Hmm ... sorry Billy but I am not so sure about any of this.
The problem with 2...e6 and 2...c6 is precisely that they commit Black to a QGD/Slav set-up, before White has committed himself to the Queen's Gambit, and so cut down on his options if White chooses to refrain from playing c2-c4.
White can, for example, play the Colle with 3.e3, when Black might want to develop his bishop outside the pawn chain and/or finachetto his king's bishop with ...g7-g6, neither of which is really an option after 2...e6. And after 2...c6, White can go into a London system with 3.Bf4, when Black has effectively ruled out an early strike back with ...c5 (if he plays this now, it will cost him a clear tempo). These systems are both very popular at club level, so I simply don't agree that White will "almost definitely" play c2-c4 at some point!
As for 2...a6, it is an interesting idea, but again I am not sure what it contributes against 3.Bf4 or 3.e3, while if White does play 3.c4, you are simply left with the options of transposing to a Chebanenko Slav with 3...c6 or a QGA with 3...dxc4. Hardly 'unkown territory' ...
I agree that 2 ... c6 is a little committal, but I'm inclined to think it's still very solid and against the Colle, it's not like the extra tempo really jeopardizes Black's play. A recent correspondence game of mine went:
1 Nf3 d5 2 d4 c6 3 c3 Bf5 4 Bf4 Qb6 5 Qc1?! Nd7 6 e3 Ngf6 7 Nbd2 e6 8 Be2 Rc8 9 Nh4 Bg6 10 Nxg6 hxg6 11 h3 c5 and Black went on to win after White hung a piece at move 20. In retrospect, Rc8 wasn't necessary and h6 probably was, but Black is solid and is bound to enjoy equality...