Reconstruction talk:Proto-Indo-Iranian/garȷ́ʰ-

Root entries

For avoidance of any further edit wars, I'd like to point out that a discussion related to this topic happened here where it was not concluded that root entries may not be used. I'm of the opinion that it ultimately comes down to preference whether to use the verb form or the root form. For most cases I don't mind the verb forms, but I strongly support root entries in cases like this where there are multiple classes. Also, the sources do the same so it won't be unique to wiktionary. @Victar, AryamanA, Kutchkutch, RichardW57, Mahagaja, Metaknowledge, Rua. -- Bhagadatta (talk) 10:13, 12 September 2020 (UTC)

For both PIE and Sanskrit, we have separate entries for roots (e.g. Proto-Indo-European *bʰer-, Sanskrit भृ (bhṛ)) and for verbs (e.g. Proto-Indo-European *bʰéreti, Sanskrit भरति (bharati)), so I don't see why we can't have separate entries for Proto-Indo-Iranian *garȷ́ʰ- and *gárȷ́ʰati. —Mahāgaja · talk 11:10, 12 September 2020 (UTC)

@Mahagaja: I agree. The issue is it may in some cases cause duplication of content, when the primary/only derivative of the root is the thematic present (or "class 1" present). But in a root like this, there's a whole paradigm of forms united solely by the root. That's why root entries are required and should not be banned. -- Bhagadatta (talk) 13:46, 12 September 2020 (UTC)

@Victar, Bhagadatta: (also going to ping @JohnC5 just in case) So I do not really see a consensus on that talk page. I really don't think we're all on the same page when it comes to this so the edit war was kind of silly--we should agree on what we are doing first.

First of all, I want to point out that reconstruction in and of itself is a formalism meant to make sense of the wordforms in the descendant languages. That is the primary purpose of reconstruction. It is not meant to be a perfect representation of PII speaker knowledge (because it can't be), it's for comparative purposes.

Towards that end, root entries, whether or not PII speakers actually were aware of them, are useful for connecting disparate forms. Sure, you can analyze a verb form and say that the ending was removed and the nouns or other verb forms were derived from that, but this is (1) inconvenient for users of PII entries, (2) not done anywhere else in the scholarly tradition of PII, and (3) needlessly pedantic. Root entries centralize diachronic inherited terms with synchronic derivations, which the verb-centred approach makes no sense for.

Finally, the Sanskrit grammatical tradition is well aware of roots and we have root entries for Sanskrit. Why would we get rid of PII root entries and leave a gap between Sanskrit and PIE? It really doesn't make sense, and this is a differentiating factor from the other Indo-European branch reconstructions. So, I'd like to keep root entries in PII as a nexus. —Aryaman^{A (मुझसे बात करें • योगदान)} 17:16, 12 September 2020 (UTC)

First off, it should be noted that, beside the two root entries Bhagadatta just created, the last time anyone created a root entry was 2017. Since then, all PII editors have been creating verb entries, which outnumber root entries, all of which were created by Bhagadatta and AryamanA.

@Mahagaja, to respond to your reply, a big difference between PIE and PII is it's much easier to prove that a term existed in PII than it is in PIE, thanks to deminisiting productivity of suffixes, morphology, etc. That means that we're far more likely to have entries for terms, while we err on the side of caution in creating non-root PIE entries. It also means that we're going to be using the related terms sections, like we do in, say, Proto-Germanic. So in the end, root entries are just a list of bluelinks, which if we're going to just have a list of entries, this can be better done with a category page. That's why we created the template {{root}}. So in this example, we have Category:Proto-Indo-Iranian terms derived from the Proto-Indo-Iranian root *garȷ́ʰ-. It's not doing away with roots, but automatting it.

@AryamanA, to answer your points:

1. Inconvenient because it's somehow harder to understand? I very much contest that subposition. It works fine in Proto-Germanic where no one is clamouring for PG root entries. If anything, it gives a false idea to the reader that all these descendants were newly built in PII from this holy magic root.
2. Again, we'll still have root category pages, just doing away with the need to have manual descendants lists, and to echo a point I made on that discussion from 2017, unlike many dictionaries we also reconstruct nouns in nominative case, as opposed to a lemmatized form. I've said this elsewhere in other contexts, how other dictionaries do things is not sacrosanct -- we are an interactive online dictionary with no physical limitations, so it makes sense that we forge our own ways of doing things.
3. You'll have to explain what exactly is "needlessly pedantic" about it.

And @Bhagadatta, "avoiding edit wars" isn't starting a discussion and then continuing to edit war. Putting aside that this is a root entry, it's creation was sloppy. You added no etymology, you added bad redlinks like *gárždi ~ *gr̥ȷ́ánti, made all the PII form null-links for some unknown reason, misattributed the inflection type of some forms, and added a bunch of unnecessary links to very attestible forms. My moving the entry was, like, 5% of my edits, but you still reverted them all. I see you reinstated some, but it's still a mess, and just *that* should have made you second guess your wheel warring. --{{victar|talk}} 06:31, 13 September 2020 (UTC)

--{{victar|talk}} 06:31, 13 September 2020 (UTC)

@Victar: I started the discussion as soon as I re-moved the entry the first time and you ignored it and continued with your moving and editing.

i) Not adding etymology can hardly be the justification for calling the entry creation "sloppy" especially when there isn't a clear etymology. I left it like that because no etymology is better than a bad/wrong one.

ii) *gárždi ~ *gr̥ȷ́ánti, is the athematic root present which Cheung says is the source of gərəzōi. It's not explicitly reconstructed by the sources so what we can do is add gərəzōi under unsorted formations.

iii) The derived terms were not shown as links because I was following the standard used by PIE entries.

iv) The citations to attested entries are gone now if that feels better. गृहते needs citation though as its exact meaning is obscure. I'll remove the cites to Middle Persian too. Not gonna lie, I think terms which do not have entries should be cited but I'm willing to compromise on that.

v) Your edits also removed the Old Persian reconstruction which is given by IAIL and Cheung. They also removed the Bactrian and Kurdish terms.

vi) Finally, as I said, it's understandable why a root entry like IIR *bhar- could be replaced by *bhárati but the best way to deal with cases like this is by having roots. Why not keep both like Mahagaja suggested? Even I by default prefer the verb form, especially for those verbs inherited from PIE; it's just in some cases I construct roots. -- Bhagadatta (talk) 06:53, 13 September 2020 (UTC)

@Victar: (Lol that first sentence is so misleading, "all PII editors" is a tiny set of people of which myself and Bhagadatta are a very significant portion of. There is no rogue editing going on here.)

• The reason I like our PIE root pages is because they list the various forms in a centralized place and they give the derivational processes involved very explicitly (e.g. "thematic reduplicated present") which category pages simply cannot do. I don't see the harm of that in PII. Also, sometimes it's difficult to attach a particular reflex to a specific verb form (for example), so we can include uncertain formations in root pages without a loss of information. Furthermore, some derivations have very few reflexes and making an isolated entry for them without the context of related root derivations (with their reflexes) present makes them not useful. One other thing I would like to point out is that there are some roots that are in PII and not able to be reconstructed for PIE. A central root page that gives a discussion of the etymological theories is nice to have in that case.
• Inconvenient because it makes information needlessly scattered and gets rid of the preferred way to centralize that information. Again, PG is not PII. Sanskrit has a productive root derivational system, with ablaut and everything well in play. We have Sanskrit root entries, and no one is clamouring for their removal. Do not see why PII should not maintain that chain. IDK what you are saying with "holy magic root", the etymology is going to explicitly given at the top if it's from PIE.
• Sure, forging our own ways of doing things is fine when we all agree it makes sense. I'm not saying to stop making lemma entries and keep everything at the root page. As an "interactive online dictionary with no physical limitations," having root pages on top of lemmas has no negative effects on our infrastructure or organization.
• It's needlessly pedantic when you have to resort to making a related terms section with descendants listed under it! That is such a terrible way to organize this. Descendants do not belong in that section at all. Why the heck would I search for a verb form and expect descendants of anything but that form (any maybe direct derivations of that form) on that page? The format of the page after you moved it leaves two lines for the actual descendant of this verb form underneath 8 lines (a quarter of my laptop screen) of related forms. This is not a good organization by any conceivable definition of "good".
• Finally, we're all well-established PII editors. We can have a normal friendly dialogue as equals. As far as I see, there was no consensus on the root pages that justified you moving it and not expecting it to be moved back. Manufacturing a consensus out of thin air isn't going to work when there's like 3 of us total. —Aryaman^{A (मुझसे बात करें • योगदान)} 02:37, 14 September 2020 (UTC)