A polyglot system needs more than one person able to read each part, and that is a lower bar than being able to write it.
what "able to review" means, concretely:
can follow the control flow
can tell whether an error is handled
can spot a missing null check
can recognise the shape of a concurrency bug
knows where the tests are, and can run them
what it does not mean: idiomatic style, library choice,
performance characteristics. those need a specialist,
and a second reviewer is not one.Setting the bar at reviewability rather than fluency is what makes a second language sustainable on a small team, and stating it explicitly stops the review becoming a rubber stamp. It also identifies what the specialist is actually needed for, which is the design conversation rather than every pull request.