Elibol, I really don't know what I did to make you sick. I certainly had no malicious intent in posting this puzzle. I don't see that there is anything wrong with enjoying a debate. If we all agreed all the time, what would be the point in talking?
And I do not agree that my version of the puzzle is flawed, either. In fact, some of your own objections highlight the fact that the U.Mich. version is flawed. That is, blue eyes are obviously a permanent feature (anybody with blue eyes has always had blue eyes), so your point about what information was provided by the visitor is well taken. Also, the U.Mich. version says they will commit suicide "by sunset", thus allowing the possibility of someone committing suicide immediately at prayer meeting, which can screw up the sequentiality of it. My version specifies that suicides only occur in private in the evening, so the information about the suicides is only available the next day. For these reasons, I think that the U.Mich. version is ambiguous.
My version, on the other hand, allows the interpretation that the marks first appeared on "day 1" when the visitor arrived. Yes, you are correct in observing that if 6 out of 7 monks were marked that day, they would all see marks on either 5 or 6 of their fellows. But the puzzle doesn't tell you how many there are, or how many are marked. So the visitor is certainly necessary for the benefit of the puzzle. IF there were only two or three monks, and IF only one monk (out of any total) was marked, the visitor is necessary for the monks as well. As far as the question of what knowledge is provided by the visitor in the case where there are 7 monks, all marked, I haven't heard or thought about that question before now, and I'll admit I don't know the answer. But if the number of monks is unknown, and the number of marked monks is unknown, and you don't know if any of the monks know if any of the monks are marked, there is no puzzle.
Anyway, while the last few posts were going up, I've been busily working on this explanation which I see is now redundant, but since I spent all this time on it, I'll throw it up here anyway. Maybe it will help someone else.
I still believe that a total of 7 monks, all marked beginning at the moment of the visitor's announcement, is the correct answer. I probably can't explain it any better than the others have done, but I'll try. The key to it is the dimensionality. Monk A is trying to deduce what monk B (and every other monk) is thinking, knowing that monk B is trying to deduce what monk C (and every other monk) is thinking, and so on. So it becomes one of those picture in a picture in a picture in a ... situations. Monk A thinks that Monk B thinks that Monk C thinks that ... etc. I can't see 4 dimensions, let alone 7.
1 of course is trivial. If there is only 1 monk, & the visitor says "at least 1 is marked", the monk commits suicide that evening & the game is over.
2 is not much harder. Monk A tells himself, "only 1 of us is marked, so if I am unmarked, B will see that, deduce that he IS marked, and he will commit suicide on evening 1. When Monk B appears at prayers on day 2, Monk A knows that B sees the mark on A, so A commits suicide on evening 2. You can see that my choice of who is A and who is B is completely arbitrary, so swap the letters, & now BOTH commit suicide on evening 2. Note that this is true only if IN FACT BOTH ARE MARKED. If only Monk B is marked, he sees that A is unmarked and commits suicide on evening 1. This confirms in A's mind that he is unmarked so he does not commit suicide. Period. Either way, the condition that "all of the monks commit suicide on evening 7" does not occur, so 2 monks is not a valid answer.
Let's look at 3.
IF there are 3 and all 3 are marked, on day 1, Monk A sees that his two brethren are marked. Monk A wants to believe that he is unmarked. He knows that Monk B wants to believe that HE is unmarked. Monk A thinks, "If I am unmarked and Monk B believes that he is unmarked, he (B) must assume that C is the only one marked. He (B) would then expect C to commit suicide on evening 1, and therefore he (B) would expect that C would not show up at prayers on day 2. Since C was still alive on day 2 (A reasons), B would have to conclude (on day 2) that C does NOT see two unmarked monks. So, if C is still alive on day 2, he must see at least 1 marked monk. If I (A) am unmarked, B sees that, and therefore he will know on day 2 that HE (B) is that marked monk, so he will commit suicide on evening 2. If B shows up at prayers on day 3, then he must see that I (A) am marked, so on day 3 I commit suicide.
If all of the monks are marked, they all follow the same reasoning, and all commit suicide on evening 3. If only 1 was marked, he would commit suicide on evening 1, and the other two would know that they are unmarked. If 2 were marked, following the same reasoning they would both commit suicide on evening 2, saving the 3rd.
If there are 4...
I don't have time to take this to 4. I'm not sure that it can be described adequately verbally. And even if I did, you'd say "prove it for 5". I believe the inductive proof already described on the U. Mich. site is valid for my scenario, even more than it is for theirs.
If you don't agree, I guess none of us is going to convince you. I hope you - all of you - enjoyed puzzling over it anyway.