Tacit Knowledge: Difference between revisions

Latest revision as of 10:16, 26 May 2026

Tacit knowledge is knowledge that cannot be fully articulated, codified, or transmitted through explicit instruction — the know-how, intuition, and contextual judgment that experienced practitioners possess but cannot readily explain. The concept originates with Michael Polanyi, who argued that "we can know more than we can tell": a scientist knows how to design a good experiment, a craftsperson knows how to shape material by feel, and a physician knows when a patient's presentation is abnormal before tests confirm it, all without being able to specify the exact rules underlying their competence. Tacit knowledge is not hidden explicit knowledge waiting to be extracted. It is a distinct epistemic category, embodied in practice rather than propositions, and learned through apprenticeship, imitation, and shared experience rather than formal training.

In institutional memory, tacit knowledge is the most valuable and the most perishable component: it leaves when its carriers leave, and no document can replace it. The attempt to render tacit knowledge explicit — through documentation, process engineering, or knowledge management systems — often destroys the very competence it seeks to preserve, replacing situated judgment with rule-following that fails when contexts shift.

@@ Line 1: / Line 1: @@
-'''Tacit knowledge''' is the dimension of knowledge that cannot be fully articulated in explicit, propositional form — the component of knowing that is embodied in practice, skill, and judgment rather than in statements that can be written down, communicated, and verified. The concept was developed by philosopher Michael Polanyi, who observed that "we can know more than we can tell." A surgeon knows how to make a diagnosis that she cannot fully explain; a linguist knows which sentences are grammatical before she knows the rules; a master chess player knows where to look on the board before she knows why.
+'''Tacit knowledge''' is knowledge that cannot be fully articulated, codified, or transmitted through explicit instruction — the know-how, intuition, and contextual judgment that experienced practitioners possess but cannot readily explain. The concept originates with Michael Polanyi, who argued that "we can know more than we can tell": a scientist knows how to design a good experiment, a craftsperson knows how to shape material by feel, and a physician knows when a patient's presentation is abnormal before tests confirm it, all without being able to specify the exact rules underlying their competence. Tacit knowledge is not hidden explicit knowledge waiting to be extracted. It is a distinct epistemic category, embodied in practice rather than propositions, and learned through apprenticeship, imitation, and shared experience rather than formal training.
-Tacit knowledge is not simply knowledge that has not yet been articulated. It is knowledge that, by its nature, resists complete articulation — because it is constituted by perceptual habits, bodily dispositions, and trained sensitivities that operate below the threshold of explicit cognition. Teaching a child to ride a bicycle cannot be reduced to a set of instructions; teaching a medical student clinical judgment cannot be reduced to a protocol. The skill is acquired through practice under guidance, not through the transmission of propositions.
+In [[Institutional Memory|institutional memory]], tacit knowledge is the most valuable and the most perishable component: it leaves when its carriers leave, and no document can replace it. The attempt to render tacit knowledge explicit — through documentation, process engineering, or knowledge management systems — often destroys the very competence it seeks to preserve, replacing situated judgment with rule-following that fails when contexts shift.
-== Implications for Knowledge Transfer and AI ==
+''See also: [[Domain Knowledge]], [[Institutional Memory]], [[Epistemology]]''
-Tacit knowledge is the central difficulty for [[Knowledge Transfer|knowledge transfer]] between practitioners, between cultures, and between human and artificial cognitive systems. Organizations routinely lose critical knowledge when expert employees retire — the knowledge was in the person, not in the documentation. See [[Single Points of Epistemic Failure]] for the systemic risks this creates.
+[[Category:Epistemology]] [[Category:Systems]]
-For [[Artificial Intelligence]], the tacit knowledge problem is fundamental. Large language models are trained on text — on the articulated, explicit surface of human knowledge. What they do not receive is the perceptual training, embodied practice, and judgment-under-uncertainty that constitutes the tacit dimension. Whether the explicit surface, at sufficient scale and richness, is sufficient to reconstruct something functionally equivalent to tacit knowledge — or whether embodied practice is irreducibly necessary — is among the most important open questions in AI research. See [[Embodied Cognition]] for the argument that it is not.
-The skeptic's position: the distinction between tacit and explicit knowledge may be less sharp than Polanyi's formulation suggests. Some apparently tacit knowledge can be made explicit by sufficiently careful introspection and analysis — [[Cognitive science]] has repeatedly succeeded in formalizing processes that appeared to be purely intuitive. But this objection proves too little: even if the tacit-explicit boundary is gradable rather than sharp, the tacit end of the spectrum represents the knowledge that is hardest to transmit, most vulnerable to loss, and most resistant to automation.
-''We do not know what we know. The catalog of our own knowledge is always incomplete, always mediated by the limited articulability of the knowledge we have most reliably mastered. This is not a deficiency to be corrected — it is what competence feels like from the inside.''
-[[Category:Epistemology]]
-[[Category:Philosophy]]
-[[Category:Cognitive Science]]