The professor has a piece of hidden information $C$. Alice is given an envelope containing information $A$, Bob is given information $B$. It is possible to deduce $C$ from $A$ and $B$. The conversation goes like this:
Alice: I don't know $C$.
Bob: I don't know $C$.
Alice: I don't know $C$.
Bob: Oh, now I know $C$!
Bob then correctly announced $C$. "We already know this game." he said.
"I've only just gotten started." said the professor. "Give me a non-negative integer. Any non-negative integer."
Alice said, "Alright, how about my birthday? It's $n$."
The professor thought for a moment, then said, "Alright, I'm thinking of a new piece of information $C_n$.". He then handed new envelopes $A_n$ and $B_n$ to Alice and Bob.
Alice: I don't know $C_n$.
Bob: I don't know $C_n$ either...
... after exactly $n$ iterations ...
Alice (her year of birth is an odd number): Finally! I know $C_n$.
"So you could have done this for any non-negative integer $n$?" asked Bob.
"Any non-negative integer." said the professor. He handed them a new pair of envelopes. "These should keep you going for exactly $1001$ iterations."
"How many!?" Bob said, "But it's almost lunch time!"
"Wait!" exclaimed Alice, "We don't actually need to play. We know that after $1001$ iterations, I'll have enough information to deduce the professor's secret - since it'll be my turn then. But the only information I'll have gained at that point is that we played the game for $1001$ iterations without yet figuring it out - but the professor's already told us that! So I should be able to deduce the information just by knowing the number of iterations..."
She stared intensely at her card for a moment, and then triumphantly announced the solution.
The professor smiled and wrote down two new cards for the two players. "But finite numbers are so provincial. Let's play another round, shall we?"
Just as the pair were about the begin, the professor interrupted: "I should probably tell you, though... no matter how many iterations you play through, you will never figure this one out."
"So you're saying $n$ is infinity?" said Bob.
"Yes, sort of - it's actually the ordinal number $\omega$, which works similar to an 'ordered infinity'. You can do a lot of the same things with it that you can with the numbers we're all familiar with. The only major issue with throwing in $\omega$ is that addition and multiplication pay attention to the order now."
"I have it!" said Alice, "There's only one possible value of $C$ compatible with the information on my card and which would keep us playing forever."
She then announced the correct value.
"Impressive." said Bob. "Got one I can solve?"
The professor handed them fresh cards. "Again, this one can't be figured out after any finite number of iterations."
Alice frowned. "I still can't deduce it."
"Me neither." said Bob.
"Indeed." said the professor, "In fact, even with the information I've given you, you still will never deduce the solution."
"Well... what about now?" said Bob.
"Now you can." the professor said. "After 24 iterations, that is."
"I see." said Bob. "So the solution is..."
"Correct." said the professor. "In this case, we had $n=(\omega\times 2)+ 24$, as it were. And in the same way, I can come up with pieces of information $A, B$ and $C$ corresponding to any number of iterations of the form $\omega n+m$, where $n$ and $m$ are positive integers."
"Wow." said Alice and Bob.
"And even $\omega^2$!" the professor said proudly. "That is, no finite number of iterations of my telling you that you still can never deduce $C$ would be enough information for you to solve the puzzle, but the previous part of this sentence would. We could go farther, too! $\omega^3$, $\omega^4$, even $\omega^\omega$!"
How can we come up with a puzzle like that of the professor? Standard ways are for $C$ to be a pair of numbers, $A$ the sum and $B$ the product, but I don't know this allows us to obtain any number of iterations, not even any finite number.
Can anyone come up with $A, B$ and $C$ corresponding to some of the professor's infinite numbers of iterations? What about a standard scheme to produce $A, B$ and $C$ corresponding to any $n$?