[lit-ideas] Grice and Garcia on reinforcement
- From: Jlsperanza@xxxxxxx
- To: lit-ideas@xxxxxxxxxxxxx
- Date: Sun, 28 Apr 2013 05:10:16 -0400 (EDT)
Ceteris Paribus, Grice refers in "Method in Philosophical Psychology" to
various experiments.
In "METHOD IN PHILOSOPHICAL PSYCHOLOGY", which Grice delivered as
President of the American Philosophical Association refers to the Yerkes
Dodson Law.
Grice was supposed to be delivering the Grand Analytic-A Priori Truth.
Instead, he chose to refer to this law
("there is, or _was_ this law") of "empirical pscychology", "based on
experiments".
"Experiments, indeed," Grice continues, "designed to test degrees of
learning compentence in rats set to run mazes under water, after varying
periods
of initial constraints."
Grice goes on to formulate the 'law' ("if it is one")
With other factors constant, degrees of learning competence are correlated
with
degrees of emotional stress by a function whose values form
a bell curve.
Grice is considering the 'form' of an empirical
psychological 'law' (so-called). Another example would be from his
colleague at Berkeley, Garcia.
As Grice notes, Yerkes and Dodson are proposing a theory, call it Th,
"from which we derive 'laws'".
The laws are formulated in psychological terms already which are
thus 'theoretical terms', NOT observational (contra, perhaps Popper -- or
not).
For each theoretical term, a correspondence "rule" is found, for the
Yerkes-Dodson law successfully predicts (if not analytically) that
over-learning
can improve performance in states of high arousal.
Since McEvoy was mentioning him (+> Garcia) and I was mentioning him (+>
Grice).
John Garcia (June 12, 1917 near Santa Rosa, California - October 12, 2012)
was an American psychologist, most known for his research on taste aversion
learning.
Curiously, Garcia studied at the University of California-Berkeley where H.
P. Grice (the Oxford philosopher, not the East Anglia one by the same
surname) where he (again Garcia, not Grice) received his A.B., M.A., and Ph.D.
degrees (Grice never received a PhD degree -- "who wants to be
overqualified?").
Garcia's first postdoctoral job was with the U.S. Naval Radiological
Defense Lab in San Francisco, California in 1955.
While Grice was interested in 'conversational intentions', Garcia began to
study the reaction of the brain to ionizing radiation in a series of
experiments on laboratory animals, mainly rats.
---
Garcia noticed that a rats avoided drinking water from a plastic bottle --
"only when in a radiation chamber."
Garcia suspected that the rat associated the “plastic tasting” water
with the sickness that radiation triggers.
During the experiments, a rat was given one taste, sight, sound as a
neutral stimulus.
Later the (same) rat would be exposed to radiation or drugs (the
unconditioned stimulus), which would make him (i.e. the male rat) sick.
Through this experiment, which McEvoy judges "seminal and Popperian",
Garcia discovers that if a rat became nauseated after presented with a new
taste, even if the illness occurred several hours later, the rat would (or as
Grice prefers, "should") avoid that taste.
This, mutatis mutandis, contradicted the belief that, for conditioning to
occur, the unconditioned response (in this case, sickness) must immediately
follow the conditioned stimulus-to-be (the taste) -- as per standard
reinforcement.
Not satisfied with that, Garcia went on to discover that the (same) rat
developed some sort of aversion to tastes, but not to sights or sounds.
This Garcia took, in a Popperian way, to disprove the previously held
theory that any perceivable stimulus (light, sound, taste, etc.) could become a
conditioned stimulus for any unconditioned stimulus.
Meanwhile, Grice was lecturing on implicature reinforcement (or "collapse"
-- "surely to explicate that p by implicating it is analytically
superfluous").
Garcia's discovery, conditioned taste aversion,[1] is considered a survival
mechanism because it allows an organism to recognize foods that have
previously been determined to be poisonous, hopefully allowing said organism
to
avoid sickness.
As a result of Garcia's work, conditioned taste aversion has been called
the "Garcia Effect." (vide "NAUSEA").
As a result of the Oxonian philosopher's work, a taste for implicature is
called the "Grice Effect."
Cheers,
Speranza
REFERENCES
Garcia J, Kimeldorf DJ, Koelling RA. Conditioned aversion to saccharin
resulting from exposure to gamma radiation. Science 1955; 122(3160): 157-8.
Grice, H. P. Logic and Conversation. Deposited at UC/Berkeley, Bancroft
Library.
McEvoy:
"...the evidence [Dr] Garcia gathered from [a rat]
... shows they will never 'associate' stomach sickness with
a flashing light nor associate an electric shock with
anything [the rat has] eaten, no matter the closeness of
the 'association' in terms of frequency and intensity
as set out in associationist pyschology."
"This tells us [a rat is] NOT 'associating' as per
associationism but are following some 'theories'
built into [the rat's] action programme that tell [the rat], for
example, that stomach sickness will be due to
something [the rat] ate.
"In a striking example, Garcia showed that [a rat] nauseated when
unconscious, using radiation, will wake up and no longer consume the sucrose
[the
rat] previously lapped up - [the rat's] inbuilt rogramme has told [the rat]
to be wary of what [the rat was] last eating given the nausea that the
programme has detected and that the programme works with a 'theory' that such
nausea will be due to something [the rat] ate ([the rat's] programme does
not have a 'theory' that anticipates the possibility of radiation)."
"In [neo-C]antian terms, this programme is not analytic nor synthetic a
posterori but is synthetic apriori."
"However, Darwin/Popper explain why this synthetic a priori knowledge, in
the form of an action programme, is conjectural rather than necessarily
true, and that the programme will have evolved under the pressure of 'natural
selection'."
"In terms of correct understanding of how we 'learn from experience',
Garcia's [rat] eat Pavlov's dogs for breakfast."
"That [Garcia's Rat is] is less well known would be comparable to Darwin
being less well known than Lamarck."
(I have reformulated, being an empiricist, each reference by McEvoy of
plural ("rats") to the "(at least one) rat" singular).
Cfr. Grice on the squirrel Toby in "Method in Philosophical Psychology" and
psychological association theory as analytically interesting (if "ceteris
paribus").
------
Extra.
"The Yerkes-Dodson law is an
empirical relationship between
arousal and performance,
originally developed by psychologists,
Robert M. Yerkes
&
John Dillingham Dodson
in 1908. The law dictates that
performance increases with
physiological or mental arousal,
but only up to a point."
"When levels of arousal become too high, performance decreases. The
process is often illustrated graphically as a curvilinear, inverted
U-shaped curve which increases and then decreases with higher levels of
arousal. Research has found that different tasks require different
levels of arousal for optimal performance. E.g. difficult or
intellectually demanding tasks may require a lower level of arousal (to
facilitate concentration), whereas tasks demanding stamina or
persistence may be performed better with higher levels of arousal (to
increase motivation). Because of task differences, the shape of the
curve can be highly variable. For simple or well learned tasks, the
relationship can be considered linear with improvements in performance
as arousal increases. For complex, unfamiliar, or difficult tasks, the
relationship between arousal and performance becomes inverse, with
declines in performance as arousal increases. The effect of task
difficulty led to the hypothesis that the Yerkes-Dodson Law can be
decomposed into two distinct factors ? compare bathtub curve. The
upward part of the inverted U can be thought of as the energizing
effect of arousal. The downward part is caused by negative effects of
arousal (or stress) on cognitive processes like attention (e.g. "tunnel
vision"), memory, and problem-solving. There has been research
indicating that the correlation suggested by Yerkes & Dodson exists
(such as that of Broadhurst 1959, Duffy 1962, Anderson, 1988), but a
cause of the correlation has not yet successfully been established
(Anderson-Revelle-Lynch 1989). A 2007 review of the effects of stress
glucocorticoids and human cognition revealed that memory performance
vs. circulating levels of glucocorticoids does manifest an upside down
U shaped curve. The authors noted the resemblance to the Yerkes-Dodson
curve. E.g. long-term potentiation (the process of forming long term
memories) is optimal when glucocorticoid levels are mildly elevated
whereas significant decreases of LTP are observed after adrenalectomy
(low GC state) or after exogenous glucocorticoid administration (high
GC state). It was also revealed that "in order for a situation to
induce a stress response it has to be interpreted as novel and/or
unpredictable, and/or the individual must have the feeling that they do
not have control over the situation. Presence of a social evaluative
threat constitutes the fourth." It has also been shown that elevated
levels of glucocorticoids enhanced memory for emotionally arousing
events but lead more often than not to poor memory for material
unrelated to the source of stress/emotional arousal. If only a few
points are presented then the theory can account for most results,
however, both performance and arousal are presented by single terms but
they are multidimensional. Also, the theory is descriptive and not an
explanatory one. The Yerkes-Dodson law predicts that over-learning can
improve performance in states of high arousal.
REFS.
Yerkes/Dodson 1908. 'The relation of strength of stimulus to rapidity
of habit-formation.
Journal of Comparative Neurology & Psychology 18:459-482.
Anderson, K. Impulsivity, and memory scanning: an explanation for the
Yerkes-Dodson effect. Motivation & Emotion, 13.
------------------------------------------------------------------
To change your Lit-Ideas settings (subscribe/unsub, vacation on/off,
digest on/off), visit www.andreas.com/faq-lit-ideas.html
Other related posts:
