linguistics geekery
Oct. 5th, 2005 01:06 pm![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
pthalo posting in ![[community profile]](https://www.dreamwidth.org/img/silk/identity/community.png){kind=small}
multiplicity_archivescross posted from our journal
I've either lost my mind or am a genius. On the bus home, I realise Chomsky's x-bar theory and the minimalist theory he replaced it with, can be reworded quite nicely to represent multiplicity. First we have x-bar theory with its rigid tree structure, D-str, S-str, movement, binding, government. This is the classic medical definition of DID. The early days when everybody thought that all systems had cookie cutters alters complete with an ISH and protectors and a clueless host and all that.
For your reference, here is a sentence diagram in traditional x-bar theory:
I'd label our clueless host as CP, which is the whole sentence, the one therapists want to talk to, the one allegedly on top of the heirarchy but c-commands nothing. Integrate, and you're left with CP.
I can't decide who should go into Spec CP, the ISH or the protectors. I think I might prefer to have them all waiting in the wings in a large NP under IP and leave Spec CP blank. Then they can move up to the front, to Spec CP or to C° (if someone else is fronting and they're waiting in the wings giving advice) whenever necessary. While they're at home in NP they can happily c-command everybody else.
Then everyone else can go under VP and various complements of VP into a nice heirarchcial structure. Maybe the angry alters could be in V° and the meeker alters could be stuck in some prepositional phrase somewhere. And of course we have it all arranged so that the appropriate people are c-commanding the appropriate people and the lils are all the way at the bottom, some of whom c-command each other.
But of course Chomsky has disgarded x-bar in favour of a minimalist approach lacking D-str and S-str and wherein sentences are built dynamically and movement happens while we're building sentences to do feature checking and the nodes are no longer labeled Spec CP, C°, NP, IP, VP, Infl, but rather words are used instead. But of course we still label them because otherwise we can't discuss them. But we now label them knowing that the labels are a tool that we impose on syntax rather than something that is a part of syntax.
So we no longer have to have a sentence tree made up of ISH and Protector and Host and all that. We can use people's names to label nodes without worrying about what sort of job they should have. We no longer have:
Instead, we can have:
So we can make little tress like:
And that can mean a bunch of different things. If you want, Jess and Allison split off of Sally. Or Jess and Allison integrated and got Sally, but you can be lots more creative than that and say that Jess and Allison symmetrically c-command each other while Sally c-commands neither of them, where c-command could mean they protect each other (And Sally protects neither) or maybe they're friends with each other but not with Sally or Jess and Allison have equal power in relationship to each other but not in relation to Sally.)
(Pthalo is Marissa's girlfriend and Jolie's mother)
This says that Pthalo and Marissa symmetrically c-command each other. In this case, I mean "are equals". Not in the "equally important to the system" sense, because who am I to judge such a thing and I'd rather say everyone's equally important. But rather in the sense that we have equal power over each other. It's not a relationship where one is always the strong one and the other is always the weak one. Responsibilities for the health of the relationship are shared, etc. Of equal status and power in relation to each other.
Jolie's my (Pthalo's) kid. She is 3 and calls me mommy. I asymmetrically c-command her. She does not c-command me. I am strong loving adult. She is inquisitive child. Sometimes scared, sometimes happy, sometime sad, sometime cranky, sometimes excited, often playful. She is in my governing category. Since she does not c-command me, I don't go to her and burden her with "mommy had a bad dream" or "mommy thinks God hates her" or whatever nonesense my brain might produce. I wouldn't bother an outside kid with my personal nonesense either. I mother her. She is mothered. No role reversals. Sometimes, Jolie tries to move into a position where she could c-command me, but due to feature checking (and my resisting it) she isn't able to do it successfully. I'm careful not to let her take on my burdens when I know she won't be able to handle them.
It would be interesting to map out an entire system and show movement and binding and c-command relationships, but I can only do this for the relationships I know best, if I want to get something even remotely accurate. And the others are less linguistics geeks than I am!
It'd also be cool to think of different applications for c-command or movement to show different aspects of system relations.
I think for our system I'd label the body CP, where we put clueless host under x-bar theory. The body is mildly sentient, cannot visit the other world and c-commands nothing. Calling it CP works.
It's also kind of neat to put body as CP because whoever's fronting could sit in Spec CP. And if you have people who don't/can't front, they could reside somewhere far down the tree where they couldn't get to Spec CP without creating a subjacency violation. The possibilities are literally endless.
If nothing else, I'll pass syntax. :p
I've either lost my mind or am a genius. On the bus home, I realise Chomsky's x-bar theory and the minimalist theory he replaced it with, can be reworded quite nicely to represent multiplicity. First we have x-bar theory with its rigid tree structure, D-str, S-str, movement, binding, government. This is the classic medical definition of DID. The early days when everybody thought that all systems had cookie cutters alters complete with an ISH and protectors and a clueless host and all that.
For your reference, here is a sentence diagram in traditional x-bar theory:
CP
/\
Spec C'
/\
C° IP
/ \
NP I'
/ / \
N' I° VP
/ +tense |
N° V'
I /\
V° NP
see /\
Det N'
| |
the N°
boy
(I saw the boy)
I'd label our clueless host as CP, which is the whole sentence, the one therapists want to talk to, the one allegedly on top of the heirarchy but c-commands nothing. Integrate, and you're left with CP.
I can't decide who should go into Spec CP, the ISH or the protectors. I think I might prefer to have them all waiting in the wings in a large NP under IP and leave Spec CP blank. Then they can move up to the front, to Spec CP or to C° (if someone else is fronting and they're waiting in the wings giving advice) whenever necessary. While they're at home in NP they can happily c-command everybody else.
Then everyone else can go under VP and various complements of VP into a nice heirarchcial structure. Maybe the angry alters could be in V° and the meeker alters could be stuck in some prepositional phrase somewhere. And of course we have it all arranged so that the appropriate people are c-commanding the appropriate people and the lils are all the way at the bottom, some of whom c-command each other.
But of course Chomsky has disgarded x-bar in favour of a minimalist approach lacking D-str and S-str and wherein sentences are built dynamically and movement happens while we're building sentences to do feature checking and the nodes are no longer labeled Spec CP, C°, NP, IP, VP, Infl, but rather words are used instead. But of course we still label them because otherwise we can't discuss them. But we now label them knowing that the labels are a tool that we impose on syntax rather than something that is a part of syntax.
So we no longer have to have a sentence tree made up of ISH and Protector and Host and all that. We can use people's names to label nodes without worrying about what sort of job they should have. We no longer have:
VP
/\
V° NP
eats \
N'
|°
N°
applesInstead, we can have:
eat / \ eat apples
So we can make little tress like:
Sally /\ Jess Allison
And that can mean a bunch of different things. If you want, Jess and Allison split off of Sally. Or Jess and Allison integrated and got Sally, but you can be lots more creative than that and say that Jess and Allison symmetrically c-command each other while Sally c-commands neither of them, where c-command could mean they protect each other (And Sally protects neither) or maybe they're friends with each other but not with Sally or Jess and Allison have equal power in relationship to each other but not in relation to Sally.)
?
/ \
Pthalo Marissa
|
Jolie(Pthalo is Marissa's girlfriend and Jolie's mother)
This says that Pthalo and Marissa symmetrically c-command each other. In this case, I mean "are equals". Not in the "equally important to the system" sense, because who am I to judge such a thing and I'd rather say everyone's equally important. But rather in the sense that we have equal power over each other. It's not a relationship where one is always the strong one and the other is always the weak one. Responsibilities for the health of the relationship are shared, etc. Of equal status and power in relation to each other.
Jolie's my (Pthalo's) kid. She is 3 and calls me mommy. I asymmetrically c-command her. She does not c-command me. I am strong loving adult. She is inquisitive child. Sometimes scared, sometimes happy, sometime sad, sometime cranky, sometimes excited, often playful. She is in my governing category. Since she does not c-command me, I don't go to her and burden her with "mommy had a bad dream" or "mommy thinks God hates her" or whatever nonesense my brain might produce. I wouldn't bother an outside kid with my personal nonesense either. I mother her. She is mothered. No role reversals. Sometimes, Jolie tries to move into a position where she could c-command me, but due to feature checking (and my resisting it) she isn't able to do it successfully. I'm careful not to let her take on my burdens when I know she won't be able to handle them.
It would be interesting to map out an entire system and show movement and binding and c-command relationships, but I can only do this for the relationships I know best, if I want to get something even remotely accurate. And the others are less linguistics geeks than I am!
It'd also be cool to think of different applications for c-command or movement to show different aspects of system relations.
I think for our system I'd label the body CP, where we put clueless host under x-bar theory. The body is mildly sentient, cannot visit the other world and c-commands nothing. Calling it CP works.
It's also kind of neat to put body as CP because whoever's fronting could sit in Spec CP. And if you have people who don't/can't front, they could reside somewhere far down the tree where they couldn't get to Spec CP without creating a subjacency violation. The possibilities are literally endless.
If nothing else, I'll pass syntax. :p
![[identity profile]](https://www.dreamwidth.org/img/silk/identity/openid.png){kind=small}
no subject
Date: 2005-10-05 02:36 pm (UTC)I had to write a computer program to generate x-bar trees from sentences once.
On a final exam.
I'm kind of afraid to take the graduate version of artificial intelligence now.
----
What about a more general structure: graphs (http://en.wikipedia.org/wiki/Graph_theory) with labeled edges. You can do fun things with graphs. Like run Google PageRank (http://en.wikipedia.org/wiki/PageRank) or the Advogato trust metric (http://www.advogato.org/trust-metric.html) over them. And that would be deeply amusing.
no subject
Date: 2005-10-05 02:56 pm (UTC)I think I'm just a big geek.
We did very basic sentence trees in 1st year in Descriptive Grammar and then we threw that out and did x-bar and all its nuances in Syntax as 2nd year students and now as 3rd year students in Syntax-Semantics we're learning to throw out x-bar and think like modern linguistics. It's all so very exciting.
Except for the part where I have to write a scary thesis contrasting negation in English, Hungarian, and ASL.
Pthalo
no subject
Date: 2005-10-05 09:03 pm (UTC)no subject
Date: 2005-10-06 01:45 pm (UTC)http://en.wikipedia.org/wiki/Minimalist_syntax (a more up to date look on syntax)
http://en.wikipedia.org/wiki/C-command c-command! The way you figure out if somethign c-commands something, for example what does "see" c-command in the first example, "I saw the boy"? You go up one node (to V') and then you look down. Everything at the same level or below "see" that is connected somehow to that first node you see (NP, in our case) is c-commanded by "see". So "see" c-commands "the boy"
can't find any simple explanations of subjacency, but the basic theory behind it under x-bar theory is that sentences always start out in one format and then to get different formats you have to move stuff around. So "What did you see?" Started out as "You did see what" and the word "what" moves to the beginning of the sentence (To Spec CP) and "did" also moves, to C°, leaving a little trace behind that no one can see. Subjacency says how far an item can move. The rule is that it can't move past more than one IP or NP. So it heads upwards along the tree and once it hits that second IP or NP, it's gone and it can't go to Spec CP or whereever it was going to. This is why sentences like "Which girl do you believe that who said he would kiss?" doesn't have any meaning. (It's trying to say "Who said he'd kiss a girl and which girl did he say he would kiss?")
So if someone can't front for some reason or another, it's analogous to something trying to move into Spec CP (where I'd placed "the front") and not being able to get there because of a subjacency violation.
I don't know if this is helping or making it worse. :/
Pthalo
no subject
Date: 2005-10-06 01:46 pm (UTC)no subject
Date: 2005-10-06 10:32 pm (UTC)no subject
Date: 2005-10-06 02:19 am (UTC)no subject
Date: 2005-10-06 04:58 pm (UTC)