A New Perspective on Learning, Memory, and Concentration

A single idea connects many of the articles on this website: the human mind is not primarily a system that reacts to the past but one that continuously predicts the future. I have explored this idea in relation to hypnosis, communication, sports psychology, learning, and memory because I believe the same underlying neural principle can be observed across many different areas of human experience. The more I have studied the subject, the more often I have found myself returning to the same question: could prediction itself be the most fundamental mechanism of learning?

My interest in this question did not begin with artificial intelligence. Long before AI became part of everyday discussion, I was fascinated by the practices of Zen Buddhist monks and their pursuit of deeper states of consciousness. I never regarded these practices merely as religious rituals. Instead, I saw them as an exploration of how the human mind functions when attention becomes exceptionally stable and free from distraction.

Over the years, I approached the same question through many different paths: transactional analysis, psychodrama, sociodrama, mental coaching for athletes, clinical hypnosis, self-hypnosis, and, more recently, mindfulness. Although these traditions use different concepts and techniques, I have gradually become convinced that they share a common goal: helping the mind focus on what truly matters while reducing internal distraction. The methods differ, but the destination may be remarkably similar.

More recently, another field has strengthened this line of thinking: language learning. I have followed the ideas of polyglot Steve Kaufmann and explored the work of language-learning researchers such as Stephen Krashen and Paul Nation. While they express themselves differently, they all point toward the same principle: learning is most effective when the input is understandable but still presents an appropriate level of challenge.

If everything is already familiar, very little learning takes place. If almost everything is unfamiliar, comprehension breaks down. Learning appears to flourish somewhere between these two extremes, where the brain's predictions are mostly correct but still require continuous adjustment.

Consider three simple examples.

A learner of English reads the sentence:

"The little boy climbed the tree because he wanted to see the birds."

If every word is already familiar, comprehension is effortless. The reader predicts the sentence almost automatically, but little new learning occurs.

Now imagine a different sentence:

"The diminutive lad ascended the sycamore to observe the migratory thrushes."

Here the learner encounters so many unfamiliar words that understanding begins to collapse. Prediction error becomes too large, and working memory is overwhelmed by decoding individual vocabulary items instead of constructing meaning.

The ideal learning situation lies somewhere between these extremes:

"The little boy climbed the oak tree because he wanted to observe the birds."

Most of the sentence is immediately understandable. Only one or two words, such as observe, require inference from context. Because the overall meaning is already established, the new word naturally finds its place within the learner's existing mental model. Learning occurs almost effortlessly.

This example illustrates what I find particularly interesting about Predictive Mind (predictive processing). Perhaps learning does not primarily occur because we consciously memorise new information. Instead, it happens because the brain generates predictions, detects a manageable prediction error, and updates its internal model accordingly.

If prediction error is too small, nothing changes.

If prediction error is too large, the internal model breaks down.

Learning seems to occur somewhere in between.

Interestingly, large language models operate according to a surprisingly similar principle. They continuously predict the most probable next token based on previous context. Human brains, however, appear to perform a far richer version of the same process. We predict not only the next word but also the next idea, the next meaning, and even the author's intention.

Eye-movement research offers fascinating evidence for this perspective. Our eyes do not move smoothly across a line of text. They pause, jump forward, and frequently move backwards. These regressions were once regarded simply as signs of reading difficulty. Today they are increasingly understood as part of normal comprehension. When incoming information does not fully match the brain's expectations, the eyes revisit earlier text while the brain updates its internal model.

Perhaps our eyes do not merely follow our thoughts.

Perhaps they actively participate in constructing them.

This raises another intriguing possibility.

What if concentration is not primarily a matter of willpower?

What if effective concentration is actually the ability to maintain a stable predictive model of whatever we are currently reading, listening to, or learning?

When that internal model remains coherent, each new sentence extends the previous one. Working memory becomes less overloaded because new information is continuously integrated into an existing structure instead of being stored as isolated facts. Memory may therefore be less about storing information than about successfully constructing meaning.

This brings me back to Zen meditation, mindfulness, hypnosis, and self-hypnosis.

Research suggests that experienced meditators are able to reduce mind wandering and maintain stable attention more effectively than most people. Some researchers have even begun to interpret meditation through the framework of predictive processing, proposing that meditation reduces competing predictions and helps attention remain focused on what is currently relevant.

This naturally leads me to another question.

Could hypnosis and self-hypnosis achieve something similar through different methods?

Zen meditation, mindfulness, hypnosis, and self-hypnosis each use different techniques, but perhaps they share an underlying objective: reducing competing predictions while strengthening the stability of the brain's internal model.

At this point, however, it is important to distinguish established research from my own theoretical reflections.

There is substantial scientific evidence supporting the effects of meditation on attention, eye movements during reading, and the importance of comprehensible input in language learning. I have not found comparable research examining hypnosis or self-hypnosis specifically through the combined framework of predictive processing, eye movements, learning, and memory.

That connection remains my own hypothesis rather than an established theory.

If I were to design a study, I would compare four groups of students. One group would simply read a text. A second group would complete a brief mindfulness exercise before reading. A third group would practise self-hypnosis to enhance concentration. A fourth group would receive self-hypnotic suggestions specifically designed to strengthen prediction of the text's meaning—for example, encouraging each new idea to connect automatically with the previous one.

Eye movements, working-memory load, reading comprehension, and later recall could then be measured and compared.

Perhaps the earliest differences would not appear in memory tests at all.

Perhaps they would already be visible in the way the eyes move across the page.

The longer I reflect on these questions, the more I see a common principle emerging across fields that are usually studied separately. Artificial intelligence predicts the next word. The reader predicts the next idea. Eye movements continuously gather the information needed to verify those predictions. Zen practitioners cultivate a stable Predictive Mind. Mindfulness gently returns attention to what matters most, while hypnosis and self-hypnosis may offer an alternative route toward the same objective.

I do not know whether this explanation is ultimately correct. At present it is a hypothesis rather than an established theory. Nevertheless, I find it compelling because it offers a coherent framework capable of connecting several apparently unrelated fields of research. If the Predictive Mind proves to be the common denominator linking language learning, eye-movement research, meditation, hypnosis, and artificial intelligence, we may eventually come to understand learning from an entirely new perspective.

Perhaps the most important learning skill is not memory itself. Perhaps it is the ability to build each new thought upon the one before it.