Post #1265404

finally in print, the last puzzle piece in a decade of thinking about communication across social networks from the perspective of ideal rational agents: how do we factor in ‚dependence‘ - the fact that the same underlying evidence may, for example, reach us via multiple different reports, giving rise to double counting? There is a strong intuition that multiple independent observations should carry more weight than dependent observations. But how much more? We show that there is no (known) answer to this normative question in the general case. This renders a fundamental feature of human testimony unsolvable, meaning that acquiring knowledge via the testimony of others is much harder than typically assumed https://link.springer.com/epdf/10.1007/s11229-025-04969-w?sharing_token=T_qZZ-qagNeAgHYmwOsO4Pe4RwlQNchNByi7wbcMAY7NNaNx5je-8CbqY3NQbXjm1PJgoB0Fw-IBoMYIKgSDlaiwownlR0py9XbHJBti1v6lLN7PlZ2uZJjf_SE6Z-Nwu8xISFeoc5h9Szl4tzerZtJWZwsR6CbwVDWSPlumzIw%3D #epistemology #SocialNetworks @philosophy