I am current reading Miller, Barnet & Grahame's (1995) review of the Rescorla-Wagner model. In it they state that "The Rescorla-Wagner model assumes that lamda during extinction is zero and that beta for extinction (beta2) is a number smaller than beta for acquisition (beta1) but larger than zero."
The difference in betas suggests that extinction happens slower than acquisition. This difference also does not seem to be necessary for extinction to happen in principle, nor can I find anything in their discussion that either explicitly or implicitly suggests that extinction is slower than acquisition. I am wondering if this is a typo, or if there are good (empirical) reasons to think that extinction is slower than acquisition. I'm not too familiar with the literature and am having difficulties finding any research that bears on the issue directly. Hoping someone might be able to 1) give me a definitive answer and/or 2) point me to relevant readings.
Among other factors, extinction depends on whether or not the cue received partial reinforcement during acquisition. Partial reinforcement increases resistance to extinction.
Haselgrove, M., Aydin, A., & Pearce, J. (2004). A partial reinforcement
effect despite equal rates of reinforcement during Pavlovian conditioning.
Journal of Experimental Psychology: Animal Behavior
Processes, 30, 240–250.
Pearce, J. M., Redhead, E. S., & Aydin, A. (1997). Partial reinforcement
in appetitive Pavlovian conditioning with rats. Quarterly Journal of
Experimental Psychology, 50B, 273–294.
Here are a couple articles directly relevant to this question with regard to assessing REPEATED reversals of acquisition and extinction:
Article Repeated extinction and reversal learning of an approach res...
Article Comparison of the Rates of Associative Change During Acquisi...
There is other research suggestive of the same type of an effect, although these studies were not designed to ask this question:
Article No impact of repeated extinction exposures on operant respon...
Article Resistance to Extinction in the Steady State and in Transition
Thus, it appears acquisition proceeds more rapidly than extinction with repeated reversals to the extent acquisition and extinction performance can be directly compared.
I hope that helps.
The standard answer is that extinction takes longer than acquisition. Extinction gets faster, however, the more it is repeated. Here is an article:
We have phobias because they are easy to acquire (it may take only once) but some phobias last for years; extinction does not always work. The famous example is prof. Watson's "little Albert" experiment; the child was conditioned to a rat phobia, which was thought to fade off by feeding him his favorite dessert while the feared rat was brought closer. There are many versions of how this experiment ended but Little Albert was most likely removed from the experiment before the extinction took place.
As a sufferer of obsessive-compulsive disorder (particularly pure-O), I can personally attest that extinction is longer than acquisition. William is quite right that extinction times get faster the more it is done.
Seems to me that it would depend on the stimulus and on the response (e.g. importance to adaptation) -- I am quite sure I have read quite a lot about such variability in animal behaviors, depending on "the behavior" (and, of course, also on "the stimulus"). Do we really have any good reasons (anymore) to believe there are general "laws of learning"? (I think, if that were the case, Skinner would be very well known and seen as very important to this day.)
Why is it we even still speak in terms of "behaviors", when biologically, and in reality, it is always BEHAVIOR PATTERNS? The nature of the various patterns and environments (once well understood) would provide one with good guesses about the quickness of acquisition of learned behavior AND how quickly extinction would likely occur (and then both could be investigated and the answers found). The PATTERNING itself is informative (of THAT itself, and of the stimuli: aspects of the environment); how and why is it that the fact** that behaviors ARE in patterns is STILL not even recognized today?
** FOOTNOTE: a fact, basically, because only thinking like that is in-line with necessarily applicable biological principles
- Badges
- Science topic
- Similar topics
- Methodology
- Research Methodology
- Research Papers