“Better Theories”

Formal, pragmatic, aesthetic, and empirical virtues

Theory
Sociology
Author

andrés castro araújo

Published

January 11, 2024

The new editors of Theory & Society have a strange new statement of goals with some minor platitudes about what makes for a good theory—e.g., “general, abstract, and amenable to empirical investigation.”

But most of the statement is quite superficial in content and angry in tone.

So, what actually makes for a good “scientific theory”?

I’m just going to copy-paste this list of good principles:

FORMAL VIRTUES

  • Testability

    The theory should express commitments about the world that in principle can be confirmed or falsified on the basis of empirical evidence.

  • Internal Coherence

    The theory should be coherent and should not contain contradictions.

PRAGMATIC VIRTUES

  • Fertility

    The theory should suggest new and exciting avenues of research. It should generate new research questions faster than it can answer them.

  • Conservatism

    The theory should retain crucial bits of what came before and not break too quickly or too dramatically with tradition without a compelling reason for doing so.

AESTHETIC VIRTUES

  • Simplicity

    The theory should posit only those entities, properties, causal relations, etc. that are necessary to account for the phenomenon.

  • Elegance

    The theory should be compact and graceful.

EMPIRICAL VIRTUES

  • Empirical Adequacy

    The theory should accommodate or fail to conflict with well-established phenomena.

  • Prediction

    The theory should make accurate predictions and retrodictions, particularly concerning phenomena that would be surprising were the theory to be false.

  • Explanation

    The theory should explain the phenomena in its domain either by showing how they follow from general laws of nature or by showing how they are produced, given rise to, or maintained by mechanisms.

  • External Coherence

    The theory should be supported by (or at least consistent with) other well-accepted non-rival theories.

  • Generality

    The theory should apply to more phenomena than its rivals.

  • Unification

    The theory should unify diverse phenomena by showing them to be instances of a common pattern.

From: Craver and Darden (2013, pp. 83–84)

Note. Craver and Darden (2013) advocate for a mechanism-based view of scientific explanation which nowadays has superseded the “covering law” approach to scientific explanation (i.e., 20th century positivism). Mechanism-based explanations have been heavily endorsed by “analytical sociology” (e.g., Hedström and Swedberg 1996; Hedstrom 2005); unfortunately, most sociologists see this literature as being exclusively committed to rational-choice theory and agent-based modeling. Read the book!

References

Craver, Carl F., and Lindley Darden. 2013. In Search of Mechanisms: Discoveries Across the Life Sciences. University of Chicago Press.
Hedstrom, Peter. 2005. Dissecting the Social: On the Principles of Analytical Sociology. Cambridge: Cambridge University Press.
Hedström, Peter, and Richard Swedberg. 1996. Social Mechanisms.” Acta Sociologica 39(3): 281–308.