The revision
was not due to the fact that this popular and successful work
was poetically immature, nor even chiefly to the desire to
add a series of new elegies ; as he himself tells us in the pre-
fixed epigram, the question was primarily one of more careful
finish and elaboration : " hoc illi praetulit auctor opus "
(cf.
was not due to the fact that this popular and successful work
was poetically immature, nor even chiefly to the desire to
add a series of new elegies ; as he himself tells us in the pre-
fixed epigram, the question was primarily one of more careful
finish and elaboration : " hoc illi praetulit auctor opus "
(cf.
Ovid - 1869 - Juvenile Works and Spondaic Period
Ciris, 1-4.
23 Plato, Aristotle, Zeno, and Epicurus; cf. Ciris, 15: sapientia . . . quat-
tuor antiquis heredibus edita censors.
24 Ciris, 12-41. The plan was later fulfilled probably in the Aetna (the
language of which I have not yet examined in all its details) and in the pro-
oemium and fifteenth book of the Metamorphoses, also perhaps in the lost
Phaenomena and in parts of the Fasti.
? ? Generated for (University of Chicago) on 2014-12-26 11:20 GMT / http://hdl. handle. net/2027/uc2. ark:/13960/t45q4vr7s Public Domain / http://www. hathitrust. org/access_use#pd
? Vol. li]
The Juvenile Works of Ovid
identical with the Sulpicia letters. The figures for the Pane-
gyric and the Ciris are those of Ehrengruber (op. cit. x, 5
and 21). The Aetna, which, according to every probability,
is also the work of Ovid, is omitted from the comparison
only because the exact figures are not accessible to me at the
present writing ; its proportions are. however, not very differ-
ent from those of the Ciris. There appears to be conclusive
evidence for including also the Culex, though I was long
prevented from examining the language of this poem by
erroneous impressions that I had at first formed respecting
the treatment of the caesura in this work. Finally, Pliny,
N. H. xxxii, 152, was wholly mistaken in his conjecture
(" fortassis ") that the Halieutica relates to the " fish of
the Black Sea " and was consequently written at the close
of the poet's life ; the schemata show conclusively that the
poem, wholly dependent as it is upon Greek books, belongs
to the Lygdamus and Sulpicia period.
THE JUVENILE WORKS OF OVID
Catalept.
TV
Halieut.
Lygdamus,
all six
Lygdamus,
four
Lygdamus,
two
Sulpicia
Letters,
IX.
62 vss.
130 hex.
elegies. 25
elegies,
i, 2, 3, 6.
elegies,
4i 5-
IV, 7-12.
290 vss.
1 60 vss.
130 vss.
40 vss.
SS. Dact. 2
53 44-8
222 42. 8
359 41-3
216 45. 0
144 36. 9
45 37-5
SS. Spond.
75 55--'
298 57. 2
5" 53. 7
264 55-o
246 63. 1
75 62. 5
I. 27 Dact.
46 71. 0
62 47. 7
1 66 57. 2
ioi 63. 1
65 50. 0
20 50. 0
I. Spond.
1 8 28. 1
68 52-3
124 42. 8
? ? 59 36-9
Generated for (University of Chicago) on 2014-12-26 11:20 GMT / http://hdl. handle. net/2027/uc2. ark:/13960/t45q4vr7s Public Domain / http://www. hathitrust. org/access_use#pd
? 164 Robert S. Radford [1920
A few words of explanation may be added to the tabular
statement. Ovid began in Catalepton, ix with a proportion
of dactyls which was wholly normal in the year 27 B. C. Thus
he has 44. 8% in the whole distich, which is as good as the
second and third books of Propertius, 28 as good also as the
eighth and third elegies of Tibullus' first book. 29 Yet he has
no limitations whatever in Catalepton, ix upon polysyllabic
endings in the pentameter; on the contrary, like Catullus
and like Propertius in his earlier work, he fairly revels in
their use (50%). In the Lygdamus elegies, however, and
in the Sulpicia letters, the ambitious and aspiring youth
seeks suddenly to pass from the longer endings to the more
elegant dissyllables of Tibullus, and is evidently preoccupied
with this problem and its difficulties. Under these conditions
throughout the Lygdamus poems he wavers greatly in his
composition, and, in one hundred and seventy verses (Lygd.
4 and 5, [Tib. ] rv, 7-12), he even sinks to the proportion of
Catullus, namely, about 37% in the distich, and to only
50% of dactylic beginnings. 30 This is of course excellent
Latin elegy, but it is not the kind that was most in vogue in
24 B. C. 31 It is only in the qualified or limited sense just ex-
28 Propertius has 44. 8% and 44. 7% in the second and third books respec-
tively. My figures are taken from Hultgren, op. cit. 23, who follows the five-
book division of Propertius.
29 Tibullus has 44. 8% in the eighth elegy (Pholoe) for the whole distich
and 45% in the third elegy. It is scarcely fanciful to see in the decline of
dactyls in the third elegy an expression of Tibullus' sadness and depression of
feeling during his illness at Corcyra ; we have the same phenomenon in Lyg-
damus' fifth elegy. My figures are taken from Cartault, Le distiqitc elegiaque
chez Tibitlle, Sulpicia, Lygdamus (Paris, 191 1), 7.
30 Figures for Catullus are given by Hultgren, op. cit. 15 ff. Catullus has
usually 58% to 6o r i of dactylic beginnings, but in Carm. 69-119 (319 vss. )
he has 36. 6^ of dactyls in the distich and only 47^ of dactylic beginnings.
31 Catullus wrote some of the best Latin elegy, and the naturalness and
directness of his style is due in part to his not exceeding the proportion named.
As I have shown in my "Licensed Feet in Latin Verse," op. cit. 251-27? ,
even the best Latin poets, such as Catullus and Horace, experience some diffi-
culty in always providing one required dactyl, and therefore they occasionally
admit without metrical ambiguity in such a foot exceptional or vulgar short-
enings and even short vovels (without m) in hiatus, as Lucilius, ix, 243 Bahr. ,
? ? Generated for (University of Chicago) on 2014-12-26 11:20 GMT / http://hdl. handle. net/2027/uc2. ark:/13960/t45q4vr7s Public Domain / http://www. hathitrust. org/access_use#pd
? Vol. li] The Juvenile Works of Ovid 165
plained that Ovid can be said to have ever been a disciple
of Catullus in the matter of dactylic proportions, but un-
doubtedly in his youth he paid this brief tribute to the freer
and more natural style of his great predecessor. A secondary
cause for the low proportion may perhaps be found in the sad
and almost despairing character of the two elegies.
It will be noted that Ovid advanced in the Sulpicia and
Messalinus poems which are composed in the elegiac metre
to about 47% of dactyls in the who! e distich ; the hexameter
lines alone of these poems show 46. 1% and 45. 9% respectively.
On the other hand, the hexameters of purely epic poems like
the Ciris fall back to 43. 5%, which is normal for epic verse
o 5 re cSrupto (necessary dactyl of the hexameter); Lucr. vi, 1133, natu 5 ra
cSruptum; Cat. 10, 26, istos c6mmoda; nam volo d Sar&pim (necessary
dactyl of the Phalaecean) ; Hor. Cann. in, 14, n, iam virum expertae, malS
6minatis (short vowel in hiatus, necessary dactyl of the Sapphic) ; Pers. 3, 66,
di'scitfi, 6 miseri (license of the first foot, with greatly preferred dactyl) ; Lux-
orius, 302, 4, magnum depre 5 nderg usum. If difficulty is experienced in supply-
ing one required dactyl for the Sapphic or the hexameter, it is clear that two
necessary dactyls, as in the pentameter, constitute a very exacting demand
upon the Roman language, and if the virtuosity is also insisted upon, a very
elegant but very artificial form of verse is the result. For example, for metrical
reasons, Ovid (like Tibullus) constantly uses ab arte and similar phrases drawn
from the vulgar language, with its analytical tendency, instead of the simple
ablative, and he also uses, by poetic license, to an unparalleled extent, the
simple ablative for the ablative of the agent with ab, as Her. 4, 64, capti
pare 5 ntg sor6r, ' my sister was captivated by your parent,' where neither
parenti nor a parente could enter the verse ; upon the whole subject see in part
Guttmann, Sogenanntes instrumentales ab bei Ovid, Dortmund, 1890. I was
mistaken therefore in my former discussion (p. 271) in thinking that Ovid had
perhaps actually used the spelling parente for the dative parenti, as he so often
uses mare and caeleste for mari and caelesti, and as Statius, Sih. iv, 2, 28, uses
glaucae certantia Do b ride saxa, and Propertius, v (rv), 8, 10, writes cum temere
anguino creditur o^rl manus (Xeue-Wagener, Formenlehre, i 3 , 301). The
honest 37% of Catullus and Lygdamus does not compel a resort to such un-
usual constructions, and is therefore by no means to be despised. Dissen and
Postgate (Selections from Tibullus, p. XLIII), on purely subjective grounds and
without consulting indices, lexicons, or Latin authors, have discovered that
Lygdamus is an author of " poor Latinity. " On the contrary, his Latinity is
more natural and in some respects better than that of the mature Ovid. Yet
even the latter like Euripides, an unrivalled master of the graceful and
pleasing forensic style could give most of us lessons in correct Latinity to
our great and lasting profit.
? ? Generated for (University of Chicago) on 2014-12-26 11:20 GMT / http://hdl. handle. net/2027/uc2. ark:/13960/t45q4vr7s Public Domain / http://www. hathitrust. org/access_use#pd
? i66 Robert S. Radford [1920
and almost precisely the proportion of Vergil. In fact that
strange and anomalous thing, the actual preponderance of
dactyls in a Latin epic poem, could never by any possibility
arise in the more vigorous epic hexameter (versus fortis or
grams) ; it must appear first in the soft and tender strains of
the elegy (versus mollis or levis) , and may be transferred thence
to epic verse, as Ovid transferred it in the Metamorphoses
and Valerius Flaccus later employed it in imitation of Ovid.
Even Ovid in his mature period has recognized this difference
to some extent; for the hexameter of elegy in the Heroides
(first series) reaches 56. 4%, while the epic hexameter of the
Metamorphoses does not exceed 54. 8%.
It is evident then that in the first edition of the A mores
which was published in 14 B. C. , 32 only a few years after the
Sulpicia elegies, the Ciris, and the Aetna, Ovid had had little
opportunity to develop a marked dactylic virtuosity and to
become a highly artistic elegiac poet. We may be sure then
that the first Amores of 14 B. C. showed only a very moderate
advance upon the 47. 4% of the Sulpicia elegies, and they
may even have contained, like the latter, a few polysyllabic
endings in the pentameter. Up to 14 B. C. Ovid was in fact
fully as much an epic as an elegiac poet (see Am. i, i, 1-18
on his early ambition to excel in epic themes), and he was
still far from having attained in the Sulpicia poems the high
virtuosity which appears in three of the four genuine elegies
of Tibullus' second book, namely, 55% or more. 33 Two of
the short Sulpicia elegies and the equally short Cornutus
poem show, it is true, an equal or a superior number of dac-
tyls, 34 but four Sulpicia poems and the long Messalinus elegy
32 Schanz, op. cit. 270 (293), suggests that these elegies probably came
first separately before the public, but later were collected in the edition of
14 B. C. This seems quite probable, especially as the spondaic Amores show
in general the predominance of the same schemata or figures of the hexameter
as the Sulpicia and Messalinus poems.
33 The first elegy of the second book has 55. 2% in the distich, the third has
50%, the fourth 55%, and the sixth 56. 8%; see the figures of Cartault, op.
cit. 7, and Ehrengruber, op. cit. x, 5.
34 iv, 6 (20 vss. ) has 50% of dactyls ; iv, 5 (20 vss. ) 58. 3% ; n, 2 (20 vss. ) 50%.
? ? Generated for (University of Chicago) on 2014-12-26 11:20 GMT / http://hdl. handle. net/2027/uc2. ark:/13960/t45q4vr7s Public Domain / http://www. hathitrust. org/access_use#pd
? Vol. li] The Juvenile Works of Ovid 167
(n, 5) have a decided preponderance of spondees. A marked
change in Ovid's whole attitude took place, however, after
14 B. C. For unlike the Ciris and the Aetna, which did not
greatly increase their author's reputation, the Amores, im-
mediately upon their first publication, achieved a prodigious
success; they at once became popular favorites, and, like
the Eclogues of Vergil, were frequently sung in the theatre
with accompanying dance (Trist. n, 519; v, 7, 25). The
favorable reception thus accorded to the elegies naturally
determined the direction of the poet'3 genius and led him to
devote himself uninterruptedly (with the exception of the
tragedy of Medea) for sixteen years (14 B. C. to 2 A. D. ) to the
perfection of an elegiac style which surpassed even the later
work of Tibullus.
Having developed a remarkable and distinctive technique
in the Heroides and the Ars, Ovid resolved to make the first
Amores conform fully to his later rules of art.
The revision
was not due to the fact that this popular and successful work
was poetically immature, nor even chiefly to the desire to
add a series of new elegies ; as he himself tells us in the pre-
fixed epigram, the question was primarily one of more careful
finish and elaboration : " hoc illi praetulit auctor opus "
(cf. Gruppe, op. cit. 377). In brief, his principal purpose
was to eliminate all polysyllabic endings, to introduce every-
where the full dactylic virtuosity, to multiply the dactylic
beginnings, and to remove or greatly reduce those schemata,
such as DSSS, SDSS, 5555, 55, SD, and the like, which
were no longer fully acceptable. The view adopted by
Heuwes in his discussion of the matter (op. cit. 31) is that
the revision " consisted in here and there substituting a word
that was more suitable for one that was less suitable to the
metre and sense, or in changing the collocation of words o 1
verses, or in doing all these things at the same time. " These
conclusions are partly correct, but there can be no doubt that
both the additions and the alterations were far more extensive
and thoroughgoing than Heuwes supposes. The first edition
? ? Generated for (University of Chicago) on 2014-12-26 11:20 GMT / http://hdl. handle. net/2027/uc2. ark:/13960/t45q4vr7s Public Domain / http://www. hathitrust. org/access_use#pd
? 1 68 Robert S. Radford [1920
was in fact subjected to a drastic revision and was almost
entirely rewritten in conformity to the new rules of art. 35
Nevertheless abundant traces of the character of the first
edition still remain ; for in not a few cases the poet has
triumphed over the versifier and the artist, and a number
of the original poems of the first edition have been retained
with very moderate changes. This fact may be shown very
clearly in two ways, if the writer correctly understands the
true nature of Ovidian versification. For the full and har-
moniously developed virtuosity in the Ovidian sense means
first that there shall be a considerable preponderance of
dactyls in the distich taken as a whole, and secondly that
there shall be a similar preponderance of dactyls in each
member taken singly, that is, in both the hexameter and the
pentameter lines. It is evident therefore that virtuosity is
lacking in the Amores in two ways: (i) when the spondees
are either equal or predominant in the whole distich ;
(2) when they are equal or predominant in the hexameter
line (versus fortis or diirus) alone. With this understanding
of the nature of virtuosity, I may summarize the facts briefly
as follows. A considerable part of the second edition is
entirely new and consists of such newly written elegies as
n, 17 and 1 8 ; in, i and 15 ; and also the last 18 lines of in.
9 (the epicedion upon Tibullus). 36 A still larger part has
been completely or almost completely revised in order to
bring the elegies up to practically the same virtuosity as the
35 The extraordinary importance which the mature Ovid attached to formal
poli. -h and elegance is not only evident from the exacting rules \vhii h he strictly
observes, but also from his own express statements. Thus he speaks of de-
stroying many works which would have won popular favor, but which he
himself considered " faulty " (I'itiosa, Trist. iv. 10, 61"), he often laments that
the poems of the exile are composed with less care and skill than was his wont
(Pont, i, 5, 15 ff. ; 57 ff. ), and he bitterly complains that the Metamorphoses
were uncorrected and lacked the finishing touches at the moment of his banish-
ment, as in Trist. in, 14, 21, illud opus . . . nunc incorrectiun populi pervenit
in ora. See also ib. i, 7, 27 ff. , defuit et scriptis ultima lima meis; i, 7, 39 ff. ;
H, 555-
36 On the new elegies, see also Schanz, op. cit. 270 ( 293).
? ? Generated for (University of Chicago) on 2014-12-26 11:20 GMT / http://hdl. handle. net/2027/uc2. ark:/13960/t45q4vr7s Public Domain / http://www. hathitrust. org/access_use#pd
? Vol. li] The Juvenile Works of Ovid 169
Ars or the Heroides (first series), namely, 57% in the dis-
tich ; 37 about a fourth part of the whole edition, however,
shows only slight and moderate changes from its original
form. That is, of the forty-nine elegies, 38 fourteen have been
only partially and imperfectly revised. A single poem of
66 verses (in, 8) remains at practically the same average as
the Sulpicia elegies, namely, 47. 98% of dactyls in the distich ;
four other poems, however, of 184 verses (i, 2. 13. 15 ; in, 10)
remain with the dactyls subordinate in the hexameter line,
and either equal (only 50%) or subordinate (49. 23%) in the
whole distich. Five elegies (i, 14 ; 39 n, 4. 8. n; in, 3), of
236 verses, also remain with the dactyls either subordinate
or equal in the hexameter line, and with 52. 4, 50. 7, 51. 2,
50. 6, 50. 7% respectively in the whole distich. Two elegies
(n, 12. 14) of 72 verses, show a weak pentameter and only
51. 2 and 52. 3% in the distich; two other elegies (in, 14 and
the first three-fourths of in, 9, the epicedion on Tibullus,
composed in 19 B. C. ), with 100 verses, are also low in their
proportion of dactyls, namely, 53. 3 and 51. 3%. Further-
more, very largely the same identical schemata are predomi-
nant in all these elegies as we find preferred in the Sulpicia
elegies (iv, 2-6) and in the imitation of Tibullus (iv, 13).
Thus fourteen elegies of 658 verses more than one-fourth
of the total number conspicuously lack the virtuosity ;
among these are included some of the most notable poems in
the collection, such as i, 2 (the poet's willing submission to
Love), i, 13 (Aurora and Tithonus), i, 14 (Corinna's use of
dyes), i, 15 (epilogue on the immortality of poesy), n, 4
(reasons for love), n, n (lament over Corinna's voyage),
37 56. c/~c in the Ars (Drobisch), and 57. 6^ in the Heroides.
3S There are really fifty elegies, as Gruppe, op. cit. 377, first pointed out,
and as Ehwald, the latest editor, obtains, by breaking up n, o into two poems.
In order, however, to avoid confusion and to be uniform with the results of
Hultgren and Drobisch, my statistics (like theirs) are everywhere based upon
the edition of Merkel, Leipzig, 1887.
39 This is the poem which, through its reference to the subjugation of the
Sigambri (15 B. C. ), dates the whole original collection.
? ? Generated for (University of Chicago) on 2014-12-26 11:20 GMT / http://hdl. handle. net/2027/uc2. ark:/13960/t45q4vr7s Public Domain / http://www. hathitrust. org/access_use#pd
? 170 Robert S. Radford [1920
n, 12 (Corinna's surrender), n, 14 (the evils of abortion,
with several Lygdamus verses repeated), in, 3 (Corinna's
perjuries unpunished), in, 9 (epicedion on Tibullus, 50 out
of 68 verses), in, 14 (Corinna's infidelities best unknown).
The percentage of the nine ' imperfect ' elegies of Books I
and in is 50. 5% of dactyls for the whole distich, of the five
'imperfect' elegies of Book n, 51. 1%, and the difference
between the perfect and the imperfect parts of the second
Amores is 6% for the four poems of Book I, 5. 3% for the
five of Book 11, and 4. 8% for the five of Book in. 40 I think
it quite reasonable, however, to assume that, even in the case
of the imperfect elegies, at least 2. 3% of actyls have been
added in the revision, and I thus reach the conclusion that
the percentage of the first Amores did not exceed 48. 5% of
dactyls for the whole distich, and therefore rose only slightly
above the 47. 4% of the Sulpicia elegies and of iv, 13-14.
Though less important than the sum of the dactyls, the
proportion of dactylic beginnings also requires mention.
The percentage of dactylic beginnings in both hexameter and
pentameter in Am. I is 80 ; in the four imperfect elegies
(i, 2. 13. 14. 15) it is 73. 4, in the remainder of the book it is
82. 2, thus giving a difference of 8. 8% between the two parts.
We may consider as normal for the mature Ovid the per-
centage in both hexameter and pentameter of the Ars, which
is 82. 3, or that of the Fasti, which is 84. 2. Similarly the
percentage of dactylic beginnings in the whole of Am. 11 is
78. 6; in the five imperfect elegies (n, 4. 8. n. 12. 14) 41 it is
71. 6, in the remainder of the book it is 80. 9, giving a difference
of 9. 3% between the two parts. With these five elegies of
Am. ii we may fitly compare not only the percentage of the
Sulpicia elegies (iv, 2-6) and iv, 13-14, which is 69. 7% in
142 verses, but also that of the early Medicamen Faciei frag-
40 In the case of in, 9, the Tibullus epicedion, only the first 50 verses out of
68 are used.
41 n, q is also spondaic if \vc cut out the six verses (23-28) which Ovid
seems to have added in the second edition in order to join the originally sepa-
rate poems 9 and 9 H.
? ? Generated for (University of Chicago) on 2014-12-26 11:20 GMT / http://hdl. handle. net/2027/uc2. ark:/13960/t45q4vr7s Public Domain / http://www. hathitrust. org/access_use#pd
? Vol. li] The Juvenile Works of Ovid 171
ment of 100 verses, which is 70, while its percentage of dactyls
for the distich is also low, namely, 53. 5- 42 The percentage
of dactylic beginnings in the whole of Am. in is 77. 1 ; in the
four imperfect elegies (in, 8. 10. 14, first 50 vs. of 9) it is
74, in the remainder of the book it is 78. 5, giving a
difference of 4. 5 between the two parts. I consider this brief
summary of usage affecting the first foot sufficient for the
practical purposes of the present study, and in my subsequent
discussion shall purposely omit this feature of the single elegies
from the tabular statements. I may add, however, that the
percentage of dactyls in the first foot in Catalepton, ix is
71. 9; owing to preoccupation with the dissyllabic close and
to imitation of Catullus, it sinks in the Lygdamus elegies to
55-8 43 and in the Sulpicia letters (iv, . 7-12, 40 verses) to 50;
it rises again to 67. 7 in the Panegyric, 66. 8 in the Ciris, and
67. 1 in the Cnlex. ^
42 Hultgren, op. cit. 28 ff. , argues that no importance should be attached
lo so small a fragment, yet at the same time (pp. 32 f. ) shows the greatest
uneasiness because of its evidence, apparently fearing that it may bring to
naught the whole elaborate fabric of Ovidian virtuosity. One wonders what
he would have thought if he had examined the single Amorcs.
4:! It varies greatly in single Lygdamus elegies from 82. 2, 61. 8, and 65. 7
in i, 5, and 6 to 53. 4, 55. 2, and 45. 8 in 2, 3, and 4. In the technical Halieutica
it sinks also to 47. 7.
44 On account of the length of the article it has been found necessary to omit
here five tables giving exact statistics for the fourteen elegies. Two articles,
which continue the present study and treat the language and schemata of
Book iv of the Tibulline Corpus, will be published shortly in the American
Journal of Philology.
? ? Generated for (University of Chicago) on 2014-12-26 11:20 GMT / http://hdl. handle. net/2027/uc2. ark:/13960/t45q4vr7s Public Domain / http://www. hathitrust. org/access_use#pd
? ? ? Generated for (University of Chicago) on 2014-12-26 11:20 GMT / http://hdl. handle. net/2027/uc2. ark:/13960/t45q4vr7s Public Domain / http://www. hathitrust. org/access_use#pd
? UNIVERSITY Ol- t A1_II-WNIN I A A I 1_<JS AIN
THE UNIVERSITY LIBRARY
This book is DUE on the last data stam^'
-f CALIFOKNTJ
AT
? ? Generated for (University of Chicago) on 2014-12-26 11:20 GMT / http://hdl. handle. net/2027/uc2. ark:/13960/t45q4vr7s Public Domain / http://www. hathitrust. org/access_use#pd
? ? ?
23 Plato, Aristotle, Zeno, and Epicurus; cf. Ciris, 15: sapientia . . . quat-
tuor antiquis heredibus edita censors.
24 Ciris, 12-41. The plan was later fulfilled probably in the Aetna (the
language of which I have not yet examined in all its details) and in the pro-
oemium and fifteenth book of the Metamorphoses, also perhaps in the lost
Phaenomena and in parts of the Fasti.
? ? Generated for (University of Chicago) on 2014-12-26 11:20 GMT / http://hdl. handle. net/2027/uc2. ark:/13960/t45q4vr7s Public Domain / http://www. hathitrust. org/access_use#pd
? Vol. li]
The Juvenile Works of Ovid
identical with the Sulpicia letters. The figures for the Pane-
gyric and the Ciris are those of Ehrengruber (op. cit. x, 5
and 21). The Aetna, which, according to every probability,
is also the work of Ovid, is omitted from the comparison
only because the exact figures are not accessible to me at the
present writing ; its proportions are. however, not very differ-
ent from those of the Ciris. There appears to be conclusive
evidence for including also the Culex, though I was long
prevented from examining the language of this poem by
erroneous impressions that I had at first formed respecting
the treatment of the caesura in this work. Finally, Pliny,
N. H. xxxii, 152, was wholly mistaken in his conjecture
(" fortassis ") that the Halieutica relates to the " fish of
the Black Sea " and was consequently written at the close
of the poet's life ; the schemata show conclusively that the
poem, wholly dependent as it is upon Greek books, belongs
to the Lygdamus and Sulpicia period.
THE JUVENILE WORKS OF OVID
Catalept.
TV
Halieut.
Lygdamus,
all six
Lygdamus,
four
Lygdamus,
two
Sulpicia
Letters,
IX.
62 vss.
130 hex.
elegies. 25
elegies,
i, 2, 3, 6.
elegies,
4i 5-
IV, 7-12.
290 vss.
1 60 vss.
130 vss.
40 vss.
SS. Dact. 2
53 44-8
222 42. 8
359 41-3
216 45. 0
144 36. 9
45 37-5
SS. Spond.
75 55--'
298 57. 2
5" 53. 7
264 55-o
246 63. 1
75 62. 5
I. 27 Dact.
46 71. 0
62 47. 7
1 66 57. 2
ioi 63. 1
65 50. 0
20 50. 0
I. Spond.
1 8 28. 1
68 52-3
124 42. 8
? ? 59 36-9
Generated for (University of Chicago) on 2014-12-26 11:20 GMT / http://hdl. handle. net/2027/uc2. ark:/13960/t45q4vr7s Public Domain / http://www. hathitrust. org/access_use#pd
? 164 Robert S. Radford [1920
A few words of explanation may be added to the tabular
statement. Ovid began in Catalepton, ix with a proportion
of dactyls which was wholly normal in the year 27 B. C. Thus
he has 44. 8% in the whole distich, which is as good as the
second and third books of Propertius, 28 as good also as the
eighth and third elegies of Tibullus' first book. 29 Yet he has
no limitations whatever in Catalepton, ix upon polysyllabic
endings in the pentameter; on the contrary, like Catullus
and like Propertius in his earlier work, he fairly revels in
their use (50%). In the Lygdamus elegies, however, and
in the Sulpicia letters, the ambitious and aspiring youth
seeks suddenly to pass from the longer endings to the more
elegant dissyllables of Tibullus, and is evidently preoccupied
with this problem and its difficulties. Under these conditions
throughout the Lygdamus poems he wavers greatly in his
composition, and, in one hundred and seventy verses (Lygd.
4 and 5, [Tib. ] rv, 7-12), he even sinks to the proportion of
Catullus, namely, about 37% in the distich, and to only
50% of dactylic beginnings. 30 This is of course excellent
Latin elegy, but it is not the kind that was most in vogue in
24 B. C. 31 It is only in the qualified or limited sense just ex-
28 Propertius has 44. 8% and 44. 7% in the second and third books respec-
tively. My figures are taken from Hultgren, op. cit. 23, who follows the five-
book division of Propertius.
29 Tibullus has 44. 8% in the eighth elegy (Pholoe) for the whole distich
and 45% in the third elegy. It is scarcely fanciful to see in the decline of
dactyls in the third elegy an expression of Tibullus' sadness and depression of
feeling during his illness at Corcyra ; we have the same phenomenon in Lyg-
damus' fifth elegy. My figures are taken from Cartault, Le distiqitc elegiaque
chez Tibitlle, Sulpicia, Lygdamus (Paris, 191 1), 7.
30 Figures for Catullus are given by Hultgren, op. cit. 15 ff. Catullus has
usually 58% to 6o r i of dactylic beginnings, but in Carm. 69-119 (319 vss. )
he has 36. 6^ of dactyls in the distich and only 47^ of dactylic beginnings.
31 Catullus wrote some of the best Latin elegy, and the naturalness and
directness of his style is due in part to his not exceeding the proportion named.
As I have shown in my "Licensed Feet in Latin Verse," op. cit. 251-27? ,
even the best Latin poets, such as Catullus and Horace, experience some diffi-
culty in always providing one required dactyl, and therefore they occasionally
admit without metrical ambiguity in such a foot exceptional or vulgar short-
enings and even short vovels (without m) in hiatus, as Lucilius, ix, 243 Bahr. ,
? ? Generated for (University of Chicago) on 2014-12-26 11:20 GMT / http://hdl. handle. net/2027/uc2. ark:/13960/t45q4vr7s Public Domain / http://www. hathitrust. org/access_use#pd
? Vol. li] The Juvenile Works of Ovid 165
plained that Ovid can be said to have ever been a disciple
of Catullus in the matter of dactylic proportions, but un-
doubtedly in his youth he paid this brief tribute to the freer
and more natural style of his great predecessor. A secondary
cause for the low proportion may perhaps be found in the sad
and almost despairing character of the two elegies.
It will be noted that Ovid advanced in the Sulpicia and
Messalinus poems which are composed in the elegiac metre
to about 47% of dactyls in the who! e distich ; the hexameter
lines alone of these poems show 46. 1% and 45. 9% respectively.
On the other hand, the hexameters of purely epic poems like
the Ciris fall back to 43. 5%, which is normal for epic verse
o 5 re cSrupto (necessary dactyl of the hexameter); Lucr. vi, 1133, natu 5 ra
cSruptum; Cat. 10, 26, istos c6mmoda; nam volo d Sar&pim (necessary
dactyl of the Phalaecean) ; Hor. Cann. in, 14, n, iam virum expertae, malS
6minatis (short vowel in hiatus, necessary dactyl of the Sapphic) ; Pers. 3, 66,
di'scitfi, 6 miseri (license of the first foot, with greatly preferred dactyl) ; Lux-
orius, 302, 4, magnum depre 5 nderg usum. If difficulty is experienced in supply-
ing one required dactyl for the Sapphic or the hexameter, it is clear that two
necessary dactyls, as in the pentameter, constitute a very exacting demand
upon the Roman language, and if the virtuosity is also insisted upon, a very
elegant but very artificial form of verse is the result. For example, for metrical
reasons, Ovid (like Tibullus) constantly uses ab arte and similar phrases drawn
from the vulgar language, with its analytical tendency, instead of the simple
ablative, and he also uses, by poetic license, to an unparalleled extent, the
simple ablative for the ablative of the agent with ab, as Her. 4, 64, capti
pare 5 ntg sor6r, ' my sister was captivated by your parent,' where neither
parenti nor a parente could enter the verse ; upon the whole subject see in part
Guttmann, Sogenanntes instrumentales ab bei Ovid, Dortmund, 1890. I was
mistaken therefore in my former discussion (p. 271) in thinking that Ovid had
perhaps actually used the spelling parente for the dative parenti, as he so often
uses mare and caeleste for mari and caelesti, and as Statius, Sih. iv, 2, 28, uses
glaucae certantia Do b ride saxa, and Propertius, v (rv), 8, 10, writes cum temere
anguino creditur o^rl manus (Xeue-Wagener, Formenlehre, i 3 , 301). The
honest 37% of Catullus and Lygdamus does not compel a resort to such un-
usual constructions, and is therefore by no means to be despised. Dissen and
Postgate (Selections from Tibullus, p. XLIII), on purely subjective grounds and
without consulting indices, lexicons, or Latin authors, have discovered that
Lygdamus is an author of " poor Latinity. " On the contrary, his Latinity is
more natural and in some respects better than that of the mature Ovid. Yet
even the latter like Euripides, an unrivalled master of the graceful and
pleasing forensic style could give most of us lessons in correct Latinity to
our great and lasting profit.
? ? Generated for (University of Chicago) on 2014-12-26 11:20 GMT / http://hdl. handle. net/2027/uc2. ark:/13960/t45q4vr7s Public Domain / http://www. hathitrust. org/access_use#pd
? i66 Robert S. Radford [1920
and almost precisely the proportion of Vergil. In fact that
strange and anomalous thing, the actual preponderance of
dactyls in a Latin epic poem, could never by any possibility
arise in the more vigorous epic hexameter (versus fortis or
grams) ; it must appear first in the soft and tender strains of
the elegy (versus mollis or levis) , and may be transferred thence
to epic verse, as Ovid transferred it in the Metamorphoses
and Valerius Flaccus later employed it in imitation of Ovid.
Even Ovid in his mature period has recognized this difference
to some extent; for the hexameter of elegy in the Heroides
(first series) reaches 56. 4%, while the epic hexameter of the
Metamorphoses does not exceed 54. 8%.
It is evident then that in the first edition of the A mores
which was published in 14 B. C. , 32 only a few years after the
Sulpicia elegies, the Ciris, and the Aetna, Ovid had had little
opportunity to develop a marked dactylic virtuosity and to
become a highly artistic elegiac poet. We may be sure then
that the first Amores of 14 B. C. showed only a very moderate
advance upon the 47. 4% of the Sulpicia elegies, and they
may even have contained, like the latter, a few polysyllabic
endings in the pentameter. Up to 14 B. C. Ovid was in fact
fully as much an epic as an elegiac poet (see Am. i, i, 1-18
on his early ambition to excel in epic themes), and he was
still far from having attained in the Sulpicia poems the high
virtuosity which appears in three of the four genuine elegies
of Tibullus' second book, namely, 55% or more. 33 Two of
the short Sulpicia elegies and the equally short Cornutus
poem show, it is true, an equal or a superior number of dac-
tyls, 34 but four Sulpicia poems and the long Messalinus elegy
32 Schanz, op. cit. 270 (293), suggests that these elegies probably came
first separately before the public, but later were collected in the edition of
14 B. C. This seems quite probable, especially as the spondaic Amores show
in general the predominance of the same schemata or figures of the hexameter
as the Sulpicia and Messalinus poems.
33 The first elegy of the second book has 55. 2% in the distich, the third has
50%, the fourth 55%, and the sixth 56. 8%; see the figures of Cartault, op.
cit. 7, and Ehrengruber, op. cit. x, 5.
34 iv, 6 (20 vss. ) has 50% of dactyls ; iv, 5 (20 vss. ) 58. 3% ; n, 2 (20 vss. ) 50%.
? ? Generated for (University of Chicago) on 2014-12-26 11:20 GMT / http://hdl. handle. net/2027/uc2. ark:/13960/t45q4vr7s Public Domain / http://www. hathitrust. org/access_use#pd
? Vol. li] The Juvenile Works of Ovid 167
(n, 5) have a decided preponderance of spondees. A marked
change in Ovid's whole attitude took place, however, after
14 B. C. For unlike the Ciris and the Aetna, which did not
greatly increase their author's reputation, the Amores, im-
mediately upon their first publication, achieved a prodigious
success; they at once became popular favorites, and, like
the Eclogues of Vergil, were frequently sung in the theatre
with accompanying dance (Trist. n, 519; v, 7, 25). The
favorable reception thus accorded to the elegies naturally
determined the direction of the poet'3 genius and led him to
devote himself uninterruptedly (with the exception of the
tragedy of Medea) for sixteen years (14 B. C. to 2 A. D. ) to the
perfection of an elegiac style which surpassed even the later
work of Tibullus.
Having developed a remarkable and distinctive technique
in the Heroides and the Ars, Ovid resolved to make the first
Amores conform fully to his later rules of art.
The revision
was not due to the fact that this popular and successful work
was poetically immature, nor even chiefly to the desire to
add a series of new elegies ; as he himself tells us in the pre-
fixed epigram, the question was primarily one of more careful
finish and elaboration : " hoc illi praetulit auctor opus "
(cf. Gruppe, op. cit. 377). In brief, his principal purpose
was to eliminate all polysyllabic endings, to introduce every-
where the full dactylic virtuosity, to multiply the dactylic
beginnings, and to remove or greatly reduce those schemata,
such as DSSS, SDSS, 5555, 55, SD, and the like, which
were no longer fully acceptable. The view adopted by
Heuwes in his discussion of the matter (op. cit. 31) is that
the revision " consisted in here and there substituting a word
that was more suitable for one that was less suitable to the
metre and sense, or in changing the collocation of words o 1
verses, or in doing all these things at the same time. " These
conclusions are partly correct, but there can be no doubt that
both the additions and the alterations were far more extensive
and thoroughgoing than Heuwes supposes. The first edition
? ? Generated for (University of Chicago) on 2014-12-26 11:20 GMT / http://hdl. handle. net/2027/uc2. ark:/13960/t45q4vr7s Public Domain / http://www. hathitrust. org/access_use#pd
? 1 68 Robert S. Radford [1920
was in fact subjected to a drastic revision and was almost
entirely rewritten in conformity to the new rules of art. 35
Nevertheless abundant traces of the character of the first
edition still remain ; for in not a few cases the poet has
triumphed over the versifier and the artist, and a number
of the original poems of the first edition have been retained
with very moderate changes. This fact may be shown very
clearly in two ways, if the writer correctly understands the
true nature of Ovidian versification. For the full and har-
moniously developed virtuosity in the Ovidian sense means
first that there shall be a considerable preponderance of
dactyls in the distich taken as a whole, and secondly that
there shall be a similar preponderance of dactyls in each
member taken singly, that is, in both the hexameter and the
pentameter lines. It is evident therefore that virtuosity is
lacking in the Amores in two ways: (i) when the spondees
are either equal or predominant in the whole distich ;
(2) when they are equal or predominant in the hexameter
line (versus fortis or diirus) alone. With this understanding
of the nature of virtuosity, I may summarize the facts briefly
as follows. A considerable part of the second edition is
entirely new and consists of such newly written elegies as
n, 17 and 1 8 ; in, i and 15 ; and also the last 18 lines of in.
9 (the epicedion upon Tibullus). 36 A still larger part has
been completely or almost completely revised in order to
bring the elegies up to practically the same virtuosity as the
35 The extraordinary importance which the mature Ovid attached to formal
poli. -h and elegance is not only evident from the exacting rules \vhii h he strictly
observes, but also from his own express statements. Thus he speaks of de-
stroying many works which would have won popular favor, but which he
himself considered " faulty " (I'itiosa, Trist. iv. 10, 61"), he often laments that
the poems of the exile are composed with less care and skill than was his wont
(Pont, i, 5, 15 ff. ; 57 ff. ), and he bitterly complains that the Metamorphoses
were uncorrected and lacked the finishing touches at the moment of his banish-
ment, as in Trist. in, 14, 21, illud opus . . . nunc incorrectiun populi pervenit
in ora. See also ib. i, 7, 27 ff. , defuit et scriptis ultima lima meis; i, 7, 39 ff. ;
H, 555-
36 On the new elegies, see also Schanz, op. cit. 270 ( 293).
? ? Generated for (University of Chicago) on 2014-12-26 11:20 GMT / http://hdl. handle. net/2027/uc2. ark:/13960/t45q4vr7s Public Domain / http://www. hathitrust. org/access_use#pd
? Vol. li] The Juvenile Works of Ovid 169
Ars or the Heroides (first series), namely, 57% in the dis-
tich ; 37 about a fourth part of the whole edition, however,
shows only slight and moderate changes from its original
form. That is, of the forty-nine elegies, 38 fourteen have been
only partially and imperfectly revised. A single poem of
66 verses (in, 8) remains at practically the same average as
the Sulpicia elegies, namely, 47. 98% of dactyls in the distich ;
four other poems, however, of 184 verses (i, 2. 13. 15 ; in, 10)
remain with the dactyls subordinate in the hexameter line,
and either equal (only 50%) or subordinate (49. 23%) in the
whole distich. Five elegies (i, 14 ; 39 n, 4. 8. n; in, 3), of
236 verses, also remain with the dactyls either subordinate
or equal in the hexameter line, and with 52. 4, 50. 7, 51. 2,
50. 6, 50. 7% respectively in the whole distich. Two elegies
(n, 12. 14) of 72 verses, show a weak pentameter and only
51. 2 and 52. 3% in the distich; two other elegies (in, 14 and
the first three-fourths of in, 9, the epicedion on Tibullus,
composed in 19 B. C. ), with 100 verses, are also low in their
proportion of dactyls, namely, 53. 3 and 51. 3%. Further-
more, very largely the same identical schemata are predomi-
nant in all these elegies as we find preferred in the Sulpicia
elegies (iv, 2-6) and in the imitation of Tibullus (iv, 13).
Thus fourteen elegies of 658 verses more than one-fourth
of the total number conspicuously lack the virtuosity ;
among these are included some of the most notable poems in
the collection, such as i, 2 (the poet's willing submission to
Love), i, 13 (Aurora and Tithonus), i, 14 (Corinna's use of
dyes), i, 15 (epilogue on the immortality of poesy), n, 4
(reasons for love), n, n (lament over Corinna's voyage),
37 56. c/~c in the Ars (Drobisch), and 57. 6^ in the Heroides.
3S There are really fifty elegies, as Gruppe, op. cit. 377, first pointed out,
and as Ehwald, the latest editor, obtains, by breaking up n, o into two poems.
In order, however, to avoid confusion and to be uniform with the results of
Hultgren and Drobisch, my statistics (like theirs) are everywhere based upon
the edition of Merkel, Leipzig, 1887.
39 This is the poem which, through its reference to the subjugation of the
Sigambri (15 B. C. ), dates the whole original collection.
? ? Generated for (University of Chicago) on 2014-12-26 11:20 GMT / http://hdl. handle. net/2027/uc2. ark:/13960/t45q4vr7s Public Domain / http://www. hathitrust. org/access_use#pd
? 170 Robert S. Radford [1920
n, 12 (Corinna's surrender), n, 14 (the evils of abortion,
with several Lygdamus verses repeated), in, 3 (Corinna's
perjuries unpunished), in, 9 (epicedion on Tibullus, 50 out
of 68 verses), in, 14 (Corinna's infidelities best unknown).
The percentage of the nine ' imperfect ' elegies of Books I
and in is 50. 5% of dactyls for the whole distich, of the five
'imperfect' elegies of Book n, 51. 1%, and the difference
between the perfect and the imperfect parts of the second
Amores is 6% for the four poems of Book I, 5. 3% for the
five of Book 11, and 4. 8% for the five of Book in. 40 I think
it quite reasonable, however, to assume that, even in the case
of the imperfect elegies, at least 2. 3% of actyls have been
added in the revision, and I thus reach the conclusion that
the percentage of the first Amores did not exceed 48. 5% of
dactyls for the whole distich, and therefore rose only slightly
above the 47. 4% of the Sulpicia elegies and of iv, 13-14.
Though less important than the sum of the dactyls, the
proportion of dactylic beginnings also requires mention.
The percentage of dactylic beginnings in both hexameter and
pentameter in Am. I is 80 ; in the four imperfect elegies
(i, 2. 13. 14. 15) it is 73. 4, in the remainder of the book it is
82. 2, thus giving a difference of 8. 8% between the two parts.
We may consider as normal for the mature Ovid the per-
centage in both hexameter and pentameter of the Ars, which
is 82. 3, or that of the Fasti, which is 84. 2. Similarly the
percentage of dactylic beginnings in the whole of Am. 11 is
78. 6; in the five imperfect elegies (n, 4. 8. n. 12. 14) 41 it is
71. 6, in the remainder of the book it is 80. 9, giving a difference
of 9. 3% between the two parts. With these five elegies of
Am. ii we may fitly compare not only the percentage of the
Sulpicia elegies (iv, 2-6) and iv, 13-14, which is 69. 7% in
142 verses, but also that of the early Medicamen Faciei frag-
40 In the case of in, 9, the Tibullus epicedion, only the first 50 verses out of
68 are used.
41 n, q is also spondaic if \vc cut out the six verses (23-28) which Ovid
seems to have added in the second edition in order to join the originally sepa-
rate poems 9 and 9 H.
? ? Generated for (University of Chicago) on 2014-12-26 11:20 GMT / http://hdl. handle. net/2027/uc2. ark:/13960/t45q4vr7s Public Domain / http://www. hathitrust. org/access_use#pd
? Vol. li] The Juvenile Works of Ovid 171
ment of 100 verses, which is 70, while its percentage of dactyls
for the distich is also low, namely, 53. 5- 42 The percentage
of dactylic beginnings in the whole of Am. in is 77. 1 ; in the
four imperfect elegies (in, 8. 10. 14, first 50 vs. of 9) it is
74, in the remainder of the book it is 78. 5, giving a
difference of 4. 5 between the two parts. I consider this brief
summary of usage affecting the first foot sufficient for the
practical purposes of the present study, and in my subsequent
discussion shall purposely omit this feature of the single elegies
from the tabular statements. I may add, however, that the
percentage of dactyls in the first foot in Catalepton, ix is
71. 9; owing to preoccupation with the dissyllabic close and
to imitation of Catullus, it sinks in the Lygdamus elegies to
55-8 43 and in the Sulpicia letters (iv, . 7-12, 40 verses) to 50;
it rises again to 67. 7 in the Panegyric, 66. 8 in the Ciris, and
67. 1 in the Cnlex. ^
42 Hultgren, op. cit. 28 ff. , argues that no importance should be attached
lo so small a fragment, yet at the same time (pp. 32 f. ) shows the greatest
uneasiness because of its evidence, apparently fearing that it may bring to
naught the whole elaborate fabric of Ovidian virtuosity. One wonders what
he would have thought if he had examined the single Amorcs.
4:! It varies greatly in single Lygdamus elegies from 82. 2, 61. 8, and 65. 7
in i, 5, and 6 to 53. 4, 55. 2, and 45. 8 in 2, 3, and 4. In the technical Halieutica
it sinks also to 47. 7.
44 On account of the length of the article it has been found necessary to omit
here five tables giving exact statistics for the fourteen elegies. Two articles,
which continue the present study and treat the language and schemata of
Book iv of the Tibulline Corpus, will be published shortly in the American
Journal of Philology.
? ? Generated for (University of Chicago) on 2014-12-26 11:20 GMT / http://hdl. handle. net/2027/uc2. ark:/13960/t45q4vr7s Public Domain / http://www. hathitrust. org/access_use#pd
? ? ? Generated for (University of Chicago) on 2014-12-26 11:20 GMT / http://hdl. handle. net/2027/uc2. ark:/13960/t45q4vr7s Public Domain / http://www. hathitrust. org/access_use#pd
? UNIVERSITY Ol- t A1_II-WNIN I A A I 1_<JS AIN
THE UNIVERSITY LIBRARY
This book is DUE on the last data stam^'
-f CALIFOKNTJ
AT
? ? Generated for (University of Chicago) on 2014-12-26 11:20 GMT / http://hdl. handle. net/2027/uc2. ark:/13960/t45q4vr7s Public Domain / http://www. hathitrust. org/access_use#pd
? ? ?
