Classical Gentzen-type Methods in Propositional Many-Valued Logics
|
|
- Harvey Rich
- 5 years ago
- Views:
Transcription
1 Classical Gentzen-type Methods in Propositional Many-Valued Logics Arnon Avron School of Computer Science Tel-Aviv University Ramat Aviv 69978, Israel Abstract A classical Gentzen-type system is one which employs two-sided sequents, together with structural and logical rules of a certain characteristic form. A decent Gentzentype system should allow for direct proofs, which means that it should admit some useful forms of cut elimination and the subformula property. In this tutorial we explain the main difficulty in developing classical Gentzen-type systems with these properties for many-valued logics. We then illustrate with numerous examples the various possible ways of overcoming this difficulty. Our examples include practically all 3-valued logics, the most important class of 4-valued logics, as well as central infinite-valued logics (like Gödel- Dummett logic, S5 and some substructural logics). 1 Introduction Many-valued logics are, above all, Logics. This means that like any other logic, the main issue with which they deal is: what formulas follow from what sets of premises (under certain assumptions). The many-valued part of their name refer to the type of semantics which determines the answers they give to this basic question. However, any logic needs efficient proof systems in order to actually be applied (as well as for a deeper understanding of its logical properties), and many-valued logics are no exception. The problem of adapting to many-valued logics the usual methods of automated reasoning (like Gentzen-type systems, tableaux, and resolution) has therefore attracted a lot of attention in recent years. The main idea which is used in most of the works on this subject is to use for any particular Ò some Ò-valued counterparts of the structures used in the usual proof systems for classical logic (like sequents with Ò components, tableaux systems with Ò signs, etc.). This fact is well reflected in the two recent survey papers [36] and [16] (see there for extensive list of references). In this paper we concentrate, in contrast, on proofs systems which use the same syntactic structures as the classical ones (with particular emphasis on Gentzen-type systems, on which all other systems are based). This approach has two advantages. First, the implementation of the proofs systems described here can be based on existing systems, because no new data structure is used. Second, two sided sequents (like those used in classical logic) can directly represent the consequence relation of any given logic, and this is, after all, what logics are all about. 2 Gentzen-type Systems A Background In what follows Ä is a propositional language, Ô Õ Ö denote atomic formulas, ³ denote arbitrary formulas (of Ä), and denote finite sets of formulas of Ä. Definition [44, 43] A (Scott) consequence relation (scr for short) for Ä is a binary relation between (finite) sets of formulas of Ä that satisfies the following conditions: R reflexivity: ³ ³ for every ³. M monotonicity: if and ¼, ¼ C cut: then ¼ ¼. if and ¼ ¼ then ¼ ¼. 2. An scr is uniform if, for every uniform substitution and every and, implies µ µ. is consistent (or non-trivial) if there exist non-empty and s.t.. 3. A propositional logic is a pair Ä Ä Ä, where Ä is a propositional language and Ä is a uniform consistent scr for Ä. Some notes concerning this Definition are in order:
2 There are exactly four inconsistent finitary scrs in any given language: the one in which iff and are non-empty; the one in which iff is nonempty; the one in which iff is non-empty; and the one in which for all and. All of them should be considered trivial, and are excluded from our definition of a logic. The notion of an scr is a natural symmetric generalization of the notion of a Tarskian consequence relation (tcr), which is defined similarly, but is a relation between sets of formulas and formulas (the monotonicity condition applies in it therefore only to the l.h.s.). A further useful generalization of the concept of a consequence relation is obtained by giving up the monotonicity condition. There are also generalizations in which the relation is taken to be between multisets (or even sequences) of formulas rather than sets thereof. A Gentzen-type calculus ([30]) over Ä is an axiomatic system which manipulates higher-level constructs called sequents, rather than the formulae themselves. There are several variants of what constitutes a sequent. Here we usually take it to be a construct of the form µ, where are finite sets of formulae of Ä and µ is a new symbol, not occurring in Ä. In other variants may be either multisets or sequences of formulae. There is no real difference, if we assume the following standard structural rules (permutation, contraction, and expansion, respectively): ½ ³ ¾ µ ½ ³ ¾ µ ½ ³ ³ ¾ µ ½ ³ ¾ µ ½ ³ ¾ µ ½ ³ ³ ¾ µ µ ½ ³ ¾ µ ½ ³ ¾ µ ½ ³ ³ ¾ µ ½ ³ ¾ µ ½ ³ ¾ µ ½ ³ ³ ¾ Definition 2.2 A Gentzen-type system is called standard if its set of axioms includes the standard axioms: µ (for all ), and its set of rules includes the standard structural rules, and the following weakening and cut rules: µ ¼ µ ¼ ½ µ ½ ³ ³ ¾ µ ¾ ½ ¾ µ ½ ¾ Definition 2.3 A rule of a Gentzen-type system is called pure ([11]), or multiplicative ([33]), if whenever µ can be inferred by it from µ ( ½ Ò) then, for arbitrary ¼ ¼ ½ Ò ¼ ½ ¼ Ò, the sequent ¼ ½ ¼ Ò µ ¼ ½ ¼ Ò can also be inferred by it from ¼ µ ¼ ( ½ Ò)1. is called pure if all its rules are pure. 1 In other words: a rule is pure if it is context-free. In practice, this means that there are no side conditions on how it can be applied. A Gentzen-type system,, directly defines a Tarskian consequence relation between sequents. Usually, however, it is mainly used as a tool for investigating scrs between the formulas of the underlying language. Any non-trivial standard Gentzen-type system naturally defines in fact an scr : Definition Let be a standard Gentzen-type system. We say that iff µ is a theorem of. is called the standard scr defined by. 2. A Gentzen-type system over a language Ä is sound and complete for a logic Ä Ä Ä if Ä. is weakly sound and complete for Ä if for all, Ä iff µ is provable in. is the most natural consequence relation associated with a given Gentzen-type system. However, it is not the only possible or useful one. Thus according to another important Tarskian consequence relation frequently associated with, a formula follows from the set ³ ½ ³ Ò µ if µ follows in from µ ³ ½ µ µ ³ Ò µ. We next describe a corresponding scr. Proposition 2.5 ([12]) If a Gentzen-type system is standard and pure then ³ ½ ³ Ò µ ½ Ñ iff µ follows in G from µ ³ ½ µ µ ³ Ò µ ½ µµ Ñ µµ This remains true even if is not closed under weakening. An ideal Gentzen-type system (of which the usual systems for classical logic provide the principal examples) is a pure, standard system in which every logical rule is an introduction rule for one connective. Moreover: it should introduce exactly one occurrence of that connective in its conclusion, no other occurrence of that connective or any other connective should be mentioned anywhere else in its formulation, and its side formulas should be immediate subformulas of the principal formula. The next definition formulates this idea in exact terms, and provides a method for describing such rules. Definition 2.6 ([15]) 1. A canonical rule of arity Ò is an expression of the form µ ½Ñ, where Ñ ¼, is either Ô ½ Ô ¾ Ô Ò µ µ or µ Ô ½ Ô ¾ Ô Ò µ for some connective (of arity Ò), and for ½ Ñ, µ is a clause such that Ô ½ Ô ¾ Ô Ò 2 2 By a clause we mean a sequent which consists of atomic formulas only. When propositional clauses are written in this way, resolution and cut amount to the same thing.
3 2. An application of a canonical rule of the form: Notes: µ ½Ñ Ô ½ Ô Ò µ µ is any inference step of the form: µ ½Ñ ½ Ò µ µ where and are obtained from and (respectively) by substituting for Ô (for ½ Ò), are any finite sets of formulas, Ë Ñ ½, and Ë Ñ ½. An application of a canonical rule of the other type is defined similarly. 1. While sequents are written in a metalanguage for Ä (which includes the extra symbol µ), a canonical rule is formulated in a meta-meta language of Ä (which includes one further symbol: ). 2. Definition 2.6 allows the case Ñ ¼, when canonical rules reduce to canonical axioms of the form Ô ½ Ô ¾ Ô Ò µ µ or µ Ô ½ Ô ¾ Ô Ò µ. This amounts to allowing canonical propositional constants. Example 2.7 The two usual introduction rules for classical conjunction can be formulated as the two canonical rules: Ô ½ Ô ¾ µµ Ô ½ Ô ¾ µ Applications of these rules have the form: µ µ µ Ô ½ µ µ Ô ¾ µ µ Ô ½ Ô ¾ µ ¼ µ ¼ ¼ µ ¼ Definition 2.8 A Gentzen-type system is called canonical if its axioms include the standard axioms, its rules include the standard structural rules, cut and weakening, and all its other rules and axioms are canonical. A canonical system is obviously standard and pure. However, in order for to really be an scr, one further condition should be satisfied: Theorem 2.9 ([15]) If is a canonical system then is consistent (and so an scr) iff the following coherence condition is satisfied: whenever Ë ½ Ô ½ Ô ¾ Ô Ò µ µ and Ë ¾ µ Ô ½ Ô ¾ Ô Ò µ are rules of, the set of clauses Ë ½ Ë ¾ is classically inconsistent (and so the empty clause can be derived from it using cuts). Moreover: any canonical system which satisfies this condition admits cut-elimination (i.e.: its cut rule is redundant). Example 2.10 The two classical rules for conjunction described in Example 2.7 form a coherent set of rules, where Ë ½ Ô ½ Ô ¾ µµ, Ë ¾ µ Ô ½ µ µ Ô ¾ µ. Ë ½ Ë ¾ is here the inconsistent set Ô ½ Ô ¾ µµ µ Ô ½ µ µ Ô ¾ µ. 3 A Classical Example The classical examples of canonical Gentzen-type systems are those for Classical Propositional Logic (CPL) and its fragments. We present now one of them, in which the primitive connectives are negation, disjunction, conjunction, and implication, and all the rules are pure. It should be noted that Gentzen s original formulation in [30] uses the additive form of these rules, in which all premises of a rule have the same side formulas. In the presence of the standard structural rules and weakening the two forms are equivalent, but otherwise the additive form is impure. In the present context it has however the advantage of being invertible. This is a very important property, and so in later sections we shall prefer to use the additive form whenever it is both invertible and equivalent to the pure one. THE SYSTEM GCPL Axioms: ³ µ ³ Structural Rules: Cut, Weakening (and the standard rules) Logical Rules: µµ µ µ µµ µµ µµ µ µ µµ µ µ µ ³ ³ µ ³ µ µ ³ ½ µ ½ ³ ¾ µ ¾ ³ ½ ¾ µ ½ ¾ ³ µ µ ³ ³ µ ³ µ ½ µ ½ ³ ¾ µ ¾ ½ ¾ µ ½ ¾ ³ ½ ³ µ ½ ¾ µ ¾ ½ ¾ ³ µ ½ ¾ µ ³ µ ³ The most important theorem concerning this calculus is Gentzen s celebrated cut-elimination theorem ([30, 48]). We present now a strong form of it, together with applications and two methods of proof. Theorem 3.1 Strong Cut-elimination for GCPL ([34, 12]): µ is derivable from ½ µ ½ Ò µ Ò
4 in È Ä iff it has a proof from these assumptions in which all cuts are done on formulas in Ë Ò ½ Ë Ò ½. Corollary 3.2 ÈÄ µ iff it has a cut-free proof. Gentzen s Method of Proof: Show by a double induction on the complexity of the cut-formula ³ and on the sum of the lengths of the proofs of the premises that if ½ µ ½ ³ ³ and ³ ³ ¾ µ ¾ have good proofs, then so does ½ ¾ µ ½ ¾. Semantic Method of Proof (outline): Let Ë be a given set of sequents, and assume Ë ½ µ ½ Ò µ Ò. A sequent µ is called Ë-saturated if: (1) It has no good proof from Ë (2) If ³ ¾ Ò Ë ½ Ò Ë ½ then ³ ¾ (3) If ³ ¾ then ³ ¾ and ¾ If ³ ¾ then either ³ ¾ or ¾ (4) Similar conditions obtain for the other connectives The theorem easily follows from the following two lemmas: Lemma 1: If µ is Ë-saturated, then the following valuation is a model of Ë, but not a model of µ : Ú Ôµ Ø Ô ¾ Ô ¾ In fact, for every ³, ³ ¾ implies Ú ³µ Ø and ³ ¾ implies Ú ³µ. Lemma 2: If µ has no good proof from Ë then it can be extended to an Ë- saturated sequent. The (strong) cut-elimination theorem for È Ä has many applications, like: decidability, the interpolation theorem, and the subformula property. It is also the basis for the two main proof search methods for È Ä: The Tableaux Method: Try to show that ÈÄ by searching for a cut-free proof of µ in È Ä. This is done by applying backwards the invertible versions of the rules. In particular one shows that ³ is valid by searching for a cut-free proof of µ ³. This gives either such a proof, or an equivalent set of clauses (i.e.: sequents consisting of atomic formulas) which can be translated into a CNF for ³. The Resolution Method: Try to show that ³ ½ ³ Ò ÈÄ ½ by showing that the empty sequent µ can be derived in È Ä from the set: µ ³ ½ µ µ ³ Ò µ ½ µµ µµ (In particular, a formula ³ is shown to be valid by proving that ³ µµ ÈÄ µ). For this replace (using tableaux) each µ ³ µ and µµ by an equivalent set of clauses. By the strong cut-elimination theorem, if µ is derivable from the original set of sequents, then it can be derived from the union of the equivalent sets of clauses using only cuts. 4 The Problem with Gentzenizing Many- Valued Logics We start by defining in precise terms what we mean by a many-valued logic : Definition A matrix Å for a language Ä is a triple Å Ç such that: (a) Å is a nonempty set (of truth-values ). (b) is a proper, nonempty subset of Å (the designated values ). 3 (c) Ç is a set of operations on Å, so that for each connective of Ä there is a corresponding operation on Å, and vice versa. 2. Let Å be a matrix for Ä. Å, The consequence relation induced by Å, is defined by: Å iff for every valuation Ú in Å which respects the operations and such that Ú µ ¾ for every ¾, we have that Ú ³µ ¾ for some ³ ¾. 3. Let ½ Ò ¼. A logic Ä Ä Ä is (weakly) Ò-valued if there exists a matrix Å for Ä such that: Notes: (a) Å has exactly Ò elements. (b) Å Ä ( Å ³ iff Ä ³ for every sentence ³). (c) For every natural number there exists a finite submatrix Å of Å such that whenever a sequent µ (a formula ³) contains occurrences of no more than atomic formulas, we have that Ä ( Ä ³) iff Å ( Å ³). 1. The main factor in this definition of a many-valued logic is the existence of a single characteristic matrix. Condition (c), on the other hand, guarantees that every propositional many-valued logic is decidable, and that even ¼ -valued logics may be viewed as semi-finite (in a certain sense). 3 In [49] and elsewhere the only condition concerning is that it should be a subset of Å. We exclude here the two extreme cases ( Å and ) because the corresponding scrs (as defined below) are not consistent.
5 2. A matrix Å is frequently used for defining consequence relations other than Å. In all cases we know the definitions of these alternative consequence relations can be reduced to the validity of certain formulas in Å, and so to Å (for example: the consequence relation usually associated with Łukasiewicz 3-valued logic can be characterized as follows: ³ ½ ³ Ò iff the formula ³ ½ ³ ¾ ³ Ò µ µµ is valid in Łukasiewicz 3-valued matrix). The following theorem is the main obstacle to providing decent Gentzen-type calculi for many-valued logics: Theorem 4.2 ([15]) Let be a consistent canonical calculus. Then either defines a logic which is a fragment of classical logic, or it is not sound and complete for any many-valued logic. It follows that a Gentzen-type calculus for a given manyvalued logic should use at least one of the following: Noncanonical rules Axioms which are neither standard nor canonical Impure rules (i.e.: rules with side conditions on their applications) A nonstandard set of structural rules In the following sections we shall see examples of all these alternatives. 5 The Use of Noncanonical Rules and Axioms The most important feature of canonical rules is that they introduce exactly one new occurrence of a connective at a time. Most of the Gentzen-type systems for many-valued logics give up this feature by allowing rules which introduce two occurrences of connectives at the same time. 5.1 Three-Valued Logics We start with the simplest type of non-classical manyvalued logics: the three-valued ones. We assume (w.l.o.g.) that those three values are Ø and Á, where Ø is designated and is not. The 3-valued logics are accordingly divided into two classes: those in which Á is designated (i.e. Ø Á), and those in which it is not (i.e. Ø). The logics inside each class differ only with respect to the expressive power of their languages. The most famous 3-valued logic with Ø is that of Kleene ([39]). It is based on the classical connectives, and, where the interpretation of is given by: Ø Ø Á Á and, are interpreted as the ÑÜ and ÑÒ operations (respectively) according to the order Á Ø. This logic has no valid sentences and no appropriate implication. One natural way to remedy this (resulting in a language equivalent to that used in the logic LPF of the VDM project ([38]) as well as to the language of Łukasiewicz 3-valued logic Ł ([40])) is to add to the language the following connective from [10], interpreted as: Ø if ¾ if ¾ The standard 3-valued logic with Ø Á is the paraconsistent logic  [19, 8, 41, 24]. Its basic connectives are again, with the interpretations defined identically as above 4. Now according to both interpretations, the classical rules for, and are valid, but one of the rules for is not (they differ w.r.t to which of them is valid and which fails). The way to get useful cut-free Gentzen-type systems for these logics is therefore to replace, first of all, the two classical rules for negation with the following rules (from [10]) for the combination of negation with the connectives of the language: 5 µµ µµ µµ µµ µµ µµ µµ µµ µ µ µ µ µ µ µ µ µ µ µ µ µ µ µ µ µ µ µ µ µ µ µ µ µ Let us call the resulting system Ë. This system is sound for both choices of. In both cases, in order to get corresponding sound and complete systems one needs 4 This means that in the case of the interpretations are not identical, because is different. The two interpretations are however definable in terms of each other using negation and the propositional constant. 5 As we stated above, we use from now on the additive form of a rule (i.e., we assume that the side formulas are the same in all premises of a rule) whenever weakening and the standard structural rules are among the rules of the system under consideration.
6 to add noncanonical axioms as well: ³ ³ µ in case Ø, µ ³ ³ in case Ø Á. The classical proofs of the generalized cut-elimination theorem and the completeness theorem can be now adapted for the two systems with very slight changes. For example, in the case where Ø Á the countermodel of a saturated sequent µ is defined by: Ú Ôµ Ø Ô ¾ Á Ô ¾ Ô ¾ Ô ¾ Unlike the two-valued case, the set of connectives used in Ë is not functionally complete (with either of the interpretaions for used above). In [13] it is proved, however, that by adding the propositional constants and Á (interpreted as the corresponding truth values), one does get a functionally complete set of connectives. To handle these constants, one only needs to add to the systems above some simple axioms. In the case of these are: µ µ What should be taken as the axioms for Á depends, on the other hand, on the choice of. In case Ø these are: Á µ Á µ while in case Ø Á the needed axioms are: µ Á µ Á It is straightforward now to generalize the completeness theorem and the cut-elimination theorem (as well as their proofs) to the extended systems. Since the set of connectives used in them is functionally complete, this allows us to find an adequate set of rules for every connective, and so for every 3-valued logic (provided its language includes ). Thus if is Łukasiewicz implication, then ³ is equivalent to ³ µ ³µ (where is defined like in the case Ø). This leads to the following four rules for : 6 µµ µµ µµ µµ µ ³ µ µ ³ ³ µ µ ³ ³ µ µ ³ ³ µ ³µ µ µ µ ³ µ ³µ 6 Note again that the sequents here do not reflect the consequence relation dictated by using as the official implication. That consequence relation is not even an scr, since contraction fails for it! The (strong) cut-elimination theorem which obtains for the Gentzen-type systems described in this section has the same applications and consequences in the 3-valued case as in the two-valued one. Examples are the interpolation theorem, and an appropriate version of the subformula property (according to which a proof of a sequent contains only subformulas of this sequent, or negations of proper subformulas of it). It also leads to versions of the tableaux and resolution methods which are very similar to those used in the classical case. The differences are as follows: 1. A clause in the present context is a sequent which contains only literals (i.e. atomic formulas or their negations) on both sides. Since all the rules of the systems above are invertible, every sequent can be reduced to an equivalent finite set of clauses (in this sense) by the corresponding tableaux rules. 2. A clause µ is valid not only when, but also when, for some atomic È, È È (in case Ø) or È È (in case Ø Á). 3. For proving ³ ½ ³ Ò ½ using resolution, the ordinary set of clauses obtained from the set µ ³ ½ µ µ ³ Ò µ ½ µµ µµ should be extended with all sequents of the form È È µ (in case Ø) or µ È È (in case Ø Á), where È is any atomic formula occurring in the original sequent. Note: In the tableaux method one should consider 4 types of signed formulas: ̳, ̳, ³, and ³. It might be more convenient to use instead four different signs: Ì, Ì Æ, and Æ. The resulting rules become very similar to the classical ones (but ̳, e.g., is reduced to Æ ³ rather than to ³). A branch is then closed if for some ³, it contains either ̳ and ³, or Ì Æ ³ and Æ ³, or ̳ and Æ ³ (in case Ø), or Ì Æ ³ and ³ (in case Ø Á). 5.2 Four-Valued Logics The methods used above for three-valued logics can be extended with very slight changes to four-valued logics in which there are exactly two designated elements. Let the truth values of such logics be Ø, where Ø and are the classical values, and Ø. is the fourvalued counterpart of the designated Á, while is the counterpart of the non-designated Á. This is the basis of what is known as Belnap four-valued logic ([18, 17]) 7. Intuitively, represents the truth-value of formulas about which there is inconsistent data (such a formula is both true and false ), while is the truth-value of formulas on which no data at 7 Belnap s structure is nowadays known also as the basic bilattice, and its logic as the basic bilattice logic (see [31, 32, 26, 25, 27, 28, 3, 4, 5]).
7 all is available ( neither true nor false ). With these intuitions, the interpretations of the connectives are exactly the same as in the 3-valued case (so,, is defined exactly as above, and, are interpreted as the ÑÜ and ÑÒ operations according to the following partial order: Ø). Now the methods described in the previous subsection can be used (see [4]) to show that the basic system Ë is sound and complete for the resulting logic, and that the (strong) cut-elimination theorem obtains for it. These results can be extended to the case when the language includes the propositional constants and, provided we add to Ë the obvious axioms for these connectives (for these are the axioms used above for Á in case Ø Á, for the axioms for Á in case Ø). Since is a functionally complete set of connectives for 4-valued logics ([13]), this again allows us to find a complete, cut-free Gentzen-type system for any 4-valued logic in which Ø. These results have the same applications here as in the 3-valued case. In fact, both of the resulting tableaux and resolution methods are simpler here than in the 3-valued case, because the set of axioms is simpler in the present system than in its 3-valued counterparts. Except for the class just dealt with, there are two other possible classes of 4-valued logics: those with Ø, and those with Ø. Similar general methods for handling these two classes have recently been found, but at present they are more complicated, and we shall not describe them here Gödel-Dummett Logics Up to now, we have only considered three- or fourvalued logics. In this subsection we show that the same method is also applicable to Ò-valued logics with Ò having bigger values, and even to infinite-valued logics. Our example will be the famous Gödel-Dummett logics. In [35] Gödel introduced a sequence Ò (Ò ¾) of Ò- valued matrices. He used these matrices to show some important properties of intuitionistic logic. An infinite-valued matrix in which all the Ò can be embedded was later introduced by Dummett in [20]. The logic of was axiomatized in the same paper, and has been known since then as Gödel-Dummett s Ä. It is probably the most important intermediate logic, which turns up in several places, such as the provability logic of Heyting s Arithmetics ([50]), and relevance logic ([21]). Recently it has again attracted a lot of attention because of its recognition as one of the three most basic fuzzy logics ([37]). The language of Ä is. The matrix is Æ Ø, where is the usual order on 8 One particular four-valued logic in which Ø belongs to the sequence of logics dealt with in the next subsection. Æ extended by a greatest element Ø, the interpretation of the propositional constant is the number 0, is Ø if and otherwise, is ¼, and and are resp. the ÑÒ and ÑÜ operations on Æ Ø 9. The matrices of Ò are similar, but the set of truth values of Ò is ¼ Ò ¾ Ø. A cut-free Gentzen-type formulation for Ä was first given by Sonobe in [47]. His approach was improved in [6] and [23]. All those systems have, however, the serious drawback of using the following rule, which introduces an arbitrary number of implications, and has an arbitrary number of premises, all of which contain formulas of essential importance for the inference: ³ µ µ Ê Here contains an arbitrary number Ñ ¼ of implicational formulas ³ ½ Ñ may also contain some other kinds of formulas. indicates the multiset consisting of exactly the Ñ implicational formulas of. is just after removal of the th formula ³. An alternative system Ä ÊË for Ä, which does not have a rule with arbitrary number of premises, has been given in [14]. Like the systems for 3-valued and 4-valued logics described above, Ä ÊË uses invertible rules which decompose formulas in a sequent into simpler ones, until a set of clauses equivalent to the original sequent is reached. Here, however, a clause is defined as a sequent consisting only of atomic formulas or implications between atomic formulas. Accordingly, the number of rules needed for each connective (other than ) is six. The rules for, for example, are: µ µ µµ µ µ µµ µ µ µµ µ ³ µ ³ ³ µ µ ³ µ µ ³ ½ µ ³ ¾ µ ³ ½ ³ ¾ µ ³ ½ ³ ¾ µ ³ ½ ³ ¾ µ µ µ ³ ½ ³ ¾ µ ³ ½ ¾ µ ³ ½ µ ³ ¾ µ ³ ½ ¾ µ µ 9 This interpretation is not the one given by Gödel and Dummett, but its dual. We note also that for the application as a fuzzy logic it is more useful ([37]) to use instead of Æ Ø the real interval [0,1], with 1 playing the role of Ø. This makes a difference only when we consider inferences from infinite theories.
8 Like in the 3-valued case, these rules are not sufficient for completeness, and we have to add non-standard axioms as well. Unfortunately, the needed axioms are more complicated than those employed in the 3-valued case. Definition 5.1 Let µ be a clause of Ä ÊË. We say that Ô Õµ ¾ µ µ iff Ô Õµ ¾. We say that Ø Õµ ¾ µ µ iff Õ ¾. We say that Ô Õµ ¾ µ µ iff Õ Ôµ ¾. We say that Õ Øµ ¾ µ µ iff Õ ¾. Let Ô Õ be either atomic formulas or Ø 10. We say that Ô» Õµ ¾ µ µ iff either Ô Õµ ¾ µ µ or Ô Õµ ¾ µ µ. A sequence Õ ½ Õ Ð (where Õ is either atomic or Ø) is called a strictly increasing sequence for µ if Õ» Õ ½ µ ¾ µ µ for ½ Ð ½, and either Õ Õ ½ µ ¾ µ µ for some ½ Ð ½, or Õ ½ Ø Õ Ð. Now the axioms needed in the case of Ä itself are those clauses for which there exists a strictly increasing sequence Õ ½ Õ Ð such that either Õ ½ Õ Ð, r Õ ½ Ø, or Õ Ð. To get a complete, cut-free formulation for a given finite k-valued Gödel-Dummett logic, one needs to have as additional axioms all basic sequents µ having a strictly increasing sequence Õ ½ Õ Ð such that for at least different Õ s, Õ Õ ½ µ ¾ µ µ. Instead of enriching the set of axioms, an alternative approach presented in [14] is to employ special analytic simplification rules. Those needed for the ¼ -valued Ä are: Transitivity: Left maximality: Right maximality: Linearity: Minimality of : Ô Õ Õ Ö Ô Ö µ Ô Õ Õ Ö µ Ô Õ Ô Õ µ Ô Õ Ô µ µ Õ Ô Ô µ Õ Ô Ô Õ µ Õ Ô µ Õ Ô Ô Ô µ Ô µ 10 Note that in this paper Ø is not an official symbol of the language of Ä. Here, however, it is used in the metalanguage. Note: All the logical rules of the systems described in this subsection have what might be called the semi-subformula property: written in Polish notation, every formula in their premises either appears in their conclusion or is obtained from some formula there by deleting some of its symbols. This is not very different from the usual subformula property. The simplification rules, in contrast, do not have this property. Nevertheless, each of them is still analytic in the sense that its possible premises (for any potential conclusion) are determined by its conclusion, they are finite in number, and can be found effectively. These rules are close in nature to what is known as analytic cut (see next section), although they are in fact simpler. 6 Systems For Weak Completeness In this section we present examples of systems which employ impure rules and systems which differ from classical logic in their structural rules. The immediate effect is that usually the consequence relation defined by a Gentzentype system of this kind cannot be identical with a consequence relation induced by a matrix. What can be achieved is weak soundness and completeness (see Definition 2.4). This is usually sufficient, since the consequence relation of the logic is frequently reducible to its set of valid sentences using some deduction theorem. 6.1 Employing Impure Rules: The Modal Ë In this subsection we present one important example of the use of an impure rule: the standard Gentzen-type system for the modal logic Ë. Like most other modal logics, Ë has a well-known (particularly simple) possible-worlds semantics. However, Ë had been shown in [45] to be a many-valued logic (in the strict sense defined above) much before its Kripke-style semantics was discovered. The many-valued semantics of Ë is based on Boolean Algebras. Let ½ ¼ be such an algebra. Define an operation ¾ on as follows: ½ ¾ ¼ ½ ½ Let ½ be the only designated value in the resulting structure. It has been proved in [45] that Ë is sound and complete for the class of these structures. Moreover: if a sentence ³ in the language of Ë is not a theorem of Ë, then it is refutable in some finite structure of this type (with at most ¾ ¾Ò elements, where Ò is the number of atomic formulas occurring in ³) All these finite matrices are embeddable in one denumerable matrix, which is characteristic for Ë.
9 Ë, the standard Gentzen-type system for Ë, is obtained from È Ä by adding to it the following two simple rules: ¾ µ ³ µ ¾³ µ µ ³ µ ¾³ µ ¾ There is, however, a side condition on the application of µ ¾, which makes this rule impure: all the side formulas (i.e., the formulas of and ) should begin with ¾ (an equivalent version demands them only to be essentially modal, in the sense that each occurrence of an atomic formula should be within the scope of a ¾). As a result, the standard consequence relation defined by Ë is different from that induced by the many-valued semantics described above (for example: ³ Ë ¾³ according to that semantics, but the sequent ³ µ ¾³ is not provable in Ë). Nevertheless, we have the following weak completeness theorem: Theorem 6.1 A formula ³ is a theorem of Ë iff Ë µ ³. Another drawback of Ë is the failure of the cut elimination theorem for Ë. Instead we have, however, the possibility to eliminate all non-analytic cuts: Definition 6.2 A cut in a proof of µ is called analytic if the cut formula is a subformula of µ. Theorem 6.3 Ë µ iff there exists a proof of this sequent in which all the cuts are analytic and are performed on formulas of the form ¾. This theorem suffices for the subformula property, and for developing a good tableaux system for Ë. 6.2 Systems without Weakening Substructural logics ([42]) are logics defined by Gentzen-type systems with an irregular set of structural rules (the regular set being the one consisting of the standard structural rules, weakening, and cut). The official consequence relation associated with a logic of this kind is therefore not a Scott consequence relation in the sense of Definition 2.4. Its set of valid formulas may still correspond, however, to some many-valued logic. We present now two substructural logics where this is indeed the case. The system ÊÅÁ Ñ is obtained from È Ä by simply deleting the weakening rule 12. It is easy to prove, using Gentzen s method, that cut-elimination still obtains for this system. Moreover: a strong form of the interpolation theorem holds for it: µ ³ is provable only if there exists 12 In the literature on relevance logic (including [7], from which the results below are taken) the symbols,, Æ, and + are used instead of,,, and (respectively). an interpolant containing only atomic formulas common to ³ and 13. This puts the corresponding logic, ÊÅÁ Ñ, in the family of Relevance logics (in which framework it was originally introduced and investigated). In [7] it was proved that it is also a many-valued logic. The corresponding matrix, called in [7], is the following: Truth Values: Ø Á ½ Á ¾ Á. All elements except are designated. Operations: Ø, Ø, Á Á ½ ½µ. Ø ÓÖ Ø Á Á otherwise. Like, can be viewed as the limit of its finite substructures. Indeed, let Ò be the substructure of consisting of Ø Á ½ Á Ò. Then we have: Theorem 6.4 A formula ³ is valid in (or in all of the finite structures Ò ) iff ÊÅÁÑ µ ³. Moreover, if ³ contains Ò atomic formulas and ÊÅÁÑ µ ³ then there is a valuation which refutes ³ in Ò. Among the Ò s, the three-valued is particularly interesting. It was first introduced and axiomatized by Sobociński in [46]. The corresponding logic is known to be the purely intensional (or multiplicative, in the terminology of [33]) fragment of Dunn-McColl semi-relevant logic ÊÅ ([1, 22]). In [9] it was shown that a corresponding cutfree Gentzen-type system, ÊÅ Ñ, can be obtained from ÊÅÁ Ñ by adding to it the following mingle rule (also called mix in the literature on Linear Logic): ½ µ ½ ¾ µ ¾ ½ ¾ µ ½ ¾ This is of course a new structural rule, which is weaker than weakening. Note: The rules of ÊÅÁ Ñ and ÊÅ Ñ are pure, but they are not invertible. It seems therefore difficult to base a reasonable tableaux system on them (although like in classical logic, there exist semantical arguments for these logics that simultaneously prove completeness and cut elimination). References [1] A. Anderson and N. Belnap. Entailment, volume I. Princeton University Press, Unlike in classical logic, such atomic formulas necessarily exist in case µ ³ is provable in ÊÅ ÁÑ. This is called the variable sharing property in the literature on relevance logics ([1, 2, 22]), and it is the most characteristic feature of these logics.
10 [2] A. Anderson and N. Belnap. Entailment, volume II. Princeton University Press, [3] O. Arieli and A. Avron. Logical bilattices and inconsistent data. In Proc. 9th IEEE Annual Symp. on Logic in Computer Science (LICS 94), pages IEEE Press, [4] O. Arieli and A. Avron. Reasoning with logical bilattices. J. of Logic, Language and Information, 5(1):25 63, [5] O. Arieli and A. Avron. The value of four values. Artificial Intelligence, 102(1):97 141, [6] A. Avellone, M. Ferrari, and P. Miglioli. Duplication-free tableaux calculi together with cut-free and contraction-free sequent calculi for the interpolable propositional intermediate logics. Logic Journal of IGPL, 7: , [7] A. Avron. Relevant entailment - semantics and formal systems. Journal of Symbolic Logic, 49: , [8] A. Avron. On an implication connective of rm. Notre Dame Journal of Formal Logic, 27: , [9] A. Avron. A constructive analysis of rm. Journal of Symbolic Logic, 52: , [10] A. Avron. Natural 3-valued logics: Characterization and proof theory. J. of Symbolic Logic, 56(1): , [11] A. Avron. Simple consequence relations. Information and Computation, 92(1): , [12] A. Avron. Gentzen-type systems, resolution and tableaux. Journal of Automated Reasoning, 10: , [13] A. Avron. On the expressive power of three-valued and four-valued languages. Journal of Logic and Computation, 9: , [14] A. Avron and B. Konikowska. Decomposition proof systems for gödel-dummett logics. Studia Logica (in print), [15] A. Avron and I. Lev. Canonical propositional gentzen-type systems Submitted to IJCAR [16] M. Baaz, C. G. Fermüller, and G. Salzer. Automated deduction for many-valued logics. In A. Robinson and A. Voronkov, editors, Handbook of Automated Reasoning. Elsevier Science Publishers, [17] N. D. Belnap. How computers should think. In G. Ryle, editor, Contemporary Aspects of Philosophy, pages Oriel Press, Stocksfield, England, [18] N. D. Belnap. A useful four-valued logic. In G. Epstein and J. M. Dunn, editors, Modern Uses of Multiple-Valued Logic, pages Reidel, Dordrecht, [19] I. L. M. D ottaviano. The completeness and compactness of a three-valued first-order logic. Revista Colombiana de Matematicas, XIX(1-2):31 42, [20] M. Dummett. A propositional calculus with a denumerable matrix. Journal of Symbolic Logic, 24:96 107, [21] J. Dunn and R. Meyer. Algebraic completeness results for dummett s lc and its extensions. Z. math. Logik und Grundlagen der Mathematik, 17: , [22] J. M. Dunn. Relevance logic and entailment. In [29], volume III, chapter 3, pages [23] R. Dyckhoff. A deterministic terminating sequent calculus for gödel-dummett logic. Logic Journal of IGPL, 7: , [24] R. L. Epstein. The semantic foundation of logic, volume I: propositional logics, chapter IX. Kluwer Academic Publisher, [25] M. Fitting. Bilattices in logic programming. pages IEEE Press, [26] M. Fitting. Kleene s logic, generalized. Journal of Logic and Computation, 1: , [27] M. Fitting. Bilattices and the semantics of logic programming. Journal of Logic Programming, 11(2):91 116, [28] M. Fitting. Kleene s three-valued logics and their children. Fundamenta Informaticae, 20: , [29] D. M. Gabbay and F. Guenthner. Handbook of Philosophical Logic. D. Reidel Publishing company, [30] G. Gentzen. Investigations into logical deduction. In M. E. Szabo, editor, The Collected Works of Gerhard Gentzen, pages North Holland, Amsterdam, [31] M. L. Ginsberg. Multiple-valued logics. In M. L. Ginsberg, editor, Readings in Non-Monotonic Reasoning, pages Los-Altos, CA, [32] M. L. Ginsberg. Multivalued logics: A uniform approach to reasoning in AI. Computer Intelligence, 4: , [33] J.-Y. Girard. Linear logic. Theoretical Computer Science, 50:1 102, [34] J.-Y. Girard. Proof Theory and Logical Complexity. Bibliopolis, [35] K. Gödel. On the intuitionistic propositional calculus. In S. Feferman and all, editors, Collected Work of K. Gödel, volume I, pages Oxford University Press, [36] R. Hähnle. Tableaux for multiple-valued logics. In M. D Agostino, D. M. Gabbay, R. Hähnle, and J. Posegga, editors, Handbook of Tableau Methods, pages Kluwer Publishing Company, [37] P. Hájek. Metamathematics of Fuzzy Logic. Kluwer Academic Publishers, [38] C. Jones. Systematic Software Development Using VDM. Prentice-Hall International, U.K., [39] S. C. Kleene. Introduction to metamathematics. Van Nostrad, [40] J. Łukasiewicz. On 3-valued logic. In S. McCall, editor, Polish Logic. Oxford University Press, [41] L. I. Rozoner. On interpretation of inconsistent theories. Information Sciences, 47: , [42] P. Schroeder-Heister and K. Došen, editors. Substructural Logics. Oxford University Press, [43] D. S. Scott. Completeness and axiomatization in manyvalued logics. In Proc. of the Tarski symposium, volume XXV of Proc. of Symposia in Pure Mathematics, pages , Rhode Island, American Mathematical Society. [44] D. S. Scott. Rules and derived rules. In S. Stenlund, editor, Logical theory and semantical analysis, pages Reidel, Dordrecht, [45] S. Scroggs. Extensions of the lewis system s5. Journal of Symbolic Logic, 16: , [46] B. Sobociński. Axiomatization of partial system of threevalued calculus of propositions. The Journal of Computing Systems, 11(1):23 55, [47] O. Sonobe. A gentzen-type formulation of some intermediate propositional logics. J. of Tsuda College, 7:7 14, [48] A. S. Troelstra and H. Schwichtenberg. Basic Proof Theory, 2nd Edition. Cambridge University Press, [49] A. Urquhart. Many-valued logic. In [29], volume III, chapter 2, pages [50] A. Visser. On the completeness principle: A study of provability in heyting s arithmetic. Annals of Mathematical Logic, 22: , 1982.
On Urquhart s C Logic
On Urquhart s C Logic Agata Ciabattoni Dipartimento di Informatica Via Comelico, 39 20135 Milano, Italy ciabatto@dsiunimiit Abstract In this paper we investigate the basic many-valued logics introduced
More information5-valued Non-deterministic Semantics for The Basic Paraconsistent Logic mci
5-valued Non-deterministic Semantics for The Basic Paraconsistent Logic mci Arnon Avron School of Computer Science, Tel-Aviv University http://www.math.tau.ac.il/ aa/ March 7, 2008 Abstract One of the
More informationHypersequent Calculi for some Intermediate Logics with Bounded Kripke Models
Hypersequent Calculi for some Intermediate Logics with Bounded Kripke Models Agata Ciabattoni Mauro Ferrari Abstract In this paper we define cut-free hypersequent calculi for some intermediate logics semantically
More informationCanonical Calculi: Invertibility, Axiom expansion and (Non)-determinism
Canonical Calculi: Invertibility, Axiom expansion and (Non)-determinism A. Avron 1, A. Ciabattoni 2, and A. Zamansky 1 1 Tel-Aviv University 2 Vienna University of Technology Abstract. We apply the semantic
More informationGeneral Patterns for Nonmonotonic Reasoning: From Basic Entailments to Plausible Relations
General Patterns for Nonmonotonic Reasoning: From Basic Entailments to Plausible Relations OFER ARIELI AND ARNON AVRON, Department of Computer Science, School of Mathematical Sciences, Tel-Aviv University,
More informationNon-deterministic Matrices for Semi-canonical Deduction Systems
Non-deterministic Matrices for Semi-canonical Deduction Systems Ori Lahav School of Computer Science Tel Aviv University Tel-Aviv, Israel Email: orilahav@post.tau.ac.il Abstract We use non-deterministic
More informationImplicational F -Structures and Implicational Relevance. Logics. A. Avron. Sackler Faculty of Exact Sciences. School of Mathematical Sciences
Implicational F -Structures and Implicational Relevance Logics A. Avron Sackler Faculty of Exact Sciences School of Mathematical Sciences Tel Aviv University Ramat Aviv 69978, Israel Abstract We describe
More informationCut-Elimination and Quantification in Canonical Systems
A. Zamansky A. Avron Cut-Elimination and Quantification in Canonical Systems Abstract. Canonical propositional Gentzen-type systems are systems which in addition to the standard axioms and structural rules
More informationWhat is an Ideal Logic for Reasoning with Inconsistency?
What is an Ideal Logic for Reasoning with Inconsistency? Ofer Arieli School of Computer Science The Academic College of Tel-Aviv Israel Arnon Avron School of Computer Science Tel-Aviv University Israel
More informationNEGATION: TWO POINTS OF VIEW. Arnon Avron. Department of Computer Science. Sackler Faculty of Exact Sciences
NEGATION: TWO POINTS OF VIEW Arnon Avron Department of Computer Science Sackler Faculty of Exact Sciences Tel Aviv University, Ramat Aviv, Israel 69978 1 Introduction In this paper we look at negation
More informationNON-DETERMINISTIC SEMANTICS FOR LOGICAL SYSTEMS
ARNON AVRON AND ANNA ZAMANSKY NON-DETERMINISTIC SEMANTICS FOR LOGICAL SYSTEMS 1.1 The Key Idea 1 INTRODUCTION The principle of truth-functionality (or compositionality) is a basic principle in many-valued
More informationAutomated Support for the Investigation of Paraconsistent and Other Logics
Automated Support for the Investigation of Paraconsistent and Other Logics Agata Ciabattoni 1, Ori Lahav 2, Lara Spendier 1, and Anna Zamansky 1 1 Vienna University of Technology 2 Tel Aviv University
More informationMultiplicative Conjunction and an Algebraic. Meaning of Contraction and Weakening. A. Avron. School of Mathematical Sciences
Multiplicative Conjunction and an Algebraic Meaning of Contraction and Weakening A. Avron School of Mathematical Sciences Sackler Faculty of Exact Sciences Tel Aviv University, Tel Aviv 69978, Israel Abstract
More informationDisplay calculi in non-classical logics
Display calculi in non-classical logics Revantha Ramanayake Vienna University of Technology (TU Wien) Prague seminar of substructural logics March 28 29, 2014 Revantha Ramanayake (TU Wien) Display calculi
More informationAN ALTERNATIVE NATURAL DEDUCTION FOR THE INTUITIONISTIC PROPOSITIONAL LOGIC
Bulletin of the Section of Logic Volume 45/1 (2016), pp 33 51 http://dxdoiorg/1018778/0138-068045103 Mirjana Ilić 1 AN ALTERNATIVE NATURAL DEDUCTION FOR THE INTUITIONISTIC PROPOSITIONAL LOGIC Abstract
More informationKripke Semantics for Basic Sequent Systems
Kripke Semantics for Basic Sequent Systems Arnon Avron and Ori Lahav School of Computer Science, Tel Aviv University, Israel {aa,orilahav}@post.tau.ac.il Abstract. We present a general method for providing
More informationA CUT-FREE SIMPLE SEQUENT CALCULUS FOR MODAL LOGIC S5
THE REVIEW OF SYMBOLIC LOGIC Volume 1, Number 1, June 2008 3 A CUT-FREE SIMPLE SEQUENT CALCULUS FOR MODAL LOGIC S5 FRANCESCA POGGIOLESI University of Florence and University of Paris 1 Abstract In this
More informationA Schütte-Tait style cut-elimination proof for first-order Gödel logic
A Schütte-Tait style cut-elimination proof for first-order Gödel logic Matthias Baaz and Agata Ciabattoni Technische Universität Wien, A-1040 Vienna, Austria {agata,baaz}@logic.at Abstract. We present
More informationFROM AXIOMS TO STRUCTURAL RULES, THEN ADD QUANTIFIERS.
FROM AXIOMS TO STRUCTURAL RULES, THEN ADD QUANTIFIERS. REVANTHA RAMANAYAKE We survey recent developments in the program of generating proof calculi for large classes of axiomatic extensions of a non-classical
More informationNon-classical Logics: Theory, Applications and Tools
Non-classical Logics: Theory, Applications and Tools Agata Ciabattoni Vienna University of Technology (TUV) Joint work with (TUV): M. Baaz, P. Baldi, B. Lellmann, R. Ramanayake,... N. Galatos (US), G.
More informationFrom Bi-facial Truth to Bi-facial Proofs
S. Wintein R. A. Muskens From Bi-facial Truth to Bi-facial Proofs Abstract. In their recent paper Bi-facial truth: a case for generalized truth values Zaitsev and Shramko [7] distinguish between an ontological
More informationProving Completeness for Nested Sequent Calculi 1
Proving Completeness for Nested Sequent Calculi 1 Melvin Fitting abstract. Proving the completeness of classical propositional logic by using maximal consistent sets is perhaps the most common method there
More informationMulti-valued Semantics: Why and How
A.Avron Multi-valued Semantics: Why and How Abstract. According to Suszko s Thesis, any multi-valued semantics for a logical system can be replaced by an equivalent bivalent one. Moreover: bivalent semantics
More informationRasiowa-Sikorski proof system for the non-fregean sentential logic SCI
Rasiowa-Sikorski proof system for the non-fregean sentential logic SCI Joanna Golińska-Pilarek National Institute of Telecommunications, Warsaw, J.Golinska-Pilarek@itl.waw.pl We will present complete and
More information185.A09 Advanced Mathematical Logic
185.A09 Advanced Mathematical Logic www.volny.cz/behounek/logic/teaching/mathlog13 Libor Běhounek, behounek@cs.cas.cz Lecture #1, October 15, 2013 Organizational matters Study materials will be posted
More informationCHAPTER 11. Introduction to Intuitionistic Logic
CHAPTER 11 Introduction to Intuitionistic Logic Intuitionistic logic has developed as a result of certain philosophical views on the foundation of mathematics, known as intuitionism. Intuitionism was originated
More informationSystematic Construction of Natural Deduction Systems for Many-valued Logics: Extended Report
Systematic Construction of Natural Deduction Systems for Many-valued Logics: Extended Report Matthias Baaz Christian G. Fermüller Richard Zach May 1, 1993 Technical Report TUW E185.2 BFZ.1 93 long version
More informationOn Sequent Calculi for Intuitionistic Propositional Logic
On Sequent Calculi for Intuitionistic Propositional Logic Vítězslav Švejdar Jan 29, 2005 The original publication is available at CMUC. Abstract The well-known Dyckoff s 1992 calculus/procedure for intuitionistic
More informationLabelled Calculi for Łukasiewicz Logics
Labelled Calculi for Łukasiewicz Logics D. Galmiche and Y. Salhi LORIA UHP Nancy 1 Campus Scientifique, BP 239 54 506 Vandœuvre-lès-Nancy, France Abstract. In this paper, we define new decision procedures
More informationAn Introduction to Proof Theory
An Introduction to Proof Theory Class 1: Foundations Agata Ciabattoni and Shawn Standefer anu lss december 2016 anu Our Aim To introduce proof theory, with a focus on its applications in philosophy, linguistics
More informationCut-free Ordinary Sequent Calculi for Logics Having Generalized Finite-Valued Semantics
Cut-free Ordinary Sequent Calculi for Logics Having Generalized Finite-Valued Semantics Arnon Avron, Jonathan Ben-Naim and Beata Konikowska 1. Introduction For at least seven decades, many-valued logics
More informationEquivalents of Mingle and Positive Paradox
Eric Schechter Equivalents of Mingle and Positive Paradox Abstract. Relevant logic is a proper subset of classical logic. It does not include among itstheoremsanyof positive paradox A (B A) mingle A (A
More informationClassical Propositional Logic
The Language of A Henkin-style Proof for Natural Deduction January 16, 2013 The Language of A Henkin-style Proof for Natural Deduction Logic Logic is the science of inference. Given a body of information,
More informationFrom Constructibility and Absoluteness to Computability and Domain Independence
From Constructibility and Absoluteness to Computability and Domain Independence Arnon Avron School of Computer Science Tel Aviv University, Tel Aviv 69978, Israel aa@math.tau.ac.il Abstract. Gödel s main
More informationFrom Frame Properties to Hypersequent Rules in Modal Logics
From Frame Properties to Hypersequent Rules in Modal Logics Ori Lahav School of Computer Science Tel Aviv University Tel Aviv, Israel Email: orilahav@post.tau.ac.il Abstract We provide a general method
More informationKRIPKE S THEORY OF TRUTH 1. INTRODUCTION
KRIPKE S THEORY OF TRUTH RICHARD G HECK, JR 1. INTRODUCTION The purpose of this note is to give a simple, easily accessible proof of the existence of the minimal fixed point, and of various maximal fixed
More informationand Proof Theory Arnon Avron which had its beginning in the work of Lukasiewicz [Luk]. Recently there is
Natural 3-valued Logics Characterization and Proof Theory Arnon Avron 1 Introduction Many-valued logics in general and 3-valued logic in particular is an old subject which had its beginning in the work
More informationFinite-valued Logics for Information Processing
Fundamenta Informaticae XXI (2001) 1001 1028 1001 IOS Press Finite-valued Logics for Information Processing Arnon Avron School of Computer Science, Tel-Aviv University Tel-Aviv, Israel aa@tau.ac.il Beata
More informationTR : Binding Modalities
City University of New York (CUNY) CUNY Academic Works Computer Science Technical Reports Graduate Center 2012 TR-2012011: Binding Modalities Sergei N. Artemov Tatiana Yavorskaya (Sidon) Follow this and
More informationMONADIC FRAGMENTS OF INTUITIONISTIC CONTROL LOGIC
Bulletin of the Section of Logic Volume 45:3/4 (2016), pp. 143 153 http://dx.doi.org/10.18778/0138-0680.45.3.4.01 Anna Glenszczyk MONADIC FRAGMENTS OF INTUITIONISTIC CONTROL LOGIC Abstract We investigate
More informationSciences, St Andrews University, St Andrews, Fife KY16 9SS, Scotland,
A Deterministic Terminating Sequent Calculus for Godel-Dummett logic ROY DYCKHOFF, School of Mathematical and Computational Sciences, St Andrews University, St Andrews, Fife KY16 9SS, Scotland, rd@dcs.st-and.ac.uk
More informationPrefixed Tableaus and Nested Sequents
Prefixed Tableaus and Nested Sequents Melvin Fitting Dept. Mathematics and Computer Science Lehman College (CUNY), 250 Bedford Park Boulevard West Bronx, NY 10468-1589 e-mail: melvin.fitting@lehman.cuny.edu
More informationTaming Implications in Dummett Logic
Taming Implications in Dummett Logic Guido Fiorino Dipartimento di Metodi Quantitativi per le cienze Economiche ed Aziendali, Università di Milano-Bicocca, Piazza dell Ateneo Nuovo, 1, 20126 Milano, Italy.
More informationMultiplicative Conjunction as an. ARNON AVRON, School of Mathematical Sciences, Sackler Faculty of
Multiplicative Conjunction as an Extensional Conjunction ARNON AVRON, School of Mathematical Sciences, Sackler Faculty of Exact Sciences, Tel Aviv University, Tel Aviv 69978, Israel. E-mail: aa@math.tau.ac.il
More informationHypersequent calculi for non classical logics
Tableaux 03 p.1/63 Hypersequent calculi for non classical logics Agata Ciabattoni Technische Universität Wien, Austria agata@logic.at Tableaux 03 p.2/63 Non classical logics Unfortunately there is not
More informationIntroduction to Logic in Computer Science: Autumn 2006
Introduction to Logic in Computer Science: Autumn 2006 Ulle Endriss Institute for Logic, Language and Computation University of Amsterdam Ulle Endriss 1 Plan for Today The first part of the course will
More informationcse371/mat371 LOGIC Professor Anita Wasilewska Fall 2018
cse371/mat371 LOGIC Professor Anita Wasilewska Fall 2018 Chapter 7 Introduction to Intuitionistic and Modal Logics CHAPTER 7 SLIDES Slides Set 1 Chapter 7 Introduction to Intuitionistic and Modal Logics
More informationA Triple Correspondence in Canonical Calculi: Strong Cut-Elimination, Coherence, and Non-deterministic Semantics
A Triple Correspondence in Canonical Calculi: Strong Cut-Elimination, Coherence, and Non-deterministic Semantics Arnon Avron and Anna Zamansky School of Computer Science, Tel-Aviv University Abstract.
More informationIntroduction to Metalogic
Philosophy 135 Spring 2008 Tony Martin Introduction to Metalogic 1 The semantics of sentential logic. The language L of sentential logic. Symbols of L: Remarks: (i) sentence letters p 0, p 1, p 2,... (ii)
More informationOn the Construction of Analytic Sequent Calculi for Sub-classical Logics
On the Construction of Analytic Sequent Calculi for Sub-classical Logics Ori Lahav Yoni Zohar Tel Aviv University WoLLIC 2014 On the Construction of Analytic Sequent Calculi for Sub-classical Logics A
More informationResolution for mixed Post logic
Resolution for mixed Post logic Vladimir Komendantsky Institute of Philosophy of Russian Academy of Science, Volkhonka 14, 119992 Moscow, Russia vycom@pochtamt.ru Abstract. In this paper we present a resolution
More informationPropositional Logic Language
Propositional Logic Language A logic consists of: an alphabet A, a language L, i.e., a set of formulas, and a binary relation = between a set of formulas and a formula. An alphabet A consists of a finite
More informationFormulas for which Contraction is Admissible
Formulas for which Contraction is Admissible ARNON AVRON, Department of Computer Science, School of Mathematical Sciences, Tel Aviv University, Tel Aviv 69978, Israel. E-mail: aa@math.tau.ac.il Abstract
More informationRestricted truth predicates in first-order logic
Restricted truth predicates in first-order logic Thomas Bolander 1 Introduction It is well-known that there exist consistent first-order theories that become inconsistent when we add Tarski s schema T.
More informationKamila BENDOVÁ INTERPOLATION AND THREE-VALUED LOGICS
REPORTS ON MATHEMATICAL LOGIC 39 (2005), 127 131 Kamila BENDOVÁ INTERPOLATION AND THREE-VALUED LOGICS 1. Three-valued logics We consider propositional logic. Three-valued logics are old: the first one
More informationRELATIONS BETWEEN PARACONSISTENT LOGIC AND MANY-VALUED LOGIC
Bulletin of the Section of Logic Volume 10/4 (1981), pp. 185 190 reedition 2009 [original edition, pp. 185 191] Newton C. A. da Costa Elias H. Alves RELATIONS BETWEEN PARACONSISTENT LOGIC AND MANY-VALUED
More informationKLEENE LOGIC AND INFERENCE
Bulletin of the Section of Logic Volume 4:1/2 (2014), pp. 4 2 Grzegorz Malinowski KLEENE LOGIC AND INFERENCE Abstract In the paper a distinguished three-valued construction by Kleene [2] is analyzed. The
More informationA refined calculus for Intuitionistic Propositional Logic
A refined calculus for Intuitionistic Propositional Logic Mauro Ferrari 1, Camillo Fiorentini 2, Guido Fiorino 3 1 Dipartimento di Informatica e Comunicazione, Università degli Studi dell Insubria Via
More informationModal systems based on many-valued logics
Modal systems based on many-valued logics F. Bou IIIA - CSIC Campus UAB s/n 08193, Bellaterra, Spain fbou@iiia.csic.es F. Esteva IIIA - CSIC Campus UAB s/n 08193, Bellaterra, Spain esteva@iiia.csic.es
More informationOn interpolation in existence logics
On interpolation in existence logics Matthias Baaz and Rosalie Iemhoff Technical University Vienna, Wiedner Hauptstrasse 8-10, A-1040 Vienna, Austria baaz@logicat, iemhoff@logicat, http://wwwlogicat/people/baaz,
More informationDeductive Systems. Lecture - 3
Deductive Systems Lecture - 3 Axiomatic System Axiomatic System (AS) for PL AS is based on the set of only three axioms and one rule of deduction. It is minimal in structure but as powerful as the truth
More informationRelevance Logic as an Information-Based Logic. J. Michael Dunn Indiana University Bloomington
Relevance Logic as an Information-Based Logic J. Michael Dunn Indiana University Bloomington The Founders of Relevance Logic Ivan Orlov (1928) Calculus of Combatibility. Orlov s pioneering work was effectively
More informationTABLEAUX VARIANTS OF SOME MODAL AND RELEVANT SYSTEMS
Bulletin of the Section of Logic Volume 17:3/4 (1988), pp. 92 98 reedition 2005 [original edition, pp. 92 103] P. Bystrov TBLEUX VRINTS OF SOME MODL ND RELEVNT SYSTEMS The tableaux-constructions have a
More informationTableau Systems for Logics of Formal Inconsistency
Tableau Systems for Logics of Formal Inconsistency Walter A. Carnielli Centre for Logic and Epistemology, and Department of Philosophy State University of Campinas CLE/Unicamp, Campinas, Brazil João Marcos
More informationDeep Sequent Systems for Modal Logic
Deep Sequent Systems for Modal Logic Kai Brünnler abstract. We see a systematic set of cut-free axiomatisations for all the basic normal modal logics formed from the axioms t, b,4, 5. They employ a form
More information1. Tarski consequence and its modelling
Bulletin of the Section of Logic Volume 36:1/2 (2007), pp. 7 19 Grzegorz Malinowski THAT p + q = c(onsequence) 1 Abstract The famous Tarski s conditions for a mapping on sets of formulas of a language:
More information3 Propositional Logic
3 Propositional Logic 3.1 Syntax 3.2 Semantics 3.3 Equivalence and Normal Forms 3.4 Proof Procedures 3.5 Properties Propositional Logic (25th October 2007) 1 3.1 Syntax Definition 3.0 An alphabet Σ consists
More informationUniform Rules and Dialogue Games for Fuzzy Logics
Uniform Rules and Dialogue Games for Fuzzy Logics A. Ciabattoni, C.G. Fermüller, and G. Metcalfe Technische Universität Wien, A-1040 Vienna, Austria {agata,chrisf,metcalfe}@logic.at Abstract. We provide
More informationComputational Logic. Davide Martinenghi. Spring Free University of Bozen-Bolzano. Computational Logic Davide Martinenghi (1/30)
Computational Logic Davide Martinenghi Free University of Bozen-Bolzano Spring 2010 Computational Logic Davide Martinenghi (1/30) Propositional Logic - sequent calculus To overcome the problems of natural
More informationOn the duality of proofs and countermodels in labelled sequent calculi
On the duality of proofs and countermodels in labelled sequent calculi Sara Negri Department of Philosophy PL 24, Unioninkatu 40 B 00014 University of Helsinki, Finland sara.negri@helsinki.fi The duality
More informationON THE ATOMIC FORMULA PROPERTY OF HÄRTIG S REFUTATION CALCULUS
Takao Inoué ON THE ATOMIC FORMULA PROPERTY OF HÄRTIG S REFUTATION CALCULUS 1. Introduction It is well-known that Gentzen s sequent calculus LK enjoys the so-called subformula property: that is, a proof
More informationREPRESENTATION THEOREMS FOR IMPLICATION STRUCTURES
Wojciech Buszkowski REPRESENTATION THEOREMS FOR IMPLICATION STRUCTURES Professor Rasiowa [HR49] considers implication algebras (A,, V ) such that is a binary operation on the universe A and V A. In particular,
More informationNONSTANDARD MODELS AND KRIPKE S PROOF OF THE GÖDEL THEOREM
Notre Dame Journal of Formal Logic Volume 41, Number 1, 2000 NONSTANDARD MODELS AND KRIPKE S PROOF OF THE GÖDEL THEOREM HILARY PUTNAM Abstract This lecture, given at Beijing University in 1984, presents
More informationPROPOSITIONAL MIXED LOGIC: ITS SYNTAX AND SEMANTICS
PROPOSITIONAL MIXED LOGIC: ITS SYNTAX AND SEMANTICS Karim NOUR 1 and Abir NOUR 2 Abstract In this paper, we present a propositional logic (called mixed logic) containing disjoint copies of minimal, intuitionistic
More informationIntroduction to Logic in Computer Science: Autumn 2006
Introduction to Logic in Computer Science: Autumn 2006 Ulle Endriss Institute for Logic, Language and Computation University of Amsterdam Ulle Endriss 1 Plan for Today Today s class will be an introduction
More informationPart 1: Propositional Logic
Part 1: Propositional Logic Literature (also for first-order logic) Schöning: Logik für Informatiker, Spektrum Fitting: First-Order Logic and Automated Theorem Proving, Springer 1 Last time 1.1 Syntax
More informationTaming Implications in Dummett Logic
Taming Implications in Dummett Logic Guido Fiorino Dipartimento di Metodi Quantitativi per le Scienze Economiche ed Aziendali, Università di Milano-Bicocca, Piazza dell Ateneo Nuovo, 1, 20126 Milano, Italy.
More informationNon-Analytic Tableaux for Chellas s Conditional Logic CK and Lewis s Logic of Counterfactuals VC
Australasian Journal of Logic Non-Analytic Tableaux for Chellas s Conditional Logic CK and Lewis s Logic of Counterfactuals VC Richard Zach Abstract Priest has provided a simple tableau calculus for Chellas
More information6. Logical Inference
Artificial Intelligence 6. Logical Inference Prof. Bojana Dalbelo Bašić Assoc. Prof. Jan Šnajder University of Zagreb Faculty of Electrical Engineering and Computing Academic Year 2016/2017 Creative Commons
More informationA Resolution Method for Modal Logic S5
EPiC Series in Computer Science Volume 36, 2015, Pages 252 262 GCAI 2015. Global Conference on Artificial Intelligence A Resolution Method for Modal Logic S5 Yakoub Salhi and Michael Sioutis Université
More informationReasoning: From Basic Entailments. to Plausible Relations. Department of Computer Science. School of Mathematical Sciences. Tel-Aviv University
General Patterns for Nonmonotonic Reasoning: From Basic Entailments to Plausible Relations Ofer Arieli Arnon Avron Department of Computer Science School of Mathematical Sciences Tel-Aviv University Tel-Aviv
More informationMaximal Introspection of Agents
Electronic Notes in Theoretical Computer Science 70 No. 5 (2002) URL: http://www.elsevier.nl/locate/entcs/volume70.html 16 pages Maximal Introspection of Agents Thomas 1 Informatics and Mathematical Modelling
More informationTR : Tableaux for the Logic of Proofs
City University of New York (CUNY) CUNY Academic Works Computer Science Technical Reports Graduate Center 2004 TR-2004001: Tableaux for the Logic of Proofs Bryan Renne Follow this and additional works
More informationPropositional Logics and their Algebraic Equivalents
Propositional Logics and their Algebraic Equivalents Kyle Brooks April 18, 2012 Contents 1 Introduction 1 2 Formal Logic Systems 1 2.1 Consequence Relations......................... 2 3 Propositional Logic
More informationTableaux, Abduction and Truthlikeness RESEARCH REPORT
Section of Logic and Cognitive Science Institute of Psychology Adam Mickiewicz University in Poznań Mariusz Urbański Tableaux, Abduction and Truthlikeness RESEARCH REPORT Szamarzewskiego 89, 60-589 Poznań,
More informationPiotr Kulicki The Use of Axiomatic Rejection
Piotr Kulicki The Use of Axiomatic Rejection 1. Introduction The central point of logic is an entailment relation, which in different contexts is expressed as an implication in a system, syntactic derivability
More informationOn the Complexity of the Reflected Logic of Proofs
On the Complexity of the Reflected Logic of Proofs Nikolai V. Krupski Department of Math. Logic and the Theory of Algorithms, Faculty of Mechanics and Mathematics, Moscow State University, Moscow 119899,
More informationCHAPTER 10. Gentzen Style Proof Systems for Classical Logic
CHAPTER 10 Gentzen Style Proof Systems for Classical Logic Hilbert style systems are easy to define and admit a simple proof of the Completeness Theorem but they are difficult to use. By humans, not mentioning
More informationBounded Lukasiewicz Logics
Bounded Lukasiewicz Logics Agata Ciabattoni 1 and George Metcalfe 2 1 Institut für Algebra und Computermathematik, Technische Universität Wien Wiedner Haupstrasse 8-10/118, A-1040 Wien, Austria agata@logic.at
More informationKrivine s Intuitionistic Proof of Classical Completeness (for countable languages)
Krivine s Intuitionistic Proof of Classical Completeness (for countable languages) Berardi Stefano Valentini Silvio Dip. Informatica Dip. Mat. Pura ed Applicata Univ. Torino Univ. Padova c.so Svizzera
More informationLecture Notes on Sequent Calculus
Lecture Notes on Sequent Calculus 15-816: Modal Logic Frank Pfenning Lecture 8 February 9, 2010 1 Introduction In this lecture we present the sequent calculus and its theory. The sequent calculus was originally
More informationSubminimal Logics and Relativistic Negation
School of Information Science, JAIST March 2, 2018 Outline 1 Background Minimal Logic Subminimal Logics 2 Some More 3 Minimal Logic Subminimal Logics Outline 1 Background Minimal Logic Subminimal Logics
More informationA Sequent-Based Representation of Logical Argumentation
A Sequent-Based Representation of Logical Argumentation Ofer Arieli School of Computer Science, The Academic College of Tel-Aviv, Israel. oarieli@mta.ac.il Abstract. In this paper we propose a new presentation
More informationGeneral methods in proof theory for modal logic - Lecture 1
General methods in proof theory for modal logic - Lecture 1 Björn Lellmann and Revantha Ramanayake TU Wien Tutorial co-located with TABLEAUX 2017, FroCoS 2017 and ITP 2017 September 24, 2017. Brasilia.
More informationA NEW FOUR-VALUED APPROACH TO MODAL LOGIC JEAN-YVES BEZIAU. In this paper we present several systems of modal logic based on four-valued
Logique & Analyse 213 (2011), x x A NEW FOUR-VALUED APPROACH TO MODAL LOGIC JEAN-YVES BEZIAU Abstract In this paper several systems of modal logic based on four-valued matrices are presented. We start
More informationSIMPLE DECISION PROCEDURE FOR S5 IN STANDARD CUT-FREE SEQUENT CALCULUS
Bulletin of the Section of Logic Volume 45/2 (2016), pp. 125 140 http://dx.doi.org/10.18778/0138-0680.45.2.05 Andrzej Indrzejczak SIMPLE DECISION PROCEDURE FOR S5 IN STANDARD CUT-FREE SEQUENT CALCULUS
More informationFirst-Order Intuitionistic Logic with Decidable Propositional Atoms
First-Order Intuitionistic Logic with Decidable Propositional Atoms Alexander Sakharov alex@sakharov.net http://alex.sakharov.net Abstract First-order intuitionistic logic extended with the assumption
More informationOverview of Logic and Computation: Notes
Overview of Logic and Computation: Notes John Slaney March 14, 2007 1 To begin at the beginning We study formal logic as a mathematical tool for reasoning and as a medium for knowledge representation The
More informationFirst-Degree Entailment
March 5, 2013 Relevance Logics Relevance logics are non-classical logics that try to avoid the paradoxes of material and strict implication: p (q p) p (p q) (p q) (q r) (p p) q p (q q) p (q q) Counterintuitive?
More informationA SEQUENT SYSTEM OF THE LOGIC R FOR ROSSER SENTENCES 2. Abstract
Bulletin of the Section of Logic Volume 33/1 (2004), pp. 11 21 Katsumi Sasaki 1 Shigeo Ohama A SEQUENT SYSTEM OF THE LOGIC R FOR ROSSER SENTENCES 2 Abstract To discuss Rosser sentences, Guaspari and Solovay
More information