From paleontology to molecular biology and back: when stochasticity and contingency overcome disciplinary differences
Résumé
In the 1970s Gould challenged the classical paradigm that evolution occurs gradually under the main guidance of natural selection. In his famous 1989 book titled Wonderful Life he emphasizes that the evolution of life is fundamentally contingent. His thought experiment is well-known: if we could replay a tape of life, the outcomes (i.e. the historical present) would be always different. His work as a paleontologist has intrigued some philosophers such as Beatty (2006) that was the first (followed by Turner 2012) to bring the attention that Gould’s idea of contingency actually referred to two distinct notions. During the presentation, I will highlight the fact that these two notions, that Beatty calls “unpredictability” and “causal dependence, are extremely similar to two other concepts of stochasticity that are used in molecular biology to explain a very different kinds of process that takes place in cell, that is to say mutually exclusive alternative splicing.
In gene expression, alternative splicing occurs after transcription and just before translation. In this process, the same mRNA can be cut and pasted in different ways and thus translated into different proteins. Alternative splicing is conceptually problematic since biologists have been struggling for years to understand how it is possible to have up to 18 000 variables from a single RNA. Most of them agree that this high level of variability is possible since the process is stochastic. But any agreement is reached in specifying what it means by “stochastic”. In trying to clarify the matter, the biologists Hiesinger and Hassan (2018) propose a definition of stochasticity that is quite similar to those made by Beatty (2006) on contingency.
The central question I will ask during the presentation is the following: can we help to clarify the fuzzy meaning of stochasticity in the alternative splicing process? My answer is yes, we can. With the conceptual tools provided by Beatty (2006) and Turner (2012), my proposal is to characterize stochasticity in alternative splicing such as an “insufficiency of causal dependence” (ICD). ICD refers to the fact that knowing certain starting point parameters could be not sufficient to understand how the process behaves and which outcomes can give arise. This definition does not only help to clarify the specific stochastic characterization of splicing but also to propose an original and general definition of stochasticity that can enrich the philosophical debate. Indeed, today, the literature on stochasticity oscillates between a metaphysical characterization (chance is something uncaused) and cognitive limitation (chance underlines our inability to explain and describe certain processes). The contribution of my ICD is to propose a third way: stochasticity in splicing is related to the simple fact that knowing a certain causal dependence is (sometimes) insufficient for explaining the performance of some processes and their outcomes. ICD could be extensible beyond splicing because a definition of stochasticity free from metaphysics and which cannot be deflated to our cognitive limitation could be a good way to rethink original relationships between chance and living been.