Zach Thornton

My philosophical research is in metaphysics, value theory, and the philosophy of language on issues related to identity and distinctness. My paper "The Identity of Necessary Indiscernibles" develops and defends novel general metaphysical explanations of identity and distinctness facts. It is forthcoming in Philosophers' Imprint. In other work, I explore issues related to identity, such as what makes survival distinctively and first-personally valuable (see "What Matters in Personal Transformations") and issues that arise due to differences between the granularity of our representations of the world and the granularity of the world itself (see "What Counts in Counting").

I also research AI alignment and safety, especially alignment interventions in pretraining informed by the Persona Selection Model of AI cognition. My manuscript "The Promise of Self-Fulfilling Alignment" details proposals and challenges for pretraining alignment, including a novel proposal. I'm presently working with AI researchers at dlab @ EPFL in Switzerland building normatively aligned base models using their synthetic persona pretraining technique.

I draw on the insights from feminist philosophy, critical theory, and non-Western philosophy in my work. In non-Western philosophy, I have special interest in the Daoist philosopher Zhuangzi and two schools of Hindu philosophy: Sāṃkhya and Vaiśeṣika. My future research includes projects that directly focus on these areas.