Complete ignorance is not especially fertile ground for questions.

If you know nothing about computer memory, you are unlikely to wonder why an SR latch has two cross-coupled gates, why one input is called reset, or how that tiny circuit becomes a register. Those questions become available after you know enough to see what is strange about the thing in front of you.

This matters because curiosity is usually described as an aid to learning: become curious about an answer and you may pay more attention to it and remember it better. A new August 2026 experiment pushes that account further by explicitly connecting curiosity with attention, encoding, and memory over time. But the more interesting possibility is recursive. What you know affects what you can become curious about. What you learn by following that curiosity changes what you know. Then a new set of questions becomes possible.

A subject can begin to open from the inside.

Curiosity needs something to work on

The intuition that ignorance produces curiosity sounds plausible, but psychological research has repeatedly found a more complicated relationship.

In a 2019 study, Shirlene Wade and Celeste Kidd examined how prior knowledge and curiosity related to later learning. Curiosity was not simply highest where knowledge was lowest. People need enough familiarity to recognize that there is something missing. Prior knowledge and curiosity each predicted subsequent learning. Wade and Kidd, 2019

That is close to the original information-gap account of curiosity. A gap can motivate information seeking only when the learner can perceive the gap. “I know nothing about this” is often less compelling than “I understand everything except this one part.”

Richárd Reichardt, Bertalan Polner, and Péter Simor tested a version of this experimentally in 2023. Participants first read short encyclopedic passages on several topics and were later shown related trivia questions whose answers were not contained in those passages. The effects differed somewhat depending on how knowledge was measured, but the reading could change subsequent curiosity about related questions and could strengthen the relationship between curiosity and memory. Reichardt, Polner, and Simor, 2023

A few paragraphs of background had altered the questions people wanted answered.

That result should not be stretched into a claim that every introductory reading creates lasting interest. It does establish something more modest and useful: learning a little can change the motivational value of what comes next.

The rich really can get richer

Prior knowledge also makes later learning easier in ways that do not depend entirely on curiosity.

Amber Witherby and Shana Carpenter tested this in three experiments involving the domains of football and cooking. Participants who began with more knowledge in a domain were better at learning additional information from that same domain. The effect did not simply generalize across domains. Knowing more about football predicted learning new football information, not new cooking information. Witherby and Carpenter, 2022

Their third experiment is particularly relevant here. Participants rated how curious they were to learn each new item, and curiosity statistically mediated part of the relationship between prior knowledge and subsequent learning.

The experiment used deliberately false facts so that the information would be genuinely novel to every participant. That makes it a strange model of ordinary education, and mediation is not the same thing as demonstrating the entire causal mechanism. Still, the pattern is useful. More domain knowledge was associated with more effective learning of new domain information, and curiosity appears to be one contributor.

There are several reasons prior knowledge could help. New information has more places to attach. Relevant concepts are easier to recognize. Contradictions and missing pieces become visible. A beginner hearing that a CPU has a cache may simply add another noun to a list. Someone who already understands the speed difference between processors and main memory can immediately ask: where is the cache, what goes into it, and how does the machine decide what to keep there?

The same fact has become a better question generator.

A new experiment follows curiosity into attention and memory

Danlu Lei and Laim Adresion's August 2026 paper, “Curiosity-driven learning: How intrinsic motivation shapes attention, encoding, and long-term memory,” tries to connect curiosity to the cognitive events between wanting an answer and remembering it later. Their experimental design combines a curiosity manipulation with measures of attention, immediate recall, and delayed memory, then models the relationships among those measures. Lei and Adresion, 2026

The paper is useful because it asks a better question than whether curious people study longer. Curiosity may alter what receives processing while learning is taking place.

It should still be read narrowly. An immediate recall score is not a direct measurement of a memory trace being “encoded,” and a statistical mediation model is not a photograph of a causal mechanism. The larger literature is also too mixed to justify a simple claim that curiosity switches the brain into a generally superior learning mode.

In fact, curiosity can make memory for other information worse.

Nicole Keller and colleagues demonstrated that in a 2024 npj Science of Learning study. Participants saw trivia questions that produced either high or low curiosity, with unrelated scholastic facts presented nearby in time. Across three versions of the task, memory for those scholastic facts was worse when they appeared around high-curiosity trivia. Keller et al., 2024

That result matters because it kills an attractive educational trick. You cannot assume that making someone intensely curious about an unrelated question will create a motivational glow that improves whatever lesson you slide in next. Curiosity can prioritize its target strongly enough that competing information loses.

For self-directed learning, that makes the choice of question more important, not less. The useful curiosity is curiosity about the material you are trying to understand.

Curiosity may help choose what to learn next

The strongest evidence for following curiosity as a path through a subject comes from a different kind of experiment.

In 2021, Alexandr Ten, Pramod Kaushik, Pierre-Yves Oudeyer, Jacqueline Gottlieb and colleagues gave 382 participants freedom to choose among learning activities of different difficulty. The researchers were interested in whether human exploration is sensitive to learning progress: not simply choosing what is novel or easiest, but gravitating toward activities where understanding is improving. Ten et al., 2021

It was. Participants' choices tracked their progress, supporting a longstanding proposal in curiosity research: curiosity-driven exploration can help people organize a learning curriculum over time.

This gives “follow your curiosity” a more defensible meaning.

It does not mean opening a textbook at random and trusting motivation to supply the prerequisites eventually. It means that once you have entered a domain, your current state of understanding may contain information about the next productive edge. Things that are already obvious offer little gain. Things far beyond your present model may offer little traction. Somewhere between them are questions you can almost formulate and answers you can almost understand.

That edge moves as you learn.

Developmental research has long observed a related preference for intermediate complexity. In the well-known “Goldilocks” experiments, infants allocated attention away from sequences that were either very predictable or very surprising. The finding is not a theory of adult curriculum design, but it illustrates a broader point: attention is not simply pulled toward maximum novelty. Kidd, Piantadosi, and Aslin, 2012

For an adult learning a field, the productive question may likewise be the one that is neither already answered nor so remote that the answer has nothing to connect to.

New knowledge can create new gaps

A 2026 theoretical paper by Alice Xu, Catherine Sandhofer, and James Stigler makes this recursive idea explicit. Their SEEK model treats curiosity as part of a longer process of conceptual development: curiosity drives exploration, exploration modifies a learner's schemas, and those changed schemas create new prediction errors and knowledge gaps that can prompt further curiosity. Xu, Sandhofer, and Stigler, 2026

SEEK is a theoretical model, not an experiment proving that every learner follows this sequence. But its proposed cycle connects several empirical findings that otherwise look separate.

Prior knowledge can influence curiosity. Curiosity can direct attention and information seeking. Domain knowledge can make subsequent learning easier. People engaged in curiosity-driven exploration can prefer activities where they are making progress. Change the learner's knowledge and you change the landscape in which the next question appears.

Imagine beginning computer architecture because you want to know how a circuit can remember one bit.

The answer leads to an SR latch. Understanding the latch requires logic gates. Gates raise questions about what a transistor is actually doing. Several latches become registers. Registers raise questions about how values move around a processor. Soon buses, clocks, caches, instructions, and memory hierarchies are not disconnected headings in a syllabus. They are answers to questions produced by the previous answers.

Nothing guarantees that this route will cover everything. You may happily learn how caches work while avoiding binary arithmetic or never acquire the mathematical foundations needed for more advanced work. Curiosity is quite capable of producing an eccentric education.

That is why it works better as a scheduler inside a chosen domain than as the sole designer of a curriculum.

Pick something you mean to learn. Within it, begin with a question whose answer you genuinely want. Learn enough to answer it properly, not merely enough to collect the fact. Then pay attention to what has become newly confusing, surprising, or incomplete. Follow those adjacent questions while occasionally checking whether the route is leaving important foundations untouched.

The first question does not have to be the most foundational question in the field. It can help build the foundation that makes the foundational questions intelligible.

Sources

Ulix app

Guten

Guten is a free Android reader for more than 70,000 public-domain books from Project Gutenberg, with offline reading, search, highlights, notes, and optional Premium features.