PIE *e inAlbanian 71

23sg. dike, disete I digje, digjete] to djeg 'to burn') and in the 3sg. ti-aor.
(Buz. engriti [engriti] to ngre'to raise'). Whereas Hamp (1971:224) has
argued that *e changed to i in front of an earlier consonant cluster,
Klingenschmitt (2000: 8) and Orel (2000: 145) explain i < *e from i­
PIE *e in Albanian mutation. The latter solution seems the more likely one, even if it is
impossible to prove the presence of "i or *1 in the following syllable for
1. There seem to be at least six different reflexes of Proto-Indo-European all the verb forms in question. The main reasons for preferring i-mutation
*e in Albanian, viz. i, e, je, ie, a and ja. Most modern treatments of the are: the 2pI.pres. originally ended in -irather than -ni (cf. Klingenschmitt
historical phonology of Albanian assume the following basic distribution 1: 1994: 226); the original ending has disappeared altogether in the 3sg.
1. PIE *e is reflected as e mainly in front of NC and after Cr- and Cl-. aorist; there is palatalization of stem-final velars in the imperfect and
2. PIE *e was diphthongized to je in the remaining environments. the present middle. Among the inherited nouns, i for *e is found in the
3. The reflex je has further developed to ja in closed syllables; in front words mish 'meat' < PIE "memso-, nip 'grandson; nephew' < "nepot-, vic
of word-final resonants, je yields ie. 'calf' < *petfios-(iJo- and Tosk vit (Gheg vjet) 'year' < *petos-. In my
4. The reflex ja yields a after some of the palatal consonants. opinion, the change of *e to i in mish and vic can be due to the influence
5. The reflex iis the result of raising of*ecaused by a high front vowel of the following palatal sibilant (for mish see de Vaan 1997: 62). The i in
in the next syllable (i-mutation). . vit can be ascribed to umlaut in the oblique forms of vjet, e.g. gen.sg.
Rule number one is safely established on the basis of etymologies such as "veti (thus B. Demiraj 1997: 420). The word nip can also be explained
pese 'five' < PIE "penkre, dhemb 'to hurt' < PIE *geml:!'-, dredh 'to turn' < from umlaut, viz. if we adopt Klingenschmitt's suggestion (2000: 11)
*cJhregn-. Some studies give more detailed rules as to the environments in that the norn.sg. PAlb. "nepo < PIE "nepots led to the incorporation of
which *e is retained, but we must be cautious here: there are indications this noun in the n-stem inflection, as shown by its plural niper < *nepenes.
that je was analogically replaced by e in part of our evidence. For instance, Since the plural suffix -er regularly causes i-mutation of *a to e (shtrat
the verb mbledh 'to collect' does not show je in any of its conjugational 'bed', pI. shtreter, vella 'brother', pI. vellezen, it seems possible that the
forms today, but it appears as enbeljedb- in the participle enbelieiune I original paradigm of 'grandson' was sg. *nep, pI. "niper, with i-mutation
enbeljedbunef 'collected' in the 16th-century text of Buzuku/. Another of *e in the plural. In order to explain i-mutation in front of the suffix ­
example has been adduced by Matzinger (1998: 234): the present of heq er, Orel 2000: 230 derives it from PIE "<ino-, but since the core of the
'to pull' is attested as biec- I hiek-I in Buzuku, which confirms the evidence nouns in pI. -eris formed by kinship terms, it is preferable to assume the
of dialects which show hjek instead of heq. Similar replacements may suffix *-en-. Possibly, its vowel was raised to "i in unstressed position in
have affected other words, for which we do not have sufficient evidence front of *n. As argued by Klingenschmitt 2000: 8, the sg. nip may have
in the older Albanian texts. acquired its i analogically from the plural, just like the sg. dreq 'dragon'
Rule number five has resulted in an alternation between e and i which < Lat. draco, -nis must have adopted e from the plural, where it arose
mainly occurs in verb forms: in the 2pI. present indicative (only attested phonetically.
as subjunctive and future after the particle te, e.g. Buz. teh enbeliinih]te Rules number two, three and four as given above contain the
enbelidbni] to mbledh, teh vdisni Ite vdisni] to vdes'to die'), in the 2sg. standard explanations for ie, je and ja. A superficial glance at the evidence
and pI. imperative (Bogdani sg. mbelii; [mbelidb], Buz. pI. enbeliini I reveals their problematic nature:
enbelidhnil to mbledh), in the imperfect (3sg. Buz. enbelil; [enbelidh] to - The reflex ja is not restricted to closed syllables.

mbledh, endis: I endiq] to ndjek 'to follow'), in the present middle (Buz. - The reflex je is found in a considerable amount of closed syllables.

- The reflex ja mainly appears in inherited words and hardly in Latin

1 Compare the surveys and discussions in Beekes 1995: 265, S. Demiraj 1996: loanwords, whereas jeis found with the same frequency in inherited
105f£., B. Demiraj 1997:41£., Topalli 1998: 75-114, Orel 2000: 3ff. words and in loanwords.
2 Words from Old Albanian authors (16'h and 17th century) are given in their - A phonetic change of je to Ie would be surprising; usually, the direction
original form, followed by a (semi-)phonological interpretation, which follows is the reverse.
the present-day orthography of Albanian. - The reflex a after palatals mainly appears in Latin loanwords.

Die Sprache 44,1 (2004),70-85
Die Sprache 44, I (2004), 70-85

(ad)rationem4.g. whereas verbs ending in -ierin the present have restored -r on the basis of the participle and maybe other forms of the paradigm. however. shtene). 72 Michiel de Vaan PIE *e inAlbanian 73 These objections lead me to reconsider the evidence for diphthongization as e. *0> ye and »e > ie are agree on the view that ie has developed out of je. and *0> "yo> yeo Since "e of PIE *e and of e in loanwords from Ancient Greek and Latin. on the other. yields ie in the same environment.1 (2004). <. "meh. tjerri!) . The same supposing that iedeveloped out of je. but also short vowels which were lengthened in stressed position show the originally complementary distribution between ie in front of in front of -11. the standard language now Threesyllabic erresye might reflect Vulgar Latin "ed-retitmem. vie11 'vomits' but participles pjelle. the vibrant is Buzuku) and ye have the stress on their first element. duar'hands' < raor« *der< PIE *f/'esr-es. The modern phenomena in the history of the Albanian language. caelum. and its alternation with je in verbal stems in -r and -11. after ie is original. Buzuku's plural form *bher. shtie 3 The Old Gheg auhtors consistenly have a threesyllabic form: Buzuku aresue 'to put into' < "sterlu: (part. The diphthongs ua (the older form of which was uo. caponem. *h2l)er.:abej 1975: 214). shpene). arsye 3 'reason' < vjelle.more details in Desnickaja 1968:269f. Topalli 1998: 99f. on the one hand. It seems to me. Klingenschrnitt 1998:95. However. there is no hint of the meanings 'teaching' or 'reproval' in erresye. and the verb erresen] cannot be the phonetic reflex of *arrati. scold. rene 'fallen'.g. -ej. < coincide with the conditions for the rise of the diphthongs ua from *0 and "stel-n-. Bogdani erresue j erresye]. in Italian (nuovo. arresenj'l reprove. it is cross-linguistically very rare condition underlies the difference between the vocalic verbs ending in Isg. and that arresye is probably derived from the verb in Albanian.(part. please' (placeo). Pekmezi 1908: 54. 'rooster' < Lat. vjerr.onllre. These two long vowels partly continue long vowels from PIE 54. we may conclude that the changes *0> uo. become mjelli.(part. where e. due to the analogical transfer of je from the participle vjele". tier 'to spin' < *ter(H). chase away'. vjerri!). However. we find three nouns: die11 'sun' (no certain etymology). S. As has been recognized by several scholars (e. although no such prefers je in the present of these verbs: bjerr. qie11 'heaven' < Lat. It is clear that the loss of *-r diphthongization by which the first part of the original vowel was raised. Cimochowski 1951: 163. whereas in others it has been lost. The alleged metathesis of *OU probably the noun which was primary in Albanian. for instance.(part.g. shtjell 'to throw' beside shtie11 dial. Most scholars of Albanian > uo. to uo is very unlikely. Compare. [okl argues that the verb can be cognate with Old French araisnier'to address' < "erretionere. I start with the discussion of Albanian lie!. e. tjerr. vier 'to hang' < aresegne jar(r)esenjej confirms the origin of arsye from a form in a nasal. Pekmezi 1908: ye from *5. as found in In part of the verbs in which *e stood in front of *-r.(part.(part. kapua word-final 11 and je in front of word-internal 11. ndegjoj'1 hear' (intellego) recently in some Albanian dialects. This yields the following relative chronology: *-e11 proposed is the restriction of ie to a limited number of phonetic >*-e11 > -fell and *-er > <er» <ier> -ie. bjerrl!).g. kembej 'I change' (cambio). Thus. and hasoccurred and those in Isg. which is attested in dialects as Isg. as the result of a retained. muaj 'month' < *muo+j < "mon < "men < PIE the Gheg presents pie11 'brings forth'.g. Demiraj 1996: 111. that In front of -11. this suggestion goes back to Jok11931: 276ff. [okl (1916: 138) has connected the verb arsen 'to instruct. shpie'to take' < "per. d. bekoj'1 bless' (Latin benedtcos. a modern present form such as vjel 'to reap' must be *arson < Lat. it is does. we find bie 'to fall. standard language contains several verbs in which both variants are The distribution of ie is as follows: it represents "e in a stressed word­ allowed: sje11 beside sie11 'to bring' < PIE *J<wel-. but bier 'to lose' < j ar(r)esyej.scold. the reverse is commonplace in languages. As argued by To­ form is continued in other Romance languages. dial. see e. mbie11 'to sow' beside pje11 final syllable in front of one of the consonants -11 and *-r. many peripheric dialects or Latin. surroundings. reprove'. shua11 'sole of the foot' < "sholl < Lat. dyer 'doors' < *dor < "dori < PIE *dhl)or-es. the change to ie and the change to je represent two independent mie11 'flour' < PIE "melluu-. Bardhi eresse. In fact. educate. d. These conditions 'to beget'. 70-85 Die Sprache 44. by a stress shift from the three members of the same process of diphthongization of long mid vowels second to the first part of the diphthong. 6 The co-occurrence of -iell and -jell has been one of the main arguments for 4 The difference between the reflexes of ceponem and retionem is caused by the presence of i in front of 0 in the second word: 0 was fronted to *lJ. piell < *peI-n-.n-. Pietro): *0> uo> ua5 . solum. Therefore. *-n and -r. as Topalli 1998:118ff. Therefore. and since a diphthongization of *eto ie would be the exact mirror image on the front vowel side of the change *0 2. miell-i'the honey' has 1981: 10M. 5 There is no need to posit an intermediate stage *ou. to bring' < PIE *lI'er.to give a lessonin how to behave. Die Sprache 44. arresl:Je.1 (2004). especially northern Tosk (according to Topalli versus pelqej ]. -oj for je to change into ie. prure 'brought'). The reason why this view was to falling diphthongs.). 70-85 i .

In front of rr. This noun was explained as a borrowing from Ancient Greek pteris 'to steal' < PIE *yedh_.g. r. they bring'. as shown by new monophthong probably was a higher vowel than the reflex of PIE *e. no diphthongization of *e has taken place. djerg'to be lying ill' < *dherHgh]-. as still found in 3. mjek 'doctor' < medicus. -r has been restored on the basis of the definite nom. but a nasalized and je in Latin loanwords exactly matches the Vulgar Latin distinction monophthong -ein Gheg. Ashta 2000: 39Of. similarly shpiem. after Cr-: grek 'Greek' < graecus. djeg 'to burn' < 13). lil/. Albanian je is found as a reflex of PIE *e in front of stops. 11. mend < mentem). *-un > -iif-u. we may assume the same for *e.sg. *dhegwh_. Proto-Albanian distinguished between two attributed to the Illyrians). Examples according to B. mjeker 'chin. it reflects the PIE root *gwher. This In front of 1. shpien The diphthongization to jealso affects Latin loanwords with Classical and shtiem. Since the nasal consonant has regularly disappeared Latin e and ae: keshtjelle 'castle' < caste11um. and in older texts such as Buzuku. the distribution of e The regular reflex of word-final *-en is not a diphthong. opposed to initial <gn-> in nje 'one' and njef'knows'. The forms which do show ie in front of word-final 1. in this case from die11 'sun'. the condition for diphthongization to Loanwords with Latin e do not undergo diphthongization: vrer 'gall. there is no diphthongization of Latin e in front of suspect. rr. This accords well with tendencies which can still be seen at work in Albanian dialects: lengthening of vowels frequently occurs in front of word-final rand 11. verbs in *-er. d. forms of pjerdh 'to fart' < *perd-. parently. monophthongization of the PIE diphthongs *oi. qete 'quiet' < quietus. but not in front 8 Sincethe Old Gheg authors consistenly spell Injeril with initial <niO)e->. dhjete'ten' < *dekrpt. 'fog' suggest that. Gheg reja 'cloud' (compare the Greek gloss pwov acc. it is more likely that <gn-> indicates a palatalized phoneme 128f.1 (2004).ind. Topalli 1996: 56ff. as of rrand 1. the diphthongization in inherited (lengthening and) diphthongization into Ie. However. but in original *be(r)me. In fier 'fern'. there was also no phonetic inherited vocabulary. pI. beard' < (1t'tEpU. As in the would have disappeared after ie too. njeri 'human being'. by means of the and *oin front of *-r and -11. whereas *e patterns with *0 may already have arisen before the Latin period.was absent. we find ie in a few Ipl. as biem and bien have analogically introduced ie from the singular. Thus.74 Michiel de Vaan PIE *e in Albanian 75 paIli. njetez < *h2ner. like Vulgar Latin. and that this was the original reflexof PAlb. je to a greater or lesser degree < Lat. somescholarshave assumed that <niO)eri> contains the diphthong lief. Therefore. from a Romanist's other stressed vowels before word-final nasal: *-an> Gheg -a I Tosk -e.sg. Put differently. Olberg 1972: 65.. booty' < praeda. vjerre 'hanged' < *h2yer-n-. S. As was seen by Romanists dealing with Albanian. nor. Possibly inherited words showing this sequence are pe. biem. the latter of discussion. We argues for the original absence of (lengthening and) diphthongization for can date this diphthongization relatively well. and 3pI. have analogically acquired ie from related forms adstrate of Proto-Albanian may have triggered this change within the in -iell.70-85 . In front of -rn and -n. it does not in front of *-n. bile' bie. -ri in most of the Romance languages (d.1 (2004). it might be expected that the nasals pjesbke 'peach' < *pesca < persica. The similar 'thread' « *pen. Demiraj 1996: 166. viz. Thus. there is no way in which initial Injef and Inief could have been graphicallydisambiguated by the Old Gheg authors. dhjes 'to shit' < *~ed-je/o-.The difference between these two vowels <in > -II-i. Die Sprnche 44. therefore.70-85 Die Spracbe 44. Demiraj 1997 are: rjep 'to peel' < PIE *h]rep-. dialects.). whereas <ni-> is used for biphonemic Inj/. shtien. these verb forms are inherited vocabulary. *h2ej and *ey to *e. ap­ which can still be found in many dialects. and -e or -e in Tosk. Bonnet 2000: 73f. Sincewe always find the word njegulle 'fog' spelled 7 A similarexplanationof ziejwas already given by Klingenschmitt 1981: 109. etc. Meyer 1904-06:1043). the endings -m and -n go back to -me and -(e)ne. the original form of the verb ziej 'to boil' was zie. with initial <nie->. bien 'we bring. pres. vjeberr 'father-in-law' < *syekuros.'to become hot". as StefanSchumacher points out to me.to PIE *(s)penhr 'to stretch' according to Olberg 1972: conditions for the development of *e in inherited and in borrowed words 29). and in front of the clusters rdh and rg. fricatives. Topalli also words and in Latin loanwords represents one and the same process.) or its Modern Greek descendant Iteri (cp'tEPTl) by Thumb (1909: "smekur-. Furthermore. point of view. the reflex of *-en is parallel to that of phonemes lei and I~I (thus Olberg 1972: 145ff. vjedh *feri. it will be safer to assume that verb forms such < venenum. nasal plus obstruent (kuvend < conventus. "ner < *ner < PIE *h~er-. vjeter 'old' < vetere. The tendency to diphthongize I~I which was present in the Vulgar Latin such as e die1 'Sunday'. *-iin > -YI-y. -ri 'bridle' which underwent diphthongization to ie.. thus e. and. A Latin loanword showing this development is ire. before the influx of the vowel *0. pre 'prey. frenum. Topalli 1998: 120ff. In front of *nj. after ua and ye from <on and *-on.8. for a between a closed mid front vowel lei and an open vowel I~/. and re. *be(r)ne.

is ambiguous Romance reflexes has preserved a trace of Greek *b. In support of when trying to explain the change of *e > ja. qarr 'oak' < cerrus. mjelte 'honey' < "meli-t-. Old Galician domea < hebdomada. Kortlandt 1987: 220. under which Die Sprache 44. for a stage *ja. it reflects Latin fabella. Hence. which does not The only remaining candidates for a Latin origin are Alb. Beekes 1995: 263. that is. ja is seldom found in Latin loanwords. B. hebdomedes via Latin. these forms cannot be used as evidence took place after the beginning of the Roman era and before the 7th century. 2. Meyer­ influence of kam. viz. and *h1sen ti is regarded as inconclusive by many scholars (e.and gjume 'sleep' Classical) pronunciation [evdomas]. gjak 'blood' < "sokro. gjarper'snake' < "serpono­ domee.1 (2004). and some of the best examples are found in Greek loanwords.) would Since we have already concluded that the change of e to ie was caused by have us expect a long vowel from the contraction of a and e. Vegliote yedma. the a in gjelpe. The evidence of jam 'I am' and jane 'they are' < *h1esmi 'seed. as djathe 'cheese' < PIE *dhedhhr . but the loss of Latin intervocalic b (which in itself is a regular development. Old French for "e > ja. qeveris 'to govern' < "kivernise < aorist lCUI3£PVll0<X. Demiraj Albanian jave is usually regarded as a borrowing of Greek bebdomes 1986: 743). Thus.70-85 . Rather. The Unlike je. Thus. pjalm 'pollen' < *pel(H)-mo-. change "e > a which was conditioned by the preceding gj. but which the change *e > ja is required. It now remains to be determined. sberre 'saw' < serra.v.g. Old Italian domada. but this is also problematic. Kortlandt (personal communication) suggests original *ja. fjale 'word' affect Latin borrowings. If the change seems to be regular. they might have occurred simultaneously. scholars. 76 Michie! de Vaan PIE *e inAlbanian 77 Slavic loanwords into the language. afterwards. it is not certain that a more probably. and it may not have spread to eastern Romance anymore. We must therefore leave these forms out of consideration an Ancient Greek loanword from before the Roman era.in front of all accented vowels (d. a PIE 4. The reason "Iebellsre. the change e> je guese (favelas 'horse teeth'). shale 'saddle' < sella. week' in Hippocrates (5th century BC). However. Fjale would then be the only Latin borrowing for developments. The explanation of fjale is disputed. the change The restriction of ja to the pre-Roman vocabulary is sufficient to dismiss of e to a in the position between q or sh to the left and lor r to the right the generally accepted view that ja has developed out of *je. In general.'to speak in public' (LI\P: 576). may have been that*eyielded a closed vowel lei in this position. Demiraj 2001: 6lf. the word is attested in Italian (favella 'speech') and Portu­ that*-eR> -ieRwould have passed through a stage *-jeR. d. under the assumption that *jwas absorbed by the preceding to me that "eviuimede may have been abbreviated to *evoo. the original forms would have Lubke (1935: 343) lists two sets of forms: 1. Latin loanwords with main problem with the Greek etymology is the unexplained loss of the a for "e after a palatal consonant are sometimes adduced as evidence for syllable containing -ma-.70-85 Die Sprache 44. jashte 'outside' in Gothic spill 'story'. which lowered to a in this position. intervocalic -v. sperm' and Greek spore 'seed' (d. neither of these two and gjashte 'six' < *s(l))eks. Kortlandt 1987: if Albanian had.g. zjarm reflect PIE "sp-. According to Meyer-Lubke resonants can still be observed in many dialects today. or it may have arisen independently in the dialects where it is attested. many scholars assume that f. fare 'fire' < *!S'hermo-. as in gjalle 'alive' < "soluo-.g. djathe'right' < *deks(i)-.can < *h1eks-t-. whence *fele. *e > ie post-dates *e > je.are also adduced as evidence Old Bolognese edema etc. S. we must conclude that it post-dates the change of "e to ja. days.(or its e. Fabella is thus a relatively recent creation of Vulgar Latin. the sequence gja. As it is unlikely (1935: 271). so that we may just as well assume that e was directly of e to je represents a single development in the history of Albanian. it would be surprising since PIE *s-yields gj. Italian edima. However.1 (2004). In theory.is rare in < "supnos.javernay be predecessor). since lengthening in front of word *fabella existed in eastern Romance. Albanian . gjalpe 'butter' < PIE "selpos. and must be derived from the verb probably did not apply in front of word-final-r. Scholars who take the tain instances of ja concern words which are inherited from PIE. and in the Septuagint. also Demiraj 1997: 56). before the 7th century (Svane and jave 'week'. they have'. < hebdoma. The distribution of ja is rather different from that of je: nearly all cer­ etymology for fjale must still be reckoned with. according to Demiraj 1997. independent "Ielle. consists of the following forms: connect fjalewith the PIE root *speIH. I note that bebdomes is already attested with the meaning 'seven to them. shartoj'to graft' < <sertzre. -11 and -n at all. The lengthening to *e in front of a resonant. Thus. gjerper and gjashte could be due to a Albanian. Orel 2000: 95) PIE evidence. mainly because of the comparison between Alb. kane 'I have. who suspect that jam and jane have acquired a under the 'week' (acc. However. According to some 1992: 290). Hence. Three words in gja.could directly reflect the Greek (post­ 219). we will return this. The latter view (e.sg. we can now be certain that the etymology might be saved by assuming (near-)haplology of *fa{3ella to changes e > je and e > ie must have been two separate. palatal: qartoj'to quarrel' < certsre. In any case. been *jem and *jene.

1 (2004). which aimed at preserving the ablaut difference between the present (e. it habe". note that the ending *-om is probably also reflected as knowledge . butte 'man'. The neuters in texts after 1944. on the next syllable contains -a.sg. of the definite article -ne (Old Olberg (1972: 72). Topalli ascribes this to systemic pressure within the verbal system. Klingenschmitt 1994: 255ff.. ending in -e. Yet there is no independent proof that the masculines in -e feminines. but not in verbs. form in -6s > -e.).which have final -e are indeed feminine Most of the words containing ja from "e end in -e: the inherited nouns. <s. there were few verb forms in which the change of *e to ja was phonetically motivated. Probably.70-85 . Die Sprache 44. Closer to Albania. fjale. Orel (2000: masculine forms with a nominative singular in zero.sg. from the outset. mjelte. (f. zero and fern. noted e. acc.sg. and Latin loans in -a have become Albanian ending *-os.javeand vocabulary djathe. The change Buchholz-Fiedler loc.g. adverbs and numerals. ending *-om regularly yielded -e in Albanian. Only part of the words in -ja. gjume 'sleep' must have the time when the Latin loanwords came in. tallow'. moreover. but not *-. as the feminines did. and feminines with 233) reconstructs an end-stressed nom. The only scholar who . most scholars had initial stress (*supnos) in order to yield the reflex gj. Since <om must have yielded *-am in the first place. ate 'himjher'. the disputed words fjale and ngjale. Accordingly. Demiraj 2002: 103-112. who also *-oh and *-a. *0 or "oi (> *e). adjectives. In fact. reconstructed in different ways by different scholars: Orel 2000 assumes As a matter of fact.sg. 78 Michiel de Vaan PIE *e inAlbanian 79 phonetic conditions ja arose.to my In support of this. 70-85 Die Sprache 44. ja is found as a reflex of "e in nouns.has previously considered a similar explanation for ja is -e in several pronominal forms: acc.cit. The two main is an adverb. -a or -e: negro 'black' (m.sg. viz. The actual endings of Proto-Albanian are which the ending -e reflects the original ending of PIE neuters. The noun has several other masculine nouns ending in nom.g. etc. masculine or neuter is dbjeme.sg. 'to pay' < PGm. Kortlandt assume that this distinction continues the difference between PIE Ir and 1987). this distinction must have been present in Proto-Albanian at had final stress and the others did not. »e. We must rather explain these masculines as original neuters. ON gjalda 'the son'.) reconstructs noun for the standard language by Buchholz-Fiedler 1987:209f. in eh2-stems. respectively.1 (2004). The noun djele which I propose is comparable to the so-called 'fracture' in Old Norse. definite emeniie 'name". Once again. and most of the remaining words (of uncertain etymology) which are djathe 'cheese' and mjalte. the noun dhjame is still acknowledged as a neuter masc. ke 'whom?'. the condition for a-mutation was not given at all in the verbal 9 Of course. kete 'this'. point out that djethe 'cheese' and mjalte are still sporadically attested as *-a. *0 (> *a). the main point being that the feminine ended in a low vowel. exceptions are pjalm and zjarm. the more change of *e to ja in the feminines can then be interpreted as a case of a­ masculines appear to have been neuters: a list of 33 nouns is given by mutation under the influence of *-a in the following syllable. see the discussion in S. but even if this were true.m. The distribution of the remaining forms is as follows: masculine jave. dialects) < *-ntom < *-m + tom. jesbte. and many of these take final -e. this It has already been claimed by Klingenschmitt (2000: 5) that the PIE development is reminiscent of the Rumanian change of e to ea if the neuter nom. *geldanan.. ngjele 'eel'. is found as a neuter with the definite article -tein Buzuku: dialete / djaletej whereby Proto-Germanic "e changed into "ee to yield ja: e. it is conceivable that the change We will start with the masculine or neuter nouns in -e. *-a and fern. djethe 'right hand' (dia~a /djatha/ in Buzuku). djath(t)e. and jashte have ja: djale 'boy'. seara 'evening' but seri 'evenings'. lume 'river'. Hence.11.acc. is possible that this ending caused the same a-mutation of*eas final *-a of The phonetic origin from a-mutation may also explain the fact that. Since Latin loans in -us have usually become that the zero ending in the majority of masculines reflects the unstressed Albanian masculines. -e instead of zero: inflexion of Albanian is characterized by the opposition between brume'dough'. but the arguments in favour of continuity contain the vowels *e. jafn 'even' < *ebnaz.) but neagra 'black' strength of examples such as Old Gheg emene.m. One might even assume masc. je) and the aorist (0). I reject an intermediate stage *je. by Topalli 1998: 133. The further back we go in time. suggesting a nom.(d. gjume 'sleep'. the assertion that the Albanian neuter really continues the PIE system: the verbal endings which we can reconstruct for Proto-Albanian neuter gender has been challenged. who suggests that "geschlossene Silbe oder Gheg.sg. whereas Klingenschmitt (1994: 223f. nachfolgender dunkler Vokal (a > e) die Offnung des je zu ja verusacht acc. Albanian of "e to ja was caused by the vowel in the following syllable. dhjame 'animal fat. respectively. see seem more compelling than the arguments against it. it would suggest that.

Zjarm has disappeared. Of course.sg.(in)def. suggests to me. because it was cold. those servants had lit a fire.yields gj.def.sg. because the conditions of both mutations are mutually exclusive: the endings *-a. and the ending(s) which bring about i-mutation cannot cause a­ in front of a stressed vowel was d-.sg.def. we only find one occurrence of zjarre.1 (2004). which would again provide the necessary condition cities' (Mt.sg.indef. -e. Buz. acc.Albanian had changed *sJcleks to *seks. nom.def. 80 Michiel de Vaan PIE *einAlbanian 81 Finally. abl. of "e.def. itss 'from here'. of the stem zjarm. it would imply that . 22:7).sg. n timit iken/ 'Qui timet ignem fumum word would regularly cause a-mutation at all. of which all sg. the numeral gjashte from PIE acc. ejermi. eierre /zjarre/ forms can also be interpreted in accordance with the rules for a-mutation Bogdani: nom. (but an indefinite for a-mutation of "e. which do not have enough data to provide a reliable survey of the dialect probably contains PIE *h1e?s 'out'. eierm /zjarm/. If this is correct.sg. gen. is conclusive. Buz. eiertnit gjalpe 'butter'. Klingenschmitt's argument (1981: 129) that zjarm. gen.ja from *ecannot be subject to i-mutation. acc.sg. n timit ichen [kusl: dro zjarre. Thirdly. ejerm. the two nouns in -ja. original *-om can explain the ending of jashte 'outside'. Greek difference in meaning.without final which we find the definite acc.def.indef. gjarper 'snake' and gjashte 'six'.sg. Die Spnche 44. probably probably due to analogical extension of i-mutation from words containing PAlb. acc. eierre.sg. As it turns out. zjermir' acc.is unlikely to be due to phonetic loss.' the basis for a few derivatives such as zjerrmi'ievex. case forms are attested except for the definite acc.sg.sg.sg. Firstly. ejere /zjarre/ acc. zjarm survives.sg. eiermit tom on the analogy with other adverbs which were derived from an acc.def. All other 37 occurrences of the word 'fire' take the stem *serpeno. 'seven'. giarpegnete / gjarpanjetej. As Hamp has argued (1992: 913).sg.indef. Pietri nde kete vend e perse isc Itoiete petne ndeeune of -e after -lm.def. Since the other Old Gheg authors all show zjannit. and the plural form is sometimes zjerm. it Latino-Epiroticum' also has zjarre only once.sg. such as ati 'pater'. The attested forms are: ElC't6<.or as *serpono-. eiettne /zjarme/. twice zjermine. warming themselves'.)eks 10 has added the suffix -te which we find in the Albanian In the text of Buzuku.and -rm. all three acc. In Bogdani's Cuneus Prophetarum.which might represent a definite acc. it is conceivable that the ending *-tos was remade to *­ gen.1 (2004). *-am cannot cause i­ 10 Pedersen (1900: 286-289) has argued that the phonetic outcome of PIE *SJcl­ mutation. eiermi. Latin intus '(from) inside'. fugiat'. Bardhi.fiarmi.indef.sg.: cus dro remains uncertain whether *a in the second syllable of a threesyllabic eisrre. and -rtne. the absence acc. which encompass many mass nouns.def. the usual word for 'fire' is Mbij Ajrit crijoj eottine eierre.occurs as air our Lord created the fire. ejermine for gjalpe may have been changed to a thematic stem in nom.sg.def. Secondly.indef. Budi's zjermit is other Indo-European languages .pl.sg.in front of stressed vowels anyway: nom.indef.70-85 . on the analogy of *septrp.sg. nom. there are only 4 occurrences of the form zjarre. In the Old Gheg texts we find zjarm and zjarre without any *-tos conveying an ablative meaning.def.7Q-SS Die Sprache 44.sg. In dialects.aate 'After the the masculine zjarr. as in Skt.indef. unfortunately. In Budi's 'Rituale'. tue ii nzeem ajo gind e sctepijsse 'Saint Peter thus went in there reverse dictionary of Albanian shows quite a number of words in -Ime and. As Stefan Schumacher Buzuku: nom.indef. eierm /zjarm/.def. the PIE neuter s-stem "selpos which must be reconstructed abl.sg. eiermi. this may well f iare kuteteuet tune / e u shtiu zjare qytetevet tyne/ 'and it set fire to their go back to PIE *-ta-.: Hini pra S. ejermi. since the ejsrre.abl. viz. acc.sg. *a. zjarre would normally have to be interpreted as the Compare the oldest attestations: nom. which is very hot and dry. acc. Bogd. ejerm. ambiguous because PIE *s. tet 'patris' in Bardhi. This is the first text in We can now discuss zjarm and pjalm.sg. eierre /zjarre/ *s(lJ. e i . which are Bardhi: nom. '(from) outside'. indef. 'that he make fire and make the fountain smoke. Bogd. However.' Bardhi's 'Dictionarium serpegneteh / gjerpanjetej. Pinal-re can go back to the PIE suffix geography.sg.sg. as a definite acc..def. Their meanings 'fire' and 'pollen' would fit in very nicely with the other The form zjarre seems to function both as an indefinite and a definite neuter nouns.acc. passion'. in eli stili numerals 'six' to 'ten'.sg. acc.sg. In the modern standard language. zjarre Above we have excepted from the evidence three forms in gja-. Budi: gen. against 82 of the stem zjarm.indef.like mutation.sg. serpenineh /gjarpeninej. gjarper has been reconstructed either as is not excluded). ejermit because of its neuter gender. eierm. giarpene / acc. issisi enscte fort i nzete. the vowel a in the Old Gheg plural forms proves PIE <ono-. of the indefinite: Rituale 43 ta beje me zjare e ta timonje Iontene gjarpene/. eierm / zjerm]. I 11 In principle. viz. *-om /zjarmine/. but a form zjarrm.

1 (2004). zjerm-: the sg. For the formal side of the opposition zjarre: zjarm.lii. phonetically from the acc. "deee. the Albanian noun may have adopted the full grade from the adj. compare shte-te 'seven' < "septtp. young woman'. zjerre. Albanian ie is the result of diphthongization of re.viz. going back to *dela(-) and "delmo-.'sterile'". uncertain whether *-ta was added early enough to cause a-mutation (recall djel-) and djelm. But it is also conceivable that zjarm does not go back to 1. This opposition is found in Latin. the other one for the plural. it is uncertain whether *a from the respectively. viz. Hamp agreement with the rules for the rise of ja as established above.sg.70-85 Die Sprache 44. Demiraj 1997: 134f. zjarre yeaned yet' to 'young animal'. and probably contemporary The noun pjalm is not attested in the 16th and 17th-century texts. which early loss of *-t. cognate with Skt. This diphthongization is relatively recent. If the earlier form was "pjalme. Albanian ja is the result of a-mutation of PIE *e. the noun can regularly vowel *r. it might have been being used for the singular. as Demiraj argues. forms only a were no inherited words which contain PIE *e > e or je in front of an minority of the attestations of 'fire' in Buzuku and Bogdani) had provided ending -e. that gjashte 'six' is ambiguous'). The other 1992: 916). Since nasals plus obstruent block the change of "e to je zjarre and zjarm. who could go back to a PIE s-stem *fjher-os.or. I find only four possible counterexamples in the evidence the basis for the present-day preponderance of zjarras the word for 'fire'. just like djelm is secondary to djale. heres. 82 Michiel de Vaan PIE *e in Albanian 83 The fact that zjermine does not occur before Bogdani might lend which. one form syllabic nasal caused a-mutation.m. Albanian je is the result of diphthongization of the open mid front apocopated or not.sg. If this is correct.at. summer'. the latter has become the only word for intervocalic -rr-.1 (2004). and I with the changes of *(j > uo and of *0> ye in front of the same resonants. with the situation with the word djele 'boy' and its plural djem < djelm. but the acc. which in its turn PIE at all. This change must have been completed before the influx of Slavic loanwords. we may reconstruct *dekrpt > *deset. the absence of ja would be regular. bear fruit'. pese 'five' I think that we are dealing with two originally different words.(or "zjerm-. in composite numerals such as njembedbjete '11' or tridhjete '30' (d. compare For dbjete < PIE *dekrpt. "zjerm-ne. singularization of the pI. probably 'fire' as a generic denomination versus a and pese. The investigation has yielded the following three main results: Albanian 'fire'. they may also have blocked *e > ja. and if it did. whence 5.'sterile.(withthe pI. initial db­ djale : djem. ghp. ending *-ne under the influence of a vowel *-a or *-a(m) in the next syllable. 'lamb'. Since PIE *gwhermos 'warm' was an adjective. They subsequently merged into one paradigm. which took place 12 There is no exact parallel of this assimilation. In analogy to syncopated at an early stage from the preform *de(Jata. it would be very The proposed explanation for ja would be even more convincing if there surprising if the definite acc. djalm-). and shtjerre 'lamb'. Ashta 2000: 203: -lne> -lle. d. H 14 The assumed semantic development goes from 'sterile' via 'which has not Alb. This does assimilate to a preceding resonant in Buzuku. Classical Latin ignis <ster-en. Compare Vedic Sanskrit stari. viz. pende 'yoke of oxen'. 'warmth. see below). -llne > -lle. intervocalic *d. Demiraj 1997: 323. He concludes that there following stages of the development are uncertain in many respects: it is have been two 'konkurrierende Stammbildungen'. They may have denoted a at a later stage. Demiraj 1997: 377 considers a derivation from concrete 'fire'.'flame' and Greek has not given birth yet. For shtjerre.(with its sg. the nasal suffix is needed to explain 'fire' versus focus 'a fire-place'. (which. we might assume that there were two words for 'fire' which represents lenited. The analysis of the evidence allows us to establish the following 13 B. 'fire' in Vl. One was zjerre« neuter *gwhernomI3. gepo<.can also derive from single intervocalic *-r-. dhjete 'ten'. according to B. + <t« The is described in detail by B. djel.to PIE "stet. oes griuiii 'kettle'. which reflected both PIE short "e and Latin e/ae in part of the go back to a neuter *pel(H)mom. 'glow (of the sun)'. this concerns pende difference in meaning. 3. I conclude that dhjete is unsuited to serve as a was zjarm. collected in Demiraj 1997. one might prefer a PIE neuter noun *gwhermom 'heat'. Demiraj (1997: 429) rightly objects that we would expect zero grade of relative chronology of developments: the root in view of Skt. maybe. -rr. Yet although this development seems possible from the phonetic point of view". zjarm might be an analogical counterexample to a-mutation. -rne > -re. underlies the verb pjell 'to support to the suggestion by B. Die Spr'IIche 44. PIE *fjhermos.'to procreate' Latin loanwords. or. do not have sufficient dialect data to say whether -e might have been 2. In fact. from the same root *pel. Demiraj (1997: 429) that zjarre arose give birth to. but is rather a secondary form built with *-m~ on the basis of reflects lengthening of PIE and Latin *e in front of word-final -11 and -r. which was probably adopted from were built on the PIE root*gwher-. change can be dated before the influx of Latin loanwords. viz.70-85 . zjerm.sg. as we have seen.

Montreal. Die SpncbD44. PIE "e > ja in front of *-a and *-a(m) in the following syllable. Innsbruck. G. *0> yein front of word-final "-n. volume I. -ue-.V. G.. In: Kratylos 44. in: Indo-European Numerals. Meyer. A. Poznan. 233. September 2000.Vaan@let. E. ed. Ko11oquium der Indogermanischen Ge­ Diphthongization of stressed »e > ie. Fonetika historike Leiden . 1908: Grammatik der albanesischen Sprache. 1935: Romanisches etymologisches WorterbucN. PAlb. Pekmezi. G. historia. 0. Handout at the XI. G. 1999:Review of: Bardhyl Demiraj: Sistemi i numerimit te gjuhes shqipe: Berlin . .Boston . BSL 66. Tirane. Gramatica. Fachtagung der indogermanischen Gesellschaft. Leipzig.1994: Das Albanische als Glied der indogermanischen Sprachfamilie. Indo­ Demiraj. 270-340. Shkoder. Topalli. Halle an der Saale.Philadelphia. [with a summary in English] 1998: Zhvillimi bistorik i diftongjeve te shqipes. Amsterdam . Grundriss der romanischen Philologie. *s in Albanian.1. 1904-06: Die lateinischen Elemente im Albanesischen. 1971: 'Fils' et 'fille' en italique: nouvelle contribution. Amsterdam Albanischen und zur diachronen Phonologie auf Grund des Vokal­ systems. Wien.1-20. Msrz 1993 in Kopenhagen. 98-164. 1992: Slavische Lehnworter im Albanischen.de. Hyrje ne historine e gjuhes shqipe. Ashta. Jokl.70-85 . Dr. Cimochowski. bis 28. 1996: Per historine e hundoresise se zanorevene gjuben shqipe. I. 1996: Fonologjia historike e gjuMs shqipe.P. K. Tirane. Olberg. H.S. Strassburg. M. M. N. 1951: Le dialecte de Dushmani. Vl!shtrim diakronik. 57-63. 1975: 1. MSS40. 93­ "eu being the source of Ie/.87-93. W. 1997: Albanische Etyrnologien.New York.A. 57-92. Pedersen. NL-2300 RA Leiden Matzinger. 131. PIE "oi. W. Sanz Ledesma. 1972: Untersuchungen zum indogermanischen Wortschatz des Beekes. C. R. Gvozdanovic. 1997: Una gramatica de la lengua albanesa (review of: M.2000: 'Albanisch und seine Lehnbeziehungen in fruherer Zeit'. 1900: Die gutturale im albanesischen. Heidelberg. *0> ye in front of word­ se11schaft vom 26. -r. Demiraj. 4. . Meyer-Lubke. Bibliography 1.Atlanta.E. 1916: Beitrage zur albanesischen Grammatik. *0> uo. *h2ei and Klingenschmitt. 221­ Diphthongization of stressed *0> uo. W. textos'.1992: Albanian. 1038-1057. Revista de actualizaci6n cientifica 17. unrounding * 0> e. Tempus.Koln. Bonnet. 1998:Review of: V. Cabej. Thumb. . Forschungen 49. Leningrad. 274-300. Comparative Linguistics (VTW) Michiel de Vaan Kortlandt. 5. Orel. Buchholz. Tirane. Aarhus. 3. i-mutation of *a to e. V. MSS61. 2000: Les mots latins de l'albanais.J. H. Orel. 2. ed. Paris .1 (2004). 70-85 Die Sprache 44. p. 1996). Madrid. 837-921. 213-227. In honorem Holger Pedersen. [with a short summary in English] summary in English] Desnickaja. Abschnitt: Romanische Sprachwissenschaft. Wiesbaden. 1987: PIE.219-226. Ras­ final -11.nl Leiden . Tirane. 'Albanian Etymological Dictionary'. Diphthongs -ua-. PIE *-eh2 > *-a > *-a. E. 1987: Albanische Grammatik. -r.84 Michiel de Vaan PIE *e in Albanian 85 1. 1968: Albanskijjazyk i ego dialekty.1995: Comparative Indo-European Linguistics. summary in English] . Fiedler. Hamp. *~> je. Studies in Slavic and General PO Box 9515 Linguistics 10. neu­ bearbeitet von W. . de Vaan. [with a Tirane. PAlb.2002: Gremetike historike e gjuhes shqipe (botim i permbledbur}.Boston . PIE "eand "e in borrowings join lei when stressed in front of -11. Die Sprache 40. 2000: Leksiku historik i gjuhes shqipe. mussen unter Mitwirkung von Benedicte Nielsen. 'El Albanes.2001: Das Meyersche Gesetz uber den Schwund der intervokalischen Svane. . 1909: Altgriechische Elemente des Albanesischen. -n. Media im Albanischen. 1981: Albanisch und Urindogermanisch. gleichende Sprachforschung 36. Indogermanische 2. J. (1lIQ4).229-241. Meyer-Lubke. acquires two different phonemes 1f#1 and lei. B.leidenuniv . K. S. A. 2000: A concise historical grammar of the Albanian language. [with a germanische Forschungen 26. F.Koln 1998. Indogermanische Forschungen 36. Zeitschrift fur ver­ e shqipes. PIE *-om > *-am 1931: Zur Geschichte des alb.