h t t p s : / / d o i . o r g / 1 0 . 1 0 1 6 / S 0 0 4 9 - 2 3 7 X ( 0 8 ) 7 1 2 0 4 - 7

Ingest: Epistemology Without a Knowing Subject

Type: types/ingest-report.md

Classification

A philosophical lecture arguing for an objective account of scientific knowledge through conceptual distinctions, thought experiments, biological analogies, and criticism of rival epistemologies. It reports no controlled learning experiment. Author: Karl R. Popper, writing as the proponent of conjectural knowledge and critical error elimination; this is primary evidence for his position, not independent confirmation of it.

Summary

Popper distinguishes physical objects, subjective acts or dispositions, and objective thought contents. The third category includes false as well as true theories, problems, arguments, and states of discussion: its membership does not depend on anyone believing or currently understanding a content. Although produced through human activity, this world has partial autonomy because constructions have unintended consequences and create unforeseen problems. Knowledge grows through a cycle from problems to tentative theories, attempted error elimination, and new problems. Scientific criticism depends, in his account, on descriptive and argumentative language, but thought contents are not identical with linguistic forms and informal argument precedes formalization. He extends trial and error to biological adaptation and subjective learning, while distinguishing the human capacity to criticize theories without perishing with them. The essay also develops its position through disputes with Plato, Hegel, empiricism, and Brouwer, and criticizes subjective approaches to probability and physics.

Quotes

(1) Without the development of an exosomatic descriptive language - a language which, like a tool, develops outside the body - there can be no object for our critical discussion. But with the development of a descriptive language (and further, of a written language), a linguistic third world can emerge; and it is only in this way, and only in this third world, that the problems and standards of rational criticism can develop. --- kb/sources/.snapshots/popper-epistemology-without-a-knowing-subject-1968.md @ sha256:11d8a52e58a660a115f877db34a4472af0b52c9a851ba7883dfdc95f49efca25 — Section 4, Language, criticism, and the third world, printed p. 346; first of the two all-important points.

The main point here I owe to Lakatos’s philosophy of mathematics. It is that mathematics grows through criticism of guesses and bold informal proofs. This presupposes their linguistic formulation, and their status in the third world. --- kb/sources/.snapshots/popper-epistemology-without-a-knowing-subject-1968.md @ sha256:11d8a52e58a660a115f877db34a4472af0b52c9a851ba7883dfdc95f49efca25 — Section 6, Appreciation and criticism of Brouwer’s epistemology, printed p. 360; opening paragraph.

Scientists try to eliminate their false theories, they try to let them die in their stead. --- kb/sources/.snapshots/popper-epistemology-without-a-knowing-subject-1968.md @ sha256:11d8a52e58a660a115f877db34a4472af0b52c9a851ba7883dfdc95f49efca25 — Section 4, printed p. 347; final paragraph before section 5.

As against this I suggest that everything depends upon the give and take between ourselves and our work; upon the product which we contribute to the third world, and upon that constant feed-back that can be amplified by self-criticism. --- kb/sources/.snapshots/popper-epistemology-without-a-knowing-subject-1968.md @ sha256:11d8a52e58a660a115f877db34a4472af0b52c9a851ba7883dfdc95f49efca25 — Section 9, Discovery, humanism and self-transcendence, printed p. 370; paragraph beginning As against this.

What may be called the second world -the world of the mind - becomes, on the human level, more and more the link between the fis t and the third world: all our actions in the first world are influenced by our second-world grasp of the third world. --- kb/sources/.snapshots/popper-epistemology-without-a-knowing-subject-1968.md @ sha256:11d8a52e58a660a115f877db34a4472af0b52c9a851ba7883dfdc95f49efca25 — Section 9, printed p. 371; paragraph beginning What may be called the second world.

Thus I do admit that in order t o belong t o the third world of objective knowledge, a book should - in principle, or virtually - be capable of being grasped (or decyphered, or understood, or ‘known’) by somebody. But I do not admit more. --- kb/sources/.snapshots/popper-epistemology-without-a-knowing-subject-1968.md @ sha256:11d8a52e58a660a115f877db34a4472af0b52c9a851ba7883dfdc95f49efca25 — Section 3, The objectivity and the autonomy of the third world, printed p. 342; paragraph beginning Thus I do admit.

My schema works through error elimination, and on the scientific level through conscious criticism under the regulative idea of the search for truth. --- kb/sources/.snapshots/popper-epistemology-without-a-knowing-subject-1968.md @ sha256:11d8a52e58a660a115f877db34a4472af0b52c9a851ba7883dfdc95f49efca25 — Section 5.2 on Hegel, printed p. 351; sentence immediately following the evolutionary schema P1→TT→EE→P2, in the contrast with Hegel's dialectic.

Language, the formulation of problems, the emergence of new problem situations, competing theories, mutual criticism by way of argument, all these are the indispensible means of scientific growth. --- kb/sources/.snapshots/popper-epistemology-without-a-knowing-subject-1968.md @ sha256:11d8a52e58a660a115f877db34a4472af0b52c9a851ba7883dfdc95f49efca25 — Section 4, printed p. 347; paragraph following the two senses of 'knowledge', immediately before the sentences on the descriptive and argumentative functions of human language.

Connections Found

The source is a conceptual foundation for tentative theory and theory builder, which cites it for condition 1 (a descriptive language as a condition of criticism), condition 2 (action guided by objective knowledge), and the reflective qualifier. Its tentative theory is a proposed solution open to criticism; it does not stipulate separately editable repair locations. Commonplace treats that finer structure as addressability, a design commitment above condition 1's minimum.

It also clarifies a boundary in representational form: a theory's content, its linguistic formulation, and a consumer's way of processing that formulation are distinct questions. Section 6 defends criticism of informal mathematical arguments while denying that mathematics is merely a formal language game. This supports distinguishing linguistic criticism from mechanically assigned consequences; it does not show that linguistic formulation is dispensable for scientific criticism.

The unread-book and computer-generated-table examples supply a contrast with the evidence that a system used and criticized a theory. Potential intelligibility suffices for the content's objective status in Popper's account. A system that consumes and criticizes a theory (conditions 2 and 3) is a further requirement for calling it a theory builder, and improvement in capacity is a separate learning claim beyond membership. A retained object can therefore count as objective knowledge without establishing that any particular system learned from it.

Learning Claims (our opinion)

Popper's mechanism is problem-directed conjecture followed by attempted error elimination, producing new problems rather than merely accumulating accepted statements. Error elimination may take the form of critical discussion or experimental tests. At the human scientific level, language enables competing theories to be considered and rejected while their proponents survive. In sections 8–9, bodily adaptations and behavioral expectations are biological analogues or predecessors of articulated theories; discovery, imitation, and habituation are treated as forms of trial-and-error learning. These broader claims should not be collapsed into the narrower claim about scientific criticism.

The theory-builder definition takes this cycle as its schema, restricted to the scientific level. Its four conditions follow Popper's account there: theories stated in a descriptive language (condition 1), guiding action (condition 2), eliminated or revised through criticism that may be argument as well as experiment (condition 3), and kept so that new problems start later work (condition 4). Replacement of a theory whole and the discovery of new problems count, and no successful patch is required; whether the process improves capacity for future action is a separate question the definition leaves to test. His content can also include consequences no thinker has yet grasped, so neither a retained document nor a successful revision exhausts the growth of knowledge he describes. The essay supports separating tentative status and builder membership from finer addressability: condition 1 asks only that some unit say something, and a finer grain is a design commitment above that minimum.

The source makes two different contributions to the occasion. It directly supports rejecting formalization as a prerequisite for critical theory work: section 6 makes informal proofs and argumentative language central. It does not support simply dropping externalization from Popper's account of scientific criticism. Section 4 explicitly requires an exosomatic descriptive language, meaning language developed outside the body as a tool; written language is a further development. This is broader than a separately stored software artifact, but it remains a substantive language requirement. His broader biological and subjective accounts allow learning beyond that scientific case. The definition settles the question this leaves for artificial systems: criticism needs a stated unit to aim at (condition 1), and its result must shape the next round (condition 4). How long that result persists is graded, not a condition: stated criticism that revises a theory within one episode and is then discarded makes a theory builder at the lowest grade, and discarding it ends that builder.

Our further inference concerns the effective update space. Popper's evolving languages, new problems, and new mathematical constructions suggest that learning can change what it is able to formulate. They offer a reason to question a permanently fixed hypothesis vocabulary, not evidence that a particular learner can revise its own representation or escape its available operations. The essay specifies no computational update space and compares no implementations.

Extractable Value

  1. Separate conjectural status from repair structure. Popper's proposed solution remains tentative because it is open to criticism, not because it has a particular storage form or edit interface. This provides a basis for reviewing where the old theory-refinement definition attached operational addressability to the broader epistemic term. [deep-dive]
  2. Separate content, formulation, and formalization. Objective consequences can exceed a thinker's grasp, while their critical examination uses linguistic formulations. Section 6 shows why informal argument belongs in this account without making natural language a formally adjudicated representation. The distinction applies across prose theories, mathematical arguments, and software encodings. [quick-win]
  3. Keep the language requirement at its stated scope. Scientific criticism receives a stronger linguistic requirement than biological adaptation or subjective learning in general. This blocks both a universal external-file requirement and an unsupported claim that Popper made scientific criticism wholly independent of linguistic formulation. [quick-win]
  4. Distinguish objective availability from demonstrated learning. Unread or machine-generated contents can be knowledge in Popper's sense. That status alone supplies no evidence of interpretation, correction, or later behavioral use by a learner. This sharpens how a KB should report the value of retained material. [quick-win]

Limitations (our opinion)

The central claims are philosophical arguments, not measurements of learning efficacy. The destroyed-library scenarios illustrate the causal importance of retained information, but do not isolate the metaphysical claim that contents form a third world: a simpler account based on recoverable information could also predict different recovery prospects. The mathematical examples more specifically motivate unintended consequences, yet do not establish every extension of the ontology.

Popper argues strongly for language in scientific criticism while also recognizing extra-linguistic objects and prelinguistic adaptation. These distinctions prevent using the essay either as a proof that all learning requires external artifacts or as a proof that fully internal, nonlinguistic scientific criticism suffices. It supplies no evidence about LLM weight updates, internal deliberation, or the reliability of model interpretation of prose.

The biological analogy does not establish a single experimentally demonstrated mechanism across evolution, habituation, and scientific argument. The essay's attacks on subjective probability and physics are announced rather than fully defended here. Its account of comparative truthlikeness likewise depends on arguments elsewhere; this observation does not independently establish that measure or its applicability to KB evaluation.

Maintain the separation among tentative theory, theory builder, and addressable theory: tentative status and builder membership rest on condition 1's minimum, not on finer addressability. Preserve Popper's distinction between linguistic criticism and formalization.