The probabilism debate that never was?
| Author | Kiel Brennan-Marquez |
| DOI | 10.1177/1365712718816748 |
| Published date | 01 April 2019 |
| Date | 01 April 2019 |
The International Journal of
Evidence & Proof
The probabilism debate that
2019, Vol. 23(1-2) 141–146
ª The Author(s) 2019
never was?
Article reuse guidelines:
sagepub.com/journals-permissions
DOI: 10.1177/1365712718816748
journals.sagepub.com/home/epj
Kiel Brennan-Marquez
University of Connecticut School of Law, Hartford, USA
Abstract
This paper responds to Ronald Allen and Michael Pardo’s essay, “Relative Plausibility and its
Critics.” In short, Profs. Allen and Pardo succeed on one front but fail on another. They succeed
in demonstrating that “probabilism” is not a conceptually or phenomenologically convincing
model of how human fact-finders in the real world carry out their task. They fail, however, in
demonstrating that probabilism is inadequate as an idealized model of human fact-finding – and to
that extent, their analysis falls short of establishing the abstract primacy of “explanationism,”
even if it clarifies the latter’s content and (limited) utility.
Keywords
epistemology, probability, jurisprudence
Humans often reason under conditions of uncertainty, in law no less than everyday life. A central ambition
of evidence theory—drawing on epistemology—is to formalise the reasoning process. Should the model be
quantitative or qualitative? When people draw conclusions about what is happening (or has previously
happened) in the world, should we conceptualise the result as a numerical assertion of likelihood, akin to
Vegas odds, or instead as a claim about the relative explanatory power of competing hypotheses?
Debate roils on, as it has for decades, sharp minds on both sides.1 And in some cases, the answer
might even make a functional difference.2 At some level, however, the most intriguing part of this debate
is that it exists at all. Why, after so long, have the two camps been unable to reach consensus—at least
about the nature of their disagreement, if not about the underlying question?
1. The citation possibilities here, needless to say, are vast. Professors Allen and Pardo’s piece itself is an exemplar of the
‘qualitative school’. See Allen and Pardo (2019). Another— framed in slightly different terms (distinguishing probability from
‘likelihood’, rather than ‘relative plausibility’—belongs to Sean Sullivan (Sullivan, 2019). For some excellent recent efforts
from ‘quantitative school’, see, for example, Nance (2016); Cheng (2013).
2. See, e.g. Allen and Pardo (2019: 7) (‘It is important to resolve this disagreement, if possible, because only with a clear
understanding of what the system is actually doing can the process be effectively critiqued. Once it is clear what the system is
doing, one can decide whether and how it can be improved’).
Corresponding author:
Kiel Brennan-Marquez, University of Connecticut School of Law, 55 Elizabeth St, Hartford, CT 06105, USA.
E-mail: kiel.brennan-marquez@uconn.edu
142
The International Journal of Evidence & Proof 23(1-2)
One possibility is that further back-and-forth is needed. It takes time to boil questions down to their
essence, and for stable battle-lines to emerge. Perhaps that moment has yet to arrive.
But another possibility—the one I explore here—is that the two sides are, in an important sense,
talking past one another. Professors Allen and Pardo are right, it seems to me, that the quantitative
approach (‘probabilism’) fails to track phenomenology. No Bayesian updating model, of either an
objective or a subjective variety, can hope to capture how most human fact-finders, in most cases, arrive
at conclusions about the truth or falsity of factual propositions.3 But this observation, powerful as it is,
only shuts the casket on probabilism if the relevant object of inquiry is factfinding-as-it-actually-plays-
out-in-most-instances (‘actual factfinding’). If, on the other hand, the relevant object of scrutiny is
factfinding-as-we-aspire-to-have-it-play-out (‘aspirational factfinding’), probabilism may have a point.
It is possible, in other words, that an updating model offers a better ideal picture of factfinding, even if
it does a poor job describing how human fact-finders in the real world actually find facts. And if that is
true, it is likewise possible that the real...
To continue reading
Request your trial