Comments on Russell's Phenomenalism

1. Existence of sense data: Sense data are supposed to be things that (a) are known directly, without inference and (b) have the characteristics they appear to have, and (c) explain illusory perception. Are there any such things?

2. Incorporating the sense data of others: My access to the sense data of others is through their bodies, for example what they tell me about their experiences. Before I can have knowledge of these bodies, I must have a pretty good idea what sense data make up such bodies. Thus before I can know what sense data other than my own are incorporated in some else's body, I must already know what that body consists of, that is, the sense data that make it up. The upshot of this is that there are no epistemological advantages to Russell's Phenomenalism unless it is reduced to solipsism, the view that the world consists of my own sense data.

3. Distinguishing Illusion from Reality: Russell recognizes that not all sense data that are "of" a thing are going to end up part of the series that constitutes the thing. The illusory ones are ruled out on the ground that they do not obey the laws of physics. This will not work as stated, since the laws of physics are about objects, not sense data. To apply the laws, he needs to know what objects exist. But which sense data are parts of objects, and thus the character of those objects, is only determined by the "laws of physics" about objects. Thus the laws need to be known to specify the objects, and the objects need to be known to discover the laws.

Another problem is whether the laws of physics should be taken to be the laws as we now believe them, or the as the true laws, now only partially known. If we exclude sense data that represent things we don't believe can happen, then we bar ourselves from discovering new phenomena, while if we exclude sense data that represent things that are barred from happening by the true-but unknown-laws, then we can not now know what sense data these objects consist of.

Summary points: Objects are supposed to consist of sense data so that those objects can be easily known. If they consist only of my own sense data, then the objects are not "objective," that is, capable of being known by others and being the basis of common activities and knowledge. If objects consist of the sense data of various people, the claimed ease of knowability disappears, since the sense data of others can only be know by interacting with their bodies. Thus you can have subjective, easily knowable objects, or objective objects that are as hard to learn about as the objects we usually accept as not consisting of sense data, but being independently real. Thus there is a fundamental tension in Phenomenalism, which come out in this subjective/objective conflict. This tension is present in all forms of Phenomenalism, not just Russell's.