ABSTRACT
Another criticism has more recently been advanced of an assumption which Locke's theory shares with many others. That is the assumption that the particulars grouped by a general word or concept owe their membership of the class to a precise point of resemblance or feature or aspect equally present in each one of them. It is in virtue of this supposed common aspect that, for Locke, anyone of them can represent the others. Yet if we consider the class of red things, it will be seen that they form a continuum within limits such that two red things, A and B, at opposite ends of the continuum, may resemble each other even in respect of their colour less that each resembles other objects falling outside the class. If, for example, A is close to orange and B to mauve, then A will resemble certain orange things more than it resembles certain red things, including B. It follows that the limits of the class cannot be set by 'resemblance'; and indeed similarity, when unspecified, is essentially indefinite and unbounded. Because it draws no line, the assertion that 'A is like B in colour' has no determinate truth-conditions. Unless in a special context, it could be taken only as the expression of a subjective response, an indication of how the speaker is struck which auditors might 'understand', and with which they might sympathize, but which they could not regard as true or false." The predicative judgement that A and B are both red cannot be reduced to a judgement of similarity: it presupposes a boundary round the class of red things. But it is evident that that boundary is essentially conventional, set by the ordinary meaning of the word 'red'. The same nominalist argument can be extended to many other concepts, although not, as we shall see, to all.
