Skip to main content
Fachgebiete
Automobil + Motoren Bauwesen + Immobilien Business IT + Informatik Elektrotechnik + Elektronik Energie + Nachhaltigkeit Finance + Banking Management + Führung Marketing + Vertrieb Maschinenbau + Werkstoffe Versicherung + Risiko
DE
EN
Bücher
Zeitschriften
Themenseiten
Organisationspsychologie Projektmanagement Marketing Smart Manufacturing
Weitere Formate
Podcasts Webinare Technik Webinare Wirtschaft Kongresse Veranstaltungskalender Awards
Jetzt Einzelzugang starten
Zugang für Unternehmen
Referenzkunden
SLX-Digitalkonferenz: Zukunftswerkstatt 2024

Mehr Umsatz mit Top-Kontakten

Am 7. Mai erfahren Sie, wie Vertriebsteams Leadgenerierung, Leadnurturing und Leadstrategien effizient steuern können.
Erweiterte Suche
Anmelden
nach oben
Minds and Machines
Erschienen in: Minds and Machines 1/2019

04.01.2019

Epistemic Entitlements and the Practice of Computer Simulation

verfasst von: John Symons, Ramón Alvarado

Erschienen in: Minds and Machines | Ausgabe 1/2019

Einloggen

Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.

search-config
loading …

Abstract

What does it mean to trust the results of a computer simulation? This paper argues that trust in simulations should be grounded in empirical evidence, good engineering practice, and established theoretical principles. Without these constraints, computer simulation risks becoming little more than speculation. We argue against two prominent positions in the epistemology of computer simulation and defend a conservative view that emphasizes the difference between the norms governing scientific investigation and those governing ordinary epistemic practices.
Vorheriger Artikel Reproducibility and the Concept of Numerical Solution
Nächster Artikel Simulation, Epistemic Opacity, and ‘Envirotechnical Ignorance’ in Nuclear Crisis

Sie haben noch keine Lizenz? Dann Informieren Sie sich jetzt über unsere Produkte:

Springer Professional "Wirtschaft+Technik"

Online-Abonnement

Mit Springer Professional "Wirtschaft+Technik" erhalten Sie Zugriff auf:

  • über 102.000 Bücher
  • über 537 Zeitschriften

aus folgenden Fachgebieten:

  • Automobil + Motoren
  • Bauwesen + Immobilien
  • Business IT + Informatik
  • Elektrotechnik + Elektronik
  • Energie + Nachhaltigkeit
  • Finance + Banking
  • Management + Führung
  • Marketing + Vertrieb
  • Maschinenbau + Werkstoffe
  • Versicherung + Risiko

Jetzt Wissensvorsprung sichern!

Jetzt informieren

Springer Professional "Technik"

Online-Abonnement

Mit Springer Professional "Technik" erhalten Sie Zugriff auf:

  • über 67.000 Bücher
  • über 390 Zeitschriften

aus folgenden Fachgebieten:

  • Automobil + Motoren
  • Bauwesen + Immobilien
  • Business IT + Informatik
  • Elektrotechnik + Elektronik
  • Energie + Nachhaltigkeit
  • Maschinenbau + Werkstoffe




 

Jetzt Wissensvorsprung sichern!

Jetzt informieren

Springer Professional "Wirtschaft"

Online-Abonnement

Mit Springer Professional "Wirtschaft" erhalten Sie Zugriff auf:

  • über 67.000 Bücher
  • über 340 Zeitschriften

aus folgenden Fachgebieten:

  • Bauwesen + Immobilien
  • Business IT + Informatik
  • Finance + Banking
  • Management + Führung
  • Marketing + Vertrieb
  • Versicherung + Risiko




Jetzt Wissensvorsprung sichern!

Jetzt informieren
Fußnoten
1
See Kuhl et al. (2000) for an introduction to the general features of computer simulation in practice. For a historically important early introductory text see Naylor et al. (1966). These texts encompass the topics that we mean to capture using the term ‘computer simulation’.
 
2
The scientific role of mathematical models is a philosophically rich topic that falls beyond the scope of the present paper. For the most comprehensive discussion of the role of mathematical models in scientific reasoning see Pincock (2011). Pincock emphasizes that our confidence in computer simulations depends on our confidence in the prior mathematical model. He writes: “It is not always trivial to ensure that this has been done correctly, especially when computational or programming limitations force adjustments […] But if we have taken reasonable steps to ensure that [hardware failures or software bugs] did not occur, we can transfer the conclusion from the simulation back to the physical system” (2011, p. 80).
 
3
Some modeling (in geophysics for example) aims to explain static phenomena (modeling magnetic or gravitational fields). In other cases the use of closed-form equations may not need iterative solutions. While many of the points we discuss here also apply to non-dynamical cases, we focus on the dynamical cases here since they are most commonly discussed in the philosophical literature. We are grateful to an anonymous referee for reminding us to emphasize the static cases in which computer simulations are also used.
 
4
Philosophical treatments of the epistemology of computer simulations are in an early stage of development and work on this topic so far has been covered well by Eric Winsberg in his Stanford Encyclopedia of Philosophy article (2015).
 
5
In this paper we give particular attention to the first two questions. Thorough discussions of the two last questions can be found elsewhere (Symons and Horner 2014; Symons and Alvarado 2016; Alvarado and Humphreys 2017).
 
6
For example, policies regarding large-scale institutional interventions, life-critical systems assessments, existential risks, and the outcome of otherwise untested medical procedures would fall within this category.
 
7
See for example Barberrouse and Vorms (2014) and Beisbart (2017).
 
8
Saam (2017), for example, thinks that given the problem of epistemic opacity can be a serious obstacle to the reliability of simulation outcomes, only those computer simulations that are generated and treated as thought experiments (as opposed to laboratory experiments) should be considered in the social sciences. Simulations treated as experiments have at their core, significant issues concerning access or lack thereof to all the relevant steps of their processes (Humphreys 2009). Further, conventional error assessment strategies may be inadequate in contexts where software use is prominent (Symons and Horner 2014; Horner and Symons, forthcoming). In contrast, simulations that are treated merely as thought experiments, according to Saam, provide “well-founded answers to what-if-things-had-been-different questions” (Saam 2017, p. 81).
 
9
Notice that explicit theoretical or justificatory assumptions from the part of the first people to trust the oracle are not a necessary component of their reliance on the oracle. They could have begun to trust the oracles by default, accident or even superstition (See Skinner’s 1947 study “‘Superstition’ in the Pigeon”. In it pigeons would continue to perform behaviors they equivocally associated with food rewards solely because the behavior and the distribution coincided successfully in the past).
 
10
A position of this kind leads Quine to concede that he would have to count successful predictions from clairvoyants and telepaths as science (1973, see especially his 1990, pp 20–21). While genuine clairvoyance would be extremely interesting and useful, on the view that we defend here it would be a phenomenon that is largely orthogonal to science.
 
11
As helpfully noted by a reviewer of this paper, in the context of computer simulation, other secondary criteria are also at play for pragmatists. These include simplicity of the algorithm and greater unification of algorithmic calculations. However, these criteria will generally be subordinate to predictive power as a mark of success for the pragmatist.
 
12
We thank one of our anonymous referees for drawing our attention to this issue.
 
13
Thus, for example in cases where an applied mathematician uses techniques that require tacit knowledge or judgment developed through practice to reach some conclusion, we could recognize and rely on this fact in the same way we might recognize that someone who successfully rode a bicycle to work arrived to work without their being able to articulate explicitly, the details of how one rides a bike. The rider’s ability to ride a bicycle will not be entirely explainable by the rider, but the fact that he rode to work successfully can be regarded as undeniable and we can believe his claim to have made the journey by bicycle in virtue of knowing that he has the ability to ride a bicycle.
 
14
As we will see, even human expert credentials are not given out the same way we give out credentials to qualify as minimally reasonable interlocutors. I can trust my son, as Burge suggests, when he says that there is no more milk in the fridge based on a minimum intelligibility requirement and an assumption of healthy perceptual capacities. We would apply very different criteria in our choice of scientific experts.
 
15
The pragmatic approach seems to presume that science using computational methods is so fluid that the methods and instruments themselves can only be assessed in an ad hoc manner. The focus on reading simulations as sets of isolated measurement tasks misses, for example, the role of carefully curated data and principled theoretical background knowledge in the judgment of the expert interpreter.
 
16
An example of the pitfalls of atheoretical Big Data studies, is discussed by Lazer et al. (2014). Ultimately, many advocates of Big Data were misguided in spite of being able to point to well-established practices, successful predictive patterns and expertise within the data science community.
 
17
Burge’s account of the Acceptance Principle (as he acknowledges) is very similar in spirit to the Principle of Charity, as it figures in Quine (1960) and Davidson (1973). The principal difference between these principles is the role that Burge’s notion of preservation of content plays in his account.
 
18
In response to a referee comment, it should be mentioned that this emphasis on the apriori and the purely formal aspects of the target system contrasts sharply with the Materiality Thesis (Morgan 2005; Parker 2009), Morgan notes that a computer model is similar to its target system only in virtue of its form, while an experimental study that involves the target system itself is more likely to be generalizable in virtue of the material similarities between the object of the study and other instances of the kind (See also Roush 2015). Note that accepting the materiality thesis means that the devices by which information is manipulated bear significant epistemic import and that there are considerations other than the purely formal at stake. Thus, holders of the materiality thesis should be cautious with respect to the role of transparent conveyers in justification. (See also Barberousse et al. 2009).
 
19
Though they acknowledge that there is a substantial difference between computer assisted mathematical proofs, such as the ones Burge focused on and complex computer simulations in terms of content preservation, they justify the aprioricity of a scientist’s entitlement to trust a simulation in virtue of a second strategy, which we will inspect in detail in section three: trusting computer simulations, they argue, is like trusting expert testimony. Though expert testimony may be fallible, for this view, casting a general doubt on the practice absent specific reasonable doubt can be seen as irrational.
 
20
Similarly, when we rely on our senses we grant that when they are working the way they are supposed to they transmit information without altering it. That is, as explained above, they are transparent conveyers. Thus, though one can acknowledge their fallibility, in the absence of a plausible reason to doubt their well-functioning, it is rational to rely on the senses (Burge 1993).
 
21
As we will see below, whether one has reason not to doubt, no reason to doubt, or reason to trust represent significantly distinct challenges for this prerequisite.
 
22
Philosophy is another social practice that sets abnormally high epistemic standards. In our case, we aim high with respect to what should count as a rationally persuasive argument.
 
23
One can think of discretization as trying to approximate a circle by drawing one regular polygon after another with more sides each time starting form a square. Of course a square is a terrible circle, but a polygon with millions of sides may be visually indistinguishable for practical purposes. Nevertheless at each point, one is not drawing a continuous curve but rather a series of straight lines at an angle from each other.
 
24
McEvoy in his response to Tymoczko and Kitcher on the aprioricity of computer assisted mathematical proofs concedes as much by saying “What determines whether a proof is a priori is the type of inferential processes used to establish the conclusion of that proof. If the method of inference for any of the steps in the proof is a posteriori, it is a posteriori” (2008, p. 380).
 
25
This is especially the case now that memory has become so inexpensive in modern computing. The ‘true’ command in Unix, for example, which originally consisted of an empty file with nothing to execute grew to nearly 23,000 bytes from 1979 to 2012.
 
26
See Horner and Symons (forthcoming) for a review of the empirical literature on software error. They show that there has been a relatively consistent level of error reported in empirical studies from 1978 to 2018—for every 100 lines of code between 1 and 2 lines of code contain errors. (See also Symons and Horner 2017).
 
27
In saying this we are not discounting the serious time constraints faced by individual scientists in their careers e.g. the need to publish novel findings, the pressure of funding agencies, or the pursuit of tenure requirements. Rather, we are referring to scientific inquiry as a series of methods aimed at furnishing the best available understanding of our world. As such, scientist (from astrophysics to geology to biology- and some social sciences) can and ought to ensure a rigor in their methods that the nature of industry and war seldom afford.
 
28
It is unclear, for example, the extent to which Barberousse and Vorms assume a negative (entitlement to not doubt) or positive (entitlement to trust) role for entitlement in their arguments. However, this particular paragraph focuses on the completely distinct categories between kinds of entitlements (a kind of warrant that gives someone a reason to believe (x)) beyond their epistemic variety.
 
29
The five-sigma standard corresponds to a p value, or probability, of 3 × 10−7, or 1 in 3.5 million. In this case it is the probability that if the particle does not exist, the CERN team would find what they observed. It is extremely unlikely that they could have generated the data by accident. How unlikely? 1 in 3.5 million unlikely!
 
Literatur
Alvarado, R., & Humphreys, P. (2017). Big data, thick mediation, and representational opacity. New Literary History, 48(4), 729–749.CrossRef
Arkoudas, K., & Bringsjord, S. (2007). Computers, justification, and mathematical knowledge. Minds and Machines, 17(2), 185–202.CrossRef
Audi, R. (1997). The place of testimony in the fabric of knowledge and justification. American Philosophical Quarterly, 34(4), 405–422.
Barberousse, A., Franceschelli, S., & Imbert, C. (2009). Computer simulations as experiments. Synthese, 169(3), 557–574.MathSciNetCrossRef
Barberousse, A., & Vorms, M. (2014). About the warrants of computer-based empirical knowledge. Synthese, 191(15), 3595–3620.CrossRef
Beebee, H. (2001). Transfer of warrant, begging the question and semantic externalism. The Philosophical Quarterly, 51(204), 356–374.CrossRef
Beisbart, C. (2017). Advancing knowledge through computer simulations? A socratic exercise. In M. Resch, A. Kaminski, & P. Gehring (Eds.), The science and art of simulation I (pp. 153–174). Berlin: Springer.CrossRef
Boschetti, F., Fulton, E., Bradbury, R., & Symons, J. (2012). What is a model, why people don’t trust them and why they should. In M. R. Raupach (Ed.), Negotiating our future: Living scenarios for Australia to 2050 (pp. 107–118). Australian Academy of Science.
Burge, T. (1993). Content preservation. The Philosophical Review, 102(4), 457–488.CrossRef
Burge, T. (1998). Computer proof, apriori knowledge, and other minds: The sixth philosophical perspectives lecture. Noûs, 32(S12), 1–37.MathSciNetCrossRefMATH
Davidson, D. (1973). Radical interpretation. Dialectica, 27(3–4), 313–328.CrossRef
Davies, M. (2004) II—Martin Davies: Epistemic entitlement, warrant transmission and easy knowledge. In Aristotelian Society supplementary volume (Vol. 78(1)). Oxford: The Oxford University Press.
Dretske, F. (2000). Entitlement: Epistemic rights without epistemic duties? Philosophy and Phenomenological Research, 60(3), 591–606.CrossRef
Fresco, N., & Primiero, G. (2013). Miscomputation. Philosophy & Technology, 26(3), 253–272.CrossRef
Frigg, R., & Reiss, J. (2009). The philosophy of simulation: Hot new issues or same old stew? Synthese, 169(3), 593–613.MathSciNetCrossRef
Gramelsberger, G. (2011). Generation of evidence in simulation runs: Interlinking with models for predicting weather and climate change. Simulation & Gaming, 42(2), 212–224.CrossRef
Horner, J., & Symons, J. (2014). Reply to Angius and Primiero on software intensive science. Philosophy & Technology, 27(3), 491–494.CrossRef
Horner, J, & Symons, J. (forthcoming). Understanding error rates in software engineering: Conceptual, empirical, and experimental approaches.
Hubig, C, & Kaminski, A. (2017). Outlines of a pragmatic theory of truth and error in computer simulation. In M. Resch, A. Kaminski, & P. Gehring (Eds.), The science and art of simulation I (pp. 121–136). Cham: Springer.CrossRef
Humphreys, P. (2004). Extending ourselves: Computational science, empiricism, and scientific method. Oxford: Oxford University Press.CrossRef
Jenkins, C. S. (2007). Entitlement and rationality. Synthese, 157(1), 25–45.CrossRef
Kuhl, F., Dahmann, J., & Weatherly, R. (2000). Creating computer simulation systems: An introduction to the high level architecture. Upper Saddle River: Prentice Hall.MATH
Lackey, J. (1999). Testimonial knowledge and transmission. The Philosophical Quarterly, 49(197), 471–490.CrossRef
Lazer, D., Kennedy, R., King, G., et al. (2014). The parable of Google Flu: Traps in big data analysis. Science, 434, 343.
McGlynn, A. (2014). On Epistemic Alchemy. In D. Dodd, & E. Zardini (Eds.), Scepticism and Perceptual Justification. (pp. 173–189), OUP Oxford.
Moretti, L., & Piazza, T. (2013). When warrant transmits and when it doesn’t: Towards a general framework. Synthese, 190(13), 2481–2503.CrossRef
Morgan, M. S. (2005). Experiments versus models: New phenomena, inference and surprise. Journal of Economic Methodology, 12(2), 317–329.CrossRef
Naylor, T. H., Balintfy, J. L., Burdick, D. S., & Chu, K. (1966). Computer simulation techniques. New York: Wiley.MATH
Newman, J. (2015). Epistemic opacity, confirmation holism and technical debt: Computer simulation in the light of empirical software engineering. In International conference on history and philosophy of computing (pp. 256–272). Springer.
Nola, R., & Sankey, H. (2014). Theories of scientific method: An introduction. Abingdon: Routledge.CrossRef
Norton, S., & Suppe, F. (2001). Why atmospheric modeling is good science (pp. 67–105). Changing the atmosphere: Expert knowledge and environmental governance.
O’Neil, C. (2016). Weapons of math destruction: How big data increases inequality and threatens democracy. New York: Broadway Books.MATH
Oreskes, N. (2004). The scientific consensus on climate change. Science, 306(5702), 1686–1686.CrossRef
Oreskes, N., Shrader-Frechette, K., & Belitz, K. (1994). Verification, validation, and confirmation of numerical models in the earth sciences. Science, 263(5147), 641–646.CrossRef
Parker, W. S. (2009). Does matter really matter? Computer simulations, experiments, and materiality. Synthese, 169(3), 483–496.CrossRef
Pincock, C. (2011). Mathematics and scientific representation. Oxford: Oxford University Press.MATH
Pryor, J. (2012). When warrant transmits. In W. Crispin (Ed.), Mind, meaning, and knowledge: Themes from the philosophy of Crispin Wright (pp. 269–303). Oxford: University Press.CrossRef
Quine, W. V. (1973). The roots of reference. La Salle, Ill: Open Court.
Quine, W. (1960). Word and object. MIT press.
Resch, M. M., Kaminski, A., & Gehring, P. (Eds.). (2017). The science and art of simulation I: Exploring-understanding-knowing. Berlin: Springer.
Resnik, M. (1997). Mathematics as a science of patterns. New York: Oxford University Press.MATH
Roush, S. (2015). The epistemic superiority of experiment to simulation. Synthese, 169, 1–24.
Ruphy, S. (2011). Limits to modeling: Balancing ambition and outcome in astrophysics and cosmology. Simulation & Gaming, 42(2), 177–194.CrossRef
Ruphy, S. (2015). Computer simulations: A new mode of scientific inquiry? In S. O. Hansen (Ed.), The role of technology in science: Philosophical perspectives (pp. 131–148). Dordrecht: Springer.CrossRef
Saam, N. J. (2017). Understanding social science simulations: Distinguishing two categories of simulations. In M. Resch, A. Kaminski, & P. Gehring (Eds.), The science and art of simulation I (pp. 67–84). Cham: Springer.CrossRef
Steadman, I. (2013). Big data and the death of the theorist. Wired Online, 25, 2013.
Symons, J. (2008). Computational models of emergent properties. Minds and Machines, 18(4), 475–491.CrossRef
Symons, J., & Alvarado, R. (2016). Can we trust big data? Applying philosophy of science to software. Big Data & Society, 3(2), 2053951716664747.CrossRef
Symons, J., & Boschetti, F. (2013). How computational models predict the behavior of complex systems. Foundations of Science, 18(4), 809–821.CrossRef
Symons, J., & Horner, J. (2014). Software intensive science. Philosophy & Technology, 27(3), 461–477.CrossRef
Symons, J., & Horner, J. (2017). On some limits to model-based proof of software correctness. In T. Powers (Ed.), Philosophy and computing: Essays in epistemology, philosophy of mind, logic, and ethics. Berlin: Springer.
Tymoczko, T. (1979). The four-color problem and its philosophical significance. Journal of Philosophy, 76, 57–82.CrossRef
Vallor, S. (2017). AI and the automation of wisdom. In T. Powers (Ed.), Philosophy and computing: Essays in epistemology, philosophy of mind, logic, and ethics. Philosophical Studies Series (Vol. 128, pp. 161–178). Berlin: Springer.CrossRef
Williams, M. (2000). Dretske on epistemic entitlement. Philosophy and Phenomenological Research, 60(3), 607–612.CrossRef
Winsberg, E. (2010). Science in the age of computer simulation. Chicago: University of Chicago Press.CrossRef
Metadaten
Titel
Epistemic Entitlements and the Practice of Computer Simulation
verfasst von
John Symons
Ramón Alvarado
Publikationsdatum
04.01.2019
Verlag
Springer Netherlands
Erschienen in
Minds and Machines / Ausgabe 1/2019
Print ISSN: 0924-6495
Elektronische ISSN: 1572-8641
DOI
https://doi.org/10.1007/s11023-018-9487-0

Weitere Artikel der Ausgabe 1/2019

Minds and Machines 1/2019 Zur Ausgabe

OriginalPaper

A Minimalist Epistemology for Agent-Based Simulations in the Artificial Sciences

Letter to the Editor

The Civic Role of Online Service Providers

OriginalPaper

Reproducibility and the Concept of Numerical Solution

OriginalPaper

The Epistemic Importance of Technology in Computer Simulation and Machine Learning

OriginalPaper

Simulation, Epistemic Opacity, and ‘Envirotechnical Ignorance’ in Nuclear Crisis

OriginalPaper

Robotic Simulations, Simulations of Robots