184 PALI sometimes also into i or u ; as isi (S. rishi), dalha (S. dridha), putha (S. prithag). E and o, representing S. a; and an respec tively, can before double consonants be further shortened into i and u, just as other long vowels may be shortened under the same circumstances ; thus ussukka (S. autsukya), rat (ha (S. rashtra). Some anomalous vowel changes are exhibited in the following examples : komjuilua (S. kaumlinya), pana (S. punar), purisa (S. purusha), iisu (S. isliu), riil iifl, (S. vijha), hcttM (S. adhastat). As regards consonants, Pali has only the dental sibilant, and replaces by anusvara most final consonants of Sanskrit words ; as manam (S. manak), sanim (S. sanais), khattiiHi (S. kritvas). Two or more consonants meeting in. the middle of a word are mostly assimilated, as umniagya (S. unmarga), pabbhdra (S. pragbhara). Other changes are jKinha (S. pragna), pall an ka (S. paryanka), ddthd (S. damshtra), and of initial consonants lattlu (S. yashti), ludda (rudra), ndngala (S. langala), kipillika (S. pipilika), khdnu (S. sthanu). Contraction is very frequent, as well as metathesis, as the following examples will show : kho (S. klialu), acccka (S. atyayika), Accra (S. acarya), cuddasa (S. caturda9an), issera (S. aigvarya), abboMra (S. avyavahara). In the Scenic Prakrits and in the Magadhi of the Jains the consonantal decay has reached a much higher stage than it has in Pali, showing that the latter holds its place between the former and the Sanskrit. This applies also to Sandhi, which in Pali is indeed sporadically and irregularly attended to, but shows a tendency to being altogether neglected. There is no dual in the declension any more than in the con jugation ; the only remnants of it appear to be to (S. tau) and ubho (S. ubhaut. The old dative case is rarely used, and the genitive takes its place. The declension of nouns has in some cases been encroached upon by the pronominal declension. According to the nature of Pali phonology, there cannot be any real consonantal stems, and therefore no regular consonantal declension. Final consonants are either dropped or have an a added to them. In the former case the final consonants reappear before the vowel terminations, in the latter the declension follows the false analogy of the a-declension. Thus, dhimd (S. dhimat) is declined as follows: Sing. nom. dhima, dhimanto ; voc. dhimam, dhima, dhima; ace. dhimantarn, dhimam; instr. dhimata, dhhnan- tena ; dat. gen. dhimato, dhimantassa, dhimassa ; abl. dhimata ; loc. dhimati, dhimante, dhimantasmim, dhimaiitamhi ; Plur. nom. voc. dhimanto, dhimanta ; ace. dhimante ; instr. abl. dhimantebhi, dhimantehi ; dat. gen. dhimatam, dhimantanam ; loc. dhimantesu. Examples of multiform cases are the loc. sing, of itadi, which exhibits the forms nadiya, nadiyam, najjam ; the voc. plur. of the. honorific pronoun bhavam (S. bhavat), which has bhavanto, bhonto, bhante ; the gen. dat. sing, of pita, which has pitu, pituno, pitussa, and in the plur. pitunam, pitunnam, pitara- nam, pitanam ; the loc. sing, of mano, manam (S. manas), which has manasi, mane, manasmim, manamhi. The personal pronouns also show a variety of forms, some of which are still traceable in the modern Prakrits. Thus aJiam has in the plural nom. vayam, mayarn, ainhe ; ace. asme, amhe, amhakam ; instr. abl. amhebhi, amhehi ; dat. gen. amhakam, amhanam, amham ; loc. amhesu. Similarly, the gen. dat. sing. fern, of the demonstrative pronoun has the forms imissa, imissaya, imaya, assa, assaya. The Pali verb shows even more than does the noun a tendency to break with the analogy of the Sanskrit. Though native gram marians arrange the conjugations on a plan similar to that of the Sanskrit, the disorganizing process which pervades the whole of Pali grammar is in no part so advanced as in this particular. Thus, the present tense of the verb tha (S. stha) is that! as well as titthati ; of dhd it is dadhati, dahati, and dhati ; of dA dadiiti, deli, dati, and (by false analogy from the optative dajjam) dajjati ; of ji jayati, jeti, and jinati; of bhl bhayati; of rudh rundhati, rundhiti, rundluti, and rundheti ; of mar (S. mri) marati and miyati ; and of knr (S. kri) the plural has karoma, karotha, karonti, and also regularly kubbanti, from which form again by false analogy a 3d person singular kubbati has been derived. The termination re of the 3d person plural perfect atmanepada has been transferred to the present tense, where it is used along with -ante. But there is a general predilection for the parasmaipada termina tions, even in the .passive. While the perfect sensibly recedes before the other tenses, and is of rare occurrence, the use of the aorist largely encroaches on that of the imperfect, the conjugation of which is in many verbs influenced by the former, as, e.g., in the verb as, in which the imperfect is: 1st sing., asim or asi; 2d and 3d, asi; 1st plur., asimha; 2d, asittha; 3d, asimsu. In the impera tive par. the 1st sing, and 2d plur. do not differ from the corre sponding forms of the present. The affixes of the future (-ssa] and passive (-ya) may also be added to the special base; thus we, have the forms dakkhati and passissati, "he will see," and gamiyati and gacchiyati, "he is gone to." In the causative verb the form with p greatly preponderates, and may even be added to the special base, as, e.g. , sunapeti (S. gravayati), "he informs"; ganhapeti (S. grahayati). Lastly, the gerund in -tvd is not only used in compound verbs in preference to the one in -ya, but may also occasionally be superadded to the latter for the sake of greater precision. Thus, sajjitvasad + ya + i + tva; and abhiruyhitva = abhiruh + ya + i ^Tva. Instead of tril the forms trAna and tfina often occur. There are two forms of the infinitive, there being besides the usual form in -turn one in -tare, which appears to have lingered in the vernacular long after it was disused in Sanskrit literature. Literature. The study of Pali by Europeans is of com paratively recent date ; in fact, our knowledge of the very existence of an extensive Pali literature dates scarcely half a century back. It is true that in 1826 Professors Burnouf and Lassen were enabled, from an examination of certain Pali MSS. which had fallen under their notice, to give a general account of the language ; but it was reserved for the late Mr G. Tumour, colonial secretary of Ceylon, to collect the first trustworthy information concerning the sacred books of the island, and to edit and translate the first Pali text of any extent. His choice of the M(thiivansa, one of the oldest chronicles, was all the more fortunate, as, in the almost total absence of historical works in Sanskrit literature, these annals were calculated to yield a vast amount of information regarding the origin and earlier history of the Buddhistic religion in India. The book had been ready for the press many years, but was not published till 1837, while a series of articles by the same author, em bodying the results of his examination of the Mahdvansa and its commentary and of the contemporaneous Dipavansa (Jour. Bengal As. Soc., vols. v. and vi.), had been received by Oriental scholars with the utmost interest. The thirty- eight chapters published by him bring the history of Ceylon down to 477 A.D.; they comprise the original work of Mahdndma. Six more chapters, ready for the press in text and translation, were found among Tumour s papers at his early death in 1842, and are now in the India office library. The whole Mahdvansa, in Pali and Sinhalese, has since been printed at the Government press, Colombo, 1877-83, and an English translation is in progress. How ever, a critical edition of the earlier part, and more especially of the commentary upon it, is still a desidera tum. There is an excellent edition and translation of the Dipavama by Professor Oldenberg (London, 1879), according to whom the work was written between the beginning of the 4th and the first third of the 5th cen tury. Among the historical works may also be classed the Dath&vansa, a poetical history of the tooth-relic of Buddha, composed by Dhammakitti early in the 13th cen tury. The work was printed at Colombo in 1882, and an English translation by M. Kumaraswami appeared in Lon don in 1874. Further, the Attanagaluvansa, the history of a temple, likewise of the 13th century, edited and trans lated by J. D Alwis at Colombo in 1866. Other historical works are described in the catalogues of Pali MSS. Lastly, there exist many mediaeval Pali inscriptions, some of considerable extent, as, e.g., those of Kalyani in Burmah, which are now in course of publication, and are likely to yield valuable historical results. Many of them are accompanied by a translation in Burmese or Taking, a language now all but extinct. It is worth noting that neither in Ceylon nor in Cambodia have any old Pali inscriptions been found ; in the island the old inscriptions are in Sinhalese, in Cambodia they are in Sanskrit, fre quently with a translation in Khmer. Though there is an old ninefold division (navanga, see Dr R. Morris s " Report on Pali Literature," in Philological Society s Proceedings, 1880) of the canonical scriptures, it is the general practice of Pali scholars to abide by the division into three " baskets " (tipitaka, pitakattaya), first specified by G. Tumour, and then more correctly in Childers s Dictionary, p. 507, viz., the Vinayapitaka, the Xuttapitaka, and the Abhidharnmapitaka, or the baskets of discipline, of discourses, and of metaphysics. Only the
