Notes from the Wired

Actions, Reasons, and Causes

May 9, 2025 | 2,252 words | 11min read

Paper Title: Actions, Reasons, and Causes
Link to Paper: https://www.jstor.org/stable/2023177?seq=1
Date: November 1963
Paper Type: Philosophy, Action Theory
Short Abstract:
This paper defends the view that rationalization is a type of causal explanation, his central thesis being reasons are causes, meaning that the reasons for an action are not merely justifications but also the causal explanations that bring about the action.

1. Introduction

The existence of reasons makes an action understandable only if it indicates what about the action appealed to the agent. It is not enough to simply say that the action appealed to them; this does not explain the action.

To characterize an action with an understandable reason, we can say that:

These attitudes can include desires, wants, urges, promises, social conventions, goals, and so on. All of these can be interpreted as ways an agent may relate to or feel about an action. Importantly, the attitude needs to cover two aspects:

An attitude is distinct from a conviction in that a conviction aims at something right or good, while an attitude does not.

Explaining why an agent performed an action often involves identifying both:

This is what we can refer to as the primary reason for an action.

Regarding the primary reason, we can make the following claims (C1):

  1. To understand the reason for an action, it is both necessary and sufficient to know how the primary reason is constructed.
  2. The primary reason for an action is its cause.

2. Why Knowing the Primary Reason Helps Us Understand the Reason for an Action

Let’s look at the following example:

I flip a switch to turn on the light. Unbeknownst to me, this alerts a robber in my room.

By saying “I wanted to turn on the light,” I rationalize my action—that is, I give a reason for it—but I do not provide a reason for alerting the home invader or for why I illuminated the room. Importantly, just because two things are physically the same action (alerting the home invader and illuminating the room) doesn’t mean the reason applies equally to both descriptions.

Thus, a necessary condition for a primary reason is:

  1. A reason R is a primary reason for why an agent performed an action A under the description d only if the agent has an attitude toward R, and believes that A, under the description d, has this attitude.

Now, let’s compare “I turned on the light” and “I wanted to turn on the light.” They differ in that desires have general aims, not particular events, while actions are specific and concrete. They do not refer to the same thing.

One can want to do something and never do it. One can also do something without having wanted to do it beforehand. Thus, “I wanted to turn on the light” ≠ “I turned on the light.” Saying “I did it because I wanted to” generally tells us that it was intentional and the agent had an attitude toward the action, but it does not tell us why they wanted to do it. It’s a kind of minimal explanation. Furthermore, “wanting” is a general category that includes feelings like lust, admiration, need, or obligation. If we already know that an action is intentional, then simply saying “I wanted to do it” doesn’t provide much new information.

A primary reason has two parts: the attitude and the belief. In practice, we often mention only one part. If one mentions the belief, the desire is usually inferred from the content. If one mentions the desire (e.g., “I want to insult you”), the belief about the action (e.g., “showing you my middle finger”) and its expected outcome (e.g., insulting you) are generally assumed. So, mentioning both parts is often unnecessary.

The same logic applies to indirect reasoning. If you want goal Y, and you believe X leads to Y, then you will come to want X as a means to that end.

When we bring up emotions as reasons for action, this works because we implicitly assume that a primary reason is at work—that is, a belief-desire structure. For example, “He left the party because of claustrophobia” implies a desire to avoid fear and a belief about how to avoid it. Emotion-based explanations are shorthand for primary reasons.

Another Example:

“He boasted out of vanity” means:

“Whenever he gets a chance to earn admiration or envy, he does what he believes will get it.”

When we know why someone acted—what their primary reason was—we understand their intention. For example, “I turned left to get to Karlsruhe” tells us both the reason and the intention.

However, knowing the intention is not necessarily the same as knowing the primary reason in full detail. For example, “James went to church with the intention of pleasing his mother” implies some attitude toward pleasing her, but not what kind of attitude:

So, we know part of the primary reason (the belief and the general goal), but not the specific motivation.

A reason explains an action if the agent believed it, even if it is false. However, a reason justifies an action only if it is both true and good. For example, if I believe “You stepped on my toes,” and step on yours in return, that belief explains my retaliation. But my action is only justified if you really did step on my toes.

Thus, explanation is foundational; justification builds on it, but not vice versa.

3. The Role of Reasons as Causes in Action Explanation

Every explanation of an action implies that the action had some desirability for the agent. This means that every rationalization of an action—that is, when the agent explains why they did something—is, in an internal, subjective sense, a weak justification.

In C1 about primary reasons, we said: An action is rationalized by a primary reason if the agent had a belief and an attitude, and the action is intentional under a description that fits those mental states. This captures the idea that an action is rational from the agent’s perspective. But importantly, rationalizations are also causal explanations. The primary reason doesn’t just make an action understandable; it is also the cause of the action.

Davidson’s famous thesis is:

Reasons are causes.

Why Primary Reasons Are Not Just Justifications

But why isn’t it enough for a primary reason to provide justification? Why must it also be a reason in a causal sense?

Suppose we say:

“He had reason R, and he did action A.”

That doesn’t explain why he did A because of R. Maybe R was irrelevant or coincidental. To explain an action by citing a reason, we need to capture the “because”:

“He did A because he believed B and desired D.”

That “because” must be understood causally, Davidson argues, or else it is meaningless. It’s not enough that the action was merely justified by the reason—we must show that the reason caused the action.

To Summarize:

Reasons as Both Causes and Interpretations

Further, reasons for action are not mere causes—they also offer an interpretation of one’s action.

For example, if someone says, “I raised my arm in order to signal,” they are redescribing the original action. However, this doesn’t tell us why the arm was raised unless we accept a causal connection between the intention to signal and the act of raising the arm. Just redescribing the action doesn’t explain why it happened—we still need to explain what brought it about.

Many philosophers before Davidson attempted to separate reasons and causes because they feared that doing so would make humans seem mechanical. However, Davidson argues that this is a mistake.

Events are often redescribed in terms of their causes. For instance, saying, “he was burned” implies that something caused the burning. Therefore, just because reasons can redescribe actions doesn’t mean they aren’t also causes.

You can have a reason for an action, but if the reason didn’t cause the action, then it’s not the reason why you acted.

Giving a reason explains an action only when the belief-desire pair (the reason) caused the action. Merely redescribing the act in different terms doesn’t explain why the agent did it—it may simply reflect how we interpret it.

4. Davidson’s Augmentation of C1 with C2

Davidson argues to augment C1 with the following condition:

C2. A primary reason for an action is its cause.

Without C2, we cannot distinguish between:

First Attack: Critique of C2

The first critique against C2 is as follows:

Davidson counters:

  1. States and dispositions can be causal explanations: For instance, “The bridge collapsed because of a structural defect” is a causal explanation, even though “structural defect” refers to a state, not an event.
  2. Mental events can be associated with primary reasons: Beliefs and desires may not be events in themselves, but there are events causally connected to subsequent actions. Even though we speak of “a desire” or “a belief” abstractly, what causes action is often the coming-into-consciousness of that state—an event. For example, you begin to want to eat a melon.
  3. Not all explanations need a single, unified cause: You can’t find one unique cause for all bridge collapses, plate breakings, or plane crashes. Different events can cause similar outcomes.
  4. Even if the agent can’t name the causal event, we can assume one exists: Just because the agent can’t name the event doesn’t mean there wasn’t one.

Second Attack: Logical Distinction Between Cause and Effect

The second objection is:

Davidson counters:

  1. Misunderstanding the role of redescription: Redescription does not equal identity. Just because an action is interpreted through a reason doesn’t mean the reason and the action are the same thing.

  2. You can describe the action without reference to the cause:

    • Action: “I flipped the switch.”
    • Reason/cause: “I wanted to turn on the light.”

    There is no logical link between these descriptions—they are empirically connected, not analytically entailed. This rebuts the claim that the reason is merely a restatement of the action.

  3. Reductive definitions don’t work for desires or beliefs: Desires are not reducible to action dispositions. Desires manifest in multiple and complex ways—not just through rationalizing actions. This deflates the idea that primary reasons are logically tied to actions in a way that would block causal explanation.

Third Attack: Causal Statements and Predictive Laws

The third objection is:

Davidson counters:

  1. Causal explanation does not require predictive laws: Just because we can’t predict the exact outcome doesn’t mean we can’t explain it. For example, “The window broke because a rock hit it” can be explained without being able to predict exactly which rocks will break which windows. Thus, predictive failure ≠ explanatory failure.
  2. Clarifying Hume’s doctrine: Davidson argues that Hume’s claim—that singular causal statements entail laws—is ambiguous. There is a stronger and a weaker version of Hume’s claim. Using the weaker version (which Davidson favors) allows causal explanations to be meaningful, even when we don’t know the laws in play. It applies just as well to rationalizations.

Fourth Attack: Self-Knowledge and Causal Relations

The fourth objection is:

Davidson counters:

  1. Self-knowledge isn’t infallible: You might misidentify your action or misidentify your motive.
  2. Lack of evidence ≠ lack of causal knowledge: Causal knowledge doesn’t always require inductive reasoning. For instance, you can know that a window broke because a rock hit it just by witnessing it happen—no law or evidence required.

Fifth Attack: Bodily Movements and Desires

The fifth objection is:

Davidson counters:

  1. False dichotomy between agent and cause: The concern stems from a false dichotomy: Either I cause my action, or something else causes it—and I’m not in control. Davidson argues this is a false choice: Not all causes require agents.
Email Icon reply via email
▲
Made with Hugo, website licensed under CC BY-NC-SA 4.0.