¹ This is one of the main criticisms that may be levelled at the pioneering attempt of Singh, Sāhib, Gurbāṇī viākaran, third ed., Amritsar, 1966Google Scholar, as opposed to the useful grammatical notes prefaced to the Śabdārath Srī Gurū Granth Sāhib jī, fourth ed., Amritsar, 1969Google Scholar, whose standard pagination of the AG is used in references here.

² Grierson, G. A., Linguistic survey of India, IX, pt. I, Calcutta, 1916, map facing p. 637.Google Scholar

³ The best general outline of this in English is Kohli, S. S., A critical study of Adi Granth, New Delhi, 1961, 27–64.Google Scholar

⁴ The origin of the term ‘Sirāikī’ and the background to its recent general currency locally is explained in my article ‘Siraiki: a language movement of Pakistan’, Modern Asian Studies, xi, 1, 1977Google Scholar. For a detailed description of the modern language the reader is referred to Shackle, C., The Siraiki language of central Pakistan: a reference grammar, London, 1976.Google Scholar

⁵ In fact only 112 śalok are included in the sample, since those numbered 13, 32, 52, 75, 82–3, 104–5, 108–11, 113 and 120–4 are editorial interpolations from the Sikh Gurus. No attempt is made here to enter into the vexed question of the identity of the Farīd of the AG, which is discussed at length in Mohan Singh Dīvāna, ‘Bābā Farīd Ganj-i Śakar, Śai Ibrāhīm aur Farīd ānī’ (in five parts), Oriental College Magazine (Urdu text), XIV–XV, 1938–1939Google Scholar, in passing in Nizami, K. A., The life and times of Shaikh Farid-u'd-Din Ganj-i-Shakar, Aligarh, 1959, 121–2Google Scholar, and during the course of general appreciations in Singh, Dīvān, Bābā Farīd darśan, Amritsar, 1951Google Scholar, Bedī, Kālā Singh, Farīd te Farīd-bāṇī, New Delhi, 1974Google Scholar, and Bhogal, Piārā Singh, Mahkavī Farīd, Jullundur, 1970Google Scholar. Whatever the authorship of the poems, it is clear that the prominence of South-Western elements must be connected, not necessarily directly, with the original Farīd's particular association with Multan.

⁶ The first members of each pair are written in Sahaskritī, or ‘AG Apabhramśa’. The stanzas themselves, like those of Mārū vār cited below, are in the more usual Western variety of the AG: collectively, therefore, these are excellent examples of deliberate linguistic variation within a single composition.

⁷ The heading is omitted from AM 23. Although South-Western elements are less prominent in this set of śalok, the omission is perhaps due to editorial oversight. AJ 1, though not of course headed, is also less South-Western than the succeeding śalok.

⁸ cf. the definition in Singh, Kānh, Mahān koś, second ed., Patiala, 1960Google Scholar. Trumpp, , in his notoriously inaccurate The Ādi Granth, London, 1877, cxxxiv–cxxxvGoogle Scholar, took ḍakhṇā as a purely metrical description (indistinguishable from the śalok!), but scans his example wrongly.

⁹ Also known as the Ādi sākhī . The edition used is that of Singh, Piār, second ed., Ludhiana, 1972, and references are to pp. 145–50.Google Scholar

¹⁰ Soḍhī, Miharbān jī, Janam sākhī Srī Gurū Nānak Dev jī, ed. Singh, Kirpāl, I, Amritsar, 1962: references are to the group of gośṭi on pp. 488–508.Google Scholar

¹¹ MJ 423–42Google Scholar, a set which has no representatives in other Janamsākhīs.

¹² For instance, the gośṭi printed in MJ 426–30Google Scholar has a high concentration of South-Western elements, entirely appropriate to its setting in Multān, but would seem to have been developed from the style of the opening verses (only) of the hymn, Āsā kāfī, mah. 1, ghar 8, aśṭapadīā (pp. 418–19)Google Scholar, on which it is a commentary.

¹³ The Janamsākhīs as a whole, and the MS sources, are discussed in McLeod, W. H., Gurū Nānak and the Sikh religion, Oxford, 1968, 8–33.Google Scholar

¹⁴ But -a is omitted in transliterations of JS words: both the language and the orthography of the JS are closer to modern usage than in the AG.

¹⁵ The relative preference for d- in F would be still more evident if such nouns (without parallels in ḍ-) as dāti ‘gift’, dāṛī ‘beard’, or dīvṛe ‘lamps’ were included: but the point is that F has more forms in ḍ- than most AG texts, and A much more. In JS d- predominates, except in the gośṭi, MJ 426–30Google Scholar, already quoted as markedly South-Western.

¹⁶ In JS, -ḍ- is confined to saḍu MJ 437 (3) versus sadu MJ 437 (1), all with the sense of śabad.

¹⁷ The special case of the sg. possessive personal pronouns is considered below in section 4.

¹⁸ The contrast between Sir. truṭ- and Pj. ṭuṭ- is thus only partially represented. Intermediate forms must also be reckoned with: thus tikhī ‘sharp’ (FS 1.4) corresponds exactly with modern Sir., which has tikhā [with Pj.], not trikhā.

¹⁹ It would anyway be pointless to pursue any investigation of this, given the archaizing tendencies of AG orthography.

²⁰ Recently at least, the failure to distinguish -ṛ- from -r- in Sir. has come to be regarded as a characteristic of (archaic ?) Hindu speech. Given their immediate context, loraü ‘I long for’ and maroraü ‘I twist’ (FS 1.1) should perhaps be explained as Easternisms [for usual loṛ-, maror-]: but bhore ‘simple’ (AM 5.3 bis) [for usual AG bholā] is probably a unique Sindhism.

²¹ It is described as ‘Lahindī’ in the grammatical notes in Śabd rath, p. viiGoogle Scholar: other examples from the AG are given in Singh, Sāhib, op. cit., 136.Google Scholar

²² But hiṅṅu dī ‘of assafoetida’ (F 33), and several instances of sg. obl. (or loc.) without postposition in -u: visu, khaṇḍu (both F 37), jindu (F 47).

²³ jhālū is traditionally described as ‘Sindhi’ in the commentaries [on the basis of Sindhi ojhal ‘light, dawn’ ?]. The first ghaṭāū might be taken as an instance of the typical Sindhi fem. pl. dir. -, while the second could also be construed as an abl. sg. [cf. infra].

²⁴ The list excludes such unremarkable forms as the verbal adjective nivāhū ‘keeping faith’ (AM 14.1) or derived nouns like pandJiāṇū ‘traveller’ (AM 13.3).

²⁵ The -ḍ- in A is in typical contrast with the -d- of F. There are several examples of adverbs in -ū in MJ, including kithāū ‘somewhere’ (439) and udū udū valahu ‘from each side’ (490).

²⁶ But the postposition kanno ‘from’ (F 112) is noteworthy [contrast Sir. kan]: kano, kanno also appears in MJ, and as kaṇo (JP 146).Google Scholar

²⁷ mākhio ‘honey’ (F 27) has the final vowel pointed as both -o and -u by an exceptional convention of AG spelling, and mākhiu is in fact demanded by the metre: sg. obl. mākhia appears in F 89.

²⁸ Against these large figures there are almost no instances in JS, confirming that this is a specifically poetic convention.

²⁹ The incidence and meaning of -r- forms in the poetry of ẉāja Farīd is examined in detail in Shackle, C., Sirāikī and Sirāikī literature, c. 1750–1900, in upper Sind and south-west Panjab (unpublished Ph.D. thesis, University of London, 1972, 236–41).Google Scholar

³⁰ There are several instances of maiḍā, taidā in the JS, sometimes pointed as maĩḍa, taĩḍā: but it is of course impossible to decide that the latter indicate not so much Sir. maida as the maĩda of Shahpuri (Shp.), intermediate between Sir. and Pj. Shahpuri is described in Wilson, J., Grammar and dictionary of Western Panjabi, Lahore, 1899, repr. Patiala, 1962.Google Scholar

³¹ Similar expansions, often as rhyme-forms, occur in later Sir. poetic language as -āh, as described in Shackle, , Sirāikī and Sirāikī literature, 110–13.Google Scholar

³² But habh is now distinctively Shp., in contrast with sabh, shared by Pj., Sir., and Sindhi.

³³ hik- is in free variation with ik-, ek- in MJ and JP, and there are several instances of habh- in MJ, including the sg. obl. habhas, as in habhas kahī kū (437), and habhas kahī dā ‘of everyone’ (506)—actually as a genitive!

³⁴ Invariable -ī forms and declined -ā forms are used side by side in MJ, as in je anek jetī sād haī aru ete hī karai aru jetī sīgār haini ete hī karai (436).

³⁵ Sir. now has 2 sg. past hāv as standard, but āh is found in older poetry and the archaic Sir. of Sind: the āh- past is still used in Shp., and instances are found in MJ, e.g. 3rd sg. mase. āhā (494), fem, āhī (491). The distinctively South-Western negative present is found in Nāuak, 's hymn (referred to above, p. 38, n. 12)Google Scholar, as nīmhī, nīmhi ‘I have not’ [Sir. nimh].

³⁶ Occasional instances of inf. and ger. with -n- are to be explained as Hindi forms: thus pekhana sunana sunāvano mana mahi driṛīai sācu (AJ 1.2), bandhana kī vela, (FS 1.2), dekhana, kī āsa (F 91).

³⁷ But JS regularly have thdā [= Sir.], etc.

³⁸ But loṛīdo ‘I long for’ (AM 12.1) is active, suggesting the normal -do of Sindhi (with trans, verbs): lajīvado (AM 19.3) is also active, with the sense of ‘shaming’, and seems to be a similar form, with infixed -v-.

³⁹ As in mākhio (cf. p. 41, n. 27), the weakening is shown by double pointing of the first syllable as both lo- and lu-.

⁴⁰ As elsewhere in the AG, the extensive use of absolute constructions and omission of postpositions sometimes makes precise grammatical definition difficult: this is particularly the case with past participles where the adverbial in -lā [modern -i] is graphically identical with mase, sg. -lā.

⁴¹ Similarly, the only pres. part, in -o in F: jīvadro muiohi ‘you are dead while alive’ (F 107). JS too only use -o as adverbial: usu murīd kagai kardo … tithāū āiā ‘sending on that disciple ahead, he came there’ (MJ 489Google Scholar), ohu lakṛīā laido pīr kol āiā ‘taking the wood, he came before the pīr’ (JP 146).Google Scholar

⁴² JS also have a few examples of past part, in -o which cannot be construed as having 2 sg. suffix, e.g. es n, sāhibu juānī vici na milio ‘her lord did not come to her when she was young’ (MJ 491).Google Scholar

⁴³ of. ḍhaṇḍholimu ḍhūḍhimu ḍiṭhumai nānaka jagu dhūe kā dhavalharu ‘I have sought and searched and seen the world to be a palace of smoke’ (Mājh kī vār, mah. 1, 1, 3Google Scholar, quoted in Singh, Sāhib, op. cit., 233Google Scholar), where the preceding suffixed forms suggest writing ḍiṭhumai as one word rather than two: but the modern convention of word-division in the AG often prefers to write two words where one might seem more appropriate, as in many nominal compounds.

⁴⁴ The (indispensable) commentary of Singh, Sāhib, Srī Gurū Granth Sāhib darpaṇ, second ed., Jullundur, c. 1970, x, 326Google Scholar, takes this as 3 sg., but MJ 491Google Scholar, reading vaḍī thī gaīāsu, takes it as 1 sg., admittedly helped by an easier reading in the next line.

⁴⁵ JS also have several instances of 2 sg. -o, e.g. kujhū samdho ihdā biānu ‘did you understand anything of his explanation?’ (MJ 489Google Scholar) = kihu samjhio ihdā biānu (JP 146).Google Scholar

⁴⁶ There are, however, a great many instances in JS, usually in -osu [Sir. -us], e.g. arju kītosu ‘he said’ (MJ 488Google Scholar): there are occasional instances of agreement with the object, i.e. fem. sg. -īasu, as in sojhī ditīasu ‘he has given awareness’ (MJ 492Google Scholar), and mase. pl. -iasu, as in kadmnū hathu lāiasu ‘he touched his feet’ (MJ 492Google Scholar). A similar pattern of concord with 3 sg. suffix is established for other parts of the AG in Singh, Sāhib, Gurbāṇī viākaran, 236–42Google Scholar, but with the agentive suffix -nu: this is used only rarely in MJ as in pahilā sunnu thī āpṇā nāũ kītonu ‘first He made his name from the void’ (MJ 496Google Scholar), and trihu loa āsaṇu badhā nu ‘He made the three worlds his seat’ (MJ 497Google Scholar). But -n- is now reserved for 3 pl.

⁴⁷ i.e. with the usual modern transfer of accent to the first syllable of a disyllabic suffix, as in the pres. part, in -dā. 1 pl. thiāse (MJ 490, 494Google Scholar) corresponds to modern thiose ‘we became’.

⁴⁸ The interesting conditional in nāhī ta hambhī dajhāhi ‘otherwise I too should have been burnt’ (F 3) is surely a South-Western construction, corresponding to Sir. karhā ‘I should have done’.

⁴⁹ The JS forms of the 1 pl., as in karh. (MJ 441Google Scholar), karhe (JP 147Google Scholar), suggest Shp. -h, versus Pj. -īe or Sir. -.

⁵⁰ kīce is now, however, replaced by the regular kanīje in Sir. 3 pl. kīcani occurs on MJ 493.Google Scholar

⁵¹ Presumably the AG forms have nasalization of the vowel. Sir. of Bahawalpur has 2 sg. -s, versus -s elsewhere.

⁵² As in the pres. (p. 47, n. 49), JS forms of the 1 pl. suggest Shp. -sh, accented finally, unlike other future terminations: cf. akhshe (MJ 490).Google Scholar

⁵³ The F examples are absolute vati, used as an adverb ‘again, then’, like the characteristic shibboleth of Sir. vat versus Pj. phir, fer: vati is used in JS to give a South-Western flavour to passages of dialogue, as in is rabābī khukmu thīvai ji maī kvati bait suṇāe, ‘let his minstrel be told to recite the couplet to me again’ (JP 146Google Scholar), or to the narration, as in tvati onī alāiā ‘and then they said’ (MJ 427).Google Scholar

⁵⁴ The JS make quite frequent use of the South-Western ghinn- ‘take’, as in maī kghinni julu uthāū ‘take me there’ (MJ 489Google Scholar). But the AG texts have only lai-, le- [F (6)].

⁵⁵ But there are no instances of kaü, the commonest AG word.

⁵⁶ cf. p. 41, n. 26, above, for kanno ‘from’.

⁵⁷ gāl- is also used, in alternation with gal-, in MJ: se evaĩ gālī kurde, haini sāhib dī gāli kāī nāhkarde ḍasde, hori gattkarde haini ji jāṇde āhai (MJ 428Google Scholar). The deliberate use of South-Western nouns is rare in JS, though meu ‘boatman’ (MJ 433Google Scholar) and māṇū ‘man’ (MJ 488Google Scholar = JP 148) could be cited.Google Scholar