Appendix talk:List of Proto-Indo-European roots

Definition from Wiktionary, the free dictionary
Jump to navigation Jump to search

Transwikied from w:Talk:List_of_Proto-Indo-European_roots

Discussion page is copied below; see also w:Talk:List_of_Proto-Indo-European_roots/Archive_1

Read this before editing[edit]

If you'd add new words, please try to follow the standards set by earlier contributors, regarding orthography, and try to have some certainty of the words you're adding, preferably check with some etymological authority beforehand, and look out for loanwords and false cognates. Don't assume you know a language's etymology, just because you speak it fluently! Although more cognates could be added, please try to stick to the languages already chosen, unless there is one root solely found in another language, not included in the particular language group. Note that this is a list of cognates derived from the same root, not translations of words' current meanings which could have diverged hugely due to semantic drift, i.e. for the root *preu (jump), Germanic words for "frog" are included, not words meaning jump derived from Germanic *springan, which originally meant the same thing. 惑乱 分からん 11:48, 29 April 2006 (UTC)

Well, apparently, *springan is believed to derive from the same root, via the S mobile, but the point should still come across. Wakuran 22:34, 2 November 2007 (UTC)


Koro is the root for war?[edit]

How can the English word harbor derive from this? They're totally unrelated concepts. - Christopher 07:19, 12 July 2006 (UTC)

Well, words change their meaning over time. Within the history of English, the primary meanings of "harbor" and "haven" have basically switched places. The primary meaning of "harbor" is a sheltered bay or the like where ships can safely anchor, while the primary meaning of "haven" is more generally "refuge". 800 years ago, the ancestors of these words had each other's meanings: Middle English herberge meant "refuge" and haven meant "harbor (for ships)". Herberge "refuge" goes back to an Old English compound herebeorg meaning "military headquarters". The here- part of the compound meant "army" and goes back to the PIE root koro- for "war". User:Angr 08:45, 12 July 2006 (UTC)

Yes indeed, it should be noted that Heer is still a German word for "army". The second part may be related to German "bergen" (to hide). The Slavic cognates "kara" mean "punishment" today.User:Cpom 27 Oct 2006

Re-inserting old roots[edit]

Jesus, I think I've finally re-inserted all the old cognates that were lost during the alphabetical sorting of the roots. Took some time... Watch out for errors that might have slipped in through the process... Anyway, now when that's finished, maybe we should come up with ideas for improving the article, like how languages should be sorted within the table, if the article should somehow be split in sub-articles etc. 惑乱 分からん 21:55, 14 July 2006 (UTC)

Pleated related to *plek-: to plait/braid?[edit]

I was just wondering whether the word "pleated" is related to the Old English "fleohtan", since as of now, there's no modern English equivalent in the chart. - Christopher 07:17, 21 July 2006 (UTC)

Pleat, plait and ply all (sometimes via French) derive originally from Latin plicāre, and are thus loanwords in English. English etymology is easy to find at, for instance, or etc. 惑乱 分からん 11:51, 1 August 2006 (UTC)

The Burushaski word 'Balk' meaning 'plank' seems to come from the same common PIE root 'plek-'

Burushaski is a language isolate, not derived from PIE, though it could have borrowed the word from some Iranian language. But I don't know of any Iranian language (or any other IE language for that matter) that changes PIE p to b. —Aɴɢʀ (talk) 10:46, 5 February 2014 (UTC)

Baltic Roots[edit]

Would someone please look at the Baltic roots that I posted in the discussion thread last month? Thanks in advance, Janis

Reposting query. I think the list still needs some work, and is still based heavily on the works of Julius Pokorny and Max Vasmer. 惑乱 分からん 08:30, 2 August 2006 (UTC)

I just recently stumbled on this article and I have a few Baltic cognates to add and some to correct. Also, some changes in spelling are necessary, especially in Latvian. I can't go through the list at once, it will take quite a lot of work, and I would like to discuss changes first. There are quite a lot of pie. roots that are missing IMHO, and some would need revision (i.e. *dákru-, why not *akru-?). With the help of the "Latviešu etimoloģijas vārdnīca" (Etymological Dictionary of Latvian) by Konstantīns Karulis (Rīga: Avots 1992) I would like to check the existing cognates. There are a few congnates mentioned here that are not listed in the dictionary mentioned above and that sound quite suspicious to me as I have never heard these words being a native Latvian speaker myself. I don't touch these, however. Here are the first few examples that are obviously wrong:

  • bher (in the meaning: to bear). The Lith. & Ltv. cognates that are mentioned here go back to the meaning of *bher as in "to cut, to rub, to split, to beat". Better examples in Latvian would be "bērns" (child), or "bērt" (to pour, to strew) and, in Lithuanian, "bernas" (young man, unmarried man, child etc.) or "berti" which both ACTUALLY go back to "to bear".
  • dra - both in Latvian and Lithuanian "drebēt" resp. "drebėti" go back to *der- (to tear, to pull, to split), not *dra-.
  • dheH₁(i)- the Lithuanian congnate ir correct, the Latvian cognate is only used dialectally in the meaning "to suck". It can be confused with "dēt" in the standard language in the meaning "to lay (an egg)" which goes back to *dheH. Furthermore, both cognates are listed under *dheH₁(i)- and *dheH₁-, but we want to keep them apart, don't we? Therefor I suggest the cognates "dēls" (son) or "dēle" (leech, sanguisuga) for *dheH₁(i)-.
  • dheub- - I agree with the Lithuanian cognate, "dibens" for Latvian is correct, but it has a late vowel shift dubens > dibens. So I suggest "dubt" instead, which corresponds with the Lituanian example. "dubļi" is another good example, but it means "mud" today instead of the more ancient "(animal) bowels, intestines".
  • dhreugh- According to Karulis, the primary meaning is "to be someone else", *dhrougho-s "another, another person", hence the opposite semantic changes to "friend" and "enemy". Good example for the same process is Latin "hostis" vs. German "Gast". If we can agree on this, we can add Ltv. "draugs" and Lith. "draugas" as cognates.
  • dhĝhu- Seems very suspicious to me in general. Greek should be "ikhthys" no "ikhthus", I see arm. "jukn" in Karulis' dictionary, and Yotvingian "żuvo" as an other Baltic cognate. He goes back to pie. *ĝhđū-, not *dhĝhu-.
  • ĝebh- I suggest the cognates "zebiekste" (weasel) for Latvian and "žebenkštė, žebenkštis" (the same) for Lithuanian
  • ĝenu- Why not Latin gena (<*genus)?
  • ĝhasto- According to Karulis, Latv. and Lith. "lazda" go back to *lēĝh "to collect, to gather".
  • ĝhelun-eH₂- What in the hell does Lith. "pušis" have to do with this????
  • ĝher- Should include Latvian "zārds" (rickstand, rack) and Latin "hortus" (garden).
  • ĝheu- If you consider dh : ģh to be one root as a result of the accomodation process (because if you dont, the "Skr. juhoti" is a false cognate), I can suggest the Latvian cognate "žaut", Lith. "džiauti" and "džiūti", Gm. "tou/Tau", Eng. "dew" etc. The reflex of both roots is identical in the Baltic.

More to come - Janis

The transcription for Greek in the article uses u, not y. Additions in "lesser known" languages by people who know their stuff is very welcome. 惑乱 分からん 08:36, 2 August 2006 (UTC)

Declination - lack of consequence[edit]

The best way to provide the proper form of word (i.e. containing the clean root) in Slavic languages, is to give its gen. sq. (in some cases gen. pl.) For instance:

n. sg. - gen. sg.
дочь - дочери [doč'er-i]
gołąb - gołębia [gouẽmb'-a]
dziecię - dziecięcia [ʒ'ećẽ(ũ)ć-a]
n. sq. - gen. pl.
niebo - niebios [ńeb'os-Ø]
telo - telesa [teles-a]

In the case of coniugation the most important rule is to avoid infinitive forms, which often alter the root's ending.

infinitive - 1st sq. - (so called "-l" form)
móc - mogę - mógł
мочь - могу - мог
kłaść - kładę - kładł
ići - idem - išao

This clearly indicates that the easiest way to extract the root is to give the form of 1st person singular (except "biti, byt', być, быть..." verb, which declines using *esmi, esti paradigm).

It makes some sense, what do others think? 惑乱 分からん 14:03, 15 October 2006 (UTC)
Certainly worth considering! --Petusek 09:54, 23 April 2009 (UTC)

Would someone explain the pronunciation rules?[edit]

How are the letters in the table of contents pronounced? Whats's the difference between h₁, h₂, and h₃, or the accent marks on the letters like ǵ vs. ǵʰ? Is there a site with sound samples I could hear? That would help a lot. Thanks. - Christopher 23:31, 24 September 2006 (UTC)

See Proto-Indo-European language. Not all of the pronunciations for the phonemes in question are precisely known; although there is broad consensus for some sounds, others, especially the laryngeals (i.e., h₁, h₂, and h₃) are quite controversial. ḱ, ǵ and ǵʰ are palatovelars: velar consonants which undergo palatilization. Velars are pronounced with the back part of the tongue against the back part of the roof of the mouth; palatovelars are pronouned with the back part of the tongue against the front part of the roof of the mouth. Basically, the difference between k and ḱ is comparable to that of the initial consonant of "cop" and the initial consonant of "keen". For the laryngeal pronunciations, see laryngeal theory. -Silence 00:14, 25 September 2006 (UTC)
Hey, Silence, thanks for your time. - Christopher 00:18, 25 September 2006 (UTC)


Ad *mor-: is this a mulberry rather than a backberry (both mûre in French..)? Dutch has moerbei for that.

Ad ar-no eagle; Dutch has arend for eagle.

Ad *dʰreugʰ- to deceive Dutch has bedriegen - bedroog - bedrogen; fallacy = drogreden (Gm Betrug)

Ad *gʰers- horror Dutch has griezel (weirdo) griezelen (to be horrified)

Ad *kʷr̥-mi- grub Is this worm in both Engl and Dutch?

Ad *bʰred- wade Is this ford in Engl. and voorde in Dutch (a place where one can wade through a river?) nl:Gebruiker:Sokpopje

  • 1. Check out an etymological dictionary before making guesses. Some might be correct, although I think "worm" isn't.
  • 2. I don't think Dutch is necessary, we already have English and German cognates, so adding a 3rd West Germanic language would only cause clutter.

惑乱 分からん 13:46, 17 October 2006 (UTC)

English ford comes from PIE *pert- (Grimm's law p>f) and is cognate with Latin portus, whence English port. There is no connection with *bʰred- or wading, though there is a deeper connection with fare, Dutch varen, from PIE *per-. Don't just guess etymologies. There is an excellent new five-volume dictionary of Dutch etymology currently appearing (two volumes are already out) if you want to do this properly. --Doric Loon 09:54, 20 October 2006 (UTC)
But anyway, do you think Dutch cognates are necessary? The way it looks now, I think it will seem cluttered. 惑乱 分からん 16:06, 20 October 2006 (UTC)


the list should go to wiktionary, Proto-Indo-European roots should remain as an article discussing PIE root structure. dab () 08:40, 20 October 2006 (UTC)

There are some other lists, as well. Should all lists be moved? 惑乱 分からん 16:04, 20 October 2006 (UTC)
no, just the lexical lists. the point isn't that this is a list, but that it is a list of dictionary entries. dab () 07:13, 24 October 2006 (UTC)
I tend to disagree. --Ghirla -трёп- 07:16, 24 October 2006 (UTC)
I think that a list makes some sense... 惑乱 分からん 11:06, 25 October 2006 (UTC)
I would very much vote for keeping the list. I don't see how it could be transferred to wiktionary as an entity. It is the list as a whole that is interesting, not so much the single words by themselves. However, the list does not need to list each and every root. The main purpose would be to serve as an extensive illustration for the PIE roots without an attempt of being complete. -- 20:42, 5 November 2006 (UTC)
  • Well, there are some lexical links that do belong on Wikipedia, and some that don't. The main distinction is how valuable they are as part of the encyclopedia, as opposed to part of the dictionary. For example, a list of expressions from a certain language, with associated historical and literary significance, would probably be valuable in the encyclopedia, since we have many articles for phrases and expressions and many non-verbal pages to link to, whereas a list of synonyms or rhyming words would be more suited to the dictionary, since it will almost exclusively be linking to articles for specific words. Consequently, the main reason to move this particular list to Wiktionary is that it provides very few pages to link to in the encyclopedia, and thousands and thousands of pages to link to in the dictionary: indeed, we can have individual dictionary articles for every single PIE root and every word derived from it, an exceedingly valuable resource. A more encyclopedic listing would focus less on having a comprehensive, expansive wordlist and more on providing historical and linguistic information (in a "notes" column, perhaps) and detailed references for the article. -Silence 12:38, 25 October 2006 (UTC)
If it's retained in list form, I guess moving makes some sense. But how should the discussion best be preserved? 惑乱 分からん 12:14, 26 October 2006 (UTC)
Who says we need to delete this page? We can simply transfer its current contents to Wiktionary, and replace them here with more encyclopedic contents that explain the linguistic and historical significance of each root more and are better-referenced. A "List of Indo-European roots" isn't inherently unencyclopedic, it's just that this specific list is more valuable on Wiktionary than Wikipedia because it exclusively deals with listing words, whereas a list that took a more in-depth look at each root might be better-suited to Wikipedia because of the potential to link to related explanatory articles. -Silence 12:52, 26 October 2006 (UTC)
If there was to be an article, the words should probably be listed thematically, rather than alphabetical... 惑乱 分からん 13:16, 27 October 2006 (UTC)

Root of 'Bid', 'Bitten' etc[edit]

Isn't the root of English 'Bid', German 'Bitten' etc Proto Indo-European *bhidh-? In which case, the Latin cognate would be 'Fidere' not 'Manifestus'.

Not according to American Heritage... 惑乱 分からん 11:09, 25 October 2006 (UTC)
Here is what we have in the Etymological dictionary (available on line) for bite: O.E. bitan (class I strong verb; past tense bat, pp. biten), from P.Gmc. *bitan (O.Fris. bita, M.Du. biten, Ger. beissen, Goth. beitan), from PIE base *bheid- "to split, crack".
That wasn't what was asked about... 惑乱 分からん 20:05, 6 November 2006 (UTC)

Russian edits[edit]

The edits by Extern looks suspicious and might be Original Research. Could someone knowledgeable in Russian etymology look them through, and weed out loanwords and false cognates?... 惑乱 分からん 11:12, 25 October 2006 (UTC)

They indeed qualify as original research. The main problem with this list is that every passerby adds something which sounds to him familiar to other words from the list. For instance, many Polish words ("gust") seem to be loans from German. I don't see any merit in enumerating them in the list. Here's a survey of Extern's additions: 09:29, 5 November 2006 (UTC)

I just joined the site (and found it exciting). Being originally trilingual (Ukrainian/Russian/Polish) I have to make some comments about some Russian words beginning from S (more to come).

  • swergʰ-( to be ill) Russ. сорога (soroga): this word does not exist in Russian; the derivative exists in Ukrainian – хворий (xvoryj), ill.
  • sewyós left Russ. шуй (šuj) – this word is from OCS, it was replaced by левый (levyj) long before the Russian language appeared and does not exist in Russian.
  • solwo-(whole) Russ. сулей (sulej) – does not exist in Russian. There is a Ukrainian word сулія (sulija), which means a big bottle (usually of alcohol) and sounds quite inspiring, but it’s most likely non-related (o>u is not known for Slavic).
  • srebʰ- (to gulp) Russ. сёрбать (sjorbat'), this is not a Russian word, it’s Ukrainian сьорбати.

None of these Russian words can be found in the Wiktionary – another indication of there absence. What are the rules – how can I correct other people’s errors? Just remove them does not sound absolutely appropriate.

By the way, I am not an expert in Albanian and Armenian, but these two examples sound quite confusing:

  • sweid- sweat Arm. k'rtink'/k'rdink', Alb. Djersë
  • swep- sleep Arm. k'nel, Alb. Gjumë

I do not see any relationship between the IE roots and the Arm/Alb words – maybe, whoever placed it here thought about this place as a dictionary?

Albanian and Armenian are for a very good reason last "known" languages to be proved being Indo-European. Both have appeared to 1) changed quite immenesly from the reconstructable forms 2) exhibited significant substratum/superstratum influence of Latin and Iranian, respectively, which have obscured the original state of many basic words quite a bit. For an example of a rather bizarre sound law, cf. PIE *#dw- yielding Armenian #erk- (that '#' signs means word-initially). Although, as far as the first point is concerned, in glottalic framework Armenian is actually the most archaic language (second archaic being Germanic...the least archaic coming Indo-Iranian and Balto-Slavic, i.e. the oposite of the "laryngeal theory" framework). But, glottalic theory has only one school of supporters nowadays (University of Leiden, headed by F. Kortlandt and R. S. P. Beekes) and most modern IEists regard it as possible state of affairs for pre-PIE (which is unprovable), or just rubbish.
As for the particular lexemes you mention:
Polish gust - indeed false cognate, LW from (Old/Middle) German, impermissible Centum reflex in Satem Slavic. Removed.
Russian soroga - could found this anywhere, seems to be hoax. Pending verification.
Russian šuj - LW from OCS, removed. Proto-Slavic *šujь appears to have been reflected only in OCS and Slovene.
PIE *sólwos "whole" does not seem to have Balto-Slavic cognates at all. Pending verification.
Russian serbátь - false alarm, confirmed reflex of PIE *srebʰ-.
Those two Albanian and Armanian words - I can confirm the Albanian as being properly derived from PIE, but have no idea of Armenian.
Note an important thing: this whole appendix was originally compiled by some industrious folk at Wikipedia, primarly basing it on somewhat outdated Julius Pokorny's , Indogermanisches etymologisches Wörterbuch, a work of outstanding scholarship that is still to be matched in its comprehensivness of IE treatment. Lots of cognates (and "cognates") were added by some passers-by on the criterion "hey, this looks familiar to the word X, let me add it..". All such words, of course, need to be purged, and dubious ones that cannot be verified tagged or simply removed. This is wiki, everyone can contribute, so for the doubtful entries you do have some modern etymological dictionary stating otherwise, please be bold and remove any nonsense you encounter. Ultimately, entries are slowely being verified one by one as the appendix pages for individual entries are being created (see those clickable blue links for PIE sources). --Ivan Štambuk 22:12, 18 January 2009 (UTC)
Russian has the word сорога meaning a fish (w:ru:плотва, see в толковом словаре Даля, w:Rutilus rutilus), but not in the meaning of sweat. —Stephen 23:01, 18 January 2009 (UTC)
I just meant, that the word сёрбать does not exist in Russian language at all - correspondent Russian word is хлебать, and if you with use this word in Moscow, you will be called "хохол" right away. That's why I believe the reference language for the word should be corrected to Ukrainian. Kachur 14:38, 19 January 2009 (UTC)
Russian dialects are really rich, the only problem is that most of local words came from non-IE sources, Ugro-Finnish or Turkish languages. Example сорога is a good one - meaning "fish" does not seem to have any relationship to *swergʰ- (to be ill), it just sounds kind of similar; in the same time Ukrainian хворий (xvoryj) means "ill" and seems to be a derivative from the IE root: *s>x is well known, the rest of consonants are the same, as well as meaning.
Actually Russian has хворать/xvoratь which means "to be ill" as well. 16:59, 6 November 2011 (UTC)

By the way, here are some other Ukrainian words, I consider relevant for s-roots:

  • (s)mer- (to remember) - Ukrainian памороки (pamoroky; root -mor- means mind)
  • (s)ter- sterile - Ukrainian терти (terty, means to rub – probably, to make something sterile)
  • (sm̥-)ǵʰéslo- thousand - OCS зѣло (zělo, means too much; s was lost to open the syllable)
  • seh₂g- to seek out - Ukrainian шукати(šukaty, to seek)
  • selp- butter - Ukrainian сало (salo, animal fat)
  • swergʰ- to be ill - Ukrainian хворий (xvoryj, ill)
  • swonos- to sound - Ukrainian дзвонити (dzvonyty, to ring), звук (zvuk, sound)

These all are original contributions - because Ukrainian was long time ignored by IE scholars in favor of Russian (by the way, ALL the Russian derivatives in the database have Ukrainian analogs), I did some research of my own in mid-90s, but never published anything. Should I just place these words into database, or there should be some prior discussion? Thank you Kachur 14:38, 19 January 2009 (UTC)

Vasmer lists серба́ть/сёрбать as Russian (also серба́ти as Ukrainian) so they shouldn't be touched. Etymologists don't care much about usage of words in spoken language, only of their etymological relationship and importance. Lots of words you'll find in this lists (of modern languages:Lithuanian, Persian, even English) are found only in the dialects, or appear only in literary language. That certainly doesn't make them somehow "lesser" in value!
As I said: lots of words, like probably the now disputed Russian сорога, were added by passers-by on WP who don't know much on their mother tongue's historical phonology, and were solely driven by the desire to promote their language even if it meant adding false cognate. (the odder common case is being added from obsolete sources). I fail to see how exactly is Ukrainian хво́рий or Russian хво́рый a reflex of PIE *swergʰ-, and both of them look highly dubious to me.
Of Slavic sources, only OCS is permitted (as an ancient "language") and Russian due its importance (etymons should probably be grouped as OCS/Russian pairs in the list, like the other ancient/modern languages are, even though in strict sense OCS in not an "ancestor" of Russian). Polish, Ukrainian and other languages should appear only if neither of the former two have a Slavic reflex of the PIE root/word in question (and they should have in 99% of cases). Of those examples you list here, some of them are obviously wrong. Please, original research is forbidden. --Ivan Štambuk 16:00, 19 January 2009 (UTC)
Good for Vasmer - he knows my languages better then me (but still his stresses are completely wrong - so, too bad for stresses. And even if the word серба́ти will sound stupid in both Ukraine and Russia, let's trust Vasmer!).
Ivan, your comment sounds sad. Especially these two phrases “Lots of words you'll find in this lists are found only in the dialects, or appear only in literary language. That certainly doesn't make them somehow "lesser" in value!” and “Of Slavic sources, only OCS is permitted (as an ancient "language") and Russian due its importance” placed in the same message.
So, even some dialects (of Russian) are important, but other Eastern Slavic languages – Ukrainian and Belorussian – are not even permitted herein! Did I understand you correctly: these two should not better even exist due to lack of their importance, right?
I though that as a Croatian you are at least a little bit familiar with a history of Eastern Slavs, and do not consider all of us as “Russians”. Sorry for my mistake. 00:36, 20 January 2009 (UTC)
Max Vasmer/Oleg Trubačev are among the greates Slavic etymologists of the 20th century, and you're just some anyonymous IP. Who do you think has more credibility? ;) If you have verifiable external evidence suggesting otherwise, please present it.
The reason why they shouldn't be listed here is because they're irrelevant. Old Church Slavonic itself is in many reflexes almost identical to Late Common Slavic. Modern words should only come into consideration to compensate the absence of OCS reflex. For Italic branch we list Latin, Osco-Umbrian et. al - no Romance languages (descendants of Vulgar Latin). For Indo-Aryan branch of Indo-Iranian, we only list Sanskrit (and no Hindi or any other modern Indo-Aryan language, which are spoken by 900M people). Of course, in some circumstances it's very convenient to group ancient language and its modern descendants (or closest match, as not so rarely they belong to slightly different dialects). For OCS in particular, that would be Bulgarian or Macedonian. Russian is, OTOH, a special case because it's much more archaic in phonology/morphology then either of them, plus being culturally the most "relevant" Slavic language, so it might be good to put it instead.
Remember, this is just this list. There are appendix pages for PIE and Proto-Slavic which are not limited by space considerations as this appendix, and which contain reflexes in all the possible languages you can imagine - even dialectalisms. Lots of those PSl. tables are missing Ukrainian cognates so you might give them a go if you feel like doing something useful and highly appreciative :P
I assure you that I am personally in no way a supporter of "Soviet scholarship" and utilise my sysop powers here to minimise the damage done by supporters of 19th century rubbish theories. Yes, even today they are users that claim that Ukrainian is "little Russian", or are more intent to call Old East Slavic (literary language of Kiven Rus') by its anacrhonistic misnomer "Old Russian" (Rus'Russian!). The funny thing is that from the modern dialectal perspective, it would be much more appropriate to call it Old Ukrainian ^_^. Hence we use politically most correct and much more proper term Old East Slavic and not "Old Russian" even if the latter is (unfortunatelly) in much higher use in Slavic studies, though it is slowly but surely being obsoleted. And the same is valid for other ones as well: there is no issue around here on whether there is Macedonian language or not (Bulgarian linguists a few yers ago apparently agreed that M "does not exist" ^_^), or whether Silesian is a "Polish dialect" or language by itself. I'm sorry that you've used the particular practice used on this page to draw the conclusion of editors taking a particular (pro-Russian) PoV, but I assure you that it is certainly not tha case here and that the usage of Russian here has to do with practical considarations such as space and relevance. --Ivan Štambuk 01:31, 20 January 2009 (UTC)

Thank you, Ivan – and sorry for a little bit erratic comment. You know, too many people still consider Russia as a center of the Universe, Ukrainian and Bielorussian as a mixture of Polish and Russian, etc. And when linguistics sounds like an extension of politics (“Russian is more important” fits the pattern), I show my teeth. Well, Russians as a nation appeared only few centuries ago (that’s why they are still so aggressive in conquering new territories – most of civilized nations went through this childhood disease, an invasion stage, long ago), and their language is a mixture of different Slavic, Finnish and Turkish substrates (Krivichies, who mostly became Bielorussians, Siverians, who are Ukrainians, Viatici and Radimici – according to Pimen, Polish people; Novgorod Slovene – also close to Ukrainian; plus Muroma, Meria, Ves’, Moskva, and other extinct Finnish tribes); all this mess was united by Church Slavic. Kind of absurd situation – Cyril and Methodius made an official Church Slavic out of the most southern Slavic dialect (definitely different from the rest of the Slavic languages); later it influenced the rest of the Slavic languages; the most affected mixture of Finnish and Turkish words ironically became a language of the most aggressive Russia. And, since Russian is very close to OCS, the rest of Slavic languages became irrelevant for Indoeuropean scholars, despite their slogan “no language left behind”. But in reality Ukrainian and Bielorussian (mostly not spoiled by official OCS) are much closer to original PIE - simply because they represent a language of original "aryans", peasants from Trypill'a - arətrom means "to plow", орати in Ukrainian.

Okay, no more philosophy. So, you don’t like amateurs – how about this: From Lexicon der Indogermanischen Verben, H. Rix, Weisbaden, 2001:

  • bhag –(to be interested) Gr.v, Ved. Bhájati, JAv. Bažat

Now, if I will add to the meanings Ukrainian “бажати” (bazhaty, to wish), will you still consider it as an original contribution, not wanted in the list? You know, I just picked the first word in the book – and it somehow has a Ukrainian derivative (by the way, not existing in Russian). Is this acceptable in this website, or you just want to have here a light copy of some heavy Indogermanischen großbuch ?

I'm am very much familiar with PIE root *bʰeh₂ǵ- "to share, distribute" (PIE had no *a), as it's the source of Slavic word for "god" - Common Slavi *bogъ < Proto-Slavic *bagu, which is usually assumed to be either an Iranian borrowing (absence of Winter's law), or influenced semantically by it (in both branches in means "deity" and "wealth, share"). Ancient Greek ἔφαγον is secondary to φαγεῖν. AFAIK, that word had no direct reflexes in Slavic, and Vasmer's entry [1] for Russian бажа́ть, Ukrainian бажа́ти, Polish zabagać, Czech bažiti se, zabahnouti dismisses the connection with the aforementioned PIE root, which is obvious as the Common Slavic form would have to have */g/ which cannot reflect PIE */ǵ/. I am really astonished that LIV lists this Slavic word as a cognate. I don't have access to LIV, but I've heard comments that the inclusion criteria in it are not too strict either - it attempts to reconstruct for PIE forms that have reflexes in only one branch, not to mention much too-complex verbal system. Sorry, that word can't go. --Ivan Štambuk 08:16, 20 January 2009 (UTC)
Actually, now I see that Sanskrit has dual reflexes bhaj-/bhag-, so there could be both variants *bʰeh₂ǵ-/*bʰeh₂g-, the latter which would have given the alleged Slavic forms (Greek and Tocharian cognates don't tell us anything). Oh well, if LIV indeed lists it, you may add it, but only the Russian cognate бажа́ть. --Ivan Štambuk 08:25, 20 January 2009 (UTC)
OK, I downloaded LIV from some Russian site [2], and the entry for the aforementioned root on the p. 65 mentions no Slavic cognates. Please, no original research! --Ivan Štambuk 10:37, 20 January 2009 (UTC)

Okay, here is not an original research, but some analysis of the same Lexicon. It used only 9 (nine) Ukrainian words. Here they are: 1. bháty<*bhewgh- sort of confusive derivative, *ew should give u/ju, not ø. 2. čérsty<*(s)ker- I did not find "čérsty" in Grinchenko’s Dictionary of Ukrainian Language (Kiev, 1909) – the most comprehensive Ukrainian dictionary, not spoiled by communist “linguists” like stalin. So, the word is kind of confusing and most likely not Ukrainian. 3. jebaty/jety <*jebh – not Ukrainian. First form is misspelled, in Ukrainian it’s always jibaty, i<ѣ; version with e< ѣ can be only Russian; the second form is exclusively Russian and never used in Ukrainian. They'd better leave this kind of words to Russians. 4. kojity<*kʷej- at least the word is Ukrainian, but I am sort of confused - *ej>i is much more likely, I would expect here *kʷej >ki>či, like in Wiktionary *kʷeiH-> OCS počiti, Ukrainian spočyty. 5. livyty<*leh1- –what is "livyty"? The root’s meaning is nachlassen – to drop, to slip, to abandon. Aha, лишити (lyšyty) means “to abandon” – probably they meant this. But it’s really confusing and the word is seriously messed up. 6. lykaty<*slewk- - unknown Ukrainian word again. Meaning “to swallow” – maybe, лизати (lyzaty), to lick? Another version of the root is *(s)lewg- - sounds closer, palatalized g gives z, although that *ew>y is completely confusing in both possibilities: should be *ew>u, I could imagine here "лузгати/лускати". Failed again – not Ukrainian word which does not fit to root. 7. s’katy<*h2ejs- okay, there is this word in our eastern dialects, but I would rather use here Russian iskat’ (where it most likely came from), which by the way keeps that initial i<*h2ej, lost in Ukrainian. 8. snit<*snejt- - what in the world is “snit”? The root means “to swallow”, the Ukrainian word is translated as Klotz… Aha, Ukrainian “гніт” (гніт) load! Now if somebody will explain me how g<*s, I will be totally happy. Looks like someone misspelled Ukrainian word, which later was confused with irrelevant IE root. 9. tjaknuty<*tek- - unknown word for me, but Grinchenko gives the meaning “to touch”, the same as in the Lexicon – okay, this one is correct.

As you see, the entire picture is really bad – most of the words are either misspelled or misused; and huge amount of words (which are totally different from OCS or Russian - you will be surprised, but most of Russians simply do not understand pure Ukrainian language) just ignored. Hopefully now you can see why I consider Ukrainian language as left behind and it’s using by IE scholars as quite improper. It looks like there was almost no research on the topic. Well, in Russian empire Ukrainian was simply ignored, and it’s sad that this tradition is continued herein. 14:49, 20 January 2009 (UTC)

Thank you for your comments. I haven't used LIV before, but now that you write this I'll pay double attention to whatever cognates they list in there. As for the Slavic/Ukrainian cognates on the Wiktonary: they are usually checked against several sources, and though a few errors slip here and there, generally are 99% reliable. There is massive Index:Ukrainian to chack against (dumped from some Ukrainian dictionary by folks at the Ukrainian Wiktionary and copied here by me), and of regulars User:Mzajac has some profficiency in the language to help when needed though he seems to be much more occupied with technical issues of this project lately. As I said: if you really want to make sure that Ukrainian cognates are properly added, go and fill Ukrainian language columns at those Proto-Slavic appendix pages :)
BTW, this paper might interest you (you need to have IP in some academic institution to acces it): the first genetic research as to the problem of establishing Slavic Urheimat (and we all know where PIE Urheimat was.. :) It pretty much corroborates traditional linguistic theories (e.g. the area with the highest density of Slavic hydronyms according to Trubačev). --Ivan Štambuk 23:41, 20 January 2009 (UTC)

Ivan, thank you - but your Proto-Slavic reference above does not work, can you please repeat it?

About Russian cognates in the database - I found few more of them definitely wrong; I will place summary later. But before it - please, look at this:

Just two examples of my "original research":

  • ph₂tḗr - father, 18 examples from different languages, but there is no Ukrainian батько (bat'ko) - the only Ukrainian word for "father". No clear origin given in Vasmer, but similarity with IE root looks obvious.
  • népot - nephew, in Wiktionary Old Russ. нестера (nestera). I simply do not see how that P was replaced by S - no clear laws, just assumptions. But, maybe, there is no need to explain it, because Ukrainian небога (neboha) actually means niece, and небіж (nebiž) - nephew; they are much closer to IE root.

You are confused with *p>Ukrainian b in both examples? Okay, just one more example:

  • skap- tool - in Wiktionary Rus. shchepa. Maybe (although, palatalization of initial consonants seems to be suspicious, as well as *a>e), but there is Ukrainian (not found in Russian) скабка (скабка) - again, ignored due to its autenticity.

I am eager to read your comments :) 01:41, 21 January 2009 (UTC)

Okay, here is a list of strange Russian derivatives. Sorry, the file is quite long.

  • bʰeh₂go- beech- Russ. бузина (buzina), - wrong plant, it’s Ukrainian бук (Fagus silvestris), while бузина is Sambucus nigra.
  • bʰer- brown, shining- Russ. бобр (bobr), - wrong: word Бобёр <*bhebhrus; I would place here Ukrainian брунатний (but it’s again original research).
  • dus- to fail - OCS дъждь (dŭždĭ), Russ. дождь (dožd'), Polish deszcz – does’t look logical and relevant.
  • dʰeb- fat - Russ. дебелый (debelyj) – suspicious; Ukr. дебелий means rather “big person”, not "fat".
  • dʰegʷh- to burn - Russ. сжигать; жгучий (sžigat'; žgučij) – I do not see any connection between root and derivatives, except meaning.
  • dʰeh₁-to place, set, put - Russ. деть; делать (det'; delat') – DET’ maybe, but delat’ – to do – seems not appropriate.
  • dʰeiǵʰ- to mold - Russ. дежа (deža) – do they mean Ukrainian dizha, “barrel”? Well, it’s pretty strange connection.
  • dʰergʰen- thorn - Russ.- дёрн (djorn), “sod”. Again, I would place Ukrainian “dereza”, quite a spiky plants (there are three of them, Caragana frutescens, Lycium barbarum, Lycopodium clavatum, all are similar).
  • dʰeub- deep - Russ. дно; дупло (dno; duplo) – both look suspicious, especially the first one.
  • gerbʰ- to scratch -Russ. жребий (žrebij),- just omophone; I would prefer Ukrainian “гребінь”
  • ǵenH₁- (*ǵénH₁ō [1ps]) - to give birth - Russ. зять (zjat') – totally irrelevant, <*ǵémH₁ter; жена looks more appropriate
  • ǵónu- knee- Russ. звено (zveno)????
  • nem- to allot/distribute - Russ. немой (nemoj), “mute” - seems to be nonrelated.
  • skap- tool - Russ. щепа (ščepa), I do not see any way to go from the root to the word.
  • solwo- whole - Russ. сулей (sulej), - we already discussed it.
  • swergʰ- to be ill - Russ. сорога (soroga), and this one is wrong, too.
  • gʰedʰ- to join, to unite- Russ. год (god),”year” – does't look relevant. I would prefer herein Ukrainian root “god” in words ugoda (agreement), pogodytys’a (agree) etc.
  • h₁edʰ- sharp- Russ. ель (jel')"fir" - do you see any connection???
  • h₁eu- to dress Russ. обуть (obut') - the same here???
  • h₁euhdʰ-r̥- udder Russ. вымя - meaning is the same, but the word and root are not connected.
  • h₁oh₁ḱ-u- quick - Russ. ястреб (jastreb)?
  • h₁reudʰ-ó- red Russ. рдеть (rdet'), - I do not see where *eu gone, but Ukrainian “рудий” fits much better.

By the way, the root *h₁rebʰ- to roof - is doubled in database.

  • akʷā- water - Russ. Ока (Oka), - name of this river is Finnish, not Slavic.
  • h₂eig- sick - Russ. яга (jaga), - person from finnish fairytales, not related.
  • h₂eiḱ- spear/pike - Russ. игла (igla) – I don’t see the connection.
  • h₂eus-dawn - Russ. утро (utro), I do not see any connection again.
  • kob- to succeed - Russ. кобь (kob') – unknown word, seems more like a sick joke.
  • ḱerberó- piebald - Russ. соболь (sobol'). Seems suspicious; more probable <собака
  • ḱlei- to lean -Russ. слой (sloj), layer - looks suspicious
  • ḱleu- clean - Russ. слеза (sleza), tear – also suspicious
  • meiǵʰ- to urinate -Russ. мезга (mezga) – irrelevant word, should be моча
  • mer-to die - Russ. мереть (meret'), - mistake, must be "умереть"
  • morwi-ant Russ. муровей (murovej), - mistake, must be "мурaвей"
  • peg- breast - Russ. пах (pax) wrong meaning, the relationship looks suspicious
  • pléu-mon- lung - Russ. плюче (pljuče), - senseless word: there is no this word in Russian
  • pray- to like -Russ. прижать (prižat') completely wrong connection: pri- (prefix pre) – zhat’ (press) means “to grasp”; nothing to do with IE root; I would rather use Ukrainian “приємний”, pleasant.
  • kʷrei- to buy - Russ. кренуть (krenut'), - again, nonexisting Russian word; somebody was playing or something.
  • gʷʰen- to press; to strike, slay, kill, - Russ. жать; гнать (žat'; gnat') – the second word is wrong.
  • leykʷ- to leave - Russ. олек (olek), - again, the word sounds just absurd.
  • lendʰo- steppe - Russ. ляда (ljada) – the same, looks like a bad joke.
  • lewk-; *lówk-o bright; light Russ. луг (lug), - senseless correlation
  • tark- to twist - Russ. торок (torok), - Russian word is suspicious, Ukrainian “torkaty” could be much better
  • teḱs- to plait Russ. тесла (tesla) – wrong word, sounds more like last name of famous physicist Nikola Tesla; misspelled Ukrainian тесля
  • yeg- ice Russ. икра (ikra) – I don’t think so.
  • yeh₁kwr̥- liver Russ. икра (ikra),- again? Kind of another sick joke?
  • weik- victory - Russ. век (vek), century???
  • wel- wool- Russ. волна (volna), - wrong, it’s "wave" – must be Ukrainian вовна, wool
  • wred- root - Russ. вред (vred), - wrong cognate, “damage”; it has nothing to do here

Sorry again for a long file. Maybe, not everything is correct, but few words (mezga, meret', murovej, pljuce, krenut', olek, ljada, torok, tesla) are really mocking, placed by some jerk with sick sense of humor. 04:45, 21 January 2009 (UTC)

OK, thanks for your comments. I can promise you that your list will be scrutinized when the time comes, but until then I must ask you for patience as this list contains thousands of etymons in very obscure languages that have not been spoken for millenias and whose individual checking is extremely time-consuming.
As for your first comment: PIE *ph₂tḗr was not reflexted in Slavic at all, Ukr. bat'ko looks like hypocoristic to me. PIE word for "nephew" was reflected as *netjь in Common Slavic, preserved as OCS and Old East Slavic netii - Russ. нестера looks indeed dubious. *skap- root is completely wrong (there was no *a in PIE), and Russian ščepa is unrelated. There is PIE (s)kop- "to dig, burry", whic Connects OCS kopati, Lith. kapoti, Latv. kapati, OP enkopts, Greek κόπτω and Albanian kep, while the rest of that entry is just rubbish.
As for the appendices you're looking for, they're in this category. --Ivan Štambuk 09:53, 21 January 2009 (UTC)
"non-existing word" Before making such claims, did you check dialectic dictionaries? Because literary Russian language, or Moscow dialect, isn't the whole Russian. And if you don't know a specific word, it doesn't mean it doesn't exist; it may be present in dialects, or maybe it was used in the past, but died out in the modern language, etc. 17:05, 6 November 2011 (UTC)

Completely worthless[edit]

  • varit' - completely different root, cognate to Eng. warm
  • pod - cognate to L рēs, pedis ("feet")
  • byk - derived from the Slavic word for "bellow, moo"
  • derźat' - cognate to SK dražaitē
  • odomašit' - a verb derived from "dom" (house), cf. Eng. dome
  • taš'it' - a dark word, no cognates known.
  • liznut' - the same root as L lingere, Eng. lick, etc.
  • dremat - related to L dormio, which is a different root.
  • čuvstvo - derived from the verb "chuyu", related to Sk kavíṣ, Gk κοέω, L caveo
  • žor, žrat' - related to another lemma, see Sk giráti
  • žostkij - usually compared to MHG kes, ON kǫs, OIr gall
  • golosit' - a verb derived from "golos" (=voice), usually connected ON "kalla", Irl "gall", Cymr. "galw", etc.
  • vzyat' - a false friend
  • gore - derived from the Russian word for "burn"; a different PIE root
  • grečiha - "buckwheat" means "the Greek one" in Russian: it was imported from Greece.
  • ževat' - the same lemma as in Eng. "chew", Arm "kiv"
  • zat' - related to Gk γέμω
  • goroh - no connection to Gk orobos and L ervum
  • uvažat - a pan-Slavic loan from OHG wa^ga
  • kon' - an abbreviation from Old Russian komon'

Perhaps there is some connection, but it is not generally accepted:

  • udirat - derived from "deru", which could be related to other words, or could be derived from another lemma, cf. L dìrti, Sk dr̥ṇā́ti, Gk δέρω, OHG zëran.
  • d'ogot - valid connection, although Vasmer denies it; sžigat' is one of numerous derivations, should be eliminated for brevity.
  • dočer - this genetive word was featured in my original list; was deleted for no apparent reason.
  • ženit - derived from žena and connected to others though the PIE lemma for woman; I suppose this lemma should be eliminated from the list as derivative.
  • izvilina - derivation from the verb "vilyat", which should be compared to OE wil, wile and Lt vylùs
  • gryzt' - the connection seems to be valid, cf. Lt gráužti (=gnaw), Arm krcem (=gnaw), and Gr βρῡχω
  • koza - a difficult word: Vasmer connects it with Old English "hacele"; Korsh postulated a loan from Turkic; Meillet seems to be correct when comparing it to Sk. ajás
  • groza - a valid comparison, although both Vasmer and Shakhmatov oppose it
  • pytat - probably correct, though Vasmer prefers to connect it to L puto. --Ghirla -трёп- 14:10, 26 October 2006 (UTC)
Good, could you remove the "worthless" edits? 惑乱 分からん 13:28, 27 October 2006 (UTC)

Sanskrit: Dakshina[edit]

I'm no expert on PIE or Sanskrit, but I was always under the impression that "Dakshina" meant "South". That is the meaning of the word in modern day Sanskrit, and at least Malayalam and Tamil, which are to a large extent derived from Sanskrit. Not suggesting a change to the article here, just want some personal clarification. Did the word Dakshina in its current form evolve from the sanskrit word for "right"? If that's the case, perhaps there was a cultural connection between "South" and the direction "Right" which may be interesting to explore. Can anyone clarify? Cheers.

Vineetmenon 23:10, 24 November 2006 (UTC)vineetmenon

Cognates often undergo semantic shifts. This could also be seen in modern languages. Wakuran 00:31, 1 January 2007 (UTC)

Sanskrit - form of verb derivatives[edit]

I wonder why Sanskrit examples of verbs derived from roots are given in a conjugated form, instead of its root (for example, bharati instead of bhṛ - which, in my opinion, highlights the relationship a bit better). As far as I'm aware, other roots are given in their dictionary form, as well. If I missed something as to the why, please enlighten me. ETA: Or, if we give a conjugated version, it should be standard for all. Baranxtu 23:38, 24 November 2006 (UTC)

roots, not stems[edit]

help to reduce this to a list of roots, not including complex stems. derived stems can be listed in all detail in entries on individual roots. I am creating a sample root entry at peḱ- now. Dbachmann 14:07, 5 December 2006 (UTC)

On, those go in the appendix namespace, where I've moved it now: Appendix:Proto-Indo-European root *pek'-. --Connel MacKenzie 16:22, 5 December 2006 (UTC)
this is a bad idea imho, and may need to be revisited (no hurry though, things can always be moved later). There is no reason wiktionary can not also be an etymological dictionary, and it is unclear why etymological entires should constitute an "appendix" to anything (other than lists such as this one, which I agree belong in an "appendix" by their nature as lists, not entries). We will (hopefully) end up with thousands of such entries, which in my opinion will constitute a valid part of wiktionary, not an "appendix". Dbachmann 11:28, 7 December 2006 (UTC)
what I could envisage is that reconstructed (etymological) entries take an asterisk in the main namespace to set them apart, that is, *peḱ- rather than peḱ-. Needs to be discussed in more detail of course. Dbachmann 11:30, 7 December 2006 (UTC)


I removed the Romanian examples, for several reasons, (although I forgot to log in, first or was logged off automatically). 1st, I personally don't think there's a need to list examples from daughter languages, in case the mother language is so well attested as Latin. 2nd, the examples given were a mess of evolutions from Latin, Slavic borrowings and likely unrelated words superficially similar to the roots. The poster didn't seem to bother checking up the etymology with a source before editing. Taken together with the implication that the examples given were somehow all from a Thracian substrate, I felt it was best just removing it. Wakuran 14:20, 1 January 2007 (UTC)

Removed languages[edit]

I just removed a couple of languages from the language list. (Mostly Germanic and Slavic.) Maybe it should have been discussed further first, but I felt that there were too many closely related languages with a lot of cognates in common. Seemed as an invitation of cluttering cognates. Nothing I'd personally like too see in the article's current state... Wakuran 22:56, 21 January 2007 (UTC)

Transcription etc.[edit]

I recently reverted some Russian transcription. I'm not against the idea, but I think a Latin transcription would be more useful for the article's scope. (I would probably have left it, if the transcription wasn't removed in the process.) The transcription seems to need some proof-checking, tho.

Anyway, we could include both original alphabet and Latin transcription for most languages, as well as meaning and probable sound evolution etc, but it would require some effort, besides the article is large, already, as it is, and would probably need some splitting. Also, there's been concerns earlier on copyright problems. The bulk of the article seems to derive from proposed roots from Pokorny's Wörterbuch, and his books are apparently still under copyright.

Languages with attested writing in alphabets other than Latin in this list include Hittite, Carian, Luwian, Lycian, Lydian, Palaic, (Old Norse), Armenian, Greek, Sanskrit, Kashmiri, Avestan, Ossetian, Persian, Pashto, (Kurdish), Oscan, Umbrian, Old Church Slavonic, Russian, Tocharian A & B, Venetic, Phrygian, Ancient Macedonian & Thracian, although I'm not sure if there are fonts for all these languages. Wakuran 12:43, 19 May 2007 (UTC)

The Cyrillic words can be looked up here, while the transliterations cannot be. Also, most people who have a serious interest in such things have already spent the half hour it takes to learn Greek and Cyrillic scripts. I’m also interested in etymologies, and I consider the present form with everything in Roman letters to be unprofessional and I would not waste my time referring to it. In the recent past, English dictionaries have had a policy of Roman only in their etymologies, but this was due to hardware and software limitations that made the original scripts very difficult or impossible. —Stephen 11:36, 20 May 2007 (UTC)
I know that, and that is a huge advantage of including them. As I see it, the problem with your edits was that you removed the transcriptions in the process, instead of keeping them in paranthesis, or something. Greek and Cyrillic scripts might not be too hard to learn for people who know Roman, but this page would require (at least) examples in Devanagari, Perso-Arabic script, Armenian alphabet etc. so in all, I think some kind of Latin transcription should be given as a default, since this is a page mainly for comparative purposes. There probably exist many options to increase the page's professionalism, correcting incorrect/ad hoc transcriptions, removing improbable cognates etc. Anyway, if you'd agree that "боб (bob)" is an acceptable format, I'd help reinserting your Cyrillic edits. The main drawback to using two different writing systems for Non-Latin alphabets is the space required, but the list probably is so large it would require some splitting, anyhow. Wakuran 14:14, 20 May 2007 (UTC)
Arabic, Cuneiform, Armenian, Georgian, Devanagari and East and Southeast Asian scripts all require transcriptions, since few can manage to read them (unlike Russian and Greek). However, I’m not against having "боб (bob)", since it makes sense to apply one rule to all non-Roman scripts. I was intending to link all of the Russian words, many of which already have articles here, all of which include a transliteration. —Stephen 14:25, 20 May 2007 (UTC)
Yes, I think "original script (Latin translit)" makes the most sense. At least, I have started a reinsert of the transliterations now. If you won't have any time, I'll get back to it during the week to finish it. This article has been a copy from Wikipedia, so it originally wasn't mede with Wiktionary in mind. When that work's finished, the list probably would need some more proof-checking, weeding out of improbable cognates, etc. Wakuran 18:18, 20 May 2007 (UTC)
Also, not everyone has Unicode, so a Latin transcription is also helpful in that matter. (Btw, I forgot Gothic, although it doesn't seem widely supported, currently.) Wakuran 11:50, 21 May 2007 (UTC)


Are you sure "name" is *H₁nōm-n̥-? Because most languages seem to derive this word regularly from *H3nōm-n̥-, with intial H3 (cf. Greek o-, Armenian a-, Anatolic a-, most others Ø). 07:01, 17 June 2007 (UTC)

Opinions differ, including *H₁nom-n̥- with short o (thus Karin Stüber, and I think she's right), *H1neH3-mn̥-, *H3neH3-mn̥-. Choosing one particuler solution (or one particular author) would qualify as personal research. What counts is the reasons why initial *H1- is better than *H3- (or conversely) in spite of the apparent difficulties, and why it is (un)necessary or (not) advisable to reconstruct a long *o or an internal laryngeal, but explaining that would, again, qualify as personal research. Hittite lāman- requires *H1- not *H3-, Greek onoma does not require *H3 because *H1 gets assimilated to o before certain consonants that are followed by o or u. --Zxly (talk) 03:21, 13 July 2014 (UTC)

Some cleanup and additions[edit]

I am fixing some errors and adding the Devanagari for Sanskrit. Azalea pomp 20:27, 11 July 2007 (UTC)

Linking words in Latin alphabet?[edit]

Perhaps we should start linking most words in Latin alphabet, as well. The languages using Latin is, as far as I know, Albanian, All Baltic languages, All Celtic languages (Or did Old Irish use Ogham?), All Germanic languages except Gothic (Some Old Norse words attested only in Runes, I think), Kurdish (?, uses several alphabets), Latin, Polish... Maybe we should start linking these words, as well. Also, maybe the article would benefit from some splitting, it is quite unwieldy... Wakuran 22:57, 23 July 2007 (UTC)

This page is an appendix of PIE roots so I am not why it needs to be split and there are already two or three other pages with roots. I think the Noun page does not have some of the same format as this page though. Azalea pomp 19:24, 24 July 2007 (UTC)
Size, mostly. What I meant was split in different parts, not into articles with different subject. Wakuran 08:55, 25 July 2007 (UTC)
I am still working on the Armenian. As you know it takes a long time. Perhaps if each person could do one language for the links it would go faster. Azalea pomp 17:20, 31 July 2007 (UTC)
Do you mean checking for correct words, writing up descriptions and etymologies in own articles, etc. or do you mean fact-checking for errors? I'm mostly an amateur on this, although I could maybe manage Germanic and Latin, if I weren't short on time, currently... Wakuran 13:44, 1 August 2007 (UTC)

Splitting article?[edit]

I'm seriously thinking on splitting up this article. It takes a huge time to load, even on a broadband connection. I'm thinking about maybe seven sub-articles(b-d, g, h, k, l-r, s, t-y), with the main page containing links to them all. In the process, I'd probably also link words from languages using the Latin alphabet, as well. Any comments? Wakuran 02:02, 23 September 2007 (UTC)

I concur. Also, wikification should not be done directly (except for English words), but via something like {{term}} (or {{l}} to which functionality of transliterations and alternative texts has not yet been added but will be; I smell major duplication of code) that can generate links to appropriate L2 langauge sections and transliteration in parentheses for lemmas in non-Latin script.
BTW, I find Dysfuntion's by-branch splitting extremely visually appealing as opposed to current layout of cognates that aren't sorted neither alphabetically nor by branches. Languages that don't have a cognate should be marked as not having them, which is a useful information as opposed to not listing it at all. --Ivan Štambuk 14:35, 7 November 2007 (UTC)
I've also noticed some entries (e.g. Russian and OCS verbs) not being written in lemma form, but in (for example) first-person singular. Basically following what the anonymous dude tells here. Unlike 'pedia, wiktionary can and does provide inflected forms in which PIE roots shine, but on the appropropriately wikilinked articles inside declension/conjugation tables. So unless we're dealing for example with Sanskrit and Old Irish where verbs are commonly given in third-person singular present indicative, I think all entries should be for the sake of uniformity be lemmatized. --Ivan Štambuk 15:29, 7 November 2007 (UTC)
I am not familiar with all the technical functions here on Wiktionary. I believe the forms chosen should be the ones that most clearly show the genetic relation. I'll write more, later. Wakuran 19:46, 7 November 2007 (UTC)
The problem with Dysfunktion's approach is that the article gets too broad on an average computer screen, A large amount of scrolling is required. I'm not against the idea, per se, just the look of its implementation. Wakuran 11:34, 8 November 2007 (UTC)
Yes indeed, but one should keep in mind that that wikitable-based approach would:
1) eliminate redundant '''''Language abbreviation''''' and thus (significantly?) reduce page size
2) provide natural grouping for genetically related languages (eg. South Slavic -> Slavic -> Balto-Slavic -> Satem) that would emphasize common evolution with languages listed in neighboring columns
3) both facilitate locating missing entries in a specific language and provide an option to mark a language/family not having a cognate from a specific PIE root (which is also a valuable information, mind you)
If you feel that fitting the article within the confines of an average computer screen outweights potential gains of wikitable-based approach, feel free to ignore anything I said :P --Ivan Štambuk 21:20, 8 November 2007 (UTC)

I find the table layout, split by category to be more informative, and as you said it's easier to see relationships between language families, but I think everything should be kept on one page because it's a pain to keep scrolling all the time. Perhaps the tables could be split up to keep it within the width of the page? ( 11:15, 12 November 2007 (UTC))

Maybe one table for Centum and the other one for Satem languages?
I also think it's redundant to have listed cognates for multiple languages of the same branch (like OCS, Russian and Polish). In fact, I would rather see them all marked as Slavic and have as entry the earliest known attested cognate (in whatever language).
The other option is to extract the data from table and create separate articles for each one inside Category:Proto-Indo-European roots (see for example **peḱ-, where hierarchical structure is simulated with appropriate indentation). Or maybe combining that kind of "horizontally hierarchical" layout in separate articles for category-based structuring? --Ivan Štambuk 11:51, 12 November 2007 (UTC)
Isn't the Centum-Satem distinction theory somewhat deprecated, currently? I agree that a clearer sorting by family could be useful, though. Wakuran 18:06, 12 November 2007 (UTC)
It was just a suggestion - any other partition will suite. Maybe approximate chronological (by time of departure from PIE "homeland") would be the best one to show how the the reconstructed roots eventually deteriorated into attested forms? --Ivan Štambuk 18:14, 12 November 2007 (UTC)
Yeah, that's better, although it still adds another level of hypothesis. Wakuran 13:37, 13 November 2007 (UTC)
The problem with IE classification is that there is not much of a consensus among linguistics about the homeland, chronology, or how the various of branches of IE related to one another. For sure there should not be a centum and satem table as that is not a branching of IE, just a sound change. Azalea pomp 22:12, 21 November 2007 (UTC)
Yeah, as far as I have understood, early on in PIE studies, some linguists considered the Centum/Satem distinction as the original PIE dialect division, but most recent scholars consider it just one IE isogloss among many. By the way, do you have any other ideas for a sorting that'll be smaller and having clearly related languages annd/or language groups next to each other? Wakuran 23:47, 25 November 2007 (UTC)

From my understanding, discovery of Tocharian invalidated geographical distribution of Satem/Centum isogloss (and some others generally thought to hold for PIE proper), and Craig Melchert proved in 1987 how triadic arrangement of PIE dorsals was retained in Luwian. Anyhow, it's still widely used today, since both Anatolian and Tocharian subgroup are divergent enough from "Central PIE" to be neglected. Basically Anatolian branch is one half of PIE, and Tocharian is the one half of the other half (considering their "divergence" with respect to others), and for the simple purpose of splitting this list Satem/Cantum isogloss would suffice just fine (putting Luwian wherever). Or do however you please - I personally don't care much :) Content is what matters, not the layout.

Another thing - I started recently continuing Dbachmann's noble work, adding more PIE stuff into Category:Proto-Indo-European language, and used information in this and the list of PIE nouns heavily. However, with 450+ PIE languages with ISO 649-3 code (that therefore merit inclusion into wiktionary), most of them will never find place in self-contained table with a limited number of columns. Grouping them hierarchically like in e.g. *ph₂tḗr, *méh₂tēr, *nókʷts would not limit the amount of languages to be added. So I'll focus mostly to promoting PIE stuff there, where it can also be linked from articles in NS:0 (using {{proto}}). --Ivan Štambuk 21:45, 20 January 2008 (UTC)

Lithuanian edits?[edit]

Could someone familiar with Lithuanian etymology check out the recent anonymous edits. I am a little suspicious that many would just be superficially similar false cognates. Wakuran 17:04, 19 December 2007 (UTC)

See User talk: - at least one of them is very, very wrong. I suggest you revert all of them and put them on the talk page, until someone reviews them and filters false cognates. --Ivan Štambuk 21:21, 20 January 2008 (UTC)
The edits were made 16 december, 2007. I'll come back, later. Wakuran 23:37, 23 January 2008 (UTC)
See also on Wikipedia this IP: This user appears to think that all PIE languages have "descended from Lithuanian" ^_^ All of his "cognates" are probably worthless; he added them just to falsely present Lithuanian as having more PIE cognates than other language. --Ivan Štambuk 20:09, 25 January 2008 (UTC)

This is the dump of dubious Lithuanian cognates, that need checking:

I'll remove them from the list, because lots of people reference it when listing cognates in Etymology sections of the entries, and we can't risk adding false cognates that were most likey deliberately added. --Ivan Štambuk 11:31, 4 February 2008 (UTC)

Good done. The user didn't remove any earlier examples, we'd need to insert? Wakuran 16:26, 4 February 2008 (UTC)
No, he didn't. Apparently the WP article on PIE root directly links to this list, so I guess there's a good potential for "insertions" such as these in the future. Slowly but steadily, outbursts of nationalism so prevalent of language-related disputes on WP, project onto wiktionary too :/ --Ivan Štambuk 21:57, 4 February 2008 (UTC)
Yeah, as could be expected. Hopefully, there are knowledgeable editore here that could weed them out quickly, though. Wakuran 14:49, 5 February 2008 (UTC)

Appendices: lists of Indo-European sound changes for all Ianguages listed here[edit]

Can I ask you something? Why is it that there are no lists of sound changes for each language listed here? I guess that would make it easier to predict which word gives place to another word, and the other way around. It would make the whole table more transparent, because everyone would be able to check whether the word forms are correct. For example, for English you would create a list of sound changes, and then so you take the Indo-European word, and then an English word comes out of the algorithm. That would make it easier to control the thing.

So all the sound changes would be listed in chronological order, and then if there are doubts as to the existence and/or chronological order of sound changes, that would be listed too. Why not do that? It would make this whole thing much more economical. An by "this whole thing" I mean the creation of lists of roots and lists of etymologies and so on. --Blog321 16:52, 30 April 2008 (UTC)

My first impression is that it sounds complicated. Anyway, there's a listing of the major sound changes on Wikipedia. Wakuran 17:52, 30 April 2008 (UTC)
Yeah, not only does it sound complicated, but it is, in fact, complicated. But think about this: Which of the two methods is more complicated: to create this table here, where it's not easy to verify the entries, or to have the rules that help you derive a derivate word from a base word? My bet is that this table here is more complicated, because there is not enough predictability. That's the whole point. If the readers of this table here had the sound change rules in addition to just the table, it'd be easier to control that the entries here are correct.
Thanks for pointing out that Wikipedia entry. I had already seen it before, and my problem with that table is that it is incomplete in the sense that it does not list the sound changes in between, the steps when you go from Proto-Indo-European to a present-day Indo-European language. The problem with that is that, first of all, it's not easy to verify that that table is right, and, secondly, when you want to derive a derivate word from a base word, you can't, because the individual steps are not listed. --Blog321 18:26, 30 April 2008 (UTC)
We don't need to "verify" anything; we just need to synthesize and present the knowledge of the last ~200 years of Indo-European studies. If any discussion is to be held, that is the one with quotes of professional linguists. Anything other would be original research and that kind of content is forbidden.
As for the "middle step" - you might want to check out the slowly growing PIE forms in the Appendix: namespace - some of them have reconstructions for individual branches. Ideally, one day each of those would have extensive ====Comment==== section that would discuss individual "laws" that gave attested forms, and that would link to wikipedia (much like {{etyl}} links to 'pedia now). Unfortunatelly, there are lots of these little "laws" that don't even have 'pedia article.
You're free and welcome to contribute to that noble goal in all possible ways ^_^ --Ivan Štambuk 18:37, 30 April 2008 (UTC)
Do all the dictionaries referenced at the bottom of the page return the same reconstructions? --Blog321 21:08, 30 April 2008 (UTC)
As far as I know, no. I think they differ both in details of the reconstructions, and opinions on whether words could be traced back to PIE roots or not. Wakuran 22:27, 30 April 2008 (UTC)
In most of the relevant and faithful cases, they do, if you count "4", "2+2" and "2*2" as different means of expressing the same thing. Linguistic reconstruction is just that - an algebraic expression that is to be interpreted, not a written speech. In most of the speculative cases, well, we'll try to mention all the mainstream interpretations and reconstructions (per Wiktionary:Reconstructed terms). Note that this excludes all the nonsense like the glottalic theory, dumping clear evidence in favour of some abstract ideals such as no */a/ in PIE or "canonical" forms of roots, Albano-Illyrian theories or Harvey E. Mayer's interpretation of Balto-Slavic --Ivan Štambuk 23:35, 30 April 2008 (UTC)
From which dictionary or dictionaries are you copying the reconstructions? --Blog321 00:38, 1 May 2008 (UTC)
Well I personally tend to combine various resources into what I think it's best ^_^ Every single resource on PIE you'll find is somehow "biased" towards some particular principles, depending on author's vision of what he thinks is the Right Thing (TM). But generally, there are only minor variations for most of the basic roots/stems presented in e.g. this appendix. Whether you write Pokorny-style *māter (which is actually more likely, assuming it originated from nursery-type words, and that there's absolutely no evidence for a laryngeal in attested reflexes in this particular case) or more "modern" version *méh₂tēr - it's really the same thing. --Ivan Štambuk 06:49, 1 May 2008 (UTC)
Why don't you copy all the reconstructions for a word and let the reader decide which one is best? --Blog321 08:49, 1 May 2008 (UTC)
I lemmatize the reconstruction which I think is the most prevalent in literature, and mention other ones in the inflection line (usually by putting symbols in parentheses). It seems as the most reasonable approach. Not so rarely, several forms can be reliably reconstructed, accounting for different PIE dialects, like the word for dream; *súpnos, *swépnos or swópnos - each having reflexes in several subfamilies, the former two even both in Indo-Iranian; in cases such as this I still think that they should both be mentioned on the same appendix page, but that's just my opinion there are no official guidelines set on matters such as this.
If you have specific comments to any of the reconstructions listed on this list or in the Appendix pages, I'm sure people would be glad to discuss. I don't think anyone here has formal linguistic training, but this ain't no rocket science either ^_^ --Ivan Štambuk 09:47, 1 May 2008 (UTC)

Greek words and question marks?[edit]

What's the reason for the question marks after the Greek words? Do they look strange or unfamiliar? Wakuran 09:23, 2 May 2008 (UTC)

There are a couple of reasons for those. The first is that the Greek definitions do not match up with the purported definitions. The second is that I couldn't find a Greek word (so you'll see both words with Greek spellings and ones without having the question marks). The former is basically just a note to check sources more thoroughly than otherwise. -Atelaes λάλει ἐμοί 13:16, 2 May 2008 (UTC)
Well, there are often sematic shifts, and someone could have moxed up ancient and modern Greek. Wakuran 21:32, 2 May 2008 (UTC)
Yes, there are a number of things that could have happened. It's possible that a grc word genuinely has a different meaning than its PIE etymon, or someone could have jacked the transliteration slightly, or any number of things. The question mark is just a signal for me to be a bit more careful. -Atelaes λάλει ἐμοί 22:23, 2 May 2008 (UTC)

General cleanup[edit]

Hi all,

I'm doing wikification of most of the entries (those that have or will have Wiktionary entries), fixing spelling errors and similar, and the following things are open to debate

  1. The need to have multiple languages of the same branch, when there's a clear cognate in much earlier extinct languages. I'd like to completely remove Polish entries, and leave Russian ones only where there's no OCS attestation.
  2. Also there's no need for 3 Baltic languages, Old Prussian is the most conservative of them and should be primarily used, with Lithuanian as a replacement (same logic as with Slavic). So Latvian goes to oblivion.

No information will be lost - all the Old Prussian entries (well, at least those I've taken care of ^_^) list other Balto-Slavic cognates (Lithuanian, Latvian and Common Slavic reconstruction - there are plenty of those inside Category:Proto-Slavic language and many more to come!), and the same holds for OCS entries.

Another thing, this *** software apparently has "template expansion limit" based on the consumption on memory while it processes wiki code, so as soon as it hits it - templates don't get expanded. This could be a problem for these bigs lists, and I already encountered it with Appendix:List of Proto-Slavic nouns, which was after that split into subpages. This list will have to be split sooner or later, for technical reasons, as more wikifications are made.

Also I'd like to propose the following arrangement of the entries:

Anatolian, Indo-Iranian, Ancient Greek, Italic, Germanic, Celtic, Tocharian, Armenian, Slavic, Baltic, Albanian

This roughly corresponds with their antiquity. Phryigian and similar scarcely attested ones would go somewhere in the middle --Ivan Štambuk 03:32, 4 May 2008 (UTC)

A couple thoughts on this. First, I find everything you suggest to be quite reasonable. However, I do wonder if this is truly worth all the work. In my opinion, list of words appendices are generally not useful long term. It seems to me that, eventually, Category:Proto-Indo-European language should really fulfill the roles which this appendix does now, and shall do a far better job. The reason why I have been linking grc words is simply to facilitate the creation of entries. However, if others disagree, I have absolutely no problem with upgrading this appendix. -Atelaes λάλει ἐμοί 05:53, 4 May 2008 (UTC)
The reason given for Polish was that there should be a representation of each sub-branch. I.e. OCS as South Slavic (most archaic attestations), Russian as East Slavic (largest ES, currently largest Slavic) and Polish as West Slavic (largest WS), although I could see why you want to have it removed. The arrangement of language groups seems reasonable, though. Wakuran 08:12, 4 May 2008 (UTC)
Well, I could also imagine a list of respective proto-forms (like Proto-Slavic, Proto-Baltic, Proto-Germanic, etc.), linked to the actual attested forms, but I guess nobody'd be willing to do this. :) Nevermind. It's good to see this appendix growing anyway!--Petusek 10:10, 23 April 2009 (UTC)
By the way, considering splitting, I quote myself from earlier:
I'm seriously thinking on splitting up this article. It takes a huge time to load, even on a 
broadband connection. I'm thinking about maybe seven sub-articles(b-d, g, h, k, l-r, s, t-y), 
with the main page containing links to them all. In the process, I'd probably also link words
from languages using the Latin alphabet, as well. Any comments? Wakuran 02:02, 23 September 2007 (UTC)

Wakuran 08:14, 4 May 2008 (UTC)

Well, the 9th century Slavic "languages" where just a bunch of more or less mutually intelligible dialects, and >50% of OCS attestation relevant for this list are identical to what we can reconstruct to 6th century Late-Proto-Slavic. I'd only leave Russian as culturally the most significant Slavic language. So maybe changing the format for Slavic to OCS/Russian ? --Ivan Štambuk 15:23, 4 May 2008 (UTC)
Well, it gives the impression that Russian is modern OCS, which seems a little like an over-simplification. Wakuran 19:04, 4 May 2008 (UTC)
I don't like that, because it would then to be hard to search by the browser (In epiphany by Ctrl-F) for a certain cognate. I instead propose using transclusions (which actually means the text is split into parts, but preferrably one article per letter) and keep this article using transclusions to include those article into this one. A typical section in [[Appendix:List of Proto-Indo-European roots]] could then look like:
  {{/letter b}} <!-- letter b is name of the subpage for letter b -->
  {{/letter bh}} <!-- or some such -->
And the article /letter b could look like:
  <!-- heading here enable us to edit by pressing the link b from
       the page [[Appendix:List of Proto-Indo-European roots]] -->
  == b ==
    ... contents of table ...
Then the article is actually split-up, but mainly behaves as one page when browsing it. Rursus 14:07, 4 May 2008 (UTC)
I just tried this with Appendix:List of Proto-Slavic nouns and this really works ^_^, i.e. template expansion limit is not hit. Fantastic idea.
Also, I'm thinking of making a template just for appendices such as these that would generate language names (and link them to 'pedia) while doing wikification. So e.g. something like
would generate:
Old Prussian swestro
It would also accept the usual sc= and tr= parameters for setting the script and providing the transliteration, and also the second unnamed parameter could be omitted and instead a cognate could be listed in the third unnamed parameter, which would prevent wikification (useful for scripts not yet encoded in Unicode, or when spelling is not provided in native script and no one can verify it). Underneath, {l-x} would call {{language}} to generate language name and link it to 'pedia and redirect the display logic to {{term}}.
This would reduce the raw list size for at least 30%; just look at the wiki code a see how much space is wasted on language names! --Ivan Štambuk 15:12, 4 May 2008 (UTC)
Yes, that's half of what I meant, for the Appendix:List of Proto-Slavic nouns, I'll produce this sample page to demonstrate how the article [[Appendix:List of Proto-Slavic nouns]] could appear. Rursus 15:18, 4 May 2008 (UTC)
The article looks humungous, but actually just contains this code:
  {{Appendix:List of Proto-Slavic nouns/Agriculture and crafts}}
  {{Appendix:List of Proto-Slavic nouns/Animals}}
  {{Appendix:List of Proto-Slavic nouns/Body}}
  {{Appendix:List of Proto-Slavic nouns/Household}}
  {{Appendix:List of Proto-Slavic nouns/Landscape and seasons}}
  {{Appendix:List of Proto-Slavic nouns/People}}
  {{Appendix:List of Proto-Slavic nouns/Vegetation}}
  {{Appendix:List of Proto-Slavic nouns/Society}}
Rursus 15:22, 4 May 2008 (UTC)
The article still takes time to load, though, and (maybe it's necessary, but still) the tables are too broad to fit one single screen. However, those technical solutions still sound rather elegant. Wakuran 16:09, 4 May 2008 (UTC)
Hmmm... that's a tough one. Then I propose that the page is written like Ivan's Appendix:List of Proto-Slavic nouns, so that surfing into Appendix:List of Proto-Indo-European roots provides links to such a page that I described, and links to the parts, like he has partitioned List of Proto-Slavic nouns. That would provide a solution that is less troublesome to low bandwidth surfers, and yet provide the opportunity for Ctrl-F searching, for those that really dare to. Rursus 16:20, 4 May 2008 (UTC)
I already suggested table-like display some time ago, much like it was done here (this appendix page was transwikied from from WP), but the argument against that was that it would be too broad to fit one screen, and not as convenient too browse as this list (all of which makes sense). I don't know about you, but these Lists of Proto-Slavic words really fit to one screen for me, even on 15-inch display, but the same definitely couldn't be said for this list. However, this all doesn't mean that one day table-like display like that one done by Dysfunction couldn't be made as an alternative.
With something like {l-x} and Latvian/Polish removed, this page size should be well under 200k which is reasonable even for low-bandwidth users. --Ivan Štambuk 16:38, 4 May 2008 (UTC)
My screen is 1280x1024, which should be considered pretty good. The tables fits one screen if I zoom down to smallest font possible in my browser (a Mozilla clone), perhaps about font size 6pt or 7pt, but the text is then barely readable. Maybe there's some space to be saved by changing the heading font to be normal, not bold. That could be achieved by
  !style="font-weight: normal;"|adding the style font-weight as made to the left
and instead mark the heading by some suitable background color, also added into the style tag, s.a.f.ex.:
  !style="font-weight: normal; background: #CCCCFF"|heading-heading
Another trick would be cutting reconstruction into recon<br>struction. Otherwise the problem you refer to is inherently hard to solve. As regards to saving space by using templates - that's of course the way to save space used by the mediawiki servers, but it saves no bandwidth for page transfer, since the templates are expanded before sent to the wikisurfer/editor. Sadly, the solution I propose for making the tables less wide increases the need for bandwidth. Some solution would be to create nice ENTITIES, but that's hard core HTML/XHTML hacking fit for the Wikimedia constructors, and I'm not sure we can create entities for mediawiki to send to the surfer/editor. I'll think a little about it. L8R. Rursus 16:58, 4 May 2008 (UTC)
Are you sure that there wouldn't be some on-thy-fly compression that would be facilitated by using templates? I don't completely follow what you're saying here. It would be much easier to visualise what you're saying if you'd create some sample page illustrating your ideas. --Ivan Štambuk 17:14, 4 May 2008 (UTC)
I'm not completely sure. I believe there is no on-the-fly compression here. What I'm speculating about is the HTML code generated by the mediawiki servers (wikipedia, wiktionary and their kin), sent via internet to the client computer. In "advanced" versions of HTML (all browsers should be able to handle such "advanced" HTML versions by now), the're is an opportunity to define entities, maybe you know of &amp; for the ampersand "&" sign. Such can be defined in HTML by preceeding the optional <!DOCTYPE html PUBLIC ...such and... heading ("SGML head" for tech-affiniados) by an <!ENTITY ...define this and that... entity declaration of &your-own-entity; so that long sequences of text can be replaced by shorter entities. Those entities are expanded in the browser, not on the mediawiki server. Rursus 17:27, 4 May 2008 (UTC)
I see what you mean. I'm not all that into web programming, but it would be great if you (or someone) could devise a way by using this new entities that this could significantly reduce the amount of raw content sent to client, possibly combining this with some specifically made template such as {l-x}, that would expand pregenerated language names and links to wikipedia.
Now I wonder whether the subst:-ed {{l}} inside those big Proto-Slavic lists would generated the same amount of web traffic as the nonexpanded versions, or the only benefit of using {{l}} is to reduce the amount of raw content in edit mode. --Ivan Štambuk 19:01, 4 May 2008 (UTC)

Looks like it's high time to split the article, because "Node-count limit exceeded" messages are being generated starting from *kob-. --Ivan Štambuk 19:49, 9 May 2008 (UTC)

I blame you Ivan. If you hadn't put so much good work into this project, we wouldn't be having this issue. :-) -Atelaes λάλει ἐμοί 19:52, 9 May 2008 (UTC)
^_^ BTW, I did some testing with the subst:-ed {{l}} and normally used one with those Proto-Slavic lists, and they really appear to generate exactly the same amount of HTTP traffic server->client when viewed. So using something like {l-x} would only facilitate editing this list (I don't know about you all, but I found those five apostrophes in a row just horrid).
After spliting it by initial phoneme ("letter"), we could decide by their individual size which optimal subset to present to the viewer. --Ivan Štambuk 20:10, 9 May 2008 (UTC)
Whoops, doesn't appear to be working, "Node-count limit exceeded" is still being generated (now on *knid-, which is just above *kob-). Too bad. If there where only a why to transclude those subpages on-demand (like clicking some button or something). --Ivan Štambuk 20:29, 9 May 2008 (UTC)
Well, I don't think it'd be the end of the world if each letter had their own page, as long as they were well linked to each other. That kind of extra space would probably allow us to do some interesting things, such as put it all in a table format, like you've got going on with the proto-Slavic stuff. -Atelaes λάλει ἐμοί 21:52, 9 May 2008 (UTC)

Strange, I don't get the same "Node-count limit exceeded" message like I did a couple of hours ago...I'll continue with the splitting and we'll see what comes out of this.. --Ivan Štambuk 00:22, 10 May 2008 (UTC)

Sorting order??[edit]

??? Rursus 16:23, 4 May 2008 (UTC)

EOR?? Rursus 16:26, 4 May 2008 (UTC)
I'd rather that we stick to some homorganic arrangement than to alphabethic (these are not signs of the alphabhet!), so something like: p, b, , t, d, , , ǵ, ǵʰ, k, g, , , , gʷʰ, s, (but, sonorants grouped together), y, w, r, l, m, n (and laryngeals) h₁, h₂, h₃.
Also, there's a problem with forms such has *atta, *albʰos, *alyos that are very often reconstructed with initial vowel and therefore don't satisfy the "canonical" form of PIE roots.. we should probably dump them at the very end, and provide them as an alternative in laryngeal-initial alternatives.. --Ivan Štambuk 17:03, 4 May 2008 (UTC)
I'm inclined to agree so far as we should use our own alphabet instead of EOR latin. But this our alphabet replacing the EOR alphabet, and for the rest using EOR rules? I'm specifically referring to putting diacritics/accents for vowels (not for ḱ nor ǵ), in a second level like EOR, so that e, é, ē, ḗ sorts as the same vowel e when compared to t, d, dʰ and such, and only in comparison to each other sorts in the order e, é, ē, ḗ? So
level 1 order p, b, , t, d, , , ǵ, ǵʰ, k, g, , , , gʷʰ, s, y, w, r, l, m, n, h₁, h₂, h₃, e, o;
level 2 order e, é, ē,
 ?? Is that OK? Rursus 17:39, 4 May 2008 (UTC)
Would that be self-explanatory? Wakuran 18:28, 4 May 2008 (UTC)
Yes of course, e, é, ē, should all be normalized as e, and that kind of alphabetization scheme is already applied to sorting the entries inside Category:Proto-Indo-European language (see Wiktionary:About Proto-Indo-European - that page needs heavy updating though). --Ivan Štambuk 18:52, 4 May 2008 (UTC)
Very good! Showing a democratic vote respect makes life easier. Approved: 11/2/3, means I shall start sorting according to the word order in Wiktionary:About Proto-Indo-European (p t ḱ k kʷ b d ǵ g gʷ bʰ dʰ ǵʰ gʰ gʷʰ s h₁ h₂ h₃ m n w r l y a e o i u r̥ l̥ m̥ n̥) and normalize e é ē ḗ as e (and similar for a o u i r̥ l̥ m̥ n̥). Unless serious objections occur in short time, I'll start sort-editing the subsections so, which may occur simultaneously as the page is cut into suitable pieces as per above discussion without serious technical consequences, and simultaneously with a resorting of the List of Proto-Indo-European roots. Opinions? Rursus 19:12, 4 May 2008 (UTC)
No, not that order, the one by place of articulation (by columns, not rows, as I gave you above). i and u are the same thing as y and w, and so are the r, l, m, m as opposed to , , , ; syllabicity is to be ignored for alphabetization purposes as much as prosody (read the PIE guideline page more carefully).
I strongly suggest that we wait a while (at least a week) before doing any major rework of this page, to give opportunity to the editors who have significantly contributed but it but don't visit WT that often (mostly hang on WP). You are free to do implement your scheme on your userspace in the meantime. Similarly, I won't be removing any content. --Ivan Štambuk 19:49, 4 May 2008 (UTC)
Message acknowledged. We wait for some time, and in the mean time I read the article a little more carefully. Rursus 20:14, 4 May 2008 (UTC)

Lag Time in edits?[edit]

I have edited a few sections, but they don't seem to stick on the main page (or right away)? Thanks Azalea pomp 17:46, 6 October 2008 (UTC)

The usual trick to purge the server-side cache is clicking on "edit" tab, and than saving without touching anything in the article. It won't show in the article history. --Ivan Štambuk 17:54, 6 October 2008 (UTC)


Isn't someone watching the sub-pages? There's an anonymous editor changing the format. I think that should at least be discussed before it's implemented. Wakuran 11:37, 20 October 2008 (UTC)

Or, what I mean is, since this format is explained at the page, it shouldn't be changed without some discussion/consensus. Wakuran 11:56, 20 October 2008 (UTC)

Russian e[edit]

I guess the decision was taken not to transliterate the Russian "y" or "i" sound before "e". "pelena", not "pyelyena". Maybe it makes it harder to see the links with the root. tony 14:44, 17 October 2009 (UTC)

We transliterate the "y" in the case of я ё ю (ja jo ju) because а о у are also common. In the case of е, it is almost the only choice, and native Russian words with э are limited to the forms of этот. е can be pronounced hard e in some few words, but these are imported words. Since е is almost always either ye or has the effect of palatalizing a preceding consonant, we can use the English e as though it were Russian е and drop the y’s. Otherwise, y’s would be everywhere and all over the place, and make the transliterations difficult to read. We still write a y word-initially and after a hard sign or soft sign: этот = étot, еда = jedá. The only places where this creates confusion is when е is pronounced hard, as in some words borrowed from English or French. —Stephen 01:32, 13 November 2009 (UTC)

Albanian roots[edit]

I will be looking through the Albanian roots again. I have noticed that Pokorny, Demiraj, Orel, etc. do not always agree on an etymology. Any suggestions? Azaleapomp2 06:43, 12 November 2009 (UTC)

The number of proto-Indo-European roots should be much more fewer.[edit]

1/How could be same roots have different meanings.

2/How could be synonims exist with different roots.

3/Many semantic shift are very lunatic.

4/Many sound shift are extremly impossible.

5/Many roots are anachronist(door,bourgh,fort...)and could not exist in the language of steppe hordes of the late neolithic.

6/Many roots are well known Semitic loans(star,three,sun,six,seven,eight,home,tree,field,barley,corn,dher,goat,buck.....)

7/Many other roots are Kartvelian,NW Caucasian,NE Caucasian,Altaic,Uralic,pre Indo-European languages of Europe.

8/Many roots are shared lislakh,eurasiatic and nostratic roots.

9/Many supoosed roots are given examples of very few Indo-European branches and sometimes only 1 ie branche,or examples from only 1-2 branch with very unbelievable sound changes.

10/The sound change du=>er in the Armenian erku=2,if included will make many languages Indo-European,for example Semitic thwa(2)is by far closer to proto Indo-European two(2)than the Armenian erku(2).

Thanks for taking into account my suggestions and consequently reduce very drastically the number or Indo-European roots listed here.

Best Regards. Richard "lionheart" Van Lancaster of Bourbon.

The Indo-European roots are given here as they have been reconstructed in form and meaning by well respected linguists. What may seem "lunatic", "impossible", "anachronistic", "unbelievable" etc., to you, do not seem so to people who have studied Indo-European linguistics in detail. The same holds for roots you think are loans from Semitic or roots that are shared with other groups. The sound change dw- > erk- in Armenian looks astonishing on the face of it, but in fact it is a well-known sound change found in more words than just the word for "two". —Angr 17:02, 12 January 2010 (UTC)
For example in երկ- (erk-, to fear) from PIE *dwi- --Vahagn Petrosyan 17:57, 12 January 2010 (UTC)

PIE root from etymology on hone[edit]

Hone: Akin to Old Norse hein ( > archaic Danish hen, spelling before the writing reform of 1948: heen , Norwegian bokmål hein), from Proto-Germanic *hainō (“‘whetstone’”), from Proto-Indo-European *k̑ōi- (“‘to be pointed’”).

How comes the name for the beaver is put under the brown/shining category (with mistake in it) while it's so well attested[edit]

Hi. How comes the name for the beaver is put under the brown/shining category (with mistakes in it) while it's so well attested. Gaulish for instance is bebros not bibrax (easy to verify). In sanskrit, babhru can be added too. If it is thought to be derived originally from the meaning "brown", it's still overwhimingly (and the association between brown and beaver could have occured in the other sense) associated with "beaver". "Beaver" should be added beside brown/shining. At the moment, it's deceitful and kind of partial (I don't think there is a consensus).

Also I think this is a mistake :

Gaulish Artioni should be artos.

jb 09:18, 31 October 2010 (UTC)


This IP root with its descendants is missing here. Does anybody know why? Nero86 22:00, 14 November 2010 (UTC)

House in Albanian[edit]

I noticed that in the part where it says House (dom) it does not also list the Albanian dhomë, which does not mean house in it of itself but rather room.

h₃owis vs. h₂owis[edit]

Most of sources (e.g. Fortson, Lühr) list this stem meaning "sheep" as 'h₂owis'. Why do we list it as h₃owis?--Anixx (talk) 04:08, 23 July 2012 (UTC)

I don't know. Ringe (which a lot of our reconstructions are based on) also lists it with h₂. —CodeCat 12:48, 23 July 2012 (UTC)
I cannot correct this because the page is protected and my account is new. Can you do so?--Anixx (talk) 18:11, 23 July 2012 (UTC)

suHnús should be moved to sewh₃nús per Hertzenberg, 2010[edit]

I renamed the entry in the list but cannot move the article.--Anixx (talk) 04:46, 27 July 2012 (UTC)

I don't think we should rename anything on the basis of a single source. We can mention the alternative reconstruction in the entry, but the name should remain at the traditional reconstruction. —Angr 05:22, 27 July 2012 (UTC)
Ringe 2006 reconstructs *suHnús. I am a bit puzzled about the ablaut grades of this word though, because there is no e-grade in it even though it ought to be athematic (as all nouns in -us are). Based on the accent placement I would expect a hysterokinetic noun *suHnéws, genitive *suHnwés. I wonder why that is not the case. —CodeCat 10:34, 27 July 2012 (UTC)
Hertzenberg gives the following declension:
  • NOM. *séu̯h₃-nu-s
  • ACC. *séu̯h₃-nu-m
  • DAT. *su̯h₃-nóu̯
  • GEN. *su̯h₃-nóu̯-s
--Anixx (talk) 02:38, 28 July 2012 (UTC)
What's his evidence for the full grade, and what's his evidence for it being specifically h3? —Angr 08:11, 28 July 2012 (UTC)
I've never seen ablaut in an athematic noun with an o-grade like that before. What is the linguistic consensus on such an ablaut pattern? —CodeCat 10:41, 28 July 2012 (UTC)
Hertzenberg gives this word as an example of proterodynamic declension with a stem ending with -u --Anixx (talk) 05:14, 29 July 2012 (UTC)
Ok, but proterokinetic nouns have e-grade in the oblique cases, not o-grade. So the expected genitive would be *suh₃-néw-s. —CodeCat 10:23, 29 July 2012 (UTC)
According to Beekes proterodynamic allows both o and e in genitive.--Anixx1 (talk) 00:48, 9 October 2012 (UTC)

Suggested move of h₃okʷ- to h₃ekʷ-[edit]

I suggest to move this root that means "eye" from h₃okʷ- to h₃ekʷ- as in the most sources.--Anixx (talk) 04:29, 14 August 2012 (UTC)

There is a difference between a root and a noun stem, though. Is h₃ekʷ- the stem of a noun, or is it the root from which the noun stem h₃okʷ- is derived? —CodeCat 10:25, 14 August 2012 (UTC)
Both. The noun derived from the h₃ekʷ form and I never encountered with the o-grade of this root.--Anixx (talk) 10:40, 29 August 2012 (UTC)
Do you know what inflection type the noun had? If it's a root noun, then it's possible that the inflection was acrostatic, nominative *h₃ókʷs, genitive *h₃ékʷs. —CodeCat 10:52, 29 August 2012 (UTC)
How can anyone tell the difference between h₃okʷ- and h₃ekʷ- anyway? —Angr 23:58, 29 August 2012 (UTC)
In theory you can't I think, unless there is something I don't know. But derivation of stems from roots generally follows well-known patterns of ablaut, so the ablaut grade can usually be determined by figuring out how the word stem was derived. —CodeCat 01:19, 30 August 2012 (UTC)
What do the dictionaries say?--Anixx (talk) 16:34, 31 August 2012 (UTC)
The fact that 'eye' from 'see' is an agent noun does not determine its vocalism, I think. – On the other hand, Hittite šakuwa has *o vocalism and this is one of the few words where Non-Anatolian *h₃ does not match a laryngeal in Anatolian, but Hitt. š, Luwian t (Gillian R. Hart). This may lead to reconstructing h₃okʷ- for Non-Anatolian. Howeer, the *h₃ : š : t thing is so strange that it may not gain general acceptance. – As for "the dictionaries", "Nomina im Indogermanischen Lexikon" has both h₃ōkʷ- and h₃okʷ-, but it writes *h2a- and *h3o- for *h2e- and *h3e- anyway, so it does not help. --Zxly (talk) 03:39, 13 July 2014 (UTC)

RFDO discussion: August 2013–November 2014[edit]

Green check.svg

The following discussion has been moved from Wiktionary:Requests for deletion/Others (permalink).

This discussion is no longer live and is left here as an archive. Please do not modify this conversation, but feel free to discuss its conclusions.

I submitted this for RFC a few months ago in the hope that someone could improve these pages. The fact that nobody has done so makes me believe that these pages are beyond saving, and that it's not worth the effort to fix them all. This page is redundant to Category:Proto-Indo-European roots, the pages of which give a much better overview of these words. These lists also have no quality control whatsoever, so they are nothing more than long and hard-to-navigate lists of cognates. But probably the most pressing problem is that a substantial number of the "roots" listed are not roots at all but word stems or even fully inflected words. —CodeCat 23:45, 2 August 2013 (UTC)

  • Keep. Do we have a separate PIE appendix entry containing every root and the corresponding reflex on that list? No. So it is worth keeping until we do because it contains (potentially) valuable information. Besides, I can spot some pretty doubtful roots and nominals inside th Category:Proto-Indo-European roots as well (*sū-, *sap-, *ǵénu-, *perḱ-, *pisḱ- - and that's just from the third column of the category!). --Ivan Štambuk (talk) 03:36, 3 August 2013 (UTC)
  • Swan diving in for keep as well. Just because no one's come to fix it doesn't mean it's beyond saving. It's just that there seem to be so few PIE linguists around- if I knew more about what I was doing myself, I'd fix it. I just don't want to risk making a worse mess of it than it already is, although I can give it a go if someone wants me to... also, Ivan brings up a good point about it needing to stay until an index is complete. Polar Night (talk) 01:35, 11 September 2013 (UTC)

Why would anyone want to delete this? Seriously ...—This unsigned comment was added by (talk) at 01:19, 24 September 2013‎.

  • Obvious keep, per Ivan and on general principles. DCDuring TALK 04:09, 24 September 2013 (UTC)
Seconded. It has its flaws, but has been very useful to me so far. David Marjanović (talk) 08:17, 7 September 2014 (UTC)
Keep for now, redirect to the category once every/most form has an entry. This is actually the only Wiktionary page whose link I’ve run into in a non-Wikimedia website, and two anons came out of nowhere to support keeping, so clearly this is a very popular page among our readership. — Ungoliant (Falai) 04:27, 24 September 2013 (UTC)

Keep: I think it's usable as is and similar resources in print can be expensive (save for Calvert Watkins) -- 04:24, 10 January 2014 (UTC)

KEEP. I use the links by sound quite often in research. The other appendix list you mention doesn't branch to those pages. If you think the page needs editing more than deleting, then be bold and edit it.

KEEP. The simplest fix is to revert to the version that was vandalized in >> this edit. << If your browser cannot handle the detail in that page, try Chrome; Chrome loads that detail in 15 seconds on an ancient 2004 laptop running Windows XP. -- Rednblu (talk) 14:36, 10 May 2014 (UTC)
Kept. Renard Migrant (talk) 02:33, 21 November 2014 (UTC)

Naglfar from *neḱ-?[edit]

The entry for *neḱ- lists ON "naglfar" as a derivative. However, the Wiki page about that ship explains it as deriving from ON "nagl" nail (anatomical). This seems right, since it was a ship made of nails. So, if anything, it should go under *h₃nogʰ- (but is arguably too specialized to merit inclusion). Can anyone confirm this, or, alternatively, give a reference? 15:17, 10 June 2015 (UTC)anon

This page's history[edit]

The part of the history of this page that was transwikied from Wikipedia is incomplete. It only goes up to 7 April 2006 (UTC) and doesn't have any history from before this cut-and-paste move over a week before that date; those edits alone comprise some 2,898 revisions. Therefore, for the full edit history of this page, see the history on the English Wikipedia. Also, the history before March 2006 on the current Wiktionary page really doesn't need to be there. Graham87 (talk) 12:42, 7 August 2015/04:31, 8 August 2015 (UTC)

Additional columns[edit]

I'm considering splitting the lists of derivatives into a couple of columns for better readability. Perhaps using the following column division:

  1. Celtic & Italic
  2. Germanic
  3. Balto-Slavic
  4. Greek
  5. Indo-Iranian
  6. Anatolian
  7. other

although I guess grouping Greek/Phrygian/Macedonian/Armenian might be reasonable (but I'm not sure if there's any snappy title for it). Possibly Tocharian could also use a column of its own.

Here is an example:

Root Meaning Italic, Celtic Germanic Balto-Slavic Greek Indo-Iranian Anatolian others
*tréyes; *tri- three Gaulish treis
Old Irish trí
Welsh tri

Latin trēs
Umbrian 𐌕𐌓𐌉𐌚 (trif)
Oscan 𐌕𐌓𐌝𐌔 (trís)
Gothic 𐌸𐍂𐌴𐌹𐍃 (þreis)
Old High German drī
German drei
Old English þrēo
English three
Old Norse þrír
Lithuanian trỹs
Latvian trīs
Old Prussian tri

Old Church Slavonic трьѥ (trĭje)
Polish trzy
Russian три (tri)
Ancient Greek τρεῖς (treîs) Avestan 𐬚𐬭𐬌 (θri)
Ossetian ӕртӕ (ærtæ)
Old Persian [script needed] (çi)
Persian [script needed] (se)

Sanskrit त्रि (tri)
Kashmiri [script needed] (tre)
Kamviri [script needed] (tre)
Hittite [script needed] (tri-)
Lycian [script needed] (trei)
Old Armenian երեք (erekʿ)

Albanian tre, tri

Illyrian Lua error in Module:links at line 70: The specified language Illyrian is unattested, while the given word is not marked with '*' to indicate that it is reconstructed

Tocharian A tre
Tocharian B trai

Phrygian [script needed] (thri-)

Thoughts? --Tropylium (talk) 00:51, 13 September 2015 (UTC)

My thoughts are that these lists duplicate the main content in Wiktionary appendices, are often outdated and should be deleted. PS. The grouping of Greek/Phrygian/Macedonian/Armenian is not universally accepted. --Vahag (talk) 12:30, 13 September 2015 (UTC)
Neither is Italo-Celtic; I'm considering them mainly as geographic units, since dedicating a single column to just Greek would not be a very efficient use of screen space. --Tropylium (talk) 13:53, 13 September 2015 (UTC)
You will have to dedicate a separate column to Tocharian anyway. --Vahag (talk) 14:47, 13 September 2015 (UTC)
Do we? Tocharian is both relatively flimsily attested and not considered especially important for PIE reconstruction. It would actually fit well together with Albanian etc. among the Restsprachen. --Tropylium (talk) 17:25, 13 September 2015 (UTC)
OK. But I still think these lists are unnecessary. --Vahag (talk) 18:00, 13 September 2015 (UTC)

RFC discussion: April 2013–August 2017[edit]

TK archive icon.svg

The following discussion has been moved from Wiktionary:Requests for cleanup (permalink).

This discussion is no longer live and is left here as an archive. Please do not modify this conversation, but feel free to discuss its conclusions.

This page and its subpages don't contain just roots, but a rather unorganised mix of roots, stems and even full words. It also seems rather redundant to Category:Proto-Indo-European roots. And the idea of listing all descendants on those pages seems doomed from the start, because there will be far more descendants than we could comfortably list in a table in many circumstances. I would prefer deleting this page outright but maybe someone else knows what to do with it. —CodeCat 18:07, 10 April 2013 (UTC)