Abstract
The previous chapter made a distinction between knowing and understanding. In computer simulations, this distinction allows us to set apart knowing when researchers trust the results and when they understand them. In this chapter, we explore different forms of understanding by means of using computer simulations. To this end, I have divided the chapter between epistemic functions that have a linguistic form, from those that are characterized for having a non-linguistic form. This distinction is meant to better help categorize the different ways in which researchers obtain understanding of the world that surround us by means of using computer simulations. Indeed, sometimes computer simulations open up the world to us in the form of symbols (e.g., by the use of mathematics, computer code, logic, numerical representation), whereas sometimes the world is accessed through visualizations and sounds. In the following, I analyze studies on scientific explanation, predictions, and exploratory strategies as linguistic forms that provide understanding of the world, and visualization as a case for non-linguistic forms.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
In some cases, we can expect an explanation by asking ‘how q’ questions. For instance, ‘how did the cat climb the tree?’ is a question that demands an explanation of how the cat managed to get up a tree. Although some theories of explanation put how-questions at their core, here we will only be interested in why-questions.
- 2.
The two major theories of explanation are ontic theories, where causality is at the center and epistemic theories, where derivation is at the center. The main advocates for the former are Salmon (1984), Woodward (2003), and Craver (2007). The main advocates for the latter are Hempel (1965), Friedman (1974), and Kitcher (1989). The reader interested in the many other theories of explanation should approach (Salmon 1989).
- 3.
Recall our discussion on computer simulations as problem-solving techniques in Sect. 1.1.1.
- 4.
In Sect. 6.2 I briefly discuss attempts to claim for causal relations in computer simulations, that is, whether researchers would be able to infer causal relations from computer simulations. This issue should not be confused with implementing a causal model, which is perfectly possible given the correct specification.
- 5.
Another important issue that speaks in favor of explanation by derivation is that results of computer simulations carry errors that causal theories are unable to account for (see Durán (2017b)).
- 6.
For an analysis of ‘brute’ and ‘independent’ facts, see Barnes (1994) and Fahrbach (2005).
- 7.
Incidentally, these same features render prediction inherently risky, since access to hidden information typically depends on shared standards in a given community, and is therefore potentially intersubjective. Now, one way to avoid problems of intersubjectivity is by demanding convergent predictions. That is, by means of having disparate areas of research all predicting towards similar results, our confidence in the prediction must inevitably increase. This is what philosopher Heather Douglas calls convergent objectivity (Douglas 2009, 120). We must keep in mind, however, that there are other forms of objectivity as well. In here, we are not going to worry about questions on subjectivity and objectivity, as they deserve a study of their own. For further references, the reader is referred to Daston and Galison (2007), Lloyd (1995), and of course (Douglas 2009).
- 8.
Clairaut was involved in several computations of the orbital period of the Halley’s Comet. For one of the first calculations, he divided the cometary orbit into three parts. First, from 0\(^{\circ }\) to 90\(^{\circ }\) of eccentric anomaly, the first quadrant of the ellipse from perihelion, which the comet took more than 7 years to traverse. Second, from 90\(^{\circ }\) to 270\(^{\circ }\), the superior half of the orbit, which the comet needed more than 60 years to traverse. Third, from 270\(^{\circ }\) to 360\(^{\circ }\), the final quadrant of the ellipse, which took about 7 years to traverse. Clairaut’s calculations were rectified in different moments throughout history (Wilson 1993).
- 9.
In epidemiology, the reproductive ratio—or reproductive number—represents the number of emerging cases that one infected individual generates on average over the course of an infectious period in an otherwise uninfected population. Thus, for \(R_0<1\), an infectious outbreak will die out in the long run, whereas for a \(R_0>1\) an infection will be able to spread out and infect the population. The larger the number, the harder it will be to control the epidemic. Thus, the metric helps to determine the speed that an infectious disease can spread through a healthy population. In the case of both simulations, the authors report that for large \(R_0\), the local epidemics become more widespread across all the layers of the population, making the population structure less and less relevant.
- 10.
This answer begs the question of what is ‘good agreement’ for results of computer simulations. Ajelli et al. do not specify how this notion should be understood. We could speculate that we have a good agreement when all results fall within a given distribution (e.g., a normal distribution). This, of course, requires specification. As we have learned from studies in the epistemology of experiment, agreement on the results of different techniques give confidence not only in the results, but also in the ability of the techniques to produce valid results (Franklin 1986, Chap. 6). The question is, then, to what extent could we consider the agent-based simulation and GLEaM two different techniques.
- 11.
By this I mean that the results of either computer simulation cannot, in principle, be empirically validated. Since each computer simulation implements sub-models, there is the possibility that some of them have been empirically validated.
- 12.
Exploratory strategies is one of those topics that attract little attention from philosophers, despite their centrality in the scientific and engineering endeavor. Fortunately, there are three excellent pieces of work that cover these issues. Friedrich Steinle on exploratory experiments (Steinle 1997), Axel Gelfert on exploratory models, (Gelfert 2016) and Viola Schiaffonati on exploratory computer simulations (Schiaffonati 2016). Here, I will take a different approach from these authors.
- 13.
This notion is used in the sense of independent of any researcher, instrument, method, or theory.
- 14.
See Brewer and Lambert (2001) and Van Helden (1974).
- 15.
The reader must be aware that there are many subtleties involved in the literature on ‘theory ladenness’ that we are not going to address here. For a good source of discussion, see Hanson (1958), and Kuhn (1962).
- 16.
These ideas can be found in García and Velasco (2013, 106).
- 17.
Ian Hacking offers a first approach to the different types and leves of theory involved in an experiment in Hacking (1992).
- 18.
This idea is discussed by Kenneth Waters in Waters (2007).
- 19.
More exploratory functions, such as starting points of scientific research, potential explanations, and other functions are discussed in (Gelfert 2016, 2018).
- 20.
There is the claim that the information that a computer simulation can provide is already contained in the models implemented. I find this claim particularly misleading for two reasons. First, because there are several cases where computer simulations produce emergent phenomena that was not strictly contained in the models implemented. Thus, the claim misinforms about the scopes of models as well as misrepresents the role of computer simulations. Second, because even if the implemented models contain all the information that the simulation is capable of offering, this fact says nothing about the knowledge that the researchers have. It is virtually impossible and pragmatically senseless to know the set of all solutions of a simulation model. Precisely for those cases, we have computer simulations.
- 21.
Recall our discussion in Chap. 4 about the reliability of computer simulations.
- 22.
Admittedly under this interpretation, many computer simulations become exploratory. I do not see this as a problem, for this characteristic fits well with the nature and uses of computer simulations. However, I do see philosophers of science objecting to my interpretation, primarily because it strips away the special status that is originally given to some kinds of experiments.
- 23.
Because we are solely interested in visualization in computer simulations, other kinds of visualizations, such as graphs, photographs, film videos, X-Rays, and MRI images, are excluded from our study.
- 24.
Perini explores this idea in Perini (2004, 2005).
- 25.
Two points to make here. First, the idea of a ‘mental cramp’ comes from Wittgenstein, who said that philosophical problems are compared to a mental cramp to be relieved or a knot in our thinking to be untied (Wittgenstein 1976). Second, there is not an ‘upside down’ world, since it is merely the way we—as humans—decided to represent it. As long as North and South are kept fixed, we are able to represent the globe the way we like—a good example of this is the logo of the United Nations. For an artistic representation of this, see the work of José Torres García, América Invertida, 1943.
- 26.
For a full video showing the development of the tornado, see http://avl.ncsa.illinois.edu/wp-content/uploads/2010/09/NCSA_F3_Tornado_H264_864.mov.
- 27.
The CAVE is a three by three by three meter pitch black room with five single-chip DLP projectors with a resolution of 1920 by 1200 pixels, each sending a respective image creating an accurate rendering for the human eye. Four cameras are installed at the corners of the ceiling for tracking the researchers inputs by the glasses and the mouse-wand.
- 28.
There are a few cases where the simulation is computed in real time during the AR session. The main concern with this kind of technology, however, is that it is too slow and time consuming.
- 29.
I thank Thomas Obst and Wolfgang Schotte at the HLRS for explaining to me the details of their interesting work.
- 30.
These issues could be overcome by computing and visualizing results in the AR environment in real time. Unfortunately, this kind of technology has high costs in terms of computing process, time and memory storage.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
Copyright information
© 2018 Springer Nature Switzerland AG
About this chapter
Cite this chapter
Durán, J.M. (2018). Epistemic Functions of Computer Simulations. In: Computer Simulations in Science and Engineering. The Frontiers Collection. Springer, Cham. https://doi.org/10.1007/978-3-319-90882-3_5
Download citation
DOI: https://doi.org/10.1007/978-3-319-90882-3_5
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-90880-9
Online ISBN: 978-3-319-90882-3
eBook Packages: Physics and AstronomyPhysics and Astronomy (R0)