I am likely missing the point here:
"The ERC requires that all peer-reviewed publications from
ERC-funded research projects be deposited on publication into
an appropriate research repository where available, such as
PubMed Central, ArXiv or an institutional repository, and
subsequently made Open Access within 6 months of publication."
To place the paper on ArXiv is not the same as to provide an
"Open Access"?
Doesn't the former and latter just mean: "availability to all"?
And if a publisher objects to a 6-month period and insists on
12-month one, for example? I should not sign the deal or I am
defended by law if I break it?
The point is that the ERC (and others) have wisely opted for the
Immediate-Deposit/Optional-Access (ID/OA) Mandate (or what Peter
Suber calls the Dual Deposit/Release Mandate), in which immediate
deposit (of the final, peer-reviewed draft, immediately upon
acceptance for publication) is mandatory, but the date at which
access to that deposit is set as Open Access (full-text and
metadata accessible webwide) rather than Closed Access (metadata
accessible webwide, but not yet the full text) may be delayed, if
there is a publisher embargo (but the allowable embargo period is
capped at a maximum of 6 months by the ERC, 12 months by NIH).