Abstract

This paper discusses one of the major problems for resemblance nominalism, posed by Bertrand Russell in 1911–12, and often referred to as Russell’s regress. It is the problem that resemblance must either be a universal, thus refuting a thorough nominalism, or must itself resemble other resemblances to count as a resemblance, which ultimately leads to an infinite regress of resemblances. I am going to discuss two solutions that have been proposed to this problem. I will then attempt to show in how far these are unsatisfactory, and propose my own solution which treats resemblance as a subject-relative phenomenon. My aim is to show that, on my account, there is no infinite regress and therefore no problem.