Caoilte wrote:
"I'm just making the point that the wording on the websites you linked doesn't seem to exclude the possibility"
The explanation on those websites regarding how to distinguish broad and slender consonants did seem a little too concise i.e. "Slender consonants are those that are followed by the letters “e” or “i”. If the consonant is in final position, it is slender if it is preceded by an “i”". This would seem to imply that a consonant is broad in all other situations but it doesn't explicitly say that. It's almost as if the writers didn't want to commit to a definitive statement.
I think they're just saying what they can be relatively sure of, while trying to avoid any generalisation which would be inaccurate. As an aid for beginners it seems fairly reasonable to provide a rule of thumb that probably holds in 95% of cases. If they'd given much more detailed information about exceptions and spelling variation, it could actually be more confusing for learners.
I was actually thinking about this thread a couple of days ago when I happened upon an interesting example from the Würzburg glosses which I think actually outright defies the statement 'Slender consonants are those that are followed by the letters “e” or “i”'.
Wb. 31b25a reads:
scéla etsenchaissi etforbandi, "stories and histories and laws".
Based on the rule above, you'd expect the cluster
ss at the end of
senchaissi would have to be slender, but this is the nom. pl. of a masc. u-stem noun. As Stifter (Sengoidelc, p. 112) notes "The root final consonants of u-stems are basically non-palatalized throughout the paradigm". Exceptions occur where syncope could have caused secondary palatalisation, but eDIL suggests that
senchas goes back to a combination of "senchae + abstract suffix -as". As
senchae is a bisyllabic Old Irish noun, which I expect post-dates syncope anyhow, and clearly the root final consonant is not palatal (i.e. slender) in it, I'm inclined to expect that it cannot be palatal throughout the paradigm for
senchas either.
Senchas did later transition into a masc. o-stem, which explains why the Modern Irish genitive form of
seanchas has a slender ending,
seanchais, but the fact that there is a vowel at the end of the nom. plural form shows that in this Old Irish example it was still very much a u-stem.
This would mean that the
ss cluster is broad, despite being both preceded and followed by
i. It's not unusual for Old Irish scribes to double a consonant in situations such as this to suggest that it should be broad despite the surrounding vowels, and the creators of the websites you linked undoubtedly are aware of this. But exceptions also occur where doubled consonants can be slender. For example, the famous poem from the Reichenau Primer begins
messe ⁊ pangur bán, "I and Pangur Bán". Because
messe becomes Modern Irish
mise, it's fairly clear the cluster,
ss, would have been slender here. Undoubtedly, the creators of the websites are aware of all of this also, but explaining it all would have been far too much to throw at beginners.