somananda e-book

vision of a Śaiva theologian and philosopher whose arguments are more than polemical remonstrations of the grammarians and the goddess-worshipers who were inspired by them. For Somānanda’s is a novel and constructive theological vision that is based in the Trika scriptures and influenced by the Krama, a tantric vision that offers a robust philosophical explanation of the nature of god, reality, and our full identity with both.

About the Edition and the Translation

16. The Manuscripts of the S´ivadr.s.t.i

M A N U S C R I P T S C O N S U LT E D

In addition to the readings of the KSTS edition of the text (Ked:), I have consulted six manuscripts in preparation of the present edition of the ŚD and ŚDVṛ, chapters one through three.

T

This is the manuscript of Trivandrum University, number 5854-H. The manuscript is a palm-leaf text in fair to poor condition. It is frayed at the edges and shows pieces of text broken off at the tops and bottoms of the folios. It is wormeaten in a number of places. The text is written in Grantha Malayalam script in a clear hand, with approximately 95 characters per line, and it regularly records 11 lines per folio. The text is written on both sides of the palm leaf. The manuscript records the mūla only, with the exception that it also records the first of the three invocatory verses of the commentary, but it breaks off in the middle of ŚD 7.50a, where the scribe stopped copying the text. The readings are largely, but not entirely, free of errors and corruptions.

C

This is the manuscript of the Calcutta Sanskrit College, CS 3, 153. The text is entitled The Śivadṛṣṭi of Durvāsamuni. The manuscript is written in devanāgarī script and is made of country paper measuring approximately 6 inches by 9 inches. The text is written on both sides of the paper, recording 40–50 characters per line, 12 lines per side. Numbering 21 double-sided folios in length, it includes the complete text of the mūla only. The maṅgala reads: śrīgaṇeśāya namaḥ. The first folio of the manuscript (folio 1r.) reads the following in large, centered script: ślokasaṃkhyā: 825 // atha śivadṛṣṭi. prāraṃbhaḥ. pṛṣṭhasaṃ(khyā):

18 // adhyāya: 1-7. (Thus, the MS begins its record of the text on folio 1v.) The MS is in good condition but is very corrupt and is full of lacunæ.

G

This is the Göttingen manuscript. Niedersächsische Staatsund Universitätsbibliothek, number: Cod.Ms.Sanscr.Vish 5 (11). It is a paper manuscript written in śāradā script. Judging from the microfilmed copy provided by the library, it measures approximately 8 inches by 10 inches. I am unsure as to whether or not the text is written on both sides of the folios, as I have only seen the microfilmed copy. It records 16 lines per folio and 17-21 characters per line. Numbering 150 folios in length, it includes Somānanda’s mūla as well as Utpaladeva’s commentary, beginning with ŚD 1.1 and up to the commentary on 4.73cd–75. The maṅgala reads: (Auspicious Symbol) svasti. śrīdevyai siddhidātryai namaḥ. śrīgaṇapataye namaḥ. oṃ. The MS is in good condition, and its readings are mostly correct.

J

This is the Jammu manuscript, owned by the Raghunāth Mandir Library, Jammu. The text is written in śāradā script on birchbark, measuring approximately 5 1/2 inches by 8 inches per folio. Numbering 142 folios in length, the manuscript is written in a clearly legible hand, but is fraying at the edges and is very fragile. It records approximately 17 characters per line and 18 lines per folio. The manuscript includes the mūla and Utpaladeva’s commentary up to ŚDVṛ ad ŚD 4.73cd–75, where the text ends at precisely the same point as does the commentary in the KSTS edition. The maṅgala reads simply: oṃ. The MS is very correct in its readings. It includes some marginal notes, particularly in the first chapter, which are also written in śāradā script. It marks intermittently, with an unidentified mark, the completion of portions of the commentary, or of portions of the mūla, particularly in the third chapter. Stein (1894: 225) has suggested that the manuscript may be dated to (Vikrama)Saṃvat 1680 (= 1624/5 C.E.).

P

This is the Pune manuscript. Bhandarkar Oriental Research Institute, number: 805 of 1891–95. The manuscript is written in devanāgarī script and is made of country paper measuring approximately 8 inches by 12 inches. The text is written on only one side of the paper, recording 25–30 characters per line, 27 lines per folio. Numbering 56 folios in length, it includes the mūla and Utpaladeva’s commentary, beginning with ŚD 1.1 and up to the first line of the commentary on ŚD 4.64cd–66. The maṅgala reads: śrīgaṇeśāya namaḥ. oṃ. The MS is in excellent condition. It is also witness to a number of corrections written in a later hand in what appears to be a ballpoint pen. These corrections are added to those of the copyist, who makes a number of corrections to the manuscript, apparently

in the course of copying the work. Subsequent to the corrections recorded in ballpoint pen, the readings of the manuscript mostly support those of Ked:.

R

This is the manuscript of the Rajasthan Oriental Research Institute, Jodhpur. The text is recorded in devanāgarī script on country paper measuring approximately 5 1/2 inches by 8 inches. The text is recorded on both sides of 44 double-sided folios in a clear hand and regularly records 21 characters per line and 24 lines per folio in a style whose consistency resembles that of printed text. The mūla as well as Utpaladeva’s commentary are recorded in full to the commentary on ŚD 4.64cd–66, where it breaks off mid-sentence. It opens with the maṅgala: oṃ śrīgaṇeśāya namaḥ. The manuscript is in excellent condition and its readings are reliably correct.

O T H E R ŚIVADṚṢṬI M A N U S C R I P T S

Aside from these six manuscripts, two other manuscripts also have some bearing, albeit indirectly, on the present edition. These are the manuscripts consulted by Kaul in the production of the KSTS edition (Ked:), including the following:

1.The Srinagar manuscript. This is a śāradā manuscript of the ŚD and ŚDVṛ in the Research Library, Srinagar. The details of this manuscript are recorded in the introduction to the KSTS edition of the text (see Kaul edition 1934: i–ii). It would be highly desirable to see this manuscript, but all of my efforts to obtain a copy failed, due in no small part to the current political instability in the Kashmir Valley.

2.The second is the manuscript of the Government Oriental Research Library, Madras. This begins on folio 79a of manuscript number 15323 of the collection. It is a complete transcription of the ŚD, without the commentary. When I went to view this manuscript in the Government Oriental Research Library, I was shown a work written in Telugu script that was copied into a modern, twentieth-century ruled and bound “copy-book.” The manuscript is in excellent condition, but obviously it is quite a recent production. Seeing that it is clearly a late copy, I decided not to collate the readings found therein. It is worth noting, however, that Kaul, in his introduction to Ked:, mentions that he based his edition on two manuscripts, the aforementioned Srinagar manuscript and the transcription of a copy of the text housed in the Egmore Manuscripts Library, Madras. (See Kaul edition 1934: i.) It is not possible that the modern copybook manuscript is the one to which Kaul refers in Ked:, as his report states that the manuscript in question is a source text of the ŚD that is housed in the Egmore Library, and is not merely a copy of it. I was not able to locate any premodern

manuscript in the Madras Library collection, however, after visits to both the Egmore Manuscripts Library located near the Egmore train station in Madras and the Government Oriental Research Library on the campus of the University of Madras on a research visit in 2003.

Apart from these is an additional pair of manuscripts that merit mention. First, there is one manuscript that at first glance appears, on the basis of the information provided in the published catalogue, to be relevant to the present study, but in fact is not related at all to Somānanda’s work. This is the manuscript of the Bodleian Library, Oxford University. The manuscript is listed as item number 168 in Aufrecht’s Catalogus Codicum Manuscriptorum Sanscriticorum Bibliothecoe Bodleianoe (p. 108). Though listed as the Śivadṛṣṭi, this short excerpt bears no resemblance whatsoever to Somānanda’s text, and it appears to be another work bearing the same name as the work under consideration in the present volume. Second, mention should be made of the manuscript of the Adyar Library, Accession Number 67455. This is a partial manuscript of the ŚD, without the commentary. The manuscript is written on palm leaves some 60–65 characters per line, 9 lines per folio. Writing appears on both sides of the palm leaves. The manuscript is written in Telugu script in a clear hand, but the manuscript is very incomplete and terribly damaged by worm-holes and tearing, and the brittle palm leaves are broken in many places, which has resulted in the loss of many pieces of text. Due to its incomplete and thoroughly fragmented condition, I have not collated this manuscript for the present edition.

17. About the Edition

T H E R E L AT I O N S H I P O F T H E M A N U S C R I P T S

Although all six of the manuscripts consulted include readings of the verses of the ŚD, only four manuscripts of the ŚDVṛ are included in the present edition, as the manuscript of Calcutta Sanskrit College (C) and that of Trivandrum University (T) record none of the commentary. The Jammu manuscript (J) and the Göttingen manuscript (G) regularly share similar or identical variant readings. The Pune manuscript (P) and the Rajasthan manuscript (R), in turn, also share many variants. Not incidentally, P and R break off at nearly the same place in the commentary (in the first lines of the ŚDVṛ ad ŚD 4.64cd–66); and the readings of G and J also break off at a nearly identical place in the text, namely, in the middle of the commentary on ŚD 4.73cd–75 at precisely the point at which the commentary published in Ked: abruptly ends. These similarities—the regular coincidence of shared variant readings and missing passages of text—suggest that one may identify two pairs of manuscripts, that of P and R, on the one hand, and that of G and J, on the other. (We shall deal with a third pair, that of T and C, below.) The published edition (Ked:), itself based on the Srinagar manuscript

and a copy of a Madras manuscript purportedly housed in the Government Oriental Research Library, both of which are currently unavailable to me, attests to a state of the text that is regularly more similar to the latter pair than the former.

Now, an examination of the variant readings of the manuscripts suggests that the major differences between P and R, on the one hand, and G and J, on the other, involve corruption of the text over time. For example, P and R offer sevinṛpādi for the sevitanṛpādi of G and J, with Ked: agreeing with the latter pair, this on ŚDVṛ ad ŚD 4.4–5, the reading of G and J being the source of that of P and R.221 Many other examples could be cited in addition to this one.222 The readings of P and R, in other words, are regularly, though not always, shown to be corruptions of the readings found in G and J.

The relative chronology of the four manuscripts in question further suggests that one should reasonably expect the readings of G and J to be the earlier ones. It is, firstly, clear that J is the oldest manuscript examined that includes the com- mentary—it is written on birchbark of obvious antiquity, and Stein (1894: 225) has suggested it may be dated to (Vikrama)Saṃvat 1680 (= 1624/5 C.E.). This is indeed a fairly old manuscript, then, as far as such matters go in South Asia. G, for its part, is a paper manuscript in śāradā script that, while apparently of some antiquity, is clearly of a more recent provenance than is J. P and R, on the other hand, are paper manuscripts written in devanāgarī, with P being of an obviously recent historical pedigree, it being copied on only one side of the folios and being orthographically of a style that is quite modern (as is, to a lesser extent, R). P and R, moreover, display evidence of a text that was at some point transmitted from (presumably older) śāradā manuscripts to devanāgarī ones. In sum, while the age of a manuscript in no way guarantees the antiquity of the text to which it attests, the direction of transmission suggested by the variant readings is confirmed by the relative chronology of the manuscripts themselves, with P and R regularly showing themselves to witness variant readings that must be corruptions of the readings found in G and J. In general, then, I take as a first principle of editing the ŚDVṛ that the readings of G and J—the latter in particular, for the reasons to be stated, below—regularly attest to an earlier state of the text and are therefore very often, though not always, superior to those of P and R.

221The former variant is unlikely because the context demands that it is a king who is served, not a king who is a sevin, or servant; and while it is possible that the compound could be read as a dvandva, listing all parties involved in a master-servant relationship, one rather expects, given the context, that the compound should refer only to the person served. It is therefore highly likely that the reading of P and R is a corruption of the more original reading of G and J.

222These include, for example, the following. G and J witness saḥ, with Ked:reading the correct sa, while P and R omit the pronoun altogether in ŚDVṛ ad ŚD 1.1; G and J, along with Ked:, read viśeṣaṇakalāpo in ŚDVṛ ad ŚD 1.2, while P and R read viśeṣeṇa kalāpo; G, J, and Ked:read susūkṣmaśaktitritaya in ŚD 1.4a, while P and R erroneously read svasūkṣmaśaktitritaya ; G and J read tadupaśamamātraṃ in ŚDVṛ ad ŚD 2.4cd–5, which is transformed into tadupadeśamātraṃ in P, R, and Ked: , the latter being easily the less favorable reading of the pair, given the context; G and J, confirmed by Ked:, read apūrvakāryābhāvāt, while P and R drop the negative prefix (a-) before bhāva and erroneously read apūrvakāryabhāvāt on ŚDVṛ ad ŚD 3.51cd–53ab; and so on.

As for the manuscripts of the root text (mūla) of the ŚD itself, the Trivandrum manuscript (T), which as mentioned records none of the ŚDVṛ, is clearly an artifact of some antiquity, though the precise date of the manuscript cannot be specified, given the speed with which South Indian palm-leaf manuscripts deteriorate.223 One can only guess, based on the appearance of the palm leaves and for orthographic reasons, that T is close to the age of, but later than, the birchbark manuscript (J) that has been dated to the early seventeenth century. The Trivandrum manuscript certainly records, in many instances, the best reading of the mūla available in any of the witnesses examined for the present edition. For example, on ŚD 2.28a, T and C (about which more will be said in a moment) read satyarūpā, while G, J, P, R, and Ked:read satyarūpān.224 This is not to say, however, that T always witnesses the best readings of the text: not only do worm holes and frayed edges litter the manuscript with lacunæ; but, more important, the manuscript shows signs of corruption, as is exemplified in, for example, ŚD 1.3c, where T reads tadecchā for the correct reading tadicchā,225 this being but one example of a number of variant and erroneous readings in T that are not found in J, G, R, or P.226 In sum, T not infrequently witnesses an earlier state of the text of the ŚD, and its variant readings are therefore often the preferable ones. Indeed, the variants of T are often preferable to readings found in G and/or J, and if either or both of the latter two manuscripts confirm the reading of T, then one must be very hesitant not to accept this reading. On the other hand, the corruptions in the manuscript require one carefully to scrutinize the variant

223See Gaur 1979: 12 (quoted in Goodall 1998: cxiii–cxiv).

224The latter reading appears, at first glance, to be the correct one, it being followed by the disjunctive particle (vā) and preceded by asatyān, suggesting the passage refers to a pair of possibilities, either the reality or unreality of the entities that paśyantī sees. However, Somānanda not infrequently uses vā as a connective particle (see, e.g. ŚD 1.32b, 1.42a, 2.29d, 2.30c, etc.), as he does here, tying thereby ŚD 2.28ab to what precedes it. To give preference to the more difficult reading—the oft-repeated maxim, lectio difficilior potior, of course has some truth in it—is to understand satyarūpā here to be the subject of the verb referring to paśyantī, ŚD 2.28ab therefore expressing the need to explain not why she sees real or unreal objects, but why she, being real, sees unreal ones. This not only follows elegantly from ŚD 2.27cd, where the unreality of the objects is considered, but it also leads beautifully into the lengthy consideration of the nature of nescience (avidyā) that follows the passage in question, beginning in ŚD 2.28cd–30ab, just as it accords with what I understand to be the correct reading of the commentary, where Utpaladeva suggests that the last part of the verse asks precisely the question as we, in the course of accepting the readings of T and C, have formulated it. The Vṛtti there reads: punar api cāsatyān arthān satyā sā paśyantī krīḍ- ādyabhāve ’pi kena prayojanaprakāreṇa sṛjatīti vimṛśyatām. (We must caution, however, that this reading of the commentary is itself based on the selection of a variant reading, that of Pa.c., which reads satyā sā for the satyān sā of G, J, Pp.c., R, and Ked:. The word order suggests this formulation of the text is rather

more appropriate, though one could also choose to preserve the reading of satyān and emend sā to vā if the readings of T and C were deemed unacceptable in ŚD 2.28a.)

225The latter reading is the correct one, first, because it is confirmed in the commentary, and, more importantly, because the correlative (tadā) recorded in T’s variant reading is attested by all the manuscripts (except C, which is corrupt) on ŚD 1.4d, rendering superfluous and awkward the purported presence of the same term in T’s reading.

226To offer but one additional example, found in ŚD 2.12d: T and C read vā for vāk, the latter, superior reading being attested in G, J, P, R, and Ked:.

readings found therein, as corruption and not just the presence of lacunæ often render the reading of T inferior to that of J and/or G, and even of R and/or P.

Related to T is the Calcutta manuscript (C), which records many of the variant readings and many of the corruptions found in T but not in the other four manuscripts. Like T, then, C is frequently corrupt, the only caveat being that it is significantly more corrupt and more riddled with lacunæ than T. Indeed, C most often appears in the apparatus for reasons of errors of mistranscription, and one has a sense that the copyist of C did not know Sanskrit well. Nevertheless, while C is clearly one of the most recently copied manuscripts of the six (with the Pune manuscript [P] probably being the most recently copied of the six, C being likely to precede it only by a short period of time), its readings, when not corrupt beyond recognition and when witnessing variants that diverge from both P and R, on the one hand, and G and J, on the other, regularly accord with those of T, which, as mentioned, often, though not always, contains readings that are likely closest of all the extant variants to the original text.227 In other words, C records a version of the text that comes to it from T, but it is significantly more corrupt than T.228

There are numerous instances where T and C record variants that are not shared by G, J, P, and R, as there are likewise many occasions when the latter four manuscripts record variants absent from both T and C. This leads one to posit the existence of two recensions of the text, a Northern Recension comprising the four manuscripts—J, G, R, and P—and a Southern Recension witnessed by two manuscripts, T and C. It is further of note that the variant readings of T, often followed by C, more commonly accord with the earlier, śāradā manuscripts (J and G) of the Northern Recension than they do with the later, devanāgarī ones (R and P) when the readings of those two pairs diverge, this being some evidence for the relative antiquity of T, given that J and G predate R and P. There is, however, also some evidence of contamination in the C manuscript of the Southern Recension, as the Calcutta witness appears in places to record readings of the Northern Recension that are absent from T.229

227For example, both T and C record vā for the reading hi of the other manuscripts in ŚD 2.23a, the former being the reading that is probably closer to that of the original text, it being the less refined expression of the idea there articulated. On the other hand, both T and C erroneously record yāvat for tāvat in ŚD 2.20a.

228This may be known, moreover, by the fact that the readings of C do not diverge from those of T, except where the manuscript is corrupt or is contaminated by readings from the Northern Recension.

229Examples include: the erroneous reading in C of yasmā for yasmāt in ŚD 2.8c, where T reads tasmāt; in ŚD 3.6c, C accords with G, J, P, R, and Ked:in reading ’pi where T reads hi; C accords with the other manuscripts and Ked:in reading dvisatyatvam for the ’pi satyatva— of T in ŚD 2.69a; etc. More substantively, C accords with all the other manuscripts excepting T in ŚD 2.46a, reading svākya for the svāniḥ of T; C, all the manuscripts of the Northern Recension, and Ked:read svaśaktyāveśanātmaka in ŚD 3.17d, while T reads svaśaktyāvedanātmaka; C accords with G, J, P, R, and Ked:in reading hy upapadyate in ŚD 2.55d, where T reads dṛṣ<?>tā—; C accords with G, J, P, R, and Ked:in reading praviramyatām for T’s praviramyate in ŚD 2.80d; C records upāyatva in accordance with G, J, P, R, and Ked:over and against the reading of upāsatva found in T in ŚD 3.15c; etc.

There can be no doubt that T witnesses an older state of the text than does C, and the direction of transmission in the Southern Recension is indisputably one from T to C.230 As for the Northern Recension, three states of the text may be identified therein. J, as noted already, is the oldest of the manuscripts that record the commentary, and it also constitutes the earliest phase of the transmission of the commentary available to us. G generally accords with J, but it sometimes witnesses readings found in P and R.231 G and J, moreover, very regularly share readings that are not found in P and R, which very regularly share readings at variance with the former pair, as already indicated, above. G, then, likely constitutes a middle phase in the manuscript transmission, albeit one rather closer to what precedes it (J) than what follows (P and R), with P and R witnessing a still later, third and most recent phase of transmission of the manuscripts here consulted.

The readings of Ked:, in turn, sometimes accord with the variants found in T and C; and they sometimes accord instead with the readings of the Northern Recension. This suggests that one of the two manuscripts that Kaul consulted in producing the KSTS edition of the ŚD, either the Śrinagar manuscript or the Madras manuscript, attests to a state of the text that is close to that of T. Indeed, one may hypothesize that it is in the Madras manuscript that Kaul found such variant readings, with the Śrinagar manuscript more regularly attesting to the readings of G, J, P, and R.

There is, furthermore, some circumstantial evidence that Kaul’s two manuscripts offered divergent readings when the texts of the Northern and Southern Recensions differed. It is not infrequently the case that, when Ked: records a reading that accords with T (and C) over the manuscripts of the Northern Recension, the concordant reading in question may be found in the errata of the KSTS edition. In other words, Kaul sometimes records two divergent readings of the ŚD, one in the body of the edition, another in the list of errata at the end of the volume. And when he does so, it is frequently the case that the readings of the Northern Recension and the Southern Recension of our manuscripts each record one of the two divergent readings found in Ked:. This suggests, though not definitively, that Kaul sometimes had occasion to choose between two plausible variant readings, and in doing so sometimes recorded the reading that he chose

230This is a transmission, moreover, that must have taken place over time, with intermediary copies of the text standing between T and C; for there can be no doubt that C was copied from a devanāgarī manuscript and not one written in Grantha Malayalam, as exemplified in, e.g., ŚD 3.26c, where C erroneously records māvaśitvena for the bhāvaśivatvena of T, G, J, P, R, and Ked:, this error being one caused by the misreading of the devanāgarī mā for the devanāgarī bhā.

231For example: J reads etat tāvan in the ŚDVṛ ad ŚD 4.1–2ab, while G accords with P, R, and Ked: in reading etāvan. (There is some doubt, it should be added, as to whether to give the reading of Ked:any weight; for one cannot know with certainty whether the manuscripts consulted for the production of the KSTS edition truly witness the reading in question, or whether the reading is instead the product of the editor’s emendation.)

not to select in the list of errata. (Of course, this analysis of Ked: amounts to educated guesswork; and only an examination of the manuscripts can reveal what is hidden behind the silent, if useful, edition of Ked:.) For example, T, C, and Ked:p.c. read yādṛśī on ŚD 3.34b for the yāvatī of G, J, P, R, and Ked:a.c.. Other examples may be found, as well.232

Ked: otherwise may be said generally, though by no means always, to agree with G and J when they differ from P and R, as one would expect, given that the commentary recorded in the published edition breaks off at precisely the same point as it does in G and J. One also suspects, however, that the editor of Ked: has taken liberties in endeavoring to correct the text without informing the reader of his editorial decisions, as was already suggested, above. Perhaps the best example of this may be found in a passage of commentary ad ŚD 4.4–5, where Ked: reads śaktimattā , while all four manuscripts of the commentary read śaktisattā . Ked:’s reading certainly suggests an editorial choice on the basis of modern research into the principles of Kashmiri Shaiva philosophy, where the idea of Śiva as the possessor of the powers is frequently mentioned.

The high degree of congruence in the readings of G and J on the one hand, and P and R, on the other, suggests that we should be able to draw a stemma representing two recensions of the text. One would involve a line of direct transmission, from T to C in the Southern Recension of the ŚD. The other would involve a recension that diverged once, with J and G witnessing one reading of the text, R and P the other. Ked:, being an edition based on one southern and one northern manuscript, witnesses elements of both recensions. Yet, the picture is not so simple. To start, G (very occasionally) witnesses variants that diverge from common readings found in J, P, and R, ones that are not likely to have been the product of simple scribal errors, errors caused by metathesis, haplology, eye-skip, and the like.233 And, contrariwise, the readings of G sometimes accord with P and R, in contrast to those of J,234 though it must be added that lacunæ in J, resulting from damage to the manuscript, are by far the most

232Thus: T and C accord with Ked:p.c. on ŚD 3.3b, reading īdṛśān for the īhate of G, J, P, R, and Ked:a.c.; T, C, and Ked:p.c. read viśiṣṭakam for the viśeṣakam of G, J, P, R, and Ked:a.c. on ŚD 3.39d; on ŚD 3.90a, T, C, and Ked:read vadane for the vedane of G, J, P, and R; etc.

233A pair of examples is as follows: in ŚDVṛ ad ŚD 2.26cd–28ab, G, along with Ked:, correctly reads asatyasṛṣṭau for the unlikely asatyadṛṣṭau of J, P, and R; and in ŚD 3.86b, G reads saugandhyakṛt, while J, P, and R, along with T, C, and Ked:, read saugandhyavat.

234J often omits final visarga or transposes the short diphthong o for the short vowel a when it varies from G, P, and R, but these are minor differences. Some more significant examples of variants include the following. J omits eva in ŚDVṛ ad ŚD 1.1 (line 30 of the present edition), while G,P, R, and Ked:do not; the same is true of strīliṅgaḥ in ŚDVṛ ad ŚD 1.3–4 (line 63 of the present edition); J, along with Ked:, reads kāryavyatirekeṇa for kāryāvyatirekeṇa of G, P, and R (the omission of the alpha-privative being a recurring variant in J, though one that could possibly be explained away by scribal corrections of the error in J in subsequent states of the text); etc.