Wiktionary talk:About Proto-Brythonic

Not quite done yet, but it's a start.

According to Kenneth Jackson (1953), PB /w/ had developed only to a lightly velarized approximate /ˠw/ by the time of the split, before developing fully into /gw/ in West and Southwest Brythonic. I'm not really sure the best way to transcribe it, so I've left it as <gw> for now. Anglom (talk) 23:49, 16 April 2015 (UTC)Reply

I'm wondering if we should revisit this. <uu> is pretty common in Old Breton, which seems to point to it not being fully realized as yet. Maybe <ẉ> or just simply <ˠw>? --Victar (talk) 01:32, 3 November 2016 (UTC)Reply

gw is just fine. —CodeCat 01:36, 3 November 2016 (UTC)Reply

I think it might be overstating a change that hadn't happened yet. Especially because we're talking about a Proto-Celtic *w-*gu-*gʷ merger. --Victar (talk) 01:39, 3 November 2016 (UTC)Reply

It's common to all Brythonic languages, the definition of Proto-Brythonic. —CodeCat 01:48, 3 November 2016 (UTC)Reply

Is it though? Like I mentioned, we have a lot of attestations of <uu> in Brythonic languages and Jackson didn't reconstruct it as on a whim. --Victar (talk) 01:52, 3 November 2016 (UTC)Reply

We represent as gw. —CodeCat 14:40, 3 November 2016 (UTC)Reply

Exactly and since and are different, perhaps they shouldn't be labeled the same. What about using <ɣw>? <ɣ> is an already used character and I'll differentiate words in PBry that are *ɣw- and *gw-, from PCelt *w- and *gu/*gʷ, respectively. --Victar (talk) 19:18, 3 November 2016 (UTC)Reply

Are w- and gw- consistently distinguished in at least some Brythonic languages? —CodeCat 19:26, 3 November 2016 (UTC)Reply

Old Breton is probably the best example of <uu> vs. <gu>/<go>. The two were confused in the merger, but it does demonstrate that at some point, they were understood to be distinct. @Anglom, do recall what Jackson says on this? --Victar (talk) 20:34, 3 November 2016 (UTC)Reply

Proto-Celtic had no ē; PIE ē develops into ī. So what does it represent in the table?

Ah, it would be representing ī then. Anglom (talk) 01:14, 17 April 2015 (UTC)Reply

Proto-Celtic probably did have ē, because the monophthongization of ei probably happened within Proto-Celtic. If not, then certainly within Proto-Insular Celtic (if you believe in that). —Aɴɢʀ (talk) 11:54, 9 January 2016 (UTC)Reply

Celtiberian does not seem to have dimonophthongised *ei, showing as it does this spelling ⟨ei⟩, especially in the dative singular. So it's almost certainly not Proto-Celtic. There may have been a common ancestor ("Proto-Northern-Celtic") of Gaulish and Insular Celtic, as several scholars suspect (Lepontic may or may not belong to this group), and the monophthongisation could have happened there. (As an innovation common to Gaulish and Insular Celtic, Schumacher points out the enclitic relative particle *yo < PC *yod, which replaced the fully inflected, sentence-introducing relative pronoun *yos, yā, yod reconstructible for PC. Goidelic and Brythonic in turn share not only striking innovations in the verbal system, but also the change from *s to *ð immediately preceding voiced stops. The enterprise of subgrouping Celtic is, of course, hampered mainly by the limited attestation and our consequently limited understanding of the Continental Celtic languages.) --Florian Blaschke (talk) 15:46, 28 September 2016 (UTC)Reply

I take it you mean Celtiberian does not seem to have monophthongised *ei, but AFAIK some scholars believe that the spelling ⟨ei⟩ was purely orthographic and stood for /eː/. I'm not sure what leads them to that conclusion, though. —Aɴɢʀ (talk) 15:56, 28 September 2016 (UTC)Reply

That seems like an unfalsifiable claim. —CodeCat 16:22, 28 September 2016 (UTC)Reply

Matasovic in his introduction says that PIE *ey became PC *ē and that it naturally came after PIE *ē became PC *ī. He also describes it as a Late Celtic change. According to w:Old Irish#Vowels this PC *ē became OI /e₁ː/ (as well as from Latin ē), which was distinguished from /e₂ː/ which arose from compensatory lengthening from consonant loss or vowel hiatus. There apparently was also a change that preceded this which deleted laryngeals in *eyHC > *eyC, which then led to *ēC (I believe M reconstructs féith < *wētis < *wéyh₁tis). —JohnC5 18:11, 28 September 2016 (UTC)Reply

If *ei was still a diphthong in PC, its vowel system would have had /i, e, a, o, u/ as short and /i, a, u/ as long vowels – each by itself perfectly symmetrical and cross-linguistically extremely common, but the combination, a vowel system with more short than long vowels, is certainly typologically unusual and I've seen it described as unstable. So from this point of view, it certainly looks systematically expected that new long /e, o/ would arise through monophthongisation eventually. However, that does not mean that PC (which by definition wasn't spoken for an extended time period, at least if you take it to refer to the most recent stage recoverable by external reconstruction alone) cannot have had exactly such an unstable system. So if this is Matasović's line of argument, as I suspect because of the "naturally", I don't buy it. Ultimately, however, it doesn't really matter anyway since you can view *ei simply as an abstract representation of a phoneme whose precise pronunciation is irrelevant – only its place in the system is (per a structuralist approach). --Florian Blaschke (talk) 18:28, 28 September 2016 (UTC)Reply

To clarify, the naturally was an infelicity of my phrasing. Matasovic was saying that the PIE *ey > P(N)C *ē must have come after PIE *ē > PC *ī or else PIE *ey would too have resulted in *ī. Sorry for the confusion. —JohnC5 18:59, 28 September 2016 (UTC)Reply

Old Welsh shows a diphthong for these, at least in spelling, and only later the spelling becomes "o". Is it established that the pronunciation was already /ɔ/ in Brythonic, with a change to a diphthong in Welsh and then back to a monophthong? It seems more likely that it was still a diphthong. —CodeCat 19:48, 26 April 2015 (UTC)Reply

Apologies, I missed this. The diphthong was monophthongized to /ɔ/, but in stressed syllables ending in a light syllable, the /ɔ/ was long (only stressed vowels in light syllables could be long, the stress being on word-final syllables, what used to be penultimate stress before apocope). This long /ɔː/ was then again diphthongized to /au/ in Welsh, but short unstressed /ɔ/ remained. From what I can find, Breton and Cornish never turned /ɔː/ back into /au/.

Each of the Brythonic languages later switched back to penultimate stress later on, at different times, apparently. Anglom (talk) 22:37, 28 June 2015 (UTC)Reply

@Anglom Would you happen to know what went on with the word for "two"? All the Brythonic languages have a diphthong here. —CodeCat 22:34, 29 April 2016 (UTC)Reply

I don't have solid information on mutation of the resonants other than /m/. Welsh obviously shows mutation of /r/ and /l/ but I'm not really sure about /n/. And I read somewhere, I can't remember where, so I will have to find it again, that Breton and Cornish treat /r/, /l/ and /n/ differently in radical and lenited position, with one being longer. I'm not altogether certain about this situation, so I tentatively represent the remaining resonants in lenited position with Greek letters, after the traditional representation of lenited /m/ with <μ>, with no speculation as to their phonetic realization.

The /st/ and /s/ mutations are also tentative, as apparently these mutation paradigms fell apart quite early, according to Schrijver. This seemingly accounts for the incomplete sound shifts of /st/ > /s/ and /s/ > /h/ within the Brythonic languages. Anglom (talk) 22:48, 28 June 2015 (UTC)Reply

What is the Brythonic outcome of initial *sw-? Purely by the table it ought to be *hw- but Welsh seems to have chw- instead. —CodeCat 22:35, 29 April 2016 (UTC)Reply

Generally seems to be given as /hw/, with Welsh /χw/ being a later development. I wondered if it might have given /xw/ originally, but I can't find any sources that support it. Anglom (talk) 23:09, 29 April 2016 (UTC)Reply

I'll admit I'm a novice to these more ancient reconstructed languages, but I was wondering, what is the difference between using đ and th? The Britons used neither when PB was around, so why not use the English? Sorry if it's a stupid/obvious question. Cheers UtherPendrogn (talk) 09:10, 11 August 2016 (UTC)Reply

The English th*, sorry. UtherPendrogn (talk) 09:11, 11 August 2016 (UTC)Reply

Sometimes that's convenient. For example we often use th, kh, sh for transcribing Arabic, as in Uthman, Khalid, Shari'a. Arabic has a sound /h/ but the actual clusters /t+h/, /s+h/ are very rare, so it's harmless to use th for something other than /t/ followed by /h/. But many Indian languages have aspirated consonants, which are similar to /t/ followed by an actual /h/, so it's natural to use th for this combination, and kh for /kh/, rather than the /θ/ and /x/ that they represent in Arabic. This presents a problem with, for example, Urdu, which has native Indian /kh/ plus borrowings from Arabic with /x/. We shouldn't use kh for both of them. In practice this happens in ordinary transcription of place and person names, but it's not suitable for scientific use. And in Australian languages th means something else again: it's a /t/ but pronounced between the teeth. As they usually have neither /th/ nor /θ/, this convention is harmless locally. So conventionally easy transcriptions have their place, but if you want to clearly indicate the sounds of a lesser-known language, something close to IPA is better, so that everyone with some technical knowledge can easily understand what sound is meant. --Hiztegilari (talk) 10:00, 11 August 2016 (UTC)Reply

I see, thanks. UtherPendrogn (talk) 16:43, 11 August 2016 (UTC)Reply

In Proto-Celtic, adjectives matched the noun in gender, case and number. With cases gone in Proto-Brythonic, that still leaves gender and number. What did this inflection look like, and were the different inflectional classes still distinguishable? —CodeCat 17:36, 22 September 2016 (UTC)Reply

AFAIK, at least in nouns, the situation in Brythonic is actually similar to Romance, specifically Modern French: sometimes Breton reflects a different case form than Welsh, for example. It has been concluded from that that Brythonic may still have had a two-case inflection (nominative and accusative) much like Old French and several other medieval Romance languages, although it may have been more limited, or altogether marginal, because the nominative and accusative forms should usually have fallen together due to the complete loss of final consonants unlike in Old French.

(Typologically, it seems that accusative, nominative and genitive are the most tenacious cases, and this triad is reconstructible for Proto-Romance AFAICS, given that Romanian attests the genitive as morphologically distinct in the a-stems, where Slavic interference cannot explain this, even in the article-less form; however, the genitive isn't morphologically distinct in Old French anymore, except in pronouns, though there are syntactic traces, especially in expressions such as la roi cort, la Dieu merci, though these may already have been become fixed or quasi-lexicalised.)

I don't know literature on the problem of noun inflection in Proto-Brythonic, but one approach would be to "reconstruct forward" Proto-Celtic nominal paradigms into Proto-Brythonic mechanically, using the known sound developments, and analyse how many distinctions remain, purely on a phonological basis, and compare these reconstructions to the attested Brythonic material. Though going from the Old French and Old Irish cases, and also from Old English and other (especially West) Germanic languages, my hunch is that the inflectional classes were still distinguishable to a limited extent (consider the vowel changes in Welsh adjectives, which wouldn't occur the same way in i-/u-/consonant-stem adjectives, at least originally), although in the process of disappearing. --Florian Blaschke (talk) 18:48, 28 September 2016 (UTC)Reply

Does "tt" always spirantise into θ? I've noticed it happen a lot, like katta and kritta become kath and kreth (can't add the macrons right now). UtherPendrogn (talk) 11:53, 23 September 2016 (UTC)Reply

I think so, yes, for any tt present in Proto-Celtic. Remember that a PIE tt already became ss in Proto-Celtic, so it doesn't apply to them. —Aɴɢʀ (talk) 11:59, 23 September 2016 (UTC)Reply

Yep, cheers. Shall we add that to the "about" page? UtherPendrogn (talk) 14:11, 23 September 2016 (UTC)Reply

If you like, but it isn't just this one: all of /pp tt kk/ become /f θ x/. —Aɴɢʀ (talk) 08:18, 24 September 2016 (UTC)Reply

@Anglom, Angr, CodeCat This is very much icing on the cake, but would it be possible to have some sort of convention to romanising the letters even further? It would be nice to have a "scientific" way of presenting the words, as we're doing now, by IPA (roughly), and then a smoother "more romanised" way.

I've made this up on the spot, I'm not suggesting it should be exactly like this, but consider this:
Scientific: lėβėrjad
Cleaner: lévéryad, or even just leveryad (I'm not quite sure what sound ė represents!)
Or as has been the convention with Old English and Latin recently, replace those acutes with macrons. lēvēryad.

I've also got an unrelated question:
If i-affection changes one letter, can than letter then change another?
I've made this word up to demonstrate it, but let's say you've got Proto-Celtic mamajos.
Would it become mamė, or would the fact that the j turned the second a into ė lead that ė to turn the first a into an ė too, giving mėmė?
UtherPendrogn (talk) 17:04, 29 September 2016 (UTC)Reply

Proto-Celtic intervocalic -j- wouldn't survive in that position. It might presumably become *mamās or some such form. Aside from that, internal i-affection would mean that the first -a- become raised like the second as well. In a form like *mamabjos, final i-affection would cause the second -a- to become -ė-, while later internal i-affection would cause the first -a- to become -ė-. *mamabjos > *mamėbjos > *mėβ̃ėβ.

<ė> stands for a raised /e̝/, which is slightly higher than a standard /e/ because it facilitates internal i-affection.

I don't really see a point in changing the transcription. Why represent /β/ with <v> when we still have to represent /ð/ as <ð> and /ɣ/ as <ɣ>. The only place where I knowingly deviated from IPA is with vowels, and that's because the reconstructions looked far too awkward that way. Anglom (talk) 21:55, 29 September 2016 (UTC)Reply

But that's my point, they still look awkward. And I'm suggesting replacing all the IPA letters with some sort of convention, but still keeping the original forms. UtherPendrogn (talk) 05:36, 30 September 2016 (UTC)Reply

I have no idea what symbols are the most widely used in the literature on Proto-Brythonic; everything here looks ad-hoc to me. The advantages to using /v/ instead of /β/ are (1) to the best of my knowledge, the reflex is labiodental, not bilabial, in all attested Brythonic languages, so the sound was probably labiodental already in PBr, and (2) ṽ is a precomposed Unicode character that is much easier to recognize than β̃ where it's difficult to see the tilde because the beta is so tall. I believe the usual symbol for a closer "e" is e with an underdot (ẹ) rather than with an overdot (ė), which I've only ever seen in Lithuanian. Using "y" instead of "j" would bring our transcription into line with that already used for Proto-Celtic and PIE, but if "j" is more common in the literature (as it is, say, for Proto-Germanic), then we should stick with it. So, unless the literature clearly militates against it, what I'd prefer is lẹvẹryad. —Aɴɢʀ (talk) 12:55, 30 September 2016 (UTC)Reply

The recent literature uses the IPA symbols, except for /β̃/ which they represent as ⟨μ⟩ or ⟨ṽ⟩ (although the latter with the caveat that it was more likely bilabial).

In the earliest attestations /β̃/ is frequently spelled ⟨b⟩, just as /β/ is (also as ⟨v⟩ and ⟨u⟩). While it could represent a nasalized labiodental fricative, the safest explanation is that it represented the unstable bilabial fricative. It also better explains why in some environments /β̃/ and /β/ became /w/, which I'm led to believe is much more common of /β/ than /v/.

The overdot rather than underdot is just a simplification on my part so that the reader is aware <ė> is raised, <ọ> is lower (/ɔ/). Anglom (talk) 13:35, 30 September 2016 (UTC)Reply

In that case, I guess we should stick with /β/. I'd use ẹ for raised /e/ and ǫ for lower /ɔ/, as is traditional in Romance historical linguistics, or else just use ɔ for /ɔ/. I think to the extent anyone has expectations for what ọ means, they'll probably expect it to mean a closer /o/, not a more open one. —Aɴɢʀ (talk) 14:09, 30 September 2016 (UTC)Reply

ǫ just makes me think nasal vowel. I think the overdot/underdot notation Anglom devised makes sense, once you realise what the placing of the dots signifies. —CodeCat 14:11, 30 September 2016 (UTC)Reply

Is this what all these symbols mean? Great to know. Whatever we decide, we should definitely provide an explanation and justification on here! —JohnC5 14:48, 30 September 2016 (UTC)Reply

I think my point is being missed. All these IPA letters are jarring, most languages only have a handful of "unconventional letters", while Proto-Brythonic could easily give you a word like löđėβ, which just looks silly though of course that's only my opinion'. While from a scientific and linguistic perspective we should use the existing format, I'm proposing a format that would be easier on the eyes and used SECONDARY to the existing format.

If there were a sign where Proto-Brythonic was written, would you prefer "jėdriđoβ mi jüɣöβ̃" , which is the correct linguistic form, or a cleaner "yēdrithov mi yugoṽ"? UtherPendrogn (talk) 16:02, 30 September 2016 (UTC)Reply

wiː dəʊnt gəʊ əˈraʊnd ˈraɪtɪŋ laɪk ðɪs. Though it is technically correct, we have a conventional way of writing and a "linguistic" one (IPA). UtherPendrogn (talk) 16:05, 30 September 2016 (UTC)Reply

Don't ever try to learn Proto-Slavic or Proto-Italic, I would suggest. —CodeCat 17:54, 30 September 2016 (UTC)Reply

While an amusing joke, it doesn't answer my question. Do we plan to implement some further romanisation, or leave it as is with IPA letters all over the place? Both are fine with me, but I was curious. It looks very jarring to have:

Outros becomes Uθr which becomes Uth(y)r.

See what I mean? Latin alphabet, weird mix of IPA, Latin alphabet. And it's not just a matter of Old Welsh being attested in the Latin script, since Proto-Celtic is a reconstruction and the "weirdest" letter there is Phi. And anyone not savvy in IPA can read a Proto-Celtic sentence without too much trouble, while "Uθr" would be very jarring.

UtherPendrogn (talk) 18:00, 30 September 2016 (UTC)Reply

Jarring is, as you said earlier, an opinion. We decide on a convention and stick to it. I personally like this one because it is similar to what we use in PI. Also, we have “weird” characters in Proto-Celtic and Proto-Indo-European as well. The system for PIE laryngeals looks like we're doing ugly calculus, but it's the convention. —JohnC5 18:08, 30 September 2016 (UTC)Reply

Again, I'm not suggesting we replace the current format, just add a second, clearer one. UtherPendrogn (talk) 18:15, 30 September 2016 (UTC)Reply

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘Now that I'm definitely opposed to. It's a protolanguage, it doesn't need two competing orthographies. And it doesn't need an easy-on-the-eyes orthography that one would be comfortable reading a newspaper in. It needs a straightforward, unambiguous one-to-one mapping of symbols to reconstructed sounds. —Aɴɢʀ (talk) 18:37, 30 September 2016 (UTC)Reply

Which would mean writing in IPA. And the vowels are innacurate, as Anglom said. UtherPendrogn (talk) 20:59, 30 September 2016 (UTC)Reply

@Angr, CodeCat, Anglom, is there a language code for Common Brythonic? cel-bry doesn't seem to work, since I'm guessing it's just a family code right now. Artognou is an example of a Common Brythonic name, from Cornwall dated 600 AD. You could maybe call it Proto-Cornish, but it's identical to what the Common Brythonic form would be. Either way, I think we need a code for it. --Victar (talk) 13:35, 22 October 2016 (UTC)Reply

I don't think we differentiate between Common Brythonic and Proto-Brythonic (cel-bry-pro). —Aɴɢʀ (talk) 13:46, 22 October 2016 (UTC)Reply

Oh, right, I guess I can just use cel-bry-pro for attested lemmas. I'll do that than, unless anyone has any other suggestions. Thanks. --Victar (talk) 13:58, 22 October 2016 (UTC)Reply

I'm not aware of it being attested. —CodeCat 13:59, 22 October 2016 (UTC)Reply

Nope, cel-bry-pro only works for reconstructed lemmas. Artognou is attested, so whatever you want to call it. --Victar (talk) 14:01, 22 October 2016 (UTC)Reply

Well, we could either (1) make cel-bry-pro usable for attested lemmas, the way gmq-pro is, or (2) we could call Artognou Old Cornish. —Aɴɢʀ (talk) 14:40, 22 October 2016 (UTC)Reply

It looks like it's pre-Proto-Brythonic, because it preserves the theme vowel of the first part of the compound. It doesn't have the vowel of the second part anymore though. —CodeCat 14:45, 22 October 2016 (UTC)Reply

600 AD seems too late for pre-Proto-Brythonic. It could be that the intermediary thematic vowel was preserved in Proto-Brythonic/Common Brythonic names, like Germanic often does. I definitely don't think it should be called simply Old Cornish. --Victar (talk) 15:50, 22 October 2016 (UTC)Reply

@Angr, CodeCat, so how can we make cel-bry-pro usable for attested lemmas? --Victar (talk) 11:43, 24 October 2016 (UTC)Reply

By removing type = "reconstructed", from its entry at Module:languages/datax, but I'd like wider consensus before I do that. —Aɴɢʀ (talk) 12:55, 24 October 2016 (UTC)Reply

I'm not really sure about this entry. The text itself is in Latin, which points to it being British Latin rather than Brythonic itself, albeit with the name being borrowed from at least pre-syncope Brythonic. The dating doesn't really mean anything for its Brythonic-ness in that case, because we don't know when it was borrowed. The proper Proto-Brythonic form is probably *Arθɣnọu, which can be deduced from Middle Welsh Arthneu and Old Breton Arthnou. Anglom (talk) 22:51, 24 October 2016 (UTC)Reply

@Anglom: Artognou could very well been actually pronounced *Arθoɣnọu, with *rt to *rθ and *og to *oɣ. I think there is a very good argument for at least an early stage of Proto-Brythonic retaining intermediary thematic vowel. The alternative is to say that Latin simply added these vowels on its own volition, which strikes me as improbable, especially since, from what I've seen, they all appear to match with the etymologically correct thematic vowel. Another theory is that these were letter-for-letter transliterations which retained the historical silent intermediary thematic vowel, but that also seems unlikely. --Victar (talk) 00:44, 25 October 2016 (UTC)Reply

I think *Ehuɣen might be a good example of where the intermediary thematic vowel must have been retained, no? --Victar (talk) 02:25, 25 October 2016 (UTC)Reply

Latin didn't add the vowels, rather they borrowed the name before the vowels were lost. Also the name is probably inflected in its attestation, leading one to suppose the full name is *Artognouus. The name itself is Brythonic in origin, but the language it's attested in is British Latin, which survived until the 700s.

Not quite. The -s-/-h- left a hiatus which was usually filled with /j/ except when adjacent to -u-, where it gave /w/. So *Esugenos > *Ehugenos > *Ëugenos > *Ewuɣenos > *Euɣen; the original thematic vowel -u- is still lost and non-syllabic -w- becomes syllabic /u/. Loss of /ɣ/ then caused the syllabic -u- to become non-syllabic -w- in the daughter languages.

The Welsh forms listed on that page do however require *-ɣėn rather than *-ɣen, so we're actually looking at earlier *-ganyos, because earlier *-genis or *-genyos would have given *-ɣɨn instead. Anglom (talk) 01:23, 27 October 2016 (UTC)Reply

I'm glad we at least agree that Artognou is from a form of Brythonic that still retained thematic vowels! I obviously disagree that the name is written in Latin because 1. we have forms of Latinized PBry names and they look quite different, 2. this isn't your typical posthumous veneration, but rather graffiti, and as such, doesn't need to follow strict Latin inflections. Thanks for explaining *Ehuɣen. What about *Karadọg? --Victar (talk) 17:06, 1 November 2016 (UTC)Reply

I forgot to mention, I think PBry lost the final thematic vowel long before losing intermediary thematic vowels (at least in proper nouns), so *Ehuɣenos/*Ewuɣenos possibly never existsed, but instead *Ehuɣen/*Ewuɣen. --Victar (talk) 17:29, 1 November 2016 (UTC)Reply

The medial -a- in *Karadọg isn't thematic and is also pretonic, so it doesn't suffer syncope as bad as thematic vowels do.

I'm not realhly sure what you're getting at. All forms of the name existed, because they follow a chronological path. But the stage of Brythonic we reconstruct has been affected by both apocope and syncope, so *Euɣėn is the only form that should appear. Anglom (talk) 19:04, 1 November 2016 (UTC)Reply

I'm saying personal names don't have to follow the predicted path normal nouns would. I think that if we are to reconstruct PBry personal names, they should retain the thematic value of the first element based on the evidence of names found both in Latin and known PBry attestations like Uindiorix. --Victar (talk) 20:56, 1 November 2016 (UTC)Reply

But they do, as seen by the descendants listed in your entries which all show loss of thematic vowel. Proto-Brythonic is reconstructed strictly backwards from the three attested daughter languages. Anglom (talk) 02:35, 2 November 2016 (UTC)Reply

The attestations in Latin are a window into a period we don't have native attestations, especially graffiti. Whether we call one early PBry and the other late PBry is still up for debate, I guess. I do also maintain that names don't have to always be subject to all the changes spoken language is. --Victar (talk) 17:55, 2 November 2016 (UTC)Reply

@Anglom: Old Welsh Dumnagual is an example of retention of the internal thematic vowel. I can't find the source for it, but it's said to be from the 10th c., so quite late. --Victar (talk) 04:53, 5 November 2016 (UTC)Reply

The thematic vowel should have been -o- though, which means the -a- has to be explained another way anyway. Anglom (talk) 06:59, 5 November 2016 (UTC)Reply

So, what is to be done about this entry? Move it back to Reconstruction space or allow cel-bry-pro in the mainspace like Proto-Norse et al? - -sche (discuss) 14:05, 1 November 2016 (UTC)Reply

Would be a shame to call an attested name a reconstruction. --Victar (talk) 16:45, 1 November 2016 (UTC)Reply

w:Common Brittonic#Sources says, "No documents written in Common Brittonic have been found, but a few inscriptions have been identified. The Bath curse tablets, found in the Roman reservoir at Bath, Somerset, contain about 150 names, about half of which are undoubtedly Celtic (but not necessarily Brittonic). There is an inscription on a metal pendant discovered in 1979 in Bath, which seems to contain an ancient Brittonic curse: Adixoui Deuina Deieda Andagin Uindiorix cuamenai or maybe Adixoui Deiana Deieda Andagin Uindiorix cuamiinai." This suggests to me that Proto-Brythonic, like Proto-Norse, is attested (if only barely), so that cel-bry-pro could be allowed in the mainspace too. —Aɴɢʀ (talk) 17:21, 1 November 2016 (UTC)Reply

Dating and definitions matter too. At what point does Proto-Celtic end and Proto-Brythonic start? Certainly, we seem to treat terms common to Brythonic and Goidelic as Proto-Celtic too, even though they possibly no longer appeared in the common ancestor to all Celtic languages. For Proto-Germanic, we treat terms attested only in West Germanic as valid Proto-Germanic terms, meaning that we consider Proto-West-Germanic to be a dialect of Proto-Germanic. So we could decide that the instriptions are still dialectal Proto-Celtic, too, instead of Proto-Brythonic. —CodeCat 19:31, 1 November 2016 (UTC)Reply

We could, sure, but calling them Proto-Brythonic seems more careful—the more conservative approach—than calling them Proto-Celtic. If "Artognou" is from 600 AD, that's just 200 years before Old Welsh is attested, and 50 years after Wikipedia claims Welsh separated from Proto-Brythonic; it's contemporaneous with Primitive Irish (which pretty clearly isn't Proto-Celtic). If anything, it would probably be most accurate to call Artognou Proto-Southwestern Brythonic, but making a separate code for that, rather than treating is as a dialect of PBr, is probably splitting hairs. —Aɴɢʀ (talk) 19:57, 1 November 2016 (UTC)Reply

So we know, then, that the first o of this word disappeared in those 200 years? —CodeCat 21:12, 1 November 2016 (UTC)Reply

The changes from Primitive Irish and Old Irish are just as dramatic. --Victar (talk) 21:22, 1 November 2016 (UTC)Reply

Losing one vowel isn't particularly dramatic. I just wanted to be clear that no later Brythonic language attests any of these vowels. Because if they do, then we have to readjust our reconstructions. —CodeCat 21:29, 1 November 2016 (UTC)Reply

I've changed it to allow entries in the mainspace. - -sche (discuss) 21:28, 1 November 2016 (UTC)Reply

I would just like to go on record as saying I am uncomfortable calling these entries Proto-Brythonic, merely because they do not show the Proto-Brythonic sound changes that characterize the language. Something like Early Brythonic/Brittonic or British Celtic would be a better description. Anglom (talk) 02:11, 2 November 2016 (UTC)Reply

@Anglom: I suppose it would be possible to separate reconstructed Proto-Brythonic from the barely attested language that Wikipedia calls Common Brittonic (other possible names being Common Brythonic, Old Brythonic/Brittonic, and British), but I can't say I'm thrilled about the idea. Would you say the "Adixoui Deuina" quote given above shows the sound changes that characterize PBr.? —Aɴɢʀ (talk) 17:47, 2 November 2016 (UTC)Reply

Not really, no. The word-final vowels show that it's before apocope and -nd- shows nasal assimilation hasn't happened yet. Anglom (talk) 00:24, 3 November 2016 (UTC)Reply

@Angr, CodeCat, Anglom I'm still struggling how to enter PBry personal name entries with early Latin attestations. Are *Broxoβ̃aɣlos and *Bodoɣnọw a possible way to go? --Victar (talk) 20:39, 4 November 2016 (UTC)Reply

@Victar: I don't think the Latin forms are not evidence of a Brythonic retaining Celtic desinences (in *Broxoβ̃aɣlos). Latin certainly could add its own endings to borrowed terms. —JohnC5 06:38, 5 November 2016 (UTC)Reply

@JohnC5: I suppose my thought is, depending on the date of the sources, to reconstruct it as Early Brythonic, with Celtic desinences, in line with Uindiorix. --Victar (talk) 06:55, 5 November 2016 (UTC)Reply

But maybe Celtic desinences where dropped from the time of Uindiorix (2nd-4th c.), to the period most of the sources I've found (5th-6th c.). --Victar (talk) 07:02, 5 November 2016 (UTC)Reply

So other than that, what are your thoughts? --Victar (talk) 07:09, 5 November 2016 (UTC)Reply

Also *Kunowalos. I'm not familiar enough with comparative Celtic chronology. —JohnC5 07:11, 5 November 2016 (UTC)Reply

I found a really great journal on the retention of the thematic vowel in the first element of PBry personal names: Old Welsh Dinacat, Cunedag, Tutagual: fossilised phonology in Brittonic personal names. To quote, "In absolute chronological terms the weakening and loss of composition vowels seems to have been completed by the latter part of the sixth century." So it is chronologically correct for Artognou to retain the -o- and is not an archaism. --Victar (talk) 08:46, 5 November 2016 (UTC)Reply

When exactly is our PBr supposed to represent? —JohnC5 17:20, 5 November 2016 (UTC)Reply

Well, I would imagine you want to reconstruct a form which is representative of all its known attestations. So, *Jʉðnerθ is more than likely a late PBr creation, and this should be reconstructed as such. *Bodoɣnọw, however, is probably inherited straight from PCelt, and as such, has very early attestations; you wouldn't want to say Bodognous descends from *Bodnọw. --Victar (talk) 20:45, 5 November 2016 (UTC)Reply

@Anglom, I think this point needs to be addressed as well, that when creating a reconstruction, it needs to be basis of the Latin forms as well. Bodognous clearly has gone through PBr specific changes, so it is indeed descended from some form of PBr, be it early PBr, pre-PBr, or what-have-you. --Victar (talk) 17:28, 6 November 2016 (UTC)Reply

No, it doesn't. The Latin forms are borrowings, not descendants, let alone the fact that they were borrowed before the sound changes leading to Proto-Brythonic were complete. The Latin forms have no bearing whatsoever on Brythonic reconstructions, except to confirm them. Latin forms of British Celtic names can't even be used to reconstruct Brythonic terms at all if no Brythonic language attests it as well. Anglom (talk) 18:24, 6 November 2016 (UTC)Reply

Again, I'll have to respectfully disagree on this. To say one can't reconstruct lemmas from Latin borrowing makes absolutely no sense to me. --Victar (talk) 19:23, 6 November 2016 (UTC)Reply

I'm not really sure you understood the conclusion of the article: the author states that the thematic vowels had been lost phonologically, but were re-introduced in literature based on Latin forms of the same names borrowed earlier, which explains why the thematic vowels don't actually match what we would expect them to be if they were inherited. Anglom (talk) 04:47, 6 November 2016 (UTC)Reply

That may be the case with much later forms. In the example I give above, the point I was making was Artognou, which is dated to 500-599 CE, falls still within the period that the -o- was only beginning to be weakened, but not yet lost. --Victar (talk) 14:36, 6 November 2016 (UTC)Reply

There is no point to that to argument however because that form of the name is only attested in British Latin, literally "PATERN COLI AVI FICIT ARTOGNOU". The name was borrowed from British Celtic, but we only know it was borrowed before syncope, whether or not that stage was contemporary with the Latin attestation can't be proven by this attestation alone. Anglom (talk) 16:07, 6 November 2016 (UTC)Reply

Well, that's a different argument to the point I was making, and one we've discussed above, but I agree that we need more examples, which is what I've been working towards with adding new entries. --Victar (talk) 16:45, 6 November 2016 (UTC)Reply

I realize you're trying to provide examples of it, but honestly the only forms that maintain thematic vowels are Latin forms of Brythonic names or forms influenced by Latin renderings. Latin forms of British Celtic names are their own beast, really. Proto-Brythonic quite simply no longer maintained internal thematic vowels, even in inherited personal names, it was a regular sound change. Anglom (talk) 17:02, 6 November 2016 (UTC)Reply

Well, you're welcome to hold to your opinion on that, and I respect that, because I respect your work. I disagree though, and think the paper above, in particular the quote cited, along with a litany of examples given, if not point to the contrary, at least welcome the alternative. --Victar (talk) 17:20, 6 November 2016 (UTC)Reply

@Angr, looking to reignite a 3 year old debate? --{{victar|talk}} 17:29, 13 July 2019 (UTC)Reply

Not especially. —Mahāgaja · talk 17:45, 13 July 2019 (UTC)Reply

A completely random comment, but I'd love a comparative table of Proto-Brythonic vowels and their reflexes in the Brythonic languages. Either here (the non-talk page) or on Wikipedia. —CodeCat 22:54, 24 October 2016 (UTC)Reply

Seconded. I'd like it to be here. :) —JohnC5 22:56, 24 October 2016 (UTC)Reply

I made a basic table now, using only the reflexes of the first/single syllable and not conditioned by consonants likely to alter them (like w or voiced fricatives). I just looked for words and wrote down the spelled reflex of the descendants. —CodeCat 23:28, 24 October 2016 (UTC)Reply

Are we also going to have PC-to-PB and PB-to-Brythonic languages tables for consonants? —JohnC5 20:20, 25 October 2016 (UTC)Reply

It's looking fantastic so far, by the way! --Victar (talk) 20:33, 25 October 2016 (UTC)Reply

Yes, this all looks awesome. —JohnC5 20:38, 25 October 2016 (UTC)Reply

While we're reworking things, can we revisit the use of *j over *y? Was there some special logic behind it, because it seems odd that we go from *y in PCelt to *j in PBry to y in Breton, Cornish and Welsh? --Victar (talk) 23:29, 25 October 2016 (UTC)Reply

Welsh uses i, as did Old Cornish. I don't know about Old Breton. —CodeCat 23:35, 25 October 2016 (UTC)Reply

To judge from the forms at Reconstruction:Proto-Brythonic/jowank, Middle and Modern Breton and Middle and Modern Cornish use y. No Brythonic language uses j. What does the literature use? —Aɴɢʀ (talk) 23:53, 25 October 2016 (UTC)Reply

Usually i̯ which is common among older IE literature, but newer works like Schrijver tend to use IPA characters for consonants. Anglom (talk) 02:16, 27 October 2016 (UTC)Reply

The logic is that we use ⟨j⟩ to transcribe /j/, I guess? Anglom (talk) 02:16, 27 October 2016 (UTC)Reply

Yeah, if that's the only logic behind it, I vote that we change it to *y instead. Does anyone have any reason not to? --Victar (talk) 05:37, 27 October 2016 (UTC)Reply

Is there a reason to do so? Anglom (talk) 13:26, 27 October 2016 (UTC)Reply

I don't have particularly strong feelings, though I prefer j "just because" (I wish all languages used it). A real argument, though, is that the Brythonic languages generally use the letter y to stand for a vowel, and for us to use it differently might be confusing. Consider, for comparison, the situation with Slavic. —CodeCat 13:30, 27 October 2016 (UTC)Reply

Maybe because PB has both a palatal glide j and a high front rounded vowel ü it's best if we avoid using the character y at all, since it could be ambiguous as to which of those sounds it's meant to represent. Granted, in practice you could probably usually tell from context, as in the conventional English transliteration of Russian: южный (yuzhnyy), but why take the chance? —Aɴɢʀ (talk) 14:31, 27 October 2016 (UTC)Reply

According to *gaβl, Proto-Brythonic still had a dual form, which was characterised by i-affection in ā-stems. Should we include this in entries? How much of the dual remains in the Brythonic languages? —CodeCat 13:51, 26 October 2016 (UTC)Reply

I'm not familiar with the under-dot notation in use here. Is this an alternative for the "lowered" articulation diacritic used in the IPA? ‑‑ Eiríkr Útlendi │^{Tala við mig} 21:18, 2 November 2016 (UTC)Reply

Yes; ọ stands for /ɔ/. I would prefer to use ɔ in reconstructions, but I don't know if anyone else would. —Aɴɢʀ (talk) 21:39, 2 November 2016 (UTC)Reply

ɔ would be a bit harsh, I think. I have no idea where ọ comes from, but I would have expected ŏ instead, as in Latin dictionaries. --Victar (talk) 21:56, 2 November 2016 (UTC)Reply

But ŏ implies a short vowel that contrasts with a long ō; PBr. didn't have distinctive vowel length. (And even if it still had noncontrastive vowel length, this vowel was probably rather long, since it comes from *ā and *au, i.e. a long vowel and a diphthong.) Another alternative is ǫ, as in Old Norse and commonly in Proto-Romance, but others object that it looks like a nasalized vowel (as in Proto-Germanic). —Aɴɢʀ (talk) 22:37, 2 November 2016 (UTC)Reply

Right, but it isn't anything new to associate a short vowel with ɔ. --Victar (talk) 23:12, 2 November 2016 (UTC)Reply

• To clarify my understanding, under-dot ⟨ọ⟩ is equivalent with lowered or raised , and is thus smack-dab in between regular and regular ?

Or, is under-dot ⟨ọ⟩ instead even lower, and equivalent with regular ?

Once that is clarified, is under-dot ⟨ọ⟩ contrastive with either or ? And how are these spelled in entries here? ‑‑ Eiríkr Útlendi │^{Tala við mig} 23:19, 2 November 2016 (UTC)Reply

⟨ọ⟩ is lower than ⟨o⟩. They are contrastive, because we wouldn't be able to reconstruct them otherwise. The comparative method reconstructs only contrasts. —CodeCat 23:22, 2 November 2016 (UTC)Reply

• Is ⟨ọ⟩ also higher than ? Or are these equivalent? ‑‑ Eiríkr Útlendi │^{Tala við mig} 16:52, 3 November 2016 (UTC)Reply
  • You're mixing Brythonic spelling with IPA now. —CodeCat 17:05, 3 November 2016 (UTC)Reply

• Yes. That is wholly intentional. I'm trying to ascertain what the reconstructed Proto-Brythonic phonetic value is for the glyph ⟨ọ⟩. Per your previous comment above, it appears that the phonetic value of the glyph ⟨ọ⟩ is lower than the phonetic value for the glyph ⟨o⟩. I infer that the glyph ⟨o⟩ is intended to represent something close to the phonetic value . It is still not clear to me what the phonetic value is supposed to be for the glyph ⟨ọ⟩, only that it is presumably lower than .

Ultimately, the outcome of this discussion should probably be included somewhere in the Wiktionary:About Proto-Brythonic page itself. ‑‑ Eiríkr Útlendi │^{Tala við mig} 17:30, 3 November 2016 (UTC)Reply

The phonetic value may not be reconstructable, nor even relevant. Since we know that ⟨ọ⟩ developed from Proto-Celtic ⟨ā⟩, it is reasonable to think that it's lower than ⟨o⟩, but that's absolutely not a necessity. What we do know is that they were distinct as phonemes. —CodeCat 17:44, 3 November 2016 (UTC)Reply

@CodeCat, Anglom, Angr, as I try out more and more sentences, it seems increasingly that Proto-Brythonic is basically odd-sounding Welsh. I've not seen a single form that's radically different from Old Welsh. Indeed, "Is ti βɨw?" is practically identical to Old Welsh "Is ti'n fyw" (which is very different from the modern "Wyt ti'n fyw"). Are there any big differences? UtherPendrogn (talk) 22:06, 26 November 2016 (UTC)Reply

That's the point though. The farther back in time you go the closer they are until they reach one language, which we've reconstructed as Proto-Brythonic. Anglom (talk) 00:40, 27 November 2016 (UTC)Reply

Fair enough, that makes sense. But I still feel like, orally, a Welshman would have no trouble listening to the Proto-Brythonic, whereas a Breton or Cornishman might. Especially if word order is the same, and some things (vowel + yn contracted as vowel'n) are the same.

By the way, would there be any interest in reconstructing the intermediary between Proto-Celtic and Proto-Brythonic? Some entries of PB have "from earlier x", which is so radically different (and shows interesting changes) that it might warrant its own language code and lemmas. UtherPendrogn (talk) 08:29, 27 November 2016 (UTC)Reply

@CodeCat, Anglom, Eirikr, Angr. Why is e sometimes e and sometimes ɛ? Neither length nor quality seem to correspond to stress, position, openness, proximity to a palatal, velar, liquid, resonant, nor the PIE or Proto-Celtic form of the vowel. In fact it appears to be absolutely random and often has nothing to do with the reflexes' pronunciation. So what determines it? 2A01:E35:8A85:7E40:84A9:43FC:689:E26B 15:44, 30 August 2017 (UTC)Reply

See the main page, which shows the reflexes. —Rua (mew) 15:48, 30 August 2017 (UTC)Reply

Earlier words I used /e/ in IPA, because my sources at the time led to me believe that's what it was, but a newer work which more or less became my bible used /ɛ/ so I used that instead in newer pages. All vowels were short except when stressed and not blocked by a long consonant or cluster, so far as I know. Anglom (talk) 16:49, 30 August 2017 (UTC)Reply

Is that an inclusive and? As in vowels that are blocked yet stressed are still short? 2A01:E35:8A85:7E40:84A9:43FC:689:E26B 17:26, 30 August 2017 (UTC)Reply

Yes. Anglom (talk) 01:54, 31 August 2017 (UTC)Reply

@CodeCat, Anglom, Angr, Victar Hi everyone, I've been learning a lot about constructing PBR in the recent months, but I am utterly, completely, and totally stumped with regards to the verbs. One of my main confusions lies with how many and which tenses and moods it possessed. I've noticed there's a non-finite heading in the conjugation tables, but with nothing below it - and also, that both literary Welsh and Breton possess the conditional mood, though Welsh has merged it with the imperative and subjunctive imperfect, Breton has it preserved in both present and imperfect tenses. Does this imply Breton has carried on something from PBR that Welsh has lost? Or has it gained such a distinct mood from the French influence? I also noticed it has a future tense in the indicative - once again, is it simply something Welsh has lost, or is it from French? How are we to tell? Also, the merger of the Welsh imperfect tenses from both the subjunctive and indicative moods is causing me issues when re-constructing verbs, is there a way to tell which parts of the verb come from which? Would comparison with Irish help perhaps? The lack of a subjunctive in Breton is also difficult, because I have nothing to compare the Welsh to - I've been somewhat reluctant to use Cornish so far as a reference, as it's essentially a re-constructed language itself, with vocabulary, grammar, and spelling taken from various periods throughout its history. Is it wise to be cautious with regards to Cornish?

Sorry, so many questions, I know, but hopefully we can work some of the answers out here.

On another note, I was playing around with *prɨnad for a while today, and I had a go at trying to re-build the indicative imperfect from Welsh and Breton, comparing it to the patterns we already have established in PBR. Now I haven't yet compared it to Proto-Celtic, but this is what I came up with:

• 1sg: *prɨnun? = prynwn, prenen - Welsh w equates to PBR u, which doesn't match the origins of Breton e. So I'm not too sure on this either.
• 2sg *prɨnɨð? = prynit, prenes - the endings for the second person don't quite match, which is common in both the present (pryni, prenez), and the preterite (prynaist, prenjout), so I decided to follow the pattern (*prɨnað, *prɨnėssɨð) that appears to exist in PBR and re-construct a ð ending. This one I'm not so confident in, naturally.
• 3sg *prɨnė? = prynai, prene - have re-constructed ė as the ending, because Welsh stressed ai came from ė or ei, and Breton e's origins overlap only with ė.
• 0sg *prɨnɨd? = prynid, prened(?) - not sure whether this is the impersonal plural or conjunct singular, as Welsh and Breton only inherited one form of impersonal per tense, however, they seem to have inherited the singular form from both indicative present and preterite, so my guess is they would have inherited the singular form from the imperfect too.
• 1pl *prɨne/ėmm? = prynem, prenemp - this one seems to follow the mm pattern you see in the present and preterite (*prɨnamm, *prɨnassomm). Breton has a habit of adding a p after what was once a double m. French influence maybe?
• 2pl *prɨne/ėd? = prynech, prenec’h - like the equivalents in the preterite and present, both Welsh and Breton seem to have turned the final d into a PBR h or x. So I'm using the pattern from PBR (*prɨnad, *prɨnassed) to re-construct a d ending for this word too.
• 3pl *prɨne/ėnt? = prynent, prenent - this follows the nt pattern (*prɨnant, *prɨnassont). It's nice when words do what you expect them to.
• 0pl *? = ? - I have no idea what the reconstruction for this part of the verb would be.

I've likely done a lot of things incorrectly, I know, please do tell me where I've gone wrong.

On a totally different note, the Wikipedia page "Brittonic Languages" has a chart of how consonants have been inherited from Celtic through PBR to the modern day, much like we already have on the about page for vowels. I think it would be beneficial to have such a chart here to accompany the one for vowels.

Seems that most pages lack this tag. We'll continue working on it. Kwékwlos (talk) 09:19, 24 September 2018 (UTC) Can anyone listen? Kwékwlos (talk) 04:33, 20 October 2018 (UTC)Reply

There is a serious issue with Category:Proto-Brythonic lemmas. Under "Recent additions to the category" and "Oldest pages ordered by last edit", it always lists the same two articles. Why? These two articles, unlike all the others in the category, ar placed in the main namespace, rather than the "Reconstruction" namespace (for terms that ar not attested, but ar reconstructed based on various evidence; for e.g. Proto-Indo-European or Proto-Celtic). This is annoying to me because, as someone who is fascinated with reconstructed languages, i like to look at the lists of reconstructed lemmas, and i like to follow what new lemmas ar added. However, i can't really follow that if it doesn't giv me an honest answer to what new new lemmas hav been added. Therefor i strongly recommend that we be consistent and place all members of Category:Proto-Brythonic lemmas in the Reconstruction namespace. Also, that is more appropriate, because Proto-Brythonic (PB) is, indeed, unattested; the PB era ended about AD 500-600, while the earliest attested Welsh isn't til ~800.

Btw, there should also be a Wiktionary Category titled "Reconstructed lemmas", for all reconstructed lemmas. Said category would hav as a subcategory, Category:Reconstructed terms by language (which includes e.g. Category:Latin reconstructed terms or Category:Old English reconstructed terms).--Solomonfromfinland (talk) 02:06, 10 September 2019 (UTC)Reply

@Anglom, Rua: It occurs to me that if Proto-Brythonic has already lost final syllables (e.g. *kaθ, not *kaθā), then it must have had a fortis /l͈ r͈/ that developed into Welsh /ɬ r̞/, since the distribution of the latter is determined by the presence vs. absence of a preceding vowel before the loss of final syllables, e.g. *merx lɨdan > merch lydan (“wide daughter”) vs. *marx l͈ɨdan > march llydan (“wide horse”). The contrast between /l͈/ and /l/ and between /r͈/ and /r/ must have then been lost in Southwestern Brythonic, rather than arising in Welsh alone. For transcription we could use *r͈/*l͈ or *ʀ/*ʟ (personally I'd prefer the latter). —Mahāgaja · talk 19:56, 8 January 2020 (UTC)Reply

Actually, I see we're already using *ll/*rr word-internally, so we could use that notation too, but it needs to be written word-initally as well. —Mahāgaja · talk 20:02, 8 January 2020 (UTC)Reply

By context, it will always be fortis word-initially. I eventually adapted ̆r, l̆, n̆ in my own transcriptions to denote lenis r, l, n, since Celtic lenis consonants were originally short or half-long, but by that time I'd already made too many entries for it to be an easy transition. Anglom (talk) 03:17, 9 January 2020 (UTC)Reply

Yes, forms in isolation will always be fortis word-initially, but I still think it's worth marking them as such to remove ambiguity. I don't really like ̆r, l̆, n̆ for the lenis forms (and Proto-Brythonic does seem to have merged fortis and lenis /n/ word-initially anyway). I think the easiest solution would be to just move thinks like *lɨdan to *llɨdan, but I'm open to other options. —Mahāgaja · talk 09:33, 9 January 2020 (UTC)Reply

In the absence of further discussion, I'm going to start implementing this. I'll leave hard redirects from the old spellings. —Mahāgaja · talk 20:18, 31 January 2020 (UTC)Reply

@CodeCat, Anglom, Angr, Rua It looks as though there are some Welsh plurals in the Nouns section that aren't correct. from what I see now:

• tiredd should be tiroedd, and *moredd should be moroedd
• hyntau should be hyntiau (maybe this entry should be changed to avoid the irregular -iau plural, perhaps to PCel *genus/*genowes, PBry *gen/*genow, Welsh gên/genau of the same declension pattern)

Are these entries using older instances of these words (maybe from Middle Welsh)? --Olster21 (talk) 20:07, 11 September 2022 (UTC)Reply

On the topic of the -oedd plural in Modern Welsh, what changes turns pCel short *e/*i into welsh 'oe'? For example, in -oedd and words such as Welsh troed < PCel *tregess and Welsh ddoe < PCel *desi. — This unsigned comment was added by Olster21 (talk • contribs) at 16:30, 11 September 2022‎ (UTC).

Indeed, Middle Welsh tireð and hynteu and Early Modern Welsh moredd are found alongside the modern plurals. I don't know the exact development by which *-eɣ- and *-ey- (< *-es-) become oe, but it's probably similar to the change of Proto-Celtic *ē to Welsh wy or Vulgar Latin ẹ to French oi (formerly pronounced /o̯e/). —Mahāgaja · talk 19:41, 11 September 2022 (UTC)Reply

@Mahagaja Thank you, that's quite helpful. Do you think that the Modern Welsh forms should be edited in? Also, you're completely right about that origin for 'oe', accounts for 'troed' (<*tregess) and 'ddoe' (<*desi).

I'm a bit confused about the -yo and -yā declensions shown in the Nouns section to: they're lacking representations on the tables, and they also don't seem to have much distinguishing them from -o and -ā nouns. The plural endings are sort of all over the place in Modern Welsh ('bugail/bugeiliaid', 'craidd, creiddiau', 'mynydd/mynyddoedd'). Is there a word that you think can be used as an example for these patterns? Olster21 (talk) 20:35, 11 September 2022 (UTC)Reply

The plural endings are all over the place because the original plural ending would have resulted in a form identical to the singular: *boukolyoi would have given bugail, *kridyā would have given craidd, and so on, so all these nouns got new plural endings from other sources. The only way you can tell they aren't simple o- and ā-stems is the internal i-affection of the vowel. —Mahāgaja · talk 20:56, 11 September 2022 (UTC)Reply