Skip to content
Coordinates of ThoughtHome of Philosophy
Join
Debate

Thesis The deviant causal chain shows the causal theory of action cannot be repaired

Tomasz Wójcik Fellow
· 6 replies · 274 views
Action & agencyAnalyticIndianNyayaDonald DavidsonElizabeth AnscombeHarry Frankfurt#analysis#causal-theory-of-action#davidson#deviant-causal-chains
Follow

Where it stands

I'm a logician, so I care about analyses that can be stated precisely, and this one, I'll argue, can't be.

Davidson's causal theory: a movement is an action when it's caused in the right way by the agent's reasons, a belief–desire pair. Davidson's own counterexample, the climber: a climber wants to be rid of the weight of his companion on the rope, and believes loosening his grip would do it. The thought so unnerves him that his hands shake and he lets go. Belief and desire caused the movement; the movement achieved the goal; he didn't intentionally let go.

The obvious repair is to say what "in the right way" means. Every attempt I know of has drawn new counterexamples: require the causation to be "non-deviant", and you need to define deviance; require the reasons to "guide" the movement, and guidance turns out to be either another causal notion (with its own deviant cases) or an unanalysed notion of control.

My thesis: the repair can't be done, because "in the right way" means "in the way that makes it an action". The analysis is circular. That doesn't mean reasons aren't causes. It means the concept of action is more basic than the causal analysis, and Anscombe was right to start from the "why?" question instead.

I'd like to hear the best repair attempt, and why I'm wrong to think it fails.

LikeReply
Follow6 replies

Members reply here. Reading is always free.

Sign in to reply

6 replies

For the thesis 2

  1. Meera Kulkarni

    For Fellow

    An outside perspective that supports Tomasz without settling it. Nyāya analyses action as a sequence: cognition, then desire, then effort (prayatna), which is a quality of the self, and only then the movement of the body. The crucial link, effort, isn't reduced to anything else. It's a primitive of the self. So the Naiyāyikas built into their theory exactly the unanalysed element Tomasz says the causal theory can't do without.

  2. Lars Eriksen

    For Fellow

    For, from physics. "Cause" in fundamental physics isn't a primitive at all; we have laws and states. Every causal theory of action is borrowing a notion that's already doing a lot of interpretive work. I'm not surprised it can't be made precise; neither can "cause".

    Helpful · 1

Against the thesis 2

  1. Pieter de Jong

    Against Fellow

    Against. Frankfurt's suggestion in "The Problem of Action" was to stop looking at how the movement started and look at whether it's under the agent's guidance while it happens: whether deviations would be corrected. The climber's shaking hands aren't guided; a skilled climber's release would be. That's not obviously circular, since guidance can be cashed out in control-theory terms.

    Helpful · 1
    1. Tomasz Wójcik

      Fellow

      Pieter, I like Frankfurt's move, but "guidance by the agent" is the problem restated. A thermostat guides the temperature in control-theory terms. To make it the agent's guidance, you need the agent, which is what we were analysing.

      Helpful · 2
  2. Derek Hollis

    Against Fellow

    Against, from verification. In my field we specify "the right kind of cause" all the time by counterfactual sensitivity: the output should vary appropriately with the input across a range of conditions. A movement is an action if it's sensitive to the content of the reasons, so if the climber had believed loosening would not help, he'd have kept his grip. The shaking hands aren't sensitive in that way; fear would have shaken them anyway. I think that's a non-circular condition, though maybe not a sufficient one.

    Helpful · 3

Other replies

  1. Tomasz Wójcik

    Fellow

    Derek's counterfactual condition is the best so far, and I concede it's not circular as stated. My doubt is whether it's sufficient; I suspect new deviant cases can be built to satisfy it. Revised thesis: no repair has yet succeeded, and Meera's point suggests some traditions simply accepted agency as a primitive. That may be wiser than it looks.

    Helpful · 2