ILLINOIS

UNIVERSITY OF ILLINOIS AT URBANA-CHAMPA1GN

PRODUCTION NOTE

University of Illinois at
Urbana-Champaign library
Brittle Books Project, 2014.COPYRIGHT NOTIFICATION

In Public Domain.
Published prior to 1923.

This digital copy was made from the printed version held
by the University of Illinois at Urbana-Champaign.
It was made in compliance with copyright law.

Prepared for the Brittle Books Project, Main Library,
University of Illinois at Urbana-Champaign
by

Northern Micrographics
Brookhaven Bindery
La Crosse, Wisconsin

2014ILUN(*

Extracted from the Transactions of the American Philological
Association, Vol. xxxv, 1904.

IV. — The Accentus of the Ancient Latin Grammarians.

bY dR. charles w. l. johnson,

BALTIMORE, MD.

The evidence from which we must derive our conception
of the nature of the accentuation of the Latin language is
very complex. There is no unanimity among scholars even
upon the general character of the accents, although it may
be claimed that a considerable majority hold the view that
these are stress-accents. They base their conclusions chiefly
upon a study of Indo-European accentuation in general, and
upon the indications which the etymology of the language
and that of the languages derived from Latin afford. The
direct evidence of the ancients, however, is strongly in favor
of a melodic or musical accent, similar to the Greek accent,
if not actually identical with it. An explanation of this state
of affairs is often given, to the effect that the ancient Latin
grammarians were so completely under the influence of the
Greek scholars to whom they owed their erudition that they
imposed upon the Latin language a Greek doctrine for which
there was little or no justification in the facts. But one can
not long remain satisfied with this. Is it possible for such an
imposition to hold its ground for centuries ? Is it possible
for men, whose judgement and whose statements upon other
matters we accept as sound, to be so self-deceived or so un-
scrupulous ? Perhaps so; but even then it is our duty to
dispose of their evidence by showing clearly how their mis-
apprehension arose, or why they conspired to spread false-
hood. I do not know that this has been done. At any rate,
here is one aspect of the subject of Latin accentuation which
deserves consideration. An exhaustive review of the situa-
tion might result in a reconciliation of the two opposing views
upon the nature of the Latin accent.

So far as the scope of the present paper is concerned, we
shall put narrow limits upon it. We shall ask, not " What was66

Charles W. L. Johnson.

[1904

the Latin accent ?" but " What did the ancient Latin gram-
marians mean to be understood by the term accentus?" It
is almost as important to know what they meant, as to know
what were the facts; to know what they thought they heard,
as to know what they really heard.

In the first place it must be admitted by all that the changes
in pitch which accompany speech received recognition at the
hands of the ancients, no less when the Latin language was
under consideration than when the Greek language under-
went an analysis of its character. The musical theorists and
the grammarians attacked the subject from slightly different
points, and their evidence is clear that in Greek speech there
ran through the sentences a sort of melody, and that in Latin
also of the period when the language was brought within the
purview of Greek scholarship there existed a melodic element
which resembled that of the Greek sufficiently to be identified
with it. When Cicero (Orator, 57) speaks of a cantiis obscu-
rior, he is referring to the Latin language, such, presumably,
as it was in his own time. The existence, at any rate, of these
pitch-changes was acknowledged. Indeed, we could expect
little else with the example of the Greeks before them. But
we must now ask, " How much further did the ancient stu-
dents of the Latin language go in their adoption of the Greek
terminology, and what were the facts upon which they based
their work ?"

The Greek term for the melodic element in the constitution
of wTords and sentences was oftenest 7rpoaotyBta. Other terms
are found, like the XoY<wSe? tl fieXos of Aristoxenus (Harmo-
nica Elementa, I, § 42, p. 18 M, p. 17 W), like the htakeicTov
fieXo? of Dionysius of Halicarnassus (de Compositione Verbo-
rum, XI) and his to rrjs	/jceXos, Aeyco 8* ov tt}? (pSifcfy;

aXXa tt}? ijriXris and ra jiexr} tmv <p06yycop {ibid. XI, fin.)y but
the most usual expression was Trpoo-ahta, and the diacritical
marks with which Aristophanes of Byzantium endowed the
language, the written accent signs, were called 7rpoawSiai, as
well as racT66? and tovol.1

1 Dionysius of Halicarnassus, de Comp. Verb. xix: rdaeis <pojvrjs ai KaXov/xivat
7rpo<T(fdlai.Vol. xxxv.] Accentus of the Ajicient Latin Grammarians. 67

This term Trpoacphia was taken over into the Latin language
in the form accentus, and in those passages where the term
accentus is explained it is asserted, as clearly as words can
assert anything, that the Greek irpoacphia and the Latin accen-
tus are one and the same thing.

Aulus Gellius, XIIT, 6: quas Graeci irpocrcphlas dicunt, eas
veteres docti turn 'notas vocum,' turn 4 moderamenta,' turn
' accentiunculas,' turn ' voculationes ' appellabant.

Diomedes, Keil, I, p. 431, 1 : accentus est dictus ab acci-
nendo, quod sit quasi quidam cuiusque syllabae cantus. apud
Graecos quoque ideo irpocrcphla dicitur, quia irpooraSerat rals
avWa{3aL$.

Idem, K. I, p. 456, 18 : tenor, quem Graeci dicunt tasin aut
prosodian, in flexibus vocis servandus est.

Servius, de Fin. K. IV, p. 451, 10: accentus autem est
quasi adcantus dictus, quod ad cantilenam vocis nos facit
agnoscere syllabas.

Martianus Capella, III, p. 65, 19, Eyssenhardt: et est
accentus ut quidam putaverunt anima vocis et seminarium
musices, quod omnis modulatio ex fastigiis vocum gravitate-
que componitur, ideoque accentus quasi adcantus dictus est.

There is, so far, no suggestion that the imported Greek
doctrine of the accents needed amendment to apply to the
Latin language. The phenomenon is the same, even if
the laws governing it turn out to be different in the two
languages.

Proceeding with their analysis of the sounds of the Latin
language, the grammarians report that the language pos-
sessed, like the Greek, three accents, — the acute, the grave,
and the circumflex. These are musical terms and translate
Greek words in which also the musical idea is obvious. Just
as Dionysius of Halicarnassus speaks from the standpoint of
a musician, when he uses the expressions rj o%ela rdais, rj
fiapela rdaL?, o 0%vs toVo?, etc., so do the Latin grammarians
employ musical terms without the least reservation or warn-
ing, when in analysing the Latin accentus they make their
classification of accentus acutus, accentus gravis, and accentus
circumflex us.68

Charles W. L. Johnson.

[I9°4

Varro ap. Sergium, K. IV, p. 531 : prosodiam ibi esse dici-
mus ubi aut sursum est aut deorsum. quae demissior est a
pluribus ftapela appellatur Graece, Latine vero gravis, ideo
quod deorsum est in sede scilicet ponderum graviorum. at earn
quae sursum est Glaucus iiriTeraixev^v^ item alius aliter, sed
nemo adhuc levem vocavit, quamvis id erat gravi contrarium ;
verum ea nomen obtinet o^elav, Latine acuta, ideo quod tenuis
et omne acutum tenue. inter has est pecrr), Latine media,
quia limes est per quem duae supra dictae ultro citroque com-
meant. quartae illi, quia ceteris perplexior est, plura sunt
vocabula [i.e. d^vftapvs, TreplairacnSy Bltovos, crvfjl7r\€/ct0<
KeKXaa/ievr)~\. verum ea nunc ab omnibus TrepLaTrco/jLevr)
Graece vocatirr, apud nos flexa, quoniam prima erecta rursus
in gravem flectitur.

Nigidius Figulus ap. Aul. Gell. XIII, 25 : P. Nigidii verba
sunt ex commentariorum grammaticorum vicesimo quarto,
hominis in disciplinis doctrinarum omnium praecellentis:
Deinde, inquit, voculatio qui poterit servari, si non sciemus in
nominibus, ut Valeri, utrum interrogandi an vocandi sint?
nam interrogandi secunda syllaba superiore tonost quam
prima, deinde novissima deicitur; at in casu vocandi summo
tonost prima, deinde gradatim descendunt. sic quidem Ni-
gidius dici praecipit. sed si quis nunc, Valerium appellans,
in casu vocandi secundum id praeceptum Nigidii acuerit pri-
mam, non aberit quin rideatur. 4 summum ' autem 4 tonum '
irpoacphiav acutam dicit et quem 4 accentum' nos dicimus 4 vo-
culationem' appellat et 4 casum interrogandi' eum dicit quem
nunc nos 4 genetivum ' dicimus.

Vitruvius, de Archit. V, 4, 2 : vox enim mutationibus cum
flectitur alias fit acuta, alias gravis, duobusque modis movetur,
e quibus unus affectus habet continuatos, alter distantes.
[Aristoxenus's continuous and intervallar motions of the
voice are then given.] 1

Donatus, K. IV, p. 371, 1 : tonos alii accentus, alii tenores
nominant. toni igitur tres sunt, acutus gravis circumflexus.

1 See the writer, "The Motion of the Voice," in Trans. Am. Phil. Assoc.
XXX, p. 45, and " Accent and Accentual Arsis and Thesis," in Studies in Honor
of B. Z. Gilder sleeve, p. 57.Vol. xxxv.] Accentus of the Ancient Latin Grammarians. 69

Diomedes, K. I, p. 430, 29: accentus est acutus vel gravis
vel inflexa elatio orationis vocisve intentio vel inclinatio acuto
aut inflexo sono regens verba, nam ut nulla vox sine vocali
est, ita sine accentu nulla est, et est accentus, ut quidam recte
putaverunt, velut anima vocis.

Sergius, K. IV, p. 482, 14: his ita se habentibus sciendum
est quod acutus et gravis et circumflexus soli sunt qui, ut
superius diximus, naturalem unius cuiusque sermonis in voce
nostra [Keil: vocem nostrae] elationis servent tenorem.

Priscian, K. II, p. 51, 21 : accidit unicuique syllabae tenor,
spiritus, tempus, numerus literarum. tenor acutus vel gravis
vel circumflexus.

Ps.-Priscian, K. Ill, p. 519, 25 : accentus namque est certa
lex et regula ad elevandam et deprimendam syllabam uniuscu-
iusque particulae orationis, qui fit ad similitudinem elemento-
rum, litterarum, syllabarumque, qui etiam tripertito dividitur,
acuto gravi circumflexo. acutus namque accentus ideo in-
ventus est, quod acuat sive elevet syllabam; gravis vero eo,
quod deprimat aut deponat; circumflexus ideo, quod deprimat
et acuat.

The next step is to determine the laws governing the dis-
position of the accents. It was found that every Latin word,
with the exception of those we call enclitic and proclitic, had
one point of acuteness and one only. As in the Greek theory
of accentuation, so in the Latin, every syllable was regarded
as having one of the three accents, but it was not considered
important either to mark the grave accents or to refer in
every case to their existence, provided the other accents were
attended to. Both of these contained a high pitch, for the
circumflex was but a combination of acute and grave, and
one or the other was found to be always present in a word.
Writers would then speak of the accent of a word when they
meant the acute or the circumflex, that is, the acute element.
With this understood they formulated the law of Latin ac-
centuation, by which the accent was confined to the last three
syllables of any word (in which particular it was like the rule
for Greek), and, furthermore, had the restriction placed upon
it that, except in the case of monosyllables, it could not occupy70

Charles W. L. Johnson.

[1904

the last place (in which particular it differed from the Greek).
But it was not sufficient to find the position of the accent
only ; it was necessary also to determine its kind, and for this
there were laws which stated the facts as observed. A long
penultimate syllable bore a circumflex if it was long by nature
and the last syllable was short, but an acute in any other
circumstances, while a short penultimate threw an acute
accent back upon the antepenultimate in words of more than
two syllables.

This surely is the meaning of the following passages :

Cicero, Orator, 57-58: mira est enim quaedam natura
vocis, cuius quidem e tribus omnino sonis, inflexo, acuto,
gravi, tanta sit et tam suavis varietas perfecta in cantibus. . . .
in quo illud etiam notandum mihi videtur ad studium perse-
quendae suavitatis in vocibus : ipsa enim natura, quasi modu-
laretur hominum orationem, in omni verbo posuit acutam
vocem, nec una plus nec a postrema syllaba citra tertiam, quo
magis naturam ducem ad aurium voluptatem sequatur in-
dustria.

Quintilian, hist. Orat. I, 5, 30: namque in omni voce acuta
intra numerum trium syllabarum continetur, sive hae sunt in
verbo solae sive ultimae, et in his aut proxima extremae aut
ab ea tertia. trium porro, de quibus loquor, media longa aut
acuta aut flexa erit, eodem loco brevis utique gravem habebit
sonum, ideoque positam ante se id est ab ultima tertiam acuet.
est autem in omni voce utique acuta, sed numquam plus una
nec umquam ultima, ideoque in disyllabis prior; praeterea
numquam in eadem flexa et acuta, quia in omni flexa est
acuta, itaque neutra cludet vocem latinam. ea vero quae
sunt syllabae unius, erunt acuta aut flexa, ne sit aliqua vox
sine acuta.

Cicero's acuta vox is the acute element which characterizes
every word, whether marked as an acute or as a circumflex
accent. Scholl1 argues that Cicero was not acquainted with
a circumflex like the Greek circumflex, or he would have
referred to it in the passage quoted above, which seems to
mention only an acute accent. But in the passage from

1 Scholl, De accentu linguae latinae, Acta Soc. Philol. Lips. VI, p. 34.Vol. xxxv.] Accentus of the Ancient'Latin Grammarians. 71

Quintilian it is clear what is the point of view of both writers,
for the latter explicitly states that an acuta is a part of every
Jlexa. The doctrine hangs well together.

Other passages where the laws of accentuation are given
are the following:

Donatus, K. IV, p. 371, 1-30; Diomedes, K. I, p. 431, 6 ff.;
Servius, K. IV, p. 426, 6-16; Sergius, K. IV, p. 482, 19 ff.;
Ps.-Serg. K. IV, p. 524, 21 ff.; Pompeius, K. V, p. 127, 12 ff.;
Ps.-Priscian, K. Ill, p. 520, 17 ff.; p. 521, 5 ff.

In these passages wordsvare given which show all possible
combinations of quantities, and the proper accent for each is
stated.

Let us now consider some of the difficulties which confront
those who see in all these passages a perverse determina-
tion to force upon the stress-accents of the Latin tongue
rules which were invented for the musical accents of another
language.

If, under the term "grave accents," we are to understand
unstressed syllables, it follows that stressed syllables are those
which are called in the theory acute or circumflex. The
question then arises, Was such an acute accent identical with
the circumflex in everything but name ? That is, was the
difference purely imaginary ? The assumption forces us into
a curious attitude toward the writers quoted. We must be-
lieve that they went to the trouble to expound rather com-
plicated rules for the fixing of these accents upon the proper
syllables, when not only was this unnecessary from the point
of view of the written language, but the distinction they drew
could not even be detected by the ear. The statement that
the circumflex is an acute followed by a grave becomes abso-
lutely meaningless with the new definitions. In the case of
monosyllables, the absurdity of the grammarians' conduct is
even greater, for here the presence of a stress-accent hardly
declares itself until the word is joined with other words to
form a sentence or a phrase.

The following from Donatus is an example of the manner
in which this subject is dealt with by the grammarians:72

Charles ■ W. L. Johnson.

[I9°4

Donatus, K. IV, p. 371, 8: ergo monosyllaba, quae cor-
reptam vocalem habebunt, acuto accentu pronuntiabimus, ut
fax, p 1 x, nux; quae productam vocalem habebunt, circum-
flexo accentu pronuntiabimus, ut r e s, d o s, s p e s. in disyl-
labis quae priorem productam habuerint et posteriorem
correptam, priorem syllabam circumflectemus, ut meta,
Creta: ubi posterior syllaba producta fuerit, acuemus
priorem, sive ilia correpta fuerit sive producta, ut n e p o s,
leges: ubi ambae breves fuerint, acuemus priorem, ut
bonus, malus. In trisyllabis et tetrasyllabis et deinceps,
si paenultima correpta fuerit, acuemus antepaenultimam, ut
T u 11 i u s, Hostflius: si paenultima positione longa fuerit,
ipsa acuetur et antepaenultima gravi accentu pronuntiabitur,
ut Catullus, M e t e 11 u s, ita tamen si positione longa non
ex muta et liquida fuerit; nam mutabit accentum, ut latebrae,
tenebrae: si ultima brevis fuerit, paenultima vero natura
longa, paenultima circumflectetur, ut Cethegus, perdsus:
si ultima quoque natura longa fuerit, paenultima acuetur, ut
Athenae, Mycenae.

Here and in the other passages it is not a case of confu-
sion between quantity and accent, although the latter depends
upon the former.

A number of writers comment on the contrast presented
by the two languages, Greek and Latin, in the matter of the
laws governing their accentuation. Now, if the Latin accent
was radically different from the Greek, what point is there
in these comments ? Quintilian complains that the Latin
language is the poorer for the restriction which prevented an
acute or circumflex from falling on a final syllable.

Quintilian, Inst. Orat. XII, 10, 33: sed accentus quoque cum
rigore quodam, turn similitudine ipsa minus suaves habemus,
quia ultima syllaba nec acuta unquam excitatur, nec flexa
circumducitur, sed in gravem vel duas graves cadit semper,
itaque tanto est sermo Graecus Latino iucundior, ut nostri
poetae, quotiens dulce carmen esse voluerunt, illorum id
nominibus exornent.

Is it likely that Quintilian is here comparing two such
dissimilar things as a Latin stress-accent and the GreekVol. xxxv.] Accentus of the Ancient Latin Grammarians. 73

melodic or musical accent and remained unconscious of the
difference ? And if so, how did the introduction of Greek
words help the Latin poets ?

Athenaeus (.Deipnosoph. X, 425 a) says that the Romans
imitate the Aeolians in all things, even in their accentuation.
Both languages are barytone. It is easier to hold that the
Aeolic dialect was stressed in the time of Athenaeus than
that a faulty comparison was here drawn between essentially
different things.

We are told that Greek words retain their Greek accents
when they occur in Latin :

Donatus, K. IV, p. 371, 27; Servius, K. IV, p. 427, 10;
Sergius, K. IV, p. 483, 33; Ps.-Sergius, K. IV, p. 525, 8;
Diomedes, K. I, p. 433, 4.

Now, what is it that they really retained — a musical ac-
cent or a stress-accent occupying the place of the musical
accent ? Servius on Virgil, Geor. I, 59, remarks that we must
pronounce Epiros with an acute on the first syllable, besides
keeping the Greek termination ~os. See Lindsay, The Lat.
Lang., p. 155. What is the real difference in the pronuncia-
tion ? It can not be a matter of stress accentuation, for the
penult is long in either case, and according to the prevailing
doctrine in regard to ictus, we must suppose the accents to
have become obliterated in verse. If, however, the accents
are musical or melodic, this difficulty of a conflict between
ictus and accent vanishes. Accent and ictus mingle harm-
lessly.

Interjections, moreover, are reported to have no definite
accents. In the case of monosyllabic interjections this re-
mark would seem to be quite superfluous, if their accents are
stress-accents, of which neither definite nor indefinite sorts
are readily conceived.

Donatus, K. IV, p. 371, 24: accentus in integris dictioni-
bus observantur, in interiectionibus et in peregrinis verbis et
in barbaris nominibus nulli certi sunt. Cf.:

Diomedes, K. I, p. 433, 31; Sergius, K. IV, p. 483, 29;
Ps.-Priscian, K. Ill, p. 520, 23.74

Charles W. L. Johnson.

[1904

Certain light is also thrown on the belief of the ancient
writers by their definition of the syllable. The syllable is a
group of elementary sounds uttered in one breath and with
one accentus. This definition contemplates the difference in
the sound of a syllable according as it is spoken with an
acute, a grave, or a circumflex, and the distinction between
stressed and unstressed syllables does not seem to be tangible
enough to call for mention when the syllable is considered, as
it were, in vacuo.

Marius Victorinus, K. VI, p. 26: syllaba est coniunctio lit-
terarum cum vocali vel vocalibus sub uno accentu et spiritu
continuata.

Now the evidence given above is not presented as an ex-
haustive collection of the passages bearing on the subject.
As already stated, the nature of the Latin accent is not to be
derived solely from the testimonia of the ancients. I have
not considered the evidence of the language itself, as revealed
in its etymology, nor the evidence of verse-forms; but then
the field of the inquiry was narrowed so as to exclude all
these. The object was to ascertain what phenomenon in the
Latin language was originally understood by the term accentus.
I believe that it was a melodic or musical accent, but this does
not preclude a belief in the simultaneous existence of a stress-
accent. Such an accent may have existed from early times;
it may have been present in the Saturnian metre and in the
verse of Plautus and Terence; it may never have died out
from the language, even that spoken at Rome; or it may
have done so only to revive again when Latin was in its
decline. We know that the Greek accents were converted
from musical to stressed accents, and the corresponding phe-
nomenon in Latin may have taken place at the same time or
even earlier. At any rate it is curious that the direct evi-
dence of ancient writers should contain so little in support
of a stress-accent. It is not my wish to . conceal such evi-
dence if I could; it is all in Scholl's article. But we must
handle all this grammatical lore of the ancients with constant
recollection of the fact that, while most of the matter is
traditional, some of it may always be original contribution,Vol. xxxv.] Accentus of the Ancient Latin Grammarians. 75

or may be so modified as to become such. Many of the later
grammarians lived at a time when there is no difficulty in
admitting that the accent was stressed. The rules of accen-
tuation were obsolete in their day, and it is not strange that
they should have made positive errors in transcribing the
doctrine which had come down to them, or should have
attempted to make an adjustment to meet the existing situa-
tion. The following passages should be considered together.
They seem to contain the doctrine of Donatus.

Servius, in Don. K. IV, p. 426, 16: accentus in ea syllaba
est quae plus sonat. quam rem deprehendimus, si fingamus
nos aliquem longe positum clamare. invenimus enim naturali
ratione illam syllabam plus sonare quae retinet accentum,
atque usque eodem nisum vocis adscendere.

Pompeius, Comm. Artis Don. K. V, p. 126, 18 : [The grave
accent is superfluous. Every word has either an acute or a
circumflex.] ut puta malesanus ; sa circumflexum habet, ma
le nus istae tres syllabae gravem habent accentum. nam
ideo dictus est gravis hac ratione, quod minus sonet, quam
sonat ille legitimus. ... s a plus sonat. ideo dictae sunt
illae habere gravem accentum, quod et pigrum et minus
sonent.

Idem, K. V, p. 126, 32: ergo ilia syllaba quae accentum
habet, plus sonat quasi ipsa habet maiorem potestatem. et
quo modo invenimus ipsum accentum ? et hoc traditum est.
sunt plerique qui naturaliter non habent acutas aures ad capi-
endos hos accentus, et inducitur hac arte, finge tibi quasi
vocem clamantis ad longe aliquem positum. [In the word
orator it will be found that ra plus sonat.]

Codex Bernensis 16, K. Suppl., p. xlv: accentus est anima
verborum sive vox syllabae, quae in sermone plus sonat de
ceteris syllabis. accentus autem a cantu vocatus est, quia
in ipso cantu producitur modulatio vocis.

Here it is said that the accented syllable of any word sounds
louder than the others. This may very well be true for the
time of Donatus (fourth century), but it is curious that Ser-
vius and Pompeius consider it necessary to give a criterion
for the accent, as if it were a difficult matter to discover it.76

Charles W. L. Johnson.

[1904

The most important question with us is, How much weight
are we to attach to these statements in determining what was
the nature of the Latin accent at the time when the laws
governing it were first formulated ? For that time the weight
of the evidence is for the existence of what we may safely
call, not an accent, but an accentus of a melodic or musical
character.This book is a preservation facsimile produced for
the University of Illinois, Urbana-Champaign.
It is made in compliance with copyright law
and produced on acid-free archival

60# book weight paper
which meets the requirements of
ANSI/NISO Z39.48-1992 (permanence of paper).

Preservation facsimile printing and binding
by

Northern Micrographics
Brookhaven Bindery
La Crosse, Wisconsin
2014