Justification Of Induction By Inference To Lesser Coincidence
Download Justification Of Induction By Inference To Lesser Coincidence PDF/ePub or read online books in Mobi eBooks. Click Download or Read Online button to get Justification Of Induction By Inference To Lesser Coincidence book now. This website allows unlimited access to, at the time of writing, more than 1.5 million titles, including hundreds of thousands of titles in various foreign languages.
Justification of Induction by Inference to Lesser Coincidence
Author: Daniel Jonathan Elstein
language: en
Publisher: Stanford University
Release Date: 2011
I begin by identifying David Hume's problem of induction. Hume argues that induction cannot be justified by a priori reasoning, because the failure of induction does not imply contradiction, or by a posteriori reasoning, because reasoning that the unobserved will resemble the observed based on observation would be circular. Hume concludes that induction cannot be justified by any reasoning. The principle that nature is uniform cannot be established without assuming that nature is uniform. But many paradigmatic instances of induction can be justified in terms of something weaker than the principle that nature is uniform, namely a form of reasoning I call "inference to lesser coincidence". This form of reasoning is meant to incorporate traditional formulations of the justification of induction expressed in terms of inference to the best explanation, statistical sampling, and Bayesian reasoning. My version of the argument is as follows: The conditional, time-invariant proposition that vast regularities in progress are likely to continue somewhat further is either true or false. If false, then the regularities we have observed are colossally coincidental. If true, they are far less coincidental. Therefore the proposition is probably true. If, in fact, vast regularities in progress are likely to continue, this has application to specific cases, such as the possibility that the Sun will rise again. I respond to three objections, which claim that time-restricted laws lessen the coincidence of observed regularities without making it likely that the Sun will rise again, that the "sample" of observed events might be biased, and that a zero prior probability assignment for dependence might be justified. I conclude by discussing the meaning of 'cause'.
Justification of Induction by Inference to Lesser Coincidence
I begin by identifying David Hume's problem of induction. Hume argues that induction cannot be justified by a priori reasoning, because the failure of induction does not imply contradiction, or by a posteriori reasoning, because reasoning that the unobserved will resemble the observed based on observation would be circular. Hume concludes that induction cannot be justified by any reasoning. The principle that nature is uniform cannot be established without assuming that nature is uniform. But many paradigmatic instances of induction can be justified in terms of something weaker than the principle that nature is uniform, namely a form of reasoning I call "inference to lesser coincidence". This form of reasoning is meant to incorporate traditional formulations of the justification of induction expressed in terms of inference to the best explanation, statistical sampling, and Bayesian reasoning. My version of the argument is as follows: The conditional, time-invariant proposition that vast regularities in progress are likely to continue somewhat further is either true or false. If false, then the regularities we have observed are colossally coincidental. If true, they are far less coincidental. Therefore the proposition is probably true. If, in fact, vast regularities in progress are likely to continue, this has application to specific cases, such as the possibility that the Sun will rise again. I respond to three objections, which claim that time-restricted laws lessen the coincidence of observed regularities without making it likely that the Sun will rise again, that the "sample" of observed events might be biased, and that a zero prior probability assignment for dependence might be justified. I conclude by discussing the meaning of 'cause'.
Optimality Justifications
Author: Gerhard Schurz
language: en
Publisher: Oxford University Press
Release Date: 2024-01-24
The leading idea of epistemology in the Enlightenment tradition was foundation-theoretic: to reach knowledge, we must not legitimize our beliefs by external authorities, but instead justify them by rational arguments. Recently, the foundation-theoretic ideal of justification has come under attack, the chief criticism being that universal standards of justification are illusory because the problem of a regress of justification is unsolvable. Alternatives to foundation theory (coherentism, externalism, or dogmatism) have been developed that give up central claims of Enlightenment epistemology such as empirical support, cognitive accessibility, or rational justifiability. Optimality Justifications develops a new account of foundation-theoretic epistemology based on the method of optimality justifications. Optimality justifications offer a solution to the regress problem. Rather than striving for a priori demonstrations of reliability, which are impossible, they show that certain epistemic methods are optimal with regard to all accessible alternatives, which is more modestly but provably possible. In particular, optimality justifications can achieve a non-circular justification of deductive, inductive, and abductive reasoning. This volume pursues two goals: a general renewal of foundation-theoretic epistemology based on the account of optimality justifications, and the advancement of methods of optimality justification in important domains of epistemology and the philosophy of science, logic, and cognition. Connected with these goals is the aspiration to develop new ideas for mainstream epistemology, as well as for formal epistemology, philosophy of science, and cognitive science, which are intended to attract researchers, students, and all other readers interested in these fields.