The nominative accusative of the neuter of visi is vise singular, visa plural. The nominative singular feminine of visi is visa, accusative singular feminine visã, nominative accusative plural feminine visẽ. The remaining cases of the neuter singular and plural and of the feminine plural, are the same as the masculine, those of the feminine singular are the same as feminine si with vi prefixed.1

209. In the nominative singular of the noun, final s is dropped, and a preceding it has become ; in the neuter, final m is dropped, and a preceding it has become o. Final ā has become a; and the stem endings u, u have become , u; ya often becomes yi, iya becomes iy.2

The vocative singular when different from the nominative reduces a to e, a to o, while it raises the final vowel of the -i stems and -u stems. In Sanskrit also these stems strengthen the final vowel in the vocative.3

The m of accusative singular is dropped without leaving any trace except in the feminine a stems; the connective a changed to e remains in the consonantal stems.3

In the genitive singular also of the consonantal stems the connective a changed to e remains while the s has been dropped. Final and are raised to u and i, as in Sanskrit the u and i are Gunated before s. The singular genitive ending -a of first and second declension is deduced by Schleicher from asya (aya, ā). To this Miklosich says that there are insuperable phonetic objections; but he does not state them. He considers that it represents the old ablative -at (4), but he does not explain how the ablative with its comparatively small range of use could have supplanted the genitive. The genitive singular of the feminine -a stems, which ends in - unless y precede the a, when it ends in -ẽ, Miklosich deduces from -ã, which has undergone the same two changes in the present participle; and this -a he deduces from the old locative -um. But as the long vowel preceding the genitive s in Sanskrit in the plural caused it to become n (9), by strengthening the sonancy and suggesting an easier passage of the breath, it may have here caused a tendency to a similar change; and the nasal would be favoured by the influence of preceding y being unfavourable to a guttural breathing. Bopp considered n in the Sanskrit genitive plural to be occasioned by hiatus,5 but if a consonant was needed to avoid hiatus how came the original s to be dropped? He deduced the Slavonic nasal in the genitive from the original final s, and compared the final in the plural and dual of the Greek verb where Sanskrit has 8 (66).

The dative singular of first and second declension ends in -u, which, as in the Irish dative (111), is doubtless due to an original b' (184). The other datives all take up i, corresponding to Sanskrit ē.

The instrumental singular ends in mi, as in Lithuanian; and this corresponds to an original bi It is absorbed as a nasalisation by

1 Miklosich, iii. pp. 52, 53. 4 Ibid. iii. p. 4.

2 Ibid. iii. p. 2.

3 Ibid. iii. p. 3.

5 Bopp, Vergl. Gram., sects. 17 b, 246.

6 Ibid. sect. p. 271.

feminine -a. The feminine i stems, and also for the most part the feminine a stems,1 add to their stem ending in the instrumental singular ya, which is a pronominal element, as may be seen in the pronominal declension. The instrumental relation is so strong an element of thought that it coalesces with the concrete idea of its object less readily than the more abstract case relations. The thought of it consequently elicits an abstract sense of the object as such, which is thought with attention directed to it; and this mental act strengthened with a sense of the feminine gender is expressed by a pronominal element (Def. 7).

A similar influence of the dative and instrumental relations, as they are thought in the dual and plural, and of the locative relation as thought in the plural, has strengthened the v- stem ending with a; and has strengthened the stem ending in the genitive locative dual of all the consonantal stems with i, which probably is pronominal (184), being needed because these stems are deficient in an abstract objective part or substance after the attributive part (Def. 4).

The consonantal stems take i also in other cases as given above. Miklosich thinks the locative endings -uve, -ene, which are in the oldest sources, are perhaps genitive. The ending i is due to the coalescence of locative with i added to the stem ending. The locative - of fourth declension seems to be a reduction of -ovi, which corresponds to the Vedic locative ending of the u stems -avi,2 and the -e of the first three declensions to the coalescence of a and a with i.

The nominative vocative accusative dual of the first declension corresponds, like the Zend -a, to Vedic -ā (12). In the others it is similar in its formation to the Sanskrit.

In genitive locative dual -u corresponds to Sanskrit -ōs; and in dative instrumental -ma to Sanskrit -b'(y)ā(m).

The nominative plural of first declension has -i corresponding to -ē in the Sanskrit pronouns; that of second declension a to Sanskrit -ā(ni); that of third declension is same as genitive singular - or -ẽ, corresponding like this to Sanskrit -us; the -e of the others corresponds to Sanskrit -as.

The accusative plural has dropped -ns; the genitive plural has reduced -um to -; the dative plural has reduced byas to -my, the instrumental plural has reduced bis to mi, the first and second declension absorbing -mi into u; the locative plural has -h for su, preceded by è in first and second declension, as in Sanskrit.

210. The stems of the personal pronouns receive remarkable additions; tebe, sebe correspond probably to Sanskrit tava, sava; but tẽ, sẽ cannot, and these show that me, mene involves a distinct pronominal element ne (155). Perhaps the heavier ending of the dative, locative, and instrumental singular weakened me-, so that under the influence of n following it it became my. In the instrumental singular the stems seem to be strengthened with ya, and their a lengthened as in composition, which seems to have affected also ta and sa; -mi was absorbed.

1 Miklosch, iii. p. 6.

2

Bopp, Vergl. Gram., sect. 199.

The nominative dual of first person has taken up i; that of the second, and the accusative dual of both, have a corresponding to original a. Is there more distinction of the individuals in i as expressing juxtaposition than in ā as expressing extension?

The nominative plural corresponds perhaps to original -ās, like genitive singular third declension; ti corresponds to te, the accusative plural to -ans.

The datives singular mi, ti, si, the genitives singular mẽ, tě, sẽ, and the accusatives plural nu, vụ, are enclitic. The latter are used also for the dative.1

In the demonstrative pronouns the genitive ending -go is probably a pronominal element ga corresponding to Sanskrit ga,2 and this is preceded by the composition vowel, changed, however, to e by y preceding it. The dative ending -mu is the dative of original sma; 2 and the feminine pronouns take -sya, whose genitive -syās is represented by -ye. Before the case endings beginning with m the stems take i, probably from the y which originally followed the b' which m represents (11).

The genitive plural -sām became -h, and as it was thus reduced almost to the same form as the locative -ihu (4), the i also was taken from the latter.

211. The adjectives, except the possessive adjectives (223), may be declined, as in Lithuanian, in composition with y, the adjective taking the case endings as well as yu, or being prefixed in its stem form. The latter takes place in all the cases whose ending begins with a consonant.3

The comparative of adjectives is formed by the suffix -iyus (5), to which y is added, except in the nominative singular masculine and neuter. If the stem ending of the adjective is dropped before the comparative element, this becomes iyusy or -yusyu; if not it coalesces with i of the comparative element into ē.1

The comparative serves also in the Slavonic languages for superlative, being then accompanied in the younger languages, and sometimes also in Old Slavonic, by a strengthening particle.5

212. The numerals for 3 and 4, triye, t'eturiye when masculine, are declined as plurals of masculine i stems, those for 5 to 10 end in -ti, and are declined as feminines in the three numbers.6

The ordinal of 3 is formed with -ti, those of 4, 5, 6, 9, 10 with -tu, those of 7, 8 with -mu, but they are all declined in composition with yu7 (151, 225).

213. The verbs have a present stem and a non-present or infinitive stem. The latter is in some verbs-I. primary, i.e., the mere root. Other verbs form the infinitive stem with a suffix; of which suffixes there are five, nã, ē, i, a, ua (ova).8

II. The stems formed with na subjoined to a root are transitive or intransitive, many of them passive; some of them can drop the

1 Miklosich, iii. p. 46.

Ibid. ii. p. 322.

Miklosich, iii. pp. 35, 37.

2 Ibid. iii. p. 47.

3 Ibid. iii. p. 55. Bopp, Vergl. Gram., ii. sect. 305. 7 Bopp, Vergl. Gram., ii. sect. 322.

8 Miklosich, ii. p. 420.

suffix. If the root forms also an e stem, the nã stem is distinguished from this by perfectivity. The deverbal nã stems are perfective (227), the denominative passive.1 Miklosich considers that -na corresponds to Sanskrit -nu, Greek -vv, Gothic -n, and that the nasal of a came from an older ; before a vowel the ã, like u, becomes ov.2

III. The primary e stems, i.e., those formed from roots, are generally neuter and durative. If the root forms also an i stem or an a stem, the ē stem is passive.

The denominative - stems are durative of inchoative, and signify to become what the noun denotes.3

To e corresponds Sanskrit -ya, which, according to Miklosich, must have become first -ay, then -ey, also Latin -ē, Gothic -ai.*

IV. The stems are all denominative, according to Miklosich, and generally durative, their signification being causative, either transitive or reflexive. But he says that when a primary stem and an i stem belong to the same root, the latter is causative of what the former signifies. The radical vowel is Gunated as in Sanskrit, and -i corresponds to Sanskrit -aya.

Miklosich thinks that in Sanskrit, and in all the other languages of the family, the causative was formed on a verbal noun to which the first a of aya belonged, the causative element being -ya.

6

Amongst the i stems, there are also some intensives and diminutives.7 V. The a stems are either primary or deverbal or denominative. The primary and denominative are durative unless they have a prefix. The deverbal are iterative. When the latter are formed on primary stems the radical vowel is strengthened, but with a different vowel increase from that of the nominal stems. The -a of the denominative a stems corresponds to Sanskrit -aya, the stem ending of the noun being dropped. The deverbal a stems, in the strengthening of the radical vowel and in their meaning, approach in some degree to the Sanskrit intensives. Even primary verbs take this -a, which, however, with some deverbals and denominatives, they keep only in the infinitive.9 No class of verbs includes so many intensives and diminutives as the a class.10

The combination sk is properly in Slavonic one of those elements by which roots are determined. It remains accordingly through all the verbal forms; thus i to go, iska to seek, př to sing, piska to sing with the flute.11

VI. The ova stems are all secondary, deverbal or denominative. The suffix consists of u and a, u arising, according to Miklosich, from the nominal stems, which correspond to the Sanskrit a stems, and spreading from them to others, and a being imperfective (227). The diminutive of these is formed with -kova, -kiva.12

214. From the infinitive stem, the aorist is formed either by subjoining to it, if it end in a consonant, the short person endings, a

connective vowel e being inserted, which, however, is different from the suffixed -e of the present, or by subjoining s (27) to it if it end in a consonant, before which t, d, z, s, are dropped, and the vowel strengthened for compensation, and is dropped without such compensation; the s may become h; r and 1, mr and ml, at the end of a stem are treated as vowels. These two formations are peculiar to the Old Slavonic and Servian.

In the latest form of the aorist, which is common to all the Slavonic languages, the s, which becomes h between vowels, is joined to a consonantal stem by a connective vowel o, which in the loss of the person endings of second and third singular became e. This o expresses the remotion of the past. The stems which end in -nã preceded by a consonant, sometimes change na before h to o.1

If the verbal stem end in a vowel, the later aorist coincides with the earlier, as the s of the latter is h between vowels, and the vowel subjoined to the stem in the former is replaced by the final vowel of the stem.

The future subjoins sy (26) to the infinitive stem, but is formed only by the verb bu.2

The first past participle active is formed with the suffix s, the second with -l, used only as a predicate. The former is declined as ending in -usy, except in nominative singular masculine and neuter (211).

The past participle passive is formed with -en or -tu.

3

The infinitive is formed with -ti, rarely tu, the supine with -tu; -ti might be a locative case or a genitive, more likely, according to Miklosich, a dative, -tu an accusative.

215. The present stem is distinguished by the suffix e, which corresponds to the Sanskrit conjugational vowel a.

Some verbs do not take this e. And the i stems and the primary ē stems take the suffix only in first singular present. Probably it coalesced in the other parts with the stem ending into i Some verbs and classes of verbs have i before the e, as in Latin cupio, and in Greek κράζω (κραγιω).4

The present stem in some classes of verbs receives also internal change, in the strengthening or nasalisation of the radical vowel, or external addition, either -i, which Miklosich considers euphonic for the prevention of hiatus, or -d, which he connects with the root dē, Sanskrit d'a (159, 192), or reduplication, which, however, is found. in only two verbs, da to give, and de 5 (189).

From the present stem is formed the imperfect, by lengthening the present suffix e to è to express the going on in past time, and subjoining hor ah, h being same as s of the aorist, and becoming euphonically s before e; the imperative by adding the old optative i; the present participle active by adding -nt to o, the prolonged present vowel by which it is absorbed into e, if y precedes, into u after any other consonant; and the present participle passive by the suffix -mụ 1 Miklosich, ii. pp. 487, 488.

3 Ibid. ii. pp. 328, 488, 489; iv. pp. 817, 844.

5 Ibid. ii. pp. 491, 492.

2 Ibid. ii. p. 488.

4 Ibid. ii. p. 489-491.