CH. VI.

Slight traces of

the forma

tion of su

perlatives.

or quantity of particular words and roots. Here we might
most naturally expect to find it in the formation of the
superlative; and such is indeed most common in savage
dialects. It is also found, though not as a rule, in San-
skrit, e.g. alpa is little, and alpálpa is very little. But if
it ever existed in Greek and Latin, it has entirely passed
away before historic times, when we find the requisite in-
crease of idea expressed by formative suffixes. Yet even
in them, as Pott has pointed out, a lingering feeling of
the possibility of the process is seen in such words as
τρισμέγιστος, τρίδουλος, and the Latin triscuria. Here
we have reduplication, or triplication rather, in the spirit
if not in the letter. A savage would have said Scuλodov-
λodovλos. The more cultivated Greek could express the
λοδουλος.
same idea with more dexterity. A further example is to
be found in the rather artificial compound used by Cal-
limachus in his very beautiful epitaph on Heracleitus,

ἀλλὰ σὺ μέν που,

ξεῖν' Αλικαρνασσεύ, τετράπαλαι σποδίη.

A sort of reduplication again may be seen in the very common ovdeìs oï, nemo non, &c. These are all superlatives: in all these it is intended to express the strongest affirmation. And though here the reduplication is certainly not of the reason, since one negative drives out the other, yet currency was probably given to the expressions by the fact of their coinciding with the popular love for repetition of the same sound.

Far more important for us, as entering more widely into the building up of the languages, are the traces still to be found in Greek and Latin of the systematic reduplication of primary roots to produce frequentative Regular formation and desiderative verbs and, more rarely, nouns. In Sanof frequen- skrit such verbs are regularly formed from every root, intensives. by reduplication. Thus budh in Sanskrit means "to know:" bo-budh (or bo-budh-ya) denotes "to know frequently," or "to know well" (i.e. is either a frequentative

tatives or

or intensive verb), bu-bodh-i-sha is "to desire to know."
It will be observed that in two out of these three verbs,
suffixes ya and sa are added, over and above reduplica-
tion; it is possible that when they were first so used,
they retained their primary sense, whatever that was;
and so modified the meaning of the root as well as
the reduplication. So much at least is certain, that ya
distinguished an intensive from the desiderative which
ended in sa.
But it does not appear that ya can be con-
nected with any root which would necessarily or even
naturally convey the idea of intensification: or that sa
conveyed to the hearer the idea of wishing apart from
these compounds. Therefore to me it is more likely that
when they were first thus employed they were purely
formal, mere grammatical machinery. The spirit was in
the reduplication. Turning now to Greek and Latin we
find-besides such onomatopoetic verbs as λaλayeîv, mur-
murare, and many others-intensives or frequentatives,
for one signification often runs into the other, thus formed;
as μapμaipei, "to flash," from pap, originally to rub,
and so to smooth down, polish. Similarly πaμḍaive is
an intensive of pav: the whole root is repeated, and the
aspirate changed to the hard, and the dental nasal to the
labial nasal according to rule. So also yapyapičew, “to
gurgle," from yap, "to swallow" (the Lat. √vor for
√gvor in carni-uoru-s, uora-re, but also gul-a, glu-tire,
&c.): compare the Latin gurgulio. Likewise πOL-TVV-W,
'to pant," from /Tvʊ, "to breathe;" Seidioooμai from √di,
"to fear;" μepunpilew, and the Latin me-mor, memor-ia,
&c. from a root which in Greek took the forms μap and
μep, and √mor in Latin, but which is to be referred
back to an Indo-European SMAR: the Sanskrit alone
has retained the s: thus peрμnρiew is "to be careful,"
intensified from the simple root which means "to re-
member."

μαρ

It will be observed that in many of these examples the reduplicated syllable is strengthened; as in TOIπVÚw,

CH. VI.

CH. VI.

Frequent loss of the original meaning.

δειδίσσομαι, μαιμάω, to pant with eagerness, κοικύλλω, ποιφύσσω, παιπάλλω, δαιδάλλω : in the nouns λαίλαψ, whirlwind, from Maß, to seize, μaμákтns, whence the Attic month paiμaктηpiwv, probably in po-pul-us, the tree, from spul1, and others. In these there is little doubt that the feeling of their origin survived until historic times; that men were conscious in using these words that they were employing intensives, and felt their relation to the simple root. It is not easy to account for the (much rarer) strengthening of the radical syllable: in ȧк-wк-ý from AK, something very sharp, in dywyós and dywyń, where the reason is not plain: though it is in ỏπwπý and ỏowdŋ; and in the irregular lengthening ed-wd- from ED, clearly on the analogy of the others. The accent, which in all is on the last syllable, may possibly have something to do with it. But there are others where we find the reduplicated syllable weakened, as in μepunpišeiv and měmor; in kíkivvos, a curl, (compared with the Latin cincinnus, whence Cincinnatus), in su-surrus; in κéкρağ, κεκρύφαλος, a thick covering, in τέτανος, in βέβαιος, and Béẞnλos (each from BA, the first that which may be gone upon with physical, the latter with ceremonial impunity), the radical vowel has been allowed to sink to e: in Tenn, in Tɩlós and Tɩáσos, tame, all probably from DHA to milk; in many Latin verbs titillare (compare Greek Tiλλew), titubare, &c., the vowel is the weakest of all. In all such words the feeling of their origin was gradually passing out of the minds of men; the emphasis had ceased to be laid on the reduplicated syllable, as it must have been at first when it was the significant part of the compound; and hence the syllable became weakened. This is a good illustration of the change which passes by degrees over all language; that which was originally formative loses its signification, and becomes only mechanical; the living principle passes out, and deadness comes on. And this

1 See p. 104.

2 See Leo Meyer, 1. 417-429.

brings us to a numerous list of verbs where this deadness is almost perfect: in some the intensive or desiderative force has merely died out; in others the no longer significant form seems to have been used to express a different idea. Such verbs are μμéoμai (Greek μe, Indo-European MA, "to measure"), which seems originally to have signified "I frequently measure myself," and thus, in a restricted sense, "to measure myself by some one, to copy or imitate;" where the frequentative force is perfectly lost. The Latin imitor and imago are not improbably blunted forms of mi-mitor and mi-mago1, and so formed originally on the same principle from the same root, which is found in me-tior, and strengthened in mensa and mensura. As μιμέομαι stands by regular phonetic change for μι-μα-οpat, it exactly corresponds in form to bo-bhud-ya, mentioned above.

In such verbs as didáσкw, apaρíoxw, &c., the intensive force seems not only to be lost, but a causal sense to have taken its place—unless indeed they are to be explained as desideratives—the terminational σкw corresponding to the sa of bubodhisha, so that ȧpapioxw should mean "I desire something to fit,” Sidάσкw, “I desire some one to think." But it is more likely that these forms should be connected with another very important class of verbs: in the formation of which this same principle of reduplication is employed, but for a more limited object. I mean such verbs as Sidwμi and TiOnμ in Greek. In these verbs it will be δίδωμι τίθημι observed that reduplication is found only in the present tense and the closely connected imperfect; not in the future Soow or the aorists edwкa and edwv, whilst the reduplication of the perfect is different in its nature. But in the intensives, which I have instanced above, the reduplication passes through all the tenses, although their occurrence is not frequent. In fact, in them the reduplicated verb is practically a new root; in these others the

1 See however Corssen, K. B. 252.

CH. VI.

reduplication is an accident of the present tense. How is this difference to be explained?

All language must of course be able to distinguish the incomplete from the complete stage of action, the "is doing" from the "is done," the yiyveolar from the eivai. But this distinction is not sufficient to express all our conceptions with sufficient clearness. In describing an incomplete action we require often to express whether the operation is over in a moment or requires time; whether it is momentary or lasting-to distinguish the "I do" from the “I am doing,” the γενέσθαι from the γίγνεσθαι. Thus then we have three stages of action, so to speak: the Momentary, the Continuous, the Completed. And each of these stages ought in a perfectly logical language to have its own three sub-divisions in time, the past, the present, and the future. That is, it should possess nine forms produced by internal modification of the root, with the help of such suffixes as have lost their original signification and have become mere parts of the machinery of grammar; not by periphrases, as in seven out of nine times in the forms by which the English denotes the nine distinct ideas; thus:

Unfortunately, historical investigation of the development of language shews that in the early stages of growth inflexions, derivatives, and the rest of the stock of grammar, are not formed to meet previously-felt logical needs. The order of the process is just reversed. A language