Wiktionary:Etymology scriptorium/2015/December

Proto Germanic heaven/steven.
I was under the impression, that /vən/ just is a northern dissimilation of West-Germanic /mn/, contrasting with an assimilation to /mm/ in the south, thus pointing to something like *stamnijo (listed) and *himnaz. We currently list *stebo and *himinaz. Having know knowledge of Proto-Germanic research: How certain are we of any of these forms? How likely is it that West Germanic had *himnaz and Proto Norse had *himinaz instead? Is Old Norse hifinn an expectable result of older *himinR? Korn &#91;kʰʊ̃ːæ̯̃n&#93; (talk) 13:19, 1 December 2015 (UTC)


 * is attractive because corresponds well to with an n/l alternation. With, since the m is attested in multiple branches, the b/v > m change seems less likely than m > b/v. But compare , for which the Swedish  went the other way (b/v > m). --WikiTiki89 14:46, 1 December 2015 (UTC)


 * There is no /mm/ in West Germanic, cf. Old High German, Old Saxon himil. The modern German -mm- is not an old geminate, but indicates the short vowel, unaffected by open-syllable lengthening as commonly before -mel, -mer, -men (cf. Hammer, Hammel < OHG hamar, hamal). I also remember having read that the paradigm had l/n-alternation. Kluge says that the /v/ is through dissimilation: "während ae. [Old English] heofon, as. [Old Saxon] heban das m dissimilatorisch zu v (stimmhafter bilabialer Reibelaut) weiterentwickelt haben." Kolmiel (talk) 15:28, 3 December 2015 (UTC)
 * You must have misread /mn/ (i.e. M and N) for /mm/. No one mentioned a geminate. --WikiTiki89 15:59, 3 December 2015 (UTC)
 * No, I haven't misread. It's in the very first sentence of this discussion: I was under the impression, that /vən/ just is a northern dissimilation of West-Germanic /mn/, contrasting with an assimilation to /mm/ in the south [...]. Kolmiel (talk) 16:07, 3 December 2015 (UTC)
 * Sorry, my bad. So how do you explain and OHG ? That couldn't have just been a short vowel. --WikiTiki89 16:13, 3 December 2015 (UTC)
 * No, no. I mean in the cognate of "heaven" there is no /mm/ in West Germanic. I forgot to specify that. -- I don't know if I understand Korn correctly, but I thought he would assimilate his Proto-Germanic *himnaz to *himmaz and then derive "Himmel" with a suffix. This is obviously wrong, but maybe he didn't mean that after all. Kolmiel (talk) 16:20, 3 December 2015 (UTC)
 * To be precise: I was wondering why *himinaz would lead to North Sea Germanic /hev-/ rather than /hemɪn/ and whether it might be reasonable to postulate *himnaz as a secondary form. For the other I was wondering where the reconstruction of *stebo is taken from. Korn &#91;kʰʊ̃ːæ̯̃n&#93; (talk) 23:23, 6 December 2015 (UTC)
 * Found the process for heaven after some reading: Old Saxon/English deleted mid vowels in open medial syllables leading to some North Sea Germanic *hemna > *hemn. So all I'm left questioning is *stebo. Korn &#91;kʰʊ̃ːæ̯̃n&#93; (talk) 20:08, 7 December 2015 (UTC)
 * do you mean ? Leasnam (talk) 22:34, 7 December 2015 (UTC)
 * what I find fascinating is that PIE *(h₁)emno- becomes PGmc, with OE and ON showing the variant with -mn: OE and ; ON  beside  (NNsk  vs. Bokml ; Swedish  vs. Danish )...this shows that the pairs -vn and -mn are somehow tied. The situation, though less pronounced, is also seen in words for "name" , where English still has the doublet  beside  Leasnam (talk) 22:45, 7 December 2015 (UTC)
 * Perhaps PG had both variants. Or maybe even a separate phoneme, such as /ṽ/ perhaps. --WikiTiki89 17:04, 8 December 2015 (UTC)
 * Yes, I think so (i.e. PG had both variants). But it still appears to be more frequent in North Germanic and Old English Leasnam (talk) 17:21, 8 December 2015 (UTC)
 * Surely not a separate phoneme, since it occurs only before /n/ and is in complementary distribution with both /b/ and /m/ there. But it is possible that the phonemic contrast between /b/ and /m/ was lost before /n/. —Aɴɢʀ (talk) 17:23, 8 December 2015 (UTC)
 * FWIW, Kroonen reconstructs *hemina ~ *hemna- and ascribes the variant form *hemila- in Continental West Germanic to either dissimilation (after Braune and others), or alternatively (after Wachter) lexical influence of *sōel- ~ *sunnōn- "sun". Moreover, he reconstructs *stimnō- (~ *stamnjō-) and *ebna- from older *-mn-, which implies an intra-Germanic (though rather irregular seeming) change -mn- > -bn-. He says that Gothic himins and Old Norse himinn (as well as Old Frisian himel, Old Saxon, Old Dutch and Old High German himil) point to *hemina-, while only Old English heofon and Old Saxon heƀan seem to imply *hemna-, and thinks that it is more attractive to explain this variation as resulting from a split of the two stems from a common paradigm nom. *hemō, gen. *hemnaz, dat. *hemeni rather than to explain it as a syncope of *hemina- to *hemna- common to Old English and (a part of?) Old Saxon (or perhaps – my own idea – as a loan from Old English into Old Saxon, besides inherited himil?). --Florian Blaschke (talk) 21:07, 8 January 2016 (UTC)

Tocharian wartto, wärt “forest”
Hello, do you think that the Tocharian words might be related to any of these: Hungarian, Mansi (from forest), Mordvin  or Germanic ? — Ivadon (talk) 10:50, 3 December 2015 (UTC)
 * Tocharian is more likely to have Turkic loanwords than Uralic ones, and I don't see how to accommodate the difference between Germanic l and Tocharian r, so neither hypothesis seems very likely. —Aɴɢʀ (talk) 11:38, 3 December 2015 (UTC)
 * Adams says on the matter:
 * “TchA wärt (pl. wärtant) and B wartto (wärttonta) reflect PTch *wärtto/wärttonta. The apparent double consonant causes some difficulty but probably we should connect this word with Old English and Sanskrit  (Lidén, 1916:139-140, and VW:562; P:1161-1162; MA:199). The semantic development might be something of the sort ‘enclosure’ > ‘sacred enclosure’ > ‘sacred grove’ > ‘forest’ (cf. Melchert, 1984:111).”
 * If that we choose to accept this, I'd have to suggest or, from  (> Sanskrit , Latin , AG ), and for the Tocharian forms we might have to assume some later *-to, *-tonta suffixation. — JohnC5 13:41, 3 December 2015 (UTC)
 * There is still the possibility of a borrowing from IE into Uralic. I think that l—r alternation, called Lambdacism, is quite common in many languages, as among Finno-Ugric, eg. between Finn. kolme, Hung. három and Khanty хәԓум “three”. — Ivadon (talk) 13:54, 3 December 2015 (UTC)
 * The l-r alternation is found in Indo-Iranian, and can be found spontaneously as assimilation or dissimilation where the two sounds are close to each other, or in metathesis (e.g. Latin > Spanish ; Latin  > Spanish ), but I don't think it's a regular feature of Tocharian historical phonology. Anyway, if we accept Adams's etymology, there are no more l-sounds around to worry about. —Aɴɢʀ (talk) 14:19, 3 December 2015 (UTC)
 * As for the Uralic forms, Mansi wōr, cognate to Fi., cannot come from anything with *-rt- (neither can the Mordvin form). There's a proposal that the first might be instead connected to PIE (as in ), but getting a Tocharian form with w- from that seems out of the question too. Hungarian -rd- meanwhile usually does indicate a loan, but even before we go into the geographic problens, that has a similar issue with w- : zero.
 * (For Finnish kolme '3' and similar "lambdacized" forms, perhaps the most plausible explaination I've seen proposed is analogy from neljä '4'.)
 * --Tropylium (talk) 00:07, 5 December 2015 (UTC)
 * Thank you for your suggestions on Tocharian, *wer- seems applicable.
 * The similarity of /vuori with Slavic gora/hora makes me think about being a variant of   (note that this root is not found in Greek). — Ivadon (talk) 14:07, 7 December 2015 (UTC)
 * I do not know to what extent this may have been proposed, but at least to me PU ~  looks more likely to be connected to PIE  than . The meaning of Greek  looks like a local innovation ('to raise' > 'elevation' > 'mountain'). --Tropylium (talk) 19:49, 7 December 2015 (UTC)

和尚
How does 和尚 relate to the Sanskrit? I don't see a semantic or phonetic connection. — justin(r)leung { (t...) 17:17, 3 December 2015 (UTC)
 * See edit history. The user who added this (Huhu9001) is not trustworthy, so... Feel free to remove. --Anatoli T. (обсудить/вклад) 03:50, 4 December 2015 (UTC)
 * Thanks. It's removed. — justin(r)leung { (t...) 03:55, 4 December 2015 (UTC)
 * Japanese sources are all consistent in pointing to a Sanskrit origin, with most clarifying that the Chinese term was a phonetic transliteration of a Sanskrit term. My dead-tree resources, the Shogakukan Kokugo Dai Jiten and the Shinmeikai dictionaries, agree with online resources such as this entry in Daijirin or these entries from the Japanese edition of the Britannica International, the Daijisen dictionary, the Nipponica encyclopedia, the Sekai Dai Hyakka Jiten encyclopedia, etc.  Where the ultimate Sanskrit etymon is specified, it is consistently given as upādhyāya, with several sources noting that the Chinese term's immediate etymon was some slang or dialectal alteration from the Sanskrit.  Daijisen gives that form as khosha, while the Sekai Dai Hyakka Jiten points to related term pwajjhaw that seems to be Tocharian.  Even the JA WP article at ja:w:和尚 lists the Sanskrit term upādhyāya as the etymon, but the article lists no sources for this.  ‑‑ Eiríkr Útlendi │Tala við mig 19:11, 4 December 2015 (UTC)
 * Turns out the ZH WP article at zh:w:和尚 also lists the Sanskrit term, as well as the Pali form. Searching in Chinese for  uncovered this academic paper (PDF), also pointing to Sanskrit.  ‑‑ Eiríkr Útlendi │Tala við mig 19:15, 4 December 2015 (UTC)
 * Thanks for finding these resources. I can see the connection now, but I'm just not sure what caused the great difference in pronunciation. The academic paper does point to Sanskrit as the "ultimate" etymon, but it does also say that it comes from a "vulgar" form. Also, while the academic paper dismisses the possibility of influence from the Western Regions (西域), some Chinese sources like Hanyu Da Cidian (mentioned in the paper) and the Revised Chinese Dictionary use something like the Western Regions to explain the etymology. — justin(r)leung { (t...) 20:28, 4 December 2015 (UTC)
 * If the Western Regions / possible Tocharian term pwajjhaw represents a shift from Sanskrit root upādhyāya, it's possible too that some other prakrit might have made similar phonetic shifts, which would make the Western Regions not a necessary step. However, from what I recall of the transmission of Buddhism to China, the Western Regions were the geographic route.
 * A shift from Sanskrit upādhyāya to Pali upajjhāya appears to be documented; from there to pwajjhaw is conceivable. Going from pwajjhaw then to MZH huadzyang is also conceivable -- pwa- to hua only needs the P to soften, and -jjhaw to -dzyang only needs the final consonant to shift, and I note that MZH terms ending in  pretty uniformly end in  when imported into Japanese, suggesting that the two sounds were close enough that a jump the other way in a potential ancient borrowing into Chinese might be within the bounds of reasonable.
 * But for my part, this is almost pure conjecture, as I know very little about Chinese diachronic phonetic shift patterns. ‑‑ Eiríkr Útlendi │Tala við mig 20:43, 4 December 2015 (UTC)
 * I added a bit to the etymology section of the entry. Wyang (talk) 09:50, 16 October 2016 (UTC)

Long vowel of ὤψ
This Ancient Greek word has a long ō, but it is said to originate from PIE (or its o-grade) which has no sequence that I recognise as giving a long vowel. What I wonder is where the long vowel came from in Greek. Is it a regular outcome of word-initial or ? Are there parallels? —CodeCat 01:59, 4 December 2015 (UTC)
 * Beekes proposes a lengthened ō-grade, and the LIN goes on to give the acrostatic nom.sg., gen.sg. , acc.sg. . This does not conform any of to the normal declensional lengthening patterns with which I am familiar. — JohnC5 10:12, 4 December 2015 (UTC)
 * Sihler (p. 120) also proposes a lengthened grade in the nominative singular which was then generalized to the other cases, then goes on to say "a secondary full grade *e/ o H₃kʷ- is thinkable". —Aɴɢʀ (talk) 10:17, 4 December 2015 (UTC)

meaning of Proto-Germanic *fatōną
According to fassen, the meaning of Proto-Germanic *fatōną is “to fetch, hold”, whereas its own article says it's “to walk, stumble, fall”. --Espoo (talk) 06:32, 4 December 2015 (UTC)
 * Look a little further down, it's to "hold, seize". I think you're looking at the Etymology and meaning of PIE *ped-. Leasnam (talk) 15:49, 4 December 2015 (UTC)
 * That semantic development from PIE to Proto-Germanic is rather odd. Do you have any idea what researchers propose for how this occurred?  ‑‑ Eiríkr Útlendi │Tala við mig 19:04, 4 December 2015 (UTC)
 * If its connected, its believed to have developed from "find a way to; reach; attain" Leasnam (talk) 23:39, 4 December 2015 (UTC)
 * Thanks, i must have been very tired... The change in meaning from PIE is fascinating. What do u mean with "if its connected"?
 * Some sources do not connect it with PIE *ped- . They say its origin is unknown. Leasnam (talk) 01:55, 7 December 2015 (UTC)
 * AHD offers no explanation despite giving details on related words. The etymology we provide for fet talks of conflation with Proto-Germanic fetaną. --Espoo (talk) 17:25, 6 December 2015 (UTC)
 * Kroonen claims that it comes from the iterative-intensive *pod-e-, ostensibly meaning “step repeatedly” > “walk to” > “fetch”. Related to OCS . — JohnC5 20:51, 7 December 2015 (UTC)

Appendix:Proto-Germanic/newun
Do the words with /ɣ/ come from *newun, and if yes, what is the cause of the change? (The alternative to that being a PGM variant *negun.) Korn &#91;kʰʊ̃ːæ̯̃n&#93; (talk) 19:52, 7 December 2015 (UTC)
 * The languages where g occurs are known to have sporadic changes of w to g in a few other words. One notable example is Low German (or Frisian?) from . —CodeCat 19:57, 7 December 2015 (UTC)
 * Though that could be the -j- which became -g-, couldn't it? Via Old Saxon *ōja? (Seemingly ōia is attested, though the source I'm currently looking at doesn't give it.) Presupposing a "zachte g" in Old Saxon/early MLG, the development would be logical. Kolmiel (talk) 22:42, 8 December 2015 (UTC)
 * The two independent processes [j] > [ʝ] and [j] > [ɣ] are a regular feature of multiple Low German variants.As is [w] > [ɣ], but only in Modern Low German times. Korn &#91;kʰʊ̃ːæ̯̃n&#93; (talk) 01:02, 9 December 2015 (UTC)
 * Yeah, /j/ > /g/ (and vice versa) is certainly common, even in High German. But why are [j] > [ʝ] and [j] > [ɣ] "independent"? Aren't [ʝ] and [ɣ] allophones for /g/? --- I would have thought that Old Saxon nigun < *nijun < *niun, but Pfeiffer also says that the /g/ is from /w/. So it should be true. Kolmiel (talk) 18:21, 9 December 2015 (UTC)
 * The difference is that [j] > [ʝ] is a narrowing/fricativisation and proceeds in some areas as far as [dʒ]. On the other hand [j] > [ɣ] is a velarisation. The resulting [ɣ] stays (palato-)velar even in palatal environments, e.g. [snɪɣɪn] (Westphalia) and [snij.ŋ] (Pomorania, phonologically /sniːɣən/), both from MLG /snijjən/. Modern Low German [w] > [ɣ] is simply an unrounding of all Middle Low German non-initial [w], but I haven't heard of a similar development (Verner alternation aside) in Old Saxon. Korn &#91;kʰʊ̃ːæ̯̃n&#93; (talk) 20:03, 9 December 2015 (UTC)
 * Had a look at the recorded forms of *awjô. Middle Low German knows *aw.jô > /ɔːjə/ and for some reason *a.wjo > /ɔwwə/. The form [oːɣ] would most likely be a regional realisation of /ɔːjə/ and perfectly normal as that. So this particular example can't be taken as a reference for spontaneous Old Saxon switching. And for the record: I know not too much about Old Saxon and even less about other Old-languages, I'm just looking to understand how this happened since I find it peculiar that all WG languages but one would have a random change in this area. Korn &#91;kʰʊ̃ːæ̯̃n&#93; (talk) 01:01, 10 December 2015 (UTC)
 * Okay, just to clarify: In the above examples you refer to different dialects with different realizations (Westphalian and Pomeranian). But there's no phonemic distinction between [ʝ] and [ɣ], is there? Kolmiel (talk) 14:02, 10 December 2015 (UTC)
 * There can be. The sound [ʝ] is a possible dialectal realisation of either of the two phonemes /j/ and /ɣ/. That is: Some dialects have it as an allophone of one or the other but never both. So if you have a dialect with /j/ = [ʝ], then you could say that /ʝ/ is phonemically distinct from /ɣ/ for that dialect. Korn &#91;kʰʊ̃ːæ̯̃n&#93; (talk) 20:15, 10 December 2015 (UTC)
 * Okay. Thanks! Kolmiel (talk) 23:57, 10 December 2015 (UTC)

Guyman
Common term in advance fee fraud and, as a result, scambaiting.

This is the term used by advance fee fraudsters to refer to themselves. The word "guy" is used to refer to themselves too, though with notably less common usage.

http://nigerianspam.com/terms-419-scammers.htm http://www.sid.in-berlin.de/nedkelly-world/fraud%20prevention.html http://www.delphifaq.com/faq/russian_marriage_scams/f1147_129.htm http://lifesmarts.org/lifesmarts-u/consumer-rights-responsibilities/consumer-glossary/

In addition, the term is used in a BBC (I think) documentary on scams. The dialogue between a scambaiter and the scammer includes an accusation by the scambaiter to the scammer that he is a guyman, which the scammer vehemently denies.

The term "joke(r)man" is used to refer to scambaiters, although the term seems to be a fairly recent coinage.

In any case, what I am trying to figure out is if "guyman" is derived from what it looks like it is derived from, or if it is derived from some African terms.

In addition, I am wondering if the BBC documentary would be considered a legitimate source for the term if it were to ever be included here. 68.0.247.51 18:23, 8 December 2015 (UTC)

Yet another PGM question (neuter of *ainaz)
Me again. Old Norse eitt, German einaz, Gothic ainata, southern Low German enet all point to a PGM *ainat-. What's up with that? (I.e. what is its origin that we do not list it as a separate reconstruction?) Korn &#91;kʰʊ̃ːæ̯̃n&#93; (talk) 15:45, 9 December 2015 (UTC)
 * All of these forms are analogical, based on the ending of the pronominal ending (see ). There is no way to determine if these analogical forms already existed already in Proto-Germanic, they could easily all have formed in the individual branches. —CodeCat 15:48, 9 December 2015 (UTC)

coach
Is the sense "trainer, instructor" really derived from the vehicle? And if so, can we add some information about how that came to be? Kolmiel (talk) 18:03, 10 December 2015 (UTC)
 * from OnlineEtymDict: Meaning "instructor/trainer" is c. 1830 Oxford University slang for a tutor who "carries" a student through an exam; athletic sense is 1861., so apparently so Leasnam (talk) 18:05, 10 December 2015 (UTC)
 * Thanks. Kolmiel (talk) 18:47, 10 December 2015 (UTC)

mix - which hypothesis is better?
Century Dictionary claims that "mix" was derived from metathesis from Germanic, and a Germanic theory is currently the chosen etymology on WT.

Other sources claim that it came from Latin.

Which path is more likely? Hillcrest98 (talk) 21:06, 16 December 2015 (UTC)
 * Hmm, good question. The Online Etymology Dictionary also gives the "back-formation from mixt from Latin " etymology, but it seems to me that if Old English is really attested (and it's not in Bosworth-Toller, only  is), then that seems more likely to be the source. I'd always pick an attested Old English verb over a back-formation. Interestingly, Etymonline also says that OE  is itself a loanword from Latin, but we say it's from Proto-Germanic and is merely cognate with, not descended from, the Latin. Sometimes I wish DNA testing on words was possible. —Aɴɢʀ (talk) 22:00, 16 December 2015 (UTC)
 * Century argued metathesis based on the whole "aks" debacle and drawing analogy. Yeah, I wish words had DNA so we can definitely trace sources (look at the whole mess that is the etymologies of iron) :] Hillcrest98 (talk) 00:40, 17 December 2015 (UTC)
 * At one time, it was generally believed that all "mix"-sounding words (German mischen, etc.) were borrowings from Latin ; but recently, the near universal presence of this word has led some to re-think that assumption: it's in nearly all branches of IE, and the related form in OE (i.e., ) supports the probability that both are inherited ablaut forms of the same PIE root. Leasnam (talk) 16:56, 17 December 2015 (UTC)
 * I think it's still the majority opinion that the Germanic words are from Latin. The Germanic verbs for "to mix" seems to have been those found in German / Dutch mengen, and English blend. "Mischen" is found chiefly in western and southern dialects of German, which makes borrowing likely. Well, be that as it may. The question concerning "mix" seems to be unrelated to this: namely is it from OE miscian by metathesis, or is it directly from Latin mixtus. I think both possibilities should be mentioned, and also that they don't exclude each other: the two words have plausibly influenced and reinforced each other. Kolmiel (talk) 01:20, 18 December 2015 (UTC)
 * It's always likely too that there were multiple words for "mix" in Germanic, not just one (btw, English also has, , & :), and I don't disagree that English  (< Latin ) may have helped in reinforcing it, sure. Leasnam (talk) 01:31, 18 December 2015 (UTC)


 * I know it's a little late, but I have created if that is of any use. — JohnC5 04:26, 28 December 2015 (UTC)
 * Thanks John, there are a lot of must-need PIE roots which need entries. Hillcrest98 (talk) 05:11, 28 December 2015 (UTC)
 * Yes, Thank you ! Leasnam (talk) 16:04, 28 December 2015 (UTC)


 * In a tangential request mostly directed at, could we figure out what is going on with Celtic *miskos etc.? Matasović and my Proto-Celtic word list both have approximately *mi(x)sko- but then Matasović says OIr. and MW.  are “deadjectival, built to the stem *mesko-. This has been dissimilated from *mik-sko- (cf. the same dissimilation in *farsk-o- 'ask' from *fark-sk-o-).” Which, if any, of the Celtic descendants reflect *miskos and which reflect *mesko- or even *mēsko- (because PIE *ey > PC *ē before a palatal)? — JohnC5 21:00, 28 December 2015 (UTC)
 * I wonder if "*mesko-" is a misprint in Matasović; it must be *miskos. The vowel i lowered to e before ā in the following syllable so mescaid can come from *miskāti with no problem, while mysgu comes from some form with a different vowel in the second syllable. A form *mēskos would have become *míasc in Old Irish and *mwysg in Welsh, and the vocalism of *meskos would be unexplainable (and probably couldn't yield Welsh mysgu either). —Aɴɢʀ (talk) 21:42, 28 December 2015 (UTC)
 * So, we are good on Celtic? Is there any resource you would recommend for PC to its descendants? I have often been mystified certain changes in Celtic vowels, particularly in cases like . — JohnC5 22:31, 28 December 2015 (UTC)

Appendix:Proto-Germanic/bi
The page for lists it as coming from PIE  (which does not list  as a descendant), while the English entry for  claims it's shortened from PGmc. I am not educated enough to have my own opinion, but I hope someone more informed can bring more consistency to these entries. Eishiya (talk) 21:40, 17 December 2015 (UTC)
 * Anglom added that etymology with reference to Kroonen (2013), but I cannot find the relevant entry in that book; where is the origin of *bi mentioned? (I tried to tag the footnote, but we don't seem to have a counterpart to w:Template:Failed verification ...) --Florian Blaschke (talk) 22:12, 8 January 2016 (UTC)
 * Page xxix in the Introduction. Anglom (talk) 23:28, 8 January 2016 (UTC)
 * That's not where anyone would look, would they? A reason to add page numbers to citations. --Florian Blaschke (talk) 23:37, 8 January 2016 (UTC)
 * Hindsight is 20/20. :) Anglom (talk) 01:34, 9 January 2016 (UTC)

rifle
The entry says that the underlying Old Dutch word meaning “to scrape” is a frequentative of PG. Instead it seems that Dutch, German (< Low German) are cognate with Old Norse , Old English , which would be a completely different root, if I understand correctly. Kolmiel (talk) 04:20, 18 December 2015 (UTC)
 * Explaining more: Dutch rijffelen in Kiliaan ("to scrape; also: to roll the dice") is iterative from obsolete or dialectal rijven ("to rake"), rijf ("rake"), see . The latter is connected, seemingly correctly, with the mentioned Old Norse and Old English words with hr-. -- In some sources, the Dutch word is also compared to repel ("ripple"), itself related with Germanic . The question is whether these stems with original hr- and r- can possibly have any connection at all, and if so: how close actually? Kolmiel (talk) 14:12, 18 December 2015 (UTC)
 * What of any connection with English, , with meanings related to to pass one thing through another? ‑‑ Eiríkr Útlendi │Tala við mig 16:14, 18 December 2015 (UTC)
 * Apparently, German riffeln merges the senses of both etyma. --Florian Blaschke (talk) 23:40, 8 January 2016 (UTC)

早月
In our current etymology for 早月, namely
 * From 早 (さ, sa), "early" + 苗 (なえ, nae, "sprouts") + 月 (つき, tsuki, "month") ,

the middle component seems spurious: there is no /nae/ (and no ‹苗›) in the word. What was meant here? Was there an older longer form of the word including /nae/, or is this solely a mis-presented attempt to explain the semantics? 4pq1injbok (talk) 00:47, 19 December 2015 (UTC)
 * According to Gogen Yurai Jiten, it's shortened from 早苗月, possibly influenced by the alternate name 皐月 (which also starts with "sa"). The etymology we have is definitely unclear. Eishiya (talk) 04:22, 19 December 2015 (UTC)


 * I  and pointed to lemma, and then I  at .  As pertains to the question from 4pq1injbok, it appears that the sa in both satsuki and in sanae is the same element, and it seems to be an ancient word referring to the rice plant, or possibly to the god of rice plants.  Per Shogakukan's Kokugo Dai Jiten, Revised Edition.  ‑‑ Eiríkr Útlendi │Tala við mig 04:54, 29 December 2015 (UTC)
 * Should the plant really be listed at ? The kanji for the plant is as far as I can see, and  is understood only as referring to the month. The two are not perfectly synonymous, they overlap in the month senses but not the plant sense. I also feel that discussion of the spellings with 早 belong in those entries, not in . They're linked from it already anyway. In summary: 五月: month senses (current); 皐月: plant (current) and month (archaic/poetic) senses; 早月: month senses (archaic/poetic). Eishiya (talk) 18:29, 5 January 2016 (UTC)

vëlla
RFV of the etymology.

This came to my attention because User:Embryomystic left it with a module error. I started to fix it up, but I wasn't sure about all of the language codes. Albanian etymologies tend to suffer from amateur guesswork and/or POV-pushers, anyway, so I thought it would be worth having someone knowledgeable take a look at it. Chuck Entz (talk) 07:04, 20 December 2015 (UTC)


 * So Demiraj and Orel support the analysis +, with Demiraj adding PAlb. *vai-láu(d) < PIE *swoi-h₁lowdʰi-. — JohnC5 07:44, 24 December 2015 (UTC)

Latin squalus and Haida sqʼahla
I cannot leave this thought unexpressed — this must be pure coincidence! Or was there a trading relationship between Europeans and the Haida sometime? — Ivadon (talk) 05:46, 22 December 2015 (UTC)

It may be left unanswered for now since there have been no proposals of trans-Pacific relationships other than Yupik and Dené-Yeniseian, nor any evidence of recent pre-Columbian contact in the North Pacific. — Ivadon (talk) 06:02, 22 December 2015 (UTC)


 * It’s only a coincidence. I believe basic Haida nouns, few as they are, are mostly one syllable. Words with more than one syllable are usually made up of individual morphemes. In the case of sḵ'aahla, it is composed of sḵ'aa + hla, with a literal meaning of "bottom fish", which means any fish that does not swim upstream. Also, Haida phonology is very different from that of Latin, and there is little similarity between Latin QU and Haida Ḵ', or Latin L and Haida HL. —Stephen (Talk) 07:16, 22 December 2015 (UTC)
 * (edit conflict) It seems odd that a Latin word for shark would make its way across oceans and continents to be borrowed without displacing any of the Haida words for shark, and without leaving any other traces in the languages in between. Not only that, the Latin qu is a labiovelar, but the Haida qʼ is an ejective, which is a type of sound not found in any European language. Given the huge number of words for fish in all the languages of the world(Haida alone has dozens), the likelihood of at least one pair of languages having words that are superficially similar is really quite high. Chuck Entz (talk) 07:34, 22 December 2015 (UTC)
 * Exactly. When we look at the number of languages in the world, and the number of words in every language, and the number of sounds: it would be odd if there were no false cognates. Coming to my mind right away: German Hilfe ("help") and Arabic . Or English bad and (provenly unrelated) Persian . None of this is mentionable, because there is nothing odd or interesting about it. It's just an expectable coincidence. Kolmiel (talk) 13:45, 22 December 2015 (UTC)

Etymologies of common English words for animals
Is there a reason that so many common animals have unknown etymologies (or not past Proto-Germanic)? shark, dog, pig, rooster, sheep, horse, lamb, rabbit (some of these are said to be unknown in the Online Etymology Dictionary, but we give purported origins) DTLHS (talk) 08:18, 24 December 2015 (UTC)


 * There are a lot of Germanic words that do not seem to have cognates in any other Indo-European languages. One possible explanation for this is that migrating speakers of Proto-Germanic encountered a non-Indo-European people living in the area of Northeast Germany/Denmark/Southern Sweden, and that relations between the Germanic tribes and this unknown seafaring people were friendly, so that many of the non-Indo-European words were adopted into Proto-Germanic, including the unusual cadence and phonology that sets the Germanic languages apart from the other European languages. Under this theory, it is thought that the Germanic tribe and the non-Indo-European people eventually melted into one homogeneous group, leaving no record of the unknown people except for its imprint on the Germanic languages. See Germanic substrate hypothesis. —Stephen (Talk) 09:34, 24 December 2015 (UTC)
 * However, these terms entered Germanic or English at different times, so it's hard to generalize. The words 'dog', 'pig', and 'sheep' all replaced words with good IE pedigrees in their primary meaning, but the old words were retained with semantic narrowing as, , and , but at different times ('ewe' was replaced by 'sheep' as the general term already in Proto-Germanic; the replacement of 'hound' and 'swine' didn't happen till Middle or Early Modern English). 'Shark' didn't appear until the 16th century (there are sharks in the waters around Great Britain; presumably there was simply no cover term for them other than fish). Roosters and rabbits weren't known to Germanic peoples until contact with the Romans, so it's not surprising there's no ancient word for them. I don't know why etymonline is skeptical of the connection of 'horse' with ; it seems quite plausible to me. That basically leaves as the only word from the list where a PIE word  was replaced in Germanic by a word of unknown etymology. (I wonder what a PG  would have looked like in Modern English; "acquen" or "acken" or something?) As Stephen points out, words of non-IE original aren't that uncommon in Germanic, but I don't think there's anything specifically about animal names that lends them to replacement. —Aɴɢʀ (talk) 14:42, 24 December 2015 (UTC)
 * I think we used to call sharks wolf-fish (cf. dogfish) before "shark". PGmc had rooster ( and ). Also rabbit . *h₂egʷnos would become something like, which suvives in the English word Leasnam (talk) 15:04, 24 December 2015 (UTC)
 * I'm not sure about that. PIE *gʷʰ would indeed surface in Germanic as *u, but I don't know about *gʷ. Under Grimm's law it would simply devoice to *kʷ and then delabialise before a consonant. Another possibility is Kluge's law, which would give . —CodeCat 16:04, 24 December 2015 (UTC)
 * , does that mean you're doubting the accuracy of the etymology of and/or the reconstruction ?, you're right about 'rooster', but  is really 'hare'. Until they got domesticated, the range of the European rabbit was far away from Germanic speakers. Our entry  does not really make it clear that rabbits are generally distinguished from hares, though both are leporids. —Aɴɢʀ (talk) 16:42, 24 December 2015 (UTC)
 * I think part of the answer is that common domesticated animals develop lots of specialized terminology for specific types, and the older terms tend to get lost in the shuffle: as Angr pointed out, we have many older terms like hound, ewe, mare, sow and farrow that became narrower, more specialized terms. Also, some names were lost because of taboos: bear probably replaced the PIE term because of fear-based avoidance, and rooster seems to have replaced cock to avoid using a word that also meant penis. I wonder if the replacement of cony had anything to do with similarity to a word for vagina? As for PIE, I suspect that Old English eoh would have lost all its consonants by the time it got to modern English and have become too phonetically indistinct from other words. I'm not sure there are any words for marine animals that go back to PIE (whale and Latin squalus come for a term for a type of catfish), but pretty much all of the neighboring Germanic languages seem to have used some variant of hai. For some reason English didn't. Chuck Entz (talk) 01:03, 25 December 2015 (UTC)
 * Add to the list for . Do other branches of IE fare better when comparing retention of original vocabulary, compared to Germanic and English or are they pretty much all the same ? Leasnam (talk) 04:19, 25 December 2015 (UTC)
 * And don't forget chickadee for tit. Etymonline seems to agree with my hunch about cony and says the change happened in England, but most of the others are due to American puritanical prudishness. As for comparisons, I think English is unique in the sheer volume of new words because of its history and its geographical spread. Other Germanic languages have much more of the older vocabulary. There are some branches such as Armenian that have been under intense pressure from other languages, and the languages of India are surrounded by a vastly different flora and fauna from the rest of Indo-European, but I don't know relative percentages between branches. Chuck Entz (talk) 06:14, 25 December 2015 (UTC)
 * Do we know that chickadee was an intentional supplantation of tit, let alone that it was a supplantation motivated by prudery?131.92.84.38
 * I'm not sure. I do remember reading that several of the new terms for animals that arose in the US were, in fact, meant to substitute for homonyms with taboo words, but I don't remember if chickadee was specifically mentioned as one of them. Given this pattern of taboo-related substitutions, the fact that tit is definitely considered vulgar in the US (it's one of George Carlin's Seven Words You Can Never Say on Television), and that the US is also the only place where the word chickadee is used for what is called a tit everywhere else, it seems quite plausible to include it in the list. Yes, it's definitely onomatopoetic, but I doubt anyone would have bothered to coin the new word unless there was a reason not to use the old one. Aside from the taboo-avoidance cases, most of the difference between US and English animal vocabulary is application of European words to different species: the elk in the US is very similar to the European red deer, while the closest relative of the European elk is called a moose. The US robin is basically a European blackbird with different coloration, and the US blackbird is in a completely different family. Chuck Entz (talk) 03:09, 13 January 2016 (UTC)
 * By the way, I wonder about the etymology at cony: it was added by Torvalu4, who tends toward dogmatic certainty in the murkiest of etymologies. Etymonline says something vague about it possibly coming from Celtic Iberian, but our etymology gives a specific Proto-Basque form as the origin, which doesn't look all that convincing. Chuck Entz (talk) 06:27, 25 December 2015 (UTC)
 * Perhaps cony was displaced in favour of more specific terms because it was so general in what it referred to, being used for leporids, but also various other animals (like the hyrax) which are apparently not closely related to them. Leasnam (talk) 18:39, 25 December 2015 (UTC)


 * Slightly off-topic, does the "Someone who exploits others" sense of shark belong to etymology 2 rather than etymology 1? - -sche (discuss) 19:03, 29 December 2015 (UTC)
 * Not sure, it may be a parallel development from the predatory fish sense, it's hard to say. The cite also uses wolves and lambs as metaphors Leasnam (talk) 19:09, 29 December 2015 (UTC)
 * Probably, but there's a lot of overlap between the two senses because sharks² are perceived as being just as predatory as sharks¹. I never knew there were two etymologies until just now; I always assumed the senses listed under etymology 2 were figurative extensions of etymology 1, and I bet I'm not the only one. —Aɴɢʀ (talk) 19:23, 29 December 2015 (UTC)

sampan
The Wiktionary entry says that it's from Mandarin with a note that requests Cantonese romanization; Wikipedia says it's from Hokkien and references Merriam-Webster which says that it's from Cantonese. Which one? —suzukaze (t・c) 04:34, 26 December 2015 (UTC)
 * It's probably not Mandarin, but Cantonese or Hokkien both seem possible. Merriam-Webster, Oxford and OED all say it's from something that literally means "three boards", so that might be 三板 instead of 舢舨. 三板 in Cantonese is saam1 baan2, and 舢舨 in Hokkien is sam-pán. Both would work as the source, but with the meaning given by other dictionaries, Cantonese seems to be the more likely candidate. — justin(r)leung { (t...) 08:36, 26 December 2015 (UTC)
 * Is it safe to say "From either Cantonese 三板 (saam1 baan2) or Min Nan 舢舨 (sam-pán)" in the etymology? (Something somewhat ambiguous, like bau.) —suzukaze (t・c) 09:02, 26 December 2015 (UTC)

Subarian
RFV of the etymology.

This entry was obviously created to host the etymology, and layers of redundant material are still being added- to the point that the bloated etymology and the lengthy list of references are dwarfing the entry itself. It's not hard to see what's going on: someone really, really wants us to believe that Turks were around the ancient Middle East as early as the third millennium BCE, and they keep adding testimonials from various scholars or near-scholars along with any references they can to fortify that position. There are a couple of lines with alternative theories, but they appear to be window dressing.

Can someone with knowledge and/or references sort through this, fact-check it, and make an actual dictionary entry out of it? Chuck Entz (talk) 03:34, 28 December 2015 (UTC)
 * I've cut it down to a simple dictionary entry on the basis of the Wikipedia article, which, to judge from its talk page, has also been subject to the same nationlistic nonsense. It's easier to eliminate here because it simply isn't dictionary material. —Aɴɢʀ (talk) 20:40, 28 December 2015 (UTC)
 * Looks good. Thanks! Chuck Entz (talk) 21:26, 3 January 2016 (UTC)

נאַר
The etymology for 🇨🇬, added by yours truly in 2011, says the word's from 🇨🇬. (Metaknowledge added that that means "bray", but it actually means "young person".) However, I don't now remember my source for giving that as the etymology. Note the existence of 🇨🇬 (where our Descendents section lists ). Both etymologies sound right (note that is pronounced exactly or almost exactly like Ashkenazic ), but which is right? Or was it influenced by both words? &#x200b;—msh210℠ (talk) 22:58, 29 December 2015 (UTC)
 * The Hebrew verb does mean "to bray". This dictionary lists the noun  under the verb, from which I guess it may have been derived. I find the Germanic etymology much more convincing for two reaons: (1) why would the spelling have been changed from נער to נאַר? And (2) most words that are derived from Hebrew verbs are derived from the active participle, i.e. present tense, so it would have come from  and would have been pronounced "noyer" (there are exceptions to this, however, such as ). --WikiTiki89 23:08, 29 December 2015 (UTC)
 * Re "bray": The verb means "bray" and doesn't share a pronunciation (or approximate meaning) with . The noun  means "young person" and does share a pronunciation (and approximate meaning) with . That also removes your reason #2. As to your reason #1, note that Hebrew loanwords change spelling e.g. in ; and there's especial reason to change the spelling here, because נער would be mispronounced ner by readers. &#x200b;—msh210℠ (talk) 23:30, 29 December 2015 (UTC)
 * They change spelling only in occasional cases where the pronunciation is too different from the original Hebrew word. Confusion has nothing to do with it, see for example. I guess you're saying that it comes from the noun, so I no longer have a problem with the form of the word, but I still have a problem with the spelling and also with the fact that the meaning of  fits much better. --WikiTiki89 00:33, 30 December 2015 (UTC)


 * I don't remember this. Anyhow, my thoughts on the matter: Wexler thinks it comes from an Iranian source; that seems to flow from his biased and ridiculous ideas, so I'll ignore that. The Hebrew origin is supported only by the plural form, which takes a characteristic Hebrew morph. However, that plural doesn't match the plural of the etymon that msh210 proposes. There must be Hebrew influence, perhaps of that very word, but I can't see that explanation as making sense when we have narro attested, and regular sound changes bring us to nar, with pretty much the same semantics. I'm not nearly as good at Yiddish etymology as my fellow Yiddish editors are, but I think we should mention both possibilities. —Μετάknowledge discuss/deeds 02:37, 30 December 2015 (UTC)
 * As for the plural, I've been looking for other similar cases. is a Hebraic term, but has a similar meaning and it has two plurals, which is the expected Hebrew plural, and , which is really the plural of Hebrew ; it's possible that the similar meaning of  influenced the plural . , which is not in the least a Hebraic term, has two unexplicably Hebraic plurals  and . The opposite also happens with Hebraic  and  having Germanic plurals  (note the umlaut!) and . Therefore, I don't think the Hebraic plural  bears any weight as to the derivation of . --WikiTiki89 16:58, 30 December 2015 (UTC)
 * Wow, nice finds! (The only one of those words with faux-Hebrew plurals I've used was tayvl, and it's been (unsurprisingly) reinterpreted as a diminutive.) I think that's good support for something systematic going on, although I've no clue what it is. If you can figure it out, we may want to create to host these. —Μετάknowledge discuss/deeds 17:22, 30 December 2015 (UTC)

cork
Can someone check whether cork (Etymology 1) is thought to have possibly ultimately been derived from Aramaic and if so, from what word in Aramaic? The page says "or from Aramaic" but doesn't elaborate; it's been that way for several years, and I can;t find anything else online that indicates it might have been. Esszet (talk) 02:02, 31 December 2015 (UTC)
 * , do you know where you got that info from? (You added it to the etymology.) &#x200b;—msh210℠ (talk) 17:01, 31 December 2015 (UTC)
 * From the Oxford English Dictionary, I think. The link I added was to an online copy of the OED. If you have access to a printed copy, I'd recommend checking that one. Or, if you have an account, or know anybody with an account, you can check the online version. --Florian Blaschke (talk) 01:50, 1 January 2016 (UTC)

Dungeon
Which etymology is more likely, the Germanic one or the Latin one? Tharthan (talk) 00:21, 1 January 2016 (UTC)
 * Phonologically, both seem possible to me; semantically, however, the Germanic etymology is much more straightforward. --Florian Blaschke (talk) 01:51, 1 January 2016 (UTC)
 * Allegedly the original meaning of "dungeon" was "keep" though. So I am unsure. Tharthan (talk) 03:42, 1 January 2016 (UTC)
 * True. I misunderstood "(castle) keep" because I wasn't familiar with the word and didn't know that it refers to a castle's main tower, not to a (subterranean) dungeon in the modern sense. Online Etymology Dictionary also supports the traditional derivation from dominio and makes the semantic transition more understandable. (Unlike the Oxford English Dictionary, which we should check, the Online Etymology Dictionary is not a professional source, it is only a lay compilation of existing sources, much like Wikipedia and Wiktionary, but it does seem to use professional sources such as the Oxford English Dictionary, so it is sensible to look there first.) Where does the Germanic derivation come from? Is it a Wiktionarian's OR? --Florian Blaschke (talk) 16:15, 3 January 2016 (UTC)
 * As this source also mentions the Germanic derivation, it's not OR, but it's not the most widespread etymology; it's mentioned only after the traditional one. It was Leasnam who added the Germanic derivation without also mentioning the Latin one. However, both in dungeon and in donjon, the traditional etymology should not only be mentioned but mentioned first, as it has not been discredited as far as I am aware. --Florian Blaschke (talk) 16:44, 3 January 2016 (UTC)
 * I don't know about that. It seems most likely that if nothing else this could be a conflation of terms. For Old English to have had dung for dungeon, for Frankish to have had *dungjo for dungeon etc., makes it seem that even if the castle keep meaning may have been derived from Latin, the dungeon meaning may well be due to the Germanic term. Who is to say that, for instance, the English word was not influenced in meaning and spelling (note the attested Middle English forms) by the existing English term?
 * Proto-Germanic had similar terms for related concepts as well, such as *dungz for "cellar", so it is by no means unreasonable to think that if the Germanic is not the origin of the term that it definitely influenced or conflated with it (at least in English).Tharthan (talk) 17:00, 3 January 2016 (UTC)
 * Precisely. In earliest English, the term already had the "subtertanean" sense, and this is undoubtedly due to association or conflation with native dung. The French tetm might have the tower sense perhaps from Old Norse dyngja Leasnam (talk) 18:15, 3 January 2016 (UTC)
 * This seems to be the most likely etymology, all things considered:
 * From Middle English "dungeon" (“castle keep, prison cell below the castle, dungeon”) a conflation of Old French donjon ("castle keep") and Old English dung ("dungeon, prison"). The Old French term is either derived from Frankish *dungjo ("dungeon, bower, underground cellar")/Old Norse dyngja ("detached flat, a lady's bower"), or from Vulgar Latin *dominio ("lord's castle"). Alternatively, the Old French term may itself be a conflation of the Vulgar Latin term and the Germanic term. The Old English goes back to Proto-Germanic *dungz ("cellar"), whereas the Frankish and Old Norse terms go back to Proto Germanic *dungijǭ ("dung heap, dung covered room, subterranean chamber, vault, treasury, flat, bower"). Both Proto-Germanic terms ultimately from Proto-Germanic *dungaz ("dung, manure"). Tharthan (talk) 18:42, 3 January 2016 (UTC)
 * I'm thinking a good way to settle this for French is to look for other words descending from Latin . If they have a different outcome in French, then can't be a descendant from . —CodeCat 18:53, 3 January 2016 (UTC)
 * The closest I can find is the ML. which might give lumignon. — JohnC5 19:05, 3 January 2016 (UTC)
 * I find it strange that -miniō became -njon. Old occitan has a byform dompnhon (h = j ?). Semantically, I always thought that the shift from "lordship" > "the place where prisoners are held" was a big step. What other Romance languages have this change ? Leasnam (talk) 03:17, 4 January 2016 (UTC)
 * It's actually -miniōnem that is the ancestor of the Old French ending. With stress on the ō. So the question is whether the -min- syllable would have syncopated; it's plausible that -mniōn gave -njon. We need other nouns with the same ending, to be able to compare the outcomes. —CodeCat 03:30, 4 January 2016 (UTC)
 * I checked a few French rhyme dictionaries and reverse dictionaries, and the only word ending in -njon I found was donjon. Searching a database dump, the only page on Wiktionary that contains miniōnem is (or was) lumignon, and the pages that contain miniō all seem to be inflected forms of -minium words. - -sche (discuss) 06:27, 4 January 2016 (UTC)
 * I've gone through and it seems that the normal outcome of -niō is -gnon . It seems to me that *domignon would be the expected outcome. — JohnC5 07:20, 4 January 2016 (UTC)
 * There's not that much distance between -gnon and -njon. I can see how preceding *dom- might inhibit the palatalization of the n. As CodeCat says, the real question is whether the *-min- gets syncopated to *-mn-. Chuck Entz (talk) 14:14, 4 January 2016 (UTC)
 * That I think is already settled in the literature. Yes, -miniōnem becomes -mnione(m) Leasnam (talk) 16:40, 4 January 2016 (UTC)
 * Ok, I was checking through Wikiling here and found this list of lemma forms “dominio* (1), domnio, domgio, dongio, dogio, donjo, doglo, dungio, dungeo, dungo, dunio, dunjo, dunlio, dangio, dunnio” meaning donjon supposedly culled from MLW 3, 1044, Niermeyer 463, Latham 159a, Blaise 277b, 324a, 329a. I do not have Niermeyer at hand to confirm this, but if this is true, then donjon definitely comes from dominiō.
 * Actually, the MLW does contain this entry verbatim. — JohnC5 15:21, 4 January 2016 (UTC)
 * We cannot conclude anything definitively based solely on the form. Yes, donjon can come from dominiōnem; but it can equally come from *dungjo. The question is not the form, but semantics. Which derivation fits the meaning better ? Leasnam (talk) 16:40, 4 January 2016 (UTC)
 * There is always too the possibility that a word of Frankish origin which has a corresponding sound-alike in a Latin word, affects the usage of the Latin word due to false association (i.e. the mistaken etymology causes users to begin using the (Mediaeval) Latin word as the Frankish word, leading to the false appearance in the historical record that this was a native sense in Latin when it was not). Can we screen for this ? This was fairly widespread in the bilingual environment of Mediaeval France Leasnam (talk) 16:54, 4 January 2016 (UTC)
 * Just out of curiosity (along a side note), does the French donjon retain the sense of "dominion, lordship"? Leasnam (talk) 17:04, 4 January 2016 (UTC)
 * According to CNTRL, apparently not. My French is rusty, but it lists those alternative Middle Latin forms (domnionus, donjo, dangio, etc.) as well and seems to say that the Old French word (judging from the variant forms in Middle Latin) probably originates in a conflation of dominiō and Old Frankish *dungjo, citing standard works on French etymology. So while it sure sounds plausible that the sense in English was influenced by Old English dung (is that word still attested in the Middle English period, though?), it seems that the influence from Germanic (whether Old Frankish or Old Norse) already happened in Gallo-Romance.
 * As for the syncopation issue: if the word was already present in Romance at an early date (I think the oldest Germanic/Frankish loanwords in French are affected too, such as Old French guisne from Old Frankish *wīhsina?), one would expect syncopation in *dominiōnem. It's quite possible that lumignon was borrowed from Latin only later, although per lumignon, it doesn't come straight from Middle Latin lūminiō anyway, but was only influenced by it, and in any case is not a direct continuation of a "Vulgar Latin"/Proto-Romance word lūminiōne(m) (if it already existed this early – it might be a medieval formation), while donjon is either inherited or was at least present early enough to behave like an inherited Romance word. So lumignon is not a reliable counter-example and cannot be used to argue against the possibility that donjon may descend straight from dominiō. (However, the fact that the nominative singular of donjon is not *donje or something along those lines might be due to Frankish/Germanic influence.) --Florian Blaschke (talk) 16:54, 8 January 2016 (UTC)
 * As for the syncopation issue: if the word was already present in Romance at an early date (I think the oldest Germanic/Frankish loanwords in French are affected too, such as Old French guisne from Old Frankish *wīhsina?), one would expect syncopation in *dominiōnem. It's quite possible that lumignon was borrowed from Latin only later, although per lumignon, it doesn't come straight from Middle Latin lūminiō anyway, but was only influenced by it, and in any case is not a direct continuation of a "Vulgar Latin"/Proto-Romance word lūminiōne(m) (if it already existed this early – it might be a medieval formation), while donjon is either inherited or was at least present early enough to behave like an inherited Romance word. So lumignon is not a reliable counter-example and cannot be used to argue against the possibility that donjon may descend straight from dominiō. (However, the fact that the nominative singular of donjon is not *donje or something along those lines might be due to Frankish/Germanic influence.) --Florian Blaschke (talk) 16:54, 8 January 2016 (UTC)

I haven't read every word of this thread, so maybe someone already said this, but if dominiōnem was syncopated to domniōnem, then donjon is the regular outcome. Vulgar Latin /mnj/ became /ndʒ/ in <, so presumably it would in this word as well. —Aɴɢʀ (talk) 17:03, 8 January 2016 (UTC)
 * Yes, but the issue is whether the syncopation can be expected to have occurred in the first place, as far as I understand. --Florian Blaschke (talk) 17:12, 8 January 2016 (UTC)
 * I think it would have or at least could have. I'm not sure syncope is always predictable; sometimes it fails to happen in environments where it might be expected. —Aɴɢʀ (talk) 18:22, 8 January 2016 (UTC)
 * I thought that in inherited Romance words it is virtually always present, and those where it seems to be missing have other explanations. (I can't think of any exception, but of course that's not saying much as I don't have all relevant examples at hand.) Compare Phonological history of French. --Florian Blaschke (talk) 18:55, 8 January 2016 (UTC)
 * So the Latin word dominiōnem "lordship" developed into VL *domniōne ("lordship"?), then got squeezed into a semantic bottleneck and narrowed to the degree that it solely refers to a castle's "keep" as donjon ...to me, that just doesn't make sense...it *had* to have been merged or conflated with a preexisting word. It's hard for me to accept that donjon is the sole inherited survivor in French of Latin dominio. It just doesn't line up that such a common and important word as Latin dominion could have been narrowed so much without leaving other senses in Old French having to do with "lordship, ownership" Leasnam (talk) 19:15, 8 January 2016 (UTC)
 * ...Unless, perhaps, it was caused by the Franks themselves. This development seems to be Gallo-Romance specific. Perhaps the Frankish Kings used domnion in place of *dungjo because it was reminiscent of *dungjo or because they confused the words (as they would have coalesced in form), assigning it an incorrect origin? So, if this is the case, what is the true etymon? Is it actually the Frankish word disguised as Latin (since donjon never means "lordship"), or a brand new use of the Latin word that emerged spontaneously in French due to whatever reason ? Leasnam (talk) 19:21, 8 January 2016 (UTC)
 * The problem is that even influence of Old Frankish or Old Norse doesn't really explain the semantic shift, either. As Tharthan pointed out, a castle keep is simply a tower, not a subterranean vault or an apartment. The semantic problem (insofar it is a problem at all; semantic shift is not nearly as regular and predictable as sound change) remains, whichever way you look at it. --Florian Blaschke (talk) 21:17, 8 January 2016 (UTC)
 * That said, considering that Old Norse dyngja apparently means "lady's bower", was the dwelling-place of a noble lady in a medieval castle usually (or at least frequently) located in the main tower? That does remind me of a motif familiar from fairy-tales such as Rapunzel and ideas of medieval culture widespread in popular culture, but is that motif based on an historical fact? --Florian Blaschke (talk) 21:38, 8 January 2016 (UTC)
 * The semantic part of the development of donjon doesn't bother me at all: the keep was the physical and symbolic center of a lord's military power. The phonological part seems plausible as well. It's only in the English part of the etymology that things get really dicey. It looks like dǒnǧǒun had three senses: keep or tower, underground cell as in modern dungeon, and subterranean cave or pit. There's also an early Middle English word dǒng, meaning "the pit of hell, abyss". The Middle English word is broader, semantically, than the Old French one, and it's tempting to attribute the to a merger with the Old English word, but I'm not sure how to read the Middle English Dictionary data. Chuck Entz (talk) 01:50, 9 January 2016 (UTC)
 * In response to Angr: you need to consider the stress as well. We're contrasting VL with  here. Vulgar Latin had contrastive stress. —CodeCat 19:36, 8 January 2016 (UTC)
 * <> The only thing I was able to find thus far is Middle English, apparently a diminutive of OE , Leasnam (talk) 19:42, 8 January 2016 (UTC)
 * Thanks. That means the semantic shift within English does not seem easy to explain, either. --Florian Blaschke (talk) 21:21, 8 January 2016 (UTC)
 * Actually, if donjon became dungeon through influence of Old English dung, then perhaps the Old English word is technically continued in "dungeon". That would explain the English spelling of the word, and why dung disappeared even though dingle (which exists as a word in Modern English too) exists: because dung was fused with donjon to make dungeon. I am convinced that the word donjon was either influenced by Germanic in French, or that it was influenced by it in English, OR that it was influenced in French, and then influenced by it again in English.
 * The fact that Old English dung disappeared suddenly whilst still evidently being productive (as dingle descends from it), points to it being replaced by dungeon. It is not farfetched to think that Old English dung was conflated with donjon to make dungeon. Tharthan (talk) 23:35, 8 January 2016 (UTC)
 * I must caution here, though: the derivation of Middle English dingle from Old English dung cannot likely have happened only in the Middle English period, considering the umlaut – if it is from (apparently unattested) Old English *dyngel, that is; but even if it is from Old English ding, the derivation can easily have happened before Middle English. There's no reason why dung cannot simply have disappeared without leaving any trace and without influencing anything, if dingle already existed before the Middle English period, which is highly likely. The apparent spelling change from Old French donjon (also spelt dongeon) to Middle English dongeon, dungeon, dungeoun, dungun need not betray the influence of dung, either; I would be careful not to read too much into it. --Florian Blaschke (talk) 23:52, 8 January 2016 (UTC)
 * Sure. It also could have equally likelily indeed influenced dungeon, and since that is what we're talking about here, I think that it is relevant to mention it. It is quite possible that it did influence it, and since neither of us were alive at the time, who can really say? The point is: dung existed in Old English, it meant "dungeon". "donjon" ("castle keep") was borrowed into Middle English, and miraculously gained the meaning of "dungeon". It also changed spelling. Old English dung may have still existed in Middle English, as evidenced by "dingle" (the alleged Old English etymon is unattested), but didn't survive past it.
 * There is no reason to think that the Old English word couldn't have influenced or been conflated with the Old French word to make "dungeon".Tharthan (talk) 00:00, 9 January 2016 (UTC)
 * As Chuck Entz' link above shows, OE *did* make it into Middle English as, , , so no need to worry about dingle. Therefore, it was possible that the Middle English word supplied the #2 and #3 senses in the above Middle English dictionary's entry at dǒnǧǒun. Incidentally, this source says that French donjon comes from Germanic Leasnam (talk) 02:05, 9 January 2016 (UTC)
 * Indeed. I think that this makes it seem very, very likely that "dungeon" is at least partially derived from Old English dung; primarily so, even, considering the only sense which survived is the one derived from Germanic. The spelling may well have been influenced by "dung", and furthermore the Old French word may have already been derived from Germanic at least partially. The point is, there is almost sheer Germanic impact on the word "dungeon". Tharthan (talk) 02:12, 9 January 2016 (UTC)
 * Considering everything here, I updated the etymology at . With the exception of listing the Middle Dutch form as a cognate of of Frankish (which it isn't...it's a descendant) (<== I revised this) I think it basically sums it all up. Please take a look and make changes if necessary. Leasnam (talk) 17:12, 11 January 2016 (UTC)
 * Looks good to me. Tharthan (talk) 18:46, 11 January 2016 (UTC)