I do not want to prove it, because I'm lazy, but I know it's trivially provable that it's possible to create a Turing machine in Excel in many ways. You could take any row or column to be the tape and store state and table in other cells. Then add a bunch of conditionals and LOOKUPs you get your Turing machine.
It would probably be easier using VBA, but it's an interesting idea. (Not that it would get you anywhere, though.)
I'm not so sure. Googling around on this brings up a lot of statements to the effect that Excel is not Turing-complete. (I mean, of course, excluding VBA.) I haven't found a proof either way, though surely one is out there.
This paper claims to prove that the authors' spreadsheet is Turing-complete in contrast to Excel:
Really? I'd find it hard to believe that it isn't Turing complete. The only reason I can think of would be limits on nested functions or otherwise, but I dunno, I've been googling around myself and some say it isn't.
Well, next thing to do: implement a Turing machine in Excel. Or Google Docs. I think I'm going for the latter.
Edit: Spreadsheets have no state. This is hard. If only I could delay evaluation :P
It is possible, not specially hard, I implemented a Turing machine and a one-dimensional cellular automaton. You can program just about anything in a spreadsheet. I will later study the inner properties of the spreadsheet as a data structure and a programming paradigm.