J

ALLITTERATIO LATINA

:

ALLITTERATIO LATINAOR-':'.

ALLITERATION IN LATIN VERSE

REDUCED TO RULE^,

WITH SPECIAL REFERENCE TO CATULLUS, HORACE, JUVENAL, LUCAN, LUCRETIUS, MARTIAL, OVID, PERSIUS, PHAEDRUS, PRIAPEIA,PROPERTIUS,STATIUS,

TIBULLUS,

AND

VIRGIL

BYJ. THE rRESBYTERIAN COLLEGE, CARMARTHEN, AND SOMETIME SCHOLAR OF JESUS COLLEGE, OXFORD

WALTER

EVANS, M.A.

PKINCIPAL OF

Fronte exile negotiumet

dignum

pueris putes

adgressis labor arduus.

Ter,

Maur. 63-5.

LONDON

WILLIAMS AND NORGATE14

HENRIETTA STREET, COVENT GARDEN, W.C.2192 1

*

MA in

TO

THE MEMORY OF

MY

EARLIEST INSTRUCTOR IN LATIN,

MY FATHER,AND TO THATOF

MY BRAVE,

SELF-SACRIFICING, DEVOTED MOTHER,

WHO

SURVIVED HIM

NEARLY HALF A CENTURY(1864-I9IO)

498281

:

PREFACEAbout twelveyears ago I rose from a perusal of Lombroso's

Manand

of

Genius with the knowledge that a tendency to alHteration in verse

and prose was a well-marked feature

of insanity in Italy

;

in

someI

of the illustrations

quoted from the writings of French mattoidsis

thought I saw a general resemblance to what

known

in

Welsh

poetry as cynghanedd

La nomade a mis

la

madonne

A la paterne deQuandLele

Petronne

grand Dacier etait diacre

caffier cultive

du fiacre.attention to the complicated

To me who had not given muchsystem inherited byhighly amusingrallying;

my

countrymen the coincidence appearedoffered, I could

and,

when opportunity

not help

my friends amongNoneof

the bards on their affinity to the lunaticshurt,I

of the South.

took

my

fun in

them seemed much such good part that when

and one

was elected

them Dean ofof

Divinity in the University of Wales he addressed some compli-

mentary verses to

me

in a local journal.

Needless to say, the

characteristic alliteration

was there

in force,I

and

I

examined

it

with

uncommon

interest.

It soI

happened that

was reading Persius at

the time, and

whenacsi

came upon the

line (iv. 33)

unctus cesses et figas in cute solem

it

occurred to

me

to

wonder why the poet had chosen that peculiarwriting could have

ending, and

how such

commanded

the popularity

viii

ALLITTERATIO LATINAsatires are

which the

known

to have immediately attained.

A

casual glance revealed the correspondence

csncts gsncts,andinI then

knew thatI

there was

more

in Latin alliteration

than the

stock illustrations had ever suggested.

Recollecting the statementsfuriously.

Lombroso,

now began

to think

somewhat

Was

the

versification of those

poor lunatics a reversion to type under the

influence of

some

cerebral derangement, or

was the feature which

they shared with the poets of both Wales and ancientelement in

Rome an?

human

nature which even disease could not destroy

I argued that

whatever the explanation might be, the phenomenonrule.

ought to admit of being reduced tounconsciousart.

There was such a thing asuntutored song withof its being imitated

The nightingale

trilled its

an exactness and uniformity which allowed

by mechanical means and perhaps reduced to notation, and I saw no reason why alliteration in Latin verse should not yield its secret,evenif

the Latins themselves did not consciously conform to rule.of puzzles,

I

had always been fondworth the solving.in so doing

and here was one which seemed

I accordiagly set tofirst

work on the Aeneid, andadapted for experimentali.

made mysome

mistake

:

for the bewildering reverberaill

tions of the Virgilian

Hexameter wereof the lines,

purposes, thoughuseful check

and particularly

7,

served as aBaffled

on

myI

successive theories for several years.

in the attempt to discover a single feature that looked like

an

approach to law,

took refuge in the Pentameter, as being a some-

what shorter

line,

and from the Pentameter was driven to

Lyrics,

where, side by side withpresence of such lines as

much

to encourage, I found myself in

cessant flamina tibiae.^

Lenaee sequi deum.^illic

bis pueri die.^

Jupiter in

Ganymede

flavo.*

The

situation

was decidedly embarrassing, but meanwhile the

conviction that there was a law had become an obsession, which,

Hor. C. III. xix.

19.

76.

lU. xxv.

19.

lb.

IV.

i.

36.

lb. IV. iv. 4.

PREFACEnotwithstanding

ix

manyI

bouts of despair, never weakened for moreof course

than a night.

was

aware throughout that nothing

could be finally settled until the pronunciation of the letters wasitself

settled

;

and here theit

difficulties

would have seemed

in-

superable, werein the texts

not for the expectation that, spite of variationsspelling, the poets

and the

would help to solve them.

A

closeI

study of Lindsay's Latin Language

workalso

owe

my

warmest acknowledgments

to which monumental introduced me to thewhomI

German

writers, Stolz,

Seelmann, Marx, and others, to

am

under deep obligations.

But

as

all

my

authorities

either

expressed uncertainty or difiered in their conclusions on points ofcardinal importance to

my

subject, I finally

had recourse to thewhich

ancient grammarians, whose pages I carefully searched for additionalinformation, fortifiedI

by a

large

number

of intractable lines

had by that time accumulated.

Daylight gradually dawned, and

hammered out the scheme which I now present to my readers. It has cost me more time and labour than I care to remember, and I have done my utmost to make it watertight but, though I have much confidence that in the main I have proceededlittle

by

little I

;

on the right

lines, I

am;

far

from thinking that theit is

last

word haslines ofeffect

been said on the subject

andI

not reasonable to suppose that,

working single-handed asverse, with Httle or

have done on over 100,000I

no ear for music,

have never misjudged andifficulty.

or

drawn a

false inference or^

overlooked a

The remarksmissed

of Aulus Gellius

warn me on the one hand thatalliteration^

I

may have

much

that nature and training have given

me no

eyes to see, and

Mons. Loth's analysis of Welshstraining after richness

on the other that, inwhere the data

and symmetryseen

in a sphere

are often uncertain, I

may haveit is

much

that a

Roman would not

have missed.

I have,

however, shirked or obscured no difficulty

that I actually noted, andlittle

on the

faith of this assurance that

my

book basesof

its

claim to the attention of students of the classics.

Such

them

as are

moved

to pursue the investigation can hardlyeasily

fail to improve on the work of a first adventurer, and they may

6 *

VI. (VII.) XX., XIII. xxi. (XX.).

La Metrique

Galloise (Paris, 1900), reviewed

by

Prof. Sir J. Morris Jones of

Bangor

in the Zeitschrift fur cdtische Philologie of 1903.

Xfind that

ALLITTERATIO LATINAwhatis

true of Latin

is also true,

mutatis mutandis, of the

sister-language Greek.

The best

line for

experimental purposes in

Latin

is

the Phalaecian, being short, simple, and available Kara

cTTixov in considerable quantity.

WALTERGrben Hill, Oabmabthen.

J.

EVANS.

CONTENTSINTRODUCTION

xiiF.

ALLITTERATIO LATINAPAGB. .

THE LETTER u {v) G. THE LETTER y H. THE ASPIRATE J. THE LETTERS b, p; g, c (k, q); d, K. THE LETTER / L. THE LETTER w M. THE LETTER n N. LIAISON AND LIGATION O. HIDDEN QUANTITIES (INCLUDING ALPHABETICAL. .

.

.127131

131.

t

.

.

.

133

.

.

.

.

.

.

,

.

.

.

.

.

.

.137 .141 .156 .164LIST)168

INDEX TO THE ILLUSTRATIONS INDEX TO THE EMENDATIONS GENERAL INDEX

186.

191 193

INTRODUCTIONThe term "aUiteration "is

said to have been invented

by Pontanus

in the fifteenth century.elastic connotation,

Like

many other words

it

has a somewhatpeople does notit is

and on theit will

lips of different

always mean the same thing.a sort of rhyme, and

None, however, would deny thatif

conduce to clearness

we

deal

first

with

the more familiar word.

Varieties of rhyme.in

Khyme may be

defined as an agreement

sound between two or more

syllables (or groups of syllables).;

Such agreementsingle word.

may

extend either to one letter or more

and,

if

weit,

hold to our definition,

may

be looked for even within the limits of a

There

is

no reason why we should not hold to

and, though rhymes in close proximityto the ear,it is

may not

always be agreeable

instance in "

only just to admit them where they are found, as for Lama," " lowly," " added," " dodo," " sentimental,"Usually, however,

" iUimitability," " abracadabra."

rhyme

is

only;

recognised between word and word. and, to narrow thefield, let

Be

it

so for present purposes

the words be monosyllables, so thatarise in their

we

may the better see how rhymesIt will1.

most elementary forms.

appear that there are at least fourfirst

varieties.

The

form

of rhyrae is

where the vowels rhyme, and theThis

vowels alone,is

e.g.

" Hke " and " sight " or " sin " and " miU."

what is known as Assonance, and in old French poetry, e.g. the " Chanson de Roland," was the only requisite for a terminal rhyme.

In Spain and Portugal such rhymes are current to this day, as they are in English when the vowel ends the word, e.g. " me " and" see."

In modern languages, however

(if

Welsh be excepted),

they are only found in accented syllables, which in the case of

xiv

ALLITTERATIO LATINAmean sonoroussyllables,

such as the2.

French, where accent is disclaimed, will first of " dire.''

The second form

of

rhyme

is

where the post- vocalic consonantse.g.

(or coalescing consonants)

rhyme, and these alone,This rhymeis

" like " and

" joke," or " send " and " found."

not

uncommon

even in English verse, particularly in hymns, where the vowelsoften

rhyme onlyIcelandic

to the eye

;

^

and

is

one of the regular featuresside with our third variety.is

of Gothic poetry,

where

it

exists side

by

As the

name

for it (skothending)

untranslatable,

it

may3.

here be distinguished as Post-sonance.

In this case also the

syllable

must be an accented syllable. The third form of rhyme is where the

pre- vocalic consonantse.g.

(or coalescing consonants)

rhyme, and these alone,

" like "

andit

" lost," or " prince " and " pray."

To balance

Post-sonance,of

might be known as Pre-sonance.(monosyllable or other)it is

At the beginning

any word

and

is

the only form

of

a strong rhyme, regardless of accent, " alliteration " recognised by some

writers.4.

The fourth form

of

rhyme, that between pre- vocalic and post-

vocalic consonants, such as " like "

and "

roll,"

does not appear to be

recognised

by any Englishit

authority, the explanation apparently

being thatof verse.

does not enter into any of the listed systems as a ruleis

Yet such a rhyme

not without

effect.

Tennyson's

line, for instance,

12The murmurwould be shorn of halfhalf.its

12

3

3

of innumerable bees,

beauty without the echoes in the latterLetit

In Latin

it

would appear to be well established.In "life

be

known

as Trans-sonance.

" and "

feel

" there are two

such rhymes.

These elementary rhymes, or half-rhymes as they are sometimescalled,

may obviously be combinedCowper's stanza:

in a variety of ways, e.g. " like,"

"look," "elk," "clay";*

"like,"

"life";

"like,"

"chyle";

Of,

O

A A

walk with God, calm and heavenly frame, light to shine upon the roadfor a closer

That leads me to the Lamb

INTRODUCTION" like," " strike."Tlie composite

xvof course richer

rhymes are

than

the others, though not rich enough to satisfy the English ear asterminal rhymes between line and line, with the single exception of the last (" like," " strike "), which has not unnaturally appropriated toitself

the generic appellation Consonance.

But

all

the

rhymes, both elementary and composite, are occasionally found inthe body of theline,

where they are undoubtedly pleasing, when

skilfully introduced, e.g.

And ice, mast-high, came floating by (Coleridge).Her look was hke the morningLightly and brightly breaksstar^

(Burns).

Sloping slowly to the west (Tennyson).

away

(Byron).

And feelsLong

its life in

every limb (Wordsworth).wistfully gazed

at the

window he stood and

on the

landscape (Longfellow).Internal rhymes of this description are sometimes called line-rhymes.

Alliteration : its

meaning, raison d'etre,and place in English. As intimatedwhich

Having thus cleared the ground, we may now return to "alliteration,"which was provisionally described as a sort of rhyme.supra, thearises

word is often used exclusivelyinitial

of the particular sort

between

and

initial,

such as we have in Tennyson's

Prudes for proctors, dowagers for deans,or in Coleridge's

The

fair

breeze blew, the white foam flew,free;

The furrow followedandit is

so understood

by the Encyclopaedia

Britannica, which

describes Churchill's attempt to illustrate the sense

by the sound

in

Apt

alliteration's artful aid

as a failure, remarking that " alliteration is never effective unless it runs upon consonants " and obviously ignoring the ^'s and the

r,

r's

and the

Vs.

This view, however, seems too restricted

;

andresist

it is

noticeable that even those2

who adopt

it

cannot always

the

Bums no

doubt rolled the

which would here be rated by many phoneticians

as a vowel.

xvilogic of facts.

ALLITTERATIO LATINAMarsh, for instance, in commenting on tlie alliterationlicence

of Piers

Plowman^ admits that by what he terms a poeticinitial syllables^

accented syllables in the middle of a word are sometimes allowed to

rank asall

;

and Sidgwick, who commonly neglectsx. 100,

but

initial

rhymes, says on Aen.

" Observe the alliterationLarousse's Dictionnaire

pat pot pot

[in jKiter, omnipotens,

potestasy

Universelle recognises

no

limitations, defining the

word simply

as a

" repetition of the same letters, the same syllables," and quoting " qui terre a, guerre a " as an example. And in the Grande EncyclopSdie thearbitrary:

narrower interpretation

is

condemned

as

altogether

" C'est tout a fait arbitrairement que les grammairiensle

ont restreintat our back,

sens de ce mot, alliteration."

With

these authorities

we need make no apologyword

for adopting the larger view,

and we

shall accordingly use the

of

any rhyme other than the

special English varietyillustration of

known as terminal.is

An excellent line for the

our meaning

provided by Tennyson'sall

Universal Ocean softly washing

her warless

isles.

Tennyson himself disclaimed aalliteration,

liking for what he understood by remarking that he had " sometimes no end of troubleit

to get rid of

";

*

and, as he pronounced the verse just quoted to

it is obvious that for him the word did not cover the congruences in " ocean " and " washing," " all " and " isles," etc. For us, however, who are influenced not by spelling,

be one of his best,^

nor, in

any undue measure, by the positioneffects,

of the

rhyming

letters,it

but only by aural

the line

is

as full of alliteration asartificiality.is

could

well be without provoking

an appearance of

That aural repetitions have a natural charm

proved abundantly("

by

the frequency of their occurrence in"),

hackneyed phrases

by

hook or by crookpolitical cries ("

proverbs ("money makes the mare to go"),"),

peace and plenty

advertisements (" pink

pills

for

pale people "),

and nursery literature (" Jack the Giant,";

"

Baby Bunting,"

etc.)

and

it is

this fact

which explains their

presence in verse.

For though the poet'sp.

distinctive faculty

may be

' StvderWs English Language (Murray, 1872), going paragraphs owe a good deal.*

390

a work to which the forei.

Memoir by

his son, vol.

ii.

p. 15.

'

lb. vol.

p.

401 n.

INTRODUCTIONvision,

xvii

and

his real

power reside

in appeals to the imagination

and theaids.

heart, he

can seldom afford to dispense with adventitious

Even,.

in that sublime Psalm,

The Lord is my Shepherd He maketh me to lie down

:

I shall not want.

in green pastures

He

leadeth

me

beside the

still

waters,

there are artificial elements

more or

less

patent to the Hebraist'smeretriciousinter alia

eye

;

and when the essenceis

of poetry is wanting,

ornament

the only resource.^Its effect

Hence the recourse to

alliteration.

having been noted in ordinary speech,

its

magic was enlisted in the service of song.There can belittle

doubt that

alliteration

was one

of the earliestit

embellishments of verse,the place of what

and that among primitive peoples

took

we now understand by rhyme using this word in the ordinary and narrower sense. To use it in the wider sense, we may say that while, generally speaking, the modern line rhymes externally and only at the end, the ancient line rhymed internally and more or less all along. The progress from the one mode to theother can often be traced, and particularly in the case of English,

where the materials available for study are very considerable.Neither in England nor elsewhere did the change come about in a

daywas

natura nihil facit per saltumfirst

used,

it

was by way

and when the terminal rhyme not of substitution, butof addition.lose

Only whentheir hold.

this

had been firmly established did the others

The

following extract from Piers

Plowman

(fourteenth century)

will illustrate

one of the old English styles which had no terminal

rhyme

':

In a Somer Season,I shope

[

when hot was the Sunne,|

me

into Shroubs,|

as I a Shepe were

;

In Habite as an Harmet

unholy of Werkes,|

Went wyde

in thys

World

Wonders tooff

heare.

" The invention of a barbarous age to set

wretched matter and lame metre,"

is

Milton's description of one of these ornaments, terminal

rhyme

(Pref. to Par.

Lost).'

only an ornament. The correspondence between " were " and " heare "

Of course metre

itself is

is

accidental.

xviii

ALLITTERATIO LATINAis

The law which

said to obtain here

is

that in every line at least

three accented syllables (usually initial syllables)

must begin with

the same letter, and that at least one of the rhyming syllables must

appear in each section of theAlliteration in

line.

Welsh.In

Welsh, alliteration

is

on

its

throne,

even to-day, though there are signs that the influences which haveprevailed in England are also operating here, for writers of

hymnsre-

and other

varieties of song are

no longer bound by the

strict rules.

As the

principles underlying internalit

rhymes have a generalthis

semblance,

may

be worth while to examine some of the forms inancient tongue,

which they have expressed themselves inwhich, asis

now

well understood,

is

derived from the same stock as

Latin and Greek, andIts love of

may be assumed to have shared their heritage.

rhyming groups

which often

extend to half a dozen

consonants or more

is

very remarkable.

Premising that a Welsh verse does not consist of a stated numberof feet, such as dactyles or spondees, but only of a statedsyllables,

number of of them observing one rule of verse, others another, and yet others two or more combined either in the same line or the same stanza weand that the metres are very numerous, some

address ourselves at once to the rules themselves.

In dealing withis

them, we shall have to neglect the bardic terminology, whichuntranslatable except in cumbrous paraphrase; and

we have noOur objectForfuller

space for the regulations relating to pause, caesura, accent, and thelike,is

though these things are of cardinal importance.

merely to bring to the notice of the reader the salient features

which have or

may have

a practical bearing on Latin.

information he wiU go elsewhere.

The four heads under which the rhymes

are usually treated

may

be conveniently thrown into the following forms.^^"

The examplesii.

Cf. Percy's Reliques of Ancient English Poetry (Wheatley's ed., 1876), vol.

pp. 377-9.

In Welsh;

may

1911, etc.)

in English, in the

bo recommended Yr Ysgol Farddol (Evans & Son, Carmarthen, Rector Edwards's Prize Essay on " The Characteristics of

Welsh Poetry "our Preface.

Carnarvon Eisteddfod Transactions

of 1886

;

and (by waya note to

of supplement) the valuable Zeitschrift article (English) referred to in

1" The Welsh examples are from Yr Ysgol Farddol ; the English from the Carnarvon Essay, the author of whioh is careful to explain that English, with its

INTRODUCTIONunderis

xix

(6)

are lines with a trochaic

rhythm at points where such(trochaic) ending

optional.I.

Between the pausal

syllable

and the

:

Cyng-

hanedd Lusg.(a)

MinauIn fact

afI

|

dros yr afon:

:

af af

he was actinglet

act act.:

Do(6)

not

go

|

the poet

o

o.

Minau groesaf|

yr afon|

:

af af:

SomeII.

discord

was afforded

ord ord.

'

Between the beginning(a)

of the line

and the ending:

:

C, Draws,

Mam 3^1 ceryddu eiI felt that he

merch

m m.:

was using

force:

f f

(b)

Cafwyd elfenau cyfoeth

cf cf.:

AIII.

lover

was then leaving

Iv Iv.

Between the pausalC. Sain.

syllables in the first

and second sections

of the line,

and (alternating therewith) between the second section:

and the third(a)

Dyddan

|

yw

can

|

yn mhob

cell.

XX

ALLITTERATIO LATINA:

IV. Between the two halves of the line(a)i

C. Groes.

mi wr

lledf

|

o

He(6)

brought a cart|

mor Hon mrll mrll. by a right course:|

:

brghtc brghtc.

Ygwirfdrddbewar'|

oGaerfyrddin:

:

grfrdd grf rdd.

how you borrow

br br.:

The following show composite rhymesar ol

Hywel

|

i'r

helyg.

INTRODUCTION8.

xxi

The

repetition of a consonant without

an intervening vowelindispensable, subject

does not vitiate the symmetry of a group.9.

Otherwise,

symmetry

in the grouping

is

to the provision that a consonant

may do

double duty, as in " Enter

a painter pointing " (supra).Miscellaneous.10.

h

may

be neglected, except init is

ch,

ph,

ih.

11. n,

wheng

the

first

consonant in a half -verse,

may

be

neglected.12. h,d,

may in certain cases rhyme with p,(e.g. ot ot)

t,

c respectively.

13. In " consonance "

short vowels

may rhymeits

with

long.14.

An

inflected

word may be treatedin

as

if it

retained

radical

vowel,15.

e.g.

the vowel in " spoke "

may rhymesome

with that in " weak."

Rhyming elements mayAt a pause

cases be in arsi or thesirule.

indifferently.16.

Other cases are governed by

or ending the concluding consonants are in certain

cases not available for internal rhyme.17.

A

concluding syllable will sometimes rhyme with the

first

syllable in the following line instead of

rhyming

in its

own line

;

and

sometimes with the pausal syllable of the followingto

line in addition

rhyming

in its

ownline.

line.

18. Liaison, ligation ( 29),

and Sandhi

( 20.

i

n.)

operate even

between

line

and

To

lovers of the strict measures, alliteration

is

a fetish which

overrides everything in a way, so that in presence of an attractive

combination of consonants asense rather than the sound.

versifier will

sometimesit

sacrifice

the

A

century or two ago

was actuallyand evenslave to the

held that poetry could not exist without the

artificial jingle,

a writer of thesuperstition.

first

rank

like

Coronwy Owen was a:

Addressing a friend in 1753, he saysis

^^

" Paradise Lostfind

a book I read with pleasure.

.

.

.

You

will

me^^

ready to subscribe to anything that can be said init,

praise of

provided you do not

call it poetry.1876), vol.ii.

... As Enghshp. 53.

Jones' 8|(?orow2^

Owen (Longmans,

xxii

ALLITTERATIO LATINApoetryis

too loose, so ours'

is'

too

much

confined and limited,

not by theit

cynghaneddau

(alliterations)

for without themetc.

would not be poetry

but by the length,"of

The author's investigations incHne him to main point and for several centuries the poetsated by a similar feeling.Alliteration in Latin:

believe that

on the

Rome

were domin-

1.

Modem

estimates. The

alliterative

character of Latin poetry has not always been recognisedwriters.

by Englishall

Marsh, for instance, afl&rms ^^ that " alliteration was wholly

unsuited to the metrical system of the ancients, which rejectedechoings of sounds, andits

accidental occurrenceall

was regarded as

a rhetorical blemish."

Macleane, too,

but ignores the subject:

in his Horace, the following being perhaps the only reference

" Dillenbr. in his Quaestiones Horatianae has drawn particularattention to the alternate arrangement of the epithets in this

passage [C. III.are'

i.

21],

.

.

.

He

gives several instances,

and they

numerousIt

enough to constitute a feature in Horace's style.'

Spiritum Graiae tenuem Camoenaeis

is

one instance out of

many.

said to arise out of the liking the Latin poets

had

for homoeoteleuton.^'Ellis, in his

elaborate:

commentary on

Catullus, confines himself

to a few brief noticesxlv.1.

Septimiossuos.

is

perhaps preferred as an assonance to

12.

The

repetition of the full vowel o in ilh purpureais

ore

no doubt

intentional.

...

Its effect is

heightened by the triple a of saviata.15.

muUo mihi. .

major, triple alliteration in answer to

Septimius's pote plurimum perire.Ixiv. 150. eripuiiv..

crevi,

assonance like [Aen.

ii.

134, 96,

374, Enn. Ann. 51].et

xcvii.

4.

mundior

melior, double alliteration as in

.

.

.

leniter et leviler, Ixxxiv.

8 (where

it is

remarked

that Apuleius*

is full

of such assonances).

Student's English Language^ p. 393.

INTRODUCTIONConington seems to regard the recurrence of anhe interprets the meaning of our wordintroduced for specialinitial

xxiii

for so as only an occasional devicethe feature in Virgil

effect, his notices of

they are not

many389.

being generally accompanied by an attempt to

explain such effect

G.

i.

The

alliteration, as in

the previous verse, gives the

effect of

monotony.

A.

iv.

460. Is doubtless intended to produce the effect of

solemnity.ix.

30. Gives the effect of slowness

and

quiet.

89. Is

intentional,

expressing

rhetorically

the

in-

tensity of the anxiety.340.

The alliteration mandit'

'

mollemutummetus

'

is

expressive,of r

xi.

627.

The recurrencetentional.

and

here

is

doubtless in-

Mr Arthur

Sidgwick, in his admirable notes on Virgil, reveals a

truer appreciation of the facts in abstaining from interpretations

which cannot be consistently applied, andwith his author's fondness for the letter vxii.

is

particularly impressed

825. Notice the alliteration of v's, the

commonest in Virgil.

Lindsay, in discussing the orthography of Martial's Epigrams{Jwirn. of Phil., 1903), remarks on the care with which the author

must have weighed{Bihl. Oxon.)

his every word,^^ and in his preface to Plautus emphasises the " Celtic assonance " which he finds:

in the poet's treatment of vowels

Curandumenunties,

est litteras,

praesertim vocales, vere et Latinefere Celtica

ut in

cum Plautus non raro assonantia Amph. 1042 (troch. septenar.)rectais

gaudet

jam ad regemPerhaps nowhere

me ducem resque ut facta

est eloquar.

there a finer tribute to the music and par-

ticularly to the sonorousness of Latin poetry

than in Verrall's chapter

on Literature in the Companion^^

to

Latin Studies,^* though the

"

Cf. Ov. Ex P. I. V. 19-20 for an indication of the same fastidious attitude. Edited by Dr Sandys (Cambridge, 1910).

xxivsensitive ear ofticular effects.

ALLITTERATIO LATINAtliis

capable writer did not always approve of par-

It

is,

however, in Munro's Lucretius that the purely

alliterative aspect is

most strongly

insisted on,

and the language

there employed

is

so

much:

to the point that

an omission to quote

would be inexcusable"

One

of theis

most marked

peculiarities of the old Latin

writers

their extreme fondness for alliteration, assonance,

repetition of the

same or

similar words, syllables,in the

and sounds,

often brought togetherfashion.

and combined

most complexnot the

In Latin, as in some other languages, this usage was

clearly transmitted

from most ancient times, and

is

invention of any one writer.it

Ennius and the serious poets use:

to produce a poetical effectit

Plautus and the comic poets..

employartifices

for

comic purposes.:

.

Cicero does not despise such

even in prose

but none scatters them about more

prodigally than Lucretius, both singly

and

in manifold

com-

bination

:

they are to be counted in his poem by hundreds, nay. . .

thousands.

His alliterations comprise almost every:

letter

of the alphabet

the more effective letters such

s.s

mpvThe.

(prolast.

nounced w) are often used with strikingsometimes expressespity, as its

effect.

sound well

fits it

to do

.

or

force or violence, because the

words indicating such.. .

effects

begin

many:

of

them with the:

letter.

Often various lettersis

are used in combination

the following

a good instance of

m p and vnon

pararepotuit, pedibus qui

pontum per vada possentdivellere mentis

transire et

magnos manibus

multaque vivendo

vitalia vincere saecla.

Compare Ennius'sMarsa manus, Peligna cohors, Vestina virumvis.

Such combinations are commondreds in Lucretius.less..

in Virgil

;

but occur by hun-

.

After Virgil's time they appear to be:

frequent in Latin literature

people probably got tired

of them, as has

happened

in other literatures." ^^

"

Vol.

ii.

pp. 15-16.

Cruttwell'8 Hist, of

Rom.

Lit. (6th ed.,

1898) has a lengthy

note on the general subject, pp. 238-0.

INTRODUCTIONon the Continent than in England.

xxv

On the whole the subject seems to have attracted greater attention Prof. KviSala of Prague, whoa weighty though not an indispensable element in Latin poetry,

confesses to a rooted conviction (feste Ueberzeugung) that alliterationis

has analysed the language of the Aeneid with meticulous care, andin his

Neue

Beitrdge zur

Erkldnmgall

der Aeneis^^ (p. 294)in the

namesfield.^'is

several investigators

who had laboured

same general

His

own

book, which takes

the letters at their face value,

mainly concerned with the registration and

classification of the

rhymes which appealed toordered system.It

his ear,

and makes no pretensions to an

would appear, nevertheless, to be a marked

advance on anything that had gone before.

Such then

is

the attitude of the moderns,

translated internal jingle into a law of verse.

who have in no case The common view ison Vowelin Latin

no doubt that expressed by Mr ClassenAlliteration in the

in his treatise

Old Germanic Languages (1913)is

that

poetry alliteration

not an essential part of the technique, but onlyalliterationis

an ornamentrules (p. 45).

;

and that the

not determined by any

Alliteration in Latin:ancients say?

It

2. Ancient allusions.What do the must be admitted that no passage can be pro;

duced which definitely points to recognition of a systemindeed alliterationis

andas a

seldom referred

to.

But there

is

evidenceit

enough to show that the grammarians were well aware ofrhetorical device.in its variousccmfricatio, etc.

They have,irapofjiOLov,

in fact,

many names

for the figurecollisio,

forms

ofxoLOTeXevrov,

adnominatio,

and they provide^^

inter alia the following illustra-

tions

:

solus Sannio servat

(Donatus).^^

quidquam quisquam cuiquam quod conveniat negetad Herenn. IV.xii.^^

(Rhet.

(18) in

Nobbe'siii.

ed. of Cicero),

non verbis sed armis16 1'

(Quint. IX.

75).

Prag, 1881.

Specially attractive among the works referred to seem Schliiter's Vet&rum Latinorum alliteratio cum nostratium alliteratione comparata (Monast., 1820) and Cadenbach's De alliterationis apud Horatium usu (Essen, 1837), both of which have

eluded the author's search.18

Ter. Eun. 780.

i

Ennius.

20 p^^til.

Lup.

2. 12.

xxvi

ALLITTERATIO LATINA^i

puppesque tuae pubesque tuorum

(ib.). ^^

machina multaK.I. 447.4).

minax minitatur maxima muris^^

(Diom.

sola mihi tales casus Cassandra canebat

(Servius).

Anchisen agnoscit amicum

^* (ib.).

In

its

most aggressive form the feature was regarded by somewith disfavour.(loc. cit.)

writers

Thus the authorit

of

the Rhetorica ad

Herennium

defines

as ejusdem Utterae nimia assiduitas.

Martianus Capella remarks, ^^ Compositionis vitiumvitare aijuslibet Utterae assiduitatem in

maximum;

est

nonis

odium repetitam

and he

supported by

Servius, who, in connection with the

two verses

last

quoted, says, Haec compositio

jam

vitiosa est quae m^joribus placuit.

Ennius's lineTite tute Tati tibi tanta turanne tulistiis

condemned byet

several of the grammarians, as for instance

by

Plotius Sacerdos under the figure Aprepia, which he defines as

absurda

indecens verborum structura.^^

Servius's

concluding words quae mujoribus placuit show that

literary taste

had undergone a change during the four centuries

which separated him from Virgilrecurrentall

at

least

on the question

of

initials,

even as the taste has changed in England.

Indeed,

the passages suggest that the writers

alliterative

law

^^

which,

however,

against the existence of such a law.Piers Ploivman

had been forgotten;

knew nothing of any is by no means conclusive The rules which obtain in among ourselves until theyandin Latin the process of

were discovered about a century agoforgetting

was helped by the changes

in pronunciation, which of

course went far to obscure the uniformities observedof earlier times.

by the poetsnot altogether

Assuming that there were some governing

rules, it is

surprising that they should have been left imnoticed

by the con-

" a *'

Virg. A. Virg. A.

i.

399.183.

^'

Anon.lb.iii.

(Keil's

minatur

is

obviously a misprint).Art. Rhetor., 33.

iii.

"

82.

" Derules,

K. VI. 454. 30. If Aulus Gellius had understood the

refer to

them

in connection with

parses a dozen lines of Virgil

he could hardly have failed to remarks on euphony. Priscian, who at great length, and even scans them, has not a word

some

of his

to say about the alliterative features.

INTRODUCTIONtemporary writers whose works have come down toearly grammarians, evenif

xxviius.

For the

they were acquainted with them, would

probably have regarded them as outside their province, or even

beneath their notice.

The chances are that they were not acquaintedIn Wales, where

with them, or at least not well acquainted.alliterative verse

has been the vogue for centuries, the rules are onlyare themselves writers.

known

to those

who

The public are indeedeffects, but, as

aware of their existence and appreciate thetiUan says,^are producedsubject,it is;

Quineffects

only the expertsin a

who understand how theis

and except

work dealing expressly with theseldommentioned.Itis

cynghanedd

(alliteration)

stated in the Zeitschrift article, referred to in a note to our Preface,

that

when Dr

Griffith Roberts,

who wrote on Welsh poetryhim the

in the

sixteenth century, asked a bard to explain to

rules of verse,

the latter refused on the ground that he was " sworn to teach no onethese secrets."

And even

to-day"

many

a Welsh poet

is

imperfectly

article,

acquainted with the canons. He writes," says the author of the " by ear rather than by rule ; he has read thousands of Huesin the bardic metres,

and

his ear has impressed their

form on

his

mind.of their

The

lines

themselves

may He

be forgotten, but the impressions

form remain, and become the moulds into which the bardtakes;

pours his new molten metal.as they

first

one and then another

happen to

suit his

purpose

but, though he uses

them

all,

he

may

not have classified them, or even counted them.

In time

of course he learns the rules,

which he easily understands, as they;

only enunciate more definitely what he already knows

but he no

more begins by studying

rules

than he begins to speak by studying

grammarancients.

"

(p.

141).

AndA-e^ts

the same was probably true of the?

What was

{JieXonoua

When

Aristotle

^^

speaks of

it

as distinguished fromit

or metrical composition, he dismissesIt

with the remark that

its

meaning was obvious to everybody.

seems not unlikely that what he had in his mind was the music ofalliteration in the sense

we have adopted.^^

And

so Horace,

when he

contrasts niodi with tempora

or numeri,^^ or speaks of distinguishingfingers,^^

a rightly constructed verse by the ear as well as by thewell have28

may274.

meant the same29

thing.

Ovid too seems to hint at more3i

IX.

iv. 116.

Poet.

c. 6.

^o I. Sat. iv. 58.

A.P. 211.

32 /^^

xxviii

ALLITTERATIO LATINAhe characterises thecolor

than mere language and metre when (without claiming much creditfor the innovation)

and

structura of hisif

verse as something distinctive

and uncommon.^^

It looks as

he

were referring

to some self-imposed

restriction in the ordering of

his rhymes, the nature oflines.^*

which was discernible even in his opening

Finally, Martial has

an epigram

^^

containing the coupletlibellos,

Lector et auditor nostros probat, Aule,sed quidam exactos esse poeta negat.If

the reference

is

not to the poet's Latinity or metrification

neither of which, so far as the author knows, has been seriouslyassailed

it

must be to some law

of alliteration

which Martial

(perhaps only occasionally) failed to observe.Alliteration in Latin : 3.

Treatment in translation.Whateverall

may be thought of the author's attempt to systematise the alliterativefeatures referred to, there can be no doubt at

that the features

themselves are

real.

In such

lines, for instance, as^^'^ ^^

verpus praeposuit Priapuscrtida Virgine

ille

Marciave mergi

chartae Thebaicaeve Caricaeve

carmina caeruleos composuisse deos

nam

didici Getice

Sarmaticeque loqui

*^**

vel anseris medtillula vel imula oricilla

they leap to the eyewillif

;

and an attentive studyaffects

of the following pages

show that they are nowhere

entirely absent.

Now, obviously,

justice is to

be done to a poet who

them, they ought to bepresented in a modernto us can have

reproduced in some form or other when hedress, particularly aslittle

is

much

that has

come downall,

had

to attract beyond the

jingle of the verse.

Translators, in so

far as they

have regarded the matter at

have usually taken the

view that the demand was adequately met by a terminal rhyme

appended to a measure which had noin so behaving they

relation to the original.;

Butit is

do the ancients a double wrongCf. 166 (infra).'

and

ExOv.

P. rV.

xiii.

M

"

jx. ixxxi.

Cf.

95

obs. (infra).

3 Cat. xlvii. 4. 3

Mart. VI.lb. III.ii.

xlii. 18.

"i

Stat. Silv. IV. 9. 26.

Ex P. IV.

xvi. 22.

"

40.

Cat. xxv. 2.

INTRODUCTIONI perhaps one of the reasons why their workis

xxixso seldom valued either

by

scholars or others. ^^justification for

Theear,

abandoning the ancient metres

is

no doubt

the fact that with few exceptions they do not appeal to the modern

when taken as they stand.

Amongis

the exceptions are the metre

of the Pervigilium Veneris,

which

that of Tennyson's Locksley

Hall:Cras amet, qui niinquam amavit:

quique amavit, eras amet'tis

Comrades, leave

me

here a

little,is

while as yet

early

morn:

;

and the metreFor

of Cat. xxv.,

which

that of Tennyson's Brook

Remitte pallium mihi

meum quod involasti men may come and men may go, but I:

go on for ever

;

and the dominant metrein

of Catullus's E'piihalamium,

which

is

found

Darby and Joan

ColHs o Heliconieicultor,

Darby

dear,

you are old and gray

:

Uraniae genus

Fifty years since our wedding day

!

But there are others which appear to resist ordinary treatment. The remedy is to break up the lines, even (if need be) to the divisionof a foot,

and to introduce;

alliterative or terminal

rhymes at the

important ictuses

whenday.

it will

be found that in at least most cases

the whole will work out in a form hardly distinguishable from the

metres of our

own

So at

least it

seems to the author, who, in

evidence of his good faith, submits a dozen illustrations to the

judgment

of his readers, claiming for

them no more than:

will

be

willingly conceded to1.

an amateur

in verse

Cat. xvii. 13-14 (Priapean),

Insulsissimus est homo, nee sapit pueri instar

bimuli tremula patris dormientis in ulna.JFoolis

he

not the wide world through:

Found would be such another Has not sense of a child of two Drowsed in arms of its mother.*2 Tennyson, who once remarked that " the benefit of translation rested with the translator," compared the Sapphic stanza to " a pig with its tail tightly curled," and parodied the pentameter with " All men alike hate slops, particularly gruel

(Memoir by

his son).

XXX2.

ALLITTERATIO LATINACat. XXX. 3-4 {Second Asclepiad),

Jam me prodere, jam non dubitas fallere, perfide ? num facta impia fallacum hominum coelicolis placentCare you,Falsest of men,

?

Dare you

Dream?

that a wrong

Thus to repay

?

Flouting the skies

Thus to betray

Vengeance

defies

And then3.

For longI. iv.

?

Hor. C.

1-2 (Fourth Archilochian),

Solvitur acris hiemps grata vice veris et Favoni,

trahuntque siccas machinae carinas.Winter's a-wing

And a breath of spring From the welcome WestAnd, winches manned,

is

coming

;

Each busy strandIs

humming.I. viii.

4.

Hor. C.

1-2 (Second Sapphic).

Lydia die per omneste deos oro, Sybarin cur properes

amando

perdere.

Hettie,Tell

by heav'n above you.you're ruining Guy,?

me why

Luring the lad to love you5.

Hor. C.

I. xxiii.

1-4 (Fifth Asclepiad).

Vitas hinnuleo

me

similis, Chloe,

quaerenti pavidam montibus aviis

matrem nonaurarumFlo,

sine

vano

et siluae metu.

you

fly

me

in fear, like a

young deer that

hies

Headlong over the hoe, seeking the doe, her eyes

Wild with

idle alarm, in

bushes visioning harm

And

skies.

INTRODUCTION6.

xxxi

Hor. C.

II. X.

1-4 {First Sapphic).

Rectius vives, Licini, neqiie altum

semper urgendo neque,iniquum.

dum

procellas

cautus horrescis, nimium premendolitus

Best

is

not,

good

friend, to

be got by heading

Out to seaPerils

perpetually, or threadingshore,

more near rock-ridden

Mid-ocean hurricane dreading.

7.

Hor. C.

III.

i.

1-4 {Akaic).et arceo;

Odi profanum vulgusfavete linguis:

carmina non priussacerdos

audita

musarum

virginibus puerisque canto.

Keep To

off,

ye

commonboys and

raff

:

from theseattempted

cloisters fly.I

A hymn no human lipsartless

e'er

girls

am

bringing.

Priest of the nine

who

are served

by

singing.

8.

Hor. C.

III. xxiv.

31-2 (Third Asclepiad).

Virtutem incolumem odimus, sublatam ex oculis quaerimusinvidi.

Hating Worth, while

it

walks on earth.

Envy

looks for the hght, only

when

lost to sight.

9.

Hor. Epod.

v.

87-90 (Iambic).

Venena magnumconverterediris

fas nefasque

non valent

humanam vicem:

agam vos

dira detestatio

nulla expiatur victima.

xxxii

ALLITTERATIO LATINANo drugor devil's art so strong

Can laws invert of right and wrong To pleasure man.*^

On you my curses hot shall hail Nor ever victim's blood avail To lift the ban.

10. Virg.

Aen.

vi.

737-43 {Hexameter).

Penitusque necesse est

multa diu concreta modis inolesceresupplicia expendunt.

miris.

Ergo exercentur poenis, veterumque malorumAliae panduntur inanes;

suspensae ad ventos

aliis

sub gurgite vasto

infectum eluitur scelus, aut exuritur igniquisque suos patimur manes.**

Much

of the mischief

they wroughtsoul have extended,

Must needs to the

Foulness in which must be fought

BySome

penance for justice ofiended.of us,

hung to the

blast,;

Are winnowed by dolorous breezes

Some

are 'neath cataracts castof diseases;

For the scouring awayOthers in furnaces burn

Out the

traces of contamination

All of us victims in turn

Of the vice we have brought for purgation.*^

Humanam

vicem, a great trouble to the commentators,

is

hominum(Epod.**

vicem, the

meaningi.

of

which

is sufficiently

illustrated

only a variation of by Helenae vicem2, I. ix. 1),

xvii. 42), illius vicem,

nostram vicem

(Cic.

Ad Fam.

IV. v.

and

publica vice (Quint. XI.

42).

Here too the commentators are perplexed, but quite

needlessly.

The divine

essence has been contaminated by the body, and carries with it into the other world the noxious elements that have to be painfully purged away ; so that the

condition of the individual manes (or soul)

is the measure of its punishment. Patimur munea is merely a contracted expression for patimur supplicia quae neceaaaria aini ad manea noatroa purgandoa.

INTRODUCTION11. Ovid,

xxxiii

Amor.

II. xix.

19-20 (Elegiac),

Tu quoque, quae

nostros rapuisti nuper ocellos,

saepe time insidias, saepe rogata nega.*^

Thou, too, whose rapturous charms

But newlyFeign often

my

senses bemuse,

fictitious alarms,.

Often thy favours refuse.12. Ter.

Maur. 142-5 (Sotadean).

Hanc

edere

vocem quotiens paramusdicere, sic citetur ortus

ore:

nitamur ut u

productius autera coeuntibus labellis

natura soni pressior altius meabit.If to utter

you seek

But shoot out theBringing the tipsSteadily nigher:

lips,

V in the

Greek

Fairly and fully.

Your mouth you must round, A note you will strike As though to sound Not much unlike, u in (say) Tulli. But thinner and higher.*^*^ " Not Ovidian," says Palmer {Heroides, Pref., p. xii), who, in agreement with Lachmann, remarks that time insidias is nonsense, and accordingly conjectures " saepe tamen sedeas.^^ Plainly the meaning is, " often pretend that your husband is on the watch, and that you are afraid of being caught." In blank verse no one has handled elegiacs more skilfully than Watson, e.g.:

Man and his Man and his**

littleness perish, erased like

greatness survive, lost

an error and cancelled. in the greatness of God.lines. So in remarks (1454, 2054) and

Syllables have been resolved in this{e.g.

and the corresponding

Ter. Maur. occasionally

1560).

Cf. his general

54

(infra).

xxxiv

ALLITTERATIO LATINAABBREVIATIONSAncient Literature.

Cat.Gell.

Catullus

:

Postgate, 1889.:

Hor. Juv.

K.Lucan.Lucr.

Aulus Gellius Teubner, 1853. Horace Oxf. Pocket Text. Juvenal: Lewis, 1882. Keil's OrammcUici Latini.:

Lucan

:

Haskins, 1887.:

LucretiusMartial:

Mart.

Bibl.

Munro, 1893. Oxon.

Ov.Pers.

Ovid

:

in Heroides, i.-xiv.,:

Palmer

;

elsewhere Teubner.ed.

Persius

Conington, 1874.:

Phaedr.Priap.Prise.

PhaedrusPriapeiaPriscian:

J.

M.

B., 1847,

supplemented by Valpy's Delphin

Bibl. Lat., Athens, 1888.

:

Krehl, 1819.:

Prop.Quint.Stat.

PropertiusStatinsVirgil

Bibl.

Oxon.:

Quintilian's Institutes:

Meister, 1886.

Valpy's Delphin ed.:

Tib.Virg.

TibuUus:

Bibl.

Bibl.

Oxon. Oxon.

ModernLindsay, L.L.:

Literature.

Latin Language, 1894.Historical Latin

L.O.:

:

Marx

Hiilfsbiichlein

Orammar (2nd ed., 1915). fur die Av^sprache der lateinischen Vokale (3rd

ed.,

1901).

Ramsay

:

Manual:

of Latin Prosody (2nd ed., 1859).lateinische Sprache, 1819.

Schneider

Grammatik der

SeelmannStolz:

:

Aussprache des Latein, 1885.

Historische:

Grammatik der

lateinische Sprache, 1894.

Walde

Lateinisches etymologisches Worterhuoht 1910.

SYMBOLSA^ A^ A^ indicate initial, medial,

and terminal

liaison or ligation respectively.

In a Line ofItalics indicate that

Verse.

a syllable (allowing for sandhi) does not rhyme independentlyline.

of

a neighbouring

InItalics indicate

the Analyses.;

annexed by liaison or ligation or an interlineal rhyme (which may also be an internal rhyme). B and the like an oblique rhyme. BB and the like (usually in brackets) an initial rhyme. 8 a broken uniped. s and the like an unbroken uniped. s, and the like ss, s 8, s . that the letters are in a position to pair. t, and the like 8'~'t, s t, s indicates a (legitimate) transposition in the elements of a group somewhere inaletter::

:

.

.

.

:

.

.

.

:

"I"

thea, e,

line.

, 0,

u

often represent a, oc,

e (y),

oe (oi),

and eu

respectively.

ALLITTERATIO LATINACHAPTERRHYME 1.(2) (6)

I

The elements(3)

of a verse in

any language are(4)

(1)

the thought,

the diction,

the syntax,

the metre,

(5)

the rhythm,

the consonantia litterarum,^ and usually (7) the rhyme.trivial or obscure,

The thought may be

the diction inelegant or

prosaic, the syntax involved or ungrammatical, the

metre irregular,

the rhythm jerky, the consonantiathe

litter

arum

frigid or harsh, ^

and

rhymeis

imperfect.

Few:

poets are concerned to observe the rules

of their art at all times

for,

with so

many

interests to attend to,

there

a constant temptation to sacrifice the less important to the;

more importantcerned,

and rhyme, with which alone

this

book

is

con-

has

sometimes been sacrificed altogether.

In Milton's

Paradise Lost, for instance, the ornament is missing. " rhyme," as commonly understood, 2. The wordterminal rhyme, such as thefirst

meanswhere:

we have

in the following quatrain,

rhyme

is

an assonance, and the second a consonanceis

All nature

but art unknown to thee.

All chance direction which thou canst not see,All discord

harmony not understood,

All partial evil universal good.^^

who(A.

The term consonantia litterarum is borrowed from Gell. VI. (VII. xx. (capitulum), also makes use of the expression consonantia vocum jproximarum (XIII. xxi.)

(xx.) 5).iii.

In speaking of

Virgil's preference for the termination

-e,

or

which are regarded (by a convention orvalue, such asjp

having the same

and

6, e.g.

praemia ponit (p p) posse negabat (p b).^ These refractory lines do not exceed 50 out of the 110,000 or so which have formed the author's field of inquiry. Only in about half of them are there no manageable variants. * Some writers use arsis and thesis in a converse sense. ' Ov. F. IV. 760. The thetic I has no significance.

4

ALLITTERATIO LATINAstrongest

Theictic

rhymes are undoubtedly those

in

which the rhyming

consonants are the same and either both precede or both follow the vowel. The English " bell," for instance, is a closer echo of" bat " thanit is

of " rub "it

;

and

so pcmit

rhymes moreif

effectively

with praemia thanthe 6 were p.

would with negabat, even

in the latter

word

Still

they were both good rhymes to the Latin ear,only noted for the sake of accuracy.

and the 8.

distinction

is

Rhymes may

subsist

between

like

vowels as well as betweenBut, thoughit

like

consonants

and on the same indulgent footing.might answer toe.

e (for instance)

ae, it is

hardly credible that

would

be admitted as a rhyme tosense of quantity than

For the Latins had a more delicateitis

weit

have, and

known

that a wrong

pronunciation on the stage would provoke an immediate uproar

on the benches.sufficient

Indeed,

would seem that short vowels had not

volume to furnish a rhyme under ordinary circumstances,accordingly be found that under our rules they are only

and

it will

recognised

when they

are lengthened

by the poet

(

60-62) or enter

into the relations described in 22

ara Dianae (a a)

laudesque manebunt

(e e).

There are

many lines whichsufficiently

as

will

be understood at a later stagee.g.

would rhyme

with their vowels alone,

tentator 6ri6n Dianae

non potui

fato nobiliore mori'violabis

quo non Romanes

vomere manes.^

9. A series of homogeneous rhymes, however short, termed a " sequence,^' which may be either simple {p

may.

be

.

.

p) or

compoundthesequence."

(pr

.

.

.

pr).

A

series of

minimum

requirements of a verse

sequences sufficient to satisfy may be termed a " line of

10. It will

be observed that in the hexameter endings quotedis

supra the rhyming consonant

not supported by a rhyming vowel,

nor the rhyming vowel by a rhyming consonant.

Such support

is

never necessary, and the rules that govern the terminal rhyme in

Hor. C. III.

iv. 71.

Mart. XI. Ixix. 12.

i

Luoan,

vii.

862.

RHYMEEnglisli

5

need not trouble

us.

Composite echoes, however, after the

EngUsh patternmeter endings(

are not exchided in Latin. 29^.)

the tendency is to avoid an icticis

But except in pentarhyme in which by the same vowel

the same consonant(6s OS, so so), the

followed or preceded

prevaihng rule being uniformity in the one and

variety in the other, or, f aiUng that, an inversion of the order of theletters. 11.

The hexameter endings which we have quoted were intended

to illustrate alliteration in its simplest form

thatis

of the single

rhyme.

The following

are examples of the double rhyme, in

some

of which, it will

be seen, the order of the recurrents:

(quite legiti-

mately) reversed

Mdrtis amore

:

mr mr:

ctira pectili

:

cti

cu

Pergama

Graiis

rg gr.

taha fatur

:

ta at.

By

combining two or more double rhymes, we get

triple,

quadruple,

quintuple, sextuple rhymes, and so on, as will be exemplified at alater stage.suffice:

For the present, instances of the

triple variety will

praemia primiexercita ciirsu 12.

:

prm prmsrc crs.

viscera

quisquam:

:

vsc vsq

:

ostroque decoriof the letters will

roq cor.

The phonetic values

be set forth in a

succeeding chapter.

Owing

to the fact that there are only sixteenitis

consonant or vowel sounds which cannot echo each other,diflScult to

construct a long line without a single rhyme.desirable,

An example,is

however,

is

and the followingearit

elegiac couplet

offered as

a curiosity.

To the Roman

would be nothing more than a

fragment of song in which every note was out of tuneAssiduene idee, pulcherrima Laodamia,aggrediebaris toUere vina

mea

?

Contrast with these the following unpretentious line

:

non meruio

lis"

tali forsitan 6sse loco ^^

so

Ov. Tr. V.

X. 50.

6

ALLITTERATIO LATINAby combeaten out hisconticuere

or this other, where the poet has replaced simple rhymes

pound, and

music with both his hands

omnes intentique ora tenebant.^^nenent v. .r

nt v.r

nene

fntcv.r.o

ntcv.Virg. A.ii.

.

ro

"

1.

CHAPTER

II

ORTHOGRAPHY 13.

One

of the difficulties connected with theitself is

problem to which

this

book addresses

the spelling, which often confronts us inof the

more than one form.

The explanationthat

phenomenon must betwo pronunciations

either (1) that the spelling

was optional while the pronunciationit

remained the same, orcorrupted by copyistsfashions of theirsuperiors.^

(2)

reflected

current at the same time, or (3) that the original texts have been

who accommodated the spelling to the own age or to the mistaken directions of theirprobability somethingis

In

all

due to eachit

of these

causes.

How muchamount

to one

and how much to another,

may

never

be possible to determine.certainof guidance helps,

But

at least on

some points we have a

from the ancient grammarians, and ourprimarily to the fewness of the lines

alliterative

scheme

though not to such an extent as might

have been expectedwhich arefree

owing

from superfluous ornament.

The

variations

may

be conveniently dealt with under three heads

:

14. (a)

Where

the variations did not affect the pronunciation to

the extent of interfering with thei.

rhymes;

(see

Chapter

III.).

-cumque, -cunqueetc.

;

umquam unquam;

quamquam quanquam,

tamtus tantussumsi sumpsi

;

;

damdus dandus eumdem eundem, etc. sumtum sumptum hiems hiemps Rodus; ;

Rhodus,

etc.

cum quum^

;

cui quoiare

;

scaena scena

;

circumeo circueo,

etc.

The

rules laid

down

still

greatly increased

by

modem

editors,

extant in some cases. The complications are who, for the sake of what they deem consist-

ency in the

spelling, are often unfaithful to the

MSS.

7

8ii.

ALLITTERATIO LATINAsulphur sulfur;

negligo neclego;

;

cycnus cygnus,;

etc.

s

maximus maxumus apud aput sed set;

lubet libet

portubus portibus,

etc.

;

baud haut,;

etc.

vulgo volgo

;

vertex vortex

faciendus faciundus, etc.

15. (6)

Where

the spelling

was optional and

reflected

two

pronunciations with different rhyming powers.i.

totiens toties

;

deciens decies, etc.;

ii.

forensia foresia

Megalensia -esiaetc.

;

intrinsecus -isecus, etc.

iii.

tonsus tosus tinguo tingoetc.

;

mensus mesus,;

iv.

unguo ungo

;

urgueo

urgeo

;

tempto tento,

V. querella querela

;

ligurrio ligurio, etc. (see Alphabetical List in

Appendix 16.

0).

Two

spellings

(and pronunciations) are also found in aof

large

number

of

words compounded with prepositions, in oneis

which the

final

consonant of the preposition

assimilated to theitself is

following consonant,

and

in the other not.

Assimilation

the

outcome

of a

tendency to avoid a combination of sounds whicheffort,

cannot be produced without a sense of

such as

is

involved in

passing rapidly from one organ of speech to another not convenientlyplaced.

and if we In other words, it is due to a sort of laziness " Harry " or " Bessie " instead of " Henry " or " Betsy," it is say;

because they come more easily to the tongue.

Now, when comformed,^that^

pounds

like ad-curro, dis-fero, in-ruo, etc.,

were

first

sense suggests

pace some weighty authorities

common

the words

were pronounced as here written.so pronouncedis

For how long a period they wereIn

another matter.

many

cases assimilationclassical

must

have become permanently established beforearose;

literature

in other cases;

two pronunciations may

well have existed side

by

side

and

in yet others, after perhaps centuries of the lazier

usage, the fashion would change,2

and there would be a harkingand

" It

is

quite a mistake to suppose the unassimilated forms to be the olderp. 313),is

the assimilated the more recent," says Lindsay {L.L.,in Prager Phil. Studien, 1887.

appealing to Dorsch

Perhaps the statement

not intended to apply to

the pre-literary period.

ORTHOGRAPHYback("

9Itis

re-composition ") to the original forms. ^

noticeable

that in

MSS. and

inscriptions the

same word

is;

sometimes spelt in

different

ways on the same page or monument

and*

in the case of

adcurroB.C.)

we have theit

distinct testimony of Lucilius

(second century

that

did not matter whether the d was assimilated or not adcurrere scribas

d-ne an

c,

nonif

est

quod quaeras eque

labores.

It certainly looks as

in at least

option.

If so,

and

in such cases,

many cases usage sanctioned an we may be sure that in elevatedis

discourse

and

official

documents the leaning would be to the moreIn poetry, where thereever a

dignified, etymological spelling.

tendency to avoid the commonplace and to introduce even outwornforms for the sake of their associations and the appeal they maketo the imagination, the preferencedirection, particularlyif

would often befell

in the

same

the less familiar spellingline.it

in best with

the alliterative requirements of the

Onis

the evidence of the

grammarians

contradictory thoughwhichresist it

often

assimilation would;

seem to have been the rule

in our period,^

and the author has not^

met manyother

lines

underlines

his

treatment

but on thefor the

hand there are many:

which plead strongly

" Applicat : secundum praesentem usum per ^ Cf. Servius's note on Aen. i. 616 d prima syllaba scribitur secundum antiquam orthographiam per p." By antiquam he no doubt means roughly (like many other grammarians) " during the: . . .

classical period."

from the"

That Lucilius did not approve of indiscriminate assimilation is evident which he sets to the assimilation of per, which, he says, could only unite with I. Cf. Vel. Long. K. vii. 65. 14, where the true reading must be*

ix. 25.

limit

Apud Lucilium

legit ur in

praepositionem

'

per

':

praepositum

nam

'per' 'liciendo' congeminatI.)

1."

{Per prefixed to Ucere doubles the^

thought that the etymological spelling was due to ignorance " Frequenter invenimus adfatur, adludo, adrideo, adnitor, adsumo. Errore tamen scriptorum hoc fieri puto quam ratione " (II. i. 7) and he cannot understand why the etymological spelling should be held more euphonious than the other. So " [Novissimam litteram praepositionum] quidam imperite semper Ter. Scaurus custodiunt, adripit et conripit et conludit " (K. vii. 25. 18). Cassiod., however, " Est ubi [d] sonet et ubi scribatur is in conflict with Priscian on one point ut adfluo, adfui, adfectus" (K. vii. 151. 16). It seems clear that in every case thePriscian:. . .

;

:

'

'

'

'

'

'

:

.

.

.

pronunciation followed the spelling.

Lucan, ix. 488, demands adligat and Stat. Silv. III. i. 73, inmaduit (unless * Libyam). With this cf. Cat. Ixi. 169 (173), which, without requiring, strongly suggests in-minentes. For adsiduus, cf. 182. 6o, 6i.^;

we read

10

ALLITTERATIO LATINA:

etymological spelling

and

his general conclusion is that, in the

absence of countervailing reasons, individual cases must be decided

by the ear. We know that Virgil's choice of turrim, was dictated solely by considerations of euphony.Obs.

urheis, trisy etc.,

The same general observations apply to other compounds suchsometimes

as idcirco {iccirco), quidquid {quicquid). 17.

Words compounded with con and in form a class apart. Beforeand sometimes(2),

liquids these prepositions behaved like other prefixes, being

assimilated

not,

and the same general considerations

apply.

Before sit

d (t), c{gk q),jy and (during our period) v the n was

retained, as

not always.tions,

sometimes was before the labials (b pf) also, though The evidence relating to these labials from inscrip-

MSS., and the grammariansit

is

confusing and conflicting, and

only by disregarding some of easyclari

can a practical rule be reached.

An

way out

of the difficulty

would be to admit the teaching of the

homines vouched for by Mar. Victorinus, that before a labial

the illustrations are Sambyx, Ampelo,intermediate betweeneither,

Lycambe

m hadignored

a sound

m

and

n, which,

without being identical with

partook of the nature of both, and was presumably capable ofeither.

rhyming with

But

as this teaching

is

by other

ancient grammarians and contestedit

by many modern

philologists,

seems safer to rely on Ter. Scaurus, who flourished near the close

of our period

and at

least recognises

(what other grammarians do

room for distinguishing, and that there were He says that cases where com and im could not be (properly) used. the n was preserved before these labials when they introduced anot) that there wassyllable containing a

vowel which was long by nature or positionit

leaving us to infer that in other cases

was a matter

of indifference

whether the n was changed intospellings

mby

or not (see 291-7).

His

may

therefore be typified

inpurus

conpelloinbellis

impia or inpia

conburoinfamis

imbuo

or inbuo

infirmus

imfimus or infimus.

They

satisfy the requirements of our alliterative scheme,

though

it

must be admitted that the

crucial lines are exceedingly few.of in

Obs. In words which are not compounds assumed to be m {umbrae semper etc.).^

and con the

spelling

is

ORTHOGRAPHY 18. (c)

11

Where

the spelling

and pronunciation may have changed

during the Golden and early Silver Ages.

During the period in which we are interested, many words under-

went a change and

of spelling side

by side with a change of pronunciation

;

this, it is clear,

should properly be taken into account in dealing

with the alliterative features of a particular author.afiected both vowels

The changesof spellings

and consonants.'

Out

of a

number

which Lindsay has

listed

as current in the time of Quintihan

(whomthe

he regards as the best model), the following are the most importantfor

present

purposes.

They do not

necessarily

represent

spellings current in earlier years.abicio,

adicio,

etc.,

better

convicium, not -tium.cotidie

oboedio, not -edio.better:

than abjicio, etc. Alexandrea, Dareus,

and

cott-, not quo-, obscenus,

than

etc.,;

dicio, better

thanf

ditio.

-scaenus

not -scoenus.

during the Republic-rius afterwards.

-ria,

faenum andfoenum.

enum, not paenitet,

not pen- ?M)r poen-.

pernicies, not -ties,better than pomoer-. proelium, not prae-.

amoenus.artus, artareearlier.;

fecundus, not foe-.arct- is

pomerium,

femina, not foe-.fetus, not foe-.

auctor,

etc.,

not autor.auct-.

Hadria, not Adria.-probably better than ar-.

setius, not seciua.

autumnus, not

harena, harundo, haruspex solacium, not -tium.

caecus, not coe-.caelebs, not coe-.

soUemnis, not -nnis.suscenseo, succenseo.-tio.

hedera, better than ed-.heres, not haeres nor eres.

better

than

caelum, not coe-.

caenum, nx)t coe-. camena, not -moena.cena, not coena.^oondicio, better than -ditio.

indutiae,

nx)t -ciae.

infitiae, not -ciae.

maereo, maestus,f/iaw

nx3t

suspicio, better than tempto, not tento. moe-. Thrax and Thraex.

multa, not mulcta

{old),

trans-

and tra-mitto,

etc.

co-necto,

etc.,

better

nactus and! nanctus.

uraQTViBandwaxoT, better than

conn-.contio, not concio.

negotium, not -cium.nuntio, not-cio.

hum-,

'

L.O., pp. 204-6.

^

Ov.

Am.

I. iv.

2

demands

coena.

CHAPTER

III

ALPHABETICAL VALUES 19.

Anything that

may

be said in this chapter on the subject

of pronunciation assumes the spelHng which

we have

recognised as

current in hterary circles at

Romewill

during the activities of the writers

named on ourwe

title-page,

and

be primarily concerned with the

values of the letters as elements of rhyme.rely will be found in the Appendices.

The evidence on which

Here we confine ourselves

to the conclusions

we have reached(e.g. U)

;

and we begin with the remark

that a double lettersingle,

has no more alliterative value than a

and that short vowels which are not lengthened by thei tt

poet have no alliterative value whatsoever, apart fromcertain positions 20.( 22).

and y

in

The phonetic value;

of a letter often

depends on the

letter

which immediately followsfeatures of a line aright,in

and

if

we

are to estimate the alliterative

it will

be necessary to discard the face-values

many

cases,n.

m, for instance, ceases to be an

m

when

it is

proin

nounced as

Such phonetic changes are sometimes expressed

the spelling by the substitution of a letter representing the true sound, as in tantus for tamtustions;

but except occasionally, in inscripof ilhterate

and certain MSS., at the hands

workmen andItis,

careless scribes, this is not

done between word and word.^

however, important to remember that the influence of a followingletter is felt

even when theis

latter is in another word,line.

and even wheninitials

that word

in

another

The terminals and

have

^ In Sanscrit, where the changes are expressed in script, the figure is known as Sandhif and the term is often used to describe the same feature in other languages. In Welsh the mutations are chiefly found at the beginning of a word, where they often effect a striking transformation, e.g. eu pen, dy ben, fy mhen, e phen (their,

thy,

my, her head).12

ALPHABETICAL VALUEStherefore to be carefully watched,in

13

and the proper adjustments madeseveralspecified in detail below.is

pronunciation.

The circumstances under which thewill

letters 21.

change their values

be

The reader will remember what

meant byby a

like letters

and

unlike (7).

Owing to the

fact that certain letters are able tolicence,

rhyme

effectively with certain disparates

and withcomlittle

yet others under the transforming influence of a neighbour, the

number

of those

which can only rhyme with their fellows

is

paratively small.different.If these

When we comebe counted, they

to sounds, the casewill

is

a

be found

discarding niceof

distinctions

to be twenty-five.

Some, however, even of these are

too hke each other to be regarded as distinct alliterative elements,

and experiment has shown that those which are incapablewith each other are just sixteen.

rhyming

They are the sounds represented by

dhcdehijlmnorsuv and their likes, or (as we may perhaps nowcall

them without danger

of being misunderstood) equivalents ( 24).

22.

Vowels, Semi-vowels,

and Diphthongs (when not;

elided),

a

e

as in father, mate, rope respectivelye

a rhymed with ai and au,

withOhs.

ae,

and

o with oe

and

oi.

When the interjectionutinam=o-v-utinam.^

was followed immediately by u, thev,

pronunciation appears to have demanded the intervention of ae.g.

Presumably the rule applied to ano.

initial

u

following

any unelided

i

1.

When

a consonant (j)

:

as in jet.

Consonantal

i

behaved

like other consonants, and,i

between two vowels, asit

though the

were doubled, which indeed

sometimeswith the

visibly was, e.g. Maiia.

The union of the

earlier i

preceding vowel had the effect of lengthening the latter,

but did not otherwise affect the pronunciation,(TpoLa)=Troi-j-a, peior=pei-j-or.^i is

e.g.

Troia

In words Hke Teia the

treated as a vowel.

2.

When

a long vowel

:

as in feelits

;

rhymed withline,

ei

and

y.

Before another vowel inj,

own

the

i

(whether

long or short) developed a

so that Pieria=Pi-j-eri-j-a,

cuius =cul-j-us, pecori apibus *=pecori-]-apibus.2^*

It

Hor. C. I. XXXV. 38. seems unlikely thatVirg. G.i.

ei

in cases like this could

have been pronounced as

i.

4.

14

ALLITTERATIO LATINA1.

u

When

a consonant

(v)

:

as in win.

Consonantal v

behaved

like

other consonants.

In late Latin the pro-

nunciation changed.2.

When

a long vowel

:

as in food

;

Before another vowel inor short) developed a

its

own

Hne, the

rhymed with eu. u (whether longheu ubi

v,

so that fuit=fu-y-it,

c

Whether it was a vowel or a consonant after hin such words as cui, anguis, aqua, huic, hui-uSy maybe an open question, but in any case the v was vocal during=heu-v-ubi.

gq

our period, except in quum.

Thus^

qui nocuere suo

y

has the v rhyme at every ictus. =the French or Welsh u (GermanLatini.

ii)

and rhymed with thelikei,

Before a vowel

it

behaved

developing a

j,

so that LyaeusIA-yaeus.Ohs.

Except ine.

a few cases where(e.g. sylva),

it

properly introduced

the letter

seems to have been imis only found in words

borrowed from the Greek.

aeai

rhymed withas insJisle

or ah^y-e^ (a being the predominant sound).if

auei *

like av in gra.vel,

the v were a

w

;

rhymed witho.

a.

In

special cases it probably

rhymed withetc.).

(Cf.

202

obs.)

I

(but not in peius, eius,u.

euoe, oi

rhymed withwith6.

as in oh^y-es or 6oy (pronounced with a long o)

;

rhymed

quoi was pronounced like cui.

23.

Consonants (andrespectively,

h).

b

dg

as in bww, done,

gun

except

(1)

(2)

when followed closely by h when closely associated with a sharp {pt;

c s),

in which cases they were pronoimced

s^s

p

I

c respectively,

with which, whether so pronounced or not, they could

always rhyme,6

b might also

rhyme with/.

Ov. Ex. P.t.

I.

ii.

136.

Priscian says

(I. 6.

32) that the ancients

for

Nigidius Figulus

deemed

it

employed this diphthong everywhere a stnpid superfluity (GelL XIX. xiv. 8).

ALPHABETICAL VALUESObs.

15f,

Final d may have been regularly pronounced asIn hand the dis:

as

was the

case in late Latin.

said to have been silent before

a consonant.'c

k q

as in catlike

rliymed with

g.

f

pwh

in;

pwha^ (an

Irish pronunciation of what) or bhv in

ah^hmc rhymed with b and p. In late Latin the pronunciaThe first finff must have been a mere p. tion changed.has in hat, except in phth chth, where thesilent.first

h must have been

It did not afiect the pronunciation ofit

anyd g)

letter\

with whichitis

was

associated, except ut supra (h

and

said to

have been pronounced more strongly with

consonants than with vowels.

The soundsaspirate

of ch, ph, th

were those in ink-horn, to^-hat, pot-house.Aspiratepositions,I

might

rhyme with

in

favourable

and perhaps with /.respectively,

p

as in

let -pin

p rhymed with 17)

b and/.

m

(wheni.

the spelling

conforms to

Like1.

m in ram.of a word.

At the beginningBefore

2. 3.

Before a vowel within the word.

bfjmpv.initial

4.

Before an

vowel or h (whether in the same or

following line),5.ii.

when

elision did

not operate.

At the end

of a line before a

marked pause.

Like n in ran.1.

Before medial or

initial(c

dlnr stz.q):

2.

Before a guttural(a)

g k

when the

letters

are in different words (jam;

queritur, remque)(6)

when

the letters are in the same word and the

m

in thesisiii.

{umqudm).

Like ng in

nng ^

(with Uberty to rh5T2ie with an ordinary

n at

least occasionally)

Before a guttural,

when the

letters are in the

same

'

Mar. Victor. K.

word and the VL 15. 21.

m in arsis

(umquam).

The sound known as agma, being that of the first y in yy a word invented by the Greek grammarians to distinguish this y from the ordinary gamma.

16iv.

ALLITTERATIO LATINASilent.1.

Almost always before anline.

initial

vowel or h in

its

own

2.

In compounds of drcum before a vowel,

e.g.

circumeo

(pronounced drcuweo).Obs. A markedj'pause cannot arise within the line, and seldom occurs elsewhere, except at the end of a poem (c/. 29). The reader is reminded that the last consonant in a line

adjusts

itself

to a following initial like 17)

any other

( 20).

n

(when thei.

spelling

conforms to

Like n in ran, subject to the following Like ng in

ii.

nng

(with liberty to

rhyme with an

ordi-

nary n at least occasionally)Before a guttural, when the letters are in the same

word and the niii.

in arsis (unquam).

Silent in

mensa

(in

the sense of " table ") and perhaps in

conjux,rs

as in hoviible as in hisswriters

the

littera

canina;

rolled asit

in Scotland.z.

strongly sibilant(e.g.

rhymed with

In the oldersilent at

Catullus, Lucretius)

was sometimes

the end of a word