Two Theories of Tacit and Implicit Knowledge
Blekinge Institute of Technology
www.athenasoft.org; [email protected]
- The notion “tacit knowing” was introduced by Michael
Polanyi (1891-1976). He used it along with “personal
knowledge” to focus on implicit social standards of science and culture integrated in persons and a key to their practical skills and performance.
- The recent discussion is far from clear as to concepts, methods or implications for learning and organization of
- There are two main theories related to the phenomenon:
A. Tacit/implicit knowlegde is a type of knowledge,
contrasting with a verbal/explicit type of knowledge.
B. Tacit/implicit is one function among several that
components of cognitive acts can take.
- Here, I will show that
A. 3 A with its implications are untenable.
B. The theory 3 B is quite reasonable.
C. The implications of the theories for learning and
knowledge transfer will be clarified.
Against confusion. Not all that is “tacit” is knowledge
- Sometimes, it is presupposed that all that is tacit consists
in true beliefs or functional procedures.
- Such a presupposition if often based on the assumption
that human experience automatically furnishes the mind
with a correct representation or that human functional
procedures spontaneously evolve from feedback from
- While this is largely true in case of perceptual learning, it is
largely false when it comes to concept learning. There, it
has been falsified by research in the cognitive sciences:
– Heuristics and biases (Kahneman and Tversky)
– Social judgment (Brunswik, Hammond, Brehmer)
– Clinical judgment (Meehl, Dawes)
– Fast and Frugal heuristics (Gigerenzer)
– Complex problem solving (Dörner, Funke)
- Therefore, we need to single out the implicit/tacit part of the
claim without assuming that all that is tacit/implicit is of
value to adequate cognitive representations or to skillful
Theory no 1. Knowledge is of two types
- There are two types of knowledge.
• Tacit or implicit knowledge.
• Verbal or explicit knowledge.
- Conversion between these two types is a
major feature in the growth of social
knowledge. Knowledge “circulates”.
“There are two types of knowledge: explicit knowledge and
tacit knowledge. […] Explicit knowledge can be
expressed in formal and systematic language and can
be shared in the form of data, scientific formulae,
specifications, manuals, and so forth.” Nonaka et al.
The classification of knowledge into ‘codified ’ and ‘tacit ’
knowledge is a recurring topic in the literature on
knowledge–based societies. Codified knowledge can
be expressed in language or symbols and may as such
be stored and/or communicated (Eurydice, 2002, p. 12).
Arguments against tacit/implicit as a type of knowledge
- The classification is dependent on time and person. What is
tacit for me here and now (as a speaker of my mother tongue)
might not be tacit for you (as a linguist) or for myself later
(when I have acquired sufficient descriptive capacity.
- The classification is language-dependent. What can be
articulated in some languages cannot be articulated in all
languages. Specialists in various domains have languages
permitting finer discriminations than ordinary language. Art
historians might have words to express nuances that we
novices do not master.
- The classification is impossible to apply. How many tests by
how many persons does it take to establish “impossible to
articulate”. Impossibilities need to be proved by a theory (e.g.
- Empirical research has not been able to establish specific
features of knowledge implicitly acquired.
The only pro-argument I know of is logically a dud but strong on
rhetoric: “Surely you cannot teach advanced new surgery
methods via instructions”. Sure you can – if the sender has
enormous linguistic capacities and the receiver enormous
background knowledge about surgery plus great skills in
learning from instructions.
The flaw in this classification is that it bases a typology of knowledge
on shifting other capacities of knowledge bearers.
Theory no 2. Knowledge is constructed from elements with
- (Almost) every human act of knowledge is about
something, it is intentional or referring.
- In each such act something is focal/explicit i.e.
represented or conceptualized. Other parts of the
act rely on elements that contribute to make the
act intentional, referring or conceptual. These
elements bear a tacit/implicit function in that very
…. the triad of tacit knowing consists in subsidiary things (B) bearing
on a focus (C) by virtue of an integration performed by a person
(A) (Polanyi, 1969, p. 64).
I have a representation of …[a]… fact and the internal, functional use
of this representation constitutes it as knowledge of mine
(rather than as a desire of mine, etc.). The central idea of how
the implicit-explicit distinction applies is that knowledge can
vary depending on what is represented (made explicit) and
which aspects remain implicit in the functional use of
representations (Dienes & Perner, 1999, pp. 735-736).
Arguments for theory no 2
• Compatible with major philosophical and psychological
concepts and research, e.g.
– Schemas and scripts
– Heuristics, both heuristics and biases and fast and
– Cognitive modeling and rules of AI-research
• Compatible with major psychological theories of implicit
• Compatible with social macro theories of institutions and
with micro reductions of them to actor conventions or
• Compatible with functional theory of mind.
• Can be related to normative studies, e.g. via evaluation of
the accuracy of professional judgment.
• Btw, what more could one want from conceptual analysis
than the fact that it demystifies a notion in a way that
connects it with empirical research?
Is theory no 2 relevant for education, organizational
learning and social growth of knowledge?
Theory no 1 said that knowledge is of two types and that it
circulates on a social scale from tacit to explicit and back
to tacit knowledge.
But it seems that much of implicit/tacitly acquired conceptual
knowledge – hence social knowledge – is false. Or else,
they might be impossible to implement well via education.
Instead, we might opt for something that replaces previous
Institutions, reforms of organisations or
Technology, e.g. computerized decision support.
Theory no 2 has normative implications saying that
institutions should be reformed and technology should be
designed so as to make them learnable for humans.
Instead of looking at tacit kowledge as a valuable store of
knowledge – that theory 1 would have us do – we should
consider whether institutions and technology of the future
can be tacitly used, whether they lend useful support and
protect us from our faults and limitations.