Monday, April 18, 2011

Lost in the Trees

The next day I attended a second year linguistics class which I believe was a prerequisite for one of the classes I was in the day before. It was on Linguistic Theory & Analysis, and was the first class that really made me feel like I had probably missed most of a term, and possibly a prerequisite, too. The professor is British, with wild woolly hair and his pet gesture is to rub his hands together at the breastbone level. He's wearing old jeans and a zippered sweatshirt. He starts by reviewing something about clauses and "DP"s from last class. Based on the answers he solicits from the class, no one who was there understands it either. He draws a diagram.

              Xmax
             /  \
    specifier   X'
                  /  \
                X    complement

Xmax --> (spec) X'
X' --> X (comp)

He states that "wh-phrases are in complementary distribution with complementizers in C," and "wh-phrases form a dependency with a gap further down the tree." From subsequent context I divine that wh-phrases are subordinate clauses such as "which I saw" "when the clock strikes" and "whether I like beets." His examples don't include anything like "whom I saw" because apparently linguists are hip to language change and prove it by eschewing moribund subordinate conjunctions.

Specifier is the subject, complement the predicate or possibly a subordinate clause or both, and I think C is short for complement. He compared:

I wonder [whether [he said [he was going to take that path.]]]

I wonder [which path [he said [he was going to take.]]]

Someone could say "he said he was going to take which path" so this theory of grammar imagines that "which path" has moved from the end up to near the beginning leaving a gap. Except that they pretend that it hasn't actually moved, but exists in both places at once. You just can't hear them both. I'm serious. That's what "The specifier and the head of the projection may not be simultaneously phonologically overt," means. And that's an actual quotation from the lecture. "Grammaticians (grammaticists?) say that evidence for the movement includes echo questions like "he said he was going to take which path?" and multiple wh-questions like "who was going to take which path?" In English only one wh-phrase can occupy the head position, so we don't say "Who where what ate?" unless we are being deliberately silly, but in Romanian and all the Slavic languages except Russian, that's a perfectly acceptable way to ask questions when there is more than one thing you don't know about a situation.

This pretending you have an unpronounced copy lurking at another point in the sentence apparently helps to unify language analysis. It's part of Chomskian contrast of the deep structure and the surface structure. Sentences that mean the same thing have the same deep structure and ones that are constructed the same way have the same surface structure and you hypothesize movements and rules to relate them. I really enjoy being dumped into the middle of something and having to put together what is going on from clues and guessing. But I also enjoy stability and regularity. I should be happy anywhere, then, right?

We did more stuff with trees, specifiers and complements. The professor pointed out that verbs don't get to define anything about their subjects, but the verb can have restrictions on what kind of complements it has. Apparently the absence of object expletives and of idioms with a free object but a lexically idiosyncratic subject is evidence supporting this.

I tried to read a library book on Chomskian grammar, that promised to explain the trees, but it was dulltastically unreadable and I put it back on the shelf.


I tried to come up with an example where you could possibly be confused if someone said who when your dialect called for whom. The best I could come up with was the two distinct questions:

Who did John wrong?
Whom did John wrong?

It's not the greatest, because both the verb "to wrong" and the expression "to do someone wrong" are bordering on archaic themselves. I believe that a speaker who did not use whom would distinguish them by stress and the rhythm of the sentence:

------------------------------ (very level tone)
Who DID JOHN WRONG

 /                    / (rising on who and wrong)
WHO did John wrong?

When someone I expect to use whom (like a university professor or an erudite character on a television show) doesn't use it, I find it mildly distracting. But I suppose there are more people who find it distracting when the word does show up. I don't think there's a problem with different people speaking the same language and having slightly different pronoun systems, though. What do y'all think?

16 comments:

Anonymous said...

You will probably enjoy reading Derek Bickerton's Adam's Tongue;
How Humans made Language, How Language made Humans.
In it he lays out his theory explaining the origin of human language.
Towards the end of the book he attacks Chomsky's theory that
the grammar of language is intrinsically recursive in nature.
Bikerton's refutation seems to me convincing on the face of it;
my reading it of course is as a layman.

In any case, the book is hightly readable.

Mike Eisenstadt
Austin Texas

Q. Pheevr said...

Hi! I'm a linguistics professor, and I've been reading your blog for some time now, although I don't comment very often. I've really enjoyed reading your posts about flying, including some of the really technical ones, partly because, like you, I enjoy figuring things out from context, and partly just because you write so well.

And now here you are writing about linguistics! I'm tempted to jump in with all sorts of explanations and examples and clarifications and caveats, but that would be presumptuous. On the other hand, if there are questions you'd like to ask, I'd be very happy to try to answer them. And I don't know what dulltastically unreadable book you got from the library (sadly, that description does not narrow the possibilities down as far as one might hope), but one textbook that I've used that would provide some background on the tree structures in this class is Andrew Carnie's Syntax: A Generative Introduction. The publisher's Web page for the book has sample chapters and a glossary of syntactic terms. (Disclaimer: I do know Andrew Carnie personally—hey, it's a small field—but I have no specific financial interest in getting people to buy his book.)

Also, I think you're entirely right about who/whom: there are very few cases in which collapsing the distinction creates any potential for ambiguity at all, and in your example, the intonation pattern would likely serve to disambiguate the two possible meanings.

david said...

Linguists and philologists write about language as it is, not as they think it should be.

Anyone who went around reciting prescriptive rules about "whom" or "split infinitives" would be taken about as seriously as a creationist at a geology conference.

david said...

Oh, and maybe you'd find Saussure a more-readable starting point than Chomsky; Saussure was, after all, Chomsky's starting point (and Derrida's as well, FWIW).

Aviatrix said...

I've been reading about de Saussure, but not reading directly what he wrote--or books by Chomsky for that matter--because I want a bit of historical perspective, not to be caught up in a controversy that has been long settled.

david said...

That's a fair point. On the other side, it's difficult to understand the arguments of most 20th-century linguists until you know what they're arguing about.

Summaries of Saussure can be glibly dismissive, falling into the "Myth of Progress" (or "Whig Historian") trap and presenting his work as merely one now-obsolete step on a road to current enlightenment.

Aviatrix said...

I've been lucky then. He seems to be cast more as The First Guy Who Had A Clue.

Aviatrix said...

And Q. Pheevr, I sent you e-mail: please go ahead your explanations and examples and clarifications and caveats, in comments or e-mail as you see fit.

Anoynmous said...

I had the interesting experience of learning Transformational Grammar in about fourth grade. It was probably twenty years before I discovered that it was not a commonly taught subject, when I was accused of making things up during a debate on grammar.

Devil In The Drain said...

Anoynmous — you are very lucky. The crap they teach for "grammar" in grade schools and high schools is the same crap they would have taught 100 years ago. Except now we know better.

It's like teaching alchemy for chemistry. Should be a crime.

Q. Pheevr said...

Sorry I didn't follow up earlier—had a bit of a busy day yesterday. So, here are some more detailed comments.

He states that "wh-phrases are in complementary distribution with complementizers in C," and "wh-phrases form a dependency with a gap further down the tree." I divine that wh-phrases are subordinate clauses such as "which I saw" "when the clock strikes" and "whether I like beets."

The wh-phrases are actually not the entire subordinate clauses, but just things like the "which" in "which I saw" and the "when" in "when the clock strikes." The reason it makes sense to talk about them as wh-phrases as opposed to just wh-words is that they also include multi-word phrases such as "which path" in "I wonder which path he said he was going to take." The bit about forming a dependency with a gap is what you describe later on, with "which path" starting out as the object of "take" and then moving to the beginning of the subordinate clause. The movement is intended to account for the fact that a phrase that we hear in one position is interpreted with crucial reference to another part of the sentence; to take a simpler example, the question "Which path did he take?" is essentially asking "Tell me a path x, such that 'He took x' is a true statement." The copy theory of movement is a relatively new theoretical development. It's not terribly obvious from looking at English why we should think that the moved phrase exists in both places when it's only pronounced in one, but we do sometimes encounter what are called 'resumptive pronouns' filling the gaps left by wh-phrases, as in "that thing your aunt gave you which you don't know what it is," where "which" and "it" are (different) pronunciations of two copies of the same phrase. In English, resumptive pronouns are non-standard, and tend to turn up only in contexts where their absence would sound even worse (*that thing your aunt gave you which you don't know what is). There are other languages that use resumptive pronouns more systematically than English does, and which provide stronger evidence for treating movement as copying.

His examples don't include anything like "whom I saw" because apparently linguists are hip to language change and prove it by eschewing moribund subordinate conjunctions.

In this case, a moribund relative pronoun. But, as David said, the point is to come up with a scientific theory of language as it is, rather than as anyone thinks it ought to be. Any given linguist might or might not use whom regularly in their own speech, but should in either case be able to cope with the fact that there are some varieties of English where the distinction between who and whom has been collapsed, and others in which it is maintained.

More to follow....

david said...

On a more relaxed note, the comment about "wh" words reminded me of something unusual about my own use of language.

In Eastern Ontario, I grew up making a phonemic distinction between initial "w" (/w/) and "wh" /hʍ/. The difference blurs in rapid speech, where both revert to /w/ (especially near vowels), but for me, "which" and "witch", "why" and "wye", "where" and "wear", etc. never sounded like the same word, any more than "orphan"/"often" pun in Pirates of Penzance works in North American English.

I understand that initial /hʍ/ is rapidly disappearing, even in the areas where it was once common (e.g. Eastern Ontario, Scottish Lowlands). My wife, from the Toronto area, doesn't use it, and neither do my daughters. Sometimes I notice myself skipping it as well, saying "wear is it" instead of "where is it", but I still hear the distinction very clearly.

Q. Pheevr said...

Specifier is the subject, complement the predicate or possibly a subordinate clause or both, and I think C is short for complement.

This is a bit too specific. The tree diagram that you quote is actually a much more general and abstract schema, which forms the basic structure of any given syntactic phrase. Xmax is the whole phrase, X is the head (e.g., the verb in a verb phrase), and the specifier and the complement are other whole phrases within this one, bearing particular relations to the head. (Each of them, if present at all, is an Xmax in its own right, with its own head and potentially also its own specifier and/or complement.)

Complements in general are phrases that are 'selected' by the head—for example, if a verb phrase is headed by a transitive verb, then it will have to have a complement; if it's headed by an intransitive verb, then it won't. Similarly, in a prepositional phrase, the preposition is the head, and its object is the complement.

In the basic structure of a simple transitive sentence, the subject is the specifier of TP (or IP; I'm not sure which name you'll have seen in the class you went to). The head of TP is a tense marker or an auxiliary verb, and the complement is the verb phrase, within which the verb is the head and the object is the complement.

C is actually short for complementizer. A complementizer is a word like the that in "He said that he was going to leave." That was originally called a complementizer because, in examples like this one, it turns an independent clause ("He was going to leave") into something that can only be an embedded clause—for example, the complement to the verb said. (It's a somewhat misleading term, because a clause with a complementizer doesn't actually have to be a complement; it could be a specifier, too, as in "That he was going to leave was obvious.") Subsequently, we've generalized the category of complementizers, so that we now say that any clause is a CP (complementizer phrase), regardless of whether it has an audible complementizer like that. The properties of C, the head of the CP, determine what type of clause it is: a CP headed by that has to be some sort of embedded clause, and a CP whose head has the feature [+Q] is a question, and so on.

When wh-phrases move, they move to the specifier position in CP. The sentence you quoted, "The specifier and the head of the projection may not be simultaneously phonologically overt," actually isn't about pronouncing only one copy of a moved phrase, but rather about a peculiarity of English. We can say "the book [that I read]," where the bracketed CP has a phonologically overt head (that), and we can say "the book [which I read]," where the bracketed CP has a phonologically overt specifier (which), but we can't say *the book [which that I read].

Aviatrix said...

david: I think they call it "the wine-whine merger." I am capable of pronouncing the /wh/ (which is actually more of a /hw/ if you listen carefully, and is still a /hv/ in Norwegian and I think probably the other northern Germanic languages) but I only habitually do it in old-fashioned words like "whilst," "whence" and "whither." I think my parent said them that way and so I learned the distinction, but lost it in my own speech for common words. I think it may come out when I am angry or excited, though "HWAT ARE YOU DOING!"

Do you pronounce Mary, marry and merry all the same or different? My dad pronounced each one distinctly, and I could hear it no problem, but they come out of my mouth all the same unless I really put effort into it. It's as if there are grooves worn into my mouth for the vowels I use every day and it's really hard to position for different vowels.

Q. Pheevr Ah so it's the "may not" that means "is not allowed to" when I thought it was the "may not" that means "is allowed not to."

In grade nine they made us memorize a list of subordinate conjunctions, I think there were forty-seven of them, and I could have sworn whom was on the list. And besides "subordinate conjunction" sounds sillier than "pronoun."

Q. Pheevr said...

In grade nine they made us memorize a list of subordinate conjunctions, I think there were forty-seven of them, and I could have sworn whom was on the list.

It's entirely possible that it was—there's often a pretty big disconnect between the traditional grammatical terminology taught in schools and the terminology used by linguists. Whom is similar to a subordinating conjunction in that it quite frequently appears at the beginning of a subordinate clause, but from a linguistic perspective it makes sense to think of it as a pronoun that starts out as the object of a verb or a preposition and then (usually) moves to the beginning of the embedded clause (the exceptions being echo questions like "You saw whom?!").

Michael5000 said...

I like "whom." It's essentially optional, and I consider it bracingly recreational. Like big words!