It is extremely common to find non-linguists espousing a kind of naive semanticism about syntax: they think everything about syntax springs from meaning, i.e., that where you can use a certain word just follows in an obvious way from what that word means. There are plenty of cases that I take to be clear counterexamples; for example, one pointed out by Richard Hudson of University College London in the 1970s is that likely seems to mean exactly what probable means, and in many contexts they are interchangeable, but I'm likely to be late is grammatical while *I'm probable to be late is not. But the layperson is often not easy to convince: they just assert that the latter sounds bad because "it doesn't mean anything" (which of course was supposed to be the datum, not the explanation!). It may be that you simply agree with them.
When reflecting on this I have often thought about the case of verbs expressing sharp or erratic movements. Deirdre Wilson (also of UCL) pointed out to me a long time ago that shaking and quaking are basically exactly the same thing — if it quaked then it shook, and conversely — but one is transitive (takes a direct object) and the other is not (Try shaking it is grammatical but *Try quaking it isn't). Below is a list of fifty verbs of shaking, rattling, rolling, juddering, twitching, jerking, oscillating, lurching, and similar movements. The question I put before you is simply whether you could truly have told in advance from knowing the semantic facts — what it means for something to shake, rattle, roll, judder, twitch, etc. — whether each of these verbs would be transitive or not. A plus sign in the intransitive column for shake means I think that It shook is clearly grammatical, and a plus sign in the transitive column means I think that Someone shook it is clearly grammatical. The intransitive mark for jog is for The road jogs left at the end of that block, and the transitive one is for You jogged my elbow. Flutter is shown as transitive because you can flutter your eyelashes. And so on. I'm sure I have missed some cases; the tabulation below is based on a quick rough-and-ready assessment, not detailed corpus searching (that is perhaps a task for the future). But see what you think. Could you predict the verbs that can take a direct object, just from the sense? Because if not, naive semanticism is wrong, and linguists' almost universal belief that meaning does not predict all syntactic properties is justified.
| Intransitive | Transitive | |
| agitate | + | |
| bob | + | |
| bounce | + | + |
| brandish | + | |
| bump | + | + |
| convulse | + | + |
| flap | + | + |
| flicker | + | |
| fluctuate | + | |
| flutter | + | + |
| jar | + | |
| jerk | + | + |
| jiggle | + | + |
| jog | + | + |
| joggle | + | + |
| jolt | + | + |
| jostle | + | |
| judder | + | |
| jump | + | + |
| leap | + | |
| lurch | + | |
| oscillate | + | + |
| palpitate | + | + |
| pulse | + | |
| quake | + | |
| quaver | + | |
| quiver | + | |
| rattle | + | + |
| roll | + | + |
| shake | + | + |
| shimmy | + | |
| shiver | + | |
| shudder | + | |
| spring | + | |
| swing | + | + |
| teeter | + | |
| throb | + | |
| tremble | + | |
| twitch | + | + |
| veer | + | |
| vibrate | + | + |
| wag | + | + |
| waggle | + | + |
| wander | + | |
| wave | + | + |
| waver | + | |
| wiggle | + | + |
| wobble | + | + |
| wriggle | + | |
| zigzag | + |
[Update for specialists: I am not unaware of semantic distinctions like that between internal and external causation, which do have some influence on the possibility of transitive uses (see Beth Levin and Malka Rappaport Hovav, Unaccusativity (MIT Press, 1995), pp. 90ff). I am not suggesting that my own position is (to use a term that Andew Koontz-Garboden mischievously suggested) naive syntacticism. I'm just saying this is going to be some complex function of both syntax and semantics.]
Posted by Geoffrey K. Pullum at January 29, 2008 07:31 AM