Epistemology: A Survey of Knowledge and Its Justification

¶Knowledge and Information: The Fundamental Difference

Information is a difference that makes a difference — a distinction that reduces uncertainty. The phrase, originating with Gregory Bateson and later connecting to Claude Shannon’s mathematical theory, captures something essential: information is objective in the sense that it does not require interpretation. A thermostat responds to information (temperature difference) without meaning or understanding. Shannon quantified information as uncertainty reduction; Bateson emphasized that information requires a receiver capable of registering distinctions. Both insights hold.

But knowledge is not the same as information. A person might possess all available true information about a domain — every fact, every law, every principle — and yet lack knowledge of how to act within it. A doctor might know all the symptoms and all the treatments but still lack knowledge of how to diagnose a patient in front of her. A scientist might know all the theory but lack knowledge of how to conduct an experiment. A student might know all the definitions but lack knowledge of how to think.

The gap between having information and having knowledge is the domain of epistemology. Information flows, is transmitted, decays. Knowledge settles in a person or a community. Knowledge is the capacity to use structured information to act, to judge, to recognize patterns in new contexts, to make warranted claims. This survey examines how philosophy has accounted for that gap.

¶The Classical Definition: Justified True Belief

Western epistemology begins with Plato’s Theaetetus, a dialogue in which Socrates asks Theaetetus directly: what is knowledge (epistēmē)? They examine two proposals and reject both: knowledge is not perception (perception is private, fleeting, individual, bound to the moment), and knowledge is not mere opinion (doxa). Knowledge requires logos — intelligible structure, rational account, something that can be articulated and examined.

From this emerges the classical definition that dominates for 2,300 years: knowledge is justified true belief. Three parts, all necessary:

• Belief: a mental commitment to a proposition
• Truth: the belief matches reality
• Justification: the belief is grounded in reasons

Each component is essential. Belief without truth is error. Truth without belief is mere luck (you might utter a true statement while meaning something else entirely). Justification without truth is illusion (you might be defended by reasons but wrong about the world). All three together constitute knowledge.

This account remained nearly unchallenged until Edmund Gettier’s 1963 paper, which is three pages and demolishes the definition. Gettier constructed counterexamples: cases where someone has a justified true belief but lacks genuine knowledge. The classic case: Smith justifiably believes that Jones owns a car because Jones has always owned one. Smith also justifiably believes that the person who gets the promotion has ten coins in his pocket. By sheer coincidence, Smith himself gets the promotion and Smith himself has ten coins in his pocket. So Smith has a justified true belief that “the person who got the promotion has ten coins in his pocket” — but intuitively, Smith does not know this. He arrived at it through good reasoning that happened to be true by accident.

Gettier’s problem opened epistemology. Justified true belief is not sufficient. Philosophers have proposed various solutions: adding a fourth condition (further constraining justification), redefining the concept of justification itself, arguing that knowledge is a relation to the world rather than a mental state, or dissolving the problem by abandoning the mental-state framework entirely.

What matters is that the Gettier problem reveals a deep structure: you can have perfect information, justified reasoning, and true conclusions and still lack knowledge if your chain of reasoning is defeated. The appearance of justification can mask hidden defects in warrant.

¶Polanyi: Knowledge We Cannot Tell

Michael Polanyi (1966) inverts the question. Rather than asking “What makes a belief count as knowledge?” he asks “What do we actually know when we know?”

The answer: we know more than we can tell. A person recognizes a face in a crowd of thousands but cannot articulate the rules by which she recognizes it. A craftsperson judges the quality of work through touch but cannot reduce that judgment to explicit criteria. A musician feels the right tempo without calculating it. A reader understands the tone of a passage without applying linguistic rules. This is tacit knowledge: embodied, practical, implicit.

Polanyi distinguishes between subsidiary awareness and focal awareness. When you ride a bicycle, you are focally aware of your destination; you are subsidiarily aware of balance, pedaling, road surface. You attend from these subsidiary cues rather than to them. Knowing is the integration of subsidiaries into a focal whole. The more expert you become, the more the mechanics recede into subsidiary awareness and the more you focus on the task itself.

This has deep implications. No explicit instruction can fully capture tacit knowledge. A manual cannot teach you to play violin; you need practice under guidance. A recipe cannot teach you to bake; you need to feel the dough’s resistance. A diagnostic procedure cannot teach you to diagnose; you need apprenticeship. The knowledge of what experienced practitioners know is structurally resistant to full articulation.

Yet Polanyi is not advocating ignorance or silence. He argues that tacit knowledge should be acknowledged and that conditions for its transmission should be created — through mentorship, through immersion in practice, through the prolonged exposure that allows knowledge to be internalized without explicit articulation. Tacit knowledge becomes articulate through practice in community, not through instruction alone.

¶The Spiral of Knowledge Creation

Ikujiro Nonaka and Hirotaka Takeuchi (1995) reconceptualize knowledge in organizations through a model they call SECI: Socialization, Externalization, Combination, Internalization. This is not a linear progression but a spiral that repeats.

Socialization (tacit to tacit) is knowledge transmission through shared practice — apprentices working alongside masters, teams building trust and shared mental models, practitioners learning from each other without explicit instruction.

Externalization (tacit to explicit) is the difficult work of articulating what was implicit — a designer sketching intuitions, a scientist writing a hypothesis, an engineer documenting procedure. Externalization is generative. The struggle to make tacit knowledge explicit often reveals new relationships and produces new insights.

Combination (explicit to explicit) is the assembly of explicit knowledge into new structures — connecting theories, synthesizing findings, merging datasets, building new frameworks from existing pieces.

Internalization (explicit to tacit) is the embodiment of explicit knowledge back into practice — a worker reading a procedure and performing it until it becomes second nature, a student reading theory and then recognizing it in the world.

These four phases are not steps in a process that terminates. Knowledge spirals: new explicit knowledge becomes practice (tacit), which generates new questions, which require new externalization, which produces new explicit knowledge. The spiral never concludes. Knowledge evolves through this endless cycle.

The model has a crucial consequence: not all knowledge transmission happens through documents or explicit instruction. Socialization — the tacit-to-tacit transmission — is irreducible. A community transmits knowledge through its practices, its ways of working, its standards of judgment. Much of what matters most in any discipline lives in this tacit layer.

¶Dewey: Knowledge as Process, Not Possession

John Dewey (1938) fundamentally reframes epistemology. Knowledge is not a mental state possessed by a mind. Knowledge is the outcome of inquiry. To know is not to possess something; to know is to have gone through a process.

Dewey begins with an indeterminate situation: a problem, a confusion, something that doesn’t fit expected patterns. Inquiry is the process of transforming that indeterminate situation into a determinate one, resolving the confusion so that action becomes possible.

The process has phases:

1. Recognition of the indeterminate situation — something is wrong, unclear, or puzzling
2. Definition — taking note of facts and ideas that seem relevant, narrowing the problem
3. Invention of hypotheses that might resolve the indeterminacy
4. Testing of hypotheses against experience, exposure to countervailing evidence
5. Settlement — the indeterminate situation becomes determinate and action becomes possible

A known fact is not a mirror of reality but a warranted assertion produced by this process. It is what you can rely on because it has been tested against resistance. It is provisional — future inquiry might refine or overturn it — but it is reliable for present purposes.

This is not relativism. Some beliefs are warranted; others are not. A warranted belief emerges from careful inquiry, exposure to critical testing, and provisional settlement. An unwarranted belief emerges from prejudice, is accepted without testing, or settles prematurely without adequate evidence.

Dewey’s vocabulary is crucial. He uses “settlement” and “warranted assertion” rather than “truth” and “correspondence with reality.” Not because truth doesn’t matter, but because the knower never has unmediated access to reality. What the knower has is the outcome of a process of inquiry conducted in a specific context. That outcome — reliable for now, subject to refinement — is what Dewey calls knowledge.

¶Peirce: Knowledge in the Community and the Long Run

Charles Sanders Peirce radicalizes Dewey’s insight. Knowledge is not what any particular inquiry has reached. Knowledge is what inquiry would converge on given unlimited time and adequate resources. It is what would be reached in the limit — the outcome toward which inquiry tends as it continues indefinitely.

This has three consequences that reshape epistemology:

Fallibilism: No particular belief is certain. All knowledge is provisional. You might believe something based on the best available evidence, but future inquiry might overturn it. This is not a weakness of knowledge but its structure. The history of science shows that what was warranted knowledge becomes recognized as incomplete or partially false. Fallibility is built in.

Communality: Knowledge is not private. It emerges through the community of inquiry — people testing each other’s claims, offering objections, refining methods. You might have a private belief; you cannot have private knowledge. Knowledge is what would survive the objections of the competent inquirers. The test is not what I find convincing but what would convince the best practitioners in the field.

The long run: Knowledge is defined by the limit, not by any particular point in time. This is not a psychological claim (we cannot wait forever). It is a definition: knowledge is “the opinion which would be fixed in the limit of their condition when the mind of man upon it should become finally saturated” (Peirce, “How to Make Our Ideas Clear”).

Peirce’s semiotics adds structure. A sign is something that stands for an object to an interpretant. The meaning is not in the sign itself but in the relation: sign, object, and interpretant co-determine each other. A statement is a sign; it refers to something (a fact, a state of affairs, a concept) to an interpreter. Its meaning is constituted by how interpreters relate to it.

Applied to knowledge: a claim is a sign. It refers to something (a phenomenon, a principle, a state of affairs). Its meaning and truth emerge through how interpreters — readers, practitioners, the community of inquiry — test it, use it, refine it. The claim’s warrant is not internal (does it follow from my premises) but relational (how does it function in the community of inquiry).

¶Bateson: The Pattern That Connects

Gregory Bateson (1972) moves epistemology from the individual mind to the relational field. He begins with information as “a difference that makes a difference” but then asks: what makes information meaningful?

The answer: pattern. Information alone — a stream of disconnected data points, facts without structure — does not constitute knowledge. What makes information meaningful is the pattern — the structure that connects multiple pieces of information into a coherent whole. You do not know something from isolated facts; you know it by understanding the pattern that makes those facts cohere.

Knowledge is ecological. It is not possessed by an individual mind alone but distributed across relationships. A scientist knows through her instruments, her laboratory, her colleagues, her field’s history of problems and solutions. A student knows through the teacher, the peers, the books, the lived experience of learning. A craftsperson knows through her materials, her tools, her community of practice. You cannot extract knowledge from the network of relationships and preserve it in isolation.

Bateson also adds: “The map is not the territory.” A document represents reality but does not contain reality. This is not a flaw of documentation but its nature. Maps are useful precisely because they are abstracted, simplified, organized for a purpose. The gap between the map and the territory is the gap between representation and reality. All knowledge involves this gap.

Understanding requires grasping the pattern, not accumulating facts. A student who learns a list of historical dates but does not understand how events connect has information but not knowledge. A technician who follows a procedure without understanding its purpose has instructions but not knowledge. Knowledge is the pattern that makes the parts intelligible.

¶Haraway: Situated Knowledge and Accountable Objectivity

Donna Haraway (1988) asks a central question: Can there be objectivity? The traditional answer — objectivity as a view from nowhere, a position of absolute impartiality — is impossible. There is no position from which you see everything without perspective. Every knower has a location: a body, a history, investments, limitations of sight.

But Haraway’s conclusion is not that objectivity is impossible. Rather, objectivity is accountable positioning. A knowledge claim is objective when it is made from a specific, named, accountable position. The scientist must declare her interests, her methods, her instruments, her prior commitments — all the things that structure what she can see and what she cannot. The map-maker must state what she is mapping for: navigation, surveying, artistic representation?

This is radical not because it claims neutrality is impossible, but because it makes that impossibility productive. By acknowledging her position, the knower becomes responsible for what her perspective reveals and what it obscures. By specifying the perspective, she makes it possible for others to evaluate the knowledge, to translate it for their own purposes, to use it with awareness of its limitations.

Haraway also introduces a political dimension. Who gets to know? Whose knowledge counts? Knowledge systems are not innocent. They include and exclude. A library that documents one philosophical tradition and ignores others has made a choice. Its objectivity is not increased by claiming neutrality; it is clarified by acknowledging that the choice has been made and stating what it costs.

Knowledge is objective not by denying perspective but by owning it, making it explicit, and taking responsibility for the resulting knowledge.

¶Indigenous Knowledge: Relational and Reciprocal

Robin Wall Kimmerer (2013) articulates how knowledge understood in Indigenous traditions differs fundamentally from the epistemologies surveyed above. Knowledge is not primarily propositional — not a collection of true beliefs about facts. Knowledge is relational.

Knowledge is constituted in the relationship between the knower and what is known, between persons and the land, between generations. In the Haudenosaunee tradition that Kimmerer draws from, knowledge is inseparable from responsibility and reciprocity. If you know the plants, you know you are responsible for them. If you know a person, you are in relationship with them and owe respect. Knowing is the establishment of a covenant, not the achievement of detached understanding.

Knowledge is place-based. The knowledge of how to live well depends on the particular ecology of the place you inhabit. General principles matter, but they must be tested, adapted, and refined in the specific conditions of your land. This is not relativism (all knowledge is equally valid in any place). It is contextualism: knowledge is always knowledge in a place, and moving knowledge to a different place requires transformation.

Knowledge is often embodied and oral. The knowledge of how to harvest cedar is in your hands as much as your mind. The knowledge of a people lives in stories, in ceremonies, in the land itself, not primarily in written texts. Recording this knowledge in text changes its character. Something essential is lost — the lived context, the embodied skill, the ongoing dialogue with the land and the community.

This is not an argument against documentation. It is an acknowledgment that different technologies for preserving knowledge have different affordances and costs. Writing makes knowledge portable and stable; it can also abstract it from its place and context. For any knowledge system, the question is: what knowledge do we capture, and what do we necessarily leave out?

¶Whitehead: Understanding as the Grasp of Pattern

Alfred North Whitehead (1929, 1938) offers a framework that integrates much of the above. Understanding is not the accumulation of facts. Facts apart from understanding are trivial. Understanding is the grasp of pattern — the recognition of how facts relate, how principles apply, how systems cohere.

Whitehead distinguishes two types of understanding:

Analytic understanding divides wholes into parts, revealing structure. It is the process of breaking down, examining components, tracing relationships.

Constructive understanding grasps how parts relate and function together. It is the process of seeing unity, recognizing how principles operate, understanding purpose.

Both are necessary. Analytic understanding without constructive understanding produces fragmentation and loss of meaning. A botanist can dissect every part of a flower but understand nothing if she cannot see how the parts function together in reproduction and pollination. Constructive understanding without analytic precision produces vagueness and confusion. Philosophical hand-waving without rigor is not understanding.

For knowledge: information provides the raw material, the individual facts and data points. Knowledge is the pattern that organizes those facts into something coherent and intelligible. The student who memorizes facts but grasps no pattern has information without knowledge. The practitioner who understands how to apply principle to case has knowledge.

¶Three Perspectives on Knowledge

The sources above suggest three ways to think about what knowledge is:

Knowledge as structure: Knowledge is information organized by principles, facts ordered by laws, propositions integrated into a coherent system. This is the emphasis of classical epistemology and Whitehead. The object of knowledge is the pattern itself.

Knowledge as process: Knowledge is the outcome of inquiry, the result of testing and refinement. This is the emphasis of Dewey and Peirce. Knowledge is not a possession but an achievement, not a state but a process that settles into a provisional state.

Knowledge as relation: Knowledge is constituted in a web of relations — relations between the knower and what is known, between propositions and the world, between individuals and community, between people and land. This is the emphasis of Haraway, Kimmerer, and Bateson. Knowledge is always knowledge in relation to.

These are not competing theories. They are three aspects of a single phenomenon, three angles on what it means to know something. Effective knowledge — knowledge that is robust, that can be used, that can be transmitted — engages all three: it is structured, it is hard-won through process, and it is embedded in relationships.

¶The Architecture of Warranted Knowing

From this survey, knowledge has four key characteristics:

Knowledge is structured and integrated. Information is raw — a distinction, a data point, a measurement, an isolated fact. Knowledge is organized — facts ordered by principles, patterns recognized, relationships understood. The move from information to knowledge is the move from scattered data to coherent structure.

Knowledge is warranted through process. Information exists; knowledge must be earned. Justification is grounded in reasons — in experience, in method, in community testing. Inquiry tests against resistance. Knowledge is what survives scrutiny.

Knowledge is practical or relational. It is not primarily the accumulation of true sentences, though true sentences can express it. It is the capacity to act intelligently, to judge wisely, to recognize a pattern in a new context, to understand one’s place and one’s responsibility. Even theoretical knowledge (pure mathematics, philosophy) is knowledge because it is grasped by a mind embedded in a tradition that keeps it alive.

Knowledge is situated. It is always knowledge from somewhere, produced in a specific time and place by a community with particular investments and limitations. This does not make it relative or arbitrary. It makes it accountable — responsible for what it reveals and what it hides.

These characteristics suggest what knowledge systems must attend to: not just the content of claims but their warrant — the process that produced them, the community that tested them, the place and time from which they emerge. A claim that has been tested, refined, integrated into a larger pattern, and found reliable in practice has different standing than a claim asserted without process.

---

See also: Documentation Theory: A Survey

Related concepts: Information, Knowledge