Only the third one really makes sense to me.

arkadio:I am recanting. "La fermita pordo," seems unambiguously to mean "The was-closed-by someone-or-something door," and not "The in-a-state-of-closedness door."

In theory you might say the presence of the passive participle in fermita unambiguously shows that the door was closed by someone. In practice however, these forms are used to show general state just as much as a result of a passive action. fermita can mean "has been closed" or just "closed" (or in a state of closeness, however you want to say it).

In this case, theoretically it would be possible to use ferma to show a general state of closedness, since adjectives essentially mean "somehow related to the root". However in all likelihood since ferm' is a verb ferma would be taken to mean characterised by the actual act of closing, rather than being closed. fermita then is the form used, which is ambiguous as to generally being closed or having actually been closed.

To me this doesn't represent too much of a problem though. In a context where fermita clearly functions as an adjective showing state (ie the pordo estas fermita) I'd say the need for a distinction between being shut and having been shut by a person is pretty minimal. Most things that are shut actually got that way by being shut, so it's no big deal to show that even if it's more information than you intended to convey. English works the same way (closed is a passive perfect form) and it doesn't cause us a problem.

arkadio:(1) "turninta," (because I turned something -- min),
(2) "turnita," (because something -- mi -- turned me), and
(3) "turniĝinta," (because I turned)?

Technically all three are grammatically correct. But the third is the only one I would use in practice, since 1) might be mistakenly read as you turned someone else, and in the same way 2) might imply someone turned you (other than yourself).

Only the third one really makes sense to me.

Technically all three are grammatically correct. But the third is the only one I would use in practice, since 1) might be mistakenly read as you turned someone else, and in the same way 2) might imply someone turned you (other than yourself).

Yeah, the third one is the least open to misinterpretation, though it doesn't preclude the possibility that something/someone else turned me: "Kiam me turnis la globeton, tiam la globeto turniĝis."

To me this doesn't represent too much of a problem though. In a context where fermita clearly functions as an adjective showing state (ie the pordo estas fermita) I'd say the need for a distinction between being shut and having been shut by a person is pretty minimal. Most things that are shut actually got that way by being shut, so it's no big deal to show that even if it's more information than you intended to convey. English works the same way (closed is a passive perfect form) and it doesn't cause us a problem.

You are absolutely right. For practical purposes, there is no problem. I have some interest in computational linguistics, and that is why I gnaw away at these formal semantic issues.

P.S.

Most things that are shut actually got that way by being shut, so it's no big deal to show that even if it's more information than you intended to convey.

The problem wasn't so much over-informing as it was misinforming. Here's the example that started me thinking about this. A building has two doors, one automatic, the other, manual. Suppose I want to say that both doors are closed. I could say

(1) La pordoj estas fermitaj,

or

(2) La pordoj estas fermiĝintaj.

But neither statement is correct if I insist that "fermita" mean "closed by someone." I do concede that there is no practical problem, just as there is none in English.

This may have already been answered, or may be a stupid question, but isn't -iĝinta redundant because the n in inta already shows the passiveness which iĝ shows?

Nope, int is active, it is passive. Additionally, iĝ dosn't show passivity, it shows a change to a state described by the root. Fermiĝinta therefore means "has become closed."