User talk:Mahagaja/Archive 2

foghlaim

I don't know the least bit about Irish verbs, but User:Gwydionleu has added one, and I'm not sure they've put in all the necessary forms, judging by the Scottish Gaelic entry. Would you mind fixing this if necessary, and explaining the layout we use for Irish to Gwydion? Thank you so much --Μετάknowledge^{discuss/deeds} 02:44, 11 April 2012 (UTC)

WT:AGRC

Thanks for the fixes. If you happen to come across anything you're not sure what to do with, or don't feel like taking the time to deal with, or just have general comments about the page, I'd love to hear about it. -Atelaes λάλει ἐμοί 08:34, 14 April 2012 (UTC)

The changes look good so far. If I have any more comments, I'll just leave them at the talk page. —Angr 13:30, 14 April 2012 (UTC)

wisha

Hi, as our top Irish dude, can you have a look at wisha and see if you can help with the etymology - this gives mhuise or muise as an origin--Itkilledthecat (talk) 21:02, 16 April 2012 (UTC)

Old Irish Babel

Would you be so kind as to knock together a translation of 'This user is able to contribute with a (x) level of Old Irish' for the Babel box? I'd try, but I am, by my own admission, a basic user. embryomystic (talk) 16:51, 30 April 2012 (UTC)

Nrrghh... I've seen that at Category:Old Irish terms needing attention and have avoided doing it because I didn't think I was up to the task. (It's not like anyone is ever taught Old Irish prose composition at university, the way people are still taught Latin prose composition.) But maybe I can cobble something together by dragging the contents of {{User ga-1}} back to their Old Irish spellings (though the idea of a 9th-century monk referring to anyone as a "user" is rather mind-boggling). —Angr 17:01, 30 April 2012 (UTC)

OK, done. I called us "writers" rather than "users". —Angr 18:07, 30 April 2012 (UTC)

Awesome! Thanks heaps! embryomystic (talk) 19:28, 1 May 2012 (UTC)

Primitive Irish

I've been thinking of adding some more Primitive Irish words (with references, of course). Although I can work out case (usually genitive, given the context), I don't know how to figure out gender. So: if a modern Irish word is a certain gender, will the Primitive cognate certainly be in the same gender? If not, can I use Old Irish and maybe other Gaelic languages to figure it out? Or is it indeterminable, and I just shouldn't mark a gender? Thanks --Μετάknowledge^{discuss/deeds} 23:45, 1 May 2012 (UTC)

If the word is attested in Old Irish, the Primitive Irish word will almost certainly have the same gender. With the modern languages it's less certain because they've lost the neuter. (It's mostly merged with the masculine, but sometimes with the feminine.) —Angr 08:55, 2 May 2012 (UTC)

Great! Then all I need is the gender of the Old Irish equivalents to the following modern Irish words: (deprecated template usage) uí, (deprecated template usage) ceíle, Lugh, and nia. Everything else is covered, AFAICT. Thanks! --Μετάknowledge^{discuss/deeds} 04:23, 3 May 2012 (UTC)

By uí do you mean the genitive of ó (“descendant”)? If so, it's from Old Irish haue, which is masculine. The Old Irish words céile (“servant; companion”), Lug, and nia (“nephew”) are all masculine too. —Angr 06:05, 3 May 2012 (UTC)

Yes, that all sounds right. I misspelled céile and gave you a genitive, but you seem to have figured it all out. Thank you so much! --Μετάknowledge^{discuss/deeds} 13:47, 3 May 2012 (UTC)

You may want to be careful though. Analogy can be a powerful force in changing the gender of words, and that can make the gender hard to reconstruct. s-stems for example were historically all neuter (in Indo-European) but I think Old Irish has some masculines and feminines as well. So there is no telling when exactly the originally-neuter nouns of that class switched over... it may have happened in before Primitive Irish or after. —CodeCat 20:32, 18 May 2012 (UTC)

Reversion on cos#Old Irish...

What was specifically wrong about that table? The forms I provided were direct from DIL, and the table was hand-rolled in the absence of any relevant template.

You may notice I haven't been editing for long, so I'm wondering if there is some style guide I've broken or other egregious faux-pas I've made. --Catsidhe (talk) 01:53, 10 May 2012 (UTC)

Some of the forms were totally wrong, like chosaib for the nominative dual (it's dative plural only). AFAICT, DIL rather surprisingly doesn't give any dual forms of this word, but the nominative dual ought to be dá chois. Also, you listed several forms in their lenited version rather than the radical "lemma" form (e.g. accusative singular chois instead of cois). DIL gives achossa as the vocative plural because that's how it's written in the MS, but the a is the vocative particle, not a prefix, so we should write a chossa. But it's great that you've joined Wiktionary and want to help out with Old Irish! Maybe you can design a declension-table template for Old Irish nouns. —Angr 20:43, 10 May 2012 (UTC)

Ha, you're right: I saw dp and read it as "dual". I take it then that what are now read as particles should be analysed out as such, rather than presented as attested or documented if there is a difference? Reading from the Mod.Ir table, I take it that only the vocative particle is included therein? (What about the dual number? And complicating that, how should the table be constructed if the dual is not attested? Merged with plural, or left blank for the reader to infer the plural while letting them know that dual is unknown?) As for creating a template... so far I've just been filling in details when I went to look up a word and thought it could use some expansion. To do templates properly would require some serious time spent figuring out how to write a template, then grovelling over Thurneysen and Old Irish Paradigms. Tell me to shut up if I'm asking too many stupid questions.--Catsidhe (talk) 00:04, 11 May 2012 (UTC)

Further: is it worth adding to the tables what mutation a noun imposes in a noun phrase? As in ns. fer; np. fir^len; as. fer^ecl; ap. firu^gem. The "gemination" mutation probably needs a different name, though, being fossilised from the P.Ir geminate which results in "h-" before vowels in OIr, but no longer gemination as such. What do you think?

Sorry if I seem too puppy-dog eager. I want to get this right.--Catsidhe (talk) 03:27, 11 May 2012 (UTC)

Stokes & Strachan and Thurneysen both include the word for "two" in their declension tables for the dual, so I suppose we can do the same. They also include the vocative particle. I wouldn't add mutation information to the template; that seems like TMI. And "gemination" really is no mutation at all; in Old Irish there's no correlation at all between the h that sometimes gets orthographically added to the beginnings of vowel-initial words and the /h/ that sometimes gets phonologically added to the beginnings of words (see the discussion under w:Old Irish#Orthography). As for the template, I wouldn't try to make it generate forms; they're too unpredictable and too sparsely attested in Old Irish. I did create {{sga-conj-simplex}} and {{sga-conj-complex}} for verbs, where the idea is that each attested form can be added individually and all other cells left blank. For nouns, only {{sga-decl-o-masc}} has been created so far. —Angr 17:58, 11 May 2012 (UTC)

Me again. I've been grovelling over Strachan and the documentation on Templates and such things.

I've got something to look at at User:Catsidhe/Old Irish Nouns, with three general templates. They include such features as auto-substitution (ie, if the gp is supplied then it is used, but if it isn't, then the ns will be used if it is supplied), and variable behaviour based on stem (ie, ns nasalises in all Neuter cases except -s stem. If the case is given as "s", then the nasalisation goes away). They will certainly need turning to get the correct behaviour for default forms, and probably the correct behaviour for no supplied stem is to not make any defaults unless it's truly universal.

I also believe that with some careful fiddling, lenition and nasalisation templates for OI can be written, so that "sga-nasal|gasúr" returns "ngasúr", or "sga-lenite|cathoir" returns "chathoir". At the least, this would enable initial mutation tables similar to those for Mod.Ir, only with less fuss on the pointy end.

What do you think? --Catsidhe (talk) 04:08, 17 May 2012 (UTC)

It looks really good! It still needs links to the forms, which should point to the radical form even when what surfaces is mutated (and note that fh as the lenition of f isn't used in Old Irish). The tables should be marked class="inflection-table" so that redlinks stay black, which is preferred for aesthetic reasons (a table full of red redlinks is harder to read). As for showing the mutations induced by the various forms, I think either we should show nothing (like Thurneysen) or we should show both lenition and nasalization, but we shouldn't follow Strachan's odd lead of showing only nasalization but not lenition. Finally, I think we should let tables show only forms that are actually attested. This may mean doing away with auto-substitution, since if the genitive plural of a word isn't attested, we probably shouldn't show it as being identical to the nominative singular, even if everyone knows it would have been the same as the nom.sg. if it had been attested. Attestation for Old Irish is always tricky because of the difficulty in deciding which texts are written in Old Irish and which one in Middle Irish, but Wiktionary treats them as separate languages, so we should too. (More about this at WT:ASGA.) Great work!! —Angr 10:22, 17 May 2012 (UTC)

Well, a template for nasalising a word is easy. Lenition, on the other hand...

There is a string manipulation template {str_left|word|n}, which returns the first n letters of the word. This, combined with {switch:} works beautifully for isolating the initial letter and doing the appropriate thing. Unfortunately, the appropriate thing for lenition is to be able to access the word without the initial letter. There is, theoretically, a similar template str_right, which does exactly that. And after grovelling over the help files across three wikis, it does exactly what is needed, and is the only thing which will do that job.

Except that that template isn't available on wiktionary. And when I try to put it back to use it (because there is no meaningful reason why it was deleted in the first place), the templates it depends on were also deleted, at least one with severe editorial prejudice. (Which is to say: I can't even recreate it.) And for all I know, the templates they depend on were also deleted. The RFD which they apparently failed is no longer findable, but it looks like they were deleted because no-one used them, and someone obviously decided that there was no possible use for them.

Now, of course, there is not only a perfectly acceptable use for them, there is arguably a need for them. Lenition done properly (ie, without needing to manually hack off the initial character when calling the template) requires this function. (And, might I add, is trivially implemented with it.)

This is hugely frustrating. Especially since with this function restored, templates which soften/aspirate/nasalise Welsh, and eclipse/lenite ModIr also become a Simple Matter of Programming.

And I have no idea where to go to see if this even can be fixed.

--Catsidhe (talk) 05:08, 18 May 2012 (UTC)

I was unaware there even was such a template, nor would I be able to use it correctly if it still existed. That's why I took the route of manually chopping off the initial consonant for the lenition templates for Modern Irish. Did you see Template talk:str sub? That's where the deletion discussion has been archived. {{str sub}} is protected from re-creation; as an admin, I can re-create it anyway, but I'd rather not unless the people who decided it should be deleted are in agreement. At WT:RFDO you can start a discussion about undeletion. —Angr 08:31, 18 May 2012 (UTC)

... OK, I think I understand how those templates work. And now I have to go take three or four showers.

The str_left template is simple enough: it just uses the #padleft: extension. Simple and easy. (That is, #padleft: takes a string (1), a number n, and another string (2). It returns (1), and then if the result isn't long enough, fills it to n characters long by inserting from (2) until it is long enough -- repeating if necessary. If (1) is empty, then it will return the first n characters of (2), which is useful in its own right.)

The str_index is clever, but evil in that cleverness: it takes the first n characters, and the first n-1 characters, and then checks the shorter block followed by a, then b, then c, then d, then... and whichever one matches against the longer block, it returns that character.

To return the characters from position a to position b, (where str_right is the special case where b = the length of the string) you get the character at index a, then a+1, a+2, a+3, ... until you reach b or you run out of string, or the function ends. (The version on Wikipedia goes up to a maximum of 100 characters.)

To that extent, I can see why they deleted them. They are quite possibly the worst conceivable way of doing this. Unfortunately, given how padright is padleft, only less useful, this is an inherent bug in the very fabric of Mediawiki as is: they are the only way of doing this. And even then, the 'grab the n first characters of the string' behaviour of padleft looks very much like a side effect. If padright were actually right-side bound, and returned the rightmost characters of a string, then none of this would be an issue. (Even the str_index would be more efficient: return the rightmost character of the leftmost n characters, done.)

To fix this properly would mean adding a new feature to MediaWiki. And that makes me very sad. --Catsidhe (talk) 12:49, 18 May 2012 (UTC)

So is the way I did it for Modern Irish, manually chopping off the first letter of the word, the only practical way of producing mutated forms? —Angr 13:24, 18 May 2012 (UTC)

Short of restoring the hideous-under-the-hood str_right, or adding functionality to MediaWiki itself, unfortunately, it seems yes. --Catsidhe (talk) 21:46, 18 May 2012 (UTC)

Mutation

Since Conjugation, Declension and Inflection are all level four headers, why would Mutation be level three not four? Mglovesfun (talk) 19:33, 17 May 2012 (UTC)

Because Mutation is independent of part of speech, and even etymology. All words with the same spelling have the same set of mutations. —Angr 19:34, 17 May 2012 (UTC)

abann

If this is just the genitive of ab, do you think that there may not have been a noun *abonā in Proto-Celtic at all? And if it did exist (which http://www.jstor.org/discover/10.2307/30007351?uid=3738736&uid=2129&uid=2&uid=70&uid=4&sid=56184057553 seems to suggest), did it survive into Old Irish? —CodeCat 15:24, 18 May 2012 (UTC)

The (first page of the) article suggests *abonā in Proto-Brittonic, but not necessarily in Proto-Celtic. If Continental Celtic has reflexes of *abonā, it's probably Proto-Celtic, but I don't know if it does. I don't think it survived into Old Irish, which AFAIK only has the reflex of *abū, genitive *abens. —Angr 15:47, 18 May 2012 (UTC)

Is it possible that the Brythonic languages regularised it by adding the (presumably more common) ā-stem ending to the existing n-stem (a process that also happened in Germanic but the other way, adding n-stem endings to ā-stems)? Do you know if there was a general tendency to do that in those languages? —CodeCat 20:28, 18 May 2012 (UTC)

It's possible, but as the article says, it's also possible that it's a suffix -onā. —Angr 21:42, 18 May 2012 (UTC)

Citation form of Celtic verbs

I'm a little mystified by the citation forms of the verbs in Celtic languages, and actually by the verbs in general. What form do you think would be appropriate for Proto-Celtic? Or would it be best to use a bare stem of some kind? —CodeCat 20:25, 18 May 2012 (UTC)

I don't know if there's an "industry standard" citation form for Proto-Celtic verbs. Modern Irish uses the 2 sg. imperative, Old Irish uses the 3 sg. present, Welsh uses the verbal noun. I'd say use the root, same as for PIE. —Angr 21:41, 18 May 2012 (UTC)

Categorising Old Irish verbs by conjugation

From what I've read about Old Irish verbs they are categorised in two main groups, A and B, which are then numbered A-I to A-III and B-I to B-V. Do you think it would be useful to categorise verbs by their conjugation type? I could personally see some use from that mainly for comparative purposes (including Proto-Celtic) but I imagine it could also be useful for language learners. —CodeCat 12:48, 19 May 2012 (UTC)

Oh... and I suppose categorising nouns and adjectives could be useful too? —CodeCat 13:36, 19 May 2012 (UTC)

Well, the classifications A-I to A-III and B-I to B-V apply only to present stems; the future, preterite, and subjunctive stems have their own classifications. Also, the A-I/B-V classification is not universal. It was created by Thurneysen for his Old Irish Grammar, and it's also used by Wim Tigges in his Old Irish Primer and by me in Old Irish Verbs and Vocabulary, but there's a different classification that Kim McCone created for his Early Irish Verb, which is used also by David Stifter in Sengoidelc. Strachan's Old Irish Paradigms uses its own system that no one else (to my knowledge) follows, and Quin's Old Irish Workbook and Lehmann's Introduction to Old Irish don't classify the verbs at all. So if we do follow Thurneysen's classification, we should at least have an appendix where we explain how it matches up with Strachan's and McCone's systems. Categorizing nouns and adjectives would be useful too, and a great deal more straightforward than categorizing verbs. (It never ceases to amaze me that Irish two-year-olds of 1200 years ago were able to acquire this language.) —Angr 14:45, 19 May 2012 (UTC)

From what I've been reading neither do I. I found a blog article that you might like... [1]. I think it may be easier to start by categorising the nouns. I do have some grammar references[2] but otherwise I'm mostly in the dark about the different types (especially the irregular ones). Just to recap there are o-stems (m/n), yo-stems (m/n), ā-stems (f), yā-stems (f), i-stems (m/f/n), u-stems (m/f/n), ī-stems (f), s-stems (should historically be all neuter but are they...?), r-stems (m/f), n-stems (m/f/n, but do these all decline alike? PIE had several types), nt-stems (m?) and others maybe? —CodeCat 18:45, 19 May 2012 (UTC)

Very good blog entry! As Warren Cowgill said, "Learning Old Irish is like mowing the lawn: it isn't something you do once."

S-stems are all neuter except mí (“month”), which is masculine. Strachan gives 5 subtypes of nasal stems. You've left out velar stems (m/f) and dental stems (m/f). Strachan also lists diphthong stems, but his only example is bó (“cow”), which Thurneysen simply calls irregular. —Angr 19:14, 19 May 2012 (UTC)

Ok I was confused because tech was shown as masculine. Does that refer to the gender it eventually took when the neuter started to disappear? And what is the difference between the five nasal stem types? And are there any remaining neuter dental or velar stems that you know of? From my experience with Germanic I know that neuter nouns overwhelmingly tended to be o-stems in that language, so I wonder if the same applies to Old Irish. I'm not sure why mí masculine but going by our entry *mḗh₁n̥s this seems like an archaism. —CodeCat 19:20, 19 May 2012 (UTC)

In Old Irish, tech is neuter; it didn't become masculine until the neuter started getting lost in Middle Irish. The nasal types include those that end in a consonant in the nominative singular (originally ending in -n̥), which includes neuters in -m (corresponding to the Latin neuters in -men and the Greek neuters in -μα); masculines and feminines ending in -(a)e in the nominative singular (originally ending in -ēn, corresponding the Greek n-stems in -ην); and masculines and feminines ending in -u in the nominative singular (originally in -ōn, corresponding to the Latin n-stems in -ō and the Greek ones in -ων). These groups are then divided into those that do and those that don't undergo fortition of final -n to -nn in oblique cases. AFAICT there are no neuter velar stems and the only neuter dental stems are a few neuter nt-stems like dét (“tooth”) and lóchet (“lightning”). I wouldn't say that neuter nouns tend overwhelmingly to be o-stems in Old Irish; there are plenty of neuter i-, u-, s- and n-stems as well. —Angr 21:57, 19 May 2012 (UTC)

Ok thank you from clearing that up. It seems that in Germanic there were only a few neuter i-, u- and n-stems, and s-stems (z-stems in Germanic) were not very common either. I think there are only one or two reconstructable neuters of other consonant stem types, and they are only reconstructible because they are irregular. Germanic only has one or two types of n-stem, corresponding to the Irish ones in -u (the nouns in -men have -mô and I don't know of any in -ēn). The fortition you describe is really a word-final lenition I imagine though, and the original form would be the one with -nn- (whatever origin that has)? —CodeCat 22:04, 19 May 2012 (UTC)

There are a few words with original -nn-, but in most of them it's fortition of -n to -nn by a weird little rule that turns -n to -nn at the end of a syllable that begins with a lenited consonant. For example, genitive Érenn has fortition of n after the lenited r; it's from earlier *éren < *īweryonos. —Angr 22:13, 19 May 2012 (UTC)

That's good to know, thank you! But how does r lenite exactly? I can't recall ever seeing rh in Irish. Or is it just that it is phonologically lenited but not on the surface? And now that you mention that word... it looks similar to 'hibernia' so it's probably the origin of it. But then where did the h come from? Do you know? —CodeCat 22:18, 19 May 2012 (UTC)

Unlenited r is spelled rr in Irish, not rh. (It's also spelled with plain r at the beginnings of words and before other coronal consonants, so rí and ard have the "fortis r" as well.) As for Hibernia, if the etymology at Ériu is right, the Proto-Celtic word started with [ɸ], which probably developed to [h] on the way to disappearing, and that [h] got preserved in the Latin name. The same thing probably happened with Hercynia < *ɸerkun- < *perk(w)unos. —Angr 22:33, 19 May 2012 (UTC)

te#Old Irish

I found a reference that writes this as té instead. Which is the correct form? This is supposedly an nt-stem, which means it's probably a participle of a verb meaning 'to be warm' like Latin tepeo. So I imagine that the Celtic form was *tefent- or something similar. —CodeCat 16:42, 20 May 2012 (UTC)

té is correct. (Modern Irish does have te with a short vowel, though.) One of the oldest attested spellings is actually tee, which some have argued maintains the hiatus created when the ɸ of *teɸent- was lost, but old manuscripts often indicate vowel length by doubling (à la Finnish), so tee could just as easily represent [teː] as [te.e]. —Angr 16:47, 20 May 2012 (UTC)

So the entry should be moved then? —CodeCat 17:10, 20 May 2012 (UTC)

Yes, definitely. —Angr 17:12, 20 May 2012 (UTC)

irrespectively

When you typed 'I would like to outlaw irregardless and lay when it means lie and beg the question when it means "raise the question"', I was stuck by the reversal of order and wondered whether you had omitted (deprecated template usage) irrespectively. If you had omitted it intentionally, we would need to add the sense (and find citations!). It seems like that definition is one someone must have thought of before. DCDuring TALK 19:46, 20 May 2012 (UTC)

Huh? I think you misunderstood me; either that or I've misunderstood you. The three terms I would like to outlaw are:

1. irregardless
2. lay when it means "lie"
3. beg the question when it means "raise the question".

I don't see how irrespectively fits in here at all, and I don't know what "reversal of order" you're referring to. —Angr 20:25, 20 May 2012 (UTC)

Templates... work in progress.

Me again. After some investigations, and some discoveries, and some stiff drinks, I've got the initial mutation template issues down to: it is impossible to take the rightmost part of a string, and this is by design. There are several ways this could be fixed, but they have been vetoed by Jimmy Himself. We are promised that "complex string manipulations" like taking the rightmost part of a string will be possible when Lua is folded into the MediaWiki Template infrastructure. This will be ready Any Year Now. In the meantime, and for more than five years since, it is impossible to do initial lenition without splitting off the first character.

So if you look at User:Catsidhe you'll see links to sga-lenite and sga-nasal. You might want to have a look at them: the lenition template does not take into account delenition, and I'm worried that I've missed cases where consonant clusters are affected differently: I don't think clusters like 'tr-' or 'cl-' behave differently in OIr, but I stand to be corrected.

The nasalisation template does not give the n or m with punctus delens as Thurneysen does. This might be added, if it's thought a worthy addition.

I have created a new declension template -- only for the masculine nouns so far. (At User:Catsidhe/Old Irish Nouns it's used for the slightly wider tables.) Strong men might look deep into the code and go mad, but it seems to work sort of maybe well enough if you treat it gently. The next move is to make it even more complicated by using the stem to change some default forms. eg., -o- and -io- stem masc voc.sing is the same as gen.sing, which is the same as nom.pl. But in (most? all?) other stems it's the same as nom.sing.

The template does do things like display the mutated form where appropriate, but still links to the unmutated root. I've also made a go at including where it lenites, but I could use some help there, I think. Thurneysen is typically German in putting forth a simple rule, and several pages of complicated exceptions.

It's a work in progress, and is already more complicated than I would like. I'm wondering if it might be better to refactor it so that there's a simple template to show the table, and per-root templates to do the default form calculations. That is a question for later.

Anyway: what do you think?

PS: Without wanting to sound stalkery: from one of your statements, I have figured out who you are. I have a copy of your Verbs and Vocabulary at home. It's an honour to meet you, sir. --Catsidhe (talk) 05:02, 24 May 2012 (UTC)

I'll take a look and leave some comments on the respective talk pages. —Angr 19:40, 24 May 2012 (UTC)

amh#Irish

Angr, do you know how to pronounce amh#Irish and its inflected form amha? Is the "mh" a lenited "m"? If so, I would have guessed the "a" made the "m" broad, and w:Irish initial mutations says broad "m" lenites to /w/, whereas slender "m" lenites to /v/... but I found a forum on Irish pronunciation that says the word "amh" is /av/. And w:Irish orthography#Vowels says the sequence "amh" is usually /əu/ — or is that only when it's word-internal, as in amhantar /ˈəun̪ˠt̪ˠəɾˠ/? - -sche (discuss) 18:36, 24 May 2012 (UTC)

I don't know for sure, but my suspicion is that it's [avˠ] in Munster (where broad mh is always [vˠ]) and Connacht (where broad mh is [vˠ] in coda position) and [aw] in Ulster (where broad mh is always [w]), but I'll have to look it up in some reference books to be sure. —Angr 19:40, 24 May 2012 (UTC)

As for amha, I really don't know. Probably /əu.ə/, but maybe /avˠə/ if the influence of the singular is strong enough. —Angr 21:07, 24 May 2012 (UTC)

Etymology at taoiseach

You changed it to include ogham and the transliteration says 'tofisaci'. But I can't find any references to that, while 'tovisaci' gives quite a few. Which one is actually attested? —CodeCat 22:32, 25 May 2012 (UTC)

And somewhat related... do you know if there is any way to verify whether the -a- in the Proto-Celtic word was short or long? —CodeCat 22:34, 25 May 2012 (UTC)

Oh and another question. I just came across tosach. Is that related? If so, how? —CodeCat 22:35, 25 May 2012 (UTC)

Presumably what's attested is ᚈᚑᚃᚔᚄᚐᚉᚓ, and the issue is how to transliterate the Ogham letter ᚃ. AFAIK it's usually transliterated f, but as Thurneysen says, "The symbol for f still denotes w or v in the earlier inscriptions; initially and medially it always represents Latin v, never f" (A Grammar of Old Irish, p. 10). So the transliteration f is just a convention; between two vowels (i.e. in a leniting position) it was certainly pronounced [w] and later disappeared. Welsh tywysog proves that the ā has to have been long, as only long ā became aw in Middle Welsh, which later became o in posttonic syllables. —Angr 22:41, 25 May 2012 (UTC)

I don't know if tosach is related. I doubt the etymology from tús+-ach, though. —Angr 22:43, 25 May 2012 (UTC)

Ok, thank you. I think it would be clearer to use w or v in the transliteration myself. Do you know what a short a would have given in Welsh? I'm not quite aware of all the sound changes yet but I'd like to learn. —CodeCat 22:53, 25 May 2012 (UTC)

Of those two I'd prefer w since the pronunciation was almost certainly [w] rather than [v]; but f is convenient in word-initial position because it became f in Old Irish (and on into the modern Goidelic languages). And I don't want to start having two context-dependent transliterations for a single Ogham letter. —Angr 22:57, 25 May 2012 (UTC)

I think short a would have stayed a in Middle and Modern Welsh. —Angr 23:00, 25 May 2012 (UTC)

If fern derives from *wernā, then it's completely regular with fer deriving from *wiros. So that's where the transliteration [F] came from: it is how it is recorded in the Auraicept na n-Éces, and is the form you'd expect in OIr for that phoneme. In [CISP], and the sources it works from, the canonical transcription is [V].

Well, you're right that it's usually transliterated with v, so I'll change it back. Looking at w:Ogham inscriptions it seems ᚍ is very often transliterated v, so I wonder if tovisaci is actually ᚈᚑᚍᚔᚄᚐᚉᚓ. —Angr 08:40, 26 May 2012 (UTC)

No it wouldn't. ᚍ is ngétal: /ŋ/, ᚒ is 'U' and ᚃ is 'V' -- as can be seen clearly on the table to the right of the example of 'AVI'. You get ᚃ on the Ogham keyboard by typing 'f', and ᚍ with 'v' for (I suspect) lack of a better place for it. I think the Wiki page has misspelled it from typing 'v' instead of 'f', and not checking the result. I had to look for an example, but [BLTAG/8/1] has 'CUNAMAQQI AVI CORBBI' clearly as ᚉᚒᚅᚐᚋᚐᚊᚊᚔ ᚐᚃᚔ ᚉᚑᚏᚁᚁᚔ, that is: 'AVI' = ᚐᚃᚔ. (That's one of the things I love about CISP: where it can it includes the drawing or photograph of the original inscription.)

I should probably correct the Wikipedia page, shouldn't I. --Catsidhe (talk) 09:32, 26 May 2012 (UTC)

Duh: beyond the phonetics of it, ᚃ is given the value [f] in Old Irish because [u] and [v] were not distinguished. (Which is why the prosthetic 'h' in húa, so that it would be read as /u:a/ rather than /wa/.) [f] was the next nearest available glyph, even taking the normal phonetic changes aside. --Catsidhe (talk) 11:04, 26 May 2012 (UTC)

uas#Old Irish

I'm not sure but did Old Irish have ua? Or was it always úa? And do you know if this originated from Celtic ouxs- (e- or o-grade) or uxs- (zero grade)? How would I tell? —CodeCat 12:59, 28 May 2012 (UTC)

Well, it's always normalized as úa, but the scribes weren't terribly consistent with marking long vowels in the manuscripts. I'd put it at úas, but we could call uas an "alternative spelling". —Angr 13:03, 28 May 2012 (UTC)

Ok I will move it. —CodeCat 13:17, 28 May 2012 (UTC)

teachín

Is this really a proper Irish word? It seems to violate the spelling rule that a consonant or cluster can't have both broad and slender vowels next to it. —CodeCat 19:24, 28 May 2012 (UTC)

It sure does, but that's the spelling found in Ó Dónall's dictionary. Must be an exception. I suspect it's pronounced [tʲaxiːnʲ], not *[tʲaçiːnʲ] (as *teaichín would imply), but the suffix -ín is never spelled *-aín, so this is the compromise. —Angr 19:28, 28 May 2012 (UTC)

There are some results on Google for teaichín and tigín, though. —CodeCat 19:33, 28 May 2012 (UTC)

I wouldn't put too much store by Google results for Irish, since the majority of Irish on the Internet is written by learners. Ó Dónall has tigín as a synonym of teachín, and Dinneen has tigín only (neither teachín nor teaichín), so that's definitely a legitimate synonym that we can add. I wouldn't add teaichín as a legitimate alternative spelling unless we can find it in published texts, but it could be added as a "common misspelling of" if the ratio of teaichín:teachín spellings is high enough. —Angr 19:39, 28 May 2012 (UTC)

I just noticed that -ín has a note saying "All nouns ending in a broad consonant change to a slender consonant before taking -ín, except words with stems ending in -ach"; another example is lachín 'duckling'. —Angr 19:41, 28 May 2012 (UTC)

So the ch is broad in this case, and the í following it is just a spelling anomaly? —CodeCat 20:19, 28 May 2012 (UTC)

I assume so, yes. If the ch were slender, there would be no reason not to spell it teaichín (and laichín for 'duckling') (cf. words like beainín and geaitín, where the consonant before the -ín has become slender. —Angr 20:24, 28 May 2012 (UTC)