What counts as knowledge?

TOK starts with the knower. Knowledge doesn’t just sit in a book or database; someone has to produce, interpret and communicate it, then decide whether it should be accepted.

One useful starting point is that knowledge is a cognitive achievement through which a knower successfully understands, represents or performs something. The definition is deliberately broad, allowing for several kinds of knowing:

• Propositional knowledge is knowledge that a claim is true, such as knowing that a particular medicine reduces infection rates.
• Procedural knowledge is knowledge of how to perform an action or process, such as knowing how to calibrate an instrument.
• Knowledge by acquaintance is familiarity gained through direct experience of a person, place, object or condition, such as knowing what severe stage fright feels like.

These forms may overlap, but they aren’t interchangeable. Someone could explain every stage of riding a bicycle and still be unable to stay upright on one. A skilled cyclist, by contrast, might struggle to describe the physical principles involved. TOK should therefore avoid reducing all knowing to a collection of true sentences.

Knowledge, belief and opinion

A belief is a mental attitude in which a person accepts a claim as true. Propositional knowledge requires belief, but belief alone isn’t enough. A person may hold a belief confidently even when it is false or poorly supported.

An opinion is a judgement expressing a person’s assessment, preference or interpretation where decisive justification is unavailable or not expected. Describing something as an opinion doesn’t make it worthless. An informed judgement grounded in extensive experience is epistemically stronger than a passing preference, although neither automatically counts as knowledge.

The traditional account of knowledge as justified true belief remains useful. It focuses attention on three tests: the knower accepts the claim, the claim is true, and adequate grounds support its acceptance. Yet people can sometimes arrive at true beliefs through luck. Consider someone who trusts a stopped station clock at the one moment when it shows the correct time. The belief is true and seems supported, but its success is accidental. Knowledge appears to require a sufficiently reliable connection between the grounds and the truth.

Truth is a property of a claim that succeeds in representing how matters actually are. Truth differs from sincerity, popularity and usefulness. A speaker can sincerely make a false statement. A community may widely accept an error, while an inaccurate model can still prove useful within a narrow range.

Certainty and doubt

Certainty is a degree of confidence at which a knower regards error as impossible or negligible. Two questions need to be kept separate: how sure is the knower, and how secure is the claim? Psychological certainty may be intense even when the evidence is poor.

Absolute certainty sets a demanding standard. Observations may be mistaken, memories can change, concepts may be ambiguous and future conditions might differ from past ones. None of this forces us into radical scepticism. In ordinary inquiry, well-supported knowledge can be dependable without being infallible. Often, the sensible question isn’t “Can I prove this beyond every imaginable doubt?” but “How much confidence is warranted for this purpose?”

Scepticism is a method or position that withholds acceptance until specified standards of justification are met. Disciplined scepticism puts knowledge to the test; indiscriminate disbelief simply avoids judgement. Doubt earns its place when it exposes a real weakness in the evidence, reasoning or method.

Usefulness, understanding and open questions

Some knowledge has immediate practical value. It may predict storms, prevent structural failure or guide public policy. Other knowledge deepens understanding without offering any obvious short-term application. Understanding is a cognitive achievement in which a knower grasps significant relationships, causes or implications rather than merely retaining isolated information.

Calling one kind of knowledge “more useful” prompts a further question: useful to whom, for what purpose, over what period, and at what cost? Technical knowledge might solve a practical problem while producing environmental or political consequences. Knowledge pursued out of curiosity may later gain unexpected applications. Usefulness is a value judgement rather than a permanent feature stamped onto a claim.

Open questions show where present understanding reaches its limits. A question may remain unresolved because evidence is unavailable or methods are inadequate. Concepts may also be contested, or different values may shape what counts as an acceptable answer. Recognising these limits doesn’t mean that inquiry has failed; it shows where knowledge currently ends.

Knowing oneself and knowing through language

Self-knowledge is knowledge in which the knower is also the object being understood. First-person access gives someone unusual evidence about their own intentions, sensations and emotions, but accuracy isn’t guaranteed. People rationalize decisions, misremember motives and adopt descriptions provided by their communities. At times, others can spot patterns in our conduct that we fail to notice ourselves.

Some forms of knowing seem possible without language, such as recognising a face, navigating a familiar room or carrying out a practised movement. Language, however, allows complex classification, reflection and communication. The categories available in one language may highlight distinctions that another language handles differently. Language doesn’t imprison thought, but it does influence which features are easier to notice, remember and discuss.

The central scope question, then, is not simply “What do I know?” It is also “What kind of knowing is this, what is it for, and what are its limits?”

What shapes a knower’s perspective?

A perspective is a standpoint from which a knower selects, organizes and interprets information. Experience, language, education, social position, disciplinary training, interests and values all shape it. Perspective isn’t simply another word for “bias”. Every act of knowing begins from a standpoint, though some standpoints are far better informed than others.

Culture is a learned system of shared meanings, practices, norms and symbols through which a community organizes experience. It can shape which questions matter, which sources are trusted and which explanations seem plausible. Yet members of the same community do not all think alike. Nor does cultural influence suggest that truth changes whenever communities disagree.

Values are standards of importance or worth that guide selection, judgement and action. Their influence begins before anyone reaches a conclusion. They help decide which research gets funded, what counts as an acceptable risk, whose testimony enters the record and which outcomes matter.

Rather than viewing perspective as one distorting lens, it makes more sense to see it as a set of interacting filters. Someone may encounter the same event through personal experience, inherited concepts, community expectations, institutional roles and the evidence available to them.

Facts, interpretations and explanations

A fact is a state of affairs that is the case, or a well-established claim reporting such a state of affairs. Perspective does not make facts true. Knowers still have to decide, however, which facts are relevant and how those facts connect.

An interpretation is an account that assigns meaning or significance to evidence. The same basic facts can support different interpretations when people emphasize different contexts, causes or implications. That doesn’t make every interpretation equally good. A strong one accounts for the available evidence, recognizes counter-evidence and avoids indefensible assumptions.

An explanation is an account that makes an event, pattern or relationship intelligible by identifying relevant causes, mechanisms, reasons or principles. The question being asked shapes the explanation. For example, an account of why a city expanded could focus on geography, economic incentives, political decisions or individual choices. Such explanations may complement each other. They may also compete when they assign incompatible causal roles.

Complete neutrality is hard to achieve because any presentation of information involves selection and framing. A technically accurate graph, for instance, may use a scale that exaggerates a change. A historical narrative has to choose where its story starts. Objectivity, then, should not require the absence of every perspective.

Objectivity is an epistemic standard that limits the influence of irrelevant personal preferences through publicly examinable evidence and procedures. People often achieve it socially, by checking one another’s methods, assumptions and conclusions. A perspective may even strengthen objectivity by revealing something the dominant viewpoint missed.

Other knowers and shared knowledge

A large part of what one person knows comes from others. No individual personally checks every map, medical claim or account of the past. Testimony is evidence conveyed by another person’s assertion or report. We cannot do without it, but its strength depends on the speaker’s access, competence, honesty, independence and record of reliability.

Communities preserve knowledge using language, institutions, archives, rituals, technologies and teaching. They also set standards for who can speak with authority. These systems allow knowledge to build over time, though they may preserve exclusions and errors alongside achievements.

Consensus needs careful treatment. Consensus is broad agreement among the relevant members of a group. It counts as evidence when those members have suitable expertise, use strong methods and are free to challenge each other. But truth is not simply whatever most people accept. A majority might rely on the same misleading source or face the same social pressure.

Power is the capacity of a person or institution to shape decisions, behaviour or the conditions under which knowledge is produced and distributed. It influences whose questions get attention, which classifications become official and who has access to information. This does not reduce every knowledge claim to an exercise of power. Instead, it gives us a reason to examine both the evidence behind a claim and the system that grants it authority.

Encountering different perspectives

Empathy is a capacity to understand another person’s experiences or feelings from that person’s standpoint. Imagination is a capacity to form possibilities or representations not immediately given in present experience. Used together, they can help a knower reconstruct how a situation might look to somebody else.

Neither one guarantees an accurate understanding. An attempt to imagine another person’s experience can turn into projection, particularly when imagination replaces listening. Empathy works most reliably when testimony and context constrain it, and when the knower is prepared to revise an initial interpretation.

Different perspectives become especially useful when they uncover hidden assumptions. A productive comparison asks:

• What does each perspective notice or leave out?
• What evidence is available from that position?
• Which concepts and values structure the account?
• What would lead its supporters to revise it?

Historical change offers another test. Knowledge isn’t automatically better just because it is newer. Improvement has to be shown through standards such as greater explanatory reach, more accurate prediction, stronger evidence, reduced contradiction or fairer inclusion of relevant experience. Since those standards can also be disputed, any judgement of progress needs explicit criteria.

How knowledge is acquired

Knowers build knowledge through capacities and practices that work together: perception, memory, reasoning, emotion, language, imagination, testimony and skilled action. They aren’t separate channels. Concepts shape how we interpret observations. Later knowledge can reconstruct memory, while reasoning relies on premises gained elsewhere. Testimony, meanwhile, has to be understood through language.

Evidence is information or experience that bears on whether a claim should be accepted. It counts as evidence only in relation to a particular claim. Muddy footprints, for instance, might support one explanation of an event while having no relevance to another. The quality of evidence depends on its relevance, reliability, sufficiency and independence.

Justification is the provision of adequate epistemic grounds for accepting a claim. The claim itself, along with the stakes involved, partly determines what counts as adequate. A casual statement about tomorrow’s lunch needs less support than a decision that could put lives at risk.

A reason is a consideration offered in support of a belief, judgement or action. For a reason to be good, it must be relevant to the conclusion and linked to it through defensible reasoning. Repetition, confidence or emotional force may make a reason sound persuasive without making it good.

When assessing a knowledge claim, follow the route from the claim back to its support:

1. State clearly what is being claimed.
2. Identify the evidence and where it comes from.
3. Make the assumptions linking the evidence to the conclusion explicit.
4. Look for plausible alternative explanations.
5. Ask what evidence would count against the claim.
6. Set the level of confidence according to the strength of the support.

Language as a tool

Language is a rule-governed symbolic system that enables people to represent, communicate and organize meaning. Through language, knowledge can move beyond immediate experience and endure across generations. Definitions, classifications and metaphors influence inquiry too, since they bring certain similarities or differences into sharper focus.

Ambiguity can enter through language. One word may have several meanings, and a technical term can be confused with its everyday counterpart. Wording also frames judgement. Calling the same policy “protection” rather than “restriction” directs attention towards different values. The answer isn’t to abandon language, but to clarify terms and compare other ways of expressing the claim.

Translation makes clear that meaning involves more than replacing one word with another. Translators interpret context, tone and cultural associations. Some loss cannot be avoided, but communication across languages is still possible. Translated knowledge claims may therefore need an explanation of the choices behind the translation.

Material and digital tools

A tool is an artefact or structured technique that extends a knower’s capacity to observe, record, calculate, model, communicate or act. A telescope extends observation, an archive extends memory, statistical software extends calculation, and digital networks broaden access and communication.

Tools don’t act as neutral containers. Design choices determine what a tool can measure or display, and its limitations affect the conclusions that can be drawn. A sensor might reveal patterns that unaided perception cannot detect, but its calibration choices and detection limits then form part of the justification. Without a particular tool, a community may be unable to investigate a question at all. Once it has that tool, even the questions worth asking may change.

A model is a selective representation of a system used to describe, explain, predict or explore it. Every model deliberately leaves out some features. Its value rests on whether that simplification fits the purpose, rather than whether the model reproduces every detail. Used under ordinary conditions, a model may work well; beyond its intended range, it may mislead.

Digital systems raise further questions. Search rankings affect which sources a knower sees. Automated recommendations reinforce some patterns of attention, while generated outputs can imitate authority without possessing understanding. Using a tool efficiently doesn’t relieve the knower of responsibility for checking sources, assumptions and limitations.

Expertise and authority

An expert is a person who has developed reliable, specialized competence through sustained training and practice in a field. Experts matter because no individual can repeat every inquiry independently. Deferring to expertise can therefore be a rational choice rather than a passive one.

That expertise still has limits. A specialist may be highly credible on a technical issue yet no better informed than anyone else when making an unrelated moral or political judgement. When assessing expert testimony, ask whether the relevant community recognizes the person’s competence and whether the expert explains the evidence behind the claim. Consider whether independent specialists broadly agree and whether any conflicts of interest have been disclosed.

Expert disagreement doesn’t make expertise worthless. It may arise from uncertain evidence, different models, distinct values or questions near the boundary of current knowledge. A responsible knower looks at how the disagreement is structured instead of choosing whichever expert offers the most convenient conclusion.

Distinguishing stronger from weaker claims

Technical vocabulary alone does not turn a claim into knowledge. Check whether the method is appropriate and whether its limitations have been acknowledged. The conclusion should remain proportionate to the evidence and open to correction.

This is where reliability is a property of a source or method that produces dependable results across relevant occasions. Reliability and truth are not the same. A reliable method may fail in one particular case, while an unreliable one may occasionally produce the right answer by chance. Knowledge becomes more secure when different, independent methods arrive at the same conclusion.

Ethical limits on inquiry

Ethics is the systematic examination of what people ought to do and what kinds of conduct are right, wrong, just or responsible. Inquiry affects participants and communities, as well as other living beings and environments, so ethics has a direct place in knowledge production. A method may produce accurate information and still be morally unacceptable.

Ethical constraints can limit the ways in which knowledge is sought. Concerns include informed consent, avoidable harm, privacy, deception, exploitation, environmental damage and the unfair distribution of risk. These constraints don’t necessarily work against knowledge. By protecting trust, they can improve inquiry and prevent people from being treated as mere instruments.

A right is a justified entitlement that places corresponding duties or limits on others. For example, a person’s right to privacy may restrict a researcher’s access to information, even if that information could answer an important question. In other cases, the need to protect the public from serious harm may justify carefully limited access. Ethical judgement requires competing claims to be weighed rather than assuming that secrecy or openness must always win.

Some knowledge should perhaps not be pursued through the methods currently available because the harm caused by obtaining it would be disproportionate to its likely benefit. That doesn’t mean the question must remain forbidden forever. New methods or stronger safeguards may change the judgement.

Does knowledge create obligations?

Responsibility is an obligation to answer for choices, actions or foreseeable consequences. Relevant knowledge can increase responsibility: it removes the excuse of ignorance and gives a person more capacity to act. Someone who knows that a system is unsafe may therefore have duties that an uninformed bystander does not.

Knowledge by itself, however, doesn’t determine what must be done. The knower may have no authority, lack resources or have no safe way to intervene. Responsibilities may conflict too. Disclosing information could protect the public while breaking confidentiality or putting an innocent person in danger.

The duty to communicate depends on several factors:

• how serious and likely the possible harm is;
• how reliable the knowledge is;
• the knower’s role and capacity to respond;
• the rights of those affected;
• the likely consequences of disclosure and silence.

People may also have responsibilities to acquire knowledge. Anyone accepting a safety-critical role ought to learn what is needed to perform it competently. A community facing a preventable hazard may have good reason to investigate. Compulsory acquisition still has limits. Forcing people to learn private facts about themselves, or gathering information about a population without consent, may violate autonomy.

Autonomy is a person’s capacity and justified authority to make informed decisions about their own life. Respecting autonomy takes more than simply presenting information. That information must be intelligible, material choices must be genuine, and there must be no coercion.

Knowledge, ignorance and power

Ignorance isn’t always innocent. Wilful ignorance is a deliberate avoidance of accessible evidence in order to escape an unwelcome conclusion or responsibility. Unlike an ordinary lack of knowledge, it involves resisting opportunities to become better informed.

Institutions may produce ignorance too. They can withhold records, fail to collect data or leave certain groups out of research. Choices about what not to investigate are therefore tied to power. When evidence is missing, this may show neglect rather than indicate that the subject is unimportant.

Unlimited collection isn’t the solution. Gathering information “in case it becomes useful” may expose people to surveillance or future misuse. Responsible inquiry considers who controls the information and who benefits from it. It also asks who carries the risks, how long the information is kept and whether affected people can challenge its use.

Evidence and moral disagreement

Evidence can settle some moral disagreements by correcting mistaken factual assumptions. Two people might disagree about a policy because they expect different consequences; stronger evidence can narrow the dispute. It may also show that a burden falls unfairly on a group that was previously ignored.

Yet evidence alone can’t always decide what ought to be done. People may agree on the facts but give different weight to freedom, welfare, fairness or loyalty. A moral judgement is a normative assessment of what ought to be done or valued. Because moral judgements involve values alongside factual beliefs, they require further ethical reasoning.

Moral disagreement is not therefore arbitrary. Participants can check for consistency, uncover hidden assumptions and compare consequences. They can also ask whether similar cases receive similar treatment and consider the perspectives of those affected. Moral knowledge can remain contestable without becoming a matter of personal taste.

A responsible knower

A responsible knower does more than avoid lying. Responsible knowing means seeking evidence in proportion to the claim, representing uncertainty honestly and acknowledging sources. It also involves listening to affected perspectives, correcting errors and thinking about foreseeable uses of knowledge.

The key TOK connection is that epistemic and ethical quality often reinforce each other. Inclusive inquiry may reveal missing evidence, while transparency can expose weak reasoning. A willingness to revise a claim can prevent harm. Tensions still remain: openness may conflict with privacy, speed with careful verification, and public benefit with individual rights. Each tension must be justified case by case rather than settled with a slogan.