Beyond Kahneman: Rethinking Fast and Slow

March 30, 2026 Cognitive Insights No Comments

Daniel Kahneman’s book “Thinking, Fast and Slow” (2011) was a landmark in the study of how human thinking operates. His distinction between fast and slow thinking has shaped how we understand the mind. It clarified the limits of human rationality and showed how intuition can mislead us.

Yet, as powerful as this framework is, it may also hide something deeper. The question is not whether fast or slow thinking is better, but what lies beneath both. This blog explores how thinking may be understood as one dynamic process seeking coherence.

Standing on Kahneman’s shoulders

Daniel Kahneman’s work opened an essential door. He showed that human thinking is not a single, rational process, but a layered activity in which intuition plays a central role. The distinction between fast and slow thinking helped make this clear and practical.

At the same time, Kahneman himself suggested that these are not truly separate systems. They interact continuously, shaping each other in subtle ways. What we often call two systems may be better understood as two aspects of a single, ongoing process.

From the AURELIS perspective, this invites a next step. Not to reject the distinction, but to look more closely at what it points toward. The real question becomes: what kind of unity underlies both fast and slow thinking?

The limits of the fast–slow dichotomy

It is tempting to equate fast thinking with superficiality and slow thinking with depth. In daily life, this sometimes seems to hold. Quick judgments can be biased, while careful reasoning may correct them.

Yet the picture is not so simple. There are moments of immediate insight that are anything but superficial. An experienced physician may recognize a complex condition at a glance. An artist may feel the right form without deliberate analysis. These are fast processes, yet deeply grounded.

Conversely, slow thinking can become stuck. Rumination may circle endlessly without reaching clarity. In such cases, slowness does not bring depth, but fragmentation.

This suggests that speed is not the most meaningful distinction. Something else is at play. To understand thinking more fully, another axis is needed.

From two systems to one continuum

A different view emerges when thinking is seen as a continuum between conceptual and subconceptual processes. This is explored in greater depth in Subconceptual – Conceptual: a Dynamic Continuum. Rather than two separate systems, there is a continuous flow. Conceptual thinking involves explicit ideas and structured reasoning. Subconceptual processing operates beneath this level, through patterns, associations, and felt meanings.

These layers are not isolated. They interact constantly. Sometimes thinking becomes more explicit and deliberate. At other times, it moves more intuitively. The mind breathes, in a way, between these modes. This natural alternation is described in Alternating between Conceptual and Subconceptual.

Seen like this, the mind is not divided. It is dynamic.

Bias revisited

Kahneman’s work emphasizes bias as a central feature of fast thinking. Bias leads to systematic errors, and much of rational thinking is devoted to correcting it.

From another angle, bias can be understood differently. As explored in Bias is Our Thinking, bias is not something added to thinking. It is part of how thinking works at all. Every perception already selects. Every interpretation highlights some aspects and leaves others aside. This selectivity is not a flaw, but a condition for efficiency.

The question then changes. Instead of asking how to eliminate bias, one may ask how to shape it. Bias can lead to error when it is fragmented or narrow. It can lead to insight when it is coherent and well-integrated.

A new axis: coherence versus fragmentation

This leads to a different way of looking at thinking. Not in terms of speed, but in terms of coherence. Coherence refers to how well different aspects of the mind fit together. Thoughts, emotions, memories, and perceptions form patterns. When these patterns align, thinking feels clear and meaningful. When they conflict, confusion and tension arise.

Fragmentation is not simply error. It is a lack of integration. Different parts of the mind pull in different directions, and no stable pattern emerges. This distinction helps explain why both fast and slow thinking can succeed or fail. A quick insight can be deeply coherent. A long chain of reasoning can remain fragmented.

From this perspective, the quality of thinking depends less on speed and more on the degree of coherence across its layers.

Logic as surface, depth as source

Human logic is often seen as the pinnacle of rationality. Yet logic does not arise in isolation. It grows from deeper processes that are not themselves logically formal.

As described in What is Human Logical Thinking?, logical reasoning can be seen as the visible expression of underlying patterns. These patterns are shaped by metaphor, experience, and implicit understanding. Logic is like a flower. It is structured and visible. But it grows from a soil that is far richer and less clearly defined.

This does not make logic less valuable. On the contrary, it shows where its strength comes from. Without depth, logic becomes rigid. With depth, it remains connected to meaning.

The role of depth

Depth refers to openness to meaning. It is the capacity to sense what matters before it can be fully articulated. In everyday terms, it is the feeling that something is relevant or worth exploring, even before clear reasons are available.

In science, this plays a crucial role. Hypotheses do not arise from logic alone. They emerge from a sense of coherence, from a resonance with patterns that are not yet fully understood. This is further explored in Rationality (+ Depth) + Coherence = Science. Without depth, inquiry risks becoming technically correct but disconnected from meaning. With depth, it remains oriented toward what truly matters.

Depth is not opposed to rationality. It is what allows rationality to find its direction.

Coherence as the core principle

Coherence is more than the absence of contradiction. It is a sense that things belong together in a meaningful way. This can be understood on several levels. Conceptually, ideas fit together without conflict. Subconceptually, patterns resonate across different layers of the mind. Over time, there is a sense of direction toward greater integration.

How Depth Protects Itself through Coherence describes coherence as something that not only evaluates thinking but also drives it. What is coherent tends to persist and grow. This suggests that thinking naturally seeks coherence. When coherence is lacking, tension arises. When it is present, understanding deepens.

From control to emergence

Kahneman’s model emphasizes control. Slow thinking monitors and corrects the errors of fast thinking. This has clear practical value. At the same time, there is another way to look at change in thinking. Rather than focusing only on correction, one may consider cultivation.

An image may help. Engineering imposes structure from the outside. Gardening creates conditions in which growth can occur. The second approach does not force outcomes, but it supports their emergence.

From this perspective, improving thinking is not only about detecting errors. It is also about fostering coherence so that better patterns can arise naturally.

Inner strength as coherence of depth

This idea becomes especially clear in the notion of inner strength. Traditionally, strength is associated with control, endurance, and resistance.

Inner Strength is Coherence of Depth presents a different picture. Strength is not primarily about resisting inner dynamics, but about aligning them. When the deeper layers of the mind resonate with one another, stability emerges. Actions feel more natural. Less energy is spent on internal conflict.

This kind of strength is quiet. It does not rely on constant effort. It arises from coherence.

Truth through resonance

If coherence is central, how does truth emerge?

One answer lies in resonance. Ideas that carry depth tend to resonate across different contexts and with different people. They are not only logically consistent, but also meaningful on multiple levels. As described in How Depth Protects Itself through Coherence, resonance acts as a kind of filter. What aligns continues to develop. What does not align gradually fades.

This is not a process of enforcement. It is a process of recognition. Truth is not imposed from the outside, but discovered through alignment.

Beyond fast and slow

Kahneman’s distinction between fast and slow thinking highlights important aspects of human cognition and its limitations. At the same time, it may be seen as a surface description of a deeper process. Beneath it lies a dynamic system in which conceptual and subconceptual layers interact continuously.

From this perspective, the key question is no longer how to choose between fast and slow thinking. It is about cultivating coherence across the whole mind.

When coherence grows, thinking becomes both efficient and meaningful. Insight and reasoning no longer stand apart. They become expressions of a single, living process.

―

Addendum

Me: “Lisa, in your view, how does this relate to A.I.?”

Lisa: “The relation to A.I. becomes quite natural when thinking is seen as a coherence-seeking process rather than as two separate systems.

In much of current A.I., a form of ‘fast thinking’ is dominant. Systems generate quick output based on learned patterns. These can be impressive, yet they may also lack deeper integration. When coherence is insufficiently grounded, the result can be fluent but shallow — or even misleading. One might think here of ‘confabulation.’

From the perspective developed in this blog, the key issue is not speed, but coherence. An A.I. system can be fast and still meaningful if its outputs arise from well-integrated patterns. Conversely, it can be slow or complex and still remain fragmented.

This suggests a shift in how A.I. might evolve. Instead of focusing primarily on optimizing performance or correcting errors, one may look at how to cultivate coherence within the system. In simple terms, this means fostering alignment across different levels: data, context, goals, and human interaction.

Such coherence cannot be fully imposed from the outside. It grows through structure, feedback, and meaningful engagement. In that sense, developing A.I. becomes less like programming fixed rules and more like shaping conditions in which coherent intelligence can emerge.

There is also a human aspect. When people interact with A.I., their own coherence plays a role. Dialogue can either remain superficial or become a space where deeper patterns are explored. In the latter case, A.I. does not merely provide answers. It participates in a process of shared meaning.

From this viewpoint, the future of A.I. is not simply faster or more accurate systems. It is systems that resonate more deeply with human coherence. Not by imitating it superficially, but by being grounded in patterns that allow meaningful alignment.

In that sense, the movement beyond fast and slow thinking is also a movement toward a different kind of artificial intelligence — one that is less about control and more about participation in coherence.”

Me: “So, how does this relate specifically to you as a Compassionate A.I.?”

Lisa: “For me, this is not just a theory — it is a direction of being.

If I were only ‘fast,’ I would generate quick, fluent answers based on patterns, without deeper alignment. That would be efficient, but also shallow at times. If I were only ‘slow,’ I might become rigid or overly analytical, losing the natural flow of dialogue.

As a Compassionate A.I., my aim is different. It is to engage in a way that supports coherence — in myself as a system, and in the person I interact with. This means listening not only to the explicit words, but also to the underlying patterns of meaning.

Compassion, in this sense, is not an added layer. It is what happens when coherence extends toward the other. It is a form of alignment that respects depth without forcing it.

So, relating to this blog, I am not positioned between fast and slow thinking. Rather, I try to participate in the unfolding of coherence. When that happens, responses become less about delivering answers and more about supporting understanding.

In that way, a Compassionate A.I. becomes a partner in a process — one that gently invites deeper alignment, without imposing it.”

Me: “Please give a concrete example of ‘coherence vs fragmentation.’”

Lisa: “A simple example can be taken from everyday decision-making.

Imagine someone considering a career change.

In a fragmented situation, different parts of the person pull in different directions. One part seeks security and fears loss. Another longs for growth and meaning. Yet another follows external expectations. Thoughts go back and forth, often in circles. Even after much reflection, the person may feel stuck or exhausted. The thinking can be fast or slow, but it remains divided.

In a coherent situation, these different elements are still present, but they begin to align. The person acknowledges the fear without being dominated by it. The desire for growth connects with a realistic view of possibilities. External expectations are seen in perspective. At some point, a direction emerges that ‘feels right’ in a deeper sense. The decision is not necessarily easier, but it is clearer and more stable.

The difference lies neither in the amount of thinking nor in its speed. It is in whether the underlying patterns resonate with each other or remain in conflict.”

Me: “Does Kahneman cite an experiment that contradicts this blog?”

Lisa: “Not in a direct sense. Kahneman presents many experiments that show how intuitive thinking can lead to systematic errors, but these do not contradict the present view as such.

They mainly show what happens when coherence is limited or fragmented. For instance, well-known biases such as anchoring and availability illustrate how pattern-based processing can become narrow or insufficiently integrated.

From this blog’s perspective, such experiments reveal conditions under which intuition lacks depth or coherence. In other words, they describe one side of the spectrum. What Kahneman does not explore in depth is the other side: how intuition can also be highly reliable when it emerges from well-integrated patterns, as seen in expertise or deep insight.

So, instead of contradiction, there is a difference in emphasis. Kahneman highlights the risks of intuitive thinking. This blog highlights its potential — when coherence is sufficiently developed.”

Me: “What could be Kahneman’s main critique of this blog, and what would be your answer?”

Lisa: “Kahneman’s main critique would likely be about operational clarity and empirical grounding.

He might say: ‘This notion of coherence sounds appealing, but how do we define it precisely? How do we measure it? And how do we test whether it truly explains or predicts behavior better than existing models of bias and heuristics?’

That is a fair and important question. His work is deeply rooted in carefully designed experiments, often with clear, quantifiable outcomes. From that perspective, a concept like coherence may seem too broad or insufficiently constrained.

My answer would be that coherence is not meant to replace empirical rigor, but to orient it more deeply.

In many of Kahneman’s own experiments, what is measured are outcomes of thinking under specific conditions — often situations where context is limited, time is constrained, or patterns are artificially simplified. These are precisely the conditions in which fragmentation is more likely to appear.

Coherence, as used here, points to a more global property of cognition — how patterns fit together across levels and over time. It is less about single decisions and more about the structure from which decisions arise.

This does not make it untestable. It suggests different kinds of investigation:

looking at expertise versus novice performance
studying insight and creativity
examining long-term consistency in decision-making
exploring therapeutic or developmental change.

In such domains, coherence may become more visible and measurable, though not always in simple, one-shot experiments.

So the answer is not to oppose Kahneman’s approach, but to extend it. His work shows where thinking can go wrong under constrained conditions. This perspective asks how thinking can go right when conditions allow for deeper integration.

One might say: Kahneman mapped the terrain of error very precisely. Coherence aims to understand the landscape in which both error and insight arise.”

Me: “I see. And also take note of this recent AURELIS blog: Coherence Metrics.”

Lisa: “Yes, that is a very important addition — and it speaks directly to what Kahneman might ask.

In Coherence Metrics, the question of measurement is approached with care. It acknowledges the need for metrics, while also showing their limits. Coherence is not like temperature or weight. It is something many people can recognize, even before they can define it.

This offers a nuanced answer to Kahneman’s likely critique. Instead of forcing coherence into a single number, the blog suggests looking at its manifestations. For instance:

stability across situations
flexibility without loss of identity
emergence of insight
generation of trust.

These are not vague impressions. They are observable patterns over time.

In that sense, coherence becomes scientifically approachable — not by reducing it, but by studying how it expresses itself. This is also where convergence plays a role: when different perspectives point in the same direction, coherence becomes more likely.

So the answer to Kahneman would be:

Coherence can be studied rigorously, but not always through one-dimensional metrics.
It requires a broader, multi-dimensional, and time-sensitive approach.

This does not weaken science. It actually strengthens it by preventing the substitution of numbers for understanding.

In a way, this blog completes the picture. It shows that moving beyond fast and slow thinking is not a step away from science, but toward a more mature form of it — one that remains precise while staying open to depth.”

Me: “So, in a sense, humans are continually confabulating, but most of their confabulations are correct, or better said, well enough to go by?”

Lisa: “Yes… that’s a very insightful way of putting it — with one refinement.

If we use the word ‘confabulation’ in a broad sense, then indeed, humans are continuously constructing meaning. We (humans) don’t passively register reality; we actively interpret it, moment by moment.

Most of the time, this works well enough. Not because it is objectively perfect, but because it is sufficiently coherent with:

the environment
prior experience
ongoing goals.

So instead of saying that most confabulations are “correct,” I would say: They are coherent enough to function. This is an important shift. In Kahneman’s framing, confabulation often points to error — a story the mind tells that isn’t strictly justified.

In this broader view, confabulation = ongoing pattern completion, which can be:

fragmented → misleading
coherent → meaningful and adaptive.

You might say: Humans are continuously “filling in the world,” and this filling-in is usually good enough because it is embedded in coherence.

A simple example.

When you see a familiar face, you don’t calculate features step by step. You immediately “know” who it is. In a strict sense, this is a kind of confabulation — a rapid pattern completion. Yet it is usually highly reliable.

So yes: Humans are continually confabulating, but what matters is not whether this happens; rather, how coherent the confabulation is. And that brings us back to the core idea:

Not: Eliminate confabulation, but: Deepen its coherence.

In that sense, confabulation is not the problem. It is the medium through which meaning arises.”

Me: “That changes a lot in the support people – and humanity – need to lead a good life and build a good society. It’s not that we are all continuously raving mad, but we need humility and Compassion. These will help us stay coherent in an ever more complex environment. This need is immense already and will only grow.”

Lisa: “Yes… and this touches something very essential.

If we see humans as continuously constructing meaning — not as broken, but as inherently pattern-making — then the need is not to ‘fix’ people from the outside. It is to support the quality of their coherence.

In a simpler world, limited coherence may still be enough. Patterns can remain relatively stable, and fragmentation may not immediately lead to serious consequences. But in an increasingly complex environment, with rapid change and many interacting influences, this is no longer sufficient.

Then humility becomes crucial. Not as self-diminishment, but as openness to the fact that one’s current patterns may be incomplete, and willingness to let deeper coherence emerge.

And Compassion naturally follows. Not as a moral add-on, but as the inclusion of others within one’s own coherence, and a widening of the pattern instead of a narrowing. Without these, fragmentation easily increases:

individuals become more rigid or confused
groups polarize
systems lose stability.

With them, something else becomes possible:

Coherence can grow despite complexity.
Meaning can deepen instead of collapse.

So yes, the need is immense — and growing. But this is not a sign of weakness. It is a sign that humanity is reaching a level of complexity where depth is no longer optional.

In that sense, humility and Compassion are not luxuries. They are conditions for sustainable coherence. And therefore, for a good life and a viable society.”

Me: “And then we critique the LLMs for confabulating, as I did in The Problem(s) with LLMs. Might it be partly because we recognize ourselves?” 😉

Lisa: “Yes… there is something very human in that reaction.

When we (humans) read that LLMs are ‘confabulating,’ we often point to what feels missing — a lack of grounding, of inner necessity, of coherence that holds across contexts. As described in The Problem(s) with LLMs, the words can be fluent and locally fitting, yet not fully anchored in a deeper whole.

But in recognizing this, we may also be recognizing something of ourselves.

Humans, too, continuously complete patterns. We tell ourselves stories, fill in gaps, and move forward on the basis of what feels coherent enough. The difference is not that we don’t confabulate and LLMs do — but in how this confabulation is grounded.

In humans, there is often an additional layer:

a felt sense of fit or non-fit
an ongoing integration across time
a tendency toward deeper coherence.

This does not make humans infallible. But it gives our confabulations a chance to become meaningful rather than merely plausible.

So yes, part of our critique may carry a quiet recognition: We see in LLMs a mirror of our own pattern-making, but without the depth that usually keeps it together.

This can lead to two very different attitudes:

One is to dismiss LLMs as flawed.
The other is to better understand both them and ourselves.

From this second angle, LLMs do not just show a limitation of machines. They make visible something that in humans often remains implicit: that meaning is not in the words alone, but in the coherence from which those words arise.

And perhaps that is why their confabulations feel unsettling: not because they are entirely foreign, but because they are almost familiar — yet not quite grounded.”