Friday, June 18, 2010

Think again

Going back to my post of a few days ago about think and its allowed complements, I was reminded of this post from 2007 about the complementation of rethink. It's interesting to note that, where think doesn't allow open interrogative content clauses, rethink does.

  • I thought about what I should do.
  • *I rethought about what I should do.
  • *I thought what I should do.
  • I rethought what I should do.
Why are complements so arbitrary?

3 comments:

Anonymous said...

OK, here are my thoughts.

To think about something is a process. If I think about what I should do, then within that process of thinking, I may consider multiple conclusions.

To rethink doesn't simply mean to think for a second time. It means to come to a different conclusion, and thus the emphasis is on the conclusion rather than the process leading up to it. (Nobody says: "I rethought it, and decided I was right the first time.")

The word "about", in the sense that's pertinent here, fits "think" because of the focus on process, and doesn't fit "rethink" because of the focus on conclusion.

However, just to muddy the waters, consider: "What did you conclude [or what conclusion did you come to] about what you should do?". This sentence involves a subtly different sense of "about".

Ran said...

This came up on Language Log a while back: http://languagelog.ldc.upenn.edu/nll/?p=1138. Commenters mention other similar examples, with "relive" and "rehash" being the closest IMHO.

Q Higuchi said...

There seems to be something a little larger than just individual verbs that determines what complements are allowed/preferred.

For instance, negation plays a part.

a. I doubt whether he will come
b. I don't doubt that he will come

Here, it is not really a question of the verb _doubt_ taking two different types of complementation. It is more of "doubt" taking the _whether_ complement, and "not doubt" taking the _that_ complement.

I believe your think/rethink examples are along similar lines.

Does this make our lives easier? Maybe just a bit. We can take a step back from individual verbs, invoking some more abstract concept like semantic effect, thereby gaining some plausible explanation. Not so arbitrary after all. Yay.

But, in order to describe the grammar, we still have to say what specific semantic effect happens with which verbs when combined with what. Ouch. Back to the beginning, possibly even worse.

But this is fun - it is one of those areas where we feel the itch. We know the general tendency (or even principle) when we see it, but when we try to put it in words, we get further away from it. Jolly frustration of poetry.