## A Logic of Argumentation for Reasoning under Uncertainty. (1995)

Venue: | Computational Intelligence |

Citations: | 107 - 3 self |

### BibTeX

@ARTICLE{Krause95alogic,

author = {Paul Krause and Simon Ambler and Morten Elvang-Gøransson and John Fox},

title = {A Logic of Argumentation for Reasoning under Uncertainty.},

journal = {Computational Intelligence},

year = {1995},

volume = {11},

pages = {113--131}

}

### Years of Citing Articles

### OpenURL

### Abstract

We present the syntax and proof theory of a logic of argumentation, LA. We also outline the development of a category theoretic semantics for LA. LA is the core of a proof theoretic model for reasoning under uncertainty. In this logic, propositions are labelled with a representation of the arguments which support their validity. Arguments may then be aggregated to collect more information about the potential validity of the propositions of interest. We make the notion of aggregation primitive to the logic, and then define strength mappings from sets of arguments to one of a number of possible dictionaries. This provides a uniform framework which incorporates a number of numerical and symbolic techniques for assigning subjective confidences to propositions on the basis of their supporting arguments. These aggregation techniques are also described, with examples. Key words: Uncertain reasoning, epistemic probability, argumentation, non-classical logics, non-monotonic reasoning 1. Introd...

### Citations

584 |
The Uses of Argument
- Toulmin
- 1958
(Show Context)
Citation Context ...se. With respect to the word "probable", consider its use in the context of epistemic probability; or perhaps in terms of the intuitive description used by Toulmin in his discussion of argum=-=entation (Toulmin 1956): "W-=-hen I say `S is probably P', I commit myself guardedly, tentatively or with reservations to the view that S is P, and (likewise guardedly) lend my authority to that view." We will not have ground... |

424 |
Introduction to Higher Order Categorical Logic
- Lambek, Scott
- 1986
(Show Context)
Citation Context ...set of `objects' and a set of `arrows'. In particular, a deductive systemsmay be modelled as a category where the objects are propositions (A, B, ...) and the arrows (f, g, ...) correspond to proofs (=-=Lambek and Scott 1986). F-=-or example, the arrow f : A �� B corresponds to a proof of A B. We may compose arrows, so that if f corresponds to a proof of B from A, and g corresponds to a proof of C from B, then the compositi... |

377 | Basic concepts of enriched category theory, volume 64
- Kelly
- 1982
(Show Context)
Citation Context ...er which a deductive system is a category can be found in, for example, (Lambek and Scott 1986). In the case of LA, (sets of) arguments are modelled as the arrows of a semi-lattice enriched category (=-=Kelly 1982). The epi-=-thet "semi-lattice enriched" means that the category structure is enriched with some ordering information on the arrows. In the case of intuitionistic logic, there is a well-trodden route to... |

352 |
Proofs and Types
- Girard, Lafont, et al.
- 1989
(Show Context)
Citation Context ...s a well-trodden route to the development of a categorical semantics. The link we have used in Section 3 between proofs and typed lambda-terms is formally established by the Curry-Howard isomorphism (=-=Girard et al. 1989-=-). Then the relationship between typed l-calculus and Cartesian closed categories, (Lambek and Scott 1986), completes the mapping of formulae to the objects, and (equivalence classes of) proofs to the... |

302 | Lambda calculus notation with nameless dummies, a tool for automatic formula manipulation, with application to the Church-Rosser theorem - Bruijn - 1972 |

173 |
Labeled Deductive Systems
- Gabbay
- 1991
(Show Context)
Citation Context ...al is not so much to generate a new logic as to augment an existing logic; we wish to annotate propositions with a representation of their proofs. That is, we wish to use a labelled deductive system (=-=Gabbay 1990-=-), where the labels represent the arguments supporting the subject of the deduction. What logic should we use to represent the underlying reasoning process? An idealised reasoning agent may construct ... |

157 | Inconsistency management and prioritized syntax-based entailment
- Benferhat, Cayrol, et al.
- 1993
(Show Context)
Citation Context ... argument. There has been a steady interest in developing models for reasoning in the presence of inconsistent data in both the AI (Dubois, Lang and Prade 1992; Fox et al. 1992; Pinkas and Loui 1992; =-=Benferhat, et al. 1993-=-a; 1993b) and philosophical logic (Nelson 1949; Priest 1987; Priest et al. 1988) communities. From a more pragmatic perspective, the sheer impracticality of ensuring consistency in large scale deducti... |

136 | The robust beauty of improper linear models in decision making
- Dawes
- 1982
(Show Context)
Citation Context ...results in a very simple theory preference model. The second case, the natural numbers augmented with infinity, corresponds to a slight refinement of the improper linear model with uniform weighting (=-=Dawes 1979-=-); basically the counting up of arguments. There are several approaches which may be taken to defining a mapping to the unit interval which we can outline here. A rigorous discussion of them can be fo... |

132 |
Contraction-free Sequent Calculi for Intuitionistic Logic
- Dyckhoff
- 1992
(Show Context)
Citation Context ...rate some form of loop checking into the theorem prover. Or one can use a modified rule set which is known to allow the same set of theorems, but which avoids the use of the problematic rule `�� �=-=���' (Dyckhoff 1992-=-). We have versions of the `ATP' which use each of these approaches. One needs to do a bit more work than just to apply the rule set in its raw form. The theorem prover should return all distinct argu... |

101 | On the Comparison of Theories: Preferring the Most Specific Explanation - Poole - 1985 |

98 | How to reason defeasibly - Pollock - 1992 |

91 |
An introduction to possibilistic and fuzzy logics
- Dubois, Prade
- 1990
(Show Context)
Citation Context ...cribed in section 5. However, to provide further insight into the properties of our system of dialectic argumentation we will look to see to what extent it satisfies the axioms of possibility theory (=-=Dubois and Prade 1988-=-a). Dubois and Prade (1988a) consider these to be the weakest axioms that a general confidence evaluation function should satisfy to ensure a minimum of coherence. Since any class of propositions, C x... |

77 |
An introduction
- Intuitionism
- 1956
(Show Context)
Citation Context ... substitute "(idealised) mathematician" and we have a situation precisely analogous to the starting point of intuitionistic logic. That is, the logic underlying argumentation is intuitionist=-=ic logic (Heyting 1956). I-=-n fact, attention will be restricted to the &, ��, ^ (and, implication, contradiction) fragment of minimal logic. Note that minimal logic is weaker than intuitionistic logic in that it does not ha... |

75 |
Constructible falsity
- Nelson
- 1949
(Show Context)
Citation Context ... models for reasoning in the presence of inconsistent data in both the AI (Dubois, Lang and Prade 1992; Fox et al. 1992; Pinkas and Loui 1992; Benferhat, et al. 1993a; 1993b) and philosophical logic (=-=Nelson 1949-=-; Priest 1987; Priest et al. 1988) communities. From a more pragmatic perspective, the sheer impracticality of ensuring consistency in large scale deductive databases has also led the logic programmin... |

47 |
Logic for computer science
- Gallier
- 1987
(Show Context)
Citation Context ... and their representation as terms in the typed l-calculus. A sequent is usually written as an expression A �� B; where �� represents an entailment relation, A and B are finite sequences of fo=-=rmulae (Gallier 1987). A-=- naive interpretation of a sequent is that the conjunction of the formulae in A implies the disjunction of the formulae in B. In our presentation, the sequents are of the form: Context �� Term: Fo... |

46 |
Arguments, contradictions, and practical reasoning
- Fox, Krause, et al.
- 1992
(Show Context)
Citation Context ...ions with the arguments which support their validity is that we wish to develop purely symbolic methods of reasoning about arguments. This is a long standing interest of the authors (Fox et al. 1988; =-=Fox et al. 1992), th-=-at linguistic confidence terms can be defined on purely logical grounds over sets of arguments. In this section we will discuss the assignment of propositions to one of a number of "acceptability... |

46 | Knowledge in Flux: Modelling the dynamics of epistemic states - G��rdenfors - 1988 |

28 | Defeasible reasoning and decision support systems - Nute - 1988 |

27 |
Non-additive probabilities in the work of Bernoulli and Lambert. Archive for History of Exact Sciences
- Shafer
- 1978
(Show Context)
Citation Context ...n reasoning, epistemic probability, argumentation, non-classical logics, non-monotonic reasoning 1. Introduction The work presented in this paper should be viewed as a study of epistemic probability (=-=Shafer 1978-=-). That is, of the probability of an event, or confidence in a proposition, as an attribute of opinion based on an appraisal of the arguments concerning that event or proposition. Epistemic probabilit... |

24 |
Reasoning from inconsistency: A taxonomy of principles for resolving conflict
- Pinkas, Loui
- 1992
(Show Context)
Citation Context ... used to construct the argument. There has been a steady interest in developing models for reasoning in the presence of inconsistent data in both the AI (Dubois, Lang and Prade 1992; Fox et al. 1992; =-=Pinkas and Loui 1992-=-; Benferhat, et al. 1993a; 1993b) and philosophical logic (Nelson 1949; Priest 1987; Priest et al. 1988) communities. From a more pragmatic perspective, the sheer impracticality of ensuring consistenc... |

24 |
Ex contradictione nihil sequitur
- Wagner
- 1991
(Show Context)
Citation Context ...cality of ensuring consistency in large scale deductive databases has also led the logic programming community to consider consequence relations which are tolerant of varying levels of inconsistency (=-=Wagner 1991-=-). In this section we will describe a system which actually encourages the use of inconsistency to provide a logical model of a form of dialectical argumentation. Given an inconsistent knowledge-base ... |

22 |
Making Inconsistency Respectable: A Logical Framework for Inconsistency
- Gabbay, Hunter
- 1992
(Show Context)
Citation Context ...should be resolved; the significance of a level of contradiction may differ in different problem solving contexts. Reasoning about and with contradiction is an active research topic (Fox et al. 1992; =-=Gabbay and Hunter 1991-=-). The work reported in section 6 provides one strategy for classifying propositions on the basis of the level of inconsistency in their supporting arguments. This is a topic of ongoing work which is ... |

16 | Self-defeating arguments
- Pollock
- 1991
(Show Context)
Citation Context ...of inconsistency in their supporting arguments. This is a topic of ongoing work which is beginning to establish a fruitful connection with much recent work in philosophical logic (see also Loui 1991; =-=Pollock 1991-=-; Rescher 1977). However, there is a limitation to the extent to which we can currently develop this approach. The work reported in this paper is founded on the established relationship between proofs... |

15 | An argumentation framework in default logic - Prakken - 1993 |

12 |
Inconsistency in possibilistic knowledge bases - to live or not to live with it
- Dubois, Lang, et al.
- 1991
(Show Context)
Citation Context ...ch as "it is possible that ...", "it is plausible that ...", "presumably ...". The usual approach is to try and map the selected terms to fuzzy modifiers, (Zadeh 1975), o=-=r to probabilistic intervals (Dubois et al. 1992-=-). However, we believe that there is an alternative non-numerical approach which may be taken to defining such terms. In (Fox 1984) it was suggested that such qualifiers may be interpreted in terms of... |

11 |
On inference from inconsistent premises
- Rescher, Manor
- 1970
(Show Context)
Citation Context ...ntially, a proposition is confirmed if no consistent case can be made against it, or the argument which supports it. It turns out that confirmed propositions correspond to the innocent bystanderssof (=-=Rescher and Manor 1970). That is-=-, they can be viewed as having no role to play in the derivation of contradiction from the database. With respect to the word "probable", consider its use in the context of epistemic probabi... |

10 |
A categorial approach to the semantics of argumentation
- Ambler
- 1996
(Show Context)
Citation Context ...idespread. So a full presentation of the semantics would unbalance and reduce the accessibility of this paper. Consequently, we only outline this work here. A more complete treatment can be found in (=-=Ambler 1992-=-). Page 8 5.2 Aggregation of arguments and confidence measures. A category C has an associated set of `objects' and a set of `arrows'. In particular, a deductive systemsmay be modelled as a category w... |

9 | Dialectic reasoning with inconsistent information - Elvang-Gransson, Krause, et al. - 1993 |

6 |
A symbolic approach to reasoning with linguistic quantifiers
- Dubois, Prade, et al.
- 1992
(Show Context)
Citation Context ...ch as "it is possible that ...", "it is plausible that ...", "presumably ...". The usual approach is to try and map the selected terms to fuzzy modifiers, (Zadeh 1975), o=-=r to probabilistic intervals (Dubois et al. 1992-=-). However, we believe that there is an alternative non-numerical approach which may be taken to defining such terms. In (Fox 1984) it was suggested that such qualifiers may be interpreted in terms of... |

6 |
Introduction to Combinators and l-Calculus
- Hindley, Seldin
- 1986
(Show Context)
Citation Context ...al theory over proofs has received a great deal of thought from logicians, and there is a well established body of work on typed proof theory which may be applied to this problem (Girardset al. 1989; =-=Hindley and Seldin 1986-=-). We use typed lambda terms as proof terms. This gives us an equational theory over proof terms which has a formal agreed basis, so that we can say when two proofs are equal (are reducible to the sam... |

6 |
Justification, computational efficiency and generalisation of the Dempster-Shafer theory
- Wilson
- 1989
(Show Context)
Citation Context ...ngth" mapping s valued in [0,1]. If we consider all possible arguments for a proposition A and take the supremum of their strengths, then we obtain a "probability of provability " for A=-=. As noted in (Wilson 1989-=-), this gives a generalisation of Dempster-Shafer belief to non-classical logic. 5.3 An example of a probability valuation. A simple example will illustrate the form of the dependency analysis which i... |

3 |
Decision making as a logical process
- Fox, Glowinski, et al.
- 1989
(Show Context)
Citation Context ...abelling propositions with the arguments which support their validity is that we wish to develop purely symbolic methods of reasoning about arguments. This is a long standing interest of the authors (=-=Fox et al. 1988-=-; Fox et al. 1992), that linguistic confidence terms can be defined on purely logical grounds over sets of arguments. In this section we will discuss the assignment of propositions to one of a number ... |

2 |
Dialectics: a controversy- oriented approach to the theory of knowledge
- Rescher
- 1977
(Show Context)
Citation Context ...cy in their supporting arguments. This is a topic of ongoing work which is beginning to establish a fruitful connection with much recent work in philosophical logic (see also Loui 1991; Pollock 1991; =-=Rescher 1977-=-). However, there is a limitation to the extent to which we can currently develop this approach. The work reported in this paper is founded on the established relationship between proofs in minimal lo... |

1 | Argumentative inference in uncertain and inconsistent knowledge bases - D, Prade - 1993 |

1 | The Development of a "Logic of Argumentation - J, Fox - 1993 |

1 |
Argument and Belief: Where We Stand
- Loui
- 1991
(Show Context)
Citation Context ... the level of inconsistency in their supporting arguments. This is a topic of ongoing work which is beginning to establish a fruitful connection with much recent work in philosophical logic (see also =-=Loui 1991-=-; Pollock 1991; Rescher 1977). However, there is a limitation to the extent to which we can currently develop this approach. The work reported in this paper is founded on the established relationship ... |