=Paper=
{{Paper
|id=Vol-18/paper-6
|storemode=property
|title=PHYSICIAN is a Role Played by an Object, whereas SIGN is a Role Played by a Concept
|pdfUrl=https://ceur-ws.org/Vol-18/6-kassel.pdf
|volume=Vol-18
}}
==PHYSICIAN is a Role Played by an Object, whereas SIGN is a Role Played by a Concept==
https://ceur-ws.org/Vol-18/6-kassel.pdf
PHYSICIAN is a role played by an object,
whereas SIGN is a role played by a concept
Gilles Kassel
LaRIA
Université de Picardie Jules Verne
5, rue du Moulin Neuf
80000 - Amiens
FRANCE
E-mail: kassel@laria.u-picardie.fr
Fax: 33 (0) 3 22 82 76 54
Tel: 33 (0) 3 22 82 76 96
KADS group [VHe97a], in that it contains the definition
of all the concepts necessary for the application SATIN to
Abstract perform its diagnostic task. It contains, in particular,
descriptions of domain concepts such as HYPOCALCEMIA
In this article we tackle controversial questions and ERYTHEMA, and of problem-solving concepts, specific
about the nature of method ontology and its
to the task at hand, such as SIGN and SYNDROME1. These
relation to domain ontology. In order to consider a
latter are usually assimilated to roles [MDe88] played by
genuine “method ontology”, and to explain the
the former during problem-solving processes. However, as
link between problem-solving concepts and
Reynaud et al. have recently recalled [Rey97], no
domain concepts, we propose a new ontological
consensus exists yet about the meaning of the notion of
framework consisting of attributing to concepts the
“role”. When constructing OnS, we were therefore faced
status of extensional objects. Within this
with the question of elucidating the nature of the link
framework it becomes possible to consider concept
which exists between domain concepts and problem-
states and to envisage different specializations of
solving concepts.
such states, in particular “roles”. The problem-
This question has recently formed the subject of a
solving concepts can therefore be defined as roles
debate, in the journal IJHCS, opposing Guarino [Gua97]
played, not by objects modelling the external
to van Heijst et al. [VHe97b]. In short, Guarino proposes
world, but by concepts. In the article we also
applying to roles (or problem-solving concepts) the same
envisage the consequences that this framework
treatment as applied to domain concepts, leading to the
entails for the formalization of an ontology.
consideration of a domain ontology and a method ontology
(a term borrowed from PROTÉGÉ-II).
1 Introduction
“... an explicit representation of the problem-solving
The framework that we propose in this paper has been vocabulary along the lines of Gennari et al., (1994) and
elaborated in reponse to problems encountered when Falasconi and Stefanelli (1994) can help to
designing an ontology for SATIN [Kas97a], an expert systematically analyse the knowledge roles (McDermott,
system in neonatalogy. This ontology will be named OnS 1988) played by the domain knowledge within a
(for: the Ontology of SATIN) in the paper. particular problem solving strategy ...” [Gua97]
OnS is an application ontology, as this term is
understood by the PROTÉGÉ group [Tu95] and the Opposed to this view, van Heijst and his colleagues
refuse to consider that a method ontology is a real
ontology, giving in particular as a pretext the example that
The copyright of this paper belongs to the papers authors. Permission to current ontological frameworks such as Ontolingua are
copy without fee all or part of this material is granted provided that the
copies are not made or distributed for direct commercial advantage. unable to render a correct account of the link which exists
Proceedings of the IJCAI-99 workshop on between domain concepts and roles. In their response to
Ontologies and Problem-Solving Methods (KRR5) Guarino, they propose instead a different conception of
Stockholm, Sweden, August 2, 1999 the method ontology (a term they borrow for the sake of
(V.R. Benjamins, B. Chandrasekaran, A. Gomez-Perez, N. Guarino, M.
Uschold, eds.)
http://sunsite.informatik.rwth-aachen.de/Publications/CEUR-WS/Vol- 1 Concepts names will be systematically noticed in capital letters
18/ and in the font Courier.
G. Kassel 6-1
the debate) corresponding to the ontological requirements epistemological primitives, that can be distinguished
of the problem-solving methods in CommonKADS. according to their function (cf. figure 1):
“The problem is that dynamic knowledge roles are not - at a first level, objects function is to model a world,
classes and should not be treated as such. The relation be it external or internal to the PS, while serving as
between dynamic roles and ontological classes is not a substitutes for things which really exist in this world.
simple specialization relation, and the vocabulary These objects (of thought) are the means by which the
available in existing ontological frameworks such as
PS can construct knowledge about this world.
Ontolingua is not sufficiently expressive to model this
relation adequately. [...] In summary then, we believe
that knowledge roles are integral components of PSMs. - at a second level, concepts function is to denote
Although they often have the same names, they cannot representation entities, in particular - but not
be replaced by ontological concepts because they are of exclusively - objects at the first level. We shall also
another epistemological type.” [VHe97b] say that a concept corresponds to the idea a PS has
about representation entities, which can be objects at
Our aim in this article is to propose a conceptual the first level. Concepts are the means by which a PS
framework allowing the reconciliation of these two points can construct knowledge about its knowledge about a
of view. We think, as Guarino, that the principles of world.
formal ontology can apply to problem-solving concepts as
they apply to domain concepts, and that the term “method
ontology” can thus be heard in the same meaning as the
term “domain ontology”. However, we agree with van c o
Heijst and his colleagues for considering that simple
specialization relations cannot render an account of the c o o
external
link which exists between the two categories of concepts c
c o world
(as suggested by Guarino), and that a piece of the puzzle
is missing which would enable us to speak of a real Universe of thought
“method ontology”.
The missing piece, or conceptual framework, that we modelling relation
c = concept
propose relies on the following idea: It is worthwhile to
o = object denotation relation
attribute to concepts the status of object in their own right,
enabling them to be referred in their turn by (meta)-
concepts; it then takes sense to consider states of concepts Figure 1: A Universe Of Thought Is Populated By Objects
and to envisage specializations of such states, among And Concepts
others roles; problem-solving concepts can then be
defined as roles played, not by objects modelling the real The notion of “concept” on which we rely comes back
world, but by concepts. to consider, according to the classical framework of the
In order to outline this idea, we adopt the following semiotic triangle (cf. figure 2a), that a concept is made up
plan. In a first part, we propose a “representation of a term, a notion (its “meaning”, or “intension”) and an
ontology” in which concepts are assimilated to individuals object (its “reference”, or “extension”). With regard to this
(section 2). Independently, we make precise our notion of framework, we add two points (cf. figure 2b):
“role”: a role corresponds to an influence, either
undergone or exercised, by an entity over another entity. - On one hand, we attribute to concepts the status of
As such, the concept ROLE is subsumed in OnS by the object, which gives them the possibility of being, in
concept STATE (section 3). These two points are then their turn, referred by (meta)-concepts. From the point
considered simultaneously in order to clarify the nature of of view of the modelling relation, we admit that a
problem-solving concepts (section 4). We begin showing concept models itself (cf. the loop of the modelling
that objects, as a modelling primitive, are suitable for relation on the concepts in figure 1).
account to models of systems in the world (domain
knowledge) whereas concepts are suitable for account to - On the other hand, we consider that the same relation
models of problem solving processes (solving methods). of denotation, which exists between concepts and
The main point here is that concepts, as opposed to objects, also exists between meta-concepts and
objects, can exercise an influence over solving processes. concepts. This point is illustrated in figure 1 by
According to this analysis, we conclude that the notion of denotation links established between concepts.
a “role played by a concept” is a natural framework for
accounting to the nature of problem-solving concepts. Let us illustrate this “pivot” role that a concept can
play - to denote or to be denoted - by taking the example
2 Treating Concepts As Individuals of the concept ONTOLOGY. Its status of concept predicts
that we can render an account of the meaning a sentence
We postulate that a Problem-Solver (PS), or such as: “An ontology is a specification of a
knowledgeable agent, has in particular at its disposal two conceptualization”. This definition is usually put forward
categories of knowledge representation entities, or for trying to circumscribe the intension, that is the
G. Kassel 6-2
meaning, of the concept ONTOLOGY. One expresses here constitute an operational framework, and in particular to
that a certain property - being a specification of a serve as a basis for the definition of a knowledge
conceptualization - is shared by all the objects-ontologies. representation language, the framework should be
The Universe of discourse is made up, in the present case, completed. However for our purpose in this article we
of ontologies developed by researchers or knowledge shall consider that this sketch is sufficient. We shall
engineers. The object status of the concept predicts for its content ourselves in the following to introduce the
part that one can render an account of the following distinction between generic concept and individual
knowledge: “The notion of 'ontology’ is not completely concept.
defined and is still controversial”. This last knowledge
bears on the idea that the Knowledge Engineering
community has about these object-ontologies. We could 3 Modelling Roles As Temporal Objects
have stated a similar knowledge about the notion of
“role”. The Universe of discourse is this time composed of In this section, we make precise our notion of “role”. To
concepts. The concept ONTOLOGY is here referred by a this end, we present and justify the place occupied by the
meta-concept ILL DEFINED CONCEPT. One can interpret concept ROLE in the top-level of OnS.
this knowledge as the fact that there does not yet exist a
consensus in the Knowledge Engineering community 3.1 The Concept ROLE Belongs To The Sub-System
about the meaning of this concept. This last knowledge TEMPORAL OBJECT
bears the status of a meta-knowledge.
When designing OnS, we have reused and adapted the
ontology developed within the AIM project MENELAS
[Bou94], in particular its top-level. We therefore consider
term object
that objects are specialized into three categories:
PHYSICAL OBJECT, ATTRIBUTE and TEMPORAL OBJECT.
The figure 3 shows the top-level of the sub-system
concept TEMPORAL OBJECT (the part of OnS subsumed by the
concept TEMPORALOBJECT). This concept denotes the set
of the objects which are located in time, i.e. objects to
whom information such as a duration or a date can be
notion 2a attached. These objects account for a snapshot of a world,
fixed in time (concept STATE), or, on the contrary, for an
evolution of the world (concept CHANGE).
term concept
TEMPORAL OBJECT
meta-concept
CHANGE STATE
notion 2b
INTERNAL STATE EXTERNAL STATE
Figure 2: A Concept Denotes An Object (2a) And Can
Be In Its Turn Denoted By A Meta-concept (2b)
COMPARATIVE STATE ROLE
The ontological framework, or representation
ontology, that we have just sketched, is situated at the
same level as the “Frame Ontology” of Gruber [Gru93].
Its contribution, in regard to the Frame Ontology and also Figure 3: The Concept ROLE Is Subsumed By The Concept
to the (implicit) representation ontology of concepts EXTERNAL STATE
languages [Hei94], is in treating concepts as a new level
of extensional objects, following thus a proposition Among the states a distinction is made between
already stated by McCarty [MCa79] 2. In order to INTERNALSTATES, which consist in an internal manner of
being (e.g., HYPOCALCEMIA, which we assimilate to an
internal state of a human being), and EXTERNAL STATES,
2 In this reference, Mc Carthy already proposed to consider which consist in a way of being with regard to external
individual concepts such as the concept MORNING STAR, objects. This latter concept is further specialized into two
refering to individuals, as new individuals. He noticed moreover
that this proposition could very well be extended to generic
concepts such as our concept ILL DEFINED CONCEPT, refering to sets of individuals.
G. Kassel 6-3
distinct categories: states of objects considered as normal physician without (happily) ceasing of exist. We therefore
or abnormal with regard to objects of the same type (e.g., notice that the concept PHYSICIAN, contrary to the concept
IMPORTANT CALCIUM RATE) and finally roles played by HUMAN BEING, is not an essential concept, or, re-using a
objects with regard to other objects. This last concept synonymous term introduced by Guarino, is not
subsumes in OnS, for example, the status of PATIENT and “semantically rigid” [Gua92].
of PHYSICIAN, as well as physiological functions fulfilled Our conception is not far distant, however, where
by organs with regard to the global physiological system. Sowa and Guarino see only one concept for accounting
The external states specialize therefore themselves for a role, we distinguish two of them. We postulate that
according to whether the relation which exists between the the particular relation characterizing the status of
entities does, or not, rely on the influence exercised by physician must be firstly defined as a concept in its own
one of the entities over the other. For example, right, that is, that we must be able at first to account for
considering the physiological function of an organ is the the meaning of this relation. This latter being fixed in
same as considering the influence exercised by the organ time, the concept which we name PHYSICIAN naturally
within a physiological system. We consider that this belongs to the sub-system TEMPORAL OBJECT. We thus
relation is of the type ROLE. On the other hand, stating that render the concept PHYSICIAN semantically rigid,
an organ is abnormal requires comparison with another denoting, not physical objects located in space as human
organ of the same type, considered as normal. We beings, but temporal objects. Once this concept is defined,
consider that we are here in presence of a COMPARATIVE nothing prevents definition of a concept such as
STATE which situates two objects, one with regard to the PHYSICIAN HUMAN BEING denoting persons having a
other, without the possibility in this case of making a physician status. This second concept, considered
reference to some influence. erroneously - we believe - by Sowa and Guarino as a role,
The definition that we have just given, in particular for will indeed be a non-semantically rigid concept.
the notion of ROLE, must be considered as an Furthermore it will be noticed that the term “role”
approximation for the meaning of this notion. It relies on designates for us a concept referring to objects, in this
distinctions which appeared both necessary and sufficient case particular temporal objects, and not a category of
for constructing the domain ontology of SATIN. concepts, i.e. a meta-concept referring to particular
However, nothing assures us that for another application concepts, as proposed by Sowa [Sow88] and Guarino
we shall not have to come back to these definitions, for [Gua94].
example in refining them while detailing the distinction
COMPARATIVE STATE versus ROLE. Therefore it is an
operational definition - for the application SATIN - of the 4 The Method Ontology Of OnS
notion of ROLE, that we have given here.
In this section, we come to the method ontology of
3.2 Discussion About The Nature Of Roles SATIN, that is the part of OnS corresponding to problem-
solving concepts. The pre-requisites are established for
We have just seen that the concepts which we consider as explaining the meaning of these concepts. Relying on the
roles are located in OnS within the sub-system TEMPORAL object status of the concepts, presented in section 2, and
OBJECT. This is the case, for example, for the concept on the notion of role that we have just seen in section 3,
PHYSICIAN. However this choice may surprise at first we are in a position to assimilate the problem-solving
sight. Indeed, considering that “every physician is a concepts into roles played by concepts.
human being” it might seem more logical to define the To that end, we focus our analysis on the concept
concept PHYSICIAN as a specialization of the concept SIGN. We first underline the importance of distinguishing
HUMAN BEING. Moreover, as is recalled by Guarino two models, the model of a physiological system and the
[Gua92], this last point of view corresponds also to the model of problem solving, and clarify the respective role
way Sowa conceives the notion of role: of the two primitives: object and concept, for taking these
models into account (4.1). Thereafter we only have to
“Subtypes of ENTITY are of two kinds: natural specify the place occupied by problem-solving concepts in
types, which have no required set of linguistic OnS (4.2).
associations, and role types, which are subtypes of
natural types in some particular pattern of 4.1 Model Of A Physiological System Versus Model Of
relationships. PERSON, for example, is a natural Problem-solving
type, and TEACHER is a sub-type of PERSON in
the role of teaching.” [Sow88] In order to perform a diagnostic task, a Problem Solver
must have knowledge about a specific case, a patient in
According to this conception, the role PHYSICIAN, just the case of a medical diagnosis. But it must also have
as the role TEACHER, must indeed be defined as a knowledge about its “domain”, that is knowledge
specialization of the concept HUMAN BEING. The accounting for regularities observed in a class of patients
differentia between the two concepts consists in the analogous to the patient-specific case [Wie93]. We shall
particular relation in which a person is engaged, and designate that knowledge by the term “model of a
which defines the status of physician. This relation physiological system”. Let us, for example, consider the
corresponds to an incidental property of a human being in
that it is possible for him (or her) to lose the status of
G. Kassel 6-4
following specific case knowledge (K S) and domain We have used this time the level of the concepts. The
knowledge (KD): expression “ the idea according to which ...” may be seen
as a means for making reference to a concept, by posing,
KS the patient presents an hypercorticism in some sort, an equality, the phrases commencing by “the
patient presents an ...” serving to designate concepts. The
KD every hypercorticism provokes an interest of resorting to concepts is to enable one to
hypocalcemia 3 designate unambiguously objects, providing for that
purpose necessary and sufficient conditions: the objects
It is easy to see that the level of the objects is well referred to in KSR are respectively the state of
adapted, and also sufficient, for rendering an account of hypercorticism and the state of hypocalcemia shown by
the meaning of this knowledge. One can for example the patient. We are here in presence of individual
propose, as a formalization of KS and KD, the following concepts, making reference to individual objects.
first order theory, in which the symbols of constants (in We have just seen the usefulness of having recourse to
large letters) represent objects and the symbols of concepts, however it is fitting to go further while noticing
predicates (in small letters) sets of objects: that the level of objects is inadequate for describing such
an inferential process and that, in consequence, the level
ϕKS1 patient(A-PATIENT) of concepts cannot be ignored. The argument that we put
ϕKS2 hypercorticism(AN- forward is that no direct causal relation exists between a
HYPERCORTICISM) physiological process occurring in the patient and an
ϕKS3 presents(A-PATIENT, inferential process occurring in the PS. The inferential
AN-HYPERCORTICISM) process relies on a model of the patient to whose
ϕKD ∀x {hypercorticism(x) ⇒ development it contributes, by trying to reconstitute a
∃y [hypocalcemia(y) ∧ causes(x, y)]} causal history of the patient. By definition, this model
does not necessarily reflect reality, and we thus cannot
The interest of domain knowledge relies in its attribute the existence of the inferential process to that of a
predicative power. By relating the domain model to the physiological process ... which perhaps has never existed.
model of the specific case, it becomes possible in the Only objects of knowledge exist for the PS, and these
present case to predict that the patient presents an objects must moreover be endowed with properties in
hypocalcemia. Again, one will note that the level of the order to be able to give rise to inferential processes. In the
objects is sufficient for performing such reasoning. From case of K SR one can say that the inferential process exists
a clausal form equivalent to ϕKD and in using ϕKS2, a due to the fact that there exists an object having the
theorem prover will introduce a Skolem function in order property of being an hypercorticism. Speaking about this
to infer the existence of an hypocalcemia, and to infer that state of knowledge refers back to an individual concept.
this hypocalcemia is provoked by the hypercorticism of The knowledge K SR that we have considered concerns
the patient. In our description of the first inference, below, a given inferential process, in the same manner as
the symbol “⁄–” corresponds to the symbol for syntactical knowledge KS was rendering an account of a given
derivation of theorems. physiological process, concerning a given patient. Let us
suppose now that - by analogy with knowledge KD - we
¬ hypercorticism(x) v hypocalcemia(f(x)) ; wished to describe a class of inferential processes
hypercorticism(AN-HYPERCORTICISM) analogous to the process we have just seen, after having
hypocalcemia(f(AN-HYPERCORTICISM)) noticed, for example, that the inference “hypercorticism
⁄– hypocalcemia” is drawn for each considered patient.
Let us now take the place of an external observer, and We are therefore seeking to express a model of a class of
let us suppose that we have to describe these inferences, problem-solving. We shall state the regularity we have
that is to state a “model of a problem-solving”. One can just mentioned in the following way:
therefore propose the formulation below. We have given it
the status of knowledge and named it KSR - by analogy KDR the idea according to which patients
with K S - in order to indicate that this knowledge bears present an hypercorticism has led the PS to
on a given reasoning process: evoke the idea according to which these
patients present an hypocalcemia.
KSR The idea according to which the patient
presents an hypercorticism has led the PS In this description the concepts which are at the origin
to evoke the idea according to which the of (or which result from) the evocation, refer, no more to
patient presents an hypocalcemia. individual objects, but to sets of objects, respectively the
set of hypercorticism and the set of hypocalcemia shown
by the different patients. These are generic concepts, in
contrast to individual concepts.
3 In principle generalizations that can be drawn from a class of To sum up, we have taken care in this section to
distinguish two types of descriptions, or models,
patients correspond rather to imprecise knowledge such as:
generally an hypercorticism provokes an hypocalcemia. depending on the nature of the modelled object: a
However, as ignoring the imprecision has no incidence on our physiological system, or a problem-solver. We have
analysis in this section, we will do so for the sake of simplicity. shown that the objects level is suitable to account for a
G. Kassel 6-5
model of a physiological system, whereas the concepts As a consequence, a concept such as HYPOCALCEMIA
level is for itself suitable to account for a model of a (or HYPERCORTICISM, if we rely on knowledge KDR
problem-solver. As a consequence, one must expect that which describes the role played by this latter concept) and
problem-solving concepts, or knowledge roles, will be the different roles played by this concept, among others
defined as states of concepts. That is what we are going to diagnostic roles, are as many distinct entities. Such an
discuss next in specifying the place of the concept SIGN in entity is the concept which we designate by the term
OnS. “hypocalcemia sign”. This concept is generic in that it
refers to a class of situations (or evocation events), as
4.2 The Sub-system CONCEPTUAL ROLE described in KSR, in which a concept evokes another. The
concepts HYPOCALCEMIA and HYPOCALCEMIA SIGN are
A knowledge such as KDR equates to taking into account linked by two relations, one being the converse of the
the influence exercised by a (generic) concept with regard other: a conceptual relation, named “has-for-role”, enables
to problem-solving processes. We thus rediscover here our to indicate what are the roles that a concept plays;
notion of role, as defined in section 3. The only difference conversely, the relation “is-played-by” indicates which
is that we have been interested up to now in roles played concept plays a given role.
by objects, whereas in the present case we are confronted Another term: “hypocalcemia syndrome” (see figure
with roles played by concepts. 4), is also linked to the notion of hypocalcemia. The
The notion of SIGN that we propose in OnS meaning of this term relies on the existence of a different
corresponds to a characterization of this conceptual conceptual role, which corresponds to the concept
influence for which knowledge KDR accounts. We state SYNDROME. The concept SYNDROME is opposed to the
that a temporal object of the type SIGN exists as soon as a concept SIGN by attributing the evocation role to a set of
concept evokes another, enabling the PS to progress concepts, and not to an isolated concept (as this set is not
toward the establishment of the diagnosis. This meaning represented in figure 4, there is no arrow leading to the
that we attribute to the term “sign” seems to us to node HYPOCALCEMIA SYNDROME).
correspond to currently accepted definitions of the term, The definition of the term “syndrome” we have
even if these do not mention the distinction we have retained may seem incomplete for the reader. Indeed, a
established between objects and concepts. The concept syndrome is usually defined as a group of signs whose
SIGN is subsumed in OnS by the concept DIAGNOSTIC origin is not known, yet our concept SYNDROME only
ROLE, itself subsumed by the concept CONCEPTUAL ROLE retains the first part of this definition. The reason is that
(cf. figure 4). the second part - having an unknown origin - corresponds
in OnS to another concept. This property amounts in
reality to define a particular “state” of a pathological state.
STATE The concept in question, SYNDROMIC STATE, is thus
located in OnS in a sub-system STATE OF TEMPORAL
OBJECT, and it is there defined by opposition with the
INTERNAL STATE EXTERNAL STATE concept DISEASE, which denotes pathological states
whose origin is known. The point of this model is that it
enables syndromic states - as well as diseases - to be
evoked by means of a group of signs. One can therefore
ROLE speak about the “hypocalcemia syndrome” and the
“meningitis syndrome”, as suggested by an analysis of the
HYPERCORTICISM domain knowledge, even if hypocalcemia and meningitis
OBJECT ROLE have a known origin. Of the two concepts: SYNDROMIC
HYPOCALCEMIA CONCEPTUAL ROLE STATE and SYNDROME, only the second belongs to the
method ontology of OnS.
PHYSICIAN PATIENT
5 Discussion
DIAGNOSTIC ROLE
We have just seen on which basis the method ontology of
SATIN has been developed. In the present discussion, we
SIGN SYNDROME return to the positions expressed by Guarino [Gua97] and
van Heijst and his colleagues [VHe97a][VHe97b], about
the nature of these ontologies, in order to clarify our
HYPOCALCEMIA SIGN position. We organize the discussion around two
has-for-role controversial points: One, the question of the link existing
HYPOCALCEMIA
between the two ontologies, this question referring in fact
is played by SYNDROME
to specifying what a method ontology really is (5.1); and
two, the question of the formalization of an application
Figure 4: The Sub-System DIAGNOSTIC ROLE ontology, and of the limits of formal frameworks such as
Ontolingua (5.2).
G. Kassel 6-6
A knowledge role, or conceptual role, such as SIGN is
5.1 Nature Of A Method Ontology defined in a manner analogous to that of a role played by
an object. We have considered in section 4.2 that a role of
As we have observed in the introduction, van Heijst and type SIGN exists as soon as a concept evokes another
his colleagues deny for roles the status of “ontological concept. Therefore the intension of this problem-solving
concept”, this denial reflecting in the title of their response concept characterizes the domain concepts playing that
to Guarino : “Roles are not classes”. Their point of view is role, without assuming that such concepts do exist. We
that an ontological approach, such as proposed by find here exactly what van Heijst et al. expect of a method
Guarino, does not provide an answer for the requirements ontology, that is that it corresponds to the requirements of
which correspond to a method ontology, or more exactly the problem solving methods of KADS, while applying to
that there exist two different conceptions for this ontology these concepts an ontological approach, as recommended
which are opposed: by Guarino.
“We can imagine at least two interpretations of the term 5.2 Formalization Of An Application Ontology
method ontology. Firstly, method ontologies can be
descriptions of the roles that domain concepts and In order to complete our proposition, we still have to show
relations can play in the reasoning process. Guarino how it is possible to formalize the kind of application
does not give a definition in his paper, but based on the
ontology that we envisage. On that point, we agree with
examples that he gives [...] we assume that this
interpretation corresponds to his view. [...] The second van Heijst et al. for noticing that a formal framework such
interpretation of the term method ontology that we can as Ontolingua is inadequate:
imagine is that a method ontology describes the domain
concepts and relations that are manipulated by a “The relation between dynamic roles and ontological
problem solving method in the reasoning process. We classes is not a simple specialization relation, and the
believe that this is how method ontologies are vocabulary available in existing ontological frameworks
understood in PROTEGE-II. Viewed in this way, such as Ontolingua is not sufficiently expressive to
method ontologies fulfill the same roles as the model this relation adequately .” [VHe97b]
ontological requirements of PSMs in CommonKADS,
the difference being that the relation between the A candidate language has to extend the representation
method ontology and the domain ontology is ontology presented in section 2, and in particular to give
implemented through ISA relations, whereas the to the concepts the entire status of objects. This
relations between the requirements and the ontology characteristic is essential, as we have seen, to account for
concepts are “satisfies” relation. We prefer the
the link between the domain concepts and the conceptual
requirements solution over the method ontology solution
roles. Languages such as KIF within the project
[...].” [VHe97b]
Ontolingua [Gru93] or languages of concepts descending
from KL-ONE [Hei94] enable one to specify and
We think that the approach proposed in this paper is
manipulate intensions of concepts, however they do not
able to conciliate the two interpretations mentioned by van
allow for considering these intensions as real individual
Heijst and his colleagues. We have seen in section 4.2
objects, by allowing one to attribute to them, in their turn,
how we are able to specify, for each domain concept, the
properties, that is by allowing them to be referred by
role(s) each one plays in the problem-solving process (by
meta-concepts. This limitation seems to us to correspond
means of the relation “has-for-role”), and, for each role,
to the distinction proposed by Hirst [Hir89] between
which domain concepts play that role (by means of the
“weakly” and “strongly” intensional languages:
relation “is-played-by”. In our view, these two relations
seem to correspond to the second interpretation quoted “It is interesting to note that, generally speaking, KR
above. Inversely, specifying the description of the roles - formalisms that treat concepts as first-class objects do
which corresponds to the first interpretation - belongs to not formally distinguish them from individuals. [...] I
the exact definition of the role. So let us now look more don’t know of any principled reason for this. Such
closely at how the concepts of the conceptual sub-system systems are weakly intensional systems, countenancing
ROLE are defined. intensions but not making anything special of them. In
Let us first consider a role played by an object, for contrast, strongly intensional systems take intensions to
example, the parental role MOTHER. As a temporal object be not just first-class objects but objects of a distinct
we define this role by assuming that this one exists as kind. [...] I suspect that a strongly intensional system
soon as a woman has a child. We equate a temporal object will be necessary for an ontologically adequate
of the type MOTHER to this situation. In other words, we treatment of intensions.” [Hir89]
consider that the fact of a woman having a child defines,
in a necessary and sufficient manner, the concept MOTHER. The language Def-* that we have been developing for
In the same way, the fact of a person undergoing a formalizing Knowledge Based Systems can be termed
medical examination defines a role of the type PATIENT. “strongly intensional” [Kas97b][Kas99]. From the point
One will note that these definitions give rules for of view of the representation of the ontological
recognizing the type of situation, but that they do not knowledge, Def-* presents two main characteristics: it
assume that there exists objects belonging in these confers to concepts and propositions the entire status of
situations. This last information belongs to a model of a objects, while offering an important propositional
world. expressive power. In this paper, for reasons of space, we
G. Kassel 6-7
must be content with illustrating these two characteristics “∃∀”4, whereas the structural links comprise quantifiers
by means of examples, without going into the details of such as: “M∃” or “M∀”, the quantifier “M” playing the
the language. We comment below on the representation of role of a “modifier”.
the concepts HYPOCALCEMIA , SIGN and HYPOCALCEMIA In figure 5a, the concept HYPOCALCEMIA is defined
SIGN (cf. figure 5). with regard to the genus INTERNAL STATE. The key-word
“=” means that the property “being an internal state
characterized by a calcium rate less than 1.85”
(def-concept #hypocalcemia corresponds to a necessary and sufficient condition. Other
= [#internal-state] -> (M∃#is-characterized-by) - properties, corresponding to additional necessary
-> [#calcium-rate] -> (M∃#has-for-value) - conditions, follow behind the kew-word “properties of the
-> [#number] -> (MI#less-than) -> [1.85],, objects”, for example : “every hypocalcemia manifests
itself by a tetany”. These object properties are separated
properties of the objects from the conceptual properties by the key-word
“properties of the concept”. The properties situated below
-> (∀∃#manifests-itself-by) -> [#shaking] the key-word then consider the concept as an object.
Three propositions bearing on the concept are thus
-> (∀∃#manifests-itself-by) -> [#tetany]
represented: “the concept HYPOCALCEMIA is expressed in
the language by some term”; “it plays the role
properties of the concept
hypocalcemia-sign during reasoning”; “it evokes the
concept HYPERCORTICISM”.
-> (#is-named-by) -> [#term123] The definitions for the concepts SIGN, in figure 5b, and
-> (#has-for-role) -> [#hypocalcemia-sign] HYPOCALCEMIA SIGN, in figure 5c, follow the same
-> (#evokes) -> [#hypercorticism] ) structure. As a temporal object, a SIGN is defined as a
5a DIAGNOSTICROLE : this role is played by a concept which
evokes another. Finally the role HYPOCALCEMIASIGN is
(def-concept #sign defined as the sign played by the concept HYPOCALCEMIA.
= [#diagnostic-role] -> (M∃is-played-by] - We have seen with these two examples that the
-> [#concept] -> (M∃#evokes) -> [#concept], language Def-* allows us to represent both the domain
concepts and the problem-solving concepts. The
properties of the concept representation of a concept contains the representation of
its intension, i.e. the representation of a set of properties
-> (#is-named-by) -> [#term78] ) verified by the objects referred to by the concept. This
5b part of the representation corresponds schematically to the
services offered by the languages of concepts, or
(def-concept #hypocalcemia-sign terminological systems [Hei94]. But the representation
= [#sign] -> (MI#is-played-by) -> [#hypocalcemia]) also contains a representation of the properties satisfied by
the concept itself, considered as an individual object. It is
5c this second part of the representation which provides, in
particular, an account of the link between the domain
concepts and the problem-solving concepts.
Figure 5: Representation In Def-* Of The Concepts
HYPOCALCEMIA (5a), SIGN (5b) and
HYPOCALCEMIA-SIGN (5c)
6 Conclusion
An entity introduced by the primitive “def-concept”
represents both a class of extensional objects and the The framework proposed in this paper mostly relies on the
concept referring to that class, and even in some cases, reification of the concepts as a new level of extensional
several co-referential concepts having for extension the objects. The precise meaning attributed to the notion of
same set of objects. This explains that different parts are “role” is, as a matter of fact, secondary in our analysis.
distinguished within such a construction (cf. figures 5a In particular, the fact for a role of being a “class” or a
and 5b): The first part is dedicated to the definition of the “meta-class” does not matter. The same meaning of our
properties of the class objects; the second part, which is concept ROLE is imbeded both in a concept such as
delineated after the key-word “properties of the concept” PHYSICIAN and the concept SIGN. In return, what
allows to state the properties of the concept. The syntax differentiates the two roles is that the former bears on an
used to formalize the properties corresponds to the linear object, whereas the latter bears on a concept.
syntax proposed by Woods for a semantic network The knowledge according to which a domain concept
[Woo91]: the links between concepts, or pairs of relations, plays a particular role in reasoning belongs to a model of
have a quantificational import, moreover assertional links problem-solving. We hope to have shown in this paper
are distinguished from structural links. The assertional
links comprise double quantifiers such as: “∀∃”, “∃∃”, 4 A link such as “[a] -> (∀∃Rel) -> []b]” is, for example,
equivalent to the following first order logical formula: ∀x [ a(x)
⇒ ∃y (b(y) ∧ Rel(x, y)) ].
G. Kassel 6-8
that languages such as KIF or similar languages of [Kas97a] G. Kassel and G. Krim. Une expérience de
concepts are inadequate for formalizing such a model, but réutilisation d’une ontologie. Internal Report
that it is necessary to resort to “strongly intensional” 97/64, HEUDIASYC, Université de
languages such as Def-* in order to attribute properties - Technologie de Compiègne.
not only to objects - but also to concepts. [Kas97b] G. Kassel and M. Traoré. Modelling with a
This latter knowledge has indeed the status of meta- strongly intensional language brings numerous
knowledge and its use in a reasoning must be assimilated advantages. 7th Workshop on Knowledge
to meta-resoning. We therefore think that our framework Engineering: Methods and Languages, KEML
offers a new way for accounting to the reflective part of ‘97, Milton Keynes, UK.
problem-solving processes. We are currently working on [Kas99] G. Kassel. Def-* : manuel de référence.
this perspective. Internal Report 99/01, LaRIA, Université de
Picardie Jules Verne.
[MCa79] J. McCarthy. First Order Theories of Individual
References Concepts and Propositions. Machine
Intelligence, 9:129-147.
[Bou94] J. Bouaud, B. Bachimont, J. Charlet and P. [MDe88] J. McDermott. Preliminary steps towards a
Zweigenbaum. Acquisition and Structuring of taxonomy of problem-solving methods. In: S.
an Ontology within Conceptual Graphs. Marcus, eds., Automating Knowledge
ICCS’94 Workshop on Knowledge Acquisition Acquisition for Expert Systems, MA:Kluwer,
using Conceptual Graph Theory , University of Boston, 225-255.
Maryland, College Park, MD, 1-25. [Rey97] C. Reynaud, N. Aussenac-Gilles, P.
[Fal94] S. Falasconi and M. Stefanelli. A library of Tchounikine and F. Trichet. The notion of role
implemented ontologies. ECAI Workshop on in conceptual modelling. 10th European
Comparison of Implemented Ontologies, Knowledge Acquisition Workshop: EKAW'97,
Amsterdam, The Netherlands, 81-91. Bonn: Springer Verlag, San Feliu de Guixolls.
[Gen94] J.H. Gennari, S.W. Tu, T.E. Rothenfluh and [Sow88] J.F. Sowa. Using a lexicon of canonical graphs
M.A. Musen. Mapping domains to methods in in a semantic interpreter. In: M.W. Evens, eds.,
support of reuse. International Journal of Relational Models of the Lexicon, Cambridge
Human-Computer Studies 41:399-424. University Press, Cambridge, 113-137.
[Gru93] T.R. Gruber. A translation approach to portable [Tu95] S.W. Tu, H. Eriksson, J.H. Gennari, Y. Shahar
ontology specifications. Knowledge and M.A. Musen. Ontology-based
Acquisition, 5:199-220. configuration of problem-solving methods and
[Gua92] N. Guarino. Concepts, attributes and arbitrary generation of knowledge acquisition tools: the
relations; Some linguistic and ontological application of PROTEGE-II to protocol-based
criteria for structuring Knowledge bases. Data decision support. Artificial Intelligence in
& Knowledge Engineering, 8:249-261. Medecine, 7:257-289.
[Gua94] N. Guarino, M. Carrara and P. Giaretta. An [VHe97a] G. Van Heijst, A.Th. Schreiber and B.J.
Ontology of Meta-Level Categories. Fourth Wielinga. Using explicit ontologies in KBS
International conference: Principles of development. International Journal of Human-
Knowledge Representation and Reasoning Computer Studies, 46:183-292.
(KR94), Morgan Kaufmann, San Mateo, CA, [VHe97b] G. Van Heijst, A.Th. Schreiber and B.J.
270-280. Wielinga. Roles are not classes: a reply to
[Gua97] N. Guarino. Understanding, building and using Nicola Guarino. International Journal of
ontologies. International Journal of Human- Human-Computer Studies, 46:311-318.
Computer Studies, 46:293-310. [Wie93] B.J. Wielinga, W. Van de Velde, G. Schreiber
[Hei94] J. Heinsohn, D. Kudenko, B. Nebel and H.-J. and H. Akkermans. Towards a Unification of
Profitlich. An empirical analysis of Knowledge Modelling Approaches. In: J.M.
terminological representation systems. David, J.P. Krivine and R. Simmons, eds.,
Artificial Intelligence, 68:367-397. Second Generation Expert Systems, Springer-
[Hir89] G. Hirst. Ontological assumptions in Verlag, 299-335.
knowledge representation. In: R.J. Brachman, [Woo91] W.A. Woods. Understanding Subsumption and
H.J. Levesque and R. Reiter, eds., First Taxonomy: A Framework for Progress. In: JF.
International Conference on Principles of Sowa, ed., Principles of Semantic Networks;
Knowledge Representation and Reasoning, Explorations in the Representation of
157-169. Knowledge, Morgan Kaufmann, 45-94.
G. Kassel 6-9