## Nonmonotonic Reasoning in the Framework of Situation Calculus (1991)

Venue: | Artificial Intelligence |

Citations: | 138 - 0 self |

### BibTeX

@ARTICLE{Baker91nonmonotonicreasoning,

author = {Andrew B. Baker},

title = {Nonmonotonic Reasoning in the Framework of Situation Calculus},

journal = {Artificial Intelligence},

year = {1991},

volume = {49},

pages = {5--23}

}

### Years of Citing Articles

### OpenURL

### Abstract

Most of the solutions proposed to the Yale shooting problem have either introduced new nonmonotonic reasoning methods (generally involving temporal priorities) or completely reformulated the domain axioms to represent causality explicitly. This paper presents a new solution based on the idea that since the abnormality predicate takes a situational argument, it is important for the meanings of the situations to be held constant across the various models being compared. This is accomplished by a simple change in circumscription policy: when Ab is circumscribed, Result (rather than Holds) is allowed to vary. In addition, we need an axiom ensuring that every consistent situation is included in the domain of discourse. Ordinary circumscription will then produce the intuitively correct answer. Beyond its conceptual simplicity, the solution proposed here has additional advantages over the previous approaches. Unlike the approach that uses temporal priorities, it can support reasoning backward...

### Citations

1586 | Some philosophical problems from the standpoint of artificial intelligence
- McCarthy, Hayes
- 1969
(Show Context)
Citation Context ...roduction The formalization of reasoning about change has proven to be a surprisingly difficult problem. Standard logics are inadequate for this task because of difficulties such as the frame problem =-=[14]-=-; nonmonotonic reasoning seems to be necessary. Unfortunately, as demonstrated by the Yale shooting problem of Hanks and McDermott [5], the straightforward use of standard nonmonotonic logics (such as... |

865 |
Circumscriptionâ€”a form of non-monotonic reasoning
- McCarthy
- 1980
(Show Context)
Citation Context ...nothing else changes. Part of the motivation behind the development of nonmonotonic reasoning was to formalize this notion, and thus to solve the frame problem; we will use McCarthy's circumscription =-=[12, 13]-=-. If A is a formula, P is a predicate, and Z is a tuple of predicates and functions, then the circumscription of P in A with Z varied is written as Circum(A; P; Z). This abbreviates a formula in secon... |

508 | Applications of circumscription to formalizing common-sense reasoning
- McCarthy
- 1986
(Show Context)
Citation Context ...nothing else changes. Part of the motivation behind the development of nonmonotonic reasoning was to formalize this notion, and thus to solve the frame problem; we will use McCarthy's circumscription =-=[12, 13]-=-. If A is a formula, P is a predicate, and Z is a tuple of predicates and functions, then the circumscription of P in A with Z varied is written as Circum(A; P; Z). This abbreviates a formula in secon... |

331 | Computing circumscription
- Lifschitz
- 1985
(Show Context)
Citation Context ...lusion sense). Besides P , only those predicates and functions in Z are allowed to vary during this minimization process. For a more extensive discussion of circumscription, the reader is referred to =-=[8]-=-. Consider the standard default frame axiom: 1 :Ab(f; a; s) oe (Holds(f; Result(a; s)) j Holds(f; s)): (1) This says that the value of a fluent persists from one situation to the next unless something... |

262 |
Non-monotonic Logic and Temporal Projection
- Hanks, McDermott
- 1987
(Show Context)
Citation Context ... for this task because of difficulties such as the frame problem [14]; nonmonotonic reasoning seems to be necessary. Unfortunately, as demonstrated by the Yale shooting problem of Hanks and McDermott =-=[5]-=-, the straightforward use of standard nonmonotonic logics (such as circumscription) for reasoning about action leads to counter-intuitive results. There have been a large number of solutions proposed ... |

145 | Reasoning about action I: A possible worlds approach
- Ginsberg, Smith
- 1987
(Show Context)
Citation Context ...ifications Often, it is impractical to list explicitly all the consequences of an action. Rather, some of these consequences will be ramifications; that is, they will be implied by domain constraints =-=[4]-=-. One of the main advantages of our method over causal minimization is that ours can handle ramifications, while causal minimization cannot. A simple example of this limitation of causal minimization ... |

123 |
Formal theories of action
- Lifschitz
- 1987
(Show Context)
Citation Context ...ill rule out this model also. 4 Sandewall [19] has proposed filter preferential entailment, a modification of chrono5 3.2 Causal minimization Another approach, developed by Haugh [6] and by Lifschitz =-=[9]-=-, is that of causal minimization. This method represents causality explicitly by stating that a fluent changes its value if and only if a successful action causes it to do so. The intuition here is th... |

91 | The logic of persistence
- Kautz
- 1986
(Show Context)
Citation Context ...arge number of solutions proposed to the shooting problem. This section discusses the two most popular groups of solutions. 3.1 Chronological minimization One idea, proposed in various forms by Kautz =-=[7]-=-, Lifschitz [10], and Shoham [20], is chronological minimization (the term is due to Shoham). This proposal claims that we should reason forward in time; that is, apply the default assumptions in temp... |

56 | A simple solution to the Yale Shooting problem
- Baker
- 1989
(Show Context)
Citation Context ..., we consider two additional scenarios in order to compare the various approaches to nonmonotonic temporal reasoning. Concluding remarks are contained in Section 7. This paper is a revised version of =-=[1]-=-. 2 The shooting problem The Yale shooting problem arises regardless of which temporal formalism is used; we will use the situation calculus [14]. A situation is the state of the world at a particular... |

50 |
Chronological ignorance: Experiments in nonmonotonic temporal reasoning
- Shoham
- 1988
(Show Context)
Citation Context ...d to the shooting problem. This section discusses the two most popular groups of solutions. 3.1 Chronological minimization One idea, proposed in various forms by Kautz [7], Lifschitz [10], and Shoham =-=[20]-=-, is chronological minimization (the term is due to Shoham). This proposal claims that we should reason forward in time; that is, apply the default assumptions in temporal order. So in the shooting sc... |

41 |
Preferential Entailment for the logic of action in almost continuous worlds
- Sandewall, Filter
- 1989
(Show Context)
Citation Context ...n is not a completely satisfactory solution. 4 3 The current formalization admits a third possibility: Fred might die during the waiting phase. Our solution will rule out this model also. 4 Sandewall =-=[19]-=- has proposed filter preferential entailment, a modification of chrono5 3.2 Causal minimization Another approach, developed by Haugh [6] and by Lifschitz [9], is that of causal minimization. This meth... |

35 | Simple Causal Minimizations for Temporal Persistence and Projection
- Haugh
- 1987
(Show Context)
Citation Context ...phase. Our solution will rule out this model also. 4 Sandewall [19] has proposed filter preferential entailment, a modification of chrono5 3.2 Causal minimization Another approach, developed by Haugh =-=[6]-=- and by Lifschitz [9], is that of causal minimization. This method represents causality explicitly by stating that a fluent changes its value if and only if a successful action causes it to do so. The... |

32 | Autoepistemic logic and formalization of commonsense reasoning - Gelfond - 1989 |

28 | The anomalous extension problem in default reasoning - Morris - 1988 |

25 | The persistence of derived information
- Myers, Smith
- 1988
(Show Context)
Citation Context ...es to location L 2 . Does the ice-cream cone's absolute position persist, or does the fact that it is being held by the robot persist? (equivalent scenarios have been discussed by many authors, e.g., =-=[4, 16, 18]-=-). Based on our understanding of the domain, we would probably guess that after the robot moves, it will continue to hold the ice-cream cone. Yet, if we simply write down the domain constraint: Holds(... |

6 | On logic and probability - Pearl - 1988 |

1 |
personal communication
- Ginsberg
- 1988
(Show Context)
Citation Context ...version of the Yale shooting 19 problem in which the gun was unloaded in the initial situation, and a Load action was performed before the waiting (the following example is based on one from Ginsberg =-=[3]-=-). Using the causal notation from [9], we assert that Load causes the gun to be loaded, and that Shoot causes Fred to be not alive (provided the appropriate preconditions are satisfied): Causes(Load; ... |

1 |
personal communication
- Lifschitz
- 1989
(Show Context)
Citation Context ...he four combinations of Alive and Loaded --- not to the eight combinations of Alive, Loaded, and the mystery fluent. Axiom (6) was not included in the preliminary version of this paper [1]; Lifschitz =-=[11]-=- pointed out that it was necessary. 5 Existence of situations In the last section, we violated the spirit of the nonmonotonic enterprise by adding an existence of situations axiom (7) that explicitly ... |

1 |
Multiple extensions, where is the problem
- Reinfrank
- 1987
(Show Context)
Citation Context ...es to location L 2 . Does the ice-cream cone's absolute position persist, or does the fact that it is being held by the robot persist? (equivalent scenarios have been discussed by many authors, e.g., =-=[4, 16, 18]-=-). Based on our understanding of the domain, we would probably guess that after the robot moves, it will continue to hold the ice-cream cone. Yet, if we simply write down the domain constraint: Holds(... |