Wiktionary talk:About Ancient Greek/Archive 6

Hapax clinomena
This has been another one of those things that has been bothering me for a long time. There are a lot of words that have, for lack of a better term, hapax clinomena—that is, inflected stems that only occur once, often as not in some backwater dialect. I don't especially like the idea of adding a whole conjugation table for them, and this is especially inappropriate when the form is one of those athematic monstrosities like λέκτο. I do however have a bit of a solution in mind already—that is, to list these forms separately, without inflection tables, and give citations for them. I had a couple entries a while back where I actually did this—e.g. σῴζω, which lacks the citations (and, for that matter, notes as to what inflected form those actually are) but otherwise is close to what I think could or should be done.

Agree? Disagree? Alternate proposals? — ObſequiousNewt — Geſpꝛaͤch — Beÿtraͤge 07:41, 2 January 2017 (UTC)


 * Are you referring to the “Other attested forms: σωάδδει and ἀποσοΐδδω ; σωννύω.” bit? What does it look like when “athematic monstrosities like λέκτο” are given conjugation tables? How do other sources treat hapax clinomena? Or does this issue not occur elsewhere because the English Wiktionary as a resource is unique in providing full conjugation tables for every Ancient Greek verb? — I.S.M.E.T.A. 19:59, 2 January 2017 (UTC)


 * Yes. From LSJ:

6 Lacon. σωάδδει· παρατηρεῖ, Hsch.: but also -σοΐδδω, aor. ἀπέσοιξεν· ἀπέσωσεν, Λάκωνες, Id. 7 σωννύω, Dinol.5:—
 * I don't know why I didn't bother with the other forms, but this is at least the idea that I have in mind. Athematic monstrosities like λέκτο are sometimes found in isolation, sometimes with a few other forms (in this case, we have ἐλέγμην and λέκτο) but supplying full conjugation tables is somewhere between misleading and impossible to predict. But in terms of other sources—the holistic dictionaries like DGE and the Great Scott just mention and cite them; other dictionaries might mention them if they are Epic but otherwise probably won't; I don't personally know of any other Greek sources than Wiktionary besides Tufts' and UChicago's Persei, which both do not have proper inflection tables (Tufts has a reverse form lookup, which is not very reliable for several reasons; UChicago has a reverse form lookup which is more reliable but much more limited.) There are admittedly a fair few problems that come out of trying to treat Greek as fully fusional rather than agglutinative-ish. — ObſequiousNewt — Geſpꝛaͤch — Beÿtraͤge 20:22, 2 January 2017 (UTC)


 * Would it be appropriate to include them in the format “Attested irregularly conjugated forms include the [grammatical form],  ([cited source]), and [grammatical form],  ([cited source]), as well as the [grammatical form],  ([cited source]).”? I didn't understand your last sentence, viz. “There are admittedly a fair few problems that come out of trying to treat Greek as fully fusional rather than agglutinative-ish.”; please explain it. — I.S.M.E.T.A. 01:39, 3 January 2017 (UTC)


 * Aye, that looks good. — ObſequiousNewt — Geſpꝛaͤch — Beÿtraͤge 02:07, 3 January 2017 (UTC)


 * Jolly good. What are those three words' grammatical glosses, and where in the corpus do they occur? — I.S.M.E.T.A. 17:55, 27 January 2017 (UTC)


 * σωάδδει is pres act ind 3sg. σοΐδδω is listed in LSJ but I evidently skimmed—the actual form is aor act ind 3sg ἀπεσόϊξεν (but listed in LSJ as ἀπέσοιξεν?). Both of these are from Hesychius. The third one, σωννύω, was used by Dinolochus, but we don't know the exact form—tracking down the citation yielded only the words of the "Anti-Atticist", who said that Dinolochus used "σωννύω ἀντὶ τοῦ σώιζω". I doubt it merits its own inflection table, regular as it may be, but the exact form is unknown. — ObſequiousNewt — Geſpꝛaͤch — Beÿtraͤge 18:21, 27 January 2017 (UTC)


 * OK, so we have “Attested irregularly conjugated forms include the third-person singular present active indicative,, and third-person singular aorist active indicative,  — both found in the Hesychian lexicon — as well as , reportedly used by Dinolochus (487 BC).” so far. Is  a sublemma (alternative first-person singular present active indicative)? — I.S.M.E.T.A. 21:53, 27 January 2017 (UTC)


 * Excellent, that looks better than I could have produced. The only thing I'd add is that the last form is specifically fragment 5. To your last question... σωννύω is presented in LSJ like a sub-lemma, but the fact that it's only attested once leaves me inclined to leave it how you have it, and not add a table. — ObſequiousNewt — Geſpꝛaͤch — Beÿtraͤge 22:03, 27 January 2017 (UTC)


 * Howzat? — I.S.M.E.T.A. 22:16, 27 January 2017 (UTC)


 * There's only one example of that present stem, and the stem itself is not a dialectal variation, so I am inclined to answer that no, it's not significant enough to gain its own table. Or was that not your question? — ObſequiousNewt — Geſpꝛaͤch — Beÿtraͤge 22:30, 27 January 2017 (UTC)


 * No, my question was simply whether you were happy with how I presented the information. :-D  — I.S.M.E.T.A. 23:00, 26 February 2017 (UTC)


 * Sorta related to this are the small handful of attested first-person dual verb forms. I created an entry for one at, and left a little note at the declension section of , but it doesn't look very elegant. Wondering how y'all would handle them. — Kleio (t · c) 19:33, 26 January 2017 (UTC)


 * That looks like a good solution to me, unless someone wants to write the module in such a way that the table accommodates first-person dual forms if and only if grc-conj has a parameter of the type  (where [X] is one of A, M, P, and [Y] is one of I, S, O, C). —Aɴɢʀ (talk) 21:00, 26 January 2017 (UTC)


 * I suppose then people would be less likely to realize there is a divergent form at all, since you'd have to look for it more specifically to notice it. Whereas if the information is presented outside the regular tables, it is less likely to be overlooked. — Kleio (t · c) 21:16, 26 January 2017 (UTC)


 * Very fascinating that such a form exists. I do like it being outside of the tables, where readers will notice it. — Eru·tuon 21:27, 26 January 2017 (UTC)


 * Yes, that is very interesting to note. Are there only three such first-person duals (viz., , and )? And are they all mediopassive indicative perfects? — I.S.M.E.T.A. 17:58, 27 January 2017 (UTC)


 * The purpose isn't (or shouldn't be) that they are there to be paid special attention to (especially considering that I suspect most of our readers are Atticists rather than philologists) but rather that they should be shown in some manner, and don't deserve whole inflection tables. We could just as easily put them in the "notes" section rather than below the tables, although I suspect the latter solution is better since many forms won't belong to any existing inflection and uniformity is good. — ObſequiousNewt — Geſpꝛaͤch — Beÿtraͤge 18:21, 27 January 2017 (UTC)

Mycenaean entries with descendants
I've noticed that a number of Mycenaean entries (which we do not have a About page for yet) give Ancient Greek entries as a "descendant", which is of course mostly incorrect. Anyone opposed to me converting these to something like "from Proto-Hellenic, compare βλαηβλαη"? --Tropylium (talk) 22:39, 12 January 2017 (UTC)
 * Yeah, that sounds wrong; I say go ahead and do it. — ObſequiousNewt — Geſpꝛaͤch — Beÿtraͤge 23:03, 12 January 2017 (UTC)
 * I think Mycenaean is supposed to be most similar to Arcadocypriot, so it can't be an ancestor of any of the literary dialects of Ancient Greek. — Eru·tuon 00:36, 13 January 2017 (UTC)

Typing convenience
&rarr; I created Module:typing-aids and, allowing us to more easily type in Ancient Greek. Module:typing-aids/data has shortcuts for macron and breve, as well as the other diacritics, which makes it much easier to type them than picking the diacritics out of the Edittools menu.

The order of diacritics described on this policy page still applies at the moment ( vs. ἀ̆́νερ, ἄ̆νερ, ά̓̆νερ...), but I should add automatic reordering of diacritics to the module. — Eru·tuon 06:54, 25 January 2017 (UTC)

Diacritic reordering is implemented! You can input the shortcuts for diacritics in any order you like, and the output will have the correct order. — Eru·tuon 07:57, 25 January 2017 (UTC)


 * Thanks, Erutuon! This is a great idea. —Aɴɢʀ (talk) 20:35, 25 January 2017 (UTC)
 * Maybe there could also be templates like l-chars, m-chars, and t-chars that would have all the functions of l, m, and t. Oh, and maybe aff-chars, cog-chars, der-chars, and inh-chars to be equivalent to the respective etymology templates. —Aɴɢʀ (talk) 20:42, 25 January 2017 (UTC)


 * I assume you mean something like . That would be so much more convenient. Module:unsubst might be useful for this, since it has a function for generating template code. Or as an alternative to creating lots of new templates, perhaps could be made to transform into another template, if you add the name of the template as the first parameter (with lang and term as second and third parameters). — Eru·tuon 21:53, 25 January 2017 (UTC)


 * Yeah, that would work too! —Aɴɢʀ (talk) 21:56, 25 January 2017 (UTC)


 * Okay, now and  work. — Eru·tuon 00:00, 26 January 2017 (UTC)
 * I added partial support for compound and affix templates too. (At the moment, no more than two morphemes are recognized, because I have not figured out how to generalize the process of printing out the template.) The parameters are printed the way I usually type them – after the part of the word that they relate to – thus,, not  . If you would prefer the other ordering, maybe I can figure something out. — Eru·tuon 21:37, 26 January 2017 (UTC)


 * This is a great idea; thanks for this. Would you mind updating Template:chars/documentation to explain the -, -, etc.-style usages, please? — I.S.M.E.T.A. 20:45, 27 January 2017 (UTC)


 * There you go! — Eru·tuon 22:35, 27 January 2017 (UTC)


 * Belated thanks, Erutuon. :-)  — I.S.M.E.T.A. 22:30, 26 February 2017 (UTC)

Fixed diacritics problem
I created a function in Module:grc-utilities that reorders diacritics for Module:grc-pronunciation, so now can take diacritics in any order you like. — Eru·tuon 20:31, 25 January 2017 (UTC)


 * Great idea. Is the module now wholly indifferent both to order and to combining-/spacing-ness? Also, since you're working on Module:grc-pronunciation and, would you like to institute audio-file auto-lookup as discussed in User talk:Erutuon/2016? — I.S.M.E.T.A. 18:22, 27 January 2017 (UTC)


 * I was also intending to do a little work on Module:grc-pronunciation to try and modernize and modularize a bit of its code. I also mentioned a syllabification error here which I may need help fixing. — JohnC5 18:30, 27 January 2017 (UTC)


 * I don't think the module will recognize spacing diacritics, except for spacing macrons and breves, since I had to recombine the letters at the end of the   function in Module:grc-utilities before Module:grc-pronunciation would recognize vowel length. Perhaps I could make Module:grc-utilities convert spacing versions of the diacritics to combining for Module:grc-pronunciation. (That would make things nearly foolproof.)
 * I'm bewildered by the pronunciation module, but I'll look into the audio-file lookup idea. — Eru·tuon 19:10, 27 January 2017 (UTC)
 * The first idea has been implemented. — Eru·tuon 19:59, 27 January 2017 (UTC)


 * Very nice. I doubt any spacing diacritics except the macra and brachia would be used very often, but there's no harm in making things foolproof. Best of luck to you with the pronunciation module; I still think that having Ancient Greek audio in our entries would be a wonderful addition. — I.S.M.E.T.A. 22:09, 27 January 2017 (UTC)
 * I'm not sure the reordering works for . — JohnC5 05:23, 28 January 2017 (UTC)
 * It looks fine to me. The Attic transcription is, with the mid-high pitch diacritic, which the module assigns to a long vowel with an acute. — Eru·tuon 05:28, 28 January 2017 (UTC)
 * However, I would prefer using the rising-pitch symbol (the caron) instead, or something else... — Eru·tuon 05:29, 28 January 2017 (UTC)
 * Sorry, at the moment, it is using, but doe snot work. — JohnC5 05:31, 28 January 2017 (UTC)
 * Ohh! I see. I'm looking into it. It may be some sort of basic coding error, to which I'm prone. — Eru·tuon 05:44, 28 January 2017 (UTC)
 * I used the wrong spacing macron: the modifier letter, not the plain one. Now it works again. — Eru·tuon 05:56, 28 January 2017 (UTC)
 * Woohoo! Thanks. — JohnC5 06:08, 28 January 2017 (UTC)
 * Something again seems to have gone sideways with . — JohnC5 06:36, 28 January 2017 (UTC)
 * It has to do with moving the data table to Module:grc-pronunciation/data. When I remove the reference to that module and preview, works again. — Eru·tuon 06:41, 28 January 2017 (UTC)
 * How do we fix it? — JohnC5 06:52, 28 January 2017 (UTC)
 * I'm not sure what's happening. I tend to just stare at things until the solution comes to me. We should add it to Module:grc-pronunciation/sandbox to try to get it to work before implementing it in the main module. — Eru·tuon 07:12, 28 January 2017 (UTC)
 * I'm very much the same way with coding in Lua. I'm also currently trying to understand how all works. First, I'd like to disentangle the IPA conversion and syllabification, so I can tackle them in turn. — JohnC5 07:16, 28 January 2017 (UTC)
 * I'm working on reformatting the IPA conversion rules so they aren't all jumbled together. They might be more understandable then. — Eru·tuon 07:48, 28 January 2017 (UTC)

Changes to transliteration module
I some of the transliteration module, using Module:grc-translit/sandbox. The transliteration in the sandbox module is slightly different, using y for the monophthongal vowel and the first element of the diphthong υι, and removing the diaeresis from a vowel that has a macron, since the macron sufficiently marks it as not being a diphthong. χ is rendered ks because English x is sometimes pronounced z. It's helpful to render what is known in Greek as a double consonant with a double consonant in the transliteration. I added the last, ξ &rarr; ks, to the main module, but commented out the rest. — Eru·tuon 23:26, 30 January 2017 (UTC)
 * &rarr; gyîon
 * &rarr; anaȳtéō
 * &rarr; sphínks


 * Where was it agreed upon to transliterate υ as y? —CodeCat 23:45, 30 January 2017 (UTC)


 * It wasn't, and I haven't added that transliteration to the main module. (I mean, I added it but commented it out.) — Eru·tuon 00:01, 31 January 2017 (UTC)


 * I would prefer to keep u as the transliteration of υ in all environments. —Aɴɢʀ (talk) 13:01, 31 January 2017 (UTC)


 * In general, I support transliteration schemes whose transliterated Greek text corresponds most closely to how Classical Latin Romanised Greek. Accordingly, I am in favour of the transliteration of monophthongal upsilon as and of diphthongal υι as, but am opposed to the transliteration of ξ (not χ) as , rather than . Unless I'm mistaken, there was a straw poll in the Beer parlour which concluded that ξ should be transliterated , not , so I ask you, Erutuon, to undo that change, please. As for omitting the diaeresis from the transliteration of long vowels, I support that if and only if it is always possible to infer from a given diaeresis-omitting transliteration that a diaeresis is required in the original Greek text. — I.S.M.E.T.A. 23:15, 26 February 2017 (UTC)
 * I'm in favor of ks because it's more intelligible, but I went and changed the transliteration of back to x since the change doesn't have consensus. — Eru·tuon 03:07, 27 February 2017 (UTC)


 * Thanks. Can you explain your intelligibility point, please? — I.S.M.E.T.A. 00:45, 11 March 2017 (UTC)
 * It's essentially what I said in my first post in this thread. Ks clearly indicates the sequence k plus s, while x word-initially would likely be interpreted as by an English speaker, and word-internally sometimes as, following the rules of English spelling. This can be seen in the pronunciations of xenophobe and Alexander. So to avoid having English speakers read xenos and Alexandros as zenos and Alegzandros, it would be clearer, more intelligible, for the transliteration to be ksenos and Aleksandros. — Eru·tuon 00:53, 11 March 2017 (UTC)


 * Shouldn't zeta therefore be transliterated, rather than ? And mightn't we expect the of a word-initial psi to be left silent by an English speaker? Anglophones are likely to have trouble, too, with monophthongal upsilon and the aspirated consonants, no matter what transliteration we choose for them. There isn't much hope for any Ancient Greek transliterations to be read faithfully if they are pronounced according to the rules of English orthography. — I.S.M.E.T.A. 00:51, 7 April 2017 (UTC)
 * No, zeta has to be transliterated z, because its pronunciation was sd only in Attic, not in Koine and Byzantine Greek (and perhaps not in other dialects in the Classical period). — Eru·tuon 00:54, 7 April 2017 (UTC)


 * Beta changes from to, delta changes from  to , phi changes from  through  to , and eta and upsilon both become . These Romanisations transliterate spellings; they don't transcribe sounds. — I.S.M.E.T.A. 23:57, 8 April 2017 (UTC)
 * I would also prefer to keep the transliteration as spelling-based (rather than pronunciation-based) as possible. I wouldn't even object to transcribing γ as g when it stands for /ŋ/, e.g. sphigx rather than sphinx. And I'd prefer to use only one Latin letter to transcribe one Greek letter wherever possible, with exceptions made only if they're unambiguous. The spellings th kh ph ps are unambiguous since Greek doesn't have spellings like τἁ κἁ πἁ or (I'm pretty sure) πσ. But ks would be ambiguous, since there are words like beside words like . That's why I strongly prefer x for ξ. —Aɴɢʀ (talk) 14:29, 9 April 2017 (UTC)


 * Why did you rewrite that much of the module? Aside from the direct use of combining diacritics, it seemed to me very simple and readable as I had it in October, and now you've almost doubled its size. — ObſequiousNewt — Geſpꝛaͤch — Beÿtraͤge 21:26, 5 February 2017 (UTC)


 * Because I couldn't understand it, and I felt there had to be a way to get the tokens that made more sense. I like searching for a whole diphthong rather than looking individually at the current and previous character. The extra length consists of longer variable names, functions to reduce repetition, comments explaining what the code is doing, and commented out code. (I removed the commented-out code just now to see how long it was.) I don't think the length involves much more processing time, though I could be wrong. — Eru·tuon 22:14, 5 February 2017 (UTC)


 * I've moved the tokenizer to grc-utilities and simplified the module. Please, in the future, if you find my code confusing, ask me about it rather than trying to rewrite it. I'm not convinced your version of the tokenizer is any simpler, and more importantly it was not a function that was prone to causing any problems. — ObſequiousNewt — Geſpꝛaͤch — Beÿtraͤge 23:57, 5 February 2017 (UTC)


 * I find it insulting that you say I "tried" to rewrite it. It worked, so how was it merely "trying"? And then you revert back to your terse and very confusing form with unintelligible variables and invisible diacritics. Sorry, perhaps I am more upset than I should be. I wish your modules were more intelligible, because then it would be easier to make the inflection templates more user-friendly. I make suggestions, but you are not interested in pursuing them. — Eru·tuon 00:11, 6 February 2017 (UTC)


 * grc-conj and grc-decl could use some work, although the problem of Greek inflection is inherently complex enough that I'm not sure how much I can really improve them. I'll make an effort after I've redone grc-pronunciation. I'd ask you to please stay calm—I didn't intend to insult you, and if you look you'll notice that I actually have followed your lead and removed the invisible diacritics. And finally, if you have recommendations as to how to make the templates more usable—or the code more readable—please talk to me; I am more than willing to help improve things. — ObſequiousNewt — Geſpꝛaͤch — Beÿtraͤge 00:25, 6 February 2017 (UTC)


 * All right, well, I'll cool down in time. One thing I can think of at the moment: a tracking category for pluperfect tables missing the third parameter (perfect passive participial stem). I just realized I've been adding pluperfects with only two stem parameters. I need to go back through and add the missing parameter. I don't understand the module enough to know how to add tracking, though I could do it in other modules. — Eru·tuon 01:24, 6 February 2017 (UTC)


 * Is there still a need to redo grc-pronunciation? Erutuon and I have significantly clarified and improved the module (though we need to move our changes out of the sandbox and into production). The only things we need to do now is (maybe) to add phonetic stuff. — JohnC5 16:22, 6 February 2017 (UTC)


 * It was something I had been working on, but I'll look at the module and see if it's still something I should bother doing. At any rate I am trying to find decent sources, though, to make sure that our pronunciation data is actually cited. — ObſequiousNewt — Geſpꝛaͤch — Beÿtraͤge 16:37, 6 February 2017 (UTC)


 * Module errors are to be avoided if at all possible, because they render the content unusable and they're alarming for site visitors. Hundreds of them constitute an emergency that needs to be resolved ASAP. Saying that a module is "working" when there are still pages at CAT:E because of it is a bit of a stretch. Chuck Entz (talk) 13:52, 6 February 2017 (UTC)

Search engine is fixed?
I can now search and find. Yay! They've finally fixed the search engine! — Eru·tuon 07:47, 10 June 2017 (UTC)

o-grade nouns
I created Category:Ancient Greek o-grade verbal nouns. Feel free to populate it (or delete it, if it's useless). --Barytonesis (talk) 11:42, 3 July 2017 (UTC)
 * After a discussion, replaced by Category:Ancient Greek words suffixed with -η (o-grade). --Barytonesis (talk) 12:15, 20 July 2017 (UTC)

Adjective templates
Category:Ancient Greek adjective inflection-table templates

Remains to be orphaned:
 * : ἑαυτοῦ
 * : φωνήεις

Orphaned:

====Issue==== cannot currently be replaced, because shifts the pitch of the neuter to the antepenult; this shouldn't be the case (the neuter retains the accent of the masculine, even when it could go further) >>, neuter , not **ἔργωδες. --Barytonesis (talk) 20:28, 12 July 2017 (UTC)
 * This is a probably one of these rules which don't always apply, though. Neuter of  seems to be correct. --Barytonesis (talk) 22:06, 12 July 2017 (UTC)

Noun templates
Category:Ancient Greek noun inflection-table templates

Almost:
 * : νόος
 * : Σάϊς
 * : υἱύς
 * : βοῦς
 * ὗς (needs to be handled)

Orphaned:
 * (but has to be fixed)
 * (but has to be fixed)
 * (but has to be fixed)
 * (but has to be fixed)
 * (but has to be fixed)
 * (but has to be fixed)

Category:Ancient Greek declension convenience templates

Orphaned:

--Barytonesis (talk) 20:07, 5 September 2017 (UTC)

Discussion
Can we please not delete the rest of the templates, and instead mark them as deprecated? Deleting them removes the record of what they contained (unless an admin can restore them), and makes earlier revisions containing them unreadable. — Eru·tuon 18:57, 5 September 2017 (UTC)
 * I personally like deleting them, but if not, we should have them throw an error / warning whenever they are not in preview mode. — JohnC5 19:21, 5 September 2017 (UTC)
 * I'd rather we continue deleting them progressively for now and keep them deleted for a few months. That way, 1) we can have a clearer idea of what still needs to be updated to or, and 2) there won't be any risk of people using the old templates if they're deleted altogether.
 * Then when the new templates are really entrenched we could restore all the old ones (I'm not really fond of that idea, even though it would indeed allow to check things).
 * I won't oppose if you really want to have them restored, but I'd like them to take as little place and be as inconspicuous as possible (I'd want to remove them from Category:Ancient Greek noun inflection-table templates and the like). And implementing John's idea of an error/warning would be nice. --Barytonesis (talk) 19:25, 5 September 2017 (UTC)
 * I can add a module error message in the main namespace for those templates that have been orphaned. Then, if someone uses them, they will appear in CAT:E, which people like to keep clean. (I was hoping that I could make an error message show up only in the most recent revision of a page, but I can't find any way to do that. There is a magic word for getting the ID of the most recent revision,, but not for getting the revision that one is currently viewing, even though that is possible in JavaScript. A CSS class could probably hide the message in old revisions, though.) I do understand how it is annoying to have the big categories of old templates that are not supposed to be used. — Eru·tuon 19:42, 5 September 2017 (UTC)
 * Looking at, I think the new template doesn't correctly generate Homeric forms. Compare:


 * The old template contains a helpful note about the oblique dual forms, and the new template generates Attic rather than Homeric forms of the genitive and dative plural. —Aɴɢʀ (talk) 15:50, 12 September 2017 (UTC)
 * Hmm, the module does have Epic forms programmed in. I've made it it replace  with  . It seems that the new template has a few forms that the old one doesn't. — Eru·tuon 17:47, 12 September 2017 (UTC)

tonos or oxia (acute)
I recently noticed that the acute accent in our Ancient Greek entries are the tonos, which is the Modern Greek accent. If I try to save an Ancient Greek term with oxia (the Ancient Greek acute), the Wikimedia process of normalization changes it to the tonos. If I copy the Ancient Greek word with tonos and try to google it, it does not find the correct Ancient Greek hits with oxia, it only finds the occasional errors or modern homographs with the tonos. I'm pretty sure there is nothing that can be done about this, I'm just griping. —Stephen (Talk) 20:18, 12 July 2017 (UTC)
 * You're right, there's nothing we can do about this unless we get Mediawiki to stop automatically changing letters with oxia to letters with tonos. And despite this problem, I don't want Mediawiki to stop doing this, because I really don't want us to have separate pages for, say, and, or even a hard redirect from one to the other. —Aɴɢʀ (talk) 00:20, 21 July 2017 (UTC)

Vocative particle in declension tables
As User:Angr suggested here, I think we should add ὦ before the vocative. I actually tried to do it myself, but to no avail. --Barytonesis (talk) 20:57, 12 July 2017 (UTC)
 * Why? It's not part of the vocative form of the noun, any more than the is a part of any English noun form, or to is part of the infinitive. —CodeCat 21:05, 12 July 2017 (UTC)
 * Neither are the articles. — Eru·tuon 21:14, 12 July 2017 (UTC)
 * Indeed, so they should also not be there. —CodeCat 22:11, 12 July 2017 (UTC)
 * I think learners find it helpful to have the article in there, though. We include articles in the German noun inflection tables as well. —Aɴɢʀ (talk) 13:40, 13 July 2017 (UTC)
 * Another (quite feeble) reason to add it: it would make copy pasting easier when we want to check that the forms are correct. --Barytonesis (talk) 12:03, 29 September 2017 (UTC)
 * How so? —Aɴɢʀ (talk) 13:43, 29 September 2017 (UTC)
 * My current method to check the vocatives is copy pasting ὦ from elsewhere, putting in front of the form the template gives me, and then copy pasting the whole for a Google query; I can then see if it's used in running text. It's necessary when the vocative is supposedly (according to the template) identical to the nominative; I have to find a way to discriminate between the two, or I won't get anywhere. I'll grant you it's not without imperfections: no result doesn't mean that the form couldn't theoretically exist, and results don't mean that there aren't other possibilities, or even that it's the more common form. Take Ἀπόλλων: "ὦ Ἀπόλλων" brings up a few results in Byzantine Greek, but it doesn't tell me that the Classical vocative form is Ἄπολλον.
 * Anyway, it's a (relatively small) waste of time having to copy paste ὦ from elsewhere first. --Barytonesis (talk) 16:19, 29 September 2017 (UTC)
 * A compromise suggestion: place the particle in brackets, to make it clearer that it's not a fixed part of the form? —Rua (mew) 18:49, 8 October 2017 (UTC)

Recent convos
--Barytonesis (talk) 12:48, 29 September 2017 (UTC)

Verbal adjectives (participles?) in -tós

 * I'd like to integrate terms like better with their corresponding verbs. Any idea? And should we consider them as passive participles? --Barytonesis (talk) 14:18, 22 September 2017 (UTC)
 * No, we shouldn't. They're related to verbs, but they're really not part of the conjugation system. The closest thing has to a past participle is . —Aɴɢʀ (talk) 14:26, 22 September 2017 (UTC)
 * As Angr says, the adjectives in 🇨🇬 were not participles but often became them in descendants. Ancient Greek did not undergo this change entirely. — JohnC5 00:31, 23 September 2017 (UTC)

What's next?

 * hello there.
 * Now that we're getting close to the finish line with the inflection templates, what will be the next big steps? Erutuon has informed me here that there was some work underway on the headword template/module. Please tell me it doesn't mean that we'll have to update five thousands entries manually ಥ﹏ಥ
 * what would be your pet projects? --Barytonesis (talk) 11:38, 2 October 2017 (UTC)
 * My next pet projects are Irish, Old Irish, and Gothic. And everything else under User:Angr. —Aɴɢʀ (talk) 12:14, 2 October 2017 (UTC)
 * Well, there are a few steps yet remaining in Ancient Greek:
 * We need to tighten up grc-IPA so that we can completely stop using grc-ipa-rows.
 * After seeing the assignment of accent-placement categories via the declension templates, I realize it should definitely be done by the headword templates and not the inflection because the differing accentuation that some tables can have from the lemma.
 * There are several changes we could make using AWB semi-automatically (removing the w param from grc-IPA, converting all instances of grc-wh-page to R:Woodhouse, some other reference cleanup).
 * My pet project at the moment is finishing up the automated Sanskrit declension module Mod:sa-decl whose output may be seen at User:JohnC5/Sandbox4. Any help you have would be nice. — JohnC5 17:57, 2 October 2017 (UTC)
 * FYI, I've removed all instances of w and calls to grc-pron. Sorry for the spam. — JohnC5 23:06, 2 October 2017 (UTC)
 * And now I've renamed all instances of grc-wh-page to R:Woodhouse. — JohnC5 23:11, 2 October 2017 (UTC)
 * other minor points that I'd like to settle are:
 * should we use the headers "Declension/Conjugation", or the sole header "Inflection"? I know CodeCat prefer the latter solution.
 * should we use or not?
 * I don't really care one way on the other, but I like uniformity. The less we have to worry about these things, the more we can work on real content. As for Mod:sa-decl, I can't help you as I don't know any Sanskrit. But it looks nice! --Barytonesis (talk) 15:10, 8 October 2017 (UTC)
 * I don't really care. I have always preferred "Declension/Conjugation", but I know cares a lot.
 * Definitely desc. I'm just lazy.
 * — JohnC5 04:08, 9 October 2017 (UTC)
 * I don't know about pet projects, but my pet peeve is a certain bit of unfinished business at Appendix:Ancient Greek declension tables/third/vowel that's been at CAT:E for a while now... Chuck Entz (talk) 13:42, 3 October 2017 (UTC)
 * I've finally dealt with that problem. — Eru·tuon 23:49, 3 October 2017 (UTC)
 * I don't know if Gothic needs work, User:KIeio has been working on that already. —Rua (mew) 18:50, 8 October 2017 (UTC)
 * Not recently; grad school is turning out to be very time-consuming. Y'all definitely shouldn't be holding your breath for me to get serious work done on Gothic in the immediate future. Glad to hear Angr intends to continue giving it some of their attention. — Kleio (t · c) 19:21, 8 October 2017 (UTC)
 * There are still over 4200 Gothic romanizations without a main entry, so yes, Gothic needs work! —Aɴɢʀ (talk) 19:38, 8 October 2017 (UTC)
 * There are still over 4200 Gothic romanizations without a main entry, so yes, Gothic needs work! —Aɴɢʀ (talk) 19:38, 8 October 2017 (UTC)

Template:R:LSJ and Module:R:Perseus
I was trying to fix an instance of in the Latin entry, and the behavior of the template is still hard to understand, despite my previous efforts to improve it. Can't use or ; have to use. It's also hard to figure out what's going on by looking at Module:R:Perseus.

This is bizarre, so I want to finally standardize the parameters. It should use only numbered parameters; w probably means "word", but I don't feel that "word" is the obvious name for the parameter in which one supplies the link text. And we replaced w with 1 in.

I'm thinking maybe the following: If there are two parameters, parameter 1 will be the Perseus entry code, and parameter 2 the link text. (The module will complain if either of the parameters are in the wrong script.) If there's one parameter, the module will detect whether it's Greek or Latin and use it accordingly. That will work for the Greek templates at least.

That means the example above will look like this:

This is the same order as a wikilink ( or  ). The thing that gets put into the URL goes before the thing that is displayed as link text. Not sure if that makes the most sense. If not, it can go the other way around.

If anyone has thoughts on what parameter order makes sense, please let me know.

I suppose a third option is for the module to detect the script of parameters 1 and 2 and decide based on that. — Eru·tuon 02:25, 8 November 2017 (UTC)

Okay, I've switched over to the system described above. I'm not totally sure if it will not cause problems in Latin reference templates. — Eru·tuon 08:31, 4 December 2017 (UTC)

Headword
Is there a way to track the entries that don't use any of the Ancient Greek headword templates, but ? --Barytonesis (talk) 10:38, 11 November 2017 (UTC)
 * I'm sure there's a way to edit Module:headword/templates to do that, but I don't know what it is. You could do it by brute force by adding  to head (and creating CAT:Ancient Greek entries using head as a hidden category), I suppose. —Aɴɢʀ (talk) 11:07, 11 November 2017 (UTC)
 * thanks. I don't have the user rights to edit though. --Barytonesis (talk) 16:39, 11 November 2017 (UTC)
 * I could edit Module:headword/templates to track this. It's also possible to do a search like  (returns a lot of form-of entries) or   (only lemmas). — Eru·tuon 16:45, 11 November 2017 (UTC)
 * Try . Chuck Entz (talk) 16:47, 11 November 2017 (UTC)

I should clarify. If you need a tracking template (for instance, for AWB), I can add that. But best to use search function if that is enough for your purposes. — Eru·tuon 16:51, 11 November 2017 (UTC)
 * The other option is to ask at the Grease Pit for someone to make a list from the dumps. Just on principle, I'm leery about adding overhead to every entry in the dictionary just so we can get information about a tiny fraction of those entries in real time. Chuck Entz (talk) 17:51, 11 November 2017 (UTC)
 * Me too. I think that is probably the best option. — Eru·tuon 18:14, 11 November 2017 (UTC)
 * Thanks to you both. I think  will work just fine, I've already found a few entries in need of overhaul. --Barytonesis (talk) 19:33, 11 November 2017 (UTC)
 * Are there dedicated headword-line templates for nonlemma forms? If so, check that category as well. —Aɴɢʀ (talk) 15:18, 12 November 2017 (UTC)
 * There aren't any yet, but I plan to work on module functions so that we can type, for instance, rather than . — Eru·tuon 20:07, 12 November 2017 (UTC)
 * I dunno, I think would be a more suitable name. —Aɴɢʀ (talk) 20:12, 12 November 2017 (UTC)
 * Oops! — Eru·tuon 20:20, 12 November 2017 (UTC)

Numbered parameters in headword templates
I've changed the behavior of numbered parameters in most of the headword templates that are powered by Module:grc-headword. To learn more and to discuss the changes, please go to. — Eru·tuon 22:21, 15 November 2017 (UTC)

Declension tables showing syncretism
In Module:grc-decl/sandbox, I'm trying out a table format in which syncretism between nominative, vocative, and accusative, and masculine and neuter, is shown: that is, if the forms are identical, the table cells are merged. (This format is used in the article on Wikipedia.) To make this possible, the articles are omitted and the cases are arranged in the order nominative, vocative, accusative, genitive, dative. The cell merging is done automatically during table creation, so it can easily be switched off and on.

An example:

This is just an experiment. I kind of like this layout, but I'm not sure if or how it should be used on Wiktionary. — Eru·tuon 02:36, 10 December 2017 (UTC)
 * I like it too. I think declension tables are easier to comprehend like this. —Mahāgaja (formerly Angr) · talk 09:23, 10 December 2017 (UTC)
 * I don't like it at all and find it much harder to understand. —Rua (mew) 14:29, 10 December 2017 (UTC)
 * I'm not terribly fond of it either, but it's probably out of habit; I could get used to it. I'm not sure either design is inherently better than the other, though; I think it's a matter of taste. So like I suggested for the ordering of the cases, it would be nice to leave the choice to the reader through some gadget (btw, thanks to you Erutuon). --Per utramque cavernam (talk) 02:25, 11 December 2017 (UTC)
 * At first sight it's a bit less clear, but I think that's mainly due to the decreased use of color coding/strongly defined borders between e.g. the singular/dual/plural columns and/or between the actual declension part at the top and the derived forms/notes part at the bottom. I like the idea of showing syncretism, but in its current form this proposed table is not as easy on the eyes as the tables we have at the moment. — Mnemosientje (t · c) 02:20, 11 December 2017 (UTC)
 * To clarify, the cosmetic changes are not what my post is about. The CSS styling and details of constructing the table can be changed. I do like having the borders that you describe, but it was easier to construct the table without them. I'm mainly curious about people's reactions to the look of merging the cells of syncretic forms (nominative and vocative; nominative, vocative, and accusative; masculine and neuter genitive singular; etc.). — Eru·tuon 02:59, 11 December 2017 (UTC)

Added some borders between singular, dual, plural, and derived forms. — Eru·tuon 03:13, 11 December 2017 (UTC)
 * Well, I for one definitely am not opposed to showing syncretism in the tables. Regarding the cosmetic side of things, I think the tables we currently use should serve as a model to emulate closely before any new design would be implemented. But yeah, that's mostly a side note. — Mnemosientje (t · c) 21:19, 18 December 2017 (UTC)


 * Ancient Greek declension is simpler than most of the other Indo European languages (Including modern ones.) That's why this new system is the better (Of course there might be an argument to add more special forms, like relic of locative case.--Yoshiciv (talk) 09:58, 21 August 2018 (UTC)
 * And if possible, I'd like to ask you to add a function to make BOT generate each pages of the forms--Yoshiciv (talk) 05:31, 3 September 2018 (UTC)

Why there are no pages of conjugated and inflected forms of Ancient Greek words, though Latin words have them?
I mean, BOT makes them when they generate the tables, doesn't it?--Yoshiciv (talk) 10:14, 21 August 2018 (UTC)
 * There are pages for many conjugated and declined forms (see Category:Ancient Greek non-lemma forms), but I think all or most of them were added by editors, not bots. The automatic tables are usually generated by modules (Module:grc-conj and Module:grc-decl). — Eru·tuon 23:29, 21 August 2018 (UTC)

Setting a standard order of inflectional categories
At the moment, inflection-of entries use several different orders when giving the categories of an inflected form. As acceleration has been added to declension tables, and will hopefully be added to conjugation tables, I would like to figure out which order should be used.

For adjectives and participles, there's a lot of variation: (which I prefer);  (which  prefers);  (which  prefers). One thing is constant in these, though: case comes before number. At the moment acceleration uses gender, case, number.

That agrees with the order that I chose for inflectional categories of nouns: case, number. I haven't specifically looked for noun form entries that use another order.

For verbs, the main variation is (I guess) whether voice comes before mood or the name of the non-finite form: for instance, present active infinitive vs. present infinitive active, first-person singular present optative active or first-person singular present active optative. I might add links to examples later.

So, and whoever else, which order do you prefer? Maybe you could also indicate the order used by textbooks, grammars, lexica, or other works that you have access to.

For what it's worth, Perseus uses number, gender, case for adjectives (a fourth order!), mood, voice for verb forms, and number, tense, voice, gender, case for participles when I look up a form in its "Word Study Tool". — Eru·tuon 19:24, 10 September 2018 (UTC)
 * For the record I prefer, 'Case Gender Number' for nouns, adjective, and participles, and for verbs 'person number tense', and although I really don't care about the order of mood and voice, I prefer 'Mood Voice'. I also prefer a system that uses 'and's instead of semicolons if the can be worked out. RexPrincipum (talk) 22:08, 10 September 2018 (UTC)


 * I prefer case and number to be kept together. In the Indo-European languages, they are a unified whole, while gender is clearly separate.
 * I definitely prefer the names of nonfinite forms to come last, thus "present active infinitive". When there's no nonfinite form, I'm less sure. In general, nouns should come after adjectives. You can speak of an "active infinitive" but not an "infinitive active".
 * As for inflections of participles, these should use the "sublemma" principle: the participle should be a sublemma, and its inflections should link back to it. Otherwise the definition lines become incomprehensible strings of technical terms.
 * For finite verb forms, person and number should be kept together, because again, these are one combined concept. Active vs passive is also closely tied to person and number in IE, but perhaps less so. Still, it might be an argument for the order "first-person singular active present indicative". —Rua (mew) 15:26, 11 September 2018 (UTC)


 * There are two ways to go about it:
 * 1) standard oral recitation as in this or that language, if there is a standard. For instance the greeks say: γ' ενικό οριστικής, ενεργητικού ενεστώτα: what is this? it is a 3rd singular of indicative, Of what?: of the active present. Or: Feminine genitive plural, of..... (adjectives before nouns). But: Genitive plural of masc, fem and neu... Or, as Rex: genitive of feminine plural. I do not know if anglophone students have a standard way to say it. But it seems there is no absolute standard.
 * 2) categories descending or ascending (as you wish) in a non verbal way:
 * [dialect]-PoS-[typedeclension, contracted, degrees, etc]-gender-number-case
 * [dialect]-PoS-[typeconjugation, contracted or not etc]-voice-diathesis-tense-mood-number-person
 * PoS-voice-diathesis-tense-gender-number-case (for participles) and so on. or ascending: ※ gen. - pl. - masc. and neu. - present perfect - mediopassive - participle
 * 2) is based on logic, not custom; I like using the ascending (person/case-number....etc). PoS is already there, headword line hosts termlabel-dialect, types, and gender for nouns. It is very safe, but it sounds like a robot.
 * 3) A nice combination as proposed by Rua could be
 * (pers/case number [+gender for adj/parts]) + of + descending: voice, tense, mood. + dialect and comments at the end.
 * infinitive of ....
 * I use and for simple co-occurences vertical or horizontal (λευκοῦ, λευκώ)
 * I use 'semicolon' when 'and' is nested only in one group (λευκόν, λύει)
 * and separate lines if very different (λύομεν) [ Perhaps 1st person 2nd person 3rd person with numbers is more economic? ]
 * handling participles (τετριμμένος: is this ok?)
 * Lots of pages coincide with modern greek (λύετε) or its polytonics. I guess it would be nice to harmonize them too. sarri.greek (talk) 19:52, 15 September 2018 (UTC)

I was curious, so I did a census of the orders of labels in instances of from Ancient Greek entries in the September 20 dump. The category names such as "person, "tense-aspect", "sound change" come from Module:form of/data. This census is imperfect because it just ignores unrecognized labels.

So as far as current practice is concerned, gender–case–number has a plurality, and voice before mood or name of non-finite form wins overwhelmingly. — Eru·tuon 20:22, 2 October 2018 (UTC)

Can you explain your reasoning for to Module:grc-decl/table changing the order from gender–case–number to case–gender–number? The previous order was my preference, but I am curious if you have some evidence that the other order is more common. — Eru·tuon 03:20, 21 May 2019 (UTC)
 * I think the order case-gender-number is more common across the whole of Wiktionary than gender-case-number, and it feels best to me, although I grant you that both orders feel reasonable to me. case-number-gender feels worse, and all the other orders much worse. Can you do a census across the whole of Wiktionary for uses of case vs. gender. vs. number? BTW the order "1st sg present active indicative" is I think overwhelmingly preferred in grammar books, and in languages whose verbs are conjugated for gender as well as person and number, the most common order is e.g. "3rd fem sg". Furthermore, state ("definite/indefinite") goes before case/number/gender, as does animacy, and I think that the order w.r.t. clusivity is "1st person exclusive plural". I actually wrote a program to sort inflection tags, and I used the following order:

"person", "clusivity", "class", "state", "animacy", "case", "gender", "number", "tense-aspect", "voice-valence", "mood", "comparison", "non-finite",
 * My sort algorithm leaves alone any tags that are not in one of the above categories, meaning that such tags effectively divide the set of tags into two, where sorting can occur only on either side of the fixed tag. Benwing2 (talk) 03:36, 21 May 2019 (UTC)
 * I can do a cross-linguistic census. (Probably the latest dump when it comes out in the next few days will be a bit cleaner and easier to process with all the work you've done.) However, I think it is better to base the order on logic or custom rather than a cross-linguistic census of existing entries. It's possible that different orders are used customarily or are more appropriate for the structure of different languages. Rua's reasoning about keeping case and number together in Indo-European languages (at least the ones I'm familiar with) makes sense to me. But I don't know what order is most commonly used in Ancient Greek textbooks and grammars written in English. (Clearly different orders are used in Greece.) I will see if there is any suggestion of a customary order in the few books I have. — Eru·tuon 06:12, 21 May 2019 (UTC)
 * I think there needs to be some consideration for the language in question too. In Dutch and German, as well as some Slavic languages, gender is tied to number in the sense that gender only exists as a category in the singular, while it's not distinguished in the plural. For such cases, it makes sense to consider gender+number as a single combined category and place them together. —Rua (mew) 09:17, 21 May 2019 (UTC)
 * I agree with Rua about Germanic and Slavic languages. If you can find evidence of a normal order in textbooks/grammars/papers etc. that's great, otherwise feel free to revert my change if you want. Benwing2 (talk) 13:49, 21 May 2019 (UTC)
 * I looked at Smyth's Greek Grammar, Hansen and Quinn's Intensive Course, and the introduction to the edition of Odyssey 1-12 by W. B. Stanford. My impression is that the order gender–case–number in what could be called context-free or neutral situations, though other orders are used in tables for presentation-related reasons, and in fluent text one or two categories are sometimes placed first or last, when they are the topic or are being contrasted with another value for that category. Granted, in Stanford I only spotted one neutral situation, in the footnote on page lxii.
 * For instance, a neutral situation from a footnote uses gender–case–number (Smyth, ): “In the fem. nom. sing. Ionic has -η, never -ᾱ; in the fem. gen. pl. Hom. has -ᾱ́ων (less often -έων); ... .” But when it has been established that the author is talking about the forms of a particular gender, case–number is placed first (Smyth, ): “The feminine is made from the stem of the masculine ... . The genitive plural feminine is always perispomenon ... .” And case is placed before gender when two cases are being contrasted (Smyth, ): “The nominative masculine ἡδεῖς is used for the accusative.”
 * Incidentally, these books also seem to use the order tense–{mood or nonfinite form}–voice in neutral situations: for instance, first-person singular present indicative active, present infinitive active. — Eru·tuon 20:56, 22 May 2019 (UTC)

"Borrowed" vs. "derived" from Pre-Greek
Compare Cat:Ancient Greek terms derived from Pre-Greek and Cat:Ancient Greek terms borrowed from Pre-Greek. Which would be the better template to use? I'm inclined to think Pre-Greek substrates had probably died out by the time Ancient Greek reached its recorded form so "derived" would make more sense than "borrowed", but we seem to have some sixty entries which use the "borrowed" template. — Mnemosientje (t · c) 16:34, 15 September 2018 (UTC)
 * We have no evidence to tell us how the Pre-Greek words got into Ancient Greek, nor even what "Pre-Greek" was, so IMO "derived" is the only responsible choice: if English were analogous, Greek, Jew, Wales, Briton, Daniel, wine, bishop, hemp, path, castle, peacock, chess, wassail, skin, them, etc. would all be from "Pre-English". Chuck Entz (talk) 18:34, 15 September 2018 (UTC)

Dialect codes
Which standard for dialect codes was used? These codes are different from the ones defined in the LINGUIST List.
 * Aeolic Greek. Wiktionary: aio, MultiTree/The LINGUIST List: grc-aeo
 * Boeotian. Wiktionary: boi, MultiTree/The LINGUIST List: grc-boe
 * Cretan. Wiktionary: kre, MultiTree/The LINGUIST List: grc-cre
 * Locrian. Wiktionary: lok, MultiTree/The LINGUIST List: grc-loc
 * Mycenaean Greek. Wiktionary: muk, ISO 639-3: gmy, MultiTree/The LINGUIST List: gmy — Лорд Алекс (talk) 23:41, 30 May 2019 (UTC)
 * Atelaes hasn't been active for a few years. I don't know if the codes are from any standard, but they seem to be the first three letters of a transliteration of the Ancient Greek place name:, , , , . — Eru·tuon 00:01, 31 May 2019 (UTC)