Model-Based Reasoning and Thomas Kuhn
Thomas Nickles has argued,
[R]ather like Polanyi (1958, 1966), Kuhn suggested that the scientific community operates surprisingly like a medieval guild:
(1) It is a community of practitioners who posses expert knowledge.
(2) The community sharply distinguishes itself from the nonexpert, lay public, including other expert scientific communities. Boundaries are maintained by the high costs of admission and expulsion, enforced by professors, journal editors, peer reviewers, and other "gatekeepers."
(3) There is a standard training procedure for novices in a given specialty area. They are trained on the same problems, using the same or similar textbooks and laboratory exercises. At advanced stages, the training typically involves something akin to a master-apprentice relation.
(4) The knowledge is imparted by example far more than by rule.
(5) Hence, the crucial knowledge that distinguishes an expert from a well-read novice remains largely tacit, inarticulate, and more knowing-how that knowing-that. It involves teaching by showing and knowing by doing.
(6) Strong personal commitment to the imparted tradition is expected. Being too critical of community presuppositions and practices threatens both the community and one's own career prospects. (146)
A guild-like community of science points to some of the following aspects of model-based reasoning:
- Science is fundamentally a set of shared activities instead of a body of knowledge. Models occur as social practices.
- One has to learn the framework of observation, the accepted methods, and to internalize the skills and procedures necessary to carry out the field's aims.
- Training stresses an emphasis on working problems built around exemplars with concrete solutions.
- Problems are only problems within a given framework; puzzles operate within a set of rules.
- To not observe these approaches is to be outside the status group.
- Contrast sets are essential to the disciplines. One is especially learning to see family resemblances across problems.
- The nature of the model always assures something like a bounded search for what is otherwise unknown.
- The search is often one of an "open texture" (Friedrich Waismann), that is not everything can be verified or even completely defined in advance; concepts cannot foresee all the possible ways they may be used or tested.
- The normal process of research is one of refinement via multiple cycles of further enrichment.
- Conceptual testing relies on "graded structures" (Eleanor Rosch) in which some examples possess better family resemblances than others. This results in taxonomies.
- The taxonomic practice has incommensurable paradigms embedded in it.
- However, new taxonomies often share aspects of the older taxonomies; therefore, comparison can take place, though evaluation of the others always occurs within a particular paradigm.
- Failure of the current practice/model to solve problems opens up the chance for a rival practice to supplant it, but the new model has to offer research problems and possibilities for it to be adopted.
- A counter-model is only taken seriously when the current one can no longer answer a large nexus of problems.
Discussion Questions
- Is it accurate to compare the scientific process to joining a guild? Why or why not?
- How has Kuhn appropriated the late Wittgenstein's insights into family resemblances?
- What distinguishes Kuhn's model of science from the standard one?
- Can a model be separated from the practice that gave rise to it?
- What distinguishes an anomaly from a counter-instance?
- Are different taxonomies incommensurable?
- Based on the above model, how might Kuhn's argument explain the failure of the Intelligent Design movement to gain acceptance in the larger scientific community?
Nickles, Thomas. "Normal Science: From Logic to Case-Based and Model-Based Reasoning." Thomas Kuhn ed. Thomas Nickles. Cambridge: Cambridge UP, 2003. 142-177.