Testimonial Evidence and Bias, a Bayesian Treatment

Bias, trust and testimony present a challenge at an academic, societal and economic level. In our daily lives, we constantly assess the trustworthiness of our sources: Is this media outlet biased?  Who are the credible expert witnesses in court cases? Which websites should we trust with our personal data? Which peer-to-peer applications should we trust?  All of these questions require urgent resolution. I therefore propose a bold yet feasible research program: a unifying logic for dealing with testimonial evidence, crafted from techniques from both logic and the flourishing field of formal epistemology. Specifically, I will develop a logical framework based on higher-order probability theory, that will enable us to capture all the nuances related to these problems. Philosophically speaking, such a framework will also result in normative epistemic principles that can help to resolve theoretical and practical issues in social epistemology (e.g. countering fake news, improving science metrics) and philosophy of science (e.g. the replication crisis) .