User talk:Mahagaja/Archive 12

I feel that these forms are somewhat superfluous. Aren't they entirely predictable from the mutation-triggering effect of the preceding article? In theory, you could include all kinds of other words that mutate a following word in the table too. —CodeCat 19:59, 6 September 2016 (UTC)Reply

I guess you're right. They also only show the situation in the standard language, not the dialects (e.g. the table at bád shows leis an mbád and don bhád, but some dialects have leis an bhád and some have don mbád). —Aɴɢʀ (talk) 10:13, 7 September 2016 (UTC)Reply

I've converted {{ga-mut-cons}} and {{ga-mut-vowel}} to use Lua instead. Nothing else has really changed about the templates, so it should all still work. However, Lua has allowed the process to be fully automatic, so I've also created {{ga-mut}}, which can just be placed on any page without any parameters, and it should work the right way. If you approve it, I assume the older templates can be deprecated. —CodeCat 21:21, 6 September 2016 (UTC)Reply

It looks good. Maybe it could take a parameter indicating if a term is a masculine singular noun, and then suppress t-prothesis for any vowel-initial word that isn't one, since masculine singular nouns are the only things that ever undergo t-prothesis. That way, the mutation table for uisce would still list t-uisce, but the tables for óige and uiscí would not list *t-óige and *t-uiscí. —Aɴɢʀ (talk) 10:18, 7 September 2016 (UTC)Reply

I do see one problem: it generates "shtát" instead of "unchanged" at stát. We normally haven't been putting mutation templates on words that don't mutate, but in principle it should give the correct results if someone does. —Aɴɢʀ (talk) 10:21, 7 September 2016 (UTC)Reply

I seem to have misinterpreted the original code a little then. Am I correct that lenition of s only occurs in the same environments as where t-prothesis can occur? That is, before a vowel or l, n, r? —CodeCat 12:08, 7 September 2016 (UTC)Reply

Yes; "s" lenites (whether to "sh" or to "ts") only before a vowel or l, n, r. It does not mutate before c, f, m, p, t. —Aɴɢʀ (talk) 13:35, 7 September 2016 (UTC)Reply

Ok, I fixed this now. —CodeCat 14:04, 7 September 2016 (UTC)Reply

I hope someone will deal with all the module errors. Chuck Entz (talk) 12:15, 7 September 2016 (UTC)Reply

I've fixed some of them, but I think the rest will need someone with better knowledge of Irish. It seems that one of the parameters is missing, which, in the old template, would display no mutated forms at all. So I wonder whether these words really don't mutate, or if this is an error.

Also, can t-prothesis occur for feminine nouns starting with s-, or is the gender distinction only relevant for vowel-initial words? —CodeCat 12:21, 7 September 2016 (UTC)Reply

T-prothesis only happens to vowel-initial words. The change of s- to ts- is considered a kind of lenition (morphologically; phonologically of course it's fortition) because it appears only in leniting environments. It can happen to feminine nouns in the nominative singular as well as to masculine nouns in the genitive singular (and, outside the standard language, in a wide range of other environments). —Aɴɢʀ (talk) 13:35, 7 September 2016 (UTC)Reply

Are there any exceptions to mutation? Words that, on account of their shape, could mutate, but actually don't? I think that mutation should be shown even if it doesn't occur, because it provides a positive affirmation to the user that we didn't just forget the table (therefore the word probably does mutate) but that there really is no mutation for this word. —CodeCat 14:04, 7 September 2016 (UTC)Reply

Unassimilated English loanwords starting with a coronal consonant or f are generally immune to lenition (but not, I think, immune to eclipsis). So in Ros na Rún, I noticed that the vocative of Tom (an unassimilated English name) is a Tom, but the vocative of Tadhg (a native name) is a Thadhg. Maybe {{ga-mut}} could thus take an optional parameter 1 that can be set to msn for masculine singular noun and to no-len for a word that looks like it ought to undergo lenition but in practice doesn't. —Aɴɢʀ (talk) 14:26, 7 September 2016 (UTC)Reply

Are these two factors mutually exclusive? Surely, there are words that are both masculine singular nouns and not-leniting. —CodeCat 14:36, 7 September 2016 (UTC)Reply

Probably, but if they're non-leniting, then it doesn't matter if they're a masculine singular noun. Being a masculine singular noun only matters for vowel-initial words anyway (the msn parameter would apply vacuously to consonant-initial masculine singular nouns), and only consonant-initial words undergo lenition (so you can't tell whether a vowel-initial word is immune to lenition or not). —Aɴɢʀ (talk) 14:41, 7 September 2016 (UTC)Reply

Ok, I've implemented this. Use msn for vowel-initial masculine singular nouns, and nolen for nouns with no lenition. —CodeCat 14:49, 7 September 2016 (UTC)Reply

In theory it could be any part of speech with no lenition. —Aɴɢʀ (talk) 14:59, 7 September 2016 (UTC)Reply

Oh, for vowel-initial words that aren't marked msn (e.g. óige), ideally there should be no "with t-prothesis" column at all, rather than one that returns "unchanged". Is that doable? —Aɴɢʀ (talk) 15:02, 7 September 2016 (UTC)Reply

I don't know. The way it is now, there is a positive affirmation to the user that there is no t-prothesis. That's useful information. —CodeCat 15:16, 7 September 2016 (UTC)Reply

That's one interpretation. The other interpretation is "in environments where t-prothesis is expected, nothing happens", but unless the word is a masculine singular noun, no t-prothesis is expected anywhere. —Aɴɢʀ (talk) 15:18, 7 September 2016 (UTC)Reply

@CodeCat: How about if instead of "unchanged" it says "does not occur" or "N/A" or something for vowel-initial words that aren't masculine singular nouns? —Aɴɢʀ (talk) 21:04, 11 September 2016 (UTC)Reply

But the t-prothesis form is unchanged for feminine nouns preceded by an. In other words, in environments where the mutation is triggered, no mutation actually occurs. It does not mean there is simply no form to use after an. —CodeCat 21:09, 11 September 2016 (UTC)Reply

Also, after analysing things a bit more, it seems pretty clear that the t- found on consonant-initial words is also t-prothesis, and has the same origin. And since it only occurs after an, it only occurs with nouns. I therefore think that this form should be moved to a separate t-prothesis column, and an extra possibility added for the parameter on {{ga-mut}}. If you specify {{ga-mut|n}}, then that indicates the word is a noun and therefore can display t-prothesis after an if it begins with s. If you specify {{ga-mut|msn}}, then t-prothesis is displayed if the word is vowel-initial. —CodeCat 21:20, 11 September 2016 (UTC)Reply

T-prothesis of vowels happens only where an precedes AND lenition of consonants is not expected, namely masculine nouns in the nominative singular:

But for feminine nouns, there is no place where they follow an AND lenition of consonants is not expected. Feminine nouns in the nominative and dative singular undergo lenition of consonants, and in all other cases (gen.sg. and thruout the plural) the article has the form na. So it isn't that the t-prothesis is unchanged for feminine nouns, but that there is no place where t-prothesis might be expected for a feminine noun.

If anything, t-prothesis occurs in the same places as h-prothesis, namely in contexts where consonants would not be lenited. The difference is that t-prothesis occurs only after an and h-prothesis occurs only after proclitics ending in a vowel sound.

The s → ts change, on the other hand, occurs in contexts where lenition of (non-coronal) consonants is expected, namely the genitive singular masculine and the nominative/dative feminine singular. The thing is, that other coronals (beside s) resist lenition after an:

And while s → ts occurs only after an in the standard language, in dialects it occurs anywhere that lenition is expected and the preceding word ends in l or n, e.g. in compounds (dialectal scoil tsaothair (“industrial school”) vs. standard scoil saothair or dialectal lántsásta (“fully satisfied”) vs. standard lánsásta) – note that sásta is an adjective. So the two t-prefixing operations are different as they occur in different environments (t-prothesis where consonants would NOT be lenited, s → ts where other consonants WOULD be lenited). —Aɴɢʀ (talk) 12:30, 12 September 2016 (UTC)Reply

I see. My way of thinking was that it doesn't matter to the mutation template which forms get the extra t and which don't. All that matters is the answer to the question: does the word have a special form when preceded by an or not? In this sense, the "special" kind of lenition is the same as the t-prothesis for vowel-initial words. In both cases, a t is added when an precedes. The patterns and underlying rules may be different, but the surface effect on a single word/case form is the same, and is always triggered by a preceding an. Therefore, certain individual words have a "special form after an" while others don't. This is my reason for considering them as one thing, at least for the purpose of the mutation template (which doesn't care about part of speech, case or gender).

As for the underlying mechanism of these changes, I'm curious where it came from. Specifically, since the special lenition type seems to be a variant of regular lenition restricted to certain combinations of consonants, I wonder where it came from. And where did t-prothesis before vowels come from? —CodeCat 15:47, 12 September 2016 (UTC)Reply

They both came from the /d/ of *sindos (“that, the”) plus /h/ from a lenited s (/d/ + /h/ = /t/). That lenited s can either be the final consonant of *sindos (t-prothesis) or the initial consonant of a following word (s → ts). In t-prothesis, an t-athair comes from *sindos (ɸ)atīr > *(h)indoh aθīr > *indh aθīr > Old Irish in t-athair. In s → ts, an tsúil (nominative singular feminine) comes from *sindā sūli > *(h)indā hūli > *ind hūlʲ > Old Irish int ṡúil /in͈ ˈtuːlʲ/ > modern an tsúil /ən̪ˠ t̪ˠuːlʲ/; and an tsaoir (genitive singular masculine) comes from *sindī sa(ɸ)irī > *(h)indī hairī > *ind hairʲ > Old Irish int ṡaír /in͈ ˈtaːi̯rʲ/ > modern an tsaoir > /ən̪ˠ t̪ˠiːɾʲ/. —Aɴɢʀ (talk) 16:33, 12 September 2016 (UTC)Reply

@CodeCat: As for there being a "special form after an", you could say that vowel- and s-initial words have a special form after an in some circumstances, but not always. The nominative singular for "the father" is an t-athair, but the genitive singular is an athar. The genitive singular for "the priest" is an tsagairt, but the nominative singular is an sagart. So an per se is not triggering the change. It always depends on the gender and case. Ó Dónaill's dictionary lists a word sraoille (“ragged person”) that can be masculine or feminine; it happens to have the genitive singular identical to the nominative singular. That means an tsraoille can be either masculine genitive or feminine nominative, while masculine nominative would be an sraoille. (Feminine genitive would be na sraoille.) So even if we were to lump t-prothesis and s→ts together in the tables, we couldn't just call it "special form after an" or "special lenition after an". And even something wordier like "special form after an under certain circumstances" wouldn't cover the dialectal cases like scoil tsaothair I mentioned above. —Aɴɢʀ (talk) 13:14, 13 September 2016 (UTC)Reply

That explanation makes sense if you treat the nominative and genitive singular as the same word, but really they're two words that coincide in form. If we have a mutation template for each of them, then this won't be an issue. We already have to have separate mutation templates for nouns and non-nouns because of t-prothesis. —CodeCat 13:19, 13 September 2016 (UTC)Reply

But we do treat them as the same word. We don't have separate entries for nonlemma forms that are identical to their lemma form. At zetten, we don't have separate entries for the present indicative plural, past indicative plural, present subjunctive plural, and past subjunctive plural of zetten. At oppidum we don't have separate entries for the accusative and vocative singular. Likewise at sraoille we wouldn't have a separate entry for the genitive singular. Or do you mean that under ===Mutation=== we could have two tables, one for the nominative and one for the genitive? That seems like more work than is necessary. It seems so much easier to just have vowel-initial words that are not tagged "msn" display "not applicable" under their t-prothesis column, and leave everything else as it is. —Aɴɢʀ (talk) 13:42, 13 September 2016 (UTC)Reply

Ok, we can leave it as it is. Do note though that some entries do have a non-lemma on the same page as the lemma, because the non-lemma inflects. This occurs with participles, e.g. vergeten. —CodeCat 13:50, 13 September 2016 (UTC)Reply

I don't want to leave it as it is; I want it to say "not applicable" rather than "unchanged" for non-msn vowel-initial words. Fair enough point about vergeten (it could be done for vergessen too, but no one's bothered), but obviously that wouldn't apply to genitive nouns. —Aɴɢʀ (talk) 13:56, 13 September 2016 (UTC)Reply

You did not warn me about it, I am apalled! It was removed without my consent or consulting me. The only descendant is not goidelic, there are also welsh, cumbric and cornish descendants. It is Proto-Brythonic, and I'm asking you to put it back. UtherPendrogn (talk) 10:40, 8 September 2016 (UTC)Reply

Sorry to intrude on Angr's talk page, but if I had seen the entry first, this would have been on my talk page because I would have deleted it on sight. I did warn you (three weeks ago) about getting the details right so people don't have to pick up after you. Also, it's not "your" page and no one has to consult you about it- this is a wiki, and no one "owns" anything.

The deleted page was a "Proto-Brythonic" entry with a Proto-Celtic header and Proto-Celtic contents. The only working Proto-Brythonic content outside of the etymology was an inflection section, which doesn't belong in a verb entry. Also, Proto-Brythonic, by definition can't have Goidelic descendants, except through borrowing between branches- very unlikely for core vocabulary such as this. The entry had nothing in it to indicate you had the slightest idea what you were doing.

You may have intended to come back and fix it later, but when you click the button to save it, it becomes part of the dictionary- so you need to preview it and get everything right before you do that. As it is, Angr deleted it a full two days after your last edit, so there was no evidence at all that you were still working on it. Chuck Entz (talk) 14:05, 8 September 2016 (UTC)Reply

Nonsense. The contents were Proto-Brythonic and the header Proto-Brythonic. The "inflection" table was your own Proto-Celtic verb table which I had modified for conjugating Proto-Brythonic. I suggest you restore the page. UtherPendrogn (talk) 14:24, 8 September 2016 (UTC)Reply

I don't even know what page you're talking about. —Aɴɢʀ (talk) 14:27, 8 September 2016 (UTC)Reply

https://en.wiktionary.orghttps://dictious.com/en/Reconstruction:Proto-Brythonic/ber Please return it? UtherPendrogn (talk) 14:41, 8 September 2016 (UTC)Reply

I guess you're talking about Reconstruction:Proto-Brythonic/ber. If there's any Brythonic content (there wasn't any when I deleted it), then just add it to Reconstruction:Proto-Celtic/bereti. —Aɴɢʀ (talk) 14:41, 8 September 2016 (UTC)Reply

That'll teach me to post when I'm late for work. My comment about the inflection section is, of course utter nonsense- I looked straight at "inflection" and read it as "declension". I stand by the rest of what I said, though. I won't belabor the rest of what you said, but let me point out that, contrary to your statement above, I had nothing to do with the entry itself, because I don't work with Celtic languages except to get rid of obviously wrong edits. I have a bachelor's degree in linguistics from UCLA, including courses in historical and Indo-European linguistics- but I'm not familiar enough with Celtic languages to feel comfortable editing them.

Probably the most important thing to learn when dealing with advanced subjects is the extent of what you don't know. It's not enough to get some of the details right, you have to get all of them right. And when you make mistakes (which everybody does), how you deal with them is almost as important as avoiding them in the first place. You seriously flunked your first test in that regard, but you're young- you still have plenty of time to learn. Chuck Entz (talk) 07:50, 9 September 2016 (UTC)Reply

Also Chuck Entz, plenty of Proto-Brythonic words have descendants or similar words in Old Irish (Goidelic). You'd know that if you had the slightest idea what you're talking about. UtherPendrogn (talk) 14:42, 8 September 2016 (UTC)Reply

You're lying through your teeth. I had used cel-bry-pro-verb or whatever it was, which *I* created. There was Proto-Brythonic content! UtherPendrogn (talk) 14:44, 8 September 2016 (UTC)Reply

But there were no Brythonic descendants. If you find any, add them to the Proto-Celtic page. That's sufficient. —Aɴɢʀ (talk) 14:51, 8 September 2016 (UTC)Reply

It's NOT. The verb needs to be conjugated in Proto-Brythonic, as it was before you deleted it. UtherPendrogn (talk) 14:58, 8 September 2016 (UTC)Reply

Why does a verb with no Brythonic descendants need to be conjugated in Proto-Brythonic? —Aɴɢʀ (talk) 14:59, 8 September 2016 (UTC)Reply

It has four Brythonic descendants and about as many Goidelic ones. I've reposted it. Besides, your logic makes no sense. The french word "aller" has no descendants, should we not give its conjugation? Or hell, the actual english word "to be" has no descendants. Should we not conjugate it? UtherPendrogn (talk) 15:14, 8 September 2016 (UTC)Reply

Modern French and English aren't protolanguages. They're subject to the rules for attested languages. Proto-Brythonic is subject to the rules (and common sense) for protolanguages. I've deleted the page again because the words you claimed as Proto-Brythonic descendants are not descendants of Proto-Celtic *bereti and Proto-Indo-European *bʰéreti (“to bear”) but rather of Proto-Indo-European *bʰuH- (“to be”). —Aɴɢʀ (talk) 15:20, 8 September 2016 (UTC)Reply

I have reported you for abuse of power. You are a moderator yet are ignoring the site's own rules "If editing can improve the page, this should be done rather than deleting the page. Vandalism to a page's content can be reverted by any user." and keep deleting the page, which contains correct information. UtherPendrogn (talk) 15:23, 8 September 2016 (UTC)Reply

Sorry, but deleting nonsense isn't vandalism. Welsh bod and its Brythonic cognates are not descendants of this verb. If you find any Brythonic descendants of bereti, add them there, but stop creating new pages for Proto-Brythonic verbs without any evidence of Brythonic descendants. —Aɴɢʀ (talk) 15:31, 8 September 2016 (UTC)Reply

My evidence is right there. I suggest you stop deleting the page. I've already reported you for vandalism and abuse of your admin powers. UtherPendrogn (talk) 15:33, 8 September 2016 (UTC)Reply

Right where? All that Reconstruction:Proto-Brythonic/ber had on it when I deleted it was non-Brythonic descendants, and all that Reconstruction:Proto-Brythonic/beri had on it when I deleted it was Brythonic descendants of a different verb. Neither of the pages was accurate. —Aɴɢʀ (talk) 15:36, 8 September 2016 (UTC)Reply

They're also descendants of Ber, not just Buta. And Proto-Brythonic words CAN have Goidelic descendants, suggesting otherwise is disingenuous and false. The details were not wrong, just incomplete, and three weeks ago they weren't any more wrong, I had just put it in the wrong category. UtherPendrogn (talk) 15:39, 8 September 2016 (UTC)Reply

No, they aren't descendants of ber; they don't have r's in them. PBr words can only have Goidelic descendants if they're loanwords, but beirid isn't a loanword from Brythonic. 15:41, 8 September 2016 (UTC)

Speaking of inaccurate pages, what is Reconstruction:Proto-Brythonic/rọj supposed to be? The ancestor of Welsh rhoi? That's a contraction of rhoddi, from Proto-Celtic *ɸrodati or the like, from Proto-Indo-European *pro- + *deh₃-. You seriously need to learn something about historical Celtic linguistics before you keep making pages for protoforms that never existed. —Aɴɢʀ (talk) 15:41, 8 September 2016 (UTC)Reply

Hello just a note from a #wikipedia IRC member. A user in there is claiming you are vandalizing by deleting pages. I made no comment, and referred him to speak with a local 'cart since I can't find your version of WP:ANI. -Cameron11598 (talk) 15:46, 8 September 2016 (UTC)Reply

Note: the user is UtherPendrogn -Cameron11598 (talk) 15:50, 8 September 2016 (UTC)Reply

Why don't YOU learn something about Celtic linguistics? It's from Proto-Celtic rāje/o and an ancestor of Welsh rôch. UtherPendrogn (talk) 15:48, 8 September 2016 (UTC)Reply

More words with Goidelic descendants. https://en.wiktionary.orghttps://dictious.com/en/Category:Irish_terms_derived_from_Proto-Brythonic https://en.wiktionary.orghttps://dictious.com/en/Category:Scottish_Gaelic_terms_derived_from_Proto-Brythonic

Ber means to carry, it doesn't even mean to go... And you're right, the goidelic forms should have gone in "cognate".

Can't apologise enough. I feel like a complete fool. Sorry.

Ber means to carry, it doesn't even mean to go... And you're right, the goidelic forms should have gone in "cognate".

Can't apologise enough. I feel like a complete fool. Sorry. UtherPendrogn (talk) 17:30, 8 September 2016 (UTC)Reply

I noticed you've been wrapping all definitions in this template, but I'd like to ask you not to do it. It causes difficulties with parsing pages, since senses are now nested inside templates. It also makes it impossible to add per-sense functions to {{def}}. —CodeCat 13:06, 9 September 2016 (UTC)Reply

I realize that {{def}} was not discussed very much and is not widely accepted yet, and a related vote is probably going to fail soon, so I take it we have some time to discuss and figure out the rules to use it. In my opinion, it's really better to wrap all definitions with {{def}} or a similar template, which is similar to wrapping all translations with {{trans-top}}, {{trans-mid}} and {{trans-bottom}}. Per-sense stuff can be added with other templates, including {{lb}} for context labels, and {{defdate}} for definition dates. --Daniel Carrero (talk) 11:12, 12 September 2016 (UTC)Reply

Shall we have a civilised discussion in a chatroom? It could be easier than talking on here. UtherPendrogn (talk) 17:42, 18 September 2016 (UTC)Reply

The discussion affects more than just you and me, so I'd prefer to keep the discussion here. —Aɴɢʀ (talk) 17:51, 18 September 2016 (UTC)Reply

Anyone else could join in. Conversation like this seems to go around in circles.

I clearly don't understand enough conventions right now. I don't want to have to make people "clean up" after me, but that seems to be what is happening. But for some things it seems to be unfair: like why delete my Lepontic stems? I quite literally lifted then off Lepontic Lexicon and added some "placeholder" forms (I didn't know it could be left blank, which I'll now do). UtherPendrogn (talk) 17:55, 18 September 2016 (UTC)Reply

Which stems do you mean? I'm pretty sure I didn't delete any Lepontic entries, I just moved them to Old Italic script. —Aɴɢʀ (talk) 19:07, 18 September 2016 (UTC)Reply

You put them in the deletion proposals. And see, you made a mistake with Matasović, and fair enough, but people seem to be far less lenient with me. UtherPendrogn (talk) 19:14, 18 September 2016 (UTC)Reply

The things I added RFD tags to were things that CodeCat had nominated for speedy deletion (i.e. deletion without discussion); I moved them to RFD precisely so that they wouldn't be deleted immediately but could first be discussed. —Aɴɢʀ (talk) 19:26, 18 September 2016 (UTC)Reply

Yes, and thanks for that. I'm not sure there's any point to this, clearly new members aren't welcome, or people who try to make any edits whatsoever that long-standing members don't like. UtherPendrogn (talk) 19:28, 18 September 2016 (UTC)Reply

Anyone is welcome to join the community, but demanding people let you do everything your way, insisting you're the only person who knows anything, and responding to any critical comment with childish insults is hardly a way to get people to listen to you. You're starting to moderate your behavior- which is good- but you have a lot to learn about getting things accomplished in a cooperative setting. We wouldn't be doing you any favors if we let you get away with this, because you're going to have to learn eventually. At least here you can't get expelled, fired or sued for acting this way. Chuck Entz (talk) 22:21, 18 September 2016 (UTC)Reply

"Matasovic made no mention of this word"

https://en.wiktionary.orghttps://en.wiktionary.org/w/index.php?title=Reconstruction:Proto-Celtic/akino&oldid=40538762

Didn't say he did. Check the ISBN, it's the addended and revised version of his dictionary, not his dictionary. Please consider reverting that edit. UtherPendrogn (talk) 18:25, 18 September 2016 (UTC)Reply

I found it later; it's in the Addenda and Corrigenda to the dictionary, which is available as a PDF online. It's not really a revised version of the dictionary, as it doesn't include entries that didn't need to be expanded or corrected. —Aɴɢʀ (talk) 19:06, 18 September 2016 (UTC)Reply

Be that as it may, can you remove the deletion tag? UtherPendrogn (talk) 19:07, 18 September 2016 (UTC)Reply

Not now that the deletion discussion is already under way. In that discussion, I did say right out that I found the Addenda & Corrigenda and that akino- is there. But I'm still uncertain what to do with this. We normally list Proto-Celtic (and Proto-Indo-European) nouns in their nominative singular form, not their stem form, but we need to know whether it was originally masculine or neuter so that we know whether to move it to akinos or akinom. And then there's the problem of the vowel: Matasović reconstructs it with a short i, but the descendants can only come from a form with a long ī. I suspect a typo on Matasović's part, but I can't prove it. —Aɴɢʀ (talk) 19:25, 18 September 2016 (UTC)Reply

Seems very odd to delete it for that. "There's a slight mistake with the i and declension, so let's permanently delete the page"? UtherPendrogn (talk) 19:27, 18 September 2016 (UTC)Reply

I'm having trouble finding actual inflection paradigms for Old Irish verbs of various classes. I'm wondering if you could tell me in what way a present in -ye- can be told apart from a simple thematic in -e-. I suspect that ibid is a ye-present and that the -iu of ibiu reflects the -y- directly, but I'm not sure. —CodeCat 19:14, 20 September 2016 (UTC)Reply

The clearest difference between the simple thematic presents (Thurneysen's class B I) and the ye-presents (Thurneysen's class B II) is in the first- and second-person singular conjunct forms. In class B I, these forms lose the final vowels the corresponding absolute forms have (absolute biru, biri; conjunct ·biur, ·bir). In class B II, these forms are identical to the absolute forms (absolute gaibiu/gaibim, gaibi; conjunct ·gaibiu/·gaibim, ·gaibi). Unfortunately, ibid isn't attested in the first- and second-person singular conjunct, so we don't know whether they would be *·iub and *·ib or *·ibiu and *·ibi. Normally the stem-final consonant shouldn't be palatalized in the 1st singular absolute in class B I (biru, not *biriu), but the presence of palatalization in ibiu doesn't really prove a lot since there is sometimes spontaneous progressive palatalization after /i/. The fact that the 1st and 3rd person plural forms are attested as ·ebam and ebait/·ebat, on the other hand, is more indicative. The lack of palatalization and the lowering of the vowel are unlikely to be analogical, which means these forms are most likely to go right back to *ɸibomos and *ɸibonti, not *ɸibyomos and *ɸibyonti. —Aɴɢʀ (talk) 20:47, 20 September 2016 (UTC)Reply

I've been using eDIL as a reference for Old Irish, but it doesn't say anything about verb classes, and while it does list attestations, it doesn't tell what form they are. Am I correct to guess that any unpalatalised form is evidence of a plain thematic, while a ye-present would expect to have only palatalised forms? —CodeCat 21:43, 20 September 2016 (UTC)Reply

In theory, yes, but in practice both palatalization and depalatalization sometimes occur in etymologically unexpected places, particularly in noninitial position (and in Old Irish spelling, palatalization isn't always reliably marked anyway). It's the "e" vowels of ·ebam/ebait/·ebat more than the nonpalatalization of the "b" that makes a plain thematic more likely. You have to be careful using eDIL also because it isn't just a dictionary of Old Irish: it lists Middle Irish and Early Modern Irish attestations as well. —Aɴɢʀ (talk) 22:08, 20 September 2016 (UTC)Reply

Which dictionary would you recommend, for the grammatical aspects in particular? —CodeCat 22:35, 20 September 2016 (UTC)Reply

Well, there isn't any other dictionary. For grammar, the classic work is Thurneysen, but he's almost exclusively synchronic. He provides the verb classification system that includes the B I and B II types, but I don't think he ever mentions that they correspond to the PIE plain thematic presents and ye/yo-presents. —Aɴɢʀ (talk) 10:34, 21 September 2016 (UTC)Reply

Would you be ok if I rearranged the forms in the Old Irish verb tables a bit? No forms will be added or removed, I just think that the layout can be improved somewhat. In particular, I'd like to change two things:

• Singular and plural forms are no longer listed together, but go in separate columns, so the order is simply 1sg, 2sg, 3sg, 1pl, 2pl, 3pl, passive sg, passive pl. Alternatively, the passive sg column can be shown after the 3sg column.
• Conjunct/deuterotonic forms are shown below the absolute/prototonic form rather than to the right. Consequently, you'll have 6 cells horizontally that show all the person-number combinations, like, say, Latin tables.

Something that I'm also curious about is why the perfect of prototonic verbs has a prototonic/deuterotonic distinction instead of an absolute/conjunct distinction like the other verbs. —CodeCat 21:01, 24 September 2016 (UTC)Reply

Can I see a mockup of how you'd like the table to look? The perfect has a prototonic/deuterotonic distinction rather than an absolute/conjunct distinction because the perfect always has a preverbal particle, which the other tenses do not necessarily always have. —Aɴɢʀ (talk) 15:59, 25 September 2016 (UTC)Reply

I created it at User:CodeCat/sandbox. —CodeCat 19:46, 25 September 2016 (UTC)Reply

I would also like to create separate parameters for the different parts that make up the inflection class. So one parameter would specify the present class (A I, II etc), one for preterite, future, subjunctive, etc. This will allow the verbs to be categorised by type, which is useful. —CodeCat 19:51, 25 September 2016 (UTC)Reply

That looks fine, and the separate declension class parameters sound good too. The table in your sandbox is very wide, too wide for my browser window at normal size, but I suspect when it's in use in actual entries it probably won't be, because so many columns will be empty in actual use. —Aɴɢʀ (talk) 20:26, 25 September 2016 (UTC)Reply

I've put in the new table. See beirid. It looks quite tidy. —CodeCat 22:45, 25 September 2016 (UTC)Reply

Looks good! Do you want to do it for {{sga-conj-complex}} as well? —Aɴɢʀ (talk) 14:24, 26 September 2016 (UTC)Reply

Yes, I just haven't had the time yet. —CodeCat 14:33, 26 September 2016 (UTC)Reply

Done now. —CodeCat 14:44, 27 September 2016 (UTC)Reply

Looks great, thanks! —Aɴɢʀ (talk) 14:46, 27 September 2016 (UTC)Reply

I'm working on the class parameters now, but I have a question. Do the classes for the various futures describe different formations that all share the same thematic vowel and endings? Or do the endings differ too? Likewise for the preterite and subjunctive classes. —CodeCat 21:44, 27 September 2016 (UTC)Reply

The different classes of future, preterite, and subjunctive generally have different endings from each other. —Aɴɢʀ (talk) 21:48, 27 September 2016 (UTC)Reply

Which types have the same endings? —CodeCat 21:50, 27 September 2016 (UTC)Reply

@CodeCat:. The f-, reduplicated, and ē-futures have basically the same endings (but with an f after the root in the f-future), but the s-future is a reduplicated s-subjunctive in form. The t-preterite has basically the same endings as the reduplicated preterite but with a t after the root; the s-preterite has different endings. The ā- and s-subjunctives have different endings. —Aɴɢʀ (talk) 19:26, 28 September 2016 (UTC)Reply

I've reimplemented the tables in Lua, with Module:sga-verbs. It doesn't change anything about how the tables work, just makes it easier to work with the parameters and add new features in the future. —CodeCat 14:59, 29 September 2016 (UTC)Reply

@CodeCat: Oh, that's great! (Except for this.) Would it be possible to write it so that it is no longer necessary to use {{l|ga|...}} in the tables? Especially, is it possible to write it so that variants could be separated by a comma or semicolon, while still being linked properly? In other words, I'd like to write |pastsubj_3p_conj=·rubaitis, ·rubtais or |subjunctive_ps_conj=·bether; ·ruibther and have the forms linked individually. And maybe some shorthand for {{q|''ro''-form}} while we're at it? It would be awesome to write something like |subjunctive_ps_conj=·bether; ·ruibther %ro and have it return "·bether; ·ruibther (ro-form)", if that's not being greedy. —Aɴɢʀ (talk) 12:01, 30 September 2016 (UTC)Reply

That's trickier. My own intuition was just to have separate parameters for when there's multiple forms, and corrsponding "qualifier" parameters. Just like {{en-verb}} already has for example. That would be more typing and less convenient, but much easier to implement and less prone to fail than a "mini-language". —CodeCat 13:30, 30 September 2016 (UTC)Reply

Well, even that would be better than what we currently have. I can live with |subjunctive_ps_conj=·bether |subjunctive_ps_conj2=·ruibther |subjunctive_ps_conjq=''ro''-form}} if need be. —Aɴɢʀ (talk) 13:41, 30 September 2016 (UTC)Reply

The numbers already worked before, so you can use them right away. The qualifiers don't yet, I need to work on those. Also, you may have noticed that I've been adding tables to entries without any forms in them, just to specify the stem classes. Presumably, in this case the table should put the entry in some category to indicate that no forms have been provided yet. We could put them in Category:Old Irish verbs needing inflection, or some new category. What do you suggest? —CodeCat 13:54, 30 September 2016 (UTC)Reply

Category:Old Irish verbs needing inflection seems good to me. What about doing without {{l}}? Is that feasible? —Aɴɢʀ (talk) 14:01, 30 September 2016 (UTC)Reply

The difficulty is that it has to support the old and new usages simultaneously. If I add linking to the module, then it's going to do the equivalent of this: {{l|sga|{{l|sga|·bether}}; {{l|sga|·ruibther}}}}. —CodeCat 14:05, 30 September 2016 (UTC)Reply

How long would it take MewBot to remove all the {{l|sga|...}}'s from existing tables? It wouldn't be a huge loss if there were a few hours when the words in the tables were left unlinked. —Aɴɢʀ (talk) 14:12, 30 September 2016 (UTC)Reply

Is there a way to tell whether a term is attested in Old Irish or in Middle Irish? —CodeCat 22:11, 24 September 2016 (UTC)Reply

Look at the date of its source. Sources like Sg, Ml, and Wb are definitively OIr, Sanas Cormaic can usually safely be assumed to be OIr, everything else needs to be read more closely, as the existing manuscripts are post-900 transcriptions of what may or may not be OIr originals, and may or may not have been "corrected" in transcription to a greater or lesser degree. Also, the Annals especially were written by professional historians, and some sources deliberately used archaic spelling and grammar. So you'll see spellings like "Áed" into the 17C, when you'd expect "Aodh", for example. --Catsidhe ^{(verba, facta)} 22:23, 24 September 2016 (UTC)Reply

Still far from ideal. Is there some kind of failsafe option I can take? —CodeCat 22:30, 24 September 2016 (UTC)Reply

An option which doesn't ultimately rely on making an educated judgement on messy primary evidence? Not really, no. --Catsidhe ^{(verba, facta)} 22:34, 24 September 2016 (UTC)Reply

Catsidhe is right. There's some more info at WT:ASGA#Defining the language, and there are some hints you can follow, but nothing I'd call failsafe. The hints include the following: (1) If you see a form ending in -a or -e where in Old Irish you'd expect another vowel like -o/-ae/-ai or -i/-iu (e.g. súla instead of súlo; unga instead of ungae and ungai, céile instead of céili and céiliu), then it's definitely a Middle Irish form (though the converse is not the case: the forms in -o,-ae, -ai, -i, -iu, etc., are not necessarily Old Irish). (2) Modern Irish spellings like ea for Old Irish e + broad consonant as well as bh dh gh mh are definite indicators of either Middle Irish or Early Modern Irish. (3) Archaic morphosyntax such as suffixed pronouns are a pretty clear indication of Old Irish, since they were dispensed with before the beginning of Middle Irish. (The infixed pronouns, on the other hand, lasted a while longer.)

I'm certainly not going to do any inflections in Old Irish, that requires way more analysing than I feel comfortable with right now. What I mainly need is something to answer the question "should there be an Old Irish entry for this". So even one OI attestation among a large number of MI ones is enough, I just need to know enough to be able to pick them out. I'm guessing that I can use the lemma form given by DIL even if that may not itself be attested (but some other form of the lemma is)? Or do I have to go hunting for an OI-attested lemma form among the lists as well? —CodeCat 19:25, 25 September 2016 (UTC)Reply

I definitely prefer to use a lemma form that's actually attested in OIr., except that I will add an acute accent to mark vowel length even when the form is attested without one. —Aɴɢʀ (talk) 19:28, 25 September 2016 (UTC)Reply

Perhaps this is obvious, but... are the sources in DIL arranged in chronological order? That is, do I only have to examine the first one to see how old it is? —CodeCat 13:51, 27 September 2016 (UTC)Reply

No, I'm afraid they aren't. —Aɴɢʀ (talk) 13:54, 27 September 2016 (UTC)Reply

*head meets desk* They really don't make it easy do they? It's great that they exist as a source, but it would be nice to be able to more easily date stuff. —CodeCat 13:58, 27 September 2016 (UTC)Reply

Yup! I think we're going to have to make Wiktionary's CAT:Old Irish language into the dictionary we would have liked DIL to be: one with dated citations that clearly differentiates between Old Irish and Middle Irish and between Middle Irish and Early Modern Irish. —Aɴɢʀ (talk) 14:03, 27 September 2016 (UTC)Reply

I just thought about something. In Middle Irish, the neuter gender merged with the masculine. So would a neuter gender be a sure pointer to Old Irish attestation? —CodeCat 20:51, 9 October 2016 (UTC)Reply

I'm not sure exactly when that happened. Early Middle Irish may still have had a neuter. And even in Modern Irish there are traces of the neuter in place names, e.g. Loch nEachach, with eclipsis after an originally neuter noun. I'd take the evidence the other way around: a clearly masculine gender of a formerly neuter noun is a good indication that a quote is Middle Irish, not Old. —Aɴɢʀ (talk) 20:58, 9 October 2016 (UTC)Reply

Hey, I wanted to get your opinion on the entry *ganskyos.

1. Should I merge *ganskyos with its root form *kankā and note the root differences in the descendants tree?
2. Does the descendants tree for *ganskyos look right to you with it's entry in Vulgar Latin?

Thanks! --Victar (talk) 17:36, 26 September 2016 (UTC)Reply

@Victar: I'm not sure I see how *ganskyos comes from *kankā? How does the initial *k- become a *g- and how does the *-k- become *-sk-? —JohnC5 17:41, 26 September 2016 (UTC)Reply

Matasovic explains the *g- by dissimilation on page 187. Also see the related is *kanxtus (“plough, ploughshare”), so I'm not sure if *ganskyos is the result of an *-s-insertion, or a new word from the PIE root. --Victar (talk) 18:09, 26 September 2016 (UTC)Reply

Perhaps the world in Proto-Celtic was *gankskyos. I don't know of any other examples of -ksk- so I'm not sure how it would manifest its descendants. --Victar (talk) 18:27, 26 September 2016 (UTC)Reply

Op, looks like I found an example in *farsko- (“ask, plead”), which dissimilated *-ksk- to *-sk-, from PIE *prek-, so *kankskyos to *ganskyos seems a legitimate shift. The question i guess still is did it happen in PIE or Proto-Celtic. --Victar (talk) 18:34, 26 September 2016 (UTC)Reply

Consider also that -ks- would become -xs-. —CodeCat 19:37, 26 September 2016 (UTC)Reply

I would definitely keep the entry at ] and mention the derivation from *kankā as only a possible etymology. It is kind of a lot to swallow: first the suffixation and gender change from *kankā to *kankskyos (and -skyos isn't exactly a common suffix in PIE, though I guess it has a parallel in the ancestor of uisce), and then the unexplained change of *k to *g (the explanation as dissimilation being really grasping at straws as far as I'm concerned). I'm not saying the etymology is wrong or utterly unbelievable, but we shouldn't treat it as an obvious and established etymology either. —Aɴɢʀ (talk) 09:40, 27 September 2016 (UTC)Reply

That makes sense to me. I'll update it accordingly. Thanks. --Victar (talk) 20:45, 27 September 2016 (UTC)Reply

Do you know anything about the etymology of this word (especially the plural roscada)? DTLHS (talk) 23:47, 28 September 2016 (UTC)Reply

Given the plural, it looks like a confusion of "an extemporaneous dithyrambic composition, used for retoiric and rosc" and its origin "a short poem, ode or chant". --Catsidhe ^{(verba, facta)} 00:22, 29 September 2016 (UTC)Reply

That's interesting. Here someone explicitly says roscada is the plural of rosc (maybe other plurals are attested). DTLHS (talk) 00:30, 29 September 2016 (UTC)Reply

I think roscada is a Middle Irish form, from a time when certain plural endings such as -acha, -anna, and -ada (modern Irish -aí) were starting to spread from the old consonant stems and being attached to nouns where they etymologically didn't belong. —Aɴɢʀ (talk) 12:07, 30 September 2016 (UTC)Reply

I have been reading through this book and wonder, is it ok if I create entries for any Old Irish lemma that's mentioned in it? Given that it's about Old Irish in particular, I wouldn't have to worry about dating, and I assume that a form is attested if it's in the book, I doubt it's unreliable. —CodeCat 13:33, 29 September 2016 (UTC)Reply

Not unreliable, but it probably uses normalized spellings that may not actually be attested. (For example, the conjunct of léicid is normalized as ·léici but AFAIK the only attested spellings are ·léci and ·léicci.) But I guess there's no reason for us to shy away from normalized spellings. —Aɴɢʀ (talk) 12:04, 30 September 2016 (UTC)Reply