Description
The present paper is concerned with classification of ‘intangibles’and what classification theory can teach us about a
‘good classification’. The present study led to three conclusions about the classification of intangibles. One is that the
value of classification lies in its function as a heuristic device, i.e. as a help construction. Accounting becomes the art of
background design. Another conclusion is that if we choose to account for intangibles it does not need to change the
transaction base of financial accounting. Rather, it is an issue how to classify and label a universe of transactions.
Finally, the dichotomy tangibleintangible should not be acknowledged as a supposition. Depending on perspective,
purpose and type of financial accounting model, other concepts may do a better job in classifying the accounting
world.
Intangibles and accounting classi?cations:
in search of a classi?cation strategy
§
Jan-Erik Gro¨ jer
School of Business, Stockholm University, S-106 91 Stockholm, Sweden
Abstract
The present paper is concerned with classi?cation of ‘intangibles’ and what classi?cation theory can teach us about a
‘good classi?cation’. The present study led to three conclusions about the classi?cation of intangibles. One is that the
value of classi?cation lies in its function as a heuristic device, i.e. as a help construction. Accounting becomes the art of
background design. Another conclusion is that if we choose to account for intangibles it does not need to change the
transaction base of ?nancial accounting. Rather, it is an issue how to classify and label a universe of transactions.
Finally, the dichotomy tangibleÀintangible should not be acknowledged as a supposition. Depending on perspective,
purpose and type of ?nancial accounting model, other concepts may do a better job in classifying the accounting
world. # 2001 Elsevier Science Ltd. All rights reserved.
1. The need of (re-) classi?cation or division
Accounting classi?cations, such as the chart of
accounts, are expressions of how we chose to
describe and, to some extent, de?ne an organisa-
tion at a certain point in time. The con-
structionÀreconstruction process of accounting
classi?cations is one important aspect of the
accounting history and, therefore, of management
history. The service functions that accounting is
supposed to ful?l are contingent on the ability of
accounting to order the world of organisations in
a (value) relevant way. This ability is challenged.
How shall we classify assets and liabilities, reven-
ues and costs and cash ?ows? What is supposed to
be ‘on-balance’ and ‘o?-balance’? Can we even
create an order between the di?erent classes?
One of today’s challenges concerning ?nancial
accounting classi?cations is to cope with an orga-
nisational world that seems to become more
immaterial than material, where resources in dif-
ferent immaterial forms act as the key production
factors. In an accounting context, this develop-
ment is re?ected in concepts such as immaterial
assets, intangibles, hybrid and intellectual capital.
Allied concepts, including virtual organisations,
relation marketing, value stars, human resource
costing and accounting, enablers and performance
drivers, are in themselves a starting point for
0361-3682/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PI I : S0361- 3682( 01) 00023- X
Accounting, Organizations and Society 26 (2001) 695–713
www.elsevier.com/locate/aos
§
This article is part of the MERITUM research program.
Meritum is the acronym for ‘‘Measuring Intangibles to Under-
stand and Improve Innovation Management’’. The research
program consists of four sub-projects: classi?cation of intangi-
bles, management control of intangibles, intangibles and the
capital market and developing and testing guidelines for
reporting intangibles. In HEC (2000) eleven preliminary papers
pertaining to the classi?cation issue are presented. MERITUM
is a joint research project among eight universities and research
institutions in Denmark, Finland, France, Norway, Spain and
Sweden. The research program has been made possible through
national research grants and a common grant through the
European Commission.
E-mail address: [email protected]
accounting classi?cations. They express a need
for altering the accounting boundaries or for
acting as a demarcation for reclassi?cation. One
way to improve the accounting art is to use new
organisational paradigms for accounting classi-
?cations. The accounting backbone is about
classi?cation and reclassi?cation, i.e. to arrange
something in a particular order. The accounting
art of portrait organisations is currently under
dramatic challenge.
The business community is experimenting with
di?erent intangible models and accounting con-
cepts (Johanson, Eklo¨ v, Holmgren, & Ma? rtens-
son, 1998). Such public authorities as Statistics
Netherlands (1998) develop and use their own
classi?cation of intangibles. Thus far, accounting
researchers seem to have developed little interest
in the ‘classi?cation of intangibles’ issue. As a
consequence, accounting policy organisations (e.g.
the IASC) tend to feel their way around like the
blind, an observation suggested by the evidence
that it took 10 years to release the accounting
standard IAS 38 for reporting intangible assets
(IASC,1998).
However, do we really have a mess of intangi-
bles
1
or do we have nothing to order, i.e. a lack of
a ‘good division’ of the intangible concept? To
divide an accounting concept into new sub-con-
cepts is an issue related to developing the
accounting language; it involves the process of
denoting phenomena in the process of describing
modern aspects of organising and organisations.
(The introduction of the concept ‘provisions’ in
the Swedish bookkeeping act of 1995 is an exam-
ple of further division of the balance sheet.) Clas-
si?cation and (logical) division are actually two
sides of the same coin. Therefore, when discussing
classi?cation in this paper, the arguments are to a
certain extent valid for division as well.
With reference to the introduction earlier, a
demand for a ‘good classi?cation’ of intangibles
from the business and research community seems
evident. The question is how to construct such a
classi?cation. Before building a classi?cation sys-
tem, however, it is necessary to know something
about what it is that entails a ‘good classi?cation’.
The aim of this paper is to formulate a strat-
egy for classi?cation of intangibles. The
approach involves the mapping of properties
of good classi?cation—its anatomy so to
speak—and confronting those properties with
corresponding accounting concepts.
The paper assumes that we can identify the
properties of good classi?cation if we can for-
mulate at least one strategy for accounting classi-
?cation (of intangibles). If we can map the
properties of a good classi?cation, we would have
at least one set of criteria (read ‘‘ideal model’’) for
the evaluation of alternative classi?cations.
1.1. Why classify?
The value of classi?cation is, according to Rud-
ner (1966), associated with its ability to function
as a heuristic device, as a help construction for
interpretation. Is that particular function a rea-
sonable description of a chart of accounts? It may
be presumed that Hegglo¨ v (1998) views account-
ing as a heuristic device when he judges traditional
accounting as a hindrance to learning. One can
suppose that Hegglo¨ v implies hindrance to learn
about management, not accounting. Hedberg and
Jo¨ nsson (1978) were most likely looking for the
heuristic aspects when they introduced the notion
of the semi-confusing accounting system. It is
probably the case that classi?cation can work in
both ways—to hamper or to help.
Nonetheless, the general idea underlying classi-
?cation is to facilitate and hence to promote
understanding. It is obvious that one aspect of
classi?cation is instrumental. Classi?cation is sup-
posed to be used as part of a certain practice. In
Rose and Miller’s (1992) conceptualisation this is
the technological aspect (elements). Classi?cation
as such is purported to facilitate bookkeeping in
1
The concept of ‘intangibles’ is used throughout the article
either with an empirical referent or as a general unde?ned con-
cept. Therefore, its contents will be given by circumstances.
Intangibles are, if nothing else is stated, discussed within a
?nancial accounting context.
696 J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713
that it brings order to a large disarray of transac-
tions. A speci?c classi?cation, however, is also an
expression of the normative aspects or values
embedded in practice, i.e. its programmatic ele-
ments (Rose & Miller, 1992). It is the outcome of
an articulated or unintentional world-view, of
political bargaining, lobbying and individual
interests.
If we can describe the world in simpler terms
(cp.), our chances to understand it increase sig-
ni?cantly. Therefore, according to Rudner (1996),
one aspect of assessing the acceptability of non-
theoretic formulation (its purpose) is based on the
simplicity of its set of primitive terms. He classi?es
(!) the uses of simplicity into six categories (the
tree structure is my own interpretation; Fig. 1).
‘‘If ‘simplicity’ is used to apply to the universe
independently of the perception of it, we shall
classify that usage as objective-ontological; if
‘simplicity’ is taken as a predicate of our (extra-
linguistic) responses to the (extralinguistic) uni-
verse, we shall classify it as subjective-ontological’’
(Ibid., p. 43). Classi?cation is an expression of
some ontological assumptions. Intangibles, as a
social construct or as empirical phenomena, are
based on certain ontological assumptions. The
ambition of this paper, however, is restricted to
the descriptive dimensions of simplicity. Objective–
notational simplicity has to do with the notational
property, such as the number and names of classes
and the number of inscriptions (e.g. of letters). To
illustrate, the accounts in a chart of accounts can
be denoted either by numbers or by names, where
numbers have a higher degree of objective–nota-
tional simplicity.
Objective–logical simplicity denotes the logical
structure of the predicates. To judge the logical
structure there is the need of such extralogical
predicates as re?exivity and symmetry to assess
the degree of simplicity of a scienti?c formulation.
There is a strong link between simplicity and sys-
tematisation. ‘‘When simplicity of basis vanishes
to zero—that is, when no term or principle is
derived fromany of the others—systemalso vanishes
to zero. Systematisation is the same thing as simpli-
?cation of basis’’ (Goodman, in Rudner, 1966, p.
44). When we assign each resource or accounting
transaction a speci?c class, we have not really sim-
pli?ed anything. To measure objective–logical sim-
plicity full formalisation is required. Because full
formalisation about intangibles is unworkable, the
question is whether partial formalisation is enough
to test this dimension of simplicity.
Subjective–notational simplicity denotes, for
example, notations that are recognised as familiar,
convenient and have some aesthetic quality. Sub-
jective–notational simplicity can be interpreted as
the link from a chart of accounts to behaviour
accounting or to the behaviour (psychological)
response to a name of an account. For instance,
how do investors perceive the concept ‘intellectual
capital’ compared with ‘intangibles’? To measure
or judge subjective–notational simplicity, it is
imperative that the appropriate criteria are
selected.
Subjective–logical simplicity denotes ‘‘how peo-
ple psychologically respond to logical properties
of theories‘‘ (Rudner, 1966, p. 43). (Rudner does
not expand on this concept because, like most
philosophers, he examines the contents and impli-
cations of objective–logical simplicity.) In my
understanding of this dimension of simplicity, it is
how the total structure, i.e. the total classi?cation,
is perceived. To exemplify, is it subjective–logical
to introduce a capital concept (intellectual, struc-
tural or human capital) for something on the asset
side of the balance sheet? Alternatively, would it
be more meaningful to introduce intangible costs
in the income statement? Here we enter into the
?elds of cognitive psychology and behavioural
accounting.
The idea of simplicity in accounting can be
traced to other concepts, such as the IASC (1998) Fig. 1. Dimensions of simplicity.
J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713 697
framework about qualitative characteristics, about
understandability and relevance (subjective sim-
plicity), and about reliability and comparability
(objective simplicity).
Introducing several dimensions of simplicity
raises the notion of trade-o?. Hoegh-Krohn and
Knivs?a? (1998) identify the main trade-o? issue of
intangibles as between relevance and reliability.
According to the authors, the present solution is
overly conservative and places too much emphasis
on reliability. And I have to agree. That is to say
that subjective simplicity (in accounting) is given
?rst priority because the classi?cation has to be
useful. Objective simplicity can be viewed in
accounting as a mean or help construction to
reach subjective simplicity.
So why do we classify? One obvious reason is
that we would like to create a certain order in
society, a certain way to direct our interpretation of
the world. Another reason is that it facilitates
understanding of the world through simpli?cation.
Focusing here on the second reason, the issue of
(subjective) simplicity must be vital when testing
di?erent accounting classi?cation constructs.
2. The nature of classi?cation
As soon as we describe something, we assign it
to a class. ‘‘Any descriptive word classi?es’’
(O’Shaughnessy, 1972, p. 23). For example,
assume that a brand name is an intangible asset.
‘‘Such a statement attributes a property to an
object, event, or state’’ (Acko?, 1968, p. 11). The
property here is what is reputed to be understood
as ‘intangible’ and ‘asset’. The statement is there-
fore equivalent to expressing the supposition that
brand name is a member of the set of intangibles.
‘‘That is, every predicational statement classi?es
its subjects’’ (ibid., pp. 11–12).
We use a language when we describe, a language
that is an expression for, and part of, a speci?c
culture. Because it is reasonable to interpret clas-
si?cations as linguistic artefacts, it follows logi-
cally that they are culture bound. The implication
is that technically the same intangibles may not
necessarily be perceived as similar in di?erent cul-
tures. Alternatively, from a cultural perspective,
identical intangibles may not be classi?ed in the
same way because they are assigned di?erent
properties. Take for example, the concept ‘knowl-
edge’. One option is to treat knowledge as some-
thing possible to separate in its own right.
Another is to treat knowledge as embedded in
human capital. A third option is that knowledge is
embedded in di?erent types of organisational
products and processes (like patents). Therefore,
every classi?cation is built upon some ideas about
the world, how it ought to be ordered.
Ruling accounting classi?cations are challenged
because ideas, culture and technology change
continuously. When we experience a technology
shift in a practice, the existing classi?cation is
challenged. This is precisely what occurred to
Linne´ ’s classi?cation of plants when instruments
for establishing the genetic code were developed.
The question is if we now see such a technology
shift in accounting through the IT development
or is it a shift of ideas about organisations and
organising?
What a classi?cation of intangibles does is to
create something based on constructs, i.e. a hyper-
real world (Baudrillard, 1988). The task of con-
structing a classi?cation of intangibles is not pri-
marily a scienti?c one. There is no logic of
discovery or construction (of the world), just of
validation. Discovery and construction is rather a
creative process. Although an intangible is a social
construct and, by de?nition, lacks tangible char-
acteristics, it nevertheless functions as such in the
imagination of the observer.
To classify something also implies de?ning it.
Therefore, a necessary requirement for classi?ca-
tion is a (set of) de?nition(s). Like most concepts,
‘intangibles’ have no generally accepted meaning
or de?nition. This is expected because objects,
states and events can be ascribed di?erent proper-
ties (or attributes). Few, if any, objects are one-
dimensional. The problem of de?nition can be
approached as part of a theory, as in Clement’s
(1997) evolutionary theory or in No¨ rreklit’s
(undated) resource-based theory in which classi?-
cation is based on the control method of the
resources, or in Gold?nger’s (1997) knowledge-
based theory in which di?erent classes of knowl-
edge are constructed. Alternatively, the de?nition
698 J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713
issue can be approached empirically as a categor-
isation of concepts used already in organisations
(see Statistics Netherlands, 1998). Stolowy and
Jeny (1999) have classi?ed the de?nitions of
intangibles stated by 21 countries into three
groups: tautological, by opposition and real. A
tautological de?nition is formulated ‘‘in order to
be included in the balance sheet category entitled
‘Intangible Fixed Assets’ an asset must be intan-
gible.’’ (Ibid. p. 7) A de?nition by opposition is
often expressed in the following way: ‘‘. . . intangi-
bles are ?xed assets other than tangible or ?nan-
cial’’ (Ibid. p. 7). Finally, the third classi?cation,
real de?nition, is an attempt to de?ne intangibles
by other concepts. According to Stolowy and Jeny
(1999), only four of the 21 countries have tried to
formulate a real de?nition.
There are two options available when one
wishes to make the de?nition of intangibles more
information rich. One is to use a deductive
approach. The other is empirical, i.e. to create a
list of intangibles based on empirical observations.
Although both approaches seem valid during the
present stage of experimentation, we need a de?-
nition, either for classi?cation or for division of a
concept into sub-classes.
In logic, the process of breaking down a class
into sub-classes is referred to as ‘logical division’.
To divide we need (to construct) an attribute (for
intangibles) that can be divided into di?erent
classes. This is a normative, deductive approach.
‘‘Classi?cation is the complementary but reverse
process to logical division. The di?erence lies in
viewpoint and the starting-point in procedure’’
(O’Shaughnessy 1972, p. 24). Classi?cation brings
objects (speci?c intangibles) together based on
certain attributes. Classi?cation also means to
reduce the number of relevant attributes of an
object. In that sense, reductionism becomes a pre-
requisite for holism. Only the attribute(s) that is
the basis for classi?cation counts. For instance,
assume that ‘knowledge’ is one class of intangibles
consisting of the sub-classes ‘human knowledge’
and ‘organisational knowledge’. In this example
the only two attributes we are looking for is
human and organisational. All other possible
attributes of knowledge are of no classi?catory
interest.
To summarise, to classify/divide something
means to describe, which implies to explain
the nature or essential qualities of something.
To classify or divide we need to construct
properties/attributes that describe an object,
event or state.
Science gives us continuously new opportunities
to describe ‘essential qualities’, and politics is set-
ting the priorities of accounting change. However,
attribute construction can be approached more
fruitfully if we ?rst formulate some ideas about what
(ideal) types of classi?cation there are and if there is
something to classify (i.e. a universe). These two
issues are addressed in the following two sections.
3. In search of a classi?cation schemata or
typology
We are in search of a classi?cation, as a formal
construct, as a non-theoretic formulation but that
is perhaps linked to theory. ‘‘The two types of
classi?cational schemata are (1) those in which a
condition that is both necessary and su?cient for
the applicability of each of the classifying terms is
given, and (2) those which such conditions are
given for none or only a few of the classi?cational
terms.’’ (Rudner, 1966, p. 32)
In accounting, we normally operate with the
second kind of classi?cational schemata (classi?-
cational analytical schemata). This is because
either they fail to classify the universe of discourse
exclusively and exhaustively, or the classi?cation is
not done by logical means but only by some con-
tingent feature of the world.
The balance sheet is an example of classi?ca-
tional analytical schemata. However, it can be
described as a speci?c form of classi?cational
schemata, i.e. as a typological system (compare
nominal versus ordinal scales). According to
Rudner (ibid., pp. 37–38), a typological system
must contain at least the following components:
1. ‘‘A concept determining the typology’s uni-
verse of discourse’’. As an example, let us
choose assets as the universe of discourse.
J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713 699
2. ‘‘Some relation(s) that determine(s) an
ordering (e.g., a serial or quasi-serial order-
ing) among the members of the universe of
discourse’’. Assets can be arranged either in
liquidity order or in purpose-of-use order.
3. ‘‘Statements implying that certain features
(e.g. transitivity) characterise the relation
(s)’’. Regarding transitivity, if accounts
receivables are less liquid than cash, then
cash is more liquid than accounts recei-
vables.
4. ‘‘A set of concepts (frequently a ‘polar-pair’
or ‘extreme opposites’) usually designating
some speci?c members of the universe of
discourse that are ‘distant’ from each other
or at opposite ends of the array into which
the ordering relation(s) order the members of
the universe of discourse’’. Cash deposits may
be viewed as one natural endpoint of liquidity
order: the question remains, however, to
determine the other endpoint? Even if the asset
side is arranged according to the purpose-of-
use order or when the assets are supposed to
be used, a problem still remains. A certain
amount of cash is needed in eternity while
land can be held for just a short period.
It follows from the earlier description of a
typology that a typological system contains more
information than classi?cational (analytical) sche-
mata in that it relates the di?erent classes to each
other. Assume that we have identi?ed three classes
of intangible asset: trademark, R&D and
employee knowledge. These three classes con-
stitute a classi?cational schema of intangibles.
However, if we can add something (through logic)
that relates the three classes to each other addi-
tional to their sub-class relation to intangibles,
then we will have a typology.
We can now add to the classi?cation construc-
tion agenda: Should the ambition be to construct
classi?cational analytical schemata or typologies?
The answer is more than logical ambition. The
former is a non-theoretic formulation that simply
groups things together. The latter still keeps the
link between theory and non-theory open, i.e.
whether some testable propositions should be
included or not.
4. Is there a universe of discourse to classify?
It was stated earlier that there is a lack of real
de?nitions of intangibles. How do we then ?nd the
intangibles to classify (or divide), i.e. a universe of
discourse? As indicated previously, we can either
identify intangibles as an empirical endeavour or
construct (divide) it as a theoretical project.
Whichever, we have two options as illustrated
with a balance sheet example. One is that the
classi?cation can start, for instance, with a classi-
?cation of economic resources and treat assets and
non-assets as a subset of economic resources.
Here, the classi?catory concept, which assigns
properties to the di?erent members of the uni-
verse, is merely used to divide accounting from
non-accounting resources. The other is that the
classi?cation starts directly with those elements
that can be assigned accounting properties and
thereafter begins a search for classifying con-
cepts. While a classi?cation of resources would
probably result in di?erent accounting classi?-
cations, it would also take into consideration
the dynamic change or transformation of prop-
erties for a given member of the universe. For
example, if trademarks become marketable in
general and, therefore, accountable, the object is
already in the classi?cation schemata as a
resource. However, trademarks will change sub-
class from non-asset to asset. The universe of dis-
course is not stable and new objects, states or
events can give rise to new classes. Accounting is,
as Hopwood (1987) suggests, under the process of
becoming.
A paradox exists in that each classi?cation or
division has to start with a (meta) division. What
Cooper (1989) and Catasus (1999) call ‘labour of
division’ emphasises the process of dividing rather
than the outcome of the process. A logical con-
sequence of their reasoning is that to get any
object, state or event classi?ed we must divide or
separate them from the rest of the world. In this
respect, it is evident that labour of division is not
only a process of seeing but also one of simpli?-
cation in which a reduction in context occurs. The
issue then is whether it is possible or even desirable
to divide intangibles and tangibles into separate
categories or classes.
700 J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713
IASC (1996) makes an attempt to separate an
accounting universe of discourse (from the rest of
the business world). ‘‘Financial statements portray
the ?nancial e?ects of transactions and other
events by grouping them into broad classes
according to their economic characteristics’’ (p.
54). However, this description or de?nition is too
vague to be operationally useful. Furthermore, the
term ‘other events’ is so broad that it can imply
almost anything.
In traditional accounting textbooks, the number
of mentioned intangibles is limited. Hendriksen
and van Breda (1992) list 13 of what they call tra-
ditional intangibles, and add to the list 13 deferred
charges related to expenditures on services. Fur-
thermore, if we search the literature for potential
intangibles relative to the balanced scorecard,
intellectual capital, etc. the list can be extended
inde?nitely. Even if only four countries have some
sort of real de?nition of intangibles, all of the 21
countries except one had a list of intangibles
(Stolowy & Jeny, 1999).
The conclusion is that there is a constructed uni-
verse of discourse of intangibles. The question then
is whether it is possible to classify or divide the
universe in some way, i.e. to simplify. In both
cases, we need a selection of attributes that helps
us to di?erentiate the objects, states, or events in
our universe of discourse. The selection of attri-
butes is contingent on the purpose of classi?ca-
tion, i.e. there are no natural selection of
attributes. Given the attributes, the construction
of a classi?cation schema is a matter of logic.
However, the real classi?cation of objects, states
or events remains a matter of interpretation.
5. Di?erent approaches to attribute identi?cation
As previously noted, one of the main issues in
classi?cation is the choice of attributes that dis-
criminate between the to-be-classi?ed objects,
states, or events. Roberts (1995) divides the search
for attributes into ?ve approaches in his classi?-
cation of ?nancial accounting systems in di?erent
countries: essentialist approaches, overall similar-
ity approaches, diachronic approaches, set theory
approaches and archetypal approaches.
The essentialist approaches, based on Aristotle’s
works, seek to identify the essence of an object
(but not in de?ning it). Essence, in Aristotle’s
world, was found through the assumed functions
of an object. The object’s attributes can be
deduced from its most essential properties.
Hempel (1952) explicates a distinction between
natural and arti?cial classi?cation. The former is
an essentialist approach, assuming that the objects
possess fundamental similarities. Super?cial
resemblance or external criteria determine the lat-
ter. A classi?cation of intangibles would be, in
Hempel’s system, an arti?cial classi?cation. How-
ever, the subject matter at issue is whether all
classi?cations are arti?cial or whether only the
selected attributes are arti?cial.
The important question here is if it would be
possible to identify and discriminate between
functions of intangibles. Can an intangible—a
social construct—have a function per se? One
possibility would be to link di?erent intangibles to
a production function or value creation process.
We would next have to formulate the production
function or value creation process before we
search for the functions of di?erent intangibles
because their functions would be identi?ed in
relation to those processes. To use a negative for-
mulation, if trademarks have no function in the
value creation process then they cannot be part of
the classi?cation.
The overall similarity approaches use all (read
many) attributes to look for similarities between
objects. A statistical procedure, such as factor
analysis, is used to search for similarities whereby
similar objects form a class. The problem, how-
ever, is the choice of possible attributes: ‘‘If we
take any set of objects there is an in?nity of attri-
butes that these objects may have or that we might
choose to give them. . .’’ (Roberts, 1995, p. 648).
The overall similarity approaches are appro-
priate and more fruitful when the set of attributes
is given and the problem is to di?erentiate and
reduce the universe of discourse to good classi?-
cation. I doubt that this could be the case for
intangibles. Moreover, an evident risk exists of
encountering complicated classi?cations that
would violate one basic idea, namely, the one of
simpli?cation.
J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713 701
The diachronic approaches introduce time rela-
tionships as attributes: ‘‘. . .the essence of an object
is no longer what it is in itself but rather how it
came to be what it is’’ (Roberts, 1995, p. 649). As
Roberts’ comment suggests, these approaches are
closely related to the essentialist approaches.
To use time relationships to classify intangibles
it must be plausible to divide the production
function or value creation processes into separate
and sequential time stages. Ahonen (1998) does
just that when dividing exploitable and generic
intangibles, where the latter are not yet exploi-
table. This dichotomy can also be interpreted as
an essentialist approach. Furthermore, the
balanced scorecard (Kaplan & Norton, 1996)
assumes an intricate relationship between, past,
present and future of the di?erent performance
drivers and ?nancial outcomes.
A diachronic approach places greater emphasis
on the origin of the intangibles while the essentialist
approaches primarily focus on the destiny of the
intangibles (their function). The gist of the problem,
then, is choosing between origin and destiny.
The set theory approaches try to avoid the pro-
blem of attribute choice by introducing some
properties of a classi?cation that are thought to
express its property of ‘goodness’. Roberts (1995)
examines four properties of classi?cation used by
the American Accounting Association: con-
sistency, exhaustibility, mutual exclusivity and
hierarchical integrity (Section 7).
It is unclear why Roberts (1995), in his debate
with Nobes (1984), (‘‘International Classi?cation
of Financial Reporting’’) classi?es ‘set theory’ as
an approach for attribute search. First, it is, per
de?nition, a non-approach or, at best, an indirect
approach. One can easily imagine other non-
approaches. Second, it is unthinkable to deduce
attributes from the properties of a ‘good classi?-
cation’. Rather, the properties are used in a testing
phase that generates feedback for reclassi?cation
(introducing new attributes), again to be tested in
relation to the classi?cation properties. Therefore,
set theory approaches are more a part of the vali-
dation processes of classi?cation schemata than
acting as possible attributes themselves.
Finally, Roberts (1995) disseminates the arche-
typal approaches. According to Roberts, ‘‘. . . the
principle which should guide attribute selection is
not similarity or di?erences between those attri-
butes per se, but similarities or di?erences in the
relationship of those attributes’’ (Roberts, 1995, p.
659).
The archetype or ideal type approaches are par-
ticularly appropriate when dealing with the classi-
?cation of systems that contain many elements
and relations. Nonetheless, it is still necessary to
de?ne the elements in the system before we can
obtain an expression (attributes) of the relations
between the elements and how they are related to
the system as a whole. Whereas earlier we dis-
cussed origin and destiny, the focus now is on
transformation processes. Perhaps the best exam-
ple of this approach is the balanced scorecard of
Kaplan and Norton (1996). The idea of these
authors is to link each performance driver, more
often than not an intangible, into a cause–e?ect
relationship ending in a ?nancial outcome. Its
relation, therefore, gives the function of each per-
formance driver.
When identifying attributes from processes, two
issues arise that are valid for all approaches con-
sidered within classi?cation theory in conjunction
with ‘exclusivity’, ‘set theory’ and ‘fuzzy sets’. The
issue is the di?cult task of how to draw (de?ne)
the border between classes.
Philosophers use the word vagueness (not to be
confused with ambiguous concepts) to denote
problems associated with border cases. When does
a number of grains of sand become a pile of sand?
How many inhabitants in a big city must be taken
away before it becomes a small city? Is it possible
to draw a de?nite border between the material and
immaterial, as well as between the di?erent intan-
gible classes? In mathematical terms, we have a
problem of selecting a cut-o? point on a con-
tinuous function for a given attribute. To reduce
this problem we have to look carefully into the
de?nitions of the attributes. Nonetheless, the
inherent vagueness of language will always be a
reason for di?erent interpretations of the same
phenomenon and, as a likely consequence, di?er-
ent classi?cations.
The second issue concerns whether objects (their
attributes) can take di?erent forms (compare with
solid, ?uid and gas forms of the same matter). The
702 J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713
problem has to do with attributes that change over
time. The most obvious example is an option that
is supposed to be turned into something else (or
nothing). Another problem is that di?erent forms
may exist at the same point in time. Several forms
of human resource and their artefacts can be
identi?ed, a situation that results in classi?cation
and double-counting problems.
We get little advice from Roberts (1995) on an
appropriate course of action in selecting attributes
and matters concerning the construction issue. ‘‘It
has been argued that classi?cation is a way of
viewing the world, a Weltanschauung with the
‘truth’ of classi?cations being relativised. From
the recognition of this idea, it has been suggested
that classi?cation should be created in line with
their purposes’’ (Roberts, 1995, p. 661). First,
‘truth’ is not an issue in classi?cation theory.
Classi?cation schemata have none empirically tes-
table proposition while a typology can include
testable ones. Second, to state that something
ought to be constructed with a ‘purpose’ has very
little information content as long as a speci?c
purpose is not stated.
The objective is to ?nd a stable classi?catory
term uncontaminated by contingent factors.
Observe that the classi?catory term itself can be
challenged by contingent factors. The dichotomy
short-term (less than a year) and long-term (more
than a year) is such a stable classi?catory term
(time horizon). We do not need to reclassify when
new objects are discovered or constructed. Reality
can prove time horizon to be an obsolete attribute,
however. An economic resource is not a stable
term in that its changing attributes are the cause of
reclassi?cation. Thus, we have to look for other
concepts or to incorporate the dynamics of
resources as an attribute.
Searching for classi?catory concepts involves a
mixture of purpose of classi?cation and creative
work. Embedded in the technocrats order creation
is the programmatic order creation. Which values
or norms are embedded in a speci?c classi?cation
schema? Is turning the balance sheet upside down
a technical matter or an expression of di?erent
priorities (e.g. land investments more important
than cash)? The key argument concerns whether
classi?cation schemata of intangibles ought to be a
part of social theory or are they based only on or
related to social theory? Immaterial resources,
intangibles, or intellectual capital are not to be
found as real objects. What constitutes a resource
is both a political and an economic matter.
So far, we have identi?ed three strategies in our
search for classi?cation attributes.
1. To use the present attributes in (?nancial)
accounting: for the balance sheet, ‘time-hor-
izon’ and ‘degree of liquidity’ would still be
the two chief attributes.
2. To reveal or construct attributes in relation to
identi?ed intangibles. The task would be to
list all intangibles used by organisations.
3. To develop new combinations of attributes
through, for example, morphologic meth-
od
2
that may or may not be related to a theory
of the ?rm. This strategy is more a matter of
division than classi?cation.
6. How to test/validate classi?cation schemata
What consequence does testing of classi?cation
schemata have on construction? If we aim for a
classi?cational schema, which is a non-theoretic
formulation, a speci?c class cannot be subjected to
empirical tests (Rudner, 1966). It makes little
sense to test, for example, if R&D ought to be a
speci?c class in a classi?cational schema. We have
to know the other classes that divide the universe
of discourse. What we can do is to test the whole
classi?cation schema, its logical properties and if it
is used (adopted).
When the dialogue of classi?cation theorists is
about ‘good classi?cation’, they are referring in
part to its logical properties. Rudner (1966) limits
the number of properties to two (except simpli-
city). One is the ability of classi?cation schemata
to be exhaustive, i.e. all objects of the universe of
discourse can be classi?ed. The other property is
that classi?cation schemata are exclusive, i.e. no
object can belong to more than one class.
2
Morphologic method involves combining di?erent known
attributes into new combinations.
J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713 703
Roberts (1995) develops his discussion around
the four properties of a ‘good classi?cation’ as
formulated by the American Accounting Associa-
tion (AAA, 1977). Exhaustive and exclusive are
two of the properties. The third is consistency. In
concurrence with Roberts (1995, p. 653) the need
for this third property is questionable. Consistency
states simply that, as soon as one attribute is cho-
sen, it should be used throughout the classi?cation
hierarchy to di?erentiate between the objects (a
classi?cation does not necessarily need to be a
hierarchy). If we use the attribute ‘marketable’ for
intangibles, then a second level must use di?erent
aspects of the marketable concept for further
classi?cation.
The fourth property introduced by the AAA is
hierarchical integrity. This property deals with
setting labels for the di?erent classes and how
those labels can be interpreted. Another aspect is
the relation between the di?erent classes (such as
specie, class and family). From my understanding
of hierarchical integrity, it is linguistic in nature
rather than logical. In Rudner’s language, it is an
issue of subjective–notational simplicity and
therefore it is treated here as part of simplicity.
Another suggestion concerning the criteria for
‘good classi?cation’ is Lazerfeld and Barton’s (in
Lerner & Lasswell, 1951). They present four cri-
teria for good classi?cation: (1) articulation, from
the general to the speci?c; (2) logical correctness,
exhaustive and mutually exclusive; (3) adaptation
to situations; and (4) adaptation to the respon-
dent’s frame of relevance. These criteria, however,
do not add anything to good classi?cation, in that
logical correctness has already been dealt with and
the other three criteria are all di?erent aspects of
simplicity.
Finally, when dealing with accounting informa-
tion, we may have to introduce information eco-
nomics as an extra property. Even if a new
classi?cation of intangibles can be helpful in pro-
ducing economic decisions that improve the capital
market e?ciency (to zero information costs), the
cost of classi?cation change and data retrieval can
be higher than the newly added value. Here the
issue involves choosing between a number of pos-
sible testing implications rather than constructing
new ones (as attitudes, behaviour, use and result).
Thus far, we have largely restricted our discus-
sion to testing the technocratic aspects of classi?-
cation schemata. Would it be possible to test or
reveal their intrinsic programmatic aspects as well?
Such a test should be based on interpretation of
the consequences using the classi?cation schema.
Accordingly, it is possible to reverse the approach,
i.e. begin to think about the consequences or
e?ects we would like to achieve through the pro-
cess of classi?cation.
The other aspect of validation concerns the
choice of an empirical test if the classi?cation
schema is adopted. However, we ?nd ourselves in
a ‘catch 22’ situation in the sense that we ?rst have
to construct a classi?cation before we can study if
it becomes adopted.
To summarise, there are two things to be learnt
about construction from the validation issue.
Construction can start from the idea of a ‘good
classi?cation’, that it should be simple and that it
should classify the universe of discourse exhaus-
tively and exclusively. However, construction can
also begin from an idea of wishful e?ects of the
classi?cation.
7. Classi?cation theory concepts meet accounting
In the previous sections, an attempt was made
to formulate a classi?cation strategy for intangi-
bles based on classi?cation theory. In this section,
a tentative confrontation between classi?cation
theory concepts and ?nancial accounting is
attempted. The three requirements placed on clas-
si?cation so far are that it has to be simple,
exhaustive and exclusive.
7.1. Simplicity and usefulness
With reference to Rudner (1966), the purpose of
classi?cation schemata is to simplify, in di?erent
dimensions, the universe of discourse. This is its
technocratic purpose. The programmatic purpose
may be expressed in terms of perceiving classi?ca-
tion schemata as reality construction, as Wel-
tanschauung, as a set of ontological–subjective
assumptions. One such assumption is to look for
utility (usefulness) rather than purpose. Utility
704 J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713
seems to bring the subject closer to context, and
the objective of this project is to examine use and
context, whatever the purpose. What is needed is
to formulate a set of normative statements about
utility (ex ante) as a device for construction. Then,
as part of testing, we can try to grasp the empirical
meaning of utility.
Usefulness, the potential to be used, is an
essential assumption about accounting data and
accounts. One way to judge the usefulness of
?nancial accounting data is to measure their
value relevance in explaining stock prices and
returns. That association has decreased between
1980 and 1996 (Lev & Zarowin, 1998), indicat-
ing the need for reclassi?cation or a change of the
accounting borders. The equivalent non-usefulness
in management control is the emergence of new
measurement packages under di?erent labels.
A problem with these value relevance studies is
that they test direct relationships between ?nancial
accounting data and stock prices. The annual
statement is mainly a historical document that
articulates about the ?nancial consequences of
past transactions and events. Therefore, its value
relevance can be sought in its stewardship func-
tion, or, in more future-oriented terms, as a back-
ground or lens for interpretation of ongoing
events in a turbulent economy (Fig. 2).
With such a view on value relevance of ?nancial
accounting data, the problem centres on how to
design an appropriate background.
3
A background
may have no value (relevance) of its own. A help
construction gives reference points. In this sense,
new designed backgrounds (read classi?cations)
can function as a heuristic device. The interpreta-
tion of a given event may change because of the
new background. (Compare the problem in colour
classi?cation when placing two colours side by
side. The perception of one colour is at the bor-
derline dependent also on the other. This phe-
nomenon is named colour induction or
simultaneous contrast).
4
The design of a new background can be done
either by introducing new accounting concepts,
such as intellectual accounting, or by rearranging
(classifying) the already recorded transactions
under new headings. A change in the names of the
accounts, the headlines, or the metaphors is likely
to have far reaching implications in the way we
perceive organisations.
5
Another aspect of value relevance concerns
studies that seek to establish a link between quali-
tative variables and business success of which the
balanced scorecard is an expression. Attempts to
integrate business success factors into the
accounting model, however, would probably be
futile. The determining factors underlying success
seem to be ?rm- or situation-speci?c, a conclusion
based on more than 400 empirical investigations
(Capon, Farley, & Hoenig, 1996).
7.2. Exclusive and consistent
As part of usefulness, much of ?nancial
accounting regulation is based on the objective to
achieve comparability in space and time. Con-
sistency, then, is assumed to be one prerequisite
for comparability. As IASC (1996, IAS1, p.76)
puts it: ‘‘3. Going concern, consistency and
accrual are fundamental accounting assump-
tions.’’ Further, IASC de?nes consistency in the
following way: ‘‘It is assumed that accounting
policies are consistent from one period to
another’’ (Ibid., p. 77). Hendriksen and van Breda
(1992, p. 142) are more precise and let con-
Fig. 2. Financial accounting data as background.
4
See Hansson and Karlsson (1974).
5
A most convincing story of reality construction has been
told by Hines (1988) and introduced into an organisational
theory context by Czarniawska-Joerges (1993).
3
In ?nancial accounting the annual statement [more pre-
cisely the balance sheet (book value) and pro?t and loss state-
ment (earnings) is supposed to constitute the background].
J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713 705
sistency refer to ‘‘. . . the use of the same accounting
procedures by a single ?rm or accounting entity
from period to period, the use of similar measure-
ment concepts and procedures for related items
within the statements of a ?rm for a single period,
. . .’’ The words ‘same’ and ‘similar’ in this context
presume an underlying classi?cation.
Changing accounting procedures implies that we
either change the properties of, for example, a
given object or change the class belonging of a
given object. In either case, comparability will be
distorted. From a classi?cation viewpoint, we are
dealing with two types of consistency, i.e. in
recognising and classifying. To be consistent yields
no information about what is ‘right’ or ‘wrong’.
Performing the same mistake repeatedly is a
demonstration of consistency. However, it is impos-
sible to be consistent if the classes are non-exclusive.
For instance, if a particular labour expense can be
treated either as part of trademark or as R&D, we
will treat that expense consistently just by chance.
Therefore, a necessary but not su?cient condition
for consistency is exclusivity of classes.
When di?erent ?rms are using the same proce-
dures, we normally denote that as uniformity
rather than as a re?ection of consistency. ‘Rigid
uniformity’ means ‘‘. . .prescribing one method for
generally similar transactions even though rele-
vant circumstances may be present’’ (Wolk &
Tearney, 1997). Then, ‘?nite uniformity’ implies
consideration of relevant circumstances, i.e. the
use of di?erent methods for similar transactions.
Most accounting theorists advocate ?nite uni-
formity because representatively it is thought of as
being more faithful. This argument is likely to
represent a form of circular reasoning. If some-
thing is relevant (i.e. a property of an object) and
is omitted, it becomes less faithful representa-
tively. Alternatively, they are dissimilar because
there are relevant circumstances that make them
dissimilar. Given the properties or attributes, a
‘good classi?cation’ is based on the idea of rigid
uniformity. What we do when we introduce some-
thing like ‘relevant circumstances’ is to create a new
attribute. The classi?cational schema was not
meant to handle that attribute. There are no rele-
vant circumstances and, if there were, that would
indicate a need for a redesigned classi?cation.
Finally, from the viewpoint of classi?cation
theory, if a classi?cation has exclusive classes, the
issue of separability is solved. IAS 38 (IASC 1998,
p. 16) is ambivalent about separability. In point 11
the following argument is expressed: ‘‘An intangi-
ble asset can be clearly distinguished from good-
will if the asset is separable’’. In point 12 the
document notes the following: ‘‘Separability is not
a necessary condition for identi?ability since an
enterprise may be able to identify an asset in some
other way’’. To separate in ‘some other way’ must
denote some sort of separation. The chart of
accounts (logical division) is based on the idea of
separation and such separation is necessary in
order to report intangible expenses or assets.
7.3. Exhaustive and residual classes
A consequence of division in accounting is that it
often leads to a residual class. One well-known resi-
dual class is goodwill. This is expressed by some as
the most intangible of all intangibles. Its relative
growing importance with respect to total assets
(Gro¨ jer, 1997) can be one explanation of the dimin-
ishing value relevance of ?nancial accounting data.
Goodwill presents the opportunity for accounting
nonsense (i.e. the ?gures add up but do not make
sense) instead of reporting vital resources or assets.
A residual class (non-class) creates a problem
because its relation to the other classes becomes
unclear. There is no room for such classes in a
‘good classi?cation. Therefore, residual classes,
such as goodwill as a result of purchase account-
ing, are to be avoided’.
The confrontation between classi?cation and
corresponding accounting theory concepts has
resulted in several requirements concerning the
construction of the classi?cation. In the next sec-
tion a confrontation between classi?cation con-
cepts and three used classi?cations is presented.
8. IAS 38, BSC and IC from a classi?cation
viewpoint
Although critical examination of IAS 38, BSC
and IC from a classi?cation viewpoint can be
made in its own right, it can also provide mean-
706 J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713
ingful ideas for future classi?cation attempts.
6
IAS
38 is important because it regulates what is pre-
sumed to be intangibles. The BSC is bene?cial
because it is widely used and makes a connection
between today’s intangibles and tangibles within a
cause–e?ect chain (a typology). Finally, IC is
relevant because it is an example of a division of
tangible concepts into intangibles.
Let us start by confronting the three models
with principal classi?cation concepts discussed in
previous sections serving as an overview of the
analyses to be followed (Table 1).
The general conclusion is that we are dealing with
three very distinguishable classi?cations, all of
which are wrought with inadequacies. The rest of
the section will discuss this conclusion in greater
detail through interpreting the three models in
relation to classi?cation concepts.
8.1. Universe of discourse
In IAS 38 a universe of intangibles is created
through subtracting the tangible universe (of
assets) from the total universe (of assets). Then, a
part of the intangible universe of discourse is
reduced through taking away ‘‘. . . goodwill,
brands, mastheads, publishing titles, customer
lists and items similar in substance’’ (IASC, 1998,
p. 3). In e?ect, this means that anything without
physical substance can be an intangible, except
for a few mentioned items and everything that
is considered similar to those items. As a con-
sequence, the universe of discourse becomes a
vague concept. IAS 38 creates a nebulous universe
by using a concrete listing of intangibles and by
stating general requirements for something to be
an intangible. From that universe, some given
intangibles are subtracted. De?ning a universe
by listing can be the only remaining solution
when the de?nitional problems become proble-
matic.
In the BSC model the distinction between tan-
gibles and intangibles is of no great importance.
Rather, the distinction is made between perfor-
mance drivers and their outcome. The total uni-
verse consists of four sub-universes. The ?nancial
sub-universe can be seen as a background towards
which the three other perspectives can be inter-
preted. The origin of the three perspectives can be
traced to a traditional demand (customer perspec-
tive) and supply model (internal business processes
and learning and growth perspective). The main
problem with the BSC model from a classi?cation
perspective is that the de?nition of each sub-universe
is poorly stated. Consequently, the boundaries
between these sub-universes are also ill-de?ned,
Table 1
Confronting IAS 38, BSC and IC with classi?cation concepts
Classi?cation concepts IAS 38 BSC IC
General concepts
Universe of discourse As a between two universes A set of four universes As a di?erence between two values
Attribute Essentialist plus by listing Archetypal plus diachronic Overall similarity
Type of classi?cation Classi?cation schema Typology Classi?cation schema
‘Good classi?cation’
Exhaustive No (Yes) No No
Exclusive No (Yes) No No
Simplicity
Objective–notational Unclear No Yes
Subjective–notational Unclear Yes No
Objective–logical No No No
Subjective–logical No Yes (No) Yes (No)
6
BSC (Kaplan & Norton, 1996) and IC (Edvinsson & Malone, 1997) are treated in the present paper as fairly well de?ned models.
For an extremely condensed description of the two models, as well as IAS 38, see Appendix A.
J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713 707
especially between internal business processes and
learning and growth. However, the idea is, as I
understand it, to use the four perspectives as a
help construction for developing cause–e?ect rela-
tionships between performance drivers and out-
come. This cause–e?ect chain is the constructed
typology of the theoretical claim by Kaplan and
Norton (1996) that a properly constructed BSC
articulates the theory of the business.
In IC, a universe (IC) is constructed by taking
the di?erence between two values, i.e. the di?er-
ence between the market and book value of the
company. From this single value a whole universe
is created. (To use a metaphor, it is like measuring
the volume of a big pear and deducting the volume
of a small apple. Because there is a di?erence, it
has to be baptised, i.e. it becomes a plum.) Fur-
ther, the amount of IC changes as soon as share
prices change, or when accounting practice chan-
ges. Moreover, as soon as an intangible is capita-
lised on the balance sheet, the amount of IC drops
(as a residual) and the universe grows gradually
less. The conclusion must be that the IC universe
can only be justi?ed in the name of rhetoric.
8.2. Attribute
In IAS 38 two attributes can be traced. The ?rst
is ‘recognition’. This concerns whether the intan-
gible can be recognised or not. The attribute can
be interpreted as an exponent of accounting ‘con-
servatism’. The other attribute is ‘origin’ (essenti-
alist)—from where the intangible originates,
internally generated and externally acquired. One
can reasonably question whether that division is
fruitful. Today, several business functions have
been outsourced or are co-produced. What 1 day
is internal becomes external the next, and vice
versa. In the BSC there is a speci?c ‘time’ (dia-
chronic) division: past, present and future. There
is also an archetypal attribute to separate di?erent
objects, states and events into the three non-
?nancial classes. The attribute in each sub-class,
however, is, and has to be, left open because the
aim of the model is that it is suitable for all types
of business logic. The attribute for division of IC
into sub-classes is also based on some hidden
properties of ‘overall similarities’ related to a
‘value to the business’ idea. The selection of attri-
butes seems to be as crucial as it is complicated and
will be dealt with in more detail later in this paper.
8.3. Type of classi?cation
IAS 38 is a typical classi?cation schema with no
relation(s) between the classes, except from being
sub-classes. In the BSC there is an assumed time
relation between the classes. In that sense BSC is a
typology. But the linkage, in the BSC, of the per-
formance drivers in a cause–e?ect chain is an even
stronger claim for typology. At ?rst glance, IC
gives the appearance of a typology because of the
neatly related classes. Nonetheless, more careful
inspection indicates straightforward division in
which one class (sub-universe) is subtracted from
the other.
To construct a new classi?cation it seems plau-
sible to start with the more limited ambition of
developing a simple classi?cation schema as
opposed to a typology.
8.4. Exhaustive and exclusive
On the one hand, the classi?cation in IAS 38 is
exhaustive because of the general de?nition of
objects, states or events in the universe of dis-
course (future economic bene?ts). On the other,
the exclusion of certain intangibles, most of them
not speci?ed, makes it hard to see the classi?cation
as exhaustive. The BSC does not pretend to be
exhaustive. One of its basic tenets is to make a
selection of objects, states and events from the
universe. The IC cannot be fully exhaustive, which
Edvinsson and Malone (1997) also recognise.
When using a subtracting technique, something
must be left over if the order of subtracting should
matter (otherwise, the last item must be exactly
equal to the residual).
None of the three models classify an object,
state or event exclusively. The vagueness of the
language (de?nition) used in IAS 38 results in
ambiguous class borders. In the BSC the border-
lines between especially internal business processes
and innovation/improvement are unclear and per-
haps such ambiguity is necessary in the modern
organisation. The same problem arises in IC, par-
708 J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713
ticularly the borderline between human capital
and customer and structure capital. Structure
capital, as well as customer capital, must be some
sort of human (capital) artefact.
In comparison with exhaustive there are prob-
ably more problems connected with making a
classi?cation exclusive. However, if the ambition
is to create a stable background, an exclusive
classi?cation is extremely important.
8.5. Simplicity
Starting with notational simplicity, IAS 38
avoids the problem (just two classes) and therefore
is described as both objectively and subjectively
7
unclear. There is a lack of objective–notational
simplicity in BSC because of multi-dimensional
concepts that can be given several meanings.
Nevertheless, the subjective–notational simplicity
is higher in that the concepts are used in the
everyday business life. For IC, the relation is the
direct opposition. The objective–notational sim-
plicity is high because it tries to establish linked
concepts but the link goes through a ‘capital’
concept. To use the word capital for di?erent
assets interferes with the traditional theoretical
and practical use of the capital concept (as the
value of the resources lent to a company).
The IAS 38 ful?ls neither objective nor sub-
jective–logical simplicity. Some intangibles have
been excluded just by listing them (and some
unspeci?ed). Furthermore, it is not obvious why
certain expenses are to be classi?ed as costs,
deferred costs or capitalised costs.
As long as there is no overt success model or
production function, it is hard to see any objec-
tive–logical simplicity in the BSC except that of
time structure (past, present and future). However,
the association between these time frames is
unknown. There is an element of subjective–logi-
cal simplicity because the structure relates, directly
or indirectly, to the traditional accounting state-
ments and the issue of investment versus con-
sumption (production).
As stated when the universe of discourse was
analysed, there is no objective logic in the IC
model or, if there is objective logic, it is not
understandable. To append to the previous argu-
ment, the subtracting algorithm does not work in
the reverse. It is not possible to add the di?erent
components to what is called a capital concept.
Still, according to the business interest, there
seems to be a subjective–logical simplicity in IC.
Perhaps it is because the algorithm orders a group
of haphazardly arranged intangibles that give rise
to a feeling of ‘being in control’. Thus, while it
adds up, it does not make sense.
It must be remembered that all three models are
used in practice. Naturally, IAS 38 has to be used
for companies following the IASC standards. Yet,
a large number of companies use the BSC volun-
tarily. To my knowledge, only a few companies
have tried to adopt the IC model of Edvinsson and
Malone (1997). Mouritsen (1999) interprets the
use of intellectual capital
8
in Danish ?rms in terms
of mobilisation of company (intangible) resources.
Roberts (1999) discusses the IC concept either as
an object (set of tools) or as rhetoric. He argues
that there exists a mutual reinforcement between
the two. One problem, however, is that when
rhetoric is transformed to tools and practise, its
mobilising capacity is hindered and an expectation
gap is often created. If IC is just rhetoric for
(intangible) resource mobilisation, there is prob-
ably no business need for its classi?cation.
Johanson, Martensson, and Skoog (1999) found
that Swedish ?rms use IC as a set of measurement
tools in the process of constructing their own
business model or production function. It is prob-
able that the intense interest of Swedish ?rms in
the BSC model has the same roots. According to
Johanson et al. (1998), however, the initial enthu-
siasm has been considerably greater than the
practical development and use of BSC (and IC).
One explanation can be found in the shortcomings
of the models as classi?cations. Once the organi-
sations start with applications and practical use,
problems like overlapping classes becomes
obvious.
7
The subjective dimension of simplicity, i.e. perceived sim-
plicity, is based here on the author’s subjective interpretation if
nothing else is stated.
8
In this paper the concept ‘intellectual capital’ denotes dif-
ferent types of model for accounting and reporting.
J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713 709
9. Conclusions: towards a classi?cation of intan-
gibles
A reasonable assumption is that new forms of
intangibles will be constructed in the future, i.e.
the universe of discourse is under constant change.
Therefore, a classi?cation of all possible intangi-
bles does not seem to be a task within the limits of
realisation.
The value of classi?cation is, according to Rud-
ner (1966), associated with its ability to function
as a heuristic device, as a help construction for
interpretation of phenomena. In that respect,
?nancial accounting data do not necessarily need
to have value relevance as such. Their value is
connected to their ability to function as a back-
drop by which ongoing business events can be
interpreted. Accounting classi?cation becomes the
art of background design while accounting, to use a
metaphor, functions as an organisational devel-
oping bath.
To construct a classi?cation schema or typology
for ?nancial accounting is an open issue. For
management control, a typology is unavoidable;
otherwise, there is no business logic. However, this
is a task for each company, i.e. to construct its
own success model.
The universe of discourse in ?nancial accounting
can be restricted to accounting transactions. If the
objective is to create a stable background for eco-
nomic interpretations, there is probably no reason
to change or expand the universe because of
intangibles. What is needed is simply to reclassify
the transactions. The sometimes–necessary reva-
luation of transactions according to an accounting
rule (standard) is a di?erent matter. It does not
change the nature of the classi?cation, just the
value of what had already been classi?ed.
Given a universe of accounting transactions, the
main challenge is to ?nd appropriate attributes. To
discover attributes is a process of creativity as
opposed to a process of science. The ?nal choice
re?ects the social (programmatic) part of
accounting. An attribute organises the universe in
accordance with a social assumption.
9
To classify
intangibles according to an attribute re?ecting
employees’ welfare would probably lead to a dif-
ferent classi?cation schema than of shareholders’
welfare.
One possible approach towards choice of attri-
bute is empirical. Classi?cations that are already
part of practice are analysed and the underlying
attributes are revealed. Classi?cations based on
those attributes can be further developed (as a
task of creativity). The BSC as described here
would be one obvious starting point.
The other approach is theoretical. One possibi-
lity is to start from di?erent equity theories (Hen-
driksen & van Breda, 1992), i.e. to whom net
income belongs and how to show equity relation-
ships in ?nancial statements. In relation to the
proprietary theory the proprietor welfare is of
main interest. Concepts such as ‘shareholder
value’ and ‘added value’ become central and
intangibles have to be objecti?ed and producti?ed
into marketable values. The attribute has to clas-
sify transactions into di?erent intangible (and
tangible) objects, including patents and trade-
marks.
Under the entity theory, the welfare concepts
are hidden because the company is assumed to
have a separate existence. Value to the business
becomes the crucial value concept and intangibles
will be a part of the production function. A classi?-
cation would probably wind up as a set of such
business functions as R&D and employee training.
A more modern entity theory is the value co-
production theory (Ramirez, 1999). In this theory,
intangibles are expressed in terms of relations.
What is then needed is a classi?cation of relations
(between di?erent value co-producers), such as the
classi?cations (typologies) developed to describe
relation marketing.
In the enterprise theory the welfare of all (many)
stakeholders is considered. Consequently, several
value concepts have to be introduced. Intangibles
will be a function of the di?erent stakeholders’ wel-
fare function. To map and classify each stake-
holder’s welfare function is, if at all possible, an
enormous task and would probably result in com-
plexity rather than simplicity (a lesson learnt from
‘social accounting’).
9
For an overview of the sociological perspective of
accounting see Jones (1995) and Roslender (1992).
710 J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713
One aspect of the confusion about intangibles
may stem from the fact that we use di?erent equity
theories during the process of describing and
interpreting intangibles. Another aspect is the
transformation of intangibles into other forms
of intangibles. One such example is the BSC,
where one type of intangible (performance dri-
ver) is supposed to be linked to another type of
intangible (outcome). A third aspect of the
confusion can be caused by the ontological
assumption that it is possible to divide the uni-
verse into tangibles and intangibles. Don’t we
often materialise the immaterial (through mea-
surements) and assign immaterial attributes to the
material? For example, we materialise trademarks
through attributing speci?c expenditures to the
(immaterial) concept. Yet, even if we could sepa-
rate them, many objects, states or events would
constitute di?erent combinations of tangibles and
intangibles. The conclusion must be to not take for
granted the division of the universe into tangibles
and intangibles. Depending on perspective, pur-
pose and type of ?nancial accounting model other
concepts may do a better job in classifying the
accounting world.
A ‘good classi?cation’ is a set of requirements; it
is something that we seldom, if ever, can fully
attain in practice, however. Presently, very little
can be said about its logical properties because
we have no fully articulated classi?cation. If the
universe consists of accounting transactions, then,
per de?nition, we have to ?nd an attribute that
makes the classi?cation exhaustive. To be exhaus-
tive, i.e. to classify all objects, states or events in
the universe of discourse, accounting often ends
up with a residual class (a non-class). That creates
a problem with exclusive classes, but it also hinders
the ambition to create a typology. The relation
between the residual class and the others becomes
unclear. We, therefore, have to settle for classi?-
cational schema with less information content.
The main problem will be to make it exclusive,
which is necessary if ‘rigid uniformity’ is to be a
means to create a stable background design. Con-
sistency is thought to be one requirement for the
basic accounting assumption of comparability. To
be consistent exclusive classes are likely necessary,
an assumption implying very precise de?nitions of
the classes. Nevertheless, the inherent vagueness
of language will always be an obstacle. In con-
trast to ?nite uniformity (there are relevant cir-
cumstances for similar transactions), exclusive
classes are also an argument for rigid uni-
formity (same classi?cation in di?erent organi-
sations without considering circumstances). If
there are relevant circumstances, they ought to
be covered by the classification attribute(s).
Consequently, there can be no relevant circum-
stances. Exclusive classes solve, in an accounting
sense, the problem of separation because logical
division is a way to keep things apart. It does not
matter if we deal with intangible assets or intangi-
ble expenses.
The simplicity concept was discussed in terms of
four dimensions. Except from technical/logical
aspects of simplicity, the important point here
concerns the concept of ‘usefulness’, i.e. how the
users perceive a classi?cation. In classi?cation
theory, simplicity has no reference to faithful
representation, which is often the strongest
argument against rigid uniformity and standar-
disation. Notwithstanding, looking upon ?nan-
cial accounting data as a background for
interpretation, the question arises regarding the
extent with which the interpreter is able to use
changing backgrounds. (For management control
purposes, it is probably necessary to design com-
pany-speci?c backgrounds, i.e. associations
between qualitative variables and organisational
success.)
The objective and subjective degree of simplicity
can be increased by using corresponding concepts
in the income statement and balance sheet to a
greater extent than are used today (e.g. R&D costs
and capitalised R&D). Finally, to judge the logical
simplicity a more developed classi?cation is needed.
Here we end with a set of ideas on how to
approach the construction phase of classi?cations.
It is about how we order the world, and account-
ing is expected to do the job.
J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713 711
Appendix A. Condensed description of IAS 38,
BSC and IC
IAS 38 (p. 3)
2. An intangible asset is an identi?able non-
monetary asset without physical substance held
for use in production or supply of goods or ser-
vices, for rental to others, or for administrative
purposes. An asset is a resource: (a) controlled by
an enterprise as a result of past events; and (b)
from which future economic bene?ts are expected
to ?ow to the enterprise.
3. IAS 38 requires an enterprise to recognise an
intangible asset (at cost) if, and only if (a) it is
probable that the future economic bene?ts that are
attributable to the asset will ?ow to the enterprise;
and (b) the cost of the asset can be measured
reliably.
4. IAS 38 speci?es that internally generated
goodwill, brands, mastheads, publishing titles,
customer lists and items similar in substance
should not be recognised as assets.
BSC
The basic idea of BSC is to translate strategy
into action. The starting point is the business
vision and strategy. Through a top-down
approach, this strategy is supposed to become
operationalised in four dimensions or perspectives:
?nancial, customer, internal processes and learn-
ing and growth. In each perspective a strategy is
formulated and key ratios (performance drivers)
expressing that strategy are constructed. The dif-
ferent key ratios are subsequently linked to each
other in a cause–e?ect chain, ending in a ?nancial
outcome.
IC
In the Edvinsson and Malone model, the start-
ing point is the market value of a ?rm. The di?er-
ence between the market and book value of the
?rm is denoted ‘intellectual capital’. Intellectual
capital is divided into human and structure capital.
Structure capital is further divided into customer
and organisational capital, and organisational
capital into innovation and process capital.
Finally, innovation capital is divided into intellec-
tual property and intangible assets.
References
AAA (1977), Report on the American Accounting Associ-
ation Committee on International Accounting Operations
and Education 1975–1976. The Accounting Review. 52,
Supplement.
Ahonen, G. (1998) Orally expressed at a meeting in Brussels,
14–15 September 1998.
Acko?, R.L. (1968). Scienti?c method—optimizing applied
research decisions, John Wiley.
Baudrillard, J. (1988). Selected writings. Stanford: Stanford
University Press.
Capon, N., Farley, J.U. & Hoenig, S. (1996). Toward an inte-
grative explanation of corporate ?nancial performance,
Kluwer Academic Publishers.
Czarniawska-Joerges, B. (1993). The three-dimensional organi-
zation—a constructionist view. Lund: Studentlitteratur.
Catasus, B. (1999). The labour of division—classi?cations in the
making, School of Business, Stockholm University, (unpub-
lished).
Clement, R. (1997). Intangible investment from an evolutionary
perspective. Chapter 2, OECD (forthcoming).
Cooper, R. (1989). The visibility of social systems. In
M. C. Jackson, P. Keys, & A. Croppers (Eds.), Operational
research and social sciences. New York: Plenum.
Edvinsson, L., & Malone, M. S. (1997). Intellectual capital.
London: Piatkus.
Gold?nger, C. (1997). Intangible economy and its implications
for statistics and statisticians. International Statistical
Review, 65(2), 191–200.
Gro¨ jer, J-E. (1997). Employee artefacts on the balance sheet: a
model illustration and implications. Journal of Human
Resource Costing and Accounting, 2(1), 27–53.
Hansson, H., & Karlsson, S.-G. (1974). bild & form, (Eng.
picture & form). Skolfo¨ rlaget: Uddevalla.
HEC. (2000). Classi?cation of intangibles. Chambre de
Commerce et d’Industrie de Paris.
Hedberg, B., & Jo¨ nsson, S. (1978). Designing semi-confusing
information systems for organizations in changing environ-
ment. Accounting, Organisations and Society, 3, 47–64.
Hegglo¨ v, E. (1998) Redovisning som hinder fo¨r la¨rande (Eng.
Accounting that hinders learning), (unpublished).
Hempel, C. (1952). Fundamentals of concept formation in
empirical science. The University of Chicago Press.
Hendriksen, E. S., & van Breda, M. F. (1992). Accounting
theory, ?fth edition. Homewood, Il: Irwin.
Hines, R. (1988). Financial accounting: In communicating
reality, we construct reality. Accounting, Organizations and
Society, 13(3), 251–261.
712 J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713
Hoegh-Krohn, N. E. J., & Knivs?a? , K. H. (1998). Accounting
for intangible assets: challenges and some solutions. Bergen:
Foundation for Research in Economics and Business
Administration.
Hopwood, A. G. (1987). The archaeology of accounting sys-
tems. Accounting, Organizations and Society, 12(3), 207–234.
IASC. (1996). International accounting standards. London:
IASC.
IASC. (1998). International accounting standard IAS 38, intan-
gible assets. London: IASC.
Johanson, U., Eklo¨ v, G., Holmgren, M. & Ma? rtensson, M.
(1998). Human resource costing and accounting versus the
balanced score card. (A report to OECD.) School of Business,
Stockholm University.
Johanson, U., Ma? rtensson, M., & Skoog, M. (1999). Measuring
to understand intangible performance drivers. Stockholm
University: School of Business.
Jones, T. C. (1995). Accounting and the enterprise. A social
analysis. London: Routledge.
Kaplan, R. S., & Norton, D. P. (1996). Translating strategy into
action: The balanced scorecard. Boston: Harvard Business
School Press.
Lerner, D., Lasswell, H. D. (Eds.). (1951). The policy sciences—
recent developments in scope and method. Stanford: Stanford
University Press.
Lev, B., & Zarowin, P. (1998) The boundaries of ?nancial
reporting and how to extend them, New York University,
working paper.
Mauritsen, J. (1999). Intellectual capital: The logic of an ‘econ-
omy of creativity’, Copenhagen Business School.
Nobes, C. W. (1984). International classi?cation of ?nancial
reporting. London: Croom Helm.
No¨ rreklit, L. (undated), The ?rm and its resources, (unpublished).
O’Shaughnessy, J. (1972). Inquiry and decision. Oxford: George
Allen & Unwin.
Ramirez, R. (1999). Value co-production: intellectual origins
and implications for practice and research. Strategic
Management Journal, 20, 49–65.
Roberts, A. (1995). The very idea of classi?cation in interna-
tional accounting. Accounting, Organizations and Society,
20(7), 639–664.
Roberts, H. (1999). Classi?cation of intellectual capital.
Norwegian School of Management, (unpublished).
Rose, N., & Miller, P. (1992). Political power beyond the state:
problematics of government. British Journal of Sociology,
43(2), 173–205.
Roslender, R. (1992). Sociological perspectives on modern
accountancy. London: Routledge.
Rudner, R. (1966). Philosophy of social science. Englewood
Cli?s, N.J: Prentice-Hall.
Statistics Netherlands. (1998). Intangible investments: de?nition
and data source for technological, marketing, IT and organi-
sational activities and rights. Statistics Netherlands (LNM-
reeks 9803).
Stolowy, H., & Jeny, A. (1999). How accounting standards
approach and classify intangibles—an international survey,
HEC School of Management (unpublished).
Wolk, H.I., & Tearney, M.G. (1997). Accounting theory—a
conceptual and institutional approach (4th ed.). South-
Western College Publishing.
J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713 713
doc_142459206.pdf
The present paper is concerned with classification of ‘intangibles’and what classification theory can teach us about a
‘good classification’. The present study led to three conclusions about the classification of intangibles. One is that the
value of classification lies in its function as a heuristic device, i.e. as a help construction. Accounting becomes the art of
background design. Another conclusion is that if we choose to account for intangibles it does not need to change the
transaction base of financial accounting. Rather, it is an issue how to classify and label a universe of transactions.
Finally, the dichotomy tangibleintangible should not be acknowledged as a supposition. Depending on perspective,
purpose and type of financial accounting model, other concepts may do a better job in classifying the accounting
world.
Intangibles and accounting classi?cations:
in search of a classi?cation strategy
§
Jan-Erik Gro¨ jer
School of Business, Stockholm University, S-106 91 Stockholm, Sweden
Abstract
The present paper is concerned with classi?cation of ‘intangibles’ and what classi?cation theory can teach us about a
‘good classi?cation’. The present study led to three conclusions about the classi?cation of intangibles. One is that the
value of classi?cation lies in its function as a heuristic device, i.e. as a help construction. Accounting becomes the art of
background design. Another conclusion is that if we choose to account for intangibles it does not need to change the
transaction base of ?nancial accounting. Rather, it is an issue how to classify and label a universe of transactions.
Finally, the dichotomy tangibleÀintangible should not be acknowledged as a supposition. Depending on perspective,
purpose and type of ?nancial accounting model, other concepts may do a better job in classifying the accounting
world. # 2001 Elsevier Science Ltd. All rights reserved.
1. The need of (re-) classi?cation or division
Accounting classi?cations, such as the chart of
accounts, are expressions of how we chose to
describe and, to some extent, de?ne an organisa-
tion at a certain point in time. The con-
structionÀreconstruction process of accounting
classi?cations is one important aspect of the
accounting history and, therefore, of management
history. The service functions that accounting is
supposed to ful?l are contingent on the ability of
accounting to order the world of organisations in
a (value) relevant way. This ability is challenged.
How shall we classify assets and liabilities, reven-
ues and costs and cash ?ows? What is supposed to
be ‘on-balance’ and ‘o?-balance’? Can we even
create an order between the di?erent classes?
One of today’s challenges concerning ?nancial
accounting classi?cations is to cope with an orga-
nisational world that seems to become more
immaterial than material, where resources in dif-
ferent immaterial forms act as the key production
factors. In an accounting context, this develop-
ment is re?ected in concepts such as immaterial
assets, intangibles, hybrid and intellectual capital.
Allied concepts, including virtual organisations,
relation marketing, value stars, human resource
costing and accounting, enablers and performance
drivers, are in themselves a starting point for
0361-3682/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PI I : S0361- 3682( 01) 00023- X
Accounting, Organizations and Society 26 (2001) 695–713
www.elsevier.com/locate/aos
§
This article is part of the MERITUM research program.
Meritum is the acronym for ‘‘Measuring Intangibles to Under-
stand and Improve Innovation Management’’. The research
program consists of four sub-projects: classi?cation of intangi-
bles, management control of intangibles, intangibles and the
capital market and developing and testing guidelines for
reporting intangibles. In HEC (2000) eleven preliminary papers
pertaining to the classi?cation issue are presented. MERITUM
is a joint research project among eight universities and research
institutions in Denmark, Finland, France, Norway, Spain and
Sweden. The research program has been made possible through
national research grants and a common grant through the
European Commission.
E-mail address: [email protected]
accounting classi?cations. They express a need
for altering the accounting boundaries or for
acting as a demarcation for reclassi?cation. One
way to improve the accounting art is to use new
organisational paradigms for accounting classi-
?cations. The accounting backbone is about
classi?cation and reclassi?cation, i.e. to arrange
something in a particular order. The accounting
art of portrait organisations is currently under
dramatic challenge.
The business community is experimenting with
di?erent intangible models and accounting con-
cepts (Johanson, Eklo¨ v, Holmgren, & Ma? rtens-
son, 1998). Such public authorities as Statistics
Netherlands (1998) develop and use their own
classi?cation of intangibles. Thus far, accounting
researchers seem to have developed little interest
in the ‘classi?cation of intangibles’ issue. As a
consequence, accounting policy organisations (e.g.
the IASC) tend to feel their way around like the
blind, an observation suggested by the evidence
that it took 10 years to release the accounting
standard IAS 38 for reporting intangible assets
(IASC,1998).
However, do we really have a mess of intangi-
bles
1
or do we have nothing to order, i.e. a lack of
a ‘good division’ of the intangible concept? To
divide an accounting concept into new sub-con-
cepts is an issue related to developing the
accounting language; it involves the process of
denoting phenomena in the process of describing
modern aspects of organising and organisations.
(The introduction of the concept ‘provisions’ in
the Swedish bookkeeping act of 1995 is an exam-
ple of further division of the balance sheet.) Clas-
si?cation and (logical) division are actually two
sides of the same coin. Therefore, when discussing
classi?cation in this paper, the arguments are to a
certain extent valid for division as well.
With reference to the introduction earlier, a
demand for a ‘good classi?cation’ of intangibles
from the business and research community seems
evident. The question is how to construct such a
classi?cation. Before building a classi?cation sys-
tem, however, it is necessary to know something
about what it is that entails a ‘good classi?cation’.
The aim of this paper is to formulate a strat-
egy for classi?cation of intangibles. The
approach involves the mapping of properties
of good classi?cation—its anatomy so to
speak—and confronting those properties with
corresponding accounting concepts.
The paper assumes that we can identify the
properties of good classi?cation if we can for-
mulate at least one strategy for accounting classi-
?cation (of intangibles). If we can map the
properties of a good classi?cation, we would have
at least one set of criteria (read ‘‘ideal model’’) for
the evaluation of alternative classi?cations.
1.1. Why classify?
The value of classi?cation is, according to Rud-
ner (1966), associated with its ability to function
as a heuristic device, as a help construction for
interpretation. Is that particular function a rea-
sonable description of a chart of accounts? It may
be presumed that Hegglo¨ v (1998) views account-
ing as a heuristic device when he judges traditional
accounting as a hindrance to learning. One can
suppose that Hegglo¨ v implies hindrance to learn
about management, not accounting. Hedberg and
Jo¨ nsson (1978) were most likely looking for the
heuristic aspects when they introduced the notion
of the semi-confusing accounting system. It is
probably the case that classi?cation can work in
both ways—to hamper or to help.
Nonetheless, the general idea underlying classi-
?cation is to facilitate and hence to promote
understanding. It is obvious that one aspect of
classi?cation is instrumental. Classi?cation is sup-
posed to be used as part of a certain practice. In
Rose and Miller’s (1992) conceptualisation this is
the technological aspect (elements). Classi?cation
as such is purported to facilitate bookkeeping in
1
The concept of ‘intangibles’ is used throughout the article
either with an empirical referent or as a general unde?ned con-
cept. Therefore, its contents will be given by circumstances.
Intangibles are, if nothing else is stated, discussed within a
?nancial accounting context.
696 J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713
that it brings order to a large disarray of transac-
tions. A speci?c classi?cation, however, is also an
expression of the normative aspects or values
embedded in practice, i.e. its programmatic ele-
ments (Rose & Miller, 1992). It is the outcome of
an articulated or unintentional world-view, of
political bargaining, lobbying and individual
interests.
If we can describe the world in simpler terms
(cp.), our chances to understand it increase sig-
ni?cantly. Therefore, according to Rudner (1996),
one aspect of assessing the acceptability of non-
theoretic formulation (its purpose) is based on the
simplicity of its set of primitive terms. He classi?es
(!) the uses of simplicity into six categories (the
tree structure is my own interpretation; Fig. 1).
‘‘If ‘simplicity’ is used to apply to the universe
independently of the perception of it, we shall
classify that usage as objective-ontological; if
‘simplicity’ is taken as a predicate of our (extra-
linguistic) responses to the (extralinguistic) uni-
verse, we shall classify it as subjective-ontological’’
(Ibid., p. 43). Classi?cation is an expression of
some ontological assumptions. Intangibles, as a
social construct or as empirical phenomena, are
based on certain ontological assumptions. The
ambition of this paper, however, is restricted to
the descriptive dimensions of simplicity. Objective–
notational simplicity has to do with the notational
property, such as the number and names of classes
and the number of inscriptions (e.g. of letters). To
illustrate, the accounts in a chart of accounts can
be denoted either by numbers or by names, where
numbers have a higher degree of objective–nota-
tional simplicity.
Objective–logical simplicity denotes the logical
structure of the predicates. To judge the logical
structure there is the need of such extralogical
predicates as re?exivity and symmetry to assess
the degree of simplicity of a scienti?c formulation.
There is a strong link between simplicity and sys-
tematisation. ‘‘When simplicity of basis vanishes
to zero—that is, when no term or principle is
derived fromany of the others—systemalso vanishes
to zero. Systematisation is the same thing as simpli-
?cation of basis’’ (Goodman, in Rudner, 1966, p.
44). When we assign each resource or accounting
transaction a speci?c class, we have not really sim-
pli?ed anything. To measure objective–logical sim-
plicity full formalisation is required. Because full
formalisation about intangibles is unworkable, the
question is whether partial formalisation is enough
to test this dimension of simplicity.
Subjective–notational simplicity denotes, for
example, notations that are recognised as familiar,
convenient and have some aesthetic quality. Sub-
jective–notational simplicity can be interpreted as
the link from a chart of accounts to behaviour
accounting or to the behaviour (psychological)
response to a name of an account. For instance,
how do investors perceive the concept ‘intellectual
capital’ compared with ‘intangibles’? To measure
or judge subjective–notational simplicity, it is
imperative that the appropriate criteria are
selected.
Subjective–logical simplicity denotes ‘‘how peo-
ple psychologically respond to logical properties
of theories‘‘ (Rudner, 1966, p. 43). (Rudner does
not expand on this concept because, like most
philosophers, he examines the contents and impli-
cations of objective–logical simplicity.) In my
understanding of this dimension of simplicity, it is
how the total structure, i.e. the total classi?cation,
is perceived. To exemplify, is it subjective–logical
to introduce a capital concept (intellectual, struc-
tural or human capital) for something on the asset
side of the balance sheet? Alternatively, would it
be more meaningful to introduce intangible costs
in the income statement? Here we enter into the
?elds of cognitive psychology and behavioural
accounting.
The idea of simplicity in accounting can be
traced to other concepts, such as the IASC (1998) Fig. 1. Dimensions of simplicity.
J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713 697
framework about qualitative characteristics, about
understandability and relevance (subjective sim-
plicity), and about reliability and comparability
(objective simplicity).
Introducing several dimensions of simplicity
raises the notion of trade-o?. Hoegh-Krohn and
Knivs?a? (1998) identify the main trade-o? issue of
intangibles as between relevance and reliability.
According to the authors, the present solution is
overly conservative and places too much emphasis
on reliability. And I have to agree. That is to say
that subjective simplicity (in accounting) is given
?rst priority because the classi?cation has to be
useful. Objective simplicity can be viewed in
accounting as a mean or help construction to
reach subjective simplicity.
So why do we classify? One obvious reason is
that we would like to create a certain order in
society, a certain way to direct our interpretation of
the world. Another reason is that it facilitates
understanding of the world through simpli?cation.
Focusing here on the second reason, the issue of
(subjective) simplicity must be vital when testing
di?erent accounting classi?cation constructs.
2. The nature of classi?cation
As soon as we describe something, we assign it
to a class. ‘‘Any descriptive word classi?es’’
(O’Shaughnessy, 1972, p. 23). For example,
assume that a brand name is an intangible asset.
‘‘Such a statement attributes a property to an
object, event, or state’’ (Acko?, 1968, p. 11). The
property here is what is reputed to be understood
as ‘intangible’ and ‘asset’. The statement is there-
fore equivalent to expressing the supposition that
brand name is a member of the set of intangibles.
‘‘That is, every predicational statement classi?es
its subjects’’ (ibid., pp. 11–12).
We use a language when we describe, a language
that is an expression for, and part of, a speci?c
culture. Because it is reasonable to interpret clas-
si?cations as linguistic artefacts, it follows logi-
cally that they are culture bound. The implication
is that technically the same intangibles may not
necessarily be perceived as similar in di?erent cul-
tures. Alternatively, from a cultural perspective,
identical intangibles may not be classi?ed in the
same way because they are assigned di?erent
properties. Take for example, the concept ‘knowl-
edge’. One option is to treat knowledge as some-
thing possible to separate in its own right.
Another is to treat knowledge as embedded in
human capital. A third option is that knowledge is
embedded in di?erent types of organisational
products and processes (like patents). Therefore,
every classi?cation is built upon some ideas about
the world, how it ought to be ordered.
Ruling accounting classi?cations are challenged
because ideas, culture and technology change
continuously. When we experience a technology
shift in a practice, the existing classi?cation is
challenged. This is precisely what occurred to
Linne´ ’s classi?cation of plants when instruments
for establishing the genetic code were developed.
The question is if we now see such a technology
shift in accounting through the IT development
or is it a shift of ideas about organisations and
organising?
What a classi?cation of intangibles does is to
create something based on constructs, i.e. a hyper-
real world (Baudrillard, 1988). The task of con-
structing a classi?cation of intangibles is not pri-
marily a scienti?c one. There is no logic of
discovery or construction (of the world), just of
validation. Discovery and construction is rather a
creative process. Although an intangible is a social
construct and, by de?nition, lacks tangible char-
acteristics, it nevertheless functions as such in the
imagination of the observer.
To classify something also implies de?ning it.
Therefore, a necessary requirement for classi?ca-
tion is a (set of) de?nition(s). Like most concepts,
‘intangibles’ have no generally accepted meaning
or de?nition. This is expected because objects,
states and events can be ascribed di?erent proper-
ties (or attributes). Few, if any, objects are one-
dimensional. The problem of de?nition can be
approached as part of a theory, as in Clement’s
(1997) evolutionary theory or in No¨ rreklit’s
(undated) resource-based theory in which classi?-
cation is based on the control method of the
resources, or in Gold?nger’s (1997) knowledge-
based theory in which di?erent classes of knowl-
edge are constructed. Alternatively, the de?nition
698 J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713
issue can be approached empirically as a categor-
isation of concepts used already in organisations
(see Statistics Netherlands, 1998). Stolowy and
Jeny (1999) have classi?ed the de?nitions of
intangibles stated by 21 countries into three
groups: tautological, by opposition and real. A
tautological de?nition is formulated ‘‘in order to
be included in the balance sheet category entitled
‘Intangible Fixed Assets’ an asset must be intan-
gible.’’ (Ibid. p. 7) A de?nition by opposition is
often expressed in the following way: ‘‘. . . intangi-
bles are ?xed assets other than tangible or ?nan-
cial’’ (Ibid. p. 7). Finally, the third classi?cation,
real de?nition, is an attempt to de?ne intangibles
by other concepts. According to Stolowy and Jeny
(1999), only four of the 21 countries have tried to
formulate a real de?nition.
There are two options available when one
wishes to make the de?nition of intangibles more
information rich. One is to use a deductive
approach. The other is empirical, i.e. to create a
list of intangibles based on empirical observations.
Although both approaches seem valid during the
present stage of experimentation, we need a de?-
nition, either for classi?cation or for division of a
concept into sub-classes.
In logic, the process of breaking down a class
into sub-classes is referred to as ‘logical division’.
To divide we need (to construct) an attribute (for
intangibles) that can be divided into di?erent
classes. This is a normative, deductive approach.
‘‘Classi?cation is the complementary but reverse
process to logical division. The di?erence lies in
viewpoint and the starting-point in procedure’’
(O’Shaughnessy 1972, p. 24). Classi?cation brings
objects (speci?c intangibles) together based on
certain attributes. Classi?cation also means to
reduce the number of relevant attributes of an
object. In that sense, reductionism becomes a pre-
requisite for holism. Only the attribute(s) that is
the basis for classi?cation counts. For instance,
assume that ‘knowledge’ is one class of intangibles
consisting of the sub-classes ‘human knowledge’
and ‘organisational knowledge’. In this example
the only two attributes we are looking for is
human and organisational. All other possible
attributes of knowledge are of no classi?catory
interest.
To summarise, to classify/divide something
means to describe, which implies to explain
the nature or essential qualities of something.
To classify or divide we need to construct
properties/attributes that describe an object,
event or state.
Science gives us continuously new opportunities
to describe ‘essential qualities’, and politics is set-
ting the priorities of accounting change. However,
attribute construction can be approached more
fruitfully if we ?rst formulate some ideas about what
(ideal) types of classi?cation there are and if there is
something to classify (i.e. a universe). These two
issues are addressed in the following two sections.
3. In search of a classi?cation schemata or
typology
We are in search of a classi?cation, as a formal
construct, as a non-theoretic formulation but that
is perhaps linked to theory. ‘‘The two types of
classi?cational schemata are (1) those in which a
condition that is both necessary and su?cient for
the applicability of each of the classifying terms is
given, and (2) those which such conditions are
given for none or only a few of the classi?cational
terms.’’ (Rudner, 1966, p. 32)
In accounting, we normally operate with the
second kind of classi?cational schemata (classi?-
cational analytical schemata). This is because
either they fail to classify the universe of discourse
exclusively and exhaustively, or the classi?cation is
not done by logical means but only by some con-
tingent feature of the world.
The balance sheet is an example of classi?ca-
tional analytical schemata. However, it can be
described as a speci?c form of classi?cational
schemata, i.e. as a typological system (compare
nominal versus ordinal scales). According to
Rudner (ibid., pp. 37–38), a typological system
must contain at least the following components:
1. ‘‘A concept determining the typology’s uni-
verse of discourse’’. As an example, let us
choose assets as the universe of discourse.
J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713 699
2. ‘‘Some relation(s) that determine(s) an
ordering (e.g., a serial or quasi-serial order-
ing) among the members of the universe of
discourse’’. Assets can be arranged either in
liquidity order or in purpose-of-use order.
3. ‘‘Statements implying that certain features
(e.g. transitivity) characterise the relation
(s)’’. Regarding transitivity, if accounts
receivables are less liquid than cash, then
cash is more liquid than accounts recei-
vables.
4. ‘‘A set of concepts (frequently a ‘polar-pair’
or ‘extreme opposites’) usually designating
some speci?c members of the universe of
discourse that are ‘distant’ from each other
or at opposite ends of the array into which
the ordering relation(s) order the members of
the universe of discourse’’. Cash deposits may
be viewed as one natural endpoint of liquidity
order: the question remains, however, to
determine the other endpoint? Even if the asset
side is arranged according to the purpose-of-
use order or when the assets are supposed to
be used, a problem still remains. A certain
amount of cash is needed in eternity while
land can be held for just a short period.
It follows from the earlier description of a
typology that a typological system contains more
information than classi?cational (analytical) sche-
mata in that it relates the di?erent classes to each
other. Assume that we have identi?ed three classes
of intangible asset: trademark, R&D and
employee knowledge. These three classes con-
stitute a classi?cational schema of intangibles.
However, if we can add something (through logic)
that relates the three classes to each other addi-
tional to their sub-class relation to intangibles,
then we will have a typology.
We can now add to the classi?cation construc-
tion agenda: Should the ambition be to construct
classi?cational analytical schemata or typologies?
The answer is more than logical ambition. The
former is a non-theoretic formulation that simply
groups things together. The latter still keeps the
link between theory and non-theory open, i.e.
whether some testable propositions should be
included or not.
4. Is there a universe of discourse to classify?
It was stated earlier that there is a lack of real
de?nitions of intangibles. How do we then ?nd the
intangibles to classify (or divide), i.e. a universe of
discourse? As indicated previously, we can either
identify intangibles as an empirical endeavour or
construct (divide) it as a theoretical project.
Whichever, we have two options as illustrated
with a balance sheet example. One is that the
classi?cation can start, for instance, with a classi-
?cation of economic resources and treat assets and
non-assets as a subset of economic resources.
Here, the classi?catory concept, which assigns
properties to the di?erent members of the uni-
verse, is merely used to divide accounting from
non-accounting resources. The other is that the
classi?cation starts directly with those elements
that can be assigned accounting properties and
thereafter begins a search for classifying con-
cepts. While a classi?cation of resources would
probably result in di?erent accounting classi?-
cations, it would also take into consideration
the dynamic change or transformation of prop-
erties for a given member of the universe. For
example, if trademarks become marketable in
general and, therefore, accountable, the object is
already in the classi?cation schemata as a
resource. However, trademarks will change sub-
class from non-asset to asset. The universe of dis-
course is not stable and new objects, states or
events can give rise to new classes. Accounting is,
as Hopwood (1987) suggests, under the process of
becoming.
A paradox exists in that each classi?cation or
division has to start with a (meta) division. What
Cooper (1989) and Catasus (1999) call ‘labour of
division’ emphasises the process of dividing rather
than the outcome of the process. A logical con-
sequence of their reasoning is that to get any
object, state or event classi?ed we must divide or
separate them from the rest of the world. In this
respect, it is evident that labour of division is not
only a process of seeing but also one of simpli?-
cation in which a reduction in context occurs. The
issue then is whether it is possible or even desirable
to divide intangibles and tangibles into separate
categories or classes.
700 J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713
IASC (1996) makes an attempt to separate an
accounting universe of discourse (from the rest of
the business world). ‘‘Financial statements portray
the ?nancial e?ects of transactions and other
events by grouping them into broad classes
according to their economic characteristics’’ (p.
54). However, this description or de?nition is too
vague to be operationally useful. Furthermore, the
term ‘other events’ is so broad that it can imply
almost anything.
In traditional accounting textbooks, the number
of mentioned intangibles is limited. Hendriksen
and van Breda (1992) list 13 of what they call tra-
ditional intangibles, and add to the list 13 deferred
charges related to expenditures on services. Fur-
thermore, if we search the literature for potential
intangibles relative to the balanced scorecard,
intellectual capital, etc. the list can be extended
inde?nitely. Even if only four countries have some
sort of real de?nition of intangibles, all of the 21
countries except one had a list of intangibles
(Stolowy & Jeny, 1999).
The conclusion is that there is a constructed uni-
verse of discourse of intangibles. The question then
is whether it is possible to classify or divide the
universe in some way, i.e. to simplify. In both
cases, we need a selection of attributes that helps
us to di?erentiate the objects, states, or events in
our universe of discourse. The selection of attri-
butes is contingent on the purpose of classi?ca-
tion, i.e. there are no natural selection of
attributes. Given the attributes, the construction
of a classi?cation schema is a matter of logic.
However, the real classi?cation of objects, states
or events remains a matter of interpretation.
5. Di?erent approaches to attribute identi?cation
As previously noted, one of the main issues in
classi?cation is the choice of attributes that dis-
criminate between the to-be-classi?ed objects,
states, or events. Roberts (1995) divides the search
for attributes into ?ve approaches in his classi?-
cation of ?nancial accounting systems in di?erent
countries: essentialist approaches, overall similar-
ity approaches, diachronic approaches, set theory
approaches and archetypal approaches.
The essentialist approaches, based on Aristotle’s
works, seek to identify the essence of an object
(but not in de?ning it). Essence, in Aristotle’s
world, was found through the assumed functions
of an object. The object’s attributes can be
deduced from its most essential properties.
Hempel (1952) explicates a distinction between
natural and arti?cial classi?cation. The former is
an essentialist approach, assuming that the objects
possess fundamental similarities. Super?cial
resemblance or external criteria determine the lat-
ter. A classi?cation of intangibles would be, in
Hempel’s system, an arti?cial classi?cation. How-
ever, the subject matter at issue is whether all
classi?cations are arti?cial or whether only the
selected attributes are arti?cial.
The important question here is if it would be
possible to identify and discriminate between
functions of intangibles. Can an intangible—a
social construct—have a function per se? One
possibility would be to link di?erent intangibles to
a production function or value creation process.
We would next have to formulate the production
function or value creation process before we
search for the functions of di?erent intangibles
because their functions would be identi?ed in
relation to those processes. To use a negative for-
mulation, if trademarks have no function in the
value creation process then they cannot be part of
the classi?cation.
The overall similarity approaches use all (read
many) attributes to look for similarities between
objects. A statistical procedure, such as factor
analysis, is used to search for similarities whereby
similar objects form a class. The problem, how-
ever, is the choice of possible attributes: ‘‘If we
take any set of objects there is an in?nity of attri-
butes that these objects may have or that we might
choose to give them. . .’’ (Roberts, 1995, p. 648).
The overall similarity approaches are appro-
priate and more fruitful when the set of attributes
is given and the problem is to di?erentiate and
reduce the universe of discourse to good classi?-
cation. I doubt that this could be the case for
intangibles. Moreover, an evident risk exists of
encountering complicated classi?cations that
would violate one basic idea, namely, the one of
simpli?cation.
J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713 701
The diachronic approaches introduce time rela-
tionships as attributes: ‘‘. . .the essence of an object
is no longer what it is in itself but rather how it
came to be what it is’’ (Roberts, 1995, p. 649). As
Roberts’ comment suggests, these approaches are
closely related to the essentialist approaches.
To use time relationships to classify intangibles
it must be plausible to divide the production
function or value creation processes into separate
and sequential time stages. Ahonen (1998) does
just that when dividing exploitable and generic
intangibles, where the latter are not yet exploi-
table. This dichotomy can also be interpreted as
an essentialist approach. Furthermore, the
balanced scorecard (Kaplan & Norton, 1996)
assumes an intricate relationship between, past,
present and future of the di?erent performance
drivers and ?nancial outcomes.
A diachronic approach places greater emphasis
on the origin of the intangibles while the essentialist
approaches primarily focus on the destiny of the
intangibles (their function). The gist of the problem,
then, is choosing between origin and destiny.
The set theory approaches try to avoid the pro-
blem of attribute choice by introducing some
properties of a classi?cation that are thought to
express its property of ‘goodness’. Roberts (1995)
examines four properties of classi?cation used by
the American Accounting Association: con-
sistency, exhaustibility, mutual exclusivity and
hierarchical integrity (Section 7).
It is unclear why Roberts (1995), in his debate
with Nobes (1984), (‘‘International Classi?cation
of Financial Reporting’’) classi?es ‘set theory’ as
an approach for attribute search. First, it is, per
de?nition, a non-approach or, at best, an indirect
approach. One can easily imagine other non-
approaches. Second, it is unthinkable to deduce
attributes from the properties of a ‘good classi?-
cation’. Rather, the properties are used in a testing
phase that generates feedback for reclassi?cation
(introducing new attributes), again to be tested in
relation to the classi?cation properties. Therefore,
set theory approaches are more a part of the vali-
dation processes of classi?cation schemata than
acting as possible attributes themselves.
Finally, Roberts (1995) disseminates the arche-
typal approaches. According to Roberts, ‘‘. . . the
principle which should guide attribute selection is
not similarity or di?erences between those attri-
butes per se, but similarities or di?erences in the
relationship of those attributes’’ (Roberts, 1995, p.
659).
The archetype or ideal type approaches are par-
ticularly appropriate when dealing with the classi-
?cation of systems that contain many elements
and relations. Nonetheless, it is still necessary to
de?ne the elements in the system before we can
obtain an expression (attributes) of the relations
between the elements and how they are related to
the system as a whole. Whereas earlier we dis-
cussed origin and destiny, the focus now is on
transformation processes. Perhaps the best exam-
ple of this approach is the balanced scorecard of
Kaplan and Norton (1996). The idea of these
authors is to link each performance driver, more
often than not an intangible, into a cause–e?ect
relationship ending in a ?nancial outcome. Its
relation, therefore, gives the function of each per-
formance driver.
When identifying attributes from processes, two
issues arise that are valid for all approaches con-
sidered within classi?cation theory in conjunction
with ‘exclusivity’, ‘set theory’ and ‘fuzzy sets’. The
issue is the di?cult task of how to draw (de?ne)
the border between classes.
Philosophers use the word vagueness (not to be
confused with ambiguous concepts) to denote
problems associated with border cases. When does
a number of grains of sand become a pile of sand?
How many inhabitants in a big city must be taken
away before it becomes a small city? Is it possible
to draw a de?nite border between the material and
immaterial, as well as between the di?erent intan-
gible classes? In mathematical terms, we have a
problem of selecting a cut-o? point on a con-
tinuous function for a given attribute. To reduce
this problem we have to look carefully into the
de?nitions of the attributes. Nonetheless, the
inherent vagueness of language will always be a
reason for di?erent interpretations of the same
phenomenon and, as a likely consequence, di?er-
ent classi?cations.
The second issue concerns whether objects (their
attributes) can take di?erent forms (compare with
solid, ?uid and gas forms of the same matter). The
702 J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713
problem has to do with attributes that change over
time. The most obvious example is an option that
is supposed to be turned into something else (or
nothing). Another problem is that di?erent forms
may exist at the same point in time. Several forms
of human resource and their artefacts can be
identi?ed, a situation that results in classi?cation
and double-counting problems.
We get little advice from Roberts (1995) on an
appropriate course of action in selecting attributes
and matters concerning the construction issue. ‘‘It
has been argued that classi?cation is a way of
viewing the world, a Weltanschauung with the
‘truth’ of classi?cations being relativised. From
the recognition of this idea, it has been suggested
that classi?cation should be created in line with
their purposes’’ (Roberts, 1995, p. 661). First,
‘truth’ is not an issue in classi?cation theory.
Classi?cation schemata have none empirically tes-
table proposition while a typology can include
testable ones. Second, to state that something
ought to be constructed with a ‘purpose’ has very
little information content as long as a speci?c
purpose is not stated.
The objective is to ?nd a stable classi?catory
term uncontaminated by contingent factors.
Observe that the classi?catory term itself can be
challenged by contingent factors. The dichotomy
short-term (less than a year) and long-term (more
than a year) is such a stable classi?catory term
(time horizon). We do not need to reclassify when
new objects are discovered or constructed. Reality
can prove time horizon to be an obsolete attribute,
however. An economic resource is not a stable
term in that its changing attributes are the cause of
reclassi?cation. Thus, we have to look for other
concepts or to incorporate the dynamics of
resources as an attribute.
Searching for classi?catory concepts involves a
mixture of purpose of classi?cation and creative
work. Embedded in the technocrats order creation
is the programmatic order creation. Which values
or norms are embedded in a speci?c classi?cation
schema? Is turning the balance sheet upside down
a technical matter or an expression of di?erent
priorities (e.g. land investments more important
than cash)? The key argument concerns whether
classi?cation schemata of intangibles ought to be a
part of social theory or are they based only on or
related to social theory? Immaterial resources,
intangibles, or intellectual capital are not to be
found as real objects. What constitutes a resource
is both a political and an economic matter.
So far, we have identi?ed three strategies in our
search for classi?cation attributes.
1. To use the present attributes in (?nancial)
accounting: for the balance sheet, ‘time-hor-
izon’ and ‘degree of liquidity’ would still be
the two chief attributes.
2. To reveal or construct attributes in relation to
identi?ed intangibles. The task would be to
list all intangibles used by organisations.
3. To develop new combinations of attributes
through, for example, morphologic meth-
od
2
that may or may not be related to a theory
of the ?rm. This strategy is more a matter of
division than classi?cation.
6. How to test/validate classi?cation schemata
What consequence does testing of classi?cation
schemata have on construction? If we aim for a
classi?cational schema, which is a non-theoretic
formulation, a speci?c class cannot be subjected to
empirical tests (Rudner, 1966). It makes little
sense to test, for example, if R&D ought to be a
speci?c class in a classi?cational schema. We have
to know the other classes that divide the universe
of discourse. What we can do is to test the whole
classi?cation schema, its logical properties and if it
is used (adopted).
When the dialogue of classi?cation theorists is
about ‘good classi?cation’, they are referring in
part to its logical properties. Rudner (1966) limits
the number of properties to two (except simpli-
city). One is the ability of classi?cation schemata
to be exhaustive, i.e. all objects of the universe of
discourse can be classi?ed. The other property is
that classi?cation schemata are exclusive, i.e. no
object can belong to more than one class.
2
Morphologic method involves combining di?erent known
attributes into new combinations.
J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713 703
Roberts (1995) develops his discussion around
the four properties of a ‘good classi?cation’ as
formulated by the American Accounting Associa-
tion (AAA, 1977). Exhaustive and exclusive are
two of the properties. The third is consistency. In
concurrence with Roberts (1995, p. 653) the need
for this third property is questionable. Consistency
states simply that, as soon as one attribute is cho-
sen, it should be used throughout the classi?cation
hierarchy to di?erentiate between the objects (a
classi?cation does not necessarily need to be a
hierarchy). If we use the attribute ‘marketable’ for
intangibles, then a second level must use di?erent
aspects of the marketable concept for further
classi?cation.
The fourth property introduced by the AAA is
hierarchical integrity. This property deals with
setting labels for the di?erent classes and how
those labels can be interpreted. Another aspect is
the relation between the di?erent classes (such as
specie, class and family). From my understanding
of hierarchical integrity, it is linguistic in nature
rather than logical. In Rudner’s language, it is an
issue of subjective–notational simplicity and
therefore it is treated here as part of simplicity.
Another suggestion concerning the criteria for
‘good classi?cation’ is Lazerfeld and Barton’s (in
Lerner & Lasswell, 1951). They present four cri-
teria for good classi?cation: (1) articulation, from
the general to the speci?c; (2) logical correctness,
exhaustive and mutually exclusive; (3) adaptation
to situations; and (4) adaptation to the respon-
dent’s frame of relevance. These criteria, however,
do not add anything to good classi?cation, in that
logical correctness has already been dealt with and
the other three criteria are all di?erent aspects of
simplicity.
Finally, when dealing with accounting informa-
tion, we may have to introduce information eco-
nomics as an extra property. Even if a new
classi?cation of intangibles can be helpful in pro-
ducing economic decisions that improve the capital
market e?ciency (to zero information costs), the
cost of classi?cation change and data retrieval can
be higher than the newly added value. Here the
issue involves choosing between a number of pos-
sible testing implications rather than constructing
new ones (as attitudes, behaviour, use and result).
Thus far, we have largely restricted our discus-
sion to testing the technocratic aspects of classi?-
cation schemata. Would it be possible to test or
reveal their intrinsic programmatic aspects as well?
Such a test should be based on interpretation of
the consequences using the classi?cation schema.
Accordingly, it is possible to reverse the approach,
i.e. begin to think about the consequences or
e?ects we would like to achieve through the pro-
cess of classi?cation.
The other aspect of validation concerns the
choice of an empirical test if the classi?cation
schema is adopted. However, we ?nd ourselves in
a ‘catch 22’ situation in the sense that we ?rst have
to construct a classi?cation before we can study if
it becomes adopted.
To summarise, there are two things to be learnt
about construction from the validation issue.
Construction can start from the idea of a ‘good
classi?cation’, that it should be simple and that it
should classify the universe of discourse exhaus-
tively and exclusively. However, construction can
also begin from an idea of wishful e?ects of the
classi?cation.
7. Classi?cation theory concepts meet accounting
In the previous sections, an attempt was made
to formulate a classi?cation strategy for intangi-
bles based on classi?cation theory. In this section,
a tentative confrontation between classi?cation
theory concepts and ?nancial accounting is
attempted. The three requirements placed on clas-
si?cation so far are that it has to be simple,
exhaustive and exclusive.
7.1. Simplicity and usefulness
With reference to Rudner (1966), the purpose of
classi?cation schemata is to simplify, in di?erent
dimensions, the universe of discourse. This is its
technocratic purpose. The programmatic purpose
may be expressed in terms of perceiving classi?ca-
tion schemata as reality construction, as Wel-
tanschauung, as a set of ontological–subjective
assumptions. One such assumption is to look for
utility (usefulness) rather than purpose. Utility
704 J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713
seems to bring the subject closer to context, and
the objective of this project is to examine use and
context, whatever the purpose. What is needed is
to formulate a set of normative statements about
utility (ex ante) as a device for construction. Then,
as part of testing, we can try to grasp the empirical
meaning of utility.
Usefulness, the potential to be used, is an
essential assumption about accounting data and
accounts. One way to judge the usefulness of
?nancial accounting data is to measure their
value relevance in explaining stock prices and
returns. That association has decreased between
1980 and 1996 (Lev & Zarowin, 1998), indicat-
ing the need for reclassi?cation or a change of the
accounting borders. The equivalent non-usefulness
in management control is the emergence of new
measurement packages under di?erent labels.
A problem with these value relevance studies is
that they test direct relationships between ?nancial
accounting data and stock prices. The annual
statement is mainly a historical document that
articulates about the ?nancial consequences of
past transactions and events. Therefore, its value
relevance can be sought in its stewardship func-
tion, or, in more future-oriented terms, as a back-
ground or lens for interpretation of ongoing
events in a turbulent economy (Fig. 2).
With such a view on value relevance of ?nancial
accounting data, the problem centres on how to
design an appropriate background.
3
A background
may have no value (relevance) of its own. A help
construction gives reference points. In this sense,
new designed backgrounds (read classi?cations)
can function as a heuristic device. The interpreta-
tion of a given event may change because of the
new background. (Compare the problem in colour
classi?cation when placing two colours side by
side. The perception of one colour is at the bor-
derline dependent also on the other. This phe-
nomenon is named colour induction or
simultaneous contrast).
4
The design of a new background can be done
either by introducing new accounting concepts,
such as intellectual accounting, or by rearranging
(classifying) the already recorded transactions
under new headings. A change in the names of the
accounts, the headlines, or the metaphors is likely
to have far reaching implications in the way we
perceive organisations.
5
Another aspect of value relevance concerns
studies that seek to establish a link between quali-
tative variables and business success of which the
balanced scorecard is an expression. Attempts to
integrate business success factors into the
accounting model, however, would probably be
futile. The determining factors underlying success
seem to be ?rm- or situation-speci?c, a conclusion
based on more than 400 empirical investigations
(Capon, Farley, & Hoenig, 1996).
7.2. Exclusive and consistent
As part of usefulness, much of ?nancial
accounting regulation is based on the objective to
achieve comparability in space and time. Con-
sistency, then, is assumed to be one prerequisite
for comparability. As IASC (1996, IAS1, p.76)
puts it: ‘‘3. Going concern, consistency and
accrual are fundamental accounting assump-
tions.’’ Further, IASC de?nes consistency in the
following way: ‘‘It is assumed that accounting
policies are consistent from one period to
another’’ (Ibid., p. 77). Hendriksen and van Breda
(1992, p. 142) are more precise and let con-
Fig. 2. Financial accounting data as background.
4
See Hansson and Karlsson (1974).
5
A most convincing story of reality construction has been
told by Hines (1988) and introduced into an organisational
theory context by Czarniawska-Joerges (1993).
3
In ?nancial accounting the annual statement [more pre-
cisely the balance sheet (book value) and pro?t and loss state-
ment (earnings) is supposed to constitute the background].
J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713 705
sistency refer to ‘‘. . . the use of the same accounting
procedures by a single ?rm or accounting entity
from period to period, the use of similar measure-
ment concepts and procedures for related items
within the statements of a ?rm for a single period,
. . .’’ The words ‘same’ and ‘similar’ in this context
presume an underlying classi?cation.
Changing accounting procedures implies that we
either change the properties of, for example, a
given object or change the class belonging of a
given object. In either case, comparability will be
distorted. From a classi?cation viewpoint, we are
dealing with two types of consistency, i.e. in
recognising and classifying. To be consistent yields
no information about what is ‘right’ or ‘wrong’.
Performing the same mistake repeatedly is a
demonstration of consistency. However, it is impos-
sible to be consistent if the classes are non-exclusive.
For instance, if a particular labour expense can be
treated either as part of trademark or as R&D, we
will treat that expense consistently just by chance.
Therefore, a necessary but not su?cient condition
for consistency is exclusivity of classes.
When di?erent ?rms are using the same proce-
dures, we normally denote that as uniformity
rather than as a re?ection of consistency. ‘Rigid
uniformity’ means ‘‘. . .prescribing one method for
generally similar transactions even though rele-
vant circumstances may be present’’ (Wolk &
Tearney, 1997). Then, ‘?nite uniformity’ implies
consideration of relevant circumstances, i.e. the
use of di?erent methods for similar transactions.
Most accounting theorists advocate ?nite uni-
formity because representatively it is thought of as
being more faithful. This argument is likely to
represent a form of circular reasoning. If some-
thing is relevant (i.e. a property of an object) and
is omitted, it becomes less faithful representa-
tively. Alternatively, they are dissimilar because
there are relevant circumstances that make them
dissimilar. Given the properties or attributes, a
‘good classi?cation’ is based on the idea of rigid
uniformity. What we do when we introduce some-
thing like ‘relevant circumstances’ is to create a new
attribute. The classi?cational schema was not
meant to handle that attribute. There are no rele-
vant circumstances and, if there were, that would
indicate a need for a redesigned classi?cation.
Finally, from the viewpoint of classi?cation
theory, if a classi?cation has exclusive classes, the
issue of separability is solved. IAS 38 (IASC 1998,
p. 16) is ambivalent about separability. In point 11
the following argument is expressed: ‘‘An intangi-
ble asset can be clearly distinguished from good-
will if the asset is separable’’. In point 12 the
document notes the following: ‘‘Separability is not
a necessary condition for identi?ability since an
enterprise may be able to identify an asset in some
other way’’. To separate in ‘some other way’ must
denote some sort of separation. The chart of
accounts (logical division) is based on the idea of
separation and such separation is necessary in
order to report intangible expenses or assets.
7.3. Exhaustive and residual classes
A consequence of division in accounting is that it
often leads to a residual class. One well-known resi-
dual class is goodwill. This is expressed by some as
the most intangible of all intangibles. Its relative
growing importance with respect to total assets
(Gro¨ jer, 1997) can be one explanation of the dimin-
ishing value relevance of ?nancial accounting data.
Goodwill presents the opportunity for accounting
nonsense (i.e. the ?gures add up but do not make
sense) instead of reporting vital resources or assets.
A residual class (non-class) creates a problem
because its relation to the other classes becomes
unclear. There is no room for such classes in a
‘good classi?cation. Therefore, residual classes,
such as goodwill as a result of purchase account-
ing, are to be avoided’.
The confrontation between classi?cation and
corresponding accounting theory concepts has
resulted in several requirements concerning the
construction of the classi?cation. In the next sec-
tion a confrontation between classi?cation con-
cepts and three used classi?cations is presented.
8. IAS 38, BSC and IC from a classi?cation
viewpoint
Although critical examination of IAS 38, BSC
and IC from a classi?cation viewpoint can be
made in its own right, it can also provide mean-
706 J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713
ingful ideas for future classi?cation attempts.
6
IAS
38 is important because it regulates what is pre-
sumed to be intangibles. The BSC is bene?cial
because it is widely used and makes a connection
between today’s intangibles and tangibles within a
cause–e?ect chain (a typology). Finally, IC is
relevant because it is an example of a division of
tangible concepts into intangibles.
Let us start by confronting the three models
with principal classi?cation concepts discussed in
previous sections serving as an overview of the
analyses to be followed (Table 1).
The general conclusion is that we are dealing with
three very distinguishable classi?cations, all of
which are wrought with inadequacies. The rest of
the section will discuss this conclusion in greater
detail through interpreting the three models in
relation to classi?cation concepts.
8.1. Universe of discourse
In IAS 38 a universe of intangibles is created
through subtracting the tangible universe (of
assets) from the total universe (of assets). Then, a
part of the intangible universe of discourse is
reduced through taking away ‘‘. . . goodwill,
brands, mastheads, publishing titles, customer
lists and items similar in substance’’ (IASC, 1998,
p. 3). In e?ect, this means that anything without
physical substance can be an intangible, except
for a few mentioned items and everything that
is considered similar to those items. As a con-
sequence, the universe of discourse becomes a
vague concept. IAS 38 creates a nebulous universe
by using a concrete listing of intangibles and by
stating general requirements for something to be
an intangible. From that universe, some given
intangibles are subtracted. De?ning a universe
by listing can be the only remaining solution
when the de?nitional problems become proble-
matic.
In the BSC model the distinction between tan-
gibles and intangibles is of no great importance.
Rather, the distinction is made between perfor-
mance drivers and their outcome. The total uni-
verse consists of four sub-universes. The ?nancial
sub-universe can be seen as a background towards
which the three other perspectives can be inter-
preted. The origin of the three perspectives can be
traced to a traditional demand (customer perspec-
tive) and supply model (internal business processes
and learning and growth perspective). The main
problem with the BSC model from a classi?cation
perspective is that the de?nition of each sub-universe
is poorly stated. Consequently, the boundaries
between these sub-universes are also ill-de?ned,
Table 1
Confronting IAS 38, BSC and IC with classi?cation concepts
Classi?cation concepts IAS 38 BSC IC
General concepts
Universe of discourse As a between two universes A set of four universes As a di?erence between two values
Attribute Essentialist plus by listing Archetypal plus diachronic Overall similarity
Type of classi?cation Classi?cation schema Typology Classi?cation schema
‘Good classi?cation’
Exhaustive No (Yes) No No
Exclusive No (Yes) No No
Simplicity
Objective–notational Unclear No Yes
Subjective–notational Unclear Yes No
Objective–logical No No No
Subjective–logical No Yes (No) Yes (No)
6
BSC (Kaplan & Norton, 1996) and IC (Edvinsson & Malone, 1997) are treated in the present paper as fairly well de?ned models.
For an extremely condensed description of the two models, as well as IAS 38, see Appendix A.
J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713 707
especially between internal business processes and
learning and growth. However, the idea is, as I
understand it, to use the four perspectives as a
help construction for developing cause–e?ect rela-
tionships between performance drivers and out-
come. This cause–e?ect chain is the constructed
typology of the theoretical claim by Kaplan and
Norton (1996) that a properly constructed BSC
articulates the theory of the business.
In IC, a universe (IC) is constructed by taking
the di?erence between two values, i.e. the di?er-
ence between the market and book value of the
company. From this single value a whole universe
is created. (To use a metaphor, it is like measuring
the volume of a big pear and deducting the volume
of a small apple. Because there is a di?erence, it
has to be baptised, i.e. it becomes a plum.) Fur-
ther, the amount of IC changes as soon as share
prices change, or when accounting practice chan-
ges. Moreover, as soon as an intangible is capita-
lised on the balance sheet, the amount of IC drops
(as a residual) and the universe grows gradually
less. The conclusion must be that the IC universe
can only be justi?ed in the name of rhetoric.
8.2. Attribute
In IAS 38 two attributes can be traced. The ?rst
is ‘recognition’. This concerns whether the intan-
gible can be recognised or not. The attribute can
be interpreted as an exponent of accounting ‘con-
servatism’. The other attribute is ‘origin’ (essenti-
alist)—from where the intangible originates,
internally generated and externally acquired. One
can reasonably question whether that division is
fruitful. Today, several business functions have
been outsourced or are co-produced. What 1 day
is internal becomes external the next, and vice
versa. In the BSC there is a speci?c ‘time’ (dia-
chronic) division: past, present and future. There
is also an archetypal attribute to separate di?erent
objects, states and events into the three non-
?nancial classes. The attribute in each sub-class,
however, is, and has to be, left open because the
aim of the model is that it is suitable for all types
of business logic. The attribute for division of IC
into sub-classes is also based on some hidden
properties of ‘overall similarities’ related to a
‘value to the business’ idea. The selection of attri-
butes seems to be as crucial as it is complicated and
will be dealt with in more detail later in this paper.
8.3. Type of classi?cation
IAS 38 is a typical classi?cation schema with no
relation(s) between the classes, except from being
sub-classes. In the BSC there is an assumed time
relation between the classes. In that sense BSC is a
typology. But the linkage, in the BSC, of the per-
formance drivers in a cause–e?ect chain is an even
stronger claim for typology. At ?rst glance, IC
gives the appearance of a typology because of the
neatly related classes. Nonetheless, more careful
inspection indicates straightforward division in
which one class (sub-universe) is subtracted from
the other.
To construct a new classi?cation it seems plau-
sible to start with the more limited ambition of
developing a simple classi?cation schema as
opposed to a typology.
8.4. Exhaustive and exclusive
On the one hand, the classi?cation in IAS 38 is
exhaustive because of the general de?nition of
objects, states or events in the universe of dis-
course (future economic bene?ts). On the other,
the exclusion of certain intangibles, most of them
not speci?ed, makes it hard to see the classi?cation
as exhaustive. The BSC does not pretend to be
exhaustive. One of its basic tenets is to make a
selection of objects, states and events from the
universe. The IC cannot be fully exhaustive, which
Edvinsson and Malone (1997) also recognise.
When using a subtracting technique, something
must be left over if the order of subtracting should
matter (otherwise, the last item must be exactly
equal to the residual).
None of the three models classify an object,
state or event exclusively. The vagueness of the
language (de?nition) used in IAS 38 results in
ambiguous class borders. In the BSC the border-
lines between especially internal business processes
and innovation/improvement are unclear and per-
haps such ambiguity is necessary in the modern
organisation. The same problem arises in IC, par-
708 J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713
ticularly the borderline between human capital
and customer and structure capital. Structure
capital, as well as customer capital, must be some
sort of human (capital) artefact.
In comparison with exhaustive there are prob-
ably more problems connected with making a
classi?cation exclusive. However, if the ambition
is to create a stable background, an exclusive
classi?cation is extremely important.
8.5. Simplicity
Starting with notational simplicity, IAS 38
avoids the problem (just two classes) and therefore
is described as both objectively and subjectively
7
unclear. There is a lack of objective–notational
simplicity in BSC because of multi-dimensional
concepts that can be given several meanings.
Nevertheless, the subjective–notational simplicity
is higher in that the concepts are used in the
everyday business life. For IC, the relation is the
direct opposition. The objective–notational sim-
plicity is high because it tries to establish linked
concepts but the link goes through a ‘capital’
concept. To use the word capital for di?erent
assets interferes with the traditional theoretical
and practical use of the capital concept (as the
value of the resources lent to a company).
The IAS 38 ful?ls neither objective nor sub-
jective–logical simplicity. Some intangibles have
been excluded just by listing them (and some
unspeci?ed). Furthermore, it is not obvious why
certain expenses are to be classi?ed as costs,
deferred costs or capitalised costs.
As long as there is no overt success model or
production function, it is hard to see any objec-
tive–logical simplicity in the BSC except that of
time structure (past, present and future). However,
the association between these time frames is
unknown. There is an element of subjective–logi-
cal simplicity because the structure relates, directly
or indirectly, to the traditional accounting state-
ments and the issue of investment versus con-
sumption (production).
As stated when the universe of discourse was
analysed, there is no objective logic in the IC
model or, if there is objective logic, it is not
understandable. To append to the previous argu-
ment, the subtracting algorithm does not work in
the reverse. It is not possible to add the di?erent
components to what is called a capital concept.
Still, according to the business interest, there
seems to be a subjective–logical simplicity in IC.
Perhaps it is because the algorithm orders a group
of haphazardly arranged intangibles that give rise
to a feeling of ‘being in control’. Thus, while it
adds up, it does not make sense.
It must be remembered that all three models are
used in practice. Naturally, IAS 38 has to be used
for companies following the IASC standards. Yet,
a large number of companies use the BSC volun-
tarily. To my knowledge, only a few companies
have tried to adopt the IC model of Edvinsson and
Malone (1997). Mouritsen (1999) interprets the
use of intellectual capital
8
in Danish ?rms in terms
of mobilisation of company (intangible) resources.
Roberts (1999) discusses the IC concept either as
an object (set of tools) or as rhetoric. He argues
that there exists a mutual reinforcement between
the two. One problem, however, is that when
rhetoric is transformed to tools and practise, its
mobilising capacity is hindered and an expectation
gap is often created. If IC is just rhetoric for
(intangible) resource mobilisation, there is prob-
ably no business need for its classi?cation.
Johanson, Martensson, and Skoog (1999) found
that Swedish ?rms use IC as a set of measurement
tools in the process of constructing their own
business model or production function. It is prob-
able that the intense interest of Swedish ?rms in
the BSC model has the same roots. According to
Johanson et al. (1998), however, the initial enthu-
siasm has been considerably greater than the
practical development and use of BSC (and IC).
One explanation can be found in the shortcomings
of the models as classi?cations. Once the organi-
sations start with applications and practical use,
problems like overlapping classes becomes
obvious.
7
The subjective dimension of simplicity, i.e. perceived sim-
plicity, is based here on the author’s subjective interpretation if
nothing else is stated.
8
In this paper the concept ‘intellectual capital’ denotes dif-
ferent types of model for accounting and reporting.
J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713 709
9. Conclusions: towards a classi?cation of intan-
gibles
A reasonable assumption is that new forms of
intangibles will be constructed in the future, i.e.
the universe of discourse is under constant change.
Therefore, a classi?cation of all possible intangi-
bles does not seem to be a task within the limits of
realisation.
The value of classi?cation is, according to Rud-
ner (1966), associated with its ability to function
as a heuristic device, as a help construction for
interpretation of phenomena. In that respect,
?nancial accounting data do not necessarily need
to have value relevance as such. Their value is
connected to their ability to function as a back-
drop by which ongoing business events can be
interpreted. Accounting classi?cation becomes the
art of background design while accounting, to use a
metaphor, functions as an organisational devel-
oping bath.
To construct a classi?cation schema or typology
for ?nancial accounting is an open issue. For
management control, a typology is unavoidable;
otherwise, there is no business logic. However, this
is a task for each company, i.e. to construct its
own success model.
The universe of discourse in ?nancial accounting
can be restricted to accounting transactions. If the
objective is to create a stable background for eco-
nomic interpretations, there is probably no reason
to change or expand the universe because of
intangibles. What is needed is simply to reclassify
the transactions. The sometimes–necessary reva-
luation of transactions according to an accounting
rule (standard) is a di?erent matter. It does not
change the nature of the classi?cation, just the
value of what had already been classi?ed.
Given a universe of accounting transactions, the
main challenge is to ?nd appropriate attributes. To
discover attributes is a process of creativity as
opposed to a process of science. The ?nal choice
re?ects the social (programmatic) part of
accounting. An attribute organises the universe in
accordance with a social assumption.
9
To classify
intangibles according to an attribute re?ecting
employees’ welfare would probably lead to a dif-
ferent classi?cation schema than of shareholders’
welfare.
One possible approach towards choice of attri-
bute is empirical. Classi?cations that are already
part of practice are analysed and the underlying
attributes are revealed. Classi?cations based on
those attributes can be further developed (as a
task of creativity). The BSC as described here
would be one obvious starting point.
The other approach is theoretical. One possibi-
lity is to start from di?erent equity theories (Hen-
driksen & van Breda, 1992), i.e. to whom net
income belongs and how to show equity relation-
ships in ?nancial statements. In relation to the
proprietary theory the proprietor welfare is of
main interest. Concepts such as ‘shareholder
value’ and ‘added value’ become central and
intangibles have to be objecti?ed and producti?ed
into marketable values. The attribute has to clas-
sify transactions into di?erent intangible (and
tangible) objects, including patents and trade-
marks.
Under the entity theory, the welfare concepts
are hidden because the company is assumed to
have a separate existence. Value to the business
becomes the crucial value concept and intangibles
will be a part of the production function. A classi?-
cation would probably wind up as a set of such
business functions as R&D and employee training.
A more modern entity theory is the value co-
production theory (Ramirez, 1999). In this theory,
intangibles are expressed in terms of relations.
What is then needed is a classi?cation of relations
(between di?erent value co-producers), such as the
classi?cations (typologies) developed to describe
relation marketing.
In the enterprise theory the welfare of all (many)
stakeholders is considered. Consequently, several
value concepts have to be introduced. Intangibles
will be a function of the di?erent stakeholders’ wel-
fare function. To map and classify each stake-
holder’s welfare function is, if at all possible, an
enormous task and would probably result in com-
plexity rather than simplicity (a lesson learnt from
‘social accounting’).
9
For an overview of the sociological perspective of
accounting see Jones (1995) and Roslender (1992).
710 J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713
One aspect of the confusion about intangibles
may stem from the fact that we use di?erent equity
theories during the process of describing and
interpreting intangibles. Another aspect is the
transformation of intangibles into other forms
of intangibles. One such example is the BSC,
where one type of intangible (performance dri-
ver) is supposed to be linked to another type of
intangible (outcome). A third aspect of the
confusion can be caused by the ontological
assumption that it is possible to divide the uni-
verse into tangibles and intangibles. Don’t we
often materialise the immaterial (through mea-
surements) and assign immaterial attributes to the
material? For example, we materialise trademarks
through attributing speci?c expenditures to the
(immaterial) concept. Yet, even if we could sepa-
rate them, many objects, states or events would
constitute di?erent combinations of tangibles and
intangibles. The conclusion must be to not take for
granted the division of the universe into tangibles
and intangibles. Depending on perspective, pur-
pose and type of ?nancial accounting model other
concepts may do a better job in classifying the
accounting world.
A ‘good classi?cation’ is a set of requirements; it
is something that we seldom, if ever, can fully
attain in practice, however. Presently, very little
can be said about its logical properties because
we have no fully articulated classi?cation. If the
universe consists of accounting transactions, then,
per de?nition, we have to ?nd an attribute that
makes the classi?cation exhaustive. To be exhaus-
tive, i.e. to classify all objects, states or events in
the universe of discourse, accounting often ends
up with a residual class (a non-class). That creates
a problem with exclusive classes, but it also hinders
the ambition to create a typology. The relation
between the residual class and the others becomes
unclear. We, therefore, have to settle for classi?-
cational schema with less information content.
The main problem will be to make it exclusive,
which is necessary if ‘rigid uniformity’ is to be a
means to create a stable background design. Con-
sistency is thought to be one requirement for the
basic accounting assumption of comparability. To
be consistent exclusive classes are likely necessary,
an assumption implying very precise de?nitions of
the classes. Nevertheless, the inherent vagueness
of language will always be an obstacle. In con-
trast to ?nite uniformity (there are relevant cir-
cumstances for similar transactions), exclusive
classes are also an argument for rigid uni-
formity (same classi?cation in di?erent organi-
sations without considering circumstances). If
there are relevant circumstances, they ought to
be covered by the classification attribute(s).
Consequently, there can be no relevant circum-
stances. Exclusive classes solve, in an accounting
sense, the problem of separation because logical
division is a way to keep things apart. It does not
matter if we deal with intangible assets or intangi-
ble expenses.
The simplicity concept was discussed in terms of
four dimensions. Except from technical/logical
aspects of simplicity, the important point here
concerns the concept of ‘usefulness’, i.e. how the
users perceive a classi?cation. In classi?cation
theory, simplicity has no reference to faithful
representation, which is often the strongest
argument against rigid uniformity and standar-
disation. Notwithstanding, looking upon ?nan-
cial accounting data as a background for
interpretation, the question arises regarding the
extent with which the interpreter is able to use
changing backgrounds. (For management control
purposes, it is probably necessary to design com-
pany-speci?c backgrounds, i.e. associations
between qualitative variables and organisational
success.)
The objective and subjective degree of simplicity
can be increased by using corresponding concepts
in the income statement and balance sheet to a
greater extent than are used today (e.g. R&D costs
and capitalised R&D). Finally, to judge the logical
simplicity a more developed classi?cation is needed.
Here we end with a set of ideas on how to
approach the construction phase of classi?cations.
It is about how we order the world, and account-
ing is expected to do the job.
J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713 711
Appendix A. Condensed description of IAS 38,
BSC and IC
IAS 38 (p. 3)
2. An intangible asset is an identi?able non-
monetary asset without physical substance held
for use in production or supply of goods or ser-
vices, for rental to others, or for administrative
purposes. An asset is a resource: (a) controlled by
an enterprise as a result of past events; and (b)
from which future economic bene?ts are expected
to ?ow to the enterprise.
3. IAS 38 requires an enterprise to recognise an
intangible asset (at cost) if, and only if (a) it is
probable that the future economic bene?ts that are
attributable to the asset will ?ow to the enterprise;
and (b) the cost of the asset can be measured
reliably.
4. IAS 38 speci?es that internally generated
goodwill, brands, mastheads, publishing titles,
customer lists and items similar in substance
should not be recognised as assets.
BSC
The basic idea of BSC is to translate strategy
into action. The starting point is the business
vision and strategy. Through a top-down
approach, this strategy is supposed to become
operationalised in four dimensions or perspectives:
?nancial, customer, internal processes and learn-
ing and growth. In each perspective a strategy is
formulated and key ratios (performance drivers)
expressing that strategy are constructed. The dif-
ferent key ratios are subsequently linked to each
other in a cause–e?ect chain, ending in a ?nancial
outcome.
IC
In the Edvinsson and Malone model, the start-
ing point is the market value of a ?rm. The di?er-
ence between the market and book value of the
?rm is denoted ‘intellectual capital’. Intellectual
capital is divided into human and structure capital.
Structure capital is further divided into customer
and organisational capital, and organisational
capital into innovation and process capital.
Finally, innovation capital is divided into intellec-
tual property and intangible assets.
References
AAA (1977), Report on the American Accounting Associ-
ation Committee on International Accounting Operations
and Education 1975–1976. The Accounting Review. 52,
Supplement.
Ahonen, G. (1998) Orally expressed at a meeting in Brussels,
14–15 September 1998.
Acko?, R.L. (1968). Scienti?c method—optimizing applied
research decisions, John Wiley.
Baudrillard, J. (1988). Selected writings. Stanford: Stanford
University Press.
Capon, N., Farley, J.U. & Hoenig, S. (1996). Toward an inte-
grative explanation of corporate ?nancial performance,
Kluwer Academic Publishers.
Czarniawska-Joerges, B. (1993). The three-dimensional organi-
zation—a constructionist view. Lund: Studentlitteratur.
Catasus, B. (1999). The labour of division—classi?cations in the
making, School of Business, Stockholm University, (unpub-
lished).
Clement, R. (1997). Intangible investment from an evolutionary
perspective. Chapter 2, OECD (forthcoming).
Cooper, R. (1989). The visibility of social systems. In
M. C. Jackson, P. Keys, & A. Croppers (Eds.), Operational
research and social sciences. New York: Plenum.
Edvinsson, L., & Malone, M. S. (1997). Intellectual capital.
London: Piatkus.
Gold?nger, C. (1997). Intangible economy and its implications
for statistics and statisticians. International Statistical
Review, 65(2), 191–200.
Gro¨ jer, J-E. (1997). Employee artefacts on the balance sheet: a
model illustration and implications. Journal of Human
Resource Costing and Accounting, 2(1), 27–53.
Hansson, H., & Karlsson, S.-G. (1974). bild & form, (Eng.
picture & form). Skolfo¨ rlaget: Uddevalla.
HEC. (2000). Classi?cation of intangibles. Chambre de
Commerce et d’Industrie de Paris.
Hedberg, B., & Jo¨ nsson, S. (1978). Designing semi-confusing
information systems for organizations in changing environ-
ment. Accounting, Organisations and Society, 3, 47–64.
Hegglo¨ v, E. (1998) Redovisning som hinder fo¨r la¨rande (Eng.
Accounting that hinders learning), (unpublished).
Hempel, C. (1952). Fundamentals of concept formation in
empirical science. The University of Chicago Press.
Hendriksen, E. S., & van Breda, M. F. (1992). Accounting
theory, ?fth edition. Homewood, Il: Irwin.
Hines, R. (1988). Financial accounting: In communicating
reality, we construct reality. Accounting, Organizations and
Society, 13(3), 251–261.
712 J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713
Hoegh-Krohn, N. E. J., & Knivs?a? , K. H. (1998). Accounting
for intangible assets: challenges and some solutions. Bergen:
Foundation for Research in Economics and Business
Administration.
Hopwood, A. G. (1987). The archaeology of accounting sys-
tems. Accounting, Organizations and Society, 12(3), 207–234.
IASC. (1996). International accounting standards. London:
IASC.
IASC. (1998). International accounting standard IAS 38, intan-
gible assets. London: IASC.
Johanson, U., Eklo¨ v, G., Holmgren, M. & Ma? rtensson, M.
(1998). Human resource costing and accounting versus the
balanced score card. (A report to OECD.) School of Business,
Stockholm University.
Johanson, U., Ma? rtensson, M., & Skoog, M. (1999). Measuring
to understand intangible performance drivers. Stockholm
University: School of Business.
Jones, T. C. (1995). Accounting and the enterprise. A social
analysis. London: Routledge.
Kaplan, R. S., & Norton, D. P. (1996). Translating strategy into
action: The balanced scorecard. Boston: Harvard Business
School Press.
Lerner, D., Lasswell, H. D. (Eds.). (1951). The policy sciences—
recent developments in scope and method. Stanford: Stanford
University Press.
Lev, B., & Zarowin, P. (1998) The boundaries of ?nancial
reporting and how to extend them, New York University,
working paper.
Mauritsen, J. (1999). Intellectual capital: The logic of an ‘econ-
omy of creativity’, Copenhagen Business School.
Nobes, C. W. (1984). International classi?cation of ?nancial
reporting. London: Croom Helm.
No¨ rreklit, L. (undated), The ?rm and its resources, (unpublished).
O’Shaughnessy, J. (1972). Inquiry and decision. Oxford: George
Allen & Unwin.
Ramirez, R. (1999). Value co-production: intellectual origins
and implications for practice and research. Strategic
Management Journal, 20, 49–65.
Roberts, A. (1995). The very idea of classi?cation in interna-
tional accounting. Accounting, Organizations and Society,
20(7), 639–664.
Roberts, H. (1999). Classi?cation of intellectual capital.
Norwegian School of Management, (unpublished).
Rose, N., & Miller, P. (1992). Political power beyond the state:
problematics of government. British Journal of Sociology,
43(2), 173–205.
Roslender, R. (1992). Sociological perspectives on modern
accountancy. London: Routledge.
Rudner, R. (1966). Philosophy of social science. Englewood
Cli?s, N.J: Prentice-Hall.
Statistics Netherlands. (1998). Intangible investments: de?nition
and data source for technological, marketing, IT and organi-
sational activities and rights. Statistics Netherlands (LNM-
reeks 9803).
Stolowy, H., & Jeny, A. (1999). How accounting standards
approach and classify intangibles—an international survey,
HEC School of Management (unpublished).
Wolk, H.I., & Tearney, M.G. (1997). Accounting theory—a
conceptual and institutional approach (4th ed.). South-
Western College Publishing.
J.-E. Gro¨jer / Accounting, Organizations and Society 26 (2001) 695–713 713
doc_142459206.pdf