6.

Deductive Arguments:

Categorical Logic

Review: Deductive Relationships

ƒ You’ll recall: Statement A deductively entails

statement B if and only if it is impossible for B
to be false given that A is true.

ƒ Logical Impossibility: A state of affairs that

cannot exist.

E.g., If x > 1, x cannot also be < 1

If Dion is a bachelor, then Dion cannot also be

married (since bachelor just means ‘unmarried
male’)

1
ƒ Recall also that what we are ultimately after in
this course is to be able to evaluate
arguments. (e.g., in terms of ARG conditions)

Deductive entailment is important to

evaluation in that, if an argument is
deductively valid, then it must satisfy R.

Moreover, if its premises also satisfy A (i.e., if

its premises are all true), then it satisfies G
and is said to be sound.

1. All pacifists are opponents of war

2. No opponents of war are arms traders
Therefore,
3. No pacifists are arms traders

Deductively valid, by virtue of its categorical

form (i.e., formally valid)

We can see more clearly what is meant by

formal validity by substituting new values for
subject and predicate in each premise …

2
1. All alephs are bleeps
2. No bleeps are cheeps
Therefore,
3. No alephs are cheeps

We can see intuitively that the argument form

is valid, even if we have no idea what an
aleph or bleep or cheep is. Similarly …

1. All A are B
2. No B are C
Therefore,
3. No A are C

3
Categorical Logic

ƒ The logic of “class membership” or

membership in categories.

Logic which uses “all,””no,””some,””not” and

so forth as its basic logical terms.

ƒ Categorical form: A subject category is

connected to a predicate category …

“All consistent vegetarians are opponents of

using animals for leather.”

A categorical statement, in which “consistent

vegetarian” is the subject and “opponents of
using animals for leather” is the predicate.

4
Four Categorical Forms

Universal Affirmation (A – Affirmo)

All S are P
(All members of the category S are included
within category P. – “All bachelors are males”)

Universal Negation (E – nEgo)

No S are P
(All members of category S are excluded from
category P. – “No bachelors are married
persons.”)

Particular Affirmation (I - affIrmo)

Some S are P
(Some members of category S are included in
category P – i.e., there is at least one S which
is P)

Particular Negation (O – negO)

Some S are not P
(Some members of category S are excluded
from category P – i.e., there is at least one S
which is not P)

10

5
The Square of Opposition

A: All S are P E: No S are P

ry
to
dic
ntra
Co Co
nt
ra
di
ct
or
y

I: Some S are P O: Some S are not P

11

Contradictories

ƒ In the square of opposition, each form of

proposition is the contradictory of the one
diagonally opposed to it.

When one statement is the contradictory of

another, the two must always opposite truth
values.

I.e., if it is true that “No C are O” then it must

false that “Some C are O”; if it is false that “All P
are Q” then it must be true that “Some P are Q”

12

6
 A: All S are P Contrary
E: No S are P

I: Some S are P O: Some S are not P

13

Contrariety

ƒ A and E statements are contraries.

Contrary statements cannot both be true, but

they can both be false.

E.g., “All dogs are black” and “No dogs are

black” cannot both be true at the same time,
but they could both be false. (And, in this
case, both statements are in fact false.)

14

7
 A: All S are P E: No S are P

Subcontrary

I: Some S are P O: Some S are not P

15

Subcontraries

ƒ I and O statements are said to be

subcontraries.

Subcontrary statements cannot both be false,

(on the ‘existential import’ assumption in
categorical logic) but they can both be true.

E.g., “some dogs are black” and “some dogs

are not black” cannot both be false, but both
can be true. (In this case, both are true.)

16

8
Distribution

ƒ A term (i.e., a subject or predicate) is said to

be distributed if the statement that contains
that term says something (makes some
assertion about) about every item falling
under that term.

ƒ A and E statements are fully universal; their

subject term is always distributed …

17

For example:

”All bachelors are males” (A)

Says something about all bachelors (namely,

that they are male), but it does not say
anything about all males.

The subject term of an A statement is

distributed, but its predicate term is not
distributed.

18

9
“No married men are bachelors” (E)

Says something about all married men

(namely, that none of them are bachelors) and
it also says something about all bachelors
(namely, that none of them are married men).

Both the subject and the predicate terms of an

E statement are distributed.

19

“Some men are bachelors” (I)

Some men ≠ all men, so the statement does

not say anything about all members of its
subject category. It also does not anything
about all bachelors (it says only the category
men and the category bachelor overlap)

So, in an I statement neither the subject nor

the predicate term is distributed.

20

10
 “Some dogs are not black” (O)

Some dogs ≠ all dogs, so the statement does

not say anything about all dogs. But this
statement does say something about the
entire category of black things (namely, that
there is at least one thing – a non-black dog –
that is excluded from that category).

So, in an O statement, the subject term is not

distributed, but the predicate term is
distributed.

21

To Recap …

A: Subject is D; Predicate is ~D

E: Subject is D; Predicate is D (both)

I: Subject is ~D; Predicate is ~D (neither)

O: Subject is ~D; Predicate is D

22

11
Translating Natural Language
Statements into Categorical Form

ƒ In the section on Natural Language and

Categorical Form (pp. 210-5), Govier offers
some advice about translating natural
language statements into categorical form.

Translation is a useful thing to do, because,

once an argument is put into categorical form,
we can make use of rules of inference that
make evaluating arguments much easier. …

23

… but, alas, not many statements in “real

world” natural language arguments are found
in perfect categorical form.

However, by paying attention to the principles

of translation, we can sometimes put
subject/predicate type statements into
categorical form …

24

12
Universal Affirmative (A)

Any S is P
Every single S is P
The Ss are all Ps
Whatever S you look at, it is bound to be a P
Each S is a P
If it’s an S, it’s a P … etc.

All of these can rightly be translated in into an

A statement of the form “All S are P”

25

Quantity of the Statement

ƒ Often, it is important to figure out if the

statement is intended to be universal or
particular:

“Nurses are retiring early due to stress and ill

health” – All of them or only some?

“Men feel uncomfortable discussing their

emotions.” – All of them or only some?

26

13
Stereotyping

ƒ While some arguers may indeed intend to

make universal claim with such statements,
we should be cautious.

The danger of attributing a universal claim

when only a particular claim is intended can
be called stereotyping.

For obvious reasons, stereotyping is not only

logically inaccurate, but ethically and
politically objectionable as well.

27

“Only”

ƒ The word “only” does typically make an

implicitly universal claim:

“Only members who have paid their dues in

full will be allowed to attend the next meeting
of the Aerosmith Fan Club.”

Only D are A = All A are D

Notice that we cannot conclude from this that

All D are A (some members might stay home,
e.g.)

28

14
Universal Negative (E)

Pigs can’t fly

No pig can fly
Pigs are not able to fly
There has never been a pig that was able to
fly.
If it’s a pig, then it can’t fly

All of these are universal negative statements

that can be translated into categorical form
as:
No pigs are creatures that can fly (No P are F)

29

Some slightly tricky cases

“Not all lawyers are rich people”

“All racial minorities are not black people”

The first looks like an E statement, the second

looks like an A statement; but both are
correctly translated as O statements:

“Some lawyers are not rich people”

“Some racial minorities are not black people”

30

15
Particular Affirmative (I)

To repeat: In categorical logic, “some” is

understood to mean “at least one” – no
distinction is made between one, several,
many or most.

ƒ So, a particular affirmative (I) statement picks

out “greater than none, less than all”

31

ƒ Particular affirmations too can create some

tricky situations:

“Some lawyers do not take bribes”

It seems natural to infer from this that some

lawyers then do take bribes, but that is not
necessarily what the arguer has asserted.

A) “Some L are not B” and B) “Some L are B”

cannot both be false, but the truth of A) does
not guarantee the truth of B) (subcontrariety).

32

16
Particular Negative (O)

ƒ The expression “not all” occurring before the

subject of a categorical statement is usually a
denial of a universal affirmation.

“Not all famous plays are written by

Shakespeare”

= there is at least one famous play not written

by Shakespeare.

Accordingly, “Some P are not S”

33

Venn Diagrams

ƒ Venn diagrams are named

after the English mathematician
John Venn (1834-1923).

ƒ Venn diagrams can be used to

show the meaning of categorical
statements, to describe rules
of immediate inference and to
work with categorical syllogisms

34

17
Areas of the Venn Diagram

S P

1 2 3

Things that are Things that are Things that are

S and not P both S and P P and not S

35

An A Statement: “All S are P”

No S are outside of P; i.e., the domain of things

that are S and not P is empty, blacked out
36

18
An E Statement: “No S are P”

No S are inside of P; i.e., the overlap between S

and P is empty, the categories have no
members in common. 37

An I Statement: “Some S are P”

S P

1 2 3
X

There is at least one thing that is both S and P;

that thing is represented by the X.
38

19
An O Statement: “Some S are not P”

S P

1 2 3

There is at least one thing that is S but is not P;

that thing is represented by the X.
39

Rules of Immediate Inference

ƒ If you have two categorical statements and

the first deductively entails the second, you
can immediately (i.e., without any
intermediate steps), infer the second from the
first.

ƒ Immediate inference, in the present context,

refers to operations involving A, E, I, and O
statements that yield valid inferences.

40

20
Conversion

ƒ The converse of a statement in categorical form is

obtained by transposing the subject and predicate
terms. So, …

Statement Converse
A: All S are P All P are S
E: No S are P No P are S
I: Some S are P Some P are S
O: Some S are not P Some P are not S

41

Logical Equivalence

ƒ E and I statements are logically equivalent to

their converse; i.e., they logically entail each
other and they are either both true or both
false. It is impossible for one to be false given
that the other is true.

E.g.,

If “no women are men,” then “no men are

women.”

If “some lawyers are not crooked,” then “some

not crooked things are lawyers”
42

21
‘E’ Conversion Venn Diagram

No S are P No P are S

43

‘I’ Conversion Venn Diagram

Some S are P Some P are S

44

22
ƒ So, to repeat, E and I statements are both
logically equivalent to their converses;
because they logically equivalent, they must
have the same truth value.

ƒ This does not hold for the conversion or A and

O statements, however, as can easily be seen
by inspecting the relevant Venn diagrams …

45

‘A’ Conversion Venn Diagram

All S are P All P are S

46

23
‘O’ Conversion Venn Diagram

Some S are not P Some P are not S

47

ƒ So, the repeat A statements and O statements

are NOT logically equivalent to their
converses.

As Govier notes, this is actually a fairly

common error in reasoning. E.g.:

Just because “All Muslims are religious

believers” we certainly cannot infer that “All
religious believers are Muslims”

48

24
Contraposition

To contrapose a statement in categorical form, one

first converts and then adds a “non” to each category.
So, …

Statement Contraposition

A: All S are P All non-P are non-S

E: No S are P No non-P are non-S

I: Some S are P Some non-P are non-S

O: Some S are not P Some non-P are not non-S

49

Contraposition & Inference

ƒ A and O statements are logically equivalent to

their contrapositives.

ƒ E and I statements are not logically equivalent

to their contrapositives.

50

25
Contraposition and A Statements

ƒ Given the truth of an A statement, its

contrapositive must also be true:

All ants are insects (All S are P)

All non-insects are non-ants (All non-P are non-S)

If the truth of A statement did not guarantee the

truth of its contrapositive, there would be a
counterexample to the A statement, a non-P that
was S: a non-insect that was an ant.

51

Contraposition and O Statements

“Some insects are not wasps”

“Some non-wasps are not non-insects”

Just as with integers in mathematics, the

double negation (“not non-”) cancels out.

To say that “some non-wasps are insects” is

clearly to say something logically equivalent to
“some insects are not wasps”

52

26
Contraposition and E Statements

E: No S are P

Contrapositive: No non-P are non-S

The first statement that categories S and P do

not intersect at all (their intersection is empty)

The second says that whatever is outside of

the P category does not intersect with
whatever is outside of the S category …

53

E: No cats are dogs

Contrapositive: No non-dogs are non-cats

The first happens to be true (in our world

there is nothing that is both a cat and a dog);
but the second is obviously false, since all
sorts of things are both non-dogs and non-
cats.

54

27
Contraposition and I Statements

I: Some S are P
Contrapositive: Some non-P are non-S

The first asserts that there is at least one

thing that is in both S and P; the
contrapositive asserts that there is at least
one thing outside of both categories.

Some numbers are prime numbers

(true a priori)
Some non-prime numbers are non-numbers
(false a priori)

55

Obversion

The obverse of a categorical statement is formed by first

adding a “non-” to the predicate category, then changing
the quality (affirmative or negative) of the statement.

Statement Obversion

A: All S are P No S are non-P

E: No S are P All S are non-P

I: Some S are P Some S are not non-P

O: Some S are not P Some S non-P

56

28
Obversion and Inference

ƒ The obverse of any categorical statement

yields a logically equivalent statement.

All guitarists are musicians

No guitarists are non-musicians

57

Contradictories

ƒ As you may have noticed, when there are

differences among rules of immediate
inference, A and O statements ‘go together’
and E and I statements go together

ƒ To see why this is the case, think back to the

Square of Opposition …

58

29
The Square of Opposition

A: All S are P E: No S are P

ry
to
dic
ntra
Co Co
nt
ra
di
ct
or
y

I: Some S are P O: Some S are not P

59

ƒ A and O are contradictories

ƒ E and I are contradictories

Which means, you’ll the recall, that the truth

of one guarantees the falsehood of the other
and vice versa.

If it is true that “All cows are ungulates,” it

must be false that “Some cows are not
ungulates”

60

30
Counterexample

ƒ This provides with another way of looking at

the idea of counterexample:

We can show definitively that the statement

“All philosophy Ph.D.s are unemployed” is
false if we can point to even a single example
of a person with a philosophy Ph.D. who has a
job.

61

False Dichotomies

1) “All refugees are non-happy persons”

It may tempting to translate this as

2) “All refugees are unhappy persons”

But these are non logically equivalent. 1)

says only some refugees are outside of the
category of happy persons; 2) says that
refugees belong within the class of unhappy
persons

62

31
ƒ “Non-happy” is the complementary predicate
of “happy”

As basic truth of logic, everything in the

universe is either happy or non-happy (T a
priori)

But it is not the case that everything in the

universe must be either happy or unhappy.

(Some things, e.g., simply be mildly bored;

for other things, the notions might not be
applicable at all, e.g., rocks, trees, oil slicks …)
63

Contrary Predicates

ƒ “happy” and “unhappy” are instead said to be

contrary predicates.

Something cannot be both happy and

unhappy, but it need not be either

64

32
 Complementary Contrary
Happy, nonhappy Happy, unhappy
Divine, nondivine Divine, satanic
Strong, nonstrong Strong, weak
Democratic, nondemocratic Democratic, undemocratic

Health, nonhealthy Healthy, unhealthy

Fat, nonfat Fat, thin
Good, nongood Good, evil

65

Fallacy of False Dichotomy

ƒ Mistaking contraries for logically exhaustive

complementaries

ƒ Also called “polarized thinking” or “binary

thinking”

Consider: “You are either with us or you are

against us.”

66

33
Categorical Logic:
Some Philosophical Background

ƒ First developed by Aristotle in

the 3rd century BCE.

ƒ (False) Assumption: That all statements can

be put into subject/predicate form.
Consider:
“Either Martin will win or Harper will win”
“If Harper wins, our country is doomed.”
These clearly have truth values, but they are
not subject/predicate statements.
67

Individuals

ƒ Also, categorical logic cannot easily deal with

individuals:

“George W. Bush has two daughters.”

ƒ We could get around this in CL, by creating a

class that contains only one individual:

“All persons identical with George W. Bush are

persons that have two daughters.” …

68

34
ƒ In propositional logic, however, this sort of
move can be avoided by symbolizing logical
individuals directly.

(B . D)

69

Existential Import

ƒ Classical (or Aristotelian) CL assumes that we

only make assertions about things that are
real (i.e., that exist).

For Aristotle, the idea of speaking about non-

existent things was irrational.

So, if we assert in traditional categorical logic:

“All humans are mortal” (A)

This carries with it the assumption that human

beings do in fact exist.
70

35
Hypothetical (Boolean) Interpretation

ƒ The existential assumption may seem

harmless enough in everyday contexts, but
consider:

“All black holes are invisible”

A scientist who makes this assertion may not

know for certain that black holes exist.
Instead, she claims only that if black holes
exist, they will be invisible. This latter claim
has a truth value (i.e., it can be true or false)
whether or not black holes exist.

71

ƒ So, in modern (non-Aristotelian) logic,

universal statements are understood as
hypothetical statements:

“All humans are mortal” ->

“If anything is a human being, then that thing

is mortal”

On the hypothetical interpretation, the

existence of the subject term is not assumed.

72

36
 Similarly,

“(All) unicorns have a single horn”

On a hypothetical interpretation, this

statement has a truth value even though no
unicorns exist (we think).

On the existential interpretation it, weirdly,

commits us to the claim that they do exist.

73

Particulars

ƒ In both modern and Aristotelian logic,

particular statements are understood to assert
existence.

ƒ BUT: On a hypothetical interpretation of, say,

an A statement, we cannot validly infer an I
statement from it.

“All mermaids have no legs” (A)

“Some mermaids have no legs” (I)

74

37
ƒ That is, on the modern view, we cannot validly
infer actual existence from hypothetical
existence.

ƒ This may seem strange in some practical

contexts. Consider Govier’s example …

75

(a) “All lawyers are rich people”

(b) “Some lawyers are rich people”

It seems natural to think that if all lawyers are

rich, then some of them are.

76

38
 Categorical Interpretation

Hypothetical Interpretation

77

How to Deal With This Difference

ƒ In this course, our main task is to gain skill in

interpreting and evaluating “real world”
natural language arguments.

So, in that context, we need not be ‘purists’

about either the modern or the Aristotelian
views.

ƒ In interpreting (and translating arguments),

Govier advises us to pay attention to the
context of the argument and, if necessary, to
‘read in’ the existential assumption if that best
makes sense of the argument.
78

39
The Categorical Syllogism

ƒ An argument with two premises and a

conclusion, all of which are statements in
categorical form.

In a valid CS, each of the categories is (must

be) mentioned in two different statements.
E.g.,

All men are mortal

Socrates is a man
Therefore,
Socrates is mortal
79

Govier’s example

1. All consistent vegetarians are opposed to

using animals for leather.
2. No opponents of using animals for leather are
fur trappers
Therefore,
3. No consistent vegetarians are fur trappers.

80

40
Formalized …

1. All C are O
2. No O are T
Therefore
3. No C are T

T = major term (predicate in conclusion)

C = minor term (subject in conclusion)
O = middle term (appears in both premises but
not in the conclusion)

81

Categorical Syllogisms and Venn Diagrams

ƒ Venn diagrams can be used both i) to

represent categorical syllogisms and ii) to
determine their deductive validity.

It’s almost magical:

Once we have successfully represented the

premises in the diagram, we can simply look
to see if the conclusion is also represented …

82

41
 S P

1 3 5

4
2 6

7
8

M
83

area 1: S, not P, not M

area 2: S, not P, M
area 3: S, P, not M
area 4: S, P, M
area 5: not S, P, not M
area 6: not S, P, M
area 7: not S, not P, M
area 8: not S, not P, not M

84

42
Diagramming Categorical Syllogisms

1. Identify and represent the major (S), minor

(P) and middle (M) terms with circles
corresponding to each category.

2. Represent universal premises

3. Represent particular premises

(This last step can be tricky)

85

1
Some examples …

1. All C are O
2. No O are T
Therefore,
3. No C are T

86

43
 2
1. All whales are mammals
2. All mammals are warm-blooded creatures
Therefore,
3. All whales are warm-blooded creatures

1. All W are M
2. All M are C
Therefore,
3. All W are C

87

3
1. All philosophers are liars
2. Some Greeks are philosophers
Therefore,
3. Some Greeks are liars

1. All P are L
2. Some G are P
3. Therefore,
4. Some G are L

88

44
 4
1. Some persons who pursue extreme sports are
snowboarders
2. Some snowboarders are persons who enjoy
taking risks
Therefore,
3. Some persons who pursue extreme sports are
persons who enjoy taking risks

1. Some E are S
2. Some S are R
3. Therefore,
4. Some E are R

89

Rules of the Categorical Syllogism

ƒ Besides using Venn diagrams, we can also

assess the deductive validity of categorical
syllogisms by applying the five rules of the
syllogism.

If none of these rules are violated (i.e.,

broken), then we can conclude that the
syllogism is a valid deductive argument.

First, a little review …

90

45
Distribution of Terms

ƒ You’ll recall that we said a term of a

categorical statement is distributed if that
statement asserts something about all
members of the category to which that term
refers.

A: Subject is D; Predicate is ~D
E: Subject is D; Predicate is D (both)
I: Subject is ~D; Predicate is ~D (neither)
O: Subject is ~D; Predicate is D

91

Recall:

”All whales are warm-blooded creatures” (A)

Says something about all whales (namely, that

they are warm-blooded), but it does not say
anything about all warm-blooded creatures.

The subject term of an A statement is

distributed, but its predicate term is not
distributed.

92

46
“No reptiles men are warm-blooded” (E)

Says something about all reptiles (namely,

that none of them are warm-blooded) and it
also says something about all warm-blooded
creatures (namely, that none of them are
reptiles).

Both the subject and the predicate terms of an

E statement are distributed.

93

“Some dogs are friendly” (I)

Some dogs ≠ all dogs, so the statement does

not say anything about all members of its
subject category. It also does not anything
about all friendly things (it says only the
category dogs and the category friendly things
overlap)

So, in an I statement neither the subject nor

the predicate term is distributed.

94

47
 “Some dogs are not friendly” (O)

Some dogs ≠ all dogs, so the statement does

not say anything about all dogs. But this
statement does say something about the
entire category of friendly things (namely, that
there is at least one thing – a non-friendly dog
– that is excluded from that category).

So, in an O statement, the subject term is not

distributed, but the predicate term is
distributed.

95

The Middle Term and the Fallacy of the

Undistributed Middle

ƒ The middle term, we just saw, is the term

which occurs in both premises, but not in the
conclusion, of a categorical syllogism.

ƒ For a categorical syllogism to be valid, its

middle term must be distributed in at least
one of its premises.

Consider …

96

48
1. No Bush supporters are pacifists
2. All pacifists are Democrats
Therefore,
3. No Bush supporters are Democrats

“Pacifists” is the middle term; it is distributed

in both premise 1 (E) and premise 2 (A).
Valid.

97

1. All logic students are persons who political

debate
2. Some Republicans are persons who enjoy
political debate
Therefore,
3. Some Republicans are logic students

“persons who enjoy political debate” is the

middle term; it is not distributed in either
premise 1 (A) or in premise 2 (I). Invalid

98

49
Rules of the Categorical Syllogism
For a syllogism to be valid,

1. Its middle term must be distributed in at least one

premise.

2. No term can be distributed in the conclusion unless it is

also distributed in at least one premise.

3. It must have at least one affirmative premise

4. If it has a negative conclusion, it must also have a

negative premise. Conversely, if it has one negative
premise, it must also have a negative conclusion.

5. If it has two universal premises, it cannot have a

particular conclusion. *
99

ƒ The fifth rule (If it has two universal premises,

it cannot have a particular conclusion), you’ll
note, follows from the hypothetical
interpretation of universal categorical
statements.

ƒ Apart from that qualification, all of the other

rules apply under either interpretation.

100

50
Govier’s (Possibly Too Cute) Example:

1. All invalid syllogisms contain an error

2. This syllogism contains an error
Therefore,
3. This syllogism is invalid

Notice that in this self-referential syllogism all

of the premises and the conclusion are true
(A is satisfied), but because it is invalid R
and, therefore, G are not satisfied.

101

Enthymemes

ƒ An incompletely expressed categorical

syllogism, one in which a premise or the
conclusion is unstated.

The bigger the burger, the better the burger

The burgers are bigger at Burger King

102

51
ƒ Enthymemes give rise to the problem (which
we have encountered several times before) of
when it is appropriate to supply a missing
conclusion or premise when translating,
formalizing or evaluating natural language
arguments.

1. All persons who love children are persons who would

make a good kindergarten teacher.
2. Sue loves children.
Therefore,
3. Sue would make a good kindergarten teacher.

103

Enthymemes & Sorites

ƒ Enthymemes are often found strung in a

pattern called a sorites (< Greek “heap”)

A typical pattern:

1. All A are B
2. All B are C
3. All C are D
4. All D are E
Therefore,
5. All A are F

104

52
ƒ Since a categorical syllogism, by definition,
has two premises, a sorites has between three
and n (i.e., there is no upward limit).

ƒ To see if a sorites is valid, we can break it up

into a succession of enthymemes …

105

1. All A are B
2. All B are C
So,
2x. All A are C (valid)

2x. All A are C

3. All C are D
So,
3x. All A are D (valid)

… and so on through to “All A are F.”

106

53