Is Forgetting Part of Understanding?

Geoffrey Chen

How a Person Builds a World · Season One: “Reading, Memory and Judgment” · Article Five

One student can reproduce a chapter’s definitions, examples and order, yet cannot apply its principle to a slightly different problem. Another has lost most of the original wording but recognises the same structure in a new situation. Which one understands better?

The comparison does not justify the slogan that more forgetting produces deeper understanding. Forgetting may simply mean that something was never learned. It can remove decisive evidence, qualifications and provenance together. But the contrast requires us to admit that complete retention and effective understanding are not identical. Understanding needs to preserve difference and to compress. It needs fidelity, but it also needs irrelevant detail to recede from the foreground.

The better question is not whether forgetting is good. It is what has been forgotten, what becomes visible as a result, and whether the lost material can be recovered when the judgement requires it.

Preservation without selection would overwhelm judgement

If every sensory feature of an experience, every word in an article and every minor event in a day returned with equal force, judgement would struggle to operate. A current problem requires some relations to become prominent and a great deal of other material to remain quiet.

The distinction between gist and peripheral detail in memory research provides a limited but useful example. In experiments involving prose recall over time, peripheral details have generally become inaccessible more quickly than central meaning. See the study of different forgetting rates for gist and peripheral details. This does not show that memory always selects the genuinely important material. “Central” information may be defined partly by a task or researcher. It does show that elements of an event do not all fade in the same manner.

Some work on categorisation goes further, proposing that reduced access to the particular features of individual examples may assist abstraction and generalisation. As surface details become less available, recurring relations may better support judgements about a new case. The research on forgetting, abstraction and the temporal dynamics of categorisation concerns controlled learning tasks, not the whole of human intellectual life. Still, it reveals an important possibility: no longer carrying the complete surface of every example at once may be a condition for forming a general concept.

The concept of a tree would be unusable if every encounter had to reactivate every detail of every tree. A concept works by ignoring some difference. Yet if it ignores too much, important distinctions among a palm, a pine and a shrub disappear. Understanding operates between compression and preservation.

Forgetting can enable learning, and it can destroy knowledge

Research on learning contains a counterintuitive result. Conditions that make retrieval somewhat difficult—spacing, testing and interleaving among them—can produce better long-term retention and transfer than continuous repetition. The phrase “desirable difficulties” describes this family of effects. Difficulty is not valuable because suffering is virtuous. A demanding but successful retrieval can strengthen and modify access to knowledge. This overview of “Remembering What We Learn” explains several of the relevant findings and their educational implications.

Temporary inaccessibility does not therefore prove that knowledge has disappeared. After a delay, the attempt to reconstruct an argument before returning to the text can reveal gaps more effectively than another immediate reading while every sentence still feels familiar.

The word “desirable” has boundaries. Fragmented attention, material far beyond a learner’s foundations, or a delay so long that no retrieval succeeds may simply produce failure. Researchers have emphasised that not every obstacle benefits learning and that the effect depends on the person, task and available support. The discussion of undesirable difficulty in high-element-interactivity materials is one reminder that a useful principle can be overgeneralised.

“It is fine to forget” cannot become a reassuring slogan. In a medicine dose, legal condition, historical date or another person’s exact statement, detail may be the centre of the judgement. A reader who loses the qualifier while retaining a vivid conclusion has not gained deep understanding. The reader has acquired a more dangerous generalisation.

Retrieving one item can make another less available

Forgetting is not always a quiet decay. New material can interfere with old material, and repeated retrieval of selected content can make related but unpractised content harder to access. The latter phenomenon is called retrieval-induced forgetting. A review of interference, inhibition and retrieval-induced forgetting shows that mechanisms remain debated, but competition and selection are plainly important parts of remembering.

The implication for intellectual life is direct. If a person always recounts a book through the same summary, that path becomes fluent while alternative understandings become less available. When public culture repeatedly quotes a single sentence from a thinker, other parts of the work lose visibility. This is not merely forgetting inside an isolated brain. Education, media, indexes and habits of citation help decide which path is repeatedly retrieved.

What feels like memory for “the most important thing” may sometimes be the advantage acquired by whatever we have most often been asked to recall.

Understanding therefore requires occasional changes of retrieval route: explain the book without using the familiar summary; seek the evidence that least fits the central account; read the chapters without annotations; or construct the strongest version of a rival interpretation. The aim is not to retain everything equally. It is to prevent one accessible path from monopolising the work.

Understanding needs reversible forgetting

The ideal knowledge life does not keep every detail permanently in conscious foreground. It lets material recede when it is not needed while preserving a path of return for verification. We might call this reversible forgetting.

A person need not remember every page number but should know where an important claim came from. Nobody must carry every item of evidence at all times, but public argument cannot rely only on a vague impression. Old views need not occupy attention indefinitely, but enough record should remain to show why they changed.

Books, notes, references, libraries and digital archives matter because they allow internal memory to put material down without making that release an irreversible severance. External preservation expands the safe range of forgetting. We need not carry everything in the head if we know how to return and if the source has not disappeared through platform failure, obsolete formats or neglect.

This changes the meaning of knowing. Knowledge is not always instantaneous recitation. It may be the ability to identify a problem, locate reliable material, rebuild an argument and state uncertainty. Distributing knowledge between memory and external structures does not diminish the human contribution. It allows a finite person to sustain more complex understanding.

Forgetting enables structure but does not guarantee truth

In Sustenesis terms, every structure capable of continuing operation must select. It strengthens some relations while allowing others to withdraw; under new conditions, it may reorganise again. A structure that selects nothing cannot develop a direction. A selection that cannot be reversed becomes rigid.

Forgetting performs a related function in understanding. It prevents all details of experience from pressing permanently upon the present, allowing concepts, patterns and questions to emerge. Yet forgetting does not automatically preserve truth. It can remove exceptions, sources and the costs borne by other people, leaving a clean story constructed from excluded evidence.

The right response is therefore double. We should allow forgetting to support compression and transfer while building external records and habits of rereading that can correct its bias.

Forgetting can be part of understanding, but only when the resulting structure remains answerable to reality and to sources. Understanding is neither remembering everything nor declaring whatever has been forgotten unimportant. It is knowing what can safely recede, what must not be lost, and how to bring background material forward when a judgement needs it again.

Primary sources and further reading

Continue reading: Explore the How a Person Builds a World series.

Discover more from Geoffrey Chen

Subscribe to get the latest posts sent to your email.