How to Tell when Simpler, More Unified, or Less Ad Hoc Theories Provide More Accurate Predictions

This page was last edited on 05/04/02 by Malcolm R Forster

  PDF files

Note: You need Adobe Acrobat Reader 3.0, or later, to read and print this article.  It is free.

This Version (287 KB) is the a PDF file, with the correct page numbers, etc.  It is as good as a reprint.

Publication Data

Forster, Malcolm R. and Elliott Sober (1994): "How to Tell when Simpler, More Unified, or Less A Hoc Theories Provide More Accurate Predictions." British Journal for the Philosophy of Science 45: 1-35.


 Traditional analyses of the curve fitting problem maintain that the data do not indicate what form the fitted curve should take.  Rather, this issue is said to be settled by prior probabilities, by simplicity, or by a background theory.  In this paper, we describe a result due to Akaike [1973], which shows how the data can underwrite an inference concerning the curve’s form based on an estimate of how predictively accurate it will be. We argue that this approach throws light on the theoretical virtues of parsimoniousness, unification, and non ad hocness, on the dispute about Bayesianism, and on empiricism and scientific realism.

Table of Contents

  1. Introduction

  2. Akaike without Tears

  3. Unification As a Scientific Goal

  4. Causal Modeling

  5. The Problem of Ad Hocness

  6. The Sub-Family Problem

  7. The Bearing on Bayesianism

  8. Empiricism and Realism

  9. Appendix A: The Assumptions Behind Akaike’s Theorem

  10. Appendix B: A Proof of a Special Case of Akaike’s Theorem