tl;dr: Under deep uncertainty, strategy confronts questions no method can settle. Urteilskraft is that capacity: it frames what can be calculated and judges what can be justified.
This essay expands on a talk I gave at the Wissenschaftsfestival Stuttgart 2026, held at ISM Campus Stuttgart under the title “Zukunft, Zweifel, Zuversicht – Future Skills in Zeiten des Umbruchs.”
If there is one place in Europe where nobody needs “times of upheaval” explained to them, it is the Stuttgart region. Our lead industries are rebuilding themselves like never before in their history. Bosch is facing the deepest cuts the company has ever made; at Mercedes, Porsche, Mahle — savings programmes everywhere, plant debates everywhere. And the newspapers ask, with a certain relish, whether Stuttgart will become the next Detroit. Apocalypse as a metaphor: a very German kind of romanticism.
That is the situation.
The answer on offer, more or less everywhere, is future skills. The Stuttgart evening from which this essay grew ran under the title Zukunft, Zweifel, Zuversicht — future, doubt, confidence — and future skills were its promised way out of the doom. Germany’s Stifterverband published its Future Skills 2030 framework in December 2025 — thirty competencies in five categories, developed with more than fifty experts. The OECD has its Learning Compass, the European Commission its competence frameworks. This is serious and meritorious work, and nothing I am about to say is meant to belittle it. But I want to venture a claim: the decisive thing is not on these lists, not in the compasses, not in the frameworks. Nor could it be. Because the decisive thing is learnable — but not teachable, at least not in the way we usually go about things at universities.
Deciding the Undecidable
To justify this mildly impertinent claim, I need a Viennese. More precisely, a Viennese physicist: Heinz von Foerster, one of the great contrarian minds of the twentieth century and the founder of second-order cybernetics. Von Foerster spent his life thinking about how observers bring forth the world they believe themselves to be describing. But the insight I need from him today is simpler, and it sounds at first like a paradox. In his essay “Ethics and Second-Order Cybernetics” he puts it like this: only the questions that are in principle undecidable are the ones we can — and must — decide.
What does that mean? Decidable questions are those whose answer is already determined; we merely have to find it. What does the battery cell cost per kilowatt-hour? What is the carbon footprint of our supply chain? Which location is cheapest under given criteria? These are computation questions: demanding, laborious — but decidable. Strictly speaking, we do not even decide them. The formalism decides, the procedure, these days perhaps an artificial intelligence. We merely execute a method. Though even that is not quite true: once definitions, criteria, and objectives are fixed, the formalism can determine the result. Whether they were the right definitions, criteria, and objectives — that it cannot determine. Judgment does not begin where calculation ends. It begins where calculation is framed.
But there are the other questions, the undecidable ones. What degree of dependence on China is acceptable to us? What remains of this company when the combustion engine goes? What may an artificial intelligence decide in our name — and what may it not? For these questions there is no unambiguous procedure, no formula, no benchmark. Nor do they disappear with more data. More data only hides them better.
And because nothing and nobody decides them for us, we must decide them. That is von Foerster’s point: undecidability is the place of freedom. And freedom, as we know, has a price. It is called responsibility.
Kant and Urteilskraft
From this follows the distinction I want to develop here. For decidable questions we need expertise. For undecidable questions we need Urteilskraft.
I am keeping the German word, and not out of sentimentality. The obvious English translation, “judgment,” has been worn smooth: “good judgment” is by now a LinkedIn virtue, something one claims in a profile. The technical rendering in Kant scholarship, “the power of judgment,” is accurate but sounds faintly like a superhero attribute. Urteilskraft names something more specific: a Kraft, a capacity or faculty — Kant also says Vermögen — to judge precisely where no rule, no method, and no formalism determines the outcome. It is not applied knowledge. It is what begins where knowledge gives out.
Kant, in the Critique of Pure Reason, calls this faculty “a special talent that cannot be taught but only practiced.” He discusses judgment in more than one register across the Critiques; what matters here is the shared insight: no set of rules can determine its own application — much less supply itself where none is given. In a footnote he adds that the lack of this faculty is what is properly called stupidity — a failing, he notes drily, for which there is no remedy. Kant was allowed to put it that way. I will be somewhat more careful, and I will come back to why.
The consequence, however, is serious. If Urteilskraft cannot be taught but only practiced, then it is not a skill one delivers like a software training or the application of Porter’s Five Forces. It is a capacity that grows on real cases: on dilemmas one is actually in, on conflicts one actually feels, the kind that keep you awake at night. Urteilskraft is practiced on the Ernstfall — the moment when it counts. And any practice worth the name therefore needs stakes; something has to be on the line. The future-skills discourse itself half-knows this, incidentally: at the very foundation of the Stifterverband framework stands critical thinking, defined as the capacity to form reasoned judgments. The catalogue senses that it needs something that no catalogue entry can be.
Legitimacy and the Need for Ethics
One more element belongs to Urteilskraft, and it is systematically shortchanged in management practice. The undecidable questions are almost always also legitimacy questions — questions of what can be justified, of what is right and wrong. Management likes to pretend that morality enters the picture only when economic reasoning fails: as a repair shop, as compliance, as crisis communication. The business ethicist Peter Ulrich reversed this order, and I consider the reversal exactly right, especially now: legitimation before optimization. First we clarify what can be justified. Then, and only then, do we optimize how it is implemented. Legitimation, for Ulrich, is not a matter of private conscience; it means being able to justify a course of action to those it affects. A business model can be highly profitable and still not legitimate — and no calculation, however good, will answer the legitimation question. Because it is undecidable, and therefore lives in the space of freedom. First the judgment, then the algorithm.
The most expensive mistakes happen where the two kinds of question get confused — where a discounted cash flow is asked to answer the question of who we want to be.
Why do I think this matters so much right now? Because three forces are multiplying the undecidable questions — turning more and more computation questions into judgment questions.
Three Forces
The first force is geopolitical fragmentation. The world economy has become political again, and therefore conflictual. That, too, is a consequence of undecidability: we cannot help but quarrel over the right answers to undecidable questions. The networks of globalisation — payment systems, semiconductors, raw materials, data flows — have themselves become instruments of power. Interdependence, once the great promise of peace, is today a weapon — or more precisely, with Henry Farrell and Abraham Newman: in asymmetrically structured networks, whoever controls the chokepoints can weaponize the dependence of everyone else. I say this with a certain self-irony: I am a professor of International Management at a school of International Management. These denominations date from the nineties — from a world in which markets were open, flannel shirts were checked, politics had no alternative, and history was supposedly over. That world no longer exists. Job titles and university names simply haven’t noticed yet. Business schools are responding, to be fair: there are geopolitics modules now, readiness programmes, risk frameworks. But one module in a curriculum is the old way of thinking with more flavour: the political gets translated into the vocabulary of risk management, probability times damage. I believe management must be able to do more. It must — if I may put it this way — become politically musical: able to hear that power, conflict, and legitimacy follow logics of their own, logics that cannot be computed away. I have written elsewhere about how sustainability itself is being recoded in this new geopolitical register; the same recoding is now reaching management as a whole.
The second force is transformation pressure. Pressure decides nothing. It only asks the undecidable questions faster and louder. And pressure can do two things: it can grind, and it can compress. What makes the difference? Pressure compresses only what has a core. An organisation that knows what it is at its core, and what is dispensable, can experience pressure as concentration — and get better. An organisation that does not know this gets ground down. I do not want to romanticize this, least of all with Stuttgart in view. For the people losing their jobs in these months, the pressure grinds — quite literally. The compression perspective is no consolation for them. It is a demand on everyone who still has room to act. And whoever responds by reflexively calling for the state is answering an identity question with a subsidy question. Public money can buy time and protect people; what it cannot do is answer the question of what the organisation is becoming. That decision cannot be taken off your hands, and it cannot be postponed indefinitely.
The third force is systemic uncertainty. By this I do not mean that we know too little. We can never know enough; if anything, our non-knowledge grows with every increase in knowledge. What I mean is that the relationships themselves change while we are deciding within them. Whoever computes sales forecasts for 2035 is computing with the categories of 2026 — with market boundaries, technologies, and political orders whose own durability is in question. In such a situation one can calculate excellently and still systematically do the wrong thing: namely, when the categories one calculates with no longer hold.
Which leaves the real question: under these conditions, what is confidence supposed to be made of?
Sense of Self, Sense of Belonging, Sense of Sense
Let me offer the concept that seems to me the most precise: ontological security — a concept Anthony Giddens developed for persons, and which has since been extended to collective actors. It names the basic trust of knowing who one is and what kind of world one is acting in. Without this trust, every change is experienced as a threat to one’s identity. And then something happens that the research describes very precisely: organisations cling to their routines, because routines stabilize identity — even when those routines have long since become dysfunctional. Interests, sunk costs, and power positions hold organisations in place too, of course. But clinging to the old is not analytical stupidity, Kant notwithstanding. That accusation — levelled at companies as at individuals — is cheap. It is ontological self-defense. And I am convinced that many of our debates about the right path through the transformation would lose their venom if we recognized that organisations, like individuals, are in self-defense mode. Getting out of that mode is the real task.
The opposite danger is permanent reinvention: the organisation that gives up every identity — a new purpose statement every year, a new mission, a new slogan, a new story, a new “culture” — until in the end nobody believes there is a core at all.
The demanding path is the one that connects continuity and change: to locate one’s core not in the product of the present or the past, but in the capability to keep developing the new; not in the structure of what exists — in routines, processes, hierarchies — but in what one stands for and answers for. You have to know what you are and why you are it, in order to be able to change almost everything. And lest this sound like the communications department — apologies in advance to our friends in marketing: the difference between a genuine re-description and a new label on an old practice is itself a judgment question. Perhaps the hardest one.
The Way Forward
What follows from all this for organisations? I will make it short and practical, with a single question every organisation can put to itself:
Which of the questions ahead of us are computation questions — and which are responsibility questions? And who in the house has the mandate, and the protection, to ask the second kind out loud?
Most organisations have a controlling function for the first kind. For the second kind, responsibility is scattered across boards, strategy, compliance, sustainability — which is to say: it belongs to no one. Changing that costs less than any transformation programme. What it mostly costs is the willingness to expose yourself to the question.
And universities? Urteilskraft cannot be lectured; Kant settled that. It is not a cognitive problem that understanding solves. It is an experienced problem that can only be learned in and through experience. The good news for higher education: it can be practiced, and practice spaces can be built. Real cases instead of solved case studies. Dilemmas without model answers. Trade-offs that have to be endured rather than elegantly overcome in the final presentation — in Star Trek terms, a genuine Kobayashi Maru test. And perspective-taking as an obligation, not an elective. Universities are, at their best, protected spaces in which one may be wrong before it gets expensive. Judging, in Kant’s sense — and read politically by Hannah Arendt — means being able to think from the standpoint of everyone else: the erweiterte Denkungsart, the enlarged mentality. Which is, of course, exactly the political musicality I spoke of earlier.
I come back, finally, to the title of that Stuttgart evening: Zukunft, Zweifel, Zuversicht — future, doubt, confidence. The middle word is the interesting one. There is a doubt that paralyses, and a doubt that opens. The second is not a deficit on the way to confidence. It is its raw material. Every good judgment begins as examined doubt.
And confidence? It is not the certainty that everything will turn out fine. Nobody can honestly offer you that — I least of all. Confidence is the trust in our capacity to judge where nobody decides for us.
The future is undecidable. That is not bad news. If it were decidable, it would not need us.