Abstract
Verbal forms of upa √ās are one of the characteristic features of Upaniṣadic diction. While several studies have investigated their semantics, very little attention has been given to their syntax. A quick comparison of different translations shows that there is no agreement among Upaniṣadic interpreters regarding the syntax of upa √ās. By considering all its occurrences in the Bṛhadāraṇyaka and Chāndogya Upaniṣads, this paper offers the first systematic study of its syntax. It is hoped that the analytic model proposed here will allow us to better parse, understand, and translate the relevant Upaniṣadic passages. The paper comprises two parts: Part I presents the results of my analysis succinctly; Part II offers a detailed analysis. Two appendixes follow: Appendix 1 examines the formula ‘ya eta- evaṃ vidvān (…) upa √ās’; Appendix 2 represents all the occurrences of upa √ās in a table.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
Scholarly works on upa √ās fall into two groups. (i) Several contributions deal specifically with upa √ās; since Oldenberg (1896, pp. 457–462), upa √ās has often been discussed in connection with the word upaniṣad. For a helpful review of the relevant literature, see Falk (1937, pp. 129–136) and, more recently, Falk (1986, pp. 80–81). (ii) Many works that deal with the Upaniṣads in general also consider upa √ās to varying degrees of details. For example, see Killingley’s (2018, pp. 134–136) recent discussion. Both kinds of scholarly works consider almost exclusively the semantics of upa √ās, while giving little or no attention to its syntax. While such works have, of course, yielded many valuable observations, it seems to me methodologically unsound to delve into the semantic and conceptual aspects of upa √ās without first having gained a solid grasp of its syntax.
I cite the text of BU and CU from Olivelle (1998). With respect to the BU passages that I cite, there are no differences between the Mādhyandina and Kāṇva recensions. I cite from the latter. As far as I can see, accent does not bear on the syntax of upa √ās.
Four our purposes, the notion of ‘section’ is unproblematic and does not need to be defined. It is immediately clear from the context in which they occur whether two or more occurrences of upa √ās belong to the same section.
Excepting instances (1.1), in this paper I render the verbal forms of upa √ās with upās. I do so to factor out semantics from my analysis, which is concerned with syntax. If the reader prefers, she can replace upās with another semantically non-committal expression such as ‘take(s) [O] as [OP]’; cf. Olivelle (1998, p. 514 [on BU 4.1.2]).
In this respect, Sanskrit is not unlike English. For the sake of the argument, let us render upa √ās plus two objects with ‘to take O as OP’ (see footnote 4). Consider the example sentence ‘I take you (O) as a friend (OP)’. In this form, this sentence follows the ordinary, unmarked order, with the O before the OP. In the marked order, however, the OP would be foregrounded, and it would thus precede the O, i.e. ‘as a friend (OP) I take you (O)’. In my translation of the instances of upa √ās, I try to reproduce the difference between unmarked and marked order of the original Sanskrit.
Cf. Delbrück’s (1878, p. 13) remarks that in the Śatapathabrāhmaṇa prose the marked order (‘occasionelle[n] Wortstellung’) is frequent in ‘animated narration and conceptual debate’ (‘in der bewegteren Erzählung und der begrifflichen Erörterung’).
Note how the two principles—(i) the grammatical case of OP; (ii) the O-OP order—operate. With respect to instance (2.1), principle (i) overrules principle (ii). While most instances follow the unmarked order (‘O “OP” iti’), and a few follow the marked order (‘“OP” iti O’), all instances have nominative plus iti as their OP. Conversely, with respect to instances (2.3), principle (ii) overrules principle (i). All instances (2.3) follow the marked order (‘OP O’), regardless of whether their OP is an accusative (i.e. instances 2.3a, ‘OP [=accusative] O’) or a locative (i.e. instances 2.3b, ‘OP [=locative] O’). As I explain below, instances (2.2) are best described in contradistinction to instances (2.1) and (2.3).
In my analysis, the eta- pronominal adjective qualifies udgītha, the O. For a fuller discussion, see my analysis of eta- in the formula ‘ya eta- evaṃ vidvān (…) upa √ās’, in Appendix 1.
Note however BU 1.5.13(2x) and BU 4.4.16. Though they are a commentary on a śloka and a śloka, respectively, both instances follow what I identify as regular syntactic patterns of upa √ās. Specifically, the former is an instance (2.1); the latter is an instance (3.2). Both will be discussed below in due course.
It is interesting to note that here upa √ās expresses subordination, that is, (a) that sits down next to (b) is subordinate to (b). Falk (1986, p. 82) cites a few passages in which verbal forms of upa-ni √sad are used in a similar sense, and emphasizes subordination as being the key semantic feature.
I take atas (ekaikam) as being equivalent with eteṣām (ekaikam). This is supported by BU 1.4.17: eteṣām ekaikaṃ. Cf. also BU 6.2.2, 3: nāham ata ekaṃ cana vedeti…tato naikaṃ cana vedeti (6.2.3) ‘I don’t know the answer to any of these…I didn’t know the answer to a single one of them’ (Olivelle 1998, p. 147). Śvetaketu first converse with Jaivali, hence atas ‘of these’; then he reports the conversation to his father Uddālaka, hence tatas ‘of those’.
On the order of nominal sentences, compare Hock’s view given in “Correlation Between Ritual Item and Marked Order” section.
With respect to CU 2.2.1, Śaṅkara (BSBh 4.1.6; p. 469, line 28) notes that even though there is inversion of the terms (nirdeśaviparyaya) the hiṅkāra is S and pṛthivī is P. This suggests that for Śaṅkara nominal sentences ordinarily follow the pattern ‘S P’.
Note the relation between the locative forms in such sentences and Śaṅkara’s theory of superimposition. As mentioned above, Śaṅkara explains the logical relation between O and OP in upa √ās sentences (as well as the relation between S and P in nominal sentences) by means of a theory of superimposition. The O is the locus upon which the OP is superimposed. That is, OP is being cognized by superimposing it on O. To put it differently, O is the prius or known item that allows one to cognize the OP, the unknown item. Note that, perhaps somewhat surprisingly, in the upa √ās sentences under discussion the O is accusative and the OP is locative, not the other way around.
I think that brahman is O also in BU 2.1(a)(12x); see “Instances (2.2), ‘O [=non-ritual item] OP’” section. It seems reasonable to think that Śaṅkara’s analysis is partly influenced by theological commitments: the ultimate principle itself cannot be the direct object of cognition, but can only be known through a symbol (pratīka). In this respect, compare my observations on how brahman is conceptualized, in footnote 32.
Since the three referents (manas (n.), vāc (f.), prāṇa (m.)) have three different genders, one would expect the neuter pronoun tāni, which is used for the same referents in BU 1.5.3: trīṇy ātmane ’kuruteti mano vācaṃ prāṇaṃ tāny ātmane ’kuruta ‘“he made three for himself” mind, speech, breath—them he made for himself’. Perhaps the irregular masculine gender ta ete (BU 1.5.13) can be explained contextually. Note that ta ete is separated from its three actual referents, but is immediately preceded by a sentence containing prāṇa (m.), āpas (f.), and candraḥ (m.). When ta- pronoun refers collectively to both masculine and feminine referents, it can be either masculine or neuter. Cf. Speijer (1886 SS, pp. 19–20, §28 b; 1896, VSS, p. 31, §101.
The word nāma is employed in BU 2.1.15 (etair nāmabhir) to refer collectively to the OP-s of BU 2.1(b)(12x), i.e. the names given to the 12 persons recorded in the debate between Gārgya and Ajātaśatru. On this point, see Acharya (2013, esp. 13) and Visigalli (2020 [forthcoming], section §1.3).
In CU 4.5.2–4.8.4(8x), the word nāma occurs in the sentence introducing the name/epithets given to the quarters of brahman. Take as example CU 4.6.3, which introduces the above-cited CU 4.6.4: ‘[the fire says to Satyakāma] Dear, this quarter of brahman consisting of four-sixteenths is named “the Infinite”’ (eṣa vai somya catuṣkalaḥ pādo brahmaṇo ’nantavān nāma). (I am aware that nāma is used adverbially.)
Compare also Jaiminīya Brāhmaṇa Upaniṣad 1.27. In this passage, three persons (puruṣa) are identified as residing in the eye, sun, and lightning, respectively. A name (nāma) is ascribed to each of them, and it is enjoined that one should upa √ās each such person (O) as their corresponding name (OP); the OP is nominative plus iti. Take as example JUB 1.27.4 (text in Oertel 1896, p. 104): sa yo’ yaṃ cakṣuṣi eṣo ’nurūpo nāma. anvañ hy eṣa sarvāṇi rūpāṇi. tam anurūpa ity upāsīta. anvañci hainaṃ sarvāṇi rūpāṇi bhavanti ‘The name of the one [person] here in the eye is “Visible-wards”. For it is to-wards all visible (things). One should upās it as “Visible-wards” [as a result] all visible (things) become turned to-wards him’.
It is noteworthy that in such occurrences upa √ās resembles ā √cakṣ, as can be evinced by comparing the parallel passage CU 4.15.2. In this passage, Satyākama teaches Upakosala the secret name of the person (puruṣa) in the eye: etaṃ saṃyadvāma ity ācakṣate ‘They call him “Lovely-uniting”’ (tr. Olivelle 1998, p. 225).
The only difference is the absence of the initial string ‘sa yo’ in CU 3.18.1. There is only one other instance in which one of the objects taken by upa √ās occurs as the first word of the sentence: BU 5.8.1: vācam (O) dhenum (OP) upāsīta ‘one should upās speech as a milch cow’ (discussed below “Instances (2.2), ‘O [=non-ritual item] OP’” section).
That manas is an accusative is confirmed by comparison with the other parallel occurrences of upa √ās in CU 7; e.g. CU 7.2.2: sa yo vācaṃ (O) brahma (OP) ity upāste ‘who upās-es speech as “brahman”’. Note that I take all these occurrences as one instance, CU 7.1.5–14.2 (28x).
The analysis of CU 3.18.1 is further complicated by the sentence that immediately follows it, ākāśo brahma [ity upāsīta]. Some versions of the CU text have ity upāsīta, some do not (see Olivelle 1998, p. 546). The textual variant is significant. For, if ity upāsīta is accepted, ākāśo brahma must be analyzed as a nominal sentence governed by upa √ās. This variant would thus provide evidence that upa √ās takes a nominal sentence, which would bear on my interpretation of CU 3.18.1.
I take this opportunity to clarify the syntax of three nominal sentences that occur in CU 3.18–19. These sentences are: mano brahma (CU 3.18.1) (assuming for the sake of the argument that this is a nominal sentence); ākāśo brahma (CU 3.18.1) (regardless of whether it is followed or not by ity upāsīta); and ādityo brahmety ādeśaḥ (CU 3.19.1). While most translators analyze these sentences as predicate-subject (P S) (see, e.g. most recently, Acharya 2017, p. 556), internal evidence rather suggests that they should be analyzed as (S P). As Śaṅkara (BSBh 4.1.5) points out, the syntax of CU 3.19.1 can be determined on the basis of CU 3.19.4, in which it is clear that the sun is O and brahman is OP: sa ya…ādityaṃ (O) brahma (OP) iti upāste ‘so who…upās-es the sun as “brahman”’. Now, it is reasonable to assume that the same syntactic-logical relation that obtains between the ‘sun’ and ‘brahman’ in CU 3.19.4 must also obtain between the same two words in the nominal sentence in CU 3.19.1. Put it schematically, ‘O : OP = S : P’. (My reasoning assumes that nominal sentences and upa √ās sentences are equivalent; see “Three Arguments to Establish the O-OP Order in Two-Object Instances” section) For this to happen, CU 3.19.1 must follow the order ‘S P’: ādityo (S) brahma (P) [ity ādeśaḥ] ‘the sun is brahman; [(this is) the ādeśa]’. Owing to both contextual and conceptual proximity, it seems likely that the other two nominal sentences (mano brahma; ākāśo brahma), too, should likewise be analyzed as ‘S P’.
One other analysis (iv) is found in literature, though I find it unlikely. This analysis takes the particle iti as having an enumerative function and the relevant words (śraddhā/ṃ tapa; iṣṭāpūrte dattam) as being the direct objects taken by upa √ās, i.e. (‘O O¹ etc.’). See, for example, Keith (1925, p. 576) ‘worship faith and asceticism’ (Keith adds: ‘unless it means “worship, holding that faith is asceticism”’); ‘those who are bent on sacrifice, gifts, and almsgiving’. As Köhler (1973, p. 55, note 107) notes, this analysis involves ascribing a meaning such as ‘to practice’ to upa √ās, in which case the accusative śraddhāṃ would seem to be preferable over the nominative śraddhā.
With respect to CU 5.10.1, Geldner (1928, p. 140, note 792) proposes reading śraddhātapas and renders it as ‘und die, welche im Walde gläubige Askese [śraddhātapas] und dergleichen [iti] üben [upāsate]’ (square brackets are my own addition). It is likely that Geldner’s proposal is influenced by tapaḥśraddhe in Muṇḍaka Upaniṣad (MU) 1.2.11: tapaḥśraddhe ye hi upavasanty araṇye ‘[b]ut those in the wilderness…who live a life of penance and fait’; transl. Olivelle (1998, p. 441). However, I doubt that comparison between CU 5.10.1 (’raṇye śraddhā(ṃ) tapa ity upāsate) and MU 1.2.11 (tapaḥśraddhe ye hi upavasanty araṇye) may cast light on the former. Rather, the replacement of upa √ās with upa √vas may suggest that a reanalysis (misunderstanding?) of upa √ās took place in MU 1.2.11; cf. Senart’s (1930, p. 67, note 2) remarks. Further study on upa √ās in the later Upaniṣads is required to clarify this point.
Three other syntactic analyses have been proposed: (i) śraddhā and satya are both direct objects of upa √ās; e.g. Müller (1884, p. 208) ‘worship faith and the True’; (ii) satyam is an adverb that modifies upa √ās; e.g. Hume (1921, p. 163) ‘who…truly worship faith’; (iii) śraddhā is OP and satya is O, i.e. the instance follows the order ‘OP O’; e.g. Olivelle (1998, p. 149) ‘who venerate truth (O) as faith (OP)’. With respect to Olivelle’s translation, it is worth noticing that his commentary on the passage seems to support my analysis (‘O OP’), rather than reflect the analysis (‘OP O’) underlying his translation. Olivelle (1998, p. 528) notes that ‘the people in the wilderness, who take faith [O] to be something internal (truth [OP] here [i.e. BU 6.2.15(a)], and austerity [OP/P] in CU 5.10.1), to people in villages, who participate in the external acts of giving’. The square brackets are my own addition.
Cf. Senart (1930, p. 47, note 1), who notes that the repetition of devatā suggests that upa √ās must govern also an OP (though he translates somewhat freely ‘enseigne moi, Seigneur, le dieu auquel to crois’). The second devatā is rendered as OP by Lüders (1940, pp. 361–362) “…die Gottheit, die (O) du als Gottheit (OP) verehrst”.
CU 5.12.1–17.1(a)(6x) can also be analyzed as following the marked order, ‘OP O’: ‘as what do you upās the ātman?’. The marked order would be consistent with the context being that of a debate. Another version of the episode is recorded in ŚB 10.6.1. In this version, Aśvapati asks the brahmins about the Vaiśvānara fire, instead of the Vaiśvānara ātman, and he uses the verb √vid, instead of upa √ās. Note that in his translation of the ŚB sentence that corresponds to CU 5.12.1–17.1(a)(6x) Eggeling (1897, p. 395) takes káṃ as the OP: káṃ tváṃ vaiśvānaráṃ vetthéti ‘as whom (OP) knowest thou Vaisvânara (O)?’. Eggeling adds in note ‘Or, perhaps, “what Vaisvânara knowest thou?”’.
For the arguments in support of this analysis, see Visigalli (2020 [forthcoming §1.2]).
In my classification, these two occurrences are classified as belonging to instances CU 2.2.(a)(7x) and CU 2.2.(b)(7x), respectively; see Appendix 2. Such instances follow the syntactic pattern (2.3b) ‘OP [=locative] O’, where saptavidha sāman features as the accusative O. Note that in such instances the word saptavidha sāman is unambiguously the O of upa √ās, regardless of whether one takes the preceding locative as OP or not.
In contradistinction to all other instances, instances (2.1b) (‘O “OP [=brahman]” iti’) regularly exhibit the link between benefits and O. As example, take CU 7.7.2. The passage states that one who upās-es perception (vijñāna) (O) as ‘brahman’ (OP), he will win worlds that consist of perception (vijñānavat) and knowledge (jñānavat), and will be able to move as he wishes as far as the domain of perception extends itself (sa yo vijñānaṃ brahmety upāste | vijñānavato vai sa lokāñ jñānavato ’bhisidhyati | yāvad vijñānasya gataṃ tatrāsya yathākāmacāro bhavati yo vijñānaṃ brahmety upāste).
It is tempting to explain this idiosyncratic feature of instances (2.1b) in connection with how brahman is conceptualized in the relevant passages. In CU 7, brahman is represented as an unknown variable (OP), which can be known through identification with a series of known counterparts (O). Knowing brahman is described as an ever-expansive process whereby provisional counterparts are successively replaced by ever more expansive counterparts. Thus, the first counterpart ‘name’ (nāma) is replaced by ‘speech’ (vāc), ‘speech’ by ‘mind’ (manas), ‘mind’ by ‘intention’ (saṃkalpa) and so on, all the way up to ‘plenitude’ (bhūman). In each stage of the process, the benefits deriving from identifying (upa √ās) a given counterpart (O) with brahman (OP) consist of features associated with that given counterpart.
The transformative or magic dimension of Upaniṣadic knowledge was already noticed by Anquetil Duperron, the first Western translator of the Upaniṣads (from Persian). Tellingly, Duperron uses as epigraph to his book his translation of Muṇḍaka Upaniṣad 3.2.9 ‘Quisquis Deum intelligit, Deus fit’ (1801, p. 393; see also his emphatic note) ‘whoever understands god, becomes god’. (The Sanskrit reads: sa ya ha vai tat paramaṃ brahma veda brahmaiva bhavati). This is the passage referred to in the title of this paper. On the links between the Upaniṣadic knowledge and the associated benefits, see the comments by Deussen (1921 [1897], p. 70, note 1) and Olivelle (1998, pp. 24–25) on CU 1.2.10–12.
There is one more passage in the identification chain in CU 1.1.8: tad vā etad anujñākṣaram | yad dhi kiṃcānujānāty om ity eva tad āha | eṣo eva samṛddhir yad anujñā. The syllable oṃ is first equated with ‘assent’ (anujñā), for whenever one assents to something he says ‘oṃ’; then assent is equated with ‘fulfilment’ (samṛddhi). So, the identification chain is oṃ > anujñā > samṛddhi.
Cf. Delbrück’s (1878, p. 27) remarks on the order ‘P S’ in nominal sentences: ‘The reason for this order is evident. That is, the subject is known, whereas the predicate brings something new, and therefore is foregrounded, according to the general rule of the occasional word-order [der occasionellen Wortstellung]’. Delbrück takes ‘P S’ to be the standard order. For a partly similar distinction between ‘Given Information’ and ‘New Information’, cf. Kobayashi’s (2013) study on the particles vái and evá.
For a review of the earlier literature on the S-P order of nominal sentences, see Gren-Eklund (1978, pp. 18–26).
I take tad as an adverb. It might be possible to take tad etat together as O.
Note that an alternative syntactic analysis of BU 4.4.16 is possible (the same applies also to CU 2.10.6 discussed below): OP¹ could stand in apposition to OP, rather than being a second distinct OP governed by upa √ās. But I do not think that this is the case; rather, I posit an asyndeton, i.e. an implied ‘and’; cf. CU 3.13(5x), where ca is mentioned in three occurrences, but is implied in two.
Note the different position of pronoun eta- in BU 4.1(b)(6x) ‘evaṃ vidvān etad upa √ās-’. Cf. Ickler’s (1973, p. 70) remarks about the order of adjectival eta-.
Cf. also the following passages, in which adjectival eta- qualifies O-s taken by verbs other than upa √ās. CU 1.4.5: sa yaetadevaṃ vidvānakṣaraṃpraṇauti ‘when someone, knowing thus, pronounces this syllable’; CU 1.7.7: atha yaetadevaṃ vidvānsāmagāyaty ‘so, who, knowing thus, sings this sāman’; CU 5.24.2: atha yaetadevaṃ vidvānagnihotraṃjuhoti ‘so, who, knowing thus, offers this agnihotra’.
Cf. Ickler’s (1973, p. 33) translation of Māṇḍūkya Upaniṣad 1.1: om ity etad akṣaram… ‘Jene bekannte Silbe “Om”…’. On the meaning and function of eta-, cf. Ickler (1973, p. 11; 68). CU 1.3.6, 7, too, are problematic. I briefly consider CU 1.3.7, which presents the formula under discussion: ya etāny evaṃ vidvān udgīthākṣarāṇy upāsta udgītha iti. Three analyses are possible: (i) upa √ās takes only one object, ‘etāny…udgīthākṣarāṇy’, the iti-clause ‘udgītha’ stands in apposition to it; (ii) it is an instance (2.1) ‘O “OP” iti’, ‘etāny…udgīthākṣarāṇy’ is the O and udgītha iti is the OP; (iii) it is a variation on instances (2.3a) ‘OP [O]’; ‘etāny…udgīthākṣarāṇy’ is the OP, the iti-clause stands in apposition to it, and the word udgītha is the implied O. Because this instance is both followed and preceded by instances (2.3a), analysis (iii) would seem to better fits with the context. However, this would mean taking the adjectival eta- to qualify the OP, like in CU 1.1.7.
References
Acharya, D. (2013). néti néti. Mening and function of an enigmatic phrase in the Gārgya-Ajātaśatru dialogue of Bṛhad Āraṇyaka Upaniṣad II.1 and II.3. Indo-Iranian Journal,56, 3–39.
Acharya, D. (2017). On the Meaning and Function of Ādeśá in the Early Upaniṣads. Journal of Indian Philosophy,45, 539–567.
Bodewitz, H. (trans.) (1973). Jaiminīya Brāhmaṇa I, 1–65 translation and commentary with a study: agnihotra and prāṇāgnihotra. Leiden: Brill.
Böhtlingk, O. (Ed., Trans.). (1889). Khāndogjopanishad kritisch herasugegeben und übersetzt. Leipzig.
Böhtlingk, O. (1891). Zu den von mir bearbeiteten Upanishaden. BKSGW,43, 70–85.
Böhtlingk, O. (1897). Bemerkungen zu einigen Upanishaden. BKSGW,49, 78–100.
Brereton, J. (1996). Yajnavalkya’s Curse. SII,20, 47–57.
BSBh = Shāstrī Paṇsīkar.
Delbrück, B. (1878). Die altindische Wortfolge aus dem Śatapathbrāhmaṇa dargestellt. Halle: Verlag der Buchhandlung des Waisenhauses.
Della Casa, C. (Trans.). (2000 [1976]). Upaniṣad vediche. Milano: Utet.
Deussen, P. (Trans.). (19213 [1897]). Sechzig Upanishads des Veda, aus dem Sanskrit übersetzt und mit Einleitungen und Anmerkungen Versehen von Dr. Paul Deussen. Leipzig: F.A. Brockhaus.
Duperron, A. (Trans.). (1801). Oupnek’hat, id est, secretum tegendum. Tomus I. Argentorati, Paris.
Eggeling, J. (Trans.). (1897). Satapatha Brahmana Part IV (Sacred Books of the East) (Vol. 43). Oxford: Clarendon Press.
Falk, M. (1937). Upāsana et Upaniṣad. Rocznik Orjentalistyczny,13, 129–158.
Falk, H. (1986). Vedisch upanisad. ZDMG,136, 80–97.
Geldner, K. F. (1928). Vedismus und Brahmanismus. Tübingen: Verlag von J.C.B. Mohr ‘Paul Siebeck’.
Gren-Eklund, G. (1978). A study of nominal sentences in the oldest Upaniṣads. Uppsala: Almqvist & Wiksell.
Hock, H. H. (2015). Some issues in Sanskrit syntax. In P. M. Scharf (Ed.), Sanskrit syntax: Selected papers presented at the seminar on Sanskrit syntax and discourse structures, 13–15 June 2013, Universite´ Paris Diderot, with an updated and revised bibliography by Hans Henrich Hock (pp. 1–52). Providence: The Sanskrit Library; New Delhi: D. K. Printworld.
Huang, B. (Trans.). (2010). Aoyishu [Upaniṣads]. Beijing: Shanwu Yinshuguan.
Hume, R. E. (Trans.). (1921). The thirteen principal Upanishads, translated from the Sanskrit with an outline of the philosophy of the Upanishads and an annotated bibliography. Oxford: Oxford University Press.
Ickler, I. (1973). Untersuchungen zur Wortstellung und Syntax der Chāndogyopaniṣad. Goppingen: Verlag Alfred Kummerle.
Keith, A. B. (1925). The religion and philosophy of the veda and upanishads. Cambridge: Harvard University Press.
Killingley, D. (2018). Knowledge and liberation in the Upaniṣads. In S. Cohen (Ed.), The Upaniṣads, A Complete Guide (pp. 133–145). London & New York: Routledge.
Kobayashi, M. (2013). Information structure and the particles vái and evá in Vedic prose. In J. S. Klein & K. Yoshida (Eds.), Indic across the millennia: From the Rigveda to modern Indo-Aryan; 14th world Sanskrit conference, Proceedings of the Linguistic Section, Kyoto, Japan, 1–5 September 2009 (pp. 77–92). Bremen: Ute Hempen Verlag.
Köhler, H.-W. (1973). Śrad-dhā in der vedischen und altbuddhistischen Literatur. Wiesbaden: Steiner.
Lüders, H. (1940). Zu den Upanisads: I. Die Samvargavidya. In Philologica Indica: Ausgewahlte kleine Schriften von Henrich Luders (pp. 361–390). Gottingen: Vandenhoeck & Ruprecht.
Max Müller, F. (Trans.). 1879. The Upaniṣads. Part I (Sacred Books of the East vol.1). Oxford: Clarendon Press.
Max Müller, F. (Trans.). (1884). The Upanishads. Part II (Sacred Books of the East, Vol. 15). Oxford: Clarendon Press.
Oertel, H. (Ed., Trans.). (1896). The Jaiminīya or Talavakāra Upaniṣad Brāhmaṇa. Journal of the American Oriental Society, 16, 79–260.
Oldenberg, H. (1896). Vedische Untersuchungen. Zeitschrift der Deutschen Morgenländischen Gesellschaft,50(3), 423–462.
Olivelle, P. (1998). The early Upanisads: Annotated text and translation. Oxford: Oxford University Press.
Senart, E. (Trans.). (1930). Chāndogya Upaniṣad. Paris: Les Belles Lettres.
Shāstrī Paṇsīkar, W. L. (Ed.). (1915). Brahmasūtrabhāshya (Text with Tippaṇis.). Bombay: Nirnaya-Sagar.
Slaje, W. (Trans.). (2009). Upanischaden: Arkanum des Veda. Leipzig: Verlag der Weltreligionen.
Speijer, J. S. (1886). Sanskrit syntax. Leiden [Reprint 1973. Delhi: Motilal Banarsidass].
Speijer, J. S. (1896). Vedische und Sanskrit Syntax. Strassburg: Trübner [Reprint 1974. Graz: Akademische Druck- u. Verlaganstalt].
Syrkin, (Trans.). (1992). Upanishady (Vol. 3). Moskwa: CCR Academy of sciences, Oriental Institute.
Thieme, P. (1966). Upanischaden: Ausgewählte Stücke. Stuttgart: Reclam.
Visigalli, P. (2020, forthcoming). The Gārgya-Ajātaśatru episode in Bṛhadāraṇyaka Upaniṣad 2.1. Zeitschrif der Deutschen Morgenländischen Gesellschaft.
Whitney, D. W. (1890). Bohtlingk’s upanishads. The American Journal of Philology,11(4), 407–439.
Xu, F. (Trans.) (1995). Wushi Aoyishu [Fifty Upaniṣads]. Beijing: Zhongguo shehui keshe chuban [Chinese Academy of Social Sciences Press].
Acknowledgements
I would like to thank Yūto Kawamura for several detailed comments on an earlier draft of this paper. Many thanks go to Joel Brereton and Patrick Olivelle for patiently going through a chaotic early draft and for offering valuable suggestions as to the style of exposition. I am indebted to the anonymous reviewer and to Diwakar Acharya for several corrections and requests to clarify my arguments. Thanks also to the fellow participants in ‘The 1st Hiroshima Sanskrit Meeting’ (December 2019) for their stimulating comments and corrections. Finally, I thankfully acknowledge Xuan Li’s help in creating the graph at the end of Part II.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
The author declares that he has no conflict of interest.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendices
Appendix 1: The Formula ‘ya eta- evaṃ vidvān (…) upa √ās’.
This formula often occurs in CU, but is absent from BU. It is typically accompanied by a description of the benefits that are said to derive from the act of upa √ās (see “Instances (2.3) Follow the Marked Order ‘OP O [=ritual item]’” section). With respect to the syntax of this formula, the issue concerns whether eta- should be construed with (i) vidvān or (ii) with upa √ās:
- (i)
‘ya | eta- (O) evaṃ (OP) vidvān | (…) upa √ās’. In this parsing, ‘eta- evaṃ vidvān’ constitutes one syntactic unit. eta- is a pronoun and is the O of vidvān. ‘Who, knowing this thus, upās-es…’.
- (ii)
‘ya eta- (O) | evaṃ vidvān | (…) upa √ās’. In this parsing, ‘evaṃ vidvān’ constitutes one syntactic unit. eta- is construed with upa √ās. It is either a pronoun, which functions as the O of upa √ās, or a pronominal adjective that qualifies the O of upa √ās. ‘Who, knowing thus, upās-es this…’.
In other Vedic texts such as BU, we find the phrase ‘evaṃ vidvān’, without eta-; (cf. Bodewitz 1973, p. 253, note 25). This suggests that ‘evaṃ vidvān’ constitutes one syntactic unit; hence, analysis (ii) is preferable.
Heuristically, it seems to me more appropriate to test whether all occurrences of the formula can be analyzed by means of one analytic model, rather than analyzing them variously, as is generally done by most translators. Excepting instance CU 1.1.7 (=1.1.8; 1.2.14), all other occurrences of the formula can be analyzed by means of (ii). We have the following cases:
- (a)
eta- is a pronoun; e.g. CU 1.9.4 (=4.11.2–13.2(6x)) sa ya etaṃ (O) evaṃ vidvān upāste parovarīya eva hāsyāmuṣmiṃ loke jīvanaṃ…, ‘who, knowing thus, upās-es this [=udgītha], he will have the most extensive life in this world…’.Footnote 39
In the other three cases, eta- is a pronominal adjective that qualifies the O in instances (2.1), (2.3b), and (3.3). Below I highlight eta- and the O it refers to. Note the hiatus that separates eta- from the O-s.
- (b)
(2.1) (‘O “OP” iti’); e.g. CU 3.19.4: sa yaetamevaṃ vidvānādityaṃ (O) brahma (OP) ity upāste ‘when one, knowing thus, upās-es this sun as “brahman”’. See also CU 1.3.7; 2.1.4; 4.5.2–8.4 (8x).
- (c)
(2.3b) (‘OP [=locative] O’); e.g. CU 2.2.3: yaetadevaṃ vidvāṃl lokeṣupañcavidhaṃ sāmaupāste ‘who, knowing thus, in/as the [five] worlds upās-es this fivefold Sāman’. See also CU 2.2.3, 3.2, 4.2, 5.2, 6.2, 7.2, 8.3.Footnote 40
- (d)
(3.3) ‘OP OP1 (ca) O’; e.g. CU 2.10.6: yaetadevaṃ vidvān ātmasaṃmitam atimṛtyusaptavidhaṃ sāmaupāste ‘who, knowing thus, as self-measured [and as] death-transcending upās-es this sevenfold Sāman’.
One exception to the model adopted here is CU 1.1.7 (=1.1.8; 1.2.14). In this occurrence, the adjectival eta- qualifies the OP (akṣara, neuter), not the O (udgītha, masculine): yaetadevaṃ vidvānakṣaram (OP) udgītham (O) upāste ‘who, knowing thus, upās-es as this syllable [= oṃ] the High Chant’.
This exception may be explained in relation to CU 1.1.1: om ityetad akṣaram (OP) udgītham (O) upāsīta ‘as this syllable “om” one should upās the High Chant’. I think that this occurrence, in which the pronominal eta- qualifies akṣara (OP), has influenced CU 1.1.7. Note that eta- seems to have different functions in CU 1.1.1 and in the formula ‘ya eta- evaṃ vidvān (…) upa √ās’. In the latter, eta- means ‘this’ in the sense of ‘as described in the passages above’. In the former, this sense is inapplicable, because CU 1.1.1 marks the beginning of CU. Rather, eta- seems to qualify the syllable ‘oṃ’ as ‘that well-known’ (prasiddha).Footnote 41
Finally, the analytic model proposed here bears on how we analyze one problematic instance, CU 1.9.2. Two variant readings are attested: ya etad/etam evaṃ vidvān parovarīyāṃsam udgītham upāste. If we accept the analytic model proposed above, then only the reading etam is correct, because eta- qualifies the O taken by upa √ās, the masculine udgītha; (cf. Ickler 1973, pp. 69–70). Accordingly, CU 1.9.2 should be translated as ‘who, knowing thus, upās-es this more extensive High Chant’.
Appendix 2
The following table gives all the instances of upa √ās in BU and CU. The first column lists the instances (‘instance’ is defined in “Introduction” section). The second column refers to the ‘group’ in which I classify the instances. The third column gives the ‘syntax’ or syntactic pattern; ‘[O]’ means that the O is implied. The fourth column gives the words that occur as the O and OP; in two instances (BU 2.1(b) (12x); CU 7.1.5–14.2 (28x), I use ‘various’, since many different words occur. The fifth column, ‘notes’, provides additional relevant information. Since I have not been able to identify reliable principles to analyze instances (1.3), I only give the ‘instance’ and ‘group’. A question mark ‘(?)’ indicates that I am not certain about one aspect of my analysis. Take as CU 1.3.8 an example. While I am confident that this instance takes an implied O and it follows a syntactic variation of the pattern of instances (2.1) ‘O “OP” iti’, I am not certain whether this implied O precedes (‘[O] “OP” iti’) or follows (‘“OP” iti [O]’) the OP. I therefore add ‘(?)’ in the column ‘syntax’. I have been conservative in my estimate: I add a question mark whenever I am not certain about an aspect of my analysis.
BU
Instance | Group | Syntax | O, OP | Notes |
|---|---|---|---|---|
1.4.7(a) | (1.3) | This instance is discussed in Part II: “Instances (1.2): OP is nominative-iti; O is implied, ‘[O] “OP” iti’” section | ||
1.4.7(b) | (1.2) [=(2.1)] | ‘“OP” iti eva [O]’ or ‘[O] “OP” iti eva’ (?) | O: ātman OP: ātman | |
1.4.8(2x) | (2.2) | ‘OP eva O’ | O: priyaṃ OP: ātman | priyaṃ is a substantive neuter Marked order is triggered by eva |
1.4.10 | (1.3) | |||
1.4.11 | (1.1) | O | O: kṣatriya | upa √ās is construed with adhastād; cf. BU 6.4.2(2x) |
1.4.15(2x) | (2.2) | ‘OP eva O’ | O: loka OP: ātman | Marked order is triggered by eva |
1.5.2 | (1.1) | O | O: etad [sādhāraṇam annaṃ ‘the food common to all’] | Instance occurs in a commentary on a śloka. On the connection between ślokas and instances (1.1), see Part II: “Instances (1.1): upa √ās has an Idiosyncratic Meaning” section |
1.5.13(2x) | (2.2) | ‘O OP’ | O: etān [= vāc; manas; prāṇa] OP: antavat; ananta | Though this instance occurs in a commentary on a śloka, the verse that is being commented upon is separated from the other verses and is commented in greater detail |
2.1(a)(12x) | (2.2) | ‘OP eva O’ (?) | O: brahman OP: etaṃ [= puruṣa] | Marked order is triggered by eva. Note the interposition of the emphatic subject aham between OP and O: etam (OP) eva aham brahma (O) upāsa iti |
2.1(b)(12x) | (2.1a) | ‘“OP” iti O’ | O: various, i.e. asau vidyuti puruṣaḥ OP: various, e.g. tejasvin | Marked order is due to the context being that of a debate |
2.1(c)(12x) | (2.2) | ‘O OP’ | O: etam [=puruṣa] OP: evaṃ | For evaṃ as OP, see CU 5.12.2–17.2(6x) and CU 5.18.1 |
4.1(a)(6x) | (2.1) | ‘“OP” iti O’ | O: enad/etad [= brahman] OP: prajñā; priyam; satyam; anantam; ānanda; sthiti | Marked order is due to the context being that of a debate. |
4.1(b)(6x) | (1.3) | |||
4.4.10 | (1.1) | O | O: avidyā | Instance occurs in a śloka. |
4.4.16 | (3.2) [=(2.2)] | ‘O OP OP1’ | O: tad [=*tam, ātman] OP: jyotir OP1: āyur…’mṛtam | Instance occurs in a śloka. Note that part of OP1 (’mṛtam) follows upa √ās. tad (O) devā jyotiṣāṃ jyotir (OP) āyur hopāsate ’mṛtam (OP1) |
5.8.1 | (2.2) | ‘O OP’ | O: vāc OP: dhenu | The O is the first word in the sentence; cf. CU 3.18.1. |
6.2.15 | (2.2) | ‘O OP’ | O: śraddhā OP: satya | |
6.4.2(2x) | (1.1) | O | O: strī | upa √ās is construed with adhas; cf. BU 1.4.11 |
CU
Instance | Group | Syntax | O, OP | Notes |
|---|---|---|---|---|
1.1(3x) | (2.3a) | ‘OP [=accusative] O’ | O: udgītha OP: akṣara [= om] | Only instance in which eta- is construed with OP, not with O; see Appendix 1. |
1.2.2–6 (5x) | = | = | O: udgītha OP: nāsikya prāṇa; vāc; cakṣus; śrotra; manas | |
1.2.7 | = | = | O: udgītha OP: tam [= mukhya prāṇa] | |
1.2.10–12(3x) | = | = | O: udgītha OP: tam [= mukhya prāṇa] | |
1.2.14 | = | = | O: udgītha OP: akṣara [= om] | |
1.3.1 | = | = | O: udgītha OP: tam [= ya eva asau tapati, i.e. āditya] | |
1.3.2 | (3.3) [=(2.3a)] | ‘OP OP1ca O’ | O: udgītha OP: imam [=mukhya prāṇa] OP1: amum [=āditya] | |
1.3.3–5(2x) | (2.3a) | ‘OP eva O’ | O: udgītha OP: vyāna | Unlike what happens with instances (2.2) (‘O OP’ vs ‘OP eva O’), eva does not disrupt the usual syntactic order followed by instances (2.3a), i.e. ‘OP O’. |
1.3.8 | (1.2) [=(2.1)] | ‘“OP” iti [O]’ or ‘[O] “OP” iti’ (?) | O: sāman; ṛc; devatā; chandas; stoma; ātman OP: upasarāṇi | |
1.5.3 | (2.3a) | ‘OP O’ | O: udgītha OP: tam [= mukhya prāṇa] | |
1.9.2 | (1.3) (?) | Variant reading etam/etad; see Appendix 1. | ||
1.9.4 | (1.3) | |||
2.1.4 | (2.1c) | ‘O “OP” iti’ | O: etad…sādhu OP: sāman | Unlike most other instances which exhibit a link between the O and the benefits said to result from the act of upa √ās, this instance exhibits a link with the O. I call the former ‘OP-oriented benefits’ and the latter ‘O-oriented benefits’; see “Instances (2.3) Follow the Marked Order ‘OP O [=ritual item]’” section |
2.2.(a)(7x) | (2.3b) | ‘OP [=locative] O’ | O: pañcavidhaṃ sāma OP: lokeṣu (2.2.1); vṛṣṭau (2.3.1); sarvāsu āpsu (2.4.1); ṛtuṣu (2.5.1); paśuṣu (2.6.1); prāṇeṣu (2.7.1); vāci (2.8.1). | In occurrence 2.7.1, the O is pañcavidhaṃ parovarīyaḥ sāma. In occurrence CU 2.8.1, the O is saptavidha sāman. |
2.2(b)(7x) | = | = | O: etad…pañcavidhaṃ sāma OP: lokeṣu (2.2.3); vṛṣṭau (2.3.2); sarvāsu āpsu (2.4.2); ṛtuṣu (2.5.2); paśuṣu (2.6.2); prāṇeṣu (2.7.2); vāci (2.8.3). | In occurrence 2.7.2, the O is pañcavidhaṃ parovarīyaḥ sāma. This occurrence exhibits ‘O-oriented benefits’; cf. CU 2.1.4. In occurrence CU 2.8.3, the O is saptavidha sāman. |
2.9 (2x) | (2.3a) | ‘OP O’ | O: saptavidham sāma OP: amum ādityaṃ | Only instance (2.3a) in which O is not udgītha. |
2.10.1 | (3.3) [=(2.3a)] | ‘OP OP1 O’ | O: saptavidhaṃ sāma OP: ātmasaṃmitam OP1: atimṛtyu | |
2.10.6 | = | = | O: etad…saptavidhaṃ sāma OP: ātmasaṃmitam OP1: atimṛtyu | |
2.21.4 | (2.1c) (?) | ‘“O OP” iti’ | O: sarvam OP: asmi | |
3.13.1 | (3.1) [=(2.1)] | ‘O “OP OP1” iti’ | O: etat [< *etaṃ = prāṅ suṣiḥ] OP: tejas OP1: annādya (n.) | *etaṃ becomes etat, because of gender attraction with OP and OP1. |
3.13.2 | (3.1) [=(2.1)] | ‘O “OP ca OP1ca” iti’ | O: etat [< *etaṃ = dakṣiṇaḥ suṣiḥ] OP: śrī OP1: yaśas | |
3.13.3 | (3.1) [=(2.1)] | ‘O “OP OP1” iti’ | O: etat [< *etaṃ = pratyaṅ suṣiḥ] OP: brahmavarcasa OP1: annādya (n.) | |
3.13.4 | (3.1) [=(2.1)] | ‘O “OP ca OP1ca” iti’ | O: etat [< *etaṃ = udaṅ suṣiḥ] OP: kīrti OP1: vyuṣṭi | |
3.13.5 | (3.1) [=(2.1)] | ‘O “OP ca OP1ca” iti’ | O: etat [< *etaṃ = ūrdvaḥ suṣiḥ] OP: ojas OP1: mahas | |
3.14.1 | (2.1a) | ‘O “OP” iti’ | O: tad [= brahman] OP: jalān | Only instance in which upa √ās is accompanied by a predicative of the subject (śānta) (tajjalān iti śānta upāsīta) |
3.18.1 | (2.1.b) | ‘O “OP” iti’ | O: manas OP: brahman | The O is the first word of the sentence; cf. BU 5.8.1 |
3.19.4 | (2.1b) | ‘O “OP” iti’ | O: etam…ādityam OP: brahman | |
4.2.2 | (2.2) | ‘O OP’ | O: yām OP: devatām | |
4.5.2–8.4 (8x) | (2.1a) | ‘O “OP” iti’ | O: etam…catuṣkalam pādam brahmaṇaḥ OP: prakāśavān; anantavān; jyotiṣmān; āyatanavān | |
4.11.2–13.2 (6x) | (1.3) | |||
5.10.1 | (2.1c) (?) | ‘O “OP” iti’ (?) | O: śraddhāṃ OP: tapas | Variant reading: śraddhā. |
5.10.3 | = | = | O: iṣṭāpūrte OP: dattam | |
5.12.1–17.1(a)(6x) | (2.2) | ‘O OP’ or ‘OP O’ (?) | O: kaṃ OP: ātman | Only instances in which O/OP is kim. |
5.12.1–17.1(b)(6x) | (2.2) | ‘O OP’ | O: yaṃ OP: ātman | For ya- as O, cf. CU 4.2.2. |
5.12.2–17.2 (6x) | = | = | O: ātman vaiśvānara OP: evam | |
5.18.1 | (2.2) (?) | ‘O OP’ (?) | O: etam…ātmanaṃ vaiśvānaram OP: evam (?) | In my provisional analysis, O is qualified by two adjectives (prādeśamātram abhivimānam). |
5.24.5 | (1.1) | O | O: agnihotra | |
7.1.4–14.1 (14x) | (1.3) | |||
7.1.5–14.2 (28x) | (2.1) | ‘O “OP” iti’ | O: various OP: brahman | |
8.12.6 | (1.3) |
Rights and permissions
About this article
Cite this article
Visigalli, P. Quisquis Deum intellegit, Deus Fit: The Syntax of upa √ās in the Bṛhadāraṇyaka and Chāndogya Upaniṣad. J Indian Philos 48, 191–228 (2020). https://doi.org/10.1007/s10781-020-09418-1
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10781-020-09418-1