Tools for Evaluating the Consequences of Prior Knowledge, but no Experiments. On the Role of Computer Simulations in Science

Eckhart Arnold

1 Introduction
2 Common features of simulations and experiments
3 Distinguishing features of experiments and simulations
4 Borderline cases
    4.1 Experimentum crucis and analog simulations
    4.2 Simulation-like experiments
    4.3 Experiment-like computer simulations
    4.4 Hybrid simulation-experiments
5 Summary and conclusion: Computer simulations as a tools for drawing conclusions from prior knowledge

4.3 Experiment-like computer simulations

The opposite case to simulation-like experiments is that of computer simulations that are experiment-like. The motivation for considering computer simulations as experiment-like is that some computer simulations resemble experiments in many different ways. As has been noted earlier (section 2) there are indeed many similarities between simulations and experiments and, naturally, these are strongly emphasized by that part of the philosophical literature on simulations that tends to liken simulations with experiments (Morrison 2009, Parker 2009, Winsberg 2010). Again, not all of these similarities may become relevant in every single case of a simulation. But as a matter of terminological convention we can call those simulation studies experiment-like which are conducted in a way that mimics experimental studies and which therefore show many of these similarities. This is often the case with simulations that are used as surrogates for experiments when it is either to costly or ethically unacceptable or for other reasons impossible to conduct a material experiment. The previously quoted example of H-tunnelling under conditions that hold in outer space is a good example for an experiment-like computer simulation, because it can be regarded as a substitute for an experiment to determine the H-tunnelling rate which is practically not feasible, because the tunnelling rates are to slow to be determined experimentally at temperatures which are as low as in outer space (Goumans/Kaestner 2010, p. 7351).

Experiment-likeness if understood in this sense describes a similarity of simulations to experimental procedures on the phenomenological level. It does not make a simulation any more empirical if it is experiment-like. Therefore, no matter how experiment-like some simulations are, there will remain some experiments (namely those, where it matters that they have an empirical content) that will never be fully resembled nor replaced by any simulation.

t g+ f @