Are you really sure about this? It sounds equally plausible to me that not knowing both the source and target languages would give one an advantage of not relying on ad-hoc, hard-to-model, human judgements. Being monolingual is more likely to enforce a discipline where one develops an algorithm which would work effectively on all natural languages.

IIRC the 'candide' group was (not intentionally) composed of scientists with no knowledge of both english and french.. http://www.cs.cmu.edu/~aberger/mt.html