Not only wrong, but not logical either I�m afraid, IM.
There are two problems with your analysis: firstly that mathematical convention is being confused with measurement; secondly that of the temperature scales being arbitrary and not absolute (which has been well aired already).
To resolve the first problem consider this: I measure my wealth and find that I owe my friend �1 (that is, I have a wealth of �minus �1�). I resolve to become twice as rich by tomorrow. So, using the same convention as you used with the temperature example that has been the subject of all this, to become twice as rich I should owe my friend �2 (because minus one times two is minus two).
As I said, the second problem has already been well discussed, but to take your heated water example simply think of this: 20 degrees C is 68 degrees F. So, to make your water �twice as hot� do you raise it to 40 degrees C (as you suggest), or 136 degrees F (which is 58 degrees C, or �2.9 times as hot�)? Of course, any other arbitrary scale you care to use will produce similarly confusing results.
Does that help at all?