Ontologies and knowledge bases: towards a terminological clarification

Full text

Ontologies and Knowledge Bases
Towards a Terminological Clarification
Nicola Guarino
National Research Council, LADSEB-CNR
Corso Stati Uniti 4, I-35129 Padova, Italy
guarino@ladseb.pd.cnr.it
Pierdaniele Giaretta
Institute of History of Philosophy, University of Padova
Piazza Capitaniato 3, I-35100 Padova, Italy
ABSTRACT
The word “ontology” has recently gained a good
popularity within the knowledge engineering com-
munity. However, its meaning tends to remain
a bit vague, as the term is used in very differ-
ent ways. Limiting our attention to the various
proposals made in the current debate in AI, we
isolate a number of interpretations, which in our
opinion deserve a suitable clarification. We elu-
cidate the implications of such various interpre-
tations, arguing for the need of clear terminolog-
ical choices regarding the technical use of terms
like “ontology”, “conceptualization” and “onto-
logical commitment”. After some comments on
the use “Ontology” (with the capital “o”) as a
term which denotes a philosophical discipline, we
analyse the possible confusion between an ontol-
ogy intended as a particular conceptual framework
at the knowledge level and an ontology intended
as a concrete artifact at the symbol level, to be
used for a given purpose. A crucial point in this
clarification effort is the careful analysis of Gru-
ber’ s definition of an ontology as a specification
of a conceptualization.
1 Introduction
1
The word “ontology” has recently gained a good pop-
ularity within the knowledge engineering community,
especially in relation with the recent ARPA knowledge
sharing initiative [14, 11, 13, 4, 6, 7, 12]. However, its
meaning tends to remain a bit vague, as the term is
used in very different ways. Limiting our attention
to the various proposals made in the current debate
within the knowledge sharing community, we can iso-
late the different interpretations reported in Fig. 1 be-
low, which in our opinion deserve a suitable clarifica-
tion.
1
Slightly amended version of a paper appeared in N.J.I.
Mars (ed.), Towards Very Large Knowledge Bases, IOS
Press 1995, Amsterdam
1. Ontology as a philosophical discipline
2. Ontology as a an informal conceptual system
3. Ontology as a formal semantic account
4. Ontology as a specification of a “conceptualization”
5. Ontology as a representation of a conceptual system
via a logical theory
5.1 characterized by specific formal properties
5.2 characterized only by its specific purposes
6. Ontology as the vocabulary used by a logical theory
7. Ontology as a (meta-level) specification of a logical
theory
Figure 1: Possible interpretations of the term “ontol-
ogy”.
The interpretation 1 is radically different from all
the others, and its implications are discussed in the
next section. The current debate regards the interpre-
tations 2-7: 2 and 3 conceive an ontology as a con-
ceptual “semantic” entity, either formal or informal,
while according to the interpretations 5-7 an ontology
is a specific “syntactic” object. The interpretation 4,
which has been recently proposed as a definition of
what an ontology is for the AI community [4, 5], is
one of the more problematic, and it will be discussed
in detail in the present paper. It may be classified
as “syntactic” but its precise meaning depends on the
understanding of the terms “specification” and “con-
ceptualization”.
According to interpretation 2, an ontology is the
(unspecified) conceptual system which we may assume
to underly a particular knowledge base. Under inter-
pretation 3, instead, the “ontology” which underlies a
knowledge base is expressed in terms of suitable for-
mal structures at the semantic level, like for instance
those described in [8, 17].In both cases, we may say

“the ontology of KB1 is different from that of KB2”.
Under interpretation 5, an ontology is nothing else
than a logical theory. The issue is whether such a
theory needs to have particular formal properties in
order to be an ontology (for instance, we may impose
it must be a “Tbox”) or, rather, it is the intended
purpose which lets us consider a logical theory as an
ontology. The latter position is being supported for in-
stance by Pat Hayes, which in recent e-mail discussions
argued that an ontology is an annotated and indexed
set of formal assertions about something: “leaving off
the annotations and indexing, this is a collection of
assertions: what in logic is called a theory”.
According to interpretation 6, an ontology is not
viewed as a logical theory, but just as the vocabulary
used by a logical theory. Such an interpretation col-
lapses into 5.1 if an ontology is thought of as a spec-
ification of a vocabulary consisting of a set of logical
definitions. We may anticipate that interpretation 4
collapses into 5.1, too, when a conceptualization is in-
tended as a vocabulary; we shall see however that the
problem is how to make clear the meaning of the term
“conceptualization”.
Finally, under interpretation 7, an ontology is seen
as a (meta-level) specification of a logical theory, in the
sense that it specifies the “architectural components”
(or “primitives”) used within a particular domain the-
ory. This point of view is maintained in [18] and, in
slightly different form, in [10]. Wielinga and colleagues
argue that it is the ontology which specifies, for a the-
ory where some formulas have the form of mathemati-
cal constraints, what a constraint is and how it differs
from a formula of another kind; Mark argues that an
ontology is “a representation of components and their
allowed interactions, with the purpose of providing an
explicit framework in which to elaborate the rest of the
system...”
We shall try to elucidate in this paper the impli-
cations of such various interpretations, arguing for the
need of clear terminological choices regarding the tech-
nical use of terms like “ontology”, “conceptualization”
and “ontological commitment” within the knowledge
engineering community. First we propose to use “On-
tology” (with the capital “o”) as a term denoting a
philosophical discipline, then we analyse a number of
possible senses of the term “ontology” (with the lower-
case “o”) where the term is somehow related to specific
knowledge bases (or logical theories) designed with the
purpose of expressing shared (or sharable) knowledge.
A starting point in this clarification effort will be
the careful analysis of the interpretation 4 adopted by
Gruber. The main problem with such an interpreta-
tion is that it is based on a notion of conceptualization
(introduced in [3]) which doesn’t fit our intuitions, as
has been noticed in [8]: according to Genesereth and
Nilsson, a conceptualization is a set of extensional rela-
tions describing a particular state of affairs, while the
notion we have in mind is an intensional one, namely
something like a conceptual grid which we superim-
pose to various possible states of affairs. We propose
in this paper a revised definition of a conceptualization
which captures this intensional aspect, while allowing
us to give a satisfactory interpretation to Gruber’s def-
inition.
2 Ontology and Ontologies
The first important distinction in the list of interpre-
tations given in the previous section is between inter-
pretation 1 and all the others. We stipulate that when
we refer to an ontology (with the indeterminate arti-
cle and the lowercase initial) we refer to a particular
determinate object (whose nature may vary in depen-
dence of the choice among interpretations 2-7), while
speaking of Ontology (without the indeterminate arti-
cle and with the uppercase initial) we refer to a philo-
sophical discipline, namely that branch of philosophy
which deals with the nature and the organisation of
reality. Ontology as such is usually contrasted with
Epistemology, which deals with the nature and sources
of our knowledge
2
.
Aristotle defined Ontology as the science of being as
such: unlike the special sciences, each of which inves-
tigates a class of beings and their determinations, On-
tology regards “all the species of being qua being and
the attributes which belong to it qua being” (Aristo-
tle, Metaphysics, IV, 1). In this sense Ontology tries to
answer to the question: What is being? or, in a mean-
ingful reformulation: What are the features common to
all beings?
This is what nowadays one would call General On-
tology, in contrast with the various Special or Regional
Ontologies (of the Biological, the Social, etc.). This
distinction corresponds to the Husserlian one between
Formal Ontology and Material Ontology [1]. But the
Husserlian notion of “formal” does not involve only
generality. For Husserl, the task of Formal Ontology is
to determinate the conditions of the possibility of the
object in general and the individuation of the require-
ments that every object’s constitution has to satisfy.
Recently, Nino Cocchiarella defined Formal Ontol-
ogy as the systematic, formal, axiomatic development
of the logic of all forms and modes of being [2]. The
connection of Cocchiarella’s definition with the Husser-
lian notion is not clear, and, in general, the genuine
interpretation of the term “formal ontology” is still a
matter of debate [16]. However, Cocchiarella’s defini-
tion is in our opinion particularly pregnant, as it takes
into account both meanings of the adjective “formal”:
on one side, this is synonymous of “rigorous”, while
on the other side it means “related to the forms of be-
ing”. Therefore, what Formal Ontology is concerned
in is not so much the bare existence of certain objects,
but rather the rigorous description of their forms of be-
ing, i.e. their structural features. In practice, Formal
Ontology can be intended as the theory of the distinc-
tions, which can be applied independently of the state
of the world, i.e. the distinctions:
2
This definition of “epistemology” is taken from
Shapiro’s “Encyclopedia of Artificial Intelligence” [15]. Re-
grettably, the entry “ontology” does not appear there. The
philosophical community prefers to use the term “theory of
knowledge” for what is here called “epistemology”.

• among the entities of the world (physical objects,
events, regions, quantities of matter...);
• among the meta-level categories used to model
the world (concept, property, quality, state, role,
part...).
In this sense, Formal Ontology, as a discipline, may
be relevant to both Knowledge Representation and
Knowledge Acquisition [7].
3 Kinds of Ontologies
Let us now refine the technical meaning of the word
“ontology” when — within the knowledge engineer-
ing community – it is used to denote a particular ob-
ject rather than a discipline. Here a possible confusion
arises between an ontology intended as a particular
conceptual framework at the semantic level (interpre-
tations 2-3) and an ontology intended as a concrete
artifact at the syntactic level, to be used for a given
purpose (interpretations 4-7). This is an important
distinction, and it is evident that we cannot use the
same technical term to denote both things. In the cur-
rent practice, however, the term “ontology” is used
ambiguously with both meanings, either to refer to
(various kinds of) symbol-level artifacts, or to their
conceptual (or semantical) counterparts
3
. Therefore,
rather than insisting on a unique precise meaning for
such a term, what we propose is to adopt different tech-
nical terms to refer explicitly to the two levels, while
tolerating an ambiguity in the interpretation of the
term “ontology” (with the lowercase initial). We shall
use the term conceptualization to denote a semantic
structure which reflects a particular conceptual sys-
tem (interpretation 3 in Fig. 1), and ontological theory
to denote a logical theory intended to express onto-
logical knowledge (interpretation 5). The underlying
intuition is that ontological theories are designed arti-
facts, knowledge bases of a special kind which can be
read, sold or physically shared. Conceptualizations,
on the other hand, are the semantical counterpart of
ontological theories. The same ontological theory may
commit to different conceptualizations, as well as the
same conceptualization may underlie different ontolog-
ical theories. The term “ontology” will be used am-
biguously, either as synonym of “ontological theory”
or as synonym of “conceptualization”. We need only
to be consistent to the choice made within the same
statement.
The details of the definitions mentioned above are
the subject of the subsequent sections; for the time
being, the meaning of statements like those listed in
Fig 2 should however be clear enough under the as-
sumptions we have made. In 1-4, the term ”ontology”
has a clear syntactic interpretations; the interpretation
of statement 5 will be discussed later.
4 Kinds of Conceptualizations
Let us notice first that the use of the term “ontology”
as related to an ontological theory is compatible with
3
The most common use is however the former one.
1. Ontological engineering is a branch of knowledge en-
gineering which uses Ontology to build ontologies.
2. Ontologies are special kinds of knowledge bases.
3. Any ontology has its underlying conceptualization.
4. The same conceptualization may underlie different
ontologies.
5. Two different knowledge bases may commit to the
same ontology.
Figure 2: Different statements making use of the term
“ontology”.
Tom Gruber’s definition of an ontology as “an explicit
specification of a conceptualization”, since it should be
clear that an “explicit” object is a concrete, symbol-
level object. The problem with Gruber’s definition,
however, is that it relies on an extensional notion of
“conceptualization” [3] which, while being compatible
with the preliminary characterization given in the pre-
vious section, does not fit our purposes of defining what
an ontology is. We have already pointed to this prob-
lem in [9] ; we shall discuss it here in detail, proposing
an alternative, intensional definition of “conceptualiza-
tion” which satisfies our needs.
Let us consider the example given by Genesereth and
Nilsson. They take into account a situation where two
piles of blocks are resting on a table (Fig 3). Accord-
ing to the authors, a possible conceptualization of this
scene is given by the following structure:
h{a, b, c, d, e}, {on, above, clear, table}i
where {a, b, c, d, e} is a set called the universe of dis-
course, consisting of the five blocks we are interested
in, and {on, above, clear, table} is the set of the rele-
vant relations among these blocks, of which the first
two, on and above, are binary and the other two, clear
and table, are unary
4
. The authors make clear that
objects and relations are extensional entities. For in-
stance, the table relation, which is understood as hold-
ing of a block if and only if that block is resting on the
table, is but the set {c, e}. It is exactly such an exten-
sional interpretation which originates our troubles.
Let us notice first that the authors used natural lan-
guage terms (like on, above) in the metalanguage cho-
sen to describe a conceptualization. This could per-
haps be seen as nothing more than a didactical device.
But such linguistic terms do convey essential informa-
tion in order to understand the criteria used to consider
some sets of tuples as the relevant relations. Such an
extra information cannot be accounted for by the con-
ceptualization itself.
Referring to the example given, consider a different
arrangement of blocks, where c is on the top of d, while
4
In the original example also a function is considered,
but for simplicity reasons we omit here to mention func-
tions as a further element in the characterization of a
conceptualization.

c
b
a
e
d
Figure 3: Blocks on a table. From [3].
a and b together form a separate stack standing on the
table (Fig. 4). The corresponding structure would be
different from the previous one, generating therefore a
different conceptualization. Of course there is nothing
wrong in such a view, if one is only interested in iso-
lated snapshots of the block world. But the meanings
of the terms used to denote the relevant relations are
still the same, since they are invariant with respect to
the possible configurations of blocks. In fact, in the
metalanguage adopted in their book, Genesereth and
Nilsson would use the same terms (on, above, clear, ta-
ble) to denote the new conceptualization. We prefer to
say in this case that the states of affairs are different,
but the conceptualization is the same. The structure
proposed by Genesereth and Nilsson seems to be more
apt to represent a state of affairs rather than a con-
ceptualization.
In order to capture such intuitions, the linguistic
terms we have used to denote the relevant relations
cannot be thought of as mere comments, informal
extra-information. Rather, the formal structure used
for a conceptualization should somehow account for
their meaning. As the logico-philosophical literature
teaches us, such a meaning cannot coincide with an
extensional relation.
Sticking to a set-theoretical framework, a standard
way to approximate such meaning is to conceive it as
an intension (intensional relation), taking inspiration
from Montague semantics. This means that a sin-
gle extensional relation is always relative to a possible
world
5
.
Formally, an intensional relation of arity n on a do-
main D is a function from a set W of possible worlds
to the set 2
D
n
of all possible n-ary relations on D.
Such a function specifies a set of admissible extensions,
relative to the domain and the set of possible worlds
considered. This means that not only the extension in
the actual world, but also those relative to the other
possible worlds are specified. We can therefore repre-
sent a conceptualization by the following intensional
5
Roughly, we can think of possible worlds like states of
affairs or situations.
b
a
e
d
c
Figure 4: A different arrangement of blocks. A differ-
ent conceptualization?
structure:
hW, D, Ri
where W is a set of possible worlds, D is a domain of
objects, and R is a set of intensional relations on D.
According to this intensional interpretation, a con-
ceptualization accounts for the intended meanings of
the terms used to denote the relevant relations. Such
meanings are supposed to remain the same if the ac-
tual extensions of the relations change due to different
states of affairs. This means that, for instance, the ac-
tual extensions of the relation on in the two examples
of Fig. 3 and 4 belong to the image of the same inten-
sional relation, applied to different worlds. Intuitively,
we can see a conceptualization as given by a set of rules
constraining the structure of a piece of reality, which
an agent uses in order to isolate and organize relevant
objects and relevant relations: the rules which tell us
whether a certain block is on another one remain the
same, independently of the particular arrangement of
the blocks. These rules can be viewed as conceptual
links which put together different extensions belonging
to the same intensional relation.
Notice that, given a set of relevant relations specified
by linguistic terms like those of our example, there
will be in general many conceptualizations of the form
given above which satisfy the natural constraints we
can attach to the meaning of such expressions. As
shown in [8] , a convenient modal theory can be used to
give an approximate characterization of such intended
meaning, with the aim of excluding deviant extensions.
For example, we can express the intuitive constraint
that a tuple like < a, a > should never belong to the
extension of the relation specified by the word “on” by
stating
2 ∀x.¬on(x, x)
Another interesting constraint which may be useful
to characterize a unary relation like block (not men-
tioned by Genesereth and Nilsson) is that such a re-
lation can be never “lost” by its instances, i.e. if it

holds of an object, it holds of that object in all possi-
ble worlds:
2 (∀x.block(x) ⊃ 2 block (x))
Such a constraint has been called “ontological rigidity”
in [8], and has been used to discriminate among various
ontological categories of unary relations.
A set of formal constraints like those above, ex-
pressed in a suitable modal language, can therefore be
used to (partially) characterize a conceptualization, in
the sense of excluding unintended extensions of the
relevant relations even for possible “worlds” different
from the one considered. Notice that in general we
cannot identify a single conceptualization by means of
a set of formal constraints, since such a set may have
many models. The set of such models is exactly what
in [9] we defined as ontological commitment
6
.
According to these considerations, we cannot see a
particular theory as a specification of a conceptual-
ization, since conceptualizations can be only partially
characterized. What we can specify is a set of concep-
tualizations, i.e. an ontological commitment.
5 A Simple Example
Having discussed in detail the various implications un-
derlying the notion of conceptualization, let us now
use a simple example to see how such a semantic notion
can be related to syntactic objects like logical theories.
Consider the following logical theory:
T 1 :
∀x.apple(x) ⊃ fruit(x).
∀x.pear(x) ⊃ fruit(x).
apple(a1).
red(a1)
If we want to isolate the ontological content of such
a theory, we can try to individuate, among its axioms,
those which we consider to be more strictly related to
the intrinsic intended meanings of the predicates used
in the language. For example, the following axioms
(which are usually related to what is called the Tbox)
may be intended as capturing part of the meaning of
apple, pear and fruit :
T 2 :
∀x.apple(x) ⊃ fruit(x).
∀x.pear(x) ⊃ fruit(x).
We shall call a set of such axioms an ontological the-
ory. An ontological theory contains formulas which are
considered to be always true (and therefore sharable
among multiple agents), independently of particular
6
In that paper we did not introduce intensions as ingre-
dients of our semantical structures, adopting instead stan-
dard modal models. Here we choose a different approach
which seems to be better suited to the perspective we want
to present.
states of affairs. Formally, we can say that such for-
mulas must be true in every possible world.
An ontological theory like the one above character-
izes very roughly the ontological content of the theory
from which it is extracted. To better grasp such a
content, we should look at the intended conceptual-
ization underlying both T1 and T2, which models (in
a much finer way) the ontologically relevant aspects
of the language used by our initial theory. According
to the discussion made in the previous section, such a
conceptualization can be characterized (in an approx-
imate way) by a suitable modal theory T3. The for-
mulas (theorems) of T2 will be true in every possible
world belonging to the intended conceptualization, and
therefore will appear as necessary formulas in T3; fur-
thermore, T3 may contain other formulas capturing
necessary facts not captured by T2. For the present
example, we choose a very simple theory like the fol-
lowing:
T 3 :
2(∀x.apple(x) ⊃ fruit(x)).
2(∀x.pear(x) ⊃ fruit(x)).
2(∀x.apple(x) ⊃ 2 apple(x)).
2(∀x.pear(x) ⊃ 2 pear(x)).
2(∀x.fruit(x) ⊃ 2 fruit(x)).
¬2(∀x.red(x) ⊃ 2 red(x)).
Such a theory expresses some very general con-
straints on the meaning of our predicates, namely the
fact that apple, pear and fruit form a hierarchy, and
that they are “rigid”, differently from red. We say that
T3 is the specification of the ontological commitment
of T1.
Notice that the same information carried by T3 can
be expressed by a meta-level theory, whose domain
is given by the nonlogical symbols used in T1. For
instance, we can write:
T 4 :
apple ≤ fruit.
pear ≤ fruit.
rigid(apple).
rigid(pear).
rigid(fruit).
¬rigid(red).
Such a theory can be usefully adopted as an alter-
nate specification of an ontological commitment, as-
suming of course that the meaning of predicates like ≤
and rigid is such that T4 can be immediately converted
into T3 by means of suitable translation rules.
6 What Is An Ontology
Let us now go back to our original problem of clarifying
the meaning of “ontology”. Our goal is to propose a
choice among the interpretations 2-7 of Fig. 1, and to
give a precise sense to at least some of the statements

listed in Fig 2. In the light of the discussion developed
so far, we shall restrict our choice to three possible
technical senses of the word “ontology”.
In sense a), “ontology” is a synonym of “ontological
theory”. In this case statements 1-4 in Fig 2 have a
unique interpretation, while statement 5 means that
the two knowledge bases may have a common subthe-
ory, which is an ontological theory. This choice is con-
sistent with interpretation 5 in Fig. 1. As discussed
in the previous section, an ontological theory differs
from an arbitrary logical theory (or knowledge base)
by its semantics, since all its axioms must be true in ev-
ery possible world of the underlying conceptualization.
This means that while an arbitrary logical theory (con-
taining for instance a statement like apple(a)∨pear(a),
expressing uncertainty about the object a) may repre-
sent a particular epistemic state, an ontological the-
ory can be only used to represent common knowledge
independent from single epistemic states. Due to this
formal difference between an ontological theory and an
arbitrary logical theory, interpretation 5.2 is therefore
discarded in favour of 5.1. T2 is an ontology according
to such an interpretation.
In sense b), “ontology” is a synonym of “specifica-
tion of an ontological commitment”. This choice is
still consistent with interpretation 5.1. In this case,
statements 1-4 still get a unique meaning, while state-
ment 5 has no sense, and it should be substituted by
“The ontological commitment of two different knowl-
edge bases may be specified by the same theory”. T3 is
an ontology according to this interpretation. The lan-
guage used by T3 is in general richer than the one used
by T1: as discussed in [8], the purposes are different,
since the purpose of T3 is to convey meaning by using
a very expressive language, while the language of T1
is the result of a tradeoff choice between expressivity
and computational efficiency. Notice that T3 is an on-
tological theory like T2, since its formulas are always
true.
In sense c), “ontology” is a synonym of “conceptual-
ization”. This choice is consistent with interpretation
3 in Fig. 1. In this case statements 1-4 in Fig 2 have
no sense, while the occurrence of “ontology” in state-
ment 5 gets a semantic interpretation. In this case,
statement 5 is equivalent to “Two different knowledge
bases may have the same conceptualization”. None of
the theories shown in the previous section is an ontol-
ogy according to this choice.
Let us now see what the meaning of Gruber’s defini-
tion “an ontology is a specification of a conceptualiza-
tion” may be. First of all, it is evident that sense c) is
incompatible with such a definition. Since we believe
we have good reasons to keep the latter, we suggest to
avoid the use of “ontology” in a semantic sense unless
it is clear from the context.
Let us now consider senses a) and b), which assign
the tag ”ontology” to T2 and T3, respectively. Strictly
speaking, none of them can be considered as a specifi-
cation of a conceptualization, and hence Gruber’s defi-
nition cannot apply. If we want to mantain its original
(good) intuitions, we must weaken Gruber’s definition,
claiming that an ontology is only a partial account of a
conceptualization. According to this choice, both T2
and T3 may be called “ontologies”.
In fact, such a weakened definition leaves space both
for senses a) and b), and this is exactly what we
want: the degree of specification of the conceptualiza-
tion which underlies the language used by a particular
knowledge base varies in dependence of our purposes:
an ontology of kind b) gets closer to specifying the in-
dended conceptualization (and therefore may be used
to establish consensus about the utility of sharing a
particular knowledge base), but it pays the price of a
richer language (and therefore, in general, undecidabile
and inefficient). An ontology of kind a), on the other
side, is developed with particular inferences in mind,
designed to be shared among users which already agree
on the underlying conceptualization.
There are still a couple of senses of “ontology”,
among those reported in Fig. 1, which are to be dis-
cussed, namely senses 6 and 7. The approach which
seems to adopt such interpretations is the one followed
in the KAKTUS project [18]; here an ontology is de-
fined as “a metalevel viewpoint on a set of possible
domain theories”. In general, such a viewpoint is a set
of metalevel definitions of the syntactic categories used
in a knowledge base. The form of such definitions is
not clear. What is interesting is that the description
of a particular knowledge base according to such met-
alevel categories may have the form of a theory like
T4. There is however an important difference: T4 uses
meta-level semantic categories, defined in the language
of T3, while Wielinga and Schreiber want to avoid any
explicit semantic notion.
In conclusion, we hope to have given a clarification of
the notion of “ontology” based on a notion of “concep-
tualization” defined in a rigorous semantic way; such
a framework allowed us to underline the difference be-
tween an ontology and an arbitrary knowledge base,
and to distinguish among various senses of “ontology”
used in the current debate.
7 A Simple Glossary
We report below the informal definitions which we sug-
gest to use as the preferred interpretations of the terms
discussed in the present paper.
conceptualization: an intensional semantic struc-
ture which encodes the implicit rules constraining
the structure of a piece of reality.
Formal Ontology: the systematic, formal, ax-
iomatic development of the logic of all forms and
modes of being.
ontological commitment: a partial semantic ac-
count of the intended conceptualization of a logical
theory.
ontological engineering: the branch of knowledge
engineering which exploits the principles of (formal)
Ontology to build ontologies.
ontological theory: a set of formulas intended to be
always true according to a certain conceptualization.

Ontology: that branch of philosophy which deals
with the nature and the organisation of reality.
ontology: (sense 1) a logical theory which gives an ex-
plicit, partial account of a conceptualization; (sense
2) synonym of conceptualization.
ACKNOWLEDGMENT
This work has been made within the CNR project
“Ontological and Linguistic Tools for Conceptual Mod-
elling”. We are grateful to Mike Uschold, Massimiliano
Carrara and Alessandro Artale for their contribute to
the final version of this paper.
References
[1] Bunge, M. Treatise on basic philosophy. Ontology
I: the furniture of the world, Dordrecht, Reidel,
1977.
[2] Cocchiarella, Nino B. Formal Ontology In
Burkhardt, H. and Smith, B. (eds.), Handbook
of Metaphysics and Ontology. Philosophia Verlag,
Munich, pp. 640-647, 1991.
[3] Genesereth, Michael R. and Nilsson, N. J. Log-
ical Foundation of Artificial Intelligence Morgan
Kaufmann, Los Altos, California, 1987.
[4] Gruber, Thomas R. A translation approach to
portable ontology specifications. In Knowledge
Acquisition, vol. 5, pp. 199-220, 1993.
[5] Gruber, Thomas R. Toward Principles for the De-
sign of Ontologies Used for Knowledge Sharing In
Guarino, N. and Poli, R. (eds.), Formal Ontology
in Conceptual Analysis and Knowledge Represen-
tation (to appear).
[6] Guarino, Nicola and Poli, Roberto. In Guarino, N.
and Poli, R. (eds.), Formal Ontology in Concep-
tual Analysis and Knowledge Representation (to
appear).
[7] Guarino, Nicola. Formal Ontology, Knowledge
Acquisition and Knowledge Representation. In
Guarino, N. and Poli, R. (eds.), Formal Ontology
in Conceptual Analysis and Knowledge Represen-
tation (to appear).
[8] Guarino, Nicola, Carrara, Massimiliano, and Gi-
aretta, Pierdaniele Formalizing Ontological Com-
mitment. In Proc. of the National Conference on
Artificial Intelligence (AAAI-94), Seattle, Morgan
Kaufmann, 1994.
[9] Guarino, N. and Carrara, M. and Giaretta, P. An
Ontology of Meta-Level Categories In J., Doyle
and Sandewall, E. and Torasso, P. (eds.), Proc.
of the Fourth International Conference Princi-
ples of Knowledge Representation and Reasoning
(KR94), Morgan Kaufmann, San Mateo, CA pp.
270-280, 1994.
[10] Mark, W. Ontology as Knowledge Base Architec-
ture In Proc. of the Banff Knowledge Acquisition
Workshop, 1995.
[11] Mars, N.J. ECAI Workshop on Knowledge Shar-
ing and Reuse: Ways and Means. European Co-
ordinating Committee for Artificial Intelligence
(ECCAI), Vienna, Austria, 1992.
[12] Mars, N.J. ECAI Workshop on Comparison of
Implemented Ontologies. European Coordinat-
ing Commitee for Artificial Intelligence (ECCAI),
Amsterdam, The Netherlands, 1994.
[13] Musen, Mark A. Dimensions of Knowledge Shar-
ing and Reuse. In Computers and Biomedical Re-
search,vol. 25, pp. 435-467, 1992.
[14] Neches, R. and Fikes, R. and Finin, T. and
Gruber, T. and Patil, R. and Senator, T. and
Swartout, W.R. Enabling Technology for Knowl-
edge Sharing. In AI Magazine, 1991.
[15] Nutter, J. T. Epistemology In Shapiro, Stuart, ed-
itor, Encyclopedia of Artificial Intelligence, John
Wyley, 1987.
[16] Poli, Roberto. Bimodality of Formal Ontology
and Mereology In Guarino, Nicola and Poli,
Roberto, (eds.), Formal Ontology in Conceptual
Analysis and Knowledge Representation, Kluwer
(to appear).
[17] van der Vet, P. E. and Speel, P. H. and Mars, N. J.
I. Ontologies for very large knowledge bases in ma-
terial science: a case study. In Proc. of the Second
international conference on building and sharing
of very large-scale knowledge bases (KB&KS ’95),
Twente, (to appear).
[18] Wielinga, B. and Schreiber, A. Th. and Janswei-
jer, W. and Anjewierden, A. and van Harmelen, F.
Framework and formalism for expressing ontolo-
gies. ESPRIT Project 8145 KACTUS, Free Uni-
versity of Amsterdam, deliverable DO1b.1, 1994.