Dia daoibh arís, a Chairde! Tá cúpla ceist agam inniu…

1. What is the difference between when to use “cér” / “cérb” / “cérbh” and “cé ar,” cé arb, cé arbh, etc. I read the section on the difference on https://www.braesicke.de/kopul5.htm …but I just got more confused.
2. Is there any particular reason why some questions words are followed by the direct relative and others by the indirect? (I hypothesize that it’s because there is sometimes an understood preposition associated with certain question words, leading to them taking the indirect relative…

Example:
Cén fáth nach raibh tú ar scoil?
(For) which reason weren’t you at school?

…however (even with that logic), I can’t quite understand why cad é mar and conas take a direct relative, while cén chaoi doesn’t.

They are the same. (cé + ar[bh] = cé + indirect relative copula form).
The combined form (cér) is used in petrified phrases as Cér díobh thú? Probably, the components aren't felt to be separate anymore.



Fáth requires an indirect relative clause
It is a so-called adverbial indirect relative. (see: https://www.braesicke.de/satz4.htm#adverb)
In English, you’d use the adverb "why" instead of a relative pronoun.
(an fáth a = the reason why ...)
The same in questions:
Cén fáth nach raibh tú ar scoil? = Which reason why weren’t you ... = or simply: Why weren’t you ...?

Mar (how) requires a direct relative (probably, because mar and indirect relative means "where", not "how").
But yes, it is adverbial in meaning, too (as well as conas). No rule without exceptions.

Caoi is like fáth. It takes an adverbial indirect relative clause. You can add the adverb "how" in English.
an chaoi a ... = the way how ...
Same in questions:
Cén chaoi a ... ? = Which way how ...? or simply How ...?

Labhrás's presentation, while good, tries to cover every single dialect at once. Not all of those forms will be found in any particular dialect. I've never seen cérb as thú - maybe it means something in Conemara.

You have to focus on whatever dialect you're learning, which for Ulster might mean exploring Tobar na Gaedhilge for Ulster literature.

As far as I know - and this is from my article published in Éigse and only relates to Muskerry Irish:

cé: cé hé féin? who is he?
cérbh é féin? who was he?

There is no cér or cérb in either of the above.

The direct relative is: cé a dhein é? cé dh'órdaigh é?
And direct relative with the copula: cé is bun leis? cé ba bhun leis? cé is feárr a dhéanfaidh é? cé ab fheárr a dhéanfadh é?

The indirect relative: well I don't recognise any thing like "cé ar" where "ar" is a supposed indirect relative particle, but that's because I'm studying Munster Irish. Munster normally has "go/gur", and it's better phrased like this:

cé leis go samhlaím é? who do I compare it to
cé air go rabhais ag cuímhneamh? who were you thinking about?
cé fé ndeár é? whose fault is/was it?
cé uaidh go bhfaighidh sí í: who she will get it from

If you're looking for weirder and weirder phrases, then they're hard to find in literature. E.g. "under whom"? Who were you sitting under? Under whom were you sitting? This is a strange sentence, meaning-wise, in English, and "cé fé" looks odd in Irish. I would avoid it. Maybe: cé hé an duine go rabhais id shuí fé? /cé hé an duine féna rabhais id shuí?

The real problem is indirect relatives with the copula.

cér leis é? who does it belong to?
cé go mba leis é? who did it belong to?

cér a díobh é? what family/place does he belong to? And this is: cér díobh é? in Standardised Irish

cér a mac é?, whose son is he?
cér' iníon í?, whose daughter is she?

There is hardly any attestation of the rarer "indirect relatives with the copula" form. And the present tense forms usually cover the past too, given the rarity of the thing anyway (this point was made by Seán Ua Súilleabháin in Stair na Gaeilge). So there is no attestation of a past tense of cér a díobh é? or cér a mac é?

Some of Labhrás's examples of cérb relate to it being used to mean "where?" in some other dialect.

Go raibh maith agaibh beirt! I’m going to reread your responses several times to make sure I understand, but it has already helped me just reading once through! THANK you as always once again, a Chairde!

Just out of curiosity, Rosie, are you the same Rosie that runs the youtube channel Irish With Rosie?

_________________
I'm an intermediate speaker of the Corca Dhuibhne dialect of Irish and also have knowledge on the old spelling
Soir gaċ síar, fé ḋeireaḋ thíar

I’ve some thoughts/comments on this “adverbial indirect relative clause” thing. Sorry for going on yet another tangent, maybe it’ll be of some use though?

The reason fáth, caoi, and some other such words take the indirect relative clause is because they are nouns that do not work as adverbials. That means, that this indirect relative clause actually hides the “true” (or at least historical) presence of a preposition.

Note that while you can say: sin í an chaoi a/go raibh sé, sin é an fáth a/go raibh raibh sé ann, you cannot say *bhí sé an chaoi sin, *bhí sé ann an fáth sin, an chaoi or an fáth on their own do not work adverbially. You have to say something like sa chaoi sin, ?ar an bhfáth sin. In fact, Nua-Chorpas na hÉireann does have a couple examples of an chaoi ina … maked as native, beside the “plain” indirect relative.

This means that sin í an chaoi a raibh sé really contains “indirect” a in the ina ‘in which’ sense, that’s why it takes the dependent form of the verb. It’s a prepositional relative clause. Same thing can be seen with fáth in historical corpus. DIL has examples like:

• ro innisemmar in fath ar nách and roadnacht Cormac (we told) the reason why C. was not buried there (lit. on which … not)
• in fáth forar crochad Ísu the reason (on which) Jesus was hanged
• ga fáth 'ma ra molais in fer sain? why did you praise that man? (lit. ‘what is (the) reason about which you praised that man?’, ’ma = um a)

FGB also gives an fáth um a ndearna é for ‘the reason why he did it’.

Now, with actual adverbials, you use direct relative clause in Irish – because there is no indirection, the antecedent is directly connected to the verb, ie. without any overt pronoun referring to it from the relative clause:

• conas atá tú?
• nuair a bhí mé ann
• ar an mbord atá sé
• sin mar a bhí sé
• an bhliain a bhí mé / a raibh mé
• an lá a bhí / a raibh mé ann

etc.

Note, that you can say bhí mé ann lá…, bhí mé ann an uair sin…, tá sé ar an mbord – ie. you can insert the antecedent directly into the relative clause and make it not relative, unlike with those (non-adverbial) nouns of reason or manner. Conas is an interesting one, because it continues a phrase with a noun (conas < cionnas < cindas < cía indas ‘what way, manner’), but it’s been calcified as a word of its own in the language centuries longer than the indirect relative clause has existed and it seems to me that indas was used adverbially on its own, judging by the II used adverbially (a) section in DIL.

Another thing to note is an bhliain/an lá a raibh mé as an option and the fact that Connacht dialects do tend to use indirect more often than other dialects with time relative clauses. But again, note that you can say either bhí mé ann an bhliain sin and bhí mé ann sa bhliain sin, similarly an lá can be used adverbially and you can say sa lá/ló.

This is pretty much the same as (an) áit a raibh mé type clauses. They just use the old form of i(n) + relative pronoun aᴺ which used to be just aᴺ.

All of this is to say that true adverbials take direct relative clause in Irish, and indirect relative clauses are used for nouns (of manner, reason, place, time) which do not work adverbially on their own.

Oh, and mar a bhfuil in the sense ‘where … is’ is actually interesting and exceptional. Mar being an adverb and conjunction meaning ‘as, because’ should not work with prepositional clause or take the indirect.

But it is an exception due to its etymology. Mar a bhfuil and similar things were actually condemned as faulty in classical grammatical tracts, and in earlier language this form does not exist at all. What’s used instead is, among others, i mbaile a bhfuil ‘in (the) place in which is’, bhaile a bhfuil, in OIr. also i fail i rraba (see below) and similar phrases. So it seems that i mbaile a bhfuil got stereotyped into something like, maybe, *maile bhfuil and then via dissimilation changed to *maire / mar and got conflated with the adverb mar. But ultimately originally it did contain a noun and was followed by a prepositional clause, just like áit, fáth, caoi.

Also compare Scottish Gaelic far a bheil of likely similar origins, but not conflated with mar; and also note that Old Irish also had fail ‘place’ for relative ‘where’ (but also followed by ‘in which’ prepositional relative), which also might be related to all this; For more see R. A. Breatnach (1973) The relative adverb mar a, in Celtica, volume 10, pages 167–170.

I think case is important.
Some nouns can act as adverbs in accusative case (= adverbial accusative). And as accusatives they can take a direct relative clause.
That is true for time nouns (am, lá, etc.) and probably for the noun indas in conas.

Nouns in dative or genitive case could perhaps have acted as adverbs as well (at least they can in other languages). But probably they would not be followed by a direct relative clause.

Of course, as O'Nolan states, the indirect relative shows a prepositional or oblique relationship. But trying to show that conas not being used with the indirect is in some sense logical ends up trying to concoct a reason for the grammatical patterns found.

As Meagrán Músgraidhe states (p108), conas can or could be found with the indirect relative in parts of Ireland: "So also cionnus a bhfuil tú? in parts of Muns. esp. in Thomond".

Ua Laoghaire used canad with the direct relative: "Canad is toil leat go n-ollamhóchaimís é?" But Amhlaoibh Ó Loingsigh used canad ann with the indirect relative: "canad ann ’na raibh an bhean le fáil".

Cathain is also used with the direct relative. It is possible to argue that this is because the answer would not contain a preposition (e.g. an lán san, and not ar an lá san), but the idiom of adverbs of time is actually a separate question from the interrogatives. In fact, conas, canad and cathain being used with direct relatives is just irregular.

Well, in general the direct relative with adverbials is pretty regular: sin mar a bhí sé, ní Aill an Tuim is mó a thugaidís uirthi (is mó being adverbial to thugaidís uirthi), dúirt sé gur ar an mbord a bhí sé…

There is some irregular stuff happening with some interrogatives, but I wouldn’t take that example of PUL to be that unexpected – it’s not “where is your will/wish?” but “what is the place that you wish…”, ie. the ‘where’ meaning does not refer to the copula, so the ionad of canad could be taken as subject of the copula (and thus direct relative).

Much more problematic is the example from Niamh: Do ghluais Gormfhlaith ar fuid an ríghtheighlaigh ag ceistiúchán ar na seirbhísigh feuchaint cé ba dhéanaighe a chonaic an tighearna óg, Amhlaoibh, agus canad a chonacthas é. And I don’t see any other explanation here than direct relative being allowed due to analogy with other interrogatives (like cathain and conas, with which it sometimes appears listed – the only examples of canad followed by a clause from Croidhe Cainnte C[h]iarraighe have connus, canad, cathain listed together).

Cá for ‘where’ is interesting (etymologically the same as cé, cía ‘who, which’) – in Old Irish cía worked either as unstressed particle before prototonic (dependent) verbs or stressed meaning ‘who/what is it’ and was followed by relative clause. And seems it was used this way also to express ‘where’ before verbs of going sometimes. But since later Middle Irish in the meaning ‘where’ it eclipses and takes dependent form (so from before the times of indirect clauses developing), which I guess is under analogy with cairm and cáit, both with nouns meaning ‘what is the place?’ and followed by prepositional clause starting with ‘in which’.

Cathain might continue cía chuin (classical cé chuin, gá chuin), and cuin seems to have never taken the prepositional clause (but rather simple relative, though seems we don’t have many OIr. examples, it seems to be attested later).

Also note that cà bheil (as well as càite a bheil) – with dependent form, as well as cuin a bhios – with relative clause, exist in Scottish Gaelic which has never had “indirect relative clause”.

At any rate, yes, I agree that the interrogatives are tricky and not fully regular. I don’t agree, though, with the characterization that Irish in general (including non-interrogatives) uses indirect relative for adverbials – the direct relative clause is much more common for them, and the indirect is used in those cases where the antecedent is not an adverbial but rather a noun of manner/place/time/reason which does not function adverbially by itself (implying “hidden” preposition).