Key Arguments

• Dreyfus generalizes from disambiguation to context recognition: 'If in disambiguation the number of possibly relevant facts is in some sense infinite so that selection criteria must be applied before interpretation can begin, the number of facts that might be relevant to recognizing a context is infinite too.'
• He illustrates the explosion of candidate features: 'How is the computer to consider all the features such as how many people are present, the temperature, the pressure, the day of the week, and so forth, anyone of which might be a defining feature of some context?'
• Even if rules for relevance are given, 'these facts would be ambiguous, that is, capable of defining several different contexts, until they were interpreted. Evidently, a broader context will have to be used to determine which of the infinity of features is relevant, and how each is to be understood.'
• Consequently, 'the programmer must either claim that some features are intrinsically relevant and have a fixed meaning regardless of contexta possibility already excluded in the original appeal to contextor the programmer will be faced with an infinite regress of contexts.'

Source Quotes

If computers must utilize the situation or context in order to disambiguate, and in general to understand utterances in a natural language, the programmer must be able to program into the machine, which is not involved in a situation, a way of recognizing a context and using it. But the same two problems which arose in disambiguation and necessitated appeal to the situation in the first place arise again on the level of context recognition and force us to envisage working down from the broadest context: (1) If in disambiguation the number of possibly relevant facts is in some sense infinite so that selection criteria must be applied before interpretation can begin, the number of facts that might be relevant to recognizing a context is infinite too. How is the computer to consider all the features such as how many people are present, the temperature, the pressure, the day of the week, and so forth, anyone of which might be a defining feature of some context?

But the same two problems which arose in disambiguation and necessitated appeal to the situation in the first place arise again on the level of context recognition and force us to envisage working down from the broadest context: (1) If in disambiguation the number of possibly relevant facts is in some sense infinite so that selection criteria must be applied before interpretation can begin, the number of facts that might be relevant to recognizing a context is infinite too. How is the computer to consider all the features such as how many people are present, the temperature, the pressure, the day of the week, and so forth, anyone of which might be a defining feature of some context? (2) Even if the program provides rules for determining relevant facts, these facts would be ambiguous, that is, capable of defining several different contexts, until they were interpreted.

How is the computer to consider all the features such as how many people are present, the temperature, the pressure, the day of the week, and so forth, anyone of which might be a defining feature of some context? (2) Even if the program provides rules for determining relevant facts, these facts would be ambiguous, that is, capable of defining several different contexts, until they were interpreted. Evidently, a broader context will have to be used to determine which of the infinity of features is relevant, and how each is to be understood.

Evidently, a broader context will have to be used to determine which of the infinity of features is relevant, and how each is to be understood. But if, in turn, the program must enable the machine to identify the broader context in terms of its relevant featuresand this is the only way a computer which operates in terms of discrete elements could proceedthe programmer must either claim that some features are intrinsically relevant and have a fixed meaning regardless of contexta possibility already excluded in the original appeal to contextor the programmer will be faced with an infinite regress of contexts. There seems to be only one way out: rather than work up the tree to ever broader contexts the computer must work down from an ultimate contextwhat Weizenbaum calls our shared culture.

Key Concepts

• the number of facts that might be relevant to recognizing a context is infinite too.
• How is the computer to consider all the features such as how many people are present, the temperature, the pressure, the day of the week, and so forth, anyone of which might be a defining feature of some context?
• these facts would be ambiguous, that is, capable of defining several different contexts, until they were interpreted.
• the programmer must either claim that some features are intrinsically relevant and have a fixed meaning regardless of contexta possibility already excluded in the original appeal to contextor the programmer will be faced with an infinite regress of contexts.

Context

In Dreyfus’s analysis of 'situation recognition', expanding the earlier disambiguation problem into a general argument about the impossibility of fact‑based context recognition without regress or arbitrary privileged features.