Skip to main content

CH6 – Slides

Slides / handouts are available on my website

Abstract
Slides for Ch4 of my logic course.

Introduction

Summary

What we’ve done thus far:

  1. Developed a language (PL)
  2. Translation from PL to English
  3. Clarified “C follows from Premises”
  4. Tables and trees show Semantic entailment
  5. Natural deduction system to show Syntactic entailment

Problem

There are valid English arguments that are not valid in PL.

  • P1: All humans are mortal.
  • P2: Socrates is a human.
  • C: Socrates is mortal.

\(H, S \not\models M\) and \(H, S \not\vdash M\)

Solution

  • We can develop a more expressive logical language!
  • Propositional logic looks at sentential logical relations
  • Quantificational logic looks at sub-sentential logical relations

The language of quantificational logic

Logical language

A logical (formal) language has three parts:

  1. Set of symbols
  2. Syntax (grammar)
  3. Semantics

graph LR
accTitle: Flowchart for QL language
accDescr: A flowchart showing that the QL language consists of Symbols, Syntax, and Semantics
  subgraph QL
    Q["QL Language"]
    Sym["Symbols"]
    Syn["Syntax"]
    Sem["Semantics"]
    Q --> Sym & Syn & Sem
end
  subgraph English
    E[English Sentences]
  end
QL <--> English

QL Symbols

  1. names: Lower case letters, \(a\) through \(v\) with or without numerical subscripts.
  2. n-place predicates: Upper case letters, \(A\) through \(Z\) with or without numerical subscripts.
  3. variables: Lower case letters, \(w\) through \(z\) with or without numerical subscripts.
  4. operators: \(\neg , \wedge , \vee , \rightarrow , \leftrightarrow\)
  5. parentheses: (, )
  6. quantifiers: \(\forall\), \(\exists\)

Exercise

Exercise 6.68, 1-8

QL Syntax

  1. Definition of a wff
  2. Formation rules
  3. Open vs closed wffs

QL Wffs

NoteDefinition

A well-formed formula (or wff) \(\phi\) is any combination of QL symbols capable of being constructed through some combination of the QL formation rules.

QL Formation Rules

NoteDefinition

Let \(P\) be a placeholder for an \(n\)-place predicate and \(x\) be a placeholder for a variable. The formation rules for QL are as follows:

  1. an \(n\)-place predicate \(P\) followed by \(n\) terms (names or variables) is a wff in QL, where \(n\geq 0\).
  2. If \(\phi\) is a wff in QL, then \(\neg (\phi)\) is a wff in QL.
  3. If \(\phi\) and \(\psi\) are wffs in QL, then \((\phi\wedge \psi)\), \((\phi\vee \psi)\), \((\phi\rightarrow \psi)\), and \((\phi\leftrightarrow \psi)\) are wffs in QL.
  4. If \(\phi\) is a wff containing a variable \(x\), then \((\forall x)(\phi)\) and \((\exists x)(\phi)\) are wffs in QL provided that there is no other quantifier containing \(x\) in \(\phi\) (that is, \((\exists x)\) and \((\forall x)\) are not already in \(\phi\)).
  5. Nothing else is a wff in QL except that which can be formed by repeated applications of the above.

Formation rules simplified

NoteDefinition
  1. an \(n\)-place predicate \(P\) followed by \(n\) terms (names or variables) is a wff in QL, where \(n\geq 0\).
  2. .
  3. .
  4. If \(\phi\) is a wff containing a variable \(x\), then \((\forall x)(\phi)\) and \((\exists x)(\phi)\) are wffs in QL provided that there is no other quantifier containing \(x\) in \(\phi\) (that is, \((\exists x)\) and \((\forall x)\) are not already in \(\phi\)).
  5. .

Example 1

Show that \(Rab\land Pa\) is a QL wff.

  1. Rab is a wff. Rule 1
  2. Pa is a wff. Rule 1
  3. \(Rab\land Pa\). Lines 1,2 and Rule 3

Example 2

Show that \(\lnot Rxx\) is a QL wff.

  1. Rxx is a wff. Rule 1
  2. \(\lnot Rxx\) is a wff. Line 1, Rule 2

Example 3

Show that \((\forall x)Px\) is a QL wff.

  1. Px is a wff. Rule 1
  2. \((\forall x)(Px)\). Lines 1 and Rule 4

Example 4

Show that \((\exists x)Rax\) is a QL wff.

  1. \(Rax\) is a wff. Rule 1
  2. \((\exists x)(Rax)\). Lines 1 and Rule 4

Example 5

Show that \((\exists x)((\forall y)(Lxy))\) is a QL wff.

  1. \(Lxy\) is a wff. Rule 1
  2. \((\forall y)(Lxy)\) is a wff. Line 1, Rule 4
  3. \((\exists x)((\forall y)(Lxy))\). Line 2, Rule 4

Examples - Not wffs

Here are some combinations of symbols that are not wffs. Let \(P\) be a one-place predicate and \(R\) be a two-place predicate.

  1. \(Ra\)
  2. \((\forall x)(Pxx)\)
  3. \((\forall x)(Ra)\)
  4. \((\forall x)((\exists x)(Pxx))\)
  5. \(\forall x\exists xPxx\) - wff in different configurations

Exercise

Exercise 6.69, 1-3

Scope of quantifiers

The scope of a quantifier is the same as negation:

  1. \((\forall x)(Lx)\land Rx\) - quantifier has narrow scope
  2. \((\forall x)(Lx\land Rx)\) - quantifier has wide scope

Formula Simplification

The rules for simplification for quantifiers is the same as negation:

  1. \(\neg (P)\) to \(\neg P\)
  2. \((\forall x)(Px)\) to \((\forall x)Px\)

Stacked negations (quantifers):

  1. \(\neg (\neg (P))\) to \(\neg\neg P\)
  2. \((\exists x)((\forall y)(Lxy))\) to \((\exists x)(\forall y)(Lxy)\)

Main operator

NoteDefinition

The main operator of a QL wff is the operator whose scope is the entire wff. It is the operator that is applied last in the construction of the wff.

  1. \((\forall x)Px\)
  2. \((\forall x)Px\land Pa\)
  3. \((\forall x)(Px\land Pa)\)
  4. \((\forall x)(\exists y)(Px\land Py)\)

Exercise

Exercise 6.72, 1-3, 8, 9

Free Variables and Bound Variables

NoteBound variable

A variable \(x\) is bound if and only if it is in the scope of an occurrence of a quantifier that has that variable to its immediate right.

NoteFree Variable

A variable is free if it is not bound.

Examples

  1. \((\forall x)Px\)
  2. \((\forall x)Pxx\)
  3. \((\forall x)(\exists y)Pxy\)
  4. \(Px\)
  5. \(Pax\)
  6. \((\forall x)Pxy\)

Exercise

Exercise 6.73, 1, 2, 10

Open and closed formulas

NoteOpen wff

An open formula is a wff containing at least one free variable.

NoteClosed wff

A closed formula is any wff that does not contain a free variable.

Examples

  1. \((\forall x)Px\)
  2. \((\forall x)Pxx\)
  3. \((\forall x)(\exists y)Pxy\)
  4. \(Px\)
  5. \(Pax\)
  6. \((\forall x)Pxy\)

Exercise

Exercise 6.74, 1-4, 6

Semantics

Semantics

We now have a formal language, the next step is to interpret it (semantics).

graph LR
accTitle: Flowchart for QL language
accDescr: A flowchart showing that the QL language consists of Symbols, Syntax, and Semantics
  subgraph QL
    Q["QL Language"]
    Sym["Symbols"]
    Syn["Syntax"]
    Sem["Semantics"]
    Q --> Sym & Syn & Sem
end
  subgraph English
    E[English Sentences]
  end
QL <--> English

What we need

QL requires:

  1. the domain of discourse: \(\mathcal{D}\)
  2. an interpretation function: \(\mathscr{I}\)
  3. a valuation function: \(v\)

flowchart LR
accTitle: Flowchart for QL language
accDescr: A flowchart showing that the QL language consists of Symbols, Syntax, and Semantics. It further shows that Semantics consists of a model, which is an interpretation and domain, and a valuation function that determines Truth and Falsity. 
  subgraph QL
    Q["QL Language"]
    Sym["Symbols"]
    Syn["Syntax"]
    Sem["Semantics"]
    Q --> Sym & Syn & Sem
    end
    subgraph Model
    D[Domain]
    I[Interpretation]
    end
    val[Valuation]
    T["T"]
    F["F"]
    Sem --> Model
    Model --> val
    val --> T
    val --> F

Domain

NoteDomain

A \(\mathcal{D}\) is simply a set (or collection) of items.

Three ways to indicate:

  1. individually listing each item in the domain
  2. indicating some property that belongs to every item in the domain
  3. indicating some method for determining every item in the domain.

Roster notation (enumeration)

State the domain by specifying a set and enumerating the items in that set.

  1. \(\mathcal{D} = \{1, 2, 3\}\)
  2. \(\mathcal{D} = \{\text{David, Liz, Renna}\}\)
  3. \(\mathcal{D} = \{2, 3, 5, 7\}\)
  4. \(\mathcal{D} = \{1, 2, 3, \ldots\}\)

Set-builder notation

State the domain by specifying a set by describing a condition or characterizing property of all members in the set.

  • \(\{x\ |\ x \text{ is a positive integer less than 4}\}\)
  • The set of all \(x\) such that \(x\) is a positive integer less than $.

Sometimes a colon : is used instead of a pipe |

  • \(\{x\ |\ x \text{ is happy}\} = \{x: x \text{ is happy}\}\)
  • The set of all \(x\) such that \(x\) is happy.

More examples

  1. \(\mathcal{D} = \{x | x \text{ is a prime less than } 10\}\) = \(\{2, 3, 5, 7\}\)
  2. \(\mathcal{D} = \{x | x \text{ is a member of my immediate family}\}\)
  3. \(\mathcal{D} = \{x | x \text{ is a positive integer}\} = \{1, 2, 3, \ldots\}\)

Sets!

There are many interesting things about sets, but we don’t need to know them.

  • Sets can contain sets as members
  • Example: \(\{A, \{A\} \}\)

We don’t need to know this.

Empty set

The empty set is a set containing no members.

  • Symbol for empty set: \(\varnothing\) or \(\{ \}\)
  • Set containing the empty set: \(\{\varnothing\}\)
  • Set containing the empty set and the set of the empty set: \(\{\varnothing, \{\varnothing\}\}\)

The domain will always be non-empty: \(\mathcal{D} \neq \varnothing\).

Preferences

Set-builder is preferred when:

  1. Domain is large (too much work to list everything)
  2. Domain is infinite
  3. Not totally sure what is and is not in the domain, e.g., all humans on earth.

However, for simplicity, we will use the roster (enumeration) approach.

Interpretation

A QL model \(\mathcal{M}\) is a two-part structure consisting of a domain \(\mathcal{D}\) and an interpretation function \(\mathscr{I}\).

  • Domain is just a set of things
  • Interpretation is a two-part function
    1. it takes names in QL and assigns them to items in the domain.
    2. it takes \(n\)-place predicates in QL and assigns them to sets of \(n\)-tuples.

Interpretation of names

What is the meaning of a name?

  • “David” means “beloved”
  • “Renna” means “born again”

Not quite right.

  • “Renna” just has its roots in a common noun that means “born again”.
  • That is not what the name “Renna” means

Direct reference

  • Your name has no meaning, it only has a reference
  • “David W. Agler” refers to me.

QL interprets names in the same way. Let \(\mathcal{D}=\{1, 2, 3\}\)

  • \(\mathscr{I}(a)=1\)
  • \(\mathscr{I}(b)=1\)
  • \(\mathscr{I}(c)=3\)

Further clarifications

  1. You can have more than one name
    • \(\mathscr{I}(a)=1, \mathscr{I}(b)=1\)
    • \(\mathscr{I}(\text{"David"})=\text{Me}, \mathscr{I}(\text{"Davey"})=\text{Me}\)
  2. A name always refers to some item in the domain (no empty names)
    • Never \(\mathscr{I}(a)=\varnothing\)

Interpretation of predicates

  • Predicates are A-Z
  • Predicates are interpreted: \(\mathscr{I}(P)\)
  • Predicates have places: one-place, two-place, three-place, etc.

One-place predicate 1

If \(E\) is a one-place predicate, then \(\mathscr{I}(E)\) is a set of elements from the domain.

  • \(\mathcal{D} = \{1, 2, 3, 4\}\)
  • \(\mathscr{I}(E)=\{2, 4\}\)

One-place predicate 2

If \(H\) is a one-place predicate, then \(\mathscr{I}(H)\) is a set of elements from the domain.

  • \(\mathcal{D} = \{\text{David, Liz, Renna}\}\)
  • \(\mathscr{I}(H)=\{\text{David, Renna}\}\)

Two-place predicate

  • If a predicate \(F\) is a 2-place predicate, it is interpreted as a set of tuples.
  • A tuple is a sequence of elements from the domain grouped together

Example

In the case of tuples, order matters.

  • (1, 2) is the tuple where 1 is first and 2 is second.
  • (2, 1) is the tuple where 2 is first and 1 is second.

Unlike sets, order matters:

  • (1, 2) \(\neq\) (2, 1)
  • \(\{1, 2\} = \{2, 1\}\)

Tuple example 1

If \(G\) is a 2-place predicate, then \(\mathscr{I}(G)\) is a set of tuples from the domain.

  • \(\mathcal{D} = \{1, 2, 3\}\)
  • \(\mathscr{I}(G)=\{(3,1), (3,2), (2,1)\}\)

Tuple example 2

If \(L\) is a 2-place predicate, then \(\mathscr{I}(L)\) is a set of tuples from the domain.

  • \(\mathcal{D} = \{\text{David, Liz, Renna}\}\)
  • \(\mathscr{I}(L)=\{\text{(David, Liz)}\), \(\text{(David,Renna)}\), \(\text{(Liz,Renna), (Renna, Renna)}\}\)

Valuation

flowchart LR
accTitle: Flowchart for QL language
accDescr: A flowchart showing that the QL language consists of Symbols, Syntax, and Semantics. It further shows that Semantics consists of a model, which is an interpretation and domain, and a valuation function that determines Truth and Falsity. 
  subgraph QL
    Q["QL Language"]
    Sym["Symbols"]
    Syn["Syntax"]
    Sem["Semantics"]
    Q --> Sym & Syn & Sem
    end
    subgraph Model
    D[Domain]
    I[Interpretation]
    end
    val[Valuation]
    T["T"]
    F["F"]
    Sem --> Model
    Model --> val
    val --> T
    val --> F

  • We have explained a QL model.
  • Our valuation function (\(v\)) takes closed wffs and assigns truth values (T or F).

NoteValuation function

Let \(\alpha_1, \ldots, \alpha_n\) be any series of names (not necessarily distinct), \(P\) any \(n\)-place predicate, and \(\phi, \psi\) be wffs in QL. Relative to a model \(\mathcal{M}\), a valuation function (\(v\) or \(V\)) of a wff in QL is a function that assigns one and only one truth value (T or F) to each closed wff in QL in such a way:

  1. if \(P\alpha_1 \ldots \alpha_n\) is an atomic wff in RL, then \(v_{\mathcal{M}}(P\alpha_1\ldots \alpha_n)=T\) iff \(\langle\mathscr{I}(\alpha_1), \ldots , \mathscr{I}(\alpha_n)\rangle\) \(\in\mathscr{I}(P)\), otherwise \(v_{\mathcal{M}}(P\alpha_1\ldots \alpha_n)=F\)
  2. \(v_{\mathcal{M}}(\neg (\phi))=T\) iff \(v_{\mathcal{M}}(\phi)=F\)
  3. \(v_{\mathcal{M}}(\phi\wedge\psi)=T\) iff \(v_{\mathcal{M}}(\phi)=T\) and \(v_{\mathcal{M}}(\psi)=T\)
  4. \(v_{\mathcal{M}}(\phi\vee\psi)=T\) iff \(v_{\mathcal{M}}(\phi)=T\) or \(v_{\mathcal{M}}(\psi)=T\)
  5. \(v_{\mathcal{M}}(\phi\rightarrow\psi)=T\) iff \(v_{\mathcal{M}}(\phi)=F\) or \(v_{\mathcal{M}}(\psi)=T\)
  6. \(v_{\mathcal{M}}(\phi\leftrightarrow\psi)=T\) iff \(v_{\mathcal{M}}(\phi)=v_{\mathcal{M}}(\psi)\).
  7. \(v_{\mathcal{M}}(\exists x)\phi=T\) iff \(v_{\mathcal{M}}\phi(\alpha/x)=T\) for at least one name \(a\) in QL.
  8. \(v_{\mathcal{M}}(\forall x)\phi=T\) iff \(v_{\mathcal{M}}\phi(\alpha/x)=T\) for every name \(\alpha\) in QL.

Definition condensed

NoteValuation function

Let \(\alpha_1, \ldots, \alpha_n\) be any series of names (not necessarily distinct), \(P\) any \(n\)-place predicate, and \(\phi, \psi\) be wffs in QL. Relative to a model \(\mathcal{M}\), a valuation function (\(v\) or \(V\)) of a wff in QL is a function that assigns one and only one truth value (T or F) to each closed wff in QL in such a way:

  1. if \(P\alpha_1 \ldots \alpha_n\) is an atomic wff in RL, then \(v_{\mathcal{M}}(P\alpha_1\ldots \alpha_n)=T\) iff \(\langle\mathscr{I}(\alpha_1), \ldots , \mathscr{I}(\alpha_n)\rangle\) \(\in\mathscr{I}(P)\), otherwise \(v_{\mathcal{M}}(P\alpha_1\ldots \alpha_n)=F\)
  2. PL wffs are determined in the normal way
  3. \(v_{\mathcal{M}}(\exists x)\phi=T\) iff \(v_{\mathcal{M}}\phi(\alpha/x)=T\) for at least one name \(a\) in QL.
  4. \(v_{\mathcal{M}}(\forall x)\phi=T\) iff \(v_{\mathcal{M}}\phi(\alpha/x)=T\) for every name \(\alpha\) in QL.

Let’s explain each new part with examples.

Valuation of atomic wffs

if \(P\alpha_1 \ldots \alpha_n\) is an atomic wff in RL, then \(v_{\mathcal{M}}(P\alpha_1\ldots \alpha_n)=T\) iff \(\langle\mathscr{I}(\alpha_1), \ldots , \mathscr{I}(\alpha_n)\rangle\) \(\in\mathscr{I}(P)\), otherwise \(v_{\mathcal{M}}(P\alpha_1\ldots \alpha_n)=F\)

Example 1

  • Model: \(\mathcal{D} = \{1, 2, 3, 4\}\), \(\mathscr{I}(E)=\{2, 4\}\), \(\mathscr{I}(a)=1\), \(\mathscr{I}(b)=2\)
  • \(v(Ea)=T\) iff \(\mathscr{I}(a)\in\mathscr{I}(E)\)

Now let’s see if the wff is true:

  • \(1\not\in \{2, 4\}\)
  • \(v(Ea)=F\)

Example 2

  • Model: \(\mathcal{D} = \{1, 2, 3, 4\}\), \(\mathscr{I}(E)=\{2, 4\}\), \(\mathscr{I}(a)=1\), \(\mathscr{I}(b)=2\)
  • \(v(Eb)=T\) iff \(\mathscr{I}(b)\in\mathscr{I}(E)\)

Now let’s see if the wff is true:

  • \(2\in \{2, 4\}\)
  • \(v(Eb)=T\)

Example 3: Two-place

  • \(\mathcal{D} = \{1, 2, 3\}\), \(\mathscr{I}(G)=\{(3,1), (3,2), (2,1)\}\), \(\mathscr{I}(a)=1\),\(\mathscr{I}(b)=2\), \(\mathscr{I}(c)=3\)
  • \(v(Gab)=T\) iff \((\mathscr{I}(a), \mathscr{I}(b)) \in\mathscr{I}(G)\)

Now let’s see if the wff is true:

  • \((1, 2)\not\in \{(3,1), (3,2), (2,1)\}\)
  • \(v(Gab)=F\)

Quantified wffs

  1. \(v_{\mathcal{M}}(\exists x)\phi=T\) iff \(v_{\mathcal{M}}\phi(\alpha/x)=T\) for at least one name \(\alpha\) in QL.
  2. \(v_{\mathcal{M}}(\forall x)\phi=T\) iff \(v_{\mathcal{M}}\phi(\alpha/x)=T\) for every name \(\alpha\) in QL.

Look at this part:

  • \(\phi(\alpha/x)\)
  • This says replace the variable \(x\) in \(\phi\) with a name \(\alpha\) in \(\phi\)

Example 1

  • Model: \(\mathcal{D} = \{1, 2, 3, 4\}\), \(\mathscr{I}(E)=\{2, 4\}\), \(\mathscr{I}(a)=1\), \(\mathscr{I}(b)=2\)
  • \(v(\exists x)Ex=T\) iff \(v(Ex(\alpha/x))=T\) for at least one name \(\alpha\) in QL.

Now let’s see if the wff is true:

  • \(v(Ea)=F\) because \(\mathscr{I}(a)\not\in \mathscr{I}(E)\)
  • \(v(Eb)=T\) because \(\mathscr{I}(b)\in \mathscr{I}(E)\)
  • So, \(v(\exists x)Ex=T\)

Example 2

  • Model: \(\mathcal{D} = \{1, 2, 3, 4\}\), \(\mathscr{I}(E)=\{2, 4\}\), \(\mathscr{I}(a)=1\), \(\mathscr{I}(b)=2\), \(\mathscr{I}(c)=3\), \(\mathscr{I}(d)=4\)
  • \(v(\forall x)Ex=T\) iff \(v(Ex(\alpha/x))=T\) for every name \(\alpha\) in QL.

Now let’s see if the wff is true:

  • \(v(Ea)=F\) because \(\mathscr{I}(a)\not\in \mathscr{I}(E)\)
  • So, \(v(\forall x)Ex=F\)

Example 3

  • \(\mathcal{D} = \{1, 2, 3\}\), \(\mathscr{I}(G)=\{(3,1), (3,2), (2,1)\}\), \(\mathscr{I}(a)=1\),\(\mathscr{I}(b)=2\), \(\mathscr{I}(c)=3\)
  • \(v(\forall x)Gbx=T\)?

Check to see if all of these are T:

  • \(v(Gba)\)
  • \(v(Gbb)\)
  • \(v(Gbc)\)
  • \(v(Gbd)\)

Exercise

Exercise 6.76, 1-3, 9, 10

Two simplifications

Our approach to semantics involved two simplifications

  1. Valuation determines truth values of closed wffs (not open wffs)
  2. We assume there is a name for every item in the domain

Exercise

Exercise 6.77. 1, 4

Translation

Translation

graph LR
accTitle: Flowchart for QL language
accDescr: A flowchart showing that the QL language consists of Symbols, Syntax, and Semantics
  subgraph QL
    Q["QL Language"]
    Sym["Symbols"]
    Syn["Syntax"]
    Sem["Semantics"]
    Q --> Sym & Syn & Sem
end
  subgraph English
    E[English Sentences]
  end
QL <--> English

Goals

  1. Be able to translate simple English sentences into QL wffs, and vice versa
  2. Identify certain ambiguities that make translation difficult

Translation Key

A translation key consists of the following:

  1. a domain of discourse,
  2. An assignment of English names to QL names
  3. an assignment of English predicates and relations to QL n-place predicates

Key Example

  1. \(\mathcal{D}\) : Living human beings
  2. a: Annie
  3. j: Jon
  4. f: Frank
  5. Txy: x is taller than y.
  6. Hx : x is happy.

Translating without quantifiers

Our approach is as follows:

  1. Translate wffs without quantifiers.
  2. Translate wffs with quantifiers.
  3. Translate wffs problematic sentences.

Example 1

  1. QL: \(Ha\)
  2. English: Annie is happy.

Notice:

  • In QL, the structure is “Predicate, Name”
  • In English, the structrue is “Name, Predicate”

Tip

It can be helpful to write out the predicate and then plug in the names

  1. \(Ha\)
  2. Predicate: \(x\) is happy
  3. Plug in name: Annie is happy.

Example 2

Key: \(\mathcal{D}\): Living human beings, a: Annie, j: Jon, f: Frank, Txy: x is taller than y, Hx : x is happy.

  1. QL: \(Hj\)
  2. English: Jon is happy.

Example 3

Key: \(\mathcal{D}\): Living human beings, a: Annie, j: Jon, f: Frank, Txy: x is taller than y, Hx : x is happy.

  1. QL: \(Taj\)
  2. English: Annie is taller than Jon.
  3. QL: \(Tja\)
  4. English: Jon is taller than Annie.

Example 4

Key: \(\mathcal{D}\): Living human beings, a: Annie, j: Jon, f: Frank, Txy: x is taller than y, Hx : x is happy.

  1. QL: \(Taa\)
  2. English: Annie is taller than Annie.
  3. Alt: Annie is taller than herself.

Example 5

Key: \(\mathcal{D}\): Living human beings, a: Annie, j: Jon, f: Frank, Txy: x is taller than y, Hx : x is happy.

  1. QL: \(\lnot Taj\to Ha\)
  2. English: If Annie is not taller than Jon, then Annie is happy.

Translate: \(\lnot, \land, \lor, \to, \leftrightarrow\) as usual.

Exercise

Exercise 6.78. 1-4

Translating with quantifiers

Translating with quantifiers is slightly more difficult.

  • Translate \((\exists)\) using “At least one” or “Some” or “There exists”
  • Translate \((\forall)\) using “Every” or “All”

Example 1

Key: \(\mathcal{D}\) : human beings, Hx: x is happy, Zx: x is a zombie, Mx: x is mortal

  • \((\forall x)Hx\)
  • Translate predicate: \((\forall x)\text{ x is happy}.\)
  • Translate quantifier: \(\text{For every x, x is happy}.\)
  • Sentence: Everyone is happy.

Example 2

Key: \(\mathcal{D}\) : human beings, Hx: x is happy, Zx: x is a zombie, Mx: x is mortal

  • \((\exists x)Hx\)
  • Translate predicate: \((\exists x)\text{ x is happy}.\)
  • Translate quantifier: \(\text{For at least one x, x is happy}.\)
  • Sentence: Someone is happy. There is a happy human. Some happy human exists.

The universal quantifier

Let’s look at some common English expressions that utilize the universal quantifier.

Sentence QL
All As are Bs.
All Bs are As.
Everything is an A and a B.
No As are Bs.

All As are Bs

  1. All As are Bs.
  2. For every \(x\), if \(x\) is an \(A\), then \(x\) is a \(B\).
  3. \((\forall x)(Ax\to Bx)\)

All bunnies are mammals

  1. All bunnies are mammals.
  2. For every object \(x\), if \(x\) is a bunny, then \(x\) is also a mammal.
  3. \((\forall x)(Bx\to Mx)\)

All mammals are bunnies

We looked at “All As are Bs”. What about “All Bs are As”?

  1. All mammals are bunnies.
  2. For every object \(x\), if \(x\) is a mammal, then \(x\) is a bunny.
  3. \((\forall x)(Mx\to Bx)\)

Clarification

There is a classic example that illustrates that “All As are Bs” is not equivalent to “All Bs are As.”

  1. All squares are rectangles. True.
  2. All rectangles are squares. False.

I used to work in a gas station:

  1. Everyone who is kind is also nice to cashiers. True.
  2. Everyone who is nice to cashiers is also kind. False (scammers).

Everything is an A and a B.

“All As are Bs” and “All Bs are As” say something different than “Everything is both an A and a B.”

  1. Everyone is both a smoker and a runner.
  2. For every object \(x\), \(x\) is both a smoker and a runner
  3. \((\forall x)(Mx\land Bx)\)

Notice that \(\land\) is used instead of \(\to\)

Clarification 1

  1. All smokers are runners. Everyone who smokes also runs.
  • Consistent with people who run that don’t smoke.
  1. All runners are smokers. Everyone who runs also smokes.
  • Consistent with many smokers who don’t run.
  1. Everyone is both a runner and a smoker.
  • Everyone is smoking and running.

No As are Bs

If we say “no As are Bs”, we are not saying that

  1. nothing is an A.
  2. nothing is a B.

We are just saying that

  1. if anything is an A, then it is also not a B, OR
  2. there is nothing that is in the intersection of A and B

No runners are smokers.

  1. No runners are smokers.
  2. For every object \(x\), if \(x\) is a runner, then \(x\) is not a smoker.
  3. \((\forall x)(Rx\to \lnot Sx)\)

Clarification

All of these are equivalent:

  1. \((\forall x)(Rx\to \lnot Sx)\) - No runners are smokers.
  2. \((\forall x)(Sx\to \lnot Rx)\) - No smokers are runners.
  3. \((\forall x)\lnot (Sx\land Rx)\) - No one is both a smoker and a runner.

Summary

Sentence QL
All As are Bs. \((\forall x)(Ax\to Bx)\)
All Bs are As. \((\forall x)(Bx\to Ax)\)
Everything is an A and a B. \((\forall x)(Ax\land Bx)\)
No As are Bs. \((\forall x)(Ax\to \lnot Bx)\)

Exercise

Exercise 6.79. 1-4

The existential quantifier

Let’s consider some sentences that are translated with the existential quantifier.

Sentence QL
Some As are Bs.
Some Bs are As.
There exists an A and a B.
Some As are not Bs.
Some Bs are not As.
No As are Bs.

Some As are Bs.

When we say say “Some As are Bs” we are saying there is at least one object that is both an A and a B.

  1. Some runners are smokers.
  2. For at least one object \(x\), \(x\) is a runner and \(x\) is a smoker.
  3. \((\exists x)(Rx\land Sx)\)

Notice the use of \(\land\)

Some Bs are As.

“Some As are Bs” and “Some Bs are As” are equivalent. We are saying that there is at least one thing that has both properties (or falls into both categories)

  1. Some smokers are runners.
  2. For at least one object \(x\), \(x\) is a smoker and \(x\) is a runner.
  3. \((\exists x)(Sx\land Rx)\)

There exists an A and a B.

“There exists an A and a B” is slightly ambiguous:

  1. There exists an A and there exists a B.
  2. There exists something that is both an A and a B.

These two sentences say different things.

Clarification

  1. There exists a nice man and a man who wears a hat.
  2. There exists a nice man who also wears a hat.

Sentence (1) is true in the following scenario:

  • Tek is a nice man.
  • Jon wears a hat.

Sentence (2) is true iff

  • There is at least one man who is both nice and wears a hat.

Example

  • There exists a nice man and a man who wears a hat.
    • Translation: \((\exists x)Nx\land (\exists x) Hx\)
  • There exists a nice man who also wears a hat.
  • Translation: \((\exists x)(Nx\land Hx)\)

Notice the two quantifiers and the placement of parentheses.

Making it precise

If you think your meaning might be misconstrued, you can be more explicit:

  • “There exists an A and a B” (Ambiguous)
  • “There exists an A and there exists a B” (Not Ambiguous)
  • “There exists something that is both an A and B” (Not Ambiguous)

Some As are not Bs

  1. Some rectangles are not squares
  2. \((\exists x)(Rx\land \lnot Sx)\)

Some Bs are not As

  1. Some squares are not rectangles.
  2. \((\exists x)(Sx\land \lnot Rx)\)

Notice that these two setences are not equivalent:

  1. “Some rectangles are not squares”. True
  2. “Some squares are not rectangles”. False

Examples

To create examples, consider cases where every kind of object A is a B, but not every B is an A.

Bunnies and mammals.

  1. “Some mammals are not bunnies.” True.
  2. “Some bunnies are not mammanls.” False.

Humans and students.

  1. “Some humans are not students.” True.
  2. “Some students are not human.” False.

No As are Bs.

We saw that you can translate “No As are Bs” with the universal quantifier

\((\forall x)(Ax\to \lnot Bx)\)

But you can also translate it with the existential quantifier:

\(\lnot (\exists x)(Ax\land Bx)\)

Making it precise

Sentence QL
Some As are Bs. \((\exists x)(Ax\land Bx)\)
Some Bs are As. \((\exists x)(Bx\land Ax)\)
There exists an A and a B. Ambiguous
Some As are not Bs. \((\exists x)(Ax\land \lnot Bx)\)
Some Bs are not As. \((\exists x)(Bx\land \lnot Ax)\)
No As are Bs. \(\lnot (\exists x)(Ax\land Bx)\)

Exercise

Exercise 6.80. 7-10

Ambiguity

Summary

graph LR
accTitle: Flowchart for QL language
accDescr: A flowchart showing that the QL language consists of Symbols, Syntax, and Semantics
  subgraph QL
    Q["QL Language"]
    Sym["Symbols"]
    Syn["Syntax"]
    Sem["Semantics"]
    Q --> Sym & Syn & Sem
end
  subgraph English
    E[English Sentences]
  end
QL <--> English

Revealing ambiguity

  • Translating English into QL reveals ambiguities in English.
  • We can tease out these ambiguities by translating them into their distinct wffs.
  • We can also rephrase our speech so the ambiguity is not present.

Revealing ambiguity

NoteDefinition

A word is ambiguous iff that word has distinct meanings.

graph LR
accTitle: Flowchart for Ambiguity
accDescr: A flowchart showing that the word bank can mean Slope of land by river or a financial institution
  Q["Bank"]
  F["Financial institution"]
  S["Slope of land by water"]
  Q --> F & S

Ambiguity in arguments

“Star” is ambiguous in the following argument:

  • P1. All stars are in outer space.
  • P2. Socrates is a star.
  • C. Therefore, Socrates is in outer space.

In P1, “star” is a celestial body, but in P2, “star” is a celebrity.

Problems with ambiguity

  • P1. All stars are in outer space.
  • P2. Socrates is a star.
  • C. Therefore, Socrates is in outer space.

In evaluating the argument,

  • C follows only if “star” has the same meaning, but
  • P1 and P2 are true only if “star” has distinct meanings.

Exploiting ambiguity

People exploit ambiguity by trying to have it both ways.

  • P1. All stars are in outer space.
  • P2. Socrates is a star.
  • C. Therefore, Socrates is in outer space.

In evaluating the argument,

  • To get validity, they insist “star” has a single meaning
  • To get truth, they insist “star” has distinct meanings.
  • If you can get both, the argument is sound.

Nuanced examples

  • No one is going to be duped by the “star” example (or care)
  • But there are plenty of harder examples to detect in more important contexts.
  • Let’s look at them super quickly.

Example 1: Safety

  • P1. If a medication has been approved by the FDA, it is safe.
  • P2. This medication has been approved by the FDA.
  • P3. A safe medication is a medication that poses no significant danger.
  • C. Therefore, it is safe.
  1. “safe” (regulatory safety) - benefits outweigh risks under specific conditions of use.
  2. “safe” (absolute safety) - no significant danger

Example 2: Evidence

  • P1. There is no evidence that my client committed the crime.
  • P2. If there is no evidence that my client committed the crime, then my client is innocent.
  • C. Therefore, my client is innocent.
  1. Admissible evidence - evidence that can be submitted to court (for P1)
  2. Evidence as proof - evidence that establishes whether my client did the crime

Example 3: Obesity epidemic

  • P1. In the early 1990s, approximately 61.7 million Americans were overweight
  • P2. In the late 1990s, approximately 97.1 million Americans were overweight
  • C. Therefore, approximately 35.4 million additional Americans became overweight during this period. Epidemic!

Change is not due to the population but to change in the definition of BMI threshold used to classify individuals who are “overweight”

Quantificational ambiguity

NoteDefinition

A quantificational ambiguity is an ambiguity that arises when a sentence is ambiguous about the presence or scope of a quantifier.

  1. Presence ambiguities
  2. Scope ambiguities

Presence ambiguity

A quantificational presence ambiguity occurs when a sentence is ambiguous about the quantity of items \(x\) from the domain that has property \(F\).

“Birds fly.”

  1. Some birds fly. \((\exists x)(Bx\land Fx)\)
  2. All birds fly. \((\forall x)(Bx\land Fx)\)

Example in argument

  • P1. Birds fly.
  • P2. This penguin is a bird.
  • C. Therefore, this penguin flies.

In this argument:

  1. P1 is T only if it is “Some birds fly”
  2. Valid only if it is “All birds fly”.

Bait and switch

  • P1. Men think they are smarter than women.
  • IC. Therefore, men are chauvinistic.
  • C. Therefore, all of the men from our office should do sensitivity training.

In this argument,

  • The person wants P1 to be “all” because (1) provocative and (2) supports C, BUT
  • P1 is not true (“not all men”).

Continued

  • P1. Men think they are smarter than women.
  • IC. Therefore, men are chauvinistic.
  • C. Therefore, all of the men from our office should do sensitivity training.

So, rhetorically,

  1. They will talk as though P1 means “all”, but
  2. If you challenge them on P1, they will claim it means “some”.

Scope ambiguity

A quantificational scope ambiguity occurs when an ambiguity arises because there is an ambiguity concerning the scope of the quantifer.

  1. Overlapping quantifiers
  2. Universally quantified wffs with negations

Example

Everyone loves someone.

  1. \((\forall x)(\exists y)Lxy\)
  2. \((\exists x)(\forall y)Lyx\)

Notice that in

  • in (1) the universal quantifier has wide scope
  • in (2) the existential quantifier has wide scope

What’s the difference?

Everyone loves someone.

  1. \((\forall x)(\exists y)Lxy\)
  2. \((\exists x)(\forall y)Lyx\)

Notice that in

  • Every person loves at least one person (could be different person – your crush)
  • There is someone (e.g., Santa Claus) that every single person loves.

Default reading

flowchart TD
    S["Everyone loves someone."]
    S --> S1["$$(\forall x) (\exists y) Lxy$$"]
    S --> S2["$$(\exists x) (\forall y) Lyx$$"]
    S1 --> S1a["Every x loves some y"]
    S2 --> S2a["There is some x whom every y loves"]

  • Without context, I always interpret the wff as “Every x loves some y.”
  • With context: Is there a some one person that everyone loves? Yes, everyone loves the Dalai Lama. So, everyone does love someone.

All Ss are not P.

This translation is straightforward:

  1. Every student passed.
  2. \((\forall x)(Sx\to Px)\)

Continued

This translation is not unambiguous:

Every student did not pass.

Two readings

  1. It is not the case that every student passed (at least one student did not pass). \(\lnot (\forall x)(Sx\to Px)\)
  2. No student passed. \((\forall x)(Sx\to \lnot Px)\)

Notice the scope of the negation

Reading 1

Scenario 1: Students are comparing their grades.

  1. I got an A!
  2. I got a B.
  3. I got a C.

Conversation:

  • Student: Did everyone pass?
  • Professor: “Every student did not pass.”
  • \(\lnot (\forall x)(Sx\to Px)\) or \((\exists x)(Sx\land \lnot Px)\)

Reading 2

Scenario 2: Professor says exam average is 20 percent.

Conversation:

  • Student: Did everyone pass?
  • Professor: “Every student did not pass.”
  • \((\forall x)(Sx\to \lnot Px)\) or \(\lnot (\exists)(Sx\land Px)\)

Exercise

Exercise 6.82. 1-2.