I am not an English native speaker, and I have a question about the following sentences:
I believe that Mary will arrive tomorrow. (grammatical)
*I believe Mary to arrive tomorrow. (ungrammatical)
Why can't we say "I believe Mary to arrive tomorrow"?
I believed (that) he was honest. = I believed him (to be) honest.
My dictionary says that both sentences are grammatically correct, so I am really confused.
When can I use "S (subject) V (verb) O (object) OC (objective complement)" structure with the verb believe?
Best Answer
Well, I guess this won't get the bounty since you want someone to explain why it's ungrammatical, but
it is grammatical.
I mean, we don't usually use this structure for the verb "arrive", but we use it for "be", "have", "possess", "mean", "hold", "understand"... (See also here.)
I can immediately parse "I believe her to arrive tomorrow" as a formal, fairly stuffy answer to the question "When will she get here?" meaning "I believe [it is the case that] she'll arrive tomorrow".
It's just not terribly common.
The archaic feeling of this way of talking that Resber picked up on comes from the way it mimics Latin grammar instead of using the more natural and versatile "that clause". Using English this way (along with not splitting infinitives, &c.) was much more common back when everyone publishing a book had sat through years of Latin classes with a tutor or at their grammar school.
Going back to that list of infinitives that more commonly show up with "believe", it seems to be the case that English speakers tend to use that structure with stative verbs rather than active ones. It's common in general to use "that clauses" for verbs reporting things (say, tell, &c.) or describing mental processes (think, know, believe, &c.). Expressing it that way gives you more freedom to mark tense and mood or add adverbs and additional clauses. That extra nuance is more important when discussing actions.
It's not common at all.
Specifically, this paper notes research that shows that English almost exclusively uses "believe [obj.] [inf.]" for stative verbs or perfective ones (i.e., "I believe her to have arrived yesterday"). There are exceptions and even their example of a "mistake"
is perfectly clear, if a bit clunky. Emending it to "to have supported" certainly sounds more natural but has a different meaning. (In that case, their support or even the entire war is already finished.)
Still, to the extent that native speakers tell you such sentences are "wrong" or "too awkward", that's the mechanism underneath why. If you want to call it a "rule" and the exceptions "ungrammatical", you can but it's not really anything they're conscious of doing and is just a side effect of wanting to have more freedom to adjust or describe active verbs, especially their tense.