Recently, Andrei Mirovan suggested to me an interesting paper by Danny Frederick called “Truth Cannot Sensibly Be Our Epistemic Aim”. (Link) Danny passed away about a year ago, so as a way to commemorate him and his work, I’d like to get people talking about this paper and debating what he had to say.
Danny argues that ‘our epistemic aim is neither… truth… nor… avoiding falsity.’ (p. 2) And that what we seek is really just ‘better theories’ where ‘better theory’ means one or more of the following:
- ‘contradict no observation statement’
- ‘give illuminating deep and simple explanations of a wide range of phenomena’
- ‘resolve many problems’
- ‘generate surprising predictions that survive empirical testing.’ (p. 2)
Popper would agree with Danny that we should never assume our theories are ‘true’ because how would we know? Nor did Popper believe our theories can be ‘more probably true’ in the sense of probability calculus because no matter how good the theory is, there is always the chance it will be overthrown by a better theory and there is no way to predict if or when that will happen. So it makes no sense to declare a theory ‘probably true.’
Danny’s Disagreement with Popper’s Theory of Verisimilitude
But Danny shows that Popper did believe we could get nearer to the truth; Popper’s called this verisimilitude. Popper spoke of how we were ‘always searching for a true theory…’ and how ‘we may have good reasons… for thinking that… we have progressed towards the truth.’ (p. 3)
Danny rejects this aspect of Popper’s epistemology and claims that we can no more know that we are ‘nearer the truth’ than we can know we have arrived at the truth.
‘[T]here is [in Danny’s theory] no pretension that [better] explanations… according to [the above] indicators are either true or probably true or approximate true or even nearer the truth than their worse-scoring rivals.’ (p. 3)
Why does Danny reject Popper’s notion of verisimilitude? Danny gives two reasons.
- ‘It seems silly to say that truth is our aim when we can have no indication that we have got the truth or even that we are approaching it.’ (p. 3) Or put another way, how can you aim at truth when you can’t recognize it? (link)
- ‘It may be claimed that a theory that has greater explanatory merits than another theory is closer to the truth than the other theory. Closeness to truth, it may be said, is what generates relative explanatory success. … But that claim seems to be refuted. According to Aristotelian Ptomemaic theory we inhabit a closed world; according to Newtonian theory, we inhabit an infinite universe; according to relativity theory, we inhabit a closed world. If Newton’s theory got us closer to the truth than its predecessor, its successor seems to have taken us further from the truth than Newton’s theory.’ (p. 3)
If we have no means by which to test how often we get to the truth while at the same time our ‘best theories’ reverse things we thought were true–only to reverse them again later–then Danny believes we can’t have good reasons for believing we are getting nearer to the truth and thus must give up that aspect of Popper’s epistemology.
Then What Is Our Epistemic Aim?
What does Danny want to replace verisimilitude with?
Danny suggests we don’t need ‘truth’ to be our ‘aim’ and we can instead simply aim for ‘better theories’ (as defined above) directly. We aim for theories that have more of the objective characteristics that we prefer theories to have, such as not having known observational counter examples. Danny claims that the fact that we prefer such ‘better theories’ alone explains our preference for such theories and saying that they get us closer to the truth adds nothing. 
This is the part of Danny’s theory I personally find the least convincing. When a skeptical epistemology like this tells me “you should prefer better rational theories for their own sake, not because they help you find the truth’ I can’t help but feel that we’re just sneaking an epistemic aim for truth in the backdoor. Surely the very reason we started preferring such ‘better’ rational theories is that we thought we had a good explanation for why such indicators (as explained above) help us reach our goal of the truth.
Sam Kuypers, for example, points out that the act of correcting mistakes seems to logically imply a move towards the truth in that you removed a falsehood from your theory. Therefore, we have an explanation that says we have, at least in some aspects, moved towards the truth. I find this a compelling argument at odds with Danny’s theory.
But Danny quickly points out that we can’t actually compare our theories with the world, only with our perceptions of the world. Thus we can’t know if we actually corrected an error and moved toward the truth or not. Therefore ‘error correction’ can’t be an objective measure by which we judge theories to be better or worse. This implies that greater explanatory power or fewer counter observations do not logically imply that the theory is truer.
What Is the Epistemic Aim of a Murder Trial?
Danny’s overall point can be illustrated well with this example:
Take a court case to determine the guilt or innocence of someone accused of murder. In court, the prosecution will attempt to present a theory explaining how and why the accused committed the murder that is consistent with the evidence that is available. In addition, the prosecution will try to use the known evidence to eliminate competing theories. By this means, the prosecution hopes to show that their theory is the ‘best theory’—or in other words that their theory is the most consistent with the available observations (evidence). The defense will attempt to demonstrate problems with the prosecution’s theory in hopes of demonstrating that some other theory is to be preferred.
So far, all of this is consistent with Popper’s theory. We are using evidence (observations) to test between competing theories and only if the best theory is that the accused did in fact commit the murder will we put them in jail.
But Why Do We Prefer Best Theories?
Now here is where things get interesting.
Danny essentially asks “why do we prefer best theories?”
If the answer to this question is “because best theories are closer to the truth” then we can see that we have an immediate problem. One is either actually guilty of a murder or they are not. There is no reasonable sense in which a ‘best theory’ about the guilt or innocence of the accused is ‘closer to the truth’ or not.
But we also know that Popper and his successors claim that best theories are not more probably true either. So, if that is correct, we can’t claim that what makes a theory about a murder the ‘best’ theory is that it’s more likely to be correct.
So whatever else we might say of the nature of our ‘best theory’ about the murder, it apparently can’t be that it’s nearer to the truth (i.e. has greater verisimilitude) nor that it is more probably true. So then what does it mean for it to be a ‘best theory’?
If you buy that a best theory about a murder case is neither truer nor more probably true, then does it even have a relationship with the truth? If it doesn’t, then Danny is correct: the truth can’t be our epistemic aim.
Justice For All
Of course, this also explains why I feel Danny’s explanation can’t be correct either. What is the aim of a murder trial if not to get the right person into jail? So at a minimum, we do aim for epistemic truth, whether sensibly or not.
Defenders of Danny’s theory will likely claim that is just a perception. We don’t really put the guilty person in jail, we put the person that seems guilty to us in jail. Danny wasn’t claiming we don’t think we aim for truth, only that we can’t sensibly aim for truth.
But that is where I think he is wrong. Ask yourself why we have things like standards of evidence or court procedures? Do we have them merely because we have a preference for a certain kind of evidence? Or do we have them because those standards really and truly do help us get the right guy into jail?
Let’s make this concrete. Take the discovery of fingerprints or the widespread use of video cameras as evidence. Did adding fingerprints and widespread camera footage only seem to help us get the right person in jail or did it actually help us get the right guy in jail? If you answered (as I do) that it actually helped us get the right guy into jail then not only are we aiming for epistemic truth, but we are sensibly doing so.
The Excitement of a New Problem
But what am I basing that judgment on?
It’s not like I can ask the gods to tell me how often we got the right guy in jail before and after the introduction of fingerprints and the widespread use of video cameras. Am I basing my judgment on anything more than my gut? Or is there some good explanation for my judgment?
Andrei issued a challenge to me: “Can you… give a non-circular explanation of why we should cherish truth [as an epistemic aim] when we don’t even seem to have an acceptable idea about how to find it?”
I admit that, for the example of our murder mystery, this proves difficult. If I can neither claim verisimilitude nor increased probability, then what is the relationship between a best theory and truth?
Popperians Do Not Agree on How to Resolve This Problem
I’ve contacted several fellow Popperians and asked them to weigh in on Danny’s problem and how it applies to a murder trial. Based on the diverse responses, I can see that we Popperians do not currently agree on how to go about addressing the problem Danny raises. So, it seems Danny has found a very real epistemological problem with Popperian epistemology that needs to be resolved.
What makes this even more interesting is that our main epistemological competitor—Bayesians—have no problem answering Andrei’s question. (Though Popperians would claim they do so wrongly.) Bayesians would claim that while we can’t know for sure that we put the right guy in jail for a murder, we can increase the probability that we did so by making sure we pick the theory most consistent with the evidence. Indeed, pretty much any non-Popperian would give a similar answer: we seek the best theory in a court case because it increases the chance we put the right person in jail. This seems so self-evidently obvious to most people that they can’t imagine why such an answer is problematic. It’s only Popperians that claim this is an incorrect answer.
It’s exciting to identify an interesting epistemological problem like this. The growth of knowledge comes from solving such problems, so I look forward to seeing other people’s attempts to resolve this one.
 An interesting side note is that Danny’s epistemology, based on Popper, seems similar—if not identical—to the epistemology of Kuhn which has usually been understood as one of Popper’s main opponents. Kuhn explains it this way:
“But need there be any such goal [in the progress of science]? Can we not account for both science’s existence and its success in terms of evolution from the community’s state of knowledge at any given time? Does it really help to imagine that there is someone full, objective, true account of nature and that the proper measure of scientific achievement is the extent to which it brings us closer to that ultimate goal? If we can learn to substitute evolution-from-what-we-do-know for evolution-toward-what-we-wish-to-know, a number of vexing problems may vanish in the process. Somewhere in this maze, for example, must lie the problem of induction.” (The Structure of Scientific Revolutions, p. 171)