to the Hebrew methegh) by `, then the word kárawat, a saw, is pronounced kárᵃwat; and kàḷakáḷaṇē̃, to be agitated, is pronounced kàḷᵃkáḷᵃṇē̃. In Konkani the vowel a assumes the sound of o in “hot,” a sound which is also heard in the language of Bengal. In dialectic speech ē is often interchangeable with short or long a, so that the standard sāṅgitᵃlē̃, it was said, may appear as sāṅgitᵃlã or sāṅgitᵃlā̃. The vowels ē and ō are apparently always long in the standard dialect, thus following Sanskrit; but in Konkani there is a short and a long form of each vowel. Very probably, although the distinction is not observed in writing, and has not been noticed by native scholars, these vowels are also pronounced short in the standard dialect under the circumstances to be now described. When a long ā, ī or ū precedes an accented syllable it is usually shortened. In the case of ā the shortening is not indicated by the spelling, but the written long ā is pronounced short like the ă in the Italian ballo. Thus, the dative of pīk, a ripe crop, is pikā̇s, and that of hāt, a hand, is hātā̇s, pronounced hătā̇s. Almost the only compound consonants which survived in the Prakrit stage were double letters, and in M. these are usually simplified, the preceding vowel being lengthened in compensation. Thus, the Prakrit kannō becomes kān, an ear; Pr. bhikkhā becomes bhīk, alms; and Pr. puttō becomes pūt, a son. In the Piśāca (see Indo-Aryan Languages) and other languages of north-western India it is not usual to lengthen the vowel in compensation, and the same tendency is observable in Konkani, which, it may be remarked, appears to contain many relics of the old Prakrit (Saurāṣṭrī) spoken in the Gujarat country before the invasion from the Midland. Thus, in Konkani, we have put as well as pūt, while the word corresponding to the Pr. ekkō, one, is ek as well as the standard ēk.
On the whole, the consonantal system is much the same as in other Indian languages. Nasalization of long vowels is very common, especially in Konkani. In this article it is indicated by the sign ~ placed over the affected vowel. The palatals are pronounced as in Skr. in words borrowed from that language or from Hindostani, and also in Marathi tadbhavas before i, ī, ē or y. Thus, caṇḍ (tatsama), fierce; jamā (Hindostani), collected; cikhal (M. tadbhava), mud. In other cases they are pronounced ts, tsh, dz, dzh respectively. Thus tsākar (for cākar), a servant; dzāṇē̃ (for jāṇē̃), to go. There are two s-sounds in the standard dialect which are very similarly distinguished. Ś, pronounced like an English sh, is used before i, ī, ē or y; and s, as in English “sin,” elsewhere. Thus, śimphī, a caste-name; śīl, a stone; śēt, a field; śyām, dark blue; but sāp, a snake; sumār (Persian shumār), an estimate; strī, a woman. In the dialects s is practically the only sibilant used, and that is changed by the vulgar speakers of Konkani to h (again as in north-western India). Aspirated letters show a tendency to lose their aspiration, especially in Konkani. Thus, bhīk (for bhīkh), alms, quoted above; hāt (Pr. hatthō), a hand. In Konkani we have words such as boiṇ, a sister, against standard bhain; gēr, standard gharī̇, in a house; āmī, standard āmhī, we. Here again we have agreement with north-western India. Generally speaking Marathi closely follows Māhārāṣṭrī when that differs from the Prakrits of other parts of India. Thus we have Skr. vrajati, Māhārāṣṭrī vaccai (instead of vajjai), he goes; Konkani votsũ, to go; Saurasēnī genhiduim, Māhārāṣṭrī ghettuṁ, to take; Marathi ghētᵃlē̃, taken. There is similarly both in Marathi and Māhārāṣṭrī a laxness in distinguishing between cerebral and dental letters (which again reminds us of north-western India). Thus, Skr. daśati, Māhārāṣṭrī ḍasai, he bites; M. ḍāsᵃṇē̃ to bite; Skr. dahati, Māhārāṣṭrī ḍahai, he burns; M. ḍādzᵃṇē̃, to be hot; Skr. gardabhas; Śaurasēnī gaddahō; Hindostani gadhā; but Māhārāṣṭrī gaḍḍahō; M. gāḍhav, an ass; and so many others. In Māhārāṣṭrī every n becomes ṇ, but in Jaina MSS. when the n was initial or doubled it remained unchanged. A similar rule is followed regarding l and the cerebral ḷ common in Vedic Sanskrit, in MSS. coming from southern India, and, according to the grammarians, also in the Piśāca dialects of the north-west. In M. a Pr. double nn or ll is simplified, according to the usual rule, to n or l respectively, with lengthening of the preceding vowel in compensation. Both ṇ and ḷ are of frequent occurrence in M., but only as medial letters, and then only when they represent ṇ or ḷ in the Pr. stage. When the letter is initial or represents a double nn or ll of Pr. it is always n or l respectively, thus offering a striking testimony to the accuracy of the Jaina and southern MSS. Thus, ordinary Māhārāṣṭrī ṇa, but Jaina Māhārāṣṭrī na, M. na, not; Māhārāṣṭrī (both kinds) ghaṇō, M. ghaṇ, dense; Māhārāṣṭrī soṇṇaaṁ, Jaina sonnaaṁ, M. sōnē̃, gold; Māhārāṣṭrī kālō, time, southern MSS. of the same kālō, M. kāḷ, time; Māhārāṣṭrī callai, M. tsālē, he goes or used to go. In some of the local dialects, following the Vedic practice, we find ḷ where d is employed elsewhere, as in (Berar) ghōḷā for ghōḍā, a horse; and there are instances of this change occurring even in Māhārāṣṭrī; e.g. Skr. taḍagaṁ, Māhārāṣṭrī taḷāam, M. taḷē̃, a pond.
The Skr. compound consonant jñ is pronounced dny in the standard dialect, but gy in the Konkan. Thus, Skr. jñānaṁ becomes dnyān or gyān according to locality.
Declension.—Marathi and Gujarati are the only Indo-Aryan languages which have retained the three genders, masculine, feminine and neuter, of Sanskrit and Prakrit. In rural dialects of Western Hindi and of Rajasthani sporadic instances of the neuter gender have survived, but elsewhere the only example occurs in the interrogative pronoun. In Marathi the neuter denotes not only inanimate things but also animate beings when both sexes are included, or when the sex is left undecided. Thus, ghōḍē̃, neut., a horse, without regard to sex. In the Konkan the neuter gender is further employed to denote females below the age of puberty, as in cēḍū, a girl. Numerous masculine and feminine words, however, denote inanimate objects. The rules for distinguishing the gender of such nouns are as complicated as in German, and must be learned from the grammars. For the most part, but not always, words follow the genders of their Skr. originals, and the abrasion of terminations in the modern language renders it impossible to lay down any complete set of rules on the subject. We may, however, say that strong bases (see below) in ā—and these do not include tatsamas—are masculine, and that the corresponding feminine and neuter words end in ī and ē̃ respectively. Thus, mulᵃgā, a son; mulᵃgī, a daughter; mulᵃgē̃, a child of so and so. As a further guide we may say that sex is usually distinguished by the use of the masculine and feminine genders, and that large and powerful inanimate objects are generally masculine, while small, delicate things are generally feminine. In the case of some animals (as in our “horse” and “mare”) sex is distinguished by the use of different words; e.g. bōkaḍ, he-goat, and śēḷī, a nanny-goat.
The nominative form of a tadbhava word is derived from the nominative form in Sanskrit and Prakrit, but tatsama words are generally borrowed in the form of the Sanskrit crude base. Thus, Skr. crude base mālin, nom. sing, mālī; Pr. nom. māliō (māḷiō); M. māḷī (tadbhava), a gardener; Skr. base mati-; nom. matis; M. mati (tatsama). Some tatsamas are, however, borrowed in the nominative form, as in Skr. dhanin, nom. dhanī; M. dhanī, a rich man. In Prakrit the nominative singular of many masculine tatsamas ended in ō. In the Apabhraṁśa stage this ō was weakened to u, and in modern Marathi, under the general rule, this final short u was dropped, the noun thus reverting as stated above to the form of the Sanskrit crude base. But in old Marathi, the short u was still retained. Thus, the Sanskrit īśvaras, lord, became, as a Prakrit tatsama, īśvarō, which in Apabhraṁśa took the form īśvaru. The old Marathi form was also īśvaru, but in modern Marathi we have īśvar. Tadbhavas derived from Sanskrit bases in a are treated very similarly, the termination being dropped in the modern language. Thus, Skr. nom. masc. karṇas, Pr. kannō, M. kān; Skr. nom. sing. fem. khaṭvā, Pr. khaṭṭā, M. khāṭ, a bed; Skr. nom. sing. neut. gr̥haṁ, Pr. gharaṁ, M. ghar, a house. Sometimes the Skr. nom. sing. fem. of these nouns ends in ī, but this makes no difference, as in Skr. and Pr. cullī, M. cūl, a fireplace. There is one important set of exceptions to this rule. In the article Prakrit attention is drawn to the frequent use of pleonastic suffixes, especially of -(a)ka- (masc. and neut.), -(i)kā(fem.). This could in Sanskrit be added to any noun, whatever the termination of the base might be. In Prakrit the k of this suffix, being medial, was elided, so that we get forms like Skr. nom. sing. masc. ghōṭa-kas, Pr. ghōḍa-ō, M. ghōḍā, a horse; Skr. nom. sing. fem. ghōṭi-kā, Pr. ghōḍi-ā, M. ghōḍī, a mare; Skr. ghōṭa-kaṁ, Pr. ghōḍa-(y)aṃ, M. ghōḍe, a horse (without distinction of sex). Such modern forms made with this pleonastic suffix, and ending in ā, ī or ē̃ are called “strong forms,” while all those made without it are called “weak forms.” As a rule the fact that a noun is in a weak or a strong form does not affect its meaning, but sometimes the use of a masculine strong form indicates clumsiness or hugeness. Thus bhākar (weak form) means “bread,” while bhākᵃrā (strong form) means “a huge loaf of bread.” The other pleonastic suffixes mentioned under Prakrit are also employed in Marathi, but usually with specific senses. Thus the suffix -illa- generally forms adjectives, while -ḍa-ka- (in M. -ḍā, fem. -ḍī, neut. -ḍē̃) implies contempt.
The synthetic declension of Sanskrit and Prakrit has been preserved in Marathi more completely than in any other Indo-Aryan language. While Māhārāṣṭrī Prakrit, like all others, passed through the Apabhraṁśa stage in the course of its development, the conservative character of the language retained even in that stage some of the old pure Māhārāṣṭrī forms. In the article Prakrit we have seen how there gradually arose a laxity in distinguishing the cases. In Māhārāṣṭrī the Sanskrit dative fell into almost entire disuse, the genitive being used in its place, while in Apabhraṁśa the case terminations become worn down to -hu, -ho, -hi, -hĩ and -hã, of which -hi and -hĩ were employed for several cases, both singular and plural. There was also a marked tendency for these terminations to become confused, so that in the earliest stages of most of the modern Indo-Aryan vernaculars we find -hi freely employed for any oblique case of the singular, and -hĩ for any oblique case of the plural. Another feature of Prakrit was the simplification of the complicated declensional system of Sanskrit by assimilating it in all cases to the declension of a-bases, corresponding to the first and second declensions in Latin.
In the formation of the plural the Prakrit declensions are very closely followed by Marathi. We shall confine our remarks to a-bases, which may be either weak or strong forms, and of which the feminine ends sometimes in ā, and sometimes in ī. In Prakrit the nom. plur. of these nouns ends masc. ā, fem. āō, īō, neut. āiṁ. We thus get the following:—
