18.4.28.23:59: TABLES AND FALCONS: THE FATE OF FINAL *L IN SLAVIC

Polish kiełbasa /kʲewbasa/ from my last two entries is spelled with ł but is no longer pronounced with an [l].

Standard Polish once had three kinds of phonetic laterals, but only two survive today: a palatal allophone before /i/ and a dental allophone elsewhere.

The reflexes of Polish laterals seem straightforward: old hard *l becomes /w/ and old soft *lʲ becomes /l/.

Hence *stolъ 'table' and *sokolъ 'falcon' became Polish stół /stuw/ and sokół /sokuw/.

(I can't predict when *o became ó /u/.)

What does not seem straightforward to me is the fate of syllable-final *l in Ukrainian, Belarusian, and Serbo-Croatian.

Ukrainian /stojaw/ 'stood' (masc. sg.) < *-l

Belarusian /stajaw/ 'stood' (masc. sg.) < *-l

Serbo-Croatian /stajao/ 'stood' (masc. sg.) < *-l

The best-known example might be Serbo-Croatian /beograd/ (cf. English Belgrade reflecting earlier *l).

Nonetheless, 'table' and 'falcon' may retain *-l:

Ukrainian /stil/, /sokil/

Belarusian /stol/, /sokal/

Serbo-Croatian /sto/ (Serbia) ~ /stol/ (Croatia), /soko/ (Bosnia, Serbia) ~ /sokol/ (Croatia)

(Countries are from Wiktionary entries.)

In Belarusian, word-final *l remains except in the past tense masculine singular (Mayo 1993: 893). (Did it erode there due to high frequency?)

The situation in Ukrainian seems similar, though I know of one case of /w/ < *l that is not a past tense masculine singular: /piw/ < *polъ 'half'.

Could /l/ retention in Croatian stol 'table' be motivated by avoiding homophony with 'hundred' which is /sto/ across Slavic? That doesn't explain Croatian sokol 'falcon', though. Browne and Alt (n.d.: 20) write,

In adjectives and nouns it [*l > o] is widespread though some words avoid it: masculine singular nominative mio [< *mil] 'nice', feminine mila, but ohol 'haughty', feminine ohola.

I assume borrowings postdating *l-shifts retain final -l in Serbo-Croatian: e.g., hotel (not †hoteo).

Ukrainian and Belarusian seem to favor borrowing foreign -l-words with /lʲ/:

U /hotelʲ/, B /hatelʲ/ 'hotel'

U /alkoholʲ/, B /alkaholʲ/ 'alcohol'

but U <mark hemill> and B <mark hèmil>, both /mark hemil/ 'Mark Hamill'. (The B form is from the B Wikipedia entry for the original Star Wars [Зорныя войны. Эпізод IV: Новая надзея].)

4.29.21:57: Added Mayo on Belarusian, Ukrainian /piw/, Browne and Alt quotation, and everything after that.

18.4.27.23:45: IRREGULARITIES IN 'KIELBASA' REVISITED

Yesterday I discovered in de Bray's (1980: 258) book on West Slavic that Polish kiełbasa /kʲewbasa/ is in fact the regular reflex of an earlier *kl̩basa (cf. Slovak klbása ~ klobása). I assume his hard *l̩ goes back to Proto-Slavic *ъl.

But I still don't know how to account for the front vowels of

Ukrainian ківбаса <kivbasa> < *kilbasa

Belarusian кілбаса <kilbasa> ~ келбаса <kelbasa>

Are they borrowings of forms resembling Polish kiełbasa or pre-Polish (proto-West Slavic?) *kl̩basa? If they are from *kl̩basa, their front vowels could have been inserted to avoid /klb/-clusters that are not possible in East Slavic.

My guess is that Belarusian келбаса <kelbasa> is a borrowing from Polish kiełbasa, whereas Belarusian кілбаса <kilbasa> is an older form with an epenthetic vowel.

Ukrainian ківбаса <kivbasa> was presumably borrowed as *kilbasa before *l > <v> /w/. I don't think it's from Polish since

- the height of the first vowel doesn't match

- Polish ł apparently became [w] in the standard language only in the early twentieth century (Wikipedia); Morfill (1884: 1) says it is "a very strong l", not [w].

- Polish ł is still [ɫ̪] and not [w] in eastern dialects of Polish in contact with Ukrainian (Wikipedia)

A recent borrowing from the modern standard pronunciation of kiełbasa would be †<kevbasa> and a borrowing from a pre-20th century standard pronunciation or an eastern dialectal pronunciation would be †<kelbasa>.

18.4.25.23:45: IRREGULARITIES IN 'KIELBASA'

Wiktionary derives Polish kiełbasa /kʲewbasa/ and its relatives from a Proto-Slavic *kъlbasa, in turn borrowed from some Turkic word similar to modern Turkish külbastı 'roasted meat', lit. 'ash-pressed'. Irregularity within Slavic implies that the word was borrowed more than once.

The Polish word and nonstandard forms like

Ukrainian ківбаса <kivbasa> < *kilbasa

Belarusian кілбаса <kilbasa> ~ келбаса <kelbasa>

have front vowels /i e/ that I would not expect from Proto-Slavic *ъ.

*külbasa < *kölbasa < *kolbasa (cf. standard Ukranian ковбаса <kovbasa>)< *kъlbasa

but *o only raises to і in standard Ukrainian before a lost weak jer (*ъ or  *ь) which wasn't in this word. Maybe the <kivbasa> dialect worked differently.

My current guess is that the /i e/ vowels in Polish, Ukrainian, and Belarusian reflect attempts to imitate Turkic ü and are not from *o or *ъ.

The Belarusian forms have /l/ instead of /w/ < *l corresponding to Ukranian <v> /w/ < *l and Polish ł < *l. This suggest that the Belarusian borrowings postdate the shift of *l to /w/ in Belarusian. But maybe I misunderstand when *l becomes /w/ in Belarusian.