I think “Mòrag glic” is because of the same sound (g) at the end of one word and the beginning of the next. Not sure about “Leagsaidh glic”, but if the dh is devoiced to ch then maybe that rule applies too?
Or it could be simply that I'm not either glic or ghlic at all, but terminally confused and just plain wrong. I think these ones are lenited, but that I was misled by having a wrong answer accepted as correct by the software. Which happens alarmingly often. If I ever encounter the question again I'll look more carefully.
As far as I can see from other questions, glic is lenited in these examples. Historically it would be Mòrag glic. Any velar pair will do - any combination of g, gh, c, ch. That is why we have surnames like MacCallum not *MacChallum in English spelling.
Officially the rules were based on when the spelling matched the pronunciation. So you should base the rule on the spelling not on the pronunciation and not apply the rule to dh_g, regardless of how it is pronounced.
But the question is, do these rules still apply? Wikipedia says
Lenition may be blocked when homorganic consonants (i.e. those made at the same place of articulation) clash with grammatical lenition rules. Some of these rules are active (particularly with dentals), others have become fossilised (i.e. velars and labials). For example, blocked lenition in the surname Caimbeul ('Campbell') (vs Camshron 'Cameron') is an incident of fossilised blocked lenition; blocked lenition in air an taigh salach "on the dirty house" (vs air a' bhalach mhath 'on the good boy') is an example of the productive lenition blocking rule.
(Note I have no idea what they are on about with taigh salach. The gh s is not homorganic so it should be air an taigh shalach. It is true that air an taigh is correct in place of
air a' thaigh.)
So basically, I think you only apply this rule to dentals, and the only example that has occurred in the course so far is, I think, words like an deoch instead of
Edit: just come across feumaidh tu instead of
feumaidh thu. The notes point this out but do not explain it.
The mhath in air a' bhalach mhath is lenited correctly. After a preposition AND the definite article you get jumping lenition, ie you don't only lenite the following noun, but also the following adjective(s), even with masculine nouns, eg bòrd beag buidhe, air bòrd beag buidhe, but air a' bhòrd bheag bhuidhe. I don't get the dirty house either though.
I'm feeling stupid now, because that is what I would have done instinctively but I made the mistake of looking it up before posting. And I must have misread my grammar because it now agrees with you. I have edited my previous post and added an example I have just come across in the course.
Or does the lenition not jump because the lenition of taigh is blocked?
I am absolutely certain that gh does not block dentals. That is the whole point of it being the homorganic rule. Sounds made in the same place block. The default assumption is that if the grammar says lenition occurs then lenition occurs. It wouldn't make any sense otherwise. The only time it is blocked should be when the grammar says it is blocked. And I have never heard anyone ever mention that homorganic can ever mean heterorganic.
Hmm, I did some more research and I was wrong, apparently blocking does jump as well (so my *air an dùn dhubh should be air an dùn dubh), thus the lenition of salach is blocked by the preceding article (regardless of what happens to the noun in between, if instead of taigh there were eg bòrd, it would be air a' bhòrd salach, even though the article has lost its final dental because of the intervening labial in bòrd, it's still implied for the salach). My bad, sorry for the confusion.