First, with a plunger mute. The adults in the Peanuts cartoons were "voiced" by a trombone with a plunger mute, which gives the impression of a voice without necessarily sounding like it's saying anything in particular. (http://en.wikipedia.org/wiki/Peanuts#Television_and_film_pro...)
Second, it's actually possible to make it sound slightly like specific words by altering the shape and size of your vocal cavity while playing a note. When this is combined with slight glissandos, it can be made to sound a lot like speech. In fact, there is a piece for solo trombone called "General Speech" which uses this technique to recreate General MacArthur's farewell speech at West Point on the trombone. You can read about it and hear it at http://artofthestates.org/cgi-bin/piece.pl?pid=11
To be fair, neither of these techniques produces anything you'd be able to understand in a conversation, but the point is that a sufficiently skilled trombonist could definitely say something that could be recognized as "I Owe You" on the instrument.
I agree with everything else you said, but your own words seem to bring us to the conclusion that pitch has nothing at all to do with distinguishing vowels (in languages in which pitch alone cannot be used to distinguish words).
Consider a couple of examples:
1. A trombone can't say "I Owe You", despite the fact that it can produce any dominant frequency, and that phrase contains only vowels.
2. Voice "Ahhhhhhhhhhhh", going from a high to a low pitch (I imagine the doppler effect of a cartoon screaming as he rides by on a cart), and then voice "Ohhhhhhhhhhhhhh" in the same fashion. At every point, they sound different. You can't take a clip from an arbitrary pitch range of an Ah and convince someone its an Oh.
3. Shape your lips like you would for oo as in Moon, Noon, or Spoon, and try make the ea vowel as in Squeak just by changing your pitch. It won't happen. You'll get the German ö, but not the English ea.