What it "looked" like is consistent with each of the hypotheses we mentioned so far. Until we see frame by frame lip-reading analysis, we'll need more info than we have to make a confident conclusion about what they were saying, let alone evaluate it after that. I know you're a "I trust my hunches" kind of guy. I'm a "trust no hunches" kind of guy. So my addition here isn't to bet against you but to disagree with your process. Maybe you're right, but you're guessing yourself across two or more steps to conclude that, and no one's hunches are always on.