I have beliefs about what you believe. I also have beliefs about what I myself believe. The big difference between the two cases is how I come by these beliefs. By and large, my beliefs about what you believe come from observations of your behaviour (understood in a wide sense, which includes the environment in which your behaviour is located). Here are two illustrations. You sell all your shares and buy gold. I infer that you believe that gold will outperform shares. You write an article saying that the Coalition will win the next election. I infer that you believe that the Coalition will win the next election. However, my beliefs about what I myself believe don’t usually come from observations by me of my own behaviour. My belief that gold will outperform shares may explain why I sell all my shares and buy gold, but it doesn’t reveal to me that I have this belief. Likewise, I don’t need to write an article saying that the Coalition will win the next election in order to discover that I have this belief. There is, to borrow some jargon, a first person–third person asymmetry in how we arrive at beliefs about beliefs.