Article

Biology and Philosophy

, Volume 6, Issue 2, pp 155-180

First online:

Must scientific diagrams be eliminable? The case of path analysis

  • James R. GriesemerAffiliated withDepartment of Philosophy, Program in History and Philosophy of ScienceCenter for Population Biology, University of California

Rent the article at a discount

Rent now

* Final gross prices may vary according to local VAT.

Get Access

Abstract

Scientists use a variety of modes of representation in their work, but philosophers have studied mainly sentences expressing propositions. I ask whether diagrams are mere conveniences in expressing propositions or whether they are a distinct, ineliminable mode of representation in scientific texts. The case of path analysis, a statistical method for quantitatively assessing the relative degree of causal determination of variation as expressed in a causal path diagram, is discussed. Path analysis presents a worst case for arguments against eliminability since path diagrams are usually presumed to be mathematically or logically “equivalent” in an important sense to sets of linear path equations. I argue that path diagrams are strongly generative, i.e., that they add analytical power to path analysis beyond what is supplied by linear equations, and therefore that they are ineliminable in a strong scientific sense.

Key words

Path analysis regression scientific diagrams Sewall Wright statistics in biology