You are here
Séminaire doctoral Philsci - Michal Hladky et François Papale
Michal Hladky (Université de Genève/MCMP Munich) – Mapping vs. Representational accounts of models and simulations
Scientific models and simulations are predominantly analysed as representational tools (Frigg and Nguyen 2017, 2016; Weisberg 2013; Giere 2010; Suárez 2010). Representational accounts expose how scientists think about models, this however does not mean that models have to be defined in terms of representation. Representationalism has recently been criticised as its ontological dimension undermines the epistemic one (de Oliveira 2018). A pragmatic approach with regard to models has been suggested.
In this paper, I argue for an alternative formal analysis of models. Contrary to the objections raised by Suárez (2003), I show that isomorphism should be used to define models. Furthermore, such definition allows to recover several aspects of representationalist analysis - the directionality and fallibility of models - by introducing an epistemic agent. Finally, mapping accounts fare better in explaining the reliability of inferences and the truth of conclusions based on models. These observations are supported by case studies (Markram et al. 2015) from the Blue Brain Project (BBP) and Human Brain Project (HBP).
1. Non-represented entities
The goal of BBP was to reconstruct the neocortical tissue of a rat with its dynamical properties. Based on simulations, it was hypothesised that the behaviour of the biological circuits depends on extracellular calcium. The justification of results relied on the assumption that the tissues were simulated correctly. However, the representation of calcium was not necessary for such simulations. This undermines the representationalist account of modelling. The mapping analysis provides a superior explanation of the success of the BBP.
Even if one accepts Suárez's (2010) directionality argument against isomorphisms in the analysis of scientific representation, it does not follow that it should be applied to models. Subsequently, it is not certain that models should be understood as representational entities.
There are two different reasons to reject the directionality argument based on scientific practice. HBP has several sub-projects in which brains are either modelled or they serve as models to develop alternative computer architectures.
If these techniques should succeed, the model relation should support the switch in directionality. The second observation comes from the building of computer systems that are intended to perform simulations of brains (BBP). In the construction phase, biological tissues serve as models for the construction of these systems. In the exploration phase, the relation is inverted. A notion of model based on isomorphism supports well this possibility. The directionality of perspective can be recovered when an epistemic agent is considered.
3. Epistemic justification and fallibility
There is a strong contrast between justification provided by representation and by isomorphism. Without qualification, anything can represent anything else. If model relations are analysed in terms of representation, the justification they provide is low. On the other hand, if an isomorphism holds between a source and a target, reliable inferences about the target can be drawn. The fallibility of modelling can be explained, as required by representationalists, by false beliefs of agents about the source-target relations. Such an explanation is compatible with mapping definition of models.
Frigg, Roman, and James Nguyen. 2016. “Scientific Representation.” In The Stanford Encyclopedia of Philosophy, edited by Edward N. Zalta, Winter 2016. Metaphysics Research Lab, Stanford University. https://plato.stanford.edu/archives/win2016/entries/scientific-representation/.
———. 2017. “Models and Representation.” In Springer Handbook of Model-Based Science, edited by Lorenzo Magnani and Tommaso Bertolotti, 25–48. Dordrecht Heidelberg London New York: Springer.
Giere, Ronald. 2010. “An Agent-Based Conception of Models and Scientific Representation.” Synthese 172 (2): 269–281.
Markram, Henry, Eilif Muller, Srikanth Ramaswamy, Michael W. Reimann, Marwan Abdellah, Carlos Aguado Sanchez, Anastasia Ailamaki, et al. 2015. “Reconstruction and Simulation of Neocortical Microcircuitry.” Cell 163 (2): 456–92. https://doi.org/10.1016/j.cell.2015.09.029.
Oliveira, Guilherme Sanches de. 2018. “Representationalism Is a Dead End.” Synthese, November. https://doi.org/10.1007/s11229-018-01995-9.
Suárez, Mauricio. 2003. “Scientific Representation: Against Similarity and Isomorphism.” International Studies in the Philosophy of Science 17 (3): 225–244.
———. 2010. “Scientific Representation.” Philosophy Compass 5 (1): 91–101.
Weisberg, Michael. 2013. Simulation and Similarity: Using Models to Understand the World (Oxford Studies in Philosophy of Science). Oxford University Press. http://www.oxfordscholarship.com/view/10.1093/acprof:oso/9780199933662.001.0001/acprof-9780199933662.
François Papale (Université de Montréal) – Redefining Units of Selections as Networks of Interactions – An Ontological Inquiry
The use of network-based models in evolutionary biology becomes more pervasive by the day. This methodological transition has important consequences, among which the need to review our ontological understanding of evolutionary individuals. In this presentation, I argue that individuals, in the context of Darwinian explanations, should be conceived as integrated networks of interactions to be described by their degree of integration. This view will be contrasted with Godfrey-Smith’s Darwinian individual framework (Godfrey-Smith 2009).
Godfrey-Smith defines Darwinian individuals, the building blocks of Darwinian populations, as genealogical entities that can be isolated more or less straightforwardly. This means that a great diversity of entities and processes can be viewed as reproducers and reproduction, respectively: “That is fine, as long as we know who came from whom, and roughly where one begins and another ends.” (Godfrey-Smith 2009, 86). In order to provide finer grained descriptions of reproducers, Godfrey-Smith organizes them into three types: scaffolded, simple and collective. The distinction between single and collective reproducers on one side and scaffolded reproducers on the other is that the formers have “the machinery of reproduction internal to [them]” (Godfrey-Smtih 2009, 88). The distinction between single reproducers and collective ones is that the latters are composed of lineages with evolutionary fates potentially independent from that of the whole.
Network analyses of evolutionary dynamics provide a different picture of biological entities: those that do reproduce are phylogenetic mosaics composed of parts that have distinct evolutionary fates; moreover, their reproduction is made possible by complex networks of interactions involving entities at various levels of organization (Bapteste et Huneman 2018). At best, then, all reproducers could be described as scaffolded collectivities. However, even this readjustment is problematic. In the presentation, a network-based analysis of cases that are considered paradigmatic reproducers (genes, organisms like us, prokaryotic cells) will show that two out of the three criteria (bottleneck, germ line, integration) provided by Godfrey-Smith for describing collective reproducers are maladapted to a Darwinian perspective.
Given these limitations, I argue for an alternative framework inspired by the work of various authors (Bouchard 2010; Brandon 1988; Dupré and O’Malley 2009; Millstein 2009): biological entities should be conceived as interactive networks whose degree of functional integration determines whether they can be considered units of selection or not. This degree can be assessed through rate of interactions within the studied biological object, whose boundaries can be drawn where the said rate drops significantly. This definition emphasizes that biological objects are collectivities and provides a more accurate reading of the part they play in evolutionary dynamics.
Bapteste, Eric, et Philippe Huneman. 2018. « Towards a Dynamic Interaction Network of Life to Unify and Expand the Evolutionary Theory ». BMC Biology 16 (1). https://doi.org/10.1186/s12915-018-0531-6.
Bouchard, Frédéric. 2010. « Symbiosis, Lateral Function Transfer and the (Many) Saplings of Life ». Biology & Philosophy 25 (4): 623‑41.
Brandon, R.N. 1988. « Levels of Selection: A Hierarchy of Interactors ». Dans The Role of Behaviour in Evolution, Plotkin, H.C., 51‑71. Cambridge, Massachussets: MIT Press.
Dupré, John, et Maureen O’Malley. 2009. « Varieties of Living Things: Life at the Intersection of Lineage and Metabolism ». Philosophy et Theory in Biology Volume 1 (décembre): 1‑25.
Godfrey-Smith, Peter. 2009. Darwinian Populations and Natural Selection. New York: Oxford University Press.
Millstein, Roberta L. 2009. « Populations as individuals ». Biological Theory 4 (3): 267–273.