I believe you are referring to "desirable difficulties" - see Robert Bjork's work. A simplified version is that effortful retrieval benefits long term retention. Rather than simply spacing study sessions, the act of taking a test is good for retention.

I think, no. I think also - it is very hard to show experimentally that one subject tottally forgot the task (e.g., riding a bycicle or swimming). 

So hard, in fact, that some people think "forgetting" may be a poor description of a phenomenon. Please see attached by Capaldi and Neath (1995).

Interesting, thank you for the resource.

Is anyone familiar with something called the "cognitive reconstruction effect" or something like that? It's related to the spacing effect. I remember reading about it but don't remember where. The idea, if I'm not mistaken, was that one possible explanation for the spacing effect is that the more difficult it is to retrieve/reconstruct a memory, the stronger it becomes and the more easily it is retrieved subsequently. Anyone?

Sounds like one of many re-phrasings of what cognitive systems may have to do with challenging fact that the stimulus seems impoverished. It sounds from your description like a remarkably unhelpful explanation for the spacing effect.

For instance, if it's correct, then the fact of your not now remembering where you read about it should mean that it will be much more easily remembered soon. So, perhaps you only need to wait! :)

Forgetting is harmful to incremental learning. For break through learnings to happen, one should be able to forget things learned in the past. In todays world we experience several paradigm shifts that require fast forgetting for survival.

I agree with Damian that "forgetting" is an ambiguous term. I think that it does not specify which processes underlie that "forgetting": decay, interference, active suppression, loss of information about specific skill components, weakening of associations or structural relations between skill components?

I also agree with Xavier that in general forgetting is harmful to learnng. However, the exception could be that, if you want to learn X, then it is useful to "forget" (i.e., inhibit, suppress) any Y that could interfere with X.

I don't have an answer but find this an interesting question and am curious for further cues.

I recently read two new'ish books that featured heavily on "spaced repetition". One, by Gabriel Wyner is called "Fluent Forever: How to learn any language and never forget it" and the second is by Peter Brown "Make it Stick: The science of successful learning" Additionally, much of the language leaning software these days uses spaced repetition to aid in vocabulary learning/acquisition. I've been extolling the virtues of spaced repetition to my students for years and years.

However, to answer your question, from what I recall about the research I've read, no, "forgetting" does not help you to remember or re-learn "better". You have to regenerate or reactivate whatever it is you want to actually retain in the mind prior to the phase of "forgetting". Apparently doing that through incrementally (time) delayed spaced repetition seems to be the to be getting a lot of attention these days, even though it's been around a long, long time.

You can read about it in general at Wiki: https://en.wikipedia.org/wiki/Spaced_repetition and if you want more "academic" style papers there are a ton out there.

I believe you are referring to "desirable difficulties" - see Robert Bjork's work. A simplified version is that effortful retrieval benefits long term retention. Rather than simply spacing study sessions, the act of taking a test is good for retention.

Sergio's contribution here is critically important in parsing out what your question is about, I think.  Learning involves more than simple memory mechanisms, and the volition behind recall has a great deal to do with how information is developed and integrated into a cognitive schema. 

I want to add on to Julia's comment about schema and Ally's comment about language. The memory research shows that spacing and desirable difficulties both help in long-term retention.  What is missing here is retrieval.  If information is in our long-term memory but it is difficult to remember it, than it is not so useful.  For example, most of us remember ( it is buried deeply for me) the sum of 6 and 8. Because I never learned that fact to automaticity, it is time-consuming for me to remember; though I do always remember with effort.  The key is knowing how automatic the learning needs to be and then using mechanisms that research has shown develop that level of retrieval.  For language learners, slow retrieval does not suffice.  Language must be automatic.  So while spacing is a good tool, it does not lead to long-term automatic retrieval of complex information such as language use.  In order to retrieve complex knowledge automatically, other learning which enhances the spread of activation; such as schema development, can lead to automatic retrieval.  

Automatic retrieval is not necessary for all learning; however.  Typically, language, math, and music are the three areas which must be automatic to be proficiently useful.  For much other learning, lower levels of retrieval proficiency are sufficient.  I think the more educators know about memory and retrieval, the better we can design instruction to be as efficient as possible.

Let us just get straight to the question:

• A knew skill X but loses memory of it due to lack of practice
• B knows skill X and remembers it
• Both A and B are taught skill X
• A and B are tested on their performance with respect to skill X
• Who will fare better at skill X, either A or B?

The answer lies in one of two possible preconditions:

In case 1, B's performance would be certainly better than that of A in the post test.

In case 2, A's performance may be eqquivalent to or even better than that of  B depending on the amount of knowledge he had before he forgot them.

Details:

In general, case 1 prevails. Research suggests that our brain fits with what seems most logical to us when we have missing neural connections related to forgotten facts. If the person has enough logical knowledge that supports the truth of the forgotten skill, then he/she would not make a mistake in the recollection process. But with reference to Dr. Sergio Morra's answer, it is highly possible for some people to have or even develop fallacies that interfere with the skill to remember. If the person merely forgot a few pointers in the skill (case 2), he would be able to fill the gap with the new learning session of the same skill. By subset of the skill, I mean part of the skill is known, the superset of the skill would mean he knows more of the skill than what is to be taught is the learning session.

You may not be able to 'forget' anything on purpose, but you can unlearn them. Deny all the facts you already know about the skill and refreshen your thoughts during the session avoiding recollecting old stimuli. This is because, true learning also involves the process of unlearning old facts, replacing them with the new. This goes with nearly all fields from researching in Neuclear Physics to updating your coding language to an entirely new one. 

I agree with Ebenezer's statement that we can unlearn in most cases.  However, typically when automaticity is reached, unlearning is near impossible.  In language acquisition, we call this fossilzation.  The person may be able to learn that a certain speech production is incorrect and be able to self-correct when monitoring his/her speech.  However, changing the unconscious production of that error seldom is possible.  

With fact learning, automatic retrieval is not typical and so replacing one fact with another seems quite possible.  Another aspect of this, too, I  think, is that the way the brain stores and processes automaticized skills seems quite different from how facts within a schema are stored and processed.  Without any neuro-imaging evidence of this, I would refer to the expert retrieval processes in memory research which clearly shows that experts, those who have automaticized knowledge and skills sets, are typically poor at explaining the component parts of their expert retrieval.  For these people, since they no longer "see" the parts of their skill/knowledge, I would think that it would be quite difficulty for them to unlearn and then relearn new data as part of their expert skill set.  

So yes, I do agree that we are in a constant state of updating our information with more accurate information.  However, I think we still have to differentiate the retrieval level of the learning goals when discussing learning or forgetting because expert knowledge and automated procedures in long-term memory typically prove resistant to change-forgetting may be impossible.  

This is a really worthy question and super relevant to my research plans!

In the circus there's a phenomenon in which a practiced skill sometimes improves while you're not practicing it. However, I'm almost entirely sure that these improvements would be even better when bolstered by daily practice.

On the other hand, we often say "you must learn a technique so well that you forget it -- and then learn it again." But that does raise the questions of why Person A would retain a technique and Person B would get rusty at it. And it'd be really hard to gauge the reasons why A vs. B would do a better job at executing this technique after a review...

Richard, have you looked at the research that shows how much better you remember when you practice and then sleep.  It works so much better than just practicing because your brain has time to process the information into long-term memory without dealing with more input.  I know it works better through personal experience.  My undergrad was in German and before I began doing this, it took me much longer to learn my vocabulary, etc.  Also, if you are looking at muscle memory phenomena, practice related to the skill enhances the skill.  This is why you hear of football players taking ballet lessons.  

True. To add on to Dr. Dayna's answer: there are, in genral, four stages that a person goes through when learning any skill.

When a point of criticality is reached with respect to the experience of a particular skill of a person, it becomes near impossible to to unlearn that skill (notice the emphasis on 'near').  This is the transition from stage 3 to stage 4. That's why a person skilled in Dvorak touch typing will be dumbstruck on a Qwerty keyboard and take longer to learn than someone with no typing practice.

"Forgetting strengthens learning" is more of a metaphor that illustrates people get stuck by the existing mental model so as to be less adapted to new knowledge.  That is, what is forgotten and what is learned are different, or even conflicting, knowledge.  To my humble understanding to how working memory functions, it takes more time to regain a forgotten skill.  The following three literatures might be of help to better understand how people learn.

Norman, D. A., & Shallice, T. (1986). Attention to action: willed and automatic control of behaviour. In R. J. Davidson, G. E. Schwartz, & D. Shapiro (Eds.), Consciousness and Self-Regulation. Advances in Research and Theory (pp. 1-18). New York: Plenum.

Schneider, W., & Shiffrin, R. M. (1977). Controlled and automatic human information processing: I. Detection, search, and attention. Psychological Review, 84(1), 1-66. doi: 10.1037/0033-295X.84.1.1

Shiffrin, R. M., & Schneider, W. (1977). Controlled and automatic human information processing: II. Perceptual learning, automatic attending and a general theory. Psychological Review, 84(2), 127-190. doi: 10.1037/0033-295X.84.2.127

Thanks everyone for weighing in. All very interesting. I'm starting to wonder if there's a sweet spot moment for deliberate non-practice. During the period between one "level" of understanding/performance and the next, right before discarding the existing mental model, using a new strategy, and making a qualitative jump to the higher level. Perhaps not practicing for a while during that period and getting rusty or forgetting could facilitate the transition to that higher level of performance once the subject returns to practice?

I would like to suggest that 'un-learning' (as opposed to forgetting) may often be involved in human endeavours to embrace novelty, e.g. a new way of working. See Nissen, et al (2007) for discussion of a double-helix relationship between use and redesign, including Nissen's discussion of Bateson's (1980) discussion of 'mind'.

Thank you, Christine. Could you please provide the title of the paper to which you're referring?

There are two papers:

Use and Redesign in IS: Double Helix Relationships? Double Helix Relationships in Use and Design of Informing Systems: Lessons to Learn from Phenomenology and Hermeneutics

Using Double Helix Relationships to Understand and Change Informing Systems.

You will find them in ResearchGate if you search on Hans-Erik Nissen (Lund University).

Learning always requires unlearning. With this in mind then we can say that not necessarily forget, but the electrochemical process causes new neuronal connections and strengthens rather than others when required to do so. Apparently we are forgotten, but in fact our brain seeks those more meaningful connections in a given situation. Besides neuronal plasticity it allows storing information and constant adaptation to new experiences and new information.

-------------------------------------------------------------------------------

El aprender siempre requiere desaprender. Partiendo de esta idea entonces podemos decir que no necesariamente olvidamos, sino que el proceso electroquímico neuronal provoca nuevas conexiones y las fortalece en vez de otras cuando requiere hacerlo. Aparentemente olvidamos, pero en realidad nuestro cerebro busca aquellas conexiones más significativas en determinada situación. Además la plasticidad neuronal permite un guardar información y una adaptación permanente a nuevas experiencias y a nueva información.

I see that my original question had to do with the Bjork's New Theory of Disuse. I see that there may have been some confusion over the word "forgetting" because that is apparently not what happens. Rather, retrieval strength reduces due to disuse. And what I was referring to, that 'forgetting' (reduced retrieval strength) leads to increased storage strength of the memories after subsequent retrieval. The harder it is to retrieve a memory (the more you 'forgot' it), the more of a gain in storage strength you get by recalling/reconstructing it.

So I guess the practical question is, other than the testing and spaced repetition, what other techniques can be used to increase both retrieval and storage strength?

Hi David,

You wrote:

 'forgetting' (reduced retrieval strength) leads to increased storage strength of the memories after subsequent retrieval. The harder it is to retrieve a memory (the more you 'forgot' it), the more of a gain in storage strength you get by recalling/reconstructing it.

Is there any empirical evidence for that? And what neural processes would account for the recovered "memory" becoming reinforced purely as the consequence of it having been recovered? Or is it, rather, that subjectively experiencing its recovery and our subsequent ease at recalling it gives us the false impression that it has become reinforced?

I'd offer an alternative explanation to account for that impression - namely, that it's the specific neural processes involved in recovering the "memory" that are reinforced, not the memory itself. To my understanding, recalling (or more accurately, re-constructing) a memory requires activation of associative, contextual markers encoded along with the specific content of the memory in the event of its formation, these markers in combination providing a unique reference value for accessing the neurally distributed components of that memory. Activating a subset of the markers could be sufficient for accessing at least some of its components, and in response, trigger generating of an updated set of markers that might be more comprehensive for accessing unrecalled components.

Of course this doesn't directly address the question of techniques at the end of your post, but it might, perhaps, offer a pointer thereto.

Kind regards!

As far as I understand it from reading Bjork's "A New Theory of Disuse" (1992), the reason for accelerated benefits from more difficult-to-retrieve memories is because the more difficult the retrieval, the more that instance of retrieval influences storage strength. Storage strength being defined as something like "how much the organism has changed as a result of exposure to the stimulus." Easier to retrieve items influence storage strength less. And the more storage strength, the less the retrieval strength diminishes with disuse.

Good question about the empirical evidence. The theory does seem to predict and explain certain results that have been found relating to the spacing effect, testing effect, and the modification of memories upon reconstruction. In general it seems highly plausible to me. The only thing I want a better explanation for is the idea that storage strength is 'unlimited.'

That being said, I haven't found any critiques of the theory, so I'm not sure how far it has been supported, and whether it has been refuted by any evidence.

http://bjorklab.psych.ucla.edu/pubs/RBjork_EBjork_1992.pdf

Thanks, David, for your reply and the attached Bjork paper - I look forward to digesting it!

Richard