Ruth Chang, DPhil
Chair and Professor of Jurisprudence
University of Oxford
Born in 1963 in Minneapolis, Minn., USA
Studied Law at Harvard University and Philosophy at the University of Oxford
Project
Making It Matter
Making It Matter proposes an alternative to dominant views of what it is to be a rational agent. Dominant views have in common the fundamental idea that rationality is a matter of recognizing reasons and values and then responding appropriately to them. In the mooted alternative, rationality centrally involves the normative power to create reasons and values through commitments that provide their grounds. Through commitments, we can quite literally – under suitable conditions – create reasons to do one thing rather than another, thereby making ourselves ultimately responsible for normative truths, such as “I have most reason to be a philosopher than a banker.” By making things matter, we write the story of our lives and create our distinctive rational identities. Setting out this view involves close examination of several major philosophical issues, such as the grounds of rational choice, the role of comparisons in practical justification, the structure of practical and theoretical normativity, different accounts of normative powers, the truth in existentialism and the truth in realism, and the determinants of agency. Instead of thinking that the world presents us with things that matter, we should allow that rational agency centrally involves making things matter.Recommended Reading
Chang, Ruth (2016). “Hard Choices.” Journal of the American Philosophical Association 3 (1): 1–21. https://doi.org/10.1017/apa.2017.7.
— (2021). “What Is It to Be a Rational Agent?” In The Routledge Handbook of Practical Reason, edited by Ruth Chang and Kurt Sylvan, 95–110. London and New York: Routledge.
— (2023). “3 Dogmas of Normativity.” Journal of Applied Philosophy 40 (2): 173–204. https://doi.org/10.1111/japp.12626.
Colloquium, 18.06.2024
The Parity Model: Putting Humans in the Loop
Value alignment, getting AI outputs to align with considered human values, is arguably the most significant open problem in AI research and development. We propose an approach to AI design rooted in a novel philosophical understanding of human values. The proposed model is an alternative to traditional forms of AI design – whether in the form of machine learning or symbolic systems – that may help us towards achieving value alignment. The numerical framework behind the model makes use of approximate differences and approximate quotients and thereby allows for the computational tractability of values outside the traditional framework of decision theory. We show how the model can then be used to facilitate communication between AI and its users when hard choices are in question.
Publications from the Fellows' Library
Chang, Ruth (Oxford, 2024)
Chang, Ruth (Oxford, 2022)
Chang, Ruth (London, 2021)
What is it to be a rational agent?
