Objections to Davidson’s Theory of Agency and Actions


Davidson’s theory of agency aims to solve the dilemma that the same action can be both intentional and not intentional. He explains primitive actions using primarily bodily movements and argues that event-causality can be described through the “accordion effect”, but not agent-causality. And Davidson uses reasons as causes to explain the actions and responds to five objections. In this paper, I critique Davidson’s argument, pointing out that he ignores certain factors in the belief-desire model, such as emotions. And his sentence holism has a problem because individual words that express complete thoughts can explain actions in language games. As a result, Davidson’s criticism of Wittgenstein is not reasonable.

Share and Cite:

Zhang, Y. (2023) Objections to Davidson’s Theory of Agency and Actions. Open Journal of Social Sciences, 11, 355-362. doi: 10.4236/jss.2023.112023.

1. Introduction

The agency involves the initiation of action by the agent. And initiation of action is compatible with intentional actions. Because intentional actions are initiated by the agent, the initiation can be explained by the causal relationship between mental states and events. According to Davidson, the initiation of action by the agent consists of causation by the belief-desire model (Schlosser, 2019) . Davidson is the proponent of anomalous monism. He claims that mental phenomena only can be explained by physical events. But mental states explain the action by making it rational and the causal potency of action belongs to mental events. He denies non-causal theories of action, which deny that reasons explain actions by causing them (Yalowitz, 2021) .

2. Objections to Davidson’s Theory of Agency

Davidson distinguished between intentional actions, actions performed without intention, and non-actions. However, there is a dilemma: if one’s intention is to spill tea into a cup, spilling coffee into the cup is not intentional. But if one’s intention is to spill the contents of the cup, then spilling coffee into the cup becomes intentional through redescription. The question is whether this redescription is reasonable or not.

Redescription means that intentional content may have different extensions in different circumstances. In a different possible world, the same sentence may have a different truth value. To make the description of the action true, the truthmaker must be the entity, so the content of the cup should be tea, not coffee, to fit the description. If we extend the description to spill the contents into the cup, the truthmaker changes, making the semantic opaque.

Davidson tries to solve the dilemma by using an extensional understanding of actions as events. However, two descriptions of the same act (letting e be a variable of events) still cannot resolve the dilemma. Redescription requires knowledge of the primitive intention and consideration of the extension of the redescription. Some intentions of actions are not crucial for explaining the intention of actions and can be ignored. Furthermore, choosing an extensional understanding of actions as events remains a challenge for the agent.

Davidson believes that the extensionality of the expression of agency suggests that the concept of agency is simpler or more basic than the concept of intention. He wants to find a way to explain agency without the help of the intention concept. The concept of cause and the outcomes or consequences of action are the reasons for responsibility. According to Davidson, primitive actions are more vital than event-causality, which “cannot help explicate the first attribution of agency on which the rest depend” (Davidson, 1971: p. 11) .

Davidson suggests that primitive actions are primarily bodily movements, which may also include “non-movements” such as stillness and mental acts. He rejects two objections about denying the attribution of an agent’s bodily movement as action.

Objection 1: Causes the finger to move requires more certain events, such as certain events (mental states) taking place in my brain, not only ordinary bodily movements.

Chisholm thinks that a person may intentionally perform an action without being aware of it. Davidson’s reply: Doing something that causes my finger to move does not cause me to move my finger, it is I who move my finger through bodily movements. The action requires that the agent acts intentionally under some description.

Objection 2: Some primitive actions involve more than a movement of the body.

Davidson’s reply: It depends on whether or not there is an appropriate description of the action. Some additional descriptions are trivial and do not provide any new information. However, even if the agent is not aware, we do not need to create new terminology to describe these actions.

However, Davidson’s reply to objection 1 is not convincing. If primarily bodily movements do not require more reason or causal relation between mental states and a movement of the body, how can we explain why the primarily bodily movements occurred? Davidson ignored the relationship between mental states and bodily movements. To defend Davidson, we would need to accept that mental states are a part of bodily movements.

Davidson’s reply to objection 2 is we do not need more terminology to describe actions. However, when solving the dilemma of an action being both intentional and not intentional, Davidson said that redescribing the action is a good solution. However, when facing the objection that some primitive actions involve more than just a body movement, Davidson said that additional descriptions are trivial and unhelpful. This is self-contradictory. To defend Davidson, we need to limit the scope of actions that require descriptions.

Davidson used Joel Feinberg’s concept of the “accordion effect” to explain the consequences of actions, but not their causes. Actions can be intentional or unintentional, and the consequences of actions may go beyond the intention of the agent. However, we cannot attribute agency to the accordion effect because we do not believe that inanimate objects can bring about things. An event is a case of the agency in which we can attribute its effects to a person (See Davidson, 1971: p. 17 ).

Therefore, event-causality can be described by the accordion effect, but agent-causality cannot. Event-causality can be elucidated in terms of the regular features of causal explanation, whereas agent-causality appears to lack such features.

Agent-causality has to take intention into account. When dealing with cases about Bismark, even though the consequences are the same, we will not count the officer as the agent. We need to consider whether the agent intentionally causes the effects. If some factors influence the choices that lead to making mistakes that comply with our primitive intention, then the agent causality cannot be explained by the accordion effect.

Finally, Davidson concluded that the distinction between primitive and non-primitive action collapses, and that all actions are primitive actions. He argues that distinguishing between actions and their consequences is a mistake and uses the example of a queen and king to explain this.

My objection to Davidson is that if agent-causality cannot be explained by the accordion effect and all actions are primitive actions, then when explaining why agents act in a certain way, it becomes difficult to distinguish between actions and their consequences, as well as between primitive and non-primitive actions. This makes it challenging to attribute responsibility when considering the intention behind the action, whether it originates from an agent or is influenced by external factors or the effects of others. Further exploration of intention analysis is necessary.

3. Objections to Davidson’s Theory of Action Explanation

To explain the intentional actions, Davidson proposes the concept of “primary reason”.

R is a primary reason why an agent performed an action A under a certain description d only if R consists of a pro attitude of the agent toward actions with a certain property, and a belief of the agent that A, under the description d, has that property (Davidson, 1963: p. 687) .

Davidson believes that both pro-attitude and belief are necessary for a primary reason and that justifying and explaining an action go hand in hand, with the primary reason for an action supporting the relevant belief or attitude of the agent.

The same process of action can be syncategorematic and have different details to explain it. “Why are you bobbing around that way?” can be answered with “I’m knitting, weaving, exercising, sculling, cuddling, training fleas” (Davidson, 1963: p. 690) .

Davidson cares about the explanation of actions by reasons, which he calls “rationalization” is “a species of ordinary causal explanation” proves reasons explain actions in the same way that causes explain their effect (Chappell, 1963: p. 700) . Davidson thinks the primary reason is the cause of action and reply five objections that deny the causal explanation of actions.

The first objection is that the primary reason is not an event and therefore cannot be a cause. Davidson’s response is that although motivation is not an event, nor is it an attitude, disposition, or object, there must exist an event that explains why we acted as we did. However, this response is seen as ambiguous and lacking in explanation.

The second objection is that a reason for an action is not logically distinct from the action and therefore reasons are not causes of actions. Davidson’s reply is that explanation by redescription does not exclude a causal explanation. For example, “My turning on the light” and “my wanting to turn on the light” are grammatical links rather than logical relationships. “Placing it in water caused it to dissolve” does not entail “It’s water-soluble”, so the latter has additional explanatory force (Davidson, 1963: p. 696) .

Therefore, the action cannot explain the property of action, which requires additional explanatory force. However, the property can be known through testing. On the other hand, the desires and beliefs behind actions are not necessarily rational or empirical. Hence, the causes of actions cannot be denied.

My objection to Davidson is that if the property of action requires additional explanatory force, then the property of desire and belief also requires additional explanatory force to explain it. Furthermore, desires and beliefs are empirical, but the property of desires and beliefs is not analytical, but synthetic, and cannot be known through testing. If desires and beliefs did not have essential properties, they could not be the causes of actions.

The third objection is that the relationship between reasons and actions cannot be explained by ordinary, singular causal statements that imply generalizations. Hume claims that of two events, A and B, if A causes B, it does not imply they occur together all the time. According to Hume, singular causal statements cannot entail causal laws. The distinction between Davidson’s and Hume’s casual theory is that Davidson holds that events that are causally related must be under some strict laws and the same event can be referred to under more than one description (Malpas, 2021) . However, Hume does not care about the problem of description and events and does not require that events must be under strict laws. Davidson believes that the explanation is not complete until a law is produced. But singular causal statements do not necessarily indicate a law. Although singular causal statements do not necessarily indicate a law, this does not mean there is no true causal statement for explaining actions. Nevertheless, the law requires reasons to be considered as causes for rationalization.

My objection is that if we need to prove that the law requires reasons to be considered as causes for rationalization, then the law may not be singular causal statements, but more general. Hume’s theory of causality means that causality is not credible based on empirical laws. Thus, the laws require a more general causal statement. However, if singular causal statements do not necessarily indicate a law, how can we guarantee that other laws must require reasons to be considered as causes?

The fourth objection is that “a person knows his own intentions in acting infallibly, without induction or observation, and no ordinary causal relation can be known in this way” (Davidson, 1963: p. 699) . Davidson argued that even if the observation of intentions may be incorrect, it still demonstrates that one knows the reasons for being wrong. The defect of Davidson’s objection is why the non-observational knowledge of causal relations is possible. Davidson had to prove that first-person knowledge of desires and beliefs causes actions, but it is difficult to prove. If the intention is non-observational or the person is unaware of their intention, then the knowledge of intention becomes questionable.

The fifth objection is that “actions are identical with bodily movements, and bodily movements have causes, yet he denies that the causes are causes of the actions.” (Davidson, 1963: p. 700) . Melden believes that only subsequent events can explain conduct through the causal link of desire and that the agent is a helpless victim who has to accept what happens. Davidson believes that the causes of actions may result in an infinite regress, even if some causes have no agents.

The defect of Davidson’s objection is that it ignores the fact that agents accept states from the past and future and that their intentions are influenced by these states. If some causes have no agents, why must agents be helpless victims or voluntary agents?

In conclusion, despite Davidson rejecting these five objections, it cannot be guaranteed that future objections will not succeed.

4. Objections to Davidson’s Belief-Desire Model

In Davidson’s theory of action explanation, it cannot be stated that an action is rationalized or rationally explained solely based on the agent’s belief and desire. Hence, the pro-attitude and belief of the agent are not enough for a causal explanation of actions. The agent’s belief and desire explain the action in a general structure, but not in more detail, and other factors have been ignored.

The most influential interpretation of the “belief-desire model” (BD) is the Humean Theory of Motivation (HTM), which states that “reason alone can never be a motive to any action of the will” (Hume, 1978: p. 413) .

The “belief-desire model” is incompatible with emotions. According to David Hume (1978) , the reason is not the only reason for motivating the action of the will. As G. F. Schueler pointed out, a person may act without having any desire at all. For example, someone listens to the news not because of their desire but due to their duty as a citizen. Additionally, the belief-desire model fails to explain expressive action in terms of goal-directedness. Emotions, on the other hand, motivate people to take action. For example, someone who takes a photo of someone they hate may feel unhappy even if they scratch out the eyes from the photo. Conversely, if they look at a picture of someone they love, they may smile at it, revealing that emotions affect expressive action. When a person is confronted with a fierce gorilla that has escaped from its cage, they will feel fear, and this emotional motivation is due to concern-based felt evaluations. These evaluations have a motivational force on emotional desires and reveal that reason is vital for desires. Desires are understood as evaluative notions that give the agent reasons for acting intelligently (Quinn, 1993) . Thus, the BD model fails to be compatible with emotions, which are important for desires.

According to Davidson’s theory of action explanation, Davidson emphasizes sentence holism, which argues that a direct connection between linguistic theory and events, actions, and objects described in non-linguistic terms must be made at the level of sentences. “It is inconceivable that one should be able to explain this relation without first explaining the role of the word in sentences; and if this is so, there is no chance of explaining reference directly in non-linguistic terms.” (Davidson, 1980: p. 135) .

Davidson criticizes the “building-block theory” of Wittgenstein. The building-block theory refers to the example from Philosophical Investigations. Davidson emphasizes that semantic terms should be explained by connecting them to “human ends and activities”. However, Wittgenstein argues that the individual words in the building-block theory do not have individual meanings. The words of command refer to propositional contents and have a clear reference in the language games between the assistant and builder. These words have a specific meaning in specific circumstances and have a pre-existing relationship to a particular kind of object. Thus, these words are complete thoughts that do not require further supplementation or augmentation. As a result, Davidson’s sentence holism is not accurate. It is possible to explain the role of words in sentences. Actions and objects can be described using individual words, which have complete thoughts within language games (See Perry, 1994 ).

5. Conclusion

In summary, Davidson refutes two objections to denying that attributing an agent’s bodily movements to the agent as an action is unsuccessful. He is accused of ignoring the causal relationship between mental states and bodily movements. Davidson’s response to the second objection may result in self-contradiction and his focus on agent-causality disregards the impact of the external world. Despite trying to make reason as causes for explaining actions and addressing five objections, Davidson’s replies are problematic. He also overlooks the important role of emotions in actions and wrongly criticizes Wittgenstein’s “building-block theory” fails to recognize that complete thoughts in language games, which do not require further supplementation to explain actions.

Data Availability

All data generated or analyzed during this study are included in this published article.

Ethical Statements

This article does not contain any studies with human participants performed by any of the authors.

Funding Statement

This research was funded by China Scholarship Council.

Conflicts of Interest

The authors declare no conflicts of interest regarding the publication of this paper.


[1] Chappell, V. C. (1963). Causation and the Identification of Actions: Comments. The Journal of Philosophy, 60, 700-701.
[2] Davidson, D. (1963). Actions, Reasons, and Causes. The Journal of Philosophy, 60, 685-700.
[3] Davidson, D. (1971). Agency. In R. Binkley, R. Bronaugh, & A. Marras (Eds.), Agent, Action, and Reason (pp. 3-25). University of Toronto Press.
[4] Davidson, D. (1980). Reality without Reference. In M. Platts (Ed.), Reference, Truth and Reality: Essays on the Philosophy of Language (pp. 134-135). Routledge Kegan Paul.
[5] Hume, D. (1978). A Treatise of Human Nature [1739/40]. Edited by L. A. Selby-Bigge, & P. H. Nidditch, Clarendon Press.
[6] Malpas, J. (2021). Donald Davidson. In N. Zalta (Ed.), The Stanford Encyclopedia of Philosophy.
[7] Perry, J. (1994). Davidson’s Sentences and Wittgenstein’s Builders. The American Philosophical Association, 68, 23-37.
[8] Quinn, W. (1993). Putting Rationality in Its Place. In R. G. Frey, & C. W. Morris (Eds.), Morality and Action (pp. 228-255). Cambridge University Press.
[9] Schlosser, M. (2019). Agency. In E. N. Zalta (Ed.), The Stanford Encyclopedia of Philosophy.
[10] Yalowitz, S. (2021). Anomalous Monism. In E. N. Zalta (Ed.), The Stanford Encyclopedia of Philosophy.

Copyright © 2023 by authors and Scientific Research Publishing Inc.

Creative Commons License

This work and the related PDF file are licensed under a Creative Commons Attribution 4.0 International License.