Orientador: Walter Alexandre Carnielli / Dissertação (mestrado) - Universidade Estadual de Campinas, Instituto de Filosofia e Ciências Humanas / Made available in DSpace on 2018-08-17T07:04:57Z (GMT). No. of bitstreams: 1 BarreroGuzman_TomasAndres_M.pdf: 584062 bytes, checksum: b969cc9456140851ae30720d7d2f5d09 (MD5) Previous issue date: 2004 / Resumo: O trabalho estuda o papel da negação na logica, abordando os fragmentos positivos da logica proposicional, de forma a atender a dois problemas: a obtenção de teoremas de completude independentes da negação e o problema de paradoxos positivos, como o Paradoxo de Curry. Para o fragmento classico, estuda-se o metodo construtivo de completude proposto por Leon Henkin. Investigam-se as razoes pelas quais este metodo nao pode ser estendido para fragmentos nao-classicos que conseguem evitar a ocorrencia da objeção de Haskel Curry como, por exemplo, os das logicas n-valentes de Jan Lukasiewicz e os (por nos denominados) intuicionistas de Wilhelm Ackermann, quer pelas características da implicação, quer pela presença de um tipo de argumento infinito. O estudo conjunto do metodo de Henkin e do fenomeno da trivialidade positiva permite estabelecer um processo de decidibilidade da logica positiva classica através de um sistema de tablos que utiliza somente recursos metalinguísticos positivos, e propor uma rediscussão a respeito do papel da negação em logica atraves do conceito de paratrivialidade. Nesse contexto discutimos, do ponto de vista conceitual, a relação da logica positiva com o infinito, as possibilidades de se obter uma logica de primeira ordem completa sem negação e o vinculo filosofico entre verdade e significado / Abstract: This work studies some problems connected to the role of negation in logic, treating the positive fragments of propositional calculus in order to deal with two main questions: the proof of the completeness theorems in systems lacking negation, and the puzzle raised by positive paradoxes like the well-known argument of Haskel Curry. We study the constructive completeness method proposed by Leon Henkin for classical fragments endowed with implication, and advance some reasons explaining what makes difficult to extend this constructive method to on-classical fragments equipped with weaker implications (that avoid Curry?s objection). This is the case, for example, of Jan Lukasiewicz?s n-valued logics and Wilhelm Ackermann?s logic of restricted implication. Besides such problems, both Henkin?s method and the triviality phenomenon enable us to propose a new positive tableau proof system which uses only positive meta-linguistic resources, and to motivate a new discussion concerning the role of negation in logic proposing the concept of paratriviality. In this way, some relations between positive reasoning and infinity, the possibilities to obtain a first-order positive logic as well as the philosophical connection between truth and meaning are discussed from a conceptual point of view / Mestrado / Mestre em Filosofia

La notion de T-radical d’un idéal permet à G.Cherlin de démontrer un Nullstellensatz dans les théories inductives d’anneaux. Nous proposons une analyse modèle-théorique de phénomènes connexes. En premier lieu, une réciproque de ce théorème nous conduit à une caractérisation des corps algébriquement clos, suggérant une version “positive” du travail de Cherlin, la théorie des idéaux T-radiciels. Ceux-ci se caractérisent par un théorème de représentation et sont associés à un théorème des zéros “positif”. Ces résultats se généralisent à la logique du premier ordre : grâce à la notion de classe spéciale, nous développons ensuite une théorie logique des idéaux. On peut encore parler d’idéaux premiers et radiciels, relativement à une classe de structures. Dans ce cadre, le théorème de représentation est une propriété intrinsèque des classes spéciales et le théorème des zéros une propriété de préservation logique, que nous appelons “complétude géométrique” et qui entretient des rapports étroits avec la modèle-complétude positive. Les algèbres basées en groupes de P.Higgins permettent d’appliquer ces résultats aux théories modèle-complètes de corps avec opérateurs additionnels. Dans certains cas “noethériens”, l’algèbre de coordonnées est un invariant algébrique des “variétés affines”. Enfin, il est possible à partir d’un ensemble de formules E de généraliser les classes spéciales et autres classes de structures. Notre théorie des idéaux logiques est de plus un cas particulier du phénomène de localisation étudié par M.Coste ; dans certaines situations, un bon choix de formules permet d’identifier les types complets d’une “algèbre” à des types de localisation / The notion of T-radical of an ideal allows G.Cherlin to prove a Nullstellensatz for inductive ring theories.We present here a model-theoretic analysis of closely related phenomena. At first, a reverse of this theorem leeds us to a characterization of algebraically closed fields, suggesting a “positive” version of Cherlin’s work, the theory of T-radical ideals. These are characterized by a representation theorem and associated to a “positive” Nullstellensatz. Those results are generalized to first order logic : thanks to the notion of special class, we then develop a logical theory of ideals. One may still speak about prime and radical ideals, relatively to a class of structures. In this setting, the representation theorem is an intrinsic property of special classes and the Nullstellensatz a logical preservation property, which we call “geometric completeness” and which is closely linked to positive model-completeness. The group-based algebras of P.Higgins allow us to apply these results to model-complete theories of fields with additional operators. In certain “noetherian” cases, the coordinate algebra is an algebraic invariant of “affine algebraic sets”. At last, it is possible from a set of formulas E to generalize special and other classes of structures. Moreover, our theory of logical ideals is a particular case of the localisation phenomenon studied by M.Coste ; in certain situations, a good choice of formulasleeds to an identification of the complete types of a given “algebra” with some localisation types