Davidson on Anomalous Monism and Reasons as Causes

Davidson argues that definitional behaviorism fails because ascribing a mental state like belief requires assuming an unlimited background of further beliefs, desires, and intentions to fix meaning. This regress shows the mental realm is holistic, which grounds both its autonomy and its anomalous character relative to physical law.

Using Nelson Goodman's grue paradox, the hosts explain that a predicate like grue is simply the wrong fit for emeralds, though it would fit an imagined object called an emerier. Davidson uses this to argue that mental and physical predicates are similarly mismatched: they are not made for one another the way homonomic concepts within a single domain are.

Davidson holds that lawlike generalizations within physics work because concepts such as length are constitutive of a system of macroscopic rigid objects, so refining these generalizations stays within the same vocabulary. Because mental and physical vocabularies are not mutually constitutive, generalizations linking them are heteronomic and can never be sharpened into strict laws.

Davidson argues that if a mental event causes a physical event, the pair must instantiate some strict law, and such laws can only be physical, which means the mental event must also have a physical description and so is a physical event. This yields token-token identity between mental and physical events while denying that mental descriptions as such ever enter into strict laws.

In Actions, Reasons and Causes, Davidson treats an action as a single event with multiple true descriptions, such as flipping a switch, turning on a light, and alerting a prowler. Giving a reason for an action, like wanting to turn on the light, rationalizes only the description under which the agent acted intentionally, not every consequence of the same physical event.

Against the Wittgensteinian or Anscombean view that giving reasons only makes behavior coherent or justified rather than causally explaining it, Davidson argues that an agent can have several equally coherent rationalizations for a behavior, yet only one actually caused the action. The bug-swatting versus signaling example shows that some fact about what really caused the arm to rise is at stake beyond mere coherence.

Davidson answers the objection that calling actions caused turns agents into passive victims by denying that all causes require a separate external causer. States and events in a person that are reasons as well as causes are precisely what make someone a voluntary agent, so being caused by one's reasons is compatible with, and constitutive of, freedom.

Mark raises a puzzle about whether a combined case of wanting to signal and coincidentally wanting to swat a bug would require the single physical event to be split into two causal parts, and finds this unclear. Chris responds that it is easier to think of only one reason actually causing the action, with a token identity to a single physical event, resisting the idea that the physical event could be divided.

Wes proposes analyzing action as a causal chain from a mental state of having a reason, through a distinct mental state of willing, to the physical act, in order to make sense of how reasons under a physical description become causes. Mark resists inserting a separate mental event of willing between intention and action except in failure cases, citing Wittgensteinian grounds and a parallel with Searle's argument against perceptual intermediaries.