If we don't know the language it's encoding, how do we know the Linear A pronunciation is correct, approximately or not? Is this done purely on the assumption t

If we don't know the language it's encoding, how do we know the Linear A pronunciation is correct, approximately or not? Is this done purely on the assumption that Linear A and Linear B might encode similar phonemes in a similar way?