Autonomy is a key feature of human nature, and it is widely understood to be a key condition for moral responsibility. In this Diverse Intelligences project, Tania Lombrozo and her team aim to provide a precise characterization of autonomy and to explain why it is central to moral evaluation. They will also document how people attribute autonomy to human and non-human agents and relate these patterns of autonomy attribution to the attribution of moral traits and intelligence. 

The project team’s proposal is inspired by recent developments in computational modeling, which introduce the idea of a “self-model” as a feature of agents that rationally pursue goals despite limited information and resources. A self-model must support the ability to evaluate one’s own beliefs and desires, evaluate one’s own capacities and limitations, and revise these evaluations in light of evidence. 

Prof. Lombrozo’s team suggests that these capacities not only underlie autonomy itself, but also explain how we attribute autonomy to others. They provide two interrelated contributions that spring from this idea. First, they address a theoretical challenge: explaining why a notion of autonomy grounded in rational, bounded action is central to moral evaluation. To do so, they draw upon long-standing ideas in philosophy on reasons-responsiveness and provide computational tools for this problem space. Second, they propose a large-scale study investigating patterns of autonomy and moral attribution to humans, non-human animals, and machines; this will not only test their theoretical proposal but also relate attributions of autonomy to a range of moral attributes and intelligence. The result will be a much richer picture of the connections between autonomy, moral evaluation, and attributions of diverse forms of intelligence, as well as a better understanding of the perceived continuities and discontinuities between humans and other forms of natural and artificial intelligence.