pre-publication version

definitive version to be published in the proceedings of:

Ljubilej – Ljubljäum – Ljubilee – Ljubljanniversaire
IG / SIES / SÉIE Arbeitstagung (100 years of comparative linguistics at the University of Ljubljana)
Ljubljana, 4–7 June 2019

THE INDO-EUROPEAN PREHISTORY OF THE

ALBANIAN WORD FOR ‘SISTER’

Marek Majer, University of Lodz

Abstract: In the present article, which builds on an almost overlooked

suggestion by Eric Hamp, it is argued that Albanian motër ‘sister’ is not a direct
reflex of Proto-Indo-European *meh2ter- ‘mother’ (with an unexpected shift of
meaning) – as assumed almost universally in the literature – but that it represents
an early derivative of this word, probably *māt(e)rā < *meh2t-(e)r-eh2 ‘of the
[same] mother’ or ‘having the [same] mother’. Such an item may have been
easily lexicalized as a replacement of the inherited word for ‘sister’; cf. Greek
αδελφ(ε)ός ‘brother’ < *sm̥ -gʷelbʰ-es-o- ‘of the same womb’ or Spanish
hermano ‘brother’ < Latin germānus ‘of [one’s own] kin, real, full’. Thus,
despite widespread claims to the contrary, Albanian ‘sister’ can probably be
explained via fairly trivial developments and does not document any
extraordinary facts about (Pre-)Albanian kinship terminology or social structure.

1. Introduction. Albanian kinship terminology and the word for

‘sister’1
As is well-known, Albanian only preserves a small part of the Proto-Indo-
European kinship terminology.2 Most of the basic kinship terms have been
replaced by innovative terms (be it new coinages, nursery words, or
borrowings), while the familiar items known from other Indo-European

1
Apart from the Ljubljana conference, this research was also presented at the 38th International
Seminar of Albanian Language, Literature and Culture (Seminari XXXVIII Ndërkombëtar për
Gjuhën, Letërsinë dhe Kulturën Shqiptare), Prishtina, 19–30 August 2019; a somewhat abridged
Albanian-language version of the present study will appear under the title Parahistoria
indoevropiane e fjalës shqipe për ‘motrën’ in the proceedings of the above Seminar. I would like to
thank the participants of both events for their helpful comments and corrections; in addition, I would
like to thank Victor Friedman, Plator Gashi, Brian Joseph and Sergio Neri, with whom I discussed
the matter at other occasions. All opinions and errors are my own.
2
For a recent overview of Indo-European kinship terminology in a typological, areal and social
context, with rich further literature, cf. Milanova 2016 and 2018.
2 MAREK MAJER

languages – and thus reconstructible for Proto-Indo-European – have not left a

single uncontroversial trace:3

Table 1. Replacements of core PIE kinship terms in Albanian

PIE term gloss sample reflexes replacements in Alb.
*ph2ter- ‘father’ Ved. pitŕ̥-, Gr. πατήρ, Lat. pater → atë, baba etc.
*bʰreh2ter- ‘brother’ Ved. bhrā́ tr̥ -, Lat. frāter, Toch. B procer → vëlla etc.
*swesor- ‘sister’ Ved. svásr̥ -, Lat. soror, Lith. sesuõ → motër etc.
*suhnu- ‘son’ Ved. sūnú-, Goth. sunus, Lith. sūnùs → bir, djalë etc.
*dʰugh2ter- ‘daughter’ Ved. dúhitr̥ -, Gr. θυγάτηρ, OArm. dowstr → bijë, vajzë etc.

Albanian kinship terms with demonstrable Indo-European ties are limited to a

few less central items, such as e.g. vjehërr ‘father-in-law’ (< PIE *sweḱuro-,
with phonological complications; cf. Ved. śváśura-, Gr. ἑκυρος, PSl. *svekrъ),
nip/mbesë4 ‘grandson, nephew’ / ‘granddaughter, niece’ (< PIE *nepot-
/*neptih2, cf. Ved. nápāt, Lat. nepōs, OLith. nepuotis). Finally, some items may
represent lexical replacements of considerable age, built of Indo-European
elements; cf. gjysh ‘grandfather’ (< Post-PIE *suhsiyos, *suhsēn or similar;5 cf.
Ved. sūṣā ́ , a form with various interpretations).6
However, a curious exception vis-à-vis the list presented in Table 1 above
is the Albanian word for ‘sister’, motër (def. -a; pl. motra, -at; dial. and OAlb.
motërë, motrë). Famously, it appears to be a direct reflex of the Proto-Indo-
European word for ‘mother’, i.e. *meh2ter-7 (Ved. mā́ tr̥ -, Gr. µήτηρ, Lat. māter

3
On the alleged derivation of vajzë ‘girl, daughter’ from PIE *swesor- ‘sister’, cf. further below in
section 3 (with footnote 14).
4
If not instead borrowed from Lat. nepōs and VLat. *nepōtia, cf. Dalm. nepoça; see Demiraj 1997:
259, 301–302.
5
We are probably dealing with a derivative of the root *sewh- ‘give birth’, replacing the inherited
PIE term *h2ewh2o- (Hitt. ḫuḫḫaš, Lat. avus, Toch. B āwe etc.). Cf. Mayrhofer EWAia: 714. I am
indebted to Sergio Neri for a discussion on the possible proto-forms here.
6
It is unlikely that the term joshë ‘grandmother’ has any links to Proto-Indo-European kinship
terminology, despite being accepted as derived from PIE *h2ewh2-ih2 e.g. in Mallory & Adams 2009:
216; cf. Demiraj 1997: 209. However, the innovative word for ‘brother’, vëlla, is probably composed
of Indo-European material; cf. further below in section 4.
7
Conventionally reconstructed in this form in modern Indo-European linguistics, and this notation
is followed here; however, assuming the ubiquitous child language syllable [ma(ː)] as the basis is
perhaps no less likely (recently e.g. Ringe 2017: 16 – attributed to Michael Weiss, who, however,
 THE INDO-EUROPEAN PREHISTORY OF THE ALBANIAN WORD FOR ‘SISTER’ 3

etc.). This is the standpoint accepted in all of the etymological dictionaries of

Albanian (Meyer 1891: 287–288; Çabej 1976: 371; Demiraj 1997: 279; Huld
1984: 96; Orel 1998: 275; Topalli 2018: 1009), in modern handbooks of Indo-
European (e.g. NIL: 457–460), as well as in authoritative treatments of Albanian
in an Indo-European context (e.g. Matzinger 2018: 1789).
Under this scenario, motër is the only basic Proto-Indo-European kinship
term directly reflected in Albanian. This isolated status need not be a problem
in itself, as it may simply reflect the extensive lexical replacement otherwise
assured for Albanian. However, the unexpected meaning of the term motër
certainly does give one pause; it is a conspicuous problem that cannot be
escaped. As such, it has received a fair amount of coverage in the literature.

2. Accounting for the apparent semantic shift: previous

approaches
For the most part, the issue of the ostensible change of meaning in the reflex of
PIE *meh2ter- has been dealt with in two ways.
2.1 ‘Pure’ semantic change.
The first theory, in essence, simply stipulates that PIE *meh2ter- ‘mother’
underwent the perplexing semantic change ‘mother’ > ‘sister’ sometime during
the development to attested Albanian. The precise trajectory is rarely discussed
at length; one could in principle assume a combination of semantic broadening
(e.g. to ‘woman’ or ‘female relative’) with subsequent narrowing (to ‘sister’).
Whatever the details, the word has been used in the literature as a showcase
example of uncanny semantic developments; cf. Eska & Ringe (2004: 572):
One may note that Albanian motër ‘sister’ continues a form of Proto-IE *meh2ter- ‘mother’,
thus demonstrating that semantic shifts in female kin terms are possible and that the
meanings cannot be taken for granted.
The explanation under discussion is of course impossible to disprove, but the
assumption that one of the most central inherited kinship terms would have been
hit by two (or more) consecutive instances of far-reaching semantic change –

opts for *meh2ter- in 2009: 42). The case for *meh2- could be made stronger if a semantically
compelling connection to an otherwise known root of this shape were established, which does not
seem to be the case so far (see a recent novel proposal in Ackermann 2018). For discussion see e.g.
NIL: 459 or Pinault 2007: 275–277 (including a general treatment of Proto-Indo-European kinship
terms in *-h2ter-; cf. also Pinault 2005).
4 MAREK MAJER

unparalleled in any other Indo-European language8 – is rather costly. The

various parallels adduced in the literature fail to replicate the development
allegedly seen here: instead, they mostly demonstrate the predictable polysemy
of assorted nursery words.9 Hence, it is not surprising that other avenues have
been explored.
2.2 Specific circumstances in (Proto-)Albanian kinship structure and
terminology
Under the alternative approach, it is assumed that the divergent denotation of
the reflex of *meh2ter- is not due to a mere haphazard semantic drift of the
‘ordinary’ variety, but rather that it reflects some particular developments in the
structure of the (Pre-)Albanian society. A typical instantiation of the theory is
the assumption that, in that society, older sisters would take custody of their
younger siblings and thus fulfil the role of mothers; the terminology would have
adapted accordingly. Thus, the change would have been largely extralinguistic.
First proposed by Meyer (1891: 287–288), the hypothesis was later widely
embraced, e.g. by Tagliavini (1965: 118), Pudić (1978: 115), Çabej (1976: 371),
and recently Topalli (2018: 1009). In fact, under this approach, the unique
meaning of motër has often been thought to reveal a substantial amount of
information on the entire system of kinship terminology of the (Pre-)Albanians.
For a telling example of how much has been inferred from it, cf. e.g. Huld (1984:
96):
In Omaha-kinship systems of terminology as is reconstructed for Proto-Indo-European, the
term ‘mother’ extends to the mother’s sister and her daughters (...). If the intermediate
kinship terminology of the pre-Albanians was of the Hawaiian type, then these cousin terms
would come to mean ‘sibling’ as well. The transfer to ‘sibling’ only was facilitated by the
rise of Eskimo-type terms in the early Middle Ages when ‘cousin’ was borrowed from
Latin.10

8
The fact that PIE *meh2ter- is also attested with a divergent meaning in Lithuanian (mótė ‘wife’,
móteris ‘woman’) is of little relevance here: the semantic gap is much smaller, while the inherited
meaning ‘mother’ is found both within Lithuanian (mótė dial. ‘mother’) and – exclusively –
elsewhere in Baltic (Latv. mãte, OPr. mūti ‘mother’).
9
E.g. Jokl 1923: 36: “ai. attā Mutter, ältere Schwester, (...) altaisch-tatarisch: ača Vater, aber auch
älterer Bruder, Onkel.”
10
For a background on the kinship terminology types mentioned here (Omaha, Hawaiian, Eskimo)
and others, see Jones 2010: 178–179.
 THE INDO-EUROPEAN PREHISTORY OF THE ALBANIAN WORD FOR ‘SISTER’ 5

Somewhat similar conclusions are reached e.g. by Barić (1955: 57–58), who
claims that the shift in meaning was caused by contact with a pre-Indo-European
substrate. He notes how the inherited Indo-European words for ‘son’ and
‘daughter’ were replaced in Albanian by bir and bijë, respectively, which
resemble Messapic bilia; subsequently, he adds:
This influence probably explains why in the Albanian language the word motër does not
have the meaning ‘mother’, as in the other Indo-European languages, but rather ‘sister’ –
because the kinship terminology of Albanian was influenced by the matriarchal
Mediterranean system, seen so clearly also in the traces of couvade.11
Yet other variants of this hypothesis are certainly thinkable,12 and disproving
any of the above theories is essentially impossible – it is conceivable that one of
them is right. Still, it is necessary to reflect on their relative probability based on
the linguistic evidence alone. It must be borne in mind that all of the above trains
of thought are based on the unexpected semantics of motër alone; there is
otherwise no evidence for any drastic makeovers of family-related terminology
from Proto-Indo-European on the way to Albanian other than mundane lexical
loss. Quite on the contrary, the few other Proto-Indo-European kinship terms
uncontroversially inherited in Albanian (cf. section 1 above) surface exactly
with the expected semantics, as do the early loans. This makes the purported
terminological revolution most unlikely.13 Scenarios like the above are rightly
labeled as doubtful by the more cautious etymological authorities (e.g. Demiraj
1997: 279). Hence, the apparent semantic change seen in motër remains an utter
mystery.

3. Hamp’s derivational approaches

11
“Po me këto ndikime shpjegohet, me shum gjasë, pse në gjuhën shqipe fjala motër nuk ka kuptimin
„nânë“ që ka në gjuhët tjera indoevropjane, por „motër“, sepsè në terminologjin familjare të shqipes
ka ushtrue ndikim sistemi matriarkal mediteran, i cili shihet aq qartë edhè në gjurmat couvade.” The
term couvade refers to the various customs and rituals that the (prospective) father is expected to
follow during the (prospective) mother’s pregnancy and puerperium in certain cultures; see Pauls
2008.
12
Brian Joseph (pers. comm. November 2019) suggests that the shift may be related to the ‘inverted’
kin-address systems found in the Balkans, especially in Bulgarian, whereby a mother refers to her
daughter as ‘mother’; accordingly, a sister could be called by the ‘mother’ word.
13
For a rich, up-to-date background and bibliography on Balkan kinship terminology, see now
Morozova 2019.
6 MAREK MAJER

The only scholar to have proposed a genuine alternative to the problem – not
requiring the assumption of either massive semantic change or purported
makeovers in family terminology – is Eric Hamp.
First (1986: 109), he considers – but ultimately rejects – the analysis of
“*mātrā as a back-formation from *mātrijā = *wesrijā ‘maternal cousin
(fem.)’”. Although not fully satisfactory in view of the required intermediate
steps, the explanation is in fact far superior to the traditional ones described
above: it involves a concrete, feasible derivational path and a down-to-earth
rationale for the lexeme’s semantics. In general, once it is admitted that motër
is not a reflex, but some kind of derivative of the Proto-Indo-European word for
‘mother’, the semantic aspect becomes far less problematic; the only task is to
determine the most probable formal means by which the derivative may have
been created.
The solution that Hamp eventually (1986: 110) opts for14 is entrenched
in his interesting but quite speculative explanation of Alb. vajzë ‘girl, daughter’,
dial. varzë etc., which he holds to be a reflex of PIE *swesor- ‘sister’.15 He
concludes that once the reflex of *swesor- came to mean ‘girl’, there arose the
need to distinguish ‘sister’ clearly, which was achieved by means of a phrase
meaning ‘mother’s girl’: “*mātrā uor- > var-, i.e. motrë (var-)”. Although not
stated explicitly, it seems that what Hamp had in mind was some sort of
appurtenance adjectival derivative of PIE *meh2ter- rather than a paradigmatic
form of the noun. A similar – though even more laconic – statement on motër
can be found in one of Hamp’s later works (1996: 140).

14
Interestingly, in addition to the theories discussed in this paragraph, Hamp came up with yet
another possible explanation (as told to Brian Joseph ca. 1985 and relayed to me in November 2019):
namely, that motër may be a form extracted out of an original dvandva compound ‘mother-sister’,
referring collectively to the females in a family. The term would have been misassigned to the
etymologically wrong female family member in the process of decomposition.
15
This particular aspect, which is the focus point of the theory, is less convincing. According to
Hamp, PIE *swesōr would have become Pre-Alb. *var via a combination of phonological and
analogical changes. The required phonological processes are not uncontroversial, however, and the
sheer number of innovations assumed (to which must be added the shift in meaning from ‘sister’ to
‘girl, daughter’ and the suffixation with -zë) makes the explanation rather cumbersome. This is not
to say, however, that a satisfactory etymology of this item is at hand; it remains rather enigmatic
(Demiraj 1997: 406–407).
 THE INDO-EUROPEAN PREHISTORY OF THE ALBANIAN WORD FOR ‘SISTER’ 7

Regrettably, Hamp’s original and insightful remarks about the word are
not discussed in any etymological dictionary of Albanian.16 It seems, however,
that his analyses clearly went in the right direction. It is possible to formulate a
fuller and less theory-specific account building on Hamp’s suggestions
(importantly, however, entirely independent of the problematic etymology of
vajzë).

4. Formal and semantic analysis

Accordingly, the central assumption must be that Alb. motër is not a direct reflex
of PIE *meh2ter-, but rather continues an early derivative – most probably in
*-eh2-. Crucially, we are not speaking of a mere regularization of the synchronic
morphological type of motër (to be sure, Albanian does not retain the Proto-
Indo-European r-stems as a separate inflectional class, and the fact that the word
mechanically reconstructs as a virtual *māt(e)rā17 is clear in any case; cf. OAlb.
and dial. motërë, motrë). Contrariwise, the present analysis operates with a
genuine derivational process, involving a change in meaning. The semantics of
motër can now be easily accounted for.
In principle, a (Post-)PIE adjective *meh2t-(e)r-o/eh2- could mean either
‘having a mother’ or ‘of the mother, associated with the mother’. Ideally, the
former (possessive) semantics could be expected to show zero-grade or retained
full-grade of the base: cf. *rot-eh2- ‘wheel’ → *rot-h2-o- ‘having wheels’ (Ved.
rátha- ‘chariot’); *h2melǵ- ‘milk’ → *h2melǵ-o- ‘having milk’ (PGerm. *melka-
‘yielding milk’ > OE melc, ModE milch [cow]).18 Conversely, the latter
(appurtenance) semantics could be expected to display vr̥ ddhi, i.e. the
strengthening of the apophonic grade: cf. *diw- ‘bright sky’ → *deyw-o-‘of the
bright sky, divine’ (Ved. devá-, Lat. deus, Lith. diẽvas ‘god’); *ped-‘foot’ →
*pēd-o- ‘of the foot’ (Gr. πηδόν ‘oar end’).19 In view of the most uncertain

16
Hamp 1986 is referenced by Topalli (2018: 1009), but the theory is not included in the discussion.
17
Cf. e.g. Schumacher & Matzinger 2014: 220. As observed by Sergio Neri (pers. comm.), it may
actually be necessary to take the ACC.SG *méh2tereh2m (or *méh2tr̥ rah2m) as the point of departure;
here, the phonological development to *māte/uran > motër(ë) is unassailable. Conversely, in the
NOM.SG *méh2tereh2, one would have to reckon with the development to *mātérā > Alb. †mëtjérë,
with accent shift.
18
For the examples see NIL: 575–578; for the process see e.g. Höfler 2017: 48–49 with references.
19
For the examples see NIL: 72–79, 529–537; for the process see e.g. Nikolaev 2015: 466 with
references.
8 MAREK MAJER

apophonic status of the root of the PIE word for ‘mother’ (*ā? *ā ̆ h2? *eh2? cf.
footnote 6), it is difficult to decide between the two possibilities on a purely
formal basis.
Either, however, could easily be understood as ‘sibling’ if read as ‘having
[the same]20 mother’ or ‘of [the same] mother’, respectively. In this way,
*meh2t(e)r-e-h2 ‘woman having/of [the same] mother’ could be lexicalized as a
replacement of the inherited PIE *swesor-. Similarly motivated replacements
can be observed in other languages; a well-known example is Gr. ἀδελφ(ε)ός
‘brother’ and ἀδελφ(ε)ή ‘sister’ < *sm̥ -gʷelbʰ-es-o/eh2 ‘of the same womb’.21
Notably, in Albanian, the inherited term for ‘brother’ (*bʰreh2ter-) was likewise
replaced with a term likely formed from Indo-European material – vëlla, pl.
vëllezër (probably *swe-lowdʰ-eh2 ‘own kin’, *swe-logʰ-eh2 ‘[from] the same
bed’, or similar; cf. e.g. Demiraj 1997: 417).
Such innovations are presumably rooted in elliptic constructions:
*sm̥ gʷelbʰesos [scil. *h2nēr] ‘[man] from the same womb’, etc. In fact, it is
conceivable that such processes also underlie some of the kinship terms
reconstructible for late Proto-Indo-European itself, for example as per Höfler’s
(2019: 4) internal reconstruction analysis of *snusos ‘daughter-in-law’. The
basis here would have been a deverbal noun *snew-es- (N) ‘connubial bond’ (cf.
the verbal root *snew+bʰ-‘to marry [a man]’), from which an adjective *snu-s-
o- ‘marriage-related’ would have been derived. Accordingly, *dʰugh2tēr snusos
would have meant ‘a daughter by marriage’ (a collocation predating the rise of
feminine grammatical agreement!), and the actually reconstructible Proto-Indo-
European term *snusos (F) ‘daughter-in-law’ would have arisen via ellipsis.
Similarly, in the case at hand, the underlying denotation was presumably
*meh2t(e)reh2 [scil. *gʷenh2] ‘[woman] having/of [the same] mother’. Still, it is
by no means necessary that the collocation was ever fully pronounced, since the
feminine agreement in *-eh2 ensured the correct interpretation.
It is also possible that the original collocation was in fact *meh2t(e)reh2
[*swesōr], ‘sister from [the same] mother’ or ‘sister having [the same] mother’

20
Admittedly, the element ‘same’ is not expressed overtly (it does not seem possible that the first
compound member *sm̥ - would have been lost without a trace my phonological change, even in the
position before the *m- of the root). Still, this does not hamper the analysis significantly, as the
reference to one’s own parent is often implied for pragmatic reasons (cf. usage such as mother
said...).
21
Beekes EDG (v. 1: 20), Chantraine DELG (v. 1: 18–19).
 THE INDO-EUROPEAN PREHISTORY OF THE ALBANIAN WORD FOR ‘SISTER’ 9

– paralleling Hamp’s original formulation mentioned in section 3 above, though

not necessarily interacting with any developments in vajzë or any other kinship
term. The ‘strengthening’ of *swesōr to *meh2t(e)reh2 [*swesōr] could be
motivated by the need to emphasize genuine immediate sibling status (cf. B/C/S
rȍđenī brȁt lit. ‘birth-brother’, as distinguished from brȁt od stríca lit. ‘brother
from uncle’, i.e. ‘cousin’).
Structurally, a thematic derivative in *-t(e)r-o/eh2- built to a kinship term
in *-ter- is unremarkable; similar (Post-)PIE derivatives with various meanings
and functions can be found in a number of different languages.22 The type
*meh2ter-o- may be reflected in PSl. *materъ/*matorъ ‘old’23 (if not from the
verbal root *meh2- ‘age’24), while *meh2tr-eh2 itself is seen in Gr. µήτρᾱ ‘womb’
(< ‘*of the mother’).25

5. Chronology and broader Albanian context

The innovation under discussion – i.e. the replacement of inherited
*swesor- with *meh2t(e)reh2 – probably occurred fairly early in the history of
(Pre-)Albanian. At the time when it was created, the word must have of course
coexisted with *meh2ter- ‘mother’ (which provided its derivational basis);
however, maintaining the distinction between *meh2ter- and *meh2t(e)reh2 was
only possible in a morphological system conservative enough to allow for a clear
distinction of a feminine r-stem from a feminine eh2-stem, which is certainly
nowhere near attested Albanian.
Notably, this would be in line with most of the other replacements
observed in the semantic sphere in question. For example, the nursery word at(ë)
‘father’, replacing *ph2tēr, displays certain archaic morphological
characteristics attested only in remains by the Old Albanian period (traces of a
GEN.SG in *-ī: OAlb. et < *attī).26 As we saw earlier, the Albanian word for
‘brother’ is a synchronically irregular, long-opaque compound formed from
Indo-European elements. Thus – precisely like the innovative term for ‘sister’ –

22
See NIL: 38, 556 for *bʰreh2tr-o/eh2- and *ph2tr-o/eh2-.
23
Thus e.g. ÈSSJa v. 17: 245.
24
Thus e.g. Yakubovich (2010: 487), who regards the association with ‘mother’ an “obvious folk
etymology”.
25
See NIL: 458, where a number of further derivatives of this kind are also discussed.
26
Matzinger 2006: 98, Klingenschmitt 1992: 102–104.
10 MAREK MAJER

these items clearly belong to a deeply archaic lexical layer of Albanian, in spite
of being wholly innovative from the Proto-Indo-European point of view.
Such status of the innovation also parallels certain processes in
grammatical structure, e.g. the rise of the postpositive definite article, whose
phonological history indicates that it predates the written attestation of Albanian
by well over a millennium.27

6. Final remarks
To recapitulate, it appears that the PIE lexeme *meh2ter- ‘mother’ is simply not
directly reflected in Albanian at all, having been replaced by nënë and similar
formations; it shared the fate of all other central Proto-Indo-European kinship
terms in Albanian. Before it was lost, however, it gave rise to a derivative
*meh2t(e)reh2 ‘of/having the [same] mother’, which, like the likewise innovative
*swelowdʰeh2 ‘[of] the same kin’, served to create a new term for referring to
siblings. Thus, motër ‘sister’ parallels vëlla ‘brother’ as an early, Indo-European
based ‘replacement item’ in the Albanian kinship terminology.
Crucially, the meaning of motër can now be dissociated from any
speculation on the potential intermediate systems of kinship terminology or
specific social circumstances ostensibly inferable for (Proto-)Albanian.28 The
lexicalization of ‘the motherly one’, ‘the one from [the same] mother’ or ‘the
one having [the same] mother’ as ‘sister’ appears no more unusual than the case
of Gr. αδελφ(ε)ός ‘brother’ < *sm̥ -gʷelbʰ-es-o- ‘of the same womb’ mentioned
above, or e.g. the Spanish substitution of hermano and hermana (< Lat.
germānus/-a ‘ ‘of [one’s own] kin, real, full’, ultimately a derivative of germen
‘seed’29) for Lat. frāter ‘brother’ and soror ‘sister’. Differently put, it could have
occurred in any language under any conditions – but in a language as innovative
in its lexicon as Albanian, it is perhaps especially unsurprising.30

27
Kümmel 2017: 79–80.
28
Note that even Hamp’s astute observations on the problem at hand are not devoid of this, as he
cites “the social structure and terminology of a Southeast European village” as the key factor
facilitating the change (1996: 140).
29
See discussion in de Vaan 2008: 260–261.
30
On a final note, we may point out that the anonymous and unsourced Wiktionary entry on motër
(https://en.wiktionary.org/wiki/mot%C3%ABr, accessed 1 May 2019) has the term derive from
“Proto-Indo-European *méh2treh2 (“mother’s sister”), variant of *méh2truh2 (compare Old English
modrige, Ancient Greek µητρυιά (mētruiá, “stepmother”)), from *méh2tēr (“mother”)” – an
 THE INDO-EUROPEAN PREHISTORY OF THE ALBANIAN WORD FOR ‘SISTER’ 11

ABBREVIATIONS
ACC accusative NOM nominative
Alb. Albanian OAlb. Old Albanian
B/C/S Bosnian/Croatian/Serbian OArm. Old Armenian
Dalm. Dalmatian OLith. Old Lithuanian
dial. dialectal OE Old English
F feminine OPr. Old Prussian
Goth. Gothic PGerm. Proto-Germanic
Gr. Ancient Greek PIE Proto-Indo-European
Hitt. Hittite PSl. Proto-Slavic
Lat. Latin SG singular
Latv. Latvian Toch. B Tocharian B
Lith. Lithuanian Ved. Vedic Sanskrit
ModE Modern English VLat. Vulgar Latin
N neuter

REFERENCES
Ackermann, Katsiaryna. 2018: Undiscovered (O)CS corpus: On the trails of certain Slavic kinship
terms in time and space. 13th Slavic Linguistics Society Annual Meeting, Eugene, 28–29 September
2018. (Presentation).
Barić, Henrik. 1955: Hŷmje në historín e gjuhës shqipe. Prishtinë.
Beekes, Robert. 2010: Etymological Dictionary of Greek. (Leiden Indo-European Etymological
Dictionary Series 10). Leiden – Boston.
Çabej, Eqrem. 1976: Studime etimologjike në fushë të shqipes. Bleu 5. K– L – LL – M. Tiranë.
Chantraine DELG = Chantraine, Pierre. 1986–80: Dictionnaire étymologique de la langue grecque.
Histoire des mots. Paris.

explanation structurally similar to the one suggested by Hamp as well as the one explored here.
However, disregarding certain formal problems with the analysis and the proposed
cognates/parallels, the attempt to bridge the gap between ‘mother’’ and ‘sister’ via the intermediate
step ‘mother’s sister’ is a somewhat naive assumption. While it may make a good impression in
purely abstract terms, it would be difficult to implement it in actual linguistic reality, where the aunt
is quite far removed from the prototypical image of a sister (incidentally, for the same reason, it does
not seem advisable to see in Alb. motër a haplologized cognate of Lat. mātertera ‘maternal aunt’,
itself from *meh2ter- suffixed with *-tero/eh2-). Nevertheless, the Wiktionary entry may also be
credited for recognizing the Albanian word as a derivative of the Proto-Indo-European term for
‘mother’ rather than its direct reflex.
12 MAREK MAJER

de Vaan, Michiel. 2008: Etymological Dictionary of Latin and the Other Italic Languages. (Leiden
Indo-European Etymological Dictionary Series 7). Leiden – Boston.
Demiraj, Bardhyl. 1997: Albanische Etymologien. Untersuchungen zum albanischen Erbwortschatz.
(Leiden Studies in Indo-European 7). Amsterdam – Atlanta.
Eska, Joseph F. & Ringe, Don. 2004: ‘Recent work in computational linguistic phylogeny’,
Language 80: 569–582.
ÈSSJa = Trubačëv, Oleg N. et al. 1974–2018: Ėtimologičeskij slovar' slavjanskich jazykov.
Praslavjanskij leksičeskij fond. Moskva.
Hamp, Eric P. 1986: Alb. vajzë, motrë. In: Roncaglia, Carlo Messori / Pellegrini, Giovan Battista
(eds.): Studi albanologici, balcanici, bizantini e orientali in onore di Giuseppe Valentini. Firenze:
109–110.
Hamp, Eric P. 1996: The nominative singular of r-stems. Baltistica 31: 139–140.
Höfler, Stefan. 2017: Der Stier, der Stärke hat. Possessive Adjektive und ihre Substantivierung im
Indogermanischen. Ph. D. dissertation, Universität Wien.
Höfler, Stefan. 2019: Story of O. On a peculiar substantivization type in PIE. 5th Indo-European
Research Colloquium, Leiden, 21–22 March 2019. (Handout).
Huld, Martin E. 1984: Basic Albanian Etymologies. Bloomington.
Jokl, Norbert. 1923: Linguistisch-kulturhistorische Untersuchungen aus dem Bereiche des
Albanischen. Berlin.
Jones, Richard R. 2010: Kinship systems. In: Birx, H. James (ed.): 21st Century Anthropology. A
Reference Handbook. Los Angeles: 174–181.
Klingenschmitt, Gert. 1992: Die lateinische Nominalflexion. In: Panagl, Oswald / Krisch, Thomas
(eds.): Latein und Indogermanisch. Akten des Kolloquiums der Indogermanischen Gesellschaft,
Salzburg, 23.–26. September 1986. (Innsbrucker Beiträge zur Sprachwissenschaft 64). Innsbruck:
89–135.
Kümmel, Martin. 2017: Is ancient old and modern new? Fallacies of attestation and reconstruction
(with special focus on Indo-Iranian). In: Goldstein, David M. / Jamison, Stephanie W. / Vine,
Brent (eds.): Proceedings of the 27th Annual UCLA Indo-European Conference. Bremen: 79–96.
Mallory, James P. / Adams, Douglas Q. 2009: The Oxford Introduction to Proto-Indo-European and
the Proto-Indo-European World. Oxford – New York.
Matzinger, Joachim. 2006: Der altalbanische Text Mbsuame e krështerë (Dottrina cristiana) des
Lekë Matrënga von 1592. Eine Einführung in die albanische Sprachwissenschaft. Dettelbach.
Matzinger, Joachim. 2018: The lexicon of Albanian. In: Klein, Jared S. / Joseph, Brian D. / Fritz,
Matthias (eds.): Handbook of Comparative and Historical Indo-European Linguistics, vol. 3.
Berlin – Boston: 1788–1800.
Mayrhofer EWAia = Mayrhofer, Manfred. 1986–2001: Etymologisches Wörterbuch des
Altindoarischen. Heidelberg.
Meyer, Gustav. 1891: Etymologisches Wörterbuch der albanesischen Sprache. Strassburg.
Milanova, Veronika V. 2016. (Proto-)Indo-European kinship terms in *-ter and (ancient) age-grade
systems. Kin, clan and community – a symposium on Indo-European social institutions,
Copenhagen, 17 May 2016. (Handout).
Milanova, Veronika V. 2018. Praindoevropejskie terminy rodstva na *-ter i vozrastnye stepeni. In:
Popov, Vladimir A. (ed.): Algebra rodstva, vol. 16 (Rodstvo. Sistemy rodstva. Sistemy terminov
rodstva). Sankt-Peterburg: 70–97.
Morozova, Maria S. 2019: Language contact in social context: Kinship terms and kinship relations
of the Mrkovići in southern Montenegro. Journal of Language Contact 12: 305–343.
Nikolaev, Alexander. 2015: ‘The Germanic word for ‘sword’ and delocatival derivation in Proto-
Indo-European’. Journal of Indo-European Studies 37: 462–488.
 THE INDO-EUROPEAN PREHISTORY OF THE ALBANIAN WORD FOR ‘SISTER’ 13

NIL = Wodtko, Dagmar S. / Irslinger, Britta / Schneider, Carolin. 2008: Nomina im

indogermanischen Lexikon. Heidelberg.
Orel, Vladimir 1998: Albanian Etymological Dictionary. Leiden – Boston – Köln: Brill.
Pauls, Elizabeth P. 2008: Couvade. In: Encyclopædia Britannica Online (URL:
https://www.britannica.com/topic/couvade, accessed 1 November 2019).
Pinault, Georges-Jean. 2005: Analyse étymologique d’un nom de parenté indo-européen. In:
Schweiger, Günter (ed.): Indogermanica. Festschrift Gert Klingenschmitt. Taimering: 465–486.
Pinault, Georges-Jean. 2007: A star is born: a “new” PIE *-ter-suffix. In: Nussbaum, Alan J. (ed.):
Verba Docenti, Studies in Historical and Indo-European Linguistics Presented to Jay H. Jasanoff
by Students, Colleagues, and Friends. Ann Arbor – New York: 271–279.
Pudić, Ivan. 1978: Verwandtschaftswörter in den Balkansprachen. In: Actes du IIe congrès
international des études du sud-est Européen (Athènes, 7–13 mai 1970), tomе IV, Linguistique еt
littérature. Athènes: 111–117.
Ringe, Don. 2017: From Proto-Indo-European to Proto-Germanic. A Linguistic History of English.
Volume I. 2nd ed. Oxford – New York.
Schumacher, Stefan / Matzinger, Joachim. 2013: Die Verben des Altalbanischen. Belegwörterbuch,
Vorgeschichte und Etymologie. Wiesbaden.
Tagliavini, Carlo. 1965: La stratificazione del lessico albanese. Elementi indoeuropei. Bologna.
Topalli, Kolec. 2018: Fjalor etimologjik i gjuhës shqipe. Durrës.
Weiss, Michael. 2009: Outline of the Historical and Comparative Grammar of Latin. Ann Arbor:
Beech Stave Press.
Yakubovich, Ilya. 2010: Indo-European *mā ‘to grow’. In: Kazanskij, Nikolaj N. (ed.):
Indoevropejskoe jazykoznanie i klassičeskaja filologija VIII.: Materialy čtenij, posvjaščennych
pamjati professora Iosifa Moiseeviča Tronskogo. 21–23 ijunja 2010. Č. II. Sankt-Peterburg: 478–
492.