Protege owl tutorialp3_v1_0

A Practical Guide To Building OWL Ontologies Using The
Prot´eg´e-OWL Plugin and CO-ODE Tools
Edition 1.0
Matthew Horridge1
,
Holger Knublauch2
, Alan Rector1
, Robert Stevens1
, Chris Wroe1
1
The University Of Manchester
2
Stanford University
Copyright c The University Of Manchester
August 27, 2004

Contents
1 Introduction 8
1.1 Conventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2 Requirements 10
3 What are OWL Ontologies? 11
3.1 The Three Species Of OWL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
3.1.1 OWL-Lite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
3.1.2 OWL-DL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.1.3 OWL-Full . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.1.4 Choosing The Sub-Language To Use . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.2 Components of OWL Ontologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.2.1 Individuals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.2.2 Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.2.3 Classes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
4 Building An OWL Ontology 16
4.1 Named Classes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
4.2 Disjoint Classes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
4.3 Using The OWL Wizards To Create Classes . . . . . . . . . . . . . . . . . . . . . . . . . . 21
4.4 OWL Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
4.5 Inverse Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
1

4.6 OWL Property Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
4.6.1 Functional Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
4.6.2 Inverse Functional Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
4.6.3 Transitive Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
4.6.4 Symmetric Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
4.7 Property Domains and Ranges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
4.8 Describing And Defining Classes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
4.8.1 Property Restrictions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
4.8.2 Existential Restrictions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
4.9 Using A Reasoner . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50
4.9.1 Determining the OWL Sub-Language . . . . . . . . . . . . . . . . . . . . . . . . . 50
4.9.2 Using RACER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.9.3 Invoking The Reasoner . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.9.4 Inconsistent Classes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
4.10 Necessary And Sufficient Conditions (Primitive and Defined Classes) . . . . . . . . . . . . 57
4.10.1 Primitive And Defined Classes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
4.11 Automatic Classification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
4.11.1 Classification Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
4.12 Universal Restrictions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
4.13 Automatic Classification and Open World Reasoning . . . . . . . . . . . . . . . . . . . . . 69
4.13.1 Closure Axioms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
4.14 Value Partitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
4.14.1 Covering Axioms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
4.15 Using the Properties Matix Wizard . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77
4.16 Cardinality Restrictions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
5 More On Open World Reasoning 84
2

6 Creating Other OWL Constructs In Protégé-OWL 91
6.1 Creating Individuals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
6.2 hasValue Restrictions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
6.3 Enumerated Classes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
6.4 Annotation Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96
6.5 Multiple Sets Of Necessary & Sufficient Conditions . . . . . . . . . . . . . . . . . . . . . . 98
7 Other Topics 100
7.1 Language Profile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
7.2 Namespaces And Importing Ontologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
7.2.1 Namespaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
7.2.2 Creating And Editing Namespaces in Protégé-OWL . . . . . . . . . . . . . . . . . 101
7.2.3 Ontology Imports in OWL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
7.2.4 Importing Ontologies in Protégé-OWL . . . . . . . . . . . . . . . . . . . . . . . . 103
7.2.5 Importing The Dublin Core Ontology . . . . . . . . . . . . . . . . . . . . . . . . . 105
7.2.6 The Protégé-OWL Meta Data Ontology . . . . . . . . . . . . . . . . . . . . . . . . 106
7.3 Ontology Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
7.4 TODO List . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
A Restriction Types 111
A.1 Quantifier Restrictions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
A.1.1 someValuesFrom – Existential Restrictions . . . . . . . . . . . . . . . . . . . . . . 112
A.1.2 allValuesFrom – Universal Restrictions . . . . . . . . . . . . . . . . . . . . . . . . . 112
A.1.3 Combining Existential And Universal Restrictions in Class Descriptions . . . . . . 113
A.2 hasValue Restrictions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
A.3 Cardinality Restrictions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
A.3.1 Minimum Cardinality Restrictions . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
A.3.2 Maximum Cardinality Restrictions . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
3

A.3.3 Cardinality Restrictions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
A.3.4 The Unique Name Assumption And Cardinality Restrictions . . . . . . . . . . . . 115
B Complex Class Descriptions 116
B.1 Intersection Classes ( ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
B.2 Union Classes ( ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
4

Copyright
Copyright The University Of Manchester 2004
5

Acknowledgements
I would like to acknowledge and thank my colleagues at the University Of Manchester and also Stanford
Univeristy for proof reading this tutorial/guide and making helpful comments and suggestions as to how
it could be improved. In particular I would like to thank my immediate colleagues: Alan Rector, Nick
Drummond, Hai Wang and Julian Seidenberg at the Univeristy Of Manchester, who suggested changes to
early drafts of the tutorial in order to make things clearer and also ensure the technical correctness of the
material. Alan was notably helpful in suggesting changes that made the tutorial flow more easily. I am
grateful to Chris Wroe and Robert Stevens who conceived the original idea of basing the tutorial on an
ontology about pizzas. I would especially like to thank Holger Knublauch from Stanford Univeristy who is
the developer of the Protégé-OWL plugin. Holger was always on hand to answer questions and provided
feedback and input about what the tutorial should cover. Finally, I would also like to thank Natasha
Noy from Stanford University for using her valuable experience in teaching, creating and giving tutorials
about Protégé to provide detailed and useful comments about how initial drafts of the tutorial/guide
could be made better.
This work was supported in part by the CO-ODE project funded by the UK Joint Information Services
Committee and the HyOntUse Project (GR/S44686) funded by the UK Engineering and Physical Science
Research Council and by 21XS067A from the National Cancer Institute.
http://www.co-ode.org
6

Chapter 1
Introduction
This guide introduces the Protégé-OWL plugin for creating OWL ontologies. Chapter 3 gives a brief
overview of the OWL ontology language. Chapter 4 focuses on building an OWL-DL ontology and using
a Description Logic Reasoner to check the consistency of the ontology and automatically compute the
ontology class hierarchy. Chapter 6 describes some OWL constructs such as hasValue Restrictions and
Enumerated classes, which aren’t directly used in the main tutorial. Chapter 7 describes Namespaces,
Importing ontologies and various features and utilities of the Protégé-OWL application.
1.1 Conventions
Class, property and individual names are written in a sans serif font like this.
Names for user interface widget are presented in a style ‘like this’.
Where exercises require information to be typed into Protégé-OWL a type writer font is used like this.
Exercises and required tutorial steps are presented like this:
Exercise 1: Accomplish this
1. Do this.
2. Then do this.
3. Then do this.
8

Tips and suggestions related to using Prot´eg´e-OWL and building ontologies are
presented like this.
Explanation as to what things mean are presented like this.
Potential pitfalls and warnings are presented like this.
General notes are presented like this.
Vocabulary explanations and alternative names are presented like this.
9

Chapter 2
Requirements
In order to follow this tutorial you must have Protégé 2.1 (or later)1
, the Protégé-OWL plugin (latest beta)
and also the OWL Wizards Plugin, which are available via the CO-ODE web site 2
. Since the release
of Protégé 2.1, the Protégé-OWL plugin and the OWL Wizards are bundled in one single download.
It is also recommended (but not necessary) to use the OWLViz plugin, which allows the asserted and
inferred classification hierarchies to be visualised, and is available from the CO-ODE web site, or can
be installed when Protégé 2.1 is installed. For installation steps, please see the documentation for each
component. Finally, it is necessary to have a DIG (Description Logic Implementers Group) compliant
reasoner installed in order to compute subsumption relationships between classes, and detect inconsistent
classes. It is recommended that the latest version of the RACER reasoner be used, which can be obtained
from http://www.sts.tu-harburg.de/~r.f.moeller/racer/.
1http://protege.stanford.edu
2http://www.co-ode.org
10

Chapter 3
What are OWL Ontologies?
Ontologies are used to capture knowledge about some domain of interest. An ontology describes the
concepts in the domain and also the relationships that hold between those concepts. Different ontology
languages provide different facilities. The most recent development in standard ontology languages is
OWL from the World Wide Web Consortium (W3C)1
. Like Protégé OWL makes it possible to describe
concepts but it also provides new facilities. It has a richer set of operators - e.g. and, or and negation. It is
based on a different logical model which makes it possible for concepts to be defined as well as described.
Complex concepts can therefore be built up in definitions out of simpler concepts. Furthermore, the
logical model allows the use of a reasoner which can check whether or not all of the statements and
definitions in the ontology are mutually consistent and can also recognise which concepts fit under which
definitions. The reasoner can therefore help to maintain the hierarchy correctly. This is particularly
useful when dealing with cases where classes can have more than one parent.
3.1 The Three Species Of OWL
OWL ontologies may be categorised into three species or sub-languages: OWL-Lite, OWL-DL and OWL-
Full. A defining feature of each sub-language is its expressiveness. OWL-Lite is the least expressive
sub-langauge. OWL-Full is the most expressive sub-language. The expressiveness of OWL-DL falls
between that of OWL-Lite and OWL-Full. OWL-DL may be considered as an extension of OWL-Lite
and OWL-Full an extension of OWL-DL.
3.1.1 OWL-Lite
OWL-Lite is the syntactically simplest sub-language. It is intended to be used in situations where only
a simple class hierarchy and simple constraints are needed. For example, it is envisaged that OWL-Lite
will provide a quick migration path for existing thesauri and other conceptually simple hierarchies.
1http://www.w3.org/TR/owl-guide/
11

3.1.2 OWL-DL
OWL-DL is much more expressive than OWL-Lite and is based on Description Logics (hence the suffix
DL). Description Logics are a decidable fragment of First Order Logic2
and are therefore amenable to
automated reasoning. It is therefore possible to automatically compute the classification hierarchy3
and
check for inconsistencies in an ontology that conforms to OWL-DL. This tutorial focuses on OWL-
DL.
3.1.3 OWL-Full
OWL-Full is the most expressive OWL sub-language. It is intended to be used in situations where very
high expressiveness is more important than being able to guarantee the decidability or computational
completeness of the language. It is therefore not possible to perform automated reasoning on OWL-Full
ontologies.
3.1.4 Choosing The Sub-Language To Use
For a more detailed synopsis of the three OWL sub-languages see the OWL Web Ontology Language
Overview4
. Although many factors come into deciding the appropriate sub-language to use, there are
some simple rules of thumb.
• The choice between OWL-Lite and OWL-DL may be based upon whether the simple constructs of
OWL-Lite are sufficient or not.
• The choice between OWL-DL and OWL-Full may be based upon whether it is important to be
able to carry out automated reasoning on the ontology or whether it is important to be able to use
highly expressive and powerful modelling facilities such as meta-classes (classes of classes).
The Protégé-OWL plugin does not make the distinction between editing OWL-Lite and OWL-DL on-
tologies. It does however offer the option to constrain the ontology being edited to OWL-DL, or allow
the expressiveness of OWL-Full — See section 7.1 for more information on how to constrain the ontology
to OWL-DL.
3.2 Components of OWL Ontologies
OWL ontologies have similar components to Protégé frame based ontologies. However, the terminology
used to describe these components is slightly different from that used in Protégé . An OWL ontology
consists of Individuals, Properties, and Classes, which roughly correspond to Protégé Instances, Slots and
Classes.
2Logics are decidable if computations/algorithms based on the logic will terminate in a finite time.
3Also known as subsumption reasoning.
4http://www.w3.org/TR/owl-features
12

Matthew Gemma
England
Italy
USA
Fluffy
Fido
Figure 3.1: Representation Of Individuals
3.2.1 Individuals
Individuals, represent objects in the domain that we are interested in5
. An important difference between
Protégé and OWL is that OWL does not use the Unique Name Assumption (UNA). This means that
two different names could actually refer to the same individual. For example, “Queen Elizabeth”, “The
Queen” and “Elizabeth Windsor” might all refer to the same individual. In OWL, it must be explicitly
stated that individuals are the same as each other, or different to each other — otherwise they might be
the same as each other, or they might be different to each other. Figure 3.1 shows a representation of
some individuals in some domain – in this tutorial we represent individuals as diamonds in diagrams.
Individuals are also known as instances. Individuals can be referred to as being
‘instances of classes’.
3.2.2 Properties
Properties are binary relations6
on individuals - i.e. properties link two individuals together7
. For
example, the property hasSibling might link the individual Matthew to the individual Gemma, or the
property hasChild might link the individual Peter to the individual Matthew. Properties can have inverses.
For example, the inverse of hasOwner is isOwnedBy. Properties can be limited to having a single value –
i.e. to being functional. They can also be either transitive or symmetric. These ‘property characteristics’
are explained in detail section 4.8. Figure 3.2 shows a representation of some properties linking some
individuals together.
Properties are roughly equivalent to slots in Protégé . They are also known as
roles in description logics and relations in UML and other object oriented notions.
In GRAIL and some other formalisms they are called attributes.
5Also known as the domain of discourse.
6A binary relation is a relation between two things.
7Strictly speaking we should speak of ‘instances of properties’ linking individuals, but for the sake of brevity we will
keep it simple.
13

Matthew Gemma
England
live
sIn
hasSibling
Figure 3.2: Representation Of Properties
Matthew
Gemma England
Italy
USA
Fluffy
Fido
livesInCountr
yhasPet
hasSibling
Pet
Country
Person
Figure 3.3: Representation Of Classes (Containing Individuals)
3.2.3 Classes
OWL classes are interpreted as sets that contain individuals. They are described using formal (math-
ematical) descriptions that state precisely the requirements for membership of the class. For example,
the class Cat would contain all the individuals that are cats in our domain of interest.8
Classes may be
organised into a superclass-subclass hierarchy, which is also known as a taxonomy. Subclasses specialise
(‘are subsumed by’) their superclasses. For example consider the classes Animal and Cat – Cat might
be a subclass of Animal (so Animal is the superclass of Cat). This says that, ‘All cats are animals’, ‘All
members of the class Cat are members of the class Animal’, ‘Being a Cat implies that you’re an Animal’,
and ‘Cat is subsumed by Animal’. One of the key features of OWL-DL is that these superclass-subclass
relationships (subsumption relationships) can be computed automatically by a reasoner – more on this
later. Figure 3.3 shows a representation of some classes containing individuals – classes are represented
as circles or ovals, rather like sets in Venn diagrams.
The word concept is sometimes used in place of class. Classes are a concrete
representation of concepts.
In OWL classes are built up of descriptions that specify the conditions that must be satisﬁed by an
individual for it to be a member of the class. How to formulate these descriptions will be explained as
8Individuals may belong to more than one class.
14

Chapter 4
Building An OWL Ontology
This chapter describes how to create an ontology of Pizzas. We use Pizzas because we have found them
to provide many useful examples.1
Exercise 2: Create a new OWL project
1. Start Protégé
2. When the New Project dialog box appears, select ‘OWL Files’ from the ‘Project
Format’ list section on the left hand side of the dialog box, and press ‘New’.
After a short amount of time, a new empty Protégé-OWL project will have been created.
4.1 Named Classes
When Protégé-OWL starts the OWLClasses tab shown in Figure 4.1 will be visible. The initial class
hierarchy tree view should resemble the picture shown in Figure 4.2. The empty ontology contains one
class called owl:Thing. As mentioned previously, OWL classes are interpreted as sets of individuals (or
sets of objects). The class owl:Thing is the class that represents the set containing all individuals. Because
of this all classes are subclasses of owl:Thing.2
Let’s add some classes to the ontology in order to define what we believe a pizza to be.
1The Ontology that we will created is based upon a Pizza Ontology that has been used as the basis for a course on
editing DAML+OIL ontologies in OilEd (http://oiled.man.ac.uk), which was taught at the University Of Manchester.
2owl:Thing is part of the OWL Vocabulary, which is defined by the ontology located at http://www.w3.org/2002/07/
owl/#
16

Figure 4.1: The Classes Tab
Create subclass (of
the selected class)
Delete selected class
Figure 4.2: The Class Hierarchy Pane
17

Figure 4.3: Class Name Widget
Exercise 3: Create classes Pizza, PizzaTopping and PizzaBase
1. Press the ‘Create subclass’ button shown in Figure 4.2. This button creates a new
class as a subclass of the selected class (in this case we want to create a subclass of
owl:Thing).
2. Rename the class using the ‘Class name widget’ which is located to the right of the
class hierarchy (shown in Figure 4.3) to Pizza and hit return.
3. Repeat the previous steps to add the classes PizzaTopping and also PizzaBase, en-
suring that owl:Thing is selected before the ‘Create subclass’ button is pressed so
that the classes are created as subclasses of owl:Thing.
The class hierarchy should now resemble the hierarchy shown in Figure 4.4.
A class hierarchy may also be called a taxonomy.
Although there are no mandatory naming conventions for OWL classes, we recom-
mend that all class names should start with a capital letter and should not contain
spaces. (This kind of notation is known as CamelBack notation and is the nota-
tion used in this tutorial). For example Pizza, PizzaTopping, MargheritaPizza.
Alternatively, you can use underscores to join words. For example Pizza Topping.
Which ever convention you use, it is important to be consistent.
4.2 Disjoint Classes
Having added the classes Pizza, PizzaTopping and PizzaBase to the ontology, we now need to say these
classes are disjoint, so that an individual (or object) cannot be an instance of more than one of these
three classes. To specify classes that are disjoint from the selected class the ‘Disjoints widget’ which is
18

Figure 4.4: The Initial Class Hierarchy
Create disjoint class from
OWL expression
Add disjointclass
Add all siblings
Remove selected class from list
Delete selected row
Remove all siblings
Figure 4.5: The Disjoint Classes Widget
located in the lower right hand corner of the ‘OWLClasses’ tab is used. (See Figure 4.5).
Exercise 4: Make Pizza, PizzaTopping and PizzaBase disjoint from each other
1. Select the class Pizza in the class hierarchy.
2. Press the ‘Add siblings’ button on the disjoint classes widget. This will make Pizz-
aBase and PizzaTopping (the sibling classes of Pizza) disjoint from Pizza.
Notice that the disjoint classes widget now displays PizzaTopping and PizzaBase. Select the class
PizzaBase. Notice that the disjoint classes widget displays the classes that are now disjoint to PizzaBase,
namely Pizza and PizzaTopping.
19

Figure 4.6: Add Group Of Classes Wizard: Select class page
Figure 4.7: Add Group Of Classes Wizard: Enter classes page
OWL Classes are assumed to ‘overlap’. We therefore cannot assume that an
individual is not a member of a particular class simply because it has not been
asserted to be a member of that class. In order to ‘separate’ a group of classes
we must make them disjoint from one another. This ensures that an individual
which has been asserted to be a member of one of the classes in the group cannot
be a member of any other classes in that group. In our above example Pizza,
PizzaTopping and PizzaBase have been made disjoint from one another. This
means that it is not possible for an individual to be a member of a combination
of these classes – it would not make sense for an individual to be a Pizza and a
PizzaBase!
20

4.3 Using The OWL Wizards To Create Classes
The OWL Wizards plugin, which is available from the Protégé web site, is an extensible set of Wizards
that are designed to make carrying out common, repetitive and time consuming tasks easy. In this section
we will use the ‘Create A Group Of Classes’ wizard to add some subclasses of the class PizzaBase.
To use the OWL Wizards you must ensure that the OWL Wizards plugin in installed and configured in
Protégé .
Exercise 5: Use the ‘Create Group Of Classes’ Wizard to create ThinAndCrispy and DeepPan as
subclasses of PizzaBase
1. Select the class PizzaBase in the class hierachy.
2. From the Wizards menu on the Protégé menu bar select the item ‘Create Group Of
Classes’.
3. The Wizard shown in Figure 4.6 will appear. Since we preselected the PizzaBase
class, the first radio button at the top of the Wizard should be prompting us to create
the classes under the class PizzaBase. If we had not preselected PizzaBase before
starting the Wizard, then the tree could be used to select the class.
4. Press the ‘Next’ button on the Wizard—The page shown in Figure 4.7 will be dis-
played. We now need to tell the Wizard the subclasses of PizzaBase that we want to
create. In the large text area, type in the class name ThinAndCrispyBase (for a thin
based pizza) and hit return. Also enter the class name DeepPanBase so that the page
resembles that shown in Figure 4.7 .
5. Hit the ‘Next’ button on the Wizard. The Wizard checks that the names entered
adhere to the naming styles that have previously been mentioned (No spaces etc.). It
also checks for uniqueness – no two class names may be the same. If there are any
errors in the class names, they will be be presented on this page, along with suggestions
for corrections.
6. Hit the ‘Next’ button on the Wizard. Ensure the tick box ‘Make all new classes
disjoint’ is ticked — instead of having to use the disjoint classes widget, the Wizard
will automatically make the new classes disjoint for us.
7. Hit the ‘Next’ button to display the annotations page. Here we could add annotations
if we wanted to. Most commonly annotations are used to record editorial information
about the ontology – who created it, when it was created, when it was revised, etc.
The basic OWL annotation properties are selectable by default. For now, we will not
add any annotations, so just hit the ‘Finish’ button.
21

If we had imported the Dublin Core ontology (see section 7.2.5) then the Dublin
Core annotation properties would have been available to annotate our classes in
step 7 Exercise 5 . Dublin Core is a set of metadata elements that, in our case,
can be used to annotate various elements of an ontology with information such as
‘creator’, ‘date’, ‘language’ etc. For more information see http://dublincore.
org/a
.
aAn ontology will be put into OWL-Full if the ontologies that are available on the Dublin Core
website are imported. We recommend that an OWL-DL version of the Dublin Core ontology
which is located in the Protégé ontology library is imported — details of this can be found in
section 7.2.5
After the ‘Finish’ button has been pressed, the Wizard creates the classes, makes them disjoint, and
selects them in the Protégé OWLClasses tab. The ontology should now have ThinAndCrispyBase and
also DeepPanBase as subclasses of PizzaBase. These new classes should also be disjoint to each other.
Hence, a pizza base cannot be both thin and crispy and deep pan. It isn’t difficult to see that if we had
a lot of classes to add to the ontology, the Wizard would dramatically speed up the process of adding
them.
On page two of the ‘Create group of classes wizard’ the classes to be created
are entered. If we had a lot of classes to create that had the same prefix or suffix
we could use the options to auto prepend and auto append text to the class names
that we entered.
Creating Some Pizza Toppings
Now that we have some basic classes, let’s create some pizza toppings. In order to be useful later on the
toppings will be grouped into various categories — meat toppings, vegetable toppings, cheese toppings
22

and seafood toppings.
Exercise 6: Create some subclasses of PizzaTopping
1. Select the class PizzaTopping in the class hierarchy.
2. Use the OWL Wizards to add some subclasses of PizzaTopping called MeatTopping,
VegetableTopping, CheeseTopping and also SeafoodTopping. Make sure that these
classes are disjoint to each other.
3. Next add some different kinds of meat topping. Select the class MeatTopping, and use
the ‘Create Group Of Classes’ Wizard to add the following subclasses of Meat-
Topping: SpicyBeefTopping, PepperoniTopping, SalamiTopping, HamTopping. Once
again, ensure that the classes are created as disjoint classes.
4. Add some different kinds of vegetable toppings by creating the following disjoint sub-
classes of VegetableTopping: TomatoTopping, OliveTopping, MushroomTopping, Pep-
perTopping, OnionTopping and CaperTopping. Add further subclasses of PepperTop-
ping: RedPepperTopping, GreenPepperTopping and JalapenoPepperTopping making
sure that the subclasses of PepperTopping are disjoint
5. Now add some different kinds of cheese toppings. In the same manner as before, add
the following subclasses of CheeseTopping, ensuring that the subclasses are disjoint
to each other: MozzarellaTopping, and ParmezanTopping
6. Finally, add some subclasses of SeafoodTopping to represent different kinds of sea
food: TunaTopping, AnchovyTopping and PrawnTopping.
The class hierarchy should now look similar to that shown in Figure 4.8 (the ordering of classes may be
slightly different).
Up to this point, we have created some simple named classes, some of which
are subclasses of other classes. The construction of the class hierarchy may have
seemed rather intuitive so far. However, what does it actually mean to be a sub-
class of something in OWL? For example, what does it mean for VegetableTopping
to be a subclass of PizzaTopping, or for TomatoTopping to be a subclass of Veg-
etableTopping? In OWL subclass means necessary implication. In other words,
if VegetableTopping is a subclass of PizzaTopping then ALL instances of Veg-
etableTopping are instances of PizzaTopping, without exception — if something is
a VegetableTopping then this implies that it is also a PizzaTopping as shown in
Figure 4.9.a
aIt is for this reason that we seemingly pedantically named all of our toppings with the suffix
of ‘Topping’, for example, HamTopping. Despite the fact that class names themselves carry no
formal semantics in OWL (and in other ontology languages), if we had named HamTopping Ham,
then this could have implied to human eyes that anything that is a kind of ham is also a kind of
MeatTopping and also a PizzaTopping.
23

Figure 4.8: Class Hierarchy
24

PizzaTopping
VegetableTopping
TomatoTopping
Figure 4.9: The Meaning Of Subclass — All individuals that are members of the class TomatoTopping are
members of the class VegetableTopping and members of the class PizzaTopping as we have stated
that TomatoTopping is a subclass of VegetableTopping which is a subclass of PizzaTopping
4.4 OWL Properties
OWL Properties represent relationships between two individuals. There are two main types of properties,
Object properties and Datatype properties. Object properties link an individual to an individual. Datatype
properties link an individual to an XML Schema Datatype value3
or an rdf literal4
. OWL also has a
third type of property – Annotation properties5
. Annotation properties can be used to add information
(metadata — data about data) to classes, individuals and object/datatype properties. Figure 4.10 depicts
an example of each type of property.
Properties may be created using the ‘Properties’ tab shown in Figure 4.11. It is also possible to create
properties using the ‘Properties Widget’ shown in Figure 4.12 which is located on the ‘OWLClasses’
tab. Figure 4.13 shows the buttons located in the top left hand corner of the ‘Properties’ tab that
are used for creating OWL properties. As can be seen from Figure 4.13, there are buttons for creat-
ing Datatype properties, Object properties and Annotation properties. Most properties created in this
tutorial will be Object properties.
Exercise 7: Create an object property called hasIngredient
1. Switch to the ‘Propterties’ tab. Use the ‘Create Object Property’ button (see
Figure 4.13 – second button on the left) to create a new Object property. An Object
property with a generic name will be created.
2. Rename the property to hasIngredient as shown in Figure 4.14 (The ‘Property Name
Widget’.
3See http://www.w3.org/TR/xmlschema-2/ for more information on XML Schema Datatypes
4RDF = Resource Description Framework. See http://www.w3.org/TR/rdf-primer/ for an excellent introduction to
RDF.
5Object properties and Datatype properties may be marked as Annotation properties
25

hasSister
Matthew Gemma
An object property linking the individual
Matthew to the individual Gemma
A datatype property linking the individual
Matthew to the data literal ‘25’,which has a type
of an xml:integer.
An annotation property,linking the class ‘JetEngine’
to the data literal (string) ‘’Matthew Horridge’’.
hasAge
Matthew “25”^^xsd:integer
dc:creator
JetEngine ‘‘Matthew Horridge’’
Figure 4.10: The Diﬀerent types of OWL Properties
26

Figure 4.11: The PropertiesTab
Figure 4.12: The Properties Widget
27

Create datatype property
Create object property
Create datatype annotation property
Create object annotation property
Delete selected property
Figure 4.13: Property Creation Buttons — located on the Properties Tab above the property list/tree
Figure 4.14: Property Name Widget
Although there is no strict naming convention for properties, we recommend that
property names start with a lower case letter, have no spaces and have the re-
maining words capitalised. We also recommend that properties are preﬁxed with
the word ‘has’, or the word ‘is’, for example hasPart, isPartOf, hasManufacturer,
isProducerOf. Not only does this convention help make the intent of the property
clearer to humans, it is also taken advantage of by the ‘English Prose Tooltip
Generator’a
, which uses this naming convention where possible to generate more
human readable expressions for class descriptions.
aThe English Prose Tooltip Generator displays the description of classes etc. in a more natural
form of English, making is easy to understand a class description. The tooltips pop up when the
mouse pointer is made to hover over a class description in the user interface.
Having added the hasIngredient property, we will now add two more properties — hasTopping, and
hasBase. In OWL, properties may have sub properties, so that it is possible to form hierarchies of
properties. Sub properties specialise their super properties (in the same way that subclasses specialise
their superclasses). For example, the property hasMother might specialise the more general property of
hasParent. In the case of our pizza ontology the properties hasTopping and hasBase should be created
as sub properties of hasIngredient. If the hasTopping property (or the hasBase property) links two
28

individuals this implies that the two individuals are related by the hasIngredient property.
Exercise 8: Create hasTopping and hasBase as sub-properties of hasIngredient
1. To create the hasTopping property as a sub property of the hasIngredient property,
right click (or ctrl click on the Mac) on the hasIngredient property in the property
hierarchy on the ‘Properties’ tab. The menu shown in Figure 4.15 will pop up.
2. Select the ‘Create subproperty’ item from the popup menu. A new object property
will be created as a sub property of the hasIngredient property.
3. Rename the new property to hasTopping.
4. Repeat the above steps but name the property hasBase.
Figure 4.15: PropertyHierarchyMenu
Note that it is also possible to create sub properties of datatype properties. However, it is not possible to
mix and match object properties and datatype properties with regards to sub properties. For example, it
is not possible to create an object property that is the sub property of a datatype property and vice-versa.
4.5 Inverse Properties
Each object property may have a corresponding inverse property. If some property links individual a to
individual b then its inverse property will link individual b to individual a. For example, Figure 4.16
shows the property hasParent and its inverse property hasChild — if Matthew hasParent Jean, then
because of the inverse property we can infer that Jean hasChild Matthew.
Inverse properties can be created/speciﬁed using the inverse property widget shown in Figure 4.17. For
29

hasParent
Matthew JeanhasChild
Figure 4.16: An Example Of An Inverse Property: hasParent has an inverse property that is hasChild
completeness we will specify inverse properties for our existing properties in the Pizza Ontology.
Exercise 9: Create some inverse properties
1. Use the ‘Create object property’ button on the ‘Properties’ tab to create a new
Object property called isIngredientOf (this will become the inverse property of hasIn-
gredient).
2. Press the ‘Assign existing property’ button on the inverse property widget shown
in Figure 4.17. This will display a dialog from which properties may be selected.
Select the hasIngredient property and press ‘OK’. The property hasIngredient should
now be displayed in the ‘Inverse Property’ widget. The properties hierarchy should
also now indicate that hasIngredient and isIngredientOf are inverse properties of each
other.
3. Select the hasBase property.
4. Press the ‘Create new inverse property’ button on the ‘Inverse Property’ widget.
This will pop up a dialog that contains information about the newly created property.
Use this dialog to rename the property isBaseOf and the close the dialog window
(using the operating system close window button on the title bar). Notice that the
isBaseOf property has been created as a sub property of the isIngredientOf property.
This corresponds to the fact that hasBase is a sub property of hasIngredient, and
isIngredientOf is the inverse property of hasIngredient.
5. Select the hasTopping property.
6. Press the ‘Create new inverse property’ button on the ‘Inverse Property’ widget.
Use the property dialog that pops up to rename the property isToppingOf. Close the
dialog. Notice that isToppingOf has been created as a sub property of isIngredientOf.
The property hierarchy should now look like the picture shown in Figure 4.18. Notice the ‘bidirectional’
arrows that indicate inverse properties.
4.6 OWL Property Characteristics
OWL allows the meaning of properties to be enriched through the use of property characteristics. The
following sections discuss the various characteristics that properties may have:
30

Create new inverse property
(Create a new property and
assigns it to be the inverse of
the current property).
Assign existing property
(Assign an existing property
to be the inverse property of
the current property).
Unassign current inverse
property
Figure 4.17: The Inverse Property Widget
Figure 4.18: The Property Hierarchy With Inverse Properties
31

4.6.1 Functional Properties
If a property is functional, for a given individual, there can be at most one individual that is related to
the individual via the property. Figure 4.19 shows an example of a functional property hasBirthMother
— something can only have one birth mother. If we say that the individual Jean hasBirthMother Peggy
and we also say that the individual Jean hasBirthMother Margaret6
, then because hasBirthMother is a
functional property, we can infer that Peggy and Margaret must be the same individual. It should be
noted however, that if Peggy and Margaret were explicitly stated to be two diﬀerent individuals then the
above statements would lead to an inconsistency.
Margaret
Peggy
Jean
hasBirthMother
hasBirthMother
Implies Peggy and Margaret
are the same individual
Figure 4.19: An Example Of A Functional Property: hasBirthMother
Functional properties are also known as single valued properties and also features.
4.6.2 Inverse Functional Properties
If a property is inverse functional then it means that the inverse property is functional. For a given
individual, there can be at most one individual related to that individual via the property. Figure 4.20
shows an example of an inverse functional property isBirthMotherOf. This is the inverse property of
hasBirthMother — since hasBirthMother is functional, isBirthMotherOf is inverse functional. If we state
that Peggy is the birth mother of Jean, and we also state that Margaret is the birth mother of Jean,
then we can infer that Peggy and Margaret are the same individual.
Margaret
Peggy
JeanImplies same individual
isBirthMotherOf
isBirthMotherOf
Figure 4.20: An Example Of An Inverse Functional Property: isBirthMotherOf
4.6.3 Transitive Properties
If a property is transitive, and the property relates individual a to individual b, and also individual b to
individual c, then we can infer that individual a is related to individual c via property P. For example,
Figure 4.21 shows an example of the transitive property hasAncestor. If the individual Matthew has an
6The name Peggy is a diminutive form for the name Margaret
32

ancestor that is Peter, and Peter has an ancestor that is William, then we can infer that Matthew has an
ancestor that is William – this is indicated by the dashed line in Figure 4.21.
Matthew
Peter
William
hasA
ncestor
has
Ancestor
hasAncestor
Figure 4.21: An Example Of A Transitive Property: hasAncestor
4.6.4 Symmetric Properties
If a property P is symmetric, and the property relates individual a to individual b then individual b is
also related to individual a via property P. Figure 4.22 shows an example of a symmetric property. If the
individual Matthew is related to the individual Gemma via the hasSibling property, then we can infer
that Gemma must also be related to Matthew via the hasSibling property. In other words, if Matthew
has a sibling that is Gemma, then Gemma must have a sibling that is Matthew. Put another way, the
property is its own inverse property.
hasSibling
Matthew GemmahasSibling
Figure 4.22: An Example Of A Symmetric Property: hasSibling
We want to make the hasIngredient property transitive, so that for example if a pizza topping has an
ingredient, then the pizza itself also has that ingredient. To set the property characteristics of a property
the property characteristics widget shown in Figure 4.23 which is located in the lower right hand corner
of the properties tab is used.
Exercise 10: Make the hasIngredient property transitive
1. Select the hasIngredient property in the property hierarchy on the ‘Properties’ tab.
2. Tick the ‘Transitive’ tick box on the ‘Property Characteristics Widget’.
3. Select the isIngredientOf property, which is the inverse of hasIngredient. Ensure that
the transitive tick box is ticked.
33

Unticking this box makes a property functional
Ticking this box makes a property Invers Functional
An inverse property may be specified here. Use the ‘O’button to create
a new property and simultaneously make the new property the inverse.
Use the ‘O+’button to add an existing property as the inverse.
Checking this box makes the property symmetric
Checking this box makes the property transitive
Figure 4.23: Property Characteristics Widgets
If a property is transitive then its inverse property should also be transitive.a
aAt the time of writing this must be done manually in Prot´eg´e-OWL . However, the reasoner
will assume that if a property is transitive, its inverse property is also a transitive.
Note that if a property is transitive then it cannot be functional.a
aThe reason for this is that transitive properties, by their nature, may form ‘chains’ of indi-
viduals. Making a transitive property functional would therefore not make sense.
We now want to say that out pizza can only have one base. There are numerous ways that this could be
accomplished. However, to do this we will make the hasBase property functional, so that it may have
only one value for a given individual.
Exercise 11: Make the hasBase property functional
2. Click the ‘Allows multiple values’ tick box on the ‘Property Charachteristics
Widget’ so that it is unticked.
34

hasTopping
isToppingOf
hasTopping
isToppingOf
hasTopping
isToppingOf
Pizza PizzaTopping
Figure 4.24: The domain and range for the hasTopping property and its inverse property isToppingOf. The
domain of hasTopping is Pizza the range of hasTopping is PizzaTopping — the domain and range
for isToppingOf are the domain and range for hasTopping swapped over
If a datatype property is selected, the property characteristics widget will be re-
duced so that only options for ‘Allows multiple values’ and ‘Inverse Func-
tional’ will be displayed. This is because OWL-DL does not allow datatype
properties to be transitive, symmetric or have inverse properties.
4.7 Property Domains and Ranges
Properties may have a domain and a range speciﬁed. Properties link individuals from the domain to
individuals from the range. For example, in our pizza ontology, the property hasTopping would probably
link individuals belonging the the class Pizza to individuals belonging to the class of PizzaTopping. In
this case the domain of the hasTopping property is Pizza and the range is PizzaTopping — this is
depicted in Figure 4.24.
Property Domains And Ranges In OWL — It is important to realise that in
OWL domains and ranges should not be viewed as constraints to be checked. They
are used as ‘axioms’ in reasoning. For example if the property hasTopping has the
domain set as Pizza and we then applied the hasTopping property to IceCream
(individuals that are members of the class IceCream), this would generally not
result in an error. It would be used to infer that the class IceCream must be a
subclass of Pizza! a
.
aAn error will only be generated (by a reasoner) if Pizza is disjoint to IceCream
We now want to specify that the hasTopping property has a range of PizzaTopping. To do this the range
widget shown in Figure 4.25 is used. By default the drop down box displays ‘Instance’, meaning that the
35

Create a class expression as the range
Add a (existing) named class as the range
Remove the selected named class from the range
Remove (delete) the selected class expression from the range
Select the ‘type’of range (Instance or class)
Figure 4.25: Property Range Widget (For Object Properties)
property links instances of classes to instances of classes.
Exercise 12: Specify the range of hasTopping
1. Make sure that the hasTopping property is selected in the property hierarchy on the
‘Properties’ tab.
2. Press the ‘Add named class’ button on the ‘Range Widget’ (Figure 4.25). A
dialog will appear that allows a class to be selected from the ontology class hierarchy.
3. Select PizzaTopping and press the ‘OK’ button. PizzaTopping should now be dis-
played in the range list.
It is also possible, but usually incorrect, to specify that a class rather than its
individuals is the range of a property. This is done by selecting ‘class’ in the drop
down box on the range widget. It is a common mistake to believe that the range
of a property is a class when the range is really the individuals that are members
of the class. Specifying the range of a property as a class treats the class itself as
an individuala
. This is a kind of ‘meta-statement’ and causes the ontology to be
in OWL-Full.
aRemember that instances of object properties link individuals to individuals
36

It is possible to specify multiple classes as the range for a property. If multiple
classes are specified in Protégé-OWL the range of the property is interpreted to
be the union of the classes. For example, if the range of a property has the classes
Man and Woman listed in the range widget, the range of the property will be
interpreted as Man union Woman. a
aSee section B.2 for an explanation of what a union class is.
To specify the domain of a property the domain widget shown in Figure 4.26 is used.
Exercise 13: Specify Pizza as the domain of the hasTopping property
1. Make sure that the hasTopping property is selected in the property hierarchy on the
‘Properties’ tab.
2. Press the ‘Add named class’ button on the Domain Widget. A dialog will appear
that allows a class to be selected from the ontology class hierarchy.
3. Select Pizza and press the OK button. Pizza should now be displayed in the domain
list.
This means that individuals that are used ‘on the left hand side’ of the hasTopping
property will be inferred to be members of the class Pizza. Any individuals that
are used ‘on the right hand side’ of the hasTopping property will be inferred to
be members of the class PizzaTopping. For example, if we have individuals a and
b and an assertion of the form a hasTopping b then it will be inferred that a is a
member of the class Pizza and that b is a member of the class PizzaToppinga
.
aThis will be the case even if a has not been asserted to be a member of the class Pizza and/or
b has not been asserted to be a member of the class PizzaTopping.
When multiple classes are specified as the domain for a property, Protégé-OWL
interprets the domain of the property to be the union of the classes .
Although OWL allows arbitrary class expressions to be used for the domain of a
property this is not allowed when editing ontologies in Protégé-OWL .
37

Add a (existing) named class to the domain
Removes the selected class from the domain
Figure 4.26: Property Domain Widget
We now need to ﬁll in the domain and range for the inverse of the hasTopping property isToppingOf:
Exercise 14: Specify the domain and range for the isToppingOf property
1. Select the isToppingOf property.
2. Use the same steps as above to set the domain of the isToppingOf property to Piz-
zaTopping.
3. Set the range of the isToppingOf property to Pizza.
Notice that the domain of the isToppingOf property is the range of the inverse property hasTopping, and
that the range of the isToppingOf property is the domain of the hasTopping property. This is depicted
in 4.24.
Exercise 15: Specify the domain and range for the hasBase property and its inverse property is-
BaseOf
2. Specify the domain of the hasBase property as Pizza.
3. Specify the range of the hasBase property as PizzaBase.
4. Select the isBaseOf property.
5. Make the domain of the isBaseOf property PizzaBase.
6. Make the range of the isBaseOf property Pizza.
38

In the previous steps we have ensured that the domains and ranges for properties
are also set up for inverse properties in a correct manner. In general, domain for
a property is the range for its inverse, and the range for a property is the domain
for its inverse — Figure 4.24 illustrates this for the hasTopping and isToppingOf.
If these steps are not taken the ontology tests (see section 7.3) can be used to spot
any discrepancies.
Although we have specified the domains and ranges of various properties for the
purposes of this tutorial, we generally advise against doing this. The fact that
domain and range conditions do not behave as constraints and the fact that they
can cause ‘unexpected’ classification results can lead problems and unexpected
side effects. These problems and side effects can be particularly difficult to track
down in a large ontology.
4.8 Describing And Defining Classes
Having created some properties we can now use these properties to describe and define our Pizza Ontology
classes.
4.8.1 Property Restrictions
In OWL properties are used to create restrictions. As the name may suggest, restrictions are used to
restrict the individuals that belong to a class. Restrictions in OWL fall into three main categories:
• Quantifier Restrictions
• Cardinality Restrictions
• hasValue Restrictions.
We will initially use quantifier restrictions. These types of restrictions are composed of a quantifier, a
property, and a filler. The two quantifiers that may be used are:
• The existential quantifier (∃), which can be read as at least one, or some7
.
• The universal quantifier (∀), which can be read as only8
.
For example the restriction ∃ hasTopping MozzarellaTopping is made up of the existential quantifier
(∃), the property hasTopping, and the filler MozzarellaTopping. This restriction describes the set, or the
7It can also be read as ‘someValuesFrom’ in OWL speak.
8It can also be read as ‘allValuesFrom’ in OWL speak.
39

hasTopping
hasTopping
hasTopping
hasTopping
Mozzarella
Things that have at least one
MozzarellaTopping
(∃ hasTopping Mozzarella)
Figure 4.27: The Restriction ∃ hasTopping Mozzarella. This restriction describes the class of individuals that
have at least one topping that is Mozzarella
class, of individuals that have at least one topping that is an individual from the class MozzarellaTopping.
This restriction is depicted in Figure 4.27 — The diamonds in the Figure represent individuals. As can
be seen from Figure 4.27, the restriction describes an anonymous (unnamed) class of individuals that
satisfy the restriction.
A restriction actually describes an anonymous class (an unnamed class). The
anonymous class contains all of the individuals that satisfy the restriction (see
Appendix A for further information about what a restriction actually represents
and an explanation of existential and universal quantification). When restrictions
are used to describe classes, they actually specify anonymous superclasses of the
class being described. For example, we could say that MargheritaPizza is a sub-
class of, amongst other things, Pizza and also a subclass of the things that have
at least one topping that is MozzarellaTopping.
The restrictions for a class are displayed and edited using the ‘Conditions Widget’ shown in Figure
4.28. The ‘Conditions Widget’ is the ‘heart of’ the ‘OWLClasses’ tab in protege, and holds virtually
all of the information used to describe a class. At first glance, the ‘Conditions Widget’ may seem
complicated, however, it will become apparent that it is an incredibly powerful way of describing and
defining classes.
Restrictions are used in OWL class descriptions to specify anonymous superclasses of the class being
described.
4.8.2 Existential Restrictions
Existential restrictions (∃) are by far the most common type of restrictions in OWL ontologies. For a set
of individuals, an existential restriction specifies the existence of a (i.e. at least one) relationship along a
given property to an individual that is a member of a specific class. For example, ∃ hasBase PizzaBase
describes all of the individuals that have at least one relationship along the hasBase property to an
individual that is a member of the class PizzaBase — in more natural English, all of the individuals that
have at least one pizza base.
40

“NECESSARY & SUFFICIENT Header”
A list of necessary & sufficient
conditions appears under this
“NECESSARY & SUFFICIENT”header.
There may be multiple sets of
necessary and suficient conditions,in
which case there will be multiple
“NECESSARY & SUFFICIENT”headers.
“NECESSARY Header”
A list of necessary conditions
appears under this “NECESSARY”header.
There is only one list of necessary
conditions
Add a named class
Remove the selected named class
Add a restriction.
Add a class expression using
the inline expression editor
Display asserted
conditions
Display inferred
conditions
Remove (delete) the
selected class expression
“INHERITED Header”
A list of conditions that have
been inherited from superclasses
will be displayed here. The class from
which a condition is inherited is
displayed next to the condition.
Figure 4.28: The Conditions Widget
41

1. Ensure that the “NECESSARY”header is selected
2. Press the “Create restriction”button
Figure 4.29: Creating a Necessary Restriction
Existential restrictions are also known as Some Restrictions.
Exercise 16: Add a restriction to Pizza that specifies a Pizza must have a PizzaBase
1. Select Pizza from the class hierarchy on the ‘OWLClasses’ tab.
2. Select the “NECESSARY” header in the ‘Conditions Widget’ shown in Figure 4.29
in order to create a necessary condition.
3. Press the ‘Create restriction’ button shown in Figure 4.29. This will display the
‘Create Restriction’ dialog shown in Figure 4.30, which we will use to create a
restriction.
The create restriction dialog has four main parts: The restriction type list; The property list; the filler
edit box; and the expression builder panel. To create a restriction we have to do three things:
• Select the type of restriction from the restriction type list - the default is an existential (∃) restric-
tion.
• Select the property to be restricted from the property list.
• Specify a filler for the restriction in the filler edit box (possibly using the expression builder panel).
42

Property list
Restriction type list
Filler edit box
Expression builder panel
Figure 4.30: The Create Restriction Dialog
Exercise 17: Add a restriction to Pizza that specifies a Pizza must have a PizzaBase (Continued...)
1. Select ‘∃ someValuesFrom’ from the restriction type list — ‘someValuesFrom’ is
another name for an existential restriction.
2. Select the property hasBase from the property list.
3. Specify that the filler is PizzaBase — to do this either: type PizzaBase into the filler
edit box, or press the ‘Insert class’ button on the expression builder panel shown in
Figure 4.31 to display a class hierarchy tree from which PizzaBase may be selected.
4. Press the ‘OK’ button to create the restriction and close the create restriction dialog.
If all information was entered correctly the dialog will close and the restriction will be
displayed in the ‘Conditions Widget’. If there were errors the dialog will not close.
An error message will be displayed at the bottom of the expression builder panel — if
this is the case, recheck that the type of restriction, the property and filler have been
specified correctly.
43

Insert class
Figure 4.31: Expression Builder Panel Insert Class Button
Figure 4.32: The Expression Builder Auto-Completion Function
A very useful feature of the expression builder is the ability to ‘auto complete’
class names, property names and individual names. Auto completion is activated
by pressing ‘alt tab’ on the keyboard. In the above example if we had typed Pi
into the inline expresion editor and pressed the tab key, the choices to complete
the word Pi would have poped up in a list as shown in Figure 4.32. The up and
down arrow keys could then have been used to select PizzaBase and pressing the
Enter key would complete the word for us.
The conditions widget should now look similar to the picture shown in Figure 4.33.
Figure 4.33: Conditions Widget: Description of a Pizza
44

Things that have at least
one PizzaBase
(∃ hasBase PizzaBase)
Pizza
PizzaBase
hasBase
hasBase
hasBase
hasBase
Figure 4.34: A Schematic Description of a Pizza — In order for something to be a Pizza it is necessary for it
to have a (at least one) PizzaBase — A Pizza is a subclass of the things that have at least one
PizzaBase
We have described the class Pizza to be a subclass of owl:Thing and a subclass of
the things that have a base which is some kind of PizzaBase.
Notice that these are necessary conditions — if something is a Pizza it is necessary
for it to be a member of the class owl:Thing (in OWL, everything is a member of
the class owl:Thing) and necessary for it to have a kind of PizzaBase.
More formally, for something to be a Pizza it is necessary for it to be in a relation-
ship with an individual that is a member of the class PizzaBase via the property
hasBase — This is depicted in Figure 4.34.
Creating Some Different Kinds Of Pizzas
It’s now time to add some different kinds of pizzas to our ontology. We will start off by adding a
‘MargheritaPizza’, which is a pizza that has toppings of mozzarella and tomato. In order to keep our
ontology tidy, we will group our different pizzas under the class ‘NamedPizza’:
Exercise 18: Create a subclass of Pizza called NamedPizza, and a subclass of NamedPizza called
MargheritaPizza
1. Select the class Pizza from the class hierarchy on the ‘OWLClasses’ tab.
2. Press the ‘Create subclass’ button to create a new subclass of Pizza, and name it
NamedPizza.
3. Create a new subclass of NamedPizza, and name it MargheritaPizza.
4. Add a comment to the class MargheritaPizza using the comment box that is lo-
cated under the class name widget: A pizza that only has Mozarella and Tomato
toppings – it’s always a good idea to document classes, properties etc. during ontology
editing sessions in order to communicate intentions to other ontology builders.
45

Having created the class MargheritaPizza we now need to specify the toppings that it has. To do this
we will add two restrictions to say that a MargheritaPizza has the toppings MozzarellaTopping and
TomatoTopping.
Exercise 19: Create an existential (∃) restriction on MargheritaPizza that acts along the property
hasTopping with a filler of MozzarellaTopping to specify that a MargheritaPizza has at
least one MozzarellaTopping
1. Make sure that MargheritaPizza is selected in the class hierarchy.
2. Select the ”NECESSARY” header in the ‘Conditions Widget’, as we want to create
and add a necessary condition.
3. Use the ‘Create restriction’ button on the ‘Conditions widget’ (Figure 4.28) to
display the ‘Create Restriction’ dialog.
4. On the ‘Create restrictions’ dialog make the restriction an existentially quantified
restriction by selecting the restriction type as ‘∃ someValuesFrom’.
5. Select hasTopping as the property to be restricted.
6. Enter the class MozzarellaTopping as the filler for the restriction — remember that
this can be achieved by typing the class name MozzarellaTopping into the filler edit
box, or by using the ‘Insert class’ button (Figure 4.31) to display a dialog containing
the ontology class hierarchy which may be used to choose a class.
7. Press the ‘OK’ button on the create restriction dialog to create the restriction — if
there are any errors, the restriction will not be created, and an error message will be
displayed at the bottom of the expression builder panel.
46

Now specify that MargheritaPizzas also have TomatoTopping.
Exercise 20: Create a existential restriction (∃) on MargheritaPizza that acts along the property
hasTopping with a filler of TomatoTopping to specify that a MargheritaPizza has at least
one TomatoTopping
1. Ensure that MargheritaPizza is selected in the class hierarchy.
2. Select the ”NECESSARY” header in the ‘Conditions Widget’, as we want to create
and add a necessary condition.
3. Use the ‘Create restriction’ button on the ‘Conditions Widget’ (Figure 4.28) to
display the ‘Create Restriction’ dialog’.
4. On the restrictions dialog make the restriction an existentially quantified restriction
by selecting the restriction type as ‘∃ someValuesFrom’.
6. Enter the class TomatoTopping as the filler for the restriction.
7. Click the ‘OK’ button on the create restriction dialog to create the restriction.
The ‘Conditions Widget’ should now look similar to the picture shown in Figure 4.35.
We have added restrictions to MargeritaPizza to say that a MargheritaPizza is
a NamedPizza that has at least one kind of MozzarellaTopping and at least one
kind of TomatoTopping.
More formally (reading the conditions widget line by line), if something is a mem-
ber of the class MargheritaPizza it is necessary for it to be a member of the class
NamedPizza and it is necessary for it to be a member of the anonymous class
of things that are linked to at least one member of the class MozzarellaTopping
via the property hasTopping, and it is necessary for it to be a member of the
anonymous class of things that are linked to at least one member of the class
TomatoTopping via the property hasTopping.
Now create the class to represent an Americana Pizza, which has toppings of pepperoni, mozzarella
and tomato. Because the class AmericanaPizza is very similar to the class MargheritaPizza (i.e. an
Americana pizza is almost the same as a Margherita pizza but with an extra topping of pepperoni) we
will make a clone of the MargheritaPizza class and then add an extra restriction to say that it has a
47

Figure 4.35: The Conditions Widget Showing A Description Of A MargheritaPizza
topping of pepperoni.
Exercise 21: Create AmericanaPizza by cloning and modifying the description of MargheritaPizza
1. Right click (ctrl click on the Mac) on the class MargheritaPizza in the class hierarchy
on the OWLClasses tab to display the class hierarchy popup menu.
2. From the popup menu select the menu item ‘Create clone’. This will create a copy
of the class MargheritaPizza named MargheritaPizza 2, that has exactly the same
conditions (restrictions etc.) as MargheritaPizza.
3. Rename the MargheritaPizza 2 to AmericanaPizza using the class name widget.
4. Ensuring that AmericanaPizza is still selected, select the “NECESSARY” header in
the conditions widget, as we want to add a new restriction to the necessary conditions
for AmericanaPizza.
5. Press the ‘Create restriction’ button on the conditions widget to display the ‘Create
restriction dialog’.
6. Select ‘∃ someValuesFrom’ as the type of restriction to create an existentially quan-
tified restriction.
7. Select the property hasTopping as the property to be restricted.
8. Specify the restriction filler as the class PepperoniTopping by either typing
PepperoniTopping into the filler edit box, or by using the ‘Insert class’ button
to display the class dialog, from which PepperoniTopping may be selected.
9. Press the OK button to create the restriction.
48

Figure 4.36: The Conditions Widget displaying the description for AmericanaPizza
Figure 4.37: The Conditions Widget displaying the description for AmericanHotPizza
The ‘Conditions Widget’ should now look like the picture shown in Figure 4.36.
Exercise 22: Create an AmericanHotPizza and a SohoPizza
1. An AmericanHotPizza is almost the same as an AmericanaPizza, but has Jalapeno
peppers on it — create this by cloning the class AmericanaPizza and adding an existen-
tial restriction along the hasTopping property with a filler of JalapenoPepperTopping.
2. A SohoPizza is almost the same as a MargheritaPizza but has additional toppings
of olives and and parmezan cheese — create this by cloning MargheritaPizza and
adding two existential restrictions along the property hasTopping, one with a filler of
OliveTopping, and one with a filler of ParmezanTopping.
For AmericanHot pizza the conditions widget should now look like the picture shown in Figure 4.37. For
SohoPizza the conditions widget should now look like the picture shown in 4.38.
49

Figure 4.38: The Conditions Widget displaying the description for SohoPizza
Having created these pizzas we now need to make them disjoint from one another:
Exercise 23: Make subclasses of NamedPizza disjoint from each other
1. Select the class MargheritaPizza in the class hierarchy on the ‘OWLClasses’ tab.
2. Press the ‘Add all siblings’ button on the ‘Disjoints widget’ to make the pizzas
disjoint from each other.
4.9 Using A Reasoner
4.9.1 Determining the OWL Sub-Language
As mentioned in section 3.1, OWL comes in three flavours (or sub-languages): OWL-Lite, OWL-DL (DL
stands for Description Logics) and OWL-Full. The exact definitions of these sub-languages can be found
in the OWL Overview, which is available on the World Wide Web Consortium website9
. Protégé-OWL
features a species validation facility, which is able to determine the sub-language of the ontology being
edited. To use the species validation facility, use the ‘Determine/Convert OWL Sub-language...’
option on the ‘OWL menu’ shown in Figure 4.39. This will report the sub-language of the ontology.
One of the key features of ontologies that are described using OWL-DL is that they can be processed
by a reasoner. One of the main services offered by a reasoner is to test whether or not one class is a
subclass of another class10
. By performing such tests on all of the classes in an ontology it is possible for
a reasoner to compute the inferred ontology class hierarchy. Another standard service that is offered by
reasoners is consistency checking. Based on the description (conditions) of a class the reasoner can check
whether or not it is possible for the class to have any instances. A class is deemed to be inconsistent if
it cannot possibly have any instances.
9http://www.w3.org/TR/owl-features/
10Known as subsumption testing — the descriptions of the classes (conditions) are used to determine if a super-
class/subclass relationship exists between them.
50

Figure 4.39: The OWL Menu
Reasoners are also known as classifiers.
4.9.2 Using RACER
In order to reason over the ontologies in Protégé-OWL a DIG11
compliant reasoner must be installed/configured
and started. In this tutorial we use a reasoner called RACER, which is available for a variety of platforms
from http://www.sts.tu-harburg.de/~r.f.moeller/racer/. RACER comes with a detailed manual
that contains installation and setup instructions. When you have installed RACER on your system, it
should be started with the default settings — RACER is typically started by double clicking on the
RACER application icon, which opens a terminal/console window and starts the reasoner running with
HTTP communication enabled.12
— Figure 4.40 shows a pruned version of the information that is dis-
played when RACER starts; The second from the bottom line indicates that HTTP communication is
running, and specifies the I.P. address and port number. If for any reason RACER needs to be started on
a different port (or computer), Protégé-OWL can be configured via the OWL preferences dialog shown
in Figure 4.41, which is accessible via the ‘Preferences...’ item on the OWL menu.
4.9.3 Invoking The Reasoner
Having started RACER, or another reasoner, the ontology can be ‘sent to the reasoner’ to automatically
compute the classification hierarchy, and also to check the logical consistency of the ontology. In Protégé-
OWL the ‘manually constructed’ class hierarchy is called the asserted hierarchy. The class hierarchy
11DIG = Description Logic Implementers Group — A DIG compliant reasoner provides the means to communicate via
the DIG interface, which is a standard interface/protocol for talking to description logic reasoners.
12By default racer runs with the HTTP service enabled on port 8080.
51

;;; RACER Version 1.7.12
;;; RACER: Reasoner for Aboxes and Concept Expressions Renamed
;;; Supported description logic: ALCQHIr+(D)-
;;; Copyright (C) 1998-2003, Volker Haarslev and Ralf Moeller.
;;; RACER comes with ABSOLUTELY NO WARRANTY; use at your own risk.
;;; Commercial use is prohibited; contact the authors for licensing.
;;; RACER is running on Mac OS Darwin computer as node Unknown
[2004-04-16 10:22:47] HTTP service enabled for: http://130.88.195.45:8080/
[2004-04-16 10:22:47] TCP service enabled on port 8088
Figure 4.40: RACER Startup Screen
Figure 4.41: The OWL Preferences Dialog
52

Check consistency
Classify taxonomy
Run ontology tests
Show TODO list
View RDF/XML source
Show Protege-OWL Syntax
Display English Prose Tooltips
Figure 4.42: The OWL Toolbar
Figure 4.43: The Inferred Hierarchy Pane, which pops open next to the Asserted Hierarchy Pane when classifi-
cation has taken place
that is automatically computed by the reasoner is called the inferred hierarchy. To automatically classify
the ontology (and check for inconsistencies) the ‘Classify Taxonomy...’ action should be used. This
can be invoked via the OWL menu (Figure 4.39), or by using the ‘Classify taxonomy’ button on
the Protégé-OWL toolbar shown in Figure 4.42. To check the consistency of the ontology, the ‘Check
consistency...’ action should be used, which can be invoked from the OWL menu, or by using the ‘Check
consistency’ button on the Protégé-OWL toolbar. When the inferred hierarchy has been computer, an
inferred hierarchy window will pop open next to the existing asserted hierarchy window as shown in
Figure 4.43. If a class has been reclassified (i.e. if it’s superclasses have changed) then the class name
will appear in a blue colour in the inferred hierarchy. If a class has been found to be inconsistent it’s icon
will be circled in red.
The task of computing the inferred class hierarchy is also know as classifying the
ontology.
53

Figure 4.44: The Conditions Widget Displaying ProbeInconsistentTopping
4.9.4 Inconsistent Classes
In order to demonstrate the use of the reasoner in detecting inconsistencies in the ontology we will create
a class that is a subclass of both CheeseTopping and also MeatTopping. This strategy is often used as
a check so that we can see that we have built our ontology correctly. Classes that are added in order to
test the integrity of the ontology are sometimes known as Probe Classes.
Exercise 24: Add a Probe Class called ProbeInconsistentTopping which is a subclass of both
CheeseTopping and Vegetable
1. Select the class CheeseTopping from the class hierarchy on the OWLClasses tab.
2. Create a subclass of CheeseTopping named ProbeInconsistentTopping.
3. Add a comment to the ProbeInconsistentTopping class that is something along the
lines of, “This class should be inconsistent when the ontology is classiﬁed.”. This will
enable anyone who looks at our pizza ontology to see that we deliberately meant the
class to be inconsistent.
4. Ensure that the ProbeInconsistentTopping class is selected in the class hierarchy, and
then select the “NECESSARY” header in the ‘Conditions widget’.
5. Click on the ‘Add named class’ button on the ‘Conditions Widget’. This will dis-
play a dialog containing the class hierarchy from which a class may be selected. Select
the class VegetableTopping and then press the OK button. The class VegetableTop-
ping will be added as a necessary condition (as a superclass), so that the conditions
widget should look like the picture in Figure 4.44.
54

Figure 4.45: The Class ProbeInconsistentTopping found to be inconsistent by the reasoner
If we study the class hierarchy, ProbeInconsistentTopping should appear as a
subclass of CheeseTopping and as a subclass of VegetableTopping. This means
that ProbeInconsistentTopping is a CheeseTopping and a VegetableTopping.
More formally, all individuals that are members of the class ProbeInconsistentTop-
ping are also (necessarily) members of the class CheeseTopping and (necessarily)
members of the class VegetableTopping. Intuitively this is incorrect since some-
thing can not simultaneously be both cheese and a vegetable!
Exercise 25: Classify the ontology to make sure ProbeInconsistentTopping is inconsistent
1. Press the ‘Classify Taxonomy’ button on the OWL toolbar to classify the ontology.
After a few seconds the inferred hierarchy will have been computed and the inferred hierarchy window
will pop open (if it was previously closed). The hierarchy should resemble that shown in Figure 4.45 —
notice that the class ProbeInconsistentTopping is circled in red, indicating that the reasoner has found
this class to be inconsistent (i.e. it cannot possibly have any individuals as memebers).
55

Why did this happen? Intuitively we know something cannot at the same time
be both cheese and a vegetable. Something should not be both an instance of
CheeseTopping and an instance of VegetableTopping. However, it must be re-
membered that we have chosen the names for our classes. As far as the reasoner is
concerned names have no meaning. The reasoner cannot determine that something
is inconsistent based on names. The actual reason that ProbeInconsistentTopping
has been detected to be inconsistent is because its superclasses VegetableTopping
and CheeseTopping are disjoint from each other — remember that earlier on we
speciﬁed that the four categories of topping were disjoint from each other using
the Wizard. Therefore, individuals that are members of the class CheeseTopping
cannot be members of the class VegetableTopping and vice-versa.
To close the inferred hierarchy use the small white cross on a red background
button on the top right of the inferred hierarchy window.
Exercise 26: Remove the disjoint statement between CheeseTopping and VegetableTopping to see
what happens
1. Select the class CheeseTopping using the class hierarchy.
2. The ‘Disjoints widget’ should contain CheeseTopping’s sibling classes: Vegetable-
Topping, SeafoodTopping and MeatTopping. Select VegetableTopping in the Disjoints
widget.
3. Press the ‘Remove selected class from list’ button on the Disjoints widget (shown
in Figure 4.5) to remove the disjoint axiom that states CheeseTopping and MeatTop-
ping are disjoint.
4. Press the ‘Classify Taxonomy’ button on the OWL toolbar to send the ontology to
the reasoner. After a few seconds the ontology should have been classiﬁed and the
results displayed.
56

It should be noticeable that ProbeInconsistentTopping is no longer inconsistent!
This means that individuals which are members of the class ProbeInconsistent-
Topping are also members of the class CheeseTopping and VegetableTopping —
something can be both cheese and a vegetable!
This clearly illustrates the importance of the careful use of disjoint axioms in
OWL. OWL classes ‘overlap’ until they have been stated to be disjoint from each
other. If certain classes are not disjoint from each other then unexpected results
can arise. Accordingly, if certain classes have been incorrectly made disjoint from
each other then this can also give rise to unexpected results.
Exercise 27: Fix the ontology by making CheeseTopping and Vegetable disjoint from each other
1. Select the class CheeseTopping using the class hierarchy.
2. The ‘Disjoints widget’ should contain MeatTopping and SeafoodTopping.
3. Press the ‘Add disjoint class’ button on the disjoint classes widget to display a dialog
which classes may be picked from. Select the class VegetableTopping and press the
OK button. CheeseTopping should once again be disjoint from VegetableTopping.
4. Test that the disjoint axiom has been added correctly — Press the ‘Classify Tax-
onomy’ button on the OWL toolbar to send the ontology to the reasoner. After a
few seconds the ontology should have been classified, and ProbeInconsistentTopping
should be highlighted in red indicating that it is once again inconsistent.
4.10 Necessary And Sufficient Conditions (Primitive and De-
fined Classes)
All of the classes that we have created so far have only used necessary conditions to describe them.
Necessary conditions can be read as, “If something is a member of this class then it is necessary to
fulfil these conditions”. With necessary conditions alone, we cannot say that, “If something fulfils these
conditions then it must be a member of this class”.
A class that only has necessary conditions is known as a Primitive Class.
Let’s illustrate this with an example. We will create a subclass of Pizza called CheesyPizza, which will
57

Figure 4.46: The Description of CheesyPizza (Using Necessary Conditions)
be a Pizza that has at least one kind of CheeseTopping.
Exercise 28: Create a subclass of Pizza called CheesyPizza and specify that it has at least one
topping that is a kind of CheeseTopping
1. Select Pizza in the class hierarchy on the ‘OWLClasses’ tab.
2. Press the ‘Create subclass’ button to create a subclass of Pizza. Rename it to
CheesyPizza.
3. Make sure that CheesyPizza is selected in the class hierarchy. Select the “NECES-
SARY” header in the conditions widget. (You may have to select the ‘Asserted’ tab
on the ‘Conditions Widget’ — the automatically shows the ‘Inferred’ tab after
classiﬁcation).
restriction dialog’.
5. Select ‘∃ someValuesFrom’ as the type of restriction to be created.
7. In the ﬁller edit box type CheeseTopping (or use the ‘Insert class’ button to display
a dialog from which CheeseTopping can be selected). Press ‘OK’ to close the dialog
and create the restriction.
Our description of CheesyPizza states that if something is a member of the class
CheesyPizza it is necessary for it to be a member of the class Pizza and it is
necessary for it to have at least one topping that is a member of the class Cheese-
Topping.
Our current description of CheesyPizza says that if something is a CheesyPizza it is necessarily a Pizza
and it is necessary for it to have at least one topping that is a kind of CheeseTopping. We have used
necessary conditions to say this. Now consider some (random) individual. Suppose that we know that
58

this individual is a member of the class Pizza. We also know that this individual has at least one kind of
CheeseTopping. However, given our current description of CheesyPizza this knowledge is not sufficient
to determine that the individual is a member of the class CheesyPizza. To make this possible we need to
change the conditions for CheesyPizza from necessary conditions to necessary AND sufficient conditions.
This means that not only are the conditions necessary for membership of the class CheesyPizza, they
are also sufficient to determine that any (random) individual that satisfies them must be a member of
the class CheesyPizza.
A class that has at least one set of necessary and sufficient conditions is known as
a Defined Class.
Classes that only have necessary conditions are also known as ‘partial’ classes.
Classes that have at least one set of necessary and sufficient conditions are also
known as ‘complete’ classes.
In order to convert necessary conditions to necessary and sufficient conditions, the conditions must be
moved from under the“NECESSARY” header in the conditions widget to be under the“NECESSARY
AND SUFFICIENT” header. This can be accomplished by dragging and dropping the conditions one-
by-one.
Exercise 29: Convert the necessary conditions for CheesyPizza into necessary & sufficient condi-
tions
1. Ensure that CheesyPizza is selected in the class hierarchy.
2. On the ‘Conditions Widget’ select the ∃ hasTopping CheeseTopping restriction.
3. Drag the ∃ hasTopping CheeseTopping restriction from under the “NECESSARY”
header to on top of the “NECESSARY & SUFFICIENT” header.
4. Select the class Pizza.
5. Drag the class Pizza from under the “NECESSARY” header to on top of the ∃ hasTop-
ping CheeseTopping restriction (note not on top of the “NECESSARY & SUFFI-
CIENT” header this time).
59

Figure 4.47: The Description of CheesyPizza (Using Necessary AND Sufficient Conditions)
NECESSARY CONDITIONS
NamedClass
Condition
Condition
Condition
Condition
NECESSARY & SUFFICIENT CONDITIONS
NamedClass
Condition
Condition
Condition
Condition
If an individual is a member of ‘NamedClass’then it must satisfy the conditions.
However if some individual satisfies these necessary conditions, we cannot say
that it is a member of ‘Named Class’(the conditions are not ‘sufficient’to be able
to say this) - this is indicated by the direction of the arrow.
If an individual is a memeber of ‘NamedClass’then it must satisfy the conditions.
If some individual satisfies the condtions then the individual must be a member
of ‘NamedClass’- this is indicated by the double arrow.
implies
implies
Figure 4.48: Necessary And Sufficient Conditions
We have converted our description of CheesyPizza into a definition. If something
is a CheesyPizza then it is necessary that it is a Pizza and it is also necessary
that at least one topping that is a member of the class CheeseTopping. Moreover,
if an individual is a member of the class Pizza and it has at least one topping that
is a member of the class CheeseTopping then these conditions are sufficient to
determine that the individual must be a member of the class CheesyPizza. The
notion of necessary and sufficient conditions is illustrated in Figure 4.48.
If you accidentally dropped Pizza onto the “NECESSARY & SUFFICIENT”
header (rather that onto the ∃ hasTopping CheeseTopping) in Exercise 29 the
conditions widget will look like the picture shown in Figure 4.49. In this case, a new
necessary and sufficient condition has been created, which is not what we want.
To correct this mistake, drag Pizza on top of the ∃ hasTopping CheeseTopping
restriction.
60

Figure 4.49: An INCORRECT description of CheesyPizza
Conditions may also be transferred from “NECESSARY” to “NECESSARY &
SUFFICIENT” and vice versa using Cut and Paste. Right click (ctrl click on a
Mac) on a condition and select Cut or Paste from the popup menu.
To summarise: If class A is described using necessary conditions, then we can say that if an individual
is a member of class A it must satisfy the conditions. We cannot say that any (random) individual that
satisfies these conditions must be a member of class A. However, if class A is now defined using necessary
and sufficient conditions, we can say that if an individual is a member of the class A it must satisfy the
conditions and we can now say that if any (random) individual satisfies these conditions then it must be
a member of class A. The conditions are not only necessary for membership of A but also sufficient to
determine that something satisfying these conditions is a member of A.
How is this useful in practice? Suppose we have another class B, and we know that any individuals that
are members of class B also satisfy the conditions that define class A. We can determine that class B is
subsumed by class A — in other words, B is a subclass of A. Checking for class subsumption is a key
task of a description logic reasoner and we will use the reasoner to automatically compute a classification
hierarchy in this way.
In OWL it is possible to have multiple sets of necessary & sufficient conditions.
This is discussed later in section 6.5
4.10.1 Primitive And Defined Classes
Classes that have at least one set of necessary and sufficient conditions are known as defined classes —
they have a definition, and any individual that satisfies the definition will belong to the class. Classes
that do not have any sets of necessary and sufficient conditions (only have necessary conditions) are
know as primitive classes. In Protégé-OWL defined classes have a class icon with an orange background.
Primitive classes have a class icon that has a yellow background. It is also important to understand that
the reasoner can only automatically classify classes under defined classes - i.e. classes with at least one
set of necessary and sufficient conditions.
61

Figure 4.50: The Asserted and Inferred Hierarchies Displaying The Classification Results For CheesyPizza
4.11 Automatic Classification
Being able to use a reasoner to automatically compute the class hierarchy is one of the major benefits of
building an ontology using the OWL-DL sub-language. Indeed, when constructing very large ontologies
(with upwards of several thousand classes in them) the use of a reasoner to compute subclass-superclass
relationships between classes becomes almost vital. Without a reasoner it is very difficult to keep large
ontologies in a maintainable and logically correct state. In cases where ontologies can have classes that
have many superclasses (multiple inheritance) it is nearly always a good idea to construct the class
hierarchy as a simple tree. Classes in the asserted hierarchy (manually constructed hierarchy) therefore
have no more than one superclass. Computing and maintaining multiple inheritance is the job of the
reasoner. This technique13
helps to keep the ontology in a maintainable and modular state. Not only
does this promote the reuse of the ontology by other ontologies and applications, it also minimises human
errors that are inherent in maintaining a multiple inheritance hierarchy.
Having created a definition of a CheesyPizza we can use the reasoner to automatically compute the
subclasses of CheesyPizza.
Exercise 30: Use the reasoner to automatically compute the subclasses of CheesyPizza
1. Ensure that a reasoner (RACER) is running.
2. Press the ‘Classify Taxonomy...’ button on the toolbar (See Figure 4.42).
After a few seconds the inferred hierarchy should have been computed and the inferred hierarchy window
will pop open (if it was previously closed). The inferred hierarchy should appear similar to the picture
shown in Figure 4.50. Figures 4.51 and 4.52 show the OWLViz display of the asserted and inferred
hierarchies respectively. Notice that classes which have had their superclasses changed by the reasoner
are shown in blue.
13Sometimes know as ontology normalisation.
62

NamePizza
Pizza
CheesyPizza
MargheritaPizza AmericanHotPizza AmericanaPizza SohoPizza
is-a is-a
is-ais-ais-a is-a
Figure 4.51: OWLViz Displaying the Asserted Hierarchy for CheesyPizza
CheesyPizza
MargheritaPizzaSohoPizza
NamePizza
Pizza
AmericanHotPizza AmericanaPizza
is-a is-a
is-a is-a
is-a is-a is-ais-a is-ais-a
Figure 4.52: OWLViz Displaying the Inferred Hierarchy for CheesyPizza
63

The reasoner has determined that MargheritaPizza, AmericanaPizza, American-
HotPizza and SohoPizza are subclasses of CheesyPizza. This is because we
defined CheesyPizza using necessary and sufficient conditions. Any individual
that is a Pizza and has at least one topping that is a CheeseTopping is a member
of the class CheesyPizza. Due to the fact that all of the individuals that are
described by the classes MargheritaPizza, AmericanaPizza, AmericanHotPizza
and SohoPizza are Pizzas and they have at least one topping that is a Cheese-
Toppinga
the reasoner has determined that these classes must be subclasses of
CheeseTopping.
aOr toppings that belong to the subclasses of CheeseTopping
It is important to realise that, in general, classes will never be placed as sub-
classes of primitive classes (i.e. classes that only have necessary conditions) by the
reasonera
.
aThe exception to this is when a property has a domain that is a primitive class. This can
coerce classes to be reclassified under the primitive class that is the domain of the property —
the use of property domains to cause such effects is strongly discouraged.
4.11.1 Classification Results
After the reasoner has been invoked, computed superclass-subclass relationships and inconsistent classes
are displayed in the ‘Classification Results’ pane shown in Figure 4.53. The ‘Classification Results’
pane pops open after classification at the bottom of the Protégé-OWL application window. The ‘spanner
icon’ on the left hand side of the pane is the ‘Assert Selected Change(s)’ button. Pressing this button
takes the superclass-subclass relationships that have been found by the reasoner and puts them into
the asserted (manually constructed) hierarchy. For example, if the ‘Assert Selected Changes’ button
was pressed with the selection shown in Figure 4.53, CheesyPizza would be added as a superclass of
AmericanaPizza.
Despite that fact that this facility exists, it is generally considered a bad idea to
put computed/inferred relationships into the ‘manually constructed’ or asserted
model whilst an ontology is being developed — we therefore advise against using
this button during the development of an ontology.
4.12 Universal Restrictions
All of the restrictions we have created so far have been existential restrictions (∃). Existential restrictions
specify the existence of at least one relationship along a given property to an individual that is a member
of a specific class (specified by the filler). However, existential restrictions do not mandate that the
only relationships for the given property that can exist must be to individuals that are members of the
specified filler class.
64

Assert selected change(s)
Figure 4.53: The Classification Results Pane
For example, we could use an existential restriction ∃ hasTopping MozzarellaTopping to describe the
individuals that have at least one relationship along the property hasTopping to an individual that is
a member of the class MozzarellaTopping. This restriction does not imply that all of the hasTopping
relationships must be to a member of the class MozzarellaTopping. To restrict the relationships for a
given property to individuals that are members of a specific class we must use a universal restriction.
Universal restrictions are given the symbol ∀. They constrain the relationships along a given property
to individuals that are members of a specific class. For example the universal restriction ∀ hasTopping
MozzarellaTopping describes the individuals all of whose hasTopping relationships are to members of the
class MozzarellaTopping — the individuals do not have a hasTopping relationships to individuals that
aren’t members of the class MozzarellaTopping.
Universal restrictions are also know as All Restrictions.
The above universal restriction ∀ hasTopping MozzarellaTopping also describes
the individuals that do not participate in any hasTopping relationships. An indi-
vidual that does not participate in any hasTopping relationships what so ever, by
definition does not have any hasTopping relationships to individuals that aren’t
members of the class MozzarellaTopping and the restriction is therefore satisfied.
For a given property, universal restrictions do not specify the existence of a rela-
tionship. They merely state that if a relationship exists for the property then it
must be to individuals that are members of a specific class.
Suppose we want to create a class called VegetarianPizza. Individuals that are members of this class
can only have toppings that are CheeseTopping or VegetableTopping. To do this we can use a universal
65

restriction.
Exercise 31: Create a class to describe a VegetarianPizza
1. Create a subclass of Pizza, and name it VegetarianPizza.
2. Making sure that VegetarianPizza is selected, click on the ”NECESSRY” header in
the ‘Conditions Widget’.
3. Press the ‘Create restriction’ button on the ‘Conditions Widget’ to display the
‘Create restriction dialog’.
4. Select the type of restriction as ‘∀ allValuesFrom’ in order to create a universally
quantified restriction.
6. For the filler we want to say CheeseTopping or VegetableTopping. First insert the
class CheeseTopping either by typing CheeseTopping into the filler box, or by using
the ‘Insert class’ button. We now need to use the unionOf operator between the
class names. The unionOf operator may be inserted using the button shown in Figure
4.54a
. Insert the unionOf symbol by pressing the ‘Insert unionOf’ button on the
expression builder panel. Next insert the class VegetableTopping either by typing
it or by using the ‘Insert class button’. You should now have CheeseTopping
VegetableTopping in the filler edit box.
7. Press the ‘OK’ button on the dialog to close the dialog and create the restriction —
if there are any errors (due to typing errors etc.) the dialog will not close and an error
message will be displayed at the bottom of the expression builder panel.
aSee section B.2 for more information about union classes.
At this point the conditions widget should look like the picture shown in Figure 4.55.
66

Insert union of
Insert class
Figure 4.54: Using the Expression Builder Panel to insert Union Of
Figure 4.55: The Description of VegetarianPizza (Using Necessary Conditions)
This means that if something is a member of the class VegetarianPizza it is nec-
essary for it to be a kind of Pizza and it is necessary for it to only (∀ universal
quantiﬁer) have toppings that are kinds of CheeseTopping or kinds of Vegetable-
Topping.
In other words, all hasTopping relationships that individuals which are members
of the class VegetarianPizza participate in must be to individuals that are either
members of the class CheeseTopping or VegetableTopping.
The class VegetarianPizza also contains individuals that are Pizzas and do not
participate in any hasTopping relationships.
Instead of using the ‘Insert unionOf’ button in Exercise 31 , we could have
simply typed or into the ﬁller edit box and it would have automatically been
converted to the union of symbol ( ).
67

In situations like the above example, a common mistake is to use an intersec-
tion instead of a union. For example, CheeseTopping VegetableTopping. This
reads, CheeseTopping and VegetableTopping. Although “CheeseTopping and
Vegetable” might be a natural thing to say in English, this logically means some-
thing that is simultaneously a kind of CheeseTopping and VegetableTopping. This
is obviously incorrect as demonstrated in section 4.9.4. If the classes CheeseTop-
ping and VegetableTopping were not disjoint, this would have been a logically
legitimate thing to say – it would not be inconsistent and therefore would not be
‘spotted’ by the reasoner.
In the above example it might have been tempting to create two universal re-
strictions — one for CheeseTopping (∀ hasTopping CheeseTopping) and one for
VegetableTopping (∀ hasTopping VegetableTopping). However, when multiple
restrictions are used (for any type of restriction) the total description is taken
to be the intersection of the individual restrictions. This would have therefore
been equivalent to one restriction with a filler that is the intersection of Moz-
zarellaTopping and TomatoTopping — as explained above this would have been
logically incorrect.
Currently VegetarianPizza is described using necessary conditions. However, our description of a Veg-
etarianPizza could be considered to be complete. We know that any individual that satisfies these
conditions must be a VegetarianPizza. We can therefore convert the necessary conditions for Vegetari-
anPizza into necessary and sufficient conditions. This will also enable us to use the reasoner to determine
the subclasses of VegetarianPizza.
Exercise 32: Convert the necessary conditions for VegetarianPizza into necessary & sufficient
conditions
1. Ensure that VegetarianPizza is selected in the class hierarchy.
2. On the ‘Conditions Widget’ select the (∀ universal) restriction on the hasTopping
property.
3. Drag the hasTopping restriction from under the “NECESSARY” header to on top of
the “NECESSARY & SUFFICIENT” header.
4. Select the class Pizza.
5. Drag the class Pizza from under the “NECESSARY” header to on top of the hasTop-
ping restriction (note not on top of the “NECESSARY & SUFFICIENT” header this
time).
68

Figure 4.56: The Conditions Widget Displaying the Definition of VegetarianPizza (Using Necessary and Suffi-
cient Conditions)
We have converted our description of VegetarianPizza into a definition. If some-
thing is a VegetarianPizza, then it is necessary that it is a Pizza and it is also
necessary that all toppings belong to the class CheeseTopping or VegetableTop-
ping. Moreover, if something is a member of the class Pizza and all of it’s toppings
are members of the class CheeseTopping or the class VegetableTopping then these
conditions are sufficient to recognise that it must be a member of the class Veg-
etarianPizza. The notion of necessary and sufficient conditions is illustrated in
Figure 4.48.
4.13 Automatic Classification and Open World Reasoning
We want to use the reasoner to automatically compute the superclass-subclass relationship (subsumption
relationship) between MargheritaPizza and VegetarianPizza and also, SohoPizza and VegetarianPizza.
Recall that we believe that MargheritaPizza and SohoPizza should be vegetarian pizzas (they should
be subclasses of VegetarianPizza). This is because they have toppings that are essentially vegetarian
toppings — by our definition, vegetarian toppings are members of the classes CheeseTopping or Veg-
etableTopping and their subclasses. Having previously created a definition for VegetarianPizza (using
a set of necessary and sufficient conditions) we can use the reasoner to perform automatic classification
and determine the vegetarian pizzas in our ontology.
Exercise 33: Use the reasoner to classify the ontology
1. Ensure that a resoner (RACER) is running. Press the ‘Classify taxonomy’ button.
You will notice that MargheritaPizza and also SohoPizza have not been classified as subclasses of Veg-
etarianPizza. This may seem a little strange, as it appears that both MargheritaPizza and SohoPizza
have ingredients that are vegetarian ingredients, i.e. ingredients that are kinds of CheeseTopping or
kinds of VegetableTopping. However, as we will see, MargheritaPizza and SohoPizza have something
missing from their definition that means they cannot be classified as subclasses of VegetarianPizza.
Reasoning in OWL (Description Logics) is based on what is known as the open world assumption (OWA).
It is frequently referred to as open world reasoning (OWR). The open world assumption means that we
69

cannot assume something doesn’t exist until it is explicitly stated that it does not exist. In other words,
because something hasn’t been stated to be true, it cannot be assumed to be false — it is assumed that
‘the knowledge just hasn’t been added to the knowledge base’. In the case of our pizza ontology, we
have stated that MargheritaPizza has toppings that are kinds of MozzarellaTopping and also kinds of
TomatoTopping. Because of the open world assumption, until we explicitly say that a MargheritaPizza
only has these kinds of toppings, it is assumed (by the reasoner) that a MargheritaPizza could have other
toppings. To specify explicitly that a MargheritaPizza has toppings that are kinds of MozzarellaTopping
or kinds of MargheritaTopping and only kinds of MozzarellaTopping or MargheritaTopping, we must add
what is known as a closure axiom14
on the hasTopping property.
4.13.1 Closure Axioms
A closure axiom on a property consists of a universal restriction that acts along the property to say that
it can only be filled by the specified fillers. The restriction has a filler that is the union of the fillers that
occur in the existential restrictions for the property15
. For example, the closure axiom on the hasTopping
property for MargheritaPizza is a universal restriction that acts along the hasTopping property, with a
filler that is the union of MozzarellaTopping and also TomatoTopping. i.e. ∀ hasTopping (Mozzarel-
laTopping TomatoTopping).
Exercise 34: Add a closure axiom on the hasTopping property for MargheritaPizza
1. Make sure that MargheritaPizza is selected in the class hierarchy on the
‘OWLClasses’ tab.
2. Select the“NECESSARY” header in the ‘Conditions Widget’.
Restriction dialog’.
4. Select the restriction type as ‘∀ allValuesFrom’ (universal restriction).
6. In the filler edit box enter MozzarellaTopping TomatoTopping. This can be done
by typing MozzarellaTopping or TomatoTopping into the filler edit box (“or” will
be automatically converted to as the filler is typed). This can also be accomplished
by using the ‘Insert class’ button and the ‘Insert unionOf’ button to insert the
class MozzarellaTopping, then insert the unionOf symbol and then insert the class
TomatoTopping.
7. Press the OK button to create the restriction and add it to the class MargheritaPizza.
The conditions widget should now appear as shown in Figure 4.57.
14Also referred to as a closure restriction.
15And technically speaking the classes for the values used in any hasValue restrictions (see later).
70

Figure 4.57: Conditions Widget: Margherita Pizza With a Closure Axiom for the hasTopping property
This now says that if an individual is a member of the class MargeritaPizza then
it must be a member of the class Pizza, and it must have at least one topping
that is a kind of MozzarellaTopping and it must have at least one topping that
is a member of the class TomatoTopping and the toppings must only be kinds of
MozzarellaTopping or TomatoTopping.
A common error in situations such as above is to only use universal restrictions in
descriptions. For example, describing a MargheritaPizza by making it a subclass of
Pizza and then only using ∀ hasTopping (MozzarellaTopping TomatoTopping)
without any existential restrictions. However, because of the semantics of the
universal restriction, this actually means either: things that are Pizzas and only
have toppings that are MozzarellaTopping or TomatoTopping, OR, things that are
Pizzas and do not have any toppings at all.
Exercise 35: Add a closure axiom on the hasTopping property for SohoPizza
1. Select SohoPizza in the class hierarchy on the ‘OWLClasses’ tab.
2. Select the ”NECESSARY” header in the ‘Conditions Widget’.
3. Press the ‘Create restriction’ button to display the ‘Create Restriction dialog’.
4. Select the restriction type as, ‘∀ allValuesFrom’, as we want to create a universally
quantiﬁed restriction.
6. In the ﬁller edit box enter the union of the toppings for SohoPizza by typing
ParmezanTopping or MozzarellaTopping or TomatoTopping or OliveTopping.
Note that the “or” keywords will automatically be converted to the unionOf symbol
( ) as you type to give “ParmezanTopping MozzarellaTopping TomatoTopping
OliveTopping”.
7. Press the OK button to create the restriction and close the dialog. If the dialog will
not close due to errors, check that the class names have been spelt correctly.
71

For completeness, we will add closure axioms for the hasTopping property to AmericanaPizza and also
AmericanHotPizza. At this point it may seem like tedious work to enter these closure axioms by hand.
Fortunately Protégé-OWL has the capability of creating closure axioms for us.
Exercise 36: Automatically create a closure axiom on the hasTopping property for AmericanaPizza
1. Select AmericanaPizza in the class hierarchy on the OWLClasses tab.
2. In the ‘Conditions Widget’ right click (Ctrl click on the Mac) on one of the exis-
tential hasTopping restrictions. Select ‘Add closure axiom’ from the pop up menu
that appears. A closure restriction (universal restriction) will be created along the
hasTopping property, which contains the union of the existential hasTopping fillers.
Exercise 37: Automatically create a closure axiom on the hasTopping property for AmericanHot-
Pizza
1. Select AmericanHotPizza in the class hierarchy on the OWLClasses tab.
2. In the ‘Conditions Widget’ right click (Ctrl click on the Mac) on one of the exis-
tential hasTopping restrictions. Select ‘Add closure axiom’ from the pop up menu
that appears.
Having added closure axioms on the hasTopping property for our pizzas, we can now used the reasoner
to automatically compute classifications for them.
1. Press the ‘Classify taxonomy’ button on the OWL toolbar to invoke the reasoner.
After a short amount of time the ontology will have been classified and the ‘Inferred Hierarchy’ pane
will pop open (if it is not already open). This time, MargheritaPizza and also SohoPizza will have
been classified as subclasses of VegetarianPizza. This has happened because we specifically ‘closed’ the
hasTopping property on our pizzas to say exactly what toppings they have and VegetarianPizza was
defined to be a Pizza with only kinds of CheeseTopping and only kinds of VegetableTopping. Figure
4.58 shows the current asserted and inferred hierarchies. It is clear to see that the asserted hierarchy is
simpler and ‘cleaner’ than the ‘tangled’ inferred hierarchy. Although the ontology is only very simple at
this stage, it should be becoming clear that the use of a reasoner can help (especially in the case of large
ontologies) to maintain a multiple inheritance hierarchy for us.
72

Pizza
MargheritaPizza
VegetarianPizza CheesyPizza
SohoPizza AmericanaPizza
NamedPizza
AmericanHotPizza
is-a
is-a is-a
is-a is-a
is-a
is-a
CheesyPizzaVegetarianPizza
SohoPizza MargheritaPizza
NamedPizza
AmericanaPizzaAmericanHotPizza
Pizza
is-ais-a
is-a
is-ais-a is-a
is-a
is-ais-ais-a
is-a
is-a is-a
Asserted Hierarchy
Inferred Hierarchy
Figure 4.58: The asserted and inferred hierarchies showing the “before and after” classification of Pizzas into
CheesyPizzas and VegetarianPizzas.
4.14 Value Partitions
In this section we create some Value Partitions, which we will use to refine our descriptions of various
classes. Value Partitions are not part of OWL, or any other ontology language, they are a ‘design pattern’.
Design patterns in ontology design are analogous to design patterns in object oriented programming —
they are solutions to modelling problems that have occurred over and over again. These design patterns
have been developed by experts and are now recognised as proven solutions for solving common modelling
problems. As mentioned previously, Value Partitions can be created to refine our class descriptions, for
example, we will create a Value Partition called ‘SpicinessValuePartition’ to describe the ‘spiciness’ of
PizzaToppings. Value Partitions restrict the range of possible values to an exhaustive list, for example, our
‘SpicinessValuePartition’ will restrict the range to ‘Mild’, ‘Medium’, and ‘Hot’. Creating a ValuePartition
in OWL consists of several steps:
1. Create a class to represent the ValuePartition. For example to represent a ‘spiciness’ ValuePartition
we might create the class SpicinessValuePartition.
2. Create subclasses of the ValuePartition to represent the possible options for the ValuePartition.
For example we might create the classes Mild, Medium and Hot as subclasses of the SpicynessVal-
uePartition class.
3. Make the subclasses of the ValuePartition class disjoint.
4. Provide a covering axiom to make the list of value types exhaustive (see below).
5. Create an object property for the ValuePartition. For example, for our spiciness ValuePartition,
we might create the property hasSpiciness.
73

6. Make the property functional.
7. Set the range of the property as the ValuePartition class. For example for the hasSpiciness property
the range would be set to SpicinessValuePartition.
It should be relatively clear that due to the number of steps and the complexity of some of the steps,
it would be quite easy to make a mistake. It could also take a significant amount of time to create
more than a few ValuePartitions. Fortunately, the OWL Wizards package contains a wizard for creating
ValuePartitions – appropriately named the ‘Create ValuePartition’ wizard.
Let’s create a ValuePartition that can be used to describe the spiciness of our pizza toppings. We will
then be able to classify our pizzas into spicy pizzas and non-spicy pizzas. We want to be able to say
that our pizza toppings have a spiciness of either ‘mild’, ‘medium’ or ‘hot’. Note that these choices are
mutually exclusive – something cannot be both ‘mild’ and ‘hot’, or a combination of the choices.
Exercise 39: Create a ValuePartition to represent the spiciness of pizza toppings
1. Select ‘Create Value Partition’ from the Wizards menu on the Protégé menu bar
to invoke the ValuePartition wizard.
2. On the first page of the wizard type SpicinessValuePartition as the name of the
ValuePartition class and press the Next button.
3. Now enter hasSpiciness for the ValuePartition property name, and press the ‘Next’
button.
4. We now need to specify the values for the value type. In the text area type Mild and
press return, type Medium and press return, and type Hot and press return. This will
create Mild, Medium and Hot as subclasses of the SpicinessValuePartition class. Press
the ‘Next’ button to continue.
5. The ValuePartition names will be verified. Press the ‘Next’ button.
6. The annotations page will be visible. At this point we could add annotations to the
ValuePartition if we wanted. However, at the moment we won’t, so press the ‘Next’
button to continue.
7. The final page of the wizard prompts us to specify a class that will act as a ‘root’
under which all ValuePartitions will be created. We recommend that ValuePartitions
are created under a class named ValuePartition, which is the default option. Press the
Finish button to create the ValuePartition.
Let’s look at what the wizard has done for us (refer to Figure 4.59 and Figure 4.60):
1. A ValuePartition class has been created as a subclass of owl:Thing.
2. A SpicinessValuePartition class has been created as a subclass of ValuePartition.
3. The classes Mild, Medium, Hot have been created as subclasses of SpicinessValuePartition.
4. The classes Mild, Medium and Hot have been made disjoint from each other.
74

Figure 4.59: Classes Added by the ‘Create ValuePartition’ Wizard
Figure 4.60: The Conditions Widget Displaying the Description of the SpicinessValuePartition Class
5. A class that is the union of Mild, Medium and Hot has been created as the subclass of Spiciness-
ValuePartition (see Figure 4.60).
6. A hasSpiciness object property has been created.
7. The hasSpiciness property has been made functional
8. SpicinessValuePartition has been set as the range of the hasSpiciness property.
4.14.1 Covering Axioms
As part of the ValuePartition pattern we use a covering axiom. A covering axiom consists of two parts:
The class that is being ‘covered’, and the classes that form the covering. For example, suppose we have
three classes A, B and C. Classes B and C are subclasses of class A. Now suppose that we have a covering
axiom that speciﬁes class A is covered by class B and also class C. This means that a member of class A
must be a member of B and/or C. If classes B and C are disjoint then a member of class A must be a
member of either class B or class C. Remember that ordinarily, although B and C are subclasses if A an
individual may be a member of A without being a member of either B or C.
75

In Protégé-OWL a covering axiom manifests itself as a class that is the union of the classes being covered,
which forms a superclass of the class that is being covered. In the case of classes A, B and C, class A
would have a superclass of B C. The effect of a covering axiom is depicted in Figure 4.61.
A
B
C
A
B
C
Without a covering axiom
(B and C are subclasses of A)
With a covering axiom
(B and C are subclasses of A
and A is a subclass of B union C)
Figure 4.61: A schematic diagram that shows the effect of using a Covering Axiom to cover class A with classes
B and C
Our SpicinessValuePartition has a covering axiom to state that SpicinessValuePartition is covered by the
classes Mild, Medium and Hot — Mild, Medium and Hot are disjoint from each other so that an individual
cannot be a member of more than one of them. The class SpicinessValuePartition has a superclass that
is Mild Medium Hot. This covering axiom means that a member of SpicinessValuePartition must be
a member of either Mild or Medium or Hot.
The difference between not using a covering axiom, and using a covering axiom is depicted in Figure 4.62.
In both cases the classes Mild, Medium and Hot are disjoint — they do not overlap. It can be seen that in
the case without a covering axiom an individual may be a member of the class SpicinessValuePartition
and still not be a member of Mild, Medium or Hot — SpicynessValuePartition is not covered by Mild,
Medium and Hot. Contrast this with the case when a covering axiom is used. It can be seen that if an
individual is a member of the class SpicinessValuePartition, it must be a member of one of the three
subclasses Mild, Medium or Hot — SpicinessValuePartition is covered by Mild, Medium and Hot.
SpicinessValuePartition
Mild
Medium
Hot
SpicinessValuePartition
Mild
Medium
Hot
Without a covering axiom With a covering axiom
(SpicinessValuePartition is covered by
Mild,Medium,Hot)
Figure 4.62: The effect of using a covering axiom on the SpicinessValuePartition
76

4.15 Using the Properties Matix Wizard
We can now use the SpicinessValuePartition to describe the spiciness of our pizza toppings. To do this
we will add an existential restriction to each kind of PizzaTopping to state it’s spiciness. Restrictions
will take the form, ∃ hasSpiciness SpicynessValuePartition, where SpicinessValuePartition will be one
of Mild, Medium or Hot. As we have over twenty toppings in our pizza ontology this could take rather
a long time. Fortunately, the Properties Matrix Wizard can help to speed things up. The properties
matrix wizard can be used to add existential restrictions along specified properties to many classes in a
quick and efficient manner.
Exercise 40: Use the properties matrix wizard to specify the spiciness of pizza toppings
1. Invoke the property matrix wizard by selecting the ‘Properties Matrix’ item from
the ‘Wizards’ menu on the Protégé menu bar.
2. The first page to be displayed in the property matrix wizard is the classes selection
page shown in Figure 4.63. By selecting toppings in the class hierarchy, and using the
buttons in the middle of the page (‘>>’ and ‘<<’) classes may be transferred to the
right hand side list. Select all of the pizza topping classes and transfer them to the
right hand side list as shown in Figure 4.63. You should only select the classes that
are ‘actual’ toppings, so classes such as CheeseTopping should not be selected. After
selecting the toppings press the ‘Next’ button on the wizard.
3. The wizard should now be displaying the page shown in Figure 4.64. Select the has-
Spiciness property and use the (>>) button to move the property to the right hand
column (as shown in Figure 4.64). Press the ‘Next’ button on the wizard.
4. In the final page on the wizard, the property fillers should be specified. This is done
by double clicking on each class that is listed and selecting a filler of either Mild,
Medium or Hot. Select fillers of Mild for everything except PepperoniTopping and
SalamiTopping, which should have fillers of Medium, and JalapenoPepperTopping
and SpicyBeef, which should have fillers of Hot. After selecting fillers, the wizard
page should resemble Figure 4.65.
5. Press the ‘Finish’ button to create the restrictions on the toppings and close the
wizard. After the wizard has closed, select some different toppings and notice that
they have restrictions on the along the hasSpiciness property, with fillers of subclasses
of the SpicinessValuePartition.
To complete this section, we will create a new class SpicyPizza, which should have pizzas that have spicy
toppings as its subclasses. In order to do this we want to define the class SpicyPizza to be a Pizza
that has at least one topping (hasTopping) that has a spiciness (hasSpiciness) that is Hot. This can be
accomplished in more than one way, but we will create a restriction on the hasTopping property, that
77

Figure 4.63: Property Matrix Wizard: Class Selection Page
Figure 4.64: Property Matrix Wizard: Property Selection Page
Figure 4.65: Property Matrix Wizard: Restriction Fillers Page
78

Figure 4.66: The definition of SpicyPizza
has a restriction on the hasSpiciness property as its filler.
Exercise 41: Create a SpicyPizza as a subclass of Pizza
1. Create a subclass of Pizza called SpicyPizza.
2. With SpicyPizza selected in the class hierarchy, select the “NECESSARY & SUFFI-
CIENT” header in the conditions widget.
3. Press the ‘Create restriction’ button on the conditions widget to show the ‘Create
Restriction Dialog’.
4. Select ‘∃ someValuesFrom’ as the type of restriction.
6. The filler should be: PizzaTopping ∃ hasSpiciness Hot. This filler describes an
anonymous class, which contains the individuals that are members of the class Piz-
zaTopping and also members of the class of individuals that are related to the members
of class Hot via the hasSpiciness property. In other words, the things that are Piz-
zaToppings and have a spiciness that is Hot. To enter this restriction as a filler type,
PizzaTopping and some hasSpiciness Hot. The “and” keyword will be converted
to the intersection symbol , the “some” keyword will be converted to the existential
quantifier symbol ∃.
7. The ‘Create Restriction Dialog’ should now appear similar to the picture shown
in Figure 4.67. Press the OK button to close the dialog and create the restriction.
8. Finally, drag Pizza from under the “NECESSARY” header to on top of the newly
created restriction (∃ hasTopping (PizzaTopping ∃ hasSpiciness Hot)).
The conditions widget should now look like the picture shown in Figure 4.66
79

Figure 4.67: Create Restriction Dialog: A Restriction Describing a Spicy Topping
Our description of a SpicyPizza above says that all members of SpicyPizza are
Pizzas and have at least one topping that has a Spiciness of Hot. It also says that
anything that is a Pizza and has at least one topping that has a spiciness of Hot
is a SpicyPizza.
In the final step of Exercise 41 we created a restriction that had the class expres-
sion (PizzaTopping ∃ hasSpiciness Hot) rather than a named class as its filler.
This filler was made up of an intersection between the named class PizzaTopping
and the restriction ∃ hasSpiciness Hot. Another way to do this would have been
to create a subclass of PizzaTopping called HotPizzaTopping and define it to be
a hot topping by having a necessary condition of ∃ hasSpiciness Hot. We could
have then used ∃ hasTopping HotPizzaTopping in our definition of SpicyPizza.
Although this alternative way is simpler, it is more verbose. OWL allows us to
essentially shorten class descriptions and definitions by using class expressions in
place of named classes as in the above example.
We should now be able to invoke the reasoner and determine the spicy pizzas in our ontology.
1. Press the ‘Classifiy Taxonomy’ button on the OWL toolbar to invoke the reasoner
and classify the ontology.
80

worksWith
Matthew
Nick
worksWith
Hai
Figure 4.68: Cardinality Restrictions: Counting Relationships
After the reasoner has finished, the ‘Inferred Hierarchy’ class pane will pop open, and you should find
that AmericanHotPizza has been classified as a subclass of SpicyPizza — the reasoner has automatically
computed that any individual that is a member of AmericanHotPizza is also a member of SpicyPizza.
4.16 Cardinality Restrictions
In OWL we can describe the class of individuals that have at least, at most or exactly a specified number
of relationships with other individuals or datatype values. The restrictions that describe these classes are
known as Cardinality Restrictions. For a givien property P, a Minimum Cardinality Restriction specifies
the minimum number of P relationships that an individual must participate in. A Maximum Cardinality
Restriction specifies the maximum number of P relationships that an individual can participate in. A
Cardinality Restriction specifies the exact number of P relationships that an individual must participate
in.
Relationships (for example between two individuals) are only counted as separate relationships if it can
be determined that the individuals that are the fillers for the relationships are different to each other. For
example, Figure 4.68 depicts the individual Matthew related to the individuals Nick and the individual
Hai via the worksWith property. The individual Matthew satisfies a minimum cardianlity restriction of
2 along the worksWith property if the individuals Nick and Hai are distinct individuals i.e. they are
different individuals.
Let’s add a cardinality restriction to our Pizza Ontology. We will create a new subclass of Pizza called
81

Figure 4.69: The Conditions Widget Displaying the Description of an InterestingPizza
InterestingPizza, which will be defined to have three or more toppings.
Exercise 43: Create an InterestingPizza that has at least three toppings
1. Switch to the OWLClasses tab and make sure that the Pizza class is selected.
2. Create a subclass of Pizza called InterestingPizza.
3. Select the “NECESSARY & SUFFICIENT” header in the conditions widget.
4. Press the ‘Create restriction’ button to bring up the ‘Create restriction dialog’.
5. Select ‘≥ minCardinality’ as the type of restriction to be created.
6. Select hasTopping as property to be restricted.
7. Specify a minimum cardinality of three by typing 3 into the restriction filler edit box.
8. Press the ‘OK’ button to close the dialog and create the restriction.
9. The conditions widget should now have a “NECESSARY” condition of Pizza, and a
“NECESSARY & SUFFICIENT” condition of hasTopping ≥ 3. We need to make
Pizza part of the necessary and sufficient conditions. Drag Pizza and drop it on top
of the hasTopping ≥ 3 condition.
The conditions widget should now appear like the picture shown in Figure 4.69.
82

What does this mean? Our definition of an InterestingPizza describes the set
of individuals that are members of the class Pizza and that have at least three
hasTopping relationships with other (distinct) individuals.
1. Press the ‘Classify Taxonomy’ button on the OWL toolbar.
After the reasoner has classified the ontology, the ‘Inferred Hierarchy’ window will pop open. Expand
the hierarchy so that InterestingPizza is visible. Notice that InterestingPizza now has subclasses Ameri-
canaPizza, AmericanHotPizza and SohoPizza — notice MargheritaPizza has not been classified under
InterestingPizza because it only has two distinct kinds of topping.
83

Chapter 5
More On Open World Reasoning
This examples in this chapter demonstrate the nuances of Open World Reasoning.
We will create a NonVegetarianPizza to complement our categorisation of pizzas into VegetarianPizzas.
The NonVegetarianPizza should contain all of the Pizzas that are not VegetarianPizzas. To do this
we will create a class that is the complement of VegetarianPizza. A complement class contains all of
the individuals that are not contained in the class that it is the complement to. Therefore, if we create
NonVegetarianPizza as a subclass of Pizza and make it the complement of VegetarianPizza it should
contain all of the Pizzas that are not members of VegetarianPizza.
Exercise 45: Create NonVegetarianPizza as a subclass of Pizza and make it disjoint to Vegetarian-
Pizza
1. Select Pizza in the class hierarchy on the ‘OWLClasses’ tab. Press the ‘Create
subclass’ button to create a new class as the subclass of Pizza.
2. Rename the new class to NonVegetarianPizza.
3. Make NonVegetarianPizza disjoint with VegetarianPizza — while NonVegetarian-
Pizza is selected, press the ‘Add named class’ button on the disjoint classes widget
(Figure 4.5).
84

We now want to define a NonVegetarianPizza to be a Pizza that is not a VegetarianPizza.
Exercise 46: Make VegetarianPizza the complement of VegetarianPizza
1. Make sure that NonVegetarianPizza is selected in the class hierarchy on the
‘OWLClasses tab’.
2. Select the “NECESSARY & SUFFICIENT” header in the ‘Conditions Widget’.
3. Press the ‘Create new expression’ button, which will ‘pop’ open an ‘inline expres-
sion editor’ in the ‘Conditions Widget’ as shown in Figure 5.1. The inline expression
editor contains an edit box for typing expressions into, and the expression builder panel
(the same one that is found in the ‘Create restriction dialog’), which can be used
to insert class names and logical symbols into the edit box.
4. Type not VegetarianPizza into the edit box. The “not” keyword will be converted
into the ‘complement of ’ symbol (¬). Alternatively, to enter the expression using the
expression builder panel, use the ‘Insert complementOf’ button shown in Figure
5.3 to insert the complementOf symbol, and the use the ‘Insert class’ button (Figure
5.3) to display a dialog from which VegetarianPizza can be selected.
5. Press the return key to create and assign the expression. If everything was entered
correctly then the inline expression editor will close and the the expression will have
been created. (If there are errors, check the spelling of “VegetarianPizza”).
A very useful feature of the expression editor is the ability to ‘auto complete’ class
names, property names and individual names. The auto completion for the inline
expression editor is activated using the tab key. In the above example if we had
typed Vege into the inline expresion editor and pressed the tab key, the choices
to complete the word Vege would have poped up in a list as shown in Figure 5.2.
The up and down arrow keys could then have been used to select VegetarianPizza
and pressing the Enter key would complete the word for us.
The conditions widget should now resemble to picture shown in 5.4. However, we need to add Pizza to the
necessary and sufficient conditions as at the moment our definition of NonVegetarianPizza says that an
individual that is not a member of the class VegetarianPizza (everything else!) is a NonVegetarianPizza.
85

Figure 5.1: Conditions Widget: Inline Expression Editor
Figure 5.2: Conditions Widget: Inline Expression Editor Auto Completion
Insert complement of
Insert class
Figure 5.3: Using the Expression Builder Panel to insert Complement Of
Figure 5.4: The Conditions Widget Displaying the Intermediate Step of Creating a Deﬁnition
forNonVegetarianPizza
86

Figure 5.5: The Conditions Widget Displaying the Definition forNonVegetarianPizza
Exercise 47: Add Pizza to the necessary and sufficient conditions for NonVegetarianPizza
1. Make sure NonVegetarianPizza is selected in the class hierarchy on the ‘OWLClases’
tab.
2. Select Pizza in the ‘Conditions Widget’.
3. Drag Pizza from under the “NECESSARY” header, and drop it onto the ‘¬ Vegetar-
ianPizza’ condition to add it to the same set of necessary and sufficient conditions as
¬ VegetarianPizza.
The complement of a class includes all of the individuals that are not members of
the class. By making NonVegetarianPizza a subclass of Pizza and the comple-
ment of VegetarianPizza we have stated that individuals that are Pizzas and are
not members of VegetarianPizza must be members of NonVegetarianPizza. Note
that we also made VegetarianPizza and NonVegetarianPizza disjoint so that if an
individual is a member of VegetarianPizza it cannot be a member of NonVege-
tarianPizza.
1. Press the ‘Classify taxonomy’ button on the OWL toolbar. After a short time
the reasoner will have computed the inferred class hierarchy, and the inferred class
hierarchy pane will pop open.
The inferred class hierarchy should resemble the picture shown in Figure 5.6. As can be seen, Margher-
87

itaPizza and SohoPizza have been classified as subclasses of VegetarianPizza. AmericanaPizza and
AmericanHotPizza have been classified as NonVegetarianPizza. Things seemed to have worked. How-
ever, let’s add a pizza that does not have a closure axiom on the hasTopping property.
Exercise 49: Create a subclass of NamedPizza with a topping of Mozzarella
1. Create a subclass of NamedPizza called UnclosedPizza.
2. Making sure that UnclosedPizza is selected in the ‘Conditions Widget’ select the
“NECESSARY” header.
3. Press the ‘Create restriction’ button to display the ‘Create restriction dialog’.
4. Select ‘∃ someValuesFrom’ in order to create an existential restriction.
6. Type MozzarellaTopping into the filler edit box to specify that the toppings must be
individuals that are members of the class MozzarellaTopping.
7. Press the ‘OK’ button to close the dialog and create the restriction.
If an individual is a member of UnclosedPizza it is necessary for it to be a Named-
Pizza and have at least one hasTopping relationship to an individual that is a
member of the class MozzarellaTopping. Remember that because of the Open
World Assumption and the fact that we have not added a closure axiom on the
hasTopping property, an UnclosedPizza might have additional toppings that are
not kinds of MozzarellaTopping.
1. Press the ‘Classify taxonomy’ button on the OWL toolbar.
Examine the class hierarchy. Notice that UnclosedPizza is neither a VegetarianPizza or NonVegetari-
anPizza.
88

Figure 5.6: The Inferred Class Hierarchy Showing Inferred Subclasses of VegetarianPizza and NonVegetarian-
Pizza
89

As expected (because of Open World Reasoning) UnclosedPizza has not been
classified as a VegetarianPizza. The reasoner cannot determine UnclosedPizza is
a VegetarianPizza because there is no closure axiom on the hasTopping and the
pizza might have other toppings. We therefore might have expected Unclosed-
Pizza to be classified as a NonVegetarianPizza since it has not been classified
as a VegetarianPizza. However, Open World Reasoning does not dictate that
because UnclosedPizza cannot be determined to be a VegetarianPizza it is not
a VegetarianPizza — it might be a VegetarianPizza and also it might not be a
VegetarianPizza! Hence, UnclosePizza cannot be classified as a NonVegetarian-
Pizza.
90

Chapter 6
Creating Other OWL Constructs In
Protégé-OWL
This chapter discusses how to create some other owl constructs using Protégé-OWL . These constructs
are not part of the main tutorial and may be created in a new Protégé-OWL project if desired.
6.1 Creating Individuals
OWL allows us to define individuals and to assert properties about them. Individuals can also be used
in class descriptions, namely in hasValue restrictions and enumerated classes which will be explained in
section 6.2 and section 6.3 respectively. To create individuals in Protégé-OWL the ‘Individuals Tab’ is
used.
Suppose we wanted to describe the country of origin of various pizza toppings. We would first need to
add various ‘countries’ to our ontology. Countries, for example, ‘England’, ‘Italy’, ‘America’, are typically
thought of as being individuals (it would be incorrect to have a class England for example, as it’s members
would be deemed to be, ‘things that are instances of England’). To create this in our Pizza Ontology we
91

Figure 6.1: The Individuals Tab
will create a class Country and then ‘populate’ it with individuals:
Exercise 51: Create a class called Country and populate it with some individuals
1. Create Country as a subclass of owl:Thing.
2. Switch to the ‘Individuals Tab’ shown in Figure 6.1 and select the class Country in
the ‘Classes’ tree.
3. Press the ‘Create Instance’ button shown in Figure 6.2. (Remember that ‘Instance’
is another name for ‘Individual’ in ontology terminology).
4. An individual that is a member of Country will be created with a auto-generated name.
Rename the individual using the ‘Name’ widget (located on the individuals tab to
the right of the class view and instances list) to Italy.
5. Use the above steps to create some more individuals that are members of the class
Country called America, England, France, and Germany.
92

Create instance
(of selected class)
Duplicate selected instance
Delete selected instance
Figure 6.2: Instances Manipulation Buttons
Figure 6.3: The SameAs Widget
Recall from section 3.2.1 that OWL does not use the Unique Name Assumption (UNA). Individuals can
therefore be asserted to be the ‘Same As’ or ‘Different From’ other individuals. In Protégé-OWL these
assertions can be made using the ‘SameAs’ and ‘DifferentFrom’ tabs shown in Figure 6.3, which are
located with the ‘Name’ widget on the ‘Individuals’ tab.
Having created some individuals we can now use these individuals in class descriptions as described in
section 6.2 and section 6.3.
6.2 hasValue Restrictions
A hasValue restriction, denoted by the symbol , describes the set of individuals that have at least one
relationship along a specified property to a specific individual. For example, the hasValue restriction
hasCountryOfOrigin Italy (where Italy is an individual) describes the set of individuals (the anonymous
class of individuals) that have at least one relationship along the hasCountryOfOrigin property to the
specific individual Italy. For more information about hasValue restrictions please see Appendix A.2.
Suppose that we wanted to specify the origin of ingredients in our pizza ontology. For example, we might
want to say that mozzarella cheese (MozzarellaTopping) is from Italy. We already have some countries
in our pizza ontology (including Italy), which are represented as individuals. We can use a hasValue
93

Figure 6.4: The Conditions Widget Displaying The hasValue Restriction for MozzarellaTopping
restriction along with these individuals to specify the county of origin of MozzarellaTopping as Italy.
Exercise 52: Create a hasValue restriction to specify that MozzarellaTopping has Italy as its country
of origin.
1. Switch to the ‘Properties’ tab. Create a new object property and name it hasCoun-
tryOfOrigin.
2. Switch to the ‘OWLClasses’ tab and select the class MozzarellaTopping.
3. Select the “NECESSARY” header in the ‘Conditions Widget’.
4. Press the ‘Create restriction’ button on the ‘Conditions Widget’ to bring up the
‘Create restriction dialog’.
5. Select hasValue as the type of restriction to be created.
6. Select hasCountryOfOrigin as the property to be restricted.
7. In the restriction filler box enter the individual Italy as a filler — either type Italy
into the filler edit box, or, press the ‘Insert individual’ button on the expression
builder panel, which will pop open a dialog box from which the individual Italy may
be selected.
8. Press ‘OK’ to close the dialog and create the restriction.
The conditions that we have specified for MozzarellaTopping now say that: in-
dividuals that are members of the class MozzarellaTopping are also members of
the class CheeseTopping and are related to the individual Italy via the hasCoun-
tryOfOrigin property and are related to at least one member of the class Mild
via the hasSpiciness property. In more natural English, things that are kinds of
mozzarella topping are also kinds of cheese topping and come from Italy and are
mildly spicy.
94

With current reasoners the classification is not complete for individuals. Use in-
dividuals in class descriptions with care — unexpected results may be caused by
the reasoner.
6.3 Enumerated Classes
As well as describing classes through named superclasses and anonymous superclasses such as restrictions,
OWL allows classes to be defined by precisely listing the individuals that are the members of the class. For
example, we might define a class DaysOfTheWeek to contain the individuals (and only the individuals)
Sunday, Monday, Tuesday, Wednesday, Thursday, Friday and Saturday. Classes such as this are known
as enumerated classes.
In Protégé-OWL enumerated classes are defined using the ‘Conditions Widget’ expression editor – the
individuals that make up the enumerated class are listed (separated by spaces) inside curly brackets. For
example {Sunday Monday Tuesday Wednesday Thursday Friday Saturday}. The individuals must first
have been created in the ontology. Enumerated classes described in this way are anonymous classes –
they are the class of the individuals (and only the individuals) listed in the enumeration. We can attach
these individuals to a named class in Protégé-OWL by creating the enumeration as a “NECESSARY &
SUFFICIENT” condition.
Exercise 53: Convert the class Country into an enumerated class
1. Switch the the ‘OWLClasses’ tab and select the class Country.
2. Select the “NECESSARY & SUFFICIENT” header in the ‘Conditions Widget’.
3. Press the ‘Create new expression’ button. The inline expression editor will pop
open.
4. Type {America England France Germany Italy} into the expression edit box. (Re-
member to surround the items with curly brackets). Remember that the auto complete
function is available — to use it type the first few letters of an individual and press
the tab key to get a list of possible choices.
5. Press the enter key to accept the enumeration and close the expression editor.
95

Figure 6.5: The Conditions Widget Displaying An Enumeration Class
This means that an individual that is a member of the Country class must be one
of the listed individuals (i.e one of America England France Germany Italy.a
More
formally, the class country is equivalent to (contains the same individuals as) the
anonymous class that is deﬁned by the enumeration — this is depicted in Figure
6.6.
aThis is obviously not a complete list of countries, but for the purposes of this ontology (and
this example!) it meets our needs.
England
Italy
America
France
Germany
Enumerated Class
(dashed line)
Country (solid line)
Figure 6.6: A Schematic Diagram Of The Country Class Being Equivalent to an Enumerated Class
The enumerated classes wizard is available for creating enumerated classes in the
above fashion.
6.4 Annotation Properties
OWL allows classes, properties, individuals and the ontology itself (technically speaking the ontology
header) to be annotated with various pieces of information/meta-data. These pieces of information may
take the form of auditing or editorial information. For example, comments, creation date, author, or,
references to resources such as web pages etc. OWL-Full does not put any constraints on the usage
of annotation properties. However, OWL-DL does put several constraints on the usage of annotation
properties — two of the most important constaints are:
96

• The filler for annotation properties must either be a data literal1
, a URI reference or an individual.
• Annotation properties cannot be used in property axioms — for example they may not be used
in the property hierarchy, so they cannot have sub properties, or be the sub property of another
property. The also must not have a domain and a range set for them.
OWL has five pre-defined annotation properties that can be used to annotate classes (including anony-
mous classes such as restrictions), properties and individuals:
1. owl:versionInfo — in general the range of this property is a string.
2. rdfs:label — has a range of a string. This property may be used to add meaningful, human readable
names to ontology elements such as classes, properties and individuals. rdfs:label can also be used
to provide multi-lingual names for ontology elements.
3. rdfs:comment — has a range of a string.
4. rdfs:seeAlso — has a range of a URI which can be used to identify related resources.
5. rdfs:isDefinedBy — has a range of a URI reference which can be used to reference an ontology that
defines ontology elements such as classes, properties and individuals.
For example the annotation property rdfs:comment is used to store the comment for classes in the Protégé-
OWL plugin. The annotation property rdfs:label could be used to provide alternative names for classes,
properties etc.
There are also several annotation properties which can be used to annotate an ontology. The ontology
annotation properties (listed below) have a range of a URI reference which is used to refer to another
ontology. It is also possible to the use the owl:VersionInfo annotation property to annotate an ontology.
• owl:priorVersion — identifies prior versions of the ontology.
• owl:backwardsCompatibleWith — identifies a prior version of an ontology that the current ontology
is compatible with. This means that all of the identifiers from the prior version have the same
intended meaning in the current version. Hence, any ontologies or applications that reference the
prior version can safely switch to referencing the new version.
• owl:incompatibleWith — identifies a prior version of an ontology that the current ontology is not
compatible with.
To create annotation properties the ‘Create annotation datatype property’ and ‘Create annota-
tion object property’ buttons on the ‘Properties’ tab should be used. To use annotation properties
the annotations widgets shown in Figure 6.7 is used. An annotations widget is located on the OWL-
Classes, Properties, Individuals and Metadata tab for annotation classes, properties, individuals and the
ontology respectively. Annotations can also be added to restrictions and other anonymous classes by right
clicking (ctrl click on a Mac) in the conditions widget and selecting ‘Edit annotation properties...’.
1A data literal is the character representation of a datatype value, for example,“Matthew”, 25, 3.11.
97

Create a new
annotation value
Add a TODO list
annotation item
Add an existing individual
as an annotation
Delete the selected
annotation
Figure 6.7: An annotations widget
6.5 Multiple Sets Of Necessary & Sufficient Conditions
In OWL it is possible to have multiple sets of necessary and sufficient conditions as depicted in Figure
6.8. In the ‘Conditions Widget’, multiple sets of necessary and sufficient conditions are represented
using multiple “NECESSARY & SUFFICIENT” headers with necessary and sufficient conditions listed
under each header as shown in Figure 6.8. To create a new set of necessary and sufficient conditions, any
“NECESSARY & SUFFICIENT” header (any that is visible) should be selected and then the condition
created (for example using the ‘Create Restriction dialog’). Alternatively, a condition should be
dragged and dropped onto a “NECESSARY & SUFFICIENT” header to create a new set of necessary
and sufficient conditions and move the condition to that new set. To add to an existing set of necessary
and sufficient conditions, one of the conditions in the set should be selected and then the condition
created (for example using ‘Create Restrictions dialog’), or an existing condition may be dragged and
dropped onto the existing set (below the “NECESSARY & SUFFICIENT” header) to add the condition
to the existing set.
NECESSARY CONDITIONS
Condition
Condition
NamedClass
Condition
Condition
implies
Condition
Condition
Implies
Im
plies
Figure 6.8: Necessary Conditions, and Multiple Sets of Necessary And Sufficient Conditions
98

Figure 6.9: The Definition of a Triangle Using Multiple Necessary & Sufficient Conditions
Exercise 54: Create a class to define a Triangle using multiple sets of Necessary & Sufficient con-
ditions
1. Create a subclass of owl:Thing named Polygon.
2. Create a subclass of Polygon named Triangle.
3. Create an object property named hasSide.
4. Create an object property named hasAngle.
5. On the ‘OWLClasses’ tab select the Triangle class. Select the “NECESSARY &
SUFFICIENT” header in the ‘Conditions Widget’. Press the ‘Create restriction
button’ on the ‘Conditions Widget’ to display the ‘Create restriction dialog’.
6. Select = cardinality as the type of restriction to be created. Select hasSide as the
property to be restricted. In the filler edit box type 3. Press ‘OK’ to close the dialog
and create the restriction
7. Select the “NECESSARY & SUFFICIENT” header in the ‘Conditions Widget’
again. Press the ‘Create restriction’ button to display the ‘Create restriction’
dialog.
8. Select = cardinality as the type of restriction to be created. Select hasAngle as the
property to be restricted. In the filler edit box type 3. Press ‘OK’ to close the dialog
and create the restriction.
9. Drag Polygon from under the “NECESSARY” header and drop it onto the hasSide
= 3 restriction.
10. Select the hasAngle = 3 restriction. Click the ‘Add named class...’ button to
display a dialog containing the class hierarchy. Select the Polygon class and click the
‘OK’ button to close the dialog.
99

Chapter 7
Other Topics
7.1 Language Profile
As explained in section 3.1 on page 11, there are three sub-languages of OWL: OWL-Lite, OWL-DL and
OWL-Full. When editing an ontology Protégé-OWL offers the ability to constrain the constructs used
in class expressions etc. so that the ontology being edited falls into either the OWL-DL or OWL-Full
sub-langauges. The desired sub-language, or ‘language profile’ to be used can be set via the Protégé-
OWL preferences dialog shown in Figure 4.41 on page 52. To choose between OWL-DL and OWL-Full,
the ‘OWL (supports one of the following OWL species)’ should be ticked, and then either the
‘OWL DL (optimized for reasoning)’ or the ‘OWL Full (supports the complete range of OWL
elements)’ should be selected.
7.2 Namespaces And Importing Ontologies
OWL ontologies are able to import other OWL ontologies rather like importing packages in java, or
including files in C/C++. This section describes namespaces, which are a general naming mechanism
and are usually used to facilitate ontology importing. It then describes how to import ontologies in
general.
7.2.1 Namespaces
Every ontology has its own namespace — this is known as the default namespace. An ontology may
also use other namespaces. A namespace is a string of characters that prefixes the class, property and
individual identifiers in an ontology. By maintaining different namespaces for different ontologies it
is possible for one ontology to reference classes, properties and individuals in another ontology in an
unambiguous manner and without causing name clashes. For example, all OWL ontologies (including the
Pizza ontology developed in this tutorial) reference the class owl:Thing. This class resides in the OWL
vocabulary ontology that has the namespace http://www.w3.org/2002/07/owl#.
In order to ensure that namespaces are unique they manifest themselves as Unique Resource Identifiers
100

(URIs)1
ending in either ‘/’ or ‘#’. For example, the default namespace in Protégé-OWL (the names-
pace that is assigned to newly created ontologies in Protégé-OWL ) is http://a.com/ontology#. This
means that all identifiers for classes, properties and individuals that are created in Protégé-OWL (by de-
fault) are prefixed with http://a.com/ontology#. For example, the full name for the class PizzaTopping is
http://a.com/ontology#PizzaTopping. The full name for the class MargheritaPizza is http://a.com/ontology#MargheritaPizz
Fortunately, Protégé-OWL hides these namespace prefixes which means that we don’t have to type in
these long winded names every time we want to use a class, property or individual identifier.
Namespaces help to avoid name clashes when one ontology references classes, properties and individu-
als in another ontology. For example, suppose an ontology about aircraft, AircraftOntology has a class
named Wing, which describes the wing of an aeroplane. An ontology about birds, BirdOntology also
has a class named Wing, which describes the wing of a bird. The namespace for the AircraftOntology is
http://www.ontologies.com/aircraft#. The namespace for the BirdOntology is http://www.birds.com/ontologies/BirdOntology
Evidently, the Wing class in the AircraftOntology is not the same as the Wing class in the BirdOntology.
Now suppose that the AircraftOntology imports the BirdOntology. Because of the namespace mechanism,
the full name for the Wing class in the AircraftOntology is http://www.ontologies.com/aircraft#Wing. The
full name for the Wing class in the BirdOntology is http://www.birds.com/ontologies/BirdOntology#Wing.
Hence, when the AircraftOntology refers to classes in the BirdOntology no name clash will occur. Note
that neither of the above namespace URIs necessarily have to be URLs i.e. they don’t necessarily have
to have a physical location (on the web) — URIs are used because they ensure Uniqueness.
In order to make referencing classes, properties and individuals more manageable when using multiple
namespaces, namespace prefixes are used. A namespace prefix is a short string, usually a sequence
of around two or three characters that represents a full namespace. For example, we could use “ac”
to represent the above ‘aircraft ontology’ namespace http://www.ontologies.com/aircraft# and the prefix
“bird” to represent the ‘bird ontology’ namespace http://www.birds.com/ontologies/BirdOntology#. When
we now use identifiers such as class names, we prefix the identifier with the namespace prefix and a colon.
For example ac:Wing or bird:Wing.
For a given ontology, the default namespace is the namespace for that ontology — in Protégé-OWL
the default namespace corresponds to the namespace of the ontology that is being edited. When using
identifiers that belong to the default namespace (the ontology being edited) a namespace prefix is not
used — classes, properties and individuals are simply referenced using their ‘local’ name. However, for
imported ontologies we must use a namespace prefix to refer to classes, properties and individuals in the
imported ontology. For example, suppose we were editing the ‘aircraft ontology’, which has a namespace
of http://www.ontologies.com/aircraft# and we wanted to refer to classes in the ‘bird ontology’ with the
namespace of http://www.birds.com/ontologies/BirdOntology# and the namespace prefix of “bird”. When
we refer to classes without a namespace prefix, for example Wing, we are talking about classes in the
aircraft ontology. When we refer to classes with a namespace prefix ‘bird’, for example bird:Wing, we are
talking about classes in the bird ontology.
7.2.2 Creating And Editing Namespaces in Protégé-OWL
Editing The Default Namespace
The default namespace can be set using the ‘Default Namespace’ widget, which is located in the top
left corner of the ‘Metadata’ tab and is shown in Figure 7.1. To change the default namespace simply
type a new namespace into the edit box. The namespace must be a valid URI and must end in either ‘/’
or ‘#’. Some examples of valid namespaces are listed below:
1Note that Unique Resource Locators (URLs), which identify physical locations of documents (e.g. web pages) are a
special form of URI.
101

Default Namespace
Add prefix
Remove prefix
Make default namespace Add to ont-policy file
Figure 7.1: The Default Namespace and Namespaces Widget
• myNameSpace#
• universityOfManchester:ontologies:pizzas#
• http://www.cs.man.ac.uk/ontologies/pizzas/
Creating Other Namespaces
As well as specifying a default namespace for the ontology it is possible to setup other namespace prefix
- namespace mappings. This makes it possible to refer to classes, properties and individuals in other
ontologies.
To create/setup namespaces and their associated prefixes in Protégé-OWL the ‘Namespace Prefixes’
widget shown in Figure 7.1 is used. This widget contains three columns: ‘Prefix’, ‘Namespace’ and
‘Imported’ — we will deal with the ‘Imported’ column later. When Protégé-OWL creates a new OWL
project it automatically creates/sets up three namespaces:
• rdf — http://www.w3.org/1999/02/22-rdf-syntax-ns# (The Resource Description Framework names-
pace)
• rdfs — http://www.w3.org/2000/01/rdf-schema# (The RDF-Schema namespace)
• owl — http://www.w3.org/2002/07/owl# (The OWL vocabulary namespace)
Let’s add a new namespace and prefix, which we can use in our ontology. For the purposes of this example
we will add a prefix and namespace for the wine ontology2
, which has the namespace http://www.w3.org/TR/2004/REC-
2The wine ontology is discussed and used as an example in the W3C OWL Guide. It contains information about various
types of wine and wineries.
102

owl-guide-20040210/wine#.
Exercise 55: Create a namespace and prefix to refer to classes, properties and individuals in the
Wine ontology
1. Press the ‘Add prefix’ button on the ‘Namespace prefix’ widget shown in Figure
7.1 to create a new namespace. A new namespace http://www.domain2.com# will be
created with a prefix of p1.
2. Double click on the prefix p1 to edit it. Change it to vin, which is the namespace
prefix used in the wine ontology.
3. Double click on the namespace http://www.domain2.com# to edit it. Change it to
http://www.w3.org/TR/2004/REC-owl-guide-20040210/wine#. If the namespace is en-
tered incorrectly i.e. if the namespace is not a valid URI and it does not end in ‘/’
or ‘#’ Protégé-OWL will reject the entry and revert to the previous value for the
namespace.
4. We can now reference concepts in the wine ontology, and create classes and properties
in the wine ontology namespace. For example create a new object property and name
it vin:myWineProperty.
The property myWineProperty resides in the vin namespace http://www.w3.org/TR/2004/REC-owl-guide-
20040210/wine# (hence the prefixed name vin:myWineProperty). The full name of the property is
therefore http://www.w3.org/TR/2004/REC-owl-guide-20040210/wine#myWineProperty.
7.2.3 Ontology Imports in OWL
OWL ontologies may import one or more other OWL ontologies. For example, suppose that the Air-
craftOntology imports the BirdOntology which contains descriptions of various birds (perhaps we want
to simulate bird strikes on aircraft) — all of the classes, properties, individuals and axioms that are con-
tained in the BirdOntology will be available to be used in the AircraftOntology. This makes it possible to
use classes, properties and individuals from the BirdOntology in class descriptions in the AircraftOntology.
It also makes it possible to extend the descriptions of classes, properties and individuals in the BirdOn-
tology by creating the extention descriptions in the AircraftOntology. Notice the distinction between
referring to classes, properties and individuals in an other ontology using namespaces, and completely
importing an ontology. When an ontology imports another ontology, not only can classes, properties
and individuals be referenced by the importing ontology, the axioms and facts that are contained in the
ontology being imported are actually included in the importing ontology. It should be noted that OWL
allows ontology imports to be cyclic so for example OntologyA may import OntologyB and OntologyB
may import OntologyA.
7.2.4 Importing Ontologies in Protégé-OWL
Ontology imports are usually co-ordinated using namespaces. The ontology being imported has its
namespace and also namespace prefix set up and is then imported. To import an ontology in Protégé-
OWL we must first locate the ontology that we want to import and determine its URL. For the purposes
103

of this example we will import the koala ontology, which is a simple ontology created by Holger Knublauch
(author of the Protégé-OWL plugin) that demonstrates the constructs of OWL. The koala ontology is
located on the web at http://protege.stanford.edu/plugins/owl/owl-library/koala.owl.
Let’s import the koala ontology — for the purposes of this tutorial the koala ontology may be imported
into a new, empty OWL ontology.
Exercise 56: Import the koala ontology into an ontology
1. Switch to the ‘Metadata’ tab.
2. Press the ‘Add prefix’ button located on the ‘Namespace prefix’ widget. A new
namespace and namespace prefix will be created.
3. Edit the namespace prefix, changing it to koala.
4. We now need to specify the namespace. When importing ontologies the
namespace should be the actual URL that the ontology is located at, fol-
lowed by ‘/’ or ‘#’. Edit the namespace for the koala prefix, changing it to
http://protege.stanford.edu/plugins/owl/owl-library/koala.owl#.
5. Now tick the ‘Import tick box’ that lies on the same line as the koala prefix and
namespace. You will be presented with a dialog stating that the changes will not take
place until the file is saved and reloaded. Click the ‘Yes’ button on the dialog. If you
haven’t already saved the project you will be presented with the save dialog box —
give the project a name and save it.
After these steps have been performed Protégé-OWL will import the koala ontology and then save and
reload the project. When the project is reloaded the ‘OWLClasses’ tab will be displayed and will
contain classes from the koala ontology – likewise, the properties tab will also display properties from the
koala ontology. It should be noted that imported classes cannot be edited (they cannot have information
retracted) or deleted – class descriptions can only have information added to them.
Alternative Locations
When it is intended that an ontology should be imported, it is usual for the namespace URI to actually
be a URL (i.e. a pointer to a physical location) that points to the location where the ontology may be
found. In most situations this is usually a web address. By ticking the ‘Imported’ tickbox Protégé-OWL
will attempt to find the ontology at the location that is specified by the namespace URI (URL). This
sounds great, but what if there is no internet connection available, or the ontology doesn’t actually exist
at the namespace URI? Fortunately, it is possible to specify an alternative URI (URL), which can point
to a ‘local’ copy of the ontology — for example a URL that points to a location on the hard disk, or
server on the local area network. Alternative locations are specified in the ontology policy file, which is
located in the Protégé-OWL plugin folder. This file does not need to be edited by hand – it can be edited
using the ontology policy dialog.
104

The (namespace) URIs of
ontologies that are in the
ontology policy file.
The alternative URIs
These URIs could point to
files on the local hard disk etc.
Figure 7.2: The Ontology Policy File Dialog
To specify an alternative location for an imported ontology follow these steps:
Exercise 57: Specifing an alternative location for an imported ontology
1. Select the ontology concerned in the ‘Namespaces Prefixes’ widget.
2. Press the ‘Add to ont-policy file’ button that is located on the ‘Namespaces
Prefixes’ widget shown in Figure 7.1. This will open the ontology policy file dialog
shown in Figure 7.2.
3. As can be seen from 7.2 the koala ontology will have been added to the import policy
(last line). To specify an alternative URI double click on the Alternative URI box
on the koala ontology row – any valid URI may be entered. Press ‘OK’ to close the
dialog.
7.2.5 Importing The Dublin Core Ontology
The Dublin Core ontology is based on the Dublin Core Meta Data Terms. The Dublin Core Meta
Data Terms were standardised/developed by The Dublin Core Meta Data Initiative3
. They are a set of
3http://www.dublincore.org/
105

elements/terms that can be used to describe resources — in our case, we can use these terms to describe
the ‘resources’ such as classes, properties and individuals in an ontology. The full set of Dublin Core
Meta Data Terms is described at http://www.dublincore.org/documents/dcmi-terms/, the following
list contains a few examples:
• title — Typically, a Title will be a name by which the resource is formally known.
• creator — Examples of a Creator include a person, an organisation, or a service. Typically, the
name of a Creator should be used to indicate the entity.
• subject — Typically, a Subject will be expressed as keywords, key phrases or classification codes
that describe a topic of the resource. Recommended best practice is to select a value from a
controlled vocabulary or formal classification scheme.
• description — Description may include but is not limited to: an abstract, table of contents,
reference to a graphical representation of content or a free-text account of the content.
• contributor — Examples of a Contributor include a person, an organisation, or a service. Typi-
cally, the name of a Contributor should be used to indicate the entity.
In order to annotate classes and other ontology entities with the above information and other Dublin
Core Meta Data Terms the Dublin Core Meta Data ontology (DC Ontology) must be imported. Because
Dublin Core Meta Data is so frequently used Protégé-OWL has an automated mechanism for importing
the Dublin Core Meta Data ontology. The ontology can be imported in following manner:
Exercise 58: Import the Dublin Core Meta Data Elements Ontology
1. From the ‘OWL Menu’ select ‘Dublin Core metadata...’.
2. A dialog box will appear. Tick the tickbox on the dialog to import the ontology.
3. A message will be displayed saying that Protégé-OWL needs to reload the ontology.
Press the ‘Yes’ button.
4. After a few seconds the Dublin Core Meta Data ontology will have been imported.
Close the ‘Dublin Core metadata’ dialog with the ‘Close’ button.
5. Switch to the ‘Properties’ tab. As shown in Figure 7.3, a number of annotation
properties (representing the Dublin Core Meta Data Terms) will have been imported.
These annotation properties may be used in the standard way (described in section
6.4).
7.2.6 The Protégé-OWL Meta Data Ontology
Several features used by the Protégé-OWL plugin (such as marking classes so that any primitive subclasses
are automatically made disjoint) rely on the use of various Protégé-OWL annotation properties. These
annotation properties are contained in the Protégé-OWL Meta Data Ontology, which is located in the
106

Figure 7.3: Imported Dublin Core Elements Available As Annotation Properties
107

Protégé-OWL plugin folder. In order for these annotation properties to be used, the Protégé-OWL Meta
Data Ontology must be imported in the following manner:
Exercise 59: Import the Protégé-OWL Meta Data Ontology
1. Select the ‘Preferences...’ item from the ‘OWL Menu’ to display the preferences
dialog shown in Figure 4.41 on page 52.
2. Tick the ‘Import protege metadata ontology’ tickbox. You will be presented with
a ‘Confirm Reload’ dialog box asking you to reload the ontology for the changes to
take effect. Press the ‘Yes’ button.
7.3 Ontology Tests
Protégé-OWL provides a test framework, which contains various tests that may be run on the ontology
being edited. These tests range from sanity tests such as checking that a property’s characteristics
correspond correctly with its inverse property’s characteristics, to OWL-DL tests, which are designed to
find constructs such as metaclasses that put an ontology into OWL-Full. The test framework is based
upon a plugin architecture that enables new tests to be added by third party programmers – check the
Protégé-OWL website for the availability of addon tests.
The various tests may be configured via the ‘Test Settings’ dialog shown in Figure 7.4, which is accessible
via the ‘Test Settings...’ item on the ‘OWL Menu’. To run the tests the ‘Run Ontology Tests...’
item should be selected from the ‘OWL Menu’, or the ‘Run Ontology Tests...’ button should be
pressed on the OWL Toolbar.
After the ontology test have been run the results are displayed in a popup pane at the bottom of the
screen as shown in Figure 7.5. The test results pane has the following columns:
• Type — The type of test result (a warning, and error etc.).
• Source — The source of the test result (e.g. a class or property). Double clicking on the source
will automatically navigate to the source, by automatically selecting a class on the ‘OWLClasses’
tab, or a property on the ‘Properties’ tab for example.
• Test Result — A message that describes the result obtained.
In some cases Protégé-OWL is able to modify/correct aspects of the ontology that the tests have found
to be at fault. In these cases, when the test is selected the small ‘spanner’ button on the left hand side
of the test results pane will be enabled. Clicking this button will repair the ontology fault that gave rise
to the test result.
7.4 TODO List
Protégé-OWL features a simple but useful TODO List mechanism. Classes, properties and the ontology
itself can be annotated with TODO items. These can be attached to classes, properties and the ontology
108

Figure 7.4: The Ontology Test Settings Dialog
Figure 7.5: Ontology TestResults
109

by using the ‘Add TODO List Item’ button that is located on the ‘Annotation Widgets’. Pressing
the ‘Add TODO List Item’ creates a new annotation property that may be filled in with a textual
description of the TODO task. To locate TODO items the ‘Show TODO List...’ item should be
selected from the ‘OWL Menu’ or the ‘Show TODO List...’ button pressed on the OWL Toolbar.
This will display a list of TODO items in a popup pane at the bottom of the screen. Double clicking on
each TODO item in this list will cause Protégé-OWL to automatically navigate to the TODO item in
the ontology.
110

Appendix A
Restriction Types
This appendix contains further information about the types of property restrictions in OWL. It is intended
for readers who aren’t too familiar with the notions of logic that OWL is based upon.
All types of restrictions describe an unnamed set that could contain some individuals. This set can be
thought of as an anonymous class. Any individuals that are members of this anonymous class satisfy the
restriction that describes the class (Figure A.1). Restrictions describe the constraints on relationships
that the individuals participate in for a given property.
When we describe a named class using restrictions, what we are effectively doing is describing anonymous
superclasses of the named class.
A.1 Quantifier Restrictions
Quantifier restrictions consist of three parts:
1. A quantifier, which is either the existential quantifier (∃), or the universal quantifier (∀).
2. A property, along which the restriction acts.
3. A filler that is a class description.
For a given individual, the quantifier effectively puts constraints on the relationships that the individual
participates in. It does this by either specifying that at least one kind of relationship must exist, or by
specifying the only kinds of relationships that can exist (if they exist).
A set of individuals that satisfy
a restriction - the restriction essentially
describes an anonymous (unnamed) class
that contains these individuals.
Figure A.1: Restrictions Describe Anonymous Classes Of Individuals
111

prop
prop
prop
prop
Class A
Anonymous class
prop
prop
pro
p
prop
Figure A.2: A Schematic Of An Existential Restriction (∃ prop ClassA)
A.1.1 someValuesFrom – Existential Restrictions
Existential restrictions, also known as ‘someValuesFrom’ restrictions, or ‘some’ restrictions are denoted
using ∃ – a backwards facing E. Existential restrictions describe the set of individuals that have at
least one specific kind of relationship to individuals that are members of a specific class. Figure A.2
shows an abstracted schematic view of an existential restriction, ∃ prop ClassA – i.e. a restriction along
the property prop with a filler of ClassA. Notice that all the individuals in the anonymous class that
the restriction defines have at least one relationship along the property prop to an individual that is
a member of the class ClassA. The dashed lines in Figure A.2 represent the fact that the individuals
may have other prop relationships with other individuals that are not members of the class ClassA even
though this has not been explicitly stated — The existential restriction does not constrain the prop
relationship to members of the class ClassA, it just states that every individual must have at least one
prop relationship with a member of ClassA — this is the open world assumption (OWA).
For a more concrete example, the existential restriction, ∃ hasTopping MozzarellaTopping, describes the
set of individuals that take place in at least one hasTopping relationship with an other individual that
is a member of the class MozzarellaTopping — in more natural English this restriction could be viewed
as describing the things that ‘have a Mozzarella topping’. The fact that we are using an existential
restriction to describe the group of individuals that have at least one relationship along the hasTopping
property with an individual that is a member of the class MozzarellaTopping does not mean that these
individuals only have a relationship along the hasTopping property with an individual that is a member
of the class MozzarellaTopping (there could be other hasTopping relationships that just haven’t been
explicity specified).
A.1.2 allValuesFrom – Universal Restrictions
Universal restrictions are also known as ‘allValuesFrom’ restrictions, or ‘All’ restrictions since they con-
strain the filler for a given property to a specific class. Universal restrictions are given the symbol ∀ –
i.e. an upside down A. Universal restrictions describe the set of individuals that, for a given property,
only have relationships to other individuals that are members of a specific class. A feature of universal
restrictions, is that for the given property, the set of individuals that the restriction describes will also
contain the individuals that do not have any relationship along this property to any other individuals. A
universal restriction along the property prop with a filler of ClassA is depicted in Figure A.3. Once again,
an important point to note is that universal restrictions do not ‘guarentee’ the existence of a relationship
for a given property. They merely state that if such a relationship for the given property exists, then it
must be with an individual that is a member of a specified class.
112

prop
prop
prop
Class A
Anonymous class
prop
Figure A.3: A Schematic View Of The Universal Restriction, ∀ prop ClassA
Let’s take a look at an example of a universal restriction. The restriction, ∀ hasTopping TomatoTopping
describes the anonymous class of individuals that only have hasTopping relationships to individuals
that are members of the class TomatoTopping, OR, individuals that definitely do not participate in any
hasTopping relationships at all.
A.1.3 Combining Existential And Universal Restrictions in Class Descrip-
tions
A common ‘pattern’ is to combine existential and universal restrictions in class definitions for a given
property. For example the following two restrictions might be used together, ∃ hasTopping Mozzarel-
laTopping, and also, ∀ hasTopping MozzarellaTopping. This describes the set of individuals that have at
least one hasTopping relationship to an individual from the class MozzarellaTopping, and only hasTop-
ping relationships to individuals from the class MozzarellaTopping.
It is worth noting that is particularly unusual (and probably an error), if when describing a class, a
universal restriction along a given property is used without using a ‘corresponding’ existential restriction
along the same property. In the above example, if we had only used the universal restriction ∀ hasTopping
Mozzarella, then we would have described the set of individuals that only participate in the hasTopping
relationship with members of the class Mozzarella, and also those individuals that do not participate in
any hasTopping relationships – probably a mistake.
A.2 hasValue Restrictions
A hasValue restriction, denoted by the symbol , describes an anonymous class of individuals that are
related to another specific individual along a specified property. Contrast this with a quantifier restric-
tion where the individuals that are described by the quantifier restriction are related to any indvidual
from a specified class along a specified property. Figure A.4 shows a schematic view of the hasValue
restriction prop abc. This restriction describes the anonymous class of individuals that have at least
one relationship along the prop property to the specific individual abc. The dashed lines in Figure A.4
represent the fact that for a given individual the hasValue restriction does not constrain the property
used in the restriction to a relationship with the individual used in the restriction i.e. there could be
other relationships along the prop property. It should be noted that hasValue restrictions are semanti-
cally equivalent to an existential restriction along the same property as the hasValue restriction, which
has a filler that is an enumerated class that contains the individual (and only the individual) used in the
hasValue restriction.
113

Anonymous class
prop
pro
p
prop
prop
prop
prop
prop
prop
abc
Figure A.4: A Schematic View Of The hasValue Restriction, prop abc — dashed lines indicate that this
type of restriction does not constrain the property used in the hasValue restriction solely to the
individual used in the hasValue restriction
A.3 Cardinality Restrictions
Cardinality restrictions are used to talk about the number of relationships that an individual may partici-
pate in for a given property. Cardinality restrictions are conceptually easier to understand than quantifier
restrictions, and come in three flavours: Minumum cardinality restrictions, Maximum cardinality restric-
tions, and Cardinality restrictions.
A.3.1 Minimum Cardinality Restrictions
Minimum cardinality restrictions specify the minimum number of relationships that an individual must
participate in for a given property. The symbol for a minimum cardinality restriction is the ‘greater than
or equal to’ symbol (≥). For example the minimum cardinality restriction, ≥ hasTopping 3, describes the
individuals (an anonymous class containing the individuals) that participate in at least three hasTopping
relationships. Minimum cardinality restrictions place no maximum limit on the number of relationships
that an individual can participate in for a given property.
A.3.2 Maximum Cardinality Restrictions
Maximum cardinality restrictions specify the maximum number of relationships that an individual can
participate in for a given property. The symbol for maximum cardinality restrictions is the ‘less than
or equal to’ symbol (≤). For example the maximum cardinality restriction, ≤ hasTopping 2, describes
the class of individuals that participate in at most two hasTopping relationships. Note that maximum
cardinality restrictions place no minimum limit on the number of relationships that an individual must
participate in for a specific property.
114

A.3.3 Cardinality Restrictions
Cardinality restrictions specify the exact number of relationships that an individual must participate in
for a given property. The symbol for a cardinality restrictions is the ‘equals’ symbol (=). For example,
the cardinality restriction, = hasTopping 5, describes the set of individuals (the anonymous class of
individuals) that participate in exactly five hasTopping relationships. Note that a cardinality restriction
is really a syntactic short hand for using a combination of a minimum cardinality restriction and a
maximum cardinality restriction. For example the above cardinality restriction could be represented by
using the intersection of the two restrictions: ≤ hasTopping 5, and, ≥ hasTopping 5.
A.3.4 The Unique Name Assumption And Cardinality Restrictions
OWL does not use the Unique Name Assumption (UNA)1
. This means that different names may refer to
the same individual, for example, the names “Matt” and “Matthew” may refer to the same individual (or
they may not). Cardinality restrictions rely on ‘counting’ distinct individuals, therefore it is important
to specify that either “Matt” and “Matthew” are the same individual, or that they are different individ-
uals. Suppose that an individual “Nick” is related to the individuals “Matt”, “Matthew” and “Matthew
Horridge”, via the worksWith property. Imagine that it has also been stated that the individual “Nick” is
a member of the class of individuals that work with at the most two other individuals (people). Because
OWL does not use the Unique Name Assumption, rather than being viewed as an error, it will be inferred
that two of the names refer to the same individual2
.
1Confusingly, some reasoners (such as RACER) do use the Unique Name Assumstion!
2If “Matt”, “Matthew” and “Matthew Horridge” have been asserted to be different individuals, then this will make the
knowledge base inconsistent.
115

Appendix B
Complex Class Descriptions
An OWL class is speciﬁed in terms of its superclasses. These superclasses are typically named classes
and restrictions that are in fact anonymous classes. Superclasses may also take the form of ‘complex
descriptions’. These complex descriptions can be built up using simpler class descriptions that are
cemented together using logical operators. In particular:
• AND ( ) — a class formed by using the AND operator is known as an intersection class. The class
is the intersection of the individual classes.
• OR ( ) — A class formed by using the OR operator is known as a union class. The class formed
is the union of the individual classes.
B.1 Intersection Classes ( )
An intersection class is described by combining two or more classes using the AND operator ( ). For
example, consider the intersection of Human and Male — depicted in Figure B.1. This describes an
anonymous class that contains the individuals that are members of the class Human and the class Male.
The semantics of an intersection class mean that the anonymous class that is described is a subclass of
Human and a subclass of Male.
The anonymous intersection class above can be used in another class description. For example, suppose
we wanted to build a description of the class Man. We might specify that Man is a subclass of the
anonymous class described by the intersection of Human and Male. In other words, Man is a subclass of
Human and Male.
B.2 Union Classes ( )
A union class is created by combining two or more classes using the OR operator ( ). For example,
consider the union of Man and ‘Woman’ — depicted in Figure B.2. This describes an anonymous class
that contains the individuals that belong to either the class Man or the class Woman (or both).
116

MaleHuman
Intersection of Human and Male
Figure B.1: The intersection of Human and Male (Human Male) — The shaded area represents the intersection
Man
Woman
Figure B.2: The union of Man and Woman (Man Woman) — The shaded area represents the union
The anonymous class that is described can be used in another class description. For example, the class
Person might be equivalent of the union of Man and Woman.
117

Protege owl tutorialp3_v1_0

More Related Content

What's hot (20)

Similar to Protege owl tutorialp3_v1_0 (20)

Recently uploaded (20)

Protege owl tutorialp3_v1_0