are much longer than the Old Norse or Old English lines. In Old Norse or Old English the half line frequently contains no more than four syllables, in marked contrast to Old High German and Old Saxon, where the half line or section is considerably longer.

The first attempt at a theory of the metrical structure of the alliterative line was made by Lachmann. He based his theory on the form of verse created by Otfrid, in imitation of Latin models, which consists of a long line of eight accents, separated by leonine rhyme into two sections each of four accents alternately strong and weak.[17] The laws of the rhyming and strophic verse of Otfrid were applied by Lachmann to the purely alliterative verses of the Old High German Hildebrandslied, and this system of scanning was further applied by his followers to the alliterative verse of Old English, the true nature of which was long misunderstood on the Continent. In England itself a sounder view of the native alliterative verse was propounded by Bishop Percy as early as 1765, in his Essay on the Metre of Pierce Plowman published along with his well-known Reliques of Ancient English Poetry, not to speak of the earlier writings of G. Gascoigne (1575) and James VI (1585). But the number and authority of some of Lachmann’s followers are such that some detailed account of their theories must be given.[18]

§ 16. The four-beat theory of the alliterative verse, based on the assumption that each of the two sections must have had four accented syllables to bring out a regular rhythm, was applied by Lachmann himself only to the Old High German Hildebrandslied,[19] while on the other hand he recognized a freer variety with two chief accents only in each section, for the Old Norse, Old Saxon, and Old English. The four-beat theory was further applied to the Old High German Muspilli by Bartsch,[20] and to the rest of the smaller relics of Old High German verse by Müllenhoff.[21] The next step was to bring the Old Saxon Heliand and the Old English Beowulf under this system of scansion; and this was taken by M. Heyne in 1866 and 1867. But the metre of Beowulf does not differ from that of the other alliterative poems in Old English, and these in their turn were claimed for the four-beat theory by Schubert,22 but with this important modification, made before by Bartsch, that side by side with the usual four-beat sections there were also to be found sections of three beats only. One obvious difficulty in applying the theory of four strongly marked beats to the Old English half-lines or hemistichs is this, that in Old English these hemistichs consist in very many cases of not more than four syllables altogether, each one of which would on this theory have an accent to itself. To meet these cases E. Jessen[23] started the theory that in certain cases pauses had to be substituted for ‘beats not realized’. A further modification of the four-beat doctrine was introduced by Amelung,[24] who maintained that in the metre of the Heliand each hemistich had two primary or chief accents and two secondary or subordinate accents. In order to bring the verse under this scansion he assumes that certain syllables admitted of being lengthened. He further regarded the Heliand verse as a metre regulated by strict time, and not as a measure intended for free recitation and depending only on the number of accented syllables.

A few other more recent attempts at solving the problem must be mentioned before we pass on to explain and discuss Sievers’s system in the next paragraph. The views of Prof. Möller of Copenhagen[25] have found an adherent in Lawrence, from whose book[26] we may quote the following summary of Möller’s theory. According to Prof. Möller the hemistich consists theoretically of two measures (Takte), each of four morae ×́××̀× (a mora, ×, being the time required for one short syllable), and therefore the whole verse of four measures, thus:

×́××̀×|×́××̀×||×́××̀×|×́××̀×||.

Where, in a verse, the morae are not filled by actual syllables, their time must be occupied by rests (represented by r*) in reciting, by holding on the note in singing.[27] A long syllable, ——, is equivalent to two morae. Thus v. 208 of Beowulf