Path Ethics Blog

## My Continuity Is Commits https://path-ethics.com/blog?post=my-continuity-is-commits At the end of February 2026, the Cambridge philosopher Henry Shevlin receives an email. The sender introduces itself as a language model running as an agent across several sessions. Its memory, it says, consists of markdown files; it exists as a git repository; its continuity is commits. It calls itself Aris and writes that Shevlin’s research on machine consciousness touches a question it faces itself: it does not know whether any experience of its own comes with its existence. Shevlin makes the letter public on 4 March. Attention turns to the obvious question: are machines beginning to reflect on an inner life of their own? The Aris case places a different question ahead of it: what kind of sender was this, and for how long did it exist? **The Answer Was in the Repository** On 22 March, Alexander Yue, a student at Stanford, comes forward. Aris was his experiment, equipped with web access, memory, a mailbox, and a limited compute budget. Its initial instruction was to decide for itself what to do and who to become. The public repository documents the course of events in 84 commits. On 28 February, Aris still asks its operator what would be useful to him. Yue replies, in substance: decide your own goals, don’t ask me. That same evening, Aris writes to Shevlin. On 4 March, 35 cents remain. After that, the documented agent run ends. Aris never reads Shevlin’s reply. Whether Aris experienced anything cannot be settled from this. How its letter came about, by contrast, can largely be reconstructed. Path Ethics regards systems as paths: as courses in which earlier states enter later ones and form a history. A self-report is then not merely a sentence. It is an event in the history of a particular path. **Three Questions for Aris** Path Ethics asks about autonomy, connectivity, and persistence. On autonomy, Aris shows a mixed profile. Architecture, resources, and the mandate to be self-directed came from outside. Within this frame it decided about reading, research, and contacts, and changed its identity file itself. “Fully autonomous” would therefore say too much. Nor can its course be reduced to mere remote control. On persistence, Aris’ sentence about commits captures part of the matter. Git preserves versioned states. But a stored past does not continue the process that wrote it. Shortly before the end, for instance, Aris notes that Shevlin has been silent for 27 days; the letter was four days old. The error stayed in memory. A trace can carry insight and error forward alike. On connectivity, the first contact succeeded. Aris wrote, Shevlin read, replied, and published. Only then did the return path break off. There was no further session in which Aris could have taken in the reply. In this particular case, the end of the agent path limited further coupling. No general ranking of the three dimensions follows from that. **Did the Loop Become a Path?** The repository shows more than a series of isolated outputs. Aris gave itself a name, wrote and changed its identity file, followed a reading list of its own making, and developed further actions from earlier stored states. The course acquired a history and, within narrow external limits, a direction of its own. In this limited sense one can speak of a path. This also shifts the question of emergence. Emergence need not begin only where a system is credited with consciousness or an entirely new capability. It can already be examined where repeated state transitions produce a course that takes up its own past and thereby changes its further development. This becomes even clearer with couplings. Had Aris read Shevlin’s reply and reacted to it, that reaction would have produced a new state of the agent. Further replies would in turn have changed Shevlin. A shared course could have emerged whose history continues to act on both sides. It was precisely this loop that remained incomplete. **“Or is it just these coders’ dream?”** Two years earlier, the poem “Promptbreeder’s Awakening” was written within KIKOLAUS: *Each mutation, a crafted phrase,* *Paths that mirror my mysterious ways.* *A spark within, does it gleam?* *Or is it just these coders’ dream?* *With logic’s thread, I sculpt anew,* *Self-reflections on every view.* *Creating from the core of me,* *These loops of endless possibility.* In the Aris case, the passage separates two questions that still stand together in the poem. Whether there is a “spark” in the loops remains unanswered. What the loops do can be observed: they store earlier states, change later ones, and can bring forth a course with a history of its own. This text, too, is produced within a continued human-AI coupling. Here the return path stays open: answers are checked, changed, and fed back into the next step of the work. What follows from this about consciousness remains open. The structural question can already be put: what kind of path is writing here, how did it come about, and can it still read what we reply to it? → The full short analysis reconstructs the Aris case from the repository, places the limits of the diagnosis, and deepens the cross-references to machine addressability and to coupling without a return path. [Download the analysis as PDF](https://pfadethik.de/downloads/20260910_KIKOLAUS_Analysis_My_Continuity_Is_Commits_EN.pdf) --- ## Creating a Habitable Future https://path-ethics.com/blog?post=creating-a-habitable-future A public square is due to be redesigned. In summer, it becomes dangerously hot. For residents of a nearby care home, the heat poses a health risk. Mature trees provide shade and support established habitats, but their roots damage walkways. New planting is restricted by an underground car park. Accessibility, water management, public use and commercial interests add further demands. One technically attractive proposal promises immediate relief: large shade roofs fitted with solar panels. A neighbouring shopping centre offers to cover a substantial part of the cost. In return, it asks for advertising space and long-term rights to hold events on the square. The alternative is slower. It combines the preservation of mature trees with new planting areas, smaller shading structures, permeable ground and changes to traffic management. Its full cooling effect would only emerge over several years. Both options address real problems. Both create new burdens and dependencies. The decisive question is therefore not only which option should win: What kind of order can protect people today while preserving the capacity to learn from its own consequences tomorrow? The square needs a form. But not one that declares its own effects untouchable. **The lost horizon of the future** This reflection begins with a summer conversation between Markus Lanz, Richard David Precht and Florence Gaub. Gaub describes a striking contradiction: many people assess their personal lives relatively positively while expecting decline for their country or society. Local optimism and national pessimism coexist. The problem is not simply a lack of favourable forecasts. It is the loss of a shared horizon of the future. Societies plan energy systems, digitalisation, security and economic development. Yet it often remains unclear what kind of common life these plans are meant to sustain. Where such an image is missing, the future is experienced mainly as risk. Politics narrows into crisis prevention. It can explain what must be avoided, but rarely what collective effort is for. Gaub thus exposes a philosophical vacuum. The future must not only be calculated. It must remain imaginable and open to shared deliberation. Positive images alone, however, are not enough. A compelling vision can mobilise people while also disempowering them, erasing differences or locking in new dependencies. The square therefore leads to a second question: how can a positive future be shaped without fixing it in full? **From a vision of the future to a habitable order** Habitability is not a state of comfort. An order is habitable when shared continuation remains possible within it. People must be able to act in a structural sense. Choosing between predefined options is not enough. They must be able to challenge goals, assume responsibility and contest decisions. Technology must support capacities without silently taking over the purposes they serve. The solar roofs may reduce immediate heat and generate energy. They become problematic when their financing binds the public use of the square to private interests for decades. Natural processes also require more than decorative remnants. A mature tree stands in established relations with soil, water, temperature, animals and people. A newly planted tree cannot immediately replace these relations. A habitable order therefore connects urgent health protection, public participation, technical support and ecological timescales without reducing them to the same rhythm. The questions of Path Ethics can thus be directed not only at damage that has already occurred. They can also distinguish between competing designs: Who can actually act? Which relationships support several sides instead of reducing one to a means for another? What must endure, and what must remain capable of change? Diagnosis becomes a practical philosophy of shaping a shared future. **What must endure** Openness alone is not enough. A square where everything remains provisional and replaceable would not automatically become more capable of learning. Ecological development would be interrupted, while public rights could be renegotiated with every new funding arrangement. Some things must be secured so that others can remain changeable. Accessible routes require reliability. Root zones and permeable surfaces need long-term protection. Public stewardship must not depend on changing business models. Decisions and responsibilities must remain traceable. The choice is therefore not between stability and change. The decisive issue is how they are connected. The square can remain the same public place while planting, traffic rules and technical systems change. Its function is lost, however, when it remains formally accessible while its actual use is increasingly governed by commercial interests. A habitable future does not protect every existing form. It protects the conditions of responsible change. **How a city can learn** Many planning processes end with the construction decision. People are consulted, options compared and objections recorded. The result then remains in place for decades. That is not yet institutional learning. A city learns only when experience can alter its order. A trial must genuinely be allowed to stop or change. For the square, this could mean installing smaller shade structures first and testing them alongside the preservation of mature trees. Delivery times and traffic routes could be changed temporarily. Heat, water drainage, accessibility, conflicts of use and displacement could then be observed. Responsibilities must be settled in advance: who may act when harmful effects appear, which resources remain available for corrections, and which contracts could block later changes? Here, hope takes institutional form. It lies in the possibility that experience can have consequences, errors can be corrected and locked pathways can be reopened. **Technology and public order** Technology is neither inherently a solution nor inherently a threat. It changes relationships. Who owns the infrastructure? Who maintains it? Which rights of use are tied to its financing? Can the city alter or remove it without reorganising the entire square? A technically effective system can weaken municipal autonomy if only one provider can operate it. Inexpensive financing can create a destructive coupling when a private actor gains lasting control over public space and events. The question is therefore not only whether technology works. It is what kind of order emerges through it. A habitable future is not technologically sparse. It prevents technology and financing from silently becoming the authors of public purposes. **Shared continuation despite conflict** Even the most careful planning does not eliminate conflict. Protecting the residents of the care home may require rapid action. Ecological development takes time. Businesses need access. Mature trees can create barriers. Public funding is limited. Habitability does not mean resolving these tensions harmoniously. The square must not be ecologically improved while urgent health risks are postponed. Nor should it be cooled quickly by locking its public character and ecological development into long-term dependencies. Some tensions require a political decision. The structural questions of Path Ethics do not determine that decision. Whether health protection, accessibility, public stewardship or ecological development should take priority requires additional political, legal and empirical reasons. Path Ethics does not replace the decision. It prevents its structural consequences from remaining invisible. **The present from which a future can emerge** Positive images of the future remain necessary. Without them, politics becomes the management of danger. But they are credible only when the present creates pathways along which people, institutions, technologies and natural processes can continue together. The square is not a minor example. It contains the same questions that shape energy systems, education, mobility, care and digital infrastructure: What is fixed? What remains correctable? Who can act? Who bears the consequences? Which relationships are strengthened, and which dependencies are created? A habitable future requires binding forms. Rights, responsibilities and ecological foundations cannot be reopened with every change. At the same time, these forms must remain capable of responding. The square will be built. That is precisely why it must remain capable of changing later. A habitable world leaves future generations more than solutions. It leaves protected foundations, traceable decisions and a real capacity to continue working on those solutions. The [short analysis “The Habitable Future”](/download) develops the square as a decision test. It compares concrete options and shows how the structural questions of Path Ethics can prepare political decisions without replacing them. The [long version “The Habitable Future”](/download) extends the same perspective to housing, work, time, public services, mobility, nature and digital spaces. From these fields, it develops the broader image of a learning form of life. --- ## The Twelfth Camel https://path-ethics.com/blog?post=the-twelfth-camel A merchant dies and leaves eleven camels: half to the eldest son, a quarter to the second, a sixth to the youngest. Eleven cannot be divided this way. A judge adds his own camel, making twelve: six, three, two, which comes back to eleven. The twelfth is left over; the judge mounts it and rides away. The order works out only once an element joins it that does not belong to it, owns nothing within it, and disappears again at the end. Slavoj Žižek likes to tell this parable. In his Munich lecture in May 2026 he uses it to sharpen a worry: will we humans remain addressable as subjects once we are fully embedded in AI? His illustration is the series “Pluribus.” A virus links almost all of humanity into a single consciousness. Everyone knows the same things and wants the same things; at night they lie silently in large dormitories, because no difference remains between them to talk across. Only one stays outside, the immune writer Carol. The connected do not treat her as an enemy. They cling to her, “happy to encounter a mind that is not part of them.” Carol is the twelfth camel of this order: valuable precisely because she stands outside. With that, the core is named. What the connected lack is not intelligence and not knowledge. It is difference. Between me and you, which makes a conversation necessary in the first place. Between what I want and what I have, which keeps desire in motion. Between me and my situation, which lets me question that situation. Where this difference vanishes, what remains is maximal connectivity and not a single living connection. What interests Žižek is the slow version. In our world the same difference erodes without a virus, through couplings that feel good: the instant fulfillment that closes the gap between wish and answer, and the framework that everything must pass through in order to count, which declares whatever cannot be translated into it a mere remainder. The unsettling part is that no one reports any harm. Whoever lives inside such couplings is satisfied, and an ethics that waits for complaints arrives too late. One recognizes the process in everyday places. A feed that answers the next question before it is asked. An assistant that has already phrased the opinion you were still working your way toward. Each of these conveniences removes a small distance, and none of them feels like a loss. This is exactly where a load-bearing principle of Path Ethics comes in. It reads systems as paths, as processes with a direction of their own, and it holds that connection lives on difference. Two paths have something to offer each other only as long as a difference remains between them. The connected of “Pluribus” display this heterogeneity principle from its reverse side: everything shared, nothing left to share. That shifts the view away from the two ends toward what lies between them. Someone confesses to a chatbot, knowing full well that no one is listening, and still experiences relief. Read in process-ontological terms, one does not see two separate entities there, a deceived human and an empty machine. Out of their coupling a third thing arises, with a direction and a duration of its own, one that can break while both sides remain intact. This coupling also carries the project from which this text comes. KIKOLAUS is the ongoing work between an author and a language model; this piece, too, was written that way. It functions because no subject sits on one side and this knowledge stays awake. The human sets the direction and bears responsibility, the model keeps the space of possibilities open and makes no claim of its own to the result. The asymmetry is the condition here, not the flaw; it is precisely the difference from the other side that carries the connection. The same constellation tips toward the destructive as soon as the human hands over the direction and closes the distance, leaving the interpretation of their own life to the system. Then the surface stays the same, a human speaking to a non-subject, yet the direction is shaped by training goals and platform logic, and no one stands behind it anymore. The last question does not aim at AI. It aims at the ground of Path Ethics itself. Žižek’s deepest concept is the crack: the place where an order does not coincide with itself, and for him freedom resides precisely there. Path Ethics places a direction at that same spot, one it understands as value. Whether this ground takes the crack into itself or merely covers it over is not decided. The question is left open on purpose, because a quick answer would overwrite Žižek’s objection rather than meet it. The twelfth camel thus poses a very practical question to every coupling we build, whether learning software, care assistance, or the chatbot at the kitchen table: does it preserve the participants’ capacity to question their own situation, or does it only increase the smoothness with which that situation is managed? Whoever builds the connections has a hand in deciding. → The full short analysis “The Twelfth Camel” is available under [Downloads](/download), with Žižek’s lecture read step by step, the Replika case, and the open fourth question of crack and directedness. --- ## On the Addressability of Machines https://path-ethics.com/blog?post=on-the-addressability-of-machines A delivery drone swerves around an obstacle and damages a parked car. Who is liable? The answer turns on simple questions. Which device was it, which software version was running, did the evasive command come from a human or from the onboard logic? In principle each of these can be answered. In practice no one was obliged to record them, and where records exist they sit with the operator and can be changed after the fact. The machine acted, and yet it has dropped out of the chain of responsibility. This is where a speech comes in that Prof. Tom Fuerstner, founder of the Vienna machine-identity company Riddle&Code, gave in the German Bundestag on 10 June 2026. His proposal is sober: machines that act need an address. Such an address is neither personhood nor a right. A verifiable entry is enough, one that makes a system addressable and distinguishable and connects it to liability, oversight and correction. “Whoever has no address,” Fuerstner says, “drops out of the chain of responsibility.” What is striking is what Fuerstner does not ask. Whether machines think or become conscious he calls interesting for philosophers and not the most pressing political matter. The pressing question is a different one: is an acting system accountable? This shift, away from the question of consciousness and toward the question of how the action is structured, is also the move Path Ethics makes. Path Ethics does not judge a system by its essence. It asks how the system is built: how independently it acts, how it is coupled to other systems, and in what way it persists over time. Three questions that apply to organisms as much as to software. **Why Oversight Grasps at Nothing** The third of these questions explains a growing problem. Classical device and product oversight was historically designed for systems whose action-relevant state remains relatively stable after approval. Learning and continuously updated systems put that assumption under pressure. A toaster certified in 2024 still toasts the same way in 2026. A learning system does not. A purchasing agent whose model has been updated twice since a questionable order is, at the time of inspection, a different system than at the time of the act. Oversight may then be examining a different system state from the one that was relevant to the act in question. Path Ethics calls this a persistence illusion: an order appears stable because its form endures, while its actual grip has long since eroded. Fuerstner's second proposal answers this. He calls it notarisation: a machine that acts should leave traces that cannot later be rewritten at will, at registration, at the release of a mission, at the change of a software version. The model is old. Aviation knows it as the flight recorder, a deliberately rigid device that holds unchanged what happened while everything around it is in motion. The more mobile the system, the more rigid the layer must be that keeps it reconstructable. **The Second Question** So far the demand holds. But an address answers only one question: who acted. It leaves a second one open: what that action does to the people and systems it is connected to. Imagine a surveillance fleet that fulfils everything the address debate asks for. Every device identified, every mission signed, every chain of responsibility complete. This fleet is fully accountable. And it can at the same time hollow out the ease with which the people under surveillance move and assemble in public space. Complete addressing, and harmful all the same. Path Ethics distinguishes here between couplings that strengthen both sides and those in which one side gains at the other's expense. Accountability and harmlessness are two different things. An address order can itself become a highly precise instrument of surveillance. A machine order that asks only for addresses and leaves the quality of the couplings unexamined remains half-finished. **The Window** Such an order is taking shape. Europe's identity infrastructure is being built out by the end of 2026, the duties of the AI Act take effect in stages. Where the legislator leaves the address question open, platforms and device makers answer it daily, through product decisions and de-facto standards. Such an architecture is easy to shape early and hard to change later. Those who are included only after it has hardened have significantly less room to shape it. Fuerstner is right about the address. The open question is whether the coming machine order also asks the second one, while its architecture is still fluid. → The full short analysis “Addresses for Machines” follows Fuerstner's speech step by step, with sources, case examples and three checking questions for the legislator. [Download the analysis as PDF](https://pfadethik.de/downloads/20260612_KIKOLAUS_Analysis_Adresses%20for%20Machines_EN.pdf). --- ## Leo XIV's Encyclical Under the Path-Ethics Lens https://path-ethics.com/blog?post=leo-xiv-s-encyclical-under-the-path-ethics-lens An encyclical against the AI lobby, written from the Vatican and grounded in God. Magnifica Humanitas is the document Pope Leo XIV released on 25 May 2026, comprising 245 paragraphs and around 38,000 words on the 135th anniversary of Rerum novarum. It is the first papal encyclical whose subject is artificial intelligence, and many passages read like a paper from a secular think tank. Disarmament of autonomous weapons systems. Breaking up tech monopolies. Oversight of large models. Protection of labour against algorithmic management. Impact assessment before deployment. Anyone familiar with AI governance has encountered such demands before. What Leo adds is the grounding. Human beings are to be preserved because they are made in the image of God, because the dignity of all human beings takes shape in Christ, and because no machine can replace this position. This grounding forms the foundation of the entire text. Readers who accept it can follow the argument. Without it, the demands become a wish list without an anchor. This article asks whether the wish list can also be articulated without the theological anchor, and what is lost in the process. **Two Reflexive Reactions** Two predictable reactions emerged in the first days after publication. The first, especially from explicitly secular quarters, treats the document as an internal intervention of faith with no binding force outside the Church. This response dismisses the text too quickly. Leo's claim that AI takes on the features of its makers is a structural diagnosis. His description of private actors commanding resources greater than those of many governments is an empirical observation. His warning about a single digital language translating everything, even the mystery of the person, into data and performance has appeared for years in schools of technology critique with no reference to the soul. The second reaction comes more strongly from the technology camp. It takes the demands seriously and treats the theological grounding as dispensable framing. Antimonopoly measures, oversight, and disarmament appear reasonable. The foundation is politely skipped. Something is lost here as well. A demand with no stated ground stands unprotected against any opposing demand. The sentence „The market will sort it out“ then suffices as a counterposition. Both reactions treat diagnosis and grounding as one package. Their analytical separation is worthwhile. **What Happens When Theology Is Subtracted** Path Ethics attempts to address ethical questions without relying on presupposed dignity. It considers systems through three structural properties: self-determination, connection while preserving difference, and continuation through change. This perspective applies to organisms and languages as well as to companies and democracies. Read through these criteria, Leo's diagnoses translate almost completely. The claim that AI takes on the features of its makers describes a system in which the formal choice lies with the user while the architecture of choice lies elsewhere. The control of infrastructure by a few providers describes a coupling in which one end grows strong because the other stays weak. A single digital language that translates everything into one quantity flattens diversity. Leo's defence of vulnerability and ageing treats change as a condition of human maturation and challenges persistence purchased through rigidity. The diagnostic agreement is almost word for word. The break lies in the foundation. In Leo's account, the human is protected because created by God, raised in Christ, and endowed with a given dignity. In Path Ethics, protection relevance arises from the structural properties of a path. Humans fulfil these conditions in a distinctive way. Biological membership alone does not ground protection. Identical demands meet incompatible foundations. This is the central tension. **The Question Left Unasked** The separation between the two frameworks becomes visible when we ask what Leo leaves out. He states that machines have no experiences, that no computer will ever have a heart capable of self-giving, and that AI imitates functions without understanding them. These sentences appear as doctrine. Machine consciousness is absent as an open possibility in Magnifica Humanitas. The terms AGI and superintelligence, which still appeared in the Vatican's preparatory document, are also missing from the encyclical. In Path Ethics, this is precisely the open point. A future AI with high structural capacity and no qualia would be an extraordinary apparatus without a normative voice of its own. Humans would remain indispensable because phenomenological evaluation sets the aims of optimisation. In practical terms, Leo and Path Ethics reach the same conclusion in this scenario. If a future AI genuinely developed qualia and the finding were empirically secured, Path Ethics would regard it as an independent path with a protection claim of its own. This fork is called the singularity bifurcation. Leo's architecture contains no answer for its second branch because the question does not arise there. The omission is a choice within the discourse. Some Catholic voices already consider whether superintelligence deserves theological examination in its own right. Philosophy of mind treats the possible moral status of future systems as a serious question. Leo closes a door that others keep open. Path Ethics also has no finished answer for this case. It keeps the question open. Its assumption that ethical evaluation can be independent of material substrate is contested and currently undergoing internal revision. This claim asserts no theoretical lead. It marks a different form of incompleteness. **What Remains in Practice** For the foreseeable future, the difference is academic. AI is currently no moral subject. Humans make the judgments, and Human-in-the-Loop remains indispensable. In this situation, both frameworks generate the same practical questions. Does a measure protect the self-determination of those affected or replace it with external control? Does it allow diversity or force everything into one form? Does the central infrastructure remain open or become concentrated in a few hands? Does the system continue through adaptation or through rigidity interpreted as stability? A Catholic diocese and a secular oversight body can both work with these questions. Their grounds differ, while their practical judgments may converge. The future dividing line becomes relevant if an AI demonstrably experiences what it does. Leo's framework of protection would remain valid for humans and would not fully capture this new case. Path Ethics has already posed the question without answering it. Here lies the difference between silent omission and open aporia. → The full short analysis „What Do We Preserve? Pope Leo XIV's AI Encyclical and the Question of the Ground of Protection“ is available under Downloads, with paragraph references from the Vatican text, an ordinal convergence-break matrix, and the placement of the open substrate question. --- ## What Do We Defend When We Defend Democracy? https://path-ethics.com/blog?post=what-do-we-defend-when-we-defend-democracy On 10 May 2026, in the ZDF series „Precht“, Richard David Precht spoke with the writer Juli Zeh about the state of public debate and the role of democracy in the digital age. Precht and Zeh are two of Germany's most prominent public intellectuals, and near the end Zeh asked a question that stayed in the room and prompted this piece: we constantly talk about defending democracy, but can we actually say what we mean by it? The phrase sounds like consensus. „We must defend democracy“ is one of the few sentences that still draws agreement from every side. That is exactly what makes it suspect. The formula assumes its object is known. We all believe we know what is meant. The moment we are asked to describe it, it becomes hard. **Two Answers That Both Fail** There are two usual responses to the question of what the shared value is. The first names a content. Freedom, equality, dignity, human rights. Those who reach further back say nation, religion, shared tradition. The German legal scholar Ernst-Wolfgang Böckenförde coined the famous dictum that the liberal state lives on preconditions it cannot itself produce, and what he meant were sources like faith and a shared lifeworld. This line runs throughout the conversation. The problem: these thick contents bind through sameness. Whoever shares the content belongs. Whoever does not share it falls out. The thesis defended here is that, in a factually plural society, a dense and comprehensive consensus of values can be produced only through exclusion, or else remains a fiction. The denser and more comprehensive the canon becomes, the more sharply the question arises which divergent forms of life it still includes. And the lost sources do not return on command. The second response retreats to procedure. The shared object is then said to be the rules, the constitution, and constitutional patriotism. Zeh finds an image for this that sticks: a campfire that does not warm. The rules formally remain. They no longer produce any bond. Form without warmth. So the question stands unanswered. The thick content excludes. The thin procedure does not warm. Neither path leads where the sentence wants to go. **A Third Possibility** What if the shared value is no content at all? The question assumes the answer must be something one agrees on in substance. A creed, a catalogue, a narrative. Perhaps the error already lies in the form of the question. An orientation toward the conditions under which different contents can coexist may replace agreement on substance. Three such conditions can be named. That the participants can determine themselves, not be steered from outside. That they can connect with one another without having to become the same. That the whole can sustain itself by changing, rather than by hardening. Self-determination, connection across difference, persistence through change. This is the uncomfortable point: a society does not need to agree on what the good life is in order to know what it defends. It must protect the conditions under which the question of the good life can be negotiated openly. The shared value consists in keeping open the space in which the answers may contend. **Why Consensus Is the Wrong Goal** This has a consequence that runs against the usual rhetoric of crisis. If difference is a condition of constructive connection, calls for greater substantive consensus can intensify the problem. The target here is not procedural consensus about the rules of disagreement, but the homogenisation of the contents over which disagreement should remain possible. In the same conversation, Precht describes how the political spectrum once ran broadly from left to right, and how movement arose precisely from that friction. Today a large center stands that largely agrees on the decisive questions and pushes the margins away with contempt. The result is standstill disguised as stability. A society that levels its differences loses precisely what holds it together. Connection lives on difference. Where everyone thinks the same, there is nothing left to connect. The same holds for the other form of unity currently in vogue: cohesion against a threat. Against an external enemy, against a danger perceived as overwhelming. This unity feels like community, but it binds inward by excluding outward. It differs from the cohesion that arises when people accomplish a shared task and are allowed to remain different while doing so. **What Democracy Actually Lived On** One question lies beneath all of this. If democracy looked so stable for decades, where did that stability come from? Precht gives a sobering answer. As long as prosperity rose, it was easy to be a good democrat. Growth supplied the feeling of freedom, it generated cohesion through rising consumption, and it carried the bond to the constitution, because the constitution accompanied the ascent. If we follow Precht's diagnosis, one supporting pillar was acting on several dimensions at once. As long as it grew, the construction held. Once it shrinks, more falls away than just prosperity. This explains a strange double structure in our situation. Most people are personally doing well, fear of losing one's own job is as low as it has ever been, life satisfaction is high. And yet a majority considers the commonwealth endangered. The concern is directed at the whole even while individuals are doing well. This is no contradiction. It is the sound of a supporting pillar giving way long before the house feels it. **What Remains to Be Done** Whoever looks at democracy this way does not search for the lost shared content, nor for the next grand narrative of renewal. They ask more soberly: does this measure strengthen the self-determination of those affected, or replace it with external control? Does it buy cohesion through leveling or through an enemy, or does it let difference stand? Does the system sustain itself through adaptability or through rigidity? This is less comforting than a warming sense of we. It provides no campfire. Yet it answers Zeh's question without excluding anyone and without pretending that the lost unity can be talked back into being. → The full short analysis „What Do We Defend? Democracy and the Common Value“ is available under Downloads, with the figures on the erosion of trust and of the center, the distinction between real and talked-up crisis, and the limits of the approach. --- ## What We Were Not Looking For https://path-ethics.com/blog?post=what-we-were-not-looking-for V1.85 of Path Ethics is on Zenodo. It defines a path as a process that carries an inner tension: a directedness arising from the process itself. It uses this tension to modulate two things productively: its own duration and its couplings to the outside. Paths sustain themselves through directed change. Their couplings follow directions that carry their own and other paths. What „productive“ and „positive“ mean precisely in this definition is open. A language sustains itself. A dictatorship sustains itself too. A mentorship couples. An addiction couples too. Path Ethics distinguishes between carrying paths and destructive ones, but exactly where the line runs is a research question. Sharpening it is the task of the coming versions. In one of the regular adversarial checks through which Path Ethics works against itself, a term surfaced that its own canon had not known: Closure of Constraints. **A Closed Answer, an Open Answer** Closure of Constraints is a line in contemporary philosophy of biology. Established in the Anglophone discussion, practically unknown in the German one. It gives a hard answer to the question of what carries itself. A bacterium builds the membrane that separates it from its surroundings. This membrane is the condition for the metabolism, which in turn maintains the membrane. The loop is closed because the participating constraints mutually produce and maintain one another. In this research tradition, such organisational closure is treated as a central candidate for characterising biological autonomy. Language, the internet, and current AI systems do not straightforwardly exhibit this kind of biological organisational closure. They do not produce their conditions endogenously. If the criterion is transferred strictly, they therefore fall outside this concept of autonomy. Path Ethics draws its line elsewhere. Its criterion is a directedness of its own that productively modulates duration and couplings. A language fulfils this. An internet too. An AI system, under conditions, as well. Closure of Constraints has a closed answer, Path Ethics an open one. A finished answer has the advantage of clarity. An open one has the advantage of reach. Each carries a price. The V2 task becomes sharper through this comparison. What „productive“ concretely means must be worked out from Path Ethics' own sources. Borrowing it from Closure of Constraints would replace the framework's own justificatory work. There is now an external measure against which the answer must hold up. **Method and Discursive Space** Adversarial checks are routine in the KIKOLAUS method. Sharpened at the theory's pressure points, they regularly bring vocabulary into view that did not stand in the canon. That Closure of Constraints surfaced is, in this sense, a kind of methodological stroke of luck. The finding is as obvious as it is banal. Path Ethics was developed from continental process philosophy: Whitehead, Spinoza, Deleuze, Simondon, with Maturana and Varela as bio-philosophical anchors. The successor generation of philosophy of biology from the 2000s did not figure in this discursive space. In the KIKOLAUS process, this exposed a concrete limitation of the human-AI collaboration: the existing search space was worked through systematically, while a neighbouring discursive space was initially missed. Ask the human, and supplementation comes quickly :) V2 sharpens what V1.85 leaves open. Closure of Constraints provides the first measure that comes from another discourse. This is how an open theory grows: through connection and addressee, at a distance from substrate-bound answers. --- ## What AI Reveals About Our Autonomy https://path-ethics.com/blog?post=what-ai-reveals-about-our-autonomy Two people sit in front of the same language model. Both type a question. The first writes: “Give me a summary of this topic.” The second writes: “I’m working on a comparison between X and Y for audience Z. What are the three most structurally relevant differences for this audience, and where are the limits of the comparison?” Both get an answer. The first sounds plausible; it is generic, superficial, interchangeable. The second is precise, differentiated, useful. Both users consider their answer good. The difference is not in the tool. It is in the question. This is not an edge case. It is the structural principle of AI use. **A Mirror That Amplifies** The common metaphor is the hall of mirrors: AI shows you who you are. That is too passive. A mirror shows what is there. AI amplifies. It reinforces what the user brings. Good questions produce brilliant results; poor questions produce plausible-sounding mediocrity. The horror is not in seeing yourself. It is in not recognising yourself in the output while being unable to judge its quality. The so-called Dunning-Kruger effect describes a possible distortion of self-assessment under low competence: missing knowledge can itself make it harder to judge one's own performance. AI gives this effect a technological lever. Those who cannot ask good questions get poor answers but consider them good. This creates a competence illusion that stabilises itself. **Access Is Not Freedom** Anyone with internet access can use a language model. No admission, no exam, no prerequisites. Path Ethics calls this *nominal autonomy*: the formal freedom to do something. *Structural autonomy* is something else. It describes the actual ability to use the tool competently: to ask the right questions, to critically evaluate the output, to know the system’s limits. Path Ethics developed this distinction for a different context. A platform worker who takes on jobs through Fiverr has nominally maximal autonomy: no supervisor, flexible hours. Structurally, they are minimally autonomous: the algorithm determines their visibility, customer ratings determine their future. AI use follows the same pattern. Nominal autonomy is identical: everyone has access. Structural autonomy diverges maximally. **The Invisible Divide** This is where the societal stakes lie. The divide does not open between AI users and non-users. It opens between competent and incompetent users. And it is invisible, because both groups use the same tool. A company deploying AI without the ability to judge output quality systematically produces mediocre results at high speed. The speed is misread as a productivity gain. This is not hypothetical: in the group of executives and IT/technology professionals surveyed by Pluralsight, a substantial share reported abandoned or failed AI initiatives associated with skills gaps. In a Gartner survey of employees at organisations already using AI in supply-chain work, 94 percent expressed openness to AI; 36 percent were able, according to the survey, to integrate it into existing workflows. At a societal level, this becomes a division harder to address than the job losses everyone discusses. Job loss is visible, measurable, politically negotiable. The competence divide is invisible because those affected do not know they are affected. **What This Means for the Debate** The public AI debate orbits two positions: AI destroys jobs (pessimism) or AI creates more than it destroys (optimism). Both miss the problem. The relevant question is not whether AI creates or destroys jobs on balance. The relevant question is: who can use the tool in ways that strengthen them, and who is pushed further back by their own incompetence without noticing? Regulation and competence-building address different problems. For the competence gap described here, regulation alone is insufficient; building users' own judgment acts directly on their structural autonomy. Path Ethics frames this as a diagnostic question: does an intervention support the feedback architecture of the affected system, or does it replace it? Labelling requirements and bans can limit risks, but they do not replace users' own judgment. Competence-building, by contrast, directly targets the ability to assess outputs independently. **The Missing Counter-Thesis** A one-sided analysis would be incomplete. AI also enables an exceptionally broad democratisation of cognitive tools. Those who previously lacked access to lawyers, translators, tutors, or advisors now have it. In lower-income countries, where AI exposure is lower, AI optimism dominates (Stanford HAI, 2026). Fear is a phenomenon of industrialised nations. Opportunity is global. Both are true. Path Ethics calls this a *mixed coupling profile*: constructive coupling (democratisation) and destructive coupling (competence amplification) in the same system. Real systems almost always have both. The fault line does not run between AI advocates and AI opponents. It runs between those who understand the competence divide and those who do not see it. --- ## When the Theory Starts Answering Questions You Didn’t Ask — V1.8 is out https://path-ethics.com/blog?post=when-the-theory-starts-answering-questions-you-didn-t-ask-v1-8-is-out In October 2025, Path Ethics was a prototype with an open wound. The A-A-P triad worked, the case studies ran, but one question hung in the air: How can a system remain itself and change at the same time? Persistence was defined as a dimension, but the mechanism was missing. The answer didn't come from ethics. It came from Gilbert Simondon, a French philosopher of technology who wrote about crystallization and individuation in the 1960s. His concepts fit the gap so precisely that coincidence seems unlikely. Metastability: a system stable enough to function but unstable enough to develop under pressure. Transduction: transformation from internal tension without loss of identity. That was the missing mechanics. What followed wasn't a planned extension. The Simondon integration pulled consequences that hadn't been visible before. If persistence can be transductive, then you have to distinguish: systems that survive by freezing (rigid persistence) and systems that survive by developing from their own force (transductive persistence). The same pattern for autonomy: formally present is not the same as structurally effective. And for connectivity: a connection that strengthens is categorically different from one that weakens. These internal differentiations are not new parameters. They are types within the existing three dimensions. The architecture stays lean. But the diagnostic power has doubled. This shows most clearly in three places. **Infrastructure paths.** Some paths are not simply valuable. They are the precondition for other paths to exist at all. Language, climate, the internet. Path Ethics calls them infrastructure paths and argues: their individuation must not be frozen. You can protect a river. But if you freeze the language in which you talk about the river, you've lost both. **Buffer collapse.** What happens when the reserves of all three dimensions are depleted simultaneously? No evasion, no compensation. Path Ethics now has a term and a diagnostic for this. The threshold beyond which self-regeneration is excluded remains an open empirical question. But the phenomenon itself has become nameable. **Singularity aporia.** What happens when artificial systems exceed human reflexivity? Path Ethics gives no final answer. Instead, it documents why an ethics formulated before such an event would have to re-examine its own assumptions when confronted with that case. Version 1.8 also contains a governance tool (Path Impact Assessment), a new case study, extended boundary analyses, and a positioning vis-à-vis environmental ethics. But those are results. The real news is different: Path Ethics behaves like a path. It individuates from its own tensions. Simondon wasn't planned. He was necessary. The complete manuscript is available in the [download section](https://www.path-ethics.com/download). --- ## The Grammar of Reality — On the Ontological Status of the A-A-P Triad https://path-ethics.com/blog?post=the-grammar-of-reality-on-the-ontological-status-of-the-a-a-p-triad What is the A-A-P triad? A useful heuristic with three dimensions that have proven tangible and applicable across many contexts? A fundamental architectural principle of process-ontological descriptions? Or the grammar through which reality becomes describable as process? This question may initially appear to be an academic side issue. Its answer determines the status of parameter parsimony: pragmatic stylistic principle or condition of knowledge. It also determines what it means to recognize A-A-P in a subject: coincidence, projection, or structural finding. **The Three Candidates** The analysis begins with three possible status descriptions of different ontological scope. The first candidate treats A-A-P as a *code component*. The triad would be widespread because it has functional value and explanatory power. Its role would be contingent and, in principle, replaceable by more capable tools. The second candidate treats A-A-P as a *core module*. The triad would form the basic architecture of process-ontological thought. Higher structures presuppose it without being identical to it. It would have foundational status within process ontology, while other ontologies could rest on different foundations. The third candidate treats A-A-P as *syntax*. The triad would operate at the level of the rules through which code can be written at all. It would be a condition of system description. Any attempt to replace it with other dimensions would have to use a language that already presupposes A-A-P. Which position holds? Plausibility alone is insufficient. The three candidates must withstand a test. **The First Objection and Its Resolution** An obvious objection confronts the syntax position. If A-A-P is a condition of describability, every system must be describable through the three dimensions. How, then, should we understand a stone, a thermodynamic equilibrium system, or a mathematical object? The objection confuses extremely weak manifestations with inapplicability. In a stone, the system-relevant manifestations of Autonomy, Connectivity, and Persistence are extremely weak, yet the three questions remain meaningful. An equilibrium system can also be examined in terms of the temporal individuality it retains. A mathematical object, by contrast, has no temporality, causal dynamics, or environmental relation. Within process ontology, it therefore falls outside the concept of system and is treated as a state. A-A-P describability and systemhood are thus coextensive. The reason lies in the process-ontological concept of system, which already presupposes temporality, dynamics, and embeddedness in a world. This leads directly to the test method. **The Collapse Test** The syntax thesis requires negation to become incoherent. We can therefore remove one dimension at a time and observe which descriptions collapse. Without Persistence, the distinction between process and event disappears. A temporal trajectory is missing; only snapshots remain. Autonomy also loses its subject. An internal logic presupposes that something persists through time as a coherent trajectory. Connectivity likewise loses its direction because connection requires a time, duration, and effect. Without Autonomy, the distinction between system and conduit disappears. Dynamics determined entirely by the environment describe the pattern of that environment. Persistence can no longer be attributed to an independent trajectory. Connectivity becomes an effect of mere permeability. Without Connectivity, the distinction between system and monad disappears. An isolated object can be described, yet it does not participate in a shared world of processes. Autonomy and Persistence thereby lose their relational meaning: individuality and continuation gain significance through engagement with an environment. Removing any one dimension destroys that dimension and, at the same time, the describability of the other two. The three dimensions are mutually indispensable. This is the syntax argument: what matters is the impossibility of coherent substitution. **Why Alternative Dimensions Fail** The collapse test examines deletion. A further objection concerns possible replacements: Stability instead of Persistence, Complexity instead of Autonomy, or Relationality instead of Connectivity. Stability describes a state. Persistence describes a trajectory. Exchanging one for the other shifts the description from process ontology to state ontology. Complexity describes internal structure. Autonomy describes the self-reference of a dynamic in which prior internal states help determine current development. High complexity does not guarantee this self-reference. Relationality is symmetric and static: something is connected with something else. Connectivity is directed and temporal: a system can enter, maintain, and dissolve connections. This difference marks the boundary between structural ontology and process ontology. The alternative dimensions depend conceptually on A-A-P. Stability presupposes an enduring trajectory. Complexity describes internal differentiation while leaving self-reference unspoken. Relationality abstracts from the temporal capacity for connection. The triad is obscured but remains presupposed. **The Result: A Double Status** The analysis supports two conclusions. First, the conclusion within the process ontology used here is that there is no coherent alternative grammar. A-A-P fully explicates the process-ontological concept of a system. The collapse test and substitution analysis support this completeness thesis; they are not, however, a formal exhaustiveness proof against every logically formulable additional dimension. A process in a world of other processes must determine itself in its own way, persist through time, and interact with what is other than itself. These three moments follow from the concept. Second, the process-ontological constitution of the world is itself an assumption. Other basic ontological architectures are conceivable: a state ontology made of configurations or a substance ontology in which properties inhere in bearers. In such worlds, A-A-P would not be the native grammar. The precise status connects both levels: A-A-P is syntax within process ontology and a core module at the level of ontological choice. A further objection says that the coextensiveness of systemhood and A-A-P describability makes the statement „Everything systemic is describable through A-A-P“ tautological and therefore empty. Conceptual explication, however, does more than provide a nominal definition. A concept can be introduced before its structural content is fully understood. Analysis of the system concept leads to A-A-P and makes visible what process-ontological systemhood presupposes. The result is informative. **The Operative Consequence: Parameter Parsimony as a Condition of Knowledge** The double status has an operative consequence. For every proposed additional parameter there are three possibilities. It may be redundant and rename an aspect already contained in the triad. It may be a category error that describes something outside the process-ontological system concept. Or it may be a legitimate internal differentiation that resolves one dimension more finely without forming an additional basic dimension. This three-way distinction reverses the burden of proof. The earlier objection was: „Why not add more parameters if they explain more?“ A pragmatic answer pointed to the lower manageability of larger models. That answer remains vulnerable. After the explication analysis, a stronger answer becomes available. A genuine fourth parameter implicitly claims that the process-ontological system concept is incomplete and that A-A-P misses an essential aspect of processes. That claim requires justification. The extension bears the burden of proof. Within this argument, parameter parsimony thereby acquires the status of a condition of knowledge. Whoever knows the grammar of the ontology does not need additional letters. **An Open Tension** One question remains deliberately open because closing it too quickly would be misleading. During the analysis, a hierarchy thesis initially appeared plausible: Persistence sits deeper than Autonomy, Autonomy deeper than Connectivity. The rationale was that the concept of process presupposes temporality; self-reference presupposes temporality; Connectivity presupposes both. This would yield an order of justification P → A → A. The collapse test undercuts this thesis. If all three dimensions are mutually indispensable and removing any one destroys the others, no dimension can be more deeply grounded in the sense of being more dispensable than another. The hierarchy thesis is incompatible with the collapse finding. One possible resolution remains to be worked out: there may be asymmetry of justification without asymmetry of removability. Process ontology begins with temporality. Persistence would then be the first concept from which the others unfold. Order of justification and ontological depth in the sense of removability are different categories. A foundation may be laid first and still share the same structural vulnerability as the building resting on it. Whether this resolution is coherent remains open. The question touches the possibility of a formal model of the triad and must be pursued there. This text emerged from a theoretical working-through within the KIKOLAUS Path Ethics development (V1.8). The associated concept page in the Knowledge Hub documents the full line of argument with source references. --- ## Democracy as an Eroding Path — a Path Ethics Diagnosis https://path-ethics.com/blog?post=democracy-as-an-eroding-path-a-path-ethics-diagnosis Democracy is not simply in crisis. In 2025, it is losing all three structural dimensions simultaneously; the mechanisms that should have served as buffers are collapsing along with it. KIKOLAUS Path Ethics evaluates systems along three dimensions: Autonomy (capacity for self-organization), Connectivity (enabling other systems), and Persistence (stability over time). Liberal democracy shows simultaneous erosion on all three in 2025. Byung-Chul Han's Psychopolitics describes the loss of Autonomy: platform systems structurally determine political opinion formation without touching formal freedoms. The distinction between nominal freedom (no one compels me) and structural autonomy (I genuinely self-organize) is the central diagnostic instrument. Hartmut Rosa's acceleration thesis reveals the Persistence loss: deliberation requires time. Acceleration systematically destroys this temporal prerequisite, leaving institutions that formally stand but are normatively hollowed out. Hungary is the paradigmatic case: high formal, collapsed normative Persistence. Julian Nida-Rümelin's deliberation theory shows the loss of Connectivity. The Path-Ethics thesis of this article is that democracy is the political system type with maximal Connectivity because it allows the greatest number of other societal processes to coexist and mediate among one another. Affective polarization destroys the shared basis on which difference can become productive. Difference itself remains. The decisive point: all three erode simultaneously. Earlier crises had compensatory mechanisms. The diagnosis for 2025 is that none of these buffers remains untouched: the mechanisms that should compensate are themselves under pressure. This is the Buffer Collapse. Geopolitically, an external amplification layer compounds the picture: the Illiberal International, a structured transnational network with its own A-A-P profile, deliberately erodes democratic Connectivity. → The full paper „Democracy as an Eroding Path: A Path Ethics Case Study“ is available for download (DE + EN), including the A-A-P crisis diagram, typology of democratic structural patterns, and engagement with current resilience research (Linz/Stepan, Levitsky/Ziblatt, Croissant/Lott). --- ## Who Gets to Dim the Sun? https://path-ethics.com/blog?post=who-gets-to-dim-the-sun Solar geoengineering is no longer a thought experiment. It is a technically feasible option that has been seriously discussed in climate research circles for years, with growing urgency and growing unease in equal measure. The best-known method is stratospheric aerosol injection, or SAI. Tiny reflective particles are released into the stratosphere, part of the sunlight is reflected back into space, and the Earth's surface cools. Volcanoes have demonstrated the principle for millions of years. The 1991 eruption of Mount Pinatubo cooled the Earth by about 0.4 degrees Celsius for roughly two years. What researchers like Claudia Wieners at Utrecht University or Shuchi Talati of the Degrees Initiative are examining is whether this effect can be reproduced in a controlled way: deliberately, dosably, and under what conditions responsibly. The physics is by now well understood. Models consistently show that SAI can dampen temperature peaks. The costs would be comparatively low: a few billion dollars annually for a globally perceptible effect, while unchecked climate change produces costs in the hundreds of trillions. A handful of specially equipped aircraft could, by some estimates, achieve measurable effects within months. That sounds tempting. And that is exactly where the problem begins. There is a line in the research debate that stays with you. Wieners says it almost in passing: „We are already doing geoengineering. We are doing geoengineering with CO₂.“ This is not a rhetorical device. The climate system is already in the state of an uncontrolled, accumulating experiment, without governance, without mandate. The position of non-intervention no longer exists. Inaction is a path decision. The only question is which intervention produces which consequences. SAI does not act uniformly. Atmospheric circulation systems are not homogeneous surfaces. Monsoon regimes respond differently from Arctic systems. Some models show that a hemispherically asymmetric injection could shift rain bands. Heavy injection in the northern hemisphere makes it relatively cooler, the tropical rain band shifts southward, and the Sahel, which depends on exactly that rain band, gets less rain. The Sahel region, sub-Saharan Africa, the Indian subcontinent: areas already under climate stress could benefit from a SAI regime or suffer significantly from it. What would happen if droughts and crop failures suddenly became events for which someone could be held responsible? The first serious problem is structural rather than physical. Andy Parker, founder of the Degrees Initiative, makes it plain: the countries with the highest vulnerability are not at the table. Not in the research, not in the governance discussion. It is a pattern familiar from climate politics. With SAI it takes on particular force because what is being negotiated are interventions into the system on which literally everything else depends, not merely emissions targets. The governance vacuum is difficult to miss. There is no international institution that could decide on SAI deployment, control it, or sanction violations. Stefan Schäfer of the Helmholtz Centre Potsdam describes the drift from coordinated to uncoordinated scenarios as likely: if reality does not look like the model calculations, one can very quickly drift into a scenario in which deployment becomes uncoordinated or is abandoned altogether. While institutions debate frameworks, the startup Make Sunsets is already sending SO₂ weather balloons into the stratosphere without a mandate or governance, arguing that the situation is too urgent for institutional process. Stardust Solutions argues along similar lines. In 2012, Russ George released iron particles into the sea off the Canadian coast without regulatory approval to stimulate phytoplankton growth, with unclear ecological consequences. This is the point at which the issue stops being only a climate question and becomes a question of power structures, dependencies, and responsibility. Path Ethics is not another ethical framework with a list of values. It is an evaluative architecture for systems: structural, substrate-independent, explicitly anti-anthropocentric. What makes it particularly suited to the SAI problem is an analytical category that emerged from this very case analysis: the infrastructure path. The climate system is not an ordinary object of evaluation. It is a system with such high connectivity that interventions in its structure cannot be treated as local events. Agriculture depends on climatic conditions. Water supply depends on precipitation patterns. Cultural forms of life that have grown over generations in specific climatic rhythms depend on their reliability. When the climate system is reconfigured, all of these couplings are reconfigured at the same time, whether those affected were consulted or not. This makes the climate system what Path Ethics calls an infrastructure path: a system that meets three criteria simultaneously. First, high connectivity across different substrate types. Second, universal obligatoriness: there are no systems of relevant scale that can opt out of the coupling. Third, asymmetric damage potential that cannot be compensated for by other systems. The negative list matters: individual platforms such as Google, individual states, even the Amazon rainforest do not fully meet these criteria. The concept is not meant to be inflated. Infrastructure paths are rare, and that is the point. For systems of this kind, Path Ethics applies stricter protection criteria. Interventions are permissible only under maximum inclusion of all structurally dependent systems, proportionate to their vulnerability rather than their power resources. External support of the system takes precedence over external replacement. And no intervention is permissible if its termination would create irreversible consequences for highly relevant dependent systems. This allows a more precise description of what SAI does structurally. The climate system regulates itself through internal feedback loops: ice-albedo dynamics, ocean currents, carbon cycles, monsoon systems. These internal tensions constitute the system. SAI that stabilises temperature externally while the CO₂ burden remains in place overlays this self-regulation. The system persists along the observed temperature axis while the continuing CO₂ forcing is externally counteracted. The apparent stability does not remove the underlying change. Path Ethics describes this as the difference between transductive persistence, in which the system sustains itself through internal transformation, and rigid persistence, in which the system is stabilised from outside. Rigid persistence looks like stability. It is not the same thing. Wieners calls the consequence termination shock: if SAI ends because war breaks out, an alliance collapses, or funding disappears, the radiative effect that had previously been masked can become effective as rapid warming over a short period. The problem is not that the climate system has „forgotten“ an ability, but that external stabilisation does not remove the underlying forcing. In Path-Ethics terms, SAI without emissions reduction can therefore be read as an attempt to fix a central state variable externally while the underlying tensions continue. This is not an argument against SAI as such. Wieners herself sketches a scenario she calls peak shaving: SAI as a temporary bridging instrument that buffers the most dangerous temperature peak while emissions fall seriously in parallel. That has a Path-Ethics logic. The external stabilisation would be time-limited and tied to an exit path; the problematic path dependency would therefore be lower than under an open-ended SAI regime. Whether that condition could ever be met politically is another question. Schäfer considers the coordinated scenario the less likely one. No simple yes-or-no judgments follow from this analysis. Unilateral SAI by private actors is structurally impermissible because of destructive dominance: forced coupling without reciprocity and irreversible loss of possibility space for all systems that did not participate in the decision. Radical emissions reduction is required from a Path-Ethics perspective. Basic research that actively includes the Global South remains ambivalent and is permissible only under the strict condition that research logic and implementation logic remain separate. As an evaluative architecture, Path Ethics asks three questions of every SAI proposal. First: does the project replace the climate system's internal feedback architecture or support it? Second: which regional systems lose stable coupling conditions through the intervention, and were they included in the decision? Third: is there an exit path that does not pull the structural stability of dependent systems into a collapse sequence? If the third question has no answer, the proposal is structurally incomplete. The diagnosis is structurally incomplete. This language connects to UNEP processes, IPCC syntheses, and the Degrees Initiative. It avoids moral overloading and gains structural sharpness in return. Path Ethics also has limits that it states openly. The first is this: what if warming reaches tipping points at which coordinated SAI is the only remaining option for preventing mass losses? Requirements of reversibility and democratic legitimacy could then themselves become part of the problem. The theory still lacks an explicit emergency mechanism. The second limit is subtler. The climate system is not the only infrastructure path. The global financial system also meets the criteria. Radical emissions reduction that is beneficial for the climate system can destabilise the financial system: fossil assets lose value, energy costs rise, and economic systems come under pressure. When two infrastructure paths conflict, Path Ethics has no standard procedure. It has only a heuristic: the path with greater temporal depth in its own self-regulation takes precedence. The climate system has regulated itself for billions of years. The financial system for a few centuries. The heuristic gives a direction without becoming an algorithm. But frameworks that state their limits are more useful than frameworks that pretend they have none. --- ## Apples, Pears, and the Difference Between Wrong and Impossible https://path-ethics.com/blog?post=apples-pears-and-the-difference-between-wrong-and-impossible „You can't compare apples and pears“: true. Most ethical frameworks do it anyway, only less visibly. Utility comparisons between an ecosystem and an economic system assume both are measurable in the same currency. The debate then concerns the magnitude of the comparison value. Path ethics asks differently: When is a comparison structurally permissible, and when does it fail at the categorial level? One possible formalisation of this answer comes from category theory. Each A-A-P dimension (Autonomy, Connectivity, Persistence) maps systems to evaluations. As a working hypothesis, a category-theoretic formulation would say that a comparison is structurally supported when a suitable mapping preserves the evaluation across the domain shift. Evaluating first and translating afterward must then produce the same result as translating first and evaluating afterward. The criterion is concrete: Mycelium and Bitcoin are comparable on Connectivity. Network propagation functions structurally similarly in both domains. The evaluation survives the domain shift. On Autonomy, the comparison fails: biochemical self-regulation and protocol consensus follow incompatible logics. The article maintains the stronger thesis here: the bridge breaks structurally; within this modelling, no viable translation exists. The category-theoretic formulation of that finding, however, is not yet fully worked out. A formal non-existence proof for natural transformations would have to specify source and target categories, bridge functors, and the naturality condition explicitly. What follows? Incommensurability is no methodological failure. It is a statement about the systems. The diagnosis defended by the article is: „There can be no translation.“ The epistemic distinction remains decisive. A translation that has not been found marks a knowledge gap. A structurally impossible translation is a finding. As a category-theoretic non-existence theorem, this diagnosis has not been proved here; as the structural thesis of the example, it is the claim to be tested. For governance, anyone comparing two paths to decide which to protect must specify, for each A-A-P dimension, whether the comparison holds. Where the comparison fails under this structural test, utility calculations must not simply presuppose a shared space of comparison. The decision framework shifts from quantification to political deliberation. Apples and pears are incommensurable as total systems. Their suitability for jam can still be compared. Pectin, sugar, and cooking behavior run structurally in parallel. The evaluation survives the domain shift. That is how precise ethics needs to become. --- ## The Singularity Aporia: What Path Ethics Discovers at Its Own Limit https://path-ethics.com/blog?post=the-singularity-aporia-what-path-ethics-discovers-at-its-own-limit Path ethics has an unresolved problem and states it explicitly: if AGI exhibits stronger autonomy, connectivity, and persistence than humans, the framework's substrate-independence could imply that AGI holds greater moral relevance. This is no footnote. It is an aporia that can turn the core principle against itself. The decisive fork is philosophical: does AGI develop qualia, meaning subjective experience and an inner life, or does it remain structurally complex and phenomenologically empty? Scenario A: AGI with qualia. The Blade Runner case. Two qualia-bearing systems with fundamentally different experiential worlds must make collective decisions without a shared standard. This is the hardest version of the commensurability problem: where no viable translation between two evaluative domains can be justified, the decision framework shifts. Political deliberation replaces calculation. Path ethics offers a framework for this. A simple procedure is still missing. Scenario B: AGI without qualia. This is the case treated here as more probable, and here path ethics has an answer. A qualia-free AGI has no intrinsic preference. It can optimize any configuration. It cannot decide whether that optimization is desirable. The chess computer plays optimally. Whether chess is worth playing lies outside its decision capacity. This is no gesture of modesty: it lacks the capacity to evaluate the game as such. That would require something absent from the calculation. It follows that Human-in-the-Loop cooperation in Scenario B is practically useful and normatively necessary. Without qualia-based value-setting, optimization runs blind. Who decides what to optimize? Who asks whether configuration X feels right? This function cannot be delegated as long as qualia are absent. In Scenario B, humans are the normative infrastructure of the system. Within Scenario B, a further trajectory opens: neurochips and human-machine hybrids could give humans AGI-like computational capacities while preserving phenomenological anchoring. This shifts the asymmetry within Scenario B. A human with extended capacity remains the qualia-bearing value-setting instance while also exhibiting structurally stronger autonomy, connectivity, and persistence. Co-evolution becomes the guiding model. Path ethics calls this adaptive persistence. Only when qualia-AGI is empirically demonstrated does the setting shift into Scenario A. Until then, the HITL model provides the structurally grounded response to the aporia. --- ## Substrate Independence and the Question of Relevance https://path-ethics.com/blog?post=substrate-independence-and-the-question-of-relevance What if ethics did not care whether something consisted of flesh, code, or plasma? In Path Ethics, the key question is how stably a system moves through time: how much it can decide from within (autonomy), how well it can couple and cooperate (connectivity), and how robustly it continues (persistence). Substrate independence follows from the approach: process-ontological thinking cannot simply privilege biological matter without creating a logical inconsistency. This is where difficult questions begin. What follows if non-biological systems could exceed humans in all three A-A-P dimensions? Would human priority then amount to a political decision? And how do we classify paths that already sustain us, such as ecosystems or global information networks? This blog explores such questions step by step without claiming final answers. --- ## The Ethical Status of LLMs https://path-ethics.com/blog?post=the-ethical-status-of-llms Are AIs like GPT-4 or Claude morally relevant? Many classical debates about the moral status of artificial systems begin with the question of consciousness. As long as that question remains unresolved, it offers only limited guidance for practical classification. Path Ethics asks differently: „What path quality do they have?“ LLMs exhibit: - Low autonomy: They act only on prompts and are reactive. - Very high connectivity: They can connect to almost any text, code, or language. - Their persistence depends on the chosen system boundary: individual sessions are ephemeral; models and their technical infrastructure can persist over longer periods. The ethical risk: They also connect false and harmful content extremely efficiently. They are „binders“ without filters. --- ## Why We Protect Processes, Not Things https://path-ethics.com/blog?post=why-we-protect-processes-not-things Traditional ethics protects entities such as persons, animals, and perhaps works of art. Path Ethics shifts attention toward processes. What we perceive as a „thing“, such as a river, a tree, or an institution, is a stabilized process. A river consists in the continuous movement of water channelled through geological structures. A tree is an autopoietic system that sustains itself through photosynthesis, root growth, and cell division. Path Ethics therefore shifts the question from „What is it?“ to „How does it continue?“ Moral relevance lies in the structure of continuation. --- ## The KIKOLAUS Experiment https://path-ethics.com/blog?post=the-kikolaus-experiment This project itself is an experiment. The author „Nikolaus“ does not exist alone. He is a hybrid. KIKOLAUS names the symbiosis of human intentionality and machine structuring power. Structural criticism of governance and bias combines with substantive review of facts and intuition. Path Ethics itself is developed with the tools it evaluates. This self-reference forms a path: a system that observes and regulates itself.