Informatics 2D ⋅ Agents and Reasoning ⋅ 2019/2020

Lecture 10 ⋅ First-Order Logic

Claudia Chirita
School of Informatics, University of Edinburgh

4th February 2020

Based on slides by: Jacques Fleuriot, Michael Rovatsos, Michael Herrmann, Vaishak Belle
Outline

• Why FOL?
• Syntax and semantics of FOL
• Using FOL
• Wumpus world in FOL

2 / 24
Propositional logic as a language
Compared to languages in Computer Science
• serves as a basis for declarative languages
• allows partial/disjunctive/negated information
⋅ unlike most data structures and databases
• is compositional
⋅ e.g. the meaning of 𝐵 , ∧ 𝑃 , is derived from that of
𝐵 , and of 𝑃 ,
⋅ unlike some instances of concurrent programming
Compared to natural languages
• meaning is context-independent
⋅ unlike natural languages, where meaning depends on
context
• propositional logic has very limited expressive power
⋅ e.g. we can say pits cause breezes in adjacent squares
only by writing one sentence for each square
3 / 24
First-order logic

Propositional logic deals with atomic facts (i.e. atomic,

non-structured propositional symbols; usually finitely many).

FOL brings structure to facts, which can be built from:

Objects: people, houses, numbers, colours, football games
Functions: father of, best friend, one more than, plus
Relations: red, round, prime, brother of, bigger than, part of

4 / 24
Example ⋅ Of brothers and kings

5 / 24
Example ⋅ Of brothers and kings

Brother(KingJohn, RichardTheLionheart)
Length(LeftLegOf(Richard)) > Length(LeftLegOf(John))

6 / 24
Syntax ⋅ Signatures
A first-order signature is a pair (𝐹, 𝑃)
𝐹 – indexed family (𝐹 ) ∈ℕ of sets of function symbols
(operations)
𝑃 – indexed family (𝑃 ) ∈ℕ of sets of relation symbols
(predicates)

For 𝜎 ∈ 𝐹 and 𝜋 ∈ 𝑃 , 𝑛 is called arity.

Constant symbols are function symbols with arity zero.

Example
functions 𝐹 = {Richard, John}, 𝐹 = {LeftLegOf}
predicates 𝑃 = {Crown, King, Person}
𝑃 = {Brother, OnHead}

7 / 24
Syntax ⋅ Sentences
Terms Least set 𝑇 such that 𝜎(𝑡 , … , 𝑡 ) ∈ 𝑇
for every 𝜎 ∈ 𝐹 and 𝑡 , … , 𝑡 ∈ 𝑇 .
In particular, 𝑇 contains all constants.
Variables Every set of (𝐹, 𝑃)-variables 𝑋 determines
an extended signature (𝐹 ∪ 𝑋, 𝑃) with the
variables in 𝑋 added to 𝐹 as new constants.

Sentences over a signature (𝐹, 𝑃) are defined by the grammar

𝜑 ∶∶= 𝜋(𝑡 , … , 𝑡 ) ∣ 𝑡 = 𝑡 atoms
∣ ¬𝜑 ∣ 𝜑 ∧ 𝜑 ∣ 𝜑 ∨ 𝜑 ∣ 𝜑 → 𝜑 ∣ 𝜑 ↔ 𝜑 boolean connectives
∣ ∀𝑋.𝜑 ∣ ∃𝑋.𝜑 quantifiers

where 𝜋 ∈ 𝑃 is a predicate symbol, 𝑡, 𝑡 , 𝑡 , … , 𝑡 are terms,

and 𝑋 is a set of variables.
Precedence: ∀𝑋, ∃𝑋, ¬, ∧, ∨, →, ↔
8 / 24
Syntax ⋅ Sentences
Sentences over a signature (𝐹, 𝑃) are defined by the grammar
𝜑 ∶∶= 𝜋(𝑡 , … , 𝑡 ) ∣ 𝑡 = 𝑡 atoms
∣ ¬𝜑 ∣ 𝜑 ∧ 𝜑 ∣ 𝜑 ∨ 𝜑 ∣ 𝜑 → 𝜑 ∣ 𝜑 ↔ 𝜑 boolean connectives
∣ ∀𝑋.𝜑 ∣ ∃𝑋.𝜑 quantifiers

where 𝜋 ∈ 𝑃 is a predicate symbol, 𝑡, 𝑡 , 𝑡 , … , 𝑡 are terms,

and 𝑋 is a set of variables.
Precedence: ∀𝑋, ∃𝑋, ¬, ∧, ∨, →, ↔
Example
Brother(John, Richard)
Brother(John, Richard) ∧ Brother(Richard, John)
¬Brother(LeftLegOf(Richard), John)
¬King(Richard) → King(John)
∀𝑥.King(𝑥) → Person(𝑥)
9 / 24
Semantics ⋅ Models

Given a signature (𝐹, 𝑃), a model 𝑀 consists of

• a non-empty set |𝑀|, called the carrier set (domain)
of 𝑀, whose elements are called objects
• a function 𝑀 ∶ |𝑀| → |𝑀| for each operation symbol
𝜎∈𝐹
• a subset 𝑀 ⊆ |𝑀| for each relation symbol 𝜋 ∈ 𝑃

Examples

10 / 24
Satisfaction relation

The satisfaction relation links the syntax and the semantics.

• We write 𝑀 ⊧ 𝜑 and read “𝑀 satisfies 𝜑”, for 𝑀 a model
and 𝜑 a sentence, both for the same signature (𝐹, 𝑃).
• To make (𝐹, 𝑃) explicit, we sometimes write 𝑀 ⊧( , ) 𝜑.
• The satisfaction relation is defined according to the
structure of sentences (in the following slides), based on
the evaluation of terms in models.

Evaluation of terms
• 𝑀 denotes the interpretation of a term 𝑡 in a model 𝑀.
• 𝑀 ( ,…, ) = 𝑀 (𝑀 , … , 𝑀 )
e.g. 𝑀LeftLegOf(John) = 𝑀LeftLegOf (𝑀John )

= 𝑀LeftLegOf ( )=

11 / 24
Satisfaction relation ⋅ 𝑀 ⊧ 𝜑

Atoms
• 𝑀⊧𝑡=𝑡 iff 𝑀 =𝑀
• 𝑀 ⊧ 𝜋(𝑡 , … , 𝑡 ) iff (𝑀 , … , 𝑀 ) ∈ 𝑀

Boolean connectives
• 𝑀 ⊧ ¬𝜑 iff 𝑀⊭𝜑
• 𝑀 ⊧𝜑 ∧𝜑 iff 𝑀 ⊧ 𝜑 and 𝑀 ⊧ 𝜑
• 𝑀 ⊧𝜑 ∨𝜑 iff 𝑀 ⊧ 𝜑 or 𝑀 ⊧ 𝜑
• 𝑀⊧𝜑 →𝜑 iff 𝑀 ⊧ 𝜑 whenever 𝑀 ⊧ 𝜑
• 𝑀⊧𝜑 ↔𝜑 iff 𝑀 ⊧ 𝜑 → 𝜑 and 𝑀 ⊧ 𝜑 → 𝜑

12 / 24
Satisfaction relation ⋅ 𝑀 ⊧ 𝜑
Quantifiers

A model 𝑀 for (𝐹 ∪ 𝑋, 𝑃) is called an expansion of a model

𝑀 for (𝐹, 𝑃) if it interprets all symbols in 𝐹 and in 𝑃 the same
as 𝑀. Expansions formalize assignments of elements from 𝑀
to the variables in 𝑋.

Example

𝑀 𝑀′
13 / 24
 Satisfaction relation ⋅ 𝑀 ⊧ 𝜑
Quantifiers

A model 𝑀 for (𝐹 ∪ 𝑋, 𝑃) is called an expansion of a model

𝑀 for (𝐹, 𝑃) if it interprets all symbols in 𝐹 and in 𝑃 the same
as 𝑀. Expansions formalize assignments of elements from 𝑀
to the variables in 𝑋.

• 𝑀 ⊧( , ) ∀𝑋.𝜑 iff 𝑀 ⊧( ∪ , ) 𝜑
for all expansions 𝑀 along the inclusion (𝐹, 𝑃) ⊆ (𝐹 ∪ 𝑋, 𝑃)

• 𝑀 ⊧( ∃𝑋.𝜑 iff there exists an expansion 𝑀 along the

, )
inclusion (𝐹, 𝑃) ⊆ (𝐹 ∪ 𝑋, 𝑃) such that 𝑀 ⊧( ∪ , ) 𝜑

14 / 24
Satisfaction relation ⋅ Example

True or False?
Brother(John, Richard) ∧ Brother(Richard, John)
¬Brother(LeftLegOf(Richard), John)
¬King(Richard) → King(John)

15 / 24
Satisfaction relation ⋅ Example

True or False?
∀𝑥.King(𝑥) → Person(𝑥)
𝑥↦ 𝑂 (i.e. 𝑀 = 𝑂 ) 𝑂 (John) is a king → 𝑂 is a person.
𝑥↦𝑂 𝑂 (Richard) is a king → 𝑂 is a person.
𝑥↦𝑂 𝑂 (John’s left leg) is a king → 𝑂 is a person.
𝑥↦𝑂 𝑂 (Richard’s left leg) is a king → 𝑂 is a person.
𝑥↦𝑂 𝑂 (crown) is a king → 𝑂 is a person.
16 / 24
Expressivity ⋅ Quantifiers

∀𝑥.King(𝑥) → Person(𝑥)

17 / 24
Expressivity ⋅ Quantifiers

∀𝑥.King(𝑥) → Person(𝑥)

17 / 24
Expressivity ⋅ Quantifiers

∀𝑥.King(𝑥) → Person(𝑥)

∀𝑥.King(𝑥) ∧ Person(𝑥)

17 / 24
Expressivity ⋅ Quantifiers

∀𝑥.King(𝑥) → Person(𝑥)

∀𝑥.King(𝑥) ∧ Person(𝑥)

17 / 24
Expressivity ⋅ Quantifiers

∃𝑥.Crown(𝑥) ∧ OnHead(𝑥, John)

18 / 24
Expressivity ⋅ Quantifiers

∃𝑥.Crown(𝑥) ∧ OnHead(𝑥, John)

18 / 24
Expressivity ⋅ Quantifiers

∃𝑥.Crown(𝑥) ∧ OnHead(𝑥, John)

∃𝑥.Crown(𝑥) → OnHead(𝑥, John)

18 / 24
Expressivity ⋅ Quantifiers

∃𝑥.Crown(𝑥) ∧ OnHead(𝑥, John)

∃𝑥.Crown(𝑥) → OnHead(𝑥, John)

18 / 24
Expressivity ⋅ Quantifiers
The order of quantifiers
∃𝑋.∀𝑌.𝜑 is not the same thing as ∀𝑌.∃𝑋.𝜑
∃𝑥.∀𝑦.Loves(𝑥, 𝑦)
There is a person who loves everyone in the world.
∀𝑦.∃𝑥.Loves(𝑥, 𝑦)
Everyone in the world is loved by someone.

Duality
𝜑 ∧ 𝜑 ≡ ¬(¬𝜑 ∨ ¬𝜑 ) and 𝜑 ∨ 𝜑 ≡ ¬(¬𝜑 ∧ ¬𝜑 )
∀𝑋.𝜑 ≡ ¬∃𝑋.¬𝜑
∀𝑥.Likes(𝑥, IceCream) ≡ ¬∃𝑥.¬Likes(𝑥, IceCream)
∃𝑋.𝜑 ≡ ¬∀𝑋.¬𝜑
∃𝑥.Likes(𝑥, Broccoli) ≡ ¬∀𝑥.¬Likes(𝑥, Broccoli)
19 / 24
Using FOL ⋅ Kinship domain

Axioms: definitions, theorems

One’s mother is one’s female parent.

∀𝑚, 𝑐.𝑚 = Mother(𝑐) ↔ (Female(𝑚) ∧ Parent(𝑚, 𝑐))
Parent and child are inverse relations.
∀𝑝, 𝑐.Parent(𝑝, 𝑐) ↔ Child(𝑐, 𝑝)
A sibling is another child of one’s parents.
∀𝑥, 𝑦.Sibling(𝑥, 𝑦) ↔ 𝑥 ≠ 𝑦 ∧ ∃𝑝.Parent(𝑝, 𝑥) ∧ Parent(𝑝, 𝑦)
Brothers are siblings.
∀𝑥, 𝑦.Brother(𝑥, 𝑦) → Sibling(𝑥, 𝑦)
The sibling relation is symmetric.
∀𝑥, 𝑦.Sibling(𝑥, 𝑦) ↔ Sibling(𝑦, 𝑥)

20 / 24
Interacting with FOL KBs
Tell/Ask interface

Assertions
Tell(KB, King(John))
Tell(KB, Person(Richard))
Tell(KB, ∀𝑥.King(𝑥) → Person(𝑥))

Queries (goals)
Ask(KB, Person(John)) true
Ask(KB, ∃𝑥.Person(𝑥)) true
Ask(KB, Person(𝑥)) {𝑥/John}, {𝑥/Richard}

Idea
Ask(KB, 𝜑) returns all substitutions 𝜃 such that KB ⊧ 𝜃(𝜑).

21 / 24
Example ⋅ Wumpus world

Suppose a wumpus-world agent is using an FOL KB and

perceives a smell and a breeze (but no glitter) at 𝑡 = 5.
Tell(KB, Percept([Smell, Breeze, None], 5))

Does the KB entail some best action at 𝑡 = 5?

Ask(KB, ∃𝑎.BestAction(𝑎, 5))
Answer: true, {𝑎/Shoot}

Perception
∀𝑡, 𝑠, 𝑏.Percept([𝑠, 𝑏, Glitter], 𝑡) → Glitter(𝑡)

Reflex
∀𝑡.Glitter(𝑡) → BestAction(Grab, 𝑡)

22 / 24
Example ⋅ Wumpus world

The environment
∀𝑥, 𝑦, 𝑎, 𝑏.Adjacent([𝑥, 𝑦]) ↔ [𝑎, 𝑏]∈{[𝑥+1, 𝑦], [𝑥−1, 𝑦],
[𝑥, 𝑦+1], [𝑥, 𝑦−1]}
∀𝑠, 𝑡.At(Agent, 𝑠, 𝑡) ∧ Breeze(𝑡) → Breezy(𝑠)

Squares are breezy near a pit.

Diagnostic rule: infer cause from effect
∀𝑠.Breezy(𝑠) → ∃𝑟.Adjacent(𝑟, 𝑠) ∧ Pit(𝑟)

Causal rule: infer effect from cause

∀𝑟.Pit(𝑟) → (∀𝑠.Adjacent(𝑟, 𝑠) → Breezy(𝑠))

23 / 24
Summary

First-order logic:
• Objects and relations are semantic primitives.
• Syntax: constants, functions, predicates, quantifiers.
• Increased expressive power – sufficient to define the
Wumpus world.

24 / 24