Corrigible Realism

This document formulates Corrigible Realism as the philosophical and academic placement of Røyndalism. It is a position paper, not a completed textbook and not a declaration of infallible system status. Its formulations must remain open to testing, specification, and correction.

The central thesis is that actuality has ontological priority while every human statement about actuality carries epistemic responsibility. What is does not depend upon the claim made about it. The claim, by contrast, depends upon being able to refer, correspond, or otherwise remain answerable to what it concerns.

Human orientation proceeds through mediations: sensation, memory, language, signs, categories, models, institutions, and technical systems. These mediations are not errors in themselves. They make distinction, communication, and cooperation possible. The problem begins when the mediation is treated as the actuality it was meant to serve.

A word may be treated as though it creates what it names. A model may be treated as though it were the modelled actuality. An institution may allow procedure to outweigh the actual condition it was established to address. An identity may be made equivalent to the whole self. A measurement may be detached from instrument, standard, situation, and interpretation. In each case, the frame begins to replace actuality.

Corrigible Realism is developed to make this displacement visible. It defends realism without making human representation sovereign, and it defends corrigibility without making actuality relative to the one who corrects.

Corrigible Realism is an ontologically realist and epistemically corrigible position. It holds that something is the case independently of whether human beings know, formulate, or accept it, and that human knowledge claims must remain open to revision in light of better access, better distinctions, stronger documentation, or disclosed contradiction.

Realism here means that actuality is not reduced to language, consciousness, power, consensus, model, or method. Corrigible means that access, representation, and practice are not identified with actuality itself. The position is therefore neither dogmatic certainty nor arbitrary doubt, but an ordering of levels: what is, and what is claimed about what is.

The position begins with an order of ground: before anything can be known, described, evaluated, modelled, or measured, something must be the case. This is not an empirical hypothesis about a particular object, but a condition for questions, claims, errors, experiences, and contradictions to occur at all.

Ontological priority does not mean that every ontological question is already settled. It means that no theory of actuality can be placed before the actuality that makes the theory, theorist, and question possible. Materialism, idealism, dualism, theism, naturalism, and other metaphysical models must all answer to what is; none becomes the ground merely by being adopted as a frame.

Human knowledge does not arise by a subject possessing actuality, but through encounter, distinction, evaluation, and recognition. Each stage may fail. Appearance may mislead, memory may change, categories may be too coarse, inferences may be invalid, and documentation may be incomplete.

Corrigibility is therefore not an addition made after knowledge has been established. It is part of the discipline of knowledge. A claim that can in principle be corrected by nothing it encounters has broken its relation to actuality and functions instead as closed belief, identity protection, or system defence.

This does not require every claim to be held with equal uncertainty. Degrees of confidence may vary with access, documentation, replication, precision, and resistance to counter-testing. Corrigible Realism permits strong knowledge, but not immunised knowledge.

Corrigible Realism distinguishes actuality from appearance without separating them into two closed worlds. Appearance is something that actually occurs: something presents, affects, is registered, or leaves a trace. Yet that which appears and the way in which it appears are not automatically identical.

Observation is situated. It proceeds through body, instrument, time, position, assumptions, and interpretation. This does not make observation worthless; it makes documentation of its conditions necessary. An instrument extends access, but also introduces calibration, resolution, error margin, and processing that must be traced.

Error is possible precisely because there is a distinction between what is claimed and what is the case. Illusion does not abolish realism. An illusion is an actual appearance that is misassigned when treated as an exhaustive account of its cause or object.

Language is an actual human practice within actuality. Words may distinguish, point, preserve experience, and make testing possible. But the sign is not the thing, the definition is not the phenomenon, and the name is not the self. Linguistic validity requires that use can be traced to something it refers to or does in an actual context.

Correspondence does not require a sentence to copy actuality picture by picture. It requires the claim to hold about what it asserts. Different linguistic forms may refer to the same condition, while the same word may be used for different conditions. Reference must therefore be clarified before semantic agreement or disagreement can be assessed.

When words are made sovereign, definitional power may replace inquiry. Corrigible Realism then requires the concept to be returned to use, referent, consequence, and conditions of correction.

A model is an ordered representation that selects, simplifies, or formalises features of an actual or possible condition. It may be mathematical, linguistic, visual, mechanical, statistical, or institutional. A model is useful precisely because it does not include everything. The same reduction that makes it manageable also establishes its limits.

Model validity must therefore be assessed relative to purpose, input data, assumptions, scale, information loss, and correspondence with observable consequences. Internal consistency shows that the model holds together with itself; it does not by itself show that the model answers to what it is used to explain.

Measurement is an act of comparison within actuality. It requires actual instruments, standards, calibrations, observers, objects, properties, and procedures. The number is a result of this relation, not a replacement for the relation. When measurements are detached from the conditions under which they were produced, precision may be simulated without corresponding traceability.

A human being may be described through names, properties, relations, diagnoses, roles, memberships, and self-narratives. Such identifications may be true and practically important. They remain limited representations of an actual person, not ownership of the whole self.

Identity capture occurs when a category ceases to serve the person and instead regulates what the person may see, say, change, or belong to. Objection to the category may then be experienced as erasure of the self, even when the objection concerns a claim, role, or social structure.

Corrigible Realism requires the person to be encountered before the category is allowed to speak completely. This does not make every self-description true merely because it is experienced, nor does it render biological, social, and legal conditions irrelevant. It requires analysis to distinguish person, experience, category, environment, action, and consequence.

Systems and institutions make coordination, continuity, accountability, and large-scale action possible. They develop language, procedures, roles, archives, and forms of decision. These structures may nevertheless begin to protect their own representation against the actual condition they were established to address.

Authority may be epistemic, legal, administrative, professional, or social. Corrigible Realism does not reject authority as such. It distinguishes authority as compressed access to competence from authority as replacement for testing. A source may provide reason for trust, but that trust must remain traceable to practice, documentation, responsibility, and possibilities of correction.

An institutional decision is actual as a decision, but this does not make its content true or just. A judgment may be validly issued and still rest upon false facts. A scientific consensus may be genuinely established and later revised. A register may contain an authorised category that does not exhaust the person or event.

The practical methodology that follows from Corrigible Realism is Reality Audit. It examines what is being treated as actual, what the frame is built from, what it excludes, how claims can be traced, and what happens when the frame encounters objection.

The method is not designed to manufacture suspicion toward everything. It must distinguish warranted trust, unresolved uncertainty, documented error, capture risk, and identified capture. An audit that has decided its conclusion in advance has itself violated corrigibility.

• clarify the object and claim being audited
• distinguish actuality, appearance, representation, model, and decision
• trace sources, premises, data, categories, and transformations
• test internal coherence without mistaking it for external correspondence
• record uncertainty, alternative explanations, and absent access
• formulate findings proportionately to the documentation
• connect correction to the concrete failure that has been demonstrated

Corrigible Realism shares features with several realist and fallibilist traditions, but is not exhausted by simple identification with any one of them. It must be assessed through its own ordering of ontological ground, reference, corrigibility, and practical responsibility.

A corrigible realist position must itself be exposed to objection. The following objections identify central pressure points rather than final dismissals of neighbouring traditions.

Corrigible Realism may be misused when ‘actuality’ becomes an honorific for preference, when traceability is demanded asymmetrically of opponents, or when correction is used for social domination. The method must therefore document its own starting point and expose its own categories to audit.

The position does not automatically resolve every empirical, legal, scientific, or moral question. It establishes requirements by which answers remain answerable. Where access is absent, the finding must remain limited or unresolved; the gap must not be filled with simulated certainty.

A strong challenge would demonstrate meaningful, true, or knowledge-bearing representation that in principle stands in no clarifiable relation to what is, or show that correction is possible without any distinction between claim and condition. Such challenges must not be rejected by definition, but examined concretely.

As a research programme, Corrigible Realism requires further work on at least four levels: precise conceptual analysis, comparison with established realist and anti-realist traditions, operationalisation of Reality Audit, and documented cases in which the method is tested against alternative forms of analysis.

Standards must also be developed for degrees of correspondence, documentary strength, uncertainty, institutional responsibility, and proportional correction. A public revision history should make it visible when formulations are changed and why.

The ambition is not to turn Røyndalism into a closed school, but to provide a disciplined framework for returning theory, language, and practice to what they claim to concern. The position stands only while this ambition itself holds in contact with what is.