Models, measures and messages: an essay on the role for induction

• Published in: Community ecology Vol. 3; no. 2; pp. 191 - 204
• Main Author: Dale, M.B
• Format: Journal Article
• Language: English
• Published: Akadémiai Kiadà 2002; Akadémiai Kiadó
• Subjects: Causality; community ecology; Dimensionality; Ecological modeling; Entropy; Falsification; Inductive reasoning; Maximum likelihood estimation; Model classes; Model selection; Modeling; Null hypothesis; plant ecology; Prediction; Series convergence; Valorisation; Vegetation
• ISSN: 1585-8553; 1588-2756
• DOI: 10.1556/ComEc.3.2002.2.6

In this essay, I examine the role of induction in developing vegetation models. Falsification is a necessary component of model building but is not itself sufficient. Induction provides a necessary complement and one that dethrones the null hypothesis from its privileged state. After examining the role of description and environment, I examine several possible criteria useful for valorising models so that we may choose the 'best'.These criteria include fit, simplicity, precision and interest. Predictability, which is given overwhelming importance in a falsification approach, is found to be ambiguous. It may be obtained by using multiple models without regard to the processes active in the real system. In addition movement towards a model which does reflect the 'real' processes can result in loss of predictivity. Finally, some comments are made on what we can infer and how this relates to our understanding of living systems.