The Causal Theory of Perception

If all we ever see are sense data, then we never see material objects. How then can we know that they exist? If sense data are conceived as being different sorts of things from what material objects are (something Ayer denies), then sense data form a barrier between us and the material world.

One family of natural proposals for crossing the barrier is what Ayer calls the causal theory of perception. The idea is that the sense data are caused in such a way that they provide empirical evidence for the existence of material objects. For example, the objects themselves cause the sense data to form, or God causes the sense data to form in such that way that they are indicative of what the material objects are. It should be obvious that what Ayer is calling a causal theory of perception is far broader than what philosophers would ordinarily be willing to call a causal theory.

He formulates such a broad notion of a causal theory of perception because he has a single criticism that applies to all such theories. On the causal theory, I see material object is to be analyzed as one of the following:

• I am aware of sense data that are caused by material object .
• I am aware of sense data that are members of a group of sense data that are caused by material object .
• I am aware of sense data that are caused (or members of a group of sense data that are caused) by and is empirically connected to material object .

Those details aren't important to Ayer's criticism, they are only important for seeing that the theories of, for example, Descartes, are causal theories in the intended sense.

Why would anyone believe the causal theory? When I say, I see my hand in front of my face, I certainly don't ordinarily take that to be short for ... . One needn't hold that that is what I think I am doing to accept the causal theory, rather, that is what, on philosophical analysis, it turns out that my statement must turn out to mean. However, adopting such a counterintuitive stance puts the burden of proof on the causal theorist. There needs to be an affirmative argument that something like that must be the case, for if there is an alternative theory that avoids such rigging, it will almost certainly be preferable. The causal theory of perception comes with such an argument, namely, the argument from illusion.

If it is an empirical fact that there are sense data and material things, as the argument from illusion claims, it becomes necessary to understand how they are related. Since we have never seen material things, we need some sort of nonempirical argument that they must be there to yield the sense data:

Here it is:

1. Every event has a cause.
2. Sense data are events.
3. Sense data are not caused by ones own internal activities or states.
4. Sense data are not caused by other sense data.

Conclusion,

• Sense data must have causes that are not internal to us and that are not themselves sense data. Those causes are what we call material objects.

Ayer's main criticism is of 1, and he also criticizes 4, but he does that in later sections. Right now, we will see how he endorses 2 and 3.

2. Sense data are events. Actually, according to the conventions we have adopted, sense data are not events but objects. However, since sense data do not exist unless they are being experienced, each sense datum is a part of a uniquely characterized event: the event of being aware of that sense datum. That's close enough. We can either modify our terminology appropriately or, as Ayer prefers, rephrase the above argument to concern, not sense data, but events of awareness of sense data.

3. Sense data are not caused by ones own internal activities or states. Well, it isn't true that anything has a single cause. There are multiple causes, and in fact there are no sense data without appropriate internal activities or states. But that isn't a serious problem with 3, only with how it is formulated: events have (as Price put it) standing and differential causes. There are causes that must always be present for a broad class of events, and the particular causes that are special to particular events. There are internal activities and states that are necessary standing causes of sense data. For example, you have to be awake for there to be any events of awareness of sense data. But the differential causes of many events of awareness of sense data are not caused by ones own internal activities or states. There may be cases of events of awareness that are caused solely by internal states, hallucinations for example, and so 3 must be corrected to read

• Events of awareness of sense data are not normally caused solely by ones own internal activities or states.

That reformulation is adequate to do the job of 3 in the argument.

-- ShaughanLavine - 26 Feb 2003 - 22 Feb 2007