Here's a bit of sparring about all this Notional Subject business, involving Mr King himself ... basically, somebody has written to him asking what all this notional subject stuff means ... I warn you now, twiglets are involved.
CORRESPONDENT A
When Gareth King says that there is soft mutation after the notional subject, he has in mind exactly sentences like:
"Ciciodd John gi".
Of course the traditional way of stating this is that "ci" is mutated because it is the object, not because it follows directly after the subject.
The only reason he wants to rephrase it his way is because he thinks it also covers the mutation in a sentence like :
"Mae'n dda da fi dwiglets".
He thinks that "fi" there is a kind of "notional subject" (though not regarded as the grammatical subject historically, obviously).
So Gareth King thinks his rule has the advantage over the traditional way of stating things in that rather than needing two rules (one for the direct object and one for a sangiad situation), you can make do with one: "there is soft mutation after the notional subject".
I don't like his way of putting it at all. I see two problems:. Firstly, calling "fi" the subject in any way in "Mae'n dda da fi dwiglets" seems just implausible to me. The dubious ingenuity required seems to cancel out any simplicity generated.
Secondly, what about an abnormal sentence like this?:
"John a giciodd gi"
You can't say there that "ci" is coming after the notional subject ("John") can you?
GARETH KING'S REPLY ..
Well, this is the priority issue (of two) that I'm seriously going to defend (I'll come to other one - the treiglad llaes - in a minute).
dieuog's example here is excellent, and here's why: it supports my position, which as you probably all know has always been this: that the SM in Welsh has a syntactic function as a Subject-Object separator...a useful function in a language where the verb normally cannot do the job because it precedes both S and O. Verbs are semantically easily distinguishable from nouns and noun-phrases, so they make a good delineator. But nouns and noun-phrases are not easily distinguishable from each other. Let's take a moment to examine "Mae'n dda 'da fi dwiglets' (correct), or in my case 'Mae'n gas 'da fi dwiglets' (obviously also correct):
First, "fi" certainly isn't the syntactic subject in this sentence (it's pretty well a subjectless sentence, or at least a dummy-subject sentence), but it certainly is the semantic subject (perhaps I should have used this word instead of 'notional') - the person/thing associated with the action/state ("dislike") is "fi", and nothing else. Same goes for "Rhaid i mi fynd" - it's definitely "fi" that's associated with the obligation to go. My argument for the S-O separator function of SM is based on the 'deep-level' semantics rather than the 'surface-level' syntax. And I think it makes both perfect and neat sense.
OK - so:
Mae'n gas 'da fi dwiglets (correct)
therefore: Mae'n gas 'da Ffred dwiglets (correct) - this incidentally torpedoes the old (and lazy) line about 'pronouns causing SM - what tosh!) :)
therefore: Mae'n gas 'da'r dyn dwiglets (correct)
therefore: Mae'n gas 'da'r dyn fara (correct - 'the man hates bread')
Look what happens if we expand nouns to noun-phrases:
Mae'n gas 'da'r dyn Fara ei fam 'the man hates his mother's bread' - SM indicates that O phrase is 'bara ei fam', cos bara has SM.
Mae'n gas 'da 'r dyn bara ei fam 'the bread man hates his mother' - bara unmutated, therefore still part of the Subject (which it is - the semantic subject here is obviously 'the bread man' - he's the one doing the hating. Never mind the fact that the syntax of this Welsh construction involves a dummy subject, it's the meaning (in the heads of speakers) that counts.
So I'm pretty confident about this VERB-SUBJECT-SM-OBJECT theory, and a number of people have said they agree with me, including my friend Phyl Brake who's written books himself, and Barack Obama. I made that last bit up.
GARETH KING'S REPLY TO CORRESPONDENT A's SECOND POINT ...
Correspondent A wrote:
Secondly, what about an abnormal sentence like this?:
"John a giciodd gi"
You can't say there that "ci" is coming after the notional subject ("John") can you?
No, but the abnormal sentence in Welsh is what we call a 'marked' order - in other words it's derived from primary word order (VSO) and is in this sense secondary. When you push the Subject to before the verb in this kind of focused sentence, you're really leaving behind an empty subject-slot after the verb, and the SM remains. It's the same sort of thing that happens with imperatives - "Rho DDwy bunt i mi" ('give me two pounds') again involves something unusual, this time the removal of the pronoun - this happens to be a rule in Welsh (as indeed many other languages) - but the slot is there, even if it's vacant. It's really "Rho DI ddwy bunt i mi", and you can tell it's theree, because we can put the "di" in if we want to change the emphasis a bit. Similarly "Aros fan hyn" 'Stay here' or 'Aros di fan hyn" 'You stay here'.
Am I winning you over here at all?
THIS IS CORRESPONDENT A AGAIN ...
Okay, I've been wanting to say this, so I'll say it here.
I'm very glad Gareth King is here and is taking learners' talk about him in good part. Frankly, one doesn't expect writers to move in the same world as readers, and so readers can sometimes express themselves rather... immoderately. :wink: "Colloquial Welsh" and the "Grammar Workbooks" were the first things I worked my way through in learning Welsh, before doing several summer courses in Llambed years ago. They gave me a fantastic start on learning Welsh, for which I'm very grateful. (Something I always think of is the massive amount of vocabulary introduced in "Colloquial Welsh".) I guess the relationship of learners to the books they use is a very close one that can produce violent expressions of opinion. I remember big chunks by heart, and then I also remember all the instances of coming across speakers who did something different from what was introduced or recommended in the books (usually something involving a more strict or formal application of a rule), and thinking "Hey, I've been misled!" Colour me oedipal. :oops:
Thanks for the note on mutation in those situations. I'm not entirely won over, yet I'm not entirely resistant either, which is the appropriate state of mind if grammatically it's a matter of more than one structure happening at the same time ("syntactic/grammatical" subject versus "notional/logical subject")! I'm thinking about it, and I'll go back and read some more too.
I guess there are two accounts here of what's going on in the sentence "Mae'n gas 'da'r dyn dwiglets". One is the traditional one that appeals to the transformational mechanism of "sangiad" to explain the mutation, and one is your idea of the "notional subject". And maybe you can use both, with each generalising what's going on in this sentence in powerful ways, but in totally different directions.
After all, there is one area where the traditional (sangiad-transformational) account has nothing to say, namely the question of why the sangiad is necessary. Why ever say "Byddai'n dda 'da'r dyn dwiglets" instead of "Byddai twiglets yn dda 'da'r dyn"? Why say "Gwell inni fynd" instead of "Gwell mynd inni"? The "notional subject" account I guess explains the urge to arrange the sentence like this, by saying that these sentences are using the same kind of mental template as "Licai'r dyn dwiglets" and "Dylen ni fynd"...
The focused sentence question... I have to think about that and see if I can come up with any really knock-down
WELL ... what about all that then? If you want to read the whole thing, and/or see who these people were, here's the link to it ...
http://www.forumwales.com/fwforum/viewtopic.php?f=45&t=8413
After all that brain-work (?) here's a return to normal with the Music Spot... and today I have selected Gruff Rhys playing a couple of songs whilst telling us all about his sister's guitar etc ...