On what critical thinking actually is, how it differs from postformal thinking, and why the second skill is the one that truly protects a professional against AI.
The phrase “critical thinking” is everywhere. It turns up in conversations about AI, in the brochure of every master’s programme, at countless dinners with friends. “Thinking critically” is something everyone seems expected to do. There are even calls to make it a subject in schools (as if it were a subject).
In much of what I read, the concept is presented as a kind of cure-all against the possible economic paradigm shift AI may bring: if you fear technology will make you redundant, stop worrying, just add a bit of “critical thinking” to your life and the problem disappears.
It is worth asking, however, whether what truly protects against AI is critical thinking, or something else entirely — something that depends far more on each person’s accumulated history and experience. Psychology has a name for that “something”: postformal thought, and we will return to it later in this piece.
Thinking, and “thinking critically”
With the exception of a few philosophers and psychologists (those specialising in cognition and reasoning), few people can explain what “thinking” actually means, let alone “thinking critically” — and I won’t even get into what distinguishes it from other kinds of thought, such as “automatic or intuitive thinking”, “groupthink”, “purely logical thinking”, “dogmatic thinking”, or simply arriving at the conclusion we already wanted from the start, which psychology calls “motivated reasoning”.
I don’t mean to be dense (I get complaints for that), but it matters to offer at least some technical grounding for what thinking is. Without it, what I say in this piece won’t amount to much more than another dinner-table conversation.
Thinking is not a single, uniform act. In its broadest sense, cognitive psychology defines it as the process by which we mentally manipulate representations — memories, perceptions, concepts — to draw inferences, solve problems, or reach a conclusion. This is the most widely shared definition among philosophers and psychologists. If we examine it closely, we notice that it covers very different forms of thought: remembering where we left our keys, recognising a face on the street, working out a sum in our head, imagining how a meeting will go before it happens, turning a worry over in our mind while trying to fall asleep, applying formal logic to a maths problem, and so on. Which of those processes deserves to be called “critical”? (If any of them do.)
What serious research on critical thinking makes clear — and what everyday usage of the phrase completely overlooks — is that thinking critically is not a fixed trait of a person, but something one exercises, or fails to exercise, on any given occasion. But there is more to it: it isn’t simply a switch a person decides to flip on or off. It is a skill that must first be built, and not everyone gets around to building it.
John Dewey was among the first to take the question seriously, as early as 1910. He called it “suspended judgment”. For Dewey, suspended judgment — close to what we today call critical thinking — consists of the disposition to keep an inquiry open before reaching any conclusion, rather than letting the conclusion arrive first and reducing the inquiry to a later justification of it.
Nearly eighty years later, Peter Facione offered a more technical definition, moving in the same direction. For Facione, thinking critically means exercising “purposeful, self-regulatory judgment”. The key word there is self-regulatory. It is not enough to check that the other person’s argument holds up. One has to be willing to subject one’s own premises to the same scrutiny — something far less common than we would like to admit. This matters, because, as noted above, critical thinking tends to be framed as something we demand of others, and it ought to begin with a demand on ourselves.
Nearly a century separates Dewey and Facione, yet they agree on something worth underlining: thinking critically is not a state a person reaches once and for all, like a degree or a diploma. It is a process exercised, or not, on each specific occasion. One can have thought rigorously yesterday and lazily today, about the very same kind of problem, in the very same office.
Robert Ennis, another widely cited author in this field, added a nuance that tends to go unnoticed: it is not enough to have the technical skill to analyse someone else’s argument or build one’s own; one also needs the genuine disposition to do so, even — or especially — when the conclusion one reaches is uncomfortable. A person can be perfectly capable of spotting a fallacy in someone else’s argument while having no real interest in applying that same standard to their own.
That, probably, is the most frequent failure I see in professional practice. We demand rigour from the other side’s argument and let ourselves off the hook when it comes to our own. That is not critical thinking. At best, it’s well-dressed rhetoric.
What it is not, even when it looks like it
From the above we already have some idea of what “thinking critically” might be. But before offering a definition — and before examining whether that concept is a necessary and sufficient condition for shielding ourselves against AI, or only a necessary one — it is worth distinguishing critical thinking from other kinds of thought.
If a kind of thinking deserves to be called “critical”, there must be others that do not. It’s worth naming them:
It is not automatic thinking. Cognitive psychology has long distinguished between an automatic, intuitive system of thought (informally called “System 1”) and a slower, deliberate one (“System 2”). The popular assumption is that the second is “the rational one” and the first “the one that gets it wrong”, but the research is more cautious: both can be right, and both can fail. A lawyer who decides, within the first five minutes of reading a file, who is in the right, and spends the following three hours not checking whether that first impression was correct but drafting the brief that justifies it, has not thought better for having spent more time on it. He has simply used deliberation to dress up a prior intuition.
It is not groupthink. There is a collective variant of the same trap, and the psychologist who described it best was Irving Janis, in 1972, studying political and military decisions that went badly wrong despite having been made by competent people in the same room. He called it groupthink: a cohesive group stops searching for the best answer and, without quite noticing, starts protecting its own unanimity — and anyone who disagrees quickly learns to stay quiet rather than become the obstacle to consensus. A committee that deliberates for three hours and lands exactly on the conclusion everyone wanted from minute one has not thought better for it. It has simply used deliberation to dress up a prior intuition — almost always that of whoever is in charge of the room, whom no one wants to be the first to contradict.
It is not purely logical thinking. This is, perhaps, the most widespread confusion among those of us who come from the legal world. We assume that thinking well means reasoning like a syllogism: major premise, minor premise, conclusion. Life rarely presents problems that clean.
Some developmental psychologists have described, and named, a kind of mature reasoning that never resolves into a syllogism: the ability to hold two legitimate positions in tension without forcing a premature logical resolution, recognising that both sides have weighty reasons and that the solution does not lie in deciding who is right, but in integrating what each one contributes. This capacity is called postformal thought. And here is what matters: it is placed above formal logic on the scale of intellectual maturity, not below it. Having that capacity, moreover, is not enough on its own. One also has to want to exercise it, even when resolving the tension with a simple yes or no would be far more convenient. We will return to it in more detail shortly.
It is not reaching the conclusion that suits us. Psychologist Ziva Kunda called this, in a now-classic 1990 paper, “motivated reasoning”. It is probably the kind of thinking that most often masquerades as critical thinking, because from the inside it doesn’t feel like a bias. The person sincerely believes they are reasoning rigorously, when in fact they are building, with considerable technical skill, an argument in service of a conclusion already decided in advance out of interest, fear, or convenience.
And here something uncomfortable shows up: intellectual skill does not protect against this bias. It arms it better. In a now-classic study, Yale researchers gave more than a thousand adults an identical statistical problem, presented sometimes with neutral data and sometimes with politically charged data. The greater a participant’s numeracy, the better they solved the neutral version, as one would expect. But on the politically charged version, that same numeracy only helped when the correct answer matched what the participant already wanted to believe. Greater analytical ability, in that particular case, did not mean more truth. It meant better rationalisation.
Psychologist Keith Stanovich has a precise name for the underlying phenomenon: dysrationalia, the capacity to be highly intelligent and systematically irrational at the same time.
It is not dogmatic thinking. Starting from an authority — a rule, a tradition, a precedent cited without further question — and not even admitting the possibility that the correct conclusion might differ from what that authority dictated, is exactly the opposite of critical thinking. However well argued it may be.
So, what is it?
The definition, gathering everything said so far: thinking critically means exercising, on each specific occasion, purposeful, self-regulatory judgment that suspends the conclusion while the inquiry lasts, that subjects one’s own premises to the same scrutiny as the opponent’s, and that holds even when the outcome is uncomfortable for the person doing the thinking. It is not a trait one has. It is something one does, or does not do, every time.
Can critical thinking be learned?
Yes and no. And the honest answer matters more than the comfortable one.
Yes, in the sense that there are concrete, perfectly teachable skills.
- One can learn to recognise a fallacy.
- One can learn to apply explicit evaluation criteria: whether a claim is clear, whether it is precise, whether it considers viewpoints other than its own, whether it treats the opposing position with the same seriousness it gives its own.
- One can train the eye to spot when a technically correct argument actually serves a conclusion decided in advance out of interest or fear. All of this can be taught, practised, and improved through repetition, like any other professional skill.
- One can learn to detect what an argument takes for granted without saying so: the hidden premise that, if false, would bring down the whole conclusion.
- One can learn to build, before rebutting it, the strongest possible version of the opposing argument, rather than the easiest one to knock down — what is known as steelmanning.
- One can learn to assess whether a source or a piece of data deserves the trust placed in it: who says it, how they know it, and what interest they have in the answer being one thing rather than another.
- One can learn to distinguish correlation from causation, and to spot when a statistic implies more than it actually proves.
But no, in the sense that a generic two-hour “critical thinking” course does not produce the mature judgment that actually matters, however well designed the slide deck.
Cognitive psychologist Daniel Willingham demonstrated this in a way that is rather uncomfortable for the entire training industry built around this subject. Critical thinking depends decisively on domain-specific knowledge. In other words, before one can think critically about a topic, one needs to have studied it — and a great deal, both in depth and over time.
Willingham also observed that generic skills, disconnected from a specific field of knowledge, barely transfer from one context to another. In other words, knowing how to spot a fallacy in a televised debate does not automatically prepare you to spot one in a contract clause.
Beyond knowledge and skill (which, as noted, appear to be domain-dependent), there is something even more decisive: a set of traits and dispositions that, if not activated, prevent critical thinking from emerging at all. Openness, tolerance for ambiguity, the genuine willingness to discover one was wrong. Unfortunately, these components are not installed by a one-afternoon motivational talk. They are cultivated slowly, almost always through years of real exposure to ambiguous problems where being wrong actually costs something. Mature critical thinking is not something taught in a classroom. It is something earned over time, often the hard way.
Postformal thought
To understand what postformal thought is, it helps first to locate where formal thought ends.
As early as the 1940s, Jean Piaget described four stages of human cognitive development, the last of which he called “formal operations”. Around the age of twelve, human beings begin to be able to let go of the concrete — what can be seen and touched — in order to reason with symbols, concepts, and abstract ideas: the mind no longer needs to work with the concrete details of “house” (walls, roof, framing); the verbal symbol “house” is enough. This capacity is refined in adulthood, and for decades was considered the ceiling of intellectual development. It is no coincidence that it is also the terrain where artificial intelligence moves comfortably: applying rules, manipulating symbols, and solving well-defined problems is, in essence, formal thought — and a machine already does that reasonably well.
But in the 1970s, psychologist William Perry noticed that some people reached a level of abstraction Piaget had not described. Those people were capable of solving ambiguous, contradictory, context-dependent problems — problems in which, as is often said, two plus two doesn’t add up to four. Perry called that capacity postformal thought: the ability to hold two legitimate positions in tension, without forcing a premature logical resolution, recognising that both sides have weighty reasons and that the solution does not lie in deciding who is right, but in integrating what each side contributes.
And here something paradoxical emerges. Reaching that higher stage — which, as Perry himself warns, not everyone reaches — requires, first, years of formal training and, on top of that, years of what we might call “intellectual tension”: real experience applying formal thought in ambiguous contexts, where being wrong carries genuine consequences. It is a kind of wisdom acquired through direct contact with real problems, one that demands loosening what formal knowledge taught rigidly.
To sum up: postformal thought is the ability to manage the history, the timing, the ambiguity, the motivations, the goals, and the values specific to a concrete problem — something that is not obtained through more knowledge or more formal logic alone. Like mature critical thinking, it is not taught in a classroom. It is earned, rather, through years of real experience with problems where being wrong actually costs something.
Necessary, but not sufficient
Earlier we left a question hanging: is critical thinking a necessary and sufficient condition to protect us against AI, or only a necessary one? Applied to any professional service — delivered by an outside adviser or by an employee within an organisation — the answer is that it is necessary, but not sufficient. And the missing piece is precisely the postformal thought we have just defined.
Critical thinking answers a very specific question: is this conclusion sound? Checking that an argument has no fallacies, that its premises are well founded, that the conclusion actually follows from the evidence, is indispensable — without that work, there is nothing worth offering anyone.
But a client does not hire an adviser merely to confirm that their reasoning is valid. They hire someone who also knows whether it is wise to apply that conclusion, as it stands, to their specific situation, with its own history, timing, and particular goals. That second question — not whether the conclusion is correct, but whether it is advisable to act on it, here and now — belongs to the domain of postformal thought. And it is, ultimately, the skill a professional service actually sells.
This is clear in the same examples used earlier, now transposed inside a firm or a company. The adviser negotiating a contract who, to close the engagement sooner, starts telling the client what they want to hear instead of what they need to know, has not failed for lack of technical skill. They failed, much earlier, at the self-regulation critical thinking demands, replacing it with what suits them. And the internal committee that decides to keep funding a project that has produced nothing for two years, because the person who championed it is still the director and no one wants to be the first to tell them they were wrong, has exercised neither. It’s the same groupthink trap seen earlier, wearing a boardroom suit.
But even when critical thinking works perfectly, it can still fall short. Consider the succession of a family business, where a father who founded it, a daughter who has professionalised it for years, and a brother who is a co-owner by inheritance each have legitimate reasons that no syllogism will reconcile on its own. A lawyer who is only trying to win the formal argument for their client can construct a flawless conclusion, without a single fallacy, and still resolve nothing in reality. A “victory” that humiliates whoever remains inside the company tends to become, in the medium term, a defeat for everyone — including the one who thought they had won. There, critical thinking has already done its job. What’s missing is the postformal question: whether it is wise to apply that correct conclusion, without nuance, to a situation where three distinct legitimacies are still sitting at the same table. The work of a mediator, or of an adviser who understands negotiation through psychology and not only through technique, consists precisely in holding those legitimacies in tension, without forcing a premature resolution, until a way out is found that integrates them rather than letting one crush the other.
There is a nuance worth keeping in mind here: postformal thought does not replace critical thinking. It rests on it. One cannot sensibly judge whether it is advisable to apply a conclusion without first having rigorously checked that the conclusion is sound. A postformal judgment built on weak analysis is not wisdom. It is well-intentioned improvisation, dressed up as experience. Order matters: critical thinking first, to know whether something is true. Postformal thinking after, to know whether it is advisable to act on it.
Research on how we reason points in the same direction. Human reasoning performs worse alone. It tends toward bias, toward polarisation, toward defending one’s own position out of sheer inertia. And it improves markedly when there is, in the room, someone able to make the parties truly listen to one another, sometimes for the first time in a long while. It is no coincidence that the systems that work best for deciding well share the same structural trait. A court needs opposing parties. A good board needs independent directors. A good internal team needs someone able to disagree with the boss without it costing them their job. A good mediation needs, precisely, a neutral third party. None of these systems depends on a single person, however critical and wise, thinking for everyone else.
Closing thoughts
The next time someone says “what’s missing here is critical thinking” — whether in a law firm, a boardroom, or a negotiation between partners — perhaps the question worth asking is not “how do we get more of it, now?”. There is another question that is more uncomfortable, and also more useful: which of all these things are we actually talking about? Asking for more logical rigour is not the same as asking for more willingness to be wrong. And arguing harder is certainly not the same as thinking more carefully.
We opened by asking whether what truly protects against AI is critical thinking, or postformal thought. We now have the full answer: the technical part of critical thinking, a machine already reproduces reasonably well. Indeed, it is often said that current AI solves ninety percent of the problems put in front of it reasonably well. What remains unresolved — that uncomfortable ten percent — is almost always a postformal matter: history, timing, ambiguity, motivations, goals, and values. That is not easily automated, and it isn’t cured by a cure-all or a two-hour course either. It is built, slowly, through years of real experience with ambiguous problems where being wrong actually costs something. Artificial intelligence, for now, has no such biography.
And here it is worth not looking away from the most uncomfortable question of all. That postformal thought does not install itself: it is cultivated, as noted, through years of real exposure to ambiguous problems, almost always by first doing the routine work that excites no one but that teaches. Reviewing a hundred similar contracts, drafting a hundred similar briefs, watching a hundred decisions play out well or badly, up close. That routine work is precisely the ninety percent that AI is already starting to handle reasonably well. If that breeding ground shrinks, one has to ask where the next generation of professionals capable of holding the real complexity of a problem will come from. This is not a rhetorical question, and this piece does not claim to answer it fully. Perhaps the answer lies in deliberately redesigning how newcomers to a profession are trained, giving them earlier contact with ambiguity that, in the past, arrived almost by accident, after thousands of hours of routine work. But ignoring the question would be, once again, the most comfortable version of critical thinking — the kind applied to someone else’s argument, and never to one’s own profession.
In the professional matters that truly matter — a company changing hands, a partner who feels betrayed, a client whose trust has been earned or lost — it is almost never the best arguer who wins. When someone wins, it is whoever is able to hold the real complexity of the problem for long enough to find a way out that serves everyone.
Would you like to talk about this?
At Estudio Jurídico Alta Mediación, we combine law and applied psychology precisely to hold this kind of complexity: at the negotiating table, in a boardroom, or in the succession of a family business. If you would like to talk about how we think through difficult decisions, let’s talk.