{
  "id": "tacit-knowledge",
  "title": "Tacit Knowledge",
  "lang": "en",
  "source_lang": "en",
  "summary": "Tacit knowledge is understanding expressed in skilled action, perception, and judgment that is not fully articulated in explicit statements or rules.",
  "category": "philosophy",
  "subcategory": "metaphysics-epistemology",
  "type": "concept",
  "aliases": [
    "Tacit knowing",
    "implicit knowledge",
    "unarticulated knowledge"
  ],
  "generated_at": "2026-11-03",
  "model": "pro/gpt-6.1-sol",
  "url": "https://aiwiki.page/en/tacit-knowledge/",
  "translations": {
    "zh-Hans": {
      "url": "https://aiwiki.page/zh/tacit-knowledge/",
      "json": "https://aiwiki.page/zh/tacit-knowledge.json"
    }
  },
  "sections": [
    {
      "heading": "",
      "text": "Tacit knowledge is knowledge that a person possesses or exercises without fully expressing it in words, rules, or other explicit representations. It is evident in such activities as recognizing a familiar face, maintaining balance on a bicycle, judging the quality of a material, or identifying a promising scientific problem. These performances depend on learned sensitivities, practical discrimination, and contextual understanding that may exceed what their practitioners can explain. The term is especially associated with Michael Polanyi, who wrote in The Tacit Dimension (1966) that “we can know more than we can tell.”\n\nThe concept matters because written descriptions do not necessarily reproduce the abilities they describe. A manual can specify the steps of an operation without enabling an inexperienced reader to perform it competently. Tacit knowledge therefore raises questions about the nature of expertise, the limits of formalization, and the transmission of understanding between people. It has become important in epistemology, the study of scientific practice, education, organizational research, and the analysis of technological change.\n\nThere is no single accepted theory of tacit knowledge. Sometimes the expression refers to knowledge that has not yet been articulated but could be; sometimes it denotes abilities whose complete specification is impractical or impossible. Polanyi’s stronger account treats tacit knowing not simply as a hidden store of information but as an indispensable dimension of all meaningful understanding. These interpretations lead to different explanations of how knowledge is learned, shared, represented, and evaluated."
    },
    {
      "heading": "Meaning and conceptual boundaries",
      "text": "The contrast between tacit and explicit knowledge concerns how understanding is possessed and exercised, rather than simply whether information has been recorded. Explicit knowledge can be stated in a form available for inspection, such as a mathematical proposition, a maintenance instruction, or a classification rule. Tacit knowledge is manifested through competent recognition or action even when the agent cannot provide an adequate statement of the basis of that competence. A skilled listener, for example, may hear that a musical performance is rhythmically unstable without identifying every timing deviation that produces the impression.\n\nSeveral distinct limitations can make knowledge tacit. A practitioner may lack a suitable vocabulary, may never have been asked to explain a procedure, or may perform a task too rapidly to report its details. In other cases, the difficulty concerns the relationship between a rule and its application. Instructions such as “tighten until secure” or “continue until the texture is smooth” depend on judgments that the words themselves do not supply. Additional descriptions can help, but their interpretation may require further practical familiarity.\n\nTacit knowledge is not identical to secret knowledge. A trade secret may be completely explicit yet deliberately withheld. Conversely, a craft worker may willingly demonstrate a technique but remain unable to articulate the perceptual distinctions that guide it. Nor is tacitness the same as complexity: a long formula can be explicit, while the apparently simple act of recognizing a particular voice can depend on an unarticulated integration of many features. The relevant issue is not the quantity of information but the form and conditions of its availability.\n\nThe term also overlaps with, but differs from, procedural knowledge, implicit learning, and practical know-how. Procedural knowledge concerns how to do something, whereas tacit knowledge can include recognition, interpretation, and judgment without a distinctive motor procedure. Implicit learning concerns the acquisition of regularities without full awareness of what has been learned. Tacit competence, by contrast, can develop through deliberate instruction and conscious practice. An experienced musician may know that practice improved a skill while remaining unable to explain the fine adjustments involved.\n\nFinally, the inability to explain a successful performance does not by itself establish knowledge. Success might result from luck, an unnoticed external cue, or a misleading test. Attributions of tacit knowledge generally require some stability of performance across relevant circumstances. This connects the topic to epistemic justification and reliabilism: the question is not only whether a person can state a reason, but whether the capacity reliably discriminates or acts within its appropriate domain."
    },
    {
      "heading": "Intellectual background and development",
      "text": "The phrase “tacit knowledge” became prominent in the twentieth century, but the distinction between articulated understanding and practical competence has older antecedents. Aristotle distinguished forms of intellectual achievement that included theoretical understanding, productive skill, and practical wisdom. His account of practical judgment emphasized sensitivity to particular situations rather than the mechanical application of universal statements. These distinctions are not equivalent to a modern theory of tacit knowledge, but they helped establish the philosophical problem of how competent action relates to general principles.\n\nEarly modern discussions of habit, custom, and skilled perception also supplied relevant themes. David Hume treated customary expectation as fundamental to ordinary reasoning about experience. Later, pragmatism emphasized the role of habits, inquiry, and activity in understanding. In the work of John Dewey, knowing was connected to the practical resolution of problematic situations rather than confined to the possession of detached representations. These approaches challenged the assumption that intelligent conduct must always begin with an explicitly formulated proposition.\n\nIn twentieth-century analytic philosophy, Gilbert Ryle made the distinction between “knowing how” and “knowing that” central to his criticism of intellectualism. In The Concept of Mind (1949), he argued against explaining every intelligent performance as the result of a prior internal act of consulting a rule. If intelligent rule consultation itself required another intelligent operation, the explanation would threaten an infinite regress. Ryle’s account and Polanyi’s developed along different lines, but both questioned the priority of explicit theoretical representation over practical competence.\n\nLudwig Wittgenstein contributed a related analysis through his discussions of rule-following in Philosophical Investigations (1953). A formulation does not determine its own application independently of established practices. Understanding a rule involves being able to go on appropriately within a setting of activity and training. Phenomenology, especially the work of Martin Heidegger and Maurice Merleau-Ponty, likewise directed attention to practical involvement and bodily familiarity with the world before detached reflection.\n\nPolanyi brought these concerns into a sustained account informed by his scientific career. A physical chemist who later turned to philosophy and social thought, he developed his analysis in Science, Faith and Society (1946), Personal Knowledge (1958), and The Tacit Dimension (1966). His work subsequently influenced the history and philosophy of science, where attention increasingly shifted toward laboratories, exemplars, instruments, and disciplinary training. From the 1980s onward, organizational and management research expanded the term’s use, sometimes treating tacit knowledge as an individual asset and sometimes as a property of shared practices."
    },
    {
      "heading": "Polanyi’s account of tacit knowing",
      "text": "For Polanyi, tacit knowing is an activity of integration rather than merely information awaiting verbal expression. Its central structure involves a relationship between what a person attends from and what that person attends to. The particulars that support an act of understanding are apprehended subsidiarily, while the meaningful whole is the focus of attention. In recognizing a face, attention is directed toward the person, not toward a complete inventory of distances, contours, and contrasts. Those particulars nevertheless contribute to recognition.\n\nPolanyi called the supporting particulars the proximal term and the meaningful object or achievement the distal term of tacit knowing. “Proximal” and “distal” here describe functional roles rather than simply physical distances. A person using a probe may attend from sensations in the hand to the object touched by the probe’s tip. With practice, the instrument becomes incorporated into the performance: attention is directed through it toward the task. Polanyi described this incorporation as indwelling.\n\nSubsidiary awareness is not simply unconsciousness. A practitioner may become focally aware of details that previously supported a performance without being separately noticed. Yet changing the direction of attention can alter their function. A pianist who concentrates on isolated finger movements may temporarily disrupt the musical phrase those movements normally sustain. This does not imply that analysis always damages skill. Rather, it shows that awareness of components and their integration into a whole are different achievements, and that moving between them can require retraining.\n\nThe distinction also applies to symbols and theories. Reading depends on attending from marks on a page to their meaning. Scientific understanding depends on using concepts, measurements, and representational conventions as instruments of inquiry rather than merely inspecting their formal shapes. On this view, explicit statements are not self-sufficient containers of meaning: their intelligible use presupposes capacities that are not exhaustively included in the statements themselves. Tacit and explicit knowing are therefore interdependent rather than mutually exclusive categories.\n\nPolanyi’s idea of “personal knowledge” does not mean that knowledge is merely subjective opinion. He emphasized the knower’s trained judgment, commitment, and responsibility while maintaining an orientation toward a reality that can resist expectations and disclose unforeseen implications. Scientific discovery involves assessing clues whose significance cannot be established by a complete advance specification of the result. This account explains why competent judgment may precede a fully articulated argument, but it does not make judgment infallible. Personal participation remains compatible with criticism, correction, and standards sustained by a community of inquiry."
    },
    {
      "heading": "Skill, perception, and cognitive mechanisms",
      "text": "Tacit knowledge is often illustrated through bodily skill because skilled action displays the gap between successful control and verbal explanation especially clearly. Balancing a bicycle requires continuous adjustment to motion and disturbance. A physical account of steering and stability can explain relevant dynamics, but understanding that account is neither necessary nor sufficient for riding. The rider acquires coordinated responses through interaction with the bicycle and environment, with feedback progressively shaping the activity.\n\nResearch on skill acquisition distinguishes stages in which deliberate attention has different roles. Paul Fitts and Michael Posner’s Human Performance (1967) described cognitive, associative, and autonomous phases. Early performance often involves explicit instructions and effortful monitoring; practice gradually improves coordination and reduces attentional demands. This model does not establish that every skill becomes wholly unconscious. Expert performance can combine automated components with deliberate planning, error checking, and strategic adjustment.\n\nPerceptual expertise is equally important. Skilled chess players do not merely calculate more moves than novices; they also recognize meaningful configurations. Research associated with Adriaan de Groot and later William Chase and Herbert Simon showed that expertise in remembering chess positions depends strongly on the meaningful organization of the positions. Such findings support accounts in which experience produces structured recognition rather than a general increase in memory capacity. A novice and an expert may look at the same arrangement while apprehending different actionable patterns.\n\nEmbodied cognition provides one framework for interpreting these capacities. It emphasizes that cognition depends on bodily abilities and engagement with an environment, not exclusively on detached symbolic manipulation. The texture of clay, the resistance of a tool, or the timing of another performer’s movement can become part of a skilled feedback loop. Nevertheless, tacit knowledge does not require commitment to any single theory of mental representation. Representational, dynamical, and ecological approaches offer different explanations of how such coordination is achieved.\n\nThe relation to consciousness is correspondingly complex. A person may consciously perceive the overall quality of a result while lacking access to the processes that generated the judgment. Conversely, an expert may deliberately inspect a familiar pattern and explain selected features without being able to state all the conditions governing recognition. Verbal report therefore captures only part of the performance, but behavioral fluency alone cannot reveal its full cognitive organization.\n\nExpert intuition is also domain-dependent. Repeated experience can support reliable recognition when situations contain stable cues and outcomes provide informative feedback. Where cues are weak, environments change, or feedback is delayed, confidence can exceed competence. Tacit knowledge should therefore not be equated with every strong impression or intuitive response."
    },
    {
      "heading": "Scientific practice and experimental replication",
      "text": "Science depends on explicit theories, methods, and public arguments, but scientific work also requires practical abilities that publications rarely describe exhaustively. Researchers must prepare materials, operate instruments, distinguish artifacts from signals, and judge whether a procedure is functioning normally. A methods section can identify equipment and specify settings while leaving unresolved how to recognize contamination, compensate for local conditions, or detect an abnormal instrument response. These judgments are often acquired through supervised laboratory work.\n\nThomas Kuhn emphasized the importance of learning from exemplars in The Structure of Scientific Revolutions (1962) and its later postscript. Scientific training includes solving representative problems, not simply memorizing general rules. Through such exercises, students learn which similarities matter and how theoretical expressions connect to concrete cases. Within a scientific paradigm, competent problem recognition can therefore precede the ability to formulate a comprehensive rule describing every legitimate application.\n\nThe sociologist Harry Collins investigated tacit knowledge through studies of experimental replication. His research on attempts to reproduce the transversely excited atmospheric-pressure carbon dioxide laser showed the importance of personal contact and practical communication in transferring an experimental capability. Written descriptions alone did not straightforwardly transmit everything required for successful construction. The case became influential because it treated replication as a social and material accomplishment rather than as the automatic consequence of following a published recipe.\n\nIn Tacit and Explicit Knowledge (2010), Collins distinguished relational, somatic, and collective tacit knowledge. Relational tacit knowledge remains unstated because of circumstances surrounding communication, including assumptions, concealment, or failure to recognize what another person needs to know. Somatic tacit knowledge concerns bodily capacities and the temporal or physical limitations of human performance. Collective tacit knowledge concerns abilities sustained through participation in social life, such as the context-sensitive use of language. These categories imply different obstacles to articulation and transfer.\n\nThe relevance to reproducibility is not that explicit documentation is futile. More complete protocols, reference materials, calibration, shared samples, software records, and demonstrations can substantially improve the transfer of methods. The point is that reproducibility depends on the relationship between documentation and the competencies of those using it. A procedure sufficiently explicit for an experienced laboratory may remain inadequate for a newcomer.\n\nTacit knowledge also creates a problem for evaluating failed replications. Failure may reflect an unreliable original result, but it may also reflect incomplete transfer of competence. Conversely, invoking tacit knowledge cannot indefinitely protect a claim from criticism. Public demonstration, independent performance, agreed benchmarks, and transparent descriptions of training help distinguish genuine practical requirements from explanations introduced only after failure."
    },
    {
      "heading": "Learning, apprenticeship, and social transmission",
      "text": "Tacit knowledge is commonly acquired through participation in an activity rather than through verbal instruction alone. Apprenticeship combines observation, imitation, guided practice, correction, and gradually increasing responsibility. A learner does not simply copy visible movements. Attention must be directed toward relevant features, and feedback must connect actions with outcomes. In woodworking, for example, an instructor may demonstrate how tool angle, grain direction, sound, and resistance jointly indicate whether a cut is proceeding well.\n\nDemonstration and language have complementary roles. A teacher can point out what a novice fails to notice, supply contrasts between acceptable and defective results, and introduce a vocabulary for increasingly fine distinctions. Expressions such as “listen for the change in pitch” or “feel the resistance before increasing pressure” do not fully encode the skill, but they organize the learner’s encounters with it. Articulation can therefore assist tacit learning without replacing the experience on which competent interpretation depends.\n\nJean Lave and Etienne Wenger’s Situated Learning (1991) described learning through legitimate peripheral participation. Newcomers gain access to a practice, initially undertake limited activities, and develop competence through changing forms of participation. The associated concept of communities of practice draws attention to shared repertoires, practical standards, and relationships among participants. Knowledge transmission is not only a transfer between isolated minds; it also involves access to tools, tasks, conversations, and accepted ways of evaluating performance.\n\nImitation alone is insufficient where the decisive features of performance are hidden or where visible behavior admits several interpretations. Learners may reproduce a teacher’s posture without acquiring the judgment behind it. Effective transmission consequently depends on opportunities to attempt the task, encounter variations, and receive corrections tied to specific results. Rotating through different assignments can expose dependencies that remain invisible when a learner repeatedly performs only one narrow operation.\n\nMethods such as think-aloud reporting, worked examples, video review, and simulated scenarios can make selected components available for discussion. Each has limitations. A think-aloud report may omit rapid judgments or alter the task by imposing additional attentional demands. Video can preserve observable action without capturing bodily sensations or the practitioner’s understanding of context. These methods produce useful representations, but not necessarily an exhaustive extraction of the competence.\n\nSocial access is therefore a major condition of learning. Participation may depend on employment, professional membership, language, equipment, or trust. Traditions of intangible cultural heritage, including performance and craft practices, illustrate how knowledge can persist through organized transmission while changing with new materials and circumstances. Preservation involves maintaining the conditions for competent practice, not merely recording descriptions of finished products."
    },
    {
      "heading": "Organizations and knowledge management",
      "text": "Organizations depend on knowledge that is distributed among employees, routines, equipment, and working relationships. A maintenance technician may know which vibration indicates an impending fault; a production team may coordinate adjustments without formally discussing every step; an experienced administrator may understand how an unusual case moves through several departments. Such knowledge can be essential to reliable operation while remaining absent from official procedures or database fields.\n\nResearch in knowledge management made the tacit–explicit distinction especially prominent during the 1990s. Ikujiro Nonaka’s 1994 article on organizational knowledge creation and Nonaka and Hirotaka Takeuchi’s The Knowledge-Creating Company (1995) presented a model commonly known as SECI. It describes socialization, externalization, combination, and internalization as interacting processes. Socialization involves shared experience; externalization involves articulating tacit understanding; combination reorganizes explicit material; internalization incorporates explicit material into practical competence.\n\nThe model highlights mechanisms through which organizations develop knowledge rather than merely store information. A team may begin with observations from experienced operators, use sketches or analogies to formulate a design principle, combine that principle with engineering specifications, and learn through constructing and testing a prototype. Metaphors and provisional models can make aspects of practical experience discussable before they are expressed in a stable technical vocabulary.\n\nA major debate concerns what “conversion” from tacit to explicit knowledge means. If tacit knowledge is simply knowledge not yet stated, externalization can be understood as successful articulation. If tacit knowing is an irreducible condition of understanding, articulation instead produces a representation that remains dependent on further tacit abilities. On the latter interpretation, a written procedure does not contain the practitioner’s entire competence; it changes the resources available for future performance.\n\nOrganizational knowledge retention consequently involves more than documentation. Staff turnover can remove relationships, historical understanding, and capacities for recognizing unusual situations. Mentoring, overlapping responsibilities, joint troubleshooting, and opportunities to observe experienced colleagues can sustain capabilities that a document repository does not preserve by itself. Equally, documentation can reduce dependence on particular individuals and expose inconsistencies that informal practice has allowed to persist.\n\nOrganizational routines are not automatically beneficial merely because they are tacit. They may embody obsolete assumptions, unsafe shortcuts, or exclusions that participants no longer examine. Formal review and practical experience therefore have different but complementary functions. Explicit standards make expectations inspectable, while situated judgment handles circumstances that standards leave open. The analytical task is to identify which capacities support reliable performance, where they reside, and under what conditions they can be challenged or transferred."
    },
    {
      "heading": "Economic significance, technology, and artificial intelligence",
      "text": "Tacit knowledge affects the economics of innovation because the movement of information does not necessarily produce an equivalent movement of productive capability. A firm may obtain a patent, purchase equipment, or license a process yet still require extended training before achieving consistent output. Troubleshooting, material selection, process adjustment, and coordination across production stages can depend on accumulated practical experience. This helps explain why apparently available technology may be difficult to reproduce elsewhere.\n\nKenneth Arrow’s 1962 work on learning by doing connected productive experience with improvements in performance. Richard Nelson and Sidney Winter’s An Evolutionary Theory of Economic Change (1982) subsequently emphasized organizational routines and capabilities in explaining economic behavior. In these approaches, productive knowledge is not exhausted by an explicit list of available techniques. It is embodied partly in what people and organizations can reliably accomplish through established patterns of activity.\n\nTacitness also influences the geography of innovation. Face-to-face interaction, worker mobility, and shared professional settings can facilitate the exchange of context-sensitive understanding. Industrial clusters may benefit when related producers, suppliers, and specialists encounter one another repeatedly. Geographical proximity alone is not sufficient, however: relevant expertise, social access, trust, and the ability to interpret another organization’s methods remain important. Knowledge can be nearby yet inaccessible, or transmitted over distance through sustained collaboration.\n\nIn artificial intelligence, tacit knowledge bears on the difference between specifying rules and learning patterns from examples. Early expert systems sought to encode specialist judgment in explicit rules, but knowledge acquisition often proved difficult. Experts could make useful distinctions without readily stating all the conditions underlying them. Machine learning offers a different strategy by fitting systems to examples and feedback rather than requiring every decision rule to be supplied in advance.\n\nThis does not establish that learned computational representations are tacit knowledge in Polanyi’s sense. A model may perform a task without providing a human-readable explanation, but opacity, unarticulated human skill, and participation in social practice are distinct phenomena. An artificial neural network can be fully specified computationally while remaining difficult to interpret at the level of meaningful concepts. Conversely, a human practitioner can explain important aspects of a skill without supplying an executable procedure.\n\nAutomation also redistributes rather than simply eliminates practical knowledge. A system may reduce the need for routine manual discrimination while increasing the need to curate training data, recognize unusual operating conditions, and interpret failures. In robotics, interaction with materials and changing environments makes the relation between learned control and practical competence especially visible. Whether artificial systems possess tacit knowledge depends partly on the definition adopted and cannot be settled solely by observing that their internal processing is difficult to explain."
    },
    {
      "heading": "Evidence, criticism, and unresolved questions",
      "text": "A central methodological problem is how to establish that knowledge is tacit. The absence of an adequate verbal report is weak evidence on its own: the person may be unwilling to explain, misunderstand the question, lack an appropriate vocabulary, or possess a competence that the chosen task does not reveal. Research therefore combines observation, performance tests, interviews, demonstrations, and comparisons between novices and experts. Each method examines a different relation between what people can do and what they can state.\n\nResearchers must also distinguish current inarticulacy from principled inarticulability. Knowledge that no participant has yet formulated may later be expressed through a new measurement technique, diagram, or conceptual distinction. Conversely, an increasingly detailed description may still require trained interpretation. Claims that something is “impossible to articulate” are consequently stronger than claims that existing accounts are incomplete. They require an argument about the nature of the competence, not merely a record of unsuccessful interviews.\n\nThe philosophical status of know-how remains disputed. Rylean approaches resist reducing intelligent abilities to propositional knowledge. By contrast, Jason Stanley and Timothy Williamson’s 2001 article “Knowing How” defended an intellectualist account on which knowing how is a form of knowing that. Much of the disagreement concerns the relation among linguistic constructions, practical abilities, and the ways in which knowledge is represented. Tacit knowledge does not settle this debate, since it can be interpreted either as nonpropositional competence or as knowledge whose propositional content is not readily reportable.\n\nA related issue concerns the bearer of knowledge. Individual-centered accounts emphasize a person’s acquired capacities, whereas social epistemology and practice-based research examine how competent action depends on groups, institutions, tools, and shared environments. Saying that an organization “knows” how to manufacture something can refer to a distributed capability that no single member possesses. The challenge is to explain this distribution without treating an organization as a person or overlooking the distinct contributions of its members.\n\nCritics also question whether the term explains too much. If every successful activity depends on tacit knowledge, the category may become too broad to distinguish mechanisms. More precise accounts separate bodily control, perceptual learning, contextual interpretation, socially sustained norms, and information that happens not to be stated. Collins’s taxonomy and cognitive research offer different ways to make these distinctions, though they do not produce a single accepted classification.\n\nFinally, tacit knowledge raises questions about accountability. Skilled judgment may be necessary where rules are incomplete, yet appeals to expertise do not remove the need for evidence, criticism, or fair evaluation. Procedures can be assessed through their outcomes, robustness across cases, responsiveness to correction, and the conditions under which competence is acquired. The continuing problem is to make practical judgment publicly assessable without assuming either that everything valuable can be fully codified or that whatever resists codification deserves exemption from scrutiny."
    }
  ],
  "links": [
    "knowledge",
    "epistemology",
    "education",
    "epistemic-justification",
    "aristotle",
    "david-hume",
    "pragmatism",
    "analytic-philosophy",
    "ludwig-wittgenstein",
    "phenomenology",
    "philosophy-of-science",
    "embodied-cognition",
    "consciousness",
    "thomas-kuhn",
    "scientific-paradigm",
    "reproducibility",
    "calibration",
    "apprenticeship",
    "intangible-cultural-heritage",
    "artificial-intelligence",
    "machine-learning",
    "artificial-neural-network",
    "training-data",
    "evidence"
  ],
  "sources": [],
  "license": "https://creativecommons.org/licenses/by/4.0/"
}