On Rorty’s “In Defense of Eliminative Materialism”

by Vered Arnon

Rorty’s article “In Defense of Eliminative Materialism” is a response to criticism of his view that neurophysiological predicates can in principle replace sensation predicates. He responds to the same point made by two critics, the charge that replacing sensation predicates with neurophysiological predicates would diminish descriptive abilities. Bernstein and Cornman argue that Rorty’s suggested use of language would either be redundant or deficient. Rorty explains how replacing sensation predicates with neurophysiological predicates is neither redundant nor deficient.

The two critics claim that replacement of the predicates would only result in a sentence that entailed or expressed the original sentence, if the predicates in question play the same descriptive role. But Rorty explains that this is not the case at all. A sentence that takes over the descriptive role of another does not necessarily entail or express the same thing as the original sentence. When replacing ‘demon’ with ‘hallucination’ one is able to describe something and also negate the original sentence. One can describe something as an hallucination and deny the existense of demons, thus the entailment does not hold. One can talk about stimulation of fibres in the brain, but ‘stimulation of c-fibres’, while describing the same state of affairs as ‘pain’, does not entail the concept of ‘pain’. None of the predicates that describe brain processes entail sensation predicates because brain processes are entirely different from sensation.

Cornman and Bernstein also claim that neurophysiological predicates would be deficient in comparison to sensation predicates if they do not entail the same things that these predicates do. They insist that neurophysiological predicates would not be able to express what’s expressed by terms such as ‘intense’ and ‘throbbing’ and so on. But Rorty says that the different vocabularies explain things in different ways, and neurophysiological predicates express causal conditions, while sensation predicates are inferential. Neurophysiological predicates are able to express things from an entirely different perspective than sensation predicates.

Rorty makes it clear that he isn’t saying this replacement is necessary or superiour. He’s merely advocating the validity of neurophysiological predicates, not arguing against the validity of other sorts of descriptive accounts.


Any article can be the object of criticism. I could argue the entailment problem further or argue that neurophysiological predicates are deficient in descriptive ability. I could argue that sensation predicates are superiour and more accurate. I could justify my argument on various different grounds. On the other hand, the same arguments in Rorty’s article could be strengthened and put forward yet again in response. The article is very short, and it’s a rebuttal to begin with, so there isn’t really much left to attack.

I can criticise and object to any argument, so long as my own argument is well-constructed. In a case such as this, that seems to me simply futile. Rorty is not the only person advocating his position. Ultimately, whether or not his argument is accepted will depend not on logical arguments, but on socio-cultural trends in language. If people gradually start using more neurophysiological predicates and fewer sensory predicates, and eventually neurophysiological predicates begin to significantly replace sensory predicates, this shift will result from the field of neurophysiology addressing and exploring significant questions. Which arguments win out in the end is not relevant, because the process of discourse will establish a new norm of vocabulary. Language is a function of human behaviour, not a function of logical arguments. Any claims for or against the adoption of a new sort of vocabulary are subjective and aesthetic. Humans created language, and humans can create a valid argument in favour of or against whatever they wish.

I think if Rorty really wanted to support his thesis and show that neurophysiological predicates are valid, then instead of using logical arguments, he ought to just point out where other people are successfully using this terminology to further their inquiries and studies.

Leave a Reply

Your email address will not be published. Required fields are marked *