Talk:List of Proto-Indo-European roots

This very imperfect listing was compiled long ago exclusively from Pokorny and Vasmer. If you have questions, please consult the online versions of their venerable dictionaries. Corrections are more than welcome, but please don't flood the list with repetitions found in sister languages and unverified entries!!!!! --Ghirlandajo 19:02, 22 July 2005 (UTC)

I think the examples given are good, but difficult to follow, I'd prefer a table with the columns giving the different IE subgroups (*Reconstructed PIE, Anatolian, Indo-Iranian, Greek/Hellenic, Italic, Celtic, Germanic, Armenian, Tocharian, Balto-Slavic (Slavic, Baltic), Albanian) instead.

Language codes

Someone might want to go through this and make the language codes a bit more user-friendly.

Two reasons:

• It's irritating to have to always scroll up to the top to look up some of the less intuitive codes (I = Avestan, B = Lithuanian, W = Gothic)

• Some of these may have confused the compiler(s). Notably, B = Lithuanian and L = Latin. Not knowing a whit of Lithuanian I cannot say for sure, but "sienas" looks to me more like Lithuanian than Latin. I know it's not Latin. It would be easier if Lithuanian were "Lith" and Latin were "Lat." --Carolus 19:17, 14 August 2005 (UTC)

Looks like you're right. I looked up the root, and šiẽnas is Lithuanian. Doesn't look like there's a Latin equivelant. - Dysfunktion 10:18, 24 October 2005 (UTC)

It seems to me that B stands for baltic not for lithuania and it also seems to me that I`m not only one who has noticed that and some latvian words are marked with B ( I might be wrong since I don`t know lituanian) so maybe it would be better to keep using B for abbreviation but ad (lv) or (lt) next to B. Xil 20:16, 24 September 2005 (UTC)

Clear-up?

I think many of these words are disputed, but it's hard to tell as it is, now, when it seems that all the words mean exactly the same thing in all languages, when that is rarely the case...

I just came to think of that when *sem-gheslo was removed, while I think that there are many remaining roots just as debatable. Also, some explanations such as akwa- (river), wedor (water), paewr- (bonfire), egni- (fire), while I read that "akwa" was "animate," referring to water as a living force and "wedor" as an inanimate substance. Likewise for "egni"(animate) and paewr(inanimate). (This makes sense, since it would be very important for a nomadic people to differ between controllable and uncontrollable water/fire, I think. Being in control of nature is very useful, not being in control is possibly dangerous...)

Focus?

This article will eventually morph into a PIE Lexicon, along the lines of IEW. That should go on wikibooks, I don't see how this can be a valid encyclopedia article. dab (ᛏ) 11:51, 23 August 2005 (UTC)

Suggested additions

• I wanted to suggest this addition that I have noticed over the years. Can somebody who knows this stuff please examine this and see if/how it should be added?: sammeln (German), assemble, sum (English), sammelan (Hindi) thanks -- Deego

"sammeln" goes back to *samo(same), Cf Greek homo(same, unrelatd to Latin Homo, human) and Latin Similis related to *sem(one), alternative word to *oinos, assemble comes from Latin ad(towards, related to En at, from PIE *ad), and simul, so it seems indirectly related, anyway, there are many etymological dictionaries which you could check out, I'd believe "sammelan" could be from the same root, as I know that the Sanskrit words "sanskrit" and "samband" (connection) are. Sum is from a completely different root *(s)up(er)-mos, related to En "over". Anyway, many of these words seem to share a common root, but they are independent creation, so they would not fit this list.

Although this list doesn't mention that root, in spite of it being very common, maybe it could be placed with "numbers", "adjectives" or "wholeness".

http://www.bartleby.com/61/roots/IE451.html

http://www.etymonline.com/index.php?search=same&searchmode=none

What does "sammelan" mean? I looked it up, and it seams it would mean a conference or gathering? (As one, together, the same(?)) I'd say it clearly looks like it's indirectly related, but I don't know much about Indo-Iranian etymology. Sorry that I mixed up the Sanskrit word earlier, it's "sambandha" (http://srath.com/lectures/sambandha.htm Sambandha defined: The word sambandha in Sanskrit means binding or joining, a close connection or union or association, conjunction, inherence, connection with or relation to. It connotes personal connection like a relationship, fellowship, friendship and even intimacy. The word ‘sambandhi’ (or as used ‘samdhi’ refers to a kinsman, a relationship brought about by a marriage or family by birth.), people often give this example of Sanskrit when they talk in Swedish about the relations of all Indo-European languages, so I gave the Swedish word (connection, conjunction etc.), by mistake.

Check out etymology before adding.

One user added several false examples, mistaking the German prefix ge- for part of the root stem. Since close appearances could deceive, I advice you to be careful to post words that you are sure are correct next time. If the article turns out any good, maybe it should be copyedited, since I believe it is likely that it will often turn worse by edits of people unfamiliar with the comparative method and basic etymology.

Is similarity and meaning important ?

Browsing trough those words I noticed that some words in IE are close to latvian, but means another thing althought those meanings are connected i.e. IE k`ak 'branch' seems to be close to latvian koks 'tree' so I think that another person maybe would ad 'koks' next to 'k`ak' and everyone would think that it means branch and I think that some people has done so in other languages i.e preu- (to jump): R prygat', S pravate, OHG frowen, G Freude, E frog does 'frog' realy means 'to jump' ? I thought that it is an animal. I also noticet that some IE werbs would be almost identical to latvian if latvian word would be writen in present singular second person - now you would write like in sentece "you are going to do something" but you would get most similar form if you were writing "you do" i.e. "you is going to laugh/go" would be "tu taisies smiet/iet" in latvian, but if you would write "you laugh/go" which is "tu smej/ej" in latvian it would be almost identical to IE 'smei' (to laugh)and 'ei' (to go) So what I want to sugest is that if one is writing word similar to IE, but with another meaning he should ad meaning next to this word and that the most similar form of word should be writen Xil 20:53, 24 September 2005 (UTC)

Yeah, I have a feeling that the semantic shift in the examples included could be quite far from the original meaning, and that many of these examples actually are quite disputable, for instance ei-s- for (ice): I isu, Oss ix, E ice, D eis >< *ei-n-: R iney, B ynis. (Meanings change from ice to frost to rime, although I guess that's semantically reasonable). Btw, the German word means "joy, happiness", so it is also far from the original meaning. (Cf. "Schadenfreude") Frog could have come from an original word like *frug-isks (Jump/Hop-ish/er). Semantic change is common, though, so it doesn't discredit the roots, per se. For instance the English word "see", is likely related to a root meaning "follow" in many other IE languages. Basically, close similarities between words' appearance and meaning in different languages are more common than what one might imagine, so genetic relationship is judged on exact sound correspondences for words with similar meaning. I think words with different meanings would be interesting, but actually they are probably so common, that it would make the list a lot harder to read. About your second point, the article mentions in the start that "all the words are shown in their most representative declinations", which basically means that "the most similar form" is used. If it's not declined already, I guess it could be changed. I have trouble understanding completely which grammatical tense in Latvian you are talking about.

Article needs some cleaning

There needs to be some order in the language cognates. The palatals needs to be marked in the PIE roots. Long vowels and other accent marks should be included for the various languages. Imperial78

As for your question whether Armenian "ov" is a *kwo cognate or not, yes it is. In Old Armenian, this was "kov"... Eiríkr Rauði 16:24, 19 October 2005 (UTC)

I'd gladly see SOME marking of the modern meaning of the words referred to. Also, maybe there should be some clean-up of some of the more widely disputed cognates... Also, some of the "original meanings" seem to be quite questionable, the root for bottom is placed among organs, though the root likely is referreing to the ground, and not the hindpart. Several things like that...

Table

I've created a program to generate a table from the existing code in the page. It's a little buggy, and there are a few typos in the code that screw up the table, but it does the trick: User:Dysfunktion/List_of_common_Indo_European_roots - Dysfunktion 22:03, 17 October 2005 (UTC)

I like the table much better. Use the latest version though since it has a lot of corrections and additions. Imperial78

Yeah, nice work! It would probably be simpler if it was arranged according to language groups, and not individual languages, (and maybe include spacing on a new line for several languages in the same group) but it's a nice start! 81.232.72.148 17:19, 22 October 2005 (UTC)

Also, I don't think the page should include one large table, as it does now, but rather several small tables for each category. Would facilitate editing. 81.232.72.148 17:33, 22 October 2005 (UTC)

Alright, I tried breaking it up into separate tables and grouping the languages by family. - Dysfunktion 07:50, 23 October 2005 (UTC)

Right, this looks better. Some things, I don't know if it's necessary to show the Greek derivations "arktos-arctic" and "mus-muscle", I wonder how *sweid is related to *glag, I think that the "See also:Indo-European copula" should be moved to somewhere else, and I think that maybe all sub-languages should be merged into their own group, abbreviated in paranthesis (It will make the tables shorter, and should still be easier to interpret than the original). But don't change anything else, until we hear what others are thinking. 85.226.122.194 12:34, 23 October 2005 (UTC)

I agree the derivations are not necessary and just confuse the table. Imperial78

I am now fixing the Armenian so it conforms to standard Eastern Armenian. Also fixing the problems with the Greek. I assume you will be making the table to the newest version of this page? The older table is riddled with errors and is not as complete. Imperial78

Yeah, I've been updating the table every now and then. Why did you change "ON meðal" to "OHG meðal", by the way? Dysfunktion 14:49, 29 October 2005 (UTC)

I didn't do that ON change, at least if I did it was on accident. Imperial78

Nice table so far, Ossetian is an Iranian language though, not Slavic. Also, the language in the table is Ancient Macedonian (not the modern Slavic Macedonian language) which should be in the other languages. Imperial78

Ah, the Cyrillic threw me off. Done. Dysfunktion 01:13, 31 October 2005 (UTC)

Here's the source code, if anyone's interested. I tried to comment it pretty well. http://xthost.info/z500/programming/tablegen.ex - Dysfunktion 21:36, 1 November 2005 (UTC)

Lithuanian is misspelled, Old Irish should not be a column. I think Persian would be better to use than Farsi, since Old Persian is also in the table column with Persian and Wikipedia has the article as Persian. Also the order should go for Indo-Iranian: Sanskrit, Kashmiri (Dardic), Kamviri (Nuristani), Avestan, Ossetian, Persian. Imperial78

Pel and pelt

Would pelt be a better example of an English word based on the root pel- than film? I wanted to check before I made any changes because I don't know whether this article is perhaps at a consensus point or if it would be considered bad to even have more than one example from a single language for a root. Theshibboleth 01:25, 19 October 2005 (UTC)

Pelt is a a borrowing from a Romance language. I'd prefer "film" or "fell"(from Bear fell), since they're native words. Fell is more common as a native word in G outside English, though. Sure, "film" as in "movie" is widespread, but that's a modern English borrowing.

I have some issues with the idea that some parts of English are borrowed and others not, but alright, I won't change the article. Theshibboleth 06:21, 27 October 2005 (UTC)

What do you mean "issues with the idea that some parts of English are borrowed and others not"? I'd prefer "native" words as far as possible... 81.232.72.148 17:59, 28 October 2005 (UTC)

Let's all do our part to put the languages in order

What kind of order would be best? I think alphabetical according to subfamily with the Old languages going before the modern ones and keeping languages of the same subfamily together. For example: Al, A, B, OCS, R, W, E, I, S, L, K Are we going to have Balto-Slavic as a unit or not? I think if everyone here works on a section, we can get the order finished quicker. Imperial78

I'd think it might be good if language groups that are more closely related (such as baltic-slavic, italic-celtic etc) would be place closer to each other, or having the language groups in chronological order, from the groups attested earliest, to the groups attested latest.

Although there is some consensus on Balto-Slavic, there isn't one on a close relationship between Celtic and Italic, so they are not needed to be placed close. Imperial78

Is there a consensus on Baltic-Slavic? ISn't that believed to be a sprachbund? 85.226.122.202 22:13, 24 October 2005 (UTC)

More linguists agree on a Balto-Slavic subgroup than an Italo-Celtic one, but I wouldn't object to Baltic and Slavic not listed together. Imperial78

Transliteration for non Latin script languages

Here is a scholary standard which we should use so there are not so many different systems: Greek Alphabet: a, b, g, d, e, z, ē, th, i, k, l, m, n, ks, o, p, rh/r, s, t, u, ph, kh, ps, ō Russian Alphabet: a, b, v, g, d, e, ë, ž, z, i, j, k, l, m, n, o, p, r, s, t, u, f, x, c, č, š, šč, ", y, ', è, ju, ja Imperial78

Languages

I was thinking of a good list of languagess to use for the list. Albanian; Anatolian: Hittite, Lycian; Armenian; Baltic: Old Prussian, Lithuanian; Celtic: Gaulish, Irish (including Old Irish), Welsh; Germanic: Gothic, Old English, English; Greek (including Old Greek); Indo-Iranian: Mitanni Indo-Aryan, Sanskrit, Avestan, Kamviri; Italic: Latin, Slavic: Old Church Slavonic, Russian; Tocharian B; Extinct language: Ancient Macedonian, Dacian, Illyrian Thracian, Phrygian, etc. Also, other languages when a cognate is not found in any of the languages listed above. Imperial78

I added a lot of Kamviri cognates, pulled from the Richard Strand website on Nuristani languages. Imperial78

We could include certain languages, like Old Norse, when there is no known cognate in related languages, (like ON "var" for "spring"). Why only Tocharian B, and not A?81.232.72.148 18:01, 28 October 2005 (UTC)

Well, we should include A and B. I guess we can label them KA and KB? I added the numerals. More to come. Imperial78

Other Languages and Irish/Old Irish

I think we should find modern Irish Gaelic terms and Old Irish. Old Irish and Irish are very close though in many terms. Also, what do you guys think about adding Latvian for Baltic, and Farsi for Indo-Iranian? Imperial78

Another Idea for the languages in the table

Perhaps some languages/dialects should share a column and be divided by a slash to save room on the table? Here is an idea: Albanian: Tosk/Gheg; Armenian: Eastern/Western; Tocharian: A/B; Baltic: Lithuanian/Latvian, Old Prussian; Slavic: Old Church Slavonic, Russian, Polish; Germanic: Old Norse, Old High German/German, Old English/English, Anatolian: Hittite, Lycian; Greek: Ancient Greek/Greek, Celtic: Old Irish/Irish, Welsh, Gaulish; Indo-Iranian: Sanskrit, Kashmiri, Kamviri, Avestan, Old Persian/Farsi, Ossetian Iron/Ossetian Digor; Italic: Latin, Oscan, Umbrian; Ancient Macedonian, Illyrian, Thracian, Dacian, etc. Imperial78

I proposed something similar earlier, grouping all languages of the same group in the same column. I think it'd make sense. I'd figure the last column would be "Other(s):" (Maybe many of these langs should have their own column, in that case, since they're not shown to be closely related.) 81.232.72.148 13:27, 31 October 2005 (UTC)

By the way, is Lycian a better choice than any other of the Anatolian languages? 85.226.122.222 06:38, 3 December 2005 (UTC)

phantom languages

OK, so I added some code to the table generator to pick out unknown language codes, and I got a whole bunch of them, aside from the unconverted Ks:

CC (might be a typo of C, but I don't know anything about Celtic langs), Sk, OG, Lt (Lithuanian?), OL, Osk, Thr, P, Nw, OS, Uk, ONG, Sw, OI, OS, Af, NI, OSw, Umb (Umbrian?), OG, Md, GM, Gl, OR, Sl, Sc, Krd, Blg (Bulgarian?), Prs, Sg - Dysfunktion 20:37, 31 October 2005 (UTC)

Sw=Swedish, OSw=Old Swedish, Prs=probably Persian/Farsi 81.232.72.148 21:52, 31 October 2005 (UTC)

Wow, there are quite a few phantom languages. I am going to now change all of the Tocharian to just T (not K, not KA, not KB). I think we should just delete all of the phantom languages. Also if a cognate does not exist or cannot be found I think we should mark it as "--" Imperial78

I don't know, I think that there should still be separate entries for Tocharian A and B. Dysfunktion 01:22, 1 November 2005 (UTC)

I think the table is getting too big and the ones on the actual page are getting rather cluttered. Each sub-branch should get its own column (prominence and relevance to what the table is showing), not an older form or a closely related language. Tocharian A and B are closer to each other than say Indic and Iranian. So I think one column would suffice. I think one column would be sufficient for Old English/English, one column for Old High German/German, and one column for Old Irish/Irish, etc. Since we have Avestan, I am not sure we need to include Ossetian, but perhaps Ossetian has cognates other Iranian languages do not have? Imperial78

Persian versus Farsi

Wikipedia's main article is called Persian, so we may as well use Persian in place of Farsi. Also, the Old Persian/Persian data isn't on the table yet.

Semantic Shift

As was discussed earlier, I think the semantic shift for several of the examples could be quite huge. Do you think we should include a translation of the meaning of the cognates, when it differs considerably from the alleged original reconstructed meaning? 85.226.122.165 14:52, 3 November 2005 (UTC)

I think listing every meaning would spam the table. There are many words where the meaning has changed slightly or a lot. If people want to know the meaning of the word, they can do their own research I think. Although it would be useful. For example PIE mei- (small yields meiu- "four" in Hittite! Imperial78

Not every word, but those where the meanings have diverged significantly... I don't know how large spaceit would take... It could be interesting... 81.232.72.148 01:16, 4 November 2005 (UTC)

English Egg is an Old Norse loan

The anon. user needs to read on his cognates before he makes changes. English egg is an Old Norse loan. The -ey in Cockney is the true cognate. Cockn (blend of cock and chicken) + ey (egg). This is in the IE roots section of Calvert Walkins in the AHC dictionary. So please do not include loans, so look up before you add or change. Imperial78

OK, I knew that egg was a borrowing from Old Norse, but I misunderstood the information... I thought that ææg was a word in the Cockney dialect, with a dialectal spelling (...) Sorry for that.. (At least the Norse word also comes from the same root, originally, I Wonder if German Ei would be a better word...) Alright, it's better now... 81.232.72.148 01:11, 4 November 2005 (UTC)

Body parts

Body parts section lists fuinneog as an Irish word for eye. As far as I know (and I admit I do not know a lot), fuinneog is Irish for window, not eye. The correct word is súil.--Ag Foghlaim 22:49, 4 November 2005 (UTC)

The word súil is not the cognate for PIE *okw-. The -og in fuinneog is a cognate. The table is of cognates, not just words for "eye" in the IE languages. Imperial78

Well, OK, I see the point, but isn't it misleading to have the word window listed as a body part? There is no indication telling the reader that only the "-og" part refers to the root, and not the whole word.--Ag Foghlaim 05:28, 5 November 2005 (UTC)

Well, that is the fun with semantic shifts, the original meaning of the root and the outcome many times diverges, sometimes greatly. So the destiny of a word in any given language is unknown. The division of the roots is just their PIE meanings. For example, English only keeps the root *dhg'hem in the word "bridegroom". So from the meaning "Earth" to "man getting married at the wedding ceremony" is quite a semantic shift. dhg'hem(PIE earth)/dhg'hm.-on-(PIE earthing) -> gumo:n (PGermanic man) -> guma (OE man) -> bridegroom (modern English). So we the authors are still working on the best way to combat the semantic shifts and the additions to the roots. I guess we could do () as in (fuinne)og. Although this can get messy, for example. Do we type Albanian nëntë as nën(të)? What would be better would to have each language have its own page about the the shift of the PIE roots in meaning and form. The table of roots just needs to be a table with roots listed with less spam junking it up. Imperial78

Btw, the ow in window does also come from the root for eye, and (bride)groom is cognate to Latin homo. 81.232.72.148 23:51, 6 November 2005 (UTC)

Well, window is an Old Norse loan, so it isn't a cognate but the -ow does mean eye, yes. Imperial78

Maybe not a true cognate, but at least from the same root. 81.232.72.148 15:21, 8 November 2005 (UTC)

Irish fuinneog, incidentally, is also an Old Norse loan, from the same word as window was borrowed from. --Angr/_{tɔk tə mi} 21:23, 11 November 2005 (UTC)

Thanks for the correction! I have found Old Irish enech/Modern Irish oineach and Welsh enep for *okw Imperial78

Well, if fuinneog is a loan-translation, then these words would probably be better choices than that word... 85.226.122.233 15:34, 13 November 2005 (UTC)

Important to have a language from each sub-branch

For some reason Ghirlandajo wants to delete the Polish data which is important for the table. Can you please stop it. Imperial78

Do we need Po, while we have OCS and Ru? Just wondering... 85.226.122.222 13:49, 18 November 2005 (UTC)

Yes we do, Slavic is divided into three sub-branches, East (Russian), South (OCS), and West (Polish). For purposes of comparison we need one from each branch just as we have in Germanic, Indo-Iranian, Baltic, etc. Imperial78

Obscure languages

I found a list containing some small dictionaries of many lesser known old IE languages. If someone's got the time, it would be nice if s/he could look through the lists to add cognates... http://www.wordgumbo.com/ie/cmp/index.htm 85.226.122.222 13:21, 19 November 2005 (UTC)

There's another good page at http://www.donbelid.com/MATN%20HATML%202/root.htm. The original seems gone, but it could be found at www.archive.org. 81.232.72.148 00:58, 22 November 2005 (UTC)

Mix-up of two different roots?

There seems to have been a mix-up of two different roots, and I'm not sure of which words belong to where...

• dhe- (to set/put):
• do- (to give):

S dadhati, I dadaiti, L facere, H dai-, P ada/, G tithenai/, B deti, P dziac', R delat', E do, W gadeths, ON dalidun, F edaes probably belongs to dhe- while R dat', Po dać, F dadón, G dōron, L dare probably belongs to do- I'm not sure about: A tal/dal, B duoti, Al dhashë, K dī... Help please! 85.226.122.222 14:22, 19 November 2005 (UTC)

Well, I think I have fixed it... 81.232.72.148 01:49, 22 November 2005 (UTC)

Another weird cognate

The example "D sarf/" from *kerp looks suspicious, since *k to *s is a very unlikely sound shift to German, I only know of one possible other example, which is "schliessen". I thought about removing it, but it allegedly is Old German, and I have no idea what it should mean, and where to look up the etymology...

PIE gWher

PIE gWher- yields burn in English, as gWhen- yields bane. Some sources say it yields warm. Calvert Watkins has burn, so I am going with this qualified authority.

Hmmm, if *gWh would yield b* in Germanic, it seems strange that it would have yielded "gunnr", since "bani" also existed in Old Norse (as well as "brenna" and "varmr"). 81.232.72.148 13:41, 26 November 2005 (UTC)

Well Watkins has the roots *gwhen- (Pokorny 2. gwhen- 491, bhen- 126) and *gwher- (Pokorny gwher- 493 bh(e)reu- 142), so perhaps the the gwh- root and the bh- root are from *gwh-.

What are you talking about? Two parallel roots? 81.232.72.148 13:41, 26 November 2005 (UTC)

There is only one root. Perhaps Pokorny thought they were two, I am not sure. Someone have Pokorny? If you look at Watkins, you will only see he has them listed with the *gWh- in the index.

So, which goes back to my original comment, it seems strange that "gunnr" would have come from *gWh, when Germanic *g regularly is believed to have come from *gh. 81.232.72.148 14:09, 27 November 2005 (UTC)

True, although gunnr is in the index of Watkins. Also, for *ghwer-, he has Old Norse brenna...so I am not sure what to make of all of this.

Well, it just looks strange. I'd see if someone could come up with a reasonable explanation. 81.232.72.148 19:29, 28 November 2005 (UTC)

*Bhel

American Heritage says that ball comes from a homonymous root meaning "swell", not white etc, there seems to be many words from that root (white) in English, though, including blue, black, blind, bleach (Maybe words based on bleach would look best, common Germanic cognate, afaik). 81.232.72.148 00:24, 1 December 2005 (UTC)

Thanks I fixed it.

Link removed

Removed the link http://www.espindle.org/roots.html, since it mostly was a list of Greek or Latin roots in English words, without any connection to PIE.

Albanian dhelpër not gjelbër is from PIE *g'hel

Please do not make changes if you do not know the material. Imperial78

Sum/Some

"Sum" is Old English for "Some", unrelated to modern English "sum". 81.232.72.53 22:41, 9 January 2006 (UTC)

Yes, sorry you are right! It is getting late. I am forgetting. lol Lately some people had been adding bad loans. And I am the one who had added sum! lol Imperial78

How would everyone feel about these additions

Most of the IE branches have an "old" and modern counterpart. How would you guys feel about a Sanskrit/Hindi...Latin/Italian...or is this just getting too many languages for the table? Also, if no cognate is found perhaps a cognate from a close language can be included with that language in () for example using the Kurish word from PIE elm in the Iranian category. Imperial78

I don't think so, the page and table is quite unwieldy already at the meantime, and I don't think it would be NPOV or practical to choose only one language from a mother language that has given birth to several daughter languages (Latin, Sanskrit, Old Norse etc). If Proto-Germanic, -Celtic, -Slavic etc. was attested, I would prefer these languages to all of these different sub-branches we have now. So, for attested mother languages with a huge lexicon, I don't think there should be daughter languages save if you're writing your own article in that language (as has been done in Spanish etc). 85.226.122.205 15:46, 13 January 2006 (UTC)

Btw, what do you mean with PIE elm? Yhe root wygh for "elm tree", or the root "elwn" for "cedar tree"? 85.226.122.205 16:09, 13 January 2006 (UTC)