Skip to main content

Full text of "Some Greek etymologies"

See other formats







This book is DUE on the last 
date stamped below 


C68 Sane Greek 

. V :* 11?. gtrr 

V . 










COUNCIL, 1893-94. 


Vice- Presidents. 

THE REV. PROF. W. W. SKEAT, Lirr.D., M.A., LL.D. 
THE REV. A. H. SAYCE, M.A., D.C.L., LL.D. 

Ordinary Members of Council. 




C. A. M. FEXXELL, M.A., LiTT.D. 





PROF. W. P. KER, M.A. 


W. R. MORF1LL, M.A. 

BEXJAMIX DAWSOX, B.A., The Mount, Hampstead, London, N.W. 

Son. Secretary. 
F. J. FURXIVALL, M.A., Ph.D., 3, St. George's Square, Primrose Hill, X.W. 

The Philological Society is formed to investigate, and to promote the study and 
knowledge of, the Structure, the Affinities, and the History of Languages. 

Each Member pays two guineas on his election, one guinea as entrance fee, and 
one guinea for his first year's subscription. The Annual Subscription becomes due 
on the 1st of January in each year. Any Member may compound for his subscription 
by the payment of Ten Guineas, exclusive of his entrance-fee. 

The Society's Transactions are published yearly. Occasional volumes are also 
issued, as the Funds allow. The Society's Xew English Dictionary is in course of 
publication, under the Editorship of two of its former Presidents, Dr. J. A. H. 
Murray, and Mr. Henry Bradley, M.A., and the auspices of the Clarendon Press. 
Parts I. to VI. (including Vol. I., A-B, and C-Cons of Vol. II.), and the 
first portion of Vol. III., E-Ev, have appeared, 12s. 6d. each. Both Part VII. of 
Vol. II., and Part II. of Vol. III. are at press. (Volunteers willing to read for the 
Dictionary should write to Dr. Murray, Oxford.) The Society has approved of a 
partial scheme of Spelling Reform embodied in a pamphlet, " Partial Corrections 
of English Spellings. 11 Triibner & Co. &d. 

Members are entitled to a Copy of all Papers issued by the Society ; to purchase at 
reduced prices (prepaid to Treasurer) the Society's publications, and the 12s. 6d. 
Dictionary Parts at 7*. ; and to attend, and introduce a friend to, the Meetings of the 
Society, held on one Friday in every month, from Xovember to June. 

Subscriptions are to be paid to the Treasurer, or to the Society's Bankers, Messrs. 
Barclay, Ransom, & Co., 1, Pall Mall Ea.-t, ^Y. 

Applications for admission should be made to the Honorary Secretary, Dr. F. J. 
Furnivall, 3, St. George's Square, Primrose Hill, London, X.W. 






(1) The modern theory that the 'prothetic,' or, as the Greek 
grammarians called it (Curtius, Grundziige, 5 720), 'prosthetic,' 
vowel is in most cases really the first vowel of an originally 
bi-vocalic root can scarcely be rejected (#) where other languages 
besides Greek have a similar vowel, as in e/>e/3o? epev^ o'Ao0v? 
besides Armenian erek orcam o\l (Persson, "Wurzelerweiterung, 
p. 246, n.), or (5) where two forms can be best explained by 
starting from a bi-vocalic root, e.g. avgia Sk. vaksh- from aveks- 
in a(F}e^w, avpa l Sk. vd- from ave- in a( /")//, Lat. unguis 
Sk. nakhds from onokhv- in owj; : though why one language 
dropt the second vowel and another the first we do not in the 
least know. But in some cases such an explanation is impossible, 
and the word must be regarded as a compound. 

A. Latin in-cunus in-clutus in-columis and I think in-vltus are 
but emphatic forms of cunus clutm columis and * vitus ' forced ' : 
' in enim saepe augendi causa adicimus,' says Festus. This in- 
may be identified with the Preposition in meaning ' upon ' (quite 
a different word from in meaning ' in,' which goes with eV), Greek 
ava in ava a/oJTnpiv (=in sceptro). The original meaning was 
' up ' and so ' upon ' (cf . German auf in both these senses), while 
in the Latin Adjectives given above and the Greek Adjectives 
to be given below we have a transition of signification from 
' up ' to ' in a high degree, quite.' The form in the TJrsprache 
would be n, represented in Latin by in-, in Greek by av- before 
a vowel, 2 a- before a consonant, as in the following words : 

-/3A)/%/ao'? '.weak, gentle,' beside /3A//X/ 30 ' 5 - The termination, 
which appears also in f)de\v-xpds /teAt-x/so? irevi-xpds, must go 

1 In the only place in which it occurs in Homer, Od. 5. 469, aSpy means the 
morning breeze, i)ia9i irp6 ; and with it in this sense Buttmann rightly connects 
aijptov 'in the morning, to-morrow.' So in my 'Etyma Grseca ' I have ex- 
plained Homer's i]a>s, Lesbian avovs, as from avo-, an Ablaut of ave- in &ijfu : to 
a people dwelling near the Mediterranean the morning breeze would be the 
natural herald of dawn. The Attic form etas takes its aspiration and accent 
from ?i\ios. 

2 In some dialects before a consonant also, Horn, av ' up,' Theocr. o/j.-ni/jiva- 
ffKo/j.evtf (in which the vocalism shows the presence of a sonant). 




with xpotd x/><i>s ' surface, skin, colour,' so that /3\i/-x/o* means 
' weak-looking, weakly ' : the root is mle- (not mla-, since /3X?/- 
Xpo? occurs in Doric), Sk. mid- 'to wither,' with Ablaut mlo- 
in Irish Uuith 'smooth, soft,' and mele- in /te'.Xeos 'useless.' 
Quite a different word is f3\ag ' slack,' in which the a must 
be due to contraction (Kretschmer K.Z. 31, 295), or we should 
have */3\j/f : as vea% or veijg comes from i/e'os, so I would explain 
ft\at~ as for *[3\aFdlj or *j3\af yg, from a simpler form */3\afd<;, 
mlvos, Lat. mollis for *molvis. The root appears in Gothic 
ga-mahjan ' to crush,' and Eng. mellow. 

a-0e'ff0aTos ' marvellous,' beside 0e'<r0aTos ' divine ' : literally 
' struck (i.e. made) by a god,' cf. Trpoa-Qmos ' made in addition, 
new,' and /-0aTos (Hesychius) <-0a<rtos ' made double.' The 
root of this -0aTo* is ghvn-, ghven-, in delvw l ' strike,' /u\/y- 
0arot ' struck by the millstone,' and, with a transfer of meaning, 
0oVos ' slaughter,' A/3//i-0aros ' slain in battle.' The first element 
of 0e'<r-0aTos appears with a ' determinative ' (Brugmann, 
Grundriss, 2. 60) in Beds, i.e. *0eaos : which, however, can hardly 
go (as Fick thinks it may, Worterbuch 4 1. 469) with Lithuanian 
dwases 'spirit,' Middle High German ge-twus 'ghost,' for the 
Greek gods were by no means spiritual beings. 

So we have a-[iavp6<s ' dim," beside fiavpd? (Photiusj ; and in 

II. 24. 753 the two readings a/jn^da\oeaaav and fi.ixOa\6eaaav t a 

word of unknown derivation (the connexion with o^fxe'io, Hoffmann, 
Bezz. Beitr. 15. 84, is absurd). So I would explain the o- as 
intensive in 

a-'ye/jw^ 09 ? *yepa,-oxos ' holding privileges.' 

a-(F}i]ffu\o<} ' wicked,' beside Sk. vdtulas ' mad.' On this 
dialectic change of TV to av see Classical Review 6. 259 : I 
connect a\o-ffvcvrj ' goddess of the sea wave ' with Irish tond 
' wave,' a<ri;0/yXos- (below) with Tw0\o's, <n;xi>o9 ' long, numerous ' 
with Tvyuv ' ordinary,' as a Litotes for ' considerable.' Thus alone 
can we fairly explain caavs and Lat. densus (i.e. *dent-tos) beside 
Albanian dent ' to make thick.' 

a-*y>os ' at the top ' (it never means ' sharp,' and therefore 
cannot go with *.-/9, Lat. acud], beside 0oXa-xy>o't ' white-headed' 2 

1 As Lat. ferio means both 'strike' and 'cheat.' and icpoCa-iy both 'striking' 
and 'cheating' (Ar. Xub. 317), so with Beivta I would connect <pfv& 'cheat': a 
Doric word, like K<50oAos ' rogue,' as the o, for ij, shows, with a dialectic <f> for 8 
as in <f>eo's for 0t6s (G. Meyer, Griech. Gramm. 2 211). 

2 The first element is bhl-n-, cf. bhl-n- in <f>a\\6s ' white ' (Hesychius) and 
Lat. fullo ' clothes-cleaner.' c 


(Schulze, Quaestiones Epicae 464), from a by-form of icdpa. On 
these by-forms see Danielsson's Grammatische iind Etymologische 
Studien pp. 1-56, Johansson K.Z. 30. 347-350, Johannes Schmidt's 
Pluralbildungen der Indogermanischen Xeutra pp. 363-379. From 
the same root as icdpd we may deduce () Kaipo? (i.e. */ca'/>/os) 
'thrum,' end or top of the thread; (i) *ra/>Tos Kpdros 'headship, 
power,' quite a different word from Kparai'-\ewv 'rooky,' Gothic 
hardus 'hard' (with which we may put Kep-rofia 'hard words'); 
(c) K\?ipo<3 'lot,' a Dissimilation for *Kpapos, cf. Arcadian Kpapiw-rat 
and Rhodian 'H.\o-Kpdp>j<! (^leister, Griechischen Dialekte 2. 104, 
G. Meyer 160), the word thus meaning 'head' as sign of in- 
dividuality, going with i/av-A.-\?y/jos vav-Kpapo? 'householder' (what- 
ever the first element of these words may be) and Hesychius' 
icpalpa 'head 1 '; (</) Kpanrd\ij 'headache,' for *Kpaffi-ira\rj ' a fight 
in one's head,' the first element being Locative of *K-/>as, while 
from another form *Kpd-Trd\rj is borrowed Lat. crapula : 

d-fiei'vwv ' better,' from /mevo-s ' strength : ' 

d-ffKijO)')? ' safe,' quasi ' supported, cared for ' (cf. with Active 
meaning, aice6p6<i 'careful'); a Dissimilation for *a-axi}di^, from 
the root of o-^?yya, e^o) (in a-^edelv the x i 8 retained through the 
analogy of <TX^ V } ' 

e's ' hastily,' from aTrepx^ ' hasten ' : 

'stiff,' from a root meaning 'to be hard,' whence also 
a?e/u.(fiv\a ' pressed grapes,' Sk. stambh- 'to restrain, hold fast ' : 

a-<rw0//Xo9 (with Aeolic accent) ' insulting,' *rv0j;\os (see above 
on ayo-i/Xos) going with Ty0Xo's ' blind,' TU0w^//s ' dull,' rt)0os 
' conceit ' : 

d-Tap-rrjpds 'baneful,' *Taprrj a Subst. from *rapro9 Part, of 
to distress ' : 
' strained,' Lat. tn-tentus, from retVw : 

' swelling,' Lat. turgidus, from a root tver-g- (Frohde 
B.B. 14. 107), an extension of the root tver- (see on aavpta^p 
sec. 5, and, on the pv, pv^w, p. 11) : 

d-va-raXeo-i 'squalid,' Od. 19. 327, cf. Sk. $ush- ' to dry,' Lat. 
SU-dtlS ' dry ' ; from *d-ffvara\eo9 as avirvo^ from *ciavTrvos. 

So, with dv- for n- before a vowel, I would explain dv-de&vo's 
'quite dowerless,' dv-de\Tno<i 'quite unlocked for,' and Hesiod's 
dv-aTrvevaTO'i ' quite breathless,' Suidas' dv-d^ivuxnos dv-aTnaicnos 
(J. Schmidt K.Z. 23. 273). 

B. The same intensive particle appears in several Yerbs : 
Lat. incito infrinyo ingemino innovo intremo etc. beside cito frango 


etc., Greek aaicapi^w dtnrai'piv 'palpitate ' beside oxapi^u a-n-alpw, 
afffapayeu} ' clank ' (Theocr. 17. 94) beside o-^a/str/eo/iat 'burst,' 
and a-Kovw ' hear ' beside ro(/ : )ew ' perceive.' So 

a{F}eicw 'sing' beside o'ca : i.e. uailia used absolutely means 
' make my meaning known,' used with an Ace. ' make known, 
celebrate.' In the original signification 'know' we have (Hoffmann 
B.B. 15. 62) Cyprian ueice 'hear' and the common word aladdvofiai 
(i.e. *d-fic-0dvofiai) 'perceive': for the transition of meaning 
from ' know ' to ' make known ' cf. ia-ropta ' knowledge ' (TO eiSevai] 
or ' narrative ' (TO elcevai 7roie1v\ and ^I'jviLaKui ' know ' beside 
ryvwpl^w ' make known ' : 

a-\e"(u) ' heed ' beside Lat. *lego in diligo intellign neglego (quite 
a different word, as the Perfects show, from lego ' gather '). From 
a Neuter Subst. *a/\67o? comes aXeryeti/o? ' demanding caution, 
troublesome ' ; which, with Ablaut, and without the intensive a-, 
appears in Hesychius' \a^eivd' Seivd. From this ^aXe'/os come 
further (a] ^va-ij\e^i'^ 'cruelly troublesome,' Homeric epithet 
of war and death ; (b} -rav-t)\e^rj^ ' intensely troublesome,' used 
of death, with the derived sense of -ravvta ' stretch,' as in the 
Homeric use with epica, fidxyv, TTOVOV, ' to intensify ' the strife, 
etc. ; (c) ajr-j/Xe'/ews ' most carefully,' the -o- heightening the 
meaning : 

a-\ei(f)w ' smear,' cf. Lat. de-libutus ' besmeared ' : 

d-XvKrafw a-\vffaia ' am in distress,' going with \vaaa (i.e. 
*\VK-JO), which in Homer means ' martial rage,' the spirit of 
a wolf, \VKOI: in II. 16. 156 and 352 warriors are compared 
to wolves, and Theocr. 4. 11 Trelaai rot M/Xwv KO.I TU>? Av*-o? avrlut 
\vaa>)v. shows that the Greeks themselves connected \vaaa with 
\vKo<t. With these words go fiopfio-\imeff0ai ' to madden as a 
hobgoblin, u.opfiw, 1 would,' and pop/no \vKeiov 'bugbear,' literally 
' hobgoblin wolf ' : 

a-[jLa\vvto ' destroy,' from */j.a\cv<s, Sk. mrdus ' soft.' 

a/ue\~/u}, cf. Lat. mulgeo. 

C. In three other Verbs we have the ' copulative ' a- of ddpdo? 

a.Tra.9 a/r\o'o9, tlialectically a- in aceX0oV U.KOI-IS dicdXovdo? uXo^oi 

aTaXaj/Tos ; representing sm-, 'dfi-a, Lat. simul, Sk. sam ' with' (as a 
Preposition) or, in compounds, ' together.' This appears as a- in 

1 With poppet! go ft,6pf,Loi ' panics ' (Hesycbius) and I think /u/p^pos ' mis- 
chievous," fj.ep/j.ripa ' trouble,' and /j.opfj.vp<av in II. 18. 402 p&os 'flKeavo'io a.(pp<f 
(topfivpcav, rightly explained by Hesychius as rapdatTtav ' making an uproar ' : it 
has nothing to do with Lat. murmur, which would give far too weak a meaning. 


(a} a-fi.apia.vta 'fail' (the root, on -which see Neisser B.B. 19. 120 sq., 
may perhaps be found in Lat. mora), with which cf. Hesychius' 

aft.ape.1v' duapidveiv and Homer's rj[i(3pOTOv ; as a- in (i) a-fieifita 

' exchange ' beside Lat. mlgrd ' remove ' and Old Slavonic miglivu. 
'mobile,' and (c) d-jievouai ' surpass' ('change places with') beside 
Lat. moved. In these Verbs the copulative prefix, like the (of 
course unrelated) Lat. com- in commaculo comminuo concito convello 
etc., merely ' gives intensity to the signification of the simple 
word,' Lewis and Short s.v. cum. With the same force it appears 
in a-o\709, which Eustathius says was 'Achaean' for aK/a,i'j 'prime' 
(as Hesiod Op. 588 uses d/.io\~{aitj of a ' prime ' cake, paa) : I 
would connect the word with mlg- in Lettish milst ' to swell,' 
Lat. multm for *mulctus (Wiedemann B.B. 13. 303 sq.}, so that it 
will mean ' swelling, climax,' and WKTOS ap.o\~/iv will be Cicero's 
Multu node ' late at night.' 

A copulative, though not an intensive, a- seems to appear in 
a-oaai]-ri]p ' helper,' which I would explain as ' one who hears a 
divine voice (Foaaa, as oaaa may everywhere be read in Homer, 
L. ICeyer K.Z. 28. 90 : the root appears in Lat. voco], and in 
obedience to it goes to help his comrades.' Hesychius has also the 
form offarjTijpa, referring, according to Moriz Schmidt, to II. 15. 254 
io1ov 101 (S)oaaift^pa ipovitov eff^&ijt rpoeifftf in which case the 
word must necessarily have had a digamma : 

a-<7/JoXo? ' soot,' quasi ' thrown together, collected,' from (3d\\w. 
The aft is the same dialectic representative of gv which Fick 
B.B. 17. 323 finds in 0e/}e'-<r/3<os 'life-giving' and aftewv^i 
'quench-' (Goth, qistjan 'to destroy'): and which I find, before 
other vowels than e or i, in () afuffil-aftaiva. ' a serpent that can 
go either way' and a0t-o-/3//Te'w 'dispute,' both from fiatvta ; 
and (b) 0\o?o-/3o5 ' din,' the clash produced by the impact of one 
heavy body on another, from a root bhleigv-, cf. 0\//3w 0\if3w 
'rub,' ~La.t.fllgo 'strike,' Welsh blif' catapult.' Another dialectic 
representative of gv was ", Eur. Phoen. 45 e7regdpei = e7reidpei, 
Hesychius VXXi gepcffpa eivafiV=/3ah.\et fidpaOpa ajBevvvpev. 
So I think in (a) ^"a\^ ' spray ' from f$d\\w ; (b) fa\o? ' jealousy ' * 
beside Lith. geld 'pain,' Old High German quula 'torment'; 
(c] %dty 'sea,' and Homer's eV<-^a0e\o5 'stormy,' from fid-mtv, 
the sea ' dipping ' the ships, cf. Eur. Orest. 706 vaui . . . eftatyei/ 

1 The a here is Ablaut of e as in itapds beside Lat. cera, /J.O.KUV beside Old 
High German mayo, see Job.ans.son B.B. 15. 306 *q. 


' the ship sank,' Old Norse kaf ' a dive, the deep ' ; (d) %uiv gww 
quasi ' walk,' and $*<7AMU tyrew quasi 'go about,' all from fiaivw : 
(<?) /5o?o9 'whistling' beside /5o?/J-gos (for the termination cf. 

D. Beside the copulative sin-, Greek a- or -, there appears a 
parallel form so-, Sk. sa-, Greek o-. We have it in Hesychius' 

o^fda-rtiyp o^yye? odpoo? o%v\ov (Schulze 495), Homer's oa-Tpo 1 } and 

II. 2. 765 o7/j<x; as oiereas (the latter, despite Schulze, a miswriting 
for *oweTeos, * e. *o-^"eVeas) ; and I think in o-cov?, Armenian 
a-tamn, each a Singular formed out of a Plural signifying ' the 
united teeth, the rows of teeth,' as perhaps aa-i'jp ua-rpov Arm. ast\ 
are Singulars formed out of Plurals signifying 'the whole body 
of stars,' the initial vowel in each word being a copulative prefix. 

E. In two important words the initial e- or t- seems reduplica- 
tive, standing for se- : 

eV0Xo9 ' brave, good ' =*a-e-oT\o'?, from the root of ore'XXw ' set 
in order,' the meaning thus being 'ready, settled, steadfast.' The 
Greeks found a difficulty in pronouncing the combination orX, which 
occurs in no old word (0-7X677/9 ' scraper ' appears first in Hippo- 
crates, offr\i*jl- ' curl ' in Attic) : they preferred either ( 1 ) to 
aspirate the T, Attic vavad\6ia ' convey by sea ' beside vav(no\ew t 
Horn. lfiaaO\rj ' whip ' with the same termination as e'^eVX^, Att. 
fidffd\rj<s ' leather ' from the same root as ^da-rt^ ' whip,' and so, I 
think, eV#Xo's ; or (2) to drop the s, 7X677/9 or the T, Sappho 
yiiao-Xj/s ' leather,' Doric and Lesbian eVXo? and Arcadian eo-XoV 
(which last must represent *eorXo's, not *eo-0Xo9, or it would have a 
smooth breathing) or both letters, Att. vau\ov ' fare ' beside 
Hesychius' vavaO\ov ; or (3) to change the X to p, 0-7/9677/9 ; or 
(4) to insert a vowel, oa-rd\i^ (as M. Schmidt reads oo~d\att; in 
Hesychius), (776X677/9. The same ore'XXw appears, I would suggest, 
in o00aX//,ds from *o7r-o-TaX-/A09 'arrangement for seeing,' the first 
element going with ofifia otrunra 6'Y^s : *o^-a\/<o'<? became ocfiOaX/uds 
as *ei^T09, the proper Participle of fyw, became e00o's : 

effTt'a ' hearth, altar ' (in Homer only in the compounds aj/e'<rno9 
e'060To9) =*o-e-<T7/a from a root sti- ' stone,' whence a-rlov 'pebble ' 
and I think vepl a-na ' lustration of the Ecclesia ' by carrying a 
victim round the altar, and, with Ablaut, Goth, stains ' stone ' 
and Old Slavonic stena ' stone wall.' On the parallel form ia-rir] 
see next section. 


(2) The greatest difficulties in Greek vocalism are connected with 
the occasional appearance of < where we should have expected e 
(G. Meyer 57-60, 29; Kretschmer K.Z. 31. 375 *?.), and of v where 
we should have expected o (G. Meyer 61, 62 ; Schulze p. 495 sq.). 

A. (o) Some of the instances quoted of i for e rest on doubtful 
or impossible etymologies : 

ai~ft'\i\fs, epithet of Tre-prj in Homer and Aeschylus (Suppl. 794), 
is of unknown meaning (Monro), and can have nothing to do 
with AeVas ' rock.' Hesychius' Xty' ^~pa a0' rj? vttap ff-rdgei may 

go with \eifiw. 

i'-/'/ 1 a- ef? and tYroj/' eV, quoted by Hesychius as Cretan, are 
too obscure to be deduced from a root sem- : M. Schmidt suspects 
both glosses. 

*'XXa<? 'rope,' t'XXo's 'squinting,' A-Xw or ei'XXw 'wrap up,' l'\\o/ 
' move to and fro ' can hardly have anything to do with Lat. volvo. 

i7rof cannot go with Lat. equus, Sk. dgvas : the aspiration is 
quite abnormal, and the Homeric form ought at least to be *IKKO<I, 
cf. 7re\eKKov from *-e\eK-Fov. The form t'/c/co? in the Etymologicum 
Magnum may be Lat. equm borrowed (with i from TTTTTOS), as 
aKKi-)/ffio<! in Athenaeus is Lat. aquipensis borrowed. 

A.-r<7o's 'ivy' cannot go with Lat. hedera from a root ghvedh.-, 
or we should have *xr<rds. 

X<*r/3/0/s 'sideways,' with a strange termination, goes with Lat. 
licinus ' with upturned horns ' and obllquus (i.e. *ob-l!c-vos) 
'crooked'; not with Xe'^/^o?, which stands for *\eg-p-io? and 
goes with Xo^o's (De Saussure, Mem. Soc. Ling. 7. 91, n.). 

X#<"o's is not directly from x^e's, i.e. ghjes, but from an 
Ablaut ghjz with sonant sibilant, Thurneysen K.Z. 30. 352 ; 
the termination is djo-, cognate with Lat. dies. 

(/3) In ryi/i/jy i^w i'dpis iff-rl)) the initial vowel is reduplicative, 
as in /A.--/S (cr/Seo?) i^Ov's (Arm. jukn) arra/uu (jreTo/uai') laOi (Zend 
zdl 'be thou') as opposed to e^^e? e^/vwica eoTrjica: in such cases 
it would be absurd to talk of a change from e to i. So in viaaofiai 
TI'KTW the i is reduplicative, as in ndi'jvt] clcw/j.1 -rnvaKeaOai beside 
tvrttvo9 ceCuiKo. TCTVKeaOai. Thus 

' hollow of the knee ' is reduplicated from the root of 

tgu} l =*i-<rc-jw from the ' reduced ' root of efo/iat, ecos, Lat. sedeo ; 
and so ify&ta (the < is short) =.*i-ac-p-vw : 

1 i'crro's, properly Participle of tfa ' set up,' is used as a subst., ' mast, loom.' 


i6pi<i' 07rau)v and edpiy' rofiias ( so M. Schmidt reads for ra\uiat] 
in Hesychius may both go with Opi'^ta ' cut off ' : 

i-ffrfi) is the Homeric form of eor/d, see above. The Homeric 
compounds of e<rria, and the evidence of Boeotian, Cretan, Hera- 
clean, Locrian, and Sicilian inscriptions (Meister 1. 46), prove that 
neither form ever had an initial F ' the Arcadian proper name 
Vistias can have nothing to do with torn/, and is as obscure as 
the name Kd/3ew<7os in the same inscription (^Meister 2. 103). 

vlaaofJLai=*vi-va-Jo/JLai, from the reduced root of i/e(o-)o^tat, voff-os '. 

T-I'KTW =*?i'-TK-Ttv, from the reduced root of re/ceiv, cf. /caovs 
' brother ' from *-K-n-n s. 

(7) We have i as Ablaut of je in Sk. vie- vidh- beside vyac- ' to 
extend' vyadh- 'to pierce,' and so I think in vTrep-ncrai'vovTo 'they 
sped on ' (Od. 23. 3) beside Sk. pra-ydk&li- ' to press on.' The 
relation of "m-epos ' jaundice ' (for the termination Ha vet, Mem. 
Soc. Ling. 4. 230, compares va-repa, ^aa-^p) to jekv-, Lak/MMr, 
is not quite clear : we should have expected *tWe/>os. 

(2) Homer's Kipvy/Ai rtrvrffu rf\vafuu (<r)*r/^fo / iiat beside Kepaaaa 
TreTaaaa ireXaaaa criceBaffa (e/ceJatrtra), Pindar's fpifuriffu l trlwH 

beside Kpe/^aaav 7re?oiffai, Attic opi'fvdouat beside ope^oiiat, OW6 

their to the analogy of reduplicated Verbs, they are formed 
after r^t^fvofuu pu-furwf as Homer's aKip-raw (root sker-, cf. 
aicaipw} is formed after W-KTW. So Pedersen Idg. Forschungen 
2. 293 says ' the i of aKi^vrjfii is due to the influence of i<mjfii 

TJfhffU 7Tl/U.7T\T]fll etC.' 

Homer's iri'crvpes beside Lesbian Treavpe^ must owe its i to the 
preceding numeral, vpla rpis T/MTOS. So (Baunack K.Z. 25. 225 sq., 
Brugmann Grundriss 2. 165 sq.} oK-a- in compounds owes its 
-- to eTTTa-, Heraclean OKTU> and Elean omw draw from eTna 
their breathing and labial respectively, efySoos borrows its consonants 
from eficofios. 

(e) There is no clear proof of any confusion in the Ionic- Attic 
dialect between e and i : the Old- Attic forms Aivcu-rai Ahiarai, 
AvXearai A.v\ia-rai only show different ways of resolving the 
diphthong ei before a vowel, Delian 0-7X177/9 is an Assimilation (see 
J. Schmidt K.Z. 32. 321 sq.} for 0-7X677/5, M<i'c/wi/ is a very late 
form for the earlier Mevdai'iav on the coins of Alende in Pallene. 

1 In the MSS. almost always written Kp-fi/j.vriiJ.1 (Kretschmer K.Z. 31. 375), see 
Aesch. Theb. 229, Eur. Here. Fur. 520. 

2 Homer's irtpvr)/j.i kept its through the influence of its cognates irepdu ' sell ' 
and irfpfjv, II. 24. 751 Trepvcurx', ovnv" (\f<rict, itfpi}v a\6s. 


But in the non-Ionic dialects the letters interchange so often that 
we can only explain the instances by supposing that in those 
dialects e was pronounced ' close,' half-way to an 'open ' i, and so 
might be represented by either letter. Before a consonant we have 
this variation in 

Arcadian lv beside lv, even in consecutive lines of the same 
inscription (Meister 2. 90) : cf. Hesychius i'^Kpof' e'ry/ce'0aX,os, 
Iff^epw' ef/ys (i.e, ev a-^epw] : 

Cyprian lv beside (Hesychius) evavov eV0es, l'(v]de beside e(v}0dSe 
(Meister 2. 210) : cf. Hesych. Tn\vdv' <paidv (= Att. ?re\\oV) : 

Cretan lv and eV in the same inscription (G. Meyer 58) : 

Locrian <y/i/os (Havet, Mem. Soc. Ling. 2. 168). 

So the Sicyonians themselves called their city 2e/cvwi> (Meister 
2. 89) : Hesychius has Xe/c/W and \iKpol ' antlers ' without designa- 
tion of dialect. Before a vowel (Solmsen K.Z. 32. 513 sq.) 
we find this variation in Cyprian, Lesbian, Boeotian, Thessalian 
(excluding Larisa), and Doric : 

Cyprian Oidv and dew, /u.1 and /*e (Meister 2. 211) : 

Lesbian ^pva-i'ia and ^pvaeia : 

Boeotian Oids and deois, doiciei but KaXe'oi/Tt : 

Thessalian Klovta and 

Laconian aias and 

Heraclean Ti/uoKpaTio? but /"e'reos : 

Cretan 6109 and 6>eo's, ajiiwv and afidwv, l&ftev and 

We must therefore conclude that, however it was written, e was 
always pronounced ' close ' in Aeolic, Doric, Locrian, and Cyprian ; 
and, at least when it stood before a consonant, in Arcadian also. 

B. () The derivation of the following words is unknown, 
and we cannot say that the v stands for o : 

irpvXees ' champions ' (Horn.), cf. Cyprian 7rpv\is ' war-dance ' 
(Hoffmann B.B. 15. 89). 

TrpvfiVT) ' stern,' 7rpvpvo<s 'hindmost': not from Trpd, which would 
give just the wrong sense. 

irpv-ravi<i ' president ' : Attic also Trpo-ravei'a TTpofaveina (Meister- 
hans, Grammatik der Attischen Inschriften, p. 19), Lesbian 
both 7rpinavi<i and Trpdravi^, the words being popularly connected 
with Trpd. 

7rv\r) 'gate': not from veXw 'move,' which gives too indefinite 
a meaning. 

TriyiaTos 'last': Sk. pu-nar 'back' (Bugge B.B. 14. 68) has 
little resemblance of meaning. 


'stercus' (Hippocrates), cf. Att. 

ls or er0iyn9 ' basket ' : Lat. sporta goes rather with a-rrap-rov 

vg ' new wine ' : Eng. dregs cannot be connected. 

'awl' (Herodotus) : Att. oirrj-nov may take its o- from OTTT'/ 

So 7\i50w 'carve' goes with Lat. glabo 'peel,' Ags. cleofan 
'split,' not with ^\a<pw 'scrape.' In u/W/os 'army,' vppat- 
'pell-mell' (Schulze 495), the first element is a dialectic form 
of avv , not a ' copulative ' o-. 

(ft) In the following words the v is formative, a stem-ending 
(Brugmann Grr. 2. 104, cf. 91, n.) : 

aiyiyws 'gathering' (a*(wv), ojifavfiis, Travrj'/vpis, Att. ayvpTT}9 

'beggar': from a stem 071;-, as ayopd a^elpw from a stem 070- 
(76-), all cognate with u^w ' drive.' 

alffvfwrjT^y ' umpire ' : stem alav-, cf. aia{F}a, see on aTtr^o?, 
sec. 4. 

a/uvdi? ' together ' (the accent, as opposed to that of dfioipySi?, 
is from owa)= stem ujiv- (on the breathing see sec. 3), as ajia 
from a stem afto-. 

Siairpvfftos ' passing through ' : stem vpv-, from the root of 
Trepaw, see Trpdfffftv, S6C. 4. 

SpvTnw ' I tear ' : stem $pv- (Sepw), as SpeTrw ' pluck ' from 
a stem fye-. 

eiraaffvrepot ' one after the other,' of obscure formation : if it 
went with the Homeric aaaorepta (by-form of aaaov) it would 

be *e7raffff07epoi. 

KpoicvSei\os 'lizard,' Hipponax 119: stem KPOKV-, as KpoKo8ei\os 
from KpoKo-, sec. 4. 

[idpwpo? Horn., fidprup Find, and Att. : stem fiap-rv-, root mr-, 
as in fipa-ftevs 'umpire ' (Kretschmer K.Z. 31. 392). 

vwvvfios Horn., vwwfivo-s Horn. Pind., fo-wwjio? ITT- oft- Horn. 

Pind. Att., 6V- TTO\V- Hes. Pind. Att., t- aw- ifsevS- 7raTpivvv[iio<} 

Att.: not 'compounds of oi/o/ia,' as Kretschmer K.Z. 31. 377 
makes them, or they would end in -fiwv : they are from a stem 
uivv- (for the ending cf. eVy/to? ijBv/u.os), as ovofia is from a stem 
ovo- (Bartholomae B.B. 17. 132). 1 

1 On ovofna and its cognates among which Lat. nomen must not be reckoned, 
it cannot be separated from co-gnomen and Old Slavonic zname ' sign ' see 
further J. Schmidt K.Z. 23. 267 sq. ; Thumb K.Z. 32. 130 sq. ; Bezzenberger 
B.B. 10. 72 ; R. Schmidt Idg. Forsch. 1. 77 ; Audouin Mem. Soc. Ling. 


' quail ' : stems op-rvy- and oprvic-, cf ., with other stems 
from the same root, Sk. vartakas vdrtika vartlras, all meaning 
' quail.' 

po$ew 'swallow' (Ionic): stem sru-, cf., with the same 'deter- 
minative ' bh, sro- in Att. poff^eta (Ablauts sre- Lithuanian sriebiu, 
sr- Lat. sorled], and, with a different determinative, sr- in Old 
Slavonic srukati. 

vTrd/Spvxa ' under water ' : stem mru-, cf . Lat. mare (Hirt Idg. 
Forsch. 1. 475). 

(7) In pvw 'growl' beside pdw po-^Oew, and pv/u/3o? 'bull- 
roarer' 1 beside pojufio?, we have different Ablauts : the pv- repre- 
sents vr- (Frohde B.B. 14. 107), the po- represents vro-. It 
seems that in one dialect of the Ursprache vr (vl) became m (In), 
in another the sonant took the same form as in other combina- 
tions : we have Sk. rue- ' shine' rudh- ' grow ' lubh- ' be lustful ' 
beside v areas ' light ' vardh- ' grow ' valbh- ' enjoy ' respectively, 
opvaaw 'dig' and (nasalised) pv~/xos 'snout' but Sk. vrh- 'to 
tear,' Av/cos Lat. lupus but Lith. wilkas, see on aavpunrfp sec. 5. 

(e) In the following cases the v like the u in Sk. dhur- ush- 
hur- beside dhvar- ' injure ' vas- ' shine ' hvar- ' be crooked ' 
respectively is Ablaut of vo or ve : 

7 yj/ '/> g" 1111 -) 2 beside gven-, Goth, qino, Old Irish ben, Old 
Slavonic zena, Sk. Jam's, and Elean fieveoi (jaia^ono -ry ^waiKi, 
Meister 2. 22). 

7ri-aKvviov ' skin over the eyes,' skim-, beside skven-to-, Old 
]S"orse skinn. 

7. 61 ; G. Meyer Gr. Gr. 2 77 and Albanesische Studien 3. 69 ; Brngmann Grr. 
1. 219 fin. and 2. 340 ; Schulze 201 sq. ; Persson 227. The forms in Celtic 
(Irish ainmm or oiiim, Welsh tine) and Old Slavonic (imf) have not yet heen 
satisfactorily explained. The root may appear in 6vo/ ' blame, disparage,' i.e. 
'name,' in our parliamentary sense, stigmatize. Arcadian K\f(ov6/, Laconian 
Trarpoy6/j.ov, seem to owe their third vowel o (for u) to a false connexion with 

1 i.e. (Andrew Lang, Custom and Myth, p. 39), a fish-shaped piece of wood 
making a hideous noise when whirled 'round by means of a piece of string. The 
Greeks themselves seem to have compared the shape of the bullroarer to that of 
the wryneck, 5firy|, with its long snake-like neck : ' the mad bird, the variegated 
wryneck of the four spokes, bound to an endless wheel,' which Aphrodite in 
Pindar (Pyth. 4. 381) brings to Jason to help him in gaining the love of Medea, 
can hardly have been a real wryneck, but a bullroarer spun round by means of a 
wheel. From the noise which this would make, not from the bird itself, which 
has not a loud cry, came the Homeric ivfa ' roar ' ; and, from the use of the bull- 
roarer in magical ceremonies, the meaning of lvy as 'charm, spell' (Pind. 
Xem. 4. 56) or 'yearning produced by a spell ' (Aesch. Pers. 989). 

2 A Velar after or before v becomes Palatal, not Labial (De Saussure, Mem. 
Soc. Ling. 6. 161 .5/7. ). 


Ovpa, dhur-, Lit. d&rys, Arm. durn, Sk. dur-, beside dhvor-, 
Zend dvara. 

KVK\O<S 'wheel,' ku'kvlos, beside kvekvlos Ags. hveohl, 1 
kvekvlos Ags. hveogel, Sk. cakrds, and kvokvl- Lat. Copies (an 
Oscan form, for *quocles, as poplna is the Oscan form of coquina] 
'ham of the knee' as being rounded (Bugge B.B. 14. 64, 65). 

KV\I 'cup,' kul-, beside kvel- weXif (Cratinus, Meineke 
2. 64), and kvol- Sk. kalagas. 

ff-vpat; ' spike at the butt-end of a spear,' (s)tur-, beside 
(s)tver-, see on cravpwTJjp, sec. 5. 

vcicp, ud- Sk. udan-, beside vod- Goth, vato, Old Slavonic voda. 

vTTi/os,, Old Slavonic siinu, Ann. khun, beside svepnos 
Ags. scefn, and svo'pnos Lith. sapnas (Lat. somnus and Sk. svdpnas 
may represent either svep- or svop-). 

iriavpe? Treffvpe? (see pp. 8 and 2), kvetnr-, Lith. keturl, 
beside kvetver- Lith. ketweri, Old Slavonic cetcerii, Irish cethir, 
and kvetvor- Dor. -re-opes. 

So apparently in 

fiv\i), mul-, beside mvel- Irish melim, Old Slavonic melja, 
Sk. mar-, and mvol- Goth, malan, Lith. tnalu, Arm. malem (Lat. 
mola mold may be either mvol- or mvl-, see below) : 

vvtl, nuk v t-, beside nvokt v - Lat. nox, Irish nocht, Goth, nahts, 
Lit. naktis, Old Slavonic nosti, Sk. ndktis : 

<J>v\\ov, bhul-jom, beside bhvcd-jom Lat. folium. 

I.e., though no extant language has initial mv, nv, or bhv, 
the Greek forms here show that such combinations existed in 
the TJrsprache. So (Brugmann Grr. 166. 170. 184) v is lost 
after initial bh in vTrep-fa'aXos, Lat fis, Old Slavonic be ' he was ' ; 
and after medial n in Att. ^ova-ra fei/o? (frdavw, and Old Slavonic 
tlnlka ' thin.' 

In five of these words we have a further Ablaut, the final liquid 
of the root becomes sonant : 

ffwrj : gvn.-, Dor. <yaf, Boeotian /3ai>a* 

Qvpd: dhvr-, Lat. foris, Old Slavonic dcW. 3 

1 So, I think, irerXos ' robe ' as being circular when spread out flat on the 
ground ; whence in Latin it was called cyclas. 

3 Also, I think, Ionic (not ' borrowed into Attic from Boeotian ' as J. Adam 
says of /Savoucros, Classical Review 7. 102) : fidvava-os 'mechanical' Herodotus 
2. 165 opposes &ava,v<rir) to rb fidxifiov = *&avB.-avT-jos (for similar shortenings 
in compounds see on 7XfA.i/s, sec. 4) ' quite womanish,' the second element being 
from ain6s in the sense of Homer's afrnas 'merely, simply.' 

3 But not Ocupos, which was the pivot of a gate, TU\OJ (II. 12. 459), not of a 


: kvl-, *K(i\i^, a form from which Lat. calix was borrowed. 
: kvetvr-, lon.-Att. -reaaape?, Hdt. -reffffepei (ep repre- 
senting r, sec. 3 C, as in epar/v beside Att. uparjv], Boeot. ire-napes, 
Osc&n petor. (So Aeolic avpice<s may be from turk-, while Att. 
ffdpKe? is from tvrk- : the root is unknown.) 

fivXij : mvj-, /nd\rj ' armpit,' cf. fiv\q in the sense of ' kneepan ' : 
both parts of the body were named from their shape, the armpit 
concave like the upper millstone, the kneepan convex like the 

Why all these forms of Ablaut were used it is hard to see : 
plainly it had nothing to do with accent, whether of pitch or of 
stress. My own theory on the subject, Etyma Latina p. xxx., has 
as yet escaped notice. 

(e) In owl; from onokhv- (p. 1) the second vowel is due to 
Dissimilation, the dislike to having the same sound in two con- 
secutive syllables : cf. with ' regressive ' Dissimilation, the dialectic 
Attic Aj//0u/3os (Kretschmer K.Z. 29. 412). In onokhv- the 
second vowel was a genuine 0, not one interchangeable with e : 
every other kind of o remains in Ionic-Attic in such a position, 
whether the preceding vowel be o, oro/Jos ororol (both onoma- 
topoeic), o/3oXo's o\o0tt>to9 opoftfta opo<fiy beside o/3eXo's eXe^alpofiat 
opefta epefoo respectively, oQofuu oXoo's beside oOfieu 6\eKiv, ovofut 
from a stem ono- (see p. 10), or w, Horn. 7re^7rwfto\ov Att. 

a0a-/3oXos 7piwftoi\oi>, Archil. Att. aTr-w/u-OTOS Att. eV- trvv- f Hom. 

Att. vTTwpdcfiio? : with the exception of Trevrwpv/a in an Attic 
inscription of 330 B.C. (Boeckh, Staatshaushaltung 3. 412), Stwpv^a 
etc. in Xenophon, Cynegeticus 2. 5, which owe their v to the 
analogy of eVuW/to<? etc., (see p. 10 fin.), since they stood to 
opd-jVLa (Ar. Fragm. 661 1 ) as eTrwwfia etc. seemed to stand to ovofia. 

But in the non-Ionic dialects an o in such a position was pro- 
nounced 'close,' inclining to an 'open' v, and might be written 
either o or v : whether the preceding vowel were o, owfia in 
Lesbian, Boeotian (as oviovfia), Thessalian, Phocian, and Doric 
(Kretschmer K.Z. 31. 377, Meister 1. 56), and o/.ivav (apparently 
for ofioae, Bezzenberger B.B. 5. 327) in Pamphylian, or w, 
Epidaurian W7ra-/3v0/a9 beside o/>o0a (Kretschmer K.Z. 378). 

(") Even apart from Dissimilation, in the non-Ionic dialects 
every un-stressed o seems to have been pronounced ' close ' and 
written indifferently o or v. Thus 

1 Find. Pyth.4. 228 6p6yviav, and Sappho 98 firropoyvwi, are mere conjectures. 


(a) in the article, which like our ' the ' was doubtless un- 
stressed, Pamphylian v but Arcadian o : 

(b) at the end of a word, Lesbian CLTTV and a/ro (the Grammarians 
give Zevpv as the Aeolic form, Sappho has ceupo). Arcadian inrv 
tea. 1 aXXi/ 2 but e\vfftrro, Cyprian Inrv ryevoiTu wplaeTu (never -TO), 
Pamphylian ef3ia\doTv eVnJXocu Ka-epepgoSv. So before a final 
consonant, Cyprian Kepdfiiv? Nom. Sing, (in other words -o?), 
Pamphylian (3iv\i)[iei>v9 and in the same inscription KCKpa^evos, 
'EtrrJ-ecitvs Nom. and Kovpaffi'wws Gen. : 

(c) in the first element of a compound, whether a monosyllabic 
Preposition, Cyprian vv-edrjKe and ov-eOrjice, both from n-, Att. 
av- ; or at the end of a disyllabic Preposition, Lesbian airv- and 
OTTO-, Larisaean OTTV-, Arcadian am- KUTV- ; or at the end of a 
stem, Rhodian 'A^aOu-fifipd-rov and Tifio-ppdcov, Pamphylian FOIKV- 
TroXts and Ne'yo-TToXets. 

Pitch-accent seems to have had nothing to do with this pro- 
nunciation of o, we have v in the oxytone syllable in FoiKviro\i<i 
and may suppose it in *A~/a0vfifipoTo<t. 3 On the other hand, except 
in the later Lesbian dialect, on which see below, this peculiar 
v seems to appear only in the unstressed syllable. 4 

In some sub-dialects o before t, whether stressed or not, seems 
to have been pronounced 'close': the instances (G. Meyer 116; 
J. Schmidt K.Z. 32. 394 sq.) are Ionic (at Abu Simbel) vh, 
Cretan vT and OTTVI, Euboean /teTwt/cos. 

(17) In the later Lesbian dialect that of Theocritus in his three 
' Aeolic ' poems, and of the poetess Balbilla, not that of Sappho 
and Alcaeus or of any early inscription even radical o seems to 
have been pronounced somewhat ' close.' The Greek grammarians, 
who say that the ' Aeolians,' i.e. the Lesbians, 'turn o into v,' 

give fiv*/is %vavov a-rvfia Tine ty<0aXo9 \nrioOa i'-pvis as the proper 
Aeolic forms, though Pittacus (in Bergk) has <ndfia, a Lesbian 
inscription ore, Alcaeus o/>i/ts ; so that we cannot be sure that 

1 For *Karb, which is to K<XT& as virb to uira- (Sappho). 

2 Homer's &\\vSis owes its v to &pv8is. 

3 So in Cyprian Kwinricrna (Hesychius) ' wine made from the refuse of pressed 
grapes,' if Meister 2. 220 is right in making this=*Keo^d-7ri(r/io, from KUVOS 
' resin' -\-wffm : but M. Schmidt suggests Kvw-irlfff/j.a ('juice pressed out'). 

4 Cyprian SoFfvcu and SvFdvoi seem to contain the same root in two forms, 
(I) dou- from 'dou- (Wiedemann, Litauisches Praeteritum, 41 sq.), cf. Old 
Slavonic davati, and (2) du-, cf. Lat. daim, with F developt before a vowel, as in 
Chalcidian rapvF6rns, Boeotian EvFdyopos, cf. Epirotic Eti&avSpos. Cf. respec- 
tively (1) do(u)- in Cyprian Scaicoi, and (2) du- in ivetvice (Meister 2. 220). 


the various readings Swei and vacwv in Sappho 40 and 4 are 
genuine. In a late inscription we have tyto/wv, in Balbilla 
(Meister 1. 53) vfiot, in Theocritus vfioiov vfidp-rr), in Hesychius 
ef v/nd\\wv' eg ofioitav, while Theocritus has <nvf.ia, Hesychius 
fivp/uvpwv. These facts seem to show (1) that the 'close' pro- 
nunciation of radical o was confined to the later Lesbian dialect, 
and (2) that even in this later dialect the radical o was pronounced 
only slightly ' close,' since in all instances but those just given 
it is spelt o and not v. 

(0} As the Romance languages, descended from the dialect of 
the Roman lower class, make every Latin u into o, so the dialects 
of Western Cyprus (Hoffmann B.B. 15. 49 sq.) seem to have 
made every v into o : the town-name So'Xot (in Cyprus) is in 
Plutarch 2vAo< (Master 2. 220), we have IdoviKrj in an inscription 
from Paphos (Deecke B.B. 6. 71), Hesychius quotes dopdva? 
/uo^oi aodva (beside Ovpa /uv^ds gv>]\i'j} as Paphian, and his 

flop/Hat; tTnoKuaev iVK<i(f)6~cve KOf.ifto's Kpd<na\\os Xo0j//s'/ep6v (beside fivp^itjg evicuaev e^[KaTa<pv-reve Ariyi/3o9 icpva-ra\\o<} 
Xi^jn's eTrvff/iiai eTrtff/u.vfepws') may belong to the same family. 
So also in some Boeotian dialect (G. Meyer 90), 'A/toWa? 6oairj<i ; 
but Boeotian Ei>0/?o<7oVi/ 20^00/w, Attic "O\o/u.7ro9. Laconian 
Kovoovpeu.'v, may be due to Assimilation (Kretschmer K.Z. 29. 
412), while Strabo's 'Opfii'va for 'Yp^rj (in Elis, II. 2. 616, 
see Meister 2. 31) seems to show the influence of 0/3/105 ' roadstead.' 

(3) Leskien's principle of ' Ausnahmslosigkeit,' that a phonetic 
law has no exceptions, is doubtless true within each dialect ; 
but in Greek there were as many dialects as there are in 
English, and every poet and each of his hearers or readers must 
have been familiar with several. So ovis and Ids, words common 
enough, were loanwords in Latin (Havet, Mem. Soc. Ling. 
6. 17 sq.\ the proper Roman forms would be *avis and *vos : 
the Romans said ' sedeo in solio,' though the I in the last 
word was Oscan. In ' the skipper met the shipper in a well- 
equipt skifE ' we have four different dialects : in the Windhill 
dialect of Yorkshire the forms misen, misel, miseln (myself) are 
used without distinction, and ' probably due to importation from 
neighbouring dialects' (Prof. Joseph Wright, Dialect of Wind- 
hill, p. 122). 


A. Whether an initial vowel should preserve its aspiration 
or not must have depended on dialect in Greek, just as it did 
in Latin and does in English: the lower orders at Rome, the 
linguistic progenitors of the Romance languages, must have 
dropt all their h's, and in England it is only the educated 
classes that keep theirs. 

ijcvs and ijco<i go together: the latter in II. 11. 318 yfiewv ?}o 
' satisfaction from us ' has an ironical sense, which I detect also 
in cognate words with short vowel, viz. () ao? II. 11. 88 
(where it seems to have a digamma), and uSy or aty (in Homer 
only in Ace.), 'satiety'; (b) acivd* or acivo? 'to repletion'; 
and (c) the post-Homeric acpds ' thick, large,' quasi ' sufficient.' 
In II. 5. 203 acqv, also written u&Bqv, may be a contraction 
of *a-aFa%i)v 'without satisfying oneself (Schulze p. 452 sq.} ; 
and so I would explain e'u> in Homer's ucyaeie a^/coVev, both 
also written aBS-, as for *a-aFacew, ' am dissatisfied, displeased ' : 

0/>o//oi> ' prelude,' beside Trpo-otpuov, must come from *oi'utov, 
as (frpovpos comes from 7rpd-\-*6pd? ' watcher ' (o/aaw), and <j)pot>cos 
from *<>pov&ou, i.e. -rrpo oSov, ' ahead on the way,' II. 4. 382. 
Homer's otfiij ' song ' will then be dialectic for *oi'/j.rj, perhaps 
meaning 'connection,' arrangement of words, from a root soi-, 
Sk. setus 'binding,' si-, 'to bind,' /-/tas 'band.' (Lat. saeculum 
then must be from some other root.) 

B. In Lesbian we have tW/> t'x^os for virep v^-os (G. Meyer 91), 
in Larisaean 'nrep (Aleister 1. 224), in Megarian ataipi>7tTa<; for 
alffvfii'ijTr'l'i (Kretschmer K.Z. 29. 412 sq.}, in Hippocrates both 
<npi(f)v os and <7T/>i;0i/o9 ' hard.' So I would explain c/rcu&ot 
' wanton ' as for *Kvvai8o<i, i.e. KVVO<S aldu> e^wv, ' with as much 
modesty as a dog,' the dialectic form being employed to disguise 
the meaning. 

C. The Lesbian representative of r was po (G. Meyer 27), 
e.g. ppo^ew^ 0poffeu>9 ffTpd-ra'/os, Homer's rjufiporov beside Att. 
ifl/jiap7ov : so /3/joT09 (cf. Sk. mrtds 'dead') must have been Aeolic, 
the true Ionic word being dvr)-r6<s, Dor. dva-rd<s. So n is represented 
by vo in Alcaeus' <^vd(f)a\\ov beside Att. /nW^aXAoc : l the root, 
as the varying initial proves, 2 was originally bi-aspirate, ghn-bh-, 

1 On va for n in Ionic- Attic see Osthoff , Morphologische Untersuchungen v. , 
preface: he quotes yvddos, vain (i.e. *vour-jia), and, for /xa from m, partvca 
beside ^era\\d(ii. 

2 Cf. Hesych. a.Ka66v aja66v, i.e. the original form was *a-xa06^, from 
' intensive ' a- (see sec. 1) + a root ghadh-, Ablaut ghadh- in Goth, gods ' good.' 


cf. ghne-bh- in Att. -/^c-ipaXXoi/, and with Metathesis (as in 
o^aXo? beside Ags. nafela] ghem-bh,-, Old Slavonic zeba 'I tear 
in pieces' (wool being carded for use). 

One dialect seems to have used e-, not a-, to represent the 
nasal sonant : Hesychius has eoaarjrrjp' uoaarj^p, and eypvirvei' 
d~(pv-ve?, the first vowel in each being the ' copulative ' prefix, 
originally sm-, sec. 1, c. (In a-^p-vTrvet, literally ' is chasing 
sleep,' 1 and in u-~/pa ' chace,' d-^pew 'seize,' gw-d-ypia 'reward 
for saving life,' the a- must be a prefix, the root gr-, ger-, as in 
Horn, gw-'fpei 'take alive,' and Hesych. e^prjvrcu- ?;/5/;i/Tt.) So 
r=e/3 in Homer's dpi- and Qepalr^ beside apt- and dupao^, Hesych. 
epe0pa and /u.ep~jte ' gobble ' beside fidpadpa and fidp'/o? ; and 
l = eX in Hesych. e\\etv pe\epov beside (3d\\eiv paXepov. Again, 
with the consonant coming first (as in fiarevu) <yi/a0o9 xpaSi-ij 
TrXaTu?), we have ni=6 in Hesych. fiea-ra^ for jidfnag, and 
r=/je in Lesbian /c/>6Tos for icpdro^ (G. Meyer 6), Hesych. ^6/t0oe 
for /ja/0o9 ' beak,' Boeotian -rpe-irecda? beside (in another inscription) 

T/JOTre'dca? (B.B. 17. 336, n.). 

D. Homer's oZjiif = Att. ooyuy, as v A5/t^T09 = Att. v Ar/tj/T09 
(Kretschmer K.Z. 29. 420) ; so Pindar's Keica/uei>o<} ( furnished 
with' =Att. KeKdfffievo?, perhaps meaning 'bound with,' and 
going with the post-Homeric *rjyco9 ' connexion by marriage ' 
quite a different word from the Homeric AT/^O? ' care, mourn- 
ing.' Homer's Kaiw^ai ' surpass,' Perf. KeKaa/aai, cannot be for 
*KaSi^vfiai or go with /ce/ca^/te'j'09, for Dental -\-v would remain 
unchanged, as in d\orrvci>r) TT^VW edvos (G. Meyer 280) : I 
would explain it as for ^Kaawfim (cf. eivvfii for *eavvfii}, from 
a root kns-, Sk. gans- ' to praise,' so that Kaiw^ai will mean 
' am praised ' for something, Od. 4. 725 TravToi'y? aperrjat, 
*:6:ffyiiJ'O9 eV Aavaoiffi. 

E. Herodotus (Ehys, P and Q, Groups, p. 16), uses *> for TT- 
in words derived from the Relative stem, e.g. tcote KOU KW? : 
thus his irpoKci 'forthwith' may stand for *-p6 ira, i.e. irpd-\-the 
Instrumental (Brugmann Grr. 2. 274) of kvo-. Thus 7r/>oV 
will be identical with Lat. prope ' near,' an Oscan form (see 
p. 12 init. on poples) for *proque from *proqua (Brugmann Grr. 1. 
973). The -Ka in av~iKa must be differently explained, 
apparently as kn, Ablaut of Kev ' then ' (see Persson Idg. Fors- 

1 Havet's (and Benfey's) connexion of &ypvirvos with eyelpw, Mem. Soc. Lin 
6. Ill, is rightly controverted by Breal, do. 17.!. 


chungen 2. 228): the first element is *avrt, Location of avs 
' ipse ' (Hesychius), while avro's is from the stem of avs-j-a 
'determinative' o (see on 0e'<r0aT05, p. 1). 

F. Before e or < a Velar ought in Greek to appear as a Dental 
(Bezzenberger B.B. 16. 254 sq., Bechtel Hauptprobleme p. 356 
sq.) : the rule-right forms of /3/os and fiia, beside Sk. jlv- and 
j'yd- respectively, would be *c/os and *c/a. The former appears in 
Siepo<; 'living' (Fick B.B. 16. 287), and I think in /e/iat ' hasten, 
am quick ' ; of which the Perfect Active would be *Beiw>ca (as 
that of a0n// is a0eWa, G. Meyer 559), whence was formed a 
Present BIWKW ' set in motion ' (G. Meyer 45). The form *c/a I 
detect in 

(a) Sia-Kovos ' servant,' quasi /3/a KOVWV, ' compelled to work ' : 
the second element, as in Hesychius' Kovelv' lirel^eaOai, Homer's 
e^Koveta ' hasten,' Att. O.KOVI-II ' without trouble ' (Schulze 353, n.), 
is from ken-, a parallel form of kven- in Trove?*, as kel- in /ce'XXw, 
Kefy?, Lat. celer, is a parallel form of kvel- in Tre'Xw ' move,' 
Lat. cold : 

(b) cia-veicr)* 'continuous,' quasi /3m evexOct?, 'brought on by 
force, not to be stopt ' : the second element being an unnasalised 
form (as in Lat. nactus] of the root of tVe7Ke2V and Lat. nanciscor. 

(4) It is often difficult to decide whether a word is a compound 
or a derivative, or of what elements an admittedly compound word 
is made up. 

afi/x'/ 9 "* Homer always connotes sound : it is used of the cries 
of pain, of a tumult, of the bleating of sheep, and of noisy eating 
and drinking. I would therefore explain it as ' dry-sounding,' 
from *o"o's, Adj. of u^tj ' dryness,' + iyx 9 'noise,' comparing 
II. 12. 160, KopvOei 'a/i0' avov atnew, of a ' harsh, grating sound ' 
(Monro), and Verg. Georg. 1. 357 aridus . . fragor 'a jarring noise.' 
Hesychius' aax[e'a] then will be the Doric form ; his afex/}s is a 
different word, cnro TOU u^tjv e-^eiv, as the Scholiast on II. 15. 25 
gives the derivation of a^xy? (which Apollonius Rhodius uses 
as = afaXe'o9 'dry,' Wackernagel K.Z. 33. 51). 

a?ffx 9 ' disgrace '= * To- <r^- o?, from ais- in Goth, aistan'to 
observe,' Old High German era ' honour,' a/(<r)-u>9 ' shame, respect ' 
(Bezzenberger B.B. 4. 313), ! and a7a-(F}a ' apportioning,' giving 

1 The 8 is probably terminational, not from 5i'5o>/ii. 


each his due share of honour. The second element of oTo-^o? 
is the ' reduced ' root of e^w, so that the whole word means 'having 
observation,' getting oneself observed. So 7ra<r^oi = *7ra0 
have suffering, I suffer,' from wa#os, -n-adelv. Hesychius' 
shows that the Greeks themselves considered Traa-^- the root. 
There is no proof that *Trdd-aKw, with an Inceptive ending, could 
give anything but TTUITKU) (which is the Elean form, G. Meyer 269) : 
fiia'ita is not for *fjn'"/-fficw, but shows the same stem, fua^- as in 
Homer's ^ia^-d^Keia 'meeting of glens,' i.e. fiiy-ff-<^-, the root of 
fii~j-vvfj.t increased by s, and with the same termination (Brugmann 
Grr. 2. 91) as ai-^rf beside ai-wrrcua 'look silent' (Kretschmer 
KZ. 31. 471). 

avw^a 'command' means 'lead up' (Lat. m-duco, see sec. 1, 
A), cause to do a thing : *w-/a is an unreduplicated Perfect, 
like o?ca, from the strong form (as in 07- 0^7-09) of the root 
of u^fia. So oifttorfl 'wailing,' from oi'/u,ot-\-*u}^^ : for the sense 
cf. KTVTTOV u"fetv ' make a noise,' yeXwra a/yet*/ ' raise a laugh.' l 
So I would explain -n-paaata as *7rpa-dy-Tj'w, ' make progress,' 
whence its Homeric use with a ' local ' Genitive, e.g. II. 24. 
264 iva Trptja-ffdufiev ooio ' advance on our way,' and so nywyos 
' business ' =*7rpa-a~(-o<} : the first element is *7rpa ' progress,' 
a Subst. formed like x/"/ an( l ofio-K\t'f (Brugmann Grr. 2. 896), 
and standing to Tropos Trepdw much as dpa- in ^pafia &pdu) stands 
to Lith. darau ' I do.' 

a/37Xeo5 ' difficult ' (the a- must be long, as it is in Att. 
0/370? ' idle,' from *a- pep-jo's) means ' impracticable, not to be 
done,' from o- negative-^/^e/^oj/. For the contraction cf. 
O.KWV from *a-F^Kwv. 

appmo-s ' unchangeable ' (Plato) = *u-Fpa-ro9, d- negative + 
*fparo9 Participle from vr- 'to turn,' whence also (a), with a 
termination -mo-, pdfio? 'worm' (see sec. 3, C), Lat. vermis (from 
*vormis), Goth, vaurms ' serpent' ; with -inn, Lat. vermina 'colic,' 
and () with a ' determinative ' t, Lat. vortd-, Sk. vart- ' to turn,' 
Elean fipcndva ' stirring-ladle ' (Hesychius), and oprvt; ' quail ' 
quasi ' dancing, turning round.' 

&i(j)depa ' dressed hide ' (Thuc. 2. 75 ceppei? ical SKfiOepas ' skins 

1 "With the literal sense of ' raise ' we find &y<a in d/cr^j ' raised land, coast, 
river-bank,' cf. Soph. Oed. Rex. 183 PV/J.IOV O.KTO.V ' the raised altar ' ; so in 
&X<)os 'weight, what one can lift,' cf. Soph. El. 119 &yeiv ou/teVi acaKia AUJTTJS 
uvTippotrov &xQ s i an( ^ the Attic use &yeiv ^.vav 'to weigh a pound, be able to 
lift it.' 


undressed or dressed ') means ' twice spoilt,' os--j-00e//w, diverted 
from its natural use as a covering for the beast, first torn off 
(Seppis from epw, with a termination -/us as in uicpis OK/H?, Brug- 
mann Grr. 2. 98) and then tanned. Hesychius has a dialectic 
form Zt^dpa, which goes with his tyei'pef (ftdee'pei (G. Meyer 
209 fin.). 

ffyxe\v9 ' eel ' = *e7xv-xe\u9, ' snake with the mouth of a 
tortoise,' %\V9: * v< y^us exactly=Lat. anguis, since eng- becomes 
in Latin any-, franco is for *freng5, cf. Goth, brikan. On similar 
shortenings in compounds, where the two middle syllables have 
the same consonant or two similar vowels concur, see G. Meyer 
302 and Schulze 427 : Aeschylus Eum. 52 has fiZeXvK-rpoTro? for 
*pBe\vK7o-7poiro<!, and so I would explain (a) 7r\a/ti/a<o? ' suppliant 

not yet purified' as for *7ra\a/to -fiv atos (as na\a/<;yc)/s = *naXa / o- 

/<?}e/;9, Fick K.Z. 22. 99), ' with a deed of violence, ira\afirj, in 
his memory,' fivrjfir] ; (1} oTo-/ia/705 ' loquacious ' as for *<nofia- 
nap<*(o<i ' mad of mouth'; and (0) virep,vrjfivice, II. 22. 491, of an 
orphan boy, as for *i//r-e-/'//->}/<yfc-e ' is bowed down in mind ' 
(quasi *vTrij^.vice ftcfunffuvo*) : though for rj^tvia ' bow down ' I 
cannot suggest any etymology. 

e!5a0o<f 'foundation' is marked by Brugmann, Grr. 2. p. 204, 
as the only Neuter word with the termination -bho- : I therefore 
take it as a compound of eos+a0y aTnofiai, l touching the seat' 
or bottom, a Dissimilation of breathing for *ea0os. 

eroi/j.09 'ready' may mean 'striving after the way,' 0710?; 
the first element being (as Prellwitz Etym. "Worterbuch der 
Griech. Sprache suggests) jet-, Sk. yat- ' seek to reach ' (Middle). 
In the Active Sk. yat- means ' to marshal, put in order ' ; and 
with this we may connect eVeo'?, ervfios, and (with the root in 
its long form reduplicated) erlj-rvfios ' true,' quasi ' regular.' 
On the difference of breathing see sec. 3, A. 

evpv<t must be a compound, or we should have *e//jws, 1 as we 
have El\ei'9via beside 'E \evOw : it seems a contraction of *ev-vpi's 
'full wide' (Sk. urus). So ev0v? 'straight' may be from ev--\-Ovw, 
1 rushing well,' going in a straight line. The first element appears 
in three different forms: (1) esu-, ev-, eV>s 'brave,' with metrical 
lengthening (Schulze 33 sq.} rjv? ; (2) 8U-, the reduced form, 
Sk. su- ' well,' Greek v- in vfipi? beside ftpiapds, vyiTj? beside 

1 The only non-compound word in Greek with v in both syllables is y\vici>s, 
apparently a by form of the *y\vic6s which appears in Hesychius' y\vKij- &o-rdrt) 


Lith. gyjti ' I get well (De Saussure Mem Soc. Ling. 7. 89, 
Zubaty K.Z. 31. 52 ?.); (3) su-, Sk. su- 'well,' and I think 
Greek *v- in *v0v<!, whence by Dissimilation (Osthoff Morph. 
Unters. 4. 190 sq., though his explanation is very different) rOvs, 
the second element being 0vw. As Zubaty points out, the 
parallelism of 

Sk. sit- ' well ' and dusk- ' ill,' 
Zend hu- and dusk-, 
Arm. h- and t-, 
Irish. SM- and du-, 

makes it difficult to separate ev-, as the correlative of va-, 
from Sk. su-, and put it with either Sk. at/us 'alive' or dvas 
'favour.' With ev I would put () eine 'when' or 'as,' in 
the latter meaning also ijine, with metrical lengthening : the -re 
being superflous, as in av-re, os re, added on the analogy of 
clauses in which it really meant ' and.' Thus II. 23. 62-65 
ev-re rov VTTVO<S UpMfnrre . . . i/\0e 6' eVt Y^X')' literally 'well 
was sleep seizing him : the ghost appeared,' came to mean ' when 
sleep was seizing him, the ghost appeared;' II. 3. 1C 13 evr' 
opco? Kopi>(ffffi \oVo9 Kare^evei' 6fii'x\i]i' . . . ws upa rwv VTTO 
Troffffi Kovt'ffa\ov tL-pw-i aeXX?}? means ' well does the south wind 
bring fog: so rose the dust,' i.e. 'as the south wind brings fog, so 
rose the dust ' ; and II. 4. 277\dv7epov, ijvre Triaaa, (fiai'veTai 
' it looks blacker, quite pitchy ' : ( b} evxopai ' boast, vow, pray,' 
literally ' use only lona verba ' about myself or the gods, the 
same termination appearing in vyx ff ^ 7 )X w o'^vu-^ta rpvx^ ty>'ix w > 
a-ov axi), cicax'i (this from the same root as ci'jta ' I will find,' Zend 
da- ' to know '). 

tX0oo7reu> 'quarrel,' II. 1. 518, means 'organise hostilities,' 
from the root of e*x0o9 e'xfyo's + dekv- Old High German gi-zeltm 
'to arrange,' with which Brugmann Grr. 1. p. 332 puts Sei-irvov 
'dinner,' quasi *cc-n-v-jov. Eng. hatred, literally 'arrangement 
(Ags. raed) of hate,' is a somewhat similar compound. 

Kl>oicdcei\o9 ' lizard,' an Ionic word (Hdt. 2. 69), = ' yellow 
coward,' AT/JO /cos ' saffron ' + cetXoV, from the colour and shyness of 
the animal. The application of the word to the crocodile must 
have been a Litotes, or joke. 

Xa^ti/09, properly used of an ox (Ar. Pax. 925), = ' with a fine 
hide,' p~ii>d<i. The first element is the intensive prefix Xo-, as in 
\aicaraTrihftaVj XaKardpaTos, Af7yx o>J > representing *Xe- i.e. *Xocre-, 


as the similar prefix \at- in Xa/wa/j'yo? Aar7ro/a9 represents *Xa<r- : * 
both are from a root las-, Sk. fos- 'to be lively,' reduplicated in 

\i\atofnai ' I desire,' i.e. *\i-\ua-jofJLai. 2 

[j&votvaw 'desire eagerly '=*fiei>o-Fotviiw (for the shortening 
of a compound see p. 20 on eiyxeXv?) 'am drunk -with desire,' 
[iei>ei otvwfiai : cf . the Attic use of peOveiv ' to be drunk with 
passion.' In II. 12. 59 fievoiveov is wrong both in form (for 
fievoivaov] and meaning ('were anxious'): Goebel, Homerische 
Blatter, p. 15 sq. proposes to read fievoiev av. 

vrjfycneos, the Homeric epithet of ^nwv and Kpi'fiefivov, may 
mean ' such as never was,' ofo? OUTTW e^/evero (as I think the 
post-Homeric aTrXero? 'immense' meant ofo? OUTTW eVXeTo), vrj- 
negative + a Participial form from ^i~(i>ouat. So Lat. ing ens 
' huge ' means ' quod nondum genitum est.' 

Trai^viij 'sport' (Hdt.) and Tral^viov 'toy' (Att.) are from 
an Adj. *7ra<-<yi/o's for *wu$-<yo'8, formed after veo-yvos, the 
second element going with "fi-/vo^ai. But the forms Trat^/fioavvr) 
jrat^oufiai TreiraiKa owe their guttural to a mistaken explanation 
of irai^w (i.e. *irai-j<a) as for *Trai^-jw, since the -fw in most 
Yerbs arose from -tfto : Curtius, Yerbum 1. 317, gives thirty 
instances of -w from -pjju>, as against nineteen of -guy from -ju>. 

TTwfiaXa 'not at all' (Att.) is a negative which was originally 
an interrogative : TTU> /j.a\a ' how, very much how ' ? The two 
words were pronounced and accented as one, to show that the 
fia\a qualified the preceding word and not anything that might 
follow. So TTW in Aesch. Again. 1507 is a negation under 
the guise of a question : Sidgwick rightly translates it ' nay.' 
In meaning it differs from z-w? ' how ? ' no more than ovrw 
differs from ovrws : in each case euphony alone determined which 
form should be employed. So OUTTW and ovwios, fiijTrw and //jyTrojy, 
are used interchangeably : in 

II. 2. 419 ovS" upa TTUJ ol eTretcpataive Kpovc'ivv (see Fasi), 

II. 3. 306 OV7TW T\r/ffO/J,', 

II. 14. 143 ffoi &' ovTTta fia\a fd*fj(V 0eol fj.u.Knp<s Koreovaiv, 
Od. 2. 118 icep^ea $' ot' OIJTTW TIV > uKovofiev ovce 7ra\.atu>v, 

1 For the difference in the final vowel cf. dpx '-' 7roA -' 5 apxi-TtKrtav (G. Meyer 

* AfXiTjjUfVor ' eager ' is not from AiXdbpoi but from *\id.o/.iat ' I am much set 
on a thins:,' cognate with \iS.v \iijv ' very much ' ; which itself seems to stand 
for *\iF-av, t ' smoothly, easily,' beside \e?f)os 'smooth,' with the same termina- 
tion as irA-oV n-A.-^f 'except,' literally 'turning from' (e.g. ir\^t> auToD='away 
from him '), from the root of WAo> ' move.' 


Soph. Oed. Rex 105 ov <yap eiaeiSdv <ye TTW, 

and again \ 

II. 4. 234 fi7j7T(v TI fieOt'ere OovpiSos dX/dJ?, 

Od. 9. 102 ftijTTuj T<9 \WTOIO (fta^fwv voatoio \d0>jTat, 

Eur. Hec. 1278 fiyjTrut fiavelij TvvSapis -roffdvSe Trats, 

we might just as well have had ovtrw-} (fiijirta^}. In many 
passages OVTTW (fiTjiria) may conveniently be translated ' not yet ' : 
but in each it is the Verb that gives the connotation of time, 
the particle denotes only manner (' not at all '). 

<ro\oiK09 'foreign' (=/3dp/3apo<i, Herodian) must be a comic 
formation from 0-0X0? 'ball of iron'+the termination of a-romo* 
cVot/cos fie-roiKo^ avvoncos, quasi ' lumpish dweller,' heavy citizen. 

<r0opws 'violently' (Od. 12. 124: <r0o'o/aa and ff<fio8p6s are 
post-Homeric) = ' acting for oneself,' from the roots of <r0o's 'their, 
his ' and Spdtv, as in o\i^ocpave'wv ' doing little, feeble.' So I 
would deduce afalavvv ' eagerly ' (II., in the phrases eVero 
afa&avov, (KpeSavov e v 067re, 'he followed on his own way') from 
the stem of a<pe-repo<s-\-\h.Q termination -Bavo-, fiyicedavos, a by -form 
of -vo- in a\<nracv6<3 ffoeSvd? fuucefvot o\o0ui'os TreXiSvos "fy-ecvos, 

as the termination -7ai>d-, eTi-tjeravd'i, is a by-form of -rvd- in 

vTrddpa, in the Homeric phrase vTrdSpa Ictav ' looking fiercely at 
him,' can have nothing to do with depKo/u,ai, which would be giving 
the same idea twice over, and in which case the word ought to be 
vTroBpd^, as the Alexandrians rightly had it. I therefore (Etyma 
Latina s.v. odium] take v-dcpa as Instrumental (for the accent cf. 
o-0o'<5/>o from <70ofy>o's) of an Adj. *vir-o8-pds ' with covert hatred,' 
from the root of ocvofiai (Schulze 341) 'am angry," Lat. odium, 
Arm. ateam ' I hate,' Old Norse otul ' fierce ' (e.g. 6'tul augu 
'fierce eyes'). With odium goes atrox 'fierce' (Lat. Consonant 
Laws 22, see Thurneysen K.Z. 32. 562) ; so that, if I may coin 
the Latin word, w7ro'8/3a = *subatrociter. 

wxpo? (apparently Neuter) ' paleness ' is in Homer the colour 
of fear, II. 3. 35 w^po? re fiiv et\e Trapeid's, Od. 11. 529 
wxpyvavra XP a > ^ a coward : it may mean ' egg-colour,' as 
yellow as the yolk of an egg, TO w^pov TOV wou (Aristotle). 
The first element is the root of w(f}dv (a post-Homeric word), 
Lat. ovum, Old Slavonic aje (which last proves the root to be 
6-, not 6v-) ; the second is a by-form of xpw? ' colour.' From 
oi^/jos was later (first in Hippocrates) formed an Adj. 
' pale, yellow.' 


(5) Some other words may best be given in alphabetical order, 
'eagle' (=*alf : eT-o9, as Pergaean a/'/^erds shows), Aratus' 
ds, 1 may mean ' mighty one,' going with afyros, epithet of 
Hephaistos in II. 18. 410, aia 'land' (the 'mighty' earth), 
and aiavJ/9 'everlasting.' The two last words Johansson, B.B. 
18. 4, puts with ate/, altav, and Sk. uyus 'living': the common 
idea then will be ' full of life, strong.' 

atVds ' terrible ' may originally have meant ' bitter, cruel,' *ufi-jd^ 
(cf. fiaivw from *fidfi-ju>, Goth, giman], m- Ablaut to om-, Lat. 
amarus 'bitter' (on the first vowel see Latin Vocalism 5), uyids 
' raw, cruel.' 

U.KWV 'javelin' may go with UKV\O<S 'acorn ' and mean ' made of 
oak.' So Schrader K.Z. 30. 461 connects ai^avey 'spear' with 
Eng. oak. 

aWy>oTJyTa ' manhood,' occurring three times in the Iliad, is of 
course unmetrical: it may be corrected in several ways. (1) The 
most impossible of all is Clemm's, who reads *c/>oT/yTa from *vfpo- 
T>/Ta (a-vfipd-i) i this in two places gives an un-Homeric caesura 
Kara T6T/>Toj/ ipoyjaiov (Monro, Homeric Grammar, 2 367. 2), 
\nrovaa *fy>OT?yra icat T)ft>]i' ) and in the third, II. 24. 6 TroOetav 

ijru 7 Kai yttej/09, will not even scan. (2) The idea that 
jra could be a ' reduction ' of *avf>o-rij?a lacks support : 
is epithet of vv% in II. 14. 78, and ap/Spo-os in Od. 11. 330, 
but the sense is different, d^pdrrj (like ufu^iftpu-rj, G. Meyer, 179) 
is ' neu componiert ' from /S/JOTOS, to mean ' void of men, unpeopled,' 

(*ra#* r)v ftporol firf (j)on)aiv Schol., eV ?J fiporo's ov Trpoeiaiv 

Eustathius), and so in Aesch. Prom. 2 ufipo^v eh eprjpiav (as 
Dindorf rightly reads for uftarov : Hesychius has dfiporov' cnrdvOpu}- 
Troi/). 2 (3) I would therefore read *apor>/Ta, as a parallel form 
to cipeTt)i>, the up- in each case representing nr- (beside a-vyp). In 
II. 2. 651 'EwaXiia av&pei(j>ovTr) we may read *upe<J)dvrtj, with the 
same stem as dpe-rrj. 

dwaKaio? ' sturgeon' was a Scythian word, Hclt. 4. 53 : if this 
means Slavonic I would connect the word with ant- in av-ri 
' against,' Lith. ant ' up to,' and akv- in Lat. aqua (the TJr- 
Slavisch form would be *aka), and explain it as ' going up 

1 For the varying quantity of the second vowel cf. dp-yen 
(Schulze 473 n.). 

2 From afrp6-rn Goebel, Homerische Blatter, p. 1 sq., derives aftpordfa in 
II. 10. 65 /u^ irtas d)3poTa|oju6)/ a.\\r]\ouv, ' walk by nijrht to each other's hurt ' : 
Hesychius has aftporfijai ' to meet by night.' 


stream,' since according to Pliny (Hist. Nat. 9. 60) the sturgeon 
swims with its scales turned forward. Lat. aquipensis may mean 
the same thing, though I cannot explain the second element. 
Another Scythian word, ua-^y ' hazelsap,' Hdt. 4. 23, may go 
with Polish oskola ' birchsap,' from a stem askhv-. 

av-rpov 'cave' may mean ' cul de sac,' and go with Sk. dntas 
' end,' Goth, andeis. 

av\at; 'furrow,' av\i'j ' court ' (within a high fence, Homer's 
/30e//9 avXi/9), awXds ' flute,' av\u)v ' hollow way ' or ' windpipe,' 
may all go together, the common idea being ' hollow.' The stem 
then is au-l-n- from ave-l-n-, p. 1, whence also ve-l-n- in Lat. 
vail is (Persson 230 : i.e. *velnos, with ' pretonic ' a), Sk. vdnl 
' reed ' and vands ' arrow made of reed.' : 

/fye'009 ' babe ' stands to Ppaxvs ' small ' much as eXafoo?, see 
note p. 11, stands to e\aj(V9 : the root of /3/>e'0o9 must be mreghv-, 
as that of fipax.v<i is mrghv-, Goth, ga-maurgjan ' to shorten ' 
(Johansson K.Z. 30. 442 #) I detect a third form, mfghv-, in 
/uufxfivo?, II. 24. 316, 'the little one,' according to Pliny (Hist. 
Nat. 10. 7) the smallest but one (the /leXaj/aeros) of the six kinds 
of eagle. 

6\ea/j 'bait,' Laconian /3X/y/> (Alcman 130), must mean 'dropt' 
into the water, from /3XXw (not, as Meister 2. 204 says, from a 
root gvel- meaning 'to split, tear'). In Od. 12. 252 the gram- 
marian Callistratus read l^Ovai TO?S oX/yoto-t oXoi> KOTO, SeiXcna 
(' bait') ySaXXwi/, for efcara (Schulze 102). 

^e<77roT^9 and Sk.jaspatis ' master of the house ' owe their t to a 
popular connexion with the words for ' lord,' iroai^ (' husband ') 
and pat is : the proper form, as Old Slavonic gospodl 'lord' shows, 
was gvgspod-, the -- perhaps appearing in eo-7roi>. The further 
derivation is obscure : the word indeed may be un-Aryan. 

ce-^ofjLiu in Homer and Att., beside Seicofiai in Sappho Pindar 
Hdt. and -COK- in compounds in all dialects, owes its % to e^w, 
a word of cognate meaning : i.e., to use Tick's convenient 
expression, ce^o^ai ' rhymes ' with e-^ofiai (iliddle). With 

fuK- may go Co^^o? (i.e., *^o/c--ff- / 09, as TrXo^/io? is for *TT\OK-(T- 

/to'?, De Saussure, Mem. Soc. Ling. 7. 91 2 ) 'aslant,' a metaphor 
from a beast turning to 'receive' the hunter, II. 12. 147 (of 

1 These must be quite different -words from Sk. vanl ' music, tone ' and vands 
' music, hundred- stringed harp,' with which Johansson, Idg. Forsch. 2. 55 n., 
puts av\6$ 

2 Cf. pa>xM<k ' cleft,' II. 23. 420, for *^u}y-a--fj.6s, from frjiyvvnt. 


boars at bay) avSpwv rffie KVVWV de-^a-rai Ko\o<rvprov iovra. 
aiaaovre irepl trtpicriv u^finnov v\rjv. 

dtjpts ' contest ' may mean ' spear-work,' from *i)pFi<i, derv- 
Ablaut to dorv-, doru, Sk. dam ' piece of wood ' : cf., with 
short vowel, derv- Lith. derwd ' pinewood,' dorv- bovpma, doru 
dopv ( spear.' This *dijpFi<i then became *O}/>/K?, e///j<9, though 
by ordinary Greek laws it should have become *epFi<i t *eDpi<s- 
much as *firjvffo<i became ^jujvvoi (Lesbian /^yi/vos), fiyvos, though 
by ordinary Greek laws it should have become *fivao<$, */teio-os. 
In other words, the law that rv became />/> was earlier in 
operation than the law that erv- became cpF, as the law that 
ns became w was earlier in operation (Brugmann GIT. 1. 611) 
than the law that ens became ei/y. 

eiKrj ' at random ' seems a Litotes for eiKo-rivs ' as we should 
have expected, simply, naturally,' Soph. Oed. Rex 979 eiKtj 
Kpariffrov %i}v, oTTWi Zvvano -its. It is then Instrumental of an 
Adj. *t:o, seen in eiKo-fto\eif ' to aim at random,' cognate with 
eoiice ' it seems.' 

eiireiv ' to say ' may mean ' to clear up,' veikv-, cf. Sk. vic- 
' to sift, separate, examine.' 

7ri-ffrafiai 'know' seems formed from the Adj. lirnnijfiuiv 
(Od. 16. 374) 'knowing,' literally 'setting oneself to a thing.' 
The Subst. eTrnntjfii} ' knowledge ' appears first in Hippocrates. 

ep[LT)vev<i ' interpreter ' must be formed from 'Epfiijv (Ace. of 
'E/j/t/ys), taken as a stem: Hermes was the god of speech, \dyto9 
(Lucian). So Ziyi/, Ace. of Zeus (II. 8. 206), being taken as a 
stem produced in the Tragedians the forms Zijva Z/;i/o's Zijvi. 

yrpov 'abdomen' ('wind' in our pugilistic sense) stands to 
Old Slavonic vetrft ' wind ' as Lat. venter (see Etyma Latina) 
stands to ventus : rppov is from (a)ve-, aypi, with a termination 
-trom, as venter, stem ventri-, is from vent- (a Participial form 
from the same root, Brugmann Grr. 1. 612), with a termination -ri-. 

Ka\\aia ' a cock's wattles' may mean ' beauties,' **raA,Xaos Adj. 
from *KO\\TI by-form of *:aX\o9 (as euxv ^ et 'x os )- 

*cXoto? /cXtt-o's ' wooden collar,' i.e. *K\u}f-i-rJv, shows the same 
stem klv- as Lat. cldva ' wooden club.' 

Kw\6w ' hinder ' is a Dissimilation for *KV\VU> (as KWKVW ' howl ' 
for *KVKUW, Lith. kukiu) : with a short vowel the root appears in 
Kv\\6u) ' cripple,' and Sk. kit n is ' crippled in the arm ' (Fortunatov 
B.B. 6. 216). 

Xao? ought in Ionic to be X?/o's, as it is in Hipponax, and perhaps 


once was in Homer (Monro, Horn. Gramm. 2 p. 390). From X^o? I 
would deduce (a) \ijtov ' crop,' the produce of ' common ' land, and 
(i) AY/S \tjir) Xeia ' booty,' public property before it was divided 
among the combatants, cf. Aei'a? aa<rra Soph. Aj. 54. 

/ncndw 'linger' (II.) and parr) 'folly' (Tragg.) are from nm-tos 
Part of fievu, with the same transition of meaning as appears in 
Eng. dwell and dull. 

pei'wv ' less,' for *fii'j-jwv (G. Meyer 391), goes with Sk. ma- 'to 
measure,' and so means ' more measured,' perpiwrepos, not so 

/Liri-vvw, Dor. /u.a-vvw 'make known,' is from inn- Ablaut of mn- 
in Lat. metis, Sk. matis ' thought.' For the transfer of meaning 
from ' think ' to ' declare ' see on aei'cw, p. 4. 

/to'/oo-To'/co?, epithet of Eileithyia in Homer, of Artemis in 
Theocritus, cannot mean, as Brugmann Grr. 1. 204 makes it, 
'causing pangs,' from ^efyoi/s Ace. Plur. of poyo? : -riic-na is not 
used metaphorically in Homer, and such a use would be peculiarly 
inappropriate in connexion with the occasion. Liddell and Scott 
rightly translate it 'helping women in hard childbirth,' protectress 
TWV [to'yis veKovauiv : the first element is an Adverb */to'<y-os (with 
the same termination as irdp-o^) from a stem ^107-, whence with 
Locative ending, and the same 9 as in &/10/-9, we get ^0749 ' with 
difficulty' (accented like the Subst.), and, with a determinative 
-o- (see on 0e'0-0aT09, p. 2), ^0709 ' labour.' 

[Lvwfy- 'gadfly, goad' means ' flylike,' stinging as a fly does: 
from fiva, the Attic form of /nvia (Theophrastus, Hist. Plant. 5. 77, 
where Liddell and Scott wrongly make it the name of a plant), 
-t- wTra. Prellwitz derives it from fivia in the sense of ' buzzing ' ; 
but this will not suit the meaning ' goad.' 

j/eaX/y? ' fresh ' is not a compound of aXifftcojuai, but a by-form of 

*f6aXo? (as <5ay<-iX'*ys is of OatyiXos, aXovp7/y? of aXoiy^/ds), from a 

Subst. *vea ' youth,' whence also veavi^, i/eat/tas. For the termi- 
nation cf. a7raT^Xo9, ffi<yij\O9. 

vouao<s=*v6a-Fo<i (Kretschmer K.Z. 31. 471), which I would 
connect with ve(<r}ofiai 'come': thus Od. 9. 411 vova-os A<09 
means 'the visitation of Zeus,' and Soph. Ant. 421 Oeia 1/00-09, of a 
whirlwind, ' the visitation of heaven.' 

7ra0Xae<j/ ' to foam ' must be from an Adj. *7ra0Xo9 (or *wa0Xa9, 
cf. ffvpuna beside ^vfivoi^, from the same root as 7r6 / /u0<f roytt0oXvf 
' bubble.' It cannot be, as Prellwitz makes it, a Reduplication 
from e(fi\aov ' they burst,' which would give *7rat(>\d%ei.v (Brug- 


mann Grr. 2. p. 1084) : Tranfailvw ' shine ' must be formed on an 
Adj. *7Tfl/t-0a///9 ' all shining,' whence also Trapfavowv, while 
Tra.Tna.ivu> ' look round ' must be from a root kvnkvth- (or whatever 
the last letter may be), kvenkvth-, whence, without the nasal, Sk. 
caksh- 'to see ' (Fick B.B. 18. 134). So rerpe^acvw must be from 
au Adj. *Te-T-pe[iavd?, not straight from Tpeftw, or we should have 
*-n-rpe[iaivw, like -n-raivw ; and -re-rpalvw 'pierce' from an Adj. 
*Te-T/)ai/o9, or we should have mpaivw (the form used by Theo- 

TT.TTvvfjievo<i ' wise,' voov TreTTvvaOai ' to have understanding ' 
Od. 10. 495 (see Goebel, Homerische Blatter, p. 24), go with 
ironrvvw ' am busy,' the common idea being that of strength : 
the root is kvneu-, Old High German pi-hniutan 'to glorify,' 
TTVV% ' meeting-place of the Ecclesia,' quasi ' enclosure, strong- 
hold.' Thus Trveiv ' blow ' must originally have been used of 
the wind blowing strong, and then transferred to the breathing 
of human beings. 

TreTnrjws' Si' uaOeveiav KOI ceiXiav 7re7nwKW9, says HesychlUS : 

in Homer the word always means 'fallen,' e.g. Od. 14. 474 
VTTO Tev^effi TreTTTij&Tes Ket'fieda ' we lay on the ground, with 
our shields over us,' and II. 2. 312 (of young sparrows) 7reT'\o9 
v7ro7re7rrt]WT'i ' at the bottom of the nest, under the leaves.' 
It has no more to do with Tni/ffau) 'crouch' than has K(na.Tn!fn]v 
in II. 8. 136, of horses falling under the car: horses do not 
crouch down when they are frightened, but struggle to get away. 

jn/77/ ' fountain ' must go with jn^ds ' big,' the* Homeric 
epithet of horses and waves, and Sk. pajrds ' strong ' : it means 
a place where the water is strong enough to force its way out. 

7ri0rjKo<s iriOwv ' ape ' must go with 7rt'0o9 ' jar,' and means 
'rotund, pot-bellied.' 

Trpoxw means ' wholly ' in Homer, as it is allowed to mean 
in Apollonius Ehodius: it has nothing to do with 701/1; (which 
would not account for the x), 1 in II. 9. 570 Trpo-^w KuOe^o^iev^ 
means ' sitting right down,' not ' sitting on her knees,' which 
would be an impossible feat. I would deduce the word from 
*7rpo^, formed from irpo as Trepi^ is formed from Trepi, and, I 
think aTraf from *aV Instrumental of *inr6<s, i.e. smkvos, from 
sm- ' together' (see 1, C) : for the sense cf. Lat. pronus ' utterly,' 

1 Sk. prnjntis, quoted by Fick, Worterbuch 4 1. 432, is unauthenticated, and 
at best only means ' bandy-legged.' 


from pro. The second element is the -EU which appears in 
Cyprian ow 'this,' Arcadian -raw 'these things,' Goth, thannu 
1 so then,' and I think TTO.VV ' altogether ' (i.e. irn-w, the first 
element going with Lat. pen-itus, Sabler K.Z. 31. 371), see 
Persson Idg. Forsch. 2. 251 : *Trp6^-w became n-po'xyv by De 
Saussure's law, Mem. Soc. Ling. 7. 90, as *Aw| r -j>os (cf. Zend 
raokhshna 'shining') became Xw^os. 

pi)*lo? ' rug ' (Horn.), ^'709 (Anacreon), pegto ' dye' (Epicharmus), 
and a\o-vp~i //9 ' dyed with sea purple ' (Att., see Schulze 498, w.), 
must go with pi/ffffw 'beat the ground' (II. 18. 571), cloth being 
stamped on to make the dye penetrate : so German walken ' to 
clean cloth ' is the same word as Eng. walk. The root must be 
vreg-, Ablaut vrg- in pdaaw ' push ' : Sk. raj- ' to be red ' cannot 
be connected, if only because the meaning would be too narrow. 

fftiTivt) ' chariot ' may be a Persian word, standing for *g(nivr) 
(as acnpcnrii<s stands for *garpa.7rt)<?, Old Persian khshatrapuvan- 
1 viceroy '), and going with Sk. Icshat-tdr- ' charioteer.' 

ffavpwTi'ip 'spike at the butt-end of a spear' must mean 'twirler,' 
from a Verb *aavp6w, itself from a Subst. aavpo?, originally, I 
would suggest, meaning ' a stirring-stick,' and hence coming to 
mean ' a lizard,' which when motionless looks like a piece of wood. 
This aaupo<i=*TFap-Fo<s, from a root tvr-, whence also Lat. trua 
' stirring-ladle' (on the ru see p. 11 med.), orpvvw ' urge,' orpa\ew^ 
'quickly' ; Ablauts (1) tur- in -ropvvrj 'ladle,' 1 a Dissimilation for 
*-rvpvvrj, as KOK>CV% is for *KVKKV!;, Lat. cuculm; (2) tvor- in Old 
Norse thvara ' stirring-stick ' ; (3) tver- in Ags. thviril ' churn- 
handle,' Eng. twirl, Old High German dweran 'to mix up,' Sk. 
tvar- ' to hasten.' From a by-form stver-, stur- (see Schrijnen, 
Phenomene de 1' S Mobile) comes (nvpag, p. 12. 

o-aym, which in Homer always means ' dead body, carcase,' 
may go with auis ' safe,' and mean ' remnant, what has escaped being 
eaten by dogs or birds ' : II. 3. 23 iaa-re. \eiav e^opy /u.^a\w eVJ 

' being,' eV<ToWa<<? ' finding ' (both in Pindar), seem 
to point to an Aeolic Yerb *Toaa~ifju ' I do so much,' from ToWo?. 

0y\| r ' guardian ' may originally have meant ' the man in 
the house,' oj/cerj/?, bhu-1- being Ablaut of bhou-1- in Old Norse 
bol 'lair' (Wiedemann Lit. Praet. p. 137, despite Kluge 

1 Pick, "Worterbuch 4 , 1. 499, adds rvp6s 'cheese'; but this is not made 
by stirring, though butter is. 


K.Z. 26. 97), bh5-l- in 0w\evw 'lurk,' 0wXa9 'in his den' (of 
a bear). So 0w\// 'tribe' may originally have meant 'house' 
in the sense of Lat. gens, 

*Xe/sj/9, used by Homer in the forms x^/"7 a ) X e/ /"/ 4 ' X 6 '/"7 e? > must 
go with \eip an( l niean ' belonging to a handicraftsman,' xW'fa 
as opposed to a warrior. Thus it is used contemptuously, II. 1. 80 
/3a<7tAet>? ore yjuas-rai avSpi X 6 V 7 / f (' a ^ ow fellow'), Od. 15. 324 
old re TCUS a<yadoifft Trapadpwtuffi ^e/jj^es ; and so in the Neuter, 
II. 14. 382 effd\a fiev eV0\o9 e8we, x^/"/ a ('those fit for an 
artisan') e -^eipovi Boiricev. In Od. 14. 176 oim x e 7"7 a waT/oo's, 
'not like a handicraftsman beside his father,' we have a Genitive- 
Ablative of comparison, such as Brugmann Griech. Gramm. 2 
183 finds in Thucydides' roXeyttov 6fio\offttnarov -rSiv Trpo^efyevij- 
fievwv ' most notable in comparison with those before.' The 
Comparative of *X^/MJ is x e P e ' wv or X e 'p wv ' niore fit for an 
artisan,' or, as in II. 14. 382 above, ' more like an artisan.' But 
in another view battle was called ' handwork,' xapfirf (with the 
' reduced ' root xj-) 5 aQ d with this 1 would put x e Pf-^ lov ' a stone 
used in battle as a missile.' 





FOR THE YEARS 1842-1853. IN 6 VOLS. CLOTH, 3. 

The Philological Society's Transactions, 1854, 1855, 1856, One 

Guinea each; 1857- Part I. only, 10s. ; 1858, 1859, 1862-63. 12s. each. 

The Philological Society's Transactions, 1860-61 (including the 

Play of the Sacrament, and a Cornish Poem, ed. by Dr. Whitley Stokes). 12s. 

The Philological Society's Transactions, 1862-63. 12s. 

The Philological Society's Transactions, 1864, including the 

Creation of the World : a Middle-Cornish Drama. Edited by Dr. "Whitley Stokes, 
and Mr. Barnes's Grammar and Glossary of the Dorset Dialect. 12s. 

The Philological Society's Transactions, 1865, with a Glossary 

of Reduplicated Words, by H. B. Wheatley, Esq. 12*. 

The Philological Society's Transactions, 1866, with a Treatise 

on the Dialect of Banffshire, and a Glossary of Words not in Jamieson's Scot- 
tish Dictionary, by the Rev. Walter Gregor ; and an Etymological Glossary of 
the Shetland Dialect, by T. Edmondston, Esq., of Buness. 12s. 

The Philological Society's Transactions, 1867, with an Essay on 

Palaeotype. by A. J. Ellis, Esq. ; and a Glossary of the Lonsdale Dialect, by the 
late R. B. Peacock, Esq. 12s. 

The Philological Society's Transactions, 1868-9, with Dr. Whitley 

Stokes' Cornish Glossary, and Mr. A. J. Ellis's edition of the Only English 
Proclamation of Henry III. 12s. 

The Philological Society's Transactions, 1870-2, Part /., 4s. 

(Parts II. and III. out of print.) 

The Philological Society's Transactions, 1873-4, Part I., 5*. 

(Parts ILandlll.out of print.) 1875-6, 15s. 1877-9, 

18s. 1880-1, 1 7s. 1882-4, 1 15s. 1885-7, 

1 5s. 1888-90, 1 10s. 1891-3. Part I. 10s. 

Index to Transactions and Proceedings, 1842-1879. 5s. 

The following may be had separately : 

The Philological Society's Early English Volume, 20s., in- 
cluding Liber Cure Cocorum, an Early English Cookery Book in Verse 
(ab. 1440 A.D.); edited by Rev. Dr. R. Morris. Hampole's Pricke of Conscience 
(ab. 1340 A.D.) ; edited by Rev. Dr. R. Morris. A Fourteenth-Century Transla- 
tion of Grosseteste'sChateaud'Amoui (ab. 1320A.D.); ed byDr, R.F.Weymouth. 

Levins's Manipulus Vocabulorum, the earliest English Rhyming 

Dictionary, 1575. Foolscap 4to. 16s. 

The Old-Irish Glosses at Wiirzbnrg and Carlsruhe, edited by 

Whitley Stokes, D.C.L. Pt. I. Glosses & Translation, os. 

On Early English Pronunciation, with especial reference to 

Shakspere and Chaucer, by A. J. Ellis, Esq.^F.R.S. 4 Pjtf ts^lOs^ch j. R. Y.25*, 



MEDIEVAL GREEK TEXTS : A Collection of the Earliest 
Compositions in Vulgar Greek, prior to A.D. 1500. With Prolegomena and 
Critical Notes by W- WAGNER, Ph.D. Part I. Seven Poems, three of which 
appear for the first time. London, 1870. 8vo. 10*. Gd. 

LIBER CTJRE COCORUM. Copied and Edited from the Sloane MS. 
1986, by the Rev. Dr. RICHARD MORRIS. 8vo. 3s. 

A Northumbrian Poem, by RICHARD EOLLE DB HAMPOLE. Copied and 
Edited from Manuscripts in the British Museum, with an Introduction, Notes, 
and Glossarial Index, by the Rev. Dr. RICHARD MORRIS. 8vo. cloth. 12*. 

CASTEL OFF LOUE (Chateau d'amour or Carmen de Creatione 
Mundi). An Early English Translation of an Old French Poem, by ROBEKT 
GROSSETESTE, Bishop of Lincoln. Copied and Edited from the MSS. with 
Notes and Glossary, by Dr. R. F. WEYSIOUTH, M.A. 8vo. cloth. 6*. 



SAYCE, D.C.L., LL.D., Professor of Assyriology, Oxford. Third Edition. 
2 vols. Crown 8vo. 9s. 


A. H. SAYCE, D.C.L.. LL.D., Professor of Assyriology, Oxford. Fourth 
Edition, revised and enlarged. Crown 8vo. 10s. 6d. 


GERMANIC LANGUAGES. By KARL BRUGMANX, Professor of Comparative Philo- 
logy in the University of Leipzig. Translated by JOSEPH WRIGHT, Ph.D. Vol. 
I. Introduction and Phonetics. 8vo. 18s. Vol. II. Morphology (Stem Forma- 
tion and Inflexion), Part I. 16s. Vol. III. Morphology, Part II. Numerals, 
Inflexion of Nouns and Pronouns. 12s. &d. 

1846-1878. By R. N. GUST, Author of " The Modern Languages of the East 
Indies." First Series. 10s. 6rf. Second Series, 1847-1887, with 6 maps. 21s. 
Third Series, with Portrait. 21s. 


SLAVIC AND LATIN. Ilchester Lectures on Comparative Lexico- 
graphy. Delivered at the Taylor Institution, Oxford. By CARL ABEL, Ph.D. 
Post 8vo. 5s. 


and its relation to the Folk-Lore of Europe during the Middle Ages. With Two 
Appendices and Plates. By M. GASTER, Ph.D. Large post 8vo. 7s. 6d. 

Lectures on the Principles of Linguistic Science. By W. D. WHITNEY. 
Fourth edition. Crown 8vo. 10s. 6d. 

LANGUAGE AND ITS STUDY, with especial reference to the 
Indo-European Family of Languages. Seven Lectures by W. D. WHITNEY, 
Professor of Sanskrit in Yale College. Edited by the Rev. R. MOERIS, M.A., 
LL.D. Second Edition. Crown 8vo. 5s. 


Sixth edition. Crown 8vo. 5*. 

8vo. Second revised and enlarged Edition. 15s.