Abstract
Floridi’s chapter on relevant information bridges the analysis of “being informed” with the analysis of knowledge as “relevant information that is accounted for” by analysing subjective or epistemic relevance in terms of the questions that an agent might ask in certain circumstances. In this paper, I scrutinise this analysis, identify a number of problems with it, and finally propose an improvement. By way of epilogue, I offer some more general remarks on the relation between (bounded) rationality, the need to ask the right questions, and the ability to ask the right questions.
Similar content being viewed by others
Notes
Unless explicitly mentioned, page-numbers refer to Floridi’s “The Philosophy of Information.”
One of the ambiguities that are removed by replacing (BC) with (BC∃) is the status of the letters i, …, l. In the original version, they could both be understood as constants and as variables. In the revised version they are clearly variables of a multi-sorted language.
Non-circular in the sense of not being defined in terms of the avoidance of the unwanted inequality.
Keep in mind that these notions only make sense in a context, at a particular LoA.
The “might” qualification is essential since the interest of a in i 1 and i 2 is captured by subsets of \({{\mathsf Q}_1}\) and \({{\mathsf Q}_2}\).
I assume here that (PR1max) is the obvious way of fixing Floridi’s probabilistic versions, just like I assumed that (BC∃) was the intended reading of (BC).
Atomic pieces of information shouldn’t be understood in the same way as atomic propositions, as this would exclude basic disjunctive information.
The question of how we should compute the sum of multiple R(i, q)’s (for dependent i’s) will come back in a different guise in the next section.
The underlying intuitive principle is that if i answers the more encompassing question, it surely also answers the less encompassing one.
This counterexample relies on the fact that relevance has different features when it applies to declarative information than when it applies to questions: A question can be less relevant (in the sense of being a worse question) than some of its sub-questions, but a piece of information is always at least as relevant as any of its parts.
Because it is unrelated to my final point, I’m deliberately ignoring the further dependence between A(i, q) and A(i, q′), which obviously should also be taken into account.
Granted, the assumption that the model concerns ideally rational agents need not follow from the description of rationality Floridi gives (p. 264), but the reliance on a probabilistic model surely pulls in that direction.
If \(\Upgamma\) is a set of declarative premises, we say that a question \(Q = ?\{A_1, \ldots, A_n\}\) is evoked by \(\Upgamma\) iff (i) \(\Upgamma \vdash_{CL} A_1 \vee \ldots \vee A_n,\) while (ii) for each A i we have \(\Upgamma \not\vdash A_i.\)
See Fitelson (2008) for this type of diagnosis in the context of evidential support and confirmation.
Such a looser connection between logical principles and probabilities need not block my arguments from “Two defences” section, for there I only relied on logical connections between questions and sets (or conjunctions) of questions. These are principles that belong to standard classical logic; not to its erotetic extension.
References
De Clercq, K., & Verhoeven, L. (2004). Sieving out relevant and efficient questions. Logique et analyse, 47(185–188), 189–216.
Fitelson, B. (2008). Goodman’s “new riddle”. Journal of Philosophical Logic, 37(6), 613–643.
Gabbay, D., & Woods, J. (2003). Agenda relevance, in ‘Agenda relevance. A study in formal pragmatics’. Elsevier, Amsterdam, pp. 155–193.
Hawthorne, J., & Bovens, L. (1999). The preface, the lottery, and the logic of belief. Mind, 108(430), 241–264.
Wiśniewski, A. (1995). The posing of questions: Logical foundations of erotetic inferences. Dordrecht, Boston: Kluwer Academic Publishers.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Allo, P. Relevant Information and Relevant Questions: Comment on Floridi’s “Understanding Epistemic Relevance”. Minds & Machines 24, 71–83 (2014). https://doi.org/10.1007/s11023-013-9325-3
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11023-013-9325-3