2. Mary gave the green coloured vase to her favorite coursin.

3. John went downtown to deposit his money in a bank.

2.2. Conceptual Graphs

A conceptual graph is a connected bipartite graph. Conceptual relation nodes represent

relations between concepts. Thus, the arcs connecting nodes are not labeled. This is one of the
main differences between conceptual graphs and semantic networks. A conceptual graph
consists of concepts, represented by boxes and conceptual relations are represented by
ellipses. Concept nodes only have arcs to conceptual relation nodes (thus the graph is
bipartite). Concept nodes represent concrete objects, e.g. cat, telephone, restaurant, etc, and
abstract objects, e.g. love, beauty and loyalty.

In order to represent a relation of arity of n a conceptual relation node must have n arcs. Let us
look at some examples to make this clearer.

1-ary relation

bird flies

2-ary relation

dog colour brown

3-ary relation

mother
child parents
father

Let us look at a more detailed example. Suppose that we want to represent the relation that
Mary gave John the book. The corresponding conceptual graph is illustrated below:

person: agent give object

Mary

person: agent
John book

5!
!
Notice that each concept node represents an individual and specifies the type of the individual.
If an individual object is unknown a unique marker can be used in place of the name of the
object. A unique marker is comprised of a hash symbol (#) followed by a number. Each object
has its own unique marker. In the first conceptual graph in Figure 7 the name of the dog is
known. However, in the second conceptual graph it is not known and a unique marker is used
for this purpose.

dog: colour brown

Emma

dog: colour brown

#1352
Figure 7

The first conceptual graph in Figure 7 is equivalent to the conceptual graph in Figure 8.

dog:
#1352
colour brown

name ”emma”
Figure 8

Note that when individuals of a particular type appear on their own (i.e. without the type
specified before the individual name) the name of the individual must appear in inverted
commas, e.g. “emma”.

Example 1

Suppose that we want to represent the following information using a conceptual graph:

Her name was McGill and she called herself Lil, but everyone knew her as Nancy.

name person: name

#941

“McGill” name ”Nancy”

“Lil”

6!
!
In addition to individual markers conceptual nodes can also contain generic markers. A generic
marker is represented by an asterisk *. A generic marker is used to represent an unspecified
individual of a type. The type dog specified in a node is equivalent dog:*. Name variables can
also be used, e.g., *X to indicate an unspecified individual. Let us look at an example of this.
Suppose that we want to represent the following information:

The dog is scratching its ear with its paw.

We need to indicate that the paw and ear belong to the same dog. We can use name variables
for this purpose. This is illustrated below:

dog: *X agent object

scratch

ear
instrument
part

paw part dog: *X

The format that we have used thus far to represent conceptual graphs is referred to as the
Display Form (DF). An alternative notation is the Linear Form (LF) which is a more compact
notation. Suppose that we wanted to represent the following statement:

John is going to Boston by bus.

The conceptual graph in DF notation:

person: agent go destination

John

instance city:
Boston

bus

7!
!
The conceptual graph in LF notation:

[Go] –
(Agnt) → [Person: John]
(Dest) → [City: Boston]
(Inst) → [Bus]

Exercise 2

Translate each of the following sentences into a conceptual graph:

1. Jane gave Tom an ice cream.

2. Basketball players are tall.
3. Paul cut down the tree with an axe.
4. Place all the ingredients in a bowl and mix thoroughly.

3. Logical Representation Scheme

We are only going to look at one logical representation scheme, namely, first-order predicate
logic. Predicate logic is more powerful than propositional logic as it allows for quantification.
Facts are represented as logical propositions and additional information is deduced from these
facts using backward chaining or resolution.

3.1 Converting facts to logical propositions

The first step is to convert facts in English to logical propositions called well-formed formulas
(wffs). The wffs include the following symbols:

• Implies: →
• And: /\
• Or: \/
• Not: ~
• For all: ∀
• The exists: ∃

Example: Consider the following facts on Italian history:

1. Marcus was a man.

2. Marcus was a Pompeian.
3. All Pompeians were Roman.
4. Caesar was a ruler.
5. All Romans were either loyal to Caesar or hated him.
6. Everyone is loyal to someone.
7. People only try to assassinate rulers that they are not loyal to.
8. Marcus tried to assassinate Caesar.
9. All men are people.

These facts can be represented in predicate logic as the following wffs:

8!
!