I agree. First person accounts have to be treated differently from third person accounts.
If I was to do something, I know whether I did it or not.
If someone else was to do something or something was to happen, I don't
have any privileged first-person knowledge about it to rely on.
Take this third-person remark as an example:
George was to be at the train station at 3 o'clock.
Supposing that I don't know if he made it there or not, there is no
implication that he made it or did not make it there by 3.
Everything about his arrival is speculation.
Someone asks
Where is George? and I answer
He was to be at the train station at 3. I say nothing here that suggests that he did not make it to the station, nor anything that suggests that he
did. I only give a vague indication of what he's up to without commiting to the idea that he has fulfilled his obligation.
On the other hand, and similar to the first-person situation, if I know that he did
not
make it to the station on time (by 3), then, at least in my own mind,
the implication is that he has not fulfilled his obligation, and my
remark includes this judgment about his behavior. Nevertheless,
if I know that he
did make it to the station on time, then my
remark is infelicitous: There is not much point in mentioning an
obligation which I know has already been fulfilled. I should be
saying,
George was at the train station at 3 o'clock (Cf. Clive's example).
In short, some of the subtleties of such modal and semi-modal
constructions depend on who knows what and when they know
it!
![Smile [:)]](/emoticons/emotion-1.gif)
Sorry, but, unfortunately, modality is dang
complicated!
CJ