Syntax I

We looked at some slides today and some chalk-drawn tree structures that featured nodes like XP, X’, and X. So, what’s the relation between XP and VP? Or X’ and V’? And what is an XP anyway?

The answer is that XP can be thought of as “some phrase or other”, an arbitrary phrase. The idea is that it doesn’t matter what the category is, what we say about XP is supposed to hold of all categories. A longer version of this might go:

Take any element from the lexicon. Assume (as is reasonable) that anything we can pick from the lexicon has a category, something like N, V, P, D, C, A, or T. We don’t know what you picked (that is, it doesn’t matter what you picked), but suppose that we wish to express that, when Merge combines this head you picked with another one, and the head you picked projects its features, then we write the label of the unit with a “P” after the name of the category. We could say this by expressing the convention with blanks to fill in with the category: “When a lexical item of category __ is Merged with another lexical item and the first one projects, the combined unit is a __P.” We then say that this is true whenever you pick a lexical item and fill in its category into the two blanks in the statement of the convention.

That’s pretty much exactly what is meant by “XP”. “X” here is not really a category, it stands for whatever category we’re talking about. By stating generalizations about X and XP, we’re basically stating generalizations about N and NP, V and VP, A and AP, D and DP, C and CP, T and TP, etc., all at once. As well as for any categories that perhaps exist but have not yet been discovered, something that just listing the generalizations (one per known category) cannot achieve.

Same goes for YP and ZP. There’s nothing in our lexicon that actually has X, Y, W, or Z as its category, but we use those letters as variables (like “blanks”) to stand in for the category of whatever it is we’re talking about. This allows us to say, for example, that the category feature of the maximal projection, intermediate projection, and minimal projection is the same, for any category you care to pick.

Adger has a very small discussion of this on p. 115, consisting essentially of the line “I have used the variables X, Z, and W to stand for arbitrary labels.” But this is what he means.

If things are easier to grasp through a math analogy (as is often the case, no doubt, in your experience), consider this: The use of X above is (in a way) quite similar to the x of algebra. That is, the x in 3x + 2x = 10. The x there stands for a number, and we’re (probably) interested in trying out whatever numbers there are in that position to see which ones meet the condition that when multiplied by 3 then added to what you get when multiplying it by 2, the result is 10. Whatever number you try, you have to put it in for x consistently, since x simply refers to “the number you’re trying.”

Helpful? I hope this clarifies things a bit if you weren’t clear what the X was doing in X-bar theory.

Concerning the question from class about where Malagasy is spoken, the answer is: Madagascar (down around Mozambique and Tanzania). There’s lots of interesting information about Malagasy on ethnologue.com.

Incidentally, if you’re ever wonder about any other language, its affiliations, where it’s spoken, the number of speakers, etc,. ethnologue.com is a great place to start.

p. 105, the penultimate line says “teminal” rather than “terminal”. Probably not a very confusing typo.

p. 127, the penultimate line refers to “a new constituent”. Its label should be “?P” not “?”. Same on the second line of p. 128.

It was pointed out to me yesterday that the bracketed structure in (9) does not quite correspond to the tree notation in (10) [pp. 64-65]. See why?

The bracketed string in (9) does not mark “might have been cracked open” as a constituent, but yet that string is indicated as a constituent in the tree in (10). There should have been a left bracket before “might” and another right bracket at the end in order to make (9) correspond to (10).

So, suppose you’ve run your constituency tests and two of your test sentences turn out ungrammatical, and the other two turn out kind of marginal, what you might mark as “?”. So, have you found a constituent or not?

In cases like this, it’s a little bit tricky, true. It depends quite a bit on how high your standards are for calling something acceptable, too. Myself, I generally treat “?” as more “grammatical” than “not”. So, I would be inclined to say this is a constituent that you’ve found.

In such a case, I think what I would advise doing is make a note of it and proceed, presuming the string in question is a constituent for the time being, and see how the rest of the problem goes under the assumption that you’ve found a constituent there. If things start to look funny in the rest of the problem, compare it to how they might look if you didn’t assume the string in question is a constituent, and if it looks better in the latter case, reverse your decision and say it isn’t.

It’s a little bit “squishy,” it’s true. There are a lot of interacting factors involved — one of those factors is described in part 6 (namely, some tests systematically fail if you test a string inside a noun phrase). Another factor is simply whether there is a word in English that can stand in for the string — just because you can’t think of a word (and perhaps there isn’t one at all) that can fit there doesn’t in and of itself mean that the string isn’t a constituent. All it means is that it’s harder to find evidence that it is. Yet another factor is simply lexical frequency — the weirder or more uncommon a word is that you need to use, the worse the test sentence will sound, as in the case of thence and whence, for example.

Anyway, do your best, and at least know that I understand it’s squishy territory.

I was asked today where specifically the Simpsons clip came from. At last, an entry that really fits in the Trivia category. The clips came from an episode called “Miracle on Evergreen Terrace“, season 9 episode 10. More than you will ever need to know about the episode is gleanable from here.

This episode also featured Bart’s use of the word “craptacular” (a nice blend, or in the terminology the Wikipedia suggests, a “frankenword.”), one of many lexical innovations spawned by this show.

The Simpsons seems to written by a relatively Linguistics-savvy bunch; examples one can use to make points in Linguistics seem to come up there more than one might expect by pure chance.

In HW2 #2, you’re asked to provide a couple of bracketed strings and their corresponding tree representations. Be sure they match!

The way the homework is set up, you test some constituents and then put brackets around those strings that are constituents. This is your bracketed string. There might be other constituents, but you shouldn’t really put brackets in unless you have evidence of a constituent contained within the brackets. Then, you translate the bracketed string into a tree representation.

Moving from brackets to trees is reasonably straightforward. You generally want to work your way from the innermost groupings to the outermost ones. Suppose you have this:

[ [Horatio’s team] [caught [the smugglers]] ]

This means the smugglers is a constituent and there should be a node in the tree that dominates the smugglers. So, you might at this point draw a triangle with the smugglers under it. Or, you could draw a node with a branch down to the and another down to smugglers.

     .                    .
    / \          or      / \
   /   \                /___\
 the   smugglers    the smugglers

Then, either mentally or actually, connect the brackets around the smugglers by crossing it out, like so:

[ [Horatio’s team] [caught XXXXXX] ]

The thing you just drew, that triangle with the smugglers under it, is the thing represented by the “XXXXXX”. And you’ll notice that “XXXXXX” and “caught” form a constituent (they are contained within the next set of brackets). That means there’s a node that has a branch down to each of them, like so:

        .
       / \
 caught   .
         / \
        /   \
      the   smugglers

Then, repeat (cross out the constituent you just made, observe what higher constituent it is a part of):

[ [Horatio’s team] XXXXXX ]

So, now the thing we just built is the sister of a different constituent Horatio’s team and before moving up any further, we should draw that constituent, which we can draw alongside what we have so far:

         .              .
        / \            / \
Horatio's team   caught   .
                         / \
                        /   \
                      the   smugglers

And cross out the constituent we just built, which (for distinguishability) I’ll call “YYYYYY”:

[ YYYYYY XXXXXX ]

The brackets indicate that the two things we recently built themselves form a constituent, yielding:

                .
               / \
             /     \
           /         \
         .             .
        / \           / \
Horatio's team  caught   .
                        / \
                       /   \
                     the   smugglers

And we’re done. There’s not much to it, really, once you see how the brackets and the trees correspond. Just be sure that you do it in that order: determine the bracketing based on the constituency tests, then draw trees representing the constituents you have evidence for. (That is, don’t do the first two steps and then leave them behind, simply drawing trees like you think they should be based on what you remember from the intro class; limit yourself to the things we’ve talked about in this class, since things will end up being different in places from what was done in the intro class).

It was just now called to my attention that my instructions for problem 1 were such that you don’t need to use C, T, or D as you label the categories — none of those categories of words were underlined. C’est la vie. If you wish to label the non-underlined words, you may, but you need not.

In homework 2, you’re asked to produce a couple of trees. I’ve given a sample of what I had in mind along with problem #3. Notice a couple of things about the example tree: i) there are no node labels, and ii) nodes can have more than two daughter nodes. These are things that will change in the future, but for now it is just as well not to try to fill in labels on the nodes, given that we haven’t really talked about how they are determined. Your intuition will probably generally not lead you astray if you try to label them anyway, but labels are not needed.

Don’t let the fact that question #3 on the homework seems easy shake your confidence. It is. It’s basically the same thing you did in part 4 of question #2 with another (and even simpler) sentence.