A note on Pahlavi lexicography: Middle Persian hassār, hassārīh

Abstract This article deals with the identification and interpretation of two rare Middle Persian words. Firstly, some attestations of the as yet unrecognized word  hassārīh are discussed, showing that it means “direction”. Then, a semantic analysis of its underived counterpart hassār is carried out, as a basis for an etymological proposal. Finally, it is argued that hassār descends from Old Persian *haçā-sāra- “(having the head) in the same direction”, and a possible reconstruction of the semantic development of the word is provided.


K43c and the word <hs'lyh> in the Zoroastrian ritual vocabulary
The fragment of Pahlavi manuscript labelled K43c consists of two non-contiguous folios (numbered 186 and 191 in Persian script) containing some liturgical indications concerning the Paragna ritual. 1 The text is still unpublished but is easily accessible thanks to the facsimile printed in Christensen (1936, vol. 2). 2ven from a superficial look, it is clear that ms.K43c provides a slightly different formulation of some ritual indications already known from the Nērangestān, which have recently been studied by Cantera (2020).In particular, f. 186 coincides to a great extent with N.30.10-11,where the procedures for drawing the water for libations are described, and ends with a fragment corresponding to N.49.14 dealing with the collection of consecrated milk (Av. gąm jīuuiiąm, Pahl. <jyw'> ǰīw) which is going to be mixed with the water (cf.Cantera 2020: 73 f.).F. 191 (starting from l. 3) contains ritual indications on the cutting of the barsom similar to those given in N.79.8 ff. (cf. Cantera 2020: 79-81).
The fact that these passages from the Nērangestān have the same content as K43c but express it in different words is very helpful in clarifying the meaning of the very first word occurring in K43c, the subject of this paper.As I shall argue below, this word, which is not registered in the main lexicographical repertoires of Pahlavi, should be read as <hs'lyh> hassārīh.Here follows a synoptical presentation of the sentences in K43c containing this term, alongside their equivalent in the Nērangestān.The edition of the passages from the Nērangestān is based on Kotwal andKreyenbroek (1992-2009, abbreviated K&K)  "Then they should go towards the barsom and recite the Avesta directed towards the barsom." Boyce and Kotwal (1971: 68) take the spelling for <hs'lyh> as a corruption of Av. ašaiia and interpret the "ašaiia of the barsom" as a reference to Y.8.2 (so Williams 1990), but there is clearly no need for such an emendation.
9 According to Alberto Cantera (personal communication), to whom I owe the reference to this passage, a comparison with N. 52.6, where ō hōytom ī barsom "the point of the barsom that is farthest to the left" is glossed with pad war ī zōhrag, could reveal a slight semantic difference between war and hassārīh, at least as far as their usage in ritual context is concerned.This is indeed conceivable and deserves further investigation, although it probably does not affect much the semantic interpretation of hassārīh outside the technical vocabulary of Zoroastrian ritual, which will be the object of the analysis carried through in the next paragraphs.
There remains one attestation of hassārīh in a difficult passage of the Pahlavi translation of the Gāϑā Ahunauuaitī (Y.28.9b):11 yōi və ̄yōiϑəmā dasəmē stūtąm "nous qui avons pris place à la cérémonie de vos éloges" (Kell.)"(we) who are standing by at the offering of praises to You" (Humb.)MNW 'w' HNA Y LKWM hs'lyh YHBWNm st'yt'l'n' kē ō ēd ī ašmāh hassārīh dahom stāyīdārān "(we) who (are) in Your direction tenth among worshippers (?)" Presumably, already in the most ancient layer of the Pahl.translation, this sentence was not correctly understood due to the ambiguity of the Old Avestan (OAv.)forms yōiϑəmā (Pf.Ind.1.Pl. of yat-"to take a position") and dasəmē (Loc.Sg. of dasǝma-"offering") and was consequently given a word-for-word translation which does not make full sense in Middle Persian.
As to the rendering of dasəmē, it seems preferable to me to accept the reading of ms.K5 <YHBWNm> (subsequently corrected to <YHBWNyt'>, which is the reading of all the other mss.) to be read as dahom "tenth", written as a Pres.1.Sg. of dādan "to give" due to homophony.This form can be explained as an erroneous translation of OAv.dasəmē (mistaken for Young Avestan (YAv.)dasǝma-"tenth" instead of YAv.dasma-"offering"), according to an interpretation already found in Y.11.9, a YAv.passage where dasmē yōi və ̄yaēθma is quoted at the end of a numerical series 1-10 (cf.Malandra and Ichaporia 2013: 16). 12The later corruption in Y.28.9b of <YHBWNm> to <YHBWNyt'> dahēd (Pres.2.Pl.rather than 3.Sg.)may have been triggered by the presence of <LKWM> ašmāh "you" in the sentence. 13he only Pahl.word which could represent the translation of OAv.yōiϑəmā is clearly hassārīh.This erroneous translation is most easily explained by assuming that yōiϑəmā was not interpreted as a verb but rather as an abstract noun in -man-or -ma-from the perfect stem of the verb yat-( yōit-, variant of yaēt-).14Since yat-is usually translated as a verb of motion in Pahlavi,15 it does not seem unlikely that the assumed abstract derivative was rendered with a word meaning "direction".
Admittedly, such an explanation must remain hypothetical to some degree, as long as the sense of the whole Pahl.sentence is unclear.However, I believe that the recognition of an attestation of Pahl.hassārīh "direction" represents a step forward in comparison with previous attempts to read the same word as asarīh "endlessness, abundance" (Dhabhar 1949: 20 in the glossary) 16 or āsārīh "encouragement" (cf.Shaked 1996: 654 fn.40), which are both otherwise unattested and show no connection with the original Av.text.Likewise, Bartholomae's emendation <hdyb'lyh> ayārīh "help, assistance" (followed by Malandra and Ichaporia 2013: 27) should be rejected as unnecessary.

The meaning of Manichaean Middle Persian <hs'r> and Pahlavi <hs'l>
The reason why the word written as <hs'lyh> has been transcribed as hassārīh in the previous paragraph is that I think it should be linked with Manichaean Middle Persian (MMP) <hs'r> hassār, translated as "likewise" (adv.), "like" (prep.)or "alike" (adj.) in the main dictionaries. 17Before I propose a new etymological interpretation for both words, it is worth analysing the few attestations of Middle Persian (MP) hassār in order to identify all of its semantic nuances.Firstly, some additional passages from the Pahlavi translation of the Gāthās will be introduced in which a spelling <hs'l> corresponding to MMP <hs'r> has mostly remained unnoticed. 18Then, the meaning of <hs'r> in its Manichaean occurrences will be discussed.

Pahlavi Yasna
The Pahlavi translation is quoted excluding the explanatory glosses which, as will be exemplified below, are later additions based only on the older Pahlavi word-for-word translation. 19Y.30.9d: hiiat̰ haϑrā manā ̊bauuat̰ yaϑrā cistiš aŋhat̰ maēϑā.
"when (our) thoughts will have become concentrated (on the place) where insight may be present" (Humb.).
the OAv.adv.haϑrā, its most plausible translation is "in the same place", so, figuratively, "concentrated" with reference to thought.Unfortunately, the sense of the whole sentence both in Avestan and Middle Persian is not fully clear, so it cannot be decided whether the equivalence of haϑrā and hassār is the result of a mechanical word-for-word translation or of a conscious interpretation of the Av.text.The explanatory gloss kū menišn pad dastwarīh ī any dārēd "i.e.(he) has his thought directed by someone else's authority" is totally unrelated to the Av.text and probably depends on a reading a-sār "without head, without guide" (cf.Av. asāra-, Pahl.transl.+ asardār, AiWb: 210).Despite being useless for the purpose of clarifying the meaning of hassār, such a gloss demonstrates that the word belongs to the oldest stage of the MP language and had become unintelligible to later commentators.21Y.32.6a-c: pourū.aēnā̊ə ̄nāxštā yāiš srāuuahiieitī yezī tāiš aϑā hātā.marānēahurā vahištā vōistā manaŋhā ϑβahmī və ̄mazdā xšaϑrōi aš ̣āicā sə ̄ṇghō vīdąm.
"Puisque, ô Maître qui conserves en mémoire, tu connais par la très divine Pensée les … des (torts) par lequels celui qui en commet beaucoup cherche, en temps de trouble, à établir sa renommée, la définition (de ces torts) doit vous être exposée, (à toi), ô Mazdā, et à l'Harmonie, au moment d'exercer l'emprise sur toi." (Kell.) "The many crimes against peace with which (the deceitful one) strives for notoriety, whether he so (strives) with these (crimes) / Thou knowest (about that) through best thought, O Ahura, Thou who rememberest (people's just) deserts./ Let praise be broadcast for You, O Wise One, and for Truth, in (the domain of) Thy power."(Humb.)"It is desired that the hate of many malicious men is announced, because if they (are) that way, (You) Ohrmazd (are) the true reckoner and the highest knower by means of Wahman.In this realm of Yours, Ohrmazd, the teaching of righteousness will become known." Leaving aside the serious exegetical problems in this OAv.passage (cf.Kellens and Pirart 1988-91, vol. 3, 84 f.), it seems to me that the Pahl.word-for-word translation reveals a coherent interpretation by the translator: the evil ones should be openly denounced because their behaviour will be judged and punished by Ohrmazd.In particular, the words yezī tāiš aϑā at the end of the first line match perfectly with ka awēšān hassār "(lit.)if they so".In this case, hassār appears to be employed in the more general meaning of "so, that way", without any local connotation.

Y.46.8a-b:
This last passage deserves discussion only because it contains a word apparently pointing to Pahl.<hs'l> (registered as <'s'l> in Dhabhar 1949 and compared to the previous attestations in Malandra and Ichaporia 2013: 204): yə ̄vā mōi yā ̊gaēϑā ̊dazdē aēnaŋhē nōit̰ ahiiā mā āϑriš šíiaoϑanāiš frōsiiāt̰ "[…] ou qui soumet mes troupeaux au tort, que le désastre (rituel) ne résulte pas pour moi de ses actes." "And if someone aims at my herds to injure (them), / may destruction not reach me through his actions."MNW ZK Y L gyh'n' YHBWNyt 'w' OLE kynyk' […]/ LA PWN ZK Y OLE kwnšn' + 'w' + mn 23 + 'slyš 24 l'nynyt'l + HWEyd 25 kē ān ī man gēhān dahēd ō ōy kēnīg […] nē pad ān ī ōy kunišn + ō man (?) + āθriš […] frāz rānēnīdār + hē "Who gives injury to my world, may (he) not direct āϑriš against me (?) as a consequence of his actions." In order to establish a plausible reconstruction of the original word-for-word translation of this passage, the manuscript text needs to be corrected in several points.In my opinion, the corruptions here are twofold.First, some mechanical errors took place: the sequence here conjecturally restored as ō man "against me" was corrupted into a spelling pointing to <ANE> an "I" 26 and the mere transcription of the OAv.hapax āϑriš <'slyš> was erroneously split into two words, originating in some mss. the pseudo-attestation of hassār <'s'l> with which we are concerned.At a later stage, maybe at the time when the explanatory glosses were added, the correct Opt.3.Sg.rānēnīdār + hē was deliberately changed into rānēnīdār ham in order to make sense of the corruptions, disregarding the fact that this would have compromised the correspondence with the original Av.text. 27The corruption of <'slyš> into <'s(')l lyš> (maybe passing through a stage where <'slyš> was written with double <l> for an alleged /l/) was already found in the text read by the later commentators, who explain asā ̆r rēš "infinite wound" as kē pad tan ud ruwān rēš kunēd "which causes a wound both in the body and in the soul" (i.e. both in life and in afterlife, forever). 28Once again, such a recostruction is rather hypothetical, since it is impossible to know to what extent the Pahlavi translator understood the OAv.text in the first place.Anyway, it should appear from this discussion that the attestation of the spelling <'s'l> in Pahl.Y.46.b is in all likelihood unrelated to the word hassār.

Manichaean texts and Armenian hasarak
In the Manichaean corpus, the passages where the meaning of <hs'r> hassār is recognizable with enough certainty are only five.They are presented below grouped according to their meaning (the references are based on Durkin-Meisterernst 2004): a) hassār as an adverb meaning "in the same place/direction as" (as in Pahl.Y. 30.9 = OAv.haϑrā).
"beyond the cosmos of earths and heavens, and outside those five hells in the south quarter, nearer (?) than there, upon the darkness of hell, from the east to the west region, corresponding to Paradise, build a new building." Although the traditional translation "corresponding to" (Andreas and Henning 1932 "entsprechend") is ambiguous, it seems difficult not to give hassār a local meaning ("in the same place as Paradise" or "in the same direction as Paradise").In fact, the whole passage is aimed at describingin a fairly cryptic way29 the place where the "New Building" (dēsmān īg nōg, see Andreas and Henning 1932: 184 fn. 1) is going to be built.
These parallel sentences follow a passage dealing with the relation between epagomenal days and constellations in the Iranian and non-Iranian (here probably Babylonian) calendar. 30According to the traditional interpretation, followed in the given translation, hassār is employed as a simple anaphoric adverb marking the continuity of discourse.However, since the purpose of these sentences is to indicate the two periods of the year in which, according to Mani's interpretation, the surplus days accumulate in the non-Iranian calendar, I regard as equally possible a translation such as "and also corresponding to the constellation of X there is one day in excess over the time of a month".From this perspective, the meaning of hassār would be more similar to (a), albeit with a temporal nuance.In any case, for practical reasons, in the following paragraph the letter (b) will be used to refer to a meaning "likewise, in the same way", since it is supported by the attestation in Y.32.6 and is somehow implied in the value (c).

Etymology
The only explicit attempt to offer an etymology of MMP hassār goes back to Henning (1935: 17), who derived it from Ir. *ham-sarda-"of the same kind", from the same stem as YAv.sarǝδa-"Art, Gattung" (AiWb: 1566 f.).This etymology is clearly based on the adjectival meaning "equal" (c), and is probably influenced by the comparison with Parth.hāwsār "similar, alike" (Arm.loanword hawasar), derived from Ir. *sarda-by Bartholomae (1906: 35 fn. 1, 233). 32In my opinion, this explanation is questionable in several respects.First, it should be emphasized that the prefix ham-undergoes assimilation before -sonly in a couple of verbal forms in which its semantic value is fairly weakened (hassāz-"to make ready" < *ham-sač-, hassūd ppt."whetted" < *ham-sauH-, cf.Henning 1947: 45). 33On the contrary, in the great number of bahuvrīhi adjectives formed with ham + noun, assimilation is never found, probably because the derivational process did not cease to be transparent and productive throughout the Middle and New Iranian period and inhibited phonological changes across morpheme boundaries.As a matter of fact, Unlike the simple hassār, the abstract derivative hassārīh seems to have preserved the old local sense "(the same) direction".Perhaps it is thanks to the retention of a more concrete connotation that the word hassārīh could survive in the Zoroastrian ritual lexicon long enough for us to be able to grasp its meaning and reconstruct retrospectively the history of such an interesting lexical item.