Learning from Non-Causal Models
- PDF / 838,065 Bytes
- 21 Pages / 439.37 x 666.142 pts Page_size
- 41 Downloads / 198 Views
Learning from Non‑Causal Models Francesco Nappo1 Received: 11 August 2019 / Accepted: 5 August 2020 © Springer Nature B.V. 2020
Abstract This paper defends the thesis of learning from non-causal models: viz. that the study of some model can prompt justified changes in one’s confidence in empirical hypotheses about a real-world target in the absence of any known or predicted similarity between model and target with regards to their causal features. Recognizing that we can learn from non-causal models matters not only to our understanding of past scientific achievements, but also to contemporary debates in the philosophy of science. At one end of the philosophical spectrum, my thesis undermines the views of those who, like Cartwright (Erkenntnis 70:45–58, 2009), follow Hesse (Models and Analogies in Science, Notre Dame, University of Notre Dame Press, 1963) in restricting the possibility of learning from models to only those situations where a model identifies some causal factors present in the target. At the other end of the spectrum, my thesis also helps undermine some extremely permissive positions, e.g., Grüne-Yanoff’s (Erkenntnis 70(1):81–99, 2009, Philos Sci 80(5): 850–861, 2013) claim that learning from a model is possible even in the absence of any similarity at all between model and target. The thesis that we can learn from non-causal models offers a cautious middle ground between these two extremes.
1 Introduction One of the central questions in the epistemology of scientific modelling is the ‘problem of learning’, i.e., the question of when and under what conditions the information gathered from a model may, if not license new inferences, at least justify changes in the degree of credence assigned to empirical hypotheses about real-world
* Francesco Nappo [email protected]; [email protected] 1
UNC Chapel Hill, Caldwell Hall CB#3125, 240 East Cameron Ave., Chapel Hill, NC 27599‑3125, USA
13
Vol.:(0123456789)
F. Nappo
targets (see, e.g., Hesse 1963; Steel 2007; Cartwright 2009; Grüne-Yanoff 2009, 2013; Knuutila 2009; Mäki 2009; Sugden 2009; Fumagalli 2015, 2016; Pietsch 2019).1 Two positions on this problem emerge from the contemporary literature. On the one hand, some authors defend a restrictive position, according to which the study of a model can prompt learning about a target only if there exist similarities between model and target with regards to their causal features, i.e., with regards to the kinds of causal connections that obtain in each domain. For instance, tests on mice can justify increasing our confidence regarding the effectiveness of drug X in humans, according to this view, only if we can expect humans to share with mice biological features (e.g., similar immune systems) that are causally connected to mice’s observed response to drug X. Advocates of this view include Hesse (1963), Sugden (2000, 2009), Cartwright (2009), Steel (2007) and Pietsch (2019). On the other hand, some authors defend a permissive position, according to which the use of some scientific models—so-called ‘
Data Loading...
