On Thursday, May 9th, the Centre for Humanities Engaging Science and Society (CHESS) featured the first of a series of lectures highlighting ideas related to its broad concerns with "Knowledge, Culture and the Public Good". A committed group of academics and students heard a talk titled, “On the Purely Instrumental Value of True Belief” given by Torsten Wilholt, Professor of Philosophy and History of the Natural Sciences at Leibniz Universität, Hannover.
Professor Wilholt, in his paper “On the Purely Instrumental Value of True Belief”, focuses on the nature of true belief, its value and how it underlies our points of inquiry. As a principal concern, he postulates, it is important to engage with this as it characterises scientists aiming for the truth. This in-depth look into truth also engaged with other debates on the overriding agenda affecting research, centred on the autonomy of science and the freedom of research.Professor Wilholt considers a Bayesian model of the value of true belief and the instrumental character of the valuations involved in the analysis. This analysis involved the principle that if something is true one better believe it and conversely, if someone believes something, it must be true. In these calculations one’s belief in “x” was a value related to a set of desires or preferences while the value of “x” being true was representative of the state of the world. In the initial diagram of the possibilities of the value of “x”, Professor Wilholt establishes what he calls vertical preferences and horizontal preferences, with the vertical preferences being essential in demonstrating the desires that a rational account of the world be well-adapted to how the world is. In the further elaborations, the discussion leads to probability (Pr(x)) and how expectations as an outcome are inherent factors on the value of truth. The probable outcome of Pr(x), with the values of x being accepting, suspending judgement or rejecting, will all be determined by an accumulation of a certain amount of evidence. Applying these additional values of probability to the Bayesian model, he deduces an outcome he calls ‘VTBx’ whereby if inequalities in this calculation maintain, then the inquiry is an open one and the preferences would be characterized as vertical. He then advances that all that matters in methodological decisions are vertical preferences. In other words, what matters is that given a state of the world, we prefer to be right about it rather than wrong; but given a belief, we need not prefer the world to be such that it accords with our belief.
Wilholt continues by critically addressing objections and other debates around his argument. He concludes that objections, raised by Eran Tal a recent Toronto PhD, are actually contextual mistakes. Wilholt discussed an example concerning one’s desire to believe in the weather forecasts as true (independently of whether one desires the weather) because one wants the source of evidence to be reliable. In this example, the contextual mistake is made clear in that the reliability can only apply to a type of source-of-belief. In some cases, such as the weather forecast example, we do have a strong desire for it to be right, but this is different from being interested in the truth. And this again points to the earlier conclusion that for epistemic matters, all that matters are the vertical preferences. Wilholt further went on to address the work of Paul Horwich regarding the value of truth and his two propositions: firstly that knowledge is valuable for its own sake and secondly that without some assumptions, the difficulty of justifying pursuits of truth in obscure fields such as ancient history, fields that may not have a payoff. Discussion into this debate brings up other works on the nature of truth by Francis Bradley, William Ross and Jonathan Kvanvig.
Ultimately, Professor Wilholt acknowledges that this work, the analysis of the value and nature of truth is a problematic one for scientists. It will make scientists address what are commonly held or cherished beliefs such as: “the practical usefulness of scientific knowledge, the aesthetic pleasure of scientific insight, and the value of scientific explanation for our understanding of the world and out place in it.” He goes on to suggest that it is impossible to account for the intellectual practice of science in terms of intrinsic value, revealing the need for an accounting of instrumental value. In order to contribute to important debates regarding the aims and means of science today, philosophers of science should perhaps be tasked with studying the instrumental value of scientific knowledge more scientifically.
After the talk, Professor Wilholt was interviewed by a CHESS affiliate, addressing two key questions.
Q: What motivated your interest in this topic?
Wilholt: I wrote a book about the freedom of scientific research and I found that the idea of truth as something that is valuable for its own sake keeps surfacing in arguments for the autonomy of science. I wanted to separate what is right about this idea—that carrying out any kind of inquiry presupposes an esteem for the truth—from what is wrong and misleading—that the value of truth is intrinsic. From there this interest developed a life of its own.