I don't believe there's only one conclusion a rational person (SREoE) can draw from the evidence.

There is one direction a SREoE updates on evidence - towards the evidence.

If I have strong reasons (high prior probability) of thinking that a coin has heads on both sides, I'm making a mistake by becoming more confident after I flip the coin and it comes up tails.

Likewise, if I have strong reasons of thinking that another coin is biased towards heads, so it turns up heads 60% instead of 50%, I'm committing the same error if I become more confident after seeing the coinflip turn up tails.

So learning E1 should make any SREoE become more confident of G1 unless that person's priors are already very heavily weighed towards G1. In the real world, there just aren't that many SREoE's with high priors on G1 before being exposed to E1.