Research and version note This is a version 0.1 research draft in The Future Has No Representative and the provisional conclusion of the season. It synthesises a working framework from the previous eleven articles and does not claim to complete a theory of intergenerational justice. The framework needs continuing tests and revision through further legal, technological, and cultural cases.
The Future Has No Representative · Article 12
Quick read
Predicting the future and preserving choice are not exclusive duties. A government that refuses to study climatic, demographic, fiscal, or technological risk leaves avoidable ignorance to later people. A government confident that one projection is sufficiently accurate may construct a path that becomes difficult to leave. Responsible institutions use projection to discover consequences, acknowledge its boundaries, and make adjustable decisions where those boundaries matter.
A duty to predict is not a guarantee of correctness. It requires reliable evidence available at the time, separation of observation from projection and scenario, disclosure of important assumptions, and updating when information changes. An unforeseeable consequence is not automatically culpable. Deliberately failing to investigate an important known risk, or treating uncertainty as zero, is different.
Preserving choice does not mean keeping every possibility open forever. Institutions invest, set standards, and close paths. Some dangerous options should be excluded, and indefinite interim arrangements can transfer management to the future. What should survive is a basic capability to form and act on later judgment: habitable conditions, usable knowledge, fiscal and professional capacity, political space to alter institutions, and a path to remedy past error.
That requires distinguishing a legal permission to change from a practical ability to change. An exit right without resources, skills, and records is an option only on paper.
Earlier articles show different forms of this responsibility. Debt must be assessed with inherited assets and capacity. A bridge depends on maintenance across terms. Nuclear waste separates passive safety from memory. Language continuation is governed by living communities. AI infrastructure needs migration and retirement. The common concern is not one fixed picture of the future. It is preventing a present choice from compressing the future into an outcome that can only be accepted.
We first owe the future reasonable foresight, and then owe preservation of judgment and correction where foresight cannot be reliable. The less reversible a decision and the more deeply it affects basic conditions, the less a claim about the “most likely” future can carry the justification alone. The future does not need us to write every answer. It needs us not to delete the questions, evidence, and capacity to answer again.
Why prediction versus choice is a false division
One response to uncertainty demands better prediction. More capable models, data, and foresight let governments prepare. Another says prediction will fail and flexibility should allow the future to decide. Each identifies a real problem and becomes hazardous as a single principle.
Without prediction, option preservation lacks direction. A government that does not investigate sea level, ageing, or asset deterioration cannot see which paths are closing. Flexibility may be delay under another name, ending only after the adjustment window has gone.
Prediction alone encourages over-optimisation. A central scenario is built into rail, energy, and welfare, and changed conditions leave the structure unable to adapt. Precise numbers create an appearance of control, making omitted communities, culture, and tail risk appear less important.
The duties arise under different epistemic conditions. Where causal evidence is strong, institutions should act rather than cite general uncertainty. Social paths, technology, and the far horizon call for scenario ranges and revision points. The task is not to choose one duty, but to align both with what can be known and what cannot readily be reversed.
What a duty of prediction requires
It begins with identifying the issue. An institution holding bridge inspection records should estimate maintenance needs. A government aware that greenhouse gases accumulate should use relevant climate projections. Foresight does not require omniscience; it requires an institution not to shorten the horizon deliberately.
Method should fit the object. A demographic projection, climate scenario, fiscal stress test, and foresight exercise on AI cannot use the same confidence language. An observed trend must be separated from a path dependent on future policy, and a central estimate should not erase the model range.
Important assumptions belong in public reasons. Growth, discounting, service life, adaptive capacity, and vendor update practice can reverse a decision. Their existence in a technical appendix is not sufficient if affected people cannot identify which assumptions drive the result.
Projection should continue after approval. Monitoring tests assumptions, while triggers determine when to strengthen, slow, or exit. A forecast that never permits observation to change a policy has ceased to function as knowledge and become an institutional belief.
The history of error should be preserved as well. Reports often retain the final scenario and remove earlier disagreement or failed models. Future scrutiny needs to know how uncertainty developed and who advanced alternatives. A record admitting error does not weaken authority; it helps prevent successors from repeating the same blind spot.
When failure to predict is not absence of responsibility
An actor should not ordinarily carry the same responsibility for a consequence that could not have been foreseen as for deliberate or negligent harm. New science can reveal a danger unknowable before, and innovation can generate genuine surprise. If every later loss were attributed backward as fault, unbounded liability would discourage building and experimentation.
“It was unknowable at the time” requires evidence. Did the institution follow the available professional standard? Did it ignore research conflicting with project interests? Was uncertainty recorded candidly? Such questions distinguish reasonable ignorance from ignorance produced for convenience.
Government and firms can preserve denial by refusing monitoring, limiting data access, or defining research too narrowly. This is not a natural state outside responsibility. A party controlling a potential hazard and capable of investigation carries at least a duty of reasonable inquiry.
Responsibility also moves through time. A new risk unforeseeable to the original designer may become clear to a later owner, who then has a duty to act. Bridges, AI, and medical systems all show that judgment cannot remain fixed at deployment. Knowledge and control at a particular time give rise to a corresponding obligation.
Which choices should be preserved?
“Leave choices to the future” sounds generous and can remain empty. Land, resources, and attention are finite, and maintaining one option often reduces another. Policy has to identify the capability under protection.
Basic living conditions come first. Drinkable water, habitability, health, and safety should not be depleted on an assumption that future technology will be better. Formal political choice has little content when its material basis collapses.
Knowledge conditions matter too. Archives, method, dissent, and histories of failure tell successors what they have inherited. Data without explanation, or a conclusion without evidence, compresses judgment. Cultural knowledge may also need controlled transmission by rights-holders; public release is not the only form of preservation.
Institutional and material capability form another layer. Without maintenance funds and skill, a future government cannot repair a bridge even if it wishes. Without portable data and human understanding, termination of an AI service is a paper option. Debt that consumes fiscal space narrows crisis response. Choice requires real resources and not merely legal wording.
Political conditions complete the structure. Successors should be able to amend a rule, contest a past decision, and obtain remedy. A constitution or durable commitment may protect basic values without placing every implementation path outside later democratic change.
Why not every option should survive
Some paths ought to close. Prohibitions on toxic emissions, dangerous products, or rights violations reduce the number of available actions while expanding practical safety and freedom. Option count is not a measure of justice.
Preservation has a cost. Maintaining retrievability of nuclear waste needs institutions and security. Several technical platforms consume budgets and people. Refusing any teaching standard for a language can prevent materials from being developed. Passing every decision onward transfers uncertainty and labour with it.
Past investment can also encourage wrongful preservation. Indefinite maintenance of a low-value asset because it was expensive is sunk-cost behaviour, not intergenerational responsibility. The future should inherit the capacity to retire, and an example of the present ending failed projects in an orderly way.
Irreplaceability helps orient judgment. The disappearance of a software product is usually substitutable. Extinction of a species, loss of the final context of an old language, or destruction of a stable groundwater system is harder to repair. The less substitutable the object, the stronger the reason for preservation, subject to current rights and distribution.
What precaution provides—and does not
Principle 15 of the 1992 Rio Declaration states that a lack of full scientific certainty should not be used to postpone cost-effective measures where there are threats of serious or irreversible environmental damage. It addresses a recurring evasion: because a model is incomplete, risk is treated as absent.
Precaution does not prescribe the strictest prohibition under every uncertainty. Seriousness, irreversibility, cost, and capacity still require judgment. Risks compete, and a measure preventing one harm can cause another.
The principle is most useful in changing the burden of explanation. When an activity may cause grave irreversible consequences, its proponent cannot require affected people to prove that disaster is certain. Safety, alternatives, and monitoring need to be demonstrated. A small and readily correctable experiment can receive a lower evidential threshold.
Precaution should be adaptive. One final approval or prohibition cannot address continuing change. Staged decisions, limits on trial scale, triggers, and scheduled review permit policy to respond as evidence develops. They make uncertainty a design variable rather than a reason for paralysis.
A common framework from twelve articles
The series began with rights. Whether unborn persons hold complete present legal rights remains contested, while current obligations need not wait for their existence. Article 2 considered institutional time and how a short term can carry long responsibility through reasons and records. Discounting, debt, and infrastructure showed that technical numbers need to be read with distribution, assets, and maintenance capacity.
Nuclear waste extended the horizon beyond institutional life and required passive safety to be layered with memory. The article on children warned against abstracting a living, expressive person into “the future”. The language article showed that continuation is not freezing and that stewardship must be governed by current communities. Climate models provide visibility without consent. AI infrastructure shows how a short-lived model can create long lock-in through its surrounding system.
Six questions can be used across these cases. Who is affected and currently unable to participate? What is known, and what exists only in scenarios? Which paths close and do they concern basic conditions? How are present benefit and later burden distributed? Who can detect and stop failure? Can successors recover reasons and obtain remedy?
They are not a decisional formula. Their purpose is to prevent one dimension from absorbing all others. Predictive accuracy does not prove fair distribution. Open procedure cannot repair every irreversible harm. Preserved records without resources for action may leave only archival choice.
The limited relation to Sustenesis
This series is an application and does not redefine Sustenesis in twelve articles. Its canonical philosophical account remains Sustenesis.com, and case writing should not shift the theory merely to produce an easy fit.
The application uses one limited methodological relation. Continuation of a structure need not mean that material, personnel, and form remain unchanged. Bridge responsibility passes through institutions; a language exists through changing use; public systems are maintained through relations, constraints, correction, and activity. Responsibility to the future therefore concerns the conditions through which later structures can continue forming and adjusting, not only the custody of present objects.
This relation remains a research path, not proof of the philosophy by example. If an institution can persist only through freezing, or transformation has destroyed relevant continuity, application should reveal the conflict and test the conceptual boundary. The theory informs questions, while facts retain the authority to require a revision of method.
Preserving choice, from this perspective, does not mean leaving the future blank. Every existence forms within inherited conditions. Present responsibility avoids compressing those conditions into a single unexplained path while providing enough structure for later people to act rather than begin among ruins.
Provisional judgment: exercise reasonable foresight, then preserve correction
We owe a duty of prediction. A society possessing science, archives, and organisation cannot use the unknown future as a reason not to investigate long effects. Where reliable evidence already identifies serious risk, waiting for complete certainty can itself be irresponsible.
Prediction always has a boundary, so responsibility cannot end when a model is completed. Beyond reliable knowledge, safety margins, reversible stages, monitoring, and practical exit are required. Successors also need original reasons, remaining resources, and a route of challenge. Preserving choice is an institutional response to limited knowledge.
The intensity varies. A small trial that can be withdrawn quickly may permit exploration. A decision affecting basic living conditions across generations and resistant to repair requires broader scenarios and stronger limits. Prudence should not have the same form for every future effect.
The future has no representative, but it is not wholly absent from present reasoning. It becomes partly visible through causal evidence, the rights of living young people, long-lived assets, and inherited risks. We still cannot know how successors will live. That boundary generates the responsibility: use what can be known, do not disguise what cannot as permission, and preserve the capacity to know and decide again.
Primary sources and further reading
- United Nations, Rio Declaration on Environment and Development, especially Principle 15.
- United Nations, Declaration on Future Generations, 2024.
- OECD, Strategic Foresight Toolkit for Resilient Public Policy, 2025.
- World Bank, Investment Decision Making Under Deep Uncertainty.
- Sustenesis, canonical source for the philosophical theory referenced in the limited application section above.
Series navigation: The Future Has No Representative — series overview
Discover more from Geoffrey Chen
Subscribe to get the latest posts sent to your email.