One point of clarification about the argument I am trying to bring to bear here, about the definition of intelligence.
I am not saying that it is wrong to *informally* say that there are different degrees of intelligence, and that thermostats are, in a sense, part of a continuum that includes all intelligent systems.
In fact, I am all for that kind of informal assessment: I think thermostats and humans really cannot be separated cleanly into two distinct classes!
The only thing I am objecting to is that anyone would claim that they can build a formalized, objective definition of intelligence that captures our intuitive use of the term.
What happens, I believe, is that some people go beyond the informal claim and try to formalize it: they then produce ridiculous definitions, like Hutter's, that allow them to do things like claim that they are studying intelligent systems when they are doing nothing of the sort.
Hutter's claim is an easy target. But what is more important are the subtler cases, where the definitions are used to distort the AI research endeavor. The ramifications of making a bad formal definition can get very subtle, but rather than get into a huge argument about what those ramifications might be, I am just drawing a line in the sand and saying that if anyone thinks they can FORMALIZE a definition of intelligence they had better obey some strict criteria and not engage in handwaving or circularity.
If someone wants to agree that no formal definition is possible, and that the best we can do is produce informal definitions, that's GREAT. I think that informal lists of characteristics can sometimes be a help. But there is a big difference between that and claiming formal definitions.
Richard Loosemore ----- This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?member_id=231415&user_secret=fabd7936
