Tacit and Explicit Knowledge

Tacit and Explicit Knowledge

by Harry Collins

View All Available Formats & Editions

Much of what humans know we cannot say. And much of what we do we cannot describe. For example, how do we know how to ride a bike when we can’t explain how we do it? Abilities like this were called “tacit knowledge” by physical chemist and philosopher Michael Polanyi, but here Harry Collins analyzes the term, and the behavior, in much greater


Much of what humans know we cannot say. And much of what we do we cannot describe. For example, how do we know how to ride a bike when we can’t explain how we do it? Abilities like this were called “tacit knowledge” by physical chemist and philosopher Michael Polanyi, but here Harry Collins analyzes the term, and the behavior, in much greater detail, often departing from Polanyi’s treatment.

In Tacit and Explicit Knowledge, Collins develops a common conceptual language to bridge the concept’s disparate domains by explaining explicit knowledge and classifying tacit knowledge. Collins then teases apart the three very different meanings, which, until now, all fell under the umbrella of Polanyi’s term: relational tacit knowledge (things we could describe in principle if someone put effort into describing them),  somatic tacit knowledge (things our bodies can do but we cannot describe how, like balancing on a bike), and collective tacit knowledge (knowledge we draw that is the property of society, such as the rules for language). Thus, bicycle riding consists of some somatic tacit knowledge and some collective tacit knowledge, such as the knowledge that allows us to navigate in traffic. The intermixing of the three kinds of tacit knowledge has led to confusion in the past; Collins’s book will at last unravel the complexities of the idea.

Tacit knowledge drives everything from language, science, education, and management to sport, bicycle riding, art, and our interaction with technology.  In Collins’s able hands, it also functions at last as a framework for understanding human behavior in a range of disciplines.

Editorial Reviews

This analysis of knowledge and its transmission is important and will shape conversation about knowledge and expertise for some time to come.”

— J. L. Croissant, University of Arizona

Technology and Culture
This is an exceptionally well-written book. . . . Collins fundamentally redefines one of the central puzzles in the philosophy and sociology of knowledge by turning the classic problem of tacit knowledge on its head.

— Wiebe E. Bijker

Evan Selinger

Tacit and Explicit Knowledge is an exceptional book. With a perfect blend of erudition, wit, and comprehensiveness, Collins succeeds in clearing up long-standing confusion about what tacit and explicit knowledge are and how they relate. His uncanny aptitude for selecting powerfully resonant examples and inventing memorable concepts should be sufficient to ensure that this text will be at the center of lively conversation among and between academics and industry leaders for some time to come.”

Donald MacKenzie

“Harry Collins is the world’s premier authority on tacit knowledge. This fine new book clarifies the concept in vital ways and is a crucial contribution to our understanding of it.”

Edgar A. Whitley

“In this book, Harry Collins applies his usual intellectual rigor to the question of tacit knowledge, a topic that has puzzled scholars for many years. Combining insights from sociological studies of scientific work with numerous, illustrative real-world examples, Collins carefully shows how what we label as tacit knowledge actually consists of three, often interlinked, elements: relational, somatic, and collective tacit knowledge. Of particular importance for management scholars is his insight that collective tacit knowledge (based on people and their living language) can only be acquired through socialization, which, in the context of business, means that if you want to make money from other businesses’ ideas, you need long-term access to their people, not just their plans or formulas.”

Trevor Pinch

“Ever since Michael Polanyi first coined the term, tacit knowledge has been rather mysterious; although now much in vogue in business schools and the social sciences more generally, it is often used in a confusing variety of different ways. It is also hard to show convincingly the presence of tacit knowledge in empirical studies since by definition it is a negative—defined by our inability to explicate it. In this profound and carefully worked-through book, leading sociologist of science Harry Collins neatly turns Polanyi on his head by showing us that the really deep mystery is how knowledge ever becomes explicit in the first place. Rich with examples including the classic ‘riding a bicycle,’ Collins teases out three different meanings of tacit knowledge.”

Robert P. Crease

“Tacit knowledge is one of the most important concepts of current scholarship in the humanities. Ambitious and important, Tacit and Explicit Knowledge is a well-written and original book.”

Choice - J. L. Croissant

“This analysis of knowledge and its transmission is important and will shape conversation about knowledge and expertise for some time to come.”

Technology and Culture - Wiebe E. Bijker

“This is an exceptionally well-written book. . . . Collins fundamentally redefines one of the central puzzles in the philosophy and sociology of knowledge by turning the classic problem of tacit knowledge on its head.”

Product Details

University of Chicago Press
Publication date:
Sales rank:
Product dimensions:
5.90(w) x 9.00(h) x 0.60(d)

Read an Excerpt

Tacit and Explicit Knowledge


The University of Chicago Press

Copyright © 2010 The University of Chicago
All right reserved.

ISBN: 978-0-226-11380-7

Chapter One

Strings and Things

It shows great folly ... to suppose that one can transmit or acquire clear and certain knowledge of an art through the medium of writing, or that written words can do more than remind the reader of what he already knows on any given subject.... The fact is, Phaedrus, that writing involves a similar disadvantage to painting. The productions of painting look like living beings, but if you ask questions they maintain a solemn silence. The same holds true of written words; you might suppose that they understand what they are saying, but if you ask them what they mean by anything they simply return the same answer over and over again. Besides, once a thing is committed to writing it circulates equally among those who understand the subject and those who have no business with it; a writing cannot distinguish between suitable and unsuitable readers. And if it is ill-treated or unfairly abused it always needs its parent to come to its rescue; it is quite incapable of defending or helping itself. (Socrates, in Plato, Phaedrus, 275d5–275e5)


When the explicit is discussed, it is usually taken as having to do with humans communicating via signs, icons, codes, or some such. As intimated in the introduction, here the analysis starts at a more basic level-interaction between physical objects, sometimes referred to as strings and sometimes as entities. At the outset, humans, animals, and other objects are treated as undifferentiated entities. The aim is to pull them apart, but the differences are to be established-not assumed. Another unusual feature of the analysis is that very little attention is paid to transmitting entities; nearly all the work of analysis concerns strings and their impact on things with the producers of strings being part of the background.

Strings, as explained, are bits of stuff inscribed with patterns. Strings in themselves are not meaningful. A string is simply anything that is neither random nor featureless. I said earlier that there were "strings" and "entities," but strings are entities and entities are strings, so what a string is and what an entity is depends on what is going on at the time.

Technically, a string always contains "information" in the sense connoted by "information theory." Information content is a physical feature of a string that refers to the number and arrangement of its elements. Strings, though they always contain information, are not always used to transmit information.

It will be useful from time to time to speak of the "elements" in the patterns pertaining to strings. Elements include the 0s and 1s of binary code, the letters of the alphabet, features of the patterns of smoke clouds, electrical states in silicon microcircuits, the positions of ratchets and cog wheels, the repeated shapes on patterned wallpaper, the notes made by a songbird, the atoms or molecules in a substance, or the features of lichen on a rock, and so forth.

The terms "string" and "strings" mean the same thing, since a long string is also a set of strings. Furthermore, elements of a string can also be strings in themselves and vice versa. For example, the letter "A" is an element in strings made up of letters of the Roman alphabet, but it is also a string when expressed in binary code as "01000001." And to illustrate the point made immediately above, "01000001" is both a string and a set of strings. Here the terms "string," "strings," and "elements" will often be used to refer to the same thing—after all, strings are just entities. In the actual analysis, the meaning of terms should emerge from the context without difficulty.

Interaction of Strings and Entities

Strings, in our manner of speaking, sometimes interact with entities. They can affect entities in four ways. (1) A string is a physical thing, so it can have a physical impact. (2) A string is a pattern, so it can impress, print, or "inscribe" a similar pattern on an entity in many different ways. (3 and 4) A string can change an entity in a more fundamental way than mere inscription—it can cause it to do something or give it the ability to do new things that it could not do before. This is called "communicating," and it can be done in two ways: (3) a string can communicate "mechanically," as when a new piece of code is fed into a computer or a human reacts to a sound in a reflex-like way; and (4) a string can communicate by being interpreted as meaningful by a human. Table 1 sets out the four ways strings affect entities.

The way a string impacts on an entity depends on the relationship between them. If the string is physically hard, it will more easily have a physical affect on an entity on which it might impact; if the entity is soft, the effect of the impact will be greater. Likewise, the way inscriptions work and the type of inscription that results from the impact of a string depend on the physical instantiation of string and entity. "Ink" can be used to enhance the effect. Whether the impact of a string results in a communication also depends on the string and the entity. The analysis will often proceed through exemplification of how differences in string and entity produce differences in outcome. For example, what will be called condition 4 and condition 5 of communication depend on different kinds of changes in the entity enabling what might be a mere impact or an inscription to become a communication.

In table 1, category 4 is emphasized because it is special. In the other cells any kind of entity can be found, but category 4 refers only to humans. Humans are also found in category 3, however, and sometimes the two categories have been confused. An example of the difference is the sergeant major's shout of "'Shun!" to soldiers on the parade ground. If the troops are thoroughly trained, they will respond to the shout in a reflex-like way. We do not say the sergeant major "told" the troops to come to attention; we say he "brought the troops to attention," implying a mechanism at work. But before the troops learned to respond in this mechanical way they would almost certainly have gone through a stage where they responded to the same shout in an interpreted way. They would have heard the "'Shun!" and interpreted it as the command "come to attention," and would have moved their bodies in response. Novice troops will hear themselves being told to come to attention. Thus, if the sergeant major assembles novice troops and tells them that, for a change, when he says "'Shun!" he wants them to relax, they will probably be able to do it. But when they are well enough trained to be causally affected by "'Shun!" they will not hear the command as an instance of "telling"; under these circumstances it would be very hard to change their normal reaction to "'Shun!" from coming to attention to relaxing. The idea of basic training is, of course, to make soldiers respond to orders on the battlefield without needing the time for conscious processing that could create the opportunity for fear.

Discussions of explicit knowledge generally concern themselves only with interpretation (column 4 of table 1) and that is why terms such as sign, icon, and code play such a central role. Unfortunately, the exclusive use of terms like these, which connote meaning and interpretation, encourages the mechanical impact of strings on entities to be confused with the interpreted impact. Thus computers are especially well represented among the entities found in the mechanical effect category (column 3 of table 1), but they are often thought to occupy the interpretation box, too. One of the advantages of starting at the basic level of strings is that this kind of confusion is less likely to occur. The three top-level categories set out in table 1 are now described.

Strings and Physical Impact

A string is always a physical thing. For example, the Bible is a string; at one time the recommended treatment for getting rid of a "ganglion"—a swelling in a tendon sheath—was to smash down on it with the family Bible. In the case of this use of a string, its information content and the fact that it is a string as well as a physical thing is not relevant.

More significantly, a string cannot have an effect that involves inscription or communication unless it also has an initial causal effect on the entity. Thus, in the course of reading this book (let us say), you have encountered a microdot containing a potentially interpretable string hidden within one of the full stops. But if this was the case it has had no causal effect on you and therefore you have not been affected by it—not even to the extent of being inscribed with its pattern. If the microdot were magnified so that you could see the string clearly then it might inscribe or communicate something. Or again, if the sergeant major whispered the command " 'shun" rather than shouted it, it would not have the initial causal impact required for subsequent causal or interpreted effects to follow—sergeants major have to shout loud if they want their strings to have mechanical or interpreted effects. As we will see below, when machines are described as "reading codes," they are transforming strings into a form that can have an appropriate causal impact on humans which, in turn, allows for the possibility of interpretation.

Strings and Inscription

A printing press contains strings in the form of the letters on plates. When these plates are introduced to the paper they produce an inscription. The same applies to my computer and its printer, except that the strings are in the form of electrical signals in the computer, which undergo various transformations prior to doing whatever is necessary to produce the paper inscription. Any forceful impact by one physical object on another is likely to cause an inscription.

Inscriptions may be temporary or lasting. When I speak, I temporarily inscribe the air with vibrations which are transformations of the movements of my larynx. Thomas Edison invented a way to transform the vibrations in the air into more permanent strings.

Humans can be inscribed with strings in various ways. For example, I might write a message on your forehead or pin a note to the back of your jacket without you noticing. More interesting is memorization and recitation; the multiplication table is a noteworthy example. Schoolchildren used to learn to recite the table in "parrot fashion." When I was young I had such a table inscribed in my person; decades on, it is still instantly accessible by chanting. Thus, if I want to know the result of multiplying 7 by 8, I simply chant the table and the result "seven eights are fifty-six" appears; it is just as though I had it written backwards on my forehead and I looked in the mirror. In this example, my "knowledge" of the multiplication table is exactly the same as it would be if I had not memorized the tables myself but had a slave who did not speak my language but who had been taught to memorize and recite the tables by a regimen of punishment and reward. A flick of the whip and the seven times table would be recited for my use without the slave needing to interpret it in any way whatsoever—the slave would be simply inscribed with the table.

Sometimes inscription is mistaken for proper teaching. A poignant example is illustrated in box 1, where a deaf boy describes his misery at having to learn meaningless words—meaningless because, without being immersed in the bath of speech from an early age, deaf children have difficulty in acquiring native spoken languages. We can assume that in the case of the deaf as represented in the box, the teachers believed they were accomplishing more than they were.

Memories are laid down in chemical and electrical pathways in that mysterious organ called the brain, so it is tempting to think that even the memories that make for mere recitation lead a somewhat more exotic metaphysical life than printed symbols. But the exotic appearance is due only to the fact that the medium on which the symbols are inscribed is three-dimensional human tissue rather than paper or silicon and that the means of "writing" is somewhat less well-understood in this case than when it is produced by a pen, typewriter or computer. Learning the multiplication table "by rote" is, however, just inscription, and it is metaphysically continuous with printing.

Strings and Communication

Taking the cue from Wittgenstein's "ask for the use, not the meaning," whether the transfer of a string counts as a communication depends on the outcome. Here a communication is defined as follows: A communication takes place when an entity, P, is made to do something or comes to be able to do something that it could not do before as a result of the transfer of a string. If the string is merely inscribed on an entity, "Alpha"—imagine it written on the forehead—then Alpha can enable a second entity, "Beta," to read the string merely by letting Beta, say, look at the inscribed forehead. But this is not what we mean here by "coming to be able to do something." On the other hand, entity Alpha will have been enabled to do something if it can use what has been transferred in some productive way—if it results in Alpha having new and useful knowledge. For example, if I can use the answer "56," which I obtain by reading the inscription on my slave, in another sum, or to affect a purchase, or some such, then the transfer of the string to me resulted in a communication as well as an inscription. A marginal case is when the entity is enabled to do something like answer a question in the game of Trivial Pursuit but is not able to do anything more with the new "knowledge." That kind of case just has to be kept in mind and will appear again.

An act of communication is like jumping across a gap between two buildings. Sometimes the gap is narrow enough to be jumped without anything special needing to be done. But very often, communication fails because the gap, as first encountered, is too wide to jump. In a subset of the difficult cases, something can be done to improve the jumping ability or to narrow the gap. One can sometimes increase the success rate by modifying the string so it can jump further or one can improve the chances by building out from the far side so there is less far to jump. There are two ways of making each kind of adjustment. Thus, including the fact that the gap might be narrow enough not to cause problems in the first place, there are five enabling conditions of communication. In this section the first four are described, while the fifth will be explained a little later.

Condition 1

Condition 1 is when everything is already in place so that the gap can be jumped and a string interacts with an entity and the result is a trouble-free communication. In condition 1, nothing has to be done to "enable" communication. An example is when a computer responds to the typed command 10 × 2.54 and produces the result 25.4 or a human is asked to multiply 10 by 2.54 and produces the result 25.4. The more interesting conditions are 2, 3, 4, and 5, because the communication fails in the first instance and something has to be changed to enable the gap to be jumped.

Condition 2.

Condition 2 has already been mentioned, but it is included here for completeness. It is a physical transformation of a string that enables it to have the causal impact on an entity which is the precondition for communication. The example that has already been discussed is magnifying a microdot. As it stands the microdot cannot "jump the gap," because it has no causal impact on the entity; when magnified, however, it can have a causal impact. For continuity, let us run through the condition as it applies to the sum 10×2.54. Imagine that the keystrokes required to get the computer to do this sum are written out on a piece of paper. For the computer to be able to actually do the sum—for the gap to be jumped—the paper string will first have to be transformed into a string that can apply physical pressure to the appropriate keys—a string of finger positions, say. This is a simple string transformation.


Excerpted from Tacit and Explicit Knowledge by HARRY COLLINS Copyright © 2010 by The University of Chicago. Excerpted by permission of The University of Chicago Press. All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.

Meet the Author

Harry Collins is a Distinguished Research Professor of Sociology and director of the Centre for the Study of Knowledge, Expertise, and Science at Cardiff University.

Customer Reviews

Average Review:

Write a Review

and post it to your social network


Most Helpful Customer Reviews

See all customer reviews >