In his 1977 article, “Second Thoughts on Paradigms,” Kuhn draws our attention to three constituent elements of a paradigm; they are: symbolic generalizations, models, and exemplars.

When I read Kuhn, I interpret “symbolic generalizations” to mean a formal language.  It can be the language of mathematics, but it doesn’t have to be.  It can be verbal propositions that follow the rules of formal logic.

When Kuhn talks about “models” in this context, he's referring to what you and I would call a theory – a collection of statements, expressed in a formal language, that describe, explain, or otherwise portray an empirical phenomenon.  A formal language alone has no empirical content, but a theory does.

Finally, there are “exemplars,” which refer to examples that researchers share for how to apply their theories in specific laboratory or field situations.  Kuhn talks about how physics students, after attending a lecture or reading a chapter in a textbook, are still unable to solve a problem set; however, after these students are given an example of how the theory is to be applied, then they are able to go ahead and solve the problem set quickly.