I always accepted that "cá bhfios dó" was simply eclipsis caused by cá, although unusual as cá should be eclipsing a verb not a noun and as the verb is the copula here the whole thing could simply be an exception.

But PUL in Niamh has " Cá bh’fhios dó", This implies the "bh" is not derived from eclipsis of the f at all, but is derived from a lenited form of the conditional copula, would you would have expected to be cá b'fhios dó.

how do people here parse that phrase?

Without any shred of evidence or proof I suspect "cá" in "Cá bhfios dom" is equivalent to the negative "cha" in the northern and Scottish dialects.

I recognise that in the question "cá bhfios duitse?" it is probably of different origin.

I look forward to more scholarly responses.

It looks like it is equivalent to cia/cá. In the DIL, under cía Ie, it says:

(e) With follg. noun (often as antecedent of rel. clause). ... gá fhios nach . . . (ḟ = bf?) ...

But under C it says the form later eclipsed. Perhaps the forms were conflated.

But Seán, where is the verb then, if cá is directly eclipsing a noun? My theory is that there is a copula there and that bh is the lenited conditional copula.The explanation in DIL is "antecedent or relative clause", but where is the relative clause?

Another alternative could be that fios is the old verbal noun fios, and not an ordinary noun?

I am not sure. The DIL seems to indicate that it is a VN, at least originally ("used as vn. of ro-finnadar, ro-fitir"). It gives many examples of fios do. Most of them were with go, so go bhfios domh with my knowledge; so far as I know. I don't know if this helps, though.

I can't explain the eclipsis in cá bhfios as one would expect the initial f to be unmarked. The Christian Brothers only say that the eclipsis is an exception in cá bhfios (13.76). Perhaps a remnant of some defunct rule?

or a simple glide /w/ for ease of pronounciation?
cá fhios /ka: is/ -> /ka wis/ -> cá bhfios

Lars, you may be right. I would put an "a" in there in that case: cá a fhios? but of course the "a" does not affect the pronunciation... Whoever edited PUL's Niamh thought there was a lenited "ba" in there, but Seán's evidence from DIL indicates that is a folk etymology.

quote:

or a simple glide /w/ for ease of pronounciation?
cá fhios /ka: is/ -> /ka wis/ -> cá bhfios

The consonant is /v'/ not /w/.

One would expect "cá fios" and not "cá fhios". Cá + lenition is only used with abstract nouns denoting degree; cá mhinice?, cá fhad?, etc.

Cá bhfios is pronounced /kous/ in Corca Dhuibhne.

Bhuel, b'fhéidir go raibh ciall éigin ag an Athair Peadair lena "cá bh'fhios"...