QueenBee: Yes, I AM serious -- but I may be partially (or completely!) wrong in trying to interpret Article 250.122(B), I am willing to admit. No harm in being wrong, at least I'm trying to uncover its intended meaning, based on the text I'm relying on -- which is mainly from "NEC 2011 Handbook". I'm a little reluctant to quote at length from its commentary passages because of copyright (unlike its NEC passages which, so I have read, because they become law can legally be quoted after that occurs).
I am given the impression that "increased in size" can (at least sometimes) have the definition I suggested based on the "Calculation Example" given where an initial installation (I assume since they wrote "installed" -- past tense) using 250-kcmil copper conductors and a 4 AWG copper EGC has its conductors increased to 350 kcmil (it doesn't say why or when) -- with the "correct code" answer coming from 250.122(B) of 2 AWG copper EGC instead of a Table 250.122 look-up answer of 3 AWG copper.
On the other hand "increased in size" may also have meaning at initial installation time, judging from the following passage: "Where the ungrounded circuit conductors are increased in size to compensate for voltage drop or for any other reason related to proper circuit operation, 250.122(B) requires that the equipment grounding conductors be increased proportionately. Increases in ungrounded circuit conductor size for the purpose of ampacity adjustment, correction, or both are not required to be considered in applying the provisions of 250.122(B).".
Getting back to the OP, my interpretation of the entire text is that using 6/2 AWG on a 40a circuit breaker is NOT a code violation of 250.122(B) -- because it does not apply. Table 250.122 applies instead, since using 6 AWG instead of 8 AWG to provide for possible future EVSE enhancement does not relate to "voltage drop or for any other reason related to proper circuit operation". I do not know if the "NEC 2008 Handbook" (or even earlier editions) have essentially the same explanatory commentary as the 2011 edition. If it does, it is hard for me to see why (even though it doesn't "constitute Formal Interpretations of the NFPA") it wouldn't have been pointed out to the authors as being incorrect -- and corrected in the 2011 edition -- if it were wrong.