Victor Henry.

A short comparative grammar of Greek and Latin for Schools and colleges online

. (page 12 of 37)
Online LibraryVictor HenryA short comparative grammar of Greek and Latin for Schools and colleges → online text (page 12 of 37)
Font size
QR-code for this ebook

expect that, in this scarcely conscious process of analogy,
he will be satisfied with a merely external and superficial
resemblance. Hence the numerous etymological deviations,
the cause and influence of which will be best illustrated by
a familiar example.

We have in French a suffix -ier^ the regular representative
of Latin -arium, -iarium^ which has been added, among other
words, to various words ending in an etymological t : lait
lait-ier, sabot sabot-ier^ clou clout-ier^ etc. But as the t has
long ceased to be pronounced in lait^ sabot^ and has even
ceased to be written in cloUy the speaker does not now iso-
late in thought, in the derivative words, the element -ier,
which he no longer perceives in them, but the element -tier^
which he fancies he perceives in them, and he transfers this
element entire to other derivatives ; hence from the words
hijou^ cafiy fer4)lanc he forms the secondary words bijou-

^ E.g. *8khid (idea of splitting) and *to (demonstrative, cf. Gk. rd), whence
*8khid'td; literally " split-it," Gk. <rxur-r6-s, " that which (is) spUt."

Digitized by



tier^ cafe-tier^ ferhlan-tier^ in which the t is to the etymo-
logist a mere monstrosity, but to the psychologist the sign of
an intellectual operation of remarkable delicacy. It is now
clear that, without the check afforded by Latin, and without
the historical evidence of the French forms, we should be
forced to admit in French the real and primitive existence of
this pseudo-suffix -tier^ the origin of which would escape us.
Now such a check and such evidence are absolutely wanting
to us in the case of the primitive Indo-European language ;
and corruptions of this kind, of which hundreds of examples
might be found in French derivatives,' and of which Greek
and Latin will afford us many instances, must necessarily have
played havoc with the Indo-European language also, from the
mere fact that this language passed through human mouths
and was thought about by human brains.^

The fact is, that linguistic analogy,* which is a special
form of the principle of association of ideas applied to language,
is not merely an indispensable element, at once creative and dis-
turbing, in the formation of the words of a language ; it may
be said to be the very essence of human speech. If we just

^ When the analogy is quite strict, as is generally the case, there is no better
way of representing it vividly to the eye than by a formula of proportion,
e.%. hijoutier I bijou^cloutier: clou{t),

^ Gf. A. Darmesteter, MoU NouveauXf passim.

3 [The English language offers many examples of the influence of analogy.
Thus the ending ation properly belongs only to words derived from Latin
verbs of the first conjugation, like contemplation^ mediation; but the ending
ation in such words being wrongly isolated, and regarded as a fit termination
for abstract words irrespective of their origin, the English languaj?e has been
enriched by the acquisition of such hybrid words &3 starvation &nd flirtation.
The word starvation is said to have been first used in the House of Commons
by Mr. Dundas in 1776, and to have earned him the nickname of '* Starvation
Dun das." The correct writers of the early part of this century recoiled from
it with horror; but it now seems to have passed into general use.]

* [Besides V. Henry's valuable Etude $ur V Analogic (Paris, 1883), which
deals chiefly with Greek, the English student may consult on the subject
of analogy in general Sayfie!^ Comparative Philology^ chap. ix. (Macmillan^
1874), Paul's Principles of Language (Swan Sonnensohein, 1888), and B. I.
Wheeler's very useful study of An^rio^^ (Cornell University, United States,
1887), which contains many Englishitiustrations and a list of authorities.
The special application of the principles of H^alogy to Greek and Latin has
been discussed by the translator in a paper Oii ** The Laws of Analogy in
Greek and Latin," published in the Transactions of^he Oxford Philological
Society for 1887-8 (Clarendon Press, Oxford, 1888, 1«.);]- .

Digitized by



think of the ease with which a child learns its own language,
of the prodigious effon of memory implied in the storing of the
hundred thousand words of a language in an ordinary brain,
of a million of words or more in that of a polyglot, we shall
be convinced that this is only possible because the words so
learned arrange themselves in our mind in families and groups,
by a continual and almost unconscious process of classification,
a process not etymological of course, but purely empirical and
based on merely external features of resemblance. Without
this phenomenon, the understanding of a language would be in-
conceivable. Pronounce for the first time the word swiftest
before a child who has not heard the word before; he will
understand, provided that he knows the positive svnft. Why ?
Because the connexion of quick quickest^ kind kindest^ big
biggest^ etc., immediately spoke to his mind far more elo-
quently than the best of dictionaries. But do not be surprised,
after that, if he should happen to say also *littlest or *bad-
dest. Suppose Demosthenes was the first to use the verb
<tii\L7nrl^€iv in the celebrated phrase " ^iX-iinrL^ei. rj TLvOia " ;
it was none the less understood, even on its first utterance,
by the most illiterate of his contemporaries, just as in our own
day the unknown French journalist was understood who first
created the word " Opportunist." Owing to this power of
analogy, it is no exaggeration to say that each individual de-
rives his language from himself, at least as much as he learns
It fi:om others ; hence it is not surprising, if language, thus
created anew by every thinking being, necessarily undergoes
from generation to generation many accretions which, while
enriching it, are incessantly changing its form.

(84) With these reservations as to the use and precise mean-
ing of the term " root," we shall apply the name root to that
element which gives the essential meaning of a word or group
of words, while we shall apply the name suffixes or affixes ^

^ In the Indo-Eoropean languages the only kind of derivation known is
derivation by means of sufl&xes. Derivation by means of prefixes is never
anything more than apparent ; for example, in certain compounds of which
the first term has ceased to be used as a separate word, e.g. dpt-yvuro-s
(well-known), in which occurs a word *dp *&pi (good, of. Ap-ia- TO'S)y or in
simple verbal combinations, ir/w-dyw, per-lego^ infra 178.

Digitized by



to those elements whose addition determines the precise shade
of meaning to be attached to the vague and general meaning
contained in the root. A suffix then is everything which,
in a given word, occurs between the root and the termina-
tions of declension or conjugation, e,g, -<rt- in ^e-o-t-s, -/xo- in
Btty-jx^^ -a-avpo- in Orj-aavpo-'*, 'fxaro- in Tt-/Ad-o-/A€i/, etc. The de-
clinable or conjugable combination thus formed, e,g, Oea-f,
OiDfiQ-j Tlfido-, is called the stem (theme, radical [or base]). A
stem is called primary, if only one suffix is attached to the
root, Tt-/Ai} ; secondary, if there are two, that is, if it is de-
rived from the primary stem just as the latter is derived from
the root, e.g. rl-fid'O- derived from rl-firj just as tZ-o- is from the
root Tt-, 1st sing. pros. ind. tI/xoo), twu ; tertiary, if there are
three, Tl-/xa-a-/A€vo-, and so on. But, as the same processes are
reproduced indefinitely in all the stages of derivation, it is
sufficient, for the purpose of studying derivation as a whole, to
distinguish between primary derivation, comprising forma-
tions derived directly from the root, and secondary deri-
vation, including all others. These, together with nominal
composition, will form the three branches of our study of

Digitized by



(85) A Stem is called nominal, e,g. Xoy-o-, or verbal, e.g,
\4y^'^ according as it is capable of attaching to itself the
terminations of declension or of conjugation respectively.
These two grammatical categories are in principle quite dis-
tinct/ but they cannot fail to react on one another, thus
mutually enriching each other. Thus from ck-koXc-o), to call
forth (cK-Kc-zcXiy-K-a, ii-^-Kkyj-drj, Ix-icXiy-To-?, etc.), the language
formed iK-KXrf-o'C'dy assembly ; from this noun, the verb €k-k\i;-
crt-a^(i>, to hold an assembly, and from this verb in iU) turn the
substantive cK-KXiT-o-t-acr-TiJ-?, orator, and the adjective cK-KAiy-o-t-
a<r-nKo-9, and theoretically the process might be continued up
to infinity. But, as in every language there are more nouns
derived from verbs than verbs derived from nouns, it seems
most natural, in approaching the study of the two systems
of derivation, to consider the verbal stems first.

Furthermore, in each system of derivation, the formations
may be distinguished, according as they go back to the Indo-
European period, or are peculiar either to Greek or Latin, and
seem to have been subsequently developed in either language.
No doubt in the latter case they are not, properly speaking,
primary ; for even when they seem to have arisen firom the
simple combination of a root and a suffix, yet, having arisen at
a time when root and suffix had long ceased to exist as separate
categories, they can only be due to a secondary and often a
very complicated operation of analogy. But, on the one hand,
as we have just seen, there is scarcely any Indo-European form

1 That is to say, \6yos is no more derived from \iy<a than Xeyw from Xcyos;
but both come, by a separate and independent process of derivation, from
a root *leg, which appears in its normal form in the one case, and in its
deflected form in the other.


Digitized by



to which a precisely similar origin may not conceivably be
assigned ; and, on the other hand, when an Hellenic form is
wanting in Latin, or vice versa^ we are not thereby justified in
thinking that it was wanting in the common language and that
the language which possesses it has formed it independently;
for it is also possible that the other language has lost it.
Hence there is no reason why we should not put on the same
level all formations, whether common or not, which are or seem
to be primary.

Section I.


§ 1. — Common Formations,

(86) The whole of this system is characterized by one fun-
damental distinction. We know that a very large number of
verbal formations, e.g, in Greek the present of verbs in -<d,
all subjunctives, all futures, and in Latin all presents, etc.,
show before the conjugation-ending a vowel o or «, alternating
according to fixed and invariable rules.^ In consequence of its
extreme frequency, the name of thematic vowel has been
given specially to this vowel o/e, and hence the name of
thematic formations is applied to those in which it is
present, n on -thematic to those in which it is absent, e.g, in
Greek the sigmatic aorist, the aorists passive, the present of
verbs in -fu, and in Latin the subjunctives, imperfects, etc. In
spite of the fundamental defect of this terminology (for i-Xv-Orf-
or legB-bd' is evidently a theme or stem just as much as Xv-o-
or leg-e-), we must needs adopt it ; for we shall see later on,
in studying the conjugation system, how necessary it is to dis-
tinguish everywhere the forms which contain the thematic ejo
from those which do not contain it.

Moreover, even at this point, the distinction is necessary.
Latin, though it kept in its conjugation a good many non-
thematic stems, retained scarcely any in the present ; in other
words, it no longer retained any verbs in -/xt. The thematic
vowel was extended in Latin by a process of analogy from
which even Greek was not entirely free, until it invaded all
' See infra 269.

Digitized by



the present steins which Greek still kept in their primitive
simple form; so that the regular correspondence between the
two languages would seem to be broken from the very begin-
ning, if we trusted merely to appearances.

(87) I. Simple root-stems (in Greek, present stems, or more
commonly aorist stems, when the stem of the present is formed
by means of reduplication, infra II). — The simple root with
no affix immediately precedes the termination, and appears
either in the normal or weakened form, according to a regular
alternation (sometimes, however, interfered with by analogy),
which will be considered among the phenomena of conjugation.^
Presents : <f>rj-f^^ <^a-/xei/ (root <^d, Dor. <^d-/xt) ; €7-/xt t-fx€v ; ct-fit,
Lesb. €/A-/At ( = *i<T'fjt.C) i(r-/jL(v, Aorists: c-Orj-v c-^c-/x€v, c-Swv
e-8o-/A€v, t-a-Tq-v (Dor. c-CTTd-v) l-o-ny-zxcv, etc.

Latin has in this class : es, es-t^ es-tis, etc., from the verb esse,
root es ; €s-t (he eats) =*ed-t, root ed; vol-t {he wishes), etc. ;
f-s, i'tj the present of the verb f-re, except 1st sing, and 3rd
pi., which are thematic ; that of the verb da-re, except dd ;
perhaps that of the verb stCL-re, except stO {stCL-s = l-ord-s, except
iu respect of the augment) ; and by a curious peculiarity some
forms of a verb which in Greek, on the contrary', is entirely
thematic, fer-s, fer-t, fer-tis, ferrte, cf. <^€p€ts, <^€/)€t, Kfycp-e-Tc,
But the Homeric form <f>€p-T€ (II. ix. 171) is no doubt a relic of
the old non-thematic conjugation of the root <l>€p.

II. Root-stems preceded by reduplication ivith the vowel i (in
Greek present and imperfect stems, in Latin lost). — The root
alternates : rt'Orj-fu Tt-^€-/x€v, t-cmj-fu Dor. l-ara-^L ( = *<rL-<rTa-fu)
L-oTTa-fia^, 8t-8<i)-/u, L-rj-fiL { = *<rL'<rrj'fJUf cf. Lat. s€-men\ impf.
i'TL-Orj'V i-rC'dc'/jLcv, etc. ; with SO called Attic reduplication
{infra 240), ovmy/xt (to benefit), aor. wvd/jii^v. In Latin si-st-0
( = Gk. L-a-rrfiu) and hi-b-6 ( = Sk. pi-bd-mi) have passed into the
thematic conjugation.

III. Root-stems preceded by reduplication with the voivel e
([)erfect stems, improperly called in Greek second perfects^). —

^ The same vowel- gradation takes place in every syllable, whether a root-
syllable or suflftx-syUable, which immediately precedes the conjagation-
ending, and does not contain the thematic «/o. Cf. infra 269.

2 The ordinary grammars have been very unfortunate in their nomen-
clature; the so-called second perfects are much more simple and primitive

Digitized by



The root alternates between the three grades :^ Gk. Fot8-a
ft8-/i€v, yi-yov-a ye-ya-ficv, A,e-A.otir-a Xc-Xci/A-fuu, vX-rfKovd'-a (Horn.)
and cX-iyXv^-a, iri-<j>€VY^^ Xi-kyjO-a, etc.; Lat. vldrlf tO'tond-lj
spchpond'lf pe-pend'l, pe-pig-f, Ifqu-lffUg-i, tulrl=te'tul'lj cf.
rettulit, fSC''l=*fe-fSC'l (cf. tOrjKo^ xc^ctKa), showing the normal
grade of the root as contrasted with the reduced grade of fdc-io
{supra 41, 3).

In Greek, roots ending in a non-aspirated guttural or labial
often show in the perfect the corresponding aspirate : vkiK-w
iri-irXex-a, Xey-co Xc-X€;(-a, )3A.a7r-Ta) /?e-/?A,a<^-a, Tpip-in ri-rpitfira,
etc. This phenomenon is by no means invariable : we have
just seen Trcc^cvya and XiKoiira, Moreover, it is somewhat late ;
the aspirated perfect is unknown to Homer; Herodotus and
Thucydides have only one instance, 7re7rofi<f>a; the tragedians
another, rcrpcx^a ; its wide extension dates from Aristophanes
and Plato. Hence it must be regarded as an analogical cor-
ruption, aided perhaps by the tendency of popular Attic to
aspiration ; e.g. ypdifyw regularly had 1st sing. perf. y€-y/oa<^-a,
and no less regularly 1st pi. perf. yc-y/oa/*-/x€v ; on the other
hand, Tpi^-w also had 1st pi. perf. Te-Tplfi-fiev, and the likeness
between yeypa/x/xtv and Terplfifiev brought about the likeness
between y€ypa<f>a and rerpZc^a (cf. supra 62 Cj*

(88) IV. Stems tvith suffix -na- {weakened -ricl-): Greek
presents. — The root is generally weakened : Safi-vrj-fxi (to sub-
due) = 8a/x-vd-/xt, 1st pi. SaLfx-va-fiev ; a-Ki^vrj-fU, KLp-vrf-fii, 8v-va-/xat,
fxdp-va-fiai. ] normal grade in Trip-vrf-yn. (to sell), cf. the deflected
grade in irop-vrj (prostitute). There is a transition to the the-
matic conjugation in 8a/A-va-«o=8a/AV7;/u.

V. Stems with suffix -new- (weakened -nO-): Greek pre-
sents. — For the regular gradation -vev- -viJ-, which Sanskrit
shows in this class, e.g. sandmi (I conquer), 1st pi. sanumds^
Greek substituted through analogy a gradation -vv- -vu- modelled
on the alternation -va- -vd- of the preceding class, e,g, SeiK-vv-fxi
S€tK-vv-/jL€Vf like Sd/jL'vd'fu Sdfi'vd'fxev, Another corruption is

than those called first perfects ; the same is the case with the second aorists
passive as contrasted with the first aorists, etc.

1 The whole subject of reduplication and vowel-gradation is further dis-
cussed in connexion with conjugation, infra 237 seq.^ 292 seq-

Digitized by



equally noticeable. As in the preceding class, the root ought
to be weakened, since the Sanskrit accent falls sometimes on
the suffix, sometimes on the termination, never on the radical
syllable. But Greek shows only a very few forms with
weakened root, op-vv-fii (I rouse), Td-vv-fiai (I stretch) =*t^-vv-
/xai, cf. rctVco and raros ; and most verbs of this class, tn^vv-fxt,
pj^-vv-fiif pmr-vv-fUy fcvy-vv-fu, 8ctic-vv-/Eu, etc., show the normal
grade. The vocalism of the sigmatic futures and aorists, in
which this grade is regular, ^fo), piy^w, Set^co, ^cv^co, must have
influenced the vocalism of the present.

There is a transition to the thematic conjugation in Greek
ra-vv-o) (I stretch), and perhaps also in the form /mt-vv-co mi-nu-O
(I lessen), which is common to Greek and Latin.

(89) VI. Stems with suffix -e-/-o- unaccented in the primi-
tive language : Greek and Latin presents. — This class is large
and well-known : Gk. A,ey-(o (Xey-o-/A€v Xcy-c-rc), <^€p-a>, Xt^O-o)
=XdO'U}, Xct?r-<i>, <^€vy-a); Lat. leg-O. fer-Oj dlc-Oj ftdrO, dUc-d.
As is indicated by the theory and shown by the examples, the
root, which was accented in the primitive language, always
assumes the normal form; we have already had occasion to
contrast XetTr-co and I-Axtt-o-v, ^cvy-oi and l-^vy-o-v, ^cr-o-ftat and
c-wT-o-fwyv. In the very rare cases in which the root seems
to be weakened in the present, Gk. apx"*^> fta;(-o-/iai, ypa<^-o>,
Lat. dIrOj scdlhd, GrsBCO-Latin 5y-a) dg-O, ayx-w ang-O^ etc., pro-
bably a second aorist stem has been substituted for a regular
present stem like ^fju^x-o-^ajLy ^ypiKfya), etc. It is not even
necessary to suppose that this substitution is due to analogy ; ^
for, just as the imperfect is the present tense augmented, it is
very possible that the so-called second aorist is the augmented
tense of another present, almost lost. In other words, the
known series l-^cvy-o-v ^cvy^o) requires a corresponding theo-
retical series l-<^iry-o-v *<^vy-a) ; but the second term of the latter
spread very little and ended by falling into disuse, whereas the
other series remained unchanged.^

» E.g. in accordance with the formula ypd<p<a : (ypa<pov(^ * e^grhh-o-m^ aorist
taken for an imperfect) «0e/)w : (<t>€pov. .

We must even go further. Given a root ^bher^ it could no doubt be
conjugated, at the option of the speaker, with no affix *bhSr-mi (cf. Lat.

Digitized by



Much more rarely the root seems to be deflected, e,g, rpdry-ta
(to gnaw), aor. l-rpay-o-v. Here it is the vocalism of the per-
fect which bend), cf . wXiK-o} (to plait), nec-tOj pleC'tO, etc. If it
seems frequent in Greek after a labial (tvttto), kotttcdj /xa/oTrro),
pLimaj etc.), the reason is that the group iry phonetically
becomes ttt ; hence all these cases belong to class XI.

XIV. Stems vnth suffix -dho- (?), Gk. -Oo-, Lat. -do-: Greek
and Latin presents. — This sufiix, which very rarely appears as
a primary suffix, forms in Greek : a-xi-Bui (to have), root <r€x ;
vrf-Oia (to spin), cf. ve-o) ; Trkr^dta (tO be full), root wXiy, cf. Trt/i-

irX-q-lxi and pU-nvrs; ta-Oia (to eat) = *l8-^a), cf. l8-(i>; ax-Oo-fxcu
(to be grieved), cf. ax-yv-fxat (same meaning), etc.; in Lat.
ten-dOj^ cf. T€tV(i> = ♦roz-yo), ^fen-dO (I strike) in offiendG, defendO,
cf. Gk. $€iv<a=*6€V'y(0j fren-do, cLfrem-d, etc. It is not known
whether pellOy tolldy etc., belong to this class or the following
one ; for from a phonetic point of view pelld may go back
equally well to *pel'dO or ^pel-nO ; the Greek correlatives have
the suffix -yo- (ttoAAo), TeAAw).

(93) XV. Stems vdth suffix -no- : Greek and Latin pre-
sents. Although we cannot assign to this suffix an Indo-

1 It will be seen that this sufidx, like the preceding, is not incompatible
with reduplication, e,g, TiraUfca (stretch) = 'rt-rn-y w, Ti'Tpd)-aK(a (to wound),
diddffKU), etc.

' It was no doubt the analogy of this suffix -laKa which introduced the t
subscript in OvicKta and other Attic spellings, supported by the best manu-

' Tendd has also been explained as *te-tn-d (reduplication and weakened

Digitized by



European origin, it is extremely common in" Greek and Latin,
in which it seems mainly to be the result of an irregular
transition of classes IV and V to the thematic conjugation:
Gk. m-v(i>, iEol. 7ro)-v(i) (to drink), cf. Lat. pO-tu-Sj SaK-vta (to bite),
T€/A-rcD (to cut), Lesb. )SoAAo/iat, Ion. PovX.ofxai, = */S6Xrvo-fiai,]
Old Lat. dornu-nt (they give),^ ne-qul-nu-nt (they cannot),
red-l-nvrYit (they return), etc., class. li-nO (to smear), si-no
(to permit), cf. supra li'tu-m^ si-tu-m. With this formation are
connected a certain number of others, much more complicated,
and apparently modified by various analogical influences.

1. In Greek we sometimes find a suffix -vco-, which, like -vo-,
occurs only in the present : iK-vio-fiou, (I come), cf . aor. iK-o-fjLrjv ;
Kv-vtii) (to kiss), cf . aor. l-KiMr-a.

2. Some verbs in -vta seem to arise from -vfw, that is, from
the suffix 'vv added to the thematic conjugation, with regular
substitution of w for u before a vowel : e.g. Slv- (to shake),
xAZvo) (to incline), Kplvn) (to distinguish, cf. Lat. ccr-n5), corre-
sponding to JEolic Stwo), xXtWo), KpLvvd), etc. ; also <^^d-F(i) (to
anticipate), tI-vw (to expiate), <^^-f(i> (to destroy), where the
radical t, always long in the time of Homer, is shortened in
later versification.

3. When the root ends in a consonant, the meeting of this
consonant with ' the nasal of the suffix seems to have usually
developed a sound which was represented as an epenthetic
vowel : ^ the suffix then took the form -avo-, e.g. a/Aa/ar-avo) (to
err, aor. y/jLopT-o-v). Moreover, in the oldest and commonest
type, the nasal of the suffix was somehow reflected in the root,
by a phonetic process not yet satisfactorily explained, though
easily conceivable : thus a root \dO (to be hidden) would give
*Xa^-vo), whence *\dv6'V<a and ^kdvO^vo), and lastly XavO-dvia.^
So also with Xayx-uvu) (root A.€yx> cf. perf. Xc-Xoyp^-a), Xa/x/S-avo),
Xifiir-dvio (to leave), TrvvO'dvofioL (to learn), and without nasaliza-
tion XrfO^voij K€v6'dv(a (to hide), auf-ava> (to increase), Bap6-dvu}

* In the very old Latin inscription known as Dedicatio Sorana: ** Donu
danwit Hercolei maxsume mereto."

a This phenomeon is exactly parallel to that of the Dutch knif (knife),
which has become in French canif^ *knnif.

' Of. fat. XiJ<ro/Aai = *Xa<?-o'o Atat.

Digitized by



(to sleep), aWd-lvo-yucu (to perceive) ; this mode of formation
was much extended by analogy.

4. In I/atin the same class of forms followed a very different
phonetic road. E,g, the root pdc (to make firm, cf. Gk. mTy-vv-fw
and Lat. pdos, treaty), by the addition of the suf&x -^o- to
the weakened form, will give successively *paC'nOj ^pag-nO
and *paiig'n6^ after which, the group fign becoming reduced to
Hg^ 1 there remains the known form pangO. In the same way
we may explain tangOy stringOj pandd, lambOj as compared
with taC'tu-s, stric-tus, pat-eO, lah-iu-in (lip), namely, through
*pat-n6 (cf. Grk. TriT-vrffiC), ^lob-nO, etc. ; and it will be noticed
that in certain verbs {jung-d junxl junc-tu-my cf. jug-Vrm^

Online LibraryVictor HenryA short comparative grammar of Greek and Latin for Schools and colleges → online text (page 12 of 37)