Chapter 7. Computers and representation

7.1 Programming as representation

7.2 Levels of representation

7.3 Can computers do more than you tell them to do?

p 92

It is possible that we might (either accidentally or intentionally) endow a machine with essential qualities we do not anticipate.

Chapter 8. Computation and intelligence

p 93

The theoretical principles developed in the earlier chapters lead us to the conclusion that one cannot program computers to be intelligent and that we need to look in different directions for the design of powerful computer technology.

8.1 Why do we ask?

Complexity of purpose

p 94

Each detail may be the result of an evolved compromise between many conflicting demands. At times, the only explanation for the system's current form may be an appeal to this history of modification.

Unpredictability

p 95

[...] in a complex computer system designed and functioning properly at a higher level there is often no way to predict how a program will act, short of running it.

8.2 Intelligence as rational problem solving

8.3 The phenomenon of blindness

p 98

We accuse people of lacking common sense precisely when some representation of the situation has blinded them to a space of potentially relevant action.

p 99

Human cognition includes the use of representations, but is not based on representations.

8.4 What about learning and evolution?

Evolution of structure

p 103

(the speed of evolution is driven by the speed of the coupling) It is highly unlikely that any system we can build will be able to undergo this kind of evolutionary change (or learning) that would enable it to come close to the intelligence of even a small worm, much less that of a person.

[Not more conclusive than that!]

8.5 Can pigs have wings?

p 104

Analytic (rationalistic) tradition: clear criteria that would be needed to design objective experiments.

p 106

We treat other people not as merely 'rational beings' but as 'responsible beings'. An essential part of being human is the ability to enter into commitments and to be responsible for the course of action that they anticipate.

Chapter 9. Understanding language

9.1 Artificial intelligence and language understanding

9.2 The problem of background

p 112

The question "Is X a bachelor?" cannot be answered without considering the potential answers to "Why do you want to know?" This is what Gadamer means by "A question behind each statement that first gives it its meaning."

9.3 Understanding as pattern recognition

p 115

The artificial intelligence literature of the 1970s heralded a move away from the traditional problem-solving orientation towards a new centered around 'frames' and 'expectations'.

9.4 What does it mean to understand?

Chapter 10. Current directions in artificial intelligence

10.1 The forking of the paths

10.2 Expert systems

p. 132

Calling a program an 'expert' is misleading in exactly the same way as calling it 'intelligent' or saying it 'understands.' The misrepresentation may be useful for those who are trying to get research funding or sell such programs, but it may lead to inappropriate expectations by those who attempt to use them.

10.3 The fifth generation computer system