The Evidence of Narrative

Been reading Miriam Shleifer McCormick's book on epistemic responsibility the past few days, and really enjoying it. One of her main contentions is: (PB) it is possible, and rational, to form a belief for purely practical, non-evidential reasons in certain circumstances.

PB echoes William James's classic claim (against Clifford and the evidentialists) that "Our passional nature not only lawfully may, but must, decide an option between propositions whenever it is a genuine option that cannot by its nature be decided on intellectual grounds." 

According to McCormick, there are really just three conditions that define circumstances in which the formation of practical beliefs are acceptable: (a) the evidence on the matter is silent or neutral, (b) the beliefs in question would contribute to the agent's sense of meaning in her life, and (c) the formation of the belief does not rely on practices that undermine truth. McCormick thinks these conditions define acceptable circumstances in which to form practical beliefs because:

  1. The norms of belief are ultimately grounded in human flourishing (and connected with truth insofar as true beliefs are generally what our proper functioning belief-forming mechanisms will produce)
  2. Meaning-making practical beliefs can contribute to human flourishing without undermining the truth
  3. Therefore, if formed in the right circumstances, we have no normative basis on which to repudiate practical beliefs

Something like that.

I agree with much of what McCormick is up to in her project. However, I want to consider a possibility that she quickly (too quickly in my mind) passes over. That is: that meaning-making beliefs are truth-conducive in a way that can go beyond the consideration of evidence available at a particular time. Consider the following case:

The Night Of: Nasir Khan has been accused of murder, and the evidence is damning. He was captured on film with the victim hours before her death, he admitted to taking drugs with her at her house just before the murder, he was found with a bloody knife (the victim was stabbed) just blocks away from the crime scene. And yet, he insists that he's innocent. After deliberating for a few days, the jury cannot come to a consensus and is finally dismissed. When asked why they failed to convict, one of the jurors who held out on the possibility of Khan's innocence puts it this way: "All the evidence was there, but something didn't seem right. Why would a kid with so much potential, and no real motivation to kill that young lady, just go berserk and randomly start stabbing her? It didn't make sense to me. He had to be framed or something."

Here, we have what I would think of as a "meaning-making" belief, but not one that is potentially justified in terms of its role in the flourishing of an individual or community. The belief seems to serve as a framework for interpreting the available evidence, rather than as the sum of all the individual bits of evidence. And this is where I think the real challenge for evidentialism lies.

When it comes to belief formation, we have at least two separable categories to consider. We have the data -- what's often called the evidence -- and the framework that makes that data intelligible (both as data, and as data potentially supporting one or another particular conclusion). Now, a lot of folks would like to see the evidential relation as going just one way -- [data --> state of affairs] -- but this ignores the crucial fact that the plausibility of the state of affairs in question, the conclusion to one's deliberation, can itself influence what appears to a subject to be data in the first place. That is to say: evidentialism within a paradigm or framework may well be true, but the foundations of that paradigm or framework are themselves in need of supporting reasons that we cannot conceive of as evidence -- at least so long as our conception of evidence depends on the paradigm or framework in question.

So what could possibly fill the gap here?

This is where narrative comes in. Narratives are complex, highly structured descriptions of events that are typically unified in terms of the exercise of individual or collective agency. They make sense of -- or render intelligible -- otherwise disparate phenomena, unifying them into a single whole with reference to a core framework that is able to explain the various constituent parts.

So narratives are not evidentially neutral, or simple "meaning-making" mechanisms, that can contribute to our overall wellbeing but are otherwise unrelated to the truth. It's not that we are simply story-telling creatures by accident, and that our flourishing depends on narratives in the same way that it depends on having spices available capable of exciting our tastebuds in various ways. Rather: our identity as believers depends crucially on our ability to "make-sense" of the world around us, to organize the disparate elements of experience into data, and to sort that data into evidence, that all of this is possible only by drawing on our capacity to narrate our (individual and collective) life.

Hopefully, it'll become clear where this general sort of view will (and where it will not) dovetail with McCormick's. On the one hand, I agree with her central claim, (PB), but this is not because I see practical reasons for belief as separable from (but normatively unified with) evidential reasons. It's because I see evidential reasons as ultimately dependent on practical reasons. The main cost of my amendment is that it radicalizes and otherwise merely odd view. But the benefit of my account is enormous: it allows the rationally permissible practical beliefs to be normatively unified with evidential beliefs at a deeper level, one that does not require us to give up their truth-conducivity, at least when considered more globally.