Norman - The Four Noble Truths (1983)

The Four Noble Truths

K.R. Norman

[Indological and Buddhist Studies (Volume for J.W. de Jong)]

K.R. Norman Collected Papers II, pp. 210 - 223

(C) Pali Text Society - Oxford 2003

ABBREVIATIONS

ABORI	Annals of the Bhandarkar Oriental Research Institute
Ai.Gr.	J. Wackernagel, Altindische Grammatik
AJP	American Journal of Philosophy
AMg	Ardha-Māgadhī
AMg Dict.	Ratnachandraji, An Illustrated AMg Dictionary
AO	Acta Orientalia
Apa.	Apabhraṃśa
AR	Abhidhāna-rājendra, Ratlam 1913-25
Aś.	Aśokan
Āv.	Āvassaya-sutta
Āyār.	Āyāramga-sutta (ed. H. Jacobi)
BCDRI	Bulletin of the Deccan College Research Institute
BD	Book of the Discipline
Be	Burmese (Chatthasañgāyana) edition
Bhav.	Bhavisatta Kaha (ed. H. Jacobi, Munich 1918)
BHS	Buddhist Hybrid Sanskrit
BHSD	F. Edgerton, BHS Dictionary
BHSG	F. Edgerton, BHS Grammar
BKS	Bṛhat-kalpa-sūtra (ed. W. Schubring, Leipzig 1905)
Bloch	J. Bloch, Les inscriptions d’Asoka, Paris 1950
BSL	Bulletin de la Société de Linguistique de Paris
BSO(A)S	Bulletin of the School of Oriental (and African) Studies
Burrow, Skt Lang.	T. Burrow, Sanskrit Language, London 1955
Burrow, Khar. Lang.	T. Burrow, Language of the Kharoṣthi Documents, Cambridge 1937
CDIAL	R.L. Turner, Comparative Dictionary of the Indo-Aryan Languages
Ce	Sinhalese edition
CII	Corpus Inscriptionum Indicarum
CPD	Critical Pāli Dictionary
cty/cties	commentary/commentaries
cū.	cūrṇi
DED(R)	Dravidian Etymological Dictionary (revised edition)
DNM	Deśīnāmamāā
DPPN	Dictionary of Pāli Proper Names
Ee	European edition
Ep. Ind.	Epigraphia Indica
Erz.	H. Jacobi, Ausgewāhlte Erzāhlungen in Māhārāṣtrī, Leipzig 1886
EV I, II	K.R. Norman, Elders’ Verses I, II, London 1969, 1971
EWA	M. Mayrhofer, Etymologisches Wörterbuch des Altindischen,
Geiger	W. Geiger, Pāli Literatur und Sprache, Strassburg 1916
Gk.	Greek
GOS	Gaekwad’s Oriental Series
Hultzsch	E. Hultzsch, Inscriptions of Asoka, Oxford 1925
IA	Indo-Aryan
IE	Indo-European
IHQ	Indian Historical Quarterly
II	Indo-Iranian
IIJ	Indo-Iranian Journal
IL	Indian Linguistics
Ind. Ant.	Indian Antiquary
Isibh.	Isibhāsiāim (ed. W. Schubring)
IT	Indologicia Taurinensia
JA	Journal Asiatique
JAIH	Journal of Ancient Indian History
JAOS	Journal of the American Oriental Society
JASB	Journal of the Asiatic Society of Bengal
JAS Bombay	Journal of the Asiatic Society of Bombay
JIABS	Journal of the International Association of Buddhist Studies
JOI(B)	Journal of the Oriental Institute (Baroda)
JPTS	Journal of the Pali Text Society
JRAS	Journal of the Royal Asiatic Society
Kapp.	The Kalpasūtra of Bhadrabāhu (ed. H. Jacobi, Leipzig 1879)
Kapp. Sām.	= part III of prec. (pp. 86-95)
Lüders, Beob.	H. Lüders, Beobachtungen über die Sprache des buddhistischen Urkanons, Berlin 1954
Lüders, Phil. Ind.	H. Lüders, Philologica Indica, Göttingen 1940
MIA	Middle Indo-Aryan
MRE	Minor Rock Edict
MSL	Mémoires de la Société de Linguistique de Paris
MS(S)	Manuscript(s)
MW	Sir Monier Monier-Williams, Sanskrit-English Dictionary, Oxford 1899
NAWG	Nachrichten der Akademie der Wissenschaften in Göttingen
Nāyā.	Nāyādhammakahāo
NIA	New Indo-Aryan
Nirayāv.	Nirayāvaliyāo
Nisīh.	Nisīha-sutta
OIA	Old Indo-Aryan
OLZ	Orientalistische Literaturzeitung
Ova.	Ovavāiya-sutta (ed. E. Leumann, Leipzig 1883)
Pā.	Pāli
Paṇh.	Paṇhāvāgaraṇāim
Paum.	Pauma-cariu
PE	Pillar Edict
PED	The PTS’s Pali-English Dictionary
Pischel	R. Pischel, Grammatik der Prākrit-Sprachen, Strassburg 1900
Pkt	Prakrit
PMWS	F.B.J. Kuiper, Proto-Munda Words in Sanskrit, Amsterdam 1948
PSM	H.D.T. Sheth, Pāiasaddamahaṇnavo, Calcutta 1928
PTC	Pāli Tipitakaṃ Concordance
PTS	Pali Text Society
Ratnachandraji	see AMg Dict.
RE	Rock Edict
ṚV	Ṛgveda
SBB	Sacred Books of the Buddhists
SBE	Sacred Books of the East
SepE	Separate Edict
Sheth	see PSM
Sinh.	Sinhalese
Skt	Sanskrit
StII	Studien zur Indologie und Iranistik
Sutt.	Suttāgame
Sūyag.	Sūyagaḍaṃga-sutta
Ṭhāṇ.	Thānaṃga-sutta
ṭī.	ṭīkā
TPS	Transactions of the Philological Society
Turner, CDIAL	R.L. Turner, Comparative Dictionary of the Indo-Aryan Languages
Turner, Nep. Dict.	R.L. Turner, Dictionary of the Nepali Language, London 1931
Utt.	Uttarajjhayaṇa-sutta (ed. J. Charpentier, Uppsala 1922)
Uvās.	Uvāsaga-dasāo (ed. Hoernle)
Vivāg.	Vivāga-sutta
Whitney, Gram.	W.D. Whitney, Sanskrit Grammar, Cambridge (Mass.) 1889
Whitney, Roots	W.D. Whitney, Roots and Verb-forms of the Sanskrit Language, Leipzig 1885
WZKS(O)	Wiener Zeitschrift für die Kunde Süd- (und Ost-)asiens

49. The Four Noble Truths¹

1. The problem

<377> 1. I want in this paper, offered in honour of Professor J.W. de Jong, to examine the grammar and syntax of something which, although fundamental to Buddhist doctrine, has never been satisfactorily explained at the linguistic level. I refer to the statement, in Pāli, of the four Noble Truths (= NTs).

1.2. This statement occurs in the Dhammacakka-ppavattana-sutta (Vin I Io foll. SV 420 foll.), [^1] which is traditionally the first sermon preached by the Buddha after his enlightenment, in the following form: idaṃ kho pana bhikkhave dukkhaṃ ariya-saccaṃ, idaṃ kho pana bhikkhave dukkha-samudayaṃ ariya-saccaṃ, idaṃ kho pana bhikkhave dukkhanirodhaṃ ariya-saccaṃ, idaṃ kho pana bhikkhave dukkha-nirodhagāminī pațipadā ariya-saccam. I shall refer to this as the ‘introduction’ set. [^2]

1.3. It is clear that there is something strange about the grammar and syntax here. The most recent statement that I have seen made about this is that of Johansson, who says:

‘Syntactically, these expressions are somewhat loosely formulated and of different types. Note that samudaya and nirodha are masculine and therefore must be acc. sg., if the compounds are not of the possessive type and therefore adjectively adapted to saccam; pațipadā can only be nom. sg. … Probably dukkhaṃ and pațipadā should be understood as nom. and translated “truth (which is) pain” = “truth about pain”; [^3] dukkha-samudayaṃ and dukkha-nirodhaṃ are probably possessive compounds adjectively related to saccaṃ and therefore nom. sg. nt., literally “pain-originating truth”, 1.e. “truth about the origin of pain”. “truth about the cessation of pain”. There are other possibilities: dukkhaṃ may also be adj., and so the same type of attribute as <378> dukkha-samudayaṃ taken as poss. compound; it may also be taken as acc. sg. of the noun, because acc. is sometimes used as a “case of reference”, although the loc. is more common in this function; dukkhasamudayaṃ and dukkha-nirodhaṃ could also be understood as acc. of reference. On the other hand, patipadā is certainly nom., if it should not simply be combined with ariya-saccaṃ to form one long compound. ²

1.4. Johansson did not quote, and possibly was unaware of, Weller’s suggestion that the statement of the four NTs in Pāli is based upon an earlier version in an Eastern dialect, where the nom. sg. of both masc. and nt. nouns was in -e.³ In that dialect, according to Weller, the 2nd and 3rd NTs would have had the form dukkha-samudaye ariya-sacce and dukkha-nirodhe ariya-sacce, and by a faulty piece of ‘translating’ on the part of the Pāli redactor, -samudaye -sacce and -nirodhe -sacce were changed to -samudayaṃ -saccaṃ and -nirodhaṃ -saccaṃ instead of the correct -samudayo -saccaṃ and -nirodho -saccam.

1.5. It seems that others, too, believed that the correct form of the statement should be -samudayo -saccaṃ and -nirodho -saccaṃ, for there is a v.l. -samudayo at D II 308,1 and M III 250,32, and a v.l. -nirodho at D II 310,4, while Weller quotes the comment of the editors of the Siamese edition, who read -samudayo -saccaṃ and -nirodho -saccaṃ, against their manuscripts, on the grounds that samudaya and nirodha are masculine nouns. ⁴ The general tendency of the manuscripts, however, to read -samudayaṃ and -nirodhaṃ indicates that this is what the Pāli tradition felt was correct, and consequently refrained from ‘correcting’.

1.6. Neither Johansson’s nor Weller’s explanation is entirely satisfactory. As we shall see (§ 2.2), the four NTs also occur in Pāli in a set where -samudayo and -nirodho are found, and Weller’s suggestion does not explain why the two compounds should appear to have different genders in different contexts. Johansson’s explanation does not take account of the fact that we should expect the grammar and syntax of each of the four NTs to be the same, and therefore the explanation must be the same for all four.

1.7. It could also be suggested that in the statement of the four NTs the gender of samudaya and nirodha is genuinely neuter, but this does not meet the objection which has been levelled against Weller’s solution, that in other sets the two words have the expected masculine gender. It would be possible <379> to suggest that -samudayo and -nirodho were changed to -samudayaṃ and -nirodhaṃ on the analogy of dukkhaṃ in the ist NT, and then the expected form of the pronoun ayaṃ was changed to idaṃ to agree with -samudayaṃ and -nirodham. This does not, however, explain why we also have idaṃ in the 4 th NT, although pațipadā is feminine.

2. Other statements of the four Noble Truths in Pāli

2.1. Later in the Dhammacakka-ppavattana-sutta we find the four NTs stated again in two sets (Vin I I I, I foll. =SV 422.3 foll.): idaṃ dukkhaṃ ariya-saccan ti me bhikkhave … āloko udapādi. taṃ kho pan’ idaṃ dukkhaṃ ariya-saccaṃ pariññeyyaṃ … pariññātaṃ … idaṃ dukkha-samudayaṃ ariya-saccan ti me bhikkhave … āloko udapādi. taṃ kho pan’ idaṃ dukkha-samudayaṃ ariya-saccaṃ pahātabbaṃ … pahīnaṃ … idaṃ dukkha-nirodhaṃ ariya-saccan ti me bhikkhave … āloko udapādi. taṃ kho pan’ idaṃ dukkha-nirodhaṃ ariya-saccaṃ sacchikātabbaṃ … sacchikataṃ … idaṃ dukkha-nirodha-gāminī pațipadā ariya-saccan ti me bhikkhave … āloko udapādi. taṃ kho pan’ idaṃ dukkha-nirodha-gāminī pațipadā ariya-saccaṃ bhāvetabbaṃ … bhāvitam. I shall call the set which is followed by ti… āloko udapādi the ‘enlightenment’ set, and that followed by pariññeyyaṃ, etc., the ‘gerundival’ set.

2.2. There are other statements of the four NTs in Pāli which differ from those in the Dhammacakka-ppavattana-sutta. One set occurs in an alternative version of the enlightenment story at M I 23.14-17, ⁵ where each item omits the word ariya-saccaṃ and is followed by ti yathābhūtaṃ abbhaññāsiṃ. I shall call this the ‘basic’ set. It is noteworthy that in this set each item has the correct gender for the nouns (-samudayo, -nirodho) and for the pronouns (idaṃ, ayaṃ, ayaṃ, ayaṃ).

2.3. We also find in Pāli versions various shortened forms of the four NTs. I shall call these the ‘mnemonic’ sets, since they were probably intended to remind the hearer of the full form of the NTs. The shortest set of all is (a): cattāri ariya-saccāni … dukkhaṃ samudayo nirodho maggo (Th 492). ⁶ This seems to be a ‘short-hand’ way of referring to the four NTs, for the ist NT is not ‘Pain’, but the realisation of the fact that ‘This is pain’. Another set, without the word ariya is (b): cattāri saccāni: dukkha-saccaṃ samudaya-saccaṃ nirodha-saccaṃ maggasaccaṃ (Pp 2,1-3). A longer version, with ariya, is found in set (c): cattāri ariya-saccāni: <380> dukkhaṃ ariya-saccaṃ, dukkhasamudayaṃ ariya-saccaṃ, dukkha-nirodhaṃ ariya-saccaṃ, dukkha-nirodha-gāminī pațipadā ariya-saccaṃ (D III 277,8-11). The 4th NT also occurs in the form dukkha-nirodha-gāmini-paṭipadā ariya-saccaṃ (Vism 494.4), where -gāmini- probably represents an attempt to write the stem form of gāminī in a compound.

2.4. It would appear that in mnemonic set (c) the Pāli tradition takes dukkhaṃ, dukkha-samudayaṃ, etc., as being in apposition to ariyasaccaṃ, so that when the latter is in an oblique case, so too is the former, e.g. dukkhaṃ ariya-saccaṃ … dukkha-samudayaṃ dukkha-nirodhaṃ dukkha-nirodha-gāmini-patipadaṃ ariya-saccaṃ pucchanti (M II 10,21 foll.); dukkhassa ariya-saccassa ananubodhā … dukkha-samudayassa … ariya-saccassa ananubodhā … dukkha-nirodhassa … ariya-saccassa ananubodhā (D II 90,12 foll.); dukkhe ariya-sacce dukkha-samudaye ariya-sacce dukkha-nirodhe ariya-sacce dukkha-nirodha-gāminiyā paṭipadāya ariya-sacce (M I 184,31 foll.).

2.5. It is interesting to note that in such contexts with an oblique case usage, the Pāli tradition was not always certain about the way in which to handle the 4 th NT. Besides reading -gāmini-, which probably represents an attempt to write a stem form (as in §2.3), we find the vv.ll. -gāminī- and -gāminim at M II 10,25. Besides the reading -gāminiyā at D II 312,2 we find the vv.ll. -gāminī- and -gāmini-. Not only is there doubt about -gāminiyā/-gāminī-/-gāmini-, but there is evidence that there was doubt about the correct form of pațipadā in such oblique usages. At Vin I 230,30 foll. we find the equivalent of D II 90,12 foll. (see §2.4), with the 4 th NT written as a compound: dukkha-nirodha-gāmini-paṭipadā- ariya-saccassa ananubodhā. An examination of more editions and manuscripts might help to settle the correct form of the 4 th NT in oblique cases, but any conclusions reached are not likely to be totally persuasive, since the possibility of scribal error or emendation can never be completely eliminated.

2.6. Despite the problems which the grammar and sytanx of the four NTs present, translators have shown little doubt about the way in which they should be translated. In the ‘basic’ set (§ 2.2), without ariya-saccaṃ, the interpretation is straightforward: ‘I understood properly “This is pain, This is the origin of pain”, etc.’ We can translate the mnemonic set (a): ‘The four NTs : pain, (its) origin, (its) cessation, the path’. Mnemonic set (b) is normally translated: ‘The four truths: <381> the truth of pain, the truth of the origin, the truth of the cessation, the truth of the path’, where dukkha-saccaṃ, etc., are translated as though they were dependent (tatpurusa) compounds. For mnemonic set (c) the same translation is given, with the addition of ‘noble’ to ‘truth’, as though the words in apparent apposition to ariya-saccaṃ were adjectives or adjectival compounds in agreement with ariya-saccam: ‘The four NTs: the NT of pain, the NT of the origin of pain, etc.’

2.7. A comparable translation is given for the ‘introduction’ set (§ I.2), and the pronoun idaṃ which occurs in each NT is taken as agreeing with -saccaṃ, so that the translation is usually given in the form: ‘This is the NT of pain, this is the NT of the origin of pain, etc.’ No-one, to my knowledge, has commented upon the strangeness of the fact that, on the basis of the translation given for the ‘basic’ set (§ 2.6), we should expect the correct translation to be: ‘The NT (that) “This is pain”, the NT (that) “This is the origin of pain”, etc.’ I presume that the syntax has always dissuaded translators from giving the interpretation which reason told them was the correct one.

2.8. It is possible to translate the ist NT in this set as ‘This pain is a NT’, and in the 4 th NT we might translate ‘This (thing, namely) the path … is a NT’, or ‘This NT (is) the path’, but such translations are not possible for the 2 nd and 3 rd NTs, since -samudayaṃ and -nirodhaṃ are not in the nom. case, unless we assume a change of gender, which is unlikely (§ I.7). In the ‘gerundival’ set (§ 2.1), it would be possible to take the pronoun taṃ, which occurs in each NT, as agreeing with ariya-saccaṃ, and the pronoun which follows it as agreeing with dukkhaṃ, etc. This gives good sense for the ist NT: ‘That truth (that) “This is pain”’, but it is not satisfactory for the other NTs because, as noted above (§ 1.7), the pronoun has the form idaṃ which is not appropriate for the expected masculine forms -samudayo and -nirodho, nor for the feminine form pațipadā.

3. The four NTs in other traditions

3.1. It might be thought that an investigation into the form which the four NTs take in BHS texts might produce a solution to this problem. In fact, such texts produce problems of their own. I quote from the Mvu. ⁷ the Lal, ⁸ and the CPS. ⁹

<382> 3.2. Where the Pāli version of the Dhammacakka-ppavattanasutta has the ‘introduction’ set (§ 1.2), Mvu and Lal have the ‘mnemonic’ set (c); CPS has neither the ‘introduction’ set nor the ‘mnemonic ’ set, although it includes the ‘mnemonic’ set later (§3.4). The equivalent of the Pāli ‘enlightenment’ set (§ 2.1) occurs in the following versions: idaṃ duhkhaṃ iti bhikṣavah … ālokaṃ prādurabhūṣi; ayaṃ duhkha-samudayo ti … ālokaṃ prādurabhūṣi; ayaṃ duhkha-nirodho ti … āloko prādurabhūṣi; iyaṃ ca duhkha-nirodha-gāminī pratipadā iti … āloko pradurabhūṣi (Mvu III 332.13 foll.); duhkha-samudaya iti … ālokah prādurbhūtah; ayaṃ duhkhanirodha iti … ālokah prādurbhūtah; iyaṃ duhkha-nirodha-gāminī pratipad iti … ālokah prādurbhūtah (Lal 417.15 foll.); idaṃ duhkhasamudayo ‘yaṃ duhkha-nirodha iyaṃ duhkha-nirodha-gāminī pratipad ārya-satyaṃ iti … buddhir udapādi (CPS 12.2-3). For the omission of the word ārya-satyaṃ in the 2 nd and 3 rd NTs we can compare the similar omission in ‘mnemonic’ set (c) in the same text (§ 3.4).

3.3. The BHS versions of the ‘gerundival’ set are as follows: taṃ khalu punar imaṃ duhkhaṃ ārya-satyaṃ parijñeyaṃ … tena khalu punar ayaṃ duhkha-samudayo ārya-satyo prahātavyo … atha khalu punar duhkha-nirodho ārya-satyo sākṣikrto … sā khalu punar iyaṃ duhkha-nirodha-gāminī pratipad ārya-satyā bhāvitā (Mvu III 333.3 foll.); yat khalv idaṃ duhkhaṃ parijñeyaṃ … sa khalv ayaṃ duhkha-samudayah prahātavy(ḥ) … sa khalv ayaṃ duhkha-nirodhah sākṣātkartavy(ḥ) … sā khalv iyaṃ duhkha-nirodha-gāminī pratipad bhāvayitavy(ā) … tat khalv idaṃ duhkhaṃ parijñātaṃ … sa khalv ayaṃ duhkha-samudayah prahīn(ḥ) … sa khalv ayaṃ duhkha-nirodhah sākṣātkrt(ḥ) … sā khalv iyaṃ duhkha-nirodha-gāminī pratipad bhāvit(ā) (Lal 418.1 foll.); tat khalu duhkhaṃ ārya-satyaṃ … parijñātavyaṃ … tat khalu duhkhasamudayaṃ ārya-satyaṃ … prahātavyaṃ … tat khalu duhkha-nirodhaṃ ārya-satyaṃ … sākṣikartavyaṃ … tat khalu duhkha-nirodha-gāminī pratipad ārya-satyaṃ … bhāvayitavyā … tat khalu duhkhaṃ āryasatyaṃ … parijñātaṃ … tat khalu duhkha-samudayaṃ ārya-satyaṃ … prahīnaṃ … tat khalu duhkha-nirodhaṃ ārya-satyaṃ … sākṣikrtaṃ … tat khalu duhkha-nirodha-gāminī pratipad ārya-satyaṃ … bhāvitaṃ (CPS 12.4-11). There is a version of the ‘basic’ set in the introduction to CPS: idaṃ duhkhaṃ ārya-satyaṃ iti yathābhūtaṃ prajānāti; ayaṃ duhkha-samudayah; ayaṃ duhkha-nirodhah; iyaṃ duhkha-nirodhagāminī pratipad ārya-satyaṃ iti yathābhūtaṃ prajānāti (CPS E.24). For the omission of the word a a aya-satyaṃ in the 2 nd and 3 rd items we can compare the similar omission in ‘mnemonic’ set (c) in CPS (§3.4).

<383> 3.4. Versions of the ‘mnemonic’ set occur as follows: catvāri … ārya-satyāni seyyathīdaṃ duhkhaṃ ārya-satyaṃ, duhkha-samudayo arya-satyaṃ, duhkha-nirodho ārya-satyaṃ, duhkha-nirodha-gāminī pratipad ārya-satyaṃ (Mvu III 331,17 foll.); catvāri … ārya-satyani duhkhaṃ duhkha-samudayo duhkha-nirodho duhkha-nirodha-gāminī pratipat (Lal 417.2 foll.); catvāri … ārya-satyāni - duhkhaṃ āryasatyaṃ duhkha-samudayo duhkha-nirodho duhkha-nirodha-gāminī pratipad ārya-satyaṃ (CPS 14.2-3). The omission of the word āryasatyaṃ in the 2 nd and 3 rd NTs in the CPS version has already been noted in the ‘enlightenment’ and ‘basic’ sets (§ 3.2-3). We find a different form of the 2 nd and 3 rd NTs at Mvu II 138.4 foll. : duhkhasamudayaṃ ārya-satyaṃ duhkha-nirodhaṃ ārya-satyam.

3.5. It would appear that in the ‘mnemonic’ set at Mvu III 331,17 foll. (§3.4) the words duhkhaṃ, etc., are in apposition to ārya-satyaṃ, although it would be possible to take the 4 th NT as a compound, since pratipad has the same form whether it is nom. sg. or the stem form. If it is a compound, however, we should have to regard -gāminī - as an irregularity, since it is the nom. sg. fem. form instead of the expected stem form. We have already seen (§ 2.5) that -gāminī - sometimes occurs in compounds in Pāli, and we find a comparable example in BHS at Mvu III 408,17 foll., where the four NTs occur as the objects of a group of verbs beginning with āciksati ‘he teaches’. The statement includes duhkhaṃ, duhkha-samudayaṃ and duhkha-nirodhaṃ, all of which could be taken as accusative in apposition to ārya-satyam. The 4th NT, however, is in the form duhkha-nirodha-gāminī-pratipad-ārya-satyaṃ (āciksati), which can only be a compound.

3.6. The problems which the syntax of the four NTs presents have led to some inconsistencies in their translation in BHS texts. Ria Kloppenburg, in her translation of CPS, ¹⁰ translates the ist NT as ‘This suffering is a noble truth’ when it occurs in the ‘basic’ and ‘enlightenment’ sets, ¹¹ as ‘Suffering, that noble truth’ in the ‘gerundival’, ¹² and as ‘The noble truth of suffering’ in the ‘mnemonic’ set. ¹³ As we have noted (§ 2.8), it is possible to take the ist NT in Pāli as ‘This suffering is a noble truth’, but it is not possible to follow her in taking the 2 nd and 3 rd NTs as ‘This origin of suffering is a noble truth’ and ‘This cessation of suffering is a noble truth’ because in the Pāli version -samudayaṃ and -nirodhaṃ cannot be nom. Nor can we translate the Pāli version of the 4th NT as ‘This path leading to the cessation of suffering is a noble truth’, because idaṃ cannot be taken as agreeing with paṭipadā.

<384> 3.7. Nor do the BHS versions of the ‘gerundival’ set help with the interpretation of the Pāli version of that set. In place of the pronoun taṃ which introduces each item in the Pāli version (§ 2.1), Mvu has taṃ, tena, atha and sā. The first three of these suggest that Pāli taṃ is the adverbial use of the pronoun in the sense of ‘then, therefore’, but sā in the 4 th item goes against this, as do yat/tat, sa, sa and sā in the Lal version. The CPS version partly agrees with Pāli in having tat in each item, but it omits the pronouns idaṃ, ayaṃ, ayaṃ and iyam. These differences present too great a problem to be solved in this short paper.

4. The problem reconsidered

4.1. If we consider the form of the ‘enlightenment’ set in Pāli (§2.1) and the other traditions (§ 3.2), we note that Pāli has ariya-saccaṃ in each item (with -samudayaṃ and -nirodhaṃ in the 2nd and 3rd NTs); Mvu and Lal omit ārya-satyaṃ in all four items (with -samudayo/ -nirodho and -samuday(ḥ)/-nirodh(ḥ) respectively); CPS omits ārya-satyaṃ from the 2nd and 3rd items (with -samudayo and -nirodh(ḥ)). In the ‘gerundival’ set the Pāli version (§ 2.1) has ariya-saccaṃ in each item (with -samudayaṃ and -nirodhaṃ); Mvu (§ 3.3) has ārya-satya in each item, but makes -satya agree with the gender of duhkhaṃ, -samudayo, etc.; Lal omits ārya-satyaṃ from each item; CPS has ārya-satyaṃ in each item (with -samudayaṃ and -nirodhaṃ). In the ‘basic’ set the Pāli version (§ 2.2) omits ariya-saccaṃ from each item; CPS omits āryasatyaṃ from the 2nd and 3rd items (with -samudayah and -nirodhah).

4.2. If we examine the form of ‘mnemonic’ set (c) in Pāli (§ 2.3) we find ariya-saccaṃ in each item (with -samudayaṃ and -nirodhaṃ); Mvu (§3.4) has ārya-satyaṃ in each item (with -samudayo and -nirodho in one version, and -samudayaṃ and -nirodhaṃ in the other); Lal omits ārya-satyaṃ from all four items; CPS omits ārya-satyaṃ in the 2nd and 3rd items (with -samudayo and -nirodho).

4.3. Woodward made a very perspicacious remark about the Pāli version of the ‘gerundival’ set (§ 2.1). With reference to the statement that the second NT should be given up (pahātabbaṃ), he noted that the word ariya-saccaṃ should be omitted, since what the Buddha meant was that the origin of pain should be given up, not the truth about it. ¹⁴ As we have seen, in the Lal version (§ 3.3) the word ārya-satyaṃ is <385> omitted from all four items, and consideration shows that this must be correct. What the Buddha said was that pain should be known, its origin given up, its cessation realised, and the path to its cessation practised. Woodward did not, therefore, go far enough. He should have suggested the removal of the word ariya-saccaṃ from all four items in the ‘gerundival’ set.

4.4. Further consideration shows that in other contexts, too, the word ariya-saccaṃ should be omitted. Following the statement of the ‘basic’ set at D II 304.26 foll., there is a series of questions about them. e.g. katamaṃ dukkhaṃ ariya-saccaṃ, etc. (D II 305,1 foll.). This is normally translated ‘What is the NT of pain?’, but since the answer is jāti dukkhaṃ, again without ariya-saccaṃ, it is clear that the original form of the question must have been katamaṃ dukkhaṃ - ‘What is pain?’ Mvu (III 332.1 foll.) and CPS (14.4-10) agree with the Pāli version in having ārya-satyaṃ in each question, although in these two texts they come after a statement of ‘mnemonic’ set (c), which includes the word āryasatyaṃ in each item. The version in Lal (417.4 foll.) has tatra katamad duhkhaṃ, etc., without ārya-satyam. A version of the questions without ariya-saccaṃ occurs in Pāli at M I 48.29 foll.

5. A proposed solution

5.1. I suggest that the original form of the ‘enlightenment’ set was the ‘basic’ set: idaṃ dukkhaṃ, ayaṃ dukkha-samudayo, ayaṃ dukkhanirodho, ayaṃ dukkha-nirodha-gāminī patipadā (to quote it in its Pāli version, without prejudice as to the actual dialect or language in which it was first uttered), as found at M I 23.14 foll. This is supported by the Mvu and Lal versions. The earliest form of the ‘mnemonic’ set was the four words dukkhaṃ samudayo nirodho maggo, without any reference to sacca, e.g. yā buddhānaṃ … dhamma-desanā taṃ pakāsesi dukkhaṃ samudayaṃ nirodhaṃ maggaṃ (Vin I 16.3). When these items became known as ‘truths’, they were so designated: cattāri ariya-saccāni dukkhaṃ samudayo maggo nirodho (Th 492).

5.2. Their designation as saccāni led to the introduction of the word -sacca into each item: cattāri saccāni - dukkha-saccaṃ samudayasaccaṃ nirodha-saccaṃ magga-saccaṃ (Pp 2.1-3). Although these items are usually translated as though they were dependent <386> (tatpuruşa) compounds (§ 2.6), they should rather be taken as descriptive (karmadhāraya) compounds: ‘The truth “pain”, etc.’, cf. uposathasaddo ‘The word “uposatha”’. They might even be taken as abbreviated forms of syntactical compounds: ¹⁵ *idamdukkha-saccaṃ, etc., ‘The truth (that) “This is pain”, etc.’, cf. idamsaccābhinivesa ‘The inclination (to say) “This is true”’, i.e. ‘The inclination to dogmatise’.

5.3. When the truths became known as ariya-saccāni, then this word was added to the ‘mnemonic’ sets. It was added to the simplest form in apposition to the four items: yā sā buddhānāṃ … dharma-deśanā tadyathā duḥkhaṃ samudayo nirodho mārgaś catvāry ārya-satyāni … samprakāśayati (CPS 16.13). The introduction of the word ariya- into ‘mnemonic’ set (b) gave a set: *dukkha-ariya-saccaṃ, etc. I suggest that the hiatus between dukkha-, etc., and -ariya-saccaṃ was avoided by the insertion of a sandhi -m-, producing dukkha-m-ariya-saccaṃ, etc. Wrong word division led to this being taken as dukkhaṃ ariya-saccaṃ ( S V 434.9-11), and then dukkhaṃ ariya-saccaṃ (D II 277.8-11), i.e. as two words in apposition. This was probably helped by the fact that dukkhaṃ could be taken as an adjective in agreement with ariya-saccaṃ. The same wrong division of dukkha-samudaya-m-ariya-saccaṃ and dukkha-nirodha-m-ariya-saccaṃ led to the appearance of dukkha-samudayaṃ and dukkha-nirodhaṃ. Despite the fact that these two words were felt to have an independent existence, ¹⁶ so that they could be declined in apposition to ariya-sacca (§ 2.4), nevertheless the Pāli tradition, with the few exceptions noted above (§ I.5), recognised that these were the correct forms, and refrained from ‘correcting’ them.

5.4. In the 4 th NT, the replacement of magga by patipadā produced a hiatus between -ā- and -ariya-saccaṃ, which was tolerated, and no sandhi -m- was inserted. Since the stem form was identical with the nom. sg. form, it was possible to take patipadā and ariya-saccaṃ as being in apposition. It was therefore possible to take dukkha-nirodha-gāminī as being a separate adjective in agreement with the nom. sg. form patipadā, although it is clear from the variety of forms we find (§ 2.3-5) that the tradition was not certain about this. It is, again, possible that we have examples of abbreviated forms of syntactical compounds here. If the original form was *ayam-dukkha-nirodha-gāminī-patipadā-ariyasaccaṃ, then the compound forms we have noted (§ 2.5) are easily understood. To this extent, Johansson’s suggestion of a long compound (§ I.3) is correct, although he did not realise that all four NTs can be taken as compounds.

<387> 5.5. Of the BHS versions of the ‘mnemonic’ set, that in Lal omits the word ārya-satyaṃ, and so the problem of hiatus does not arise there. At Mvu II 138.4 (§ 3.4) we find duhkhaṃ ārya-satyaṃ duhkhasamudayaṃ ārya-satyaṃ duhkha-nirodhaṃ ārya-satyaṃ duhkha-nirodha-gāminī pratipad ārya-satyaṃ, which in the light of the discussion in the previous paragraph can all be taken as compounds, with sandhi -m- in the first three items. In the 4 th NT pratipad, which can be either nom. sg. or stem form, creates no hiatus. At Mvu III 331,17 foll., however, we find duhkha-samudayo ārya-satyaṃ and duhkha-nirodho ārya-satyaṃ. It would seem most likely that these forms represent late attempts to ‘correct’ what was thought to be faulty grammar when wrong word division led to the appearance of the anomalous forms -samudayaṃ and -nirodhaṃ, just as we have seen occasionally in the Pāli tradition (§ I.5). The CPS version omits ārya-satyaṃ from the 2nd and 3rd items, where we find -samudayo and -nirodho.

5.6. As suggested above, the word ariya-sacca is not appropriate in the ‘enlightenment’ (§5.1) or the ‘gerundival’ (§4.3) sets, but its presence in the ‘mnemonic’ set doubtless led to its introduction there by analogy. Theoretically, its introduction should have led to syntactical compounds: *idam-dukkha-m-ariya-saccaṃ, *ayam-dukkha-samudaya-m-ariyasaccaṃ, *ayam-dukkha-nirodha-m-ariya-saccaṃ, *ayam-dukkha-nirodha-gāminī-patipadā-ariya-saccaṃ, but just as a misunderstanding of the structure of the compounds in the ‘mnemonic’ set led to a faulty word division, so another misunderstanding led to the separation of the pronouns from the beginning of the compound. Since in the first item in the Pāli version idaṃ seemed to agree with saccaṃ, the other three pronouns were changed to idaṃ to agree in the same way.

5.7. Lal does not include the word ārya-satyaṃ in either the ‘enlightenment’ or the ‘gerundival’ set (§4.1). Mvu does not include the word in the ‘enlightenment’ set, and that it is an addition to the ‘gerundival’ set is clearly shown by the fact that the syntactical problem of fitting it into each item was solved by making -satya agree in gender with duhkhaṃ, -samudayo, etc. The CPS version of the ‘enlightenment’ set omits ārya-satyaṃ in the 2 nd and 3 rd items, as it does in the ‘mnemonic’ set, and has -samudayo and -nirodh(ḥ) as in the same set. The CPS version of the ‘gerundival’ set has ārya-satyaṃ in each item, with -samudayaṃ and -nirodhaṃ in the 2 nd and 3 rd items. It is not obvious why CPS sometimes includes ārya-satyaṃ in the 2nd and 3rd items, and sometimes omits the word. It is, however, clear that when ārya-satyaṃ <388> is included we find -samudayaṃ and -nirodham; when it is omitted we find -samudayo (-ah) and -nirodho (-ah). We do not find -samudayo ārya-satyaṃ or -nirodho ārya-satyaṃ, which indicates that tradition felt that this combination of words was incorrect. In the CPS versions of the ‘enlightenment’ and the ‘basic’ sets we find idaṃ-duhkha-m-ārya-satyaṃ and iyaṃ-duhkha-nirodha-gāminī -pratipad-ārya-satyaṃ in the ist and 4th items respectively, which are precisely the forms which are expected as syntactical compounds.

5.8. The ‘introduction’ set, found only in the Pāli version, resembles the Pāli form of the ‘enlightenment’ and ‘gerundival’ sets, with the words kho pana bhikkhave inserted between the pronoun and the noun. Its absence from the three BHS versions and the fact that it is replaced in the Mvu and Lal versions by the ‘mnemonic’ set suggests that it is not an original feature of the sutta. It is possible that in the earliest version there was no set of four NTs at the beginning of this portion of the narrative at all. When the idea of the NTs became more widespread, and the word ariya-saccaṃ was inserted into the ‘basic’ set which, as suggested above (§5.1), was the original form of the ‘enlightenment’ set, a statement of the four NTs was prefixed to the story as a heading or rubric, in some traditions. If this was so, then it is likely that the Mvu and Lal versions independently prefixed the ‘mnemonic’ set as being a very appropriate introduction to what was to follow.

6. Conclusions

6.1. A number of problems remain. The precise relationship between the different versions is not clear. The reasons for the inconsistencies in some texts, e.g. in the form of the 2 nd and 3 rd NTs in the CPS, are unknown, but in some cases they may be due to a mixture of material from various sources. The relative chronology of the changes which must be assumed to have taken place in the form of the NTs is hazy. Nevertheless it seems possible to come to some conclusions.

6.2. The correct form of the NTs in Pāli is: idaṃ dukkhaṃ, ayaṃ dukkha-samudayo, ayaṃ dukkha-nirodho, ayaṃ dukkha-nirodha-gāminī patipadā - ‘This is pain, this is the origin of pain, this is the cessation of pain, this is the path leading to the cessation of pain’. When the word ariya-saccaṃ is included in the statement, we should translate: ‘The NT (that) “This is pain”, etc.’

<389> 6.3. The grammatical form of the four NTs when the word ariyasaccaṃ is included is a syntactical compound. This was not understood by the tradition, with the result that faulty division of the compounds led to the apparent production of nom. sg. forms -samudayaṃ and -nirodham. The belief that in the ist NT idaṃ was an independent pronoun agreeing with -saccaṃ led to the pronouns in the other three NTs being changed to idam.

6.4. The earliest forms of the ‘enlightenment’ and ‘gerundival’ sets did not include the word ariya-saccam. Since the ‘introduction’ set is an addition to the Dhammacakka-ppavattana-sutta, we may conclude that the earliest form of this sutta did not include the word ariya-saccam.

6.5. Nevertheless, as John Brough stated in his note on pamādo/ *pāmado, ¹⁷ these readings are so well entrenched in the Pāli tradition that, even if agreement could be reached upon the original form of the four NTs, no editor would think of inserting an emendation of them into his text.

Footnotes

1. Indological and Buddhist Studies (Volume in honour of Professor J.W. de Jong). Canberra 1982, pp. 377-91. [^1] Abbreviations of titles of texts are those of the CPD; in addition: CPS = Catuṣparisatsūtra. [^2] The names ‘introduction’, etc., are given merely for convenience of reference, without prejudice as to the original form or function of the sets to which they refer. [^3] For consistency, I translate dukkha/duhkha hereafter as ‘pain’, except when quoting other persons’ translations, without implying that this is necessarily the best translation. ↩

2. Rune E.A. Johansson, Pali Buddhist Texts, Lund 1973, p. 24. ↩

3. F. Weller, “Über die Formel der vier edlen Wahrheiten”, in OLZ, XLIII/3-4 (1949), pp. 73-79. ↩

4. Ibid., p. 73, note 3. ↩

5. I normally give a single reference for each Pāli quotation. Other references, if they exist, can be found in PTC. ↩

6. At Th 492 the order of the last two items is reversed for metrical reasons. ↩

7. É. Senart, Le Mahāvastu, I-III, Paris 1882-97. According to the text itself (I 2.13-14), it is of the Vinaya-piṭaka according to the text (pāthena) of the Lokottaravādins of the noble Mahāsānghikas of the Middle Country. ↩

8. S. Lefmann, Lalita Vistara, Halle 1902. M. Winternitz, (History of Indian Literature, Vol. II, p. 248) quotes the Chinese tradition that this Mahāyāna text originally contained the life story of the Buddha for the Sarvāstivādins of the Hīnayāna. ↩

9. E. Waldschmidt, Das Catuspariṣatsütra, ADAWB, 1960, I, Berlin 1962. The CPS is a Sarvāstivādin text, but is identical with the Sanghabhedavastu of the Mūlasarvāstivādins, from which the Introduction in Waldschmidt’s edition of CPS was taken, if I understand the situation correctly. ↩

10. Ria Kloppenburg (tr.), The Sütra on the Foundation of the Buddhist Order, Leiden 1973. ↩

11. Ibid., pp. 4, 24. ↩

12. Ibid., p. 24. ↩

13. Ibid., p. 28. ↩

14. F.L. Woodward, The Book of the Kindred Sayings, Part V, London 1930, p. 358 , note I. ↩

15. For such compounds in Pāli see G.V. Davane, Nominal Composition in Middle Indo-Aryan, Poona 1956, pp. 135-39. For syntactical compounds in Sanskrit se Wackernagel, Ai.Gr., II.I, §§ I2 I-24 and Whitney, Gram., § I314. ↩

16. In a version of the ‘mnemonic’ set found in an inscription in Brāhmī characters of the second or third centuries A.D. at Sarnath, we find the word bhikkhave inserted between dukkhaṃ and ariya-saccaṃ in the ist NT. See Sten Konow, “Two Buddhist inscriptions from Sarnath” in Ep. Ind., IX (1907-08), pp. 29193. ↩

17. John Brough, The Gāndhārī Dharmapada, London 1962, p. 194. ↩