Some of the above suggestions are obviously of little importance, and many must fairly remain open to question. I have been guided sometimes merely by the overwhelming majority of the MSS. collated, when no critical principle seemed to be involved, since the concurrence of many MSS. having no special tendency otherwise to agree seemed to have some weight. In other cases my reasons will be found stated either in the notes to the collation of the Inferno, or in the discussions of the selected passages.

APPENDIX V.

ON THE METRE OF THE DIVINA COMMEDIA.

[Mr Tozer has kindly contributed the following Essay on the principles of metre and scansion observed by Dante. These can only be determined by such a minute study of the poem itself from this special point of view, as Mr Tozer has carefully carried out. I regret that I had not given sufficient attention to the subject to enable me to employ as often as I might have done the arguments available from this source for the purposes of textual criticism. It is evident that when it is once clearly established what is admissible and what is not in respect of metre and scansion, we should be able sometimes to reject summarily a reading otherwise well supported, as not conforming itself to the rules and principles observed by the poet.]

The subject of metre is necessarily a technical one, but, notwithstanding this, it is of importance, not only to the student, but also to the general reader, because of its intimate connection with the power of appreciating poetry. In order to discover a faulty line, if such occurs, and to distinguish a more rhythmical verse from a less rhythmical one— and thus much it is desirable that an intelligent reader should be able to do-it is necessary to understand something, at least, of the principles on which the metre is constructed. This is admitted on all hands with regard to the classical languages, but it is true also, though in a somewhat less degree, of modern poetry. Without this knowledge, a person may read a poem with an irregular, vague impression of its rhythm; but this will probably be something different from what the author intended, and occasionally will be extremely lame. Now, in order to be able to test the metre of the successive lines as we meet them, and to know why they should be read in this, and not in that way-a process which a very little practice renders easy-it is necessary to determine such questions as the following. To what extent is it allowable for the poet to depart, for the sake of variety or for other causes, from the exact metre and rhythm of the lines? How many vowels in the same word may be pronounced together, so as to form one syllable for metrical purposes?

Is it ever allowable to pronounce two vowels, with a consonant between them, as if they were one syllable? How many vowels at the end of a word can be elided before a vowel at the beginning of the next word? Under what circumstances may a vowel be left open and unelided at the end of a word before a vowel at the beginning of the next? It is the object of the present paper to throw some light on these and similar points with reference to the Divina Commedia; and they are the more worthy of consideration, because metrical difficulties are steadily ignored by the commentators on that poem. The study of this subject is important also for the emendation of the text, because by this means in some cases we can test the accuracy of a reading, and in others we can discover the origin of a corruption.

I. The rule of eleven syllables to the line.

Dante's metre is the regular heroic verse of eleven syllables to the line, and the verses are arranged in stanzas of three lines each. To the rule of eleven syllables there are certain exceptions in the form of tensyllable and twelve-syllable lines. Those with twelve syllables are rare, for I do not include such as end with words in -io, -ia, &c., as ufizio, rabbia, where the two last vowels, if they occurred in the middle of a verse, would usually be pronounced as one syllable. Eliminating these, we find twelve-syllable lines only in six passages; viz. Inf. xv. 1, Ora cen porta l'un de duri margini, with a corresponding rhyme in argini, but no third rhyme, because it is a doublet at the beginning of a Canto; Inf. xxiii. 32, where the endings are scendere, rendere, prendere ; Inf. xxiv. 62, malagevole, fievole, and disconvenevole; Inf. xxviii. 80, cattolica, Maiolica, Argolica; Par. xxvi. 125, inconsumabile, razionabile, durabile; Par. xxviii. 125, girano, rimirano, tirano. One of these instances, it will be seen, contains proper names, in which all poets allow themselves a certain license; while the rest, with one exception, are forms, the last vowel of which might be omitted in the middle of a line. The exception is margini, argini; but in these words the two last syllables belong to the inflexion, and are therefore weak.

The ten-syllable lines are more numerous, occurring in fourteen passages, besides two, which we need hardly count, in the Provençal passage at the end of Purg. xxvi. Seven of these are accounted for by one of the rhymes in the terza rima being a proper name with the final syllable accented; viz. Inf. iv. 56, D' Abel suo figlio, e quella di Noè, with corresponding rhymes re, fe'; xx. 74, può, co, Po; xxviii. 32, All, qui, così; xxxii. 26, Osteric, Tambernic, cric; xxxii. 62, Artù, più, fu; Purg. iv. 68, Sion, orizzon, Feton; xii. 41, Gelboè, te, fe. Three end with Hebrew or Latin words, viz. Purg. xxiii. 74, Che menò Cristo lieto a dire: El, with corresponding rhymes dì, qui; xxxiii. 8, Modicum et non videbitis me, with piè and me; Par. vii. 1, Osanna sanctus Deus Sabaoth, with malacoth, a doublet at the beginning of a Canto. Four others remain to be accounted for, viz. Inf. xxiii. 143, where the terminations

are udï', gì, parti'; xxxi. 143, sposò, levò, a doublet at the end of the Canto; Purg. vii. 8, fè, sè, è; Par. xxv. 98, udì, schiarì, dì. It might perhaps be sufficient to say, that the poet has here availed himself of a permissible license, in order to express his meaning more exactly and directly. But it is worth considering, whether they may not all be explained as adaptations of sound to sense, since in every case a sudden or abrupt movement is implied, to which the abrupt termination corresponds. This is most plainly seen in the second passage, where Anteo is depositing the two pocts in the pit of Hell;

Ma lievemente al fondo, che divora
Lucifero con Giuda, ci sposò;
Nè sì chinato lì fece dimora,

E come albero in nave si levò.

In the other three passages we also find,

Appresso il Duca a gran passi sen gì;
Qual è colui che cosa innanzi sè
Subita vede, ond' ei si maraviglia;
E prima, appresso al fin d' este parole,
Sperent in te, di sopra noi s'udì,

A che risposer tutte le carole;

Poscia tra esse un lume si schiarì.

But, however this may be, the number of these exceptions is so small, that we may feel confident that they were not introduced to modify the regular recurrence of the hendecasyllable line, as if it were monotonous. Dante, indeed, himself maintains in the De Vulgari Eloquio (Bk. 2, c. 5), that the ten-syllable lines which are found in Provençal and Old French poetry are in many cases to be regarded as elevensyllable lines, because the accented syllable in which they end implies that an additional syllable belonged to the word, as bonté for bonitatem; and that this syllable was pronounced in poetry, as bontée. But Karl Bartsch has shewn (Dante's Poetik, in the Jahrbücher der deutschen Dante-Gesellschaft, vol. 3, pp. 327, 328) that Dante was in error here, for the monosyllabic rhyme, and consequently the ten-syllable verse, was the earlier; and that the eleven-syllable verse became prevalent in Italian owing to the great number of disyllabic endings in that language. Anyhow this explanation could not be made to apply to some of Dante's own ten-syllable verses; e.g. to those which end in me, se, qui, and proper names like AN.

II. Rhythmical arrangement of feet in the verse.

The metre of the Divina Commedia, in its most exact form, if described by the terms we apply to classical poetry—though these terms can only be used approximately in reference to modern verse-is purely iambic; that is, the lines naturally break into feet of two syllables each,

the stress or beat being on the latter of the two syllables. This is the case with the two first lines of the poem :

Nel mézzo del cammín di nóstra víta

Mi rítrovái per úna sélva oscúra ;1

and when the poet desires to give the greatest smoothness to his verses, he keeps to this; as in the well-known description of the vesper bell,

Che paía il giorno piánger ché si muóre.

But the regularity of this rhythm is constantly modified in one part or other of the line, the most common modification being the substitution of a trochee for an iambus in the first foot-in other words, the shifting of the stress from the second to the first syllable in that foot—as

Ché la veráce vía abbándonái.

The same change is also very frequent in the third foot, as
Per mé si vá nélla città dolénte;

and in the second, as

Perchè tánta viltà nel córe allétte;

and it is in the temperate use of these and similar changes that the melody of the verse consists. But in many of his lines Dante departs much further from the original metre, so that it is not easy at first to reconcile oneself to their rhythm. Thus in such lines as

Éd avánti che sían di là discése;

É alquánto di lúngi dálla spónda;

we feel a harshness in the two first feet, arising mainly from there being no word-accent on either of the two first syllables of the line. This can best

be removed by throwing a stress on the first monosyllable.

So, too, there

is something questionably rhythmical, when the former half of the verse is constructed in the following manner;

Tále imagine quívi facéan quélle;

Chè l' imágine lór vie più m' asciuga;

and when the latter half is constructed thus;

Che Itália chiúde e i suói términi bágna ;

E sé tu bén la túa Física nóte.

Now in the two last instances a certain relief is introduced by the monosyllable at the end of the third foot; when this is absent, the harshness is greater; as in

E vídivi éntro térríbile stĺpa.

I have given the references to the quotations that follow, only when that which they illustrate is peculiar or exceptional.

The least pleasing of all verses are those, the latter half of which is arranged as follows:

Ch' ío mi sia tárdi ál soccórso leváta;

Perchè ardíre é franchézza non hái.

In these, however, again the monosyllable in the third foot brings some relief without it the lines seem almost unrhythmical; as in

Questi non hanno spéránza di morte;
Quél della Préssa sápéva già cóme;
Invér la Spágna rívólse lo stuólo;

Ma nón con quésta módérna favella.

We can test this question of rhythm by examining corresponding lines in English. Thus, while we at the present day accent 'perspective' on the second syllable, we can be sure from the rhythm only, independently of other proofs, that Drummond accented it on the first syllable in the line

While sense's light mind's pérspective kept blind.

Again, we accent 'crystalline' on the first syllable; but we can be sure from the rhythm that Shelley accented it on the second, when he wrote

Lúlled by the coíl of hís crystalline streams.

If, however, these two words were accented 'perspéctive' and 'crystalline' in these verses, we should have the exact metre (allowing for the absence of the eleventh syllable) of some of Dante's lines; thus

Per mé si vá néll' etérno dolore:

(While sénse's light mind's perspective kept blind)

Bátte col rémo quálúnque s'adágia;

(Lúlled by the coíl of his crystalline streáms).

Thus we see that the rhythm which we find in Dante would be intolerable in good English verse. Certainly, the most unrhythmical English poet would not think of introducing as verses such quotations as Mánibus ó dáte lília plénis; Módicum ét nón vidébitis mé.

Still, all these rhythms, which seem to our ears more or less objectionable, are found in Petrarch also, though perhaps less frequently than in Dante; evidently therefore they were not offensive to an Italian ear. It may serve in part to explain this apparent discrepancy in taste, if we remember that the Italian pronunciation, which throws the accent of each word less definitely on any one syllable than is the case in English-or (to express the same thing in other words), admits the use of secondary as well as primary word-accent, which English does not-somewhat helps out these lines. But in modern verse opinions as to what is rhythmical differ so much in different periods and in different languages, that it is necessary to speak with great diffidence on the subject.