Why do we forget what we learned at school?

Do you remember what you learned at school? Chances are you have a vague idea but the details are a bit fuzzy. Or you might not recall anything. Surely, however, if learning really is an ‘alteration in long-term memory’ then we shouldn’t forget what we’ve learned? So, perhaps, we forgot because we didn’t learn it in the first place, despite getting all those grade A grades in our exams? Like everything else in cognitive psychology (and cognitive science more widely), there are several possible explanations. Here are just a few of them.

You haven’t really forgotten

This might seem a rather arrogant suggestion, but stay with me for a moment. It’s likely you’ve lost much of the detail of what you learned, but retained some of the gist (see here for a discussion about gist memory), with varying degrees of accuracy. Chances are, if you decided to study, say Physics, later in life having seemingly forgotten everything you learned in school, the process might be easier (dependent on many factors, including, unfortunately, age related cognitive decline). There may well be a familiarity to what you learn because past learning is playing an important role. You recall more detailed information if you were provided with nudges, prompts, cues and clues. Often, recalling the context will also give our memory a nudge; oh, that was what Miss Connor was teaching us when Kyle Reese fell off his chair and broke his arm. Alternatively, significant words, phrases or providing the bare bones of a concept can be enough to shift attention towards a more detailed description. 

You have actually forgotten

No matter how well we learn something, time is always going to conspire against us. Items in long-term memory can potentially last a lifetime, the important word here being ‘potentially’. It’s an unfortunate truth that our ability to accurately recall some information declines with the passing of time. In this respect Ebbinghaus was certainly correct, but as I’ve already pointed out here, the forgetting curve can be somewhat misleading if we neglect the role of meaning and context. This forgetting might be due to decay or interference (we can’t scientifically prove that it’s definitively one or the other). Certainly, remote memories (memories of events that occurred many years ago) can become more vulnerable to forgetting and researchers are beginning to come around to the idea that forgetting is a process that is intentionally built into brain networks (Gravitz, 2019).

Even when we believe we correctly recalled something, chances are we haven’t. One good example of this pertains to what cognitive psychologists refer to as flashbulb memories. A flashbulb memory is one that was, at the time of experiencing it, so emotionally charged that the details appear to have been etched onto our brains. Examples might be the assassination of John F. Kennedy or the 9/11 terrorist attacks, in that people can generally recall when and where they heard the news in great detail. Unfortunately, so-called flashbulb memories seem just as prone to error than more mundane ones and often rely on context. One of the founders of cognitive psychology, Ulrich Neisser, tells the story of when he first heard about the Japanese attack on Pearl Harbor. Neisser had, for years, been convinced that he was listening to a baseball game on the radio when he heard the news, realising many years later that the attack on the 7 December was outside the baseball season. Indeed, the more confident we are that our memories are accurate, the less our actual memories appear to be (I’ve discussed this is greater deal in The Emotional Learner). On the other hand, rather than decay, your memory might have fallen foul of interference, in that things you learned after your exam have corrupted and permanently altered the nature of the original memory. The problem here is that, despite their best efforts, psychologists have never really managed to decide amongst themselves if forgetting occurs through decay or interference. 

Errors have occurred during the memory cycle

Encoding is the first stage of the memory cycle. Somehow, the newly learned information needs to be transformed into something that can be stored and retrieved later, and it’s likely this process involves different brain regions, including the hippocampus and the medial frontal lobe. After encoding, the information is consolidated, probably during sleep or periods of wakeful rest (see here for more on wakeful rest) and connections between new and older memories will be strengthened. Consolidation may also involve the stripping away of episodic information, such as the time and the place of where the information was first encountered. One suggestion is that the information eventually becomes ‘hippocampally independent’, that is, the hippocampus plays no further role in storage or retrieval, and the memory is dispersed amongst several different areas of the cortex. Another theory proposes that the hippocampus only plays a role in retrieval under specific circumstances.

Whatever the process, the very act of retrieval will alter the information again, whereby it will be re-consolidated. There’s still a great deal we don’t about this process, but it seems that some memories will go through several rounds of re-consolidation, perhaps for weeks, months or even years. Through this process of encoding, consolidation, retrieval, and re-consolidation, schemas emerge (see below). There are many reasons why this information may have been inadequately encoded. If you crammed for your exam the night before, rather than spreading your revision over a much longer period, you may well have done enough to pass the exam, but not retain the information for very long. In an ideal world, learning would have taken place over time, and topics returned to at different stages and revision would have started early. However, learning more often than not takes place in blocks, so one topic is covered before moving onto another with little opportunity to refresh. But errors may have been encountered at any stage, for example, if the information retrieved following consolidation is inaccurate it will then be re-consolidated and the errors strengthened.  

The success of this process is dependent on several factors, including rate of refreshing (returning periodically to the to-be-learned information) and the depth of processing, including elaboration. We can learn something superficially, for example, learning a list of words or dates by rote, or deeply by making connections between related information or through techniques like elaborative interrogation. You might not remember what you learned over the long-term because learning was superficial and, therefore, was less able to remain relatively permanently in long-term memory.

Inadequate formation of schemas or assimilation into existing ones

You may have realised already that this is also related to consolidation. Schemas are curious things. While they represent a very useful explanatory tool, they remain hypothetical, a little bit fuzzy in terms of definition and detail and virtually impossible to prove. Bartlett (1932) described schemas as ‘a structure that people use to organise current knowledge and provide a framework for future understanding’. From the perspective of cognitive neuroscience, van Kesteren and Meeter explain them in terms of ‘a network of neocortical representations that are strongly interconnected and that can affect online and offline processing.’ This rather computer metaphor heavy definition simply means that these connections can affect processing when we are actually learning (online) and when we aren’t (offline). When we talk about consolidation and sleep, this is a process that takes place offline. 

If we adopt a broad definition of learning; to form an internal model of the external world, schemas, then, represent these internal models. When we learn something new, this information is said to form related clusters of knowledge, probably hierarchical is nature. The number of schemas increase as we experience the world, and they provide useful short-cuts and prevent us from having to learn new information from scratch if the information is already somehow represented in one of these many schemas. If we learn about a new kind of insect, we can slot this new find into our insect schema. Tse et al. (2007) discovered that rats were more quickly able to encode and consolidate new information when they fitted with a spatial schema, equating to the ‘more you know, the faster you learn’ view. Neurobiological explanations of schema formation are still in the early stage, and several theories currently exist. But cognitive psychology has been looking at them for some time. 

One very influential view is that the creation and maintenance of schemas represents accumulated learning, so the more schemas we have (or the more we know), the easier it is to learn more (an important notion within Cognitive Load Theory). It would, therefore, make sense that if new learning hasn’t been adequately integrated, assimilated (or whatever) into a schema, the memory trace is going to remain pretty weak. 

Completion (the Zeigarnik effect)

The final proposal is perhaps a little more speculative and involves a phenomenon known as the Zeigarnik effect, after the Lithuanian psychologist Bluma Zeigarnik (1901-1988). The theory centres around an incident involving Gestsalt psychologist Kurt Lewin (Zeigarnik’s supervisor). Although there are a couple of different versions of the story, the gist of it involves Lewin and some friends at a restaurant. In one version from Boring (in his 1957 History of Experimental Psychology), Lewin called the waiter over, asked how much he owed and was told instantly without the waiter having to check. Lewin paid, the waiter left and the friends continued their discussion. He then had an ‘insight’, called the waiter back, and asked him how much he’d been paid. The waiter could no longer recall. In a second version, Marrow (in his 1969 book on Lewin) quoted Donald MacKinnon, who was one of Lewin’s party that day. MacKinnon states that someone called the waiter over and that the waiter knew exactly what everyone had ordered, even though he hadn’t kept a written record. The bill was paid, and they continued their discussion. Then, about half an hour later, Lewin called the waiter over and asked him to write the bill again. The waiter explained that the bill had been paid, so he no longer remembered what everyone had ordered.

From this tale, Zeigarnik embarked on her own investigation to establish the cause of the waiter’s forgetting, concluding that a task that has already been started, establishes a task-specific tension, which improves cognitive accessibility of the relevant contents. However, once the task in complete, the tension is relieved but will continue even if the task is interrupted. It’s a stretch, but we could conclude from this that students continue to remember information they’ve learned throughout the period of study, but once the exam is over, the learning is seen as complete and therefore no longer requires remembering.

No single explanation (sorry)

These are only a few possible suggestions, and I’m sure there are more. Unfortunately, there is no complete answer to the question and the best we can do is offer the best explanations based on what we know. This is without entering into an extended debate over individual differences, the fact that we won’t all forget for the same reason or at the same rate. It’s also worth stressing that we do, indeed, remember a great deal of what we learned in school, such as how to read, write and get to grips with basic arithmetic. Furthermore, the gist of what we learned may well be enough to fuel a passion in some subject or topic, so we may love literature without being able to recall very much about the books we studied at school.