Many philosophers consider teleology, explanations in terms of purpose, necessary for describing nature. Whereas most biologists discuss the purposes of organisms only to interpret their results. We believe that practical utility will allow teleology to step out of the shadows in science.
Our core idea is to fit empirical data in a setup where control objectives represent system purposes. Goodness-of-fit analyses of the data then evaluate these purposes. This allows objective tests of what a system is trying to do, what aims it is prioritizing, what functions it is implementing, etc. Beyond asking “how” by fitting mechanistic models, scientists then can ask “why” by fitting control schemes.
We make innovative use of Model Predictive Control (MPC), which uses a model of the system and its environment to predict control outcomes. Practically all mechanistic models used in science can be adapted for this. The core technical development of our project is hence to re-tool MPC to enable scientific purpose fitting.
We will study platelets; blood cells that have the simple purpose of forming thrombi to stop bleeding. Yet platelet behavior offers layers of complexity, from a binary choice between pro-aggregatory and pro-coagulant phenotypes to thrombus structure being shaped by communication through soluble mediators and gap junctions. Our experiments will scaffold the MPC development, produce scientific advances, and attract peer attention to the new method.
The project will be reflected in philosophical analysis. Are scores delivered by this computational procedure really all there is to purpose? The strains put on an ancient concept by a novel scientific method provide an excellent opportunity for philosophical progress.
To maximize adoption, we will publish open access with all software tools released open source and run workshops and public outreach. Our long-term ambition is simply to make purpose as “real” in biology as fields are in physics.
