[From Bruce Gregory (2004.1229.1912)]
Have any PCT models been developed to explain the phenomenon of
"Priming"?
The enemy of truth is not error. The enemy of truth is certainty.
[From Fred Nickols (2004.12.29.1954 EST)] --
"Whoa there, white man!" said Tonto.
"Priming" has a specific meaning in the behaviorism school (at least the one
to which I was exposed). To "prompt" is to hint or suggest or otherwise
"elicit" or "draw out" a response. In contrast, td "prime" is to physically
walk someone through the appropriate response (e.g., position someone's
limbs in a preparatory position).
(As an aside, I can see how "priming" might contribute to establishing new
reference signals.)
(As an another aside, I was taught that "priming" applies to psychomotor but
not to cognitive skills.)
So, Bruce G, how are you using the term "priming"?
Regards,
Fred Nickols
nickols@att.net
···
-----Original Message-----
From: Control Systems Group Network (CSGnet)
[mailto:CSGNET@listserv.uiuc.edu] On Behalf Of Bruce Gregory
Sent: Wednesday, December 29, 2004 7:13 PM
To: CSGNET@listserv.uiuc.edu
Subject: Priming[From Bruce Gregory (2004.1229.1912)]
Have any PCT models been developed to explain the phenomenon of
"Priming"?The enemy of truth is not error. The enemy of truth is certainty.
[From Bruce Gregory (2004.1229.2015)]
Fred Nickols (2004.12.29.1954 EST)
So, Bruce G, how are you using the term "priming"?
Priming
Priming is an increase in the speed or accuracy of a decision that occurs as a consequence of a prior exposure to some of the information in the decision context, without any intention or task related motivation. One of the original demonstrations (Meyer and Schvaneveldt, 1971) of priming occurred in a lexical decision task in which a series of decisions is made about whether letter strings are words or not. Priming was shown to occur in cases where two successive letter strings were semantically related words. For example the decision that 'doctor' is a word was faster when the preceding letter string was 'nurse' as compared to 'north' or the non-word 'nuber'.
The enemy of truth is not error. The enemy of truth is certainty.
[From Bjorn Simonsen (2004.12.30,10:30 EST)]
From Bruce Gregory (2004.1229.2015)
Priming is an increase in the speed or accuracy of a decision that
occurs as a consequence of a prior exposure to some of the
nformation in the decision context, without any intention or task
related motivation. One of the original demonstrations
(Meyer and Schvaneveldt, 1971) of priming occurred in
a lexical decision task in which a series of decisions is made about
whether letter strings are words or not. Priming was shown to occur in
cases where two successive letter strings were semantically related
words. For example the decision that 'doctor' is a word was faster
when the preceding letter string was 'nurse' as compared to 'north'
or the non-word 'nuber'.
I think "priming" is a concept with many associations. Fred Nickols
mentioned one refering to behaviorism. I also see behaviorism through "Meyer
and Schvanveldt, 1971" (I have not studied their work), therefore I may be
wrong.
A more modern use of the concept "priming", you find in Wikipeda; Priming in
psychology is activating particular connections or associations in memory.
These associations are often unconscious.
I said a more modern use, because brain researcers are engaged in
connections between (what they call) "association centers" in the occipatal
lobe and cortex. Susan A. Greenfild says that it seems as if the the
"association centers" links sensing systems and motorfunction. They are also
engaged in connections between (what they call) "association areas" in the
frontal lobe and motorfynctions in cortex.
An interesting flight of ideas (my words) is that very small children have
an explicit memory. They can copy a play they have seen other children play,
but they can't play it after someone has explaines the game with words. The
explicit memory has something with hippocampus and hypothalamus to do and
the implicit memory (concepts) has something with later developed
associative areas in the frontal lobe to do ("priming)".
Back to HPCT.
* Have any PCT models been developed to explain the phenomenon of "Priming"?
The way I understand HPCT is that a higher level, e.g. the sixth level (the
relationship level) is later developed than the third order levels we find
in thalamus. I think HPCT is a good direction indicator (model) that
explaines "Priming".
But I may be wrong.
Bjorn
[From Rick Marken (2004.12.30.0950)]
Bruce Gregory (2004.1229.2005)
Have PCT models been constructed to explain the data
on priming? That would be a good place to start.
There are no PCT models of priming for the same reason that there are no PCT
models of any conventional psychological research. The reason is that this
research is generally not done in a way that makes it possible to know how
to model it. We have tried to model various findings of operant conditioning
research, where the relevance to PCT seems obvious (in operant conditioning,
organisms act to produce a goal result: reinforcement) and where we often
have data from individual subjects. But the data that is reported is simply
inadequate for use as a basis for PCT modeling.
Priming (of the type you describe) is hard to model mainly because the
results are based on group data. But even if we had individual data, there
is still the problem of determining what each subject is controlling for.
This makes a difference in how one would model the effect of the "priming"
word on response time and accuracy. But perhaps we could get some hints from
looking at the individual data. One thing I would want to know is how they
measured accuracy in these priming tasks. I presume that they used a measure
like d' from signal detection theory, which can (if used properly) remove
the effect of "response bias" from measures of accuracy. Measures of
reaction time would also have to take this into account.
If, in an individual case, it is true that the time to say that 'doctor' is
a word is shorter when the preceding letter string was 'nurse' as compared
to 'north' or the non-word 'nuber', then I would think that such a result
could be explained in terms of the PCT hierarchy. For example, one
possibility is that consciousness remains focused on the system where the
last identification was mode. So when the preceding letter sting is "doctor"
consciousness remains at he category level system that recognizes "medical
professional" words. This model suggests that identification of non-words
should be faster when preceded by non-words, and slower when preceded by
words. Is this what is found?
···
--
Richard S. Marken
MindReadings.com
Home: 310 474 0313
Cell: 310 729 1400
--------------------
This email message is for the sole use of the intended recipient(s) and
may contain privileged information. Any unauthorized review, use,
disclosure or distribution is prohibited. If you are not the intended
recipient, please contact the sender by reply email and destroy all copies
of the original message.
[From Bruce Gregory (2004.1230.1616)]
Rick Marken (2004.12.30.0950)
Bruce Gregory (2004.1229.2005)
Have PCT models been constructed to explain the data
on priming? That would be a good place to start.There are no PCT models of priming for the same reason that there are
no PCT
models of any conventional psychological research. The reason is that
this
research is generally not done in a way that makes it possible to know
how
to model it.
Pity.
The enemy of truth is not error. The enemy of truth is certainty.
Re: Priming
[Martin Taylor 2004.12.30.13.27]
[From Rick Marken (2004.12.30.0950)]
Bruce Gregory (2004.1229.2005)
Have PCT models been constructed to explain the data
on priming? That would be a good place to start.
There are no PCT models of priming for the same reason that there are
no PCTmodels of any conventional psychological research. The reason is that
thisresearch is generally not done in a way that makes it possible to know
howto model it.
As control systems. How about modelling it as a property of
perceptual systems?
Priming (of the type you describe) is
hard to model mainly because theresults are based on group data.
Disputable. Here’s a quote from the book my wife and I wrote in
1983 (The Psychology of Reading, Academic Press, p257).
“The semantic priming effect
can come under top-down control. In Tweedy and Lapinski’s (1981) study
the effect was strong when subjects expected a semantic relationship
between the prime and the target. [There’s that word “expect”
again. MMT now]. Eisenberg and Becker (1982) asked people to read or
to skim an easy and a difficult passage. Some maintained their speed
in both passages, whereas others slowed drastically when skimming the
difficult passage. Thes two groups of subjects were then given a
lexical decision task [Is this a word or not] with semantic priming.
The speed maintainers showed mainly facilitation with semantic priming
and littel interference from an unrelated prime, whereas those who
slowed showed the reverse, namely, some facilitation and much
interference.”
Sounds as if they considered individuals separately, based on the
way they performed in a different task.
But even if we had individual data,
thereis still the problem of determining what each subject is controlling
for.
It is a basic assumption of all psychological experiments that
the subjects are controlling for pleasing the experimenter, the
mechanism for which is to perform the task in the manner the
experimenter requests. However, sometimes the experimenter measures
something that to the subject may be irrelevant to the task (such as a
physiological effect, or the likelihood of using words of specific
classes). That’s not the case in the priming studies, where the
experimenter measures the output that the subject believes is being
measured.
From the experimenter’s point of view, the observable behaviour
in a priming study is the speed or accuracy of an overt response. For
the PCT analytic point of view, that output supports the control of
something. If that “something” is to perceive oneself as
doing what the experimenter wants one to do, I don’t see how one would
apply “The Test” for the controlled perception. It seems
obvious that if one prohibits the subject from perceiving what the
experimenter wants, the subject won’t be taking part in the
experiment.
So I think we have to rely on the assumption that the subject is
controlling for perceiving him/herself as doing what the experimenter
wants, unless the subject’s overt actions lead the experimenter to
believe otherwise (as sometimes does happen :-).
One thing I would want to know is
how they
measured accuracy in these priming tasks.
I presume that they used a measurelike d’ from signal detection theory, which can (if used properly)
removethe effect of “response bias” from measures of accuracy.
Measures ofreaction time would also have to take this into account.
True. A lot of experimenters don’t do this properly.
Nevertheless, if a theory provides a qualitative prediction, as PCT
often can, and the experimental data give a qualitatively different
result, even badly done data can be useful if one is clear on the
kinds of effect the mistakes would cause.
If, in an individual case, it is true
that the time to say that ‘doctor’ isa word is shorter when the preceding letter string was ‘nurse’ as
comparedto ‘north’ or the non-word ‘nuber’, then I would think that such a
result
could be explained in terms of the PCT
hierarchy. For example, onepossibility is that consciousness remains focused on the system where
thelast identification was mode. So when the preceding letter sting is
“doctor”consciousness remains at he category level system that recognizes
"medicalprofessional" words.
Priming is one situation in which a clear difference is found between
conscious and unconscious perception. Here’s some more quotes from the
book (p258-9).
"Flash a prime briefly,
following it with a pattern of random lines and curves. If the
conditions are carefully chosen, the prime cannot be reported, but it
nevertheless facilitates recognition of a related target [several
references here]. In a different procedure, an unattended word flashed
together with a word to be recognized canaffect how well subjects
categorize, even when they are unaware of the existence of the
unattended word (Underwood, 1981). The effect even works for pictures
as well [harping back to a discussion of a month or so ago. MMT] A
picture used as a semantic prime can have an effect even when it is
flashed at too short a duration to be named. At first glance, these
results smack of the discredited notion of “subliminal
perception.”
" This ‘magical’ priming
effect actually resolves some conflicting results on ambiguous words
used as primes. When the prime is an ambiguous word such as PALM, both
senses of the word facilitate the recognition of related targets, such
as HAND and TREE. That this effect happens shows that neither sense is
inhibited, even though subjects may be aware of only one sense of the
ambiguous prime (Holley-Wilcox & Blank, 1980). If, however, one
sense of the ambiguous prime has been sensitized by prior presentation
of a related word (e.g. MONEY-BANK, versus RIVER-BANK), then only that
sense serves to prime the target, and targets related to the other
sense are inhibited. For example, MONEY-BANK facilitates CASH, whereas
RIVER-BANK inhibits it (Schvaneveldt, Meyer & Becker,
1976).
“Marcel (1980a) used the
double priming techniquw of Schvaneveldt et al, but masked the
ambiguous word in some of his conditions. His results agreed with
those of Schvaneveldt et al if the ambiguous prime was not masked, but
with those of Holley-Wilcox and Blank if it was masked. With masking,
both senses of the ambiguous prime facilitated recognition of the
target, whereas without masking, the sense incongruent with the
sensitizer was inhibited. Masking the prime can eliminate the
inhibitory effect of an unrelated prime. An unrelated prime strongly
inhibits recognition of the target if the prime is easily seen, but
not if the prime is masked. A related prime facilitates recognition of
the target, whether the prime is masked or not.”
I think studies of the perceptual component of control systems
as being potentially valuable inthe study of control, just as one
ought to consider studies of any other component of control
systems.
To my mind, what these studies say is that there are (at least)
two different modes of perception, and that (at least for word
perception) it is only in conscious perception that mutually
contradictory possibilities are disallowed – only in conscious
perception is an ambiguous pattern restricted to just one of the
possible variants. If this is generally true,one might speculate that
all the many ways the Necker Cube can be seen are seen,
unconsciously, and it is only in the process of bringing them to
consciousness that one dominates, inhibiting the others.
This model suggests that identification of non-words
should be faster when preceded by non-words, and slower when preceded
by
words. Is this what is
found?
Another thing the studies suggest is that the priming effect IS
controlled, and that priming is stronger if it is expected by the
subject. So I would guess that your prediction would pan out for
studies in which subjects expected that a non-word prime would be
followed by a non-word target.
I’m suere lots of studies have been done in this area in the 22
years since the book was written. I haven’t followed them. But I have
been told that Marcel’s unexpected result was at first dismissed as
incredible and probably due to bad experimental technique, but has
since been shown to be reliable by people who had been sceptics. And
it is Marcel’s result that is the crux of this posting.
Martin
[From Bill Powers (2004.12.31.0705 MST)]
Martin Taylor 2004.12.30.13.27–
Some maintained their speed in both passages,
whereas others slowed drastically when skimming the difficult passage.
Thes two groups of subjects were then given a lexical decision task [Is
this a word or not] with semantic priming. The speed maintainers showed
mainly facilitation with semantic priming and littel interference from an
unrelated prime, whereas those who slowed showed the reverse, namely,
some facilitation and much interference."
Sounds as if they
considered individuals separately, based on the way they performed in a
different task.
It still sounds like two groups (treatments) to me: those who showed one
pair of effects, and those who showed the other pair. It’s true that
dividing them according to whether their speed remained the same or
decreased introduces individual measurements, but (tell me if I’m wrong),
I don’t expect that everyone who maintained speed was facilitiated by
related priming and unaffected by unrelated priming, and vice versa for
the other group. How did the authors explain those who showed the
opposite effects – and do you happen to know how many there
were?
Best,
Bill P.