Full Idea
A Realistic version of a Regularity theory of laws will have to postulate universals. How else will it be possible to say that the different instances of a certain uniformity are all instances of objectively the same phenomenon?
Gist of Idea
Realist regularity theories of laws need universals, to pick out the same phenomena
Source
David M. Armstrong (What is a Law of Nature? [1983], 02.4)
Book Reference
Armstrong,D.M.: 'What is a Law of Nature?' [CUP 1985], p.16
A Reaction
I disagree. We may (or may not) need properties, but they can be have a range. We just need stable language. We use one word 'red', even when the shade of redness varies. Non-realists presumably refer to sense-data.
Related Ideas
Idea 223 If you deny that each thing always stays the same, you destroy the possibility of discussion [Plato]
Idea 227 You must always mean the same thing when you utter the same name [Plato]