So then your argument is purely academic, since you believe there's only a minority supporting a fork anyway. Correct?
The answer here is both yes and no. Currently I believe that only a minority supports the fork. I'm not saying that this can't change though, nor am I saying how many people are actually in support of a block size increase (where supporting a block size increase doesn't imply that you support a different proposal (fork)). I wouldn't mind 2 MB blocks right now (if they properly fixed that attack vector without a workaround as Gavin proposed). However, in regards to Segwit or 2 MB blocks, I'd take Segwit.
I'm not sure that increasing it to 2MB would cause re-centralisation of the bitcoin network.
I am also not sure that increasing it to 2MB would do very much! We could probably do wih increasing it to 10MB to make it more viable and a faster network (also offering greater difficulty and with greater difficulty can come a faster speed of hardware).
All wrong. 2 MB might cause some centralization due to less nodes being online (the number is dropping anyways though). We can't do 10 MB blocks because of propagation delay, validation time and orphans. An increase of the block size limit won't make the network 'more viable and faster'. The speed of the network will remain the same.