No CrossRef data available.
We use cookies to distinguish you from other users and to provide you with a better experience on our websites. Close this message to accept cookies or find out how to manage your cookie settings.
We use cookies to distinguish you from other users and to provide you with a better experience on our websites. Close this message to accept cookies or find out how to manage your cookie settings.
Published online by Cambridge University Press: 24 December 2009
Among Sinhalese verbal forms, those which may be called nominal or participial verb-forms have received little attention hitherto from grammarians, either in Europe or in Ceylon. They present problems both of form and of usage.
The basic forms considered here are of the following types in early Sinhalese prose:
The existing grammars of Sinhalese give no adequate account of the difference between the pair balanu—balana in col. 1, nor of the syntactical contexts in which these forms occur.
1 Verbs are here given in forms in which they appear in Sorata, W., Śrīsumangalaśabdakoṣaya, Colombo, 1952–1956Google Scholar.
2 Gunasekara, A. M., A comprehensive grammar of the Sinhalese language, Colombo, Ceylon Govt. Press, 1891Google Scholar.
3 Geiger, W., A grammar of the Sinhalese language, Colombo, RAS, Ceylon Branch, 1938Google Scholar—a revised version of his Sinhalese grammar in Litteratur und Sprache der Singhalesen (I-A, Grundrissder Phil., I, 10. Ht.), Strassburg, 1900Google Scholar. This is supplemented by Studien zur Geschichte und Sprache Ceylons, München, 1941 (Studien)Google Scholar, and Beiträge zur singhalesischen Sprachgeschichte, München, 1942Google Scholar.
4 Paranavitana, S., Sīgiri graffiti, London, 1956Google Scholar.
5 Introduction, p. xiv.
6 See Geiger, , Grammar, §§ 146 + addenda, 149, 177Google Scholar, but cf. § 95.
7 -nā is confined to verbs of Geiger's conjugation II (i.e. those which later have a thematic i-vowel), and to the early period of attested literature. In such verbs there is, as shown in the table above, no formal difference between abstractor and present relator. In all other cases these two forms can only coincide when the abstractor is employed in the ‘passive periphrasis’. The other forms listed as havkiriya in SidS, i.50, viii.10 (ed. Kumāraṇatunga, M., posthumous edition, 1954)Google Scholar are not here considered as verbal, since, unlike the forms in -nu and -nā, they are not normally used transitively.
8 I use Geiger's nomenclature for case names.
9 See above, n. 7.
10 Certain common verbs use past relators of an irregular pattern; e.g. ya- ‘go’, e- ‘come’, de- ‘give’, kara- ‘do’, dan- ‘know’, ve- ‘become’, which use as past relators giya, ā, dun, koḷa, dat, and vū respectively.
11 The term ‘middle voice’ is used for Geiger's conjugation III, which he himself calls ‘das Medium’ (Studien, 26). A past relator form in -unu can theoretically be attached to any verb which already has a past relator in -u or -i. This middle past relator is rare in the works of Gurulugomi, on which this article is chiefly based, but it occurs: kärunu, 26.17.
12 Günther, H., ‘Über das ka-Suffix im Sinhalesischen’, ZDMG, XCVII, 1, 1943, 89–124Google Scholar. The poetical examples he quotes for ‘Adjektivbildung’, however, are hardly sufficient. If karana represents Pkt. karanaka, Geiger's etymological accounts of karanu and karana must be revised. There are other cases where final -a and -u of noun-forms seem to alternate in early prose, e.g. kumara and kumaru, raja and raju.
13 op. cit., introduction, § 350.
14 No formal criteria for division of Sinhalese into words have been established, but a phrase such as mā kī vadan (corresponding to a Sanskrit *mayā kathitavacanam) is most usually written as three words (though practice varies). Writing it as three words does not prevent grammarians from classing kī vadan as a samāsa, and conversely in Sanskrit texts printed in Sinhalese script kathita vacanam would often be thus printed as two words. Stem-forms (forms with a zero-inflexion) often stand as complete words in Sinhalese, in contexts where they cannot be considered part of a samāsa.
15 Ed. D. B. Jayatilaka, 1932.
16 In Studien, 33, Geiger appears to posit two separate origins for -na forms: ‘dadurch dass der Typ vasana = alt vasana mit dem Pprs. vasana = alt vasana lautlich zusammengefallen ist, entsteht manche etymologische Unsicherheit’. This does not accord with the account of the participle given in his Grammar, § 137, and referred to in Studīen, 29, but seems rather to revert to the account given in §§ 55 and 57 of the Grammatik of 1901.
17 For Amāvatura, I give only a page and line reference, according to the edition of Richard de Silva Mahānāma, 1912–15. For Dharmapradīpikāva (DhPr.), references are to page and line of the 1951 reprint of the edition of Dharmakīrti Śrī Dharmārāma.
18 Ed. D. B. Jayatilaka, 1924.
19 Here later writers would use a causative verb-form.
20 As also in modern Sinhalese. SidS, vii.36, calls this a passive imperative, which leads Paranavitana (op. cit., p. cxliv) to connect it, implausibly, with MI forms in -neyya.
21 Saňdahā is occasionally preceded by a relator: see below.
22 The commonest equivalent of the Sanskrit infinitive.
23 Corresponding to the Tamil periphrasis with patu. See Geiger, § 155. It is not in fact always passive even in Guruḷugōmi. Laduyēya in the quoted example is a tense form from laba-.
24 In later prose the contexts where the ending -na occurs here can be specified, but this is not so at the earlier period.
25 ‘Éventuellement eṉu’, JA, CCXXXVIII, 2, 1950, 188Google Scholar.
26 ‘Verbe sans racine’, ibid.
27 In modern speech this difficulty hardly exists, since (1) the subject of the relator does not appear in speech in the oblique case but in the direct; and (2) the oblique case has an optional distinctive termination.
28 I treat this example as parallel to keḷavaraṭa takāgat ‘regarded as a last resort’, 26.16, lōkayan gat pariddaṭa koṭa ‘according to the world's beliefs’, 53.4; but the kāraka relationship is disputable.
29 And sometimes other words, e.g. unu: ekak'hu unu dahasak minisun ‘a thousand men less one (a-one lacking a-thousand men)’, 122.30.
20 For examples of these, see 78.32, 170.19, 20.25; Dhpg., 66.29; 92.11, 54.10, 30.26.
31 The -ṭa in this word is not a case-termination; see Kumāraṇatunga's, edn. of Amāvatura, 5–8 (1923), p. 13Google Scholar.
32 In later literature also tek, piṇisa, sēk, sulu.
33 This use of tāk is very uncommon in Guruḷugömi. Tek, which is used instead of tāk in such contexts in later literature, is no t attested in Guruḷugōmi, whose normal usage here would be mā noena tāk kalhi.
34 Saňdahā is more normally preceded by an abstractor. The suffixed -ya turns a clause into a complete sentence.
35 The suffixed -ya turns a clause into a complete sentence.
36 See Sīgiri graffiti, introduction, § 505.
37 See below, p. 149.
38 There is no Sinhalese equivalent to the Skt. pronominal yaḥ, though there is an adverbial particle yam.
39 Origin and development of the Bengali language, Calcutta, 1926, I, 172–7Google Scholar.
40 See Bloch, J., BSL, XXXIII, 1932, 299Google Scholar, for an occurrence in Marathi; Mr. T. W. Clark has given me similar usages in Nepali.
41 The (lost) pre-Buddhaghosa Sinhalese commentary on the Tripiṭaka.
42 See Hendriksen, , Syntax of the infinite verb-forms of Pali, Copenhagen, 1944, 146 sqqGoogle Scholar. Even in Guruḷugōmi we occasionally find a genitive case-form instead of the kartṛ oblique form, e.g. Brahmapajāpatiyagē pā devi diyen ‘with the water with which Brahmapajāpati has washed her feet (of-B. feet washed with-water)’. Here the Heḷaṭuvā may have been something like Brahmapajāpatiya pā devi diyen, put into Pali, as Brahmapajāpatiyā pādadhovanena (Papañcasūdanī (PTS), III, 78)Google Scholar, and then put back into Sinhalese by Guruḷugōmi with a genitival form Brahmapajāpatiyagē;. Two lines further on, however, we find äya pā devi diya ‘the water with which she washed her feet’, where äya is an oblique case-form.
43 And perhaps should always be; the texts are not reliable on such points.
44 SidS, vii.4, gives as the endings for the future tense, 3rd sing., -annē, -ennē, -onnē, -nē, as opposed to the endings in the present tense, which appear as -ī, -ō, -ē, -ā.
45 See, for instance, Sīgiri graffiti, introduction, §§ 441–450, an d also Geiger, , Grammar, § 149 (2)Google Scholar. This is in fact also recognized by SidS in vii.19 (kal no-äraba).
46 Thus the formation of these tense-forms from the present and past relators is comparable to the formation of the periphrastic future in Skt.
47 This exceptional form, which looks like a present nominator form, is in fact an irregular past form.
48 Paranavitana (op. cit., p. cxliv) ingeniously connects such instances with MI forms in -neyya. It is true that the form in the works of Guruḷugōmi seems always to appear as -nnēya (in 152.19 we should read dakvannēyayi, as in Jayatilaka's, H.editīo princeps of 1887, instead of dalcvanneyiGoogle Scholar); but the -ya is the terminating particle which makes the participial phrase a complete sentence. Paranavitana's further equation of the -nu; forms in certain contexts with MI -neyya forms is certainly unacceptable.
49 As the case-ending -gen is not found in prose of this date, no examples are given here of this inflexion. The printed texts of Amāvatura contain one instance, Bōsataṇangen, 13.1; this should be emended to Bōsatāṇangē.
50 Geiger, , Grammar, § 117Google Scholar.
51 In its context, this is a question: ‘is it you …?’
52 On p. xl, however, Paranavitana writes: ‘-ā, the termination of the nom. s., is a development of the earlier -e through *-aya. … Verbal forms such as yana-vā (he goes) … go back to an earlier yana ve (is one who goes)’. (Yana ve seems to be unattested.)
53 In poetry, inflexional endings are generally omitted. Furthermore, Sinhalese prose is attested earlier than poetry (except for the problematical Sīgiri graffiti).
54 But not before van, badu ‘like’; see below.
55 The oblique case of animates often replaces the genitive.
56 See Wijayaratne, D. J., History of the Sinhalese noun, Colombo, Univ. of Ceylon, 1956, ch. iiiGoogle Scholar. Geiger has not appreciated the position here.
57 Günther's, excellent article ‘The conditional mood in Sinhalese’, JAOS, LXIX, 2, 1949Google Scholar, seems to be over-unwilling to allow this past conditional, which occurs already in the eleventh-century Mihintale Tablets (Epigraphia Zeylanica, I, p. 85, 1. 7) tubuva vaḥi, and before yutu throughout the inscription.
58 See Sīgiri graffiti, introduction, §§ 524–526, where this is called bhāve prayoga. This is, I believe, the correct explanation of the hav-piyō of SidS, vii.3 (although Kumāranatunga's edn., of which the commentary is incomplete, prints here a text which differs from the traditional version); the Tamil Vīracōḷiyam uses bhāve prayoga in yet another sense, differing from the usage of SidS as much as from Skt. usage, see Sastri, P. S. Subrahmanya, History of grammatical theories in Tamil, Madras, Oriental Research Inst., 1934, 144Google Scholar.
59 This inverted form becomes standard for interrogations in later language.
60 Mänava may also follow an abstractor: avavāda karanu mänava ‘please give advice’, 167.10. The past-tense nominator does not occur in this context.
61 This form, present tense, with ending -nuvā, is probably responsible for the modern spoken verb-form in -navā. The forms in -navā have been the usual dictionary forms of verbs, although they do not appear in old literature at all, and not often in modern literature.
62 In later prose this is typified with the past nominator and the particle ma (Gunasekara, , p. 189, n.)Google Scholar.
63 e.g. Gunasekara, § 289: ‘The verb vū, to become, takes two nominatives, of which the first is the subject and the second the subjective complement of vū’. But cf. for example: tō osap'hu vehida? ‘are you in your courses (utusamayo) ?’, 67.1.
64 Before the participial vū, this nominator-form continues in modern literary usage, e.g. āvā-vū, = ā, past relator of e- ‘come’.
65 Geiger explicitly quotes (Studien, 29) liyana as the stem-form of liyannā ‘one who writes’; and this agrees with Gunasekara, § 118 (10). But phonologically speaking it is clear that the base of all the inflected -nnā forms is a form ending in -nu, since -nnā arises from -nuvā (see Geiger, , Grammar, § 97 (1))Google Scholar.
66 It is probably from contexts like these that the ‘future tense’ of SidS derives, since it is only in the 3rd person plural that the personal terminations unmistakably differ from the requisite forms of ve-.
67 cf. with the bhāvahriyā: äya … miyuru sarayen kiyanu äsi ‘he heard her singing sweetly (her … sweet with-voice singing he-heard)’, DhPr., 228.18.
68 Instrumental case-form only.
69 In later literature similarly viṇa ‘time’, pasu ‘after’, atara ‘between’, häṭi, lesa, andama ‘manner’, taram ‘amount’, etc.
70 For the genitive, see above, p. 138, n. 42.
71 Heyin ‘by reason’, which is often found in later language following a relator in this way, is not attested in this usage at this period.
72 This use with van is common in Guruḷugōmi, but appears to be unknown in later literature. The different usage before väni and sē (in similes) is noted above.