On the vocabulary of the Khwarezmian
~f -| Muqaddimatu I-Adab, as edited by J. Benzing
by Maetin Schwartz, New York City
^70';
The chief source for our Itnowledge of the vocabulary of Khwarezmian,
^ the Arabic-Khwarezmian text of Mahmüd az-Zamahsari's 12th cent,
dictionary Muqaddimatu l-Adab, was published in facsimile by Z. V.
Togan in 195P. Four years later W. B. Henning announced" that he
would soon publish his glossary of the material, the product of many
years of study, but was not able to begin work on the final draft of the
dictionary until fifteen years later. Meanwhile Iranists had abstained
from any similar project, because of the difficulty of the subject matter
and in deference to the great authority's plans. In January 1967, as he
was beginning a sabbatical leave to be devoted to the dictionary, and
with only the entries ' to 'kw ready for publication, a harsh fate ended
Henning's days. Barely a year later booksellers' catalogues bore the
surprising announcements of a transliterated edition of the Khwarez¬
mian Muqaddimah by an Altaist, Professor Dr. Johannes Benzing
Vol. I of this work by Benzing contains a transcription of the text
of J. G. Wetzstein's edition of the Arabic-Persian Muq. redaction*,
and added to each Arabic entry are Latin translations (extracted from
Wetzstein's Arabic-Latin index) ; the Persian glosses which follow are
accompanied by Benzing's German translations, and to these in turn
are added Benzing's transliterations of the corresponding Khwarezmian
glosses from Togan's text. All this is a great convenience, since it
eliminates the necessity of collating Togan's text with that of Wetz¬
stein's scarce edition, and (for the most part) does away with the need
for looking up the translations of the Arabic or Persian. A further aid is
1 Horezmce Tercümeli Muqaddimat al-Adab / Khorezmian Glossary of tlie
Muqaddimat al-Adab, Istanbul.
2 AM V/1 (1955) 43.
21 Johannes Benzing. Das chwaresmische Sprachmaterial einer Hand¬
schrift der "Muqaddimat al-Adab" von ZamaxSari; I: Text. Franz Steiner
Verlag GmbH, Wiesbaden, 1968. XX -f 403 S.
3 Samachscharii Lexicon Arabicum Persicum . . . Lipsiaj 1850.
On the vocabulary of the Khwarezmian 289
provided by Benzing's preface, with its bibhographical survey of
Khwarezmian studies, notes on the history of the Muqaddimah, and
generalities concerning the Khwar. text. The utility of this expensive
book is greatly reduced by the lack of word indices, although such are
promised for a second volume. Another shortcoming is the total omission
of translations for the Khwar. glosses, which often differ in detail from
the Persian and/or Arabic, and manifest a rather intricate "aggluti¬
native" structure. Literal translations would have been helpful toward
indicating just why the Editor chose a particular reading.
B.'s readings of these glosses, the raison d'etre of this work and its
sole truly original contribution, are of great concern to the future of
Iranian studies. Anyone who has seen the erratic, unevenly pointed
scrawls which make up the glosses will agree that a transliteration of the
entirety is a bold undertaking, requiring diligent and intelligent applica¬
tion. The results therefore merit the most vigilant scrutiny. Before
evaluating B.'s execution of his task, I shaU first call attention to some
characteristics of the modified Arabic script employed for Khwarezmian,
and point out the consequences for B.'s scheme of transliteration.^
B.'s exposition of the idiosyncrasies of the script (pp. xiv — xvii) is
essentially correct, but disregards what may be called a principle of
abbreviated pointing: On those characters (except Mn and Od) which
ideally bear a diacritical of three points, a single point may be substituted
for the three. This principle is known e.g. for certain Persian manuscripts,
including the text used by Wetzstein, e.g. ~J•>^- = c'dr (B. 6ddir
p. 292 ;jädur p. 288), or = cap (B. 6ab p. 189), cf ~ = MpiS,
supni& (B. SupuS p. 198, SubuS p. 172). Hence in Khwar. often v = vP.
^ = J c, i-j = '-^ ß, and sometimes J = J z. As examples for the latter,
notc>- (B. xz "good" 278. 2 = the usual >• xz, or ^jLi J^j (B. fy'zy'dc)
"silvered" (ppptc. f ) 401. 6 = *ßy'zy'dc, < *ßy'zy- denom. vb. (with
ßy- < *abi-) formed from *^z{y)d "silver" 35. 6 (misanalyzed *'z- plus
-d, or rather by dissimilation or haplology [haplography?] from *ßy'z-
dy'dc).
The principal extends even to ^ c {ts), which is occasionally written ^.
To make matters worse, c is also noted by ^, properly = 6, e.g. i'w'k
"prepared, ready" 365. 5; 502. 6. 7 against c'w'k "id." 348. 3; the latter
is justified by GAv. ^kav- (aor. tdv'is-) "versehen", cf OInd. d-kuvate
"er beabsichtigt", Lat. caveo etc. Also '6w'n "wide, spacious, com¬
fortable", abstract noun Wn'wk 304.2; ~ 'cw'm(-) 249.5; 463.6;
* For Arabic and Arabo-Persian words I find it necessary to depart from
the system of transliteration recommended by this Journal, as follows:
I use Ö for t (g) ; ffoT g; X for h ; 6 ior d (z) ; y for g.
290 Mabtin Schwartz
475.7 (~ 'c'm- 343. 5; 475. 8 with c < cw).^ Hence e.g. even though
only O^yr- "sharp" is attested, and this fairly frequently, one may
assume from Olr. tigra- that 6yr- = *cyr [cf. cy- "to enter" < ati-(a)i-]^;
note further 'kw6 457. 2 ~ "'kwx" 454. 3 "immersed" = *'A;wc'. And since "/" may represent 6, it may also serve for c; among very abundant
examples one may cite the sentence "fy bfy'f w[-Ynbzyd f-fwh^n. xrd"
315. 2 for fy pcy'c iv-'nbzyd f-cwb'n xrd "he lay down on his face to drink
water".
Benzing fails to understand the principle of abbreviation*, and
regards e.g. the very common jJjjIj "became" as & phonetic variant of
p'rwzd as due to the operation of undiscovered soundlaws ("eventuelle,
noch zu findende Lautgesetze", p. xviii, bottom). Even if one approaches
the Khwar. Muqaddimah ignorant of Persian scribal habits, the fre¬
quency of such pairs as s'w p'rtvzd 459. 6 ~ s'w b'rwzd 459. 7 "became
black", the alternation of p- with b- even at the beginning of a gloss or
in clusters with s (e.g. sp'O, sb'O), the absence of a parallel mysterious
voicing of t etc., should all indicate the impossiblity of finding, and the
absurdity of seeking, such sound laws. B.'s misunderstanding of the
elementals of Khwar. orthography and phonology has grave repercus¬
sions for B.'s text-critical/transliterational scheme of notation.
This scheme uses bold Roman type-face for sure, unambiguous
readings of letters ; lower case italics for propable readings (as may be
established from other examples, or by morphological criteria); and
italic capitals for poorly written and/or unpointed letters whose readings
cannot be established from other occurrences of the word in question.
^ For cw > c cf. *6w in fw'kd 'y 's "the camel knelt" (sic!) 373. 5 against y' 6'k'k-mync r'wyn "locus ubi procumbit camelus" 373. 6 {6'kd "he knelt"
is fairly common) ; doubtful if connected with Turk, öök- or 6oq- ; conceivably
< *haöa- (or aii-) plus *Yvak- as in OInd. vakra- "bent, crooked", MIrish feccaid "bückt sich". Also 6w'xy- (c//) "to suckle, give milk" 35.5. 2, 3; 6'xy'k (/-) verbal noun 497. 2; < Av. ^vaxS- "wachsen lassen" (or ^caxs- "sprühen"?).
iy:6 is shown by iywrk'r "famous" 394. 5, iwrk'r 436. 3 bis etc. Perhaps also 'c'dyk ppptc. to rn^cyd "suffered from cold, froze" 340. 3 (cf. Yidya 6iy- etc.), if for 3ct/-äd- (with prothetic vowel before cluster) and not < Olr.
6äta-; pres. stem prob. *3cy- from Olr. *ö{a/i)yd- and not contracted from
*6äya-, cf. below on myyd'Ii.
° It is therefore difficult to say whether Khwarezmian possoed a native
phoneme 6 distinct from c even, say, in if- "steal" < *trf-, iy- "to thirst"
< *träya-, ifk "sour" < trfi-.
' = äküc; cf. m'kwcyd- "ho dipped" 359. 4; to Sogd. "kwi- "to suspend".
* A concise, but incisive discussion of this "principle"and its origin was
in fact given by Henning, "Mitteliranisch" [in B. Spulee, Handbuch der
Orientalistik: Iranistik: Linguistik (1958)] 82.
On the vocabulary of the Khwarezmian 291
The latter category uses
.ßfor = p, b, t, 6, n, y;
c „ = c, Ö, ], x;
D „ = d, S, and in effect w, r, z, z
R „ = r, z, z, w, d, S,
F „ = f, ß [ y; also
s „ = s, s BBB],
and so forth.
This automatic transliteration is reminiscent of the method employed
by C. Babtholomae for Book Pahlavi words of uncertain reading.* As
outlined above, B.'s scheme would serve adequately if modified only to
accomodate the principle of abbreviated pointing (e.g. p when v = the
sound p). Unfortunately B. complicates matters by employing P, T,
N, X and the like "wenn eine gewisse Wahrscheinlichkeit für ein solche
Lesung spricht, ohne daß vergleichbare Wortbelege Klarheit darüber
schaffen könnten" (p. xix). Ostensibly such "a certain probability" is to
be decided by etymological considerations, as distinguished from other,
well written, attestations of the same Khwar. word. The desirability of
such a distinction is moot. But with B.'s mode of indicating this refine¬
ment, it is impossible to tell if B, for example, means "etymologicaUy
probable 6-sound", or "B — one of the following letters: p, b, t, Ö, n, y".
As complicated and impractical as B.'s scheme is, it is assiduously
betrayed by its inventor. Often, readings with italic capitals are given
where B.'s system demands lower case italics, since the correct reading
may be ascertained from other attestations. Among numerous examples
one may mention'" p'cxR'zd "was fragrant" 136. 8; 248. 6 (where also
ptc. pcxR'znyk), but p'cxr'ZD 305.7; correctly pcxw'znyk 255.5;
prob. < *pati-x^äd(a)ya-, cf. OInd. svädu- "sweet, tasty", NPers. x^ä
"good taste", with semantics as in NHGerm. schmecken "to have a
taste": Yiddish (etc.) Smekn "to have an odor or fragrance". pCXäS
"excrement (of birds and quadrupeds)" 95. 8, 102. 3, despite pcxbS
302. 7; < *pati-xard- [JAOS 89/2 (1969) 447, to which I now add NPers.
pizdl, paixal^^ "excrement (of birds and quadrupeds)"; further NPers.
äxäl "rubbish"].
» For this type of transliteration see the remarks of Heitning, "Mittel-
iranisch" 126 seq.
" For the rest of Section I of this paper I shall cite the Khwar. forms
according to Benzing's system, but where I give what I assume to have been
the ideal, fully pointed form I shall use bold-face type preceded by asterisk.
" The Persian pronunciation is pixäl, but (as Dr. D. N. MacKenzie
obligingly informs me, with references to the dictionaries of Bellew, Forbes,
and a native work from Peshawar), the expected paixal occurs in Paätö
(where also paixära, cf. xaral) and Hindustani. Thus the analysis plx plus
-äl given by Hom, Grundriss d. Ir. Phil. I 176 must be rejected.
292 Mabtin Schwartz
SFR "stream, canal" 91.7 is ascertained by ""äyE- "to be wet", see 24. 8; 445. 2, caus. Hy'rij- 208. 8; < *fra-gar- cf. NPers. faryar "stream (-bed)", faryarda "moist", faryäridan "to make wet"'", 'xn/i "sixth"
180. 3 is fully pointed, while 'xTyu 90. 4 should be read '.ttm, and
'xyyM 34. 7 is an odd misreading (or misprint) of 'xtyu.
Conversely B. employs lower case italics where his system demands
italic capitals. An interesting instance is the hapax "spider" at 55. 1,
which B. reads es. Even eS would be moot; the word looks to me more
like or but is possible. Among the many words known in
New and Middle Iranian for "spider", none clarifies the Khwar. word.
However, Av. ardOna- may be of help. In YaSt 5.90 it occurs in a list of
noxious creatures: aziS6a ("snakes") ardQnäiSca vawzakäi&ia ("wasps")
var9naväiS6a vardna.viSäiSia (?). Not only the context, but the resem¬
blance to Gr. OLp^yyTf, apixyyoc,, Lat. aranea, araneus^^, make the meaning
"spider" for arddna- an attractive possibility. It is true that the last
word(s) in the list are comparable with OInd. ürnäväbhi- "spider" (lit.
"wool-weaver")'*, but var° may mean "tarantula" or some similar
creature. Av. arddna-, as the homonymous word for "elbow" (cf. OPers.
araSni-, OInd. aratni-) may be for Olr. *ara6na-. The ideal Khwar.
'" Here belong also NPers. äyär "moist"; Sogd. 'ny'rth "muddy, juicy, fertile" and "yrtystr (Henning, Sogdica 30 on 6); NPers. zayär "moist soil"
etc. (Asadi); *gar- from *gal- as in OInd. galati "drips", jala- "water", NPers. paiyäla, Sogd. pty'd "cup" Oss. zycelyn "to flow" cf. Benveniste, JA 1934, 234), also Mazandaräni gal- "to flow" (Grundriss I, 364). NPers.
paiyäla is the cognate (not loanword) to Sogd. pty'ö (Chr. pty'd BST ii
851. 18), which has d from i(~r). Cf. Chr. wdrz- "to tremble" against wlrz-
and Bu. wr'rz- ; NPers. larz- but Man. MPers. and Parth. rrz-, to which add
Khwar. (')bez(y)k "fever with trembling" 20. 6; 341. 4.
" V. Pisani, RSO 18 (1939) 92—94 attempted to show, via some contro¬
versial phonological assumptions, that araOna- is the precise correspondent
of the Gr. and Lat. words for "spider". If tho development Icsn > Olr.
sn > Av. On is rejected, one may still derive arddna- from PIE *araJc-nV-
alongside *aralc s-nV-. For Av. 6 instead of s, see Grundriss I, § 282; cf.
Khwar. dyd = NPers. saxt (Henning, "Mitteliranisch" 109 fn. 2). A probable
example for Khwar. 6 = Av. 0 < PIE fc is found in *mb6md "he woke up"
200. 3, caus. mrö'myd- 406. 2 < *fra-Oam- (with *'r- < fra-, cf. Henning, Togan'a Armagan 432) to Av. Oamnah- "care, vigilance, attentiveness", OInd. ni-sam- "to observe, perceive", Gr. xofjteu "to tend".
'* For this and tho comparable Avestan word(s), see M. Mayrhofer,
Kurzgef.Etym. Wörterbuch des Altindischen, I (1956) 559. From two natives
of Shiraz I learn that the speech of that city has an obsolescent word vol to
denote a large and apparently dangerous kind of spider. This word may
possibly be an old dialect form from Olr. *varnaav-, with i < rn as in
NPers. dalman < *darnu-mani- (cf. further Henning, "Mittleliranisch"
p. 39, fn. 4 with reff.), to which add NPers. galla 'flock' = OInd. gana- <
*garna-, PIE *ger- 'to gather'.
On the vocabulary of the Khwarezmian 293
reflex would be *byn (pausal), but this is difficult as a reading for the
word at 55. 1. In descending order of likelihood one may propose *bnyn
(rani) < *ara6ni- or *aradniya- ; *e0yn < *aradan(a)- ; or *es < *aradya-.
In some instances the notation in italic capitals seems to be a device
for evading the difficulties of the actual pointing. E.g. c:B0Ny "lati-
tudinally" 206. 8; in the facsimile the top of "B" (*P, cf pdnyk 328. 1)
is broken but the point preserved, "0" has its lower points visible, and
"iV" is clearly pointed. At 292.7 by'r-KWrYK "night-bhnd" (lit.
"evening-blind"), where by'r is (as in the prec. line) unpointed, is
correctly transcribed, but at 292. 7 B. has JSy'r-kwryk where the MS
has fully pointed y'bn'b, which merely represents binär, nasalization
of biyär; cf the discussion in Henning, "Mitteliranisch" 117. [At
292. 7 read perhaps b' yn'b, mispointing of *y' bn'b].
Further caprices: mruxsyd "slept" 167.8, but correctly uzMXsyd
288. 7; uzMxsd 292. 4 and note zuydyK (ppptc. m.) 200. 3; for z cf. zmxc
(ppptc. f.) 490. 6, and caus. mzmxsyd'h 413. 8; all from *uz- + fmauk
"to loosen"'^ (cf. NHGerm. schlaff-.schlaf en) —■ a remarkable replace¬
ment of Olr. x^ap-. At 194. 3 B. reads xv/nydyd "pushed it aside, put it
off"; in reality the pointing is xwnb- as at 342. 3; 378. 8; 383. 8; cf.
Sogd. Bu. ßsly'wnp-, Chr. fstxump- (on which see Benveniste, JA 1955,
325-326); perhaps also Suyni xumböxSz "a swelling, surging; wave";
xum 6iddu> "to heap up"; Sogd. *anxöß- in Chr. mxwbynt "they (the
crowds) moved forth" BST ii 851. 9, although the -y- renders the word
suspect. [Differently now MacKenzie, BSOAS XXXIII/1 (1970) 124.]
Benzing's interpretation of many entries suffer not only from his
disregard of the other entries, but from hasty or otherwise deficient
analysis. At 402. 1 B. restores [w]y'nk iiYBr ß'NCYD "he tied a girdle
on him" (Pers. kamar bast ü-rä). Here B. seems wrongly to have seen a
loan-word from NPers. miyän "midriff, waist". The Khwar. requires
"girdle"; this occurs isolated at 44. 6, where B. reads uy'nyK. In point
of fact the word at 402. 1 has as its first letter b, of which the dot alone
is visible, but unmistakable, while at 44. 6 we have an unpointed by'nk;
one may also note that Pers. miyän-band is translated by the native
mö^-s-ßncyK at 27. 1. B. has here ignored 470. 8, where Pers. kamar
bi-bast is rendered by by'nk ß'NCYK (complete sentence: ß'NCYK scribal
error for ß'NCYD; B. indicates a cmpd. followed by a lacuna), "'by'nk
< *upa-yähana- ; cognates with different proverbs are familiar: Av.
'5 For \fmauk in Iranian, soe H. W. 'Bailey, Khotanese Texts VI: Prolexis
to the Book of Zambasta (1967) 153. Incidentally, Sogd. ßö6 "to liberate, save", ßök "liberation" etc. are conveniently explained by "contamination" of
ybaug "to liberate" by ymauk; regular is Sogd. ßöy "liberation".
21 ZDMG 120/2
294 Mabtin Schwartz
a^wyärjhana- , NPers. hamyän, pirähan, but the dosest relative is to be
found in Rustam's tiger-skin garment babr-bayän\
B. invents a word 6wJ "mit Rang und Würden" 335. 7; read *6-wc
lit. "with place" (Pers. bä-ß) as at 405. 7. kynkwb "malevolent" 80. 2
is a misreading (prob, due to Pers. kina-var) of kynkwz; correctly 128.7,
fuUy pointed 434. 8; from *kyn "vengeance" plus kwzy- "to wish, to
seek", cf. NPers. kin[a)-x^äh (further Justi, Namenbuch 163), kin(a)-jü.
"Cover, bedding, rug" (Pers. nihäli, yiläf) 30. 1 is misinterpreted by B.
as 'z(5'n«/k, apparently since Pers. yiläf -i nar "covering or sheath of the
male genitalium (of stallions and camels)" is Khwar. '^S'wk 39. 1,
2S'«fc 46.7; these however are from *rzi-däna- (orig. "scrotum"), cf.
Henning, Togan'a Armagan 432; in 30. 1 I read ""bS'fyk.i^ Other in¬
stances where B. shows disregard of Henning's work are the misreadings
syyr 58. 7 {sxyr or scyr, Togan'a Armagan 435) or ny'k "Pfand" 317. 4
bis; 208. 7, despite the pointing nb*k of the last two occurrences;
further on nb'k H. BSOAS XXVII/2 (1965) 248 fn. 2.
At 36. 6, in the section on weights and measures, the weight Arab.
istär, Pers. yak astir "one stater" is glossed by what B. transcribes as
DStBRc prxrs. Of the first word tho shapes of D and R are certain,
and S virtually so. The entire gloss is parallel to that of p'cm'nc prxrs
in the same line, which renders Pers. paimäna-yi yak man-l "the measure
(or capacity) of [the weight] one mann", so that the gloss under discussion
means "the measure of [the weight] one stater". Oddly B. was hesitant
even in the translation of the well-known Pers. weight astir. More under¬
standable is his abstention from interpreting the Khwar. as *dld-s'vyrc,
first because of the troublesome initial, and second because the Hellenistic
monetary term stater gives Khwar. (')stk(y)k 35. 6; 201. 7; 379. 2; 497. 6
which, moreover, translates Arab, dirham, Pers. diram (cf. Henning,
"Mitteliranisch" 53; Togan'a Armagan 430). One may accordingly
distinguish (')stk(y)k "drachm" from the Pers. loanword '''(')styb(k)
"(the weight) stater". This still leaves the difficulty of dfö-, which con¬
veys no meaning in Khwar. Since the dai-shaped character plus '''-styr(k)
corresponds to Pers. yak astir, we should have here somehow as ymbol
for "one" (Khwar. 'yw). At this point it should be remembered that the
native Chorasmian script was still in use at least until a century before
the composition of the Muqaddimah (Henning, "Mitteliranisch" 58).
It is conceivable that even if this script had fallen into general desuetude
by the time our text was written, its numerals were retained marginally,
just as we stiU employ the Roman numerals. In the latest Chorasmian
>» See my "Miscellanea Iranica" (to appear in the W. B. Henning
Memorial Volume, ed, M. Boyce and I. Gershevitch) 386 fn. 9.
On the vocabulary of tho Khwarezmian 295
inscriptions (7th cent.?) the numeral for "one" has a decidedly curved
or angular appearance ; an Islamic scribe could easily misread some of
the attested forms as däl. The old numeral may have attained the appear¬
ance of däl with the passage of time and the predominance of the Arabic
ductus. Such a transition is very mild in comparison with the fate of the
native Sogdian ideogram ZY "and" in the Christian texts 4 and 5 of
STii, where it is indistinguishable from Sjt. '«/ (correctly H.H. Schaedeb,
whose view is now affirmed by other forms of this ideogram in Sogdian
texts in Syriac script ; see my Studies in the Texts oj the Sogdian Christianas
[doctoral dissertation, Berkeley 1967], pp. 122—123 with suppl.).
In a great number of instances B. fails to utilize etymological pro¬
babilities toward establishing a reading. E.g. wämÖyk "a drill" 17. 8,
where it is not clear whether the dot is over the first or the second letter ;
in either event read *NZMeYK < niz- -f |/wi«(w)ö "to twirl"*'. For
'BDRO "there" 64. 4, 5 read »'ydbÖ < Av. *aeta6ra {avadä : avaOra ::
aetadä : X), cf. the Sogd. prepositions in -rd, CMS § 440i8.
At 254. 4 we have the translation of Arab, baxira "had halitosis" and
huwa abxaru "the foul-breathed", Pers. ganda-dahan (sud) "(he became)
stinking-mouthed", by Khwar. ijjjU U and ^^l resp. which B.
reads ßwd-S' p'rwzd and 'y ßwd-S'nw. It is likely that the second
member of the compound means "breath"; read *§' = Säforsän < *fräna-
(Bu. Sogd. ßr"7i "breath"; OInd. prärina-^^, with -w poss. or qualitative
suffix, cf. XYEW "stupid". Now ßwd means, "odor, aroma, fragrance,
incense", but not "stinking" ; neither in 254. 4 nor in 255. 4^" «-TIjjj =
Pers. ganda-bayal "having fetid armpits" does the initial look like
Read *pwd, cf. "bwdyk" = Pers. ganda 276.7; to MPers. pUdag
"rotten, putrid", Oss. D. fyd, I. füd "bad"; vb. *pwy- "be putrid",
imperf. pwnd (for pwyd) 364. 4, 8; 276. 6 < Av. puya-. Less interesting
*' For ymanB see Bailey, Prolexis 274; E. Benveniste, Mudes sur la
langue ossete (1959) 87—88.
But the derivation of Sogd. wysp(')r8(y) "everywhere" from Olr.
*mspaOra (QMS § 440) is now open to doubt; Chr. Sogd. has wsprd BST ii
878. 7, which lends support to Bailey's etymology with -arda-, Prolexis 413
(the reflex of *vispaOra appears to have been wysp8r, at least in the older
Sogd. of tho Ancient Letters, cf. QMS 440 fn. 1.) Note however myd BST ii
878. 7 ("day" confirmed by the Syriac) against myO passim. Tho demotic
orthographical traits of BST ii argue against influence by S. spelling (where
(5 = Chr. 0 and d).
1" Also Yaynobi jirön "odor", and cf. P. Thieme's remarks on Av. jrma-,
BSOAS 23 (1960) 271—272.
2° Possibly *(5by'k < *adaryä-; the latter may also be attested in NPers.
lar "armpit" Jahängiri < Rüni, verse); from Sogd. Cf. also Oss. dad-arm
"armpit" (lit. "under-arm"; dml < adari).
21«
296 Martin Schwartz
is 'wr'c "pregnant" 31. 6; 45. 8; 269. 7 which should be read *'wz*c
< *ava-zätikä.
Instances of patently wrong scribal transmission are sometimes noted
by B., e.g. at 402. 5 "8'byd'h [r. S'wdh?, v. Zm. 391, 8]", but most
.such errors go ignored. E.g. B. 'nJry "wrinkle" 276. 7 = NPers.
anjüy, thus *'njwy. The NPers. form represents *ham-6auka-, as is
shown by the Oss. cognate I. cendzyg, D. cendzug "shriveling, shrinking,
contraction"; cf. also Khwar. *'ncxs- in mncxst "became wrinkled"
454.4 ; 249.2; "it contracted" 467.3,6; ppptc. 'ncydyk 249.3". Yet
another instance of dotted e for w is "MrsKEPo" "people" 454. 4; read
*MESKWYD. Here the scribe seems to have accidentally placed a dot
under the wrong "b", whUe a fortuitous speck under *y best explains
the "p". The word is written correctly in the next line.
Certain of the mistakes in B.'s book are perhaps due to typographical
errors; one clear instance is is 'mk'b 352. 2 for 'nk'e. Other possible
misprints are turanjuhin (MS -1»), turangubin (for ta°) ad 20. 4, although
one notes various editorial lapses in the apperception and transliteration
of the Persian glosses, e.g. "rö yd roy" "von Angesicht zu Angesicht!"
ad 71. 4: for rö-bä-röy (rü-bä-rüy).
Many more errors of reading and interpretation will be corrected by
a detailed series of review articles on Benzing's Chwaresmische Sprach¬
material by D. N. Mackenzie, to appear in the BSOAS (drafts of which
Dr. Mackenzie most generously sent me). Despite the shortcomings of
B.'s edition, if used with extreme caution it provides a labor-saving
vade mecum to vocabulary of Khwarezmian, and it will no doubt stimu¬
late publications on this fascinating language for years to come. For
this reason alone Orientalists and Indo-Europeanists owe a great debt
to Professor Benzing.
II
The following are further examples of etymologically interesting
words found in the Khwarezmian Muqaddimah. All readings, except
those otherwise noted, are guaranteed by all or several of the citations
given for each word.
^' *6auk- represents the older form of tho root *kauk- seen in OInd.
sarnkucati "shrinks, contracts", sarnkoca- "contraction, fear", Khot. harng- güjs- "to fear", see further R. E. Emmbrick, Saka Orammatical Studies (1968)
137. ycauk is of interest for the etymology of OInd. cukra- "(tamarind-)
vinegar", Bengali cuk, Hindi cük "sour", for which an ultimate connection
with ykuc has often been proposed, cf. Mayrhofer, Etym. Wb. I 393.
On the vocabulary of the Khwarezmian 297
Oßzwk f. "saliva" 187. 3; 260. 7; 468. 8 shows partical resemblance
to Chr. Sogd. yztwq "id.", which was itself connected with NPers. xadü,
xayü "id." by Henking, BSOAS XI/4 (1949) 719. yztwq should be added
to the examples in GMS § 55 for xVz > yVz in Sogdian; *xazdu- <
*hazdu- to OInd. syedu- was proposed attractively by Bailey, BSOAS
XX (1957) 59; further Khotanese Studies VI: Prolexis to the Book of
Zambasta (1968) 401. From *hazdäkä- I would also derive Waxi yizg
"saliva" (via *äz°, with early compensatory lengthening?). Assuming,
with Moegenstiebne (IIFL II 470, 532), that WaxT nezd- "to sit" has
Olr. zd, it would follow that Waxi, like Sogdian, has both zd and z
(GMS § 379) for Olr. zd (for Persian, see Henning, loc. cit.). Since
Khwar. retains zd in 'zd'k "intelligent", yzdk "rich", 'zd- "fat" (cf.
Henning, Togan'a Armagan 434; "Mitteliranisch" 115 fn. 1), the
derivation of (')ßzwk < hazdükä would hint at a two-fold Khwar.
treatment of the cluster, but further evidence is required. The (')ß-
may be attributed to contamination by (^)fcw- "to spit, eject from the
mouth", which occurs with (')ßzwk at 187. 3; 468. 8; see also 201. 1;
77. 7.
*{')fö'w- "to hunt" occurs in *(')ß'wc 'my f. "bird of prey" 56.3;
301.6; 'ft'wk'nyk "hunting dog" 53.2; *f6'wyn (pausal) f "game,
prey" 93.6; (')f6'wnyk (pausal) "id." 438.1, pl. *f6'wnc (-/) 301. 6;
also fö'wy "game, prey" 358.3; (')f6'wy (hy) mk(y)d "hunted (it)"
437. 5, 6; 447. 1; 137. 6; 92. 6; 93. 6. Fully pointed -6- at 53. 2; 437.5,6;
447. 1; 358.3 etc.; otherwise -/- or unpointed, (')fi'w- can hardly be
separated from Oss. I. cuan, D. caucen "the hunt" < Olr. *6yäva7iä- <
*cyäva(ya)- "to chase, cause to go", for which see V. I. Abaev, Istoriko-
etimologi6eskij slovar' osetinskogo jazyka (1958) 318; Benveniste,
Etudes sur la langue ossete 24, but from *6y one would expect Khwar. s
(Henning, "Mitteliranisch" 111 fn. 8). Positing an Olr. preverb abi-
or fra- leaves the problem unchanged. One is however at liberty to
assume *fraca-6ydv-^^, which would result in *(d)fcdw- via haplology or
assimilation, or else {d)f6dw- with genuine Khwar. -6- from -66-.
'figyd (pausal) f. "grape" 471. 7, 8: *-d- is found also in NPers. an-
gurda, Suyni (and the NPers. of Mobgenstiebne's Turk from Yarkand)
22 For Khwar. /- < fra- see Henning, Togan'a Armagan 433. Av. has
frada with ykarat "to cut"; the same proverb may occur in Chr. Sogd.
fcm[rzt] (ST i 6. 11), imperf f'cmrz (BST ii 897. 11), both = Syr. htp "to
snatch up, carry off, plunder"; ppptc. fcmJty; see my Studiea in the Texts of
the Sogdian Chriatiana 16. Possibly fSywS "*atumhling-block, cxavSaXov"
BST ii 832. 12 may be from *f6yöS with *fraia- and the same final element
as in Bu. ary'wS "troubled, disturbed" (QMS § 551). A variant form of
fSywS may be the Bu. "'ßs'wSh of the three evil roads" P 9. 35.
298 Mabtin Schwaetz
angürd, Yidya agidro, MunJi aglero. Mobgenstieene regarded the Yidya-
Munji forms as early borrowings, with metathesis, from *angurdä
{IIFL 77 81 and cf. 188). Now neither NPers. angurda nor MPers.,
NPers. angür show the normal Persic development of old *angurd-,
which would give *angül-. But angür is readily derived via *anguhr from
*angudrä, of which agidro, aglero may represent the genuine Yidya-
Munji reflexes, while angurda and angürd (for *angurd, with possible
influence of angür) are probably borrowings from an EIr. dialect in
which dr becomes rd. Dr. R. Pj. Emmebick kindly sent me this valuable
note: "I should think only [Olr.] *angudrä [and not an^urdä] would be
possible, in light of Khot. güra- (on which see Bailey, BSOAS 1957, 51);
... güra- may be as Khot. niürä- ["jewel, coin"] beside OInd. mudrä,
but rti > Z in Klhot." The Khwar. cognate of Khot. mürä- is mök f.
"coral bead, Glasperle" (= Pers. muhra) 454. 6. Apparently Olr. *mudrä
became mard- (cf. also Yaynobi mirda, Iskäsmi mdrdik "glass bead")**
and then mad{d)-; the evolution of *''ngö from *angudrä would be per¬
fectly parallel. Elsewhere in Khwar. rd from internal dr is retained
(see fn. 24 with ref.), but note the sporadic loss of r before d e.g. in
'pd(y)k "large axe" < p{ä)rt-, mdk "dead" < mrta- etc.
The word for "saddle" is 'stry{y)k f. 37. 6, 7 his; 346. 7 (at 39. 7 B.
has twice misread d for k written without the diagonal stroke ; similarly
at 36. 6 wrongly Mym:(5-p'cm'wc for nymk p'cm'nc "one half moww").
Here we have a cognate of PaStö sary, Oss. sary, Pers.-Arab, sarf, Bu.
Sogd. syrh (on which see Benveniste, JA 1936, 228), Mugh Sogd.
s'yr'; therefore Khwar. dstrVy- < *srVy- < sary-, with secondary
Khwar. sr > str as independently in Germanic and Balto-Slavic.
Khwar. also has sym f. "camel-saddle" 47. 5, 6; if this is connected with
the latter group we would have Olr. *sag-md or *sag-man- alongside
*sag-rd (metathetized in all languages but Sogdian) from a root *sag-.
*{')wrdncy- "to entrust, bequeath", imperf. stem w'rdncy- 386.3;
418.4; 493. 1; 507.6 should be added to the examples of rd < *dr;
from *ava-dranfaya-, cf. Av. ädrdnfayeiti "setzt fest"; Bu. Skt. < Ir.
avadranga- "earnest money" etc.^*
2ä For further forms from Olr. *mudrä see Bailey, Prolexis 283—284;
Moegenstiebne, IIFL // 63, 226. To the "pearl" words discussed in JAOS
89/2 (1969) 445 I now add Chr. Sogd. mrywrt, which occurs in a fragment
of I Timothy 2, 9: "Modest {pr Sfrwqy') ... be their adornment {h't wy&nty py't'qc) . . . not with pearls and with gold {ny pr mrywrt't pr zyrn)", T iii B : C 51, 3—5. This fragment is mentioned ("nicht identifiziert") by O. Hansen,
Handbuch der Orientalistik: Iranistik: Literatur (1968) 94 fn. 2.
For rä < dr, and cognates of avadranga-, seo my "Miscellanea Iranica"
538—387; further apud B. Porten, JAOS 89/1 (1969) 154—155, and my
Studies in the Texts of tlie Sogdian Christians 112, 153.
On the vocabulary of the Kliwarezmian 299
*C)wzß'hy- "to bore, annoy", imperf. stem mwzß'hy. 362.5; 368.2;
403. 7; imperf. passive stem mwzßh- 272. 7 (which guarantees -ß-) ;
473. 2. Cf. Bu. Sogd. wyzp-, wzp-, Man. wfp', Chr. 'wzf, wyzp- {BST ii)
' 'terror". Some years ago Dr. Gershevitch suggested to me that the Sogd.
may be from Av. ubjya- (" )/ao6gr") "to press down, hold down' (differently
OMS § 377). One may accordingly assume a noun *ub]yä "oppression,
af3iction", with divergent semantic developments for both the Sogd.
and Khwar. *'wzß'hy- may be explained as its denominative in *-aya-,
with -h- from hiatus. The -y- of Sogd ivyzp- would be due to Umlaut, cf.
Chr. nyywynt < *ni-gundaya-, Bu. 'pyw'yz- < *apa-gauzaya- {QMS
§§ 186, 191).
^bk's- "to be stupefied, dumbfounded", imperf. b'k'syd (pausal) 247.4;
ppptc. bk'dyk (pausal) 247.5; caus. b'k'syd'h "he astonished him"
300. 4: < apa- (or upa-) + ]/M; this furnishes the root of Sogd. Chr.
qH, Bu. k'dn "dumb, mute".
*bSzy- "avoid, shrink back from, abstain", imperf. stem b'Szy- 466. 2;
473.8; 98.7: < apa- + Av. sizdya- "zurücktreiben", mid. "zurück¬
weichen, sich entfernen von —". But mßzyd'n "he let them flow"
395. 4 is from *'ßSzy- < *abi-Sarzaya- < Av. ^haraz- {hdrazaya- e.g. Vd.
6. 2). The preservation of the mid. sense of sizdya- is of interest; from
the act. sense are known Parth. syzdyn "overwhelming, tyrannic", abstr.
syzdyft (MirMan. Ill; Hymn Cycles), Zor. MPers. syzd (Andarz-i Ädurpäd
Mahr. 26), szd (Ardä Viräz I 4), szdyn' (ibid. XVI 2) "id.". From Av.
syazda- comes Khot. ssaysda, Tumäuq sazda- "snake".
*by'ßy- in b'y'ßyd'h "he made her pregnant" 361. 1 may be derived
from *upa-yäbaya-, and would thus represent an Iranian cognate of
OInd. yabhati "futuif^^.
yw- 342. 2—3 corresponds to Arab, axta'a "go astray, miss, err".
Thus ywd'h 'y pnd'k = Pers. gum kard räh-rä == axta'a t-tariqa "he lost
the road"; this suggests that NPers. gum is from Olr. *gu-md-. In the
same lines (and elsewhere) yd(y)k m. is "mistake, sin" < *g(a)u-ta-,
which cannot be separated from ywnd(y)k m. "sin" 343. 5-6 etc. <
*gav-ant-. The root idea is "lack, fall short", cf. Sogd. yw- "to lack, need", yw'n "sin", yw'ncyk "necessary"; further Ir. cognates in Emmeeick,
Saka Gram. Stud. 144. Olr. *^gav "to lack" can be from PIE *gh(e)Hw-
"to gape", which accounts at least as well as *gheu, *ghdu- for the various
forms, including the Baltic, listed by Pokoeny, IEW 449. From a deri¬
vative "wide open area" may come not only Goth, gawi, NHGerm. Gau,
2* [Ir. *yäb-, OInd. yabh- in NPers. fäfi and {äf-fäf "whorish" according to Hoen, Orundriss I 113. But perhaps *by'ßy- "to make pregnant" is merely
euphemistic "to impose, make obligatory" (viz. the birth of a child); cf.
b'y'ßyd'hybr 'y w'c "he imposed the matter upon him" 368. 7].
300 Mahtin Schwabtz
Arm. gavar, but Av. gäva-, cf. AIrWb. 509 with notes; Benveniste,
Etudes sur la langue ossete III. [That Gava may have been an Sogdian
toponym (cf. Henning, "Mitteliranisch" 26) hardly argues against this
explanation.]
In JAOS 89/2 (1969) 445 I proposed "yyck ['book, letter'] < Aram.
gittä (Accad. gittu) 'document'??". The Sumero-Accadian loanword
occurs in Jew. (Bab., Pal.) Aram, as gittä "document", Syr. getM "testa¬
ment", Mand. gita "passport, document". Its presence in "Reichs-
arämaisch" is attested by KYT' (for *GYT') as an ideogram for MPers.
nämag "book, letter" (Farhang-i Pahlavig XV). It also appears as a
loanword gty, gtky in. the MPers. inscriptions of Kartir (Naqs-i Rustam,
Naqs-i Rajab, Ka'ba-yi Zardust), as was recognized independently by
Schaeder^' and Henning^'. Armenian ktak 'testament, treaty' further
attests West Middle Iranian *git(t)ak. Since old -it- > Khwar. -c-, yyc- is
< giT- for gilt-, it is only fitting that the source of Khwarezm's early
script should have contributed the word for "book" as well.
b'rk f. "udder" 51. 2 may be united with Gr. Oy^Xy) under a PIE form
*dhelä (*dheE-le-A). Henning, Togan'a Armagan 426 contemplated a
derivation from Olr. *üdar-.
kw'zc "jokery, jesting, facetiousness" 419.3; 420.5 109.7 etc.,
probably pl. of *kw'zk, may be analysed as representing an Olr. pejora¬
tive ku- plus arfa- "worth, value", therefore *" triviality". Alternatively this may be explained from pej. ka- plus mrza- (or *värza-'i) "performance, activity", therefore "''"what behavior!", "what doings!". Cf. *'wzy- <
Av. V9rdzya- "to effect" in mwzyd'h 'y kcy'k "he exerted the effort"
497.4—5; mwzyd'h 'y kcy'kh "he exerted his efforts" 495. 2; caus. (< Av.
vardzaya-) mwzyd'h "he caused him to perform (something)" 361.7;
498. 6. It is not clear whether the ' of kw'zc is mater lectionis for o (to
avoid ambiguity) or d.
A certain instance of vrddhi is *h'wcr "friend": pl. h'wcrn 181. 3;
h'wcry'd "friendship" 425. 4 (both with / for c) ; < *häu6ara- from *hu- cara- "benign(ity)".
*pcmyz- "to pour, flow copiously" (= Arab, ^drr): imperf. stem
p'cmyz- 119. 4 (of milk from the breast); -y-caus. 365. 8; 500. 4 (of the
wind's action on rain[-clouds]). Prob. < *pati-maiz- "to make water"
i.e. "urinate"; the latter sense (inherited from PIE) is found for this
root throughout Iranian, hut as I. Gershevitch, Unvala Memorial
Volume (1964) 89—91 has shown, there are West Iranian traces of *maiz-
2' Apud E. Ebelino, Das aramäisch-mittelpersische Olossar Frafiang-i-
Pahlavik (1941) 34.
2' Prof. Jackson Memorial Volume 41 fn. 3, where, however, gty is derived
from Syr. getlä (ref. kindly sent me by Prof. M. Dresden).
On the vocabulary of the Khwarezmian 301
with the meaning "drizzle, be misty". "Urinate" (and apparently
"defecate") is Khwar. 6liy)-, see 132. 5; 223. 6; 512. 1, caus. 6Hy- 376. 1;
6l.-n pl. "urine" 258. 4. Cf Iskäsmi 6al 'mud, mire (Russ. grjaz'), swamp'
(thus Paxalina; Sköld has 'soaking wet'), Rösani 6äl 'mire', Suyni 6äl
'puddle, mire'; these may well represent a loanword from Persic *6al,
cf Baxtiyäri 6el, 6il 'puddle'. Alternatively one may compare Yaynobi
caldska, 6Uiska 'a drop, a splash', which may have been influenced by
Yayn. saSka 'dew' (Sogd. SySky 'a drop') or Täjik siriSk (all < Oir. sras-
ifca-).28 [Note also Kurd. 6ilk "filth", 6Um "snot".]
In JAOS 89/2 445 I regarded pcpk "facing, opposing" as a metathesis
of the synonj^mous pckb. I would now prefer to derive pcpk from *pati-
paka-, cf Av. paHipa- "against the stream". The identical semantic
development is shown by OInd. pratipa- "facing, opposing". Note also
pcy'b "upstream" 348.2 < Av. paHydpam; Pb "downstream" 349.7
< *frdp3ni. [frapa- adj. of Lake Satavaesa (MacKenzie).]
On p. 446 of the same article I connected r'wyn (pausal) f. "land,
place" 355. 8; 373. 6; 497. 4 with Av. ravan- (and Sogd. r'wHh). A closer
comparison may be made with Oss. I. ran, D. raucen "place", cf. also
MPers. rön "direction". From *rävan- raun- comes Av. ravan- raon-
prob. "vaUey" or "lowland" (compared by Benveniste with Oss. ran,
raucen Etudes sur la langue ossite 31 and with Sogd. r'w'th Tsp 221),
with V > ^ before resonant. Sogd. r'w'th = räuat ; Dr. A. Xromov kindly
confirmed my suspicion that the Yaynobi place-names Rout and Pullaro-
w(u)t reflect such a Sogd. word, andnotröi "river" (wrongly Andreev, Yagn.
teksty). Dr. Xeomov writes me that Rout is indeed located in a "Seiten-
bergthal" (Yayn. rdwut, rout); Piillarout ~ Piillarowut "Little Rout)
derives its name from the aforementioned village, which lies nearby.
Xromov further defines the Yayn. word rdwui, rout as "ein enges Berg¬
thal, durch das ein Flüßchen fließt und in dem es viel Gras gibt"; the
term corresponds to Tajik tagov and Turk, säy {soy). Among the other
toponyms Xromov adduces are Röutak (a ravine in Yaghnob),
and Kisrovüt, Ki§rovät, Kusrovät, Fuzrovüt, Kadrovät, Narovüt,
Vosrovut, Mandarovüt, and Xarkisrovit, all on the upper Zaravshan
valley. Sogd. räucd finds a correspondent in NPers. rävad which Asadi
(where verse of 'Asjadi) and BQ understand as a verdant hill-and-dale
region; the same word occurs in a verse of Firdausi SN (cited by
2' I take this occasion to point out a clear Yaynobi-Khwarezmian corre¬
spondence: Yay. öirdak : Khwar. crd 56. 8 bis (unpointed) "sparrow". Cf.
also Skt. ca{aka-, Pkt. cada{y)a-, Hindi cirä etc. "sparrow", which hitherto
have been explained as onomatopoetic or — dubiously — of Dravidian origin,
see Maybhofeb, Etym. Wb. 368 with lit. ; T. Bubeow and M. B. Emeneau,
Dravidian Etymological Dictionary (1961) 178, no. 2273.
302 Mabtin Schwartz
VuLLEKS Lex. s.v.) ; 'Abd ul-Qädir's SN Glossary has for rävad "a place
which has a meadow with greenery and flowing water". For the Sogd.,
Yayn., and NPers. forms one may reconstruct Olr. *rävali f., from a
formation in *-ant- besides that in -an-.
*rnf- "to mate, couple with" in rnfd 'y nrk my f-'Scyk "the male bird
mounted the female" 80. 6 my copy of Togan's ed. has a possible
trace of the dot of n). A likely cognate is Sogd. rnß- P 3. 183—191; the
context suggests that the animals (and humans) are to be depicted as
copulating, rather than battling*^. So too mß- in P 2. 325—326 seems to
refer to making love, not war; ZKZY yönyw rnß'nt rty myrty is hardly
"ceux qui sont battus entre eux et meur(en)t", since this is expressed
later in the list by ZKZY FAnyw ptyw'y'nt 320—330. Both the Khwar.
and Sogd. forms represent ranf- (not ranß-; old *ramb- would give
Kliwar. *mb-, Sogd. *rn'p-). It is therefore difficult to connect with these
Khwar. p'crßdyd "concubuit cum ea" 152. 6; 352. 8, which may belong
with OInd. ^rabh "to clasp, embrace". Incidentally "futuere" is *'yy-
in myyd'h (not mfyd'hl) 188.5; this agress with Oss. qcein < *gäya-
against *gäya- elsewhere in Iranian ; *gäya- would give Khwar. *myyt'h^.
An Iranian cognate of OInd. sUnya- "empty" is found in Khwar. snk
256. 3—4. In the Käfir languages we have ASkun cuni, Waigeli cüne-,
Kati cui "id."
"Lip" occurs unpointed at 37. 8; 46. 6; 294. 8 bis; Benzing reads Syt
from 250. 6, where t is pointed, and 38. 5, where S is assured and the
second letter interpreted as pointed y. If Syt is correct, one could take
it as Syt, comparing Bu. Sogd. snth "elephant's trunk", Suyni Sand
"lip" (with Umlaut < Sundä'i), NPers. Sand "bird's beak" (Asadi and
Faxr-i Samsi, with verses); cf. Benveniste, Vessantara Jätaka 92 on
36b, who dismisses borrowing from Indic suridä "elephant's trunk",
whence however [via Dardic] Pastö Sund{a), Paräöi Sund "lip". Other
forms without retroflexion are Waneci Sond, Sund "lip" (ELfenbein;
also Sund), Waxi "Sind", Sandik "gum". At least the lack of w of the
Sogd., and the vocalization of the NPers., argue somewhat against
borrowing from Indic, which would in any event be excluded for the
Khwar. OInd. suridä has been compared with Dravidian and Munda
words (see Mayehofee, Etym. Wb. s.w. tundam and cancuh). It is con¬
ceivable that we have here an old substratic Wanderwort, cf. Henning
2' This possibility was mentioned to me by Prof. Henning, when I read
this text with him.
And pres. stem *y'h-, like other stems with Olr. -ä{-ya-) e.g. Sm'h-,
Smy-t; prm'h-, prmy-t; pcm'h-, pcony-t, cf. Henning, AM N.S. V/1 (1955) 45.
For *äyV > 'h in nouns, note t'h "thief" 3. 4 < täyu-; yr'h "mud" 476. 8
< *gräya-, cf. Sogd. yr(')yk, Munp yaray, Oss. (Bryoe < *graya-.
On the vocabulary of the Khwarezmian 303
on the Indic and Iranian words for "mustard" and "lion", AIQN Sez.
Ling. VI (1965) 42—46. These show irregular correspondences of initial
sibilant, found also in the Indic and Iranian words for "sand" (Bailey,
Indo-Scythian Studies 361 with lit.), to which one may perhaps add
Eüiwar. cy "sand" 464. 5, cyn'n genetive plural 201. 3 (both with /-)
instead of *Sk (Dr. Emmerick informs me that he has made this com¬
parison). Incidentally Finnish Mekka "sand" {h- < *s- or 6-), apparently
without etymology, may belong with this group.
Thus the correspondence of s to S in the "lip" words is not without
precedence. Unfortunately the Khwar. form is not decisive; one would
expect *Sytßnd; although the other Khwar. word for "lip", 't (< Av.
aosta-) may have had some role here. But at 38. 5 the faint diagonal
points(?) under the second letter suggest not so much y as p, with the
top right point missing. Hence &pt, metathesis of Sogd. p{r)M But the
latter may be pl. of *p{r)S, cf. Yidya porSik "lip". Possibly other MSS
of the Khwar. Muq. will settle this and other tantalizing problems of the
present text.
Some interesting examples of i < Olr. fr: *Pzy- grant, bestow (e.g.
a blessing)", 3rd sg. subj. Pzy'c (not S'dy'cl) 425.6 < Av. fräöayät;
cf. S'd "nine" < fräd- (Henning, Togan'a Armagan 433). &mnk "hope"
402. 6; 448. 5 etc., cf. Parth. frmnywg "hope" (M. Boyce, Manichaean
Hymn Cycles, 187); Toch. A and B pärmank "id." prob. < Bactrian
*fr3maiyg < *framanaka-; thus Av. framanah- is prob, "hopeful" rather
than "cheerful". Sn'zk "vertigo" 218. 1—2 is from Olr. *näj- found in Sogd. näz- "to turn round" (GMS §§ 266,1068) and ßn'f- "to 8tumble"(?),
(Henning, Sogdica p. 33), Suyni nöz- (nöy) "to revolve", caus. nez-.
*t's- "to melt" (intrans.), imperf t'sd 213. 6 "it melted": < jtä in Oss. I. tain, D. tayun "to melt", ppptc. tad = Av. täta- (Benveniste,
Etudes sur la langue ossete 45). Henning, "Mitteliranisch" 112 fn. 1 has
ts- < ta6ya- as ps- < pa6ya-, but the present stem ts- seems undocumented,
and (as Dr. MacKenzie points out to me) one would expect an imperf.
*tsd, since "it was cooked" is psd 249. 8 etc. The caus. t'syd'h "he melted it" 372. 6 follows the pattern oi b'k'syd'h; cf. also mzmx'syd'h.
"Sleeve" is zwfyk (pausal) 27. 7, cf. p'rzwfyk (pausal) "front part ofa
sleeve" 28. 8, pl. p'rzwfc 405. 4 from a form in para- with vrddhi. These
words are from Av. zuS-, zuSa "sleeve" (ef. frazus- "having long sleeves"),
whence also Suyni zuz, Ormuri zü, Sangleci zöl, etc., see Gershevitch,
Avestan Hymn to Mithra 220 fn. with ref Phonologically, like mwf <
müS-, frwf < fruSä etc. (Henning, Togan'a Armagan 434). It is very
tempting to connect zus- with OWNorse ermar-kios "a (baggy) sleeve",
Swed. kju^ "end of a sack", Norw. keysa, kysa "a kind of hood for woman"
under PIE *g{e)us-. These and other Scandinavian words, indicating
304 Maktin Schwartz, On the vocabulary of the Khwarezmian
geographical and anatomical concavities, were united with NPers. güsa
"corner, end, extremity" under PIE *geu-s- by Lid^n, IF 19 (1906) 327.
However güSa may be a develoment of Olr. gauSa- (< *ghausa-) "ear"
as in Av. 6a0ru-gaoSa- , cf. OInd. karrjMka-.
In my Studies in the Texts of the Sogdian Christians I discussed a
number of Khwarezmian-Sogdian lexical correspondences ; the following
are among the clearest :
'wd 376. 3: Chr. 'wt (BST ii), Sogd. script "wt "religious faith" <
ä-wat-. *{')wdSj/- (see Benzing, p. 397 no. 6, where in a) read y'wdzy'k
for 'y wdzy'k): Chr. 'wdyz-, Bu. 'wd'yz- (sic P 2. 284) "to strangle" <
*ava-darzaya-. 'wrzyw "tyranny" 366.8; 367.11 439.8 etc.: Chr.
wrzyw "haughty" [BST ii 902. 4), wrzy.wy' "arrogance" (BST ii 826. 9;
894.19), Bu. wrz'yw (sic) "staunch, unyielding" (P 2.1077, 1101);
perhaps cf. Av. vazra-, vdzardt, OPers. vazrka- etc. plus *-aiva- as in the
Oss. abstract suffix I. -iuceg, D. -ewceg'???
'wz'cyk "crimen imputatum" 152. 8; 441. 3 < *'u!zH-cyk : Chr., Man.
'wz'ni- "to condemn", pass, 'wzins-, past stem 'wzt- and 'wzmt- (GMS
§ 584); aU from *ava-zam(H)-, with past stems -zät- ~ -zat- ~ -zamt- as
Sogd. -yät- ~ -yat- ~ -yamt- from \/yam(H)- (GMS § 539 fnn. 3 and 4).
prftk "greedy" 370. 6—7; 93. 2;prftk'wk "greedy, acquisitiveness" 93. 1,
452. I : Chr. prp- "seize" (BST ii 827. 54; 858. 11; 915 bottom 1), cf!
also Khwar. wrfk'wk (prob, for *wrftk'wk) "greed" 370. 1; hence ^rap-,
to Lat. rapio, rapax. Since apa- and upa- give Khwar. b- (and the lack
of a verb *prp- in Khwar., and of an adj. *prft- in Sogd., eliminate the
likelihood of borrowing from Sogd.), it may bc best to derive prftk and
prp- from *pa-rap-^^.
The existence of a preverb *pa- has been proven convincingly by
Benveniste, Etudes sur la langue ossete 98—102, at least for East Iranian. The
Armenian examples there proposed may at least as conveniently have pa-
upa- ; certainly upa- is reflected in pastar (Av. upastarana-, NPers. bistar) and
pa6ar (see Henning, "Mitteliranisch" 39 fn. 4).
L'origine directe et indirecte de tokharien B laks «poisson»
Par A. J. Van Windekbns, Louvain
C'est ä Wolfgang Kjiause, Nachrichten der Akademie der Wissen¬
schaften in Qöttingen, I. Philologisch-historische Klasse, 1961, Nr. 4,
p. 83 SS., que nous devons l'etude la plus approfondie qui jusqu'ioi ait
et6 consacr6e au nom du «saumon» tel qu'on le trouve non seulement en
germanique et en balto-slave, mais aussi en tokharien, oü le mot en
question a pris le sens general de «poisson». On peut meme dire que 1'
enquete de Kbause porte en premier lieu sur le terme tokharien B laks,
dont la döcouverte en 1914 par S. Levi (MSL 18, p. 389) n'a cesse de
stimuler sensiblement tous les savants, linguistes et historiens, qu'in-
teresse le probleme de la patrie primitive des Indo-Europeens.
Krause a certainement eu le merite d'avoir examine minutieusement
tokh. B laks — en general ses devanciers s'6taient tous contentes d'etablir
purement et simplement l'equation tokh. B laks = germ. *laxsaz, sans
contröle aucun — et d'avoir prouve que phonetiquement et morphologi-
quement le mot tokharien präsente une structure qui oblige ä y attribuer
une origine indo-europeenne directe, non pas, comme 1' a cru ä. tort
R. VON Heine-Geldern, Saeculum 2 (1951) p. 247, une origine indo-
europ6enne indirecte par l'intermediaire du germanique*.
Seulement il reste des difficultes qui sont surtout d'ordre phonötique
et qui ont ete reconnues en partie par Krause lui-meme.
En effet, tandis que germ. *laxsaz (cf. v. h. a. lahs, etc.) remonte ä
i.-e. *loicso-s, lit. la^is ä i.-e. *loJcsi-s et russe losös ä i.-e. HoUso-Ui-s ou
*loJcso-jcio-s (cf. Krause, p. 85 et 90 ss.)*, la forme tokharienne B laks
devrait continuer i.-e. *lh{i-s) ou %h(i-s) qui d'apres Krause, p. 90,
prouverait pour le tokharien «dasz das behandelte Lachsetymon in sehr
alter Zeit, nämlich vor den Erscheinungen des Ablauts, entstanden sein
muß». C'est lä cvidemment une hypothese incontrölable. D'ailleurs
Krause lui-meme reconnait le caractere forcö de cette reconstruction en
ecrivant ä la meme page : «so möge der Leser den Ansatz Hlcsis für nicht
' M. Maybhofeb, ZDMG 105 (1955) p. 179, a egalement ou tort de ne pas
rejeter cette idöe.
2 Je pröfere ne pas faire ötat ni de ossöte loesccg (peut-etre emprunt), ni
de skr. laksd- "100000" (extremement douteux).