Labhrás wrote:
As originally foreign names they don't change for case, so Síon in all cases.
This isn't a hard and fast rule, though. Various case forms of
Sion have been used since at least the 8th century, and there seems to have been the same interplay between the palatalisation and non-palatalisation of the final consonant for about that long.
In the
Amra Coluim-cille you get
in faith De de deis Sion suidiath "The prophet of God has sat down on the south of Zion", glossed
.i. suidestar for deis in tSióin nemda "he has sat down on the south of the heavenly Zion. Here, I believe both
Sion and the later gloss form,
in tSióin, are in the genitive case, but represented completely differently.
You also get a dative in
Amrad inso ind rig ro dom-rig firdib-snaidfe Sione "this is the eulogy of the king who made me king, who will convey us to Zion". This too is glossed
ron-snáidfe co sliab Sion nó co Sion nemda "he will protect us to Mount Zion or to heavenly Zion". Oddly, in the first example, it was the - presumably older - main text which retained the "nominative" form in a genitive construction, while the presumably later gloss showed a specific "genitive" form. In this case it's the main text which shows a distinct dative form. The two forms in the gloss appear to be in the expected accusative case following the preposition
co, though perhaps you would still expect a genitive form of
Sion after
sliab in
co sliab Sion.
In more recent examples palatalisation seems to be the rule for final consonants, albeit, with continued variation in lenition. Logainm gives
Shióin as the genitive for places named
Sión in counties
Waterford and
Wexford,
an tSiáin as the genitive of
An Sián between Leixlip and Maynooth (anglicised Sion and attested in sources going back to 1752). There's also a
Cnoc Shióin in Kilkenny, a
Cnoc Sióin in Waterford, a
Bóthar Chnoc Shióin in
Cork and
Waterford, as well as
Ascaill Chnoc Sióin and
Ardán Radharc Chnoc Sióin in Waterford. And, of course, djwebb also mentioned thinking there was one instance of
Shíoin in PUL too, but this just makes his typical use of
cnuc Síon even more interesting.
At the least, these forms show that
cnoc is common in place of
sliabh when referring to Mount Zion. I do have doubts as to the antiquity of some of the examples though. The likes of
Ardán Radharc Chnoc Sióin seem particularly contrived.
Labhrás wrote:
Usually, definite nouns in genitive are lenited today. But in the older language, there are numerous examples without lenition. And names of foreign orign resist lenition as well-
It's never really possible in the Early Irish to say that the lack of a mark of lenition, particularly on an S, means that it was not lenited. Often marks of lenition were omitted where we know lenition would have taken place. In fact, the presence of t in the genitive form above,
in tSióin, would suggest that the S would have to be lenited. If there were consistent lack of lenition being marked in examples from the Early Modern period, that might be more telling, but I still wouldn't rely on not seeing it as indicative that it wasn't intended.
My guess would be that PUL was using some sort of long-established, compound proper noun when he writes
cnuc Síon, which preserves a historical lack of distinction between nominative and genitive forms, but I'd be very interested to know the context in which he uses the form
Shíoin now to see if this breaks that pattern.