From: Richard Loosemore (rpwl@lightlink.com)
Date: Fri Apr 28 2006 - 10:36:24 MDT
Ben Goertzel wrote:
> Richard,
>
> If you think that
>
>> what is probably happening in STM is that
>> there are several thousand things in "short term memory" at any given
>> moment,
>
> then either
>
> a) you are flat wrong, and making an assertion refuted by experimetns
> already,
>
> or
>
> b) you are not defining STM in the same way as is standard in
> cognitive psychology currently
>
> I assume that b) is probably the case...
>
> Now, this may be fine: perhaps, for example, your definition of STM may
> correspond better to brain structure than the standard definition. The
> area of STM/working-memory/whatever-you-want-to-call-it is the
> subject of plenty of debate in contemporary cog psych.
>
> But I would be curious for you to clarify exactly what your definition
> is...
>
> Please note, it is not obvious that a mind needs to have a rigid
> STM/LTM distinction. Novamente does not exactly have such a
> distinction, though it has a couple different distinctions that are
> somewhat related to this one.
It is b), of course, but I don't think the literature is as clear cut as
you suggest: the presence of a large number of "partially-activated"
concepts is what I have in mind, and although people do not normally
refer to this as STM, the distinctions that *they* make between STM and
whatever place these partial activations are situated is not so clear.
They would like it to be LTM, but partialy activated, and I have
(repeatedly) chalenged them to be clear about the whys and wherefores of
that way of looking at it, without much success.
I am stretching the concept of STM, but I am doing so in a way that most
cognitive psychologists would not say is wrong, just open to
interpretation. They do not often want to get into a discussion about
the distinction I am making here, because it is difficult for them to
think of experiments to clarify the distinction, but if I take a <let's
build a system that actually works"> perspective, I find that this issue
comes up immediately, whether or not they can address it experimentally.
Richard
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:56 MDT