Key TOK terms with definitions, categories, and session links.
Categories
Match:
Date range
Introduced–
Used in–
Importance
Sort
View
▸
Knowledge claim
Core
Core
Any assertion that something is true or that one knows something.
Whenever we say we know something — or make an assertion that something is true — we are making a knowledge claim. TOK asks us to examine these claims carefully: what type are they? How are they justified? What counts as adequate evidence?
A classical definition of knowledge: a belief that is both true and supported by adequate justification.
According to this theory, knowledge is a justified belief that is true. Associated with Plato's Meno and Theaetetus. Edmund Gettier argued in 1963 that justified true belief is not sufficient for knowledge, producing scenarios where all three conditions are met yet we would hesitate to say the person "knows." The debate continues.
Knowledge that depends on the individual's own experience, insight, or skill.
Personal knowledge is difficult to transfer directly to others — it is built through lived experience. Compare with shared knowledge, which is produced and tested collectively. The distinction matters because personal knowledge can influence how we interpret shared knowledge.
Knowledge developed, tested, and held collectively by a community or discipline.
Shared knowledge is not the sum of individuals' personal knowledge — it is knowledge that has been collectively constructed, challenged, and refined. Science, history, and mathematics are paradigmatic examples. Shared knowledge has greater authority than personal knowledge in many contexts, but it is not immune to the influence of culture, power, and perspective.
Knowing how to do something; practical, skill-based knowledge.
Also called "know-how." Procedural knowledge is demonstrated through action rather than statement. You know how to ride a bicycle or solve an equation not by being able to describe it perfectly, but by doing it. This type of knowledge raises interesting questions: can it be fully articulated? Can it be shared as effectively as propositional knowledge?
Knowing that something is the case; factual knowledge expressed as a statement.
Also called "know-that." Propositional knowledge is expressed in declarative sentences: "I know that the Earth orbits the Sun." It is the form of knowledge most directly evaluated by truth conditions, justification, and evidence. Most academic subjects produce propositional knowledge.
Information or data that supports or challenges a knowledge claim.
The key TOK question is: what counts as evidence? Different areas of knowledge have different standards — what counts as good evidence in history (a primary source) may differ from what counts in natural science (a replicable experiment). Evidence is never self-interpreting; it always requires a framework to be understood.
The degree of confidence we have that a knowledge claim is true.
Certainty exists on a spectrum. Mathematics can achieve a form of certainty through proof that empirical sciences cannot match — yet even mathematical certainty depends on the axioms we start from. How certain must a belief be to count as knowledge? This is one of TOK's oldest questions.
A claim that corresponds to reality or is universally valid.
There are multiple theories of truth: correspondence (a claim is true if it matches reality), coherence (a claim is true if it fits with other accepted beliefs), consensus (a claim is true if a community of knowers agrees), and pragmatic (a claim is true if it works). Each captures something important but also faces objections.
The act of explaining or making sense of knowledge or meaning.
Interpretation is unavoidable: data, texts, artworks, and historical events do not speak for themselves. The knower always brings a perspective, framework, and set of assumptions to the act of interpretation. The TOK question is: what makes an interpretation justified rather than arbitrary?
The ability to influence what is accepted as knowledge.
Power shapes what gets counted as knowledge, who gets to produce it, and whose interpretations are taken seriously. Institutions, governments, publishers, and funding bodies all exercise power over knowledge production. The TOK question is: to what extent should we simply accept knowledge on the basis of authority?
Reasoning that shows why a knowledge claim is valid.
Justification is what transforms a mere belief into knowledge — or at least into a well-grounded belief. The standards for justification vary across areas of knowledge: what justifies a claim in mathematics (proof) differs from what justifies one in history (documentary evidence) or in the arts (informed critical judgment).
A statement that clarifies why something is the way it is, resolving a puzzle.
A good explanation moves us from puzzle to understanding. In natural science, explanations may be constitutive (what a thing is made of) or causal (what brought it about). In history, explanations are typically causal and contextual. In the arts, "explanation" may shade into "interpretation." What makes one explanation better than another is a central TOK question.
Seeing or producing knowledge without personal or cultural bias; independent of the observer.
Complete objectivity may be more of an ideal than a reality. The scientific method is designed to cancel out the human element through replication and peer review — but scientists are human beings with perspectives, funding interests, and cultural assumptions. The gap between the natural sciences and other AoKs on objectivity may be smaller than it first appears.
A particular position or angle from which knowledge is viewed.
All knowers have a perspective — shaped by culture, language, experience, and position. This does not mean all perspectives are equally valid, but it does mean we should ask: who produced this knowledge, from where, and with what assumptions? The TOK question is whether some perspectives are more justified than others, and if so, why.
Shared practices, values, and beliefs of a community that shape how knowledge is produced and understood.
Culture influences what questions get asked, which methods are trusted, and what counts as a satisfying answer. Mathematical truths may be culture-independent in their content, but how mathematics developed, which problems were pursued, and who got credit are deeply culturally shaped. The TOK question is: does culture determine knowledge, or merely influence it?
Principles or standards of behaviour that guide how knowledge is used and what counts as worthwhile.
Values are not separate from knowledge — they shape what we choose to study, how we interpret findings, and what we do with the results. The TOK question is: is knowledge influenced by ethical considerations, and if so, does that undermine its validity, or is it unavoidable?
The duties knowers have in creating, sharing, and applying knowledge.
Knowledge brings responsibility: the obligation to seek it honestly, share it accurately, and use it wisely. In mathematics, the user who presents data bears the greatest ethical burden. In science, the researcher who publishes results, and the policymaker who acts on them, share responsibility for outcomes. Where does responsibility begin and end?
A truth check: a claim is true if it corresponds to (matches) observable reality.
The most intuitive theory of truth. "The cat is on the mat" is true if and only if there actually is a cat on the mat. Challenges: how do we access reality independently of our perceptions to check correspondence? What about abstract claims (mathematical truths, ethical principles) that have no obvious physical correlate?
A truth check: a claim is true if it fits consistently with other accepted beliefs.
Coherence theory holds that truth is a matter of how well a claim fits within a web of other accepted beliefs. It avoids the problem of direct comparison with reality but faces a different problem: a coherent set of beliefs could all be false together. Pure mathematics and some areas of ethics may rely more on coherence than correspondence.
A truth check: a claim is true if it is agreed upon by a relevant community of knowers.
Scientific consensus (e.g., climate change, evolution) carries significant weight. But consensus alone is not truth — history is full of cases where the consensus was wrong. The question is: whose consensus, arrived at through what process, counts? Expert consensus reached through open inquiry differs from forced agreement or popular opinion.
A truth check: a claim is true if believing it produces good outcomes in practice.
Associated with American pragmatists William James and John Dewey. A claim is true if it "works" — if acting on it leads to successful outcomes. This has intuitive appeal (Newton's physics is "true" because bridges built with it stay up) but faces the objection that a useful belief can still be literally false.
The TOK framework element exploring the nature and extent of each theme or area of knowledge.
Scope asks: what does this area of knowledge study? What questions can it answer — and what is beyond its reach? How does it fit within the totality of human knowledge? Exploring scope helps us understand what each AoK is uniquely suited for and where it runs into limits.
The TOK framework element exploring how knowledge is produced in each area.
Methods and Tools asks: how do practitioners in this area produce knowledge? What practices, conceptual frameworks, and material or cognitive tools do they use? How have these changed over time with technological development? Different AoKs have strikingly different methods — and those methods shape what kind of knowledge is possible.
The TOK framework element exploring how context, culture, and position shape knowledge.
The Perspectives element asks: who is producing knowledge in this area, and from where? How do different groups or individuals approach the same knowledge differently? How has knowledge in this area changed over historical time, and what drove those changes? It includes reflection on the student's own perspective and what informs it.
The TOK framework element exploring ethical considerations woven into the pursuit of knowledge.
The Ethics element asks: what are the ethical implications of producing, sharing, or applying knowledge in this area? This includes the relationship between facts and values, questions of justice and inequality, and the responsibilities of knowers. TOK discussions of ethics must focus on knowledge questions rather than debating ethical issues for their own sake.
A mathematical claim supported by examples but not yet proven.
A conjecture is a hypothesis in mathematics — it seems to be true based on evidence (examples, patterns), but it has not been established with a proof. Fermat's Last Theorem was a conjecture for 358 years before Andrew Wiles proved it in 1995. Before the proof, was it already true? This is a fascinating question about the nature of mathematical truth.
A mathematical claim proven to be true from accepted axioms using logical reasoning.
A theorem has been established beyond doubt — within the axiom system. It cannot be shown false by any counterexample (if the proof is correct). Mathematical proof is unique among areas of knowledge in offering this kind of certainty. But note: theorems are only certain relative to their axioms, which are themselves accepted without proof.
A logical argument demonstrating, beyond doubt, that a statement is true for all cases.
A mathematical proof must be completely rigorous — a single error can invalidate the entire argument, no matter how much of it seems right. This is why Andrew Wiles had to fix a gap in his initial proof of Fermat's Last Theorem before it was accepted. The requirement for proof is what gives mathematics its distinctive form of certainty.
A systematic distortion of the truth, often present in how data are selected or presented.
Bias is distinct from perspective. A perspective is simply a viewpoint; bias is a misleading representation. A graph that starts its y-axis at 90 instead of 0 introduces visual bias. In mathematics and the sciences, ethical issues often arise not in the mathematics itself but in how results are presented — this is where bias most often enters.
A record or narrative of past events, always shaped by the author's perspective and evidence.
No historical account is a neutral transcript of the past. Every account reflects the perspective, purpose, and available evidence of the person who produced it. This does not mean all accounts are equally valid — some are better supported and more honestly presented than others — but it does mean we should always ask: who produced this, and why?
The process of explaining the meaning or significance of historical evidence.
Historical interpretation is always partial and subject to revision as new evidence emerges or interpretive frameworks change. Historians must bridge the gap between a document produced in the past and a reader in the present — with all the language, cultural, and conceptual differences that entails. "To what extent is interpretation a reliable tool in the production of historical knowledge?" is a canonical TOK question.
A document, object, or testimony produced at the time of the events under study.
Primary sources are the historian's raw material: letters, diaries, photographs, laws, artefacts. They are not automatically reliable — they reflect the purposes and perspectives of their authors — but they are direct evidence from the period in question. A secondary source (an account produced later) relies on primary sources and its own interpretation of them.
An account produced after the events under study, drawing on primary sources and earlier interpretations.
Secondary sources include textbooks, biographies, academic articles, and documentaries. They are not raw evidence — they are interpretations of primary sources, themselves shaped by the perspective, purpose, and historical moment of their authors. A secondary source can also become a primary source: a Cold War history textbook is primary evidence about Cold War-era educational ideology.
In aesthetic judgment, drawing on features of the observer rather than the artwork itself.
Subjectivity in the arts does not mean that aesthetic judgments are arbitrary. It means that the observer's personal experience, emotional response, and cultural background are part of the judgment. The key TOK question: which aspects of artistic value are purely subjective, and which are grounded in objective features of the work itself? Experts in art often converge in their judgments — suggesting that aesthetic value is not entirely in the eye of the beholder.
The quality that makes an artwork meaningful, beautiful, or significant — partly subjective, partly objective.
Aesthetic value is at the heart of the TOK Arts discussion. If it were purely subjective, there would be nothing to argue about — "I like it" ends the conversation. But art criticism, curation, and the art market all behave as if some artworks have more value than others, and experts can give reasons. What criteria are being applied, and how much do they depend on the observer?
A puzzling event or situation in the natural world that calls for scientific explanation.
Science begins with a phenomenon — something that does not fit, something that demands explanation. Lightning, the perihelion precession of Mercury, the double-slit interference pattern: each was a "that's funny…" moment that drove scientific inquiry. A phenomenon is distinct from a fact: it is a fact that calls for an explanation.
A fact that holds independently of any observer, perspective, or human practice.
"Water is H₂O" is a brute fact — it is true regardless of culture, religion, or belief. Contrast with social facts (e.g., "this piece of paper is £20") which depend on human practices and agreements. The natural sciences aim to produce brute facts; this is part of what makes their knowledge feel so authoritative.
Karl Popper's criterion: for a claim to be scientific, it must be possible to show it is false.
Popper's key insight: a theory that can explain every possible outcome is not explaining anything — it is just a story. A genuinely scientific claim sticks its neck out and says: "If this does not happen, I am wrong." The theory of evolution is falsifiable (it would be falsified by confirmed anachronistic fossils in the wrong geological strata); astrology is not (its predictions are vague enough to fit any outcome).
A set of claims that resembles science but cannot be properly tested or falsified.
Pseudoscience is not merely wrong — wrong science can be corrected. Pseudoscience is systematically insulated from correction: when tested, it shifts the goalposts; when challenged, it invokes untestable forces. Examples include astrology and homeopathy. The danger is that pseudoscience can be mistaken for genuine knowledge, especially when it uses the language and visual trappings of science.
Scientific inquiry driven by curiosity alone, without an immediate practical application.
Pure science seeks general models and theories — knowledge for its own sake. It aims to fit knowledge to the world (the "direction of fit" points toward reality). Applied science then uses those theories to change the world to meet a specification. The distinction matters for funding, ethics, and responsibility — but note that the line is often blurry, and "useless" pure science has repeatedly turned out to have profound applications.
The use of scientific knowledge to solve practical problems; the world is changed to fit a specification.
In applied science, the direction of fit is reversed: instead of adjusting knowledge to match the world, the engineer or technologist adjusts the world to match a safe specification. The Bangkok Skytrain must be protected from lightning — the engineer applies the general theory but must account for messy local variables. Applied science depends on pure science but has its own distinctive challenges.
Thomas Kuhn's term for a revolution in science that changes theory, method, scope, and the kinds of questions that can be asked.
Kuhn observed that science does not progress by smooth accumulation of facts. Instead, a dominant paradigm governs "normal science" until anomalies accumulate to a breaking point, triggering a revolutionary shift to a new paradigm. Examples: Copernican revolution (geocentric to heliocentric), miasma to germ theory. Kuhn controversially argued that paradigms are incommensurable — they cannot be directly compared.
Repeating an experiment in a different context or by a different team to verify a result.
Replication is a cornerstone of scientific objectivity. A result obtained once might be a fluke, an artefact of the equipment, or a product of the researcher's assumptions. When independent teams using different methods obtain the same result, confidence in that result increases dramatically. The failure to replicate results is a current crisis in some fields of psychology and medicine.
The error of concluding that because two things vary together, one causes the other.
Correlations are symmetrical; causation is asymmetrical. If A correlates with B, that relationship is the same as B correlating with A — but if A causes B, the reverse is not usually true. Distinguishing correlation from causation requires theory about the causal mechanism. Observation alone cannot establish causation; that requires controlled experiments or careful theoretical argument.
An explanation that rules out the possibility of itself being shown wrong.
A closed explanation is one that can explain any outcome: if the prediction comes true, the theory is confirmed; if it doesn't, some ad hoc reason is given for the failure (the test was badly designed, negative energies interfered, etc.). This is precisely what Popper's falsifiability criterion is designed to rule out. Uri Geller's claimed inability to bend spoons under controlled conditions — because "the scientists' negative energies" interfered — is a classic example.
An explanation that draws only on matter, energy, and natural forces — excluding supernatural causes.
The natural sciences are defined by their commitment to natural explanation: when lightning occurs, the explanation refers only to electrostatic discharge, not to divine intervention. This commitment defines the scope of natural science and is the basis for distinguishing science from pseudoscience and religion.
In pure science, knowledge must fit the world; in applied science, the world must be changed to fit the specification.
The phrase captures a fundamental difference in purpose. Pure scientists revise their theories when the world contradicts them — knowledge follows reality. Applied scientists (engineers, technologists) change the world to meet a pre-specified goal — reality must follow the design. The distinction matters for understanding the different responsibilities of researchers and practitioners.
A legitimate area of knowledge that does not study the natural world and is not expected to meet scientific criteria.
Non-science is not lesser knowledge — it is different knowledge. History, the arts, and ethics produce genuine, valuable knowledge using methods appropriate to their subject matters (which inherently involve human beings). The critical distinction is between non-science (which has its own standards of rigor) and pseudoscience (which pretends to be science while failing all scientific criteria).
An explanation of what a phenomenon is made of — what kind of thing it is.
A constitutive explanation answers "what is it?" Lightning is an electrical discharge. Water is H₂O. Constitutive explanations work by identifying a phenomenon with something we understand better, or by specifying its components. They are often paired with causal explanations.
An explanation of what caused a phenomenon — how it came to be.
A causal explanation answers "how did it happen?" Lightning occurs because air particles rub together, charge builds up, and the air can no longer resist the potential difference — discharge occurs. Unlike constitutive explanations (which identify what something is), causal explanations trace the sequence of events that produced the outcome.
Explaining a whole system by the behaviour of its component parts.
The classic approach of physics and chemistry: explain the behaviour of a system by tracing the interactions of its smallest components. A reductive explanation goes "down" — from the whole to the parts. It works well for simple mechanical systems but faces challenges in ecology and climate, where the interactions of parts change the environment, which changes the interactions — a dynamical system that cannot be fully reduced.
Explaining a system by the interactions among its parts, where the whole cannot be fully reduced.
Holistic explanation goes "up" — to the whole — rather than down to components. It is appropriate when a system's behaviour cannot be predicted from its parts alone because the interactions change the environment. Ecology, climate science, and evolutionary biology often require holistic explanation. James Lovelock argued that applying reductive explanations where they do not fit contributed to the climate crisis.
The principle that, given two equally good explanations, the simpler one should be preferred.
Named after the 14th-century philosopher William of Ockham. Occam's Razor is a methodological guideline, not a law: the simpler of two explanations should be preferred when they explain the data equally well. Newton's simpler laws of motion work for everyday scales; Einstein's more complex general relativity is necessary near black holes. Simplicity is a virtue, but not at the cost of explanatory accuracy.
An explanation invented specifically for one situation, covering no other phenomena.
An ad hoc explanation (Latin: "for this purpose") is one constructed to explain a single anomalous case without being derivable from any general principle. "God Thor struck his hammer" explains only lightning — it generates no further predictions and covers no other phenomena. Ad hoc explanations are the worst kind of scientific explanation: they have no generative power and cannot be tested further.
A scientific model that is self-contradictory or literally false yet still functions as a good explanation.
Hans Vaihinger's term for models that science knowingly uses even though they are false — or even logically contradictory. In quantum mechanics, photons are modelled as particles passing through two slits, while the screen is modelled as a solid wall — but the screen is also made of particles. The model is internally inconsistent, yet it perfectly predicts the interference pattern. A real fiction is a useful lie that gets the relevant features right.
The finding of a valuable or interesting result by chance, while looking for something else.
Science does not always proceed by deliberate hypothesis-testing. Percy Spencer (1945) noticed the chocolate bar in his pocket had melted while testing radar equipment — and helped invent the microwave oven. Penzias and Wilson (1964) detected a persistent background hiss while testing a horn antenna — and discovered the cosmic microwave background radiation, the echo of the Big Bang. Serendipity works because good scientists notice when something doesn't fit.
Formal evaluation of research findings by other specialists before they are accepted as knowledge.
Peer review is a key mechanism for quality control in science. Before research is published, it is evaluated by independent experts who assess the methodology, data, and conclusions. It is not infallible — reviewers can miss errors, and strong paradigms can make unorthodox findings difficult to publish — but it provides a check on individual error and bias that distinguishes scientific knowledge from personal opinion.
In science, a systematic body of knowledge — concepts, laws, methods, and standard examples — with strong evidential support.
The word "theory" in everyday speech means something tentative or speculative. In science and TOK, it means something much stronger: a systematic, well-tested framework for explaining and predicting a range of phenomena. The theory of evolution has as much evidential support as the theory that water is H₂O. Confusion between the two meanings has been exploited in arguments against evolution.
A general, experimentally confirmed description of a feature of reality.
A scientific law describes a pattern in nature that holds universally — Newton's second law (F = ma), Boyle's law, the law of conservation of energy. Laws differ from theories in that they describe what happens, rather than explaining why. An anti-realist holds that laws are useful idealisations invented by scientists — they apply only to idealised conditions (perfectly closed systems, ideal gases) that do not actually exist.
An idealised, usually simplified (and often false) representation that illuminates real phenomena.
A model takes a familiar structure and uses it to illuminate an unfamiliar one — like a metaphor. The Bohr atom (electrons orbiting the nucleus like planets) is false, but useful for predicting chemical properties. Newton's treatment of the apple as a point particle ignores its actual shape, but the simplification lets us make accurate predictions. A good model gets the causally relevant features right — like a map that omits hedgerows but includes roads.
Kuhn's claim that competing paradigms cannot be directly compared because each defines what counts as good evidence and method.
Thomas Kuhn argued that paradigm shifts are not rational in the way scientific progress is usually imagined. Two paradigms — geocentric and heliocentric astronomy, or miasma and germ theory — operate with different assumptions about what the phenomena are, what counts as a good explanation, and what methods are trustworthy. There is no neutral standpoint from which to compare them. Peter Galison disputes this: different layers of a paradigm change at different rates, giving footholds for comparison.
Scientific activity governed by an accepted paradigm, focused on solving puzzles within the framework.
Thomas Kuhn's term for ordinary scientific work: applying established methods to solve well-defined problems within an accepted framework. Normal science does not question the paradigm — it takes its concepts, methods, and standards for granted and works within them. The productive puzzle-solving of normal science is what most scientists do most of the time. When anomalies accumulate that the paradigm cannot explain, normal science gives way to crisis and eventually a paradigm shift.
The view that one world implies one true unified theory; multiple scientific perspectives are unfinished business.
Realists tend to hold this view: since there is only one world, all true scientific theories should eventually converge into a single unified account. Multiple competing perspectives in physics or biology signal that we have not yet found the complete truth — not that pluralism is acceptable. This view fits the notebook metaphor of knowledge: a growing list of true sentences that will eventually fit together.
The view that multiple overlapping theories can legitimately co-exist in science — a patchwork quilt rather than one unified account.
Pluralists are comfortable with theories that do not fully cohere with one another. Different disciplines, and even different sub-disciplines, can use different models that serve different purposes without any obligation to unify them. This view makes honest sense of current science, where theories in physics, chemistry, and biology often sit side by side without being fully integrated. It fits the map metaphor of knowledge: many maps of one territory, each useful for different purposes.
Popper's thesis that science progresses by attempting to falsify hypotheses, not by confirming them.
Karl Popper (1902–1994) argued that the hallmark of science is not the accumulation of confirming evidence but the willingness to formulate claims that can be shown false and then test them rigorously. The black swan problem shows why: no number of white swans confirms "all swans are white" — one black swan refutes it. Scientists should therefore try to prove their hypotheses wrong. The practical implication: scientists formulate a null hypothesis (H₀) and try to falsify it, not directly confirm their experimental hypothesis.
An assumption required to run an experiment that can absorb a falsification — insulating the main hypothesis from refutation.
When an experiment fails, the result can be attributed to a faulty auxiliary hypothesis (e.g., "the equipment was malfunctioning") rather than to the main hypothesis being false. This is not dishonest; it is a rational response to the fact that experimental apparatus relies on its own theoretical assumptions. But it means no single hypothesis can be definitively falsified by any single experiment. Pierre Duhem and W.V.O. Quine formalised this insight in what is now called the Duhem-Quine thesis.
The hypothesis (H₀) of no relationship between variables; scientists aim to falsify it, not directly confirm H₁.
In Popperian scientific practice, the experimental hypothesis (H₁) is not directly tested. Instead, scientists formulate its opposite — the null hypothesis (H₀: no relationship between the independent and dependent variables) — and try to show it is false. Only by demonstrating that H₀ is false can H₁ be provisionally accepted. This is why "proof" in science is technically inappropriate: we only ever reject H₀ with varying levels of statistical confidence.
The property of light whereby it exhibits wave-like behaviour in some experiments and particle-like behaviour in others.
One of the most striking scientific examples of competing perspectives eventually synthesising: Huygens's wave theory (17th C), Newton's particle theory (18th C), and Fresnel's improved wave theory (early 19th C) gave way to Einstein's 1905 proposal that light is quantised into photons — which led to the conclusion that light behaves as both wave and particle depending on the experimental context. Quantum mechanics ultimately suggests that asking what light really is (wave or particle?) may be the wrong question.
An observation that a current theory cannot explain; drives science forward without immediately overthrowing the theory.
Anomalies are not automatically fatal to a theory. The history of science shows that successful theories survive many anomalies, which scientists treat as puzzles to solve rather than refutations to accept. When noble gases were discovered in the 1890s and did not fit Mendeleev's periodic table, scientists did not abandon the table — they added a new column. Anomalies accumulate over time and eventually contribute to a paradigm crisis, but a single anomaly rarely overthrows an established theory.
The unavoidable act of choosing which variables to observe before investigation begins.
Observation cannot be neutral: to observe, you must already know what to look for. This creates a circularity — to know which variables are relevant, you need to observe them; to know which to observe, you need to know which are relevant. In practice, scientists make guided guesses based on prior experience, theory, and intuition. The danger is that other potentially relevant factors are overlooked from the outset.
All observation is structured by prior concepts and theoretical commitments — there is no purely neutral perception.
The claim that observation is "theory-laden" means that what we see is always shaped by the concepts and theoretical framework we bring to it. We cannot see mass without the concept of mass; we cannot see that a plant is flowering without the concept of a life cycle; a cat can see the light reflected from a laptop but cannot see it as a computer. Once you have a theoretical framework, it restructures what you perceive — mostly unconsciously.
The active nature of perception: we see things as something, not neutrally; shaped by prior concepts and expectations.
Formulated by Norwood Russell Hanson in Patterns of Discovery (1958), drawing on Wittgenstein's duck-rabbit example. "Seeing-as" describes how all perception is concept-dependent: we do not passively receive data from the world but actively impose conceptual structure on it. Once you learn the constellation of Orion, it is nearly impossible to see those stars as a random pattern. In science, expectations can result in change blindness — failure to notice changes — and can lead trained observers to see what they expect to see (as with Percival Lowell and Martian canals).
The change in a system caused by the act of observing it.
In particle physics, observing a particle requires bouncing photons off it, which changes the particle's state. This is a specific form of the Heisenberg uncertainty principle: the more precisely you observe one property of a particle, the less precisely you can know another. The observer effect also operates in social science (the Hawthorne effect) and in everyday experimental contexts, where the presence of an observer changes the behaviour of the subject.
The change in a system caused by physical contact with a measuring device.
The probe effect is a practical manifestation of the observer effect: the measuring instrument must be in contact with the system being measured, which changes the system. The classic example is inserting a thermometer into a cup of tea — the thermometer absorbs some of the tea's heat, changing the very temperature it is measuring. In more sophisticated experiments, the effect can be reduced but not eliminated. It is a reminder that "objective measurement" is always an approximation.
The measurable tendency to weight confirming evidence more heavily than disconfirming evidence.
Confirmation bias is a well-documented feature of human cognition that operates in scientific contexts as well as everyday reasoning. The N-ray scandal (René Blondlot, 1903) is the definitive scientific case study: Blondlot continued to observe N-ray effects even after Professor Robert Wood secretly removed the prism central to the experiment. His observations were a product of expectation, not physical reality. Confirmation bias is managed — imperfectly — through replication, blind and double-blind experimental design, and peer review.
An unexamined assumption needed to start a scientific investigation; can skew the entire inquiry if wrong.
Before any experiment begins, scientists must make assumptions that are rarely examined: that the equipment behaves as expected, that the theoretical context is sound, and — most dangerously — that certain kinds of result are possible and others are not. The H. pylori case illustrates the last point: the background assumption that the stomach is too acidic for bacteria blocked the correct explanation of stomach ulcers for decades. It took Marshall and Warren's courageous self-experimentation (1982) to overcome it.
No finite number of confirming observations can establish a universal claim with certainty; a counterexample is always possible.
Induction — drawing general conclusions from particular cases — is the fundamental reasoning tool behind all scientific laws. Every time you add water to anhydrous copper sulphate and it goes blue, your confidence increases. But you cannot be certain: there is no set standard for how many observations are enough, and scientific laws generalise to cases that have never been observed, which conflicts with the empiricist ideal. The problem was systematised by David Hume and is central to Popper's falsificationism: because induction cannot achieve certainty, scientists should at least be trying to falsify rather than confirm.
Scientific hypotheses cannot be tested in isolation; auxiliary hypotheses can always absorb a falsification.
Named after physicist Pierre Duhem (1861–1916) and philosopher W.V.O. Quine (1908–2000). Both independently observed that testing a hypothesis requires testing dozens of background assumptions simultaneously — about the equipment, about the underlying theory of the apparatus, and about the experimental design. When a result is anomalous, it is always rational (if not always correct) to blame an auxiliary hypothesis rather than the main hypothesis. This is why no single crucial experiment can definitively refute any hypothesis.
Theoretical entities posited by science that cannot be directly observed even in principle.
Quarks, gravitational fields, energy, and genes-as-units-of-inheritance are central to scientific theories but cannot be directly observed. Anti-realists (empiricists) argue these entities do not exist — they are useful fictions. Realists respond with the no-miracles argument: if quark theory makes accurate predictions but quarks do not exist, this would be a remarkable coincidence. Most working scientists proceed as if theoretical entities exist without worrying too much about the philosophical status of that assumption. The debate raises the question: to what extent is science actually empirical in the strict sense?
One of the distinct groups of knowers in an area of knowledge, each bearing different ethical responsibilities.
In the natural sciences, three groups are identified: pure scientists (who produce knowledge), applied scientists and technologists (who translate knowledge into practical technologies), and users (who employ those technologies). Each group owns the part of the ethical chain it controls. The framework is useful in TOK because it prevents both the over-attribution of responsibility to a single party (the pure scientist who had no knowledge of the application) and the under-attribution that occurs when everyone passes responsibility down the chain.
The principle that persons must always be treated as ends in themselves and never merely as instruments.
Derived from Immanuel Kant's categorical imperative. In scientific ethics, the Kantian principle establishes absolute prohibitions on non-consensual experimentation, regardless of the potential benefits. It is the philosophical foundation of informed consent requirements and the Nuremberg Code. It is always in tension with the utilitarian principle: cases where both apply simultaneously (the patient consents to a painful but beneficial procedure) are the norm; cases where they conflict are the hard ethical problems.
The principle that an action is ethical if it maximises overall well-being; some harm may be acceptable when benefits clearly outweigh it.
Associated with Jeremy Bentham and John Stuart Mill. In scientific ethics, the utilitarian principle permits research that causes some discomfort or risk if the expected benefits — to the subject, or to society — outweigh those costs. It is the philosophical basis for permitting clinical trials, animal testing, and other research that involves some harm. It is in constant tension with the Kantian principle: the utilitarian calculation can be used to justify violations of individual autonomy if the aggregate benefit is large enough.
The requirement that research subjects freely agree to participate with full understanding of what is involved and the right to withdraw.
Informed consent sits at the intersection of the Kantian and utilitarian principles: it honours autonomy (Kantian) while permitting beneficial research that involves some risk (utilitarian). Established as foundational by the Nuremberg Code (1947) and codified in the Declaration of Helsinki. Consent must be voluntary (free from coercion), informed (the subject understands what is involved), and ongoing (the subject may withdraw at any time). The requirement extends to human sciences as well as natural sciences.
Harm to the quality of knowledge — from falsified data, inadequate evidence, or suppressed results — that can cascade into physical harm when applied.
The concept draws attention to the ethical dimension of knowledge quality beyond direct physical harm. A scientist who publishes a causal claim without adequate evidence (e.g., Wakefield's MMR/autism paper) is guilty of epistemic harm even before any physical consequence follows. When the flawed knowledge is applied — in this case, falling vaccination rates leading to measles outbreaks — the epistemic harm becomes physical harm. Epistemic harm can also result from funding pressure to exaggerate positive results or suppress negative ones.
The ethical norms governing honest knowledge production: accurate data, transparent methods, proper attribution, and resistance to funding pressure.
Academic integrity is the institutional expression of knowledge virtues. Its requirements include: honest data collection and analysis; transparent reporting of methods, limitations, and negative results; proper attribution of sources and co-authors' contributions; and independence from funders who might prefer different conclusions. Violations range from minor (incomplete attribution) to serious (data fabrication), but all undermine the reliability of knowledge. Peer review and replication exist partly to enforce these norms, though neither is infallible.
An epistemic disposition — honesty, carefulness, objectivity, openness — that makes a knower a reliable producer of high-quality knowledge.
Knowledge virtues are the personal qualities that underpin academic integrity. They include honesty (not falsifying or misrepresenting), carefulness (not publishing without adequate evidence), objectivity (not letting desires shape conclusions), openness (sharing methods and data), and accountability (taking responsibility for errors). The concept is important in TOK because it locates ethical responsibility in the character of the knower, not just in the rules of the institution. A system of rules can be gamed; virtuous dispositions are harder to fake.
The first TOK assessment: three real-world objects linked to a knowledge question, with a 950-word commentary.
The TOK Exhibition is completed at the beginning of DP2. You choose three objects and write a commentary (up to 950 words total) showing how each object illustrates a different perspective on one of the IB's 35 prescribed prompts. The key: objects must be real and specific, with genuine knowledge contexts — not symbols or generic images. The Exhibition is worth 33% of your final TOK score.
The real-world situation in which an exhibition object is used or significant — the "so what?" that connects it to knowledge.
The knowledge context answers: what knowledge is at stake regarding this object? A photograph of a generic microscope has no knowledge context; a photograph of a specific attosecond electron microscope at the University of Konstanz, used in a particular research programme, has a rich one. Many students describe objects well but omit this crucial step.
The written explanation (up to 300 words per object) linking each exhibition object to the prompt through its knowledge context.
The commentary must do four things: identify the object (who made it, when, where, for what purpose); establish its knowledge context (what knowledge is at stake?); make an explicit link to the prompt keywords; and justify the object's place in the exhibition (what perspective on the prompt does it illustrate?). The most common failure is spending too much space on the object and too little on knowledge.