Prasun historical linguistics: new etymologies and the fate of the Indo-Iranian morphemes *-am and *-ka-

Abstract

In this article, I touch on some lexical and morphological aspects of Prasun historical linguistics. I propose six new etymologies for Prasun words that have not been etymologized at all (üžóg “resin”, ćəwā́ “rhubarb”, wulóg “footprint”, žíma “tent, camp”) or differently (wuzógrog, zógrog “knee”, wuẓnúg, wuẓéŋ “salt”), and add further remarks to three words (üzǖ́ etc. “ice; cold”, lümī́, lümǖ́ “tail”, wəs “day”) with whose traditional etymologizations I basically agree. Furthermore, it is argued that the common epenthetic wu- ~ ü- and the final (usually) -u ~ -ü have the same origin and largely go back to the acc.sg.m/n, nom.sg.n *-am of the Indo-Iranian a-stems. Additionally, while the *-ka-suffix is present in all Nuristani languages in various functions, there is a noticeable split between Prasun, where *-ka- is added to many nouns of the inherited basic vocabulary while it is absent in the cognates in the other Nuristani languages.

Sources and transcription

The Prasun words are generally cited from the vocabulary list in Buddruss and Degener (2015), which is the most precise work available on Prasun now. Morgenstierne (1949) is used when it contains additional information. Comparative data from other Nuristani languages stem, mostly, from Ashkun (described by Morgenstierne 1929, 1934; Strand 2011c), the Kati varieties (described by Strand 2011a, 2011b) and Waigali (described by Morgenstierne 1954). Regarding Kati, I usually cite Kati (that is, western Kati) and Kamviri (eastern Kati). While I do not want to imply that Kati and Kamviri are distinct languages, the research by Strand (2011a–2011d) has made available to us some Kati varieties in such detail that it seems appropriate to make use of the data where needed. Data from my own research with native speakers of Nuristani languages are only cited when they can provide clarification or further information not found in previously published literature.

The transcription follows, by-and-large, the academic transcription also used by other authors for Nuristani languages. Where the authors disagree – specifically for the sounds [ts] and [dz] – I have opted to use ć for [ts] and ȷ́ for [dz].

Abbreviations

I use abbreviations for some language names when they have been established in the literature.

Av

Avestan

IA

Indo-Aryan

MIA

Middle Indo-Aryan

OIA

Old Indo-Aryan

PIE

Proto-Indo-European

PIIr

Proto-Indo-Iranian

Pkt

Prakrit

PNur

Proto-Nuristani

üzǖ́ etc. “ice; cold“

The general etymological connection of the Prasun word has been clear for a long time, but the Proto-Indo-European reconstruction shows some variation in the literature. Because the Nuristani data can help clarify the situation, I shall make some brief remarks on üzǖ́ “ice; cold”. Its Nuristani cognates are Waigali yoz “coldness; cold”, Kati yuz “cold”, Kamviri üć “cold” (Turner 1973: no. 10396), and the word can further be connected with various words denoting “ice” or the like in other Indo-European languages (for example, Hittite eka- “cold, frost, ice”, Old Norse jaki “icefloe”, Old Irish aig “ice”), going back to Proto-Indo-European *yeǵ- (Morgenstierne 1949: 280). As kindly pointed out by an anonymous reviewer, Wakhi yaz, Sariqoli yoz “glacier” also point to *yeǵ-. But this root has also been prominently reconstructed as *yeg- in the Indo-European literature, such as in Pokorny (1969: 503; Kloekhorst 2008: 279; Matasović 2009: 435). In light of the Nuristani and Iranian data, I would like to emphasize that the correct PIE root can only be *yeǵ-. From *g (assuming it precedes *e or *i), we could expect PIIr *ǰ > PNur *ǰ (possibly [dʑ]), whence Kati, Waigali ǰ and Ashkun, Prasun ž. PIE *ǵ, on the other hand, yields PIIr *ȷ́ > PNur *ȷ́ [dz], giving southeastern Kati/Kamviri ȷ́ / ċ#, western Kati z ~ ȷ́ and Waigali, Ashkun, Prasun z (cf. also Halfmann 2022). While there can certainly be – especially in Prasun, as I argue in the etymology for žíma below – individual oscillations between z and ž, the situation for this etymon is clear: all cognates in Nuristani and Iranian clearly point to PIIr *ȷ́. Therefore, PIE *yeǵ- must be reconstructed with a voiced palatal stop *ǵ and not with a plain *g – using the traditional interpretation of the PIE phoneme inventory.¹

üžóg “resin”

The Prasun word for “resin”, üžóg (Buddruss and Degener 2015: 617) is related to the “resin” word in Waigali, ǰöw < *ǰatu-. The word is of Indo-European origin with a good but not overwhelming representation in Indo-Iranian. We have, for example, OIA játu- “lac”, with a few continuants in MIA and beyond (Turner 1973: no. 5093; Mayrhofer 1992: 565) and Pashto žā́wla “resin” on the Iranian side (Morgenstierne 2003: 105). The root also seems absent in the Nuristani languages other than Waigali and Prasun – or, at least, any possible continuant is not used in the sense of “resin”.² The development in Prasun is regular, as PIIr *ǰ yields the voiced sibilant ž in Prasun, and intervocalic *t generally undergoes lenition in all Nuristani languages. The initial ü constitutes an epenthetic vowel, the addition of which is quite common in Prasun (see discussion under wuzógrog, zógrog “knee” below).

Note the *-ka- suffix in the Prasun continuant, which is also present in the OIA variant jatuka- “lac”, but absent in all other known Indo-Iranian cognates. It is common for Prasun to use the *-ka-suffix even when other Nuristani languages lack it, as shown in Table 1. Although there are counterexamples – with the *-ka-suffix present in all or some other Nuristani languages, with or without Prasun showing *-ka-, depending on the word – there is a strong general tendency. A few examples should be enough:

Table 1. The *-ka- suffix in Nuristani languages

It is unclear how recent the addition of *-ka- in these Prasun words is. Additionally, it must be emphasized that the morpheme has remained productive for a long time in all Nuristani languages, leading to several waves of *-ka- which is therefore preserved in different shapes in one and the same language. But despite the complex questions involving *-ka-, there is a striking difference between Prasun and non-Prasun Nuristani regarding the usage of the morpheme in inherited (often basic) vocabulary. Given the many other peculiarities of Prasun, this could be yet another feature which shows the primary (genetic) or secondary (areal) split between Prasun and the other Nuristani languages. Nevertheless, we cannot say if the presence of the *-ka-suffix in üžóg is as old as in OIA jatuka-: it could go back to Old Indo-Iranian times, or it could be somewhat recent.

ćəwā́ “rhubarb”

ćəwā́ “rhubarb”, recorded by Buddruss and Degener (2015: 642) for the dialect of Paṣki, is certainly a loanword from Kati ćawó. The fact that Kati o shows up as Prasun ā should probably not be overinterpreted, but it suggests that the word was loaned at a time when Indo-Iranian (and Proto-Nuristani) *ā was still *ā in Kati and not yet raised to o.⁵ The change PNur *ā > Kati o is confined to the Kati varieties, as other Nuristani languages generally do not undergo *ā > o, which means that it is an inner-Kati change. In any case, the genuine Prasun word is ućápar, ućápər, known from all Prasun dialects (Buddruss and Degener 2015: 516). Morgenstierne (1949: 256) derives the native Prasun word from *ćwātwara-, and the Kati form from *ćwātara-, both possibly related with OIA śvātrá- “invigorating” (of soma and other food; Turner 1973: no. 12762). In line with ancient Indo-Aryan evidence, but also with both Kati and Khowar (išpār) – the latter being the only Dardic language for which this etymon is attested so far – we can reasonably assume that the original shape of the word in Nuristani had initial *ćw and medial simple *t. Therefore, Prasun ućápar, ućápər, which requires a medial cluster *tw (compare čpū “four” < *čatwāra-), is a secondary form. It cannot be totally ruled out that the protoforms in Prasun and Kati also differ from each other regarding the vocalic situation, although this is unlikely. Prasun ućápar, ućápər presupposes *ćātwara- (or, with Morgenstierne 1949, *ćwātwara-), with a short vowel (already cautiously proposed by Morgenstierne 1949: 207), but the fact that the rhotic element was lost in Kati could suggest that *r was preceded by a long vowel *ā in Kati (cf., for example, štawó “four” < *čatwāra-, do “mountain; forest” < *dhārā- “edge”, versus kyur “leg” < *khura-, sur “fountain” < *sáras- “lake”, but then again bor “burden, load” < *bhārá-). In theory, the Kati word might therefore go back to *ćwā̆tāra-, but this would remove it further from OIA śvātrá-. Therefore, as an anonymous reviewer suggested, we could assume that the long vowel before *r is a straightforward result of the loss of the intervocalic dental in *ćwātara- > *ćwāra-. This line of development has the advantage that we do not need to create another protoform which is not directly attested anywhere.⁶

In my opinion, it is not necessary to assume an original shape *ćwātwara- (Morgenstierne 1949) for the Prasun word, as *ćātwara- (with a simple metathesis of w) would yield the same result.

lümī́, lümǖ́ “tail”

The difficulties in reconstructing the precise protoform and the dissemination of this very etymon throughout the Indo-Iranian world could well be the topic of an article on its own, so I shall only make some short remarks on the “tail” word regarding Prasun. lümī́, lümǖ́ “tail”, with cognates in other Nuristani and Dardic languages (Turner 1973: no. 6419), is obviously related to Iranian *dum(b)a- “tail”,⁷ but a protoform *dumbha- given by Turner (1973) cannot be accepted for Prasun. PIIr *mb(h) always yields b in Prasun, for example, üštyüb “tree” < PIIr *stambha- “stem, post” (while it gives m in the other Nuristani languages, for example, Kati štum, Waigali üštüm “tree”). The forms in the Nuristani and Dardic languages are not easily reconcilable – not, in any case, if we assume that they are all genuine inherited words, neither regarding their initial consonant (l appears also in some Dardic varieties without *d > l) nor their medial consonant(s) (simple m versus cluster mb(h)) nor the presence versus absence of suffixes. If we suppose that lümī́, lümǖ́ “tail” is an inherited word from Proto-Indo-Iranian, then only a protoform *duma- is possible for Prasun. Morgenstierne (1949: 257) also records a form dəmū́ from a man from Pronj, but it is not found in Buddruss and Degener (2015). Because of the initial dental stop, it cannot be genuine Prasun and must be a loanword after the change *d > *δ > l. It might be a borrowing from a Kati dialect, which could point to a more widespread dispersal of direct continuants of *duma- in Nuristani, but none of the recorded “tail” words in the Kati varieties (dramří in Strand [2011b], damáři in Strand [2011a], under “Zoology: External Body Parts”) seem particularly close. Loanwords from Ashkun and Waigali into Prasun are less frequent, but in both languages the “tail” words look more like that of Kati than that of Prasun anyway (for example, Ashkun dimašī́ also with some suffix, Morgenstierne [1934: 89]; Zhönchigali düməř with the same suffix as in Kati, Tāza [1396/2017: 1079]); therefore, they can be excluded as donor languages.

Phonologically, lümī́, lümǖ́ “tail” could be a loan from a neighbouring Iranian language, either before or after the change *d > *δ > l, because a number of Eastern Iranian languages (Bactrian, Munji-Yidgha, Pashto) also undergo lambdacism.⁸ Geographically, both Bactrian and Munji are potential source languages for Prasun,⁹ but it is unclear if we can assume that such a loanword would also reach Gawar-Bati, Shumashti and Pashai (which all have “tail” words with a lateral instead of a dental), while at the same time seemingly leaving out all other Nuristani languages. Possibly, the Dardic words are not even cognates of the Iranian “tail” etymon: Morgenstierne (1973b: 109) cautiously derives Pashai līm “tail” from OIA *lumbī- “bunch” (cf. Pkt luṁbī-, Turner [1973: no. 11089]). Alternatively, we could assume that the Pashai and other Dardic cognates of *dumba- with initial lateral have acquired the l from association with words such as Pashai lūm “hair, wool” < OIA lṓman- “hair on body of animals or men” (Turner 1973: no. 11154).

Be this as it may, two things are certain: the default word for “tail” in Prasun is lümī́, lümǖ́, which, if genuine and not borrowed, requires a protoform with simple medial *-m-.

wulóg “footprint”

Phonologically and semantically, a derivation of Prasun wulóg, wológ “footprint” (Buddruss and Degener 2015: 839) from *pada-ka-, cf. OIA padá- “footstep, track, place” is obvious. Turner (1973: no. 7747) also lists under padá- a Prasun continuant wəl as in ṭə wəl “foot sole” (ṭə being the general term for “foot, leg” in contemporary Prasun). Despite the fact that it is notoriously difficult to come up with general rules regarding the exact development of the Indo-Iranian vowels in Prasun, we may be allowed to make further assumptions about wulóg and wəl. There seems to be a tendency that Prasun ə more often goes back to old *ā (including *ā < *aH), while u is the more common outcome of *a:

In Table 2, I have not included words whose exact derivation or origin remain unclear, for example, sətə́ “seven” could be a loanword or inherited from PIIr *saptá- and either directly from *saptá- > *sat(t)á- or via *sātá- with compensatory lengthening, and words that underwent umlaut, for example, misü in t-əwa misü “fish” < *mátsya- (?),¹³ and words in which *a or *ā are, at least historically, in a nasal environment, for example, (wu)zógrog “knee” < *zang(h)ra-ka- (see below), wuč, wučū́ “five” < PIIr *pánča.

Table 2. The outcomes of *a and *ā in Prasun

But even if we do not count such special environments or cases with unclear etymology, numerous examples remain where, for example, *ā yields neither ə nor u, but other vowels: for example, āw, ā “water” < *Hā́p-, čpū “four” < *čatwā́ras. Nevertheless, if we are willing to follow the argument that there is a tendency *a > u, *ā > ə, then we might be allowed to assume that wulóg “footprint” derives from *pada-ka- with short *a, while wəl “sole” continues *pād(a)- with long *ā.

wəs “day”

The Prasun word for “day”, wəs, is clearly related to the word for “day” in the other Nuristani languages, for example, Ashkun wās, Waigali wās, Kati wos. Already Morgenstierne (1929: 284) compares them to OIA vāsá- “abode; staying”. Phonologically, this etymologization is attractive because the vowels in this word – Ashkun ā, Kati o, Waigali ā – clearly and undoubtedly point to *ā. However, the semantic connection between the Nuristani “day” words and *Hwāsá-, if it indeed had the meaning “abode; staying” also in Proto-Nuristani, is not immediately obvious. Because of this, some specialists (for example, Fussman 1972: 199, Degener 1998: 559, Buddruss and Degener 2015: 858) have considered a derivation from or influence by *Hwasar- “dawn”, cf. OIA vasar- “dawn”. But phonologically, *Hwasar- is a bad match, which is already noted by Fussman (1972), who thinks that the missing *-r- in the contemporary Nuristani words is the biggest obstacle in connecting the Nuristani “day” words with OIA vasar-. I would not rule out that, regarding the disappearance of *-r-, we may have a good parallel in Prasun syus “(younger) sister”, Ashkun, Waigali sos “sister”, Kati, Kamviri sus “sister” < PIIr *swásar- (OIA svásar-). However, I agree with the objection of an anonymous reviewer who notes that the disappearance of *-r- in the “sister” word can easily be attributed to the nom.sg *swásā or other cases without *-r- in the paradigm of feminine (and masculine) r-stems, which we cannot assume for a neuter r-stems like *Hwasar-. The problem could be solved by reconstructing an n-stem *Hwasan-, as n-stems are often influenced by, or merge into, the a-stems in Early Middle Indo-Iranian times (von Hinüber 1986: 153). But even there, the problem that the protoform requires a long *ā would remain.

As a solution, we could assume the existence of an ancient *Hwāsá- with the meaning “dawn” in the PIIr dialect that would eventually become Nuristani. This *Hwāsá- would have been homonymous with *Hwāsá- “abode; staying” which lived on in OIA. Just as the latter is a derivative of PIE *h₂wes- “to remain, to stay”, the former would be a derivative of the homonymous but semantically different *h₂wes- “to become bright, to dawn” (Rix et al. 2001: 292f). This approach would enable us to combine the advantages of both OIA vāsá- “abode; staying” and vasar- “dawn”, while eliminating the downsides which come with a connection to either one.

Whatever approach one might prefer, Prasun wəs is in line with its cognates in the other Nuristani languages which unanimously point towards ancient *ā in the etymon.

wuzógrog, zógrog “knee”

The Prasun word for “knee” differs from that of all other Nuristani varieties (for example, Ashkun zā̃, Kati zũ, Kamviri ȷ́õ, Waigali zā̃). A connection with the latter is unlikely for phonological reasons. PIIr *ȷ́ā́nu-, the source of the non-Prasun “knee” words, would yield something like †zən in Prasun, which is far removed from the actual wuzógrog, zógrog. Morgenstierne (1949: 253) was the first one who cautiously connected it with OIA jáṅghā- “lower part of the leg, shin”, Av zəṇga-, zaṇga- “ankle”. Turner (1973: 1660a) reconstructs *ujjaṅghura- “above the shank” as a more precise OIA counterpart of the Prasun term, and points to a structurally similar údbāhu- “with raised arms”. The (likewise unattested) *jaṅghura- is probably conceived of as a derivative of jáṅghā- by means of the morpheme -ra-, but the precise process and motivation behind its addition to the root are unclear. We do not know if Buddruss and Degener (2015) agree with the more general connection with OIA jáṅghā- or the specific form *ujjaṅghura-, since the entries for wuzógrog and zógrog (Buddruss and Degener 2015: 881, 888) do not, unlike other generally agreed-on etymologies, contain a reference to Turner.¹⁴

I generally agree with the direction taken by Morgenstierne and, subsequently, Turner. However, in light of Av caθβarǝ.zǝṇgra- “four-legged”, one may wonder if the reconstruction *ujjaṅghura-, especially regarding the prefix úd-, but also the vowel -u- between the root and -ra-, is unnecessarily complicated.

Prasun often adds an epenthetic vowel *u (in the variants wu ~ ü) to etyma in which such an initial syllable is clearly unetymological. For example, Prasun ügür “hoof” versus, for example, Ashkun kur “foot, hoof” and OIA khura- “foot”, Prasun wəltə́m versus Prasun lətə́m, letém “tooth”, Prasun wuẓnúg “salt” (whether to *rawka- or *rawana-ka-, see below). It could well be that this epenthetic vowel was in origin the final -ū̆ we see in so many Prasun words, and that this final -ū̆ was analysed as the anlaut of the following word instead of as the auslaut of the preceding one. Possibly, this -ū̆ is a remnant of some case ending, perhaps the acc.sg.m/n, nom.sg.n *-am > *-um > *-u of the a-stems and the acc.sg *-um > *-u of the u-stems. We see the same sound change in the Sogdian -w /u/ as the marker of the acc.sg.m/n and nom.sg.n of the light stems, the Kushan Bactrian -ο /-u/ and Khotan Prakrit -u < *-am.¹⁵

Therefore, Prasun wuzógrog, zógrog might simply go back to the same PIIr form as Av ̊zəṇgra- (< *ȷ́anghra-). Such a connection would have the advantage that we do not need to suppose yet another, albeit related, form PIIr *ȷ́anghura- or, with prefix, *ud-ȷ́anghura- (> Turner's *ujjaṅghura-) for Prasun only. Prasun -og shows that *ȷ́anghra- was later extended by the common suffix *-ka-. Given that *-ka- is much more frequent in Prasun than in other Nuristani languages, the addition of *-ka- could, of course, be a rather late, for example, post-Proto-Nuristani, feature (see discussion further above regarding *-ka-).

wuẓnúg, wuẓéŋ “salt”

Prasun wuẓnúg, wuẓéŋ “salt” is traditionally, since Morgenstierne (1949: 254), connected with the other Nuristani words for “salt”, for example, Ashkun ẓōk, Kati ẓuk, Waigali wřək.¹⁶ Morgenstierne (1949) compares Skt rucaka- “sharp, acid; sochal salt”, followed by Turner (1973: no. 10761) and Buddruss and Degener (2015: 884). Fussman (1972: 333), too, basically agrees but thinks that a direct derivation from a protoform *ručaka- is impossible and assumes an intermediary *ručka-, because “[l]a conservation de -k- [in the contemporary Nuristani languages, A/N] ne peut s'expliquer que par la disparition ancienne du -a-”. Even then, however, it remains unclear if ancient *-čk- could even yield simple k in non-Prasun Nuristani and -g- in Prasun. Additionally, old *ū̆ gives Ashkun ü, Kati yü (after b, p changed to i), Kamviri ü, Waigali ü,¹⁷ so the vowels in the abovementioned Nuristani “salt” words strongly disagree with the assumption that the etymon contained *ū̆. Therefore, it may be feasible to look for alternative etymologies.

Phonologically, the “salt” word in the Nuristani varieties could go back to PIIr *rawká-, cf. OIA roká- “light, brightness”. Such a derivation would cause fewer problems and would, in fact, constitute an entirely regular development (other than the initial variation in some contemporary Nuristani dialects; see footnote 16) from *rawká- > *rōk > Ashkun ẓōk, Kati ẓuk, Nishey, Kegali wřuk, etc. If the Nuristani word for “face” (Ashkun mok, Kamviri mük, Prasun müg, Waigali mük) is an inherited word from PIIr *mukha- (cf. OIA múkha-) and not a borrowing (albeit old) from Indo-Aryan, *rawká- > *rōk > Ashkun ẓōk, etc. could show that *-k(h)- is occasionally preserved as a velar stop in contemporary Nuristani (depending on stress and syllable structure). In the case of the “salt” and “face” words, the velar stop might have been preserved because of the monosyllabic quality of the etyma in Proto-Nuristani or another early stage of the languages.

Concerning Prasun wuẓnúg, wuẓéŋ, however, a protoform *rawka- only constitutes a slight improvement over *ručaka-, since in both cases, the nasal is left unexplained. We could assume that *rawka- > *ẓug with the regular sound changes of initial *r > ẓ (for example, ẓət, ẓət “night” < *rātrī-) and monophthongization of *-awa- (for example, wulús “earlier” < *dawsa- ~ *dawša-). But the reason for the nasalization in *ẓug > *ẓung would be unclear. Nasalization is a well-attested feature in Prasun (Buddruss and Degener 2015: 61), but the actual insertion of a nasal consonant – not simply the nasalization of the preceding vowel – is somewhat rare. If we nevertheless want to pursue this path, we would arrive from *ẓung at wuẓnúg via metathesis and addition of the epenthetic wu- (like in wuzógrog; see above). For wuẓéŋ, we would need to assume that the vowel in ẓung was palatalized at some point.¹⁸

However, given the mystery of the inserted nasal in this word, we should not rule out a different source for Prasun wuẓnúg, wuẓéŋ, even if this comes at the cost of splitting the Prasun word from the other Nuristani “salt” words. Looking at the various Indo-Iranian words for “salt”, there is one word which seems phonologically promising: OIA lavaṇá- “salt” (Turner 1973: no. 10978), whose early Nuristani counterpart could be *rawana-. The changes *r > ẓ and *-awa- > -u- are again regular; *n generally remains (for example, 2pl ending -n < *-thana). Adding the *-ka-suffix, which is so common in Prasun, we would arrive at *ẓunúg. From there, it is only a small step to wuẓnúg via the common initial epenthetic wu-. wuẓéŋ requires the assumption of further changes, namely the shift of stress from the final to the medial syllable, umlaut and elision (possibly *ẓinúg > *wuẓíŋ), but given what we know from Prasun, these are hardly insurmountable obstacles.¹⁹

While it is impossible to give a final answer regarding the etymology of the Nuristani “salt” words, there is good reason to be sceptical about the traditional derivation from *ručaka-. At least for most Nuristani languages, a derivation from *rawká- is preferable.

žíma “tent, camp”

Buddruss and Degener (2015: 895) record žíma “Zelt” for the dialect of Dewa, noting that the word was unknown in Paṣki and Iṣṭewi and obsolete in Dewa in 1970, the time of Buddruss's second field trip to the Prasun valley. The word does not seem to have any cognates in other Nuristani languages. Given that our knowledge of the Nuristani vocabulary is – depending on the variety – still limited, it could be that the word does indeed exist in one or the other dialect and simply has not been recorded yet. But it seems we can exclude the possibility that the word is related to any common word in any other Nuristani language meaning “tent”, “house”, “camp” or, if the original meaning referred to the material, “cloth” or the like.²⁰

However, it may still be possible to find cognates of this word in Indo-Iranian, albeit outside of Nuristani. Prasun ž in genuine vocabulary typically goes back to PIIr *ǰ > PNur *ǰ (for example, žur- “to grieve” < PIIr *ǰvara-, OIA jvárati “is feverish”, Turner 1973: no. 5304), PIIr *ǰʰ > PNur *ǰ (for example, žüt “leopard” < PIIr *ǰʰántar-, Av jaṇtar- “killer”, OIA hantar- “killer”, Turner 1973: no. 13969) and PIIr *č > PNur *č (for example, žiṭig, žiṭəg “spindle” < PIIr *čā̆ttra-, OIA cā̆ttra-, Turner 1973: no. 4743), but also PIIr *ȷ́ (for example, ižéŋ, žeŋ, etc. “snake” < PIIr *ȷ́antu-, OIA jantú- “living being”, Av zaṇtu- “county, region”, Turner 1973: no. 5110). Intervocalic m can go back to PIIr *m (zəmá, zemá, zimá < PIIr *ȷ́ʰimá-, Av ̊zəma- in hazaŋrō.zəma- “thousand winters”, OIA himá- “cold, frost, snow”, Turner 1973: no. 14096) and certain combinations like *rm (īmə, yəmə “blacksmith” < PIIr *karmín-, OIA karmín- “doing”, Turner 1973: no. 2900) and *mr (omo- “to die” < PIIr *mriyátay “dies”, Av ̊miriieite, OIA mriyáte “dies”, Turner 1973: no. 10383).

In light of this, I propose that Prasun žíma “tent, camp” can be connected with OIA harmiyá- “large house” and Av zairimiiå “cover; house” < PIIr *ȷ́harmiya-. The word is also found in MIA (Prakrit hammia-, Pali hammiya- “large building with an upper story”) and Sindhi hamiya “palace” (a borrowing from Pali?, Turner 1973: no. 13998). It is not known from any other New Indo-Aryan language. The word has also been marginalized in the Iranian group, only surviving in Khotanese as ysīmā “covered place: pavilion, roofed building” (Bailey 1979: 351)²¹ and in Pashto as zérma “preparation, stockpiling, reserve” (Kreidl 2019).

Phonologically, PIIr *ȷ́harmiyá- could yield žíma through the sound laws of Prasun. Nevertheless, we need to emphasize that žíma would not be the only possible outcome of a protoform *ȷ́harmiyá-. This is because PIIr *ȷ́h > PNur *ȷ́ [dz] usually yields z [z] in Prasun, for example, zəmá, zemá, zimá < PIIr *ȷ́ʰimá- (Av ̊zəma- in hazaŋrō.zəma- “thousand winters”, OIA himá- “cold, frost, snow”); zir “heart” < PIIr *ȷ́ʰŕ̥daya- (Av zərəδaiia-, OIA hŕ̥daya-). However, Proto-Nuristani *ȷ́ can also show up as ž [ʒ] in Prasun, as I have mentioned above.²² PIIr *a in *ȷ́harmiyá- underwent umlaut in Prasun because of the palatal element in the following syllable (just like in the Khotanese and Pashto cognates). Despite the known difficulties of Prasun historical phonology, umlaut of *a in palatal environments is actually rather well-attested, for example, misü in t-əwa misü “fish”, literally “water fish” < PIIr *mátsya- (Av masiia-, OIA mátsya- ²³) and iží, ižĩ́ < PIIr nom.du *HákšiH (Av aši < *axši,²⁴ OIA nom.du akṣī, also akṣiṇī, akṣyau) (cf. also Morgenstierne 1949: 213; Hegedűs 2012: 151f). The assumption of a development *-rm- > -m- is likewise unproblematic (for a further example, see above). Therefore, phonologically, the derivation of žima “tent” from *ȷ́harmiyá- seems permissible.

Let us now discuss the semantic side. In light of Rigvedic harmiyá- (translated as “ein festes Gebäude: Burg, Schloss, Herrenhaus; Wohnhaus, Vorrathshaus” in von Böthlingk and Roth 1855–1875: 1560), it is traditionally believed that already in the oldest attestations, PIIr *ȷ́harmiyá- refers to a solid structure. Av zairimiiå is more ambiguous, as it could potentially mean both “house” and a more general “cover”. The common example zairimiiaŋura-, most likely “tortoise”, could be translated more literally as either “who has its toes in a house” or “who has its toes in/under a cover”, and it seems the adjective zairimiiāuuaṇt-, referring to the moon, does not necessarily require a literal translation “with a (firm) house” either, as a slightly different “with a cover” may be permissible too. Despite the fact that with all other factors being equal, evidence from OIA such as Rigvedic weighs heavier than data from modern languages such as Pashto and Prasun, it could be that PIIr *ȷ́harmiyá- in origin did not, or at least not exclusively, refer to a “firm structure”, but to any structure that can serve as housing. This is already hinted at in Kreidl (2019: 223), where the author notes that Pashto zérma f. “stockpiling etc.” < Old Iranian *zairmyā n.pl. “huts, houses” does not imply an especially fancy or elaborate house, as the contemporary meaning emerged via “storages” or “storehouses”.

The discrepancy between the Rigvedic and the Prasun meaning is even bigger, because Prasun žíma “tent” definitely does not refer – at least not in the material recorded by Buddruss during his field trips – to a firm structure built with stones or the like. However, it is not uncommon that words for “tent” are etymologically close to buildings of a firmer nature; compare, for example, Latin taberna “shop, store, hut” and the diminutive tabernaculum “tent” or Spanish tienda “tent; shop, store”. Therefore, both the phonological and semantic connections seem strong enough that we may see in Prasun žíma “tent” one of the few continuants of PIIr *ȷ́harmiyá- in modern Indo-Iranian.

Conclusion

Prasun, like other Nuristani languages, offers a treasure trove of ancient inherited vocabulary which still awaits retrieval and etymologization. But Prasun is especially interesting in this respect as it is spoken in the centre of Nuristan, surrounded only by other Nuristani languages – and Munji to the north – thus exhibiting less Indo-Aryan influence than, for example, Waigali. Most recorded Prasun words are still without clear and unambiguous etymologies, which certainly has to do with the many difficulties regarding Prasun historical phonology, but also its exact relationship to the other Nuristani languages. Even when a given Prasun word has been correctly identified with a certain OIA or PIIr root for decades, an exact etymologization can lead to more clarity. For example, we know now that Prasun wuzógrog, zógrog “knee” goes back to PIIr *ȷ́anghra-, otherwise only attested in Avestan in this very shape. Entirely new etymologizations can shed light on some issues that are even relevant for both Indo-European linguistics at large and ancient Indo-Iranian substrate studies. For example, in the case of žíma “tent” < *ȷ́harmiyá-, which is considered a loan from a BMAC language by, for example, Lubotsky (2001: 311), we might need to rethink the original PIIr meaning of the etymon.

Outside of etymology studies, Prasun can offer us some insights into the early history of Nuristani. While many aspects of Proto-Indo-Iranian morphology are lost forever in modern languages such as Prasun, a careful comparative study of Nuristani can bring to light at least some hitherto unknown forms and phenomena, such as the acc.sg.m/n *-am > *-um, which shows up as wu-, ü-, -ū̆, etc., in the anlaut and auslaut of many Prasun nouns. Additionally, there is a noticeable split between Prasun and the other Nuristani languages regarding the usage of the morpheme *-ka-. Although it is unclear in most cases how old the addition of *-ka- is, it constitutes an interesting dichotomy, whether we think this is yet another argument for a primary split of Nuristani into Prasun and non-Prasun, or whether we think it merely shows a later, independent streak of Prasun.

Obviously, the present work is only a modest contribution to Prasun and Nuristani historical linguistics, but it is hoped that the article shows that a modern Hindukush language like Prasun can contribute to Indo-Iranian historical linguistics despite the complex, even contradicting, sound laws we find in the language.