The argument from surprise

Abstract

I develop an account of productive surprise as an epistemic virtue of scientific investigations which does not turn on psychology alone. On my account, a scientific investigation is potentially productively surprising when (1) results can conflict with epistemic expectations, (2) those expectations pertain to a wide set of subjects. I argue that there are two sources of such surprise in science. One source, often identified with experiments, involves bringing our theoretical ideas in contact with new empirical observations. Another, often identified with simulations, involves articulating and bringing together different parts of our knowledge. Both experiments and simulations, then, can surprise.

Keywords:

• Simulation
• model
• surprise
• experiment

Acknowledgements

This paper has benefited greatly from comments by Andrew Buskell, Alice Murphy, Emily Parke, David Colaco, Arnon Levy, Mark Migotti, Marta Halina, Kirsten Walsh and Mihail Cernea. It was presented at the University of Calgary Philosophy of Biology Research Group, at the University of Washington philosophy department, and the Simplification & Distortion as Scientific Strategy workshop at the IRH in Bucharest – I am grateful of the feedback provided. Some of the research for this publication was made possible through the support of a grant from Templeton World Charity Foundation. The opinions expressed in this publication are those of the author(s) and do not necessarily reflect the views of Templeton World Charity Foundation. I’d also like to thank two anonymous referees from this journal.

Notes

1. One well-travelled philosophical path where surprise matters is discussion of novel facts. However, as this is geared towards theory-confirmation and issues of realism it is tangential. The same is true of discussions of ‘anomalies’ in, for instance, Kuhn’s work.

2. There is a growing body of work comparing the epistemic status of simulations with experiments. Many philosophers approach the difference in ontological terms. Perhaps experimental subjects are materially continuous with their targets, while simulations and their targets are ‘made of different stuff’ – and perhaps this makes a difference to the kinds of epistemic tasks they can perform (see Morgan 2002, 2003, 2005; Guala 2002, 2005; Parker 2009; Winsberg 2010; Parke 2014). Others, such as Parker (2008) and Winsberg (2003, 2009) compare the two via their epistemic capacities. Although this is similar to my approach, neither discusses surprise.

3. There are several discussions in the philosophy of modelling which I avoid here. One important question concerns the metaphysical nature of models and their relationship to the world (Godfrey-Smith 2009; Frigg 2010; Levy 2012; Weisberg 2012). Another question focuses on the potential roles models can play: building better theories (Wimsatt 1987), providing explanation (Weisberg 2007), or traction on complex systems (Mitchell 2002, 2003). A final question asks after the relationship between models and other scientific tools, such as whether models are autonomous of theory (Morgan and Morrison 1999; Winsberg 2010). I will restrict myself to this last kind of question.

4. An area where philosophical attention has focused on the epistemic potential of simulations is in the philosophy of climate science. See, for instance, Parker (2009, 2011), Epstein and Forber (2013), Werndl and Steele (2013), Lloyd (2009, 2010). These papers have a narrower focus than mine, and I expect much of what I say is complimentary.

5. See, for instance, the papers collected in Klein (2011).

6. The effects on body size against locomotor performance depend upon whether an activity is power limited, such as jumping, or force limited, such as standing. As muscle power increases roughly proportionately to muscle mass, increases in size will be matched by increases in power. However, muscle force is (roughly) proportional to muscle area – and the ratio of muscle-area to body size decreases as size increases. The result is that some activities, such as walking, become increasingly trickier at larger sizes.

7. Sellers et al.’s (2013) own account is gratifyingly clear, and I refer the reader to them.

8. For a fuller defence of the account implied here, see Currie and Levy (under review).

9. In other contexts, we may want to distinguish between experiments and more ‘passive’ observations (Currie and Levy under review). However, this is unnecessary for the purpose of distinguishing between experiments and simulations in terms of surprise.

10. Mark Migotti suggested the term ‘productive’.

11. Thanks to Arnon Levy for help with notions of surprise.

12. As such, we might want to put some restrictions on how surprising a result ought to be, but I leave that discussion for future work.

13. Of course, sometimes surprising experimental results to turn out to be artifacts – due to the experimental design itself – here I am referring to successful experiments.

14. I am being deliberately ambiguous between externalist and internalist readings, epistemically speaking. That is, we could read my notion of surprise internally, referring to scientific representations alone, or externally – including facts or truth. My account is amenable to a range of such readings, and I don’t want (or need) to arbitrate between epistemic internalism or externalism here.

15. Thanks to David Colaco for pressing me on this.

16. It is worth pointing out that my sense of ‘validity’ (and, I suspect, many philosophers’) is much wider than that meant by simulationists. Where scientists often refer to validity testing as a stage in an investigation which involves comparing their simulation’s behaviour (or some components’ behaviour) to the world, I mean any aspect which provides epistemic links between object and target.

17. Although in my view establishing external validity works differently in experiments and simulations (Currie and Levy under review).

18. This is a somewhat unusual understanding of ‘epistemic virtue’, and I don’t mean for it to be tied to virtue epistemology (see Greco and Turri 2015) necessarily. Roughly, I consider a scientific epistemic virtue to (1) be a property of an investigation (as opposed to a hypothesis, evidence or agent) which (2) possession of provides epistemic goods, or makes the provision of such goods likely.