European Balkan(s), Turkic bal(yk) and the Problem
of Their Original Meanings
Marek Stachowski
Jagiellonian University, Poland
Since most European students of Turkic linguistics start their
university training with a course of Turkish they first learn that the
Turkish word balyk ‹balık› means ‘fish.’ It therefore stands to reason
that they feel most surprised when they are told in an Old Turkic
course that balyk means both ‘wall’ and ‘town, city’ in Old Turkic.
One can hardly understand how three meanings as different as ‘fish’
and ‘wall’ or ‘town’ are expressed by one and the same word, and
thus the idea immediately suggests itself that at least one of these
words is probably (or certainly) borrowed into Old Turkic from a
foreign language that is apparently unknown to the student. Besides,
there is only one possible phonetic association, that with Turkish bal
‘honey’ but this does not seem to make much sense.
Also a look into ÈSTJa II does not solve the problem. First of
all, the difference between Turkic balyk ‘wall; town’ and bālyk ‘fish’
(ÈSTJa II 59) is somewhat artificial. One cannot see why the former
word should be read with a short vowel if, in the Arabic script, only
one attestation without elif (i.e. ‹bliq›) and as many as six notations
with elif (i.e. ‹bāliq› ~ [twice:] ‹bālq› for bālyk, ‹bāliḫ› ~ ‹bālḫ› for
balyh, and ‹bāliġ› for balyg) are adduced in ÈSTJa. Of course, a
graphematic argument cannot be considered conclusive; nevertheless,
this proportion does not suggest the reading with a short -a-, as
opposed to bālyk ‘fish.’ Rather, one is tempted to posit a long -ā- in
both words. But, then, the question arises whether we have to deal
Kim Juwon and Ko Dongho eds. Current Trends in Altaic Linguistics: A Festschrift for Professor
Emeritus Seong Baeg-in on his 80th Birthday. Seoul: Altaic Society of Korea, 2013.
613
614 Current Trends in Altaic Linguistics
with polysemy (bālyk ‘1. wall; town; 2. fish’) or homonymy (bālyk 1
‘wall; town’; bālyk 2 ‘fish’). Because the vowel length in these words
can be also secondary (see [1] and [4] below) we are going to write
just ‹a› and to ignore the question of vowel length in what follows.
Three ideas suggested by four scholars concerning the origins
of balyk ‘fish’ have been published hitherto. The oldest is by László
Rásonyi who connected this word with the verb balk(y)- ‘to shimmer,
glimmer, twinkle.’ Èrvand V. Sevortjan is absolutely right when
dismissing this explanation as morphologically well nigh impossible
(ÈSTJa II 60) since a derivative of balk(y)- would have been *balkyk
rather than balyk. Martti Räsänen (VEWT 61b) does not, in fact,
suggest any solution at all because his reconstructions *baluk or
*balyk say nothing about the original structure and meaning of the
word (in essence, they do not even convey any real information about
its original phonological make-up) and the rapprochement of the
Turkic word with both Kalmuk balγă zaγăsn ‘eine Fischart; ? Brassen’
and Manchu falu ‘eine Art Brassen’ does not help a Turkologist in
any way. The third etymology was first suggested by S. P. Tolstov
(1947; cited after ÈSTJa II 60) who thought of a connection of balyk
‘fish’ with words like bal or bar ‘water reservoir, esp. one of stagnant
water on swampy and clay soil.’ È. V. Sevortjan picked up this idea,
pursued it with further examples (e.g. Kazakh dial. bal ‘small ditch’
~ balgy ‘foam, scum’; Azeri dial. balax ‘sedge at a pond or stream
bank,’ Yakut baltyj- ‘to churn up water,’ Kirghiz balčylda- ‘to squelch
in water’) and concluded that a nominal-verbal root *√bal(-) must
have existed in Proto-Turkic whose nominal variant is represented by
Kazakh dial. bal ‘small ditch’ (ÈSTJa II 60). The problem, however,
is that the verbal variant *bal- is not reflected as such in any Turkic
language.
As for balyk ‘wall; town,’ a popular opinion is that this is a
derivative formed from a word for ‘clay.’ In ÈSTJa II 59 a word bal
‘clay’ is admittedly adduced, probably attested in Saryg Uyghur (the
notation in ÈSTJa l.c. is not entirely unambiguous) but it seems to be
a hapax. On the other hand, the standard words for ‘clay’ are balyk
and balčyk in Turkic, both clear derivatives of *bal whose meaning
must, indeed, have somehow been connected with ‘clay.’
European Balkan(s), Turkic bal(yk) and the Problem of Their Original Meanings 615
Clauson (1972: 335sq.) does not give any etymology at all for
either of these words, this being unfortunately a standard situation
rather than an exception in what he has called an etymological
dictionary.
For Turkic bal ‘honey’ Clauson (1972: 330a) says “[i]t is generally
agreed that this is a very early l.-w. fr. [= loan-word from] some
Indo-European language, dating fr. [= from] a period when m-
was so unacceptable as an initial that it was replaced by b-.” This is
generally a correct report on the state-of-the-art although nobody
knows when this mysterious “period” began and, what is even worse,
when it ended, because even Russian words borrowed into Siberian
Turkic languages – a process that began only in the 17th century –
were sometimes subject to the m- > b- change (cf. Russian matrác
‘mattress’ > Tuvinian matrās ~ batrās id. [Schönig 2002: 263]) but
also a reverse change of b- > m- is known (e.g. Russian peč’ ‘oven’
> Tobolsk-Tatar müts id. [Anikin 2003: 448]). And this means that
nobody can establish the terminus ante quem of the borrowing of the
Indo-European word for ‘honey’ into Turkic, so we have no grounds
to say this was a very early loanword.
Worse still, what Clauson (op. cit.) says in the remainder of his
entry is in no way acceptable: “The closest IE parallel is Latin mel; the
Sanskrit form is madhu.” First of all, Latin mel and Sanskrit madhu are
two different words so that madhu cannot just be viewed as a Sanskrit
variant of Latin mel (cf., e.g., Smoczyński 2007: 382). Besides, the
Proto-Indo-European form is generally reconstructed as *melit (>
Hittite militt-, Greek méli, -tos id.) whereas Latin mel, mellis ‘honey’
goes back to Proto-Italic *meli. The origin and the exact mechanism
of the emergence of Latin mel are still unclear and in dispute (de
Vaan 2008: 370). At any rate, both the geography and the chronology
of a Latin or even a Proto-Italic influence upon Proto-Turkic cannot
be possibly viewed as self-evident; on the contrary, it requires proof.
However, Clauson does not elaborate on this idea at all.
In addition, nobody seems to have ever explained why Proto-Indo-
European *-e- should have been changed into Proto-Turkic *-a-.
All in all, against this background, the Indo-European origin of
Turkic bal ‘honey’ appears more than highly questionable.
616 Current Trends in Altaic Linguistics
* *
*
My own opinion on the etymology and interrelations of the words
presented above can be put as follows:
[1] Turkic
bal ‘honey’ is a genuine Turkic word whose protoform
was *bal (or, maybe, *bāl, if the notation ‹bāl› in the Arabic
script in 11th century Turkic comparative dictionary by
Mahmud Kashgari is to be taken literally; compare, however,
Turkmen bal, not *bāl; see also [4] below). Its modern meaning
‘honey’ is secondary, the original one being rather *‘mud, clay;
thick, gluey and kneadable substance.’
The
word bal produced a few derivatives, and some other
words were wrongly supposed, in the history of Turkic
linguistics, to go back to bal ‘honey.’ Two such examples
are adduced in ÈSTJa II 47, unfortunately amongst correct
derivatives and without a clear explanation of their
incorrectness. These are: [1a] Hungarian bálmos, Ukrainian
bánuš ‘a sort of dish, resembling Rumanian mămăligă’
(although no honey is used in its production) that in reality
reflects Turkish bulamaç ‘puree, mash, pulp, porridge’ <
bula- ‘to roll in flour; to stir, mix’ (Stachowski 2012: 194)
and [1b] Rumanian bali(e)mez ‘a very heavy howitzer’ <
Turkish balyemez id. < German Faule Metze, an ironical name
of a heavy howitzer in Brunswick (Kissling 1951: passim); if
this Turkish word was connected with bal ‘honey’ the word
balyemez would have literally meant ‘it does not eat honey’
– indeed, a weird name for a howitzer, one is inclined to ask
what it is that is distinguished from a howitzer solely by the
fact that does eat honey.
[2] One
of the derivatives of *bal is Turkic balčyk ‘mud’ with its
+čyk suffix forming diminutives that have both the standard
meaning ‘small X’ and a less frequent one: ‘quasi X.’ Thus,
the etymological meaning of balčyk ‘mud’ was approximately
‘quasi bal,’ that is something resembling a thick and gluey
European Balkan(s), Turkic bal(yk) and the Problem of Their Original Meanings 617
substance.
[3] The
Turkic word for ‘wall; town’ is generally adduced as
balyk or, sometimes, also baluk or balyg, at any rate, as a
bisyllabic word. Mahmud Kashgari, however, attests also a
monosyllabic variant balk, typical of the pronunciation of
some Oghuz tribes (ÈSTJa II 59). Even if he criticizes it as
incorrect we should reckon with the possibility that the vowel
in the second syllable is secondary, i.e. *bal > *balk > balyk.
This is, however, somewhat uncertain because an epenthetic
vowel would generally be expected to become a moveable
one (as, for instance, in Turkish oğul [acc. oğlu, not *oğulu]
‘son’ < Proto-Turkic *ogul < *okl < *ok > Oyrot, Tuvinian
uk ‘progeny, offspring’) which is not the case with balyk ‹balık›
(acc. balyğy ‹balığı› not *balgy). Thus, the protoform of this
word may tentatively be suggested as *bal(y)k *‘wall made of
clay’ > ‘1. wall(s) > 2. town,’ its etymological-morphological
meaning being something like *‘clay-thing.’
[4] For
Turkic balyk ‘fish’ I would like to suggest that the
word originally concerned the tench or some other species
preferring waters with muddy or clayey bottom. This meaning
was generalized over the course of time, a process due to
migrations of the Turkic peoples that made more and more
fish genera known to them.
The fact that Turkmen bālyk and Kyzyl-Khakas pālyx point
both to a long vowel in the first syllable (cf. [1] above) does
not prove much because Turkic low vowels often tend to
lengthening before syllables with -y- or -i-. Thus, the vowel
length can be secondary in both examples, particularly in
view of the fact that the length in Turkmen bālyk is not
corroborated by the short vowel in Turkmen bal ‘honey.’
In short: The Proto-Turkic word *bal ‘mud; clay; thick and gluey
substance’ produced three derivatives (*balčyk ‘mud’; *bal(y)k ‘wall;
town’; *balyk *‘(?) tench’ > ‘fish’) and then changed its meaning into
‘honey.’
However, before the semantic shift of Proto-Turkic *bal took
618 Current Trends in Altaic Linguistics
place one more derivative must have come into being, namely *balyk
*‘swamp(land), marsh’ > *‘swampy forest.’ It was the latter meaning
that formed a semantic basis of a new derivative: *balyk+an >
balkan ‘wooded mountain(s)’ (this is the traditional translation
but a collective meaning like *‘mountains with swampy forests’ is
probably more correct), and this derivative lives on even today in the
internationally known geographical term: the Balkan(s) (Eren 1987:
118 sq.).
References
ÈSTJa II = Sevortjan È. V. 1978. Ètimologičeskij slovaŕ tjurkskich jazykov
(Obščetjurkskie i mežtjurkskie osnovy na bukvu «B»). Moskva.
VEWT = Räsänen M. 1969. Versuch eines etymologischen Wörterbuchs der
Türksprachen. Helsinki.
Anikin, A. E. 2003. Ètimologičeskij slovaŕ russkich zaimstvovanij v jazykach
Sibiri. Novosibirsk.
Clauson, G. 1972. An etymological dictionary of pre-thirteenth-century
Turkish. Oxford.
Eren, H. 1987. Balkan adının kökeni. – Türk dili 429. 113-124.
Kissling, H. J. 1951. Baljemez. ‒ Zeitschrift der deutschen morgenländischen
Gesellschaft 101. 333-340.
Schönig, C. 2002. Anmerkungen zum b ~ m-Wandel in den modernen
Türksprachen. – Hauenschild I. et al. (eds), Scripta ottomanica et res
altaicae, 257-275. Wiesbaden.
Smoczyński, W. 2007. Słownik etymologiczny języka litewskiego, Wilno.
Stachowski, M. 2012. Ein neues Wörterbuch der türkischen Lehnwörter in
Balkan- und Karpatensprachen und ukrainisch bánuš ‘Maisbrei.’ ‒
Studia etymologica cracoviensia 17(2012). 191-197.
Tolstov, S. P. 1947. Goroda guzov. – Sovetskaja Ètnografija 1947/3. 71-75.
de Vaan, M. 2008. Etymological dictionary of Latin and the other Italic
languages. Leiden – Boston.
Marek Stachowski
Institute of Oriental Studies
Faculty of Phiology
Jagiellonian University in Kraków, POLAND
<stachowski.marek@gmail.com>