Moral agency—the capacity to understand and respond to moral reasons—is typically required for moral responsibility. Paradigmatically, adult humans are agents, but membership extends to entities that can deliberate, form intentions, and govern their behavior accordingly. The boundaries are disputed: are infants, the severely cognitively disabled, animals, or artificial intelligences moral agents? These questions determine who bears moral responsibility and to whom we owe direct moral duties.
Consider paradigm cases of agency and non-agency (adult deliberating, infant acting on impulse, animal responding to stimulus) and ask what capacities distinguish moral agents.
From your work on normative versus metaethical questions, you know that normative ethics asks what we ought to do, while metaethics asks deeper questions about the nature and grounds of moral claims. Moral agency sits at the intersection: it is a metaethical question (what makes something a moral agent?) with direct normative consequences (who can be held responsible? to whom do we owe duties?). Getting this analysis right matters practically, because attributing moral responsibility to the wrong kind of entity — or denying it to the right kind — distorts the entire moral framework.
The concept of moral agency should be distinguished from the related but distinct concept of moral patiency. A moral agent is an entity that can bear moral responsibility — can be praised, blamed, required, and obligated. A moral patient is an entity that can be wronged — that has interests or status that moral agents are required to take into account. These categories can come apart. An infant is not a moral agent (it cannot deliberate or respond to moral reasons) but is a moral patient (it can be harmed, and we have strong duties toward it). A thermostat is neither — it lacks both deliberative capacity and morally relevant interests. An adult human paradigmatically combines both. The interesting philosophical cases are entities that seem to have moral patiency without agency, or that fall in gray zones for both.
What capacities constitute moral agency? The core candidate is responsiveness to reasons: a moral agent is an entity that can recognize moral reasons, deliberate about them, and govern its behavior accordingly. This is why infants and severely cognitively disabled individuals present hard cases — they act, they even have preferences, but whether they can recognize and respond to moral reasons is contested. Animals present a different hard case: some exhibit apparent deliberation and social norm-following (chimpanzees, elephants, some corvids), which makes pure capacity-based criteria difficult to apply cleanly. The harder you look at the edges, the more the "paradigm case" of adult deliberating human begins to look like one point on a continuum rather than a bright line.
Artificial intelligence has made this debate newly urgent. Large language models and autonomous systems exhibit sophisticated language use, apparent reasoning, and consistent behavior patterns that superficially resemble deliberation. Whether any of this constitutes genuine intention formation or responsiveness to moral reasons — or merely processes that mimic the outputs of such capacities — is genuinely unresolved. The philosophical stakes are high: if an AI system is a moral agent, it can bear responsibility for its outputs. If it is a moral patient, we owe it direct moral consideration. If it is neither, then questions about AI ethics reduce entirely to questions about the humans who design, deploy, and use it. How you answer the question of what agency requires will determine which conclusion you reach — which is why careful analysis of the capacity conditions matters before applying the concept to novel cases.
Topics in reflective domains aren't scored by quiz answers. Read, reflect, and mark when you've thought it through.