[Martin Taylor 951129 14:00]
Rick Marken (951129.1000) to Bruce Abbott
I have no problems with the main thrust of your message on the behavioural
illusion, but I have a bone to pick on one issue that tends to recur
on CSG-L. (If it didn't, I'd let it pass in this context).
1) You observe that some behavior, o, (such as pecking at key A or B) is a
function of some stimulus situation, s, (such as pictures). That is, you
observe:(1) o = f(s)
2) You conclude that the relationship between o and s (f()) reflects a
characteristic of the organism. For example, you conclude that differential
responding to the pictures reflects the organism's ability to discriminate
the pictures; different o's result from different values of s because the
bird can disciminate different values of s.
Actually, the only legitimate aspect of that comment comes from your
introduction of the unnecessary word "because". I think that most
people would have phrased the part after the last semicolon differently:
"different o's can result from different values of s only if the bird
can discriminate different values of s," to which some might add "and
the bird wants to use that discrimination to produce different o's."
Or even "and the different q's result in distinguishable error levels
within the bird's control systems."
If the bird dosn't respond differentially, you can't conclude that it
doesn't discriminate between the pictures, but if it does respond
differentially to them, it must be able to discriminate between them.
There are three possibilities. I think you may be arguing that there
might be four:
1. Bird discriminates, bird behaves differentially.
2. Bird discriminates, bird does not behave differentially.
*3. Bird does not discriminate, bird behaves differentially
4. Bird does not discriminate, bird does not behave differentially.
The third of these four cannot happen.
The experimenter can see the differences between 1or3 and 2or4. If the
experimenter observes 1or3, the case must be 1, because 3 is impossible.
But if the experimenter observes 2or4, nothing can be said about the
bird's ability to discriminate. If the bird behaves differentially
under any circumstances, the experimenter can put at least an upper bound
on the bird's ability to discriminate, but can't say how much better the
discrimination ability might actually be.
This comes up again and again when you treat of psychophysics. You assert
that psychophysical experiments tell nothing about the abilities of
people to discriminate sensory variation, whereas in fact they put
upper bounds on those abilities--people may be able to do better than
the experiment suggests, but their best ability is not worse than their
differential behaviour discriminates among the sensory conditions.
Martin