p. 287 It is not less than ten years that some writers in Bengal have begun to use quite a new word, sandhya-bhasa, or 'twilight language, ' as it is translated by them into English. It was, however, first introduced by Mahamahopadhyaya Haraprasad Shastri (-HPS) in the introduction to his Bauddha Gana O Doha.(1) This volume contains four Buddhist works, viz. (1) Caryacaryaviniscaya (an anthology in the oldest Bengali), (2) Dohakosa of Sarojavajra, (3) Dohakosa of Krsnacarya or Kanhapada, and(4) Dakarnave, each of the first three being accompanied with a commentary in Sanskrit. They are all edited by him. In the commentaries on the Caryacaryaviniscaya, and Sarojavajra's Dohakosa we come across the following words: (1) sandhya, pp.6, 11, 29, 32 ; (2) sandhya-bhasa, pp. 5, 13, 16,18, 19, 23, 24, 26, 30, 51, 83, 93 ;(2) (3) sandhya-vacana, p. 37 ; and (4) sandhya-sariketa, p. 9. These are used as synonymous and with reference to them HPS writes in Bengali in the introduction (p.8) referred to above: "All the works of the Sahajayana are written in the Sandhya-bhasa. Sandhyabhasa is a bhasa (language) of light and darkness ("alo-a dhar1"), partly light, partly darkness; some parts can be understood while others cannot. In other words, in these discourses on dharma which are of a high order there are references also to different things. This is not to be openly explained." Let us see how far this view can be accepted. Mr. Panchkawri Banerjee does not subscribe to it and offers his new interpretation saying that Sandhya-bhasa is the bhasa or language of the country known by the name of Sandhya, i.e., the border land between the old Aryavarta and Bengal proper.(3) This is, in my ---------------------- 1. Vangiya Sahitya-Parisat, series No.55, Calcutta, 1323 B.S. - 1916 A.D. 2. Once the reading is bhasa. 3. "Pandit Haraprasad Shastri came to the conclusion that the language used by the Siddhacaryas was called Sandhya because it p. 288 opinion, mere imagination, there being absolutely nothing to support it. The following words which occur in the Saddharmapundarika (Bib. Budh., 1912) deserve to be noted in order to find out if there is any connection between these two sets of expressions: (1) sandha-bhasita, pp. 125, 199, 233 ; (2) sandha-bhasya, pp. 29, 34, 60, 70, 273 ; and (3) sandha-vacana, p. 59. In fact, these are synonyms. But what is the sense in which they are used ? Burnouf in his French translation of the work, Le Lotus de la Bonne Loi(Paris, 1852, p. 342) has, perhaps for the first time, discussed the word used there (p.29) in the following sentence: " durvijneyam sariputra sandha-bhasyam tathagatanam."--'O Sariputra, the sandhabhasya of the Tathagatas is very difficult.' His conclusion is that the word means 'enigmatical talk' (''le language enigmatique"), as supported by the Tibetan version reading ldem por dgons te bsad pa ni which, according to him, means 'explanation of the thought expressed enigmatically' ("le'xplication de la pensee exprimee enigmatiquement "). Kern in his English translation of the work (SBE, vol. XXI) has translated the term throughout by 'mystery' (see p.59, note3). Max Muller takes the term to mean 'hidden saying' on Chinese authority (SBE, vol. XLIX, p. 118; Vajracchedika, p.23, note 5). Let us, however, discuss the point once again. ______________________ was a kind of twilight language which sought to give a mere glimpse of the high truths of Buddhism, not in their pure original form, but in such modified shape as could be understood by the common people leaving deliberately vague what was not deemed safe or useful for them to worry about. With this conclusion 1 cannot agree. The tract to the S.E. of Bhagalpur comprising the western portion of Birbhum and Santhal Perganas is the borderland between the old Aryavarta (the Indian domicile of the Aryans) and Bengal proper, and was called Sandhya country. Any one, who is familiar with the several dialects all closely resembling one another spoken in that region, cannot have any doubt as to their near relationship to the language used by the Siddhacaryas." Visvabharati Quarterly, 1924, p. 265. p. 289 That sandha of sandha-bhasya, sandha-bhasita, etc. is in reality a shortened form of sandhaya, a gerund from sam+ dha is beyond doubt. This is suggested even from the simple fact that on p.70 of the printed text of the Saddharmapundarika two of the MSS. used by Kern, K and W, actually read sandhaya for sandha in the word sandha-bhasya.(1) Though the former cannot claim to be the actual reading on the metrical ground, we cannot discard it altogether, for it clearly indicates the sense of the word in which it was taken by the scribes of the two MSS. referred to above. That the original form of sandha cannot be other than sandhaya is shown below. But in that case the only question that presents itself is with regard to the dropping of ya of sandhaya. This is, however, easily solved. See the following Pali forms: anna < annaya (Skt. ajnaya), Dhammapada, 56;abhinna < abhinnaya (Skt. abhijnaya), Sumangalavilasini, pp.173 313; upada < upadaya, Dhammasamgani, 877, 960. See Geiger: Pali Literature und Sprache, 1916, 27'2 (aya = a). The import of sandhaya may be expressed in Sanskrit by the words such as abhisandhaya, abhipretya, uddisya, etc. which can be translated into English by 'meaning,' 'aiming at,' 'having in view,' 'intending,' 'with regard to,' etc. In support of this the following passages, both Sanskrit and Pali, may be quoted here: (1) punar api mahamatir aha / yad idam uktam bhagavata yam ca ratrim tathagato' bhisambuddho yam ca ratrim parinirvasyati atrantara ekam apy aksaram tathagatena nodahrtam na pratyaharisyti/ avacanam buddhavacanam iti/ tat kim idam(2) sandhayoktam tathagatenarhata samyaksambuddhenavacanam buddhavacanam iti/ bhagavan aha/ dharmadvayam mahamate sandhaya mayedam uktam/ katamad dharmadvayam yad uta pratyatmadharmatam ca sandhaya pauranasthitidharmatam ca/ idam mahamate dharmadvayam sandhaya mayedam uktam.-Lankavatara, ed. B. Nanjio, Kyoto, 1923, p. 143. 'Mahamati asked again: "It is said by the Blessed One that between two nights, one on which the Tathagata attained to perfect enlightenment and the other on which he would attain complete nirvana, he did not utter even a single syllable; nor would he utter it. The utterance of the Buddha is non-utternance. Meaning what has the perfectly enlightened, venerable, Tathagata sdid that ---------------------- 1. Sandha-bhasyena bhasanto buddhabodhimm anuttamam. 2. In the Vajracchedika, p.23, note 5, read idam with the Ms J, and not iyam as suggested by Max Muller. p. 290 the utterance of the Buddha is non-utterance? " Replied the Blessed One: "Meaning two dharmas, O Mahamati, I said it. And what are these two dharmas? They are pratyatmadharmata and pauranasthitidharmata. These are two dharmas, O Mahamati, meaning which I have said it.' (2) caturvidham samatam sandhaya mahamate tathagata vacam niscarayanti / katamam caturvidham samatam sandhaya / imam mahamate caturvidham samatam sandhaya tathagata vacam niscarayanti/ Op. cit., p. 141. 'Meaning four-fold equality, O Mahamati, the Tathagatas utter their words. Meaning what four-fold equality? Meaning this four-fold equality, O Mahamati, the Tathagatas utter their words.' (3) anutpattim sandhaya mahamate sarvadharma nihsvabhavah / 'It is meaning their non-origination, O Mahamati, that all the things are (said) to be without their nature.' (4) Yah sandhayaham evam vadami. -- Dasabhumakasastra, ed. T. Radher, p.5. 'Intending these (Bodhisattvabhumis) I say so.' (5) Idam nu te etam magandiya sandhaya bhasitam bhunahu samano gotamo. -- Majjhima Nikaya, I, p.503. 'Intending this, O Magandiya, it is said to you that the recluse Gotama is destroyer of beings.' (6) yam sandhaya vuttam.--Jataka, 1, 203. 'Meaning which (the following) is said.' (7) Idam kira bodhisatto attano abbhantare nanavudham sandhaya kathesi.--Op. cit., p. 274. 'This was said by the Bodhisattva meaning the weapon of knowledge which was within him.'(1) ---------------------- 1. The following may also be referred to : Sumangala- vilasini, pp. 163: (i) r a j o d h a t u o' ti rajaokkinntthanani hatthapithapadapithadini sandhaya vadati. (ii) s a t t a s a n n i g a b b h a' ti otthago- nagadrabhaajapasumigamahise Sandhaya vadati. (iii) a s a n n i g a b b h a' ti saliyavagodhu- mamuggakanguvarakakudrusake sandhaya vadati. (iv) n i g g a n t h i g a b b h a' ti velunala- dayo sandhaya vadati. (See also pp. 161, 165. Kathavatthupakarana-Atthaka- tha, JPTS, 1889), p.3: annam sandhaya bhanitam. p. 291 From the above uses it is perfectly clear that so far as the sense is concerned, the word sandhaya of the Buddhists is nothing but abhisandhaya found in Brahmanic and Buddhist works as the following passages will bear out: (1) abhisandhaya tu phalam. Bhagavadgita, XVII, 12. 'Having in view(1) the consequence'. (2) abhisandhaya yo himsam. Bhagavata Purana, III, 29, 8(2). 'Who having in view injury.' (3) visayan abhisandhaya. Op. cit., III, 29. 9. 'Having in view the object of senses.' (4) labham abhisandhaya. Bodhicaryavatarapanjika, p.214. 'Having in view gain.' (5) isvaram eva abhisandhaya. Op. cit., p.554. 'Just with reference to God.' With regard to this significance of sandhaya we have support also from the Tibetan sources. The following line is quoted from the Lankavatara in the Madhyamakavrtti (p. 555) by Candrakirtti: svabhavanutpattim sandhaya mahamate sarvadharmah sunya iti desitah. 'Intending the natural non-origination, O Mahamati, it is taught by me that all things are void.' Here sandhaya is translated into Tibetan by dgons nas which simply means 'having in view' (abhipretya, abhisandhaya, uddisya). We have already seen, as Burnouf has quoted, that the Tibetan expression for sandha-bhasya in the Saddharmapundarika (p.29) is ldem por dgons te bsad pa ni. Here ldem por dgons means nothing but abhisandhi (or abhipraya) though each of these two expressions, ldem po and dgons or dgons pa has the same sense. For example, gzug pa la ldem por dgons pa = avataranabhisandhi; mtshan nid la ldem por dgons pa=laksanabhisandhi; and so on. With te added to it ldem por dgons means here only sandhaya =abhisandhaya. And bsad pa means bhasita, bhasana, etc. 'speech,' 'talk,' 'explanation,' etc. The same (ldem por dgons tebsad or gsuns sometimes omitting te) ---------------------- 1. The commentators, Sankara, Nilakantha, Dhanapati, Sridhara, Abhinavagupta and Madhusudana explain the word abhisandhaya by uddisya. See also Gopicandana Upanisad, 5:vedartham abhisan- dhaya. 2. Sridhara here says that abhisandhaya means sanka- lpya. p. 292 or similar (ldem por nag) Tibetan expression in the sense of sandha is found in other cases in the Sadharmapundarika (see pp. 60, 199, 233, 59). Now, the Sanskrit and Pali passages quoted above will show us that sandhaya is used with a verb, express or understood, meaning 'to say'(û vad, û kath, û dis etc.). We further see that in some of these passages sandhaya is employed with past participles, e.g., sandhaya desitam, sandhaya bhasitam, sandhaya vuttam. In such cases these pairs of words are not compounded, but gradually with the change of the meaning of the second member they began to be compounded; in other words, the past participles lost their own sense, and assumed that of a verbal noun. Thus bhasita did not then mean 'said,' but 'saying,' 'utterance,' bhasa, bhasya, bhasana; similarly ukta, Pali vutta, began to be used in the sense of vacana 'speaking,' 'speech,' 'utterance'. Thus the word sandha-bhasita with its other synonyms already alluded to came into being. We have also the use of abhiprayika vaccana (or vacas) in that very sense. For example, we read in Vasubandhu's Vimsatika (Vijnaptimatratasiddhi), ed. Levi, Paris, 1925, p.5, Karika 8: rupadyayatanastitvam tad vineyajanam prati / abhiprayavasad uktam upapadukasattvavat// (1) Here runs the commentary on the word abhiprayavasad (Tib. dgons pahi dban gis): cittasantatyanucchedam ayatyam abhipretya (Tib. dgons nas) abhiprayikam tad vacanam (Tib. bkah de ni dgons pa can no). 'Having in view the non-annihilation of the continuity of citta in the future, the speech is intentional. The following line from the Tattvasamgraha (GOS, 1926, p. 868, sl. 3, 331) may also be cited here: abhiprayikam etesam syadva dadi vaco yadi.(2) ---------------------- 1. Poussin translates it (Museon, vol. VIII, 1912, p.74) : C'est intentionnellement (abhisandhi, abhiprayavasat) que Bhagavat a enseigne l'exis- tence des ayatanas aux hommes que cet [enseigne- ment] doit convertir; -- comme [il a enseigne l'existence] des "etres de naissance surna- turelle" (aupapaduka sattva). 2. Here is the Tib. version (Tanjur, Narthang, ed., Mdo He, fol. 133a of the Visvabharati Library): gal te de dag hgyur smra sogs/ tshig ni dgos pahi don yin na/ Literally it reads prayojanarthaka for abhiprayika. We should, p. 293 'If their utterance such as s y a d v a d a, etc. is intentional.' It is thus perfectly clear that sandha-bhasita or bhasya is nothing but abhiprayika vacana which can be translated by 'intentional speech.' We can gather also from the Chinese sources(1) that in reality sandha-bhasya is abhiprayika vacana ('intentional speech'). In Chinese there are three translations of the Saddharmapundarika, first by Fahu or Dharmaraksa (286 A.C.), second by Kumarajiva (406 A.C.), and the third by Jnanagupta and Dharmagupta (601 A.C.); besides other two translations which do not contain the passage bearing on sandhabhasya. It is found in these translations that only on two occasions (Saddharma-pundarika, p. 34=Kumarajiva's tr., Tokio ed. xi. 1. 11a. 15-16; and Sadharmapundarika, p. 233=Kumarajiva's tr., xi. 30b. 2-3) the word (sandha-bhasya or bhasita) is translated, in the first case by wei and in the second by yu. In other passages it has either been omitted or translated by an expression, sui i. Now, both wei and yu mean 'subtle,' 'obscure,' 'secret,' 'hidden'; and this is, I think, just the opposite of nitartha which is explained in the Abhidharma- kosavyakhya, as we shall presently see, by vibhaktartha 'of clear sense'. In other words, these two Chinese words may be explained in such cases by avibhaktartha 'that of which the sense is not clear,' 'subtle,' 'obscure' =vibhajyartha, 'that of which the sense is to be made clear'=neyartha (see below). With regard to the second exppression sui i, as explained by Couvreur, it means 'suitable' ("d'apres ce qui convient). If we, however,consider how the Sanskirt term under discussion (sandhabhasya) is explained in Tibetan, as has already been shown, it seems to me, we may take the Chinese expression as equivalent of abhiprayika, 'intentional.' Chinese sui means 'to follow' (anu+ ûgam) and is frequently used for the Sanskrit prefix anu (for instance, sui tso, anuvidhana; sui sheng, anujata). And i means (i) 'suitable,''proper'; (ii) 'should,' 'ought.' Thus sui i may mean ______________________ however, like to read dgons for tgos in the second line, and in that case dgons pahi don would mean abhiprayarthaka which is in fact the same as abhiprayika. 1. I am grateful to Dr. Prabodh Chandra Bagchi for the help I have derived from him on this point. p. 294 'anusaraniya 'that which is to be followed,' and this is, in fact, abhipreya (=gamya =abhiprayika, from ûi 'to go'), figuratively 'that which is to be intended.' Now, the beauty of the instruction (desanavilasa) of the Buddhas, or their skill in showing the means for realization of truth (upayakausalya) is that their instructions (desana) differ according to the degrees of fitness of their disciples.(l) Those instructions are mainly of two kinds, (i) one, the object of which is to show the real state of things directly (tattvartha), and (ii) the other, 'intentional' (abhiprayiki) meaning thereby that it is intended to imply or suggest something different from what is expressed by the words (yatharuta).(2) The object of the former is to lead one to the path of nirvana (margavatara), while that of the latter is to lead one to the fruit (or final result) of nirvana (phalavatara). These two kinds of instruction or discourse or sutra as often styled are also called nitartha and neyartha respectively. By the former is meant the kind of instruction which is clearly expressed, and by the latter the kind which is implied. In other words, the first gives us the literal sense while the second the implied sense.(3) ---------------------- 1. desana lokanathanam sattvasayavasanugah/ bhidyante bahudha loka upayair bahubhih kila// Bodhicittavivarana quoted in the Sarvadar sanasamgraha, Anandasrama ed., 1906, p.11 durbodhyam capi taj jnanam sahasa srutva balisah/ kamksam kuryuh sudurmedhas tato bhrasta bhrameyu te// yatha visayu bhasami yasya yadrsakam balam/ anyamanyehi arthehi drstim kurvami ujjukam// Saddharmapundarika, p. 125. 2. See Madhyamakavrtti, pp. 42-43. 3. See the commentary of the Nettipakarana, PTS, p.218: N i t a t t h a n t i yatharutavasena natabbattham, N e y y a t t h a n t i niddharetva gahetabbattham. The Manorathapurani (Siamese ed.) on the Anguttara Nikaya, I.10, says: Yassa attho netabho tam neyyattham suttantam, Nitattho...... khathitattho. Abhidharmakosavyakhya quoted by Poussin in the Madhyamakavrtti p. 597: nitartha = vibhaktartha (Abhid. K V. 230b)...de sense clair; tandis que "neyarthasya sutrasya nanamukhapra- krtarthavibhago'niscitah sandehakaro bhavati." p. 295 It follows then from what is said above that the term sandha-bhasya with all its variants such as sandha-bhasita, etc., is synonymous with abhiprayika vacana and neyartha vacana or sutra. Let us now take up the question as to the origin of the words such as sandhya-bhasa, sandhya-vacana, etc. reading sandhya for sandha in HPS's Bauddha Gana O Doha already referred to. Burnouf says (Lotus, p. 343) in discussing the meaning of sandha-bhasya that he too found in his MSS. of the Saddharmapundarika the reading sandhya for sandha. It is, however, to be noted that in the edition of the Sanskrit text of the work by Kern and Nanjio not less than eight MSS. are used but in none of them is the word sandhya found even in a single instance. There is, in my opinion, sufficient ground for saying that though the materials collected by the editor were not sufficient, the present edition of the Baudha Gana O Doha is not as critical as it could have been if the materials that were at the disposal of the editor had fully been made use of. I do not, therefore, accept his readings. He says(1) that in the Asiatic Society of Bengal there is a copy (No. 8063) of the palmleaf manuscript on which he based his edition. In that copy there are a number of variant readings. Without comparing that copy with the original palmleaf MS. it cannot be ascertained whether the variants were in the original or made by the scribe of the copy. However, as regards the reading sandhya in the printed text, there is at least one case (p. 29, 1, 13) where the copy of the Asiatic Society of Bengal(p. 38) does not read sandhyaya as edited, but sandhaya.(2) With regard to the second work in which sandhya-bhasa occurs, viz. Advayavajra's commentary on the Dohakosa of Sarojavajra, the scribe himself tells us that the original from which he made his copy was a very corrupt one, yet with a view to collecting the work he copied it.(3) As there is no other MS of the work at our disposal we can only verify its readings by comparing it with its Tibetan translation. The following is found in the commentary of Sarojavajra's ---------------------- 1. A Descriptive Catalogue of Sanskrit Manuscripts in the Government collection under the care of the Asiatic Society of Bengal, vol.I, Buddhist Manus- cripts, 1917, P. 144. 2. The actual writing looks just like. That it is not sandhyaya is quite clear. 3. astavyastapado bhati grantho'yam lekhadosatah/ tathapi likyate'smhabhir. granthasamgrahakamksaya// p. 296 Dohakosa edited by HPS, p.83: Sandhyabhasam ajananatva [t] ca. And here runs its Tibetan version (Tanjur, Rgyud hgrel, Mi, Narthang edition,(1) fol 194a:(2) 'dgons pas gsuns pahi no bo mi ses pahi phyir/ It may be translated into Sanskrit: sandhabha- sabhavajnanat, 'on account of not knowing the nature of the intentional speech.' Again, we read in the same work (p.93): sandhyabhasam ajanadbhih. And here is the Tibetan version (fol, 194a): dgons te bstan pahi skad ma ses pas. We may literally translate it thus into Sanskrit: sandhaya upadistam bhasam ajanadbhih (=sandhabhasam ). Thus there is nothing of sandhya 'twilight' in these cases. And it is now not difficult to say how sandhya has crept into the MSS. in place of sandha (=sandhaya). It is quite possible that scribes not knowing the true significance of sandhaya or its shortened form sandha changed it into sandhya with which they were very fmiliar. ---------------------- 1. Visvabharati Library. 2. Cordier, II, p, 214 (42), 199a-231a, 5.