In the existing literature, the concepts of ‘trust’ and ‘trustworthiness’ are built upon a central assumption: trust and trustworthiness are intimately linked because trust is agent-directed. The two concepts are reciprocal: all else being equal, an agent should trust the trustworthy agent. Most accounts assess trust and trustworthiness based on a normative requirement, treating them as two sides of the same coin. In contrast to these views, my account begins with the observation that trust and trustworthiness can come apart, both on the descriptive level and on the normative level. Descriptively, agents can and do sometimes trust untrustworthy agents or fail to trust a trustworthy agent. Normatively, we should sometimes trust untrustworthy agents. due to the ‘feedback-loop’ of trust, where trust encourages trustworthiness. Conversely, we should sometimes not trust otherwise trustworthy agents because trustworthiness is often domain- (or situation) specific.

I propose that we first understand trust and trustworthiness descriptively, as distinct phenomena. By developing an accurate descriptive account based on recent findings in psychology, I aim to provide a novel, empirically grounded metaphysical view. Once both descriptive and metaphysical accounts are established, we can then address the normative question of when an agent should trust another.