Mental events follow no laws, yet reasons can still be causes that leave us free
Davidson argues that mental events fall under no strict laws and are identical to physical events only as particulars; but in the explanation of action, reasons are genuinely causes, and when a reason causes an action the agent is precisely then free.
The video won't play here. Listen to the audio instead:
The argument · tap a timestamp to hear it
Reducing belief to behavior collapses into an infinite regress
Davidson points out that reducing beliefs and desires to behavior — defining belief, for instance, as assenting to a proposition — collapses into an infinite regress: to judge whether a given act of assent is genuine assent, you have to appeal to a further intention, and every layer requires a new condition. Beliefs and desires therefore have content only inside a whole made up of other beliefs and desires. This holism in the mental domain establishes both that the mental is autonomous and that it lacks the kind of lawlikeness the physical domain has, which is what grounds the claim that the mental is anomalous.
Between the mental and the physical there are correlations, never laws
Homonomic generalizations stay inside one and the same conceptual scheme — length measurement, say, built out of a set of postulates and axioms applied to macroscopic objects, and refinable step by step toward a strict law. Heteronomic generalizations instead link physical vocabulary to mental vocabulary. Davidson holds that cross-domain generalizations of this kind cannot be tightened into laws by adding further conditions, because describing the correlation necessarily means switching into the other vocabulary. So between the mental and the physical there are only loose correlations, no strict laws — and that is the core of anomalous monism.
If the mental is to cause movement, it must be a physical event
Davidson establishes the identity of mental and physical events in three steps: suppose a mental event M causes a physical event P; then under some description M and P must fall under a strict law; and the only strict law available is a physical law. So M falls under a physical law, which means M has a physical description — M is itself a physical event. Refusing this identity leaves only a retreat to Cartesian interactionism, with mental substances intruding on the closed circuit of physical causation for no reason at all. A variant of Laplace's demon is used to reinforce the conclusion: knowing every physical fact still gets you to no mental event whatsoever.
Wanting the light on is not wanting my arm to move just so
Davidson uses the light-switch example to pull apart the object of an intention: wanting to turn the light on is not the same as wanting my arm to travel along one precise path; every detail of what actually happens (angle, speed, timing) is a contingent token. As he writes, the details of the action are of course all fixed, but it makes no sense to demand that my wanting be aimed at one unique moment or manner of acting. The point answers behaviorism and volitionism at once: no surplus ghost of the will needs to be introduced, because a general pro attitude plus a belief already explains why this particular token action occurred on this occasion.
Whether physical causation is really closed is not obvious
If one and the same event has several mental descriptions (wanting to signal, wanting to turn on the light, wanting to light up the room) and also several layers of physical description (molecules, cells, muscles), then which physical event does token-token identity match the mental event to? Davidson gives no answer. Participants note that the language games of the biological and the physical levels are likewise irreducible to each other, and may even involve emergence between levels; which means it is not obvious that the closed circuit of physical causation is genuinely closed. This is one of the few places in the episode the author never defends, and the largest doubt the discussion leaves standing.
When an action succeeds, the act of will is a redundant step
Host Mark offers one way to read it: separate the mental state that corresponds to the reason from the mental state of resolving to act, give each its own brain state, and let the brain states form a physical causal chain; then "reasons are causes" becomes: the reason, as a physical event, causes bodily movement by way of the physical correlate of the volition. The Wittgensteinian objection is that an extra mental event of willing is needed only in cases where the will fails (paralysis, for instance); adding it in the successful case is redundant. The disagreement exposes that the levels of reason, intention and causation in Davidson's text are not fully sorted out.
The plate broke because of a hidden crack: states can be causes too
The first objection: beliefs and desires are only states or dispositions, causes must be events, so states are not causes. Davidson replies that we are perfectly willing to say "my head hit the door frame because I am tall" or "the plate broke because it had a hidden crack" — states and structures can do causal work. More importantly, mental states such as perception are not static dispositions; they change over time, they are dynamic. So there is no reason to insist that only events can enter causal explanation; states can too, as long as they sit at the right place in the physical causal chain.
Only when reasons are causes is the agent a willing agent
The sharpest objection is that attributing an action to a cause turns the person into a victim of physical necessity. Davidson turns the question around: why should having a cause turn an action into a mere happening and the person into a victim? It is when a reason causes an action that the agent is a willing agent. Granting that reasons are causes does not erase agency; on the contrary, only by making reasons causal factors that bring actions about can a person act meaningfully in the physical world, rather than being reduced to a shell that gets redescribed after the fact. This is the heart of his answer to the worry about free will.
In their own words · checked verbatim
beliefs and desires issue in behavior only as modified and mediated by further beliefs and desires, attitudes and attendings without limit.
mental and physical predicates are not a priori made for one another
There cannot be tight connection. Between the realms, if each is to retain its allegiance to its proper sorts of evidence.
Every detail is fixed, but it makes no sense to demand that my want be directed at an action performed at any one moment or done in the same unique manner
why on earth should a cause turn an action into a mere happening and a person into it help as a victim
Figures
| Publication years of the two papers | Mental Events: 1970; Actions, Reasons and Causes: 1963 | 0:00 |
| Length of the appendix to Davidson's paper | 3 pages | 6:57 |
| Pages left undiscussed | 8 pages | 43:38 |
Glossary
- token identity
- Mental and physical events are one and the same event as concrete particulars, not a lawlike correspondence between types.
- homonomic generalization
- A generalization within a single vocabulary that can be refined toward a strict law, as in length measurement.
- heteronomic generalization
- A generalization spanning mental and physical vocabulary that cannot be refined into a law.
- pro attitude
- Covers desires, wants, principles — every inclination that pushes toward action.
- primary reason
- A pro attitude plus the relevant belief, together constituting the complete causal explanation of an action.
- rationalization
- Describing an action as having a reason, that is, clothing it in an intentional description.
How to listen
Readers in philosophy and cognitive science who want to work out why reasons can be causes, and how anomalous monism squares with freedom; product people and researchers curious about physicalist accounts of mind and body.