Full Idea
Armstrong's theory holds that what makes certain regularities lawful are second-order states of affairs N(F,G) in which the two ordinary first-order universals F and G are related by a certain dyadic second-order universal N.
Clarification
'Dyadic' means needing two variables
Gist of Idea
Regularities are lawful if a second-order universal unites two first-order universals
Source
report of David M. Armstrong (What is a Law of Nature? [1983]) by David Lewis - New work for a theory of universals 'Laws and C'
Book Reference
'Properties', ed/tr. Mellor,D.H. /Oliver,A [OUP 1997], p.213
A Reaction
[see Lewis's footnote] I take the view (from Shoemaker and Ellis) that laws of nature are just plain regularities which arise from the hierarchy of natural kinds. We don't need a commitment to 'universals'.