Open main menu

chow mein

Is "chow mein" from Mandarin or from Taishanese? According to the Wikipedia article chow mein, it's from Taishanese. Justinrleung (talk) 01:41, 1 September 2015 (UTC)

I think that version of the etymology dates to when we split Chinese into all the dialects, and changed Chinese to Mandarin as the default when we didn't have any dialect specified. It looks like the etymology originally said "Chinese", and pinyin was used as the romanization just because we did that for all the Chinese entries. I'll leave the question of which lect the English word came from to others who know more than I do. I will say that The Mandarin form doesn't fit all that well. Chuck Entz (talk) 01:57, 1 September 2015 (UTC)
I replaced it with zh, which at least isn't inaccurate. The OED doesn't specify, and Wikipedia doesn't have a source for saying it's from Taishanese; from sound alone, I reckon Taishanese or Hakka would be good bets, but I don't know. As an added problem, yue-toi (for Taishanese) doesn't even seem to work as an etymology-only language. @WyangΜετάknowledgediscuss/deeds 20:28, 1 September 2015 (UTC)
Taishanese seems right: [tsɔu55 mein32]. Wyang (talk) 04:00, 2 September 2015 (UTC)
You might have better luck with yue-tai instead of yue-toi. Either that or change the language name to Toishanese so they match... ;-p Chuck Entz (talk) 06:34, 2 September 2015 (UTC)
The langrev subpages were what had that error; I have now fixed them and updated the entry. Thanks, Wyang. —Μετάknowledgediscuss/deeds 00:08, 6 September 2015 (UTC)

Illyrian entries

There are a few recent entries in the as yet uncreated Category:Illyrian lemmas, but I'm a bit leery of entries in a basically unattested language like this. Yes, there are mentions and other indirect means of reconstructing terms in the language, but I'm not so sure that the person who added these knows anything at all about the language, let alone can sift through the uncertainties associated with indirect attestation. Probably the most clear-cut example of what I'm concerned about is Dalmatae, which may be based on an Illyrian demonym, but which looks very Latin to me.

I should mention that this person has been adding a lot of etymologies to Albanian entries that seem to be somewhat indiscriminately pulled out of various references. Given that the question of whether Albanian is descended from Illyrian is rather emotionally loaded, I'm concerned: I don't know much about Albanian etymology, but the pattern makes me nervous.

I realize that it only takes a mention to verify one of these, but I would like to know if there's anything reliable that corroborates them as they're currently constituted. Thanks! Chuck Entz (talk) 03:28, 1 September 2015 (UTC)

The user who created these, 8mike, seems to be a problematic editor. @Etimo, Vahagn Petrosyan: What is the story behind his edits? Is he as relentlessly POV-pushing as he seems to be? If so, he should be warned that he will be blocked if he continues; I don't know the linguistic context well enough to judge.
As for the entries: I made Dalmatae into a perfectly good Latin entry, which is of course what it should have been. I'm not clear how the other two are sourced, so I'm going to RFV them and see what happens there. —Μετάknowledgediscuss/deeds 05:12, 1 September 2015 (UTC)
I appreciate that he's added a lot of Albanian entries. I can't speak to the accuracy or inaccuracy of the etymologies in general (in the case of adur specifically, I'll comment on RFV), but hopefully we can resolve this in such a way that he gets to continue adding Albanian entries/translations. - -sche (discuss) 06:52, 1 September 2015 (UTC)
I have interacted with the editor in question. I don't think he is a relentless POV-pusher but his (good-faith) etymologies are unreliable original research and should be removed. The Illyrian entries too are probably unreliable and hopefully will fail rfv. --Vahag (talk) 08:03, 1 September 2015 (UTC)
What is it with Albanian editors recently? —JohnC5 12:54, 1 September 2015 (UTC)
Probably something in their water. --Vahag (talk) 14:06, 1 September 2015 (UTC)
Not just recently. Albania is in a part of the world that's been through a lot of ethnic conflicts where they haven't fared too well, they're still recovering from an extremely harsh, isolationist dictatorship, and they belong to a very distinct branch of the Indo-European languages with no sources going back very far and relatively little attention from scholars. The latter factor means that there's an obvious void that it's tempting to fill with guesswork, and the rest means that there's an emotional vested interest in connecting to something grand and glorious in the ancient past. There have been problems with all of the more prolific adders of Albanian etymologies that I can think of Chuck Entz (talk) 14:26, 1 September 2015 (UTC)
Very interesting. I wish I knew more about Albanian so that I could help. I just have a copy of Orel and this, which are useful from a PIE standpoint but not from an adding entries standpoint. —JohnC5 01:17, 2 September 2015 (UTC)

I honestly don't know. I went through this user's contributions, but I didn't find emotional writing (perhaps some unreferenced etymological explanations which smell of folkish), perhaps you could bring some examples that I have overlooked. Some of the entries and edits are not referenced (pishë, llohë, haromë, kursej, kërnalle), others are, some others seem quite arbitrary or dubious (brerore, katërshor, ish, ëzë), some even wrong (tokë, prrar, mpiks, vervele), while others look ok. As far as the toponym Dalmatia (or Delminum) goes, there are many scholars who have linked it with Albanian delme, as with other toponyms related to animals and vegetation throughout the Balkans which are better explained through Albanian (Ragusa, Ulqin, Dardania etc), BUT, considering the lack of Illyrian writings and the dearth of inscriptions, these are to be considered only as assumptions, not facts, and this should always be mentioned. Etimo (talk) 19:28, 1 September 2015 (UTC)


The etymology of 因特網 in Chinese is currently as follows:

However, 因特網 comes from the transliteration of "inter-" (not "Internet") + the translation of "net". Is there a special term for this type of translation? Justinrleung (talk) 02:06, 2 September 2015 (UTC)

A partial calque. DTLHS (talk) 02:59, 2 September 2015 (UTC)
We have a template {{calque}}, maybe we should have one for partial calques, or modify {{calque}} so it accommodates partial calques. A famous English example is liverwurst, where we translated the Leber part of Leberwurst but borrowed the Wurst part. —Aɴɢʀ (talk) 09:26, 2 September 2015 (UTC)
And conversely, Germans might wash their Wurst down with a glass of Milchshake. Smurrayinchester (talk) 15:33, 3 September 2015 (UTC)

Old English þon

Campbell´s Old English, love it though I do, seems only to say of this form that its 'classification as instrumental is traditional, but reflects neither its origin nor prevailing use.' Perhaps he finishes his thought elsewhere in the book, but I have no patience to scour it. Am I correct in assuming the following: þon < earlier þan < *þáne (unaccented Indo-European final 'ĕ' lost in Germanic, cf. G οἶδε, Goth wait) < IE *to(i)ne (whence Sanskrit तॆन tēna, OBulgarian тѣмь tēmĭ)?

A trifling question, I know, but when I can't find it on this site I have no other recourse. —Colin

Puerto Rico

Why is "Puerto Rico" borrowed from Spanish and pronounced as if it were Portuguese or something? ばかFumikotalk 01:43, 3 September 2015 (UTC)

Probably because monolingual English speakers recognize puerto as a cognate of port, and pronounce it like the latter. Besides, the normal US English way of sounding out spellings would probably produce something pronounced like pure + toe, which is kind of odd, and the original Spanish pronunciation doesn't fit very well with English phonotactics. Chuck Entz (talk) 03:20, 3 September 2015 (UTC)
It even used to be spelled Porto Rico; see note a at w:Puerto Rico#Notes. And it still is spelled that way in French, Italian, and Portuguese, of which it could be considered a translation only in Portuguese (to be a translation in French it would have to be "Port-Riche" and in Italian it would have to be "Porto Ricco" with a double c). —Aɴɢʀ (talk) 15:25, 3 September 2015 (UTC)
I don't buy the phonotactic explanation; it could easily have been pronounced p-ware-to. It must be a historical reason. Perhaps (I'm 100% speculating here) the dialect of Spanish that was first brought there pronounced the word more like puorto or porto. Do we have any experts here on historical Spanish dialects? --WikiTiki89 15:41, 3 September 2015 (UTC)
Some Asturian dialects have puortu. — Ungoliant (falai) 15:44, 3 September 2015 (UTC)
Yes, I noticed that (some apparently even have puorto), but it would have to match up historically, and I'm not an expert on the history of the settling of Puerto Rico. Another theory I just thought of is that it was meant to be pronounced /pwɚ-/, which could much more easily have morphed into what we have now, but this does not explain the spellings in other languages that Angr mentioned. --WikiTiki89 15:57, 3 September 2015 (UTC)
  • FWIW, the island of Vieques (administered as a county of Puerto Rico) has a distinct accent where they drop a lot of "s" sounds, perhaps via a process similar to what happened in French. For instance, the island's name of Vieques is often pronounced (and has been written historically) as Bieke. The common Spanish verb form está comes out more as /e̥hta/. I've been told that this accent is a remnant of the Asturian accent of the island's Spanish settlers. I note too that the Puerto Rico WP article mentions that Asturians were a sizable proportion the Spanish migrants to the islands. ‑‑ Eiríkr Útlendi │Tala við mig 08:55, 29 September 2015 (UTC)
For what's it worth, the few times I've heard the name said by British English speakers, they've pronounced it the Spanish way. Smurrayinchester (talk) 15:31, 3 September 2015 (UTC)
If you pronounce the t in the British way, it is not difficult to say puerto like the Spanish. But if you use the American pronunciation of the t, it is difficult to pronounce that word in English. I speak Spanish (I have a degree in Spanish), and I have no problem saying Puerto in Spanish; it is difficult for me to say puerto in English, however, and I would never use that English pronunciation unless I was pretending to be pompous and pedantic. In English I say porto, regardless of the spelling. —Stephen (Talk) 16:40, 3 September 2015 (UTC)
In older books, other Puertos are (not always consistently within a work) also rendered Porto, including Porto Bello (Portobelo, Panama, but with older works giving Puerto... as the Spanish name), Porto Cabello (Puerto Cabello, Venezuela), and Porto Principe (Puerto Principe = Camagüey, Cuba). This suggests a general process may have been at work on Porto Rico, rather than one-off Asturian influence. That process might have been anglicization, or the 'levelling' of the Spanish Puertos and Portuguese Portos to one word.
- -sche (discuss) 19:23, 29 September 2015 (UTC)

arrop i talladetes

Can anyone reference or confirm this etymology? A direct borrowing from Catalan seems more likely. — Ungoliant (falai) 03:47, 5 September 2015 (UTC)

Icelandic klifa from *klibjaną?

I don't pretend to know much about Proto-Germanic, but looking at the descendants in other languages from *klibjaną (to stick) one would expect an Old Norse/Icelandic form klifa. There is such a word in Icelandic which is nowadays only used in the set phrase klifa á (to harp on about (something)). It doesn't seem too unlikely to me that a figurative meaning such as this could have arisen from an original meaning of "to stick". Opinions? BigDom 07:58, 5 September 2015 (UTC)

Icelanidic klifa comes from the Old Norse klifa (to repeat, harp on"; also "to climb), which seems indeed to come from *klibjaną. Leasnam (talk) 20:39, 5 September 2015 (UTC)
added. Leasnam (talk) 20:44, 5 September 2015 (UTC)
FYI, "to climb" is klífa (from *klībaną), not klifa. BigDom 06:24, 6 September 2015 (UTC)
I don't quite see how any of the attested forms could have come from *klibjaną. Where did the -j- go in Old Norse? Shouldn't it be klifja? And why no gemination in West Germanic? Shouldn't it be clibban in Old English? —Aɴɢʀ (talk) 21:13, 5 September 2015 (UTC)
The reconstruction of class 3 weak verbs is not really settled. The forms we have are based on Ringe's reconstruction, which admittedly leaves a lot of question unanswered. But the most common alternative, to reconstruct -ēną as the ending, is equally puzzling as the long ē apparently fails to become ā in West Germanic. Also, some class 3 weak verbs do have a clear -j- in both Old Norse and West Germanic, such as *sagjaną. —CodeCat 23:38, 5 September 2015 (UTC)
This verb is iterative, meaning 'to repeat' as BigDom mentioned (from Vigfusson-Cleasby). It is, then, in all likelihood a class 2 weak, especially because it lacks the features pointed out by Angr. The Proto form would be *kliƀōjōn, and so by regular rules Norse would lack the -j- ("blocked" by the bimoric vowel) and, therefore also, the geminate -f-. If it really derives from the mentioned *kliƀjaną, it was perhaps removed early to this class in ON because of its meaning, as sometimes happens. —Colin
I have updated the entry to show that klifa may not be a direct descendant Leasnam (talk) 18:17, 7 September 2015 (UTC)
Gerhard Koebler's Old Norse dictionary has this: klif-a, an., sw. V. (3): nhd. wiederholen. So the "repeat" sense dates to Old Norse, but it was also still a class 3 weak verb. It must be a direct descendant. —CodeCat 18:23, 7 September 2015 (UTC)
Thats what I used in the first place...okay so the parentheses can come off (barring any objections, of course). Excellent. Leasnam (talk) 04:11, 8 September 2015 (UTC)

mee#Etymology 2

I believe this should be Min Nan, not Cantonese. (see 麵#Pronunciation)—suzukaze (tc) 06:50, 6 September 2015 (UTC)

This seems possible, since it is pronounced as in Min Nan. Justinrleung (talk) 22:46, 20 September 2015 (UTC)
I agree. In Cantonese the word is pronounced "meen". Smuconlaw (talk) 20:41, 10 October 2015 (UTC)


The etymology section (English) is confusing and reads like it was written by several arguing people (who have nothing but contempt for each other). WurdSnatcher (talk)

Pretty dreadful - and also most of the "information" pertains to Niger, not Nigeria. I've shorn it down to a barebones "From Medieval Latin" etymology, and directed people to Niger#Etymology for further information. Smurrayinchester (talk) 09:40, 11 September 2015 (UTC)
Thanks! WurdSnatcher (talk)

Proto-Germanic *ēlaz

Why do we reconstruct an e-vowel when all the descendants seemingly have a reflex of an a-vowel? --WikiTiki89 22:02, 11 September 2015 (UTC)

If this is a general rule, then let's say this question is about Germanic phonology in general. --WikiTiki89 22:08, 11 September 2015 (UTC)
I believe the derives from PIE / *eh₁ and is continued in Gothic as ē; though not for this etymon (See *dēdiz). —JohnC5 22:51, 11 September 2015 (UTC)
Also, more information here. —JohnC5 23:00, 11 September 2015 (UTC)
In all the descendants, the vowel is long. That points to an original long ē. There was no long ā in Proto-Germanic. —CodeCat 23:08, 11 September 2015 (UTC)
If this word were attested in Gothic it would be els (e is always long in Gothic). The vowel that shows up in Gothic as e and in North and West Germanic as ā (but ǣ in Old English and ē in Old Frisian) is usually reconstructed as *ē, sometimes written *ē₁ or *ǣ. —Aɴɢʀ (talk) 06:18, 12 September 2015 (UTC)
For words like this where there are clear differences, it might be useful to list the reconstructible Northwest Germanic form (*ālaz) somewhere, perhaps in the list of descendants, or under a ===Reconstruction=== header.
(Also, PGmc did have *ā, from earlier *aja, in e.g. *stāną (to stand).) --Tropylium (talk) 15:05, 12 September 2015 (UTC)
Is there a case of this *ā that has a Gothic descendant? We don't seem to have one listed at *stāną. --WikiTiki89 19:33, 12 September 2015 (UTC)
It's an invention particular to Ringe, used to explain these verbs along with class 3 weak. So the only descendant of ā would be in the class 3 weak verbs in Gothic, if this explanation is valid. Other than that, the only source of Gothic ā is the sequence -anh-. —CodeCat 19:39, 12 September 2015 (UTC)
So you're saying that the Gothic reflex of this hypothetical *ā is ā? --WikiTiki89 19:55, 12 September 2015 (UTC)
Yes. —CodeCat 21:07, 12 September 2015 (UTC)


RFV of the Latin etymology.

This shows signs of too much rearranging by people who didn't understand what they were rearranging. As User:Dave crowley pointed out on the talk page, the un-metathesized hypothetical Latin equivalent isn't compatible with the metathesized Proto-Italic form, and the current wording is very confused about what changes happened when and why. Chuck Entz (talk) 06:54, 13 September 2015 (UTC)
Besides the incorrect *volquus, I don't see much that's wrong. I cleaned it up a bit and added a reference for what I changed. (The putative form that should have come about would be *luquus/*lucus, but I don't see much point in mentioning it. —Μετάknowledgediscuss/deeds 07:40, 13 September 2015 (UTC)

Terms with both known and unknown etymology

We normally categorize words in multiple etymological categories if their history can be traced thru multiple languages (e.g. Category:Spanish terms derived from Latin has, or at least should have, substantial overlap with Category:Spanish terms derived from Proto-Indo-European).

However, this practice might not work very well when it comes to "non-source" etymology categories, e.g. Category:Spanish terms with unknown etymologies. Consider matar: the page presents several possible etymologies, two from Latin and one from Arabic. Yet it is also classified together with words like lacha, with no known origin. But clearly we're dealing with two different things here: actually unetymologized words, and words with competing etymological proposals. (We do have e.g. Category:Spanish terms with multiple etymologies, but this appears to be actually for etymologically distinct homonyms, not for cases of disputed etymologies.) — Another type of example: Finnish hiili, which derives from a Proto-Finnic root; but it is also categorized as being of unknown origin (since the PF root is as well).

I suggest that the use of {{unk.}} should be restricted for terms whose origin is actually non-trivially unknown:

  • if a single term has multiple etymologies, we could bring in a category like "LANG terms with disputed etymologies" (or, I guess, leave the conflict only in the text).
  • if a term in a parent language is of unknown origin, only that entry should be categorized as being of unknown origin, not the descendants in the daughter languages.

("Non-trivially" in that we do not know the "ultimate" etymology of most words; all non-neologisms can at best only be traced back to oldest known proto-languages like Proto-Indo-European.)

I'm not sure if any editors are actively following either of the usages I'm questioning, but I'll raise this here for the record, before editing the template documentation and/or WT:ETYM. --Tropylium (talk) 15:05, 13 September 2015 (UTC)

I definitely agree on point 2. All words can only be traced so far, so there is always a point where we don't know anymore. As for point 1, I think "uncertain" or "unclear" is better than "disputed". —CodeCat 15:09, 13 September 2015 (UTC)


This is so obvious it discredits professional etymologists for not figuring it out. It is related to Latin digitus "finger" and Greek deiknumi "show." So dog originally meant "pointer." —This unsigned comment was added by The Sage of Main Street (talkcontribs).

  • Ho ho. We have a place for these but I've forgotten what it's called. Anyway, since most etymology is makebelieve I might as well leave it. SemperBlotto (talk) 14:09, 14 September 2015 (UTC)
  • Funnily enough, these words do appear to have common relatives in modern English... but these are teach (*taikijaną), token (*taikną) and toe (*taihwǭ). There was a strong tendency for the d sound to become t – hence why we say two when the Romans said duo. Smurrayinchester (talk) 15:21, 15 September 2015 (UTC)
    • Not so much a strong tendency as a law. —Aɴɢʀ (talk) 15:59, 15 September 2015 (UTC)


(see this rather hasty example of Australian politics)

Noun sense 8 is:

(Australia, politics) A declaration that the leadership of a parliamentary party is vacant, and open for re-election. Short form of leadership spill

Which sense of spill does this derive from? I've found this memo from February 2015, which talks about "a motion to spill the leadership positions [of the party]", which sounds like it's a reference to throwing away the current leadership. But that still doesn't really explain why the word "spill" is used, rather than, I don't know, "election"? Any Aussies know where this jargon comes from? Keith the Koala (talk) 15:58, 14 September 2015 (UTC)

Could it have something to do with take a spill? Chuck Entz (talk) 01:51, 15 September 2015 (UTC)
My guess would be noun-sense-2: "A fall or stumble." (same sense as take a spill). If it was called an "election" that would imply a public election, but the election was internal (voted on by elected members of the party), so it's a spill. A public election is not a spill. No idea where it originates. Didn't realise it was an Australian thing, and really I'm not even sure any more which part of the process gets called a spill or how broadly the term can apply. Also I notice there's also an obsolete sense of "spill" meaning "to overthrow a person" listed in OED which might be related, but might be a coincidence. —Pengo (talk) 15:26, 15 September 2015 (UTC)
Our definition probably needs some work. Neither OED nor Macquarie definitions even mention re-election or leadership. OED: "A vacating of other posts after one important change of office." Macquarie: "the declaring vacant of a number of positions when one above them falls vacant" These definitions make it sound like "spill" refers to the change "spilling over" to the less senior positions, though I don't feel like any of the news coverage was really referring specifically to this. It was more about the contest ("Turnbull wins spill"), and treating the spill as the overall event rather than just the declaration ("Tony Abbott's love of onions recalled on social media during leadership spill") and to refer to the subsequent change of leadership ("the spill changes nothing"). But overall I'm more confused now than when I started looking into it. —Pengo (talk) 15:55, 15 September 2015 (UTC)

source in Appendix:Proto-Slavic/sinjь

I find most interesting and credible the explanation for the meaning of Ἄξεινος as "Black sea" and I am quoting it on French wiktionnaire, Ἄξεινος, but it lacks a source... --Diligent (talk) 13:37, 18 September 2015 (UTC)

The source is Max Vasmer, but his explanation is possibly wrong. Read The Name of the Black Sea, an article by De Blois. --Vahag (talk) 17:13, 18 September 2015 (UTC)


A lot of craziness was going on in the etymology, so I notched it all down, but it would be nice to have something that could be referenced. —Μετάknowledgediscuss/deeds 17:46, 20 September 2015 (UTC)

Proto-Germanic entries of

This IP obviously knows a thing or two about the sound changes between Proto-Germanic and Old English, but a significant number of their Proto-Germanic entries have exactly one descendant, which doesn't exactly inspire confidence. I hope they're not going through an Old English word list and running the sound changes backwards to arrive at a "Proto-Germanic" reconstruction- but that's exactly what it looks like. Chuck Entz (talk) 21:37, 20 September 2015 (UTC)

Having only a single descendant doesn't automatically invalidate an entry — if there's a higher-level i.e. 'parent' i.e. PIE root with other descendants, then the Proto-Germanic step probably existed, unless loaning happened — but if there's no 'parent' root and only one descendant, it is suspicious, and it becomes downright implausible when there's ae process that would have created the form within the 'descendant' language. The scattered references I looked at analyse botettan as a compound formed in Old English, suggesting the Proto-Germanic page bōtatjaną should go. The situation is similar for Grīmaz (Köbler's etymological note on Grímr is to see gríma), and I can't find evidence of aikulaz either, nor ballukaz, baswaz, bōkōną, hampijaną or snigilaz (though I can find references for snagilaz), only some of which assert (without citing references) descent from PIE. þrawwaz is interesting (not necessarily invalid) in that it doesn't list regular descendants, only loans. - -sche (discuss) 17:44, 21 September 2015 (UTC)

Proto-Slavic *velťi or *velkti?

See Russian воло́чь (volóčʹ) (source: Vasmer) and влечь (vlečʹ), Polish wlec (source: Derksen). Pls fix appropriately. (I used *velkti from the Polish entry in влечь). --Anatoli T. (обсудить/вклад) 13:25, 21 September 2015 (UTC)

ť and kt are essentially equivalent in Proto-Slavic. We used to use kt, but a while ago we switched to using ť. --WikiTiki89 14:49, 21 September 2015 (UTC)


Discussion moved from WT:FB.

Where did you find fokka? I’m having trouble finding it anywhere. -- 05:30, 21 September 2015 (UTC)

I sometimes wonder about etymologies marked "dialectal Swedish" (it sometimes seems to be a way for the linguist to give an otherwise obscure Germanic term a root), but in this case it's in Merriam-Webster, and I did find a couple of books that make the same claim: 1, 2. I don't know enough Swedish to verify whether it existed (if a dialectal word for "fuck" would be recorded hundreds of years ago), but it seems plausible - the closely related language of Danish has the cognate fokken. Smurrayinchester (talk) 20:54, 22 September 2015 (UTC)
Verifiable from Svenskt dialektlexikon (1862-1867), p. 188 ("Bhl." = Bohuslän). --Tropylium (talk) 15:34, 23 September 2015 (UTC)
Good find! Smurrayinchester (talk) 17:25, 23 September 2015 (UTC)

I made the Swedish section for it recently. fokka#Swedish --Romanophile (talk) 11:51, 25 September 2015 (UTC)

"Bahusia"? Can we use common terms please? I've never heard of this one. —CodeCat 18:31, 28 September 2015 (UTC)
  • Indeed, that context label bordered on the useless. After poking around Wikipedia a bit, I think I've fixed the label. Or, I've completely screwed it up. Someone more familiar with Swedish should please have a look. ‑‑ Eiríkr Útlendi │Tala við mig 23:58, 1 October 2015 (UTC)

*vesti vs. *vezti

(Notifying CodeCat, Ivan Štambuk, Atitarev):

A while ago CodeCat decided that the infinitive of Proto-Slavic *vezǫ should be spelled *vesti rather than *vezti claiming in this discussion that "z automatically devoiced to s before t in Proto-Slavic". Only now I realize, how do we know that this is true and that devoicing didn't occur at later stages? Old East Slavic spelled this word as везти (vezti), unlike Old Church Slavonic, which spelled it вести (vesti) (at least based on our entries that was the case), indicating that it may have been pronounced with a [z], and Ukrainian still pronounces it with a [z]. So why are we so sure that it was devoiced in Proto-Slavic? --WikiTiki89 19:22, 24 September 2015 (UTC)

(Notifying CodeCat, Ivan Štambuk, Atitarev): Re-pinging people, since no one responded. --WikiTiki89 18:12, 28 September 2015 (UTC)
You're right. Vezti and vesti are two related but different verbs with different senses, inflections, descendants and pronunciations, including the infinitive. --Anatoli T. (обсудить/вклад) 21:53, 28 September 2015 (UTC)
That wasn't the point. The question was only about the infinitive and only about how it should be spelled, which for a reconstruction reflects how we think it would have been pronounced. For example, if we were reconstructing Russian, we would spell both infinitives as vesti, because вести and везти are pronounced exactly the same. --WikiTiki89 22:28, 28 September 2015 (UTC)
Then the modern Ukrainian pronunciation should be taken into account. Why the two inflection paradigms shouldn't matter? The two verbs should be split in any case. --Anatoli T. (обсудить/вклад) 23:36, 28 September 2015 (UTC)
The two verbs are split (see the entry), and if you read my original post, I already did mention Ukrainian. --WikiTiki89 02:47, 29 September 2015 (UTC)
  • You can't pronounce [zt] without some kind of (glottal) break, which either Ukrainian inserted (by analogy to present stem or under the influence of written forms) or the reconstruction of Proto-Slavic is wrong with respect to voiced-unvoiced sequences (which are unassimilated in uk in general). Another possibility is that uk (or perhaps OEsl.) recreated the form везти after that kind of assimilation ceased to be functional). Historical grammars of uk should be consulted if anyone has access to one. --Ivan Štambuk (talk) 08:53, 30 September 2015 (UTC)
    • No, it's phonetically quite pronouncable, especially if you admit a degree of variation in where the voicing switches exactly (ranging from [z͡st] to [zd͡t]). --Tropylium (talk) 09:50, 30 September 2015 (UTC)
    • I agree that it is very possible that OES used везти as an etymological spelling, rather than a phonetic one, and that Ukrainian may have extended this to the pronunciation as well. Strong evidence of voicing in PS would be if we found a word with root-internal [zt] in OES or Ukrainian. If no such word is found, we are left to speculate. Do we know what respective infinitives of *wedʰ- and *weǵʰ- would have looked like in PIE? --WikiTiki89 14:59, 30 September 2015 (UTC)
      • PIE possess no infinitive form. Many descendant, including PS, take their infinitives from the PIE nominal suffix *-tis. This is, however, misleading because the descendants of original *-tis-stems all have the stressed zero-grade root in all of their descendants (e.g. *méntis > *mń̥tis > PS *pamętь, *wéh₁itis > *uh₁ítis > PS *vitь). If we found the PS descendant of *wédʰ-tis (which does give PG *gawissiz, we would expect to see *wédʰ-tis > *wét-tis > *wétstis > *útstis > PS ?*ъsti. In this situation we would not only see a different root grade in the stem but also dental assimilation and assibilation in PIE. Similarly *wéǵʰ-tis should probably produce PS ?*ъzti, though I do not know whether the *z should assimilate in this case. As such, either form *vesti or *vezti should be a PS formation. —JohnC5 15:43, 30 September 2015 (UTC)
        • @JohnC5: The intent of my question was to determine whether voicing assimilation occurred before PS in each case. You seem to be saying that with *wédʰ-tis it occurred early on, before the frication of the *-dʰ-, while for *wéǵʰ-tis, it may be that no voicing assimilation occurred at all before PS. Is that correct? Also, why do you suggest *weC- > *uC- > *ъC-? There is no word-initial ъ in PS. --WikiTiki89 15:59, 30 September 2015 (UTC)
        • (edit conflict) There was voicing assimilation already in PIE. Anything spelled -ǵʰt- for morphological reasons would have been pronounced -ḱt- in PIE and would (barring analogical re-formation) have developed exactly like any other -ḱt-. If any Slavic language truly has phonetic [zt] in this (or any other) word, it can't be very old, but must be a relatively new case of analogy/paradigm leveling. —Aɴɢʀ (talk) 16:02, 30 September 2015 (UTC)
        • P.S.: word-initial ъ got prothetic v in PS, so *ъsti and *ъzti would be *vъsti and *vъzti respectively. —Aɴɢʀ (talk) 16:04, 30 September 2015 (UTC)
          • @Wikitiki89: Sorry for the confusion on my part. My guess of word-initial stemmed from a lack of knowledge of how word initial PIE *u descended into PS. My long-winded and tangential answer was to say that the full-grade stem form prevented either infinitive from being a direct PIE descendant regardless of PIE assimilation rules. Also, Aɴɢʀ is absolutely correct that *-ǵʰt- would become *-ḱt- in PIE. —JohnC5 16:15, 30 September 2015 (UTC)
            • There is no evidence that assimilation occurred in PIE already. In fact, w:Bartholomae's law requires its absence. —CodeCat 16:47, 30 September 2015 (UTC)
              • …yes. Regardless, the point about the root grade still stands. —JohnC5 16:57, 30 September 2015 (UTC)
                • But *vъsti / *vъzti > *vesti / *vezti through leveling is also possible. Anyway, the change *ved-ti > *vesti could not have occurred within PS. --WikiTiki89 17:30, 30 September 2015 (UTC)


Claims that it came from PIE, but it did not name a specific root. Hillcrest98 (talk) 21:52, 30 September 2015 (UTC)

According to zolfo (Italian), it's from *swelplos, from the root *swel- (to burn, smoulder). Sounds plausible (it would correlate with Proto-Germanic *sweblaz) - any PIE experts able to back that up? Smurrayinchester (talk) 13:39, 1 October 2015 (UTC)
According to Kroonen[1], we are dealing with a pre-IE Wanderwort reflected in Latin, Germanic, Armenian, Hebrew, Mongolian, Turkish. --Vahag (talk) 14:36, 1 October 2015 (UTC)
So De Vaan[2] would like to derive the original form sulpur from an unattested ablaut r/n-stem *sólp-r from *selp- (fat) and compares it to ὄλπη (ólpē). This root does have an s-stem *sélpos that gives AG ἔλπος (élpos, olive oil)/ἔλφος (élphos, butter)[3], Albanian gjalpë (butter)[4][5], and Tocharian A/B ṣälyp/ṣalype[6]. Also *solpéh₂ > PG *salbō, a secondary -ís formation > Sanskrit सर्पिस् (sarpís), and *sl̥prós > सृप्र (sṛprá).[7] Beekes, however, does not believe that ὄλπη and ἔλπος are related, though the LIN[8] derives ὄλπη from *solpéh₂. None of the sources discussing *selp- derived sulpur from it except De Vaan, though the semantics of fat, oil, ointment > sulfur do not seem plausible to me. With this information, I am more inclined to trust Kroonen's claim that it is a non-IE Wanderwort. —JohnC5 16:18, 1 October 2015 (UTC)
Let's try to map this out. So far, this is mostly guesswork, so feel free to add to it or modify it:
Still not sure where to put Slavic forms *sěra and OES цѣрь (cěrĭ) (and whether they belong here at all). --WikiTiki89 17:30, 1 October 2015 (UTC)


  1. ^ Kroonen, Guus (2013) Etymological Dictionary of Proto-Germanic (Leiden Indo-European Etymological Dictionary Series; 11), Leiden, Boston: Brill, page 497
  2. ^ De Vaan, Michiel (2008), “sulpur”, in Etymological Dictionary of Latin and the other Italic Languages (Leiden Indo-European Etymological Dictionary Series; 7), Leiden, Boston: Brill, page 598
  3. ^ Beekes, Robert S. P. (2010), “ἔλπος”, in Etymological Dictionary of Greek (Leiden Indo-European Etymological Dictionary Series; 10), volume I, with the assistance of Lucien van Beek, Leiden, Boston: Brill, →ISBN, pages 415-516
  4. ^ Demiraj, Bardhyl (1997), “gjalpë”, in Albanische Etymologien: Untersuchungen zum albanischen Erbwortschatz [Albanian Etymologies: Investigations into the Albanian Inherited Lexicon] (Leiden Studies in Indo-European; 7) (in German), Amsterdam, Atlanta: Rodopi, page 182
  5. ^ Orel, Vladimir (1998), “gjalpë”, in Albanian Etymological Dictionary, Leiden, Boston, Köln: Brill, page 129
  6. ^ “ṣalype-” in Douglas Q. Adams's “A Dictionary of Tocharian B.” Leiden Studies in Indo-European 10 (1999).
  7. ^ Kroonen, Guus (2013) Etymological Dictionary of Proto-Germanic (Leiden Indo-European Etymological Dictionary Series; 11), Leiden, Boston: Brill, page 424
  8. ^ Wodtko, Dagmar S.; Irslinger, Britta; Schneider, Carolin (2008) Nomina im indogermanischen Lexikon [Nouns in the Indo-European Lexicon] (in German), Heidelberg: Universitätsverlag Winter, pages 612-613