Question: Should a person's credibility matter in judging his empirical claims? Is it rational to use a person's track record – meaning the frequency or consistency of truth in his past statements – in judging the likely truth of his current statements? In Ayn Rand's Normative Ethics, Tara Smith explains that to believe something just because someone said it is a violation of the virtue of independence. Also, to judge an argument based on the speaker is known as the fallacy of "ad hominem." However, doesn't the character of the speaker matter when considering whether to believe his claims? For example, when Thomas Sowell makes an empirical claim, my knowledge that he vigorously tests his hypotheses against the facts makes me more likely to judge his claim as true, even before I've confirmed his statement. Likewise, if a person is frequently wrong in his factual claims, I'd be sure to require lots of evidence before believing him. Is that rational? Or should all factual claims be treated equally regardless of who makes them?