A distinctive feature of Professor Armstrong's approach to the topic of universals is that it combines different philosophical trends that have often been regarded as mutually opposed. His conception is basically Lockean, in founding universals on the causal efficacy of particulars, conceived of as entities with powers. Unlike Locke, however, he does not regard particulars (or thin particulars, in the terminology of the latter part of his paper) as the basic ontological category, but takes a sort of Wittgensteinian turn in regarding states of affairs as the fundamental ontological category, and construing particulars (as well as universals) as abstractions from states of affairs. But, unlike Wittgenstein, he regards states of affairs as consisting of universals — properties and relations — as well as particulars, and he presents a fully realistic view of universals. But, again, unlike many other realists considering universals, he does not regard them as abstract entities but fully "concrete" (he does not use the word this way) and real parts of the space-time continuum. As such, the existence and nature of universals are determined a posteriori by empirical, scientific investigation, and not by a priori reasoning about the meaning of terms.