Great Ideas from the Brain Sciences: The Work Behind Working Memory
By: Mirela C. C. Ramacciotti
How much work do you think students have to put into tasks to learn? If this question baffles you, let me break it down. For starters, there is a variety of the kinds of work learning demands. When students deal with information that has to be available, accessible and processed–in a conscious manner–to perform a task, we are talking about a type of short memory that lasts around 30 seconds. The name, Working Memory, or WM in short, is the topic of this Great Ideas in Brain Science.
In 1956, before the term WM existed, the psychologist George Armitage Miller published a study where he described the limited memory capacity of adults. The magical number seven plus or minus two soon became a hit. It brought about the notion that humans’ abilities are limited to about 7 things. And this number may change, that is, it might be nine or five things depending on the task. You can only imagine the uproar that piece of information caused. Publicity soon took hold of the notion. However, what publicity failed to divulge is that the seven verbal items that Miller proposed need some support to be effectively used. That support would be strategies like verbal rehearsal or chunking–a term coined by Miller to name a meaningful unit of information. A clear example is when single items, like the letters M; B; E; J; A; L; T, are more easily remembered when put together in acronyms, like MBE and JALT.
Miller’s idea that there is a limit to how much we can store (an item limit) soon expanded from capacity only and went on to be explored in relation to time. That is when Alan D. Baddeley, and Graham J. Hitch, both of them English psychologists (more on these two researchers in a forthcoming Great Ideas article), proposed the notion that WM was made of different components jointly employed to manipulate the information so that it remained available. The three components are: a verbal storage system called the phonological loop, a visual storage system called the visuospatial sketchpad, and a central executive.
A first piece of information very relevant for language teachers is that the phonological loop, which can hold memory traces for a few seconds, involves an articulatory rehearsal process. Therefore, two processes, retrieval and re-articulation, need to be used to refresh memory traces. As Miller forewarned, there is a limit here. It has to do with articulation. WM is limited by the amount of material that can be articulated before the first stored item fades. As you probably noticed, time was a question that Baddeley explored well.
A second piece of information relevant for language teachers is that phonological loop capacity is a good predictor of second language learning. This goes in tandem with theories of language development that take the phonological loop as an evolutionary feature that facilitated language acquisition. For example, take a look at this article on Aboitiz’s model and this other one on the Connectionist model.
Now that we have explored two of the WM components in the Baddeley and Hitch’s model, let’s move onto the third: The central executive, or the overarching processor. It accounts for how we focus, divide and switch attention. Notice that these are different processes that involve a cognitive cost. The main areas involved in this processor are in the frontal lobes and answer for executive control. Now, that is an intriguing idea for executive control is the supramodal function of attention. In other words, it works to optimize how we use attention. As attention is central to how we perform actions, and those depend on the information that we store (verbally or visually) and manipulate (in articulation or rotation), then it makes sense to have a mechanism that centralizes and optimizes WM, much like the work of a conductor in an orchestra.
To recap, Baddeley and Hitch’s WM model works with two distinct modules, one for verbal items and another for spatial and visual items. These modules depend on attention optimization performed by a third component, the central executive. Now, you may be wondering about how information that we manipulate gets transformed into our treasure trove of knowledge. That is exactly why a fourth component was added to the model.
In 2000, Baddeley introduced the episodic buffer to account for how we chunk information into episodes–thus limited in time–that interact with our long-term memory.
Now, let’s take a step back to recall Miller’s original quest involving limits. You remember that two limits emerged: one for time, another for storage. While much work advanced the quest for how long items remained in WM, the question on how much we could store and manipulate–chunk limit–was not as well researched till the turn of the century. In 2001, Nelson Cowan published groundbreaking work showing that, when strategies for rehearsal are not used, only 3 or 4 separate objects or chunks would be remembered when presented at the same time.
Cowan remains an active and very productive mind in the area of WM. In recent work, and with several colleagues, he has demonstrated that WM performance changes, either in plus or minus conditions as per Miller’s work. These changes accompany development and are of different kinds. Quantitative changes mean an increase in how much information we hold while qualitative changes mean that the mechanisms involved in WM also evolve, i.e., they mature. That’s why, in regular classrooms, teachers usually notice a marked increase in WM during adolescence, with a peak at young adulthood, and a decline in old age.
Talking about variance, there is a final recommendation from Cowan’s work (2014, 2022) that is gold for teachers: instead of pushing learners to overburden their WM and use up precious resources, teachers should pay heed to learners’ cognitive level and processing limits and adapt learning materials accordingly.
In closing, I myself often wonder about how we, as children, change from being guided by constrained processes to deliberate actions and behaviors. Cognitive levels and control make for a really fascinating notion. But that is a topic for another article. I hope you remember to check that out in the near future.
Sources and More
Baddeley, A. D., & Hitch, G. J. (1974). Working memory. In G. Bower (Ed.), The psychology of learning and motivation (Vol. 8, pp. 47-90). Academic Press.
Baddeley, A. D., & Hitch, G. J. (1994). Developments in the concept of working memory. Neuropsychology, 8(4), 485-493. https://doi.org/10.1037/0894-4105.8.4.485
Baddeley, A. (2003). Working memory: looking back and looking forward. Nature Reviews Neuroscience 4(10), 829–839. https://doi.org/10.1038/nrn1201
Cowan, N. (2001). The magical number 4 in short-term memory: A reconsideration of mental storage capacity. The Behavioral and Brain Sciences, 24(1), 87–114. https://doi.org/10.1017/s0140525x01003922
Cowan, N. (2014). Working memory underpins cognitive development, learning, and education. Educational Psychology Review, 26(2), 197–223. https://doi.org/10.1007/s10648-013-9246-y
Cowan, N. (2015). George Miller’s magical number of immediate memory in retrospect: Observations on the faltering progression of science. Educational Psychology Review, 122(3), 536–541. https://doi.org/10.1037/a0039035
Cowan, N. (2022). Working memory development: A 50-year assessment of research and underlying theories. Cognition, 224, 105075. https://doi.org/10.1016/j.cognition.2022.105075
Superbia-Guimarães, L., & Cowan, N. (2023). Disentangling processing and storage accounts of working memory development in childhood. Developmental Review, 69, 101089. https://doi.org/10.1016/j.dr.2023.101089
Mirela C. C. Ramacciotti is presently engaged as an external lecturer on the topic of Mind, Brain, and Education at the Graduate Level Course with the Psychology Department at the University of São Paulo. She holds a PhD in Neuroscience and Behavior and another in Human Communication Disorders.