Detect wording that admits multiple plausible interpretations where the intended reading is not recoverable from context.
