I don't think the time-line has as much to do with it as the scope of the changes do. Again, this is personal opinion; and I'm speaking from the angle of someone who has spent many years doing consulting work on the Microsoft stack, and having to deal with figuring out which ways the 9-to-5'ers are going to break things next.

Version 1.0 of the C# specification clocks in at 405 pages. There was an addendum in 1.2 that I'll leave out, because it was a respecification of 1.0.

2.0 came along and added 117 pages of specification - including generics, partial classes, iterators, and a handful of other things. On pure specification, that's a growth of 25% to the core language. (I know this isn't really scientific, but I think you get my drift).

Then in 3.0, just 2 years later, the spec is back to over 500 pages (just for the language!). This version added a bunch of functional paradigm stuff to the language, lambdas, linq and the like.

And now in 4.0, they're talking about making C# essentially a dynamic language.

Now, all this being said, I really do like C#; but I think that it behooves everyone who likes C# to admit that it's a large and evolving language; the paradigms that Microsoft keeps adding to the language don't really seem to fit from one version to the next.