- This event has passed.
Phyllis Illari (Senior Lecturer, University College London, UK), Why do we need evidence of mechanisms?
17 December 2019 | 11 h 00 min - 12 h 30 min
Phyllis Illari is a Senior Lecturer at University College London (Dept of Science & Technology Studies; Faculty of Maths & Physical Sciences) in the UK. She works in the philosophy of science, with particular expertise in causality, mechanisms and information.
I will present a view of evidence of mechanisms as evidence of the activities, entities, their organization, and the phenomena they explain, using the idea of ‘minimal mechanism’ (Glennan and Illari, 2018). I will argue that this view allows us to theoretically organize an incredibly diverse array of forms of evidence and empirical practices. I will then home in on a specific way in which evidence of mechanism is crucial, arguing that it is important even for solid evidence of correlation. Any clinical study, even a well-conducted RCT, which is still one of our best methods of establishing a reliable correlation, needs decent answers to two questions: (i) what are the variables for disease, treatment and outcome? and (ii) how and when are they measured and why?
I will use the case of ‘vitamin D deficiency’ to show how important these questions are, even when they are not explicitly addressed in published results, because they are regarded as sufficiently standardised to be unimportant. Until recently, vitamin D deficiency was regarded as well-understood, reliably measurable in standardised ways, and linked to diseases such as rickets by well understood mechanisms. However, recent research has linked vitamin D deficiency to other diseases, in ways that expose the fact that different measuring techniques measure slightly different forms of Vitamin D. Those differences are now relevant.
Considering the case shows how deeply integrative our evidential pluralism needs to be, and therefore how complex our practices of reasoning about evidence are. Philosophical accounts need to be responsive to this.