I largely agree with much of your comment. However, I think it is possible for a less intelligent person to identify the most intelligent among 2-3 disagreeing

I largely agree with much of your comment. However, I think it is possible for a less intelligent person to identify the most intelligent among 2-3 disagreeing smarter people if all of them is rational and accepts solid evidence/results of thought experiments AND the most intelligent is good at articulating his/her rationale for having a belief. With this method, the perceiver must have a pretty high threshold of intelligence and rationality themselves in order to identify good reasoning and detect issues with bad reasoning/biases though. Another method is looking at the results of several predictions in different domains. Assuming roughly equal initial expertise and data, the more intelligent should get more predictions right. The latter method only requires knowledge/intuition of basic statistics.