-->
The Frame Problem (Stanford Encyclopedia of Philosophy)
Stanford Encyclopedia of Philosophy
Menu
Browse
Table of Contents
What's New
Random Entry
Chronological
Archives
About
Editorial Information
About the SEP
Editorial Board
How to Cite the SEP
Special Characters
Advanced Tools
Contact
Support SEP
Support the SEP
PDFs for SEP Friends
Make a Donation
SEPIA for Libraries
Entry Navigation
Entry Contents
Bibliography
Academic Tools
Friends PDF Preview
Author and Citation Info
Back to Top
The Frame Problem<br>First published Mon Feb 23, 2004; substantive revision Mon Feb 8, 2016
To most AI researchers, the frame problem is the challenge of<br>representing the effects of action in logic without having to represent<br>explicitly a large number of intuitively obvious non-effects. But to many<br>philosophers, the AI researchers' frame problem is suggestive of<br>wider epistemological issues. Is it possible, in principle, to limit<br>the scope of the reasoning required to derive the consequences of an<br>action? And, more generally, how do we account for our apparent ability<br>to make decisions on the basis only of what is relevant to an ongoing<br>situation without having explicitly to consider all that is not<br>relevant?
1. Introduction
2. The Frame Problem in Logic
3. The Epistemological Frame Problem
4. The Metaphysics of Common Sense Inertia
5. The Frame Problem Today
Bibliography
Academic Tools
Other Internet Resources
Related Entries
1. Introduction
The frame problem originated as a narrowly defined technical problem<br>in<br>logic-based artificial intelligence<br>(AI).<br>But it was taken up in an embellished and modified form by<br>philosophers of mind, and given a wider interpretation. The tension<br>between its origin in the laboratories of AI researchers and its<br>treatment at the hands of philosophers engendered an interesting<br>and sometimes heated debate in the 1980s and 1990s.<br>But since the narrow, technical problem is largely solved, recent<br>discussion has tended to focus less on matters of interpretation and<br>more on the implications of the wider frame problem for<br>cognitive science.<br>To gain an understanding of the issues,<br>this article will begin with a look at the frame problem in its<br>technical guise. Some of the ways in which philosophers have<br>re-interpreted the problem will then be examined. The article<br>will conclude with an assessment of the significance of the frame<br>problem today.
2. The Frame Problem in Logic
Put succinctly, the frame problem in its narrow, technical form is<br>this (McCarthy & Hayes 1969). Using mathematical logic, how is it<br>possible to write formulae that describe the effects of actions without<br>having to write a large number of accompanying formulae that describe<br>the mundane, obvious non-effects of those actions? Let's take a look at<br>an example. The difficulty can be illustrated without the full<br>apparatus of formal logic, but it should be borne in mind that the<br>devil is in the mathematical details. Suppose we write two formulae,<br>one describing the effects of painting an object and the other<br>describing the effects of moving an object.
Colour(x, c) holds after<br>Paint(x, c)
Position(x, p) holds after<br>Move(x, p)
Now, suppose we have an initial situation in which<br>Colour(A, Red)<br>and Position(A, House) hold. According to<br>the machinery of deductive logic, what then holds after the action<br>Paint(A, Blue) followed by the<br>action Move(A, Garden)? Intuitively, we<br>would expect Colour(A, Blue) and<br>Position(A, Garden) to hold. Unfortunately,<br>this is not the case. If written out more formally in classical<br>predicate logic, using a suitable formalism for representing time and<br>action such as the situation calculus (McCarthy & Hayes 1969), the<br>two formulae above only license the conclusion<br>that Position(A, Garden) holds. This is<br>because they don't rule out the possibility that the colour of<br>A gets changed by the Move action.
The most obvious way to augment such a formalisation so that the<br>right common sense conclusions fall out is to add a number of formulae<br>that explicitly describe the non-effects of each action. These formulae<br>are called frame axioms. For the example at hand, we need a<br>pair of frame axioms.
Colour(x, c) holds after Move(x, p) if Colour(x, c) held<br>beforehand
Position(x, p) holds after Paint(x, c) if Position(x, p) held<br>beforehand
In other words, painting an object will not affect its position, and<br>moving an object will not affect its colour. With the addition of these<br>two formulae (written more formally in predicate logic), all the<br>desired conclusions can be drawn. However, this is not at all a<br>satisfactory solution. Since most actions do not affect<br>most properties of a situation, in a domain comprising<br>M actions and N properties we will, in general, have<br>to write out almost MN frame axioms. Whether these formulae<br>are destined to be stored explicitly in a computer's memory, or are<br>merely part of the designer's specification, this is an unwelcome<br>burden.
The challenge, then,...