The Sogdian epitaph of Shi Jun and his wife

Abstract Three bilingual Chinese and Sogdian epitaphs have been discovered and published in recent years. The first of these forms part of the tomb of Wirkakk, alias Shi Jun 史君 “Mr Shi”, and his wife Wiyusi, who were buried in 580 ce in Chang'an, the capital of the Northern Zhou. This article provides some corrections to the previous editions of the Sogdian version and attempts to elucidate points which have so far appeared obscure.

having been found in situ and unearthed in a controlled archaeological excavation; the tomb which it accompanies is elaborately decorated with carved panels depicting the life and afterlife of the deceased, which have led to animated discussion concerning their Zoroastrian or Manichaean religious beliefs. 4 Amongst the problems are the fact that both the Chinese and the Sogdian texts seem to contain numerous errors and that neither can be described as a translation of the other, each providing a different selection from the same set of facts. The Sogdian text also has a few lacunae where the stone was broken, but these do not present a major problem. Yutaka Yoshida, the first editor of this text, already found plausible restorations for most of these lacunae, and some additional suggestions will be presented below.
My reading of the Sogdian text, which differs only slightly from Yoshida's, and my translation, which differs slightly more, may be presented first. without being aware (of it), a husband and wife see one 25 ʾyδ myδ wyrʾ-wδwh prwh Lʾ-ptzʾn another (for the first time) the same year, the same month, the same day, 26 prʾʾw wy-nʾnt ms kʾw ʾxwštmʾxw in the human world (and) also in paradise, 27 pr ʾyδ zmnʾ-ʾnγwth prʾyw ʾʾ(zw)n (so that) the beginning 28 βʾy of (their) life together (in each place) may be at the same period. 29 ʾkrty ywnʾk snkynʾk βγktʾk This stone tomb was made by 30 βrʾyšmnβntk δrymtβntk Vreshman-vande, Zhimat-vande (and) 31 prʾwtβntk wsʾʾn ʾBYʾ Prot-vande, desiring 32 mʾδryh sʾcʾw 5 wyʾʾkh a suitable place for (their) father (and) mother.

Commentary
The main purpose of this commentary is to justify the novelties in the text and translation given above. In general, matters which have already been fully explained by Yoshida will not be discussed. Lines 1-3. The day of the burial of Wirkakk and Wiyusi is given as year two of (the period) daxiang of the Great Zhou (dynasty), a rat year, the 23rd day of the first month. Yoshida, using the calendar table compiled by Chen Yuan (1956), calculates that this date corresponds to 23 February 580. 6 However, as Bill Mak has pointed out to me, this part of Chen's table is based on the calendar used by the southern dynasties based in Nanjing, whereas our inscription is explicitly dated according to the calendar of the Northern Zhou. According to the latter, the day concerned should rather be 24 February 580 (see the table in Xue and Ouyang 1956: 410).
Line 5. Yoshida (2005: 64;2016: 79) restores ( ʾ)[zw]ʾntk "living". However, this word corresponds to the English word "living" in the sense "alive" rather than in the required sense "dwelling, resident". No objection can be made to Yoshida's reading of the antepenultimate letter as -n-, but the execution of the Sogdian letters by the stonemason is so irregular (as can be seen from Yoshida's notes to the text) that it is not implausible to interpret it as a slightly careless variant of the similar letter -ʾ-. I therefore prefer to read and restore ( ʾ)[nc]( ʾʾ)tk, or perhaps ( ʾ)[n](cʾ)tk, the past participle of ʾncʾy "to stop, stay, reside". The use of the past participle in this sense is clearly attested in a passage from the martyrdom of St George: wdy wyʾq qw xwny trsʾq ʾncʾty sty "the place where that Christian is dwelling" (Hansen 1941: 11-12, lines 173-4).
[p]tβry[ δʾ]rt "he has the honour, holds the rank". Yoshida (2005: 64) seems to take it for granted that the form to be restored here is a 3 sg. preterite form with auxiliary verb -δʾrt and a meaning such as "he obtained", but he admits that "no suitable verb is known to me". However, δʾrt is also 3 sg. present of the verb "to have, hold", and the use of the present tense here with reference to the deceased would be no more surprising than that of nʾmt "he is named" in line 8. On the noun ptβr-, more commonly written pδβr-, and its meaning "honour, rank", see Yoshida 2019: 168. The spelling ptβr-does not seem to be attested elsewhere in Sogdian script, but is the expected equivalent of Manichaean Sogdian pṭfr-.
Lines 8-10. "He is named Wirkakk, the son of Wanuk, (namely) Wanuk, the son of Rasht-vande the sabao". The formulation of the genealogy, with the repetition of the name of the father, is reminiscent of that found in the late Old Persian inscriptions, e.g. "I am Darius . . ., son of King Artaxerxes, (namely) Artaxerxes, the son of King Xerxes, (namely) Xerxes, the son of King Darius" (Darius II, Hamadan inscription, see Schmitt 2009: 183-4). The name wyrkʾk is clearly a hypocoristic derived from wyrk-"wolf". I transcribe it as Wirkakk, since -akk (often, though not here, spelled with double kk) is clearly the original form of the suffix concerned (see Sims-Williams 1992: 34).
Lines 10, 12. synpyn was identified by Yoshida (2005: 65) as a Sogdian spelling of the place name Xiping 西平 (Early Middle Chinese *sεi biajŋ), 7 though he notes that "it is not easy to see why we have a nasal element in the first syllable". In native Sogdian words the voiced [b] occurs only after a nasal, where it is an allophone of /p/ and is normally written with the letter <p>. Thus, if Yoshida's identification is correct, a possible explanation for the presence of the nasal might be that it was intended to indicate that the following <p> here stands for [b], in the same way that Modern Greek uses μπ for [b] in words of foreign origin such as μπαρ "bar" or μπύρα "beer". On the other hand, Wang Ding (2011: 235-6) has argued that Xiping was known as Ledu 樂都 in the Northern Zhou period and that synpyn should rather represent Xinping 新平 (EMC *sin biajŋ), a town some 150 km to the northwest of Chang'an. 8 Lines 10, 14, 17. ktyʾʾβr "wife" is not attested elsewhere, though a variant qṭyβryy may occur in the Manichaean fragment M110ii, V10: "But the Hearers, with their wives (qṭyβryy δβʾmbʾn) and children and all (their) concubines, when they dwell(?) in the monastery, then the Elect begin . . .". 9 I take the difference to be that the first component of ktyʾʾβr is a collective noun meaning "household" (attested with an additional -k-suffix in Christian Sogdian qtyʾq "id."), while that of qṭyβryy is merely the underlying *kata-or *kata-ka-"house". The expression qṭyβryy δβʾmbʾn, lit. "lady/wife in charge of the house", which has the advantage of being more specific than δβʾmbʾn "lady, wife" alone, may be compared with a phrase such as xʾnʾkh pʾtxšʾwnh wδwh 7 Early Middle Chinese is cited following Pulleyblank 1991. 8 I am grateful to Wang Ding, who kindly sent me an English summary of the relevant part of his article. 9 This text will be published by Federico Dragoni and Enrico Morano. The interpretation suggested here is my own. Elsewhere ktyβry or ktyβryh is attested as an abstract noun "worldliness" and ktyβryk as an adjective "worldly" (all of these exclusively in Manichaean texts), but the underlying notion is that of the "householder", who is by definition "worldly", i.e. a layman, a Hearer rather than an Elect.
"a wife with authority over the house" in the Sogdian marriage contract Nov. 3, lines 10-11; 10 ktyʾʾβr "wife" may be an abbreviation of a similar expression. Lines 11-14. The day of the marriage is given as a pig year, the 7th day of the 6th month, a hare day. As noted by Yoshida (2005: 61), the only day which accords with the data given is 19 July 519 (cf. the table of dates according to the calendar of the Northern Wei in Xue and Ouyang 1956: 406). Lines 14-16. As Yoshida rightly indicates, the date given for the death of Wirkakk, the 7th day of the 5th month in a pig year, must correspond to 16 June 579. 11 Lines 17-19. The day of Wiyusi's death is given as the 7th day of the 6th month, evidently in the same year as that of her husband. According to the calendar table used by Yoshida (2005: 62), this day should be 15 July 579, a tiger day, while the Sogdian text states that it was the day of the hare (or rabbit), that is, the day following the day of the tiger. As we shall see, the fact that this was a hare day seems to have been a point of some significance for the authors of the inscription, so it is unlikely to be a mere mistake. Since an interruption or variation in the regular count of the cycle of the twelve animals is hard to envisage, it seems certain that the day in question was in fact the hare day 16 July 579. 12 Once again I am grateful to Bill Mak for showing me that there is a straightforward solution to the apparent discrepancy. As mentioned above in the note to lines 1-3, one must consult the calendar of the Northern Zhou dynasty (Xue and Ouyang 1956: 410), in which the 7th day of the 6th month was indeed 16 July 579.
Lines 19-20. pr ʾyδ pcβγtk srδ ʾyδ mʾxw ʾyδ myδy, lit. "in this pcβγtk year, this month, this day". Similarly in lines 24-5: ʾyδ srδ ʾyδ mʾx(y) ʾyδ myδ "this year, this month, this day". Yoshida (2005: 66) is clearly right to assume that ʾyδ "this" is used here in the sense "this same". The matter is complicated by the presence of another new word in line 19, pcβγtk, presumably at least in origin the past participle of a verb *pcβxš-. In his 2005 article, Yoshida translates pr ʾyδ pcβγtk srδ as "in this *given year", comparing the form pty-βxšʾ "gave away (as a bride)" in the marriage contract just cited, 3 sg. imperfect of an equally unattested verb *ptβxš-. 13 He assumes that the phrase "in this *given year" refers to the year mentioned immediately before, that of the death of Wirkakk, the point being that Wiyusi died in the same year as her husband. That she did indeed die in that year, i.e. 579 CE, is clear from the fact that she was buried together with her husband early in the following year. However, this interpretation of the phrase is hard to square with the immediately following 10 Livshits 2015: 25, also cited in connection with ktyʾʾβr by Yoshida (2005: 65;2016: 72). 11 Here the calendars of the southern and northern dynasties happen to coincide (Xue and Ouyang 1956: 116, 410). 12 As Professor Yoshida kindly informs me, Iwami (2016: 41-2) also comes to this conclusion. One could suppose that the 6th month started a day later than is indicated in the standard calendar table because one of the preceding months was counted as a "big" month of 30 days instead of a "small" month of 29 days. 13 Nov. 3, R6-7 (unfortunately misprinted pry-in Livshits 2015: 25). Although the preverbs pt-and pc-are etymologically related, they are not usually interchangeable. However, a verb ptbynt "to answer" is attested in Christian texts beside the well-attested noun pcbnt "answer". words ʾyδ mʾxw ʾyδ myδy "this (same) month, this (same) day", since the date given in the text indicates that Wiyusi died a month later than her husband. It seems to me therefore that "this pcβγtk year" as well as the following "this month, this day" must refer to some other date than that of the death of Wirkakk. Lines 17-19 inform us that Wiyusi died on a hare day which was the 7th day of the 6th month; although this is not specifically stated at this point in the text, we know that the year was a pig year, since this was specified as part of the date of the death of Wirkakk, just one month earlier. It is a remarkable fact that another hare day which was the 7th day of the 6th month of a pig year was mentioned earlier in the inscription, namely, the date of the marriage of Wirkakk and Wiyusi. This surely is the very point to which the author of the inscription is trying to draw attention. If we accept Yoshida's comparison of pcβγtk with *ptβxš-"to give away (a bride in marriage)", and if we suppose that this verb, which is not attested in any other context, was a technical term, it seems possible to understand pcβγtk as a noun meaning "giving away (of a bride), marriage", as indeed proposed by Yoshida in his second article (2016: 70). This interpretation is confirmed by the second passage which refers to "the same year, the same month, the same day" (lines 24-5). Here the point seems to be that the beginning of the life of the husband and wife together in this world (i.e. the day of their marriage) and in the other world (i.e. the day on which the wife died and rejoined her husband in paradise) took place on an identical day, i.e. a hare day which was the 7th day of the 6th month of a pig year. It is perhaps significant that it is only in the case of these two datesnot in the case of the date of Wirkakk's death or that of the burialthat the animal of the day is specified. I would suggest that it is only mentioned in order to emphasize the extraordinary coincidence between the day of the marriage and that of the death of Wiyusi, exactly one sexagenary cycle later. Admittedly, the authors or commissioners of the inscription (presumably the three sons named at the end) did not make their point as clearly as they might have done, so that I have had to add some words in parentheses for clarification. In addition, they perhaps took some slight liberties with the facts. It seems unlikely, though not impossible, that it is literally true that Wirkakk and Wiyusi saw one another for the first time on the day of their marriage. Moreover, it would hardly have been supposed that Wiyusi would reach paradise on the very day of her death. According to orthodox Zoroastrian ideas, the soul of the deceased would remain close to the body for three nights, after which it had to cross the Chinvat bridge, an episode clearly depicted on the tomb of Wirkakk and Wiyusi. However, an epitaph is not a legal or doctrinal treatise, and strict logic or consistency in such matters is hardly to be expected.
Line 25. wyrʾ wδwh "husband and wife". As subject of the sentence, the nominative case is required, for which the expected form in the case of the masculine light stem wyr-would be *wyry. The form wyrʾ is also attested as nom. sg. in the marriage contract Nov. 3, R19. The reason for this is unclear. Since the related document Nov. 4, R11, written by the same scribe Ramtish, has the equally remarkable gen. sg. form myδrʾ "Mithra" for expected *myδry, one might think of a phonetic change, e.g. a vowel lowering provoked by the preceding -r-. 14 In the case of our epitaph, an alternative might be to explain wyrʾ-wδwh as a dvandva (copulative) compound, as kindly suggested to me by Timothy Barnes. In this case the ending of wyrʾ will be that of the nom.-acc. dual, as in Vedic mātárā-pitárā "mother and father" (surviving in adapted form in Sogdian mʾt-ptry, Khotanese mārāpätara "parents"), Avestan pasu vīra "cattle and men", etc. Since such a compound would be an ancient relic, this would explain the use here of the traditional word for "wife", wδwh, rather than the neologism ktyʾʾβr found elsewhere in this inscription. Moreover, the same explanation could be applied to another surprising collocation, zmnh-ʾnw(γ)wth (line 22), zmnʾ-ʾnγwth (line 27). Yoshida (2005: 67) noted that the rare word ʾnγwt means "period", but did not comment on the strange zmnh or zmnʾ, though he evidently regarded it as a form of the wellattested noun zmnw "time". 15 Since it seems certain that this noun derives from a neuter stem žamn-< *ǰam-ana-(Sims-Williams 1979: 341 n. 37), zmnh or zmnʾ, i.e. [žamnā], would be the correct nom.-acc. dual. Thus zmnʾ-ʾnγwth could be a dvandva consisting of two near synonyms, cf. Sogdian rʾδpntʾʾk "way (and) path" (SCE 258, in MacKenzie 1970: 16) or English "time and tide". It should be noted that the form zmnh also occurs in the Mug document A12, column 6, where it is suffixed to the names of the planets and other luminaries to give the names of the seven days of the week. 16 In the Manichaean Sogdian texts -jmnw is used in the same way, e.g. mʾhjmnw "Monday", but clearly zmnh does not represent this form. If zmnh is indeed a Sogdian form (which is not quite certain, since the preceding names of the heavenly bodies are all West Iranian), it may perhaps represent an enclitic variant [-žamn], with the common silent -h.
Line 31. The form wsʾʾn is not found elsewhere. Yoshida (2005: 69) assumes that it must be a spelling variant of wsn "on account of, for", but admits that it is hard to justify such a spelling in the light of the derivation of wsn from Old Iranian *wasnā. In his second article (2016: 68) he refers to the form wsʾn in a passage from the Mahāyāna-mahāparinirvāṇa-sūtra, 17 where wsʾn ywnʾk (or ZKZY wsʾn, or perhaps wsʾn alone) seems to translate shigu 是故 "for this 14 For both passages see Livshits 2015: 25-6. It is difficult to be sure whether myδr-was still pronounced as written at this period or whether it is merely a historical spelling for [miš-]. -A nom. sg. wyrʾ also occurs in the "Sutra of Causes and Effects", line 174 (MacKenzie 1970: 10), but this is of doubtful significance in view of the extremely common use of -ʾ in place of -y in this text (Sims-Williams 1979: 337 n. 4). Tremblay (2001: 68 n. 112) takes wyrʾ here as nom. pl. on the basis of the following pl. verb, but it seems clear that the subject of the latter is the combination wyrʾ δʾwn wδwyh "man and wife" (cf. Gershevitch 1954: §1659): "If the man and wife lie (pl.) at night in the vihāra, he is born (sg.) a pigeon". 15 In his later article Yoshida (2016: 62 n. 21) queries whether the form zmnh may imply that the Sogdian word for "time" was originally a feminine noun, like Choresmian zmʾnyk "id.". However, forms which contain -ān-(cf. Middle Persian zmʾn, Parthian jmʾn, pl. jmnyn) cannot derive from *ǰam-ana-but rather from an ablauting stem *ǰam-ā n-, so they are only distant relatives of Sogdian zmnw. The history and inter-relationships of this group of words are discussed in detail in Panaino 2017. 16 In the edition (Frejman 1962: 48) zmnh is consistently read zmʾn, but a glance at the photo shows that this reading is extremely unlikely. 17 The passage (So 15200(5)+, V7) is cited in Reck 2016: 221.
T H E S O G D I A N E P I T A P H O F S H I J U N A N D H I S W I F E reason, accordingly". However, while wsʾn could be a variant spelling either of wsʾʾn or (less likely) of wsn, there seems to be no possibility that wsʾʾn and wsn can both represent one and the same form. I therefore suggest interpreting wsʾʾn as "willing, desiring" < *wasāna-, present participle middle of the root VAS "to wish".