Syntax random

[zompist]

(50) Mary persuaded John to hit Fred

Now, I realise that I am horribly ignorant of syntactic matters, but to me this seems to contradict what everyone was saying earlier about raising.

It shouldn't, because it's not Raising. It's Subject and Object Control.

The difference is that, as Dixon says, John really is the direct object of "persuaded". Who did Mary persuade? John. "Mary persuaded John" is a complete, valid sentence. So yes, John is object in the main clause, subject in the subclause.

Compare "Mary wants him to hit Fred." The meaning is not "Mary wants John"; what Mary wants here is not John, but [John hits Fred]-- the object is the entire subclause.

[bradrn]

(50) Mary persuaded John to hit Fred

Now, I realise that I am horribly ignorant of syntactic matters, but to me this seems to contradict what everyone was saying earlier about raising.

It shouldn't, because it's not Raising. It's Subject and Object Control.

The difference is that, as Dixon says, John really is the direct object of "persuaded". Who did Mary persuade? John. "Mary persuaded John" is a complete, valid sentence. So yes, John is object in the main clause, subject in the subclause.

Compare "Mary wants him to hit Fred." The meaning is not "Mary wants John"; what Mary wants here is not John, but [John hits Fred]-- the object is the entire subclause.

But Dixon’s analysis still holds even when using a proper example of raising — both “Mary wants herself to hit Fred” and “Mary wants him to hit himself” are grammatical.

(Oh, and by the way, I didn’t get a notification for your post. Any idea what might have gone wrong?)

[zompist]

But Dixon’s analysis still holds even when using a proper example of raising — both “Mary wants herself to hit Fred” and “Mary wants him to hit himself” are grammatical.

Only the second one is grammatical for me. The first would be "Mary wants to hit Fred."

Though, confusingly, there's another reflexive construction, e.g. "Mary wants to hit Fred herself." The implication is that she needs to do it personally.

(Oh, and by the way, I didn’t get a notification for your post. Any idea what might have gone wrong?)

I had to fix up the quotes and it probably got confused.

[bradrn]

But Dixon’s analysis still holds even when using a proper example of raising — both “Mary wants herself to hit Fred” and “Mary wants him to hit himself” are grammatical.

Only the second one is grammatical for me. The first would be "Mary wants to hit Fred."

Well, you’re right, I suppose — the first one is not quite as acceptable to me as the second one. Nevertheless, I would still call both of them grammatical.

(Oh, and by the way, I didn’t get a notification for your post. Any idea what might have gone wrong?)

I had to fix up the quotes and it probably got confused.

It seems to be working now, so that must have been it.

[zompist]

Looking at Dixon again, I'm not sure what your contradiction is? Dixon looks right to me.

Getting this right is a matter of rule ordering, and the cycle. (Terms and mechanisms will differ in Minimalism.)

Mary persuaded he_i [him_i to hit him_i]
Reflexivization in subclause: Mary persuaded him [him to hit himself_i]
Control: Mary persuaded him [t to hit himself]

Mary_j persuaded her_j [her_j to hit him_i]
Reflexivization in subclause can't apply
Control: Mary persuaded her [t to hit him]
Reflexivization in main clause: Mary persuaded herself [to hit him]

(I've shown the subject in the subclause as oblique, since arguably all pronouns in non-finite subclauses are oblique. That's what we actually see in sentences like "For him to leave town is impossible.")

[bradrn]

Looking at Dixon again, I'm not sure what your contradiction is? Dixon looks right to me.

Getting this right is a matter of rule ordering, and the cycle. (Terms and mechanisms will differ in Minimalism.)

Mary persuaded he_i [him_i to hit him_i]
Reflexivization in subclause: Mary persuaded him [him to hit himself_i]
Control: Mary persuaded him [t to hit himself]

Mary_j persuaded her_j [her_j to hit him_i]
Reflexivization in subclause can't apply
Control: Mary persuaded her [t to hit him]
Reflexivization in main clause: Mary persuaded herself [to hit him]

(I've shown the subject in the subclause as oblique, since arguably all pronouns in non-finite subclauses are oblique. That's what we actually see in sentences like "For him to leave town is impossible.")

The ‘contradiction’ I was talking about is in the underlying sentence: Dixon (and you) say it’s ‘Mary persuaded he_i [him_i to hit him_i]’, but I was under the impression that it was ‘Mary persuaded [he_i to hit him_i]’, with the ‘him’ coming from raising-to-object. (I’m probably completely wrong in this though, since as I mentioned I don’t know much syntax.)

[akam chinjir]

Yeah, it's not raising, it's control, as zompist said. Though it's not as simple as deletion, I think.

Supposedly you've got something like this:

Mary persuaded John [John to leave]

And then you delete the second John because it corefers with the object of the matrix clause.

That makes it seem like you could have something like this:

Mary persuaded John [Fred to dance]

And then you wouldn't delete "Fred," because it doesn't corefer with "John" The problem is, of course, that the result isn't grammatical. I don't think the issue is just semantic (maybe John has authority over Fred's dancing behaviour).

With "persuade," in English, I'm pretty sure(*) the embedded subject has to corefer with the matrix object. So the rule here can't just be that you delete the embedded subject whenever it happens to corefer with the matrix object. So this is "obligatory control."

In some languages there's another possibility, topic drop. If you're free to drop topics, then you'll often get embedded clauses with no overt subject, but this won't really be about control, it'll just be topic drp. (Or pro drop, which I guess would work the same in these contexts.)

That's actually an issue that arises with Dixon's analysis of Dyirbal. Control is one of the issues people look at when thinking about syntactic ergativity: in most languages, even ones that have other syntactically ergative patterns, only the subject of an embedded clause can be controlled. Dixon argued that Dyirbal is an exception to this, and allows control of absolutive arguments generally, but apparently Dyirbal allows topic drop and there's some question of what's really going on here. (Citation: Aissen, Dyirbal ergativity. She also argues against Dixon's conclusions about coordination structures. I don't have an opinion about either topic.)

(*NOTE.) Does anyone thing something like "Mary persuaded John Fred to be fired" could be ok? (If so, would it entail that Mary persuaded John to fire Fred, or could it be someone else who's to do the firing?)

Edit. Oh, and similar issues come up in arguments about whether Pirahã has recursion, since a control structure would be recursive but a topic-drop one maybe wouldn't have to be. For example: https://ling.auf.net/lingbuzz/004928

[zompist]

The ‘contradiction’ I was talking about is in the underlying sentence: Dixon (and you) say it’s ‘Mary persuaded he_i [him_i to hit him_i]’, but I was under the impression that it was ‘Mary persuaded [he_i to hit him_i]’, with the ‘him’ coming from raising-to-object. (I’m probably completely wrong in this though, since as I mentioned I don’t know much syntax.)

Righto. But yeah, it's Control, not Raising. There are syntactic arguments for this, but it's easiest just to look at the semantics-- as I said, "persuade" has a human direct object, as well as an optional sentential adjunct.

[Richard W]

There's something different between persuades in Mary persuades John to hit Fred and Mary persuades John. It's even greater with the verbs command and encourage. I fear the former structure is quite capable of having its own grammar rules that don't derive from the claimed constituents, e.g. for when the infinitive is in the progressive.

[bradrn]

The ‘contradiction’ I was talking about is in the underlying sentence: Dixon (and you) say it’s ‘Mary persuaded he_i [him_i to hit him_i]’, but I was under the impression that it was ‘Mary persuaded [he_i to hit him_i]’, with the ‘him’ coming from raising-to-object. (I’m probably completely wrong in this though, since as I mentioned I don’t know much syntax.)

Righto. But yeah, it's Control, not Raising. There are syntactic arguments for this, but it's easiest just to look at the semantics-- as I said, "persuade" has a human direct object, as well as an optional sentential adjunct.

Alright, so I used an invalid example. It doesn’t matter in this case — my argument applies equally if you substitute ‘persuade’ with a raising predicate such as ‘want’: Dixon and you say it’s ‘Mary wants he_i [him_i to hit him_i]’, but I thought it was ‘Mary wants [he_i to hit him_i]’.

That's actually an issue that arises with Dixon's analysis of Dyirbal. Control is one of the issues people look at when thinking about syntactic ergativity: in most languages, even ones that have other syntactically ergative patterns, only the subject of an embedded clause can be controlled. Dixon argued that Dyirbal is an exception to this, and allows control of absolutive arguments generally, but apparently Dyirbal allows topic drop and there's some question of what's really going on here. (Citation: Aissen, Dyirbal ergativity. She also argues against Dixon's conclusions about coordination structures. I don't have an opinion about either topic.)

After researching this for my next ergativity post, I’ve come to the conclusion that ergative control most probably does not exist. For Dyirbal, I’m inclined to agree with Manning about the purported ‘control’ constructions: they’re actually an instance of coordination rather than control, making their ergative alignment consistent with the S/O pivot for conjunction reduction. But I hadn’t come across that Aissen paper before — that’s an interesting argument against the conjunction reduction argument, though I’m not sure it successfully argues against Dyirbal having conjunctoin reduction.

(*NOTE.) Does anyone thing something like "Mary persuaded John Fred to be fired" could be ok? (If so, would it entail that Mary persuaded John to fire Fred, or could it be someone else who's to do the firing?)

For me, that’s about as ungrammatical as you can get.

[zompist]

The ‘contradiction’ I was talking about is in the underlying sentence: Dixon (and you) say it’s ‘Mary persuaded he_i [him_i to hit him_i]’, but I was under the impression that it was ‘Mary persuaded [he_i to hit him_i]’, with the ‘him’ coming from raising-to-object. (I’m probably completely wrong in this though, since as I mentioned I don’t know much syntax.)

Righto. But yeah, it's Control, not Raising. There are syntactic arguments for this, but it's easiest just to look at the semantics-- as I said, "persuade" has a human direct object, as well as an optional sentential adjunct.

Alright, so I used an invalid example. It doesn’t matter in this case — my argument applies equally if you substitute ‘persuade’ with a raising predicate such as ‘want’: Dixon and you say it’s ‘Mary wants he_i [him_i to hit him_i]’, but I thought it was ‘Mary wants [he_i to hit him_i]’.

Where does Dixon say that? In the passage you cited, he uses "persuade". The verbs don't behave the same, that's the whole point!

With "want", you should have

Mary wants [him_i to hit him_i]
Reflexivization: Mary wants [him_i to hit himself_i]
Raising: Mary wants him_i [t to hit himself_i]

Raising is needed for sentences like "Mary wants herself to win", to move "her" to the main clause so it can be reflexivized.

Informally: Raising moves things into positions that make no sense semantically, but are comfortable syntactically. Another example is "Mary is likely to win". Mary isn't being called likely. What's likely is the sentence "Mary (will) win".

[zompist]

There's something different between persuades in Mary persuades John to hit Fred and Mary persuades John.

It's different, but only like "Mary vacationed" and "Mary vacationed at a chalet overlooking a lake last November".

There's a pragmatic difference between giving the additional information or not, of course. Normally we're quite interested in saying what the proposition was. But it could have been stated earlier, or implied from the context.

Of course you can say "Mary wanted John," but the difference there is not the absence of a sentential object. It's a very different sort of wanting, and has no sentential object at all.

Edit: Well, the GS people actually said it had one: it derived from "Mary wanted [(Mary) to have John]". But that only underlines the difference from the other sentences we've been discussing-- John is the object there, not the subject.

[bradrn]

I suppose the only conclusion I can draw from this is that Dixon is so remarkably bad at syntax that he can’t even distinguish raising and control predicates, and so it’s probably best to not worry too much about anything he says on the subject…

(I did try one more time to figure out what he was trying to say, but I think he got hopelessly muddled up by trying to give an example of raising using the control predicate ‘persuade’, so that particular section doesn’t seem to make much sense in the light of what everyone else has said.)

_______


Anyway, I have another (unrelated) question: After talking to everyone on this thread, you’ve all managed to convince me that syntax is much more interesting than I used to think, so I decided to try learning a bit of syntax. Specifically, I’m thinking of starting with Minimalism, as it seems to be the most current theory. Does that seem like a good idea? As for resources, I managed to find Radford’s Minimalist Syntax: Exploring the structure of English — is that a good book to learn from? And if not, does anyone have any other recommendations?

[zompist]

I suppose the only conclusion I can draw from this is that Dixon is so remarkably bad at syntax that he can’t even distinguish raising and control predicates, and so it’s probably best to not worry too much about anything he says on the subject…

Ah, I didn't notice he said "raising" in there. I'm not sure when the distinction was introduced, but it dates back to the 1970s at least.

Anyway, I have another (unrelated) question: After talking to everyone on this thread, you’ve all managed to convince me that syntax is much more interesting than I used to think, so I decided to try learning a bit of syntax. Specifically, I’m thinking of starting with Minimalism, as it seems to be the most current theory. Does that seem like a good idea? As for resources, I managed to find Radford’s Minimalist Syntax: Exploring the structure of English — is that a good book to learn from? And if not, does anyone have any other recommendations?

I haven't read Radford, but someone did recommend him when I asked about Minimalism awhile back. (I read David Adger's book instead. And Chomsky's, but god oh god do not start with Chomsky.)

I think you have my syntax book... I'd suggest re-reading it, because all the basics are there and it will introduce Minimalism and other schools as well. This stuff can be hard to retain if you're not using it and reading more about syntax, so a review can help a lot.

[bradrn]

I suppose the only conclusion I can draw from this is that Dixon is so remarkably bad at syntax that he can’t even distinguish raising and control predicates, and so it’s probably best to not worry too much about anything he says on the subject…

Ah, I didn't notice he said "raising" in there. I'm not sure when the distinction was introduced, but it dates back to the 1970s at least.

Well, Basic Linguistic Theory was apparently published in 2010, so he doesn’t have much of an excuse in that case!

Anyway, I have another (unrelated) question: After talking to everyone on this thread, you’ve all managed to convince me that syntax is much more interesting than I used to think, so I decided to try learning a bit of syntax. Specifically, I’m thinking of starting with Minimalism, as it seems to be the most current theory. Does that seem like a good idea? As for resources, I managed to find Radford’s Minimalist Syntax: Exploring the structure of English — is that a good book to learn from? And if not, does anyone have any other recommendations?

I haven't read Radford, but someone did recommend him when I asked about Minimalism awhile back. (I read David Adger's book instead. And Chomsky's, but god oh god do not start with Chomsky.)

It looks like Adger has published quite a few books — which one are you talking about? (But I certainly had no intention of reading Chomsky!)

I think you have my syntax book... I'd suggest re-reading it, because all the basics are there and it will introduce Minimalism and other schools as well. This stuff can be hard to retain if you're not using it and reading more about syntax, so a review can help a lot.

I do have your syntax book; I wasn’t thinking of re-reading it, but maybe I will.

Oh, and now that you mention it: whatever happened to all those old schools of syntax (e.g. X̄-theory)? Have they been completely invalidated, or are they still used now and again?

[Kuchigakatai]

Anyway, I have another (unrelated) question: After talking to everyone on this thread, you’ve all managed to convince me that syntax is much more interesting than I used to think, so I decided to try learning a bit of syntax.

formalist syntax*, you mean

A good part of what I was trying to get with our discussion of the definitions of "word order" vs. "syntax" was that you already knew a bit of syntax, it just doesn't happen to be formalist.

Oh, and now that you mention it: whatever happened to all those old schools of syntax (e.g. X̄-theory)? Have they been completely invalidated, or are they still used now and again?

X-bar continues to be widely taught and used in the US at least. Wouldn't surprise me if more so than Minimalism, but I don't really know.

[zompist]

It looks like Adger has published quite a few books — which one are you talking about? (But I certainly had no intention of reading Chomsky!)

The one I have is Core Syntax: A Minimalist Approach. It does a great job introducing Minimalism.

Oh, and now that you mention it: whatever happened to all those old schools of syntax (e.g. X̄-theory)? Have they been completely invalidated, or are they still used now and again?

Someone said that in Cambridge (or was it Oxford?) they teach X' in the first year, Minimalism in the second.

When you have a ream of course notes, exercises, and primary research in X' theory, it's kind of hard to throw it all out! An actual research paper will normally dive into quite a few older theories, including those of Chomsky's opponents. Partly this is because when you're a graduate student studying a topic, you are expected to read all the papers on that topic. But it's also because old papers will have examples and analyses that are still important.

[bradrn]

It looks like Adger has published quite a few books — which one are you talking about? (But I certainly had no intention of reading Chomsky!)

The one I have is Core Syntax: A Minimalist Approach. It does a great job introducing Minimalism.

I’ll see if I can find that one then!

Oh, and now that you mention it: whatever happened to all those old schools of syntax (e.g. X̄-theory)? Have they been completely invalidated, or are they still used now and again?

Someone said that in Cambridge (or was it Oxford?) they teach X' in the first year, Minimalism in the second.

When you have a ream of course notes, exercises, and primary research in X' theory, it's kind of hard to throw it all out! An actual research paper will normally dive into quite a few older theories, including those of Chomsky's opponents. Partly this is because when you're a graduate student studying a topic, you are expected to read all the papers on that topic. But it's also because old papers will have examples and analyses that are still important.

So do you think it’s still worth learning X̄-theory, or should I just focus on Minimalism if I want to learn syntax today?

[zompist]

So do you think it’s still worth learning X̄-theory, or should I just focus on Minimalism if I want to learn syntax today?

It's quite all right to say X'-- that' what my textbook on X' uses.

You don't have to study X' as well, and if you're going to study two theories, I advise not making them both Chomsky's. My book has lots of recs for other books; see especially Van Valin in the bibliography.

[bradrn]

So do you think it’s still worth learning X̄-theory, or should I just focus on Minimalism if I want to learn syntax today?

It's quite all right to say X'-- that' what my textbook on X' uses.

Well, I’ve already made my own keyboard to input diacritics, so I might as well use it! (You’ll notice that I also use proper left and right quotes rather than straight quotes, and em/en dashes rather than hyphen-minus, for the same reason.)

You don't have to study X' as well, and if you're going to study two theories, I advise not making them both Chomsky's.

Well, what other theories are there? I know you listed a couple in your syntax book, but I wouldn’t know which one to focus on.