Notes from the Wired

Rational Animals by Donald Davidson

June 20, 2025 | 1,578 words | 8min read

Paper Title: Rational Animals
Link to Paper: https://www.jstor.org/stable/42969807?seq=1
Date: 1982
Paper Type: Philosophy, Action Theory
Short Abstract:
In this paper, Davidson tries to determine the distinguishing characteristics between rational and non-rational animals. For him, the key difference lies in having propositional attitudes, such as belief and desire—which require language and communication as prerequisites.

1. Propositional Attitudes

We think of some beings as rational and some as not—rational beings are those that act on reasons, sometimes deliberate, and imagine consequences. Rationality does not imply perfection; rational beings can make errors and act against their own best interests.

But what differentiates non-rational animals (animals without reasons) from those that can reason?

Davidson proposes that the key lies in propositional attitudes, such as belief, desire, and intention. Importantly, these form a matched set—they cannot exist in isolation within a being but must appear together. A belief presupposes other beliefs, and both beliefs and desires require basic attitudes like intention, as well as language.

Although Davidson wants to avoid relying too much on specific examples, doing so cannot be entirely avoided due to the nature of the discussion.

2. An Exemplary Case for Animal Rationality

Norman Malcolm tells the following story to show that dogs think:

Suppose our dog is chasing the neighbor’s cat. The latter runs full tilt toward the oak tree, but suddenly swerves at the last moment and disappears up a nearby maple. The dog doesn’t see this maneuver and, on arriving at the oak tree, he rears up on his hind feet, paws at the trunk as if trying to scale it, and barks excitedly into the branches above. We who observe this whole episode from a window say, “He thinks that the cat went up that oak.”

Malcolm claims that we can attribute belief to the dog in this case.

We are asked to infer from what we observe that the dog wants to catch the cat. He runs where he does because of his desire and his belief about where the cat has gone. The point is: if we are justified in inferring beliefs, we are also justified in inferring intentions, desires, and more.

But what about the dog’s supposed belief that the cat went up the tree? Suppose the oak tree was the oldest tree around—did the dog think the cat went up the oldest tree? Or that the cat always goes up the same tree when being chased? These questions don’t make sense.

One way to test for propositional attitudes is through substitution failure—if a sentence changes its truth value when we substitute one term with another that refers to the same thing, then the subject holds a propositional attitude.

For example, the belief that “the cat went up the oak tree” is not necessarily the same as “the cat went up the oldest tree,” even if they refer to the same tree. But this makes it hard to pinpoint what the dog actually believes, or whether it has the conceptual apparatus to hold such differentiated beliefs.

One might respond that the right way to describe the dog’s belief is: “The dog thinks, with respect to the oak tree, that the cat went up it.” But even this assumes that the believer—in this case, the dog—can refer to or pick out specific objects.

What kind of description would that be? For instance, does the dog believe it is a tree? That seems unlikely, because that would require the dog to have general knowledge about trees: that they grow, need soil and water, can burn, and so on.

Without a range of general beliefs, there’s no reason to identify the dog’s belief as a belief about a tree.

3. Network of Beliefs

Thoughts are located within a dense network of beliefs. A single belief not only requires other beliefs to support it but also presupposes the presence of other propositional attitudes.

For example, in order to believe that the cat went up the tree, I must already hold many other true beliefs about cats and trees—their appearance, their behavior (e.g., that cats can climb trees), and so on. Beliefs also imply other logically connected beliefs, which must also be held for the original belief to make sense.

A belief cannot stand entirely on its own—radically incoherent thoughts are impossible. This interconnectedness extends to other attitudes: intentions and intentional actions can be explained in terms of beliefs and desires. Similarly, emotions are connected to and depend on certain beliefs.

This does not mean that irrational beliefs or emotions are impossible—but rather that the possibility of irrationality depends on already having rational capacities. One can only be irrational if one is in some sense rational to begin with.

4. The Importance of Language

At this point, one might wonder about the role of language in relation to all these propositional attitudes.

Davidson assumes that an observer can identify what beliefs, desires, and intentions an agent has. Importantly, this does not mean that we can always state explicitly what evidence is needed to determine the presence of a particular thought, nor that observation is the only way to detect thoughts, nor that all thought is reducible to observable behavior.

As we’ve seen, the dependence of a belief on other beliefs and attitudes forms a very complex pattern. For us to reasonably attribute thought to a being, we need strong evidence that this pattern is present. Unless this complex pattern exists, there can be no thought. Davidson argues that such a pattern can only arise in creatures that have language.

He acknowledges that this idea is not entirely new, but he distinguishes his position from others by making several clarifications:

Instead, Davidson’s thesis is that a creature cannot have thoughts unless it has language. In order for a creature to be rational, it must be able to express a wide range of thoughts and be capable of interpreting both the speech and thoughts of others.

But can’t we say that, in observing animals, we do ascribe beliefs, desires, and intentions to them—just as we do with humans? And since we lack a better method of explaining animal behavior, isn’t this good enough?

Davidson replies: Just because we don’t have a better explanation doesn’t mean the one we have is true.

5. Language as a Requirement for Rationality

Davidson’s argument proceeds in three main steps:

  1. All propositional attitudes (such as belief, desire, intention) require a network of other attitudes. He established this in the previous section.
  2. In order to have a belief, one must possess the concept of belief.
  3. In order to have the concept of belief, one must have language.

Davidson draws an important distinction about thoughts:

This is close to the distinction I have in mind between believing that p and believing that one believes that p. The second is a belief about a belief, and so requires the concept of belief. […] I hold that in order to think, one must have the concept of a thought, and so language is required in both cases.

In essence, he claims:

In order to have any propositional attitude at all, it is necessary to have the concept of belief—to be capable of having a belief about a belief.

To illustrate this, Davidson offers a thought experiment:

Imagine you believe you have a coin in your pocket. You check your pockets and find no coin—you would be surprised. Now imagine you had no belief that there was a coin in your pocket, and upon checking, you also found none. Clearly, you would not be surprised.

Thus, surprise presupposes belief. More specifically, it presupposes awareness of the contrast between what one believed and what actually occurred. This awareness is a belief about a belief. More generally, Davidson suggests that surprise is a basic condition for having thought at all.

So, what is the concept of belief? A belief is a mental state that can be either true or false. Therefore, to have the concept of belief is to possess the concept of objective truth—the ability to distinguish between subjective perspective and how things actually are.

A creature may interact with the world in complex ways, but still not possess this subjective-objective contrast required for belief.

What would be needed for this contrast? Davidson’s answer is: linguistic communication.

Communication requires that each communicator has a concept of a shared, intersubjective world. Each must understand that others can see the world differently and that these perspectives refer to the same objective reality.

To complete his argument, Davidson would need to show that language is the only way to achieve this kind of intersubjectivity. He acknowledges, however, that he has no direct argument for this final step.

He concludes with:

Our sense of objectivity is the consequence of another sort of triangulation, one that requires two creatures. Each interacts with an object, but what gives each the concept of the way things are objectively is the baseline formed between the creatures by language. The fact that they share a concept of truth alone makes sense of the claim that they have beliefs, that they are able to assign objects a place in the public world. The conclusion of these considerations is that rationality is a social trait. Only communicators have it.

Email Icon reply via email
▲
Made with Hugo, website licensed under CC BY-NC-SA 4.0.