Argthtjtr

I heard this lyric in a song the other day and it just sounded so wrong that I assumed it must be incorrect grammar, but I can't find any specific prohibition that applies.

That's what it's.

That rolls off your tongue with the grace of a moose in a tutu, but I can't figure out why.

There is clearly no problem with ending other sentences with a contraction. These sound fine.

I thought I could, but I can't.

Stop touching that, it will fall off if you don't.

You say that the sky is green, but it isn't.

Also, it sounds just fine if you remove the contraction:

That's what it is.

So what's up with this construction? Should it be avoided?

This is covered in the Cambridge Grammar of the English Language (CGEL), as it turns out, in Chapter 18, “Inflection Morphology and Related Matters”, section 6, “Phonological reduction and liaison”.

The form ’s, representing either has or is, along with ’m (am), ’re (are), ’ve (have), ’ll (will), and ’d (had or would) are called clitics, and they are a variant of what are known as weak forms of words, which are pronunciations of words like a, have, from, you, etc. (about fifty in total) with a reduced vowel, such as schwa.

In the discussion of weak and strong forms, CGEL points out that there are certain grammatical contexts that require strong forms, and one of those contexts is something called stranding, where the object of a phrase is preposed (moved before the phrase). These are examples they give of stranding requiring strong forms:

a. Who did you give it [to __ ]?

b. We’ll help you if we [can __].

c. They want me to resign, but I don’t intend [to __].

In each of these cases, the word in the brackets has a weak form, but it cannot be used in this context because its object has been stranded. Of course, in written English, there is no difference between weak and strong forms—it’s only a spoken difference—but clitics are distinguished in written English, and the restriction on weak forms also extends to clitics. (There are additional restrictions on clitics, but they are not relevant to this discussion).

So, thus we can say that the second is in the sentence It is what it [is __] cannot be reduced to either a weak form or to a clitic because of the restriction to strong forms in cases of syntactic stranding.

You can end a sentence with a NEGATIVE contraction (Is he here? No he isn't), but not a POSITIVE one (Is he here? Yes he's).

"It's" is only used when something more follows, you wouldn't use that at the end of a sentence. It sounds like there is something missing if you just say, "it's" or end a sentence with it.

Probably what makes the it's man so funny.

My guess would be that your ear is tripping over the homophone -

its = belonging to it

it's = it is

That's why, when you try the same thing with the other non-homophone contractions, it sounds fine. When you try it with another homophone contraction, look at what happens:

If you want to know if they're crazy, well, they're.

If you want to know if they're crazy, well, they are.

(Typing like Yoda, I am)

Your ear may not recognize whether you're hearing it's or its, they're or their (or there...). So the ear expects another noun. I think that's why it sounds fine with other contractions that aren't homophones. "Can't," "don't," and "isn't" clearly sound like the words they are. "It's" and "they're" are the sources of endless confusion.

There's nothing wrong with the grammar (in terms of 'prescriptive grammar').

But it "rolls off the tongue" poorly because of its scansion. In particular, people tend to fall into "rhythmic schemes", where short phrases have a certain pattern of stresses. They tend to avoid phrases that don't fit whatever scheme they are currently in. They tend to construct phrases that fit the scheme.

I don't have a reference for this, but it does seem that these patterns don't usually end with stresses when they have an odd number of syllables. So it sounds like there is a word missing.

http://en.wikipedia.org/wiki/Prosody_(linguistics)

I think we're so preconditioned to expecting a word/phrase after a contraction ("it's" in particular), that it strikes the listener/reader as a rather unexpected/out-of-place stop.

I'm not sure if there's technically anything grammatically wrong with that sort of usage. (Obviously contractions should never be used in formal writing.) Still, I think we all agree it sounds very unnatural.