PhilSci Archive

The logic of scientific debate: Epistemological quality control practices and Bayesian inference – a neoPopperian perspective

Skoyles, John (2008) The logic of scientific debate: Epistemological quality control practices and Bayesian inference – a neoPopperian perspective. [Preprint]

[img]
Preview
PDF
Download (190Kb) | Preview

    Abstract

    Science is about evaluation, persuasion and logic. In scientific debate, scientists collectively evaluate theories by persuading each other in regard to epistemological qualities such as deduction and fact. There is, however, a flaw intrinsic to evaluation-by-persuasion: an individual can attempt and even succeed in persuading others by asserting that their reasoning is logical when it is not. This is a problem since, from an epistemological perspective, it is not always transparent nor obvious when a persuasive assertion is actually deductively warranted. Empirical research upon reasoning, indeed, supports the notion that assertions are often strongly persuasive for reasons other than their logic. The unreliability of the link between persuasion and logic raises an important methodological issue: how do scientists debate in a manner such that claimed but false “logical” arguments are ignored, and only warranted arguments get to determine theory preference? This need for soundness in debate is a particularly important epistemological concern in cases where the deductive qualities of persuasive argument are not overt, and so cannot be directly checked --such as when they are founded upon Bayesianism probabilistic coherence. The argument presented here is that scientists make the qualities of probabilistic and nonprobabilistic inference sound (and so warranted) through how they organize their debate. Scientists, I argue, abide by “epistemological quality control practices” that limit the persuasive power of unsound arguments upon theory evaluation. Examples of such debate quality control practices are publicness, clarity, openness to criticism, and the collective promotion of attempts at theory conjecture and refutation. Methodologically, these quality control practices are extralogical since they do not directly provide scientific inferences with additional logical warrantedness. They function instead in science to generate an epistemological evaluative environment in which persuasiveness is due, and only due, to logic (i.e. sound). Their methodological role is therefore to make what is warranted de papyri (in our principles of rationality – epistemologically competence) also what is persuasive in evaluation and debate de cognitio (in our exercise of such principles – epistemologically performance). Several limitations exist upon the soundness of Bayesian inferential coherence – surreptitious revision, logical omniscience, uncertain evidence, old evidence, and new hypotheses. Bayesianism, as a result, can only exist if it is pursued in a debate that is regulated by quality control practices over its inferences (for instance, practices that ensure there is autonomy of inference, diligence of inference, probity of evidence, auditability of inference and assiduousness of conjecture). In this context, one can reinterpret Popper’s concern with criticism, openness, refutation and conjecture, as deriving not (as he thought) directly from the needs of logic, but, indirectly, from the need of scientists to create epistemological soundness.


    Export/Citation:EndNote | BibTeX | Dublin Core | ASCII/Text Citation (Chicago) | HTML Citation | OpenURL
    Social Networking:

    Item Type: Preprint
    Keywords: Bayesianism, persausion, scientific debate, theory evaluation, methodology, soundness, robustness, clarity, criticism, conjecture, refutation, rationality, Popper.
    Subjects: Specific Sciences > Probability/Statistics
    General Issues > Confirmation/Induction
    Specific Sciences > Psychology/Psychiatry
    General Issues > Science vs. Pseudoscience
    General Issues > Theory Change
    General Issues > Rhetoric of Science
    General Issues > Values In Science
    Depositing User: John Skoyles
    Date Deposited: 02 Nov 2008
    Last Modified: 07 Oct 2010 11:17
    Item ID: 4230
    URI: http://philsci-archive.pitt.edu/id/eprint/4230

    Actions (login required)

    View Item

    Document Downloads