rem%UCLA-LOCUS@houxa.UUCP (11/08/83)
THE MUELLER MEASURE If an AI could be built to answer all questions we ask it to assure us that it is ideally human (the Turing Test), it ought to be smart enough to figure out questions to ask itself that would prove that it is indeed artificial. Put another way: If an AI could make humans think it is smarter than a human by answering all questions posed to it in a Turing-like manner, it still is dumber than a human because it could not ask questions of a human to make us answer the questions so that it satisfies its desire for us to make it think we are more artificial than it is. Again: If we build an AI so smart it can fool other people by answering all questions in the Turing fashion, can we build a computer, anti-Turing-like, that could make us answer questions to fool other machines into believing we are artificial? Robert E. Mueller, Bell Labs, Holmdel, New Jersey houxa!rem
STORY%MIT-MC@sri-unix.UUCP (02/02/84)
This message is empty.
"J.R.COWIE%rco"@ucl-cs.arpa (07/10/84)
Use of if in natural language:
The following is a brief description of the project proposal by one
of our students on a M.Sc. course in Information Technology.
This student is originally a philosopher by profession, but has
decided to move over into Computer Science. He is interested in
using prolog to test out his ideas.
If you have any suggestions or references send them to me and I will pass them
on to him. (j.r.cowie%rco@ucl-cs.arpa)
---------------------------------------------
It is arguable that contraposition is not a universally
valid principle of inference for empirical conditionals and
yet we use it, apparently successfully, all the time. An
obvious suggestion is that we are discriminating and select
a subclass of cases to contrapose. We then ask what
characterizes that subclass.
The approach to be adopted attempts to isolate several
components of a conditional 1) a truth-functional component
2) an inferential component 3) an explanatory component. An
attempt is to be made to explain features of the logic of
conditionals in terms of the relations between these
components and in particular the relation between the
explanatory direction of a conditional (antecedent-to-
consequent or consequent-to-antecedent) and the inferential
direction.
In the philosophical literature questions about the
validity of contraposition are generally associated with
questions about the validity of (the invalid principles)
"strengthening the antecedent" (i.e. the logic of "if" in
English is not monotonic) and transitivity. And all these
questions are generally asked under the headings "Subjunctive
Conditionals","Counterfactuals" or "Contrary-to-fact
Conditionals". It may well be appropriate to cover these
topics to some extent.
Since what is envisaged is of the nature of an empirical
hypothesis concerning the logic of natural language
statements, and that hypothesis will take the form of a set
of principles of natural inference, it is expected that it
will be desirable to construct a (PROLOG) inference machine
employing these principles for test purposes. It has not been
decided how the machine should work or how it should be
employed.
I am not acquainted with the psychological literature or
artificial intelligence literature on these topics and would be
grateful for any references.
Ian Wilson.