Wiktionary talk:Proto-Indo-European entry guidelines

This creative page needs to specify that all these entries need to start with "Appendix:Proto-Indo-European ". --Connel MacKenzie 20:47, 18 December 2006 (UTC)Reply

Looking at the sources being cited, perhaps we need a blanket prohibition against all this recent research. (Recent, meaning not before 1923.) Absolutely none of this can be free of copyright? --Connel MacKenzie 20:51, 18 December 2006 (UTC)Reply

Started vote for wholesale removal instead, at Wiktionary:Votes/pl-2006-12/PIE. --Connel MacKenzie 21:22, 18 December 2006 (UTC)Reply

Some of the English meanings are direct copies from Pokorny 1959. Compare:

*peḱ, pēḱ-, pōḱ- "to make pretty; to be joyful"

with

all of the other info has to be either copyvio or original research. (IMNSHO) Robert Ullmann 21:28, 18 December 2006 (UTC)Reply

There was research during the 1800's, but it seems to be considered outdated for various reasons by the modern proponents; even Pokorny is apparently outdated. It is a bit disturbing that this page advocates taking entries from it: "One main source for PIE entries will be Pokorny's 1959 dictionary, ..." which would be clear copyio. Robert Ullmann 21:40, 18 December 2006 (UTC)Reply

by your logic, all material is either copyvio or original research, there is no middle ground left. Dbachmann 16:14, 22 December 2006 (UTC)Reply

Correct...for all P-I-E material. That is another reason there should be no P-I-E information anywhere on en.wiktionary.org. --Connel MacKenzie 03:12, 28 December 2006 (UTC)Reply

I've removed a suggestion to use Pokorny "as a main source". Most of it reconstructions are obsolete anyway. --Ivan Štambuk 16:50, 13 January 2008 (UTC)Reply

To clarify, using knowledge contained in a copyrighted work, without copying part of the work, is never an infringement. 68.173.113.106 05:28, 10 December 2012 (UTC)Reply

Kept. See archived discussion. 09:32, 19 January 2008 (UTC)

As has been suggested here - This would leave only single-language descending families listed, and would miss the whole point of the ===Descendants==== section: to list all the ancient and modern descendants of a particular proto-word (or a root/word family), including borrowings. This would enable the creation of a giant, unparalleled semantic network that would ease the memorization of both etymologies and the meanings of words. If the descendants section grows too big (which I doubt, but which would be awesome if we get to that point!), collapsibility of the families at the first level of indentation could be implemented in javascript to ease the browsing. If the duplication by itself is the problem, it should be solved at the lower levels of inheritance, with entries of proto-languages using either labeled section transclusion to selectively map the content of the master descendant section to that of inheriting proto-language, or by plain periodic bot-synchronization. --Ivan Štambuk 13:26, 16 April 2011 (UTC)Reply

I think redirects should be allowed for PIE, because there are common spellings of certain known PIE roots, such as *ekwos or *kmtom. This would also make it easier for people to find the actual PIE root, because not everyone knows how to spell h₁éḱwos or ḱm̥tóm on a keyboard. EliasAlucard / Discussion 04:54, 8 May 2013 (UTC)Reply

As far as I know, such redirects are already allowed, and there are quite a few of them. —CodeCat 12:47, 8 May 2013 (UTC)Reply

I don't agree with splitting of stems and derived terms on separate pages. For example on Reconstruction:Proto-Indo-European/dʰeh₁- we have four derivations, each of which is on a separate page. What is the point of ====Descendants==== section there anyway? To list unsorted terms that need to moved to the derived terms pages? It should also be noted that for most of the roots such derivations generally come with a big question mark because they are often postulated on the basis of very little evidence (e.g. reflexes in a single branch, usually Sanskrit or Greek which preserve PIE verbal system the best). This layout makes it very hard to navigate and observe the evidence in the daughter langauges.

My preferred layout would have the root listed under the ===Root=== section, and have all of the derivations listed either as as subsections. (e.g. for *dʰeh₁- we'd have ===Root=== and then ====*dʰédʰeh₁-==== or ====Present====), or just in a one big two-column table with derivations in the first column, and reflexes (which generally won't be many, a dozen tops) in another. Inflection would then come at the very end, after the evidence for particular stems has been presented. --Ivan Štambuk (talk) 01:33, 6 August 2013 (UTC)Reply

I have wondered about this as well. Most descendants don't clearly preserve the original primary verb inflection. Germanic and Slavic for example have almost only thematic verbs left. I think your solution may work well; alternatively we can list the verbs on the same page but under a ===Verb=== section. But we should keep secondary formations distinct as these were still clearly derived verbs. —CodeCat 01:43, 6 August 2013 (UTC)Reply

Over the past year, I have encountered quite a few words that say they derive from PIE roots that contain either a turned e (ǝ) or a schwa (ə); for example, Special:WhatLinksHere/Appendix:Proto-Indo-European/wenǝ-, Special:WhatLinksHere/Appendix:Proto-Indo-European/wenə-. (I have AWB set to alert me to turned es because they are frequently misused in IPA transcriptions.) PIE mentions schwas, but this page's list of canonical symbols does not. Should it? If not, how should the pages that currently link to wenǝ-/wenə- be changed? - -sche (discuss) 02:53, 17 February 2014 (UTC)Reply

Schwas are an alternative notation for laryngeals that appear between consonants and therefore become "vocalised" in many descendant languages. So there are really three schwas to match the three types of laryngeal. The notation is a bit outdated though, and most modern sources just write the laryngeal. That's what our practice page reflects. —CodeCat 03:11, 17 February 2014 (UTC)Reply

The pages that currently use wenǝ- or wenə- should be changed to use wenh₁- if fine is correct. —Aɴɢʀ (talk) 08:24, 17 February 2014 (UTC)Reply

In Wiktionary:About Latin and About Ancient Greek, I suggested creating categories for present tense forms created by nasal infix. Responses to my suggestion lead me to think that this suggestion should be handled here. Here's a proposal. How about there be a PIE root template that allows for showing the particular affixed, ablauted, or otherwise modified form of a PIE root that gives rise to a given Indo-European form? For instance, in the Etymology section of λαμβάνω (lambánō), this description could be added: the present stem λαμβάν- originates from *lh₂⟨n⟩gʷn̥-: zero-grade of *sleh₂gʷ- without s-mobile, with nasal infix *n, and with suffix *n̥. (I'm just guessing αν is n̥.)

In this case, the template would need to be able to remove the s-mobile, insert the nasal infix, remove the *e of ablaut, and add the suffix n̥. In other cases, it would need to replace the default *e of ablaut with *o or add a different suffix. Perhaps it could even include accent (although this might only be relevant in Ancient Greek, Germanic forms with Verner's Law operating, etc.). It would ideally also add categories based on the morphological and phonological changes (something like Latin forms with PIE nasal infix, Ancient Greek forms from PIE zero-grade, Latin forms with PIE s-mobile.).

Is this a crazy idea? Unfortunately I'm not that much of a PIE scholar or a template coder, so I cannot judge for sure whether this makes sense in either sphere, but I am interested in having words categorized by their origin from PIE forms with PIE morphological process x or y, and in having more detailed PIE etymologies, at least for a few words. — Eru·tuon 17:49, 5 January 2015 (UTC)Reply

Nobody has responded, so a simpler idea: a template with a bunch of morphological categories, like nasal infix, reduplicated present or perfect, and ablaut grade, that you could put before the root, and which would automatically categorize words as being derived from a particular category of PIE morphological form.

Let's say it was called {{ine-pro-morph}}. So, in δίδωμι (dídōmi):

From {{etyl|ine-pro|grc}} {{m|ine-pro||*de-deh₃-}}, {{ine-pro-morph|grc|redupl|pres}} of {{m|ine-pro|*deh₃-||to give}}.

resulting in

From Proto-Indo-European *de-deh₃-, reduplicated present of *deh₃- (“to give”).

and categorizing the word in Category:Ancient Greek reduplicated presents or something like that.

Any thoughts? I didn't get a response last time, probably because my idea was too fantastical. — Eru·tuon 07:24, 8 March 2016 (UTC)Reply

I think it's fine to categorize Category:Ancient Greek reduplicated presents and the like, but not all Ancient Greek reduplicated (nasal-infix, ske/sko, etc.) presents necessarily come from PIE reduplicated (nasal-infix, ske/sko, etc.) presents. —Aɴɢʀ (talk) 12:50, 8 March 2016 (UTC)Reply

Interesting. Meaning some of these morphological processes were still productive after the PIE period, perhaps in Proto-Hellenic? Is there a way to determine whether the forms originate in PIE or not? — Eru·tuon 19:54, 8 March 2016 (UTC)Reply

Not really. At best, it's unlikely that a form is of PIE date if it's only found in one branch. But if a form is found in many branches, it may mean that it came from PIE, but it may also mean that the derivation process was still productive in the individual branches and they innovated the same form in parallel. It's very hard to tell these cases apart. —CodeCat 20:05, 8 March 2016 (UTC)Reply

Okay, so I shouldn't keep assuming that Greek forms have PIE etyma... oh well. Probably have to change a few entries then. — Eru·tuon 20:32, 8 March 2016 (UTC)Reply

If a certain present stem is attested in two branches, it's probably old; if it's attested in three or more, it almost certainly is. (I don't think anyone seriously doubts that *yewg- had a nasal infix present in PIE.) —Aɴɢʀ (talk) 13:32, 9 March 2016 (UTC)Reply

We currently have a lemma form for nouns and adjectives, but not for verbs. The reason is presumably that there is a wide variety of lemma forms in the PIE languages, so it's not possible to choose one that neatly reflects them all, except by using the basic stem. However, I am finding that omitting the verb ending leaves out important details that are not supplied by the stem alone, namely whether the verb took primary (present), secondary (default) or stative endings. PIE also had deponent verbs, and this would be shown with middle endings in the lemma. These details are often reconstructible, and I think they should be included. Note that this question applies primarily to derived terms listed under roots. We only have a few actual PIE verb entries anyway.

The choice of lemma form is a bit more difficult, so I'd like to ask others for their opinions here. We should choose something that is used by PIE and at least some of the IE languages, so non-finite forms like infinitives are automatically out. That leaves only two contenders: 1st and 3rd person singular.

For the 1st person singular:

1. Endings:
   • Active primary athematic *-mi, thematic *-oh₂.
   • Actuve secondary athematic *-m, thematic *-om.
   • Middle primary athematic *-h₂or (Sihler)/*-h₂er (Ringe, Fortson), thematic *-oh₂or/*-oh₂er.
   • Middle secondary athematic *-h₂o (Sihler)/*-h₂e (Ringe, Fortson), thematic *-oh₂o/*-oh₂e.
   • Stative *-h₂e
2. Used by two big ancient IE languages, Latin and Greek, as well as by modern Greek. Also used by Bulgarian.
3. Clearly distinguishes thematic from athematic verbs, at least in the active primary endings.
4. Consensus on what the final vowel in the middle ending is?

For the 3rd person singular:

1. Endings:
   • Active primary athematic *-ti, thematic *-eti.
   • Active secondary athematic *-t, thematic *-et.
   • Middle primary athematic *-(t)or, thematic *-etor.
   • Middle secondary athematic *-(t)o, thematic *-eto.
   • Stative *-e
2. Used by two big ancient IE languages, Sanskrit and Avestan, though both are in the same branch and are a bit less well known to English Wiktionary editors. Also used by Old Irish, which is slightly more obscure, and by modern Macedonian.
3. The most frequently used form of a verb, generally, and the only one that is present for even defective/impersonal verbs.
4. Probably less likely to result in funny alternations like Szemerenyi's law, which might obscure the stem.
5. If we decide to list the other alternant of athematic verbs, the 3rd person plural is an ideal candidate as it is well reconstructed. The 1st person plural is somewhat more murky terrain, at least to me.
6. Should we use the middle endings with t or without? The t-less forms are very rare among IE languages, attested only in a few Sanskrit, Anatolian, Celtic and Tocharian verbs, but they appear to be the early PIE endings, replaced with forms with t within PIE itself already.

Which one do you think we should use?

—CodeCat 21:10, 1 April 2016 (UTC)Reply

I doubt that these forms actually existed as such in true PIE, due to some recent articles by Roland Pooth. However, I haven't read much about PIE and this is probably a minority view at the moment. Despite that I'm an Ancient Greek enthusiast, if a form has to be chosen, I'd support the third person singular, since reconstructions of that form seem to be less uncertain, and the active ones all contain t, which makes it more recognizable. — Eru·tuon 21:32, 1 April 2016 (UTC)Reply

My preference is leaning towards the 3rd person forms as well. —CodeCat 15:05, 3 April 2016 (UTC)Reply

Passer-by from beer parlour here. I recently thought about this and I'm wondering: Why do we/dictionaries in general see the need to have the verb class reflected by the lemma for PGM/Latin but not any descending Germanic/Romanic languages? Korn (talk) 11:31, 14 April 2016 (UTC)Reply

The Proto-Germanic lemma is the infinitive, the same as for all the descendants. So any information it gives about verb class is coincidental. And the Romance infinitive does reflect verb class in many languages. The Proto-Slavic infinitive doesn't reflect verb class entirely accurately, as Slavic verbs have at least two principal parts (present and infinitive). —CodeCat 15:43, 14 April 2016 (UTC)Reply

The ordering of the various formations in root entries is currently a bit haphazard, so I propose the following standard, based in part on what is already common practice:

• Verb formations first, then nominals and other formations.
• Derivations are listed in alphabetical order by the suffix. The grade of the root and suffixes are ignored (i.e. assume zero grade for ordering). Simple thematic nominal derivations are ordered as "o". Thus, *sth₂-tós comes between *stéh₂-tis ~ *sth₂-téy-s and *stéh₂-tus ~ *sth₂-téw-s.
• Root verbs are listed before all other verbs, root nominals before all other nominals.
• More derived formations after basic formations. Thus, a verb derived from a noun comes after that noun.

What do you think? —CodeCat 20:39, 3 May 2016 (UTC)Reply

The Origins of Proto-Indo-European: The Caucasian Substrate Hypothesis (revised October 2017)

I have saved this public list of PIE roots here: http://archive.is/WCZyL

--80.187.105.82 15:00, 21 March 2018 (UTC)Reply

If we're gonna keep listing Anatolian descendants on PIE pages, the following paragraph on WT:AINE on noun gender is unacceptable as is:

The template {{ine-noun}} takes one positional parameter: gender. Wiktionary uses Post-Anatolian stage of Proto-Indo-European as is usual, which had masculine, feminine and neuter genders which can be specified as m, f, n respectively.

Not only is the "Wiktionary uses Post-Anatolian stage of Proto-Indo-European as is usual" contradicted by us listing Anatolian descendants in the first place (hence any PIE form with an Anatolian descendant must be at a pre-Anatolian stage, by definition), the forced 3-gender system leads to patent nonsense like *nókʷts, *-os, *dʰéǵʰōm etc. being listed with masculine and feminine genders despite having Anatolian descendants, falsely implying that PIE masculine/feminine nouns descended into Anatolian common-gender nouns, when in fact it was PIE common/animate-gender nouns that descended into Anatolian.

I hence propose we simply delete this entire paragraph, and actually write accurate and updated gender documentation on {{ine-noun}} (which currently has no documentation) instead.

I do not propose removing masculine and feminine genders; but instead my preference is for common/animate gender to be added as a fourth available gender option, to be used when an Anatolian descendant of a nominal is attested. — Ceso femmuin mbolgaig mbung, mellohi! (投稿) 02:31, 11 February 2025 (UTC)Reply

Pinging @Caoimhin ceallach @Exarchus @Mahagaja @Ganjabarah @ElkandAcquerne @Djkcel @RwiTexx @Vahagn Petrosyan @Dragonoid76 @Svartava @Agamemenon for comment. — Ceso femmuin mbolgaig mbung, mellohi! (投稿) 02:37, 11 February 2025 (UTC)Reply

I don't have that much knowledge regarding the evolution of the Proto-Indo-European gender system, but from what I know, I believe that the inanimate noun class became the neuter gender and the animate class split into the masculine and feminine genders. If this is true, I propose that we should use an animate/inanimate distinction, similar to what @Mellohi! is saying, while also keeping the three-gender system. For example, the entry for pékʷtōr should have "animate -> masculine" instead of just "masculine". RwiTexx (talk) 04:10, 11 February 2025 (UTC)Reply

Like many other issues around our representation of PIE, this of course comes from treating a very wide timespan with poorly defined boundaries as one single, synchronically analyzable, internally consistent language. It's not an unsolvable problem, since for any problematic entry we can just put two or more heads, two or more inflection tables and so on, which we already do in some cases and not just for PIE. I support having "Proto-Indo-Anatolian" and "Proto–Nuclear Indo-European" (or similar) as dialect labels for Proto-Indo-European, and heads on separate lines when absolutely necessary. Two starting examples that come to mind for this sort of primary division are *h₁éḱwos and *h₂éwh₂os, which for the "PIA" stage one assumes athematic *h₁éḱus and *h₂éwh₂s respectively, since that is how they are reflected in Anatolian and everyone agrees thematicization was already a trend within PIE as far back as we can directly reconstruct.

For the sake of thoroughness, I agree that removing gender is a bad idea since it's standard practice to assume genders for PIE nouns at least for late PIE or "PNIE", and citing at least a best guess about gender (e.g. *h₂ep- was a feminine noun) is an important part of the reconstruction. I would be very surprised if anyone advocates for abolishing the m/f genders completely. Moreover most reconstructions don't involve Anatolian, and so they're reconstructible to a stage when masculine and feminine had emerged or were emerging (cf. Tocharian where it is fully developed but the suffixes are ever so slightly different); for example, it would be absurd not to mention that *h₂ékʷeh₂ is feminine, or that *-eh₂ and *-ih₂ had feminizing function in the non-Anatolian branches. I don't know if we're quite at the point where there's even widespread agreement among Indo-Europeanists about the origin of the tri-gender system post–Anatolian splitoff (and the other presumptions required for that), but it's an emerging consensus in a comparable way to how certain Leiden school theories (Beekes–Kortlandt glottalic+laryngeal model, origin of Balto-Slavic accent, accent–ablaut paradigmatic evolution) might be regarded as at the core of the academic buzz of recent decades. Consensus in IE studies is generally slow. That said, I'm all for Wiktionary not lagging behind and demonstrating that we understand the current state of thinking.
(The extreme alternative would be to split PIE into two languages, which would be ideal if we knew about the relevant chronologies beyond a shadow of a doubt. I'm sure PIA and PNIE were separated by over a millennium, and to draw a box around everything considered PIE would probably span more like 3 if not 4 millennia, so we are certainly not dealing with "the same language" through and through, but of course we're not ready for something so radical and separating the two could be a practically impossible nightmare. Come back in a few decades and maybe.) — Ganjabarah (talk) 04:19, 11 February 2025 (UTC)Reply

@RwiTexx I can get behind listing both the (if a noun or suffix is attested in Anatolian) pre-Anatolian (animate/inanimate) and post-Anatolian (m/f/n) genders in a headword. — Ceso femmuin mbolgaig mbung, mellohi! (投稿) 04:39, 11 February 2025 (UTC)Reply

Support. This is a good idea and I've had basically the same idea in mind for years. Shouldn't be hard to implement. See my above comment for cases where we might need two noun heads, one with common gender and one not. — Ganjabarah (talk) 04:19, 11 February 2025 (UTC)Reply

Support, I'm always in favor of cutting down on misleading reconstructions. DJ K-Çel (contribs ~ talk) 14:32, 11 February 2025 (UTC)Reply