Quote:
As for sula/sular, unless I've missed something somewhere over the years, it has absolutely nothing to do with the possessive pronouns. The r is appended to sula before regular verbs (verbs that would take níor in the negative, not ní).
Correct.
The
a in
sula is the relative particle
a ‘that/which/who’.
Sul (or
sar) is a kind of noun (I think technically it’s more of a preposition, really, but it’s acts most like a noun) that means ‘beforeness’ or something like that. So the sense is something like ‘(the) beforeness that…’.
It’s like with
an áit used to mean ‘where’: when you say ‘where’ in English, it’s a simple conjunction, but
áit itself is a noun in Irish, not a conjunction, so it needs something to make a relative clause after it:
an áit ar tháinig sí ‘where she came (to/from)’, lit. ‘the place
that she came (to/from)’. So when you say
sular tháinig sí, you’re very over-literally really saying something like ‘(the) beforeness that she came’ = ‘before she came’.
It’s a fairly convoluted thing to explain, but it does kind of make sense, once you get your head around it.
Breandán wrote:
Next, you have the his and her back-to-front: a (his) makes no change to a vowel; a (her) appends an h to a vowel. (This is the opposite to the lenition rule for consonants.)
Actually, it’s not really the opposite at all, if you just think about it like this (which is also the way the concepts were historically related). Careful, this may be extremely long and go a lot wider than just the topic at hand … please forgive me.
A is a
possessive adjective. So it’s really a kind of genitive.
Now think about how the definite article in the genitive works:
(Form: effect on initial consonants / effect on initial vowels)Masculine: lenition /
nothingFeminine: nothing /
prefixed hPlural: eclipsis /
prefixed n-See how this is exactly the same? The masculine
a causes lenition to consonants, but nothing to vowels; the feminine
a does nothing to consonants, but prefixes h to vowels; and the plural
a causes eclipsis on consonants and prefixes n- to vowels.
The article and the possessive adjectives in the third person are actually completely parallel, and there’s good reason for that: in Proto-Irish (i.e., at the time when the mutations were just started to come to life), they were declined very similarly.
Very basically, this is how mutation came about:
- If a consonant ended up between two vowels, it was weakened. This is very common—it happens in English a bit, too, and it’s very common (and completely parallel) in Spanish as well. Just think about how the v in vaso sounds when you say un vaso [um 'baso], and how it sounds when you say tu vaso [tu 'βaso], where [β] is the IPA sign for a ‘v’ that’s made by just letting air slip through your lips, rather than between your upper teeth and lower lip. This is called ‘lenition’ because the consonant becomes lenis (‘soft’ in Latin—though technically, some of the consonants do other things than just become lenis).
- If a consonant ended up after a nasal sound (n, m, or ng), the two kind of melted together into a mixture of the two sounds. So unvoiced consonants (p, t, c, f) became voiced (b, d, g, β) (*); and voiced consonants (b, d, g) became continuants (m, n, ng), i.e., they merged with the nasal itself.
- If a consonant came after an s, it was left alone and nothing happened to it.
(*) At this stage in the Irish language, f was pronounced with air coming through between the lips, not between the upper teeth and lower lip. So the voiced version of f was actually just β. The broad and slender versions of β are what have developed into w and v (both written bh) nowadays.
These very simple general rules are actually all quite common in many languages, particularly the first one.
Now, since most words in Irish at that time ended in either a vowel, an s, or a nasal (a, e, i, o, u, s, n, m, and ng make up a good part of the Irish alphabet), this meant that the initial consonant of a following word would be somehow affected.
So what happened, then, if the second word began with a vowel, not a consonant?
Well, basically speaking: nothing. That is, the second word wasn’t affected, ’cause there was no consonant to affect.
BUT: if the
first word ended in vowel + consonant (very common), and the second word started in a consonant, then suddenly the consonant that finishes the first word ends up between two vowels—and voilà, you’ve got the right circumstances for lenition!
So let’s start with words that ended in
-s. There were many of these. What happens to an s if it’s between vowels? It gets lenited to
ṡ/sh, pronounced as a regular h. You end up with a word that ends in an h, and the following word that starts with a vowel.
Now, not long after this stage, the Irish started getting
really sloppy with their pronunciation. They started making unstressed vowels really short and dropping off entire syllables like nobody’s business. And especially final syllables in words were ruthlessly cut off—basically, in order for a final syllable to survive, it
had to end in a long vowel plus at least one consonant (that wasn’t a nasal). Everything shorter than that was just dropped entirely, and even the ones that were kept were reduced just a short schwa ([ə]) vowel. But the changes that had happened across the boundaries of words had become permanent already, and they stayed even after the factors that made them possible were gone. That basically leaves us with what we have now: words that just
cause some kind of mutation on following words, for no apparent reason.
So if we do a bit of (perhaps not quite correct, but still illustrative) reconstruction, we can make a few scenarios that can show how these things worked. Now remember that Old Irish had case endings. I’ll just use six here: the (regular, thematic) nominative and genitive singular, and the genitive plural, in both masculine and feminine. They had the following endings:
Masculine: Nsg
-os / Gsg
-ī / Gpl
-ōnFeminine: Nsg
-ā / Gsg
-ās / Gpl
-ōnAs you can see, of these three case endings, only the feminine genitive is long enough to be kept at all (the line over the vowels mean long vowel). All the others are lost completely (though the masculine genitive in
-ī was there long enough to make sure that the consonant before it became slender, which is why the genitive singular is now formed in the first declension by making the final consonant slender;
an t-arán ->
an aráin).
These case endings were used for (many, though not all) nouns, and also for both the article and the possessive adjectives (which were formed with the genitive case endings). So we can reconstruct those more or less like this:
ArticleMasculine: N
indos / G
indīFeminine: N
indā / G
indāsPlural: G
indōnPossessive adjectivesMasculine (‘his’):
eyīFeminine (‘her’):
eyāsPlural (‘their’):
eyōnIf we add some words to that, we can start to see how the results come about. Let’s just use
arán ‘bread’, which is nice and simple to reconstruct as a masculine word
arānos, and
maith ‘good’, which, for the sake of simplicity, we’ll reconstruct here as
matios (it was really just
matis, but that has different forms, so we’ll pretend it was
matios):
Nsg
indos arānos matios ‘the good bread’
Gsg
indī arānī matiī ‘of the good bread’
Gpl
indōn arānōn matiōn ‘of the good breads’
Masculine:
eyī arānos matios ‘his good bread’
Feminine:
eyās arānos matios ‘her good bread’
Plural:
eyōn arānos matios ‘their good bread’
So, what happens first? Well, consonants between vowels are weakened. This applies both inside words and across word boundaries. I’ll use the modern forms with h to indicate weakened consonants. Consonants after nasals are eclipsed here as well, but there aren’t any of those here, so we’ll leave that be.
Nsg
indosh arānos mathios ‘the good bread’
Gsg
indī arānī mhathiī ‘of the good bread’
Gpl
indōn arānon mathiōn ‘of the good breads’
Masculine:
eyī arānos mathios ‘his good bread’
Feminine:
eyāsh arānos mathios ‘her good bread’
Plural:
eyōn arānos mathiōs ‘their good bread’
Next thing that happens is that final syllables start to drop rapidly—
but they leave a reminder of their quality (broad/slender) on the consonant before the last vowel before leaving. And if the lost syllable ended in a (now lenited) s or nasal, that consonant was reinterpreted as being the beginning sound of the next word, rather than ending sound of the first. I’ll indicate the slenderisation in the modern Irish way, too, by adding an i before the consonant if it would otherwise be broad, and the re-segmentation of final consonants from first to last word by putting a hyphen after the consonant.
Nsg
ind sh-arān maith ‘the good bread’
Gsg
ind arāin mhaith ‘of the good bread’
Gpl
n(d)o n-arān maith ‘of the good breads’
Masculine:
e arān maith ‘his good bread’
Feminine:
e sh-arān maith ‘her good bread’
Plural:
e n-arān maith ‘their good bread’
Now, both the article and the possessive adjectives don’t have much stress. Actually, they’re about as unstressed as you can get. So finer details are easily lost. (See how
indōn weirdly doesn’t give
ind above, but rather
no(n)? That’s probably the same—in such a short word in an unstressed position, the long vowel probably attracted what little stress the word did have, and the first syllable was lost instead)
As such, it can be hard to tell here whether the possessives were really
e or
a—it was probably neither, but rather a simple schwa [ə], just as nowadays. The same goes for the article, which was probably just [ənd], rather than really [ind]. For ease of reading, I’ll write them with a henceforth, as in the modern language. Similarly, it can be a bit of a pain pronouncing nd if you’ve got another consonant following it (which you very often had, since most words began with consonants), so the d in the article was lost quite quickly after this, too.
But have a look at the article in the nominative masculine and genitive feminine. It is now
ind, and because it used to be
indos(h), there’s a lenited s before the following vowel—a sound that at the time must have seemed like it just popped out of nowhere. Before the last syllable was lost, they would still have been aware that this lenited s (which was pronounced like a regular h, most likely) was really an s, because if the following word began with a consonant, there would still be a regular s (it wouldn’t be between vowels, so no weakening). Once the final syllable was lost, though, there was
nothing if the following word began with a vowel, so there was no connection to anything. It was just an h now.
And what happens when you have a word that ends in a d + a word that begins in an h? In Irish, the two basically collided, and d+h gave (quite logically) simply a t—remember, Irish dh is a weakened form of d; it has nothing to do with the sound h! So they were basically left with a t that came from out of nowhere.
So when you apply these little spelling changes (and change the lengthening mācrōns for fádás), turn the sh’s into regular h’s, and turn the d+h into a t, you actually get exactly what Modern Irish has:
Nsg
an t-arán maith ‘the good bread’
Gsg
an aráin mhaith ‘of the good bread’
Gpl
na n-arán maith ‘of the good breads’
Masculine:
a arán maith ‘his good bread’
Feminine:
a h-arán maith ‘her good bread’
Plural:
a n-arán maith ‘their good bread’
The
only thing that’s changed in the spelling since this is that there’s no longer a hyphen after the h.
But if you look at this together, and remember how things came about, you’ll never more be confused about whether lenition and a prefixed h are the same or not:
- Words that cause lenition must originally have ended in a vowel (otherwise the lenited consonant would not have been weakened)
- Words that cause eclipsis must originally have ended in a nasal consonant
- Words that prefix h to vowels must originally have ended in an s, and that s is actually the h
So if a word causes h to be prefixed to a word, you can tell easily that the same word cannot possibly cause lenition on a similar word. Simple as that.
And if you always remember that those four endings in the singular:
-os/-ī | -a/-ās and associate them with mutations, then you’ll easily remember if it’s the masculine or the feminine that causes the lenition or the prefixing.
Phew. Well, that took ages to type out and think through. I hope someone finds it helpful, at least.
Login
Register
FAQ
Search
Féabar Mac
Wow, that is the most long-winded misinterpretation of a comment I've seen to date. 
You pretty much verified that at the beginning of your reply. 
