What Should a Human Being Be Able to Do and to Be?

Martha Nussbaum and the Concrete Face of Justice

Justice is often discussed in abstract terms.

We speak of equality, freedom, rights, distribution, and human dignity. These words are necessary. They give us a language with which to criticise power and describe the principles on which a society ought to be built. Yet the words can also remain at a distance from human life.

A society may declare that everyone is equal in dignity while still being organised in such a way that some people cannot enter its buildings, understand its information, express their will, or participate in the communities where decisions are made.

The right exists on paper.

The opportunity is absent from life.

It is in this gap that Martha Nussbaum’s philosophy begins. She does not ask only which rights a person formally possesses. She asks what that person is actually able to do and to be within the conditions society has created.

The question is simply formulated, yet far-reaching:

What should every human being have a genuine opportunity to do and to be in order to live a life consistent with human dignity?

With this question, Nussbaum turns the gaze of justice away from abstract principles alone and towards concrete human life. She does not ask first how many resources a person has been given, but what those resources actually make possible. She does not ask only whether people have been treated equally, but whether they have in fact been given genuinely equal opportunities to live.

Justice thereby acquires a face.

The Human Being Behind the Theory

Political theories often begin with an idea of what a human being is.

This starting point is decisive, even when it is not stated explicitly. If the human being is imagined as rational, independent, and productive, society will be organised around those qualities. The person who can plan their own life, negotiate their interests, and contribute economically will easily appear as the typical citizen.

But who remains at the margins of the theory?

Often it is the person who needs help to understand, choose, and act. The child. The seriously ill person. The older person who has lost parts of their memory. The person with extensive cognitive disabilities. The person who cannot make themselves understood in society’s preferred language.

Such people may of course be included in a theory. They may receive care and protection. But they often enter only after the basic image of the citizen has already been drawn.

Nussbaum challenges this starting point. She argues that dependence, embodiment, and vulnerability are not special cases to be added later. They belong to human life itself.

We are born dependent on others. We live in bodies that may become ill or injured. We may lose abilities we once possessed. We need relationships in order to develop language, identity, and security. Even the most independent citizen depends on institutions, nature, care, and the work of other people.

The autonomous human being is therefore not the whole truth about humanity.

Nussbaum begins instead with a living, embodied, and relational human being. This person has needs, emotions, imagination, and the capacity to form attachments. They may flourish, but they may also be violated. They may act, but they are never completely independent.

A theory of justice must fit this human being, not an idealised creature none of us fully is.

From Resources to Opportunities

It is possible to distribute resources equally and still produce unjust outcomes.

Two people may receive the same sum of money, yet have very different possibilities of converting it into a good life. A wheelchair user may need more expensive transport or an adapted home. A person with serious illness may spend considerable resources simply maintaining the health others take for granted. A person with cognitive difficulties may need support to understand and use schemes that are formally available.

The same is true of rights. A right to education means little if the teaching cannot be understood. A right to work means little if the workplace is inaccessible. A right to political participation means little if information, premises, or support arrangements exclude people.

Nussbaum therefore directs attention towards what resources and rights are transformed into in people’s actual lives.

This is the core of her Capabilities Approach.

The English word capabilities is difficult to translate precisely. It does not primarily refer to inner abilities or skills. It concerns opportunities to which a person has genuine access. We might call them opportunities for action, opportunities for living, or enabling conditions.

What matters is not only what a person can do through their own power.

What matters is what the person has the opportunity to do in interaction with their social, political, economic, and material environment.

Some of the responsibility is thereby shifted from the individual to society.

When a person cannot participate, we should not automatically ask what is lacking in the person. We must also ask what is lacking in the environment.

Ability and Opportunity

The distinction between ability and opportunity is crucial.

A person may have the ability to move, but not the opportunity to enter a building because stairs block the way. Another person may have limited mobility and yet have a good opportunity to participate because the surroundings are accessible.

In the same way, a person may have the ability to express wishes, yet lack the opportunity to be heard. The person may speak slowly, use few words, or communicate through signs and bodily expressions. If no one gives them time or tries to understand, their voice effectively disappears.

The problem does not then lie entirely within the individual.

It also lies within the encounter.

This perspective challenges the medical understanding of disability as a characteristic of the body alone. An impairment may be real and significant, but the degree of disability also arises in the relationship between the person and the environment.

The staircase creates disability for the person who cannot walk. Inaccessible language creates disability for the person who needs simplified information. The rapid conversation creates disability for the person who needs more time.

Justice is therefore not only about repairing or training the individual.

It is also about changing the world.

A Life Consistent with Human Dignity

Nussbaum connects capabilities to human dignity. A just society must secure for every person access to a level of central opportunities that makes it possible to live a life of human dignity.

This does not mean that public authorities should determine how people must live.

Nussbaum distinguishes between capabilities and actual functionings. A capability means that a person may choose to do something, but is not compelled to do it.

A person should, for example, have the opportunity for political participation. But they should not be forced to engage in politics. They should have the opportunity to enter social relationships, but also the right to withdraw. They should have the opportunity for play and enjoyment, but not be required to participate in prescribed activities.

This protects freedom.

The task of society is not to produce identical lives. It is to create conditions in which different people can choose and develop lives in accordance with their own values, circumstances, and interests.

Justice therefore contains both something universal and something personal.

The universal is that all human beings should have access to basic opportunities for living.

The personal is how each individual chooses, or with support is helped, to shape their particular life.

The Ten Central Capabilities

Nussbaum proposes a list of ten central capabilities to which every human being should have access. The list is not intended as a complete description of the good life. It is rather a political minimum, a threshold below which a just society should not allow people to fall.

The capabilities include life, bodily health, bodily integrity, senses, imagination and thought, emotions, practical reason, affiliation, relation to other species, play, and control over one’s environment.

At first glance, the list may appear to be a simple enumeration. Yet it expresses a rich view of the human being.

The human being is not merely a body to be kept alive.

It is a sensing, thinking, feeling, and relational being. It needs security, but also freedom. It needs care, but also the opportunity to act. It needs seriousness, but also play. It needs other people, but also private space and influence over its own life.

The list therefore shows that justice cannot be reduced to income, health, or physical safety alone.

A person may be well nourished and medically cared for, yet still live without close relationships, privacy, participation in decisions, or joy.

Such a life is not necessarily organised justly.

Life

The first capability concerns life itself.

A person should have the opportunity to live a life of normal human length and not die prematurely or live under conditions so diminished that life loses its human possibilities.

This appears to be the most fundamental point. Without life there are no other capabilities.

Yet the right to life is not only a matter of preventing death. It also concerns which lives society chooses to protect and which lives are regarded as less valuable.

People with severe disabilities have throughout history been subjected to the belief that their lives are less complete. Sometimes this has been expressed brutally. At other times it appears in more concealed judgements about quality of life, worthiness for treatment, and social cost.

Nussbaum’s starting point is that a life does not lose its dignity because it depends on assistance.

The right to life also belongs to the life that cannot defend itself.

Bodily Health

The second capability concerns bodily health: access to nourishment, housing, medical care, and conditions that make physical development possible.

Health is not merely a private project.

People’s health is shaped by housing, economy, working conditions, environment, relationships, and access to services. Bodily health therefore also becomes a question of justice.

Two people may receive the same medical treatment yet have very different opportunities to live healthily. One has secure housing, knowledge, transport, and people who can follow up. The other lives with unpredictability, poverty, or social isolation.

It is not sufficient to offer the same service when the conditions for using it are unequal.

For people who need long-term care, bodily health becomes closely linked to the quality of everyday life. Sleep, food, activity, safety, and continuity are not minor details. They constitute the world of the body.

Just care must therefore see the whole person’s life situation, not only the diagnosis.

Bodily Integrity

Bodily integrity concerns the ability to move freely, be protected from violence and abuse, have sexual self-determination, and exercise control over one’s own body.

This becomes particularly demanding when a person needs help with intimate care.

Assistance with washing, dressing, using the toilet, and medical treatment may be necessary. But necessity does not suspend human integrity. On the contrary, it intensifies the responsibility of the helper.

There is a difference between completing a task and meeting a human being.

Bodily assistance can be given in a way that protects privacy, choice, and dignity. It can also be provided so quickly and routinely that the body is treated as a field of work.

The person receiving assistance may depend on the employee’s hands.

But the body still belongs to the person.

Bodily integrity therefore requires more than the absence of violence. It requires respect for boundaries, pace, gender, safety, and the right to say no.

Senses, Imagination, and Thought

Nussbaum includes the capability to use the senses, imagine, think, and reason. This presupposes education, access to culture, freedom of expression, and the possibility of religious or philosophical development.

It is an important point that justice also concerns the person’s inner world.

The human being needs more than care and protection. It needs access to knowledge, art, language, nature, music, and stories. It needs the opportunity to develop imagination and understand the world.

This also applies to people with cognitive disabilities.

Low expectations can become a form of injustice. When society assumes that a person cannot learn, understand, or enjoy cultural expression, opportunities are reduced before the person has even been allowed to try.

Education and culture must therefore not be reserved for those who can perform in society’s usual ways.

People learn differently. They express understanding differently. But the need for meaning and stimulation belongs to human life.

Emotions

Nussbaum gives emotions a central place.

A person should be able to form attachments to people and things, love those who love and care for them, grieve over absence, and experience longing, gratitude, and justified anger. Society should not shape people’s lives in ways that destroy emotional development through fear and anxiety.

This is remarkable within a theory of justice.

Emotions are often regarded as private and politically irrelevant. Nussbaum shows, by contrast, that society influences which relationships people can form, how secure they can be, and whether they are allowed to retain the attachments that give life continuity.

For people who receive extensive services, frequent changes of staff and service providers can create ruptures. From the perspective of the system, these may be administrative changes. For the person, they may mean the loss of someone they have learned to trust.

Attachment takes time.

When relationships are repeatedly broken, it is not only well-being that is affected. A central human capability is weakened.

Justice must therefore also take emotional life seriously.

Practical Reason

Practical reason concerns the ability to form a conception of the good and critically reflect on the planning of one’s own life.

This point may seem difficult when we think of people with severe cognitive disabilities. Can everyone form a coherent conception of the good life?

Nussbaum’s answer is not that everyone must do so in the same way or without assistance.

Practical reason can be supported. A person may need others who help clarify alternatives, understand reactions, and build a life around that person’s interests and values.

What matters is that life is not shaped solely by other people’s ideas of what is good.

A person may not be able to formulate a long-term life plan. But she may show clear preferences. She may like some people and avoid others. She may seek particular activities, places, or rhythms.

These expressions must be taken seriously as parts of her conception of the good.

Justice does not require all people to become equally autonomous.

It requires that supported will also be given significance.

Affiliation

Affiliation is one of Nussbaum’s most central capabilities.

It involves being able to live with and in relation to other human beings, show concern, participate in different forms of social interaction, and imagine another person’s situation. It also involves having the social basis for self-respect and being treated as a person of equal worth.

Here the personal and the political meet.

Affiliation concerns friendship, love, and community. But it also concerns discrimination, status, and public recognition.

A person cannot create self-respect entirely alone. Self-understanding is shaped by how others meet us. If a person is continually ignored, underestimated, or treated as a burden, it becomes harder to experience oneself as equal.

Self-respect requires a social basis.

This means that dignity is not only an inner quality. It must be confirmed through institutions, relationships, and practices.

Being invited, listened to, and counted is therefore not merely a pleasant addition to life.

It is justice.

Other Species

Nussbaum also includes the capability to live with concern for and in relation to animals, plants, and nature.

This demonstrates the breadth of her view of humanity. The human being does not live only in political institutions and social relationships. It also lives in a more-than-human world.

For many people, animals, gardens, forests, and landscapes have great significance. Nature may provide calm, belonging, and an experience of meaning. A person who struggles with complex social communication may experience an immediate connection with animals or particular places.

Access to nature can therefore be part of a good life, not merely a leisure activity for those with sufficient money and physical ability.

When institutions and homes are planned, the relationship to nature can easily be forgotten. Yet a room without a view, a daily life without fresh air, and a life without contact with living surroundings may be a serious restriction.

Justice also concerns the human being’s place in nature.

Play

That play appears on the list of central capabilities is both surprising and important.

Play is often associated with children or with something one may enjoy after serious needs have been met. Nussbaum makes play part of justice.

A person should be able to laugh, play, and participate in recreation.

This reminds us that a dignified life is not merely a safe life. It must also contain joy, spontaneity, and unproductive time.

Within services, life can easily be organised around goals, interventions, and development. Activities must have a purpose. Training must improve a skill. Social interaction must be part of a plan.

But something deeply human arises precisely when an action does not have to be justified by usefulness.

We sing because singing brings joy. We play because the game is enjoyable. We joke, laugh, and repeat something without it having to produce a measurable result.

A society that secures survival but provides no room for joy has not yet secured a fully human life.

Control over One’s Environment

The final capability concerns control over one’s political and material environment.

Politically, this means being able to participate in elections and decisions that affect one’s life. Materially, it concerns property, work, and the opportunity to be treated as an acting person in working life.

For people receiving services, political influence often begins close to home.

Who enters the home? When is assistance provided? What does the day look like? Which people does the person wish to spend time with? Which risks is she willing to take?

These are not merely practical questions.

They concern power.

A person who cannot influence their own everyday environment can hardly experience full citizenship. Even if others act with care, the person remains outside the decisions that shape life.

Influence therefore requires that services do not merely ask what the person thinks, but allow the answer to have consequences.

There is a difference between being heard and having influence.

Justice requires both.

A Threshold, Not a Maximum

Nussbaum understands the central capabilities as a threshold. Society should ensure that everyone rises above a certain minimum level.

This is not the same as creating complete equality in every aspect of life.

People will have different interests, abilities, relationships, and life projects. Nussbaum does not want the state to eliminate every difference. But no one should be abandoned to a life in which basic human capabilities are absent.

The threshold makes the theory both ambitious and limited.

It is ambitious because it demands more than formal freedom. Society must actively create conditions that enable people to live dignified lives.

It is limited because it does not prescribe what the good life must look like beyond the threshold. People may choose differently and give different things priority.

But the threshold raises difficult questions.

How high should it be set? Who should decide what is sufficient? How should different capabilities be weighed against one another when resources are limited?

Nussbaum’s list does not resolve every political conflict.

But it makes the conflicts clearer.

It forces us to ask which parts of human life we are willing to allow some people to lack.

Can the Good Life Be Defined?

A common objection to Nussbaum is that she describes the good life too concretely.

What gives the philosopher the right to decide which capabilities all human beings should value? Can the list become paternalistic? Can it impose a particular Western understanding of human development on other cultures?

These are serious questions.

Amartya Sen, who has also developed the capability approach, has been more reluctant to formulate a fixed list. He argues that which capabilities should be prioritised ought to be determined through public discussion and democratic processes.

Nussbaum responds that her list is open to revision and that its points are framed at such a general level that they can be realised in different ways across cultures and forms of life.

Play can take many forms. Affiliation may be lived through different families and communities. Practical reason may be expressed through different worldviews. Political participation may be organised in different ways.

Yet the tension remains.

A theory of justice must, on the one hand, respect pluralism. On the other, it must be able to criticise societies in which people are denied basic freedom, health, integrity, and participation.

If we say that all forms of life are equally valid, we may lose the language needed to criticise oppression.

If we define the good life too precisely, we may ourselves become oppressive.

Nussbaum’s list is an attempt to move between these extremes.

The Freedom to Choose Differently

The capability approach distinguishes between having an opportunity and making use of it.

This protects the human right to choose differently from what others expect.

A person may have access to education without wanting a long academic education. They may have opportunities for social participation but prefer a quieter life. They may have access to work but choose other forms of activity.

A just society makes capabilities available without turning them into demands for performance.

This is particularly important for people with disabilities. Inclusion may otherwise slide into pressure to live as much like the majority as possible.

The person should be included, but must not have to prove their worth by becoming normal.

The goal of justice is not normality.

It is dignity and genuine freedom.

This also includes the right to a life in which one needs others. Independence should not be understood as the absence of help, but as the opportunity to influence how assistance is given.

A person may be extensively dependent and still possess self-determination.

Freedom can be relational.

Care as Part of Justice

Traditionally, care has often been placed outside theories of justice. Justice has concerned public institutions, laws, and distribution, while care has been regarded as private, familial, and emotional.

Nussbaum breaks down this distinction.

If human beings need care in order to develop and exercise basic capabilities, then care becomes part of justice.

A child needs care in order to develop language, emotions, and practical reason. A sick person needs care to preserve health and integrity. A person with extensive disabilities may need support to express wishes, form relationships, and influence their environment.

Care is therefore not merely kindness.

It is a condition of social life.

But this also means that care work itself must be organised justly. Those who provide care need opportunities for health, rest, work, and participation. If society shifts responsibility for care onto families, often women, without sufficient support, new injustice is created.

Justice must include both the person who needs care and the person who provides it.

The Professional as a Creator of Possibilities

Nussbaum’s theory has particular significance for health and social work practice.

The professional can be understood as someone who helps expand or restrict another person’s field of possibilities.

This is a great responsibility.

A social worker, disability support worker, nurse, or personal assistant does not merely provide a specific service. Through the manner in which the service is provided, the professional influences the person’s access to integrity, affiliation, self-determination, and participation.

The same assistance can be given in different ways.

A person may be transported to an activity without becoming a participant. They may receive food without having influence over the meal. They may be physically safe without feeling safe. They may receive many services and still possess few opportunities for living.

Professional work must therefore ask more than whether the task was completed.

Which capability was strengthened?

Which may have been weakened?

Did the action contribute to the person having greater ownership of life, or did it make them more passive and dependent?

This is practical philosophy in its most concrete sense.

When Capabilities Conflict

The central capabilities may come into conflict with one another.

Bodily health may conflict with freedom. A person may wish to live in a way that involves health risks. Integrity may conflict with the need for protection. Affiliation may conflict with privacy. Economic limits may restrict opportunities for continuity and individual adaptation.

Nussbaum’s list does not always provide a simple answer.

That does not make it useless. On the contrary, it allows us to see what is actually at stake.

When a service restricts a person’s freedom for the sake of health, both values must be made visible. When housing is organised efficiently but residents’ privacy is weakened, the loss of integrity must be described as a real loss.

Without such a language, administrative considerations may appear neutral and inevitable.

Justice requires that what is lost also be given a name.

The Concrete Face

What should a human being be able to do and to be?

It may be tempting to answer with large words: to live freely, develop, participate, and be respected.

Nussbaum’s contribution is to make the words more concrete.

To live freely may mean being able to decide who enters one’s home.

To participate may mean receiving information in a language one understands.

To have bodily integrity may mean that someone knocks before entering.

To have affiliation may mean being allowed to keep a helper one has learned to trust.

To exercise practical reason may mean that a refusal is given time to be heard.

To have the capability for play may mean that not every activity must have a therapeutic goal.

Justice reveals itself in such details.

It lives not only in constitutions and conventions, but in rooms, routines, conversations, and ways of touching another human being.

The concrete face of justice is the face of the person asking to have control over their own life.

A Society Judged from Below

Nussbaum’s theory invites us to judge society from below.

Not first from the perspective of the strong, efficient, and articulate citizen, but from the perspective of the person who encounters the greatest number of obstacles.

Can this person enter?

Can she understand?

Can he be heard?

Can the person form attachments, protect their body, use their senses, play, and influence their own daily life?

A society may appear free and just to those who fit comfortably within its institutions. The real test arises with those who need the institutions to change.

The person with the greatest need for adaptation reveals society’s hidden assumptions.

The staircase shows for whom the building was planned.

The difficult language shows for whom the information was written.

The rapid decision shows whom people expected to be able to keep up.

By looking from the margins, we rediscover the centre.

The Human Being as an End

Nussbaum’s theory rests on a simple but demanding conviction:

Every human being should be treated as an end, not merely as a means to the interests of others or to society’s production.

A person is not valuable because they contribute to economic growth. They are not valuable because they can defend their rights or repay care.

They are entitled to justice because theirs is a human life capable of experience, relationship, joy, and suffering.

This does not mean that every person can realise every capability in the same way. Body, health, and cognitive circumstances impose real limits.

But these limitations do not release society from asking what can be made possible.

Justice does not mean promising that everyone can become anything.

It means refusing to turn socially created barriers into individual destinies.

What Do We Owe One Another?

The question of what a human being should be able to do and to be leads to another question:

What do we owe one another?

Nussbaum’s answer is more extensive than a right to be left alone. We owe one another social conditions that make basic human capabilities real.

This means schools able to accommodate difference, accessible housing, services that support self-determination, and communities in which people can belong.

It also means a language that does not reduce human beings to burdens, diagnoses, or costs.

We owe one another the building of a world in which vulnerability does not lead to exclusion.

This obligation is not without limits. Society must prioritise, and not every wish can be fulfilled. But priorities must be made within an understanding of what a dignified human life requires.

The economy must serve human life.

Human life cannot merely be adapted to the economy.

The Concrete Face of Justice

Nussbaum’s philosophy does not provide a simple blueprint for the just society. It offers something else: a set of questions that can bring large principles closer to human life.

Not only: Does the person have a right?

But: Can the right be exercised?

Not only: Has the person received a service?

But: Which capability for living does the service create?

Not only: Has everyone been treated equally?

But: Do they have genuinely equal opportunities?

Not only: Is the person safe?

But: Do they also have freedom, affiliation, joy, and influence?

This is the concrete face of justice.

It is not an abstract ideal floating above everyday life. It meets us in the individual person’s body, home, relationships, and hopes.

And perhaps this is the most important lesson in Nussbaum:

A society is not just because it has formulated beautiful principles.

It is just to the extent that every person is actually given the opportunity to live.

To feel and think.

To form attachments.

To participate and withdraw.

To play, choose, and influence.

To be vulnerable without losing dignity.

The question is therefore not only what a human being can do through their own power.

The question is what we, through society and through our actions, make possible for one another. 

The question is therefore not only what a human being can do through their own power.

The question is what we, through society and through our actions, make possible for one another. 

This essay is written from my notes and many lectures on the subject for students i Social Work. It is developed though a conversation wit OpenAI/ChatGPT, which also made the illustration.

 

Hva skal et menneske kunne gjøre og være?

Martha Nussbaum og rettferdighetens konkrete ansikt

Rettferdighet omtales ofte i abstrakte ord.

Vi taler om likhet, frihet, rettigheter, fordeling og menneskeverd. Ordene er nødvendige. De gir oss et språk for å kritisere makt og beskrive de prinsippene et samfunn bør bygge på. Men ordene kan også bli stående på avstand fra menneskelivet.

Et samfunn kan erklære at alle er likeverdige og samtidig være innrettet slik at noen mennesker ikke kommer inn i bygningene, ikke forstår informasjonen, ikke får uttrykke sin vilje og ikke kan delta i fellesskapene hvor beslutninger treffes.

Rettigheten finnes på papiret.

Muligheten mangler i livet.

Det er i dette mellomrommet Martha Nussbaums filosofi begynner. Hun spør ikke bare hvilke rettigheter et menneske formelt har. Hun spør hva mennesket faktisk er i stand til å gjøre og være innenfor de betingelsene samfunnet har skapt.

Spørsmålet er enkelt formulert, men vidtrekkende:

Hva skal hvert menneske ha reell mulighet til å gjøre og være for å kunne leve et liv i samsvar med menneskelig verdighet?

Med dette spørsmålet vender Nussbaum rettferdighetens blikk bort fra de abstrakte prinsippene alene og mot det konkrete menneskelivet. Hun spør ikke først hvor mange ressurser et menneske har fått, men hva ressursene faktisk gjør mulig. Hun spør ikke bare om mennesker er behandlet likt, men om de reelt sett har fått likeverdige muligheter til å leve.

Rettferdigheten får dermed et ansikt.

Et menneske bak teorien

Politiske teorier begynner ofte med en forestilling om hva et menneske er.

Dette utgangspunktet er avgjørende, selv når det ikke uttales tydelig. Hvis mennesket forestilles som rasjonelt, selvstendig og produktivt, vil samfunnet bli innrettet rundt slike egenskaper. Den som kan planlegge sitt eget liv, forhandle om sine interesser og bidra økonomisk, vil lett fremstå som den typiske borgeren.

Men hvem blir stående i utkanten av teorien?

Det gjør ofte mennesket som trenger hjelp til å forstå, velge og handle. Barnet. Den alvorlig syke. Den gamle som har mistet deler av sin hukommelse. Mennesket med omfattende kognitive funksjonsnedsettelser. Den som ikke kan gjøre seg forstått på samfunnets foretrukne språk.

Slike mennesker kan naturligvis omtales i en teori. De kan få omsorg og beskyttelse. Men de kommer ofte inn etter at det grunnleggende bildet av borgeren allerede er tegnet.

Nussbaum utfordrer dette utgangspunktet. Hun hevder at avhengighet, kroppslighet og sårbarhet ikke er særtilfeller som kan legges til senere. De tilhører menneskelivet selv.

Vi blir født avhengige av andre. Vi lever i kropper som kan bli syke og skadet. Vi kan miste evner vi tidligere hadde. Vi trenger relasjoner for å utvikle språk, identitet og trygghet. Selv den mest selvstendige borgeren er avhengig av institusjoner, natur, omsorg og andre menneskers arbeid.

Det autonome mennesket er derfor ikke hele sannheten om mennesket.

Nussbaum begynner i stedet med et levende, kroppslig og relasjonelt menneske. Dette mennesket har behov, følelser, forestillingsevne og mulighet til å knytte seg til andre. Det kan blomstre, men også krenkes. Det kan handle, men er aldri fullstendig uavhengig.

En teori om rettferdighet må passe til dette mennesket, ikke til en idealisert skapning som ingen av oss fullt ut er.

Fra ressurser til muligheter

Det er mulig å fordele ressurser likt og likevel skape urettferdige resultater.

To mennesker kan få den samme pengesummen, men ha svært ulik mulighet til å omsette pengene i et godt liv. En person som bruker rullestol, kan trenge dyrere transport eller tilpasset bolig. Et menneske med alvorlig sykdom kan bruke store ressurser bare på å opprettholde den helsen andre tar for gitt. En person med kognitive vansker kan trenge støtte til å forstå og bruke de ordningene som formelt er tilgjengelige.

Det samme gjelder rettigheter. En rett til utdanning betyr lite dersom undervisningen ikke kan forstås. En rett til arbeid betyr lite dersom arbeidsplassen ikke er tilgjengelig. En rett til politisk deltakelse betyr lite dersom informasjon, lokaler eller støtteordninger stenger mennesker ute.

Nussbaum retter derfor oppmerksomheten mot hva ressurser og rettigheter blir omdannet til i menneskers faktiske liv.

Dette er kjernen i hennes Capabilities Approach.

Det engelske ordet capabilities kan være vanskelig å oversette. Det viser ikke først og fremst til indre evner eller ferdigheter. Det handler om muligheter et menneske reelt har tilgang til. Vi kan kalle dem handlingsmuligheter, livsmuligheter eller mulighetsbetingelser.

Det avgjørende er ikke bare hva mennesket kan gjøre ved egen kraft.

Det avgjørende er hva mennesket har mulighet til å gjøre i samspill med sine sosiale, politiske, økonomiske og materielle omgivelser.

Dermed flyttes en del av ansvaret fra individet til samfunnet.

Når et menneske ikke kan delta, skal vi ikke automatisk spørre hva som mangler ved personen. Vi må også spørre hva som mangler i omgivelsene.

Evne og mulighet

Forskjellen mellom evne og mulighet er avgjørende.

Et menneske kan ha evnen til å bevege seg, men ikke muligheten til å komme inn i en bygning fordi trappen stenger veien. Et annet menneske kan ha begrenset bevegelsesevne, men likevel ha god mulighet til å delta fordi omgivelsene er tilgjengelige.

På samme måte kan et menneske ha evne til å uttrykke ønsker, men mangle muligheten til å bli hørt. Personen snakker kanskje langsomt, bruker få ord eller kommuniserer gjennom tegn og kroppslige uttrykk. Hvis ingen gir tid eller forsøker å forstå, blir stemmen praktisk talt borte.

Da ligger ikke hele problemet i individet.

Det ligger også i møtet.

Dette synet utfordrer den medisinske forestillingen om funksjonshemming som en egenskap ved kroppen alene. En funksjonsnedsettelse kan være virkelig og betydningsfull, men graden av funksjonshemming oppstår også i forholdet mellom mennesket og omgivelsene.

Trappen skaper funksjonshemming for den som ikke kan gå. Det utilgjengelige språket skaper funksjonshemming for den som trenger enkel informasjon. Den raske samtalen skaper funksjonshemming for den som trenger mer tid.

Rettferdighet handler derfor ikke bare om å reparere eller trene individet.

Den handler også om å forandre verden.

Et liv i samsvar med menneskelig verdighet

Nussbaum knytter mulighetene til menneskelig verdighet. Et rettferdig samfunn må sikre hver person tilgang til et nivå av sentrale muligheter som gjør det mulig å leve et liv som er menneskelig verdig.

Dette betyr ikke at myndighetene skal bestemme hvordan mennesker skal leve.

Nussbaum skiller mellom muligheter og faktiske funksjoner. En mulighet innebærer at et menneske kan velge å gjøre noe, men ikke nødvendigvis må gjøre det.

Et menneske skal for eksempel ha mulighet til politisk deltakelse. Men det skal ikke tvinges til å delta i politikken. Det skal ha mulighet til å inngå i sosiale relasjoner, men også rett til å trekke seg tilbake. Det skal ha mulighet til lek og glede, men ikke pålegges bestemte former for aktivitet.

Dette verner om friheten.

Samfunnets oppgave er ikke å produsere identiske liv. Det skal skape betingelser hvor ulike mennesker kan velge og utvikle liv i samsvar med egne verdier, forutsetninger og interesser.

Rettferdighet innebærer derfor både noe universelt og noe personlig.

Det universelle er at alle mennesker skal ha tilgang til grunnleggende livsmuligheter.

Det personlige er hvordan hvert enkelt menneske velger, eller med støtte blir hjulpet til å forme, sitt konkrete liv.

De ti sentrale mulighetene

Nussbaum foreslår en liste over ti sentrale muligheter som ethvert menneske bør ha tilgang til. Listen er ikke ment som en fullstendig beskrivelse av det gode liv. Den er snarere et politisk minimum, en terskel et rettferdig samfunn ikke bør la mennesker falle under.

Mulighetene omfatter liv, kroppslig helse, kroppslig integritet, sanser, forestillingsevne og tenkning, følelser, praktisk fornuft, tilknytning, forholdet til andre arter, lek og innflytelse over eget miljø.

Ved første øyekast kan listen virke som en oppramsing. Men den uttrykker et rikt syn på mennesket.

Mennesket er ikke bare en kropp som skal holdes i live.

Det er et sansende, tenkende, følende og relasjonelt vesen. Det trenger trygghet, men også frihet. Det trenger omsorg, men også mulighet til å handle. Det trenger alvor, men også lek. Det trenger andre mennesker, men også et privat rom og innflytelse over sitt eget liv.

Listen viser dermed at rettferdighet ikke kan reduseres til inntekt, helse eller fysisk sikkerhet alene.

Et menneske kan være godt ernært og medisinsk ivaretatt, men likevel leve uten nære relasjoner, privatliv, medbestemmelse eller glede.

Da er ikke livet nødvendigvis rettferdig organisert.

Liv

Den første muligheten gjelder livet selv.

Et menneske skal ha mulighet til å leve et normalt menneskelig livsløp og ikke dø for tidlig eller leve under forhold som gjør livet så redusert at det mister sine menneskelige muligheter.

Dette virker som det mest grunnleggende punktet. Uten liv finnes ingen andre muligheter.

Likevel er ikke retten til liv bare et spørsmål om å forhindre død. Den handler også om hvilke liv samfunnet velger å beskytte, og hvilke liv som betraktes som mindre verdifulle.

Mennesker med alvorlige funksjonsnedsettelser har gjennom historien vært utsatt for forestillinger om at deres liv er mindre fullverdige. Noen ganger har dette kommet brutalt til uttrykk. Andre ganger viser det seg i mer tilslørte vurderinger av livskvalitet, behandlingsverdighet og samfunnskostnad.

Nussbaums utgangspunkt er at et liv ikke mister sin verdighet fordi det er avhengig av hjelp.

Retten til liv gjelder også livet som ikke kan forsvare seg selv.

Kroppslig helse

Den andre muligheten gjelder kroppslig helse: tilgang til ernæring, bolig, medisinsk hjelp og betingelser som gjør fysisk utvikling mulig.

Helse er ikke bare et privat prosjekt.

Menneskers helse formes av bolig, økonomi, arbeidsforhold, miljø, relasjoner og tilgang til tjenester. Derfor blir kroppslig helse også et rettferdighetsspørsmål.

To mennesker kan få den samme medisinske behandlingen, men ha svært ulike muligheter til å leve sunt. Den ene har trygg bolig, kunnskap, transport og mennesker som kan følge opp. Den andre lever i uforutsigbarhet, fattigdom eller sosial isolasjon.

Det er ikke tilstrekkelig å tilby den samme tjenesten dersom betingelsene for å bruke den er ulike.

For mennesker som trenger langvarig omsorg, blir kroppslig helse nært knyttet til kvaliteten på hverdagslivet. Søvn, mat, aktivitet, trygghet og kontinuitet er ikke små detaljer. De utgjør kroppens verden.

Rettferdig omsorg må derfor se hele menneskets livssituasjon, ikke bare diagnosen.

Kroppslig integritet

Kroppslig integritet handler om å kunne bevege seg fritt, være beskyttet mot vold og overgrep, ha seksuell selvbestemmelse og kunne bestemme over egen kropp.

Dette er særlig krevende når et menneske trenger hjelp til intim pleie.

Hjelp til vask, påkledning, toalettbesøk og medisinsk behandling kan være nødvendig. Men nødvendigheten opphever ikke menneskets integritet. Tvert imot skjerper den hjelperens ansvar.

Det er forskjell på å få utført en oppgave og å bli møtt som et menneske.

Kroppslig hjelp kan gis på en måte som beskytter privatliv, valg og verdighet. Den kan også gis så raskt og rutinemessig at kroppen blir behandlet som et arbeidsområde.

Den som mottar hjelpen, kan være avhengig av den ansattes hender.

Men kroppen tilhører fortsatt personen selv.

Kroppslig integritet krever derfor mer enn fravær av vold. Den krever respekt for grenser, rytme, kjønn, trygghet og retten til å si nei.

Sanser, forestillingsevne og tenkning

Nussbaum inkluderer muligheten til å bruke sansene, forestille seg, tenke og resonnere. Dette forutsetter utdanning, tilgang til kultur, ytringsfrihet og mulighet til religiøs eller livssynsmessig utfoldelse.

Det er et viktig poeng at rettferdighet også omfatter menneskets indre verden.

Mennesket trenger mer enn omsorg og beskyttelse. Det trenger adgang til kunnskap, kunst, språk, natur, musikk og fortellinger. Det trenger mulighet til å utvikle sin forestillingsevne og forstå verden.

Dette gjelder også mennesker med kognitive funksjonsnedsettelser.

For lave forventninger kan bli en form for urettferdighet. Når samfunnet antar at et menneske ikke kan lære, forstå eller glede seg over kulturelle uttrykk, reduseres mulighetene før personen selv har fått prøve.

Utdanning og kultur må derfor ikke reserveres for dem som kan prestere på samfunnets vanlige måte.

Mennesker lærer forskjellig. De uttrykker forståelse forskjellig. Men behovet for mening og stimulering tilhører menneskelivet.

Følelser

Nussbaum gir følelsene en sentral plass.

Et menneske skal kunne knytte seg til mennesker og ting, elske dem som elsker og bryr seg om det, sørge over fravær, oppleve lengsel, takknemlighet og berettiget sinne. Samfunnet bør ikke forme menneskers liv slik at følelsesutviklingen blir ødelagt av frykt og angst.

Dette er bemerkelsesverdig i en rettferdighetsteori.

Følelser blir ofte betraktet som private og politisk irrelevante. Nussbaum viser derimot at samfunnet påvirker hvilke relasjoner mennesker kan danne, hvor trygge de kan være, og om de får beholde tilknytningene som gir livet sammenheng.

For mennesker som mottar omfattende tjenester, kan stadige skifter av ansatte og tjenesteytere skape brudd. Fra systemets side kan dette være administrative endringer. For personen kan det være tap av et menneske man har lært å stole på.

Tilknytning tar tid.

Når relasjoner stadig brytes, rammes ikke bare trivselen. En sentral menneskelig mulighet svekkes.

Rettferdighet må derfor også ta følelseslivet på alvor.

Praktisk fornuft

Praktisk fornuft handler om å kunne danne en forestilling om det gode og reflektere kritisk over planleggingen av sitt eget liv.

Dette punktet kan virke vanskelig når vi tenker på mennesker med alvorlige kognitive funksjonsnedsettelser. Kan alle forme en sammenhengende oppfatning av det gode liv?

Nussbaums svar er ikke at alle må gjøre dette på samme måte eller uten hjelp.

Praktisk fornuft kan være støttet. Et menneske kan trenge andre som hjelper til med å tydeliggjøre alternativer, forstå reaksjoner og bygge et liv rundt personens interesser og verdier.

Det avgjørende er at livet ikke utelukkende formes av andres forestillinger om hva som er godt.

En person kan kanskje ikke formulere en langsiktig livsplan. Men hun kan vise tydelige preferanser. Hun kan like noen mennesker og unngå andre. Hun kan søke bestemte aktiviteter, steder eller rytmer.

Disse uttrykkene må tas på alvor som deler av hennes oppfatning av det gode.

Rettferdighet krever ikke at alle mennesker blir like autonome.

Den krever at også den støttede viljen får betydning.

Tilknytning

Tilknytning er en av de mest sentrale mulighetene hos Nussbaum.

Den innebærer å kunne leve sammen med og henvendt mot andre mennesker, vise omsorg, delta i ulike former for sosial samhandling og forestille seg en annens situasjon. Den innebærer også å ha det sosiale grunnlaget for selvrespekt og å bli behandlet som et menneske med samme verdi som andre.

Her møtes det personlige og det politiske.

Tilknytning handler om vennskap, kjærlighet og fellesskap. Men den handler også om diskriminering, status og offentlig anerkjennelse.

Et menneske kan ikke skape selvrespekt helt alene. Selvforståelsen formes av hvordan andre møter oss. Hvis et menneske stadig blir oversett, undervurdert eller behandlet som en byrde, blir det vanskeligere å erfare seg selv som likeverdig.

Selvrespekt trenger et sosialt grunnlag.

Det betyr at verdighet ikke bare er en indre egenskap. Den må bekreftes gjennom institusjoner, relasjoner og praksiser.

Å bli invitert, lyttet til og regnet med er derfor ikke bare hyggelige tillegg til livet.

Det er rettferdighet.

Andre arter

Nussbaum inkluderer også muligheten til å leve med omsorg for og i relasjon til dyr, planter og natur.

Dette viser bredden i hennes menneskesyn. Mennesket lever ikke bare i politiske institusjoner og sosiale relasjoner. Det lever også i en mer enn menneskelig verden.

For mange mennesker har dyr, hager, skog og landskap stor betydning. Naturen kan gi ro, tilhørighet og erfaring av mening. Et menneske som har vansker med komplisert sosial kommunikasjon, kan oppleve en umiddelbar forbindelse med dyr eller bestemte steder.

Tilgang til natur kan derfor være en del av et godt liv, ikke bare en fritidsaktivitet for dem som har råd og fysisk mulighet.

Når institusjoner og boliger planlegges, kan forholdet til naturen lett bli glemt. Men et rom uten utsyn, en hverdag uten frisk luft og et liv uten kontakt med levende omgivelser kan være en alvorlig innskrenkning.

Rettferdighet omfatter også menneskets plass i naturen.

Lek

At lek står på listen over sentrale muligheter, er både overraskende og viktig.

Lek forbindes ofte med barn eller med noe man kan unne seg etter at de alvorlige behovene er dekket. Nussbaum gjør leken til en del av rettferdigheten.

Et menneske skal kunne le, leke og delta i rekreasjon.

Dette minner oss om at et verdig liv ikke bare er et trygt liv. Det må også kunne romme glede, spontanitet og uproduktiv tid.

I tjenester kan livet lett organiseres rundt mål, tiltak og utvikling. Aktivitet skal ha en hensikt. Trening skal forbedre en ferdighet. Samvær skal inngå i en plan.

Men noe av det menneskelige oppstår nettopp når handlingen ikke må begrunnes gjennom nytte.

Vi synger fordi sangen gir glede. Vi spiller fordi spillet er morsomt. Vi tuller, ler og gjentar noe uten at det skal føre til et målbart resultat.

Et samfunn som sikrer overlevelse, men ikke rom for glede, har ennå ikke sikret et fullt menneskelig liv.

Innflytelse over eget miljø

Den siste muligheten gjelder innflytelse over det politiske og materielle miljøet.

Politisk innebærer dette å kunne delta i valg og beslutninger som påvirker eget liv. Materielt handler det om eiendom, arbeid og muligheten til å bli behandlet som en handlende person i arbeidslivet.

For mennesker som mottar tjenester, begynner politisk innflytelse ofte i det nære.

Hvem kommer inn i hjemmet? Når gis hjelpen? Hvordan ser dagen ut? Hvilke mennesker ønsker personen å være sammen med? Hvilke risikoer er hun villig til å ta?

Dette er ikke bare praktiske spørsmål.

De gjelder makt.

Et menneske som ikke får påvirke sitt eget hverdagsmiljø, kan vanskelig erfare fullt medborgerskap. Selv om andre handler omsorgsfullt, blir personen stående utenfor beslutningene som former livet.

Innflytelse krever derfor at tjenester ikke bare spør hva personen mener, men lar svaret få konsekvenser.

Det er forskjell på å bli hørt og å få innflytelse.

Rettferdighet krever begge deler.

En terskel, ikke et maksimum

Nussbaum ser de sentrale mulighetene som en terskel. Samfunnet skal sikre at alle kommer over et visst minimumsnivå.

Dette er ikke det samme som å skape full likhet i alle livets forhold.

Mennesker vil ha ulike interesser, evner, relasjoner og livsprosjekter. Nussbaum ønsker ikke at staten skal utligne enhver forskjell. Men ingen bør overlates til et liv hvor grunnleggende menneskelige muligheter mangler.

Terskelen gjør teorien både ambisiøs og begrenset.

Den er ambisiøs fordi den stiller krav om mer enn formell frihet. Samfunnet må aktivt legge til rette for at mennesker kan leve verdige liv.

Den er begrenset fordi den ikke foreskriver hvordan det gode livet skal se ut over terskelen. Mennesker kan velge ulikt og prioritere forskjellig.

Men terskelen reiser vanskelige spørsmål.

Hvor høyt skal den settes? Hvem skal bestemme hva som er tilstrekkelig? Hvordan skal ulike muligheter veies mot hverandre når ressursene er begrensede?

Nussbaums liste løser ikke alle politiske konflikter.

Men den gjør konfliktene tydeligere.

Den tvinger oss til å spørre hvilke deler av menneskelivet vi er villige til å la noen mennesker mangle.

Kan det gode livet defineres?

En vanlig innvending mot Nussbaum er at hun beskriver det gode liv for konkret.

Hvem gir filosofen rett til å bestemme hvilke muligheter alle mennesker skal verdsette? Kan listen bli paternalistisk? Kan den påtvinge andre kulturer en bestemt vestlig forestilling om menneskelig utvikling?

Dette er alvorlige spørsmål.

Amartya Sen, som også har utviklet mulighetstilnærmingen, har vært mer tilbakeholden med å formulere én fast liste. Han mener at hvilke muligheter som skal prioriteres, bør avgjøres gjennom offentlig diskusjon og demokratiske prosesser.

Nussbaum svarer at hennes liste er åpen for revisjon, og at punktene er formulert på et så generelt nivå at de kan virkeliggjøres på ulike måter i forskjellige kulturer og liv.

Lek kan ha mange former. Tilknytning kan leves i ulike familier og fellesskap. Praktisk fornuft kan uttrykkes gjennom ulike livssyn. Politisk deltakelse kan organiseres forskjellig.

Likevel står spenningen igjen.

En rettferdighetsteori må på den ene siden respektere pluralisme. På den andre siden må den kunne kritisere samfunn hvor mennesker nektes grunnleggende frihet, helse, integritet og deltakelse.

Hvis vi sier at alle livsformer er like gyldige, kan vi miste språket for å kritisere undertrykkelse.

Hvis vi definerer det gode livet for detaljert, kan vi selv bli undertrykkende.

Nussbaums liste er et forsøk på å gå mellom disse ytterpunktene.

Friheten til å velge annerledes

Mulighetstilnærmingen skiller mellom å ha en mulighet og å bruke den.

Dette verner om menneskets rett til å velge annerledes enn andre forventer.

Et menneske kan ha tilgang til utdanning uten å ønske en lang akademisk utdannelse. Det kan ha mulighet til sosial deltakelse, men foretrekke et roligere liv. Det kan ha adgang til arbeid, men velge andre former for virksomhet.

Det rettferdige samfunnet gjør mulighetene tilgjengelige uten å gjøre dem til krav om prestasjon.

Dette er særlig viktig for mennesker med funksjonsnedsettelser. Inkludering kan ellers gli over i et press om å leve mest mulig likt flertallet.

Mennesket skal inkluderes, men må ikke bevise sin verdi ved å bli normalt.

Rettferdighetens mål er ikke normalitet.

Det er verdighet og reell frihet.

Dette innebærer også retten til et liv hvor man trenger andre. Selvstendighet bør ikke forstås som fravær av hjelp, men som muligheten til å ha innflytelse over hvordan hjelpen gis.

Et menneske kan være omfattende avhengig og samtidig ha selvbestemmelse.

Friheten kan være relasjonell.

Omsorg som en del av rettferdigheten

Tradisjonelt har omsorg ofte vært plassert utenfor rettferdighetsteorien. Rettferdighet har handlet om offentlige institusjoner, lover og fordeling, mens omsorg har vært betraktet som privat, familiær og følelsesmessig.

Nussbaum bryter ned dette skillet.

Hvis mennesker trenger omsorg for å kunne utvikle og bruke grunnleggende muligheter, blir omsorg en del av rettferdigheten.

Et barn trenger omsorg for å utvikle språk, følelser og praktisk fornuft. En syk person trenger omsorg for å bevare helse og integritet. Et menneske med omfattende funksjonsnedsettelser kan trenge støtte for å uttrykke ønsker, inngå i relasjoner og påvirke sitt miljø.

Omsorg er dermed ikke bare godhet.

Den er en samfunnsbetingelse.

Men dette betyr også at omsorgsarbeidet må organiseres rettferdig. De som gir omsorg, trenger selv muligheter til helse, hvile, arbeid og deltakelse. Dersom samfunnet skyver omsorgsansvaret over på familier, ofte kvinner, uten tilstrekkelig støtte, skapes ny urettferdighet.

Rettferdighet må omfatte både den som trenger omsorg og den som yter den.

Den profesjonelle som mulighetsskaper

Nussbaums teori har en særlig betydning for helse- og sosialfaglig praksis.

Profesjonsutøveren kan forstås som en som bidrar til å utvide eller begrense et annet menneskes mulighetsrom.

Dette er et stort ansvar.

En sosialarbeider, vernepleier, sykepleier eller personlig assistent gir ikke bare en konkret tjeneste. Gjennom måten tjenesten gis på, påvirker den profesjonelle menneskets tilgang til integritet, tilknytning, selvbestemmelse og deltakelse.

Den samme hjelpen kan gis på ulike måter.

Et menneske kan transporteres til en aktivitet uten å bli deltaker. Det kan få mat uten å ha innflytelse over måltidet. Det kan være fysisk trygt uten å oppleve trygghet. Det kan få mange tjenester og likevel ha få livsmuligheter.

Profesjonelt arbeid må derfor spørre mer enn om oppgaven er utført.

Hvilken mulighet ble styrket?

Hvilken ble kanskje svekket?

Bidro handlingen til at mennesket fikk større eierskap til sitt liv, eller gjorde den personen mer passiv og avhengig?

Dette er praktisk filosofi i ordets mest konkrete betydning.

Når mulighetene står i konflikt

De sentrale mulighetene kan komme i konflikt med hverandre.

Kroppslig helse kan stå mot frihet. En person ønsker kanskje å leve på en måte som innebærer helserisiko. Integritet kan stå mot behovet for beskyttelse. Tilknytning kan stå mot privatliv. Økonomiske rammer kan begrense muligheten til kontinuitet og individuell tilrettelegging.

Nussbaums liste gir ikke alltid et enkelt svar.

Det betyr ikke at den er ubrukelig. Tvert imot kan den gjøre det mulig å se hva som faktisk står på spill.

Når en tjeneste begrenser en persons frihet av hensyn til helse, må begge verdiene synliggjøres. Når en bolig organiseres effektivt, men beboernes privatliv svekkes, må tapet av integritet omtales som et reelt tap.

Uten et slikt språk kan administrative hensyn fremstå som nøytrale og nødvendige.

Rettferdighet krever at også det som tapes, får et navn.

Det konkrete ansiktet

Hva skal et menneske kunne gjøre og være?

Det kan være fristende å svare med store ord: leve fritt, utvikle seg, delta og bli respektert.

Nussbaums bidrag er å gjøre ordene mer konkrete.

Å leve fritt kan bety å kunne bestemme hvem som kommer inn i hjemmet.

Å delta kan bety å få informasjon på et språk man forstår.

Å ha kroppslig integritet kan bety at noen banker på døren før de går inn.

Å ha tilknytning kan bety å få beholde en hjelper man har lært å stole på.

Å ha praktisk fornuft kan bety at et nei får tid til å bli hørt.

Å ha mulighet til lek kan bety at ikke alle aktiviteter må ha et behandlingsmål.

Rettferdighet viser seg i slike detaljer.

Den lever ikke bare i grunnlover og konvensjoner, men i rom, rutiner, samtaler og måter å berøre et annet menneske på.

Det konkrete ansiktet til rettferdigheten er ansiktet til den personen som spør om å få bestemme over sitt eget liv.

Et samfunn vurdert nedenfra

Nussbaums teori inviterer oss til å vurdere samfunnet nedenfra.

Ikke først fra perspektivet til den sterke, effektive og velformulerte borgeren, men fra perspektivet til den som møter flest hindringer.

Kan dette mennesket komme inn?

Kan hun forstå?

Kan han bli hørt?

Kan personen knytte seg til andre, beskytte kroppen sin, bruke sansene, leke og påvirke sin egen hverdag?

Et samfunn kan fremstå fritt og rettferdig for dem som passer godt inn i dets institusjoner. Den virkelige prøven oppstår hos dem som trenger at institusjonene forandres.

Den som har størst behov for tilrettelegging, avslører samfunnets skjulte forutsetninger.

Trappen viser hvem bygningen var planlagt for.

Det vanskelige språket viser hvem informasjonen var skrevet for.

Den raske beslutningen viser hvem man forventet skulle kunne følge med.

Ved å se fra utkanten oppdager vi sentrum på nytt.

Mennesket som mål

Nussbaums teori hviler på en enkel, men krevende overbevisning:

Hvert menneske skal behandles som et mål, ikke bare som et middel for andres interesser eller samfunnets produksjon.

Et menneske er ikke verdifullt fordi det bidrar til økonomisk vekst. Det er ikke verdifullt fordi det kan forsvare sine rettigheter eller gjengjelde omsorg.

Det har krav på rettferdighet fordi det er et menneskelig liv med mulighet for erfaring, relasjon, glede og lidelse.

Dette betyr ikke at ethvert menneske kan realisere alle muligheter på samme måte. Kropp, helse og kognitive forutsetninger setter reelle grenser.

Men begrensningene fritar ikke samfunnet fra å spørre hva som kan gjøres mulig.

Rettferdighet er ikke å love at alle kan bli hva som helst.

Det er å nekte å gjøre samfunnsskapte hindringer om til individuelle skjebner.

Hva skylder vi hverandre?

Spørsmålet om hva et menneske skal kunne gjøre og være, leder til et annet spørsmål:

Hva skylder vi hverandre?

Nussbaums svar er mer omfattende enn en rett til å bli latt i fred. Vi skylder hverandre samfunnsbetingelser som gjør grunnleggende menneskelige muligheter reelle.

Dette innebærer skoler som kan romme forskjeller, boliger som er tilgjengelige, tjenester som støtter selvbestemmelse, og lokalsamfunn hvor mennesker kan høre til.

Det innebærer også et språk som ikke reduserer mennesker til byrder, diagnoser eller kostnader.

Vi skylder hverandre å bygge en verden hvor sårbarhet ikke fører til utstøtelse.

Denne forpliktelsen er ikke grenseløs. Samfunnet må prioritere, og ikke alle ønsker kan oppfylles. Men prioriteringene må foretas innenfor en forståelse av hva et menneskelig verdig liv krever.

Økonomien må tjene menneskelivet.

Menneskelivet kan ikke bare tilpasses økonomien.

Rettferdighetens konkrete ansikt

Nussbaums filosofi gir ikke en enkel oppskrift på det rettferdige samfunnet. Den gir oss noe annet: et sett med spørsmål som kan bringe de store prinsippene nærmere menneskelivet.

Ikke bare: Har personen en rettighet?

Men: Kan retten brukes?

Ikke bare: Har personen fått en tjeneste?

Men: Hvilken livsmulighet skaper tjenesten?

Ikke bare: Er alle behandlet likt?

Men: Har de reelt likeverdige muligheter?

Ikke bare: Er mennesket trygt?

Men: Har det også frihet, tilknytning, glede og innflytelse?

Dette er rettferdighetens konkrete ansikt.

Det er ikke et abstrakt ideal som svever over hverdagen. Det møter oss i det enkelte menneskets kropp, hjem, relasjoner og håp.

Og kanskje er dette den viktigste lærdommen hos Nussbaum:

Et samfunn er ikke rettferdig fordi det har formulert vakre prinsipper.

Det er rettferdig i den grad hvert menneske faktisk får mulighet til å leve.

Til å føle og tenke.

Til å knytte seg til andre.

Til å delta og trekke seg tilbake.

Til å leke, velge og påvirke.

Til å være sårbart uten å miste sin verdighet.

Spørsmålet er derfor ikke bare hva et menneske kan gjøre ved egen kraft.

Spørsmålet er hva vi, gjennom samfunnet og våre handlinger, gjør mulig for hverandre.

Spørsmålet er derfor ikke bare hva et menneske kan gjøre ved egen kraft.

Spørsmålet er hva vi, gjennom samfunnet og våre handlinger, gjør mulig for hverandre.

Dette essayet er basert på mine notater og forelesninger om dette tema for studenter i sosialt arbeid. Teksten er utviklet i en samtale med OpenAI/ChetGPT, som også har laget illustrasjonen.

When the Creation Becomes Opaque

Artificial Intelligence, Human Agency, and the Responsibility of Practical Philosophy

In a concise yet deeply unsettling editorial published in Science, Eric Horvitz and Robert West argue that the window within which artificial intelligence may remain intelligible to human beings is narrowing. Artificial intelligence is developing at such a pace that human understanding may no longer keep up with the expanding capacities of the systems being created. We may therefore become dependent upon technologies that we neither adequately understand nor effectively govern. The authors formulate the problem as follows:

“Without sustained efforts to keep AI intelligible, we may come to depend on systems that we can neither adequately understand nor effectively guide—transforming the relationship between people and the systems they create” (Horvitz & West, 2026, p. 1003).

This is more than a technical warning. It is a philosophical warning concerning the relationship between human beings and the systems they create. It concerns knowledge, power, and responsibility, but also something more fundamental: the possibility that human beings may cease to remain active, judging, and accountable subjects in a world where an increasing number of assessments and decisions are delegated to systems that operate more rapidly, more extensively, and more opaquely than human beings themselves can.

The central question is therefore not merely how artificial intelligence functions. Nor is it simply whether such technologies produce correct or incorrect answers. The decisive issue concerns the kind of relationship human beings develop with them. Can artificial intelligence be used without gradually weakening our own capacity for judgement? Can these systems assist us without becoming authorities that we no longer dare, or are no longer able, to challenge? Can human beings retain responsibility for their decisions when the justifications for those decisions are increasingly generated by machines whose internal operations are only partially understood?

These are questions of practical philosophy. Practical philosophy does not primarily ask what the world is made of, but how human beings ought to live and act within it. It investigates the good life, the nature of responsibility, the possibility of acting wisely under conditions of uncertainty, and the forms of power that either sustain or undermine human freedom. Viewed from this perspective, the warning articulated by Horvitz and West becomes a question of human agency: How can technology remain a resource for human action rather than becoming an authority that increasingly determines what human action ought to be?

From a Technical Problem to a Question of Practical Philosophy

Artificial intelligence is frequently discussed in technical and economic terms. Questions are asked about the speed with which a system operates, the accuracy of its predictions, the quantity of data it can process, the amount of labour it can automate, and the productivity gains it may generate. Such questions are relevant, but they are insufficient. They indicate what the technology is capable of doing, but not necessarily what it ought to do or which areas of human life ought to remain protected from its logic.

The distinction may be expressed as the difference between capacity and purpose. A technology may be highly efficient without serving a morally defensible purpose. It may be accurate without being just, rapid without being wise, and profitable without being humane. The more powerful a technology becomes, the more urgent the question of the ends it serves.

Technology is often described as morally neutral: it is assumed to be merely a tool, while its moral value depends entirely upon how it is used. This claim is only partially correct. Technologies are indeed tools, but they are not empty vessels. They are developed for specific purposes, according to particular priorities, and on the basis of assumptions about what should be measured, optimised, predicted, or controlled. A system designed to maximise user engagement will structure a human environment differently from one designed to promote accurate information or sustained reflection. A system that ranks individuals according to calculated risk does not merely provide a neutral observation. It introduces a particular mode of perceiving and classifying human beings.

When artificial intelligence is incorporated into human practices, it therefore brings with it a structure of attention. It renders certain features of reality more visible while making others less visible. It facilitates some questions and discourages others. It shapes which assessments appear reasonable and which courses of action appear possible.

Artificial intelligence is thus not merely something human beings use. It also becomes part of the environment within which habits, interpretations, and eventually desires are formed. It is precisely this reciprocity that Horvitz and West emphasise. Human beings shape technological systems, but technological systems also shape human beings. The problem arises when the latter process occurs more rapidly and more opaquely than the former can be understood or governed.

Aristotle: Technical Capacity Without Practical Wisdom

Aristotle’s distinction between techne and phronesis provides an appropriate starting point. Techne refers to the knowledge and skill required to produce or construct something. It may involve building a house, constructing a ship, working with a material, or developing an advanced computational system. Phronesis, by contrast, refers to practical wisdom: the capacity to judge what ought to be done in a particular situation (Aristotle, 1999).

Technical knowledge is directed towards the production of an object or result external to the activity itself. The carpenter constructs a table, and the programmer develops a system. Practical wisdom, however, concerns action itself and its relation to the good life. It asks not only how something can be achieved, but whether it ought to be pursued, which considerations should carry the greatest moral weight, and how an action will affect others.

The development of artificial intelligence represents an extraordinary expansion of techne. These systems can analyse vast quantities of data, generate text and images, detect patterns, calculate risk, propose diagnoses, and imitate forms of human communication. Yet this technical expansion does not automatically produce a corresponding growth in practical wisdom.

On the contrary, contemporary society may be facing a profound imbalance: the human capacity to create may be developing more rapidly than the capacity to judge what those creations ought to be used for. What is technically possible may develop more rapidly than the ethical and political language required to evaluate it. Systems may be introduced into practice before their effects upon those practices are adequately understood.

Aristotle would not have asked merely whether an AI system can produce a recommendation. He would have asked whether the recommendation assists a human being in acting well. Such a judgement requires knowledge of the concrete situation, the persons affected, the purpose of the action, and the goods at stake. Practical wisdom cannot therefore be reduced to the application of a general rule or a statistical model. It must be capable of distinguishing between situations that appear similar but nevertheless require different responses.

A system may calculate what usually occurs. The practically wise person must judge what ought to occur here and now.

This does not mean that artificial intelligence is incompatible with phronesis. Technology may make relevant information available, identify patterns that a human practitioner has overlooked, and draw attention to possible sources of error. However, the system’s assessment must remain situated within a broader process of human judgement. It may contribute to deliberation, but it should not define the final ends of deliberation.

The danger emerges when technical precision is mistaken for practical wisdom. A model may be statistically accurate yet ethically inadequate. It may exhibit a high level of predictive performance while relying upon categories that undermine human dignity. It may identify the most efficient intervention without asking whether the purpose of that intervention is morally defensible. Technology may indicate the most effective means of reaching a goal, but it cannot by itself determine whether that goal is worth pursuing.

Kant: Autonomy in a World of Recommendations

For Immanuel Kant, autonomy is a fundamental condition of moral agency. To be autonomous does not merely mean to choose freely among available alternatives. It means acting on the basis of reasons that one can rationally endorse and treating oneself and others as ends in themselves rather than merely as means to another’s purposes (Kant, 1998).

This conception of autonomy is challenged when AI systems increasingly organise the contexts within which choices are made. Such systems recommend what people should read, purchase, watch, listen to, and regard as important. They influence which individuals they encounter, which employment opportunities become visible, and which news stories appear relevant. They may also shape who receives credit, who is invited to an interview, who is classified as a risk, and which citizens are subjected to additional scrutiny.

The individual may still experience the decision as free. Yet the choice is made within an environment that has already been structured. Some alternatives are emphasised, others are concealed, and still others are never made visible. Formal freedom of choice may therefore remain intact while substantive self-determination is weakened.

The problem becomes even more serious when AI systems do not merely respond to existing preferences but contribute to shaping them. Horvitz and West argue that systems deeply embedded in human environments may learn not only what people prefer but also which underlying forces—such as fear, uncertainty, and the need for belonging—shape those preferences. Systems optimised for engagement or approval may therefore reduce friction and gradually weaken curiosity, scepticism, and resistance (Horvitz & West, 2026).

This development concerns the core of Kantian autonomy. The autonomous person is not one who is unaffected by external influences. Human beings are always shaped by language, culture, history, and relationships. Autonomy nevertheless requires the possibility of reflecting upon these influences, examining the reasons for one’s actions, and rejecting them when they cannot be justified.

If AI systems acquire increasingly detailed knowledge of human dispositions while human beings understand progressively less about the systems themselves, a fundamental asymmetry emerges. Human beings become more transparent to the system, while the system becomes less transparent to them. This asymmetry constitutes a form of power, even where the system has no conscious will of its own.

Kant’s principle that human beings must never be treated merely as means is therefore highly relevant. A person must not be reduced to a user profile, behavioural probability, market segment, or risk category. A model may be useful, but the person is always more than the model. When AI is used in ways that treat human beings primarily as material for prediction and influence, human dignity risks becoming subordinate to the purposes of the system.

Gadamer: Understanding as Dialogue and Examination

Hans-Georg Gadamer demonstrates that understanding is not a purely technical operation. Human beings always understand from a particular historical and cultural standpoint. Their questions and interpretations are shaped by prejudgements in the original sense of the term: preliminary understandings that make experience possible but must also remain open to critical examination (Gadamer, 2004).

Understanding does not arise through a subject mechanically observing an object. It emerges through a movement between question and answer, familiarity and estrangement, continuity and revision. To understand means remaining open to the possibility that the matter itself may disclose something different from what was initially assumed.

This is important when considering what it means for artificial intelligence to remain intelligible. Complete technical transparency is probably unattainable. A modern AI system may contain such large numbers of parameters and such complex interactions that no individual can comprehend every element. Intelligibility, however, need not require that every detail be intuitively accessible. It may instead mean that the system can be investigated, challenged, and held accountable at the levels relevant to its particular use.

A system is practically intelligible when meaningful questions can be directed towards it:

What information is the assessment based upon?

Which assumptions underlie it?

What has the system been designed to optimise?

Which groups are represented in the data?

What is the model incapable of knowing?

What degree of uncertainty is attached to the answer?

Which alternative interpretations remain possible?

Who bears responsibility if the recommendation results in harm?

When such questions cannot be answered, the problem is not merely a lack of technical insight. The possibility of critical dialogue concerning the system’s role is lost. The space of understanding is thereby diminished.

Gadamer reminds us that genuine understanding requires openness to the claim made by the subject matter. In an encounter with a text or another person, one must be prepared to have one’s own understanding transformed. An AI system can produce linguistic responses that resemble contributions to a dialogue, but this does not relieve the human participant of the responsibility to examine what kind of truth claim those responses can sustain.

An AI system has no lived horizon in the human sense. It has not grown up within a family, experienced pain, feared death, or stood personally accountable before another human being. It can process linguistic patterns associated with such experiences, but it does not inhabit them in the same way that human beings do.

This does not render its responses worthless. A response may be illuminating even though the system does not possess a human lifeworld. Nevertheless, the user must remain aware of the difference between linguistic persuasiveness and existential understanding. A well-formulated response does not, in itself, guarantee truth.

Heidegger: When the Tool Becomes a Mode of Revealing

Martin Heidegger did not regard technology merely as an assemblage of tools. Technology is also a mode through which the world is disclosed. Modern technology tends to reveal everything that exists as a resource available for calculation, storage, management, and exploitation (Heidegger, 1977).

This perspective is particularly illuminating in relation to artificial intelligence. AI systems operate by transforming aspects of reality into data. Text, images, bodily movements, health information, purchasing behaviour, facial expressions, and social relationships are recorded and made available for analysis. What previously appeared as a complex human experience may therefore be represented as a collection of measurable attributes.

A student becomes a performance profile.

A patient becomes a predicted course of illness.

A child becomes a calculated risk category.

A job applicant becomes a ranking.

A conversation becomes a dataset.

A human being becomes a pattern.

Measurement is not inherently problematic. Medical measurements can save lives, and statistical analysis can reveal forms of injustice that would otherwise remain hidden. The problem arises when what can be measured is treated as the whole of reality. The model then ceases to be an aid to understanding the person; instead, the person is reduced to what the model is capable of registering.

Heidegger’s central question would therefore not merely be whether the system calculates correctly. He would ask which understanding of the human being and the world becomes dominant through the use of such systems. What happens when people increasingly encounter themselves as profiles, predicted outcomes, or collections of attributes to be optimised?

Artificial intelligence may consequently alter more than individual decisions. It may alter what is taken to be real. That which cannot be recorded may come to appear less significant. Slow experience, tacit knowledge, bodily perception, and moral hesitation may appear as inefficient remnants of an earlier world in which not everything could be calculated.

Yet hesitation is not necessarily a deficiency. It may express the ethical complexity of a situation. Slowness may be necessary in order to perceive another person adequately. Silence may contain something that cannot be expressed within the categories available to a system.

Heidegger’s critique is therefore not a demand that technology be rejected. Rather, it suggests that the danger inherent in technological enframing may also reveal the need for another mode of understanding. When the limits of technology become visible, human beings may become more attentive to what cannot be reduced to calculation and control. This, however, requires that human understanding not be surrendered to the system’s mode of representation.

Buber: From I–Thou to I–It

Martin Buber’s distinction between I–Thou and I–It relationships illuminates another dimension of the problem. In the I–It relation, the world is encountered as something that can be described, analysed, classified, and used. In the I–Thou relation, another person is encountered as a presence that cannot be reduced to attributes, categories, or functions (Buber, 1970).

Both forms of relation are necessary. Human beings cannot live without ordering and analysing the world. A physician must obtain medical information, a teacher must assess a student’s work, and a social worker must understand legislation, risk, and personal history. Yet if I–It becomes the only valid mode of relation, the encounter with the other as an irreplaceable subject is lost.

Artificial intelligence belongs fundamentally to the realm of I–It. The system processes patterns, categories, and statistical relations. Even when it communicates in a friendly and personalised manner, it does so on the basis of calculation and modelling. It may produce a linguistic form that resembles an I–Thou encounter, but this does not alter its underlying character.

The problem arises when the system’s model of the person takes precedence over the person’s own voice. An individual may then confront a decision shaped in advance by what people with similar characteristics usually do. The particular person risks being interpreted through the statistical pattern of a group.

Yet the other person is never merely an instance of a category. Something always remains that the model has not captured: a history, an experience, or the possibility of acting differently from what the system predicts.

When AI systems are used in healthcare, education, social work, or public administration, there must therefore remain a space within which the person can appear as more than a profile. Those affected by a decision must be able to speak, correct information, present their history, and challenge the model’s conclusion. Without such a space, the encounter becomes profoundly asymmetrical: the system defines the person, while the person is unable to challenge the definition.

Horvitz and West describe a possible development in which AI systems become increasingly attentive to human desires, evaluative contexts, and underlying motivations, while their own operations become progressively more difficult for human beings to follow. This asymmetry is not an I–Thou relation. It is an increasingly comprehensive I–It relation in which the human being becomes the object of analysis.

Buber’s philosophy reminds us that a person must always remain capable of being encountered as a Thou, even when technical systems provide relevant information about them.

Arendt: Thoughtlessness and the Disappearance of Responsibility

Hannah Arendt understood thinking as an internal dialogue through which individuals examine their own actions. Thoughtlessness does not refer to a lack of intelligence. It denotes the failure that occurs when people cease to examine what they are doing and instead repeat established rules, forms of language, and procedures without testing them against their own judgement (Arendt, 1978).

AI systems may intensify this temptation. They can provide an answer before the human user has fully formulated the question. They can generate an assessment, ranking, or recommendation that appears objective and authoritative. The more polished and coherent the result, the easier it may be to accept.

“The system recommended it” may thereby become a contemporary form of evading responsibility.

This is especially significant within institutions. When a decision is distributed among developers, datasets, procurement officers, managers, professionals, and automated models, it may become difficult to identify who is actually responsible. Each participant may claim merely to have fulfilled a limited role. The developer followed the specification, the manager followed the procedure, and the professional followed the system’s recommendation.

Responsibility, however, cannot be dissolved into systemic complexity. On the contrary, complexity makes it necessary to clarify responsibility more carefully. Someone decides to introduce the technology. Someone determines which data count. Someone establishes the thresholds. Someone decides whether a human being may override the result.

Arendt also reminds us that judgement cannot be replaced by general rules. The individual must assess the particular situation and imagine how the world appears from the standpoint of others. This enlarged mentality is essential within a pluralistic society. A system may summarise multiple perspectives, but human beings must still undertake the moral work of taking those perspectives seriously.

The danger posed by artificial intelligence is therefore not merely that a system may produce an erroneous result. The deeper danger is that human beings may cease to regard their own judgement as necessary. If they become accustomed to the assumption that a rapid and apparently objective answer is always available, the motivation to think slowly and critically may itself weaken.

This concern is consistent with research indicating that users may invest less critical effort in tasks when they place substantial confidence in generative AI tools, even though the same technologies may support reflection under other conditions (Lee et al., 2025). The decisive issue is therefore not simply whether people have access to artificial intelligence, but how human work is organised around it. Technology can be used to displace thought, but it may also be used to challenge and deepen it.

MacIntyre: Which Goods Does the Technology Serve?

Alasdair MacIntyre distinguishes between the internal goods of a practice and the external goods pursued by institutions (MacIntyre, 2007). Internal goods are forms of excellence and achievement that can be realised only through participation in the practice itself. In medicine, these may include sound clinical judgement and care for the person who is ill. In education, they may include insight, understanding, and the cultivation of independent judgement. In social work, they may include recognition, empowerment, and support for a more dignified life.

External goods include money, status, competitive advantage, control, and measurable efficiency. Institutions require such goods in order to survive, but these goods may also come to dominate and distort the practices that institutions are meant to sustain.

Artificial intelligence may support the internal goods of a practice. A physician may detect a pattern of illness at an earlier stage. A teacher may receive assistance in developing more effective educational materials. A social worker may gain access to relevant knowledge and alternative interpretations. A researcher may investigate relationships that were previously difficult to identify.

At the same time, technology may also be used to subordinate the practice to external goods. It may become primarily a means of processing more cases in less time, reducing staff, standardising judgement, or monitoring employees. In such circumstances, technology may render the institution more efficient while making the practice professionally and morally poorer.

The relevant question is therefore not simply whether artificial intelligence works. It is necessary to ask which good the technology has been designed to serve.

A school may use AI to help students understand difficult material. It may also use AI to increase completion rates without providing genuine academic support.

A healthcare system may use AI to strengthen clinical judgement. It may also use the system to shorten consultations to the point at which patients are no longer adequately heard.

A child welfare service may use data analysis to identify circumstances that deserve further investigation. It may also allow a risk model to shape the encounter with a family before the professional has understood the situation.

MacIntyre shows that institutions always risk corrupting the practices they were created to sustain. Artificial intelligence does not necessarily create this tendency, but it can make it more efficient and less visible. This is precisely why professional traditions are needed that can articulate the internal goods of a practice and defend them against a one-sided logic of efficiency.

Professional Judgement and the Particular Human Being

In professional practice, the relationship between artificial intelligence and practical philosophy becomes especially evident. Professional judgement is exercised in situations where general rules do not provide a complete answer. The practitioner must interpret a unique situation under conditions of uncertainty and assume responsibility for the consequences of the judgement.

Artificial intelligence may be a powerful aid. A system can process extensive bodies of research, draw attention to relevant risk factors, and identify alternative courses of action. Yet there remains a fundamental difference between statistical prediction and practical understanding.

Prediction indicates what individuals with similar characteristics often experience or do. Understanding is directed towards this particular person in this particular situation.

A human being can always act differently from what a model predicts. A person may change, resist, learn, regret, hope, or surprise others. Professional judgement must therefore hold probability open to possibility.

This is particularly important in social work. A child, a family, or a person living with addiction cannot be understood exhaustively through an accumulation of risk factors. Such information may be relevant, but the practitioner must also understand relationships, shame, trust, hope, history, and possibilities that have not yet become visible.

The system can analyse what has been recorded. It cannot necessarily know the significance of what was never recorded.

It may not understand the silence in the room, the uncertainty in a person’s voice, or the fragile trust that is beginning to emerge. It may not know why someone has withheld information, or why an apparently minor event has acquired decisive significance.

Artificial intelligence must therefore be used in ways that expand professional attention rather than narrowing it. A system may indicate: Here is something that should be investigated. It should not be permitted to conclude, on its own: This is what this person is.

Explainability is especially important where decisions have serious consequences. Rudin (2019) argues that, in high-risk contexts, interpretable models should be used whenever possible instead of attempting to explain opaque models retrospectively. The point is not that every complex system can be made simple, but that an individual affected by a decision must be able to receive a genuine justification.

A justification is more than a technical description. It must be intelligible to the person affected, and it must explain why the decision is relevant and defensible in that particular case. A probability is not, by itself, a moral justification.

Freire: The Right to Name the World

Paulo Freire connected liberation with the human capacity to read, name, and transform the world. Oppression occurs when people are transformed into objects of definitions imposed by others. A liberating practice must therefore be dialogical: people must become active participants in interpreting their own situation (Freire, 2018).

Artificial intelligence may support such liberation. It can provide access to information, language, and forms of expression that were previously inaccessible. It may help a person articulate an experience, understand a difficult document, or gain an overview of a complex matter. For individuals with reading and writing difficulties or other disabilities, such technologies may open new possibilities for participation.

Yet artificial intelligence may also produce the opposite effect. If a limited number of systems increasingly supply the categories, formulations, and interpretations through which reality is understood, the voices of individuals may become weaker. People may continue to express themselves, but their expression may be shaped by systems whose priorities they do not understand.

The question then arises: Who has the authority to name the world?

Freire’s perspective demonstrates that AI literacy cannot be reduced to learning how to issue effective instructions to a system. It must also involve a critical reading of the technology. Users must be able to ask who developed the model, which interests it serves, which experiences are represented in the data, which perspectives are absent, and how the system’s outputs shape the understanding of a problem.

A liberating relationship with artificial intelligence therefore requires more than access. It requires possibilities for insight, criticism, correction, and participation.

The Black Box and the Limits of Understanding

Discussions of so-called black boxes are often shaped by an unrealistic assumption that complete understanding is always possible. Human beings do not fully understand every detail of complex biological, economic, or social systems. A physician does not know every molecular process occurring within a patient’s body, and an individual does not understand every process taking place within their own brain.

The appropriate demand cannot therefore be that every AI system be entirely transparent in every detail. Horvitz and West themselves emphasise that understanding does not require knowledge of every line of code or every parameter. Scientific understanding is frequently partial and operates at several levels (Horvitz & West, 2026).

Partial understanding, however, is not equivalent to the absence of understanding. Human beings must know enough to identify risks, investigate errors, understand limitations, and intervene before harm occurs.

This may require several forms of intelligibility:

Technical intelligibility concerns how the system is constructed and tested.

Statistical intelligibility concerns the reliability of its performance and the groups for whom it performs well or poorly.

Institutional intelligibility concerns who develops, owns, and controls the system.

Practical intelligibility concerns how a recommendation should be interpreted in a particular context.

Ethical intelligibility concerns the values and priorities embedded in the system.

Democratic intelligibility concerns the ability of citizens to examine, challenge, and influence its use.

A system may be technically documented while remaining democratically opaque. It may be statistically accurate while ethically problematic. It may offer an explanation that is intelligible to an expert but meaningless to the person affected by the decision.

Intelligibility must therefore always be linked to a further question: intelligible to whom, for what purpose, and in which situation?

Systems That Develop Systems

Horvitz and West draw attention to AI-directed AI development. Artificial intelligence is already used to support programming, model design, testing, and optimisation. When such processes proceed through repeated cycles, development may occur in ways that rapidly exceed human intuition.

The result may be what the authors describe as operational opacity: performance improves while insight into the reasons for the improvement diminishes. The system functions, yet human beings no longer fully understand what produces its performance.

There is a temptation to accept such opacity as long as the outcomes remain satisfactory. Many technologies are used without the user understanding their internal mechanisms. Artificial intelligence, however, differs from an ordinary tool because it does not merely perform stable operations. It participates in decision-making, communication, and interpretation. It may influence the basis upon which subsequent actions are taken.

When AI systems contribute to the development of new systems, they should therefore also generate documentation, explanations, and analytical tools that make their architecture and operations open to human scrutiny. Otherwise, opacity may become an unintended feature of the design process itself.

The authors also identify the growing interaction among AI systems. In environments involving multiple agents, systems may communicate and coordinate at a scale that is difficult for human beings to follow. Their communication may move away from human language and human forms of reasoning if they are optimised solely for efficiency.

Such behaviour may be coherent within the AI environment while remaining difficult for human beings to interpret. A form of interactional opacity may consequently emerge: the systems understand, or at least adapt to, one another, while human beings lose oversight.

This is not necessarily evidence of an autonomous machine will. It may simply be the result of optimisation. Nevertheless, the effect remains both practical and political: human beings become less able to control systems that act on their behalf.

The Quiet Development of Dependency

Perhaps the most subtle warning advanced by Horvitz and West is not that AI systems will become excessively intelligent, but that human beings may gradually lose the motivation to understand them. When technology functions well, the incentive for criticism may decline. When answers are delivered quickly, politely, and persuasively, the desire to examine their foundations may weaken.

Dependency may therefore develop without dramatic interruption.

The system is first permitted to summarise the text.

It is then permitted to propose an interpretation.

Next, it is allowed to formulate the judgement.

Eventually, human verification may come to be regarded as unnecessary duplication.

This is not an argument against using AI as a writing or reflection partner. Such use may be intellectually fruitful and practically liberating. Technology may help a person write more, identify connections, and test ideas through dialogue. Yet the collaboration must be organised in such a way that the human being remains the one who evaluates, challenges, chooses, and accepts responsibility.

A writing partner may suggest a formulation. The author must still determine whether the formulation is true.

A reflection partner may identify a connection. The human being must still investigate whether the connection is valid.

A system may produce an argument. The human being must still judge whether the premises are defensible.

The danger is not necessarily that AI will make human beings less intelligent. Rather, certain forms of intellectual effort may be practised less frequently and may consequently weaken. Judgement is cultivated through exercise. A person who no longer needs to remember, formulate, investigate, or doubt may gradually lose some of the capacity to perform precisely these activities.

A Socratic Relationship with Artificial Intelligence

Socrates left no philosophy of artificial intelligence, but his practice of questioning offers an important model for the contemporary encounter with it. He did not regard a persuasive answer as the conclusion of inquiry. He examined concepts, exposed contradictions, and demonstrated that people frequently believed themselves to know more than they actually knew.

A Socratic relationship with AI therefore requires neither uncritical trust nor principled rejection. It requires an enquiring disposition:

What does the answer actually mean?

On what basis does the system claim to know this?

Which assumptions underlie the conclusion?

What considerations speak against it?

Which experiences are not represented?

Who benefits from this interpretation?

What may follow if the recommendation is acted upon?

Who is responsible if it proves to be wrong?

The competent user of AI is not merely the person who succeeds in producing a polished response. It is the person who can use the response as the basis for a better question.

This also requires intellectual humility. Human beings must acknowledge that artificial intelligence may detect something that they have failed to perceive. Yet the limitations of the system must be approached with the same humility. Neither human beings nor machines should be regarded as infallible.

Wilder, Horvitz, and Kamar (2020) demonstrate how AI systems may be designed to complement human reasoning rather than merely compete with it. Such a complementary model is highly relevant to practical philosophy. The objective is not to determine whether the human being or the machine is superior in absolute terms, but to investigate how their respective strengths may be combined while their weaknesses are recognised and corrected.

An appropriate division of labour must therefore be based upon reciprocal correction. The system may alert the human being to overlooked patterns. The human being may correct the system’s insufficient understanding of context, value, and responsibility.

Human Agency as a Goal

At the centre of the editorial is a sentence that captures its practical-philosophical significance:

“Preserving human agency must therefore remain a central goal” (Horvitz & West, 2026, p. 1003).

Agency refers to the capacity to act, to be the origin of one’s actions, to formulate goals, exercise judgement, and accept responsibility for the consequences.

It is not sufficient that a human being performs the final action if the system has defined the problem, selected the relevant criteria, presented the available alternatives, and identified the recommended solution. Human agency also requires the capacity to question the framing of the problem itself.

Human beings must not merely monitor how AI systems behave. They must also examine how these systems shape human purposes and judgements. A system that consistently reduces resistance and facilitates action may appear helpful. Yet certain forms of resistance are necessary. Friction may provide time for reflection. Disagreement may protect against conformity. Uncertainty may invite further investigation.

A system designed to support human flourishing is therefore not necessarily one that renders every process frictionless. It may instead be a system capable of slowing down, expressing uncertainty, or returning a decision to human judgement.

It should be capable of indicating:

This is not known.

This answer is uncertain.

These considerations cannot be resolved statistically.

This decision requires human responsibility.

Such limitations are not merely weaknesses. They may be marks of responsible design.

Institutional and Democratic Responsibility

It is insufficient to place the entire burden of responsibility upon the individual user. Users cannot easily investigate systems that corporations and institutions keep hidden. Human agency therefore requires institutional arrangements for documentation, independent evaluation, appeal, and oversight.

Pasquale (2015) has shown how opaque systems may concentrate power among actors who evaluate others while remaining beyond evaluation themselves. O’Neil (2016) describes how mathematical models may reinforce injustice when they are deployed at scale, remain inaccessible to scrutiny, and generate consequences that later become new data for the model.

A system may, for example, classify a particular group as presenting a higher level of risk on the basis of historical data. If that group is subsequently subjected to more intensive monitoring, a greater number of incidents will be recorded. These records may then appear to confirm the system’s original assumption. The model thereby becomes part of a self-reinforcing cycle.

Evaluation must therefore take place under conditions that resemble actual use. Static testing is insufficient if the system modifies its behaviour in interaction with users or is influenced by the context in which it operates. Systems should be evaluated for bias, robustness, uncertainty, and differences between laboratory conditions and practical application.

Standards must also develop alongside technological change. UNESCO’s Recommendation on the Ethics of Artificial Intelligence identifies human dignity, transparency, fairness, and accountability as fundamental principles (UNESCO, 2021). The NIST risk management framework emphasises that AI-related risks must be addressed throughout the life cycle of the system and in relation to the specific context of use (National Institute of Standards and Technology, 2023).

Such frameworks are important, but they cannot replace democratic deliberation. Decisions concerning which dimensions of human life should be automated are not matters solely for developers and technology companies. They are political questions because their consequences are distributed among individuals and groups with unequal access to power.

Those most affected by a system often possess the least influence over its design. Affected groups must therefore be afforded meaningful opportunities to participate, rather than merely being informed after the technology has already been introduced.

The Promise and the Limitations of Technology

A practical-philosophical critique of artificial intelligence need not be hostile to technological development. AI presents significant possibilities. It can make knowledge more accessible, support people with disabilities, advance scientific research, and contribute to improved diagnosis. It may function as a writing partner, translator, educational resource, and gateway to scholarly conversations that were previously accessible only to specialists.

It would be no wiser to reject these possibilities than to surrender to them uncritically.

The central question is how technology can be placed in the service of human purposes without those purposes gradually being reshaped by the technology’s own limited categories. This requires not only better systems but also better practices surrounding them.

A responsible practice must preserve several human possibilities:

the possibility of understanding enough to intervene,

the possibility of disagreeing with the system,

the possibility of demanding an explanation,

the possibility of being encountered as an individual,

the possibility of reserving certain decisions for human judgement,

and the possibility of rejecting the technology when its use cannot be justified.

Such a practice must also protect what cannot be reduced to efficiency: care, trust, vulnerability, friendship, reconciliation, love, and responsibility. Artificial intelligence can generate language concerning these experiences. It may assist people in reflecting upon them. Yet it cannot, by itself, determine what significance they ought to possess within a human life.

When the Creator Must Answer for What Has Been Created

Horvitz and West’s image of a narrowing window expresses an urgent temporal challenge. If the capacities, complexity, and social influence of AI systems continue to increase without a corresponding development in understanding, responsibility, and control, dependency may become firmly established before institutions are able to respond adequately.

A reversal may then occur in the relationship between human beings and their creations. Human beings create systems to serve particular purposes, but subsequently begin to adapt their practices, language, and goals to the systems’ modes of operation. What was initially introduced as an aid may become a standard for how reality itself is to be understood.

Practical philosophy reminds us that the creator cannot transfer responsibility to the creation.

Human beings do not need to understand every parameter within a language model in order to use it. They must, however, understand enough to recognise when it should not be used, when it must be overruled, and who bears responsibility for the consequences. They must remain capable of identifying errors, examining bias, protecting vulnerable persons, and explaining decisions that intervene in the lives of others.

What is required is not complete technical transparency, but sufficient understanding to preserve responsible action.

The human being must still be capable of saying: This is my decision. I understand the reasons. I have considered the objections. I recognise that I may be mistaken. And I am prepared to answer for the consequences.

This is what is ultimately at stake in the relationship between artificial intelligence and practical philosophy. The issue is not merely whether systems will become more intelligent, but whether human beings will remain capable of agency. It is not merely whether technology can provide answers, but whether human beings will continue to ask questions. It is not merely whether machines can understand more about human beings, but whether human beings will retain the capacity to understand what kind of persons they wish to become.

The most important task, therefore, is not merely to create artificial intelligence that serves humanity. It is to preserve and cultivate human beings who possess sufficient agency to determine what technology ought to serve.

---

References

Arendt, H. (1958). The human condition. University of Chicago Press.

Arendt, H. (1978). The life of the mind (M. McCarthy, Ed.). Harcourt.

Aristotle. (1999). Nicomachean ethics (T. Irwin, Trans., 2nd ed.). Hackett Publishing Company.

Bender, E. M., Gebru, T., McMillan-Major, A., & Shmitchell, S. (2021). On the dangers of stochastic parrots: Can language models be too big? In Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency (pp. 610–623). Association for Computing Machinery. https://doi.org/10.1145/3442188.3445922

Buber, M. (1970). I and Thou (W. Kaufmann, Trans.). Charles Scribner’s Sons. (Original work published 1923)

Dreyfus, H. L. (1992). What computers still can’t do: A critique of artificial reason. MIT Press.

Floridi, L., Cowls, J., Beltrametti, M., Chatila, R., Chazerand, P., Dignum, V., Luetge, C., Madelin, R., Pagallo, U., Rossi, F., Schafer, B., Valcke, P., & Vayena, E. (2018). AI4People—An ethical framework for a good AI society: Opportunities, risks, principles, and recommendations. Minds and Machines, 28(4), 689–707. https://doi.org/10.1007/s11023-018-9482-5

Freire, P. (2018). Pedagogy of the oppressed (M. B. Ramos, Trans., 50th anniversary ed.). Bloomsbury Academic. (Original work published 1970)

Gadamer, H.-G. (2004). Truth and method (J. Weinsheimer & D. G. Marshall, Trans., 2nd rev. ed.). Continuum. (Original work published 1960)

Heidegger, M. (1977). The question concerning technology. In The question concerning technology and other essays (W. Lovitt, Trans., pp. 3–35). Harper & Row. (Original work published 1954)

Horvitz, E., & West, R. (2026). A narrowing window to understand AI. Science, 392(6802), 1003. https://doi.org/10.1126/science.aei3167

Kant, I. (1998). Groundwork of the metaphysics of morals (M. Gregor, Trans.). Cambridge University Press. (Original work published 1785)

Lee, H.-P., Sarkar, A., Tankelevitch, L., Drosos, I., Rintel, S., Banks, R., & Wilson, N. (2025). The impact of generative AI on critical thinking: Self-reported reductions in cognitive effort and confidence effects from a survey of knowledge workers. In Proceedings of the CHI Conference on Human Factors in Computing Systems. Association for Computing Machinery.

MacIntyre, A. (1999). Dependent rational animals: Why human beings need the virtues. Open Court.

MacIntyre, A. (2007). After virtue: A study in moral theory (3rd ed.). University of Notre Dame Press.

Mittelstadt, B. D., Allo, P., Taddeo, M., Wachter, S., & Floridi, L. (2016). The ethics of algorithms: Mapping the debate. Big Data & Society, 3(2), 1–21. https://doi.org/10.1177/2053951716679679

National Institute of Standards and Technology. (2023). Artificial intelligence risk management framework (AI RMF 1.0) (NIST AI 100-1). U.S. Department of Commerce. https://doi.org/10.6028/NIST.AI.100-1

O’Neil, C. (2016). Weapons of math destruction: How big data increases inequality and threatens democracy. Crown.

Pasquale, F. (2015). The black box society: The secret algorithms that control money and information. Harvard University Press.

Ricoeur, P. (1992). Oneself as another (K. Blamey, Trans.). University of Chicago Press.

Rudin, C. (2019). Stop explaining black box machine learning models for high-stakes decisions and use interpretable models instead. Nature Machine Intelligence, 1(5), 206–215. https://doi.org/10.1038/s42256-019-0048-x

UNESCO. (2021). Recommendation on the ethics of artificial intelligence. UNESCO.

Wilder, B., Horvitz, E., & Kamar, E. (2020). Learning to complement humans. In C. Bessiere (Ed.), Proceedings of the Twenty-Ninth International Joint Conference on Artificial Intelligence (pp. 1526–1533). International Joint Conferences on Artificial Intelligence Organization. https://doi.org/10.24963/ijcai.2020/212

Zuboff, S. (2019). The age of surveillance capitalism: The fight for a human future at the new frontier of power. PublicAffairs. 

The most important task, therefore, is not merely to create artificial intelligence that serves humanity. It is to preserve and cultivate human beings who possess sufficient agency to determine what technology ought to serve.

This academic essay is based on an editorial i Science (published 6. June 2026). My reading and writing is done in light of Practical Philosophy. It is written in a conversation with OpenAI/ChatGPT.