On Cause


     In this paper I will explain Hume’s skeptical attack on causation alongside Kant’s defense of the concept. First, I will elaborate on Hume’s 2 pronged attack on the notion of cause and effect. Then, I will move to Kant’s argument that it is a pure concept of the understanding, arguing for why he believes that two events which exist in objective succession necessitate a causal rule. Then, I will field an objection to Kant’s theory and consider whether he still satisfies the requirement to justify causal claims.

     The first prong of Hume’s attack on causation begins with a separation of two classes of events: (1) those that follow each other in succession and (2) those that are necessarily connected. Note that (2) is a proper subset of (1). Hume points out that while necessarily connected events follow each other in succession, events can exist in succession without being necessarily connected.

     From this, Hume argues that we are not justified in making causal claims based upon singular events. He writes that “the first time a man saw the communication of motion by impulse, as by the shock of two billiard-balls, he could not pronounce that the one event was connected; but only that it was conjoined with the other” (Hume 50). Hume instead argues that our notion of cause and effect is generated when we observe a repeated number of instances where event1 and event2 occur in succession.

     Yet, he adds that “there is nothing in a number of instances, different from every single instance” (Hume 50) from which we are any more justified in making causal claims. Repeatedly observing some events in succession only justifies correlatory claims, not causal ones. Yet, whether within natural science or in day-to-day life, Hume notes that we still generate causal notions. So, he concludes that these notions must only be a symptom of the mind’s propensities. Thus, Hume believes that causal connections are something of which we only “feel in the mind” (Hume 50). Hence, Hume’s central point in the first prong is that causation is not of a necessary relation between events, but rather a psychological phenomenon – a mental habit – that relates events which succeed another with high correlation.

     Now, the second prong of Hume’s attack involves the relation between the past, present, and future. As noted above, our causal claims are justified by repeated experience of events following another in succession. Yet, when we make causal claims we are not only talking about the necessary connection between events in the past and present. Rather, we are extending the claim that these events are necessarily connected to all points in the future.

     Hume realizes that any claim of this sort relies on the following principle: The future will be like the past. Naming it the Uniformity Principle, he argues that we are never entitled to it, simply because it is possible – and often likely – that the future will be different from the past.

     It is possible that the world could be configured in a way such that before and only before a time tn, event2 always follows event1. Before tn, if an individual were to observe event1 and event2 occur in succession for a sufficient number of times, then it seems (s)he can reasonably justify a causal claim that asserts a necessary connection between these events. Now, after tn, event2 no longer follows event1. So, there must not be a causal connection between those two events. Yet, this cannot be known at any time before tn.

     Hume argues that we never know where we are in time with respect to tn. Hence, it is possible that a causal connection that is seemingly justified in the present can be shown as unjustified at a future time if the conditions of that future time are different from the past and present. Hume’s central point is that such a future time is always possible even if it may not be likely with respect to a certain causal relation. So, we are never justified in accepting the Uniformity Principle. Hence, it must follow that we are never justified in making causal claims, because they rely upon the assumption that future will be like the past and present.

     Now, just as with Leibniz and spatial relations, Kant utilizes a transcendental argument to push back on the first prong of Hume’s argument attack on causation. Like Hume, Kant believes that justifying a causal connection requires illustrating a necessary connection between two events. Thus, Kant holds that proving that two events are causally connected involves demonstrating a rule that requires event2 to follow event1 i.e. given the initial conditions that characterize event1, event2 must necessarily follow. I will utilize Kant’s two illustrations to put this in view.

     Case1 involves an individual looking at a house. (S)he looks from the ground up, representing the grass at t0, the first floor at t1, and then the roof at t2. Intuitively, these representations are not causally related.

     Case2 involves an individual watching a ship pass by a waterway. (S)he represents a ship upstream at t0, midstream at t1, and then downstream at t2. Intuitively, these representations are causally related.

     Kant first notes that the apprehension or the “run[ning] through and tak[ing] together” (Kant 229) of our representations is always successive. As shown in the Transcendental Aesthetic, our representations are subject to the formal conditions of sensibility. Thus, they follow another in time. Given that apprehension is the synthesis of these appearances with the faculty of imagination, then it must be the case that our apprehensions follow another in time.

     Yet, he adds that “the representations of the parts succeed one another. Whether they also succeed in the object is a second point for reflection, which is not contained in the first” (Kant 305). By this he means that we can distinguish between two forms of succession: the subjective succession of our perceptions (apprehension) and the objective succession of of what our perceptions regard (the object).

     We can now revisit the two cases to get an understanding of Kant’s point. In both examples, the apprehensions of the appearances – that is, the parts of the house and the positions of the ship – were subjectively successive. In Case1, (s)he first perceives the ground before she perceives the first floor. In Case2, (s)he first perceives the ship upstream before midstream. Yet, Kant’s point is that this doesn’t say anything about the temporal succession in the manifold of the object.

     To this, Kant adds that the order with which the house was apprehended was arbitrary. Instead of looking from the ground up, (s)he could have looked from the top down or left to right. Yet, in the case of the ship, there is no other way with which the parts could’ve been apprehended. Given the direction of the current, the forces of the water, etc. the ship must’ve first been represented upstream before it could’ve been represented downstream. Hence, “the apprehension of one thing (that which happens) follows that of the other (which precedes) in accordance with a rule” (Kant 307).

     This rule must be causal, because it requires that a certain event follow another necessarily. Now, Kant generalizes from this specific case to any general sequence of events because they all involve an objective temporal succession. All sequences of events involve a change – an alteration – and this necessitates one to first represent an unaltered state-of-affairs before representing the altered state-of-affairs. Thus, there must some a causal law that obligates that necessity. Hence, all events presuppose a causal law, because a certain outcome must obtain given the particular set of initial conditions.

     Now, I will consider one objection: an event where nothing changes. One case of this is a ship that stops in the middle of a flowing river at t1 and maintains the same position through some later time t2. To an observer standing from the shore, the representation of the ship’s position at t1 is the same as the representation of the ship’s position at t2. We already know that there must be a subjective succession, yet – in Kant’s words – is there a corresponding objective succession? An objector could say no, arguing that the representations of t1 and t2 are the same. Hence, they could subjectively be ordered arbitrarily, and it will follow that there is no causal law relating the events.

     Kant has two ways to respond. The first route is to deny that such a situation is an event. Yet, I think this is counter to intuition. When we take a ferry ride across a river and the ship stops in the middle of the trip for longer than a few seconds we are inclined to think that something wrong is happening.

     Instead, Kant’s other option involves denying that the representations at t1 and t2 are the same. Although the position of the ship is the same in both representations, the position of other components of the representation are not. For one, the water that was beside the ship at t1 is further downstream at t2. That could not have been represented another way, and thus, must follow a causal rule. Alongside this, Kant can take a step outside of this subjective/objective succession argument and ask why the ship itself does not move between t1 and t2. There must be some reason for this, and it seems that any explanation will have to invoke a causal law that justifies the ship’s preservation of its initial position. Hence, even though the position of the ship does not change in this case, there is still a causal law that accounts for the preservation of its initial conditions. So, Kant’s account remains intact in the face of this objection.

     In this paper, I have broken down Hume’s two pronged attack on causation. While the first prong argued that causation is a psychological phenomenon, the second prong argued that causal claims depends upon the unjustified Uniformity Principle. After this, I developed Kant’s defense of causation and showed that all events derive their representation of a subjective succession from an objective succession. This was shown to be sufficient to demonstrate the presence of a causal law. Finally, I addressed an objection regarding an event where there is no change, and showed that Kant’s position is not compromise even in the face of it.


Works Cited


     Kant, Immanuel. Critique of Pure Reason. Ed. Paul Guyer and Allen W. Wood. Cambridge: Cambridge UP, 1998. Print.

     Hume, David. An Enquiry Concerning Human Understanding; A Letter from a Gentleman to His Friend in Edinburgh; An Abstract of a Treatise of Human Nature. Ed. Eric Steinberg. Indianapolis: Hackett, 1997. Print.