Rabbit

From OWLED-Wiki

Contents 1 Introduction 2 Language Principles 3 Language Overview 3.1 Concepts 3.2 Relations, Instances, Values and Sentences 3.3 General assumptions and rules 4 Semantics 4.1 Concepts 4.2 Instance 4.3 Sentences 4.4 Declarations 4.4.1 Concept declaration 4.4.2 Instance declaration 4.4.3 Relationship declaration 4.5 Assertions 4.5.1 Concept Assertions 4.5.2 Is a kind of 4.5.3 That 4.5.4 Prepositions – of, by, for, from 4.5.5 Number restrictions: exactly, at least, at most 4.5.6 Every / concept name 4.5.7 No 4.5.8 Simple List 4.5.9 Non-Exclusive List: <n> or more 4.5.10 Usually 4.5.11 Only 4.5.12 Only… or nothing 4.5.13 Mutually Exclusive 4.5.14 Are the same thing 4.5.15 Are different 4.5.16 anything that 4.5.17 Are equivalent 4.5.18 Can only be 4.5.19 Everything that, will also 4.6 Relationships 4.6.1 Relationship Sentences 4.6.2 Specialisation 4.6.3 Equivalent Relationships 4.6.4 Mutual relationship exclusion 4.6.5 Object Property Domain 4.6.6 Object Property Range 4.6.7 Inverse 4.6.8 Transitive 4.6.9 Can only have one object 4.6.10 Can only have one subject 4.6.11 Reflexive 4.6.12 Irreflexive 4.6.13 Symmetric 4.6.14 Asymmetric 4.7 Working with other Ontologies 4.7.1 Import 4.7.2 Refer to 4.7.3 Referencing as a secondary concept

Introduction

This provides an overview of the Rabbit language, its grammar and semantics. Rabbit is a controlled natural language designed to represent the Web Ontology Language OWL in a manner that is more understandable to domain experts with little knowledge of description logics. A “domain expert” is a person with a high level of knowledge in a particular subject area, such as hydrology, medicine, transport etc.

Language Principles

The fundamental principles underlying the design of Rabbit are:

1. To be understandable to domain experts with little or no prior knowledge of the language;
2. To enable domain experts to be able to actively participate in authoring domain ontologies;
3. To have a well defined grammar and semantics and be sufficiently formal to enable Rabbit to expressed as OWL-DL.

In line with the first two principles, Rabbit statements are as natural as possible, and user testing in both comprehension and authoring have been used extensively during the design of Rabbit. However, this testing has highlighted that certain aspects of Rabbit, necessary to support OWL, cannot be made easily understandable. Such aspects are mostly related to Rabbit expressions for OWL properties such as reflexive and transitive. In these cases it is accepted that there is a limit to what any language can achieve. Overall though, Rabbit has been shown to be far more comprehensible than OWL even when OWL is expressed in its most friendly Manchester Syntax.

A good way of thinking about the relationship between Rabbit and OWL is that Rabbit is a way of writing OWL in a way that has the appearance of English, although at times the English may seem a little stilted.

Lastly, in line with the overarching design principle of simplicity for the sake of comprehension, a good Rabbit sentence is a short Rabbit sentence.

Language Overview

Concepts

Relations, Instances, Values and Sentences

Rabbit has five basic types of things that form the language. These are concepts, relationships, instances, values and sentences.

Concepts represent the way we classify the domain. Concepts are things, processes, events and abstract ideas such as laws, boundaries and so on. Concepts represent the way we classify our domain of interest. Examples of concepts are: roads – a thing, flooding – a process, carnival- an event, 1841 Ordnance Survey Act – an abstract concept. In OWL concepts are represented as classes.

Rabbit distinguishes between two forms of concept: core concepts and secondary concepts. Core concepts are those concepts that are directly associated with the domain of interest. For example in the domain of hydrology, a pond would be a core concept. Secondary concepts are those concepts that are not directly related to the domain but which are required to help to describe the core concepts within the domain. For example and again using the example of hydrology, a duck pond is a pond created and maintained to support ducks. In order to properly describe a duck pond we need to use a secondary concept, duck. Since duck is not a concept core to the domain of hydrology, we would typically say nothing or very little about ducks, leaving this to a bird ontology.

An instance is a specific occurrence or realisation of a concept. For example the M25 is an instance of the concept road. In OWL instances are called individuals.

A value is a number, string or some other form of quantification. For example the concept boiling point of water could have the value 100ºC. In OWL data type properties have values as objects.

Relationships are associations that exist between concepts, instances and values. Examples of relationships are “is near” as in “Southampton is near Portsmouth”, “is part of” as in “a wheel is part of a car” and so on. Rabbit has a number of predefined relationships, these are:

“is a kind of”

“is a”

“of”

“for”

“by”

These are described in more detail below.

In OWL relationships are known as properties.

Sentences represent basic Rabbit statements. There are two forms of sentences: those that are related to adding knowledge to an ontology and those that help to manage the ontology. The former are called assertions or axioms since they provide factual information about the ontology and the latter are know as auxiliary sentences.

Rabbit assertions are a way of linking concepts, instances and values together though relationships. A simple assertion is:

Every car is a kind of Vehicle. HN51 005 is a Car.

Which says that the concept “car” is a special type of (subclass of) of the concept vehicle and that HN51 005 is an instance of a car. Because the first assertion says that all cars are also vehicles. We know that HN51 005 is not only a car but also a vehicle.

Auxiliary sentences are there to help piece together an ontology. Example are import sentences which enable the current ontology to access all the descriptions in another ontology.

All Rabbit sentences are terminated by a full stop.

General assumptions and rules

• Plural concepts add an ‘s’ as default (otherwise the appropriate sentence must be written)
• Plural relationships – remove an ‘s’ as default (otherwise the appropriate sentence must be written)
• Plurals – phrases written in the singular imply the plural, except for materials (e.g. water, rock) (if it applies to one, it applies to all).
• “Or” – in the conceptual ontology, the use of the term “or” in a collection means an exclusive or (this or that, but not both) unless stated otherwise.
• Sentences describing the same concept are implicitly concatenated together with AND in the OWL translation.
• The indefinite article will be “an” in cases where the noun phrase begins with a vowel

• No punctuation should be used, with the following exceptions:
  • A full stop must be used to denote the end of a sentence.
  • Commas are used in the sentence structure “Tributary is a concept, plural tributaries.” And in lists or collections: “a sea, a river, or a lake”
  • Semi-colon and colon are used in definition or reference sentence structure, for example

“A Source is anything that: 
                  is a kind of Spring or a Wetland; 
                  feeds a River or a Stream.”

• Round Brackets are used only to for disambiguation of terms, for example Spring (Season) and Spring (Water) and for disambiguation of and/or conjunctions.
• Square brackets are only used for denoting references (provenance of sentences reused from other ontologies).

Semantics

Concepts

Concepts represent the way we classify the domain. Concepts are groups of things, processes, events and abstract ideas such as laws, boundaries and so on. Concepts represent the way we classify our domain of interest. Examples of concepts are: roads – a thing, flooding – a process, carnival- an event, 1841 Ordnance Survey Act – abstract concept. In description logic concepts are represented as classes.

Syntax

Concepts are represented by a capitalised noun phrase. The capitalisation is intended to make it easier for the user to spot which are the concepts and which the relationships, however Rabbit can still be parsed in lowercase.

<concept> ::= <capitalised noun> [ <space> <concept body>]

<concept body> ::= <capitalised noun or integer> [<space> <concept body>]

<capitalised noun or integer> ::= <capitalised noun> | <integer>

Examples

Fish

River Stretch

Agricultural Use

Core and secondary concepts

Rabbit distinguishes between two forms of concept: core concepts and secondary concepts. Core concepts are those concepts that are directly associated with the domain of interest. For example in the domain of hydrology a pond would be a core concept. Secondary concepts are those concepts that are not directly related to the domain but which are required to help to describe the core concepts within the domain. For example and again using the example of hydrology, a duck pond is a pond created and maintained to support ducks. In order to properly describe a duck pond we need to use a secondary concept, duck. Since duck is not a core concept we would typically say nothing or very little about ducks, leaving this to a bird ontology.

Instance

An instance is a specific occurrence or realisation of a concept. For example the M25 is an instance of the concept road. In Description Logics instances are called individuals.

Syntax An Instance can be either upper or lowercase noun phrases.

<instance> ::= <noun or integer> [<space> <instance body>] <instance body> ::= <noun or integer> [<space> <instance body>]

Examples England osgb0001234598768874 The Isle of Wight

Sentences

A sentence is the way in which Rabbit expresses things within an ontology. There are four main different type sentence:

• Declarations, used introduce concepts, instances and relationships.
• Assertions, used to state facts about concepts and instances.
• Relationship modifiers, used to modify the interpretation of a relationship.
• Import sentences, used to specify other ontologies and the objects used within those ontologies.

Syntax

All sentences are terminated by a full stop.

<sentence> ::= <sentence body> .

<sentence body> ::= <concept declaration> 
                  | <concept assertion> 
                  | <instance declaration> 
                  | <instance assertion>
                  | <relationship declaration> 
                  | <relationship modifier> 
                  | <import sentence>

Examples

Car is a concept.

Every Sheep has at most four Legs.

London is located in England.

Declarations

Concept declaration

Concept declarations introduce concepts to the ontology. By default concepts are assumed to be core concepts. Secondary concepts are introduced using the “secondary” key word. By default plurals are constructed by adding an “s” to the last noun of the concept name. Non-standard plurals can be stated explicitly using the “plural” keyword.

Syntax

<concept declaration> ::= <concept> is a [secondary] concept [, plural <concept>][, synonym <synonym list>]

<synonym list> ::= <synonym> [, <synonym list>]

<synonym> ::= <concept> [, plural <concept>]

Examples

Ship is a concept.

River Stretch is a concept, plural River Stretches, synonym Reach, plural Reaches.

Duck is a secondary concept.

Rhyne is a concept, synonym Rhine, Reen

Semantics

This introduces a concept to Rabbit. It has the effect of creating an OWL class as a subclass of Thing. Concepts are not disjoint unless stated otherwise by a concept assertion.

Instance declaration

These introduce instances and specify the concept of which they are an instance.

Syntax

<instance declaration> ::= <instance> is a <concept>

Examples

England is a Country.

Semantics

This introduces an instance (OWL individual) and assigns a class to it. Multiple instance declarations are allowable. E.g.

Brian is a Man.

Brian is a Teacher.

Relationship declaration

This introduces relationships to the ontology.

Syntax

<relationship declaration> ::= <relationship term> is a relationship [, plural <relationship term>]

Examples

“near to” is a relationship.

“has part” is a relationship, plural have part.

Every House has an Address. (This becomes in OWL House -> hasAddress some Address)

Note that relationships can be enclosed in double quotes for clarity but this is not mandatory. When the relationship “has” is used on its own, this is translated in OWL to has + object concept.

Semantics

In OWL it introduces a property, without specifying if it’s a datatype or object property – this distinction must come from the context of the sentence.

Assertions

Concept Assertions

Concept assertions are the bread and butter of Rabbit. They relate to OWL axioms. There are a number of different types of assertion in terms of the syntax used. All are converted to OWL expressions. Usually each sentence converts to one OWL axiom, but occasionally (for example with “only”) there may be two or more.

Syntax

<concept assertion> ::= <concept subsumption> 
                      | <positive assertion> 
                      | <negative assertion> 
                      | <concept definition> 
                      | <Equivalent concepts> 
                      | <probable assertion> 
                      | <value partition> 
                      | <general concept inclusion> 
                      | <complex role inclusion> 
                      | <mutual concept exclusion> 
                      | <closure assertion> 
                      | <all values assertion>

Examples

Every School has purpose Education.

Agriculture is a kind of Activity.

No Backwater has a Current.

Is a kind of

This sentence type enables an assertion to be made that one concept is a sub-class of another.

Syntax

<concept subsumption> ::= <subject> is a kind of <concept>

Examples

Every Car is a kind of Vehicle.

Semantics

OWL Subclass Of

That

Objects are concepts that are associated with the subject of the sentence by a relationship. An object may be concept or it may be a concept that has some criteria applied to it.

Syntax

that <simple relationship phrase>

Examples

Every School has a part a Building that has purpose Education.

Semantics

The semantically That translates to And in OWL. Syntactically the position of That is constrained so that it can only occur as a modifier for an object and cannot form part of a long chain of conjunctions.

OWL: and (position restricted) School -> hasPart some (Building and hasPurpose some Education)

Prepositions – of, by, for, from

Objects are concepts that are associated with the subject of the sentence by a relationship. An object may be a concept in isolation, or it may be a concept that has some criteria applied to it.

Syntax

<preposition modifier> ::= <preposition> <concept>

<preposition> :== of | by | for | from

Examples

Every School has purpose Education of Children.

Semantics

Like That, the prepositions of, from, for and by are also implemented as And and are similarly constrained. However unlike that, which simply substitutes for And, the prepositions serve as object properties:

OWL: ObjectProperies of, by, for and from

School -> hasPurpose some (Education and of some Child)

Number restrictions: exactly, at least, at most

These enable the number of objects concepts to be specified with respect to a particular relationship with the subject.

Syntax

<number restriction> ::= exactly <n> | at least <n> | at most <n> <concept>

Examples

Every car has at least three Wheels.

Every Bike has at most two Wheels.

Every Person has exactly two Parents.

Semantics

OWL: ObjectMinCardinality, ObjectMaxCardinality, ObjectExactCardinality

The examples above convert to the following OWL expressions:

Car -> hasWheel min 3 Wheels

Bike -> hasWheel max 2 Wheels

Person -> hasParent exactly 2 Parent

Every / concept name

This is used to specify that an assertion applies to all instances on a concept. Normally these sentences begin with the word “Every” but this can be omitted in cases where the use of Every would make the Rabbit look very odd English. It is used before all positive and subclass assertions.

Syntax

<positive assertion> ::= <subject> <compound relationship>

<subject> ::= [Every ] <concept name>

Examples

Every Car has part at an Engine.

Agriculture is the responsibility of DEFRA.

Semantics

Universal quantifier, as applied to the subject class.

No

This allows you to say that something is not true about a concept.

Syntax

<negated assertion> ::= No <concept> <compound relationship>

Example

No Backwater has a Current.

Semantics

OWL: not

Backwater -> not (hasCurrent some Current)

Simple List

A simple list allows one of a number of concepts to be select as the object. The list comprises two or more object alternatives. It is very rarely used, and there is often a more appropriate way to model the knowledge than using exclusive or.

Syntax

<simple list> ::= <object> <or> <simple list body>

<simple list body> ::= <object> [<or> simple list body>]

<or> ::= or | ,

Example

Every River terminates in a Sea, River or Lake.

Every Pizza has exactly one Thin Base or exactly one Deep Base.

Semantics

The list imposes the semantics of an exclusive or on the list members.

Non-Exclusive List: <n> or more

This precedes a list and modifies it to specify that the subject can be associated to more than one of the concepts in the list.

Syntax

<object list> ::= <n> or more [of] <object> <or> <object list body>

‘’’Example’’’

Every River flows into one or more of a Sea, River or Lake.

Semantics

The list imposes the semantics of a non-exclusive or on the list members.

River -> flowsInto some (Sea or River or Lake)

Usually

Usually allows the use to specify a relationship that normally exists but not always. It cannot be translated into OWL.

Syntax

<probable assertion> ::= <uc indefinate article> <concept> usually <compound relationship>

Examples

A School usually has part a Car Park.

Semantics

Specifies the typical case. (currently no OWL translation)

Only

This is used to modify a relationship so that it says that the object of the sentence has to apply to the subject, and that will be the only object that can be linked to the subject via that relationship.

Syntax

<closure assertion> ::= <subject> only <relationship phrase>

Example

Every Giant Panda only eats Bamboo.

Semantics

In the example above we’re saying that a Giant Panda has to eat something, can only eat Bamboo and cannot eat anything else. Note that this converts to two OWL axioms, and corresponds to OWL closure.

GiantPanda -> eats some Bamboo and eats only Bamboo.

Only… or nothing

This is used to modify a relationship so that it says that if subject of the sentence has a specific relationship with an object it can only have that relationship with that object and no other concept. This is similar to the use of Only above but allows the subject to not have the relationship at all.

Syntax

<all values assertion> <subject> only <relationship phrase> or nothing

‘’’Example’’’

Every Giant Panda only eats Bamboo or nothing.

In the example for Only, what it says is that all giant pandas have to eat, and when they eat, they only eat bamboo. Add “or nothing” to the end of the sentence would change its meaning to: giant pandas don’t have to eat, but when they do, they only eat bamboo. Not very realistic since pandas that don’t eat won’t remain pandas for very long.

Semantics

Universal quantifier

GiantPanda -> eats only Bamboo.

Mutually Exclusive

This enables us to say that two concepts cannot coexist – in other words something could not be both concepts at the same time. For example, a person can’t be both a man and a women and so they are mutually exclusive concepts.

Syntax

<mutual concept exclusion> ::= <concept> and <concept> are mutually exclusive

‘’’Example’’’

Father and Mother are mutually exclusive.

Semantics

OWL disjoint classes

Are the same thing

enables us to say two instances are the same thing. For example Uncle George might also be know as George Entwistle.

Syntax

<instance> and <instance> are the same thing.

Examples

UK and United Kingdom are the same thing.

Old Bill and William Sidebottom are the same thing.

Semantics

OWL sameAs

Are different

enables us to say two instances cannot be the same thing. For example Uncle George is not Aunty Mabel.

Syntax

<instance> and <instance> are different.

Examples

Great Britain and United Kingdom are different.

George and Mabel are different.

Semantics

OWL differentFrom

anything that

Enables defined concepts to be implemented. (In OWL this is identical to equivalent class) Normally where we make a number of statements about a concept, we are saying that the concept has those properties. We can however make a stronger statement and say that anything that has the properties must be an example (or instance) of that concept. These concepts are known as defined concepts.

For example we could say that a butcher’s shop is where you buy meat. But you can buy meat at other types of shop too, so a Butcher’s shop is not a defined concept. However, a Doctor (in the Academic sense) is anyone that holds a PhD so having a PhD must mean you are a doctor. Normally more than one criterion is used to specify a defined concept, so the Rabbit sentence is made up of a number of clauses, each separated by a semi-colon, and, as usual, the whole thing is terminated by a full stop.

Syntax

<concept definition> ::= <uc indefinate article> <concept> is anything that: <definition relationship list>

<definition relationship list> ::= <definition relationship phrase> [; <definition relationship list>]

<definition relationship phrase> ::= <compound relationship> | does not <compound relationship>

Examples

A City is anything that:

is a kind of settlement;

has a City Charter.

Semantics

OWL equivalent class, for defined classes.

Are equivalent

States that two concepts are the same (have exactly the same set of instances). to be implemented. (In OWL this is identical to equivalent class)

Syntax

<equivalent concept> ::= <concept> and <concept> are equivalent.

Examples

Loch and Scottish lake are equivalent.

Semantics

OWL equivalentClass.

Can only be

This enables the definition of a concept as a list of other concepts. That is it enables a concept to be defined as only being one of a number of other concepts.

Syntax

<value partition> ::= <subject> can only be <simple object list>

Examples

Every Season can only be Spring or Summer or Autumn or Winter.

Semantics

OWL:one Of and value Classes are disjoint.

Everything that, will also

This enables us to say that if a particular relationship holds for a concept then it also follows that another specific relationship will also hold too. In Description Logics this is known as a general concept inclusion (GCI) and complex role inclusion (CRI). CRI is a more general version of GCI.

Syntax

<general concept inclusion> ::= Everything that <relationship> that <relationship> some <concept> will also <relationship> some <concept>

And

<complex role inclusion> ::= Everything that <relationship> that <relationship> something will also <relationship> that thing

Examples

GCI:

Everything that has a Part that contains some Water will also contain some Water.

CRI:

Everything that has a Part that contains something will also contain that thing.

Semantics

In OWL this is expressed as: C subClassOf D

where C and D can both be complex (e.g. anonymous and not named) classes.

e.g.

hasPart and contains some Water o contain some Water

The CRI is a property chain:

hasParent o hasBrother -> hasUncle

Relationships

Relationship Sentences

Relationships are the means by which we link concepts together. They can comprise one or more words. The meaning of relationships can be modified so that they have additional semantics. For example “The relationship “friend of” is reflexive” means that it is valid if the subject and object are swapped over. So if Fred is a friend of Jane and we say that “friend of” is reflexive, they we can infer than Jane is a friend of Fred. The following sentence types enable these modifications. Note that it’s the relationship itself that is being modified – and its new characteristic will apply whenever it’s used, rather than only applying in specific sentences to certain objects.

Syntax

<relationship modifier> ::= <literal relationship modifier> 
                          | <functional relationship modifier> 
                          | <inverse functional relationship modifier> 
                          | <reflexive relationship modifier> 
                          | <irreflexive relationship modifier> 
                          | <relationship subsumption modifier> 
                          | <equivalence relationship modifier> 
                          | <mutual relationship exclusion modifier> 
                          | <relationship domain modifier> 
                          | <relationship range modifier> 
                          | <inverse relationship modifier>

Specialisation

This enables us to say that one relationship is a more specialised form of another relationship.

Syntax

<relationship subsumption modifier> ::= The relationship <relationship term> is a special type of the relationship <relationship term>

Examples

The relationship “is directly connected to” is a special type of the relationship “is connected to”.

Semantics

sub-object property.

Note that unlike subclasses of concepts, restrictions on the object-property are not inherited by the sub-object property. For example if “connect to” is transitive and “touches” is a sub-object property of “connected to” then “touches” is not transitive unless this is explicitly stated.

Equivalent Relationships

This states that two relationships are equivalent and can be used interchangeably. All the instances that are related by one relationship, will also be related by the other, and vice versa

Syntax

<equivalence relationship modifier> ::= The relationships <relationship term> and <relationship term> are equivalent

Examples

The relationships “is contained in” and “is inside” are equivalent.

Semantics

EquivalentObjectProperties.

Mutual relationship exclusion

This states that two relationships cannot be applied to the same subject and object pair at the same time.

Syntax

<mutual relationship exclusion modifier> ::= The relationships <relationship term> and <relationship term> are mutually exclusive

Examples

The relationships “contains” and “is contained in” are mutually exclusive.

Semantics

DisjointObjectProperties

Object Property Domain

This restricts a relationship by saying it can only apply to a particular subject.

Syntax

<relationship domain modifier> ::= The relationship <relationship term> must have the subject <concept>

Examples

The relationship “is the capital city of” must have the subject Capital City.

Semantics

OWL ObjectPropertyDomain

Using the above capital city example, if we say “Paris is the capital city of France”. Paris must be an instance of a Capital City.

Object Property Range

This restricts a relationship by saying it can only apply to a particular object.

Syntax

<relationship range modifier> ::= The relationship <relationship term> must have the object <concept>

Examples

The relationship “is the capital city of” must have the object Country.

Semantics

OWL ObjectPropertyRange

Using the above capital city example, if we say “Paris is the capital city of France”. Therefore, France must and instance of a Country.

Inverse

This allows us to say that one relationship has the reverse effect of another.

Syntax

<inverse relationship modifier> ::= The relationship <relationship term> is the inverse of <relationship term>

Examples

The relationship “contains” is the inverse of “is contained in”.

Semantics

OWL: Inverse Object Property

So for the above example if:

England contains Hampshire then it can be inferred that:

Hampshire is contained in England.

Transitive

This enables us to say that a relationship can be used to form a chain of inference. That is to say if x has a relationship to y and y had the same relationship to z then x also has that relationship with z.

Syntax

<literal relationship modifier> ::= The relationship <relationship term> is transitive

Examples

The relationship “is part of” is transitive.

Semantics

OWL: Transitive Object Property

So in the above example if:

Warsash is part of Hampshire. and

Hampshire is part of England.

Then it can be inferred that :

Warsash is part of England.

Can only have one object

This enables us to say that the relationship between subject and object means that only one instance of an object can apply to the subject for that relationship.

Syntax

<functional relationship modifier> ::= The relationship <relationship term> can only have a single object

Examples

The relationship “is the capital of” can only have one object.

Semantics

OWL: Functional Object Property

In the example above the statements:

London is the capital of The United Kingdom.

And

London is the capital of The UK.

The inference is that the United Kingdom and the UK are the same thing, because London can only have one object, so the two objects must be the same.

Can only have one subject

This enables us to say that the relationship between subject and object means that only one instance of a subject can apply to the object for that relationship.

Syntax

<inverse functional relationship modifier> ::= The relationship <relationship term> can only have a single subject

Examples

The relationship “is capital city of” can only have one subject.

Semantics

OWL: Inverse Functional Object Property

In the example above the statements:

Tripoli is the capital of Libya.

And

Tarabulus is the capital of Libya.

The inference is that Tripoli and Tarabulus are the same thing.

Reflexive

This enables us to say that wherever a relationship is used, the subject of the sentence must also have the relationship reflexively applied to itself.

Syntax

<reflexive relationship modifier> ::= For the relationship <relationship term> [,] everything <relationship term> itself

Examples

For the relationship "flows into" everything flows into itself.

Semantics

ReflexiveObjectProperty

For example, if the relationship “flows into” is reflexive, then if we say “Every River flows into the Sea” we are also implying that “Every River flows into itself”

Irreflexive

This says that a relationship cannot apply reflexively to the subject.

Syntax

<irreflexive relationship modifier> ::= For the relationship <relationship term> [,] nothing <relationship term> itself

Examples

For the relationship “eats”, nothing “eats” itself.

Semantics

OWL: IrreflexiveObjectProperty

For example, if the relationship “eats” is reflexive, then if we say “Every Panda eats Bamboo” we are also implying that “No Panda eats itself”

Symmetric

This states that if a relationship is true between a subject and object then it is also true if the subject and object are swapped over.

Syntax

<literal relationship modifier> ::= The relationship <relationship term> is symmetric

Examples

The relationship “is adjacent to” is symmetric.

Semantics

OWL: Symmetric Object Property

In the example above if we say:

Dorset is adjacent to Hampshire.

Then it can be inferred that Hampshire is adjacent to Dorset.

Asymmetric

This states that if a relationship is true between a subject and object then it is never true if the subject and object are swapped over.

Syntax

<literal relationship modifier> ::= The relationship <relationship term> is asymmetric

Examples

The relationship “larger than” is asymmetric.

Semantics

OWL: Asymmetric Object Property

In the example above if we say:

England is larger than Wales.

Then we know that we can never say:

Wales is larger than England.

Working with other Ontologies

Import

This enables another Rabbit ontology to be accessed.

Syntax

<import sentence> ::= <use ontology sentence> | <reference concept sentence> | <reference instance sentence> | <reference relationship sentence>

<use ontology sentence> ::= Use <single import> | <multiple import>

<single import> ::= ontology: <url ref>

<url ref> ::= <label> from <url>

<label> ::= ‘[‘ <noun> ‘]’

<multiple import> ::= ontologies: <url ref> ; <url list>

<url list> ::= <url ref> [; <url list>]

Examples

Use ontology: Wildfowl from http://ontology.ordnancesurvey.co.uk/Wildfowl.

and

Use ontologies: Wildfowl from http://ontology.ordnancesurvey.co.uk/Wildfowl;

Transport from http://ontology.ordnancesurvey.co.uk/Transport.

Semantics

owl:imports

The statement makes accessible to the ontology all the concepts, relationships and instances contained in the imported ontology. However, these concepts, relationships and instances cannot be used by the importing ontology until they are specifically referenced by means of the Refer to statement.

Refer to

This enables the importing ontology to use an imported concept or instance from an imported ontology.

Syntax

<reference concept sentence> ::= refer to <concept> <label> as <concept>.

<reference instance sentence> ::= refer to <instance> <label> as <instance>.

Examples

Refer to Duck [Wildfowl] as Duck.

Semantics

The semantics of this are to create a local concept Duck and to make an equivalent concept to the Duck concept found in the wildfowl ontology.

Referencing as a secondary concept

Syntax

<reference concept sentence> ::= refer to <reference name> <label> as secondary <reference name>

Examples

Reference <concept> as Secondary from [ref]

Reference River as Secondary from [OSHydro]

OWL Source annotation:

OWL: http://ontology.ordnancesurvey.co.uk/CNL

Semantics

There are no corresponding OWL semantics at present.