User talk:Arafsymudwr

Welcome!

Hi! You're clearly not new and know what you're doing, so I'll just say hi! Good work with the Welsh entries so far! – Guitarmankev1 ^(talk) 20:03, 27 October 2023 (UTC)

Thanks! New to Wiktionary but not to Wikis. Looking forward to seeing you around :) Arafsymudwr (talk) 19:03, 30 October 2023 (UTC)

May I also respectfully draw your attention to Wiktionary:Requested entries (Welsh)? We don't have many Welsh speakers around, but the requests page is quite tiny too. You could clear in a weekend mate :) Equinox ◑ 00:15, 7 January 2024 (UTC)

Sure, I'll look at it now! Arafsymudwr (talk) 12:07, 7 January 2024 (UTC)

Cadw draenog yn ei boced

Hello - I would firstly like to thank you for your great work in creating these Welsh pages!

This idiom, however, is correctly 'cadw draenog yn ei boced'. The one under the entry 'cadw draenog yn boced un' is incorrect gramatically in its lack of aspirate mutation after 'yn' (in the sense of 'in') and also its use of 'un' in a manner equivalent to the English. I think the current wording is a literal (mis)translation, or a misremembered version of the correct idiom.

Cofio n 2A00:23C7:7C9B:AB01:8CB8:97E1:9C41:CFC3 15:35, 8 January 2024 (UTC)

Thanks! I was sure that “un” does have uses like English “one” but just checked and yeah not in a genitive sense whoops.

About the mutation, do you mean lack of nasal mutation? Because that's what I would expect. Arafsymudwr (talk) 16:55, 8 January 2024 (UTC)

Nasal mutation is most likely what they meant: ‘ym mhoced’ would certainly be correct in this instance. 2A00:23C7:7C9B:AB01:1C71:4047:8AC4:8F0B 07:58, 9 January 2024 (UTC)

Welsh rhymes

Why do you only put the rhymes from the northern pronunciation? Stríðsdrengur (talk) 18:38, 9 January 2024 (UTC)

Because it is an established practice on Wiktionary. See e.g. grudd and rhudd which I have never edited and also show the northern rhyme only. The links for the rhyme lists include the necessary information about Southern rhymes. Arafsymudwr (talk) 18:54, 9 January 2024 (UTC)

Taro

Shwmae gyfaill,

I notice that the initial and already-existing article for 'taro' had the incorrect verbal root: it is 'traw' (and irregular). See: https://www.geiriadur.ac.uk/gpc/gpc.html?trawaf

I have therefore changed the stem in the articles you have recently made, and hope that I have not stepped on your toes!

Also, the idiom is 'taro bargen' rather rhan 'taro'r fargen' (see Geiriadur Prifysgol Cymru).

Thanks for all the good work you've done recently: you've increased the number of Welsh entries monumentally, and many, many people are grateful for your efforts! 2A00:23C7:7C9B:AB01:9014:85B3:58F6:1893 14:48, 19 January 2024 (UTC)

It's ok - I copied that from taro and later noticed that had the same error! Thanks for picking it up :) Arafsymudwr (talk) 14:50, 19 January 2024 (UTC)

aspirate mutation of yr?

My very limited knowledge of Welsh makes me very confused about mutation (like every other learner, I imagine), but can there really be aspirate mutation of the article in any grammatical circumstance? You've included mutation tables for yr undydd and yr unnos. --Hiztegilari (talk) 17:39, 19 January 2024 (UTC)

You're correct, there cannot be. Arafsymudwr (talk) 17:41, 19 January 2024 (UTC)

Equative

Hello there, I was wondering if you could help with something, as you seem to be well-versed in the templates used here.

For Welsh adjectives, most of the equatives are incorrect: they are automatically rendered as [cyn-] + [adjective, with correct soft mutation], whereas there should be and [-ed] after the adjective, thus giving:

[cyn-] [adjective, with soft mutation] [**-ed]**. Without the necessary -ed ending, it is incorrect, and worse, implies 'former' (cf. cyn-fyfyriwr - alumnus/former student).

That is to say, for 'bras', the equative, thus formed, is 'cyn frased', not 'cyn fras', which means nothing.

Somebody would therefore have to adjust the template, and I feel that you might know how to do that!

• I should add also that the 'cyn' is not strictly required when using the equative degree, but is usually included.

Warm regards 2A00:23C7:7C9B:AB01:DB6:CE07:AD45:7BF3 14:36, 26 January 2024 (UTC)

Most of these templates are locked so only accounts with a certain level of clearance (that I don't have) can edit them. I think you should either talk to Mahagaja (Wiktionary administrator who recently edited that template) or post your question in the grease pit (forum dedicated to discussing templates).

Sorry I can't be more helpful :) Arafsymudwr (talk) 15:07, 26 January 2024 (UTC)

psygodle & na'i gilydd

Hey. First, just to say "Diolch yn fawr" for all the work you've done on the Welsh entries.

Second, I noticed a typo in the entry psygodle. I'm guessing it's best to redirect and get it deleted, as its not a common spelling mistake.

Also, I wondered whether na'i gilydd really needs its own entry. It's just na (“than”) + ei gilydd (“each other”). Or is there some idiom I'm not remembering that uses it?

Thanks again! Llusiduonbach (talk) 21:23, 19 February 2024 (UTC)

Duw duw! Yes, it's redirected now.

About na'i gilydd, I feel there is more point to having it than creating a page for every mutated form of a word like I see getting created all over the place, but I wouldn't object to you binning it. Arafsymudwr (talk) 21:37, 19 February 2024 (UTC)

Thanks. I agree with you the mutated forms are not priority, especially as they're findable in search results. It's a shame the same can't be said for mutated non-lemma forms and that there's not a bot that can automatically create these entries as for other languages. Llusiduonbach (talk) 13:40, 20 February 2024 (UTC)

Links to Middle Welsh terms with the dot

Hi - following on from what you said at WT:RFDN, it sounds like the combining dot should be ignored in link targets for Middle Welsh, in a similar way that macrons and breves are for Latin (e.g. the input {{m|wlm|teẏrn}} should output teẏrn). Are there any other orthographical differences that need to be accounted for? What about Old Welsh? Theknightwho (talk) 14:14, 5 March 2024 (UTC)

@Theknightwho hi, I don't have much to do with either Middle or Old Welsh actually - mostly confined to recategorising terms mistakenly entered as Modern Welsh. But afaik Middle Welsh (as normed by modern scholars) also uses the ð symbol, which should be redirected to d. Although it seems that this is already accounted for.

Btw, on an unrelated topic, could we see about having a "literary form of" template made? Welsh could use such a template a lot, and atm manually adding "literary" to the "form of" template doesn't automatically sort an entry into the literary terms category. See: rhedyn y dwfr and dwfr where two different workarounds have been attempted. Arafsymudwr (talk) 16:14, 5 March 2024 (UTC)

@Arafsymudwr Thanks - these seem like odd decisions to me, especially the hard redirects, because they're somewhat different from ignoring diacritics. @Mahagaja Could you please elaborate? Also pinging @Benwing2, who may have views on this. Theknightwho (talk) 16:18, 5 March 2024 (UTC)

ð → d doesn't involve a hard redirect; rather, using ð in a Middle Welsh link automatically links to the term with d. Thus, typing {{m|wlm|dyð}} links to [[dyd#Middle Welsh]] automatically; it's a kind of diacritic stripping. As for dotted y, I don't think we should be using it at all. I've only seen it in Morris Jones's book as way of indicating when Middle Welsh y corresponds to Modern Welsh i, but I've never seen it used in editions of Middle Welsh texts or in other pedagogical works such as Evans's grammar – both of which do use ð. —Mahāgaja · talk 16:24, 5 March 2024 (UTC)

@Mahagaja Thanks. I should've been clearer: that page says that ⟨ỻ⟩ and ⟨ỽ⟩ are handled via hard redirects, which is what seemed wrong to me. Theknightwho (talk) 16:26, 5 March 2024 (UTC)

@Theknightwho: yes; I wrote that in WT:AWLM because AFAIK those letters aren't used in any language other than Middle Welsh, so a hard redirect would do no harm. Having a separate page for ỻyuyr that just says it's an alt form of llyuyr seems like a waste of the reader's time; taking them straight to llyuyr is more efficient. And since ỻyuyr won't be a word in any other language, the hard redirect won't be taking someone to the wrong entry. —Mahāgaja · talk 16:36, 5 March 2024 (UTC)

@Mahagaja Can ⟨ll⟩ occur outside of the digraph? Because if so then it's probably not a good idea. Theknightwho (talk) 16:46, 5 March 2024 (UTC)

Not an expert, but I believe that callon is an example of just that. In Modern Welsh, calon is pronounced with a short /a/ even in dialects that normally have a long vowel in stressed open syllables. Arafsymudwr (talk) 16:49, 5 March 2024 (UTC)

Having hard redirects from ỻ spellings to ll spellings implies only that ỻ always corresponds to ll, not that ll always corresponds to ỻ. —Mahāgaja · talk 17:02, 5 March 2024 (UTC)

I'm not keen on hard redirects because it implies they're exactly equivalent when they aren't. Theknightwho (talk) 18:04, 5 March 2024 (UTC)

We could also simply say that entries shall not use ⟨ỻ⟩ and ⟨ỽ⟩, just as we already don't use wynn for Old English or long s for any language (I think). —Mahāgaja · talk 16:45, 5 March 2024 (UTC)

@Arafsymudwr In terms of a literary form of template, that's definitely possible. Please see the new {{literary form of}}. Theknightwho (talk) 16:25, 5 March 2024 (UTC)

Thanks, much appreciated! Arafsymudwr (talk) 16:33, 5 March 2024 (UTC)

@Mahagaja I tend to dislike hard redirects in the mainspace although I understand some people use them as an easier way of creating an entry than a soft-redirect entry. I think however your suggestion that ⟨ỻ⟩ and ⟨ỽ⟩ should simply not be used is a good one; it's unlikely someone will want to go out of their way to type them and unless they frequently show up in modern editions found on the Internet (where someone might reasonably cut and paste a term with those characters in them), there's no reason to create redirects. Benwing2 (talk) 23:10, 5 March 2024 (UTC)

I've updated WT:AWLM to state that ⟨ỻ⟩ and ⟨ỽ⟩ are not used. I don't think we have any hard redirects using them at this point. —Mahāgaja · talk 07:36, 6 March 2024 (UTC)

IPA for malltod

Hi. Just checking why you reverted malltod. Phonolgically it's /ˈmaɬdɔd/ but phonetically [ˈmaɬtɔd]. Thanks. Llusiduonbach (talk) 10:58, 6 March 2024 (UTC)

Phonetically the fortis/lenis distinction in Welsh is primarily aspirate vs unaspirated. The lenis stops are unvoiced word-finally in isolation, and voiced when a sonorant follows Arafsymudwr (talk) 12:00, 6 March 2024 (UTC)

Have you got any references for that? I'm thinking here the devoicing begins during the consonant rather than the vowel, e.g. /taːd/ is more [tʰaːd͡d̥] rather than [tʰa͡ḁt], but I'd love to see a study about that in particular (and learn more) if I'm wrong. I'm guessing a number of a number of phonetic transcriptions may have to be changed if I am. Thanks. Llusiduonbach (talk) 13:14, 6 March 2024 (UTC)

@Llusiduonbach sure, check out Grawunder and Asmus and Asmus, Jaworski and Baran. John Well’s paper on the relationship between Welsh phonology and orthography takes it for granted too (not open access, but I have a copy if you’d like it). The Journal of the International Phonetic Association’s description of Northern Welsh phonology seems to go even further and treat voice in plosives as not even phonemic.

I can’t find a description specifically for Southern Welsh, but the Asmus Jaworski and Baran paper is cross-dialectal, and I’ve been told in person that Northern Welsh tends to lean more on an aspiration distinction while Southern Welsh devoices final stops but is otherwise much like English. Certainly my flawed perception as a native English speaker mostly exposed to Southern Welsh hasn’t noticed anything particularly different from English in the stops. Arafsymudwr (talk) 14:27, 7 March 2024 (UTC)

{{t}} vs. {{t+}}

Hi, just FYI — we use {{t+}} when the page exists on the foreign-language Wiktionary, and {{t}} when it doesn't. If you use the 'Add translation' interface, it will magically handle that for you. (But don't worry, it's not a big deal if you use the wrong one; I run a bot every so often that switches between the two templates as necessary. The only reason I'm even mentioning it is that I noticed that you mistakenly changed a {{t}} to a {{t+}}, which made me suspect that you'd misunderstood the difference, and might want to know about it.) —Ruakh_TALK 06:42, 9 March 2024 (UTC)

Hi @Ruakh thanks, I actually was aware of this and the bot - I just default to adding t+ because I never really saw the harm in it. I don't generally bother with translations already added by others unless I'm changing another parameter like gender. Arafsymudwr (talk) 10:42, 9 March 2024 (UTC)

OK, fair enough! —Ruakh_TALK 16:48, 9 March 2024 (UTC)

Amrhydlwyd (amhrydlwyd)

By the way, the entry for ‘amrhydlwyd’ in y Bywiadur and the associated publications is incorrect. It is ‘amhrydlwyd’, as in GPC. This one has been lingering for a while the Cymdeithas’ otherwise brilliant work; but then again, given that it is an uncommon noun, the misspelled variant is ptobably by far the most cited, given the prominence given to it by y Bywiadur. 2A00:23C7:7C81:9001:6826:35D:CA5B:DC3 22:19, 25 March 2024 (UTC)

The GPC lists both amhrydlwyd and amrhydllwyd though, and ammrhydllwyd was the spelling used in Hugh Davies' Welsh Botanology, cited in the GPC. The intention has has always been to list variant spellings as alternatives when an entry for the word is eventually made, with prominence given to the form that is more common, regardless of my feelings about which spelling is best on etymological grounds. Arafsymudwr (talk) 22:36, 25 March 2024 (UTC)

I can’t see that GPC has a record for the spelling ‘amrhydlwyd’? 2A00:23C7:7C81:9001:5034:5825:4034:B031 13:24, 27 March 2024 (UTC)

I don't see why it needs to. The GPC is not the ultimate arbiter of whether something exists or not in Welsh. Arafsymudwr (talk) 17:24, 27 March 2024 (UTC)

I don't think that this is a matter of GPC being an "arbiter" of what "exists or not" in Welsh: it is a descriptive, not a prescriptive, dictionary. The fact the 'amrhydlwyd' has no entry shows that it is a recent variant (or, in this case, a recent misspelling).

Further, the difference in spelling is not a stylistic choice: it is a difference of mutation. The actually attested 'amhrydlwyd' shows that is from 'an' and 'prydlwyd'; the misspelling would indicate that is from 'an' and 'rhydlwyd' (though missing the required mutation), which means something entirely different (and rather doesn't make much sense at all, as 'rhydlwyd' is not an attested word).

The only reasonable conclusion here is that Cymdeithas Edward Llwyd has given rise to a mispelling which has, through the very useful Bywiadur resource, become entrenched; and with it not being a word used very often, the mispelling can practically instantly become the most commonly found name. 2A00:23C7:7C81:9001:A4C5:4642:1D6B:F2D1 15:39, 27 June 2024 (UTC)

Misspellings

Hi, the best template to delete an unwanted redirect is {{d}}. Ultimateria (talk) 17:46, 29 July 2024 (UTC)

What's this in reference to? I've used that template a few times. Arafsymudwr (talk) 20:58, 29 July 2024 (UTC)

Heads-up

Hello! A new account has made two nonsensical and incorrect entries under ‘dad-goron’ and ‘coronaf’. I’m not sure how these are to be removed, but I thought that you might now. I think the user in question means well, but has no knowledge of the Welsh language. 2A00:23C7:21B4:FD01:B9F0:233E:7A30:1188 08:13, 8 September 2024 (UTC)

Thanks, I've just modified those entries and left a note on the editor's talk page.

In future you may want to raise concerns with users who have admin privileges, like @Benwing2 or @Chuck Entz. Arafsymudwr (talk) 11:16, 8 September 2024 (UTC)

Revamping parts of Template:cy-mut

Hi - I'm reworking a lot of the back-end for Template:cy-mut to begin the process of creating a centralised mutation module for the Celtic languages, so that we don't have to reinvent the wheel for each language. However, there are a few things I just wanted to pick your brains about in respect of Welsh in particular:

1. I think it makes sense to give both the literary and colloquial mutations for terms beginning with tsi-, as there's an obvious discrepancy between the written form and what many people say when speaking. This should be doable by showing both unchanged and the colloquial mutated form, with footnote labels for each one. I appreciate that it's rare to see the colloquial forms written down, but that also goes for many inflections in lots of languages.
2. For terms beginning with vowels, we currently replace the "aspirate mutation" column with "h-prothesis", and show unchanged for the soft and nasal mutations. This implies that the aspirate mutation and h-prothesis are two names for the same process, which is wrong, in my view, since they only share one trigger: ei (“her”). Looking at the others, ein and eu trigger h-prothesis but not the aspirate mutation, whereas a, â, gyda, tua, tri and chwe trigger the aspirate mutation but not h-prothesis, suggesting that these are separate processes. To fix this, I suggest that we instead only display a radical and h-prothesis column for vowel-initial terms, perhaps with a note below explaining that vowel-initial terms don't undergo any other type of mutation.
3. We don't acknowledge the limited soft mutation, even though it has different triggers to the full soft mutation. I think it would be better to have the subheadings "full" and "limited" under "soft", where the "full" column displays the same as what we show now, but the "limited" column gives unchanged for terms starting with ll- and rh-. For terms where the two mutations are the same, we should still give the subheadings, but should merge the two cells which actually give the term so that we don't show it twice.
4. We also don't have the mixed mutation either, which I appreciate is down to the fact that learning materials usually treat it as a mix of the soft and aspirate mutations. While this is fine for learners, I think it would be better for readers to display it in its own column as well, given that it has its own set of triggers (ni, na, oni) independent from either, so omitting it feels a bit like omitting the Latin vocative on the basis that it's usually the same as the nominative, when in reality it's a different process in its own right. This is another place where it might make sense to give literary and colloquial mutations, since the soft forms are commonly used in place of the aspirate ones for t-, p- and k- initials in colloquial language (especially in the south).

Theknightwho (talk) 00:12, 3 November 2024 (UTC)

Hi, I think I agree with all of these except 4, and even then, I don't disagree much.

I'd suggest having a column for radical, soft, mixed, aspirate and nasal mutations in that order. The headers would be separate but the mixed mutation always a double width column shared with either soft or aspirate like so:

Welsh mutation
radical	soft	mixed (literary)	aspirate	nasal
daw	ddaw		unchanged	naw

Welsh mutation
radical	soft	mixed (literary)	aspirate	nasal
poena	boena	phoena		mhoena

I think that notes on when the literary and the colloquial language have different mutation rules ought to appear on the entry for the trigger, not as part of the mutation template. E.g.

Particle

na (triggers mixed mutation (literary) or soft mutation (colloquial))

The only situation where I would support adding detail about Colloquial Welsh to the template would be for words beginning with ts-, that soft-mutate to j-, set up in such a way that it ONLY appears for these words, and can be blocked for words that don't do that. The header for the soft mutation in such cases might say “soft (colloquial)”. Arafsymudwr (talk) 17:01, 3 November 2024 (UTC)

Reordering it as "radical, soft, mixed, aspirate, nasal" is a good idea. We still need to slot the limited soft mutation in there to show the distinction, but that's not a problem if we order it "limited, full".

With the extra complexity this creates, I think we should flip the table to the conventional vertical format used for most declension templates: in my experiments with the current layout, it's so wide that it's very diffuclt to show a term spanning across two columns in a way that doesn't look really stupid, since each column inherently has to be at least as wide as the term itself. On the other hand, merging cells between rows looks fine (and is very common in declension/conjugation templates). The overall table still wouldn't be very tall, either.

In terms of generating the colloquial ts- mutation, it currently uses the following logic:

1. It must occur before i, which may have a diacritic (e.g. tsips or tsîc, but not tswnami).
2. If the i is followed by a consonant, it's retained (e.g. tsips → jips, tsîc → jîc).
3. If the i is followed by a vowel, it's absorbed as part of the mutation (e.g. tsiec → jec, tsiarjo → jarjo).
4. Except if the i has a diacritic, in which case it's retained (e.g. tsïars → jïars). Realistically, this exception only matters for ï, since î and ì only occur before consonants (so they're covered by rule 2 already), and í doesn't occur in initial syllables. However, if any of them did occur before a vowel for some bizarre reason, I'm pretty certain they wouldn't get deleted either.

Now I think about it, I have a feeling the i should probably also be retained if it's part of the diphthong iw, too (e.g. tsiwawa). What do you think?

Theknightwho (talk) 18:08, 3 November 2024 (UTC)

Yes to all of that. By the way I overlooked something else in your initial post.

In 3. you said about limited mutation "For terms where the two mutations are the same, we should still give the subheadings, but should merge the two cells which actually give the term so that we don't show it twice."

I disagree with this because this is every word that doesn't start with ll- or rh-. It would be much less cluttered to only acknowledge the limited mutation with the subheadings for words beginning with ll- or rh-.

By the way, are you going to take this to the Grease Pit or Beer Parlour? I'd like to see what other people say about all this, most of which I support btw.

Arafsymudwr (talk) 20:16, 3 November 2024 (UTC)

I'll do a mock-up first (or maybe two or three), but yes. I just wanted to brainstorm things first. Theknightwho (talk) 17:10, 4 November 2024 (UTC)

@Arafsymudwr I added another exception to the i-deletion rule: it's not deleted if it's followed by a semivowel (e.g. tsiwawa → jiwawa). In this case, w is a semivowel because it's followed by a vowel. Collectively, I think those rules together cover every possible case, though it's tricky due to the paucity of written materials (e.g. I can only find one instance of jiwawa, but it's clearly real: [1]). Theknightwho (talk) 16:17, 5 November 2024 (UTC)

Can I add a further suggestion? If you want to add non-standard forms, Northern Welsh can (nonstandardly) do aspirate mutation on m, n so they become mh, nh.

This also might be a good opportunity to rename the aspirate mutation - because it isn't truly aspiration, that's just an artefact of older terminology. I'd suggest spirantisation, which might be rarer than, say fricativisation but is close enough to the older term for some continuity. Arafsymudwr (talk) 21:17, 7 November 2024 (UTC)

@Arafsymudwr Yes to the including the Northern Welsh aspirate mutations (appropriately labelled), but changing the name to "spirantisation" is problematic since it's preferable to use the most common name, even if it's not strictly accurate to the underlying phonological processes. Theknightwho (talk) 22:36, 7 November 2024 (UTC)

@Arafsymudwr Here's one possible layout:

Mutation

radical		tad
soft	limited	dad
	full
nasal		nhad
aspirate		thad
mixed		thad
h-prothesis		unchanged

Some explanatory points:

1. This follows the style used for many declension tables, so it's tried and tested.
2. A vertical layout is much more mobile-friendly than the current horizonstal one.
3. I don't object to only including the limited/full split in the soft mutation when they actually differ - I've simply included it here for demonstrative purposes.
4. Unlike most declension templates, I've not made it collapsible, as it's still only slightly taller than the current mutation template (especially when "soft" isn't split).
5. I've included h-prothesis at the bottom, since I realised that only including it for terms starting with vowels has the potential to confuse readers when terms start with "w" used as a consonant, and there's really no harm in just including it as an extra row. The main issue is that I want to make it clear that it's not the aspirate mutation as applied to vowels, which our current template (wrongly) implies.
6. I experimented with reordering things, placing mixed between soft and aspirate (as you suggested) and then positioning the mutated form spread across both rows, but what I found was that the difference was relatively subtle, such that an inexperienced learner might not spot that the mixed mutation merges with different rows depending on the term. It seems safer to simply include it as its own row below aspirate.
   Mutation

   radical	tad
   soft	dad
   mixed	thad
   aspirate
   nasal	nhad
   h-prothesis	unchanged

   Mutation

   radical	mam
   soft	fam
   mixed
   aspirate	unchanged
   nasal	unchanged
   h-prothesis	unchanged

Theknightwho (talk) 19:01, 9 November 2024 (UTC)