Open AccessStrength and Stability
Author(s) -
Paula Teijeiro
Publication year - 2021
Publication title -
análisis filosófico
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.111
H-Index - 1
eISSN - 1851-9636
pISSN - 0326-1301
DOI - 10.36446/af.2021.459
Subject(s) - stability (learning theory) , mathematical economics , mathematics , epistemology , positive economics , computer science , economics , philosophy , machine learning
In this paper, I present two presumed alternative definitions of metavalidity for metainferences: Local and Global. I defend the latter, first, by arguing that it is not too weak with respect to metainference-cases, and that local metavalidity is in fact too strong with respect to types. Second, I show that although regarding metainference-schemas Local metavalidity is always stable, Global metavalidity is also stable when the language satisfies reasonable expressibility criteria (and that in fact, both concepts collapse in those cases).