3

The Concept of Information

Information seems to be everywhere. We talk of its being encoded in the

genes … disseminated by media of communication … exchanged in con-
versation … contained in all sorts of things … Libraries are overflowing
with it, institutions are bogged down by it, and people are overloaded with
it … [yet] no one seems to know exactly what information is.
Christopher Fox (1983, p. 3)

Information, usually seen as the precondition of debate, is better understood

as its by-product.When we get into arguments that focus and engage our
attention, we become avid seekers of relevant information. Otherwise we
take in information passively — if we take it in at all.
Christopher Lasch (1995, p. 162)

Chapter Outline

3.1. Searching for a Definition of Information

3.1.1. Explicating “Information”
3.1.2. The Concept of Information
3.1.3. Typologies of Information Concepts
3.2. Definitions of Information and Their Problems
3.2.1. The Influential and Restrictive “Information Theory”
3.2.2. Five Problematic Issues in Defining Information
3.2.3. Utility as a Requirement
3.2.4. Physicality as a Requirement
3.2.5. Structure/Process as a Requirement
3.2.6. Intentionality as a Requirement
3.2.7. Truth as a Requirement
3.3. Must There Be a Universal Definition of Information?
3.4. Distinctions among Information, Knowledge, and Data
3.5. Summary

39
40 3.The Concept of Information

3.1
Searching for a Definition of Information

“Information” is a fairly old English word, making an early appearance in

one of Chaucer’s tales sometime between 1372 and 1386 (Schement, 1993a,
p. 177); Capurro and Hjørland (2002) take its orgins back to Latin and Greek
terms of the pre-Christian era. One would think that hundreds of years of usage
would tend to settle a word and result in a consensus on its meaning.This has
not been the case with the term “information.” Especially in the last five
decades, as the various phenomena that people call information began to be
objects of empirical study, meanings of the word have proliferated. Schrader
(1983, p. 99) goes so far as to complain about “the multiplicity of vague, contra-
dictory, and sometimes bizarre notions of the nature of the term ‘information’.”
One of the problems of studying any phenomenon — or merely talking
about a thing — is reaching an agreement on what to call it.Words are ambiguous,
the same string of characters often having multiple meanings. Each meaning
may identify a distinct concept, in the way that the noun “port” can refer to a for-
tified wine, the left side of a ship, or a gateway or opening for passage (a harbor,
a modem port, a valve port, etc.).The case of the word “information” is much
more complex, as it has been used to denote various overlapping concepts, rather
than neatly distinct phenomena as is the situation with “port.”
Unless otherwise stipulated, in this book “information” will be taken to
mean any difference that makes a difference to a conscious, human mind (Bateson,
1972, p. 453). In other words, information is whatever appears significant to a
human being, whether originating from an external environment or a (psycho-
logically) internal world.This definition was chosen by the anthropologist Gregory
Bateson, after he had struggled for two decades with the inadequacies of mathemat-
ical definitions of information. A perceived difference, according to Bateson, is a basic
“unit of mind” that can be inferred through study of both humans and animals.
Other authors have employed definitions of information that are similar
to Bateson’s. A popular version originated with psychologist George A. Miller
(1968): information is any stimuli we recognize in our environment. Others (e.g.,
Dervin, 1976a; Farace, Monge, & Russell, 1977; Higgins, 1999; Johnson, 1997;
Rogers, 1986) have generalized such statements to mean the recognition of patterns
in the world around us.
Obviously, the characterization of information as a difference implies a
very broad definition for a common word that has been defined in several dis-
tinct ways — with virtually all other definitions implying more restrictions on
meaning.That is, many authors have used other words to define a concept that
they have called “information,” some of them incorporating specific requirements,
such as information must always be true or useful, or it must be embodied in a
form or object, or it must be intentionally transmitted, and so forth.
3.1. Searching for a Definition of Information 41

The reason for adopting this broad (some would say vague) definition is
because this book reviews a great many studies from several disciplines and needs
to cover a variety of more restrictive concepts. The reason some other defini-
tions are less appropriate will be made clearer in the remainder of this chapter,
which first discusses the nature of conceptual explication, and then examines
a variety of definitions for “information,” each of which identifies a somewhat
different concept than the rest.

3.1.1 Explicating “Information”

To discuss and study any concept, we first need to define it. In the process,
we may identify and define other ideas that are related to (and sometimes derived
from) the concept under study. In the case of information, related concepts are
“knowledge” and “data.” “Information behavior,” “information seeking,” “infor-
mation source,” and “information use” are among several higher-order concepts
that build upon the concept of information.
Social scientists call the process of defining a concept explication. As dis-
cussed by Steven Chaffee (1991), explication is the intellectual process that
relates theory to research, that links a focal concept to the ways in which it is
studied. For researchers who aspire to direct observation of phenomena, expli-
cation eventually results in an operational definition of a concept, a set of proce-
dures used to observe and measure instances of a concept. In this chapter, we
will deal only with the initial stages of explication — reviewing and analyzing
existing definitions — and leave issues of measurement for later chapters.
The process of explication often starts with a word for which we have
only a general meaning.At this stage we have only a nominal, or dictionary, def-
inition for a term — that is, a word is defined by other words. Explication con-
tinues by examining what has been written about the concept; we review the
publications about it, with an eye toward how different authors have defined
and used the concept. In doing this we may not only find multiple definitions
for the term, but discover that some authors have studied the same concept but
called it by a different name.
The next step in explication is to analyze the meaning of a term by one or
both of two approaches. The first possible approach is a top-down procedure.
We distill the discussions of many authors to their abstract core: what is the heart
of what they say about the concept? In the case of information, a core idea may
be that it is a message expressed in some medium, and/or that it has the poten-
tial of altering a person’s consciousness.As Chaffee points out (1991, pp. 26–27),
finding a single, central meaning is unusual, particularly when distinct research
literatures are examined.
In the second alternative, the bottom-up approach to meaning analysis, we list
all of the subsidiary concepts that make up the focal concept. For “information,”
42 3.The Concept of Information

we might attempt to list exhaustively all of the possible forms that information
could take — a Web page, a book, a radio broadcast, a conversation, a handwrit-
ten note, e-mail, and so on. This is a massive task that would be subject to
change over time as new forms of information appear or are identified. Listing
all examples of information has been one way that researchers have guided the
observation of the concept.
Whichever means of analysis is chosen, the eventual result is a more
abstract definition than one defined by near synonyms. The definition may be
expressed as a series of critical distinctions between the focal concept and related
concepts; for example, what is the difference between information and data?
Or, instead, it may simply identify attributes that serve to identify something as
an instance of the concept; for example, a book is an instance of information
because it contains symbols that can, or are intended to, inform someone.
The remainder of this chapter will consider, through literature review and
distillation, the various definitions of information and their key distinctions.

3.1.2 The Concept of Information

Ordinarily, we both use and hear the word “information” without much
concern for its definition; we know what we mean when we use the word. At
first glance, the Oxford English Dictionary definition seems adequate: “(1) the
action of informing. The action of telling or fact of being told of something.
(2) That of which one is apprised or told; intelligence, news.” This nominal
definition reveals at least one important distinction: the term may be used to
indicate either a process (informing) or a kind of message (news).
Further distinctions lay buried in the nominal definition, as a series of
publications have made obvious. One explication of the term (Wellisch, 1972)
uncovered eight distinct definitions of information, without any common
elements. Not long afterward,Wersig and Neveling (1975) identified 17 unique
definitions, which they grouped into six broad categories. Summarizing
30 years of commentary, Levitan (1980) declared that 29 different concepts
had been associated with the term information.A review by Schement (1993b)
includes a selection of 22 definitions written between 1968 and 1989. How
has the concept of information been used such that so many definitions have
resulted?
The central difficulty is that the word “information” has been used to
denote several different concepts. The adoption of the term by multiple dis-
ciplines is part, but not all, of the problem. The same term has been used to
refer to, among other phenomena, sensory stimuli, mental representations,
problem solving, decision making, an aspect of human thinking and learning,
states of mind, the process of communication, judgments about the relevance
3.1. Searching for a Definition of Information 43

of information to information needs, the content of subject specialties,

recorded knowledge, and particular objects that carry information such as
documents. It is no surprise that scholars have struggled to come up with a
formulation that promises to condense most of these meanings into one universal
principle or attribute.

3.1.3 Typologies of Information Concepts

Let’s first examine four parallel attempts to identify different “families” of

information definitions.Two attempts to distinguish types of information con-
cepts, one from 1976 and the other from 1992, illustrate how periodically we
revisit the problem of defining information. Two articles by Brenda Dervin
(1976a, 1977) set the stage for the development of the sense making school of
thought (see Chapters 4 and 7) regarding information seeking. In her articles,
Dervin posited three types of information, based on the writings of philosopher
Karl Popper (1972):
1. Objective, external information is that which describes reality (but never
completely so).
2. Subjective, internal information represents our picture or cognitive map
of reality, the structures we impute onto reality.
3. Sense-making information reflects the procedures and behaviors that
allow us to “move” between external and internal information to
understand the world, and usually to act on that understanding as well.
Dervin argues that to look at information in such a way has several advan-
tages. For example, it acknowledges that legitimate inputs may come from
inside us, rather than viewing the only important information as arising from
external sources. In a similar way, this view does not privilege formal informa-
tion systems (e.g., books) over informal sources (e.g., friends, relatives, or
coworkers); consulting the latter is a much more common approach to under-
standing than are the former channels.
In a manner reminiscent of Dervin, Brent Ruben (1992, pp. 22–24) places
information conceptualizations into three “orders.” The first of these captures
information as “environmental artifacts and representations; environmental data,
stimuli, messages, or cues.” This environmental (Ie) sense of information consists
of “stimuli, messages, or cues, waiting to be attended to.” Second-order informa-
tion is that which is “internalized, individualized appropriations and representa-
tions.” Here Ruben identifies information as something that is “transformed and
configured for use by a living system,” internal (Ii) representations that include
“semantic networks, personal constructs, images, rules or mind.” And the third
type of information is that which is “socially constructed, negotiated, validated,
44 3.The Concept of Information

sanctioned and/or privileged appropriations, representations, and artifacts.”

Third-order information, then, is the social (Is) context of information.
Dervin’s and Ruben’s types are parallel but not identical, especially in the
terms and examples they use to describe their third category, which for Dervin
is decidedly intrapersonal, abstract, and process oriented. For Ruben, the social
context is external, is socially constructed, and may encompass physical objects
like books (which seem to fall under Dervin’s objective category).
Two other typologies, both from the 1990s, are also somewhat parallel,
but they bear only a modest resemblance to the Dervin and Ruben schemes.
Michael Buckland’s (1991a) widely cited typology portrays uses of the term
“information” as falling into three categories. The first category is information-
as-process, which refers to the act of informing, the communication of informa-
tion, and how a person’s state of knowledge is changed. A second sense of
information is information-as-knowledge, a usage of the term denoting that which
is perceived in the first category (i.e., the knowledge communicated).The final
sense of the term is information-as-thing, in which “objects, such as data and doc-
uments … are referred to as ‘information’ because they are regarded as being
informative.”
Buckland takes great pains to explain the difficulties inherent in such a
typology, pointing out the intangible nature of the first two categories (which
makes them difficult to observe), the issue of intentionality (some definitions of
information take for granted an intention to communicate), and the problem
that any object in the world might potentially be informative (“if everything is
information, then being information is nothing special”). He concludes that it
is essential to investigate information-as-process, even though information-
as-thing cannot be dismissed as a focus of study.
The second typology is similar to that of Buckland but breaks out his two
categories of information-as-thing and information-as-knowledge into three
overlapping conceptions of information. Altogether, McCreadie and Rice
(1999, pp. 47–58) identify four distinct “conceptualizations,” the first of which
is information as a resource or commodity. Under this conceptualization, informa-
tion is something that can be “produced, purchased, replicated, distributed, sold,
manipulated, passed along, controlled” — such as a message that travels from
sender to receiver, with or without some kind of payment in exchange.
The second type of information is characterized as data in the environ-
ment, that is, “objects, artifacts, sounds, smells, events” that may be perceived in
the environment. This category takes into account the potential for uninten-
tional communication of information, such as when one observes and interprets
natural phenomena.
McCreadie and Rice’s third type of information concept is that
expressed as a representation of knowledge, such as that expressed in “documents,
books, periodicals.” Finally, their fourth type of information is as a part of the
3.2. Definitions of Information and Their Problems 45

communication process. That is, information is meanings that are created as

people go about their lives and try to make sense of their world.
While at first glance it may seem that these latter two typologies charac-
terize types of information in similar ways, there are several differences, partic-
ularly in the distinctions they make between representation, thing, and (in the
case of McCreadie and Rice) resource. McCreadie and Rice use “documents,
books, periodicals” among their examples of representations, whereas Buckland
uses those as examples of “things.” It seems that McCreadie and Rice are trying
to make finer distinctions than Buckland regarding possible embodiments of
information.
Ultimately, the typologies of Dervin, Ruben, Buckland, and McCreadie
and Rice are each distinct from one another in several ways.At least the distinc-
tions these authors make are useful in illustrating the many ways one could
parse the attributes of the information concept.

3.2
Definitions of Information and Their Problems

The typologies discussed above fall short of providing specific definitions

of “information.” Rather, their intention is to show that there are distinct usages
of the term rather than a single universal usage.
Nevertheless, many authors have attempted to create a general definition
of information that at least would be adequate for some areas of investigation.
As we shall see, many scholars have incorporated into their definitions specific
and powerful assumptions regarding the nature of information.
We will begin our examination with the most influential definition of
information, one that developed a half-century ago for the study of signal
transmission in broadcasting and telephony. Now it might seem odd that a
definition for such a mundane concept as information should come to us
from a highly specialized field as telephone engineering. In fact one writer
(Tor Nørretranders, 1991; English translation 1998) jokes that

there are plenty of grounds for a conspiracy theory of the most devious kind:
that the notion of information was invented and developed by engineers from big
private corporations who then made a profitable business out of having the rest of
us talk about truth, beauty, meaning and wisdom — on the phone. (p. 96)

In Nørretranders’s view, this development was unfortunate because it

shifted our attention away from the more important elements involved in
information — the senders and receivers of messages — and toward characteristics
of the carrier.
46 3.The Concept of Information

3.2.1 The Influential and Restrictive “Information Theory”

The first widely recognized attempt to define information, the misnamed

“Information Theory” (properly called “The Mathematical Theory of
Communication”) is still frequently invoked to describe the nature of informa-
tion.The popularity of information theory cannot be overemphasized: a review
of two decades past (Zunde, 1984) listed over 400 selected citations to this
theory; by now, the number of references to information theory surely runs into
the thousands.
Fifty years ago, the works of Claude Shannon and Warren Weaver (1949)
on communication of messages gave rise to a popular conception of informa-
tion. Shannon, an engineer at Bell Labs, was concerned with the fidelity of
telecommunications signals, such as those sent over radio waves, and the deter-
mination of the effective capacities of telecommunication channels. It was
Shannon who came up with a model of communication as a process of signal
transmission. His became the basis for applying measures to parts of messages
based on the statistical probability of their appearance — a technique that led to
improvements in signal transmission because it helped to predict the likelihood
of errors and decide how to correct them, such as the sending of redundant
portions of a message.
In Shannon’s famous diagram (Figure 3.1), the source and destination of a
message were seen as being at the opposite ends of a chain, linked by a message
converted by a transmitter into a signal sent over some kind of channel to
the receiver, which converts the signal back to a message for delivery to the
destination.The channel was acted on by sources of noise, which could disrupt or
distort the message.
Along with the diagram came both a definition and measure for the con-
cept of information, as it is encoded in a message. Shannon’s definition of infor-
mation was based on the notion of entropy, a measure of the degree of
disorganization in a system which reflected a tendency for any state of affairs to
lose order and become more random. In signal transmission, noise is the vehicle
for the effects of entropy, that is, noise degrades the signal to some degree.
Messages are organized exchanges (e.g., grammatical sentences) based on selec-
tions from an agreed-upon set of signals (phonemes, words, letters, etc.). The
requirement that the message elements are selected from a fixed universe of

Noise

SOURCE TRANSMITTER [] RECEIVER DESTINATION

Message Signal sent Signal received Message

Figure 3.1
Shannon’s model.
3.2. Definitions of Information and Their Problems 47

possible elements has led some scholars to refer to it as a theory of “selective

information,” to contrast it with theories of “semantic information,” that is, theories
concerned with the meaning of messages.
The effects of entropy lead to more randomness in messages that, in turn,
leads to higher levels of uncertainty. In Shannon’s view, these higher levels of
uncertainty imply the potential for more information in the message. At the opposite
end of the entropy scale would be messages that are highly organized — and thus
familiar to the receiver, but which tend to carry little “new” information. Shannon’s
theorems dealt with statistical probabilities associated with the selection of signals
from a well-defined set. However, subsequent applications of Shannon’s work
tended to interpret the theory in terms of uncertainty reduction for the receiver
of the signals — what Ritchie (2003) calls “epistemological probability.”
Such a definition of information is somewhat counterintuitive because
we tend to associate information with certainty, rather than uncertainty (Miller,
1983a). In fact, Shannon had been advised by computer scientist John von
Neuman to call his concept “entropy” rather than “information” (or “uncer-
tainty,” another near synonym) because entropy was a vague term less likely to
be confused with the vague, everyday meanings associated with the word
“information” (Campbell, 1982, p. 32; Machlup & Mansfield, 1983, p. 48).
Indeed, these opposite or “negative” forms of Shannon’s definition appeared
in the writings of physicist Leo Szilard in 1929 and philosopher Charles Pierce
in 1878 (Morowitz, 1991).
To demonstrate how easy it is to misunderstand Shannon’s notion of
uncertainty when we apply it to human communication, Miller (1983a, p. 495)
provides the example of the sentences “Rex is a dog” and “Rex is a mammal.” The
latter sentence contains terms less likely to appear in everyday usage, so according
to Shannon’s measure it would carry more “information” that is, a rarer, more sur-
prising, message. But the term “dog” is more specific than “mammal” (which
could be a dog, a bat, a dolphin, or many other creatures); semantically, there-
fore, we would judge that “Rex is a dog” carries more information, reversing
the logic of Shannon’s measure. It is all too easy to misinterpret Shannon’s
definition of information outside the realm of signal transmission (Losee, 1997).
Common misunderstandings of Shannon’s Information Theory are partly
attributable to his coauthors and advocates.Warren Weaver, a physicist, was invited
to write an introduction to two journal articles by Shannon that the University
of Illinois was publishing under the title The Mathematical Theory of Communication.
In his introduction to Shannon’s work, Weaver speculated on how Shannon’s
model of signal transmission might be applied to human communication.Weaver
acknowledged in a later publication that the model could be taken too literally:
“Information must not be confused with meaning,” Weaver said (1949, p. 8).Thus,
he anticipated that some scholars would attempt to extend the theory to the
subjective interpretation of signals by humans.
48 3.The Concept of Information

Nevertheless, Weaver’s broad-ranging analogies and speculations became

conflated with the very limited theorems devised by Shannon, much to the
latter’s chagrin; according to Ritchie (1986, 1991), it is Weaver’s extrapolation
of Shannon’s model to which most writings refer, not to Shannon’s original
explanation and theorems. To confuse matters further, Shannon himself was
inconsistent in his use of the terms uncertainty and entropy (Cole, 1993).
Despite its flaws, Shannon’s simple depiction of signal transmission as linear,
one-way process was seen by many scholars as an adequate model of human
communication.Additional interpretations (e.g., Berlo, 1960) resulted in inevitable
simplification and distortion of the model. David Berlo’s famous “Source-
Message-Channel-Receiver” model (Rogers, 1986, pp. 86–90) dropped the
“signal” component of Shannon’s model. Conflating the concepts of message
and signal ignored an important distinction between meanings (messages) and
their encodings (signals).
For several decades, various simplified versions of Shannon’s model became
the basis for studying the exchange of messages among people (Rogers, 1994).
As Jesse Shera and Donald Cleveland (1977) put it,

Everybody tried to get into the act, hopeful that Shannon’s magical formula
would unlock countless information secrets and give a quantitative measure for
laying a scientific theoretical foundation for practically every major field lacking one.
Unfortunately this overextension was generally an intellectual get-rich-quick
scheme and, in the long run, most of the hopefuls fell to the wayside. (p. 261)

A 1974 article by James Watt and Robert Krull can serve as an example
of how some researchers applied Shannon’s concepts. In a study of television
viewing habits,Watt and Krull noted that other researchers had no classification
system in common for program contents. Some would use categorizations like
“news, mystery drama, situation comedy, quiz-audience,” while others used
“documentary, crime-detective, comedy-variety, game shows” to cover the same
content. Obviously, these variations posed problems for researchers trying to test
for effects of television viewing, given that the results independent investigations
were not directly comparable.
Instead of subjective classifications of content,Watt and Krull argued that
“structural or form characteristics of the program may also have an effect on the
audience”; therefore, they proposed a “content-free measure of television pro-
gram form” that applied “information theory entropy terms” to features that
appeared on television screens (1974, p. 44–45). In a study of adolescent view-
ing, they developed several formulas for measuring various aspects of a broad-
cast. For instance, “verbal time entropy is defined as the degree of randomness
of the time of audible behavior on the part of characters in a program”; a formula
measured this in terms of a negative, logarithmic function of the time that a
series of television characters produced audible sound. I do not need to repro-
duce Watt and Krull’s various formulas to convince you that, although they did
3.2. Definitions of Information and Their Problems 49

indeed carefully measure “nonrandom viewing patterns,” their results are diffi-
cult to interpret in terms of what normally concerns us about television viewing
and its effects.
Nevertheless, Ritchie (1991) describes Watt and Krull’s work as among
the more successful applications of Shannon’s entropy measure, along with Seth
Finn’s studies of unpredictability in news articles (1985 and 1986). Less success-
ful, in Ritchie’s view, were attempts by Garner (1962) and Hsia (1968) to apply
entropy to neurologic and cognitive information processing.
Eventually, the Shannon and Weaver model came to be seen as inadequate
for expressing many of the important features of human communication.
As early as 1969, Donald MacKay complained that

communication engineers have not developed a concept of information at all.

They have developed a theory dealing with only one particular feature or aspect of
messages “carrying” information — their unexpectedness or surprise value. (pp. 56–57)

Nørretranders (1991/1998, p. 96) observes that Shannon’s view of infor-

mation equated it with “something completely meaningless, something closely
related to disorder … quite unlike what the rest of us understand by the
everyday word ‘information’ — meaning, content, overview, order.” It is what
Søren Brier (1992) calls a “mechanistic concept of information,” which reduces
human cognition to the level of computer processing. Shannon’s so-called
“Information Theory” simply did not adequately reflect the way in which
people interpret and assess the “meaning” of messages. As the Canadian
sociologist Orin Klapp (1982) concluded:

Meaning, being subjective, and referring to synthetic or holistic properties

that cannot be reduced to the sum of parts, might be called a higher sort of
information that does not come easily, let along inevitably, from a growing heap of
mere information. (p. 58)

3.2.2 Five Problematic Issues in Defining Information

Despite its popularity, Shannon and Weaver’s implied definition of infor-

mation contains several assumptions and requirements that differ with the ways
we usually think about and experience similar phenomena in everyday life.
Their definition is useful only in a very limited sense.
The Shannon and Weaver definition is not alone in posing such problems.
Most information concepts contain assumptions regarding five issues that often
turn out to be problematic when we try to apply their definitions.The five types
of assumptions are about the following:
Utility: Does information, in order to be information, have to have some
kind of effect, some sort of usefulness for humans? If not, what would be the
50 3.The Concept of Information

point of talking about it? In particular, must information reduce uncertainty

about something? If information does not reduce uncertainty, must it be useful
in some other way — e.g., in providing entertainment or some other kind of
useful stimulation?
Physicality: Must information always take on some physical form, such
as a book, the sound waves of human speech, or a natural object that embodies
some kind of data? Is it even proper to discuss what people know, or believe, as
being information? A related, and perhaps distinct, issue is whether information
(or at least its effects) must be directly observable. If the effects are not directly
observable, then how can it be the subject of scholarly study? This latter ques-
tion bears on Belkin’s (1978) seventh and eighth requirements of an information
concept, which he calls methodological, and will be addressed in Chapter 6.
Structure/Process: Must information be structured in some way? That
is, must it be composed of elements in fixed relations to one another, or in some
way consist of a complex “whole,” such as an image? Or is information a process,
some kind of function, a series of steps — a sort of recipe?
Intentionality: When studying information, is it necessary to assume that
someone (or something) intends to communicate it to another entity? Or is some
information simply out there in the environment, to be perceived and inter-
preted by a sentient organism? For instance, we can imagine circumstances in
which information is not communicated with a purpose in mind (requirement 1);
a glance at threatening clouds informs us that rain is imminent, but in this case the
generator of the message is an aspect of the natural world and has no intentions.
Truth: Must information, in order to be information, be true? Is it
improper to call something information if it is demonstrably false? If so, then we
need another term for that which is untrue, such as misinformation.
Let’s examine each of these issues in turn, and consider what various
authors have written about them.

3.2.3 Utility as a Requirement

As Fred Dretske says (in “Putting information to work,” his essay in a

volume about language and cognition), the concept of information isn’t fruitful
if it doesn’t account for an effect of some kind:

Information isn’t much good if it doesn’t do anything. … a difference that

doesn’t make a difference isn’t really a difference at all .… I mean a causal difference,
a difference in the kinds of effects it has. (Dretske, 1990, p. 112).

Dretske’s concern is the relationship of “information” to causation and hence

explanation of behavior; he doesn’t think that information is the same thing
3.2. Definitions of Information and Their Problems 51

as a “belief ” or a “meaning.”We can accept Dretske’s claim at a superficial level:

it is not worth discussing a concept that makes no “difference” in the world. Yet
this leaves many questions unanswered: Exactly what, how, and when are these
effects?
First, it is necessary to say something about the relationship between
information (or knowledge) and power (see Braman, 1989, 2006, for a fuller
discussion of this topic). Suffice it to say that information can have powerful
effects on humans, but usually does not. Yes, it is easy to imagine that, if we
knew in advance which team would win the World Cup or which stock would
increase in value, we could easily turn that information into a large sum of
money — which in turn could be used in various, powerful ways — but the
accurate prediction of future events is rare. Of course if we know other
people’s secrets we may be able to compel them to do our bidding by threat-
ening to disclose the secrets — an unusual and perverted kind of power. And
it is true that having specialized knowledge (e.g., of medicine) may grant one
certain privileges and enable one to attract money. But such “formal knowl-
edge” is closely bound to issues of performance, competence and (especially)
institutions and social relations. As Eliot Freidson, (1986) says, knowledge must
have agency in order to exercise power. Many discussions in the “information
is power” vein tend to underplay the complex social relations (e.g., organi-
zations like universities, governments, and corporations) or material objects
(e.g., properties or weapons) typically involved in accumulating, maintaining,
and using power.
In contrast, most of the “information” in our possession has little value or
effect in and of itself. Most of what we know results only in subtle changes or
uses in the real world. Some examples of relatively “powerless” information: the
current time and temperature; the names of capitals of the world’s nations; or
what the person next to me just said.
Another utility of information is one in which there has been much more
scientific interest and evidence: the ability of information to reduce uncertainty.
From the 1950s through the 1970s, definitions of information proliferated, most
of them incorporating the uncertainty aspect of Shannon’s model. A review by
Bouazza (1989) reflects the majority view regarding the role of uncertainty in
these definitions: “The most cited and perhaps the most useful definition of
information is ‘that which reduces uncertainty’” (p. 145).
In the interest of brevity, we will ignore definitions that try to preserve
major portions of Shannon’s definition (e.g., those discussed by Artandi, 1973,
and Fairthorne, 1975), as they are too narrow to be compatible with common-
sense notions of information. The most common drawback of these early
attempts is that they insist on defining information in terms of uncertainty
reduction — typically in the execution of a choice or decision. For example,
Wersig and Neveling (1975) declare that
52 3.The Concept of Information

The basic term “information” can be understood only if it is defined in

relation to … the information needs of people involved in social labour …. Either
as reduction of uncertainty caused by communicated data. Or as data used for
reducing uncertainty. (p. 138)

Like many such definitions,Wersig and Neveling’s implies that information

must be useful (“involved in social labour”) and intentional (“communicated”).
Similarly, Everett Rogers defines information in terms of reducing
uncertainty in a decision task, as “patterned matter-energy that affects the
probabilities of alternatives available to an individual making a decision”
(Rogers, 1986, p. 85). Other writers leave out the assumption that a task is being
performed, but they still cling to the uncertainty component, in which infor-
mation is whatever “removes the doubt, restricts the uncertainty, reduces the
ignorance, curtails the variance” (Nauta, 1972, p. 179).
The emphasis on uncertainty continues to current times. For example,
one book on the nature of information begins with the words “The concept of
information, which is the subject of this book, is intimately connected with the
concept of uncertainty … information in a given context is obtained by a cog-
nitive agent whenever relevant uncertainty is reduced” (Klir, 1996, p.VII).This
is a typical approach for defining information among economists. Hirshleifer
(1973) for example, discusses the probability of signals but draws a contrast with
the way that probability is applied (p. 31):

Uncertainty is summarized by the dispersion of individuals’ subjective

probability [or belief] distributions over possible states of the world. Information,
for our purposes, consists of events tending to change these probability distributions.
A rather different concept of ‘information’ is employed in communications and
statistical theory, according to which a dispersed probability distribution is called less
‘informative’ than a concentrated one. This latter concept uses the term
‘information’ merely as a negative measure of uncertainty.

It is important to note once again that while this writer refers indirectly
to Shannon’s work, he is actually discussing epistemological probability —
likelihood estimates in the “real world” of human communication.
The idea that information must be useful to be information has been
undermined in several critiques, most notably by Fox (1983) and Losee (1997).
The latter observes that

a good definition or theory of information … should bear some resemblance

to the natural language notion of information but need not adhere to it when the
natural language definition loses its generality and explanatory power.This happens
when the common language definition of information, for example, becomes
conflated with the notion of useful information, that is, information is understood
to be in all cases useful. For those accepting this concept of information, if it is not
useful, it is not information. Requiring that all information be useful limits the domain
3.2. Definitions of Information and Their Problems 53

of discussions about information to cognitive processes that can “use” something; it

excludes the information carried by a subatomic particle which is not sensed by a
cognitive process.We try to avoid excluding information phenomena. (p. 257)

Neither must information (intentionally communicated or otherwise)

automatically reduce uncertainty.To use an example from Fox (1983), suppose
that I tell you something but you do not believe me. I have not reduced your
uncertainty. Or suppose I tell you that the stock market crashed this morning.
Last you heard, the market was going up, so perhaps I have reduced your uncer-
tainty about the overall direction of the market (assuming you have been won-
dering about it), and yet I have probably created uncertainty within you
regarding collateral knowledge (e.g., the values of your individual stock hold-
ings). As Fox points out, there are other scenarios in which the communicating
of information might actually increase uncertainty, rather than decrease it.
Some other disciplines that initially exploited uncertainty reduction have
recently questioned the ubiquity of such a psychological drive. For instance, in
the interpersonal communication literature, we have the emergence of Problematic
Integration Theory (Babrow, 1992) and Uncertainty Management Theory
(Babrow, Kasch, & Ford, 1998). Both of these theories question the assumption
that humans always strive to reduce uncertainty.While most discussions empha-
size the negative effects of uncertainty, some other scholars (e.g., Brashers,
Goldsmith, & Hsieh, 2002; Huber & Sorrentino, 1996) identify positive psycho-
logical effects of uncertainty — or instance, it may provide “hope.”
Uncertainty Management Theory, in particular, highlights how people
sometimes deliberately increase uncertainty. Uncertainty Management Theory
holds that uncertainty is experienced “not simply as an uncomfortable tension
demanding reduction” (Bradac, 2001, p. 463) but as feelings and cognitions that
can be managed in other ways as well; these may include “seeking instead ambi-
guity and even confusion” (Bradac, 2001, p. 471). This is because “individuals
may use uncertainty as a tool … sometimes this cognitive state will be cultivated,
rather than eradicated” (Bradac, 2001, p. 464).
Two examples of the deliberate increasing of uncertainty in interpersonal
conversation are apparent when a physician must deliver a threatening diagno-
sis to a patient. One party to the dyad, the physician, might choose to provide
an uncertainty increasing message when he or she believes the patient is certain
of bad news (Ford, Babrow, & Stohl, 1996); the patient, in turn, might avoid
information in order to maintain uncertainty, or even seek out uncertainty
increasing information (Brashers et al., 2002). In both cases, increased uncer-
tainty might actually provide some increase in comfort for the patient, even
though in a way that might compromise his or her treatment. Afifi and Weiner
(2004) give other examples in the health care context whereby individuals
deliberately “avoid relevant information” (p. 182).
54 3.The Concept of Information

Other scholars have also debated the effects of new information on

uncertainty. Lang, Newhagen, and Reeves (1996) imply that cognitive capacity
limits may prevent new information from reducing uncertainty. Berlo (1977)
notes that information always reduces uncertainty in the “now,” but in the long
term it may have the opposite effect. Yovits and Foulk (1985) conducted an
empirical study in which they tested the assumption that information always
reduces uncertainty; they found that sometimes new information made their
subjects less sure that their evaluations of a problem were correct. Similar con-
clusions are reached by Kellermann and Reynolds (1990) and Robertson (1980).
Although uncertainty is not satisfactory as a basis for defining informa-
tion itself, it is nevertheless an important concept for information seeking. Even
though information can be encountered in a passive way, actively acquiring
information implies recognition of uncertainty or anomalies at some level.Kuhlthau
(1993b) makes good arguments for considering uncertainty as a beginning stage
in the process of finding information, and Yoon and Nilan (1999) demonstrate that
one cannot study uncertainty without considering what informants already know
(i.e., certainty).
What about other effects of information? That is, if information does not
always reduce uncertainty, must it have some other utility to be considered
“information”? Some authors include in their definitions of information almost
any kind of stimulation that humans find useful. Examples of such stimulation
could include sound (from music or from a waterfall); sight (the words of a
novel, the images of a painting, a photograph, or a film), or touch (the feel of
warm sunshine or cool water on skin). Each of these sensations “tells us” some-
thing. In Chapter 5, and obliquely below under “Truth,” I argue that such stim-
uli include potentially useful information. Therefore, if information must have
an effect, it should extend to outcomes beyond “reducing uncertainty.”

3.2.4 Physicality as a Requirement

Everyone acknowledges that information can have a physical form (e.g.,

see Buckland, 1991a, and McCreadie & Rice, 1999), but few explicitly argue
that it must. Indeed, many scholars take pains to state that a more useful con-
ceptualization of information is as a phenomenon that exists apart from physi-
cal media; that is, that we should not think about information as primarily
something found in human-created messages like printed texts.
However, as Klaus Krippendorff (1984) has pointed out, at some level
information must have a physical form. He argues in favor of information defined
as “a change in an observer’s state of uncertainty caused by some event in his world”
(sic, p. 49). This, of course places Krippendorff ’s definition in the “uncertainty
3.2. Definitions of Information and Their Problems 55

requirement” camp we have already discussed. But the interesting aspect is how
he analogizes information to energy:

First, neither energy nor information exists in a vacuum. Both are embodied
in material processes to which one must refer. Just as one can speak of energy only
in conjunction with some specific resource, fuel or storage capacity … so one can
speak of information only in conjunction with a physically identifiable source, a
message or a situation as described by an observer, and relative to what he already
knows. Second, energy and information are measures of work … information is a
measure of the (intellectual) work required to distinguish, to a degree better than
chance, among a set of initially uncertain possibilities. (pp. 49–50)

Certainly information in the sense of thoughts has a physical dimension:

the electrical impulses of a human nervous system. In any event, it does not pose
much of a restriction to contend that information must have a physical compo-
nent, in the sense that energy does. In Krippendorff ’s case, it is much more of a
restriction to require that information reduce uncertainty — as his Shannonesque
conceptualization makes clear.

3.2.5 Structure/Process as a Requirement

Other families of definitions avoid the uncertainty concept through use of

analogy — typically to a structure or process — and sometimes require inten-
tionality to do so. Kenneth Boulding (1956) used the analogy of an image, or a
“picture in our head” in his popular characterization of messages and meaning.
In Boulding’s view (1956, p. 7),“the meaning of a message is the change which
it produces in the image” — i.e., the image of reality (or a portion of it) that
exists in someone’s mind.This conception of information is similar to the one
offered by Bateson at the beginning of this chapter, defining “change” as “difference
which occurs across time” (1972, p. 452).
Expanding on Boulding’s analogy, Pratt (1977) defines information as an
event: “That which occurs within the mind-upon-absorption-of-a-message”
(p. 215).That is, information (or an “informative event”) is what we call a change
in one’s mental image. For Pratt, then, information is the event that changes
someone’s image of reality. Like those of Boulding and Bateson, Pratt’s definition
has ties to the “Internal Information” described by Dervin and Rubin earlier
in this chapter.
The “image” metaphor is evoked by Donohew and Tipton (1973) as well.
But, for them, the idea of “image” is not something like a holistic picture, but
rather a complex mental structure of parts and subparts:

An individual’s “image of reality” is divided into three parts. First are the goals,
beliefs, and knowledges [sic] which an individual has compiled as a result of his lifetime
56 3.The Concept of Information

of experiences.These cognitive “objects” are defined as any concepts, issues, material

objects, or ideas which exist psychologically for a person …. The second part of an
individual’s image or reality is the concept of self.This includes an evaluation of his
ability to cope with various situations …. The third part of the image of reality is
an information-handling “set” developed out of past experiences.The “set” probably
controls the selection of information used by the individual to cope with the
environment. here we are talking about an individual’s information-seeking and
processing “styles.” (pp. 246–247)

Similarly, MacKay argued that information must be considered in the

contest of a hierarchy of goals (Cornelius, 2002, p. 413).
Other authors have conceptualized information as a structure or organiza-
tion of experience and sensory data, for example,Thompson (1968) and Belkin
and Robertson (1976). Following Thompson’s definition, Belkin and Robertson
(1976) state that “information is that which is capable of transforming structure”
(p. 198) — in other words, it changes the knowledge state of the recipient.
A parallel characterization comes from MacKay (1969), in which information is
“that which does logical work on the organism’s orientation” (p. 95).
Belkin (1978) notes that characterizing information as something that
transforms knowledge structures has its problems, but it relates well to informa-
tion as it has been defined in a variety of disciplines. Given that information-
as-process assumes that a process has an effect on some entity — as an alteration
of a mental image, or the creation of meaning in a human mind — the process
and structure views of information are analogically similar. In some definitions
of this type, individual authors have added one or more restrictions to serve
their purposes. Belkin and Robertson (1976), for example, are concerned with
document retrieval systems and therefore assume that messages are intentional
and that messages are represented by texts: “a collection of signs purposefully
structured by a sender with the intention of changing the image-structure of a
recipient” (p. 201).
Charles Cole (1994) notes a dilemma that accompanies the assumption
that information changes a cognitive structure: if “new” information can
modify knowledge structure, then it must be so that “old” or “expected” infor-
mation does not modify knowledge structures in the same way, or at least at the
same time. New, or “pure,” information must be extremely rare because such
information is completely unanticipated, and there is a natural tendency not to
recognize, see, or perceive that which is unanticipated. Information, then, has
the quality of being unexpected and expected, old and new, at the same time.
Therefore, for information to be unexpected and expected, old and new, at the
same time, information must enter the perceptual system in at least a two-stage
process.
More recently, Robert Losee (1997) has advanced a general and coherent
definition of information. He attempts to resolve some of the conflicts between
3.2. Definitions of Information and Their Problems 57

definitions of belief, knowledge, information, and misinformation by viewing

information generically as processes that produce outputs. The processes, or
functions, may be invoked by humans, or machines, or other entities.The inputs
into the process can be perceived from the environment or retrieved from
human memory.The output of the process (e.g., the value taken on by a vari-
able) is informative about both the original inputs and the process that
produced the output from those inputs. Thus, Losee defines information as
“the values of characteristics in the processes’ output” (p. 256).
In Losee’s view, by examining any output we can usually infer something
about the process that created it. Examining a tree, for example, informs us
about its origins, soil, moisture, and growth process. Or a cake, which is created
through a procedure that includes ingredients, instructions, and heating, may be
inspected to determine some, but not all, of the ingredients and the process by
which it was created.
Defining information in terms of the output of a process, Losee con-
cludes, moves beyond discipline-specific definitions (such as the tendency for
decision theorists to define information in terms of uncertainty reduction)
and provides a link between various studies of information.Yet, while Losee’s
definition sidesteps some of the criticisms of Fox regarding information-
as-process, it substitutes the vagueness of “process” with a mysterious
“function” that takes input and returns a value to be attached to variables
(or “characterizations”). This in turn begs the question of how variables
emerge in the first place and what determines the nature of the functions.
These are not necessarily questions that Losee is responsible for answering,
but rather problems to be faced by researchers trying to apply his definition
of information. Hörz (1996) is yet another commentary on the notion of
information as structure or process.
Fox (1983) criticizes all structure-based definitions as failing to provide
a clear definition of “structure” itself. Fox also makes compelling arguments
that information cannot be considered either an event or a process —
although he also provides counterexamples supporting the process view of
information.
Fox himself favors defining information as a type of “telling,” as repre-
sented in propositions. It would be impossible to fairly convey Fox’s arguments
in any shorthand version, presented as they are in 213 pages. It may suffice to
say that his is the most extended dissertation on the subject thus far. Fox sum-
marizes his conclusions in this way:
Information need not be true, though misinformation must be false; information
need not be believed by anyone; information need not originate with a reliable
informant, but it must originate with someone in an appropriate position to know.
Ontologically, information is propositions [the identification of which] depends on
contextual factors. (pp. 212–213)
58 3.The Concept of Information

Fox admits that his conclusions leave several issues unresolved, including the

crucial notion of the amount of information carried by a set of sentences

remains unanalyzed.The notion of informativeness remains unanalyzed …. The details
of how meaning determines propositional content as a function of context is not
well understood. (p. 213)

In the intervening years since Fox’s book,Tor Nørretranders has published

(1991/1998) a text that addresses the relation of context to content.
Nørretranders introduces the term exformation to describe the ways in which
messages may refer to a “mass of information” that is “not present” and “explic-
itly discarded” but nevertheless is understood to be relevant by the receiver and
is used in construing the meaning of a message (p. 92).
Nørretranders provides two examples to illustrate the concept of exfor-
mation, one involving an extremely short message and the second no message
at all.The first example is Victor Hugo’s famous query to his publisher regard-
ing the appearance of his latest novel, Les Misérables, in 1862. On vacation and
out of touch with news about public reaction to his work, Hugo mailed a
letter consisting of a single character: “?” His publisher replied, simply, “!”
Without prearrangement, both parties understood these exchanges to mean
something like the question “How is my book selling?” and the response,
“Surprisingly well!”
Nørretranders’s second example corresponds to the saying “no news is
good news.” When parents do not receive a phone call from their son away at
college, they assume that he is OK and that things are fine. Information has been
conveyed without sending a message at all. This example echoes Cole’s (1994)
observation about “new” information versus “old” or “expected” information:
can “no news” be viewed as merely preserving the original knowledge struc-
ture — or does it still modify the structure, but in a different way?

3.2.6 Intentionality as a Requirement

The manner in which Fox characterizes information has a critical limita-

tion. His analysis is based, by necessity, on propositions expressed in the form of
sentences. (“In this work I deal only with information carried by sentences,”
1983, p. 7). Fox notes that his propositions represent “what is asserted to be the case
by (someone who writes or utters)” (p. 77). One problem with this limitation is that
it takes us back into the assumption of a message intentionally sent by a sender
to a receiver.
We could call this type of intentionality the “communication assumption” —
that information necessarily involves communication, and hence, intention to
communicate. Bowers and Bradac (1982) see the presence of intentionality as a
3.2. Definitions of Information and Their Problems 59

key dividing point among rival definitions of “communication.” Their exami-

nation of 27 metatheoretical discussions of communication finds that 18 of their
authors hold that intentionality is a requirement for communication to exist.
Few of those theorists have an unusual definition for “intention”; most mean
the concept in its usual sense: a “purposeful activity … [that] must be explained
by ‘in order to’ as well as ‘because’ statements” (Bowers & Bradac, p. 7).
Although the restriction of intentionality may hold true for what is the
most important sense of information — the exchange of information between
humans (e.g., see Buckland, 1998) — it does not apply to all senses in which we
use the word. Information may originate outside natural language propositions,
for example, as signs occurring in our environment. Whether we are viewing
the natural world (e.g., trees, animals, rocks) or the human-made world (e.g.,
what people are wearing and doing, or a printed sign that says “exit”), we can
take in stimuli that have meaning.The only way to retain the notion of inten-
tionality is to assume that it can refer to either a “sender” (“someone who writes
or utters”) or “receiver”(the viewer of the world), but does not necessarily
involve both ends of a communication process.
If we believe that people must intend to receive in order to take in informa-
tion, then information is, in this more limited sense, intentional. Intentionality
solely on the part of the receiver was suggested by Westley and Maclean in
1957, and in a discussion of news-seeking behavior, by Westley and Barrow
(1959).The latter described “the need of the selecting receiver to be oriented in
his extended environment” (p. 431); this assumption would take in the kind of
“viewing” that I discussed previously. Theirs was a rather radical conception of
communication, because it did not assume that a sender’s intentions were involved;
hence, Bowers and Bradac count it among the “nonintentional” definitions of
communication. However, it is an intentional view of information behavior.
This is a different view from that of Stonier (1990, p. 21), who claims that

Information exists. It does not need to be perceived to exist.

It does not need to be understood to exist.

3.2.7 Truth as a Requirement

Losee (1997) also considers the notion of “misinformation.” He notes

that information can have various flaws, including inaccuracy, incompleteness,
lack of justification, and intent to deceive. Do we need a special label for
information that is so flawed as to be untrue?
Traditionally, philosophers have made a distinction of this type regarding
knowledge, the common stance being that knowledge is “justified true belief.” In
this definition, belief is taken to be “the most elementary of our opinions …
characterized by two qualities: … either true or false ... arrived at either rationally
60 3.The Concept of Information

or nonrationally” (Cherwitz & Hikins, 1986, p. 31). By justified is meant that the
believer has sufficient, relevant evidence that his or her belief is true.The “justified
true belief ” definition of knowledge has been criticized since the analyses of
Russell (1959) and Gettier (1963) but continues to have many advocates. More
to the point, no major philosophers have extended the requirements of truth
and justification to the concept of information.
Patrick Wilson’s quote at the beginning of the next chapter suggests that
the truth or falsity of information is something that we can ignore in discussing
information in the abstract sense. For one thing, it could be argued that we
rarely know for sure if something (a statement or perception) is true or not;
even if a “fact” is demonstrated to be true at this moment, it may be possibly
proven wrong a few moments later. Fox (1983, p. 212) and Derr (1985, p. 496)
also hold that information need not be true, based upon analyses of usage of the
term in ordinary discourse. Buckland (1991b) concludes that

the question of whether specific bits of knowledge are true is not central to
our concerns. We adopt the position that the process of becoming informed is a
matter of changing beliefs. Whether these beliefs are held or denied by others and
whether they are compatible with some a priori or fundamental assertion need not
detain us. (p. 43)

Some would disagree with this point of view and instead argue that a
true–false distinction is worth keeping in defining “information.” Frické (1997),
for example, argues that information should be “truthlike” in order to “fit the
world” so that we can “succeed in our interactions with the world” (p. 888).
Dretske (1981, 1983) makes a similar case, that information must tell us truly
about a state of affairs, such that we can learn from it; he concludes that
false information and misinformation cannot be considered to be varieties of
information, but rather distinct concepts.
However, we are concerned here with a broad view of information phe-
nomena that fits both real life and empirical studies of real life, not with estab-
lishing a philosophical distinction. Studies of information seeking provide many
examples in which people value information that they know not to be entirely
true. For purposes of this text, then, we will generally ignore any distinction
between the truth or falsity of information, unless such is the focus or finding
of a given investigation.

3.3
Must There Be a Universal Definition of Information?

All of the definitions we have examined have taken a stance on one or

more of these issues. The distinctions and disagreements among reviewers of
3.3. Must There Be a Universal Definition of Information 61

definitions are too many to resolve; in short, there is as yet no single, widely
accepted definition for the concept of information. At least among recent
reviews, however, there has been some agreement on the types of definitions of
information that exist.
To summarize the chapter thus far, we can see that there have been many
attempts to characterize information, some of them quite broad (e.g., the
Image/Event/Structure/Process definitions), while others have been very
narrowly focused (e.g., the view of information as a selection of signals from a
well-defined set of symbols, the reception of which may reduce uncertainty for
the receiver of the signals).
Narrow definitions assume one or more of the restrictions discussed earlier
in the article.They hold that information must be useful, or that its transmission
is intentional, or that it must be represented in an recordable medium (in written
or spoken language, or images), and/or that information must be true (or at least
easily verifiable). Shannon’s model contains examples of all of these assumptions.
The vast majority of the early definitions and investigations of information
seeking include at least the first three of these assumptions.
Before we go any further, it is important to note that defining informa-
tion in an absolute and final sense is not entirely necessary for the study of
information phenomena to proceed.As Artandi (1973) and others have pointed
out, all we need are useful conceptualizations of information. Belkin (1978) makes
this point most effectively when he says,

we are not concerned with definitions of information, but rather with concepts
of information. The distinction is that a definition presumably says what the
phenomenon defined is, whereas a concept is a way of looking at, or interpreting,
the phenomenon … by accepting the idea of a concept one becomes free to look
for a useful concept, rather than a universally true definition of information. (p. 58)

If we wish to keep talking about “information,” we may have to give up

trying to define it rigorously. Carl Hempel (1952) notes that there are some
terms in any conceptual scheme that are so basic that they need not be fully
explicated. Hempel calls these basic concepts “primitive terms.” Primitive terms
are simply accepted as they are commonly understood. Chaffee (1991, p. 7) pro-
vides the example of the concept of a “person,” or a “human.” Perhaps there are
some fields in which what we mean by “human” needs to be carefully defined
and is subject to debate — in zoology, for example. But, for most purposes of
study, we do not need to explain that particular notion. It is when we deal with
concepts that build upon the notion of humans — family, community, society —
that we are in need of careful definitions for those concepts.
Information can be, and has been, treated as a primitive term as well.
Some writers believe this approach is problemmatic — see Frohmann (2004,
p. 86) and Nunberg (1996, p. 110).Yet it has been a common practice in the
62 3.The Concept of Information

IB literature; commenting on a variety of recent studies, Pertti Vakkari (1997)

notes that

one of the striking features in many studies was the use of the central concepts,
like information, knowledge, information need, seeking, and use as primitive concepts,
i.e., without definition. (p. 460)

Can we reconcile the various definitions of “information” with one

another? It does not seem so, and perhaps it is not necessary. Although there is
scholarly disagreement over the “most rigorous,” or “most easily quantifiable,”or
“most productive,” or “most parsimonious” meaning of “information,” these
debates have done little to promote a fuller understanding of the concept
among a community of scholars. In fact, if anything such discourse has resulted
in a fracturing of scholarly effort in studying the phenomenon of information;
it has resulted in too many definitions that defy comparison and that provide
no common basis for understanding.Thomas Allen (1969) and Brenda Dervin
(1977) discussed using the concept of “communication” instead (see the insight-
ful discussion by Bernd Frohmann, 2004, pp. 53–67). Similarly, Jonathan Furner
(2004) argues that there are so many productive substitutes for the concept of
“information,” that we could do without it all together:

… philosophers of language have modeled the phenomena fundamental

to human communication in ways that do not require us to commit to a separate
concept of “information.” Indeed, we can conclude that such a concept is
unnecessary for information studies. Once the concepts of interest have been labeled
with conventional names such as “data,” “meaning,” “communication,” “relevance,”
etc., there is nothing left (so it may be argued) to which to apply the term
“information.” (p. 428)

Instead, let us treat “information” as a primitive term, as a phenomenon

that we all recognize when we see it in its various forms (Fox, 1983, p. 16).
Information would then be treated as “anything that exists psychologically for
a person” (Carter, 1965; Chaffee, 1991, p. 9).We have only to look around us to
establish the fact that information exists in the form of physical objects (what
Buckland, 1991a, calls “information-as-thing”); and hundreds of studies have
documented that people believe that information exists as a psychological
object as well — a disembodied result of “becoming informed.”
Fox (1983) observes that the “ordinary notion of information” is one
through which “information scientists apparently do succeed in communicat-
ing with one another quite effectively regarding information and related con-
cepts” (p. 5). Nunberg (1996, p. 110), as well, notes that “ ‘Information’ is able to
perform the work it does precisely because it fuzzes the boundaries between
several genetically distinct categories of experience.”
3.3. Must There Be a Universal Definition of Information 63

Allowing a broad definition of information poses problems for opera-

tionalization and measurement of concepts, as shall be seen as individual stud-
ies are reviewed in later chapters. Yet to argue for any tighter definition of
information would be to limit the scope of this book, which is intended to
review a broad spectrum of investigations having to do with information seek-
ing and sense making.Therefore, in this text, I allow for any definition of infor-
mation, however vague or difficult to study. Where more restricted definitions
of information apply in the review of individual theories, methods, or studies,
they will be made explicit.
It should be noted that the definition “any difference that makes a differ-
ence” places at least one important restriction on the scope of information: it
rules out the possibility of information existing independently of a knowing
mind. For the purpose of this text, we will assume that a conscious brain must
be engaged at some point for information to be said to exist. Otherwise, we are
back to the unhelpful stance that “everything is information.”
Two examples, the first suggested by Fox and the second by Buckland,
will help to make the restriction clear. First, Fox provides the example of some-
one who keeps a secret diary that no one else is ever allowed to read. Some def-
initions of information would imply that, since the content of the diary was
never communicated, that it cannot be considered information. (Fox uses this
straw man to defeat the requirement that a message must be transferred — i.e.,
received — to qualify as information.) Of course, a diary is a clear example of
information-as-thing, and the symbols written in the diary are an expression of
a human mind — it is some kind of message, even if never received by any other
than its creator. So, yes, a secret diary (an unviewed, human-created record) can
be safely considered to be information.
In the second example, trees could be viewed as carrying information in
the form of their growth pattern of rings, which among other things tell us
about the amount of rain that fell in a past season. Even if no person has viewed
that information, is it not still information? For our purposes, the answer is “no.”
It is nothing more than wood until someone both encounters and makes some
sense of it. So, if a tree falls in the forest and there is no one there to see it, then
it conveys no information.
To conclude this portion of the chapter, it bears emphasizing that in this
text we will consider only human information behavior, and the making or
meaning or sense, and therefore we interpret information as requiring the
involvement of a human mind. (It could be easily argued that animals use infor-
mation, but they are simply outside the scope of this book; see Bates, 2005b, for
a distinction between animal and human use of information.) As will be seen in
later chapters, a broad conceptualization of information is in keeping with the
way the term has been employed in studies of information needs, uses, seeking,
and sense-making.
64 3.The Concept of Information

3.4
Distinctions among Information, Knowledge, and Data

A side issue as regards information seeking research is worth noting.

Much attention has also been granted to defining the concepts of “data” and
“knowledge.” Machlup (1983) examines the issue of whether information is syn-
onymous with data and knowledge, noting that there has been a tradition to treat
the three as a hierarchy, with data at the bottom and knowledge at the top.
Raber (2003) raises a troubling question about such an arrangement when he
says “But at what moment and how does information become knowledge?”
Machlup holds that historical usage of the three terms does not fully jus-
tify the distinction that information is data that has been processed and/or
organized. The origins of the term in the Latin dare, “to give,” along with the
history of its usage, imply that the word data (“the givens”) can be assumptions,
facts, measurements, and so forth, expressed in either words or numbers. As
Machlup points out (p. 647), many writers claim that data are a “raw” type of
information, while a few others see information as a type of data. Machlup con-
cludes that there is neither precedent nor need to establish a hierarchy between
the two words.
The common notion that knowledge is information that has been sifted,
organized, and understood by a human brain is on firmer ground. Brown and
Duguid (2000) complain that the two concepts are unfortunately conflated:

People are increasingly eager that their perfectly respectable cache of information
be given the cachet of knowledge. Such redefinitions surreptitiously extend the
overlapping area where knowledge and information appear as interchangeable
terms. Nevertheless … there do appear to be some generally accepted distinctions
between knowledge and information …. For example, it sounds right to ask,“Where
is that information?” but odd to ask, “Where’s that knowledge?” (p. 2)

Machlup (1983) makes the useful point that “information is acquired by

being told, whereas knowledge can be acquired by thinking” (p. 644).Through
our inner experience of thought, we can form new knowledge without taking
in new information from the external environment. Information implies transfer,
says Machlup, while knowledge is a state (“knowing”). Knowledge and infor-
mation are therefore not usually the same, except that “information in the sense
of that which is being told may be the same as knowledge in the sense of that
which is known, but need not be the same” (p. 644). Robert Hayes (1993) makes
a somewhat different point when he says that “knowledge is internal; it cannot
be received but must be internally created” (p. 5).
Other authors sometimes raise the issue of truth, discussed earlier, in the
knowledge-information distinction back to the truth issue discussed earlier,
3.5. Summary 65

e.g.,:“The relationship between knowledge and truth is especially problematic”

Raber (2003, p. 8). However, Dretske (1981, p. 45) makes no distinction between
the two in regards to a truth requirement: “information is what is capable
of yielding knowledge, and since knowledge requires truth, information
requires it also.”
Decades of arguments about distinctions among the words data, informa-
tion, and knowledge (and sometimes “wisdom,” too) have not prevented the
continued use of terminological hierarchies. As Marcia Bates (2005b) points
out, a number of authors make such distinctions, a recent example being the
paper by Houston and Harmon (2002). Bates herself sees five categories of
information-like concepts, although these are not strictly hierarchical to one
another. Bates uses the terms information 1 (“the pattern of organization of
matter and energy”), information 2 (“some pattern of organization of matter and
energy that has been given meaning by a living being”), knowledge (“informa-
tion given meaning and integrated with other contents of understanding”),
data 1 (“that portion of the entire information environment available to a sensing
organism that is taken in, or processed, by that organism”), and data 2 (“informa-
tion selected or generated by human beings for social purposes”).
In this book, the usage of the terms data, information, and knowledge will
generally be used synonymously, because they are usually not clearly delineated
in studies of information behavior. Knowledge, however, is strictly a phenome-
non of the human mind, whereas data and information are often represented by
tangible, physical objects.That information usually has a physical manifestation
has often been the key consideration in past studies of information seeking.The
way that information seeking is typically approached under the new paradigm,
though, is in the sense of knowledge — something in someone’s mind — not
primarily as a physical object.
The usage of data, information, and knowledge outlined above represents
a necessary simplification of the many definitions and examples that have been
discussed in dozens of scholarly works. However, the fine distinctions made
between data, information, and knowledge are of little value in most studies of
information seeking. This book will treat information as a broad concept,
encompassing instances that would be considered unusual by some scholars.

3.5
Summary

This chapter has explored the central concept employed in studying

information seeking: information. We have seen that there are widespread dis-
agreements about what would constitute a general definition of information.
Most of these disagreements concern the issues of truth, physicality, intentionality,
66 3.The Concept of Information

uncertainty, and utility.The most common types of definitions that have emerged
assume that information is something that either reduces uncertainty or changes
one’s image of reality. In this chapter, I provide examples that suggest that a truly
universal concept of information would need to fulfill at least the following
requirements:
1. allow for common-sense notions of information used in everyday
discourse;
2. allow for unintentional origins of information (e.g., observations of the
natural world) as well as for purposeful communication among people;
3. allow for internally generated information (e.g., memories, construc-
tions) as well as externally generated information (e.g., reading a text);
4. allow for types of information beyond that needed for “solving a prob-
lem” or “making a decision”;
5. admit the importance of informal sources (e.g., friends) as well as
formal sources (e.g., data or documents); and
6. involve the human mind, either in the creation, perception, or inter-
pretation of information; to leave out such a requirement is to declare
that anything is information and that would leave us with no focus in
our investigations.
I have considered numerous distinctions made over the years, but I argue
in favor of treating information as a primitive concept that is so basic to human
understanding that it does not require a tight definition. To the extent that
information needs a definition it must be a broad one, such as “any difference
that makes a difference” — in essence, implying a change to the structure of a
human mind. Such a characterization, vague though it is, would allow us to
consider what many authors have said about information seeking without
having to worry about whether they restricted their observations to phenome-
non that must be true, observable, physical, intentional, and so forth.
In the next chapter, I will build on the initial discussion of information
to define information needs. Following that, Chapter 5 ventures farther afield
to consider more peripheral concepts and behaviors related to information
seeking. A review of these other concepts is important in addressing several
vexing questions about information-related behavior:
● Why do people seek information?
● What makes information relevant?
● Can information be found without intentionally searching for it?
● Is it possible to have too much information?
● Why do people sometimes avoid information?
● How does information differ from entertainment?
These and other issues are taken up in Chapters 4 and 5.
Recommended for Further Reading 67

Recommended for Further Reading

Belkin, N. J. (1978). Information concepts for information science. Journal of Documentation, 34,
55–85.
This lengthy article explains Belkin’s view of information and articulates his concept of the Anomalous
State of Knowledge (ASK) and how it is resolved.
Boulding, K. (1956). The image: Knowledge in life and society. Ann Arbor, MI: University of Michigan
Press.
An oft-cited book by the unconventional thinker Kenneth Boulding. Boulding uses the analogy of an
“image”to discuss how we come to know our world.
Capurro, R. and B. Hjørland (2002).The concept of information. In B. Cronin (Ed.), Annual Review
of Information Science and Technology, (vol. 37, pp. 343–411). Medford, NJ, Information Today.
A comprehensive discussion of the history of definitions of information.
Frohmann, B. (2004). Deflating information: From science studies to documentation. Toronto:
University of Toronto Press.
While focused on the discourse surrounding the concept of “scientific information,” the initial chapters
of Frohmann’s book give a fascinating account of how information seeking, use, and other behaviors have
been studied over several decades.
Schement, J. R. (1993). Communication and information. In J. R. Schement & B. Ruben (Eds.),
Information and behavior (vol. 4, pp. 3–33). New Brunswick, NJ:Transaction Books.
Jorge Schement examines a variety of definitions of information. A companion article in the same
volume supplies the history of the word itself.
Thayer, L. (1987). How does information “inform”? In B. D. Ruben (Ed.), Information and behavior,
(vol. 2, pp. 13–26). New Brunswick, NJ:Transaction Books.
Lee Thayer’s amusing essay considers several problematic aspects of information as the term is commonly
used. He convincingly makes the point that information is always “from the perspective of some observer.”