Papers from the Colloquium

Theme 4: Evidence and Context

Respondent: Philip Dawid
Department of Statistical Science, University College London

I would like to start by outlining recent developments relating to Evidence in the specific context of University College London.

I have spent most of the last year putting together an interdisciplinary programme on "Evidence, Inference and Enquiry", bidding for a programme grant from the Leverhulme Trust (now augmented by ESRC). We have now heard that this has been successful, and are about to set the programme in motion.

When I first circulated colleagues around UCL for expressions of interest, I was overwhelmed with responses, from about 25 different departments. The final programme involves 11 departments (in 7 faculties): Statistical Science, Crime Science, Geography, Science and Technology Studies, Economics, Hebrew and Jewish Studies, Primary Care and Population Studies, Psychology, Computer Science, Educational and Professional Development, and Laws. Many others will be involved informally.

It has been fascinating to me to have this opportunity to meet and interact with people I would not normally have much contact with, and to have my eyes opened to the wide range of approaches and assumptions that different disciplines bring to intellectual enquiry. In particular, it has been very educational -- and challenging -- to try to come to some appreciation of the differences between the approaches of the "two cultures", Science and Humanities. If I can be allowed an obviously over-simplified generalization, I would say that the sciences struggle for unification, while the humanities celebrate diversity. However, there need not be conflict between these approaches, and I would hope that the interdisciplinary nature of our Evidence Programme might be captured through the slogan "Unity in Diversity".

Many participants in the Evidence Programme rightly wish to consider the special features of the construal and uses of evidence in their particular disciplines. What I personally want to do is to emphasise common logical features and principles of thinking about evidence, which can be applied irrespective of the underlying field of discourse, assumptions and conclusions. Just as Aristotelian logic can be applied correctly to ridiculous premises, or badly but nevertheless to a true conclusion, we should to be able to identify and study the logical and methodological principles that underlie valid analysis of evidence, over and above variations in its subject matter and specific applications. With this in mind, I was particularly interested in Jason Davies' extract from Herodotus. Here we have a rare (perhaps unique?) example of the experimental method applied to historical research. As Jason says, we would nowadays regard the whole enterprise as ludicrous, but I can't help but be impressed at what -- within its own terms -- seems to be a carefully conceived and entirely appropriate logical approach to the gathering and interpretation of evidence to throw light on a hypothesis of interest.

Jason has pointed out the importance of hypotheses in relation to the interpretation of evidence. Indeed one could say that without explicit consideration of hypotheses we have no right to speak of "evidence" at all, but merely have "data". This is perhaps the most fundamental principle of "subject-blind" evidence.

Jason also argues for the importance of narrative -- and, at a still broader level, "discourse" -- in making sense of evidence. This I find more questionable. It seems to me that there are many important applications where narrative is irrelevant, or at any rate takes a back seat -- for example, in the statistical design and interpretation of clinical trials to determine whether a newly formulated pharmaceutical preparation works better at curing patients than the standard medical treatment. And even if we agree that it is human nature to wish for a coherent narrative and achieve "closure" in our understandings of evidence, that still leaves open the question whether this is a mere psychological peculiarity of the human brain, or whether it has a more fundamental philosophical and methodological place.

This general question of the role of narrative, both in general and in particular, is one of a number of themes that we hope to explore within the Evidence Programme. Watch this space!