INICIO | 24 de mayo de 2024

# (Pure) logic out of probability.

Título inglés (Pure) logic out of probability. Lógica (pura) obtenida de la probabilidad. Sales, Ton Dep. Lleng. Sist. Inform. Fac. Inform. Univ. Politèc. Catalunya, Barcelona, España 1134-5632 1996, 3 (1-2): 137-147, 23 Ref. articulo Inglés Today, Logic and Probability are mostly seen as independent fields with a separate history and set of foundations. Against this dominating perception, only a very few people (Laplace, Boole, Peirce) have suspected there was some affinity or relation between them. The truth is they have a considerable common ground which underlies the historical foundation of both disciplines and, in this century, has prompted notable thinkers as Reichenbach [14], Carnap [2] [3] or Popper [12] [13] (and Gaifman [5], Scott & Krauss [21], Fenstad [4], Miller [10] [11], David Lewis [9], Stalnaker [22], Hintikka [7] or Suppes [23]) to consider connection-building treatments of Logic and Probability as desirable. Indeed such a line of thinking can be pursued (this author, for one, attempted it in [15-19]). In so doing, one straightforwardly obtains a logic based on -as the simple unifying concept- an additive non-functional truth valuation which, though technically indistinguishable from (axiomatic) Probability, can however be totally decontaminated from parasitical probabilistic interpretations (such as the usual readings of event, probability or conditioning) and be given instead a strictly logical reading and justification (in terms of sentence, truth or relativity). Once some deeply-ingrained reading habits are overcome, the required concepts and formulas flow easily, and the resulting assertion-based sentential calculus becomes a very natural extension of ordinary two-valued reasoning. Furthermore, in the process we get: (a) intuitive geometrical and information-related interpretations of the concepts, (b) a simple theoretical explanation for some poorly justified formulas (intermittently advanced by various authors, some mentioned above),and (c) a semantics -and a proof theory- for general assertions that is unproblematically derived and also fully consistent with empirical or ad hoc approximate-reasoning Bayesian formulas found by Artificial Intelligence researchers. 120304 ; 110403 Lógica inductiva ; Algebras de Boole ; Incertidumbre MR1414261 Zbl 0861.03016 Acceso al artículo completo
Equipo DML-E
Instituto de Ciencias Matemáticas (ICMAT - CSIC)
rmm()icmat.es