Basic Syllogistic Logics

Larry Moss, Indiana University

ESSLLI 2010, Copenhagen

1/64

Logic and Language: Traditional Syllogisms

All men are mortal.

Socrates is a man.
Socrates is mortal.
Some men are mortal.
Socrates is a man.
Socrates is mortal.
All frogs are reptiles.
All reptiles are animals.
All frogs are animals.

2/64

More examples

All frogs are reptiles.

All frogs are animals.
All reptiles are animals.
All sagatricians are maltnomans.
All sagatricians are aikims.
All maltnomans are aikims.
The first point is that there is an exact definition of validity for
arguments.
The second point here is that the form is as important, even more
important, than the particular words.

3/64

More examples

All
All
All
All

X are Y .
Y are Z .
W are X .
W are Z .

So valid arguments can have more than two premises.

Our plan is to start with sentences containing all, some, and no.

4/64

Syntax and Semantics

Probably the key point of logic is that there is a distinction between
syntax

and

semantics.

The idea is that syntax is the raw symbols.

Syntax is usually painful: think of computer programming.
Ill try to avoid it as much as possible to concentrate on the ideas.
The semantics is where we get the meaning.
So in our examples, we need some context or model to give a
meaning.
In our examples, the syntax will start with some variables p, q, n,
n1 , . . ..
The our sentences are expressions of the form
All p are q, Some p are q, and No p are q

5/64

Semantics
To say whether
All sagatricians (s) are maltnomans (m).
is true or not needs a model.
This is given by a few things:
First, a set M called the universe.
Second, for the words sagatrician and maltnoman,
we need sets [[sagatrician]] M and [[maltnomanan]] M.
Definition
For the language of All,
a model M is a set M together with sets [[p]] M
for all nouns p.

6/64

Syntax and Semantics

Syntax: We begin with atoms (nouns) p, q, X , Y , Z , etc.
The sentences in our first fragment are the expressions All p are q
We usually use letters like for sentences.
Semantics: A model M is a set M,
and for each variable p we have an interpretation [[p]] M.
M |= All p are q

iff

[[p]] [[q]]

M |= is read as M satisfies .
A statement like M |= All p are q could also be read as
All p are q is true in M

7/64

Review

Is All X are Y true or not?

8/64

Review

Is All X are Y true or not?

Without a model (also called a context), the question makes no
sense.
So lets take an example model, and ask whether our sentence is
true in that model or not.
Let M = {1, 2, 3, 4, 5}
Let [[X]] = {1, 2, 4}.
Let [[Y]] = {3, 4}.

8/64

Review
Is All X are Y true or not?
Without a model (also called a context), the question makes no
sense.
So lets take an example model, and ask whether our sentence is
true in that model or not.
Let M = {1, 2, 3, 4, 5}
Let [[X]] = {1, 2, 4}.
Let [[Y]] = {3, 4}.
In this model, All X are Y is false!
But if we change the model by re-setting [[Y]] to {1, 2, 3, 4}, then
our sentence is true.

8/64

A Quirk
One fine point on the definition is that if [[X]] is the empty set ,
then our sentence All X are Y is true!
So in this room now,
All people in the room over 7 feet tall are standing
is (on this definition) true.
This strange point will lead us to various issues over the next days.
The standard reply is to say that its true because there are no
exceptions.
But we again admit that the semantics of All that we are giving
is not what most people would agree to in cases where [[X]] = .

9/64

Validity of Arguments
At this point, we know how to give the semantics of single
sentences.
follows from 1 , . . . , n
This means that every model that makes all of the s true also
makes true.
We write this as
1 , . . . , n |=
and we also say that the s semantically imply .
To argue that 1 , . . . , n |= we need some reasoning.
Usually, we do this in English and in an informal way,
just as one would do ordinary mathematical reasoning.
But to argue that 1 , . . . , n 6|= we can produce a
counterexample.
The main thing is that we have a rigorous definition, using a
semantic notion (models).

10/64

A small note on notation

We use letters like (Greek letter Gamma) for sets of sentences.
|=
This means that every model of all the sentences in
is also a model of .
Whenever is a set that we have listed out, say
= {1 , 2 , . . . , 104 }.
then usually we would write |= as
1 , 2 , . . . , 104 |=
rather than as
{1 , 2 , . . . , 104 } |= .
That is, we drop the set braces on the left of the |= symbol.
We do this to make things a little more readable.

11/64

Validity: the idea

premises
z
}|
{
1 , 2 , . . . , n

|=

conclusion
z}|{

The intuition is that

1 , 2 , . . . , n |=
means that
any circumstance in which the premises 1 , 2 , . . . , n are all true
is also a circumstance in which the conclusion is true

12/64

Examples

All frogs are reptiles.

All frogs are animals.
All reptiles are animals.
We can take M = {1, 2, 3, 4, 5, 6}.
[[F ]] = {1, 2},
[[R]] = {1, 2, 3, 4}.
[[A]] = {1, 2, 4, 5, 6}.
In this context, the assumptions are true but the conculsion is false.
So the argument is invalid.

All frogs are reptiles, All frogs are animals, 6|= All reptiles are animals.

13/64

Some points

Note the difference between syntax and semantics.

|= is intended to mean follows by general-purpose reasoning.
We can check whether our definitions match with our intuitions.
In the case of our very simple fragment, this mostly is right.
The main exception is that people usually wouldnt say
All X are Y
in a context where they know that there are no X .

14/64

Some points
A secondary point is that a computer should be able to decide
whether
1 , . . . , n |=
or not.
The entailment problem should be decidable.

Another way to make these points:

the definitions and theory should be tight enough so that the
decision can be made without semantics (!),
by only looking at the form of the argument.

15/64

Proof Trees and `

Definition
Let be a set of sentences {1 , . . . , n }.
A proof tree over is a tree labeled with sentences,
and with the following property:
Every node is either labeled with a sentence from ,
or matches one of the rules of our system (see the next slide).
We draw proof trees with the root at the bottom and the leaves at
the top.
`
This means that there is a proof tree over whose root is labeled
,
We say that is provable from in our system.

16/64

The rules for building trees

All p are p

All p are n All n are q

All p are q

17/64

Example
Here is an example: Let be the set
{All A are B, All Q are A, All B are D, All C are D, All A are Q}
Let be All Q are D. Here is a proof tree showing that ` :
All A are B All B are D
All Q are A
All A are D
All Q are D
All of the leaves belong to .
Note also that some elements of are not used as leaves.
This is permitted according to our definition.
The proof tree above shows that ` .

18/64

Another proof tree for the same assertion

`

All A are B All B are B

All A are B
All B are D
All Q are A
All A are D
All Q are D
One of the leaves is justified not because it belongs to ,
but because it matches the reflexivity rule.
All p are p

All p are n All n are q

All p are q

19/64

What are we doing here?

The idea is that proof trees are our model of basic reasoning
using the words all, some, no.
A proof tree is like a caricature of a real proof.
It can be examined (and even constructed) by a person or
computer
who has no understanding of anything but the rules!
There are several hopes about this work:
? The whole thing will scale up to include many more words.
(This would call on linguistic semantics to provide the correct
notion of context.)
? The formal relation ` should have something to do with |=
(logic)
? The proof system ` should have something to do with actual
human reasoning (psychology)
? A computer should be able to work with ` without
understanding anything.

20/64

Soundness
A computer could check whether a purported tree actually satisfies
our definition, even if it didnt understand All.
So one important question is: what is the relation between
` A and

|= A ?

Soundness Lemma
If ` , then |= .
This means that proof trees do not lead us astray:
if ` , then in any context where the sentences of all hold,
too must hold.
Our proof system will not lead us to believe that bogus syllogisms
are in fact valid.

21/64

Soundness

Here is the basic idea of why the Soundness Lemma holds.

The two most basic facts about are:
1

X X for all sets X .

For all sets X , Y , and Z : if X Y and Y Z , then X Z .

(Probably the third would be that X for all X .)

22/64

Soundness Sketch, Continued

Lets go back to our example proof tree.
All A are B All B are D
All Q are A
All A are D
All Q are D
Take any model, say M.
Assume that in M, [[A]] [[B]], etc.
We have to show that in this same model M, [[Q]] [[D]].
The idea is to use our proof tree and specialize it to M:
[[A]] [[B]]
[[B]] [[D]]
[[A]] [[D]]
[[Q]] [[A]]
[[Q]] [[D]]
And then going downward mirrors intuitively valid reasoning in the model.
Since the model M was arbitrary (had no special features),
the conclusion |= holds.

23/64

A question?

All
All
All
All
All
All
All
All
All

A are B,
A are C,
B are C,
C are B,
C are D,
B are E,
D are G,
F are G,
G are F

We see that ` All B are G .

Do you think that ` All D are E ?

Is there an algorithm to tell yes or no?

24/64

Preorders

Definition
A preorder is a pair (P, ),
where P is a set
and is a relation on it with the following properties:
reflexive p p
transitive If p q and q r , then p r .
We need not have the following property:
anti-symmetric if p q and q p, then p = q.
An anti-symmetric preorder is a partially ordered set (poset).

25/64

A picture of a preorder

All
All
All
All
All
All
All
All
All

A are B,
A are C,
B are C,
C are B,
C are D,
B are E,
D are G,
F are G,
G are F

F,G
D?

??
??
??







B, C
A
The set P here is {A, . . . , G }.
The order is given by
X Y

iff

` All X are Y .

26/64

Downsets in preorders
In a preorder, p = {x : x p}.

F = G = {A, B, C , D, F , G }

F,G
D?

??
??
??






B, C
A

D = {A, B, C , D}, E = {A, B, C , E }

B = {A, B, C } = C
A = {A}

is monotone: if p q, then p q.

27/64

Proof of completeness
Suppose that |= All X are Y.
Let M be the set of variables.
(Yes, the model is built from the syntax!)
Define A B to mean that ` All A are B.
Check that this is reflexive and transitive, using the logic.
The semantics is via downsets:
[[A]]

{B : B A}

By transitivity, M |= .
In more detail, suppose contains All C are D.
Then if W C , then also W D.

28/64

Proof of completeness
Suppose that |= All X are Y.
Let M be the set of variables.
(Yes, the model is built from the syntax!)
Define A B to mean that ` All A are B.
Check that this is reflexive and transitive, using the logic.
The semantics is via downsets:
[[A]]

{B : B A}

By transitivity, M |= .
In more detail, suppose contains All C are D.
Then if W C , then also W D.

28/64

An example of how the proof works

All
All
All
All
All
All
All
All
All

A are B
A are C
B are C
C are B
C are D
B are E
D are G
F are G
G are F

[[A]] = {A}
[[B]] = {A, B, C }
[[C ]] = {A, B, C }
[[D]] = {A, B, C , D}
[[E ]] = {A, B, C , E }
[[F ]] = {A, B, C , D, F , G }
[[G ]] = {A, B, C , D, F , G }

Getting back to the question of whether or not

` All D are E .
Since [[D]] is not a subset of [[E ]], 6|= All D are E .
By soundness, 6` All D are E .

29/64

Syllogistic Logic of All and Some

Syntax: All p are q, Some p are q
Semantics: A model M is a set M,
and for each noun p we have an interpretation [[p]] M.
M |= All p are q
M |= Some p are q

iff
iff

[[p]] [[q]]
[[p]] [[q]] 6=

Proof system:
All p are p
Some p are q
Some q are p

All p are n All n are q

All p are q

Some p are q
Some p are p

All q are n Some p are q

Some p are n

30/64

Example
If there is an n, and if all n are p and also q, then some p are q.

Some n are n, All n are p, All n are q ` Some p are q.

The proof tree is
All n are p Some n are n
Some n are p
All n are q
Some p are n
Some p are q

31/64

Completeness
We show that if |= Some p are q, then also ` Some p are q

Suppose that |= Some p are q.

We shall construct a particular model M of
(hence of Some p are q),
and then read off a proof Some p are q.
We take for M the set of pairs
{x, y }
such that ` Some x are y.
Note that if {x, y } M, then also {x} M.
We declare
{x, y } [[u]]

iff

x u or y u.

This defines a model M.

32/64

Completeness, continued
{x, y } [[u]] iff x u or y u

M |=
This is fairly routine.

Thus M |= Some p are q

Suppose that {x, y } [[p]] [[q]].
At this point, we can prove the completeness of the system.
Again, we assume that |= Some p are q.
and we show that ` Some p are q.
(What about |= All p are q?)

33/64

End of the proof of completeness for the

language of All and Some
{x, y } [[u]] iff x u or y u

We know that there is some {x, y } [[p]] [[q]].

There are four possibilities:
1

x p and y q

x p and x q

y p and x q

y p and y q

Lets just go through the first case: x p and y q

The proof tree below shows that ` Some p are q:
..
..
..
..
..
All
y
are
q
Some
x
are y
..
All x are p
Some x are q
Some p are q

34/64

Where we are
All p are p
Some p are q
Some q are p

All p are n All n are q

All p are q

Some p are q
Some p are p

All q are n Some p are q

Some p are n

We now know that this is complete for the language of All and
Some.

35/64

Where we are
All p are p
Some p are q
Some q are p

All p are n All n are q

All p are q

Some p are q
Some p are p

All q are n Some p are q

Some p are n

We now know that this is complete for the language of All and
Some.
We can also add names to the fragment, interpreted as points in
M.
J is J

M is J
J is M

J is M M is F
J is F

J is an X J is a Y
Some X are Y

All X are Y J is an X
J is a Y

M is an X J is M
J is an X

35/64

The languages S and S add noun-level

negation
Let us add complemented atoms p on top of
the language of All and Some,
with interpretation via set complement: [[p]] = M \ [[p]].
We always have p = p.
So we have

All p are q

Some p are q
All p are q No p are q

Some p are q Some p arent q

Some non-p are non-q

36/64

A complicated semantic fact

Let be
{All B 0 are X , All X are Y , All Y are B, All B are X , All Y are C }.
We claim that |= All A are C .

37/64

A complicated semantic fact

Let be
{All B 0 are X , All X are Y , All Y are B, All B are X , All Y are C }.
We claim that |= All A are C .
Here is the reasoning, done informally.
Since all B and all B 0 are X , everything whatsoever is an X .
And since all X Y B, we see that everything is a B.
But also B X Y C .
In particular, all A are C .
But the last two premises and the fact that all X are Y
also imply that all B are C .
So all A are C .

37/64

The logical system for S

All p are p

Some p are q
Some p are p

All p are n All n are q

All p are q

All q are q
Zero
All q are p
All p are q
Antitone
All q are p

Some p are q
Some q are p
All n are p Some n are q
Some p are q

All q are q
One
All p are q
Some p are p
Ex falso quodlibet

38/64

An example of what can be done in this logic

We saw this yesterday as one of our goal examples

All xenophobics are yodelers.
All zookeepers are non-yoders.
All zookeepers are non-xenophobics.

39/64

Another exercise for us

Show that
{All B are X , All B 0 are X } ` All A are X .

40/64

A formal proof tree in our S

Let be
{All B 0 are X , All X are Y , All Y are B, All B are X , All Y are C }.
Here is a derivation showing that ` All A are C .
All X are Y All Y are B
All B 0 are X
All X are B
All X are Y All Y are C
All B 0 are B
All B are X
All X are C
All A are B
All B are C
All A are C

41/64

Facts

Some X are X 0 ` S (a contradiction fact)

All X are Z , No Z are Y ` No Y are X (Celarent)

No X are Y ` No Y are X (E-conversion)

Some X are Y , No Y are Z ` Some X are Z 0 (Ferio)

All Y are Z , All Y are Z 0 ` No Y are Y (complement

inconsistency)

42/64

A fine point on the logic

The system uses

Some p are p
Ex falso quodlibet

and this is prima facie weaker than reductio ad absurdum.

(Reductio will be discussed next time in detail.
It is closer to what Aristotle used, and for more expressive
fragments, it is needed.)

43/64

I wont go through the completeness in this

lecture

All p are p

Some p are q
Some p are p

All p are n All n are q

All p are q

All q are q
Zero
All q are p
All p are q
Antitone
All q are p

Some p are q
Some q are p
All n are p Some n are q
Some p are q

All q are q
One
All p are q
Some p are p
Ex falso quodlibet

You can read it in the rest of this slide set, and in the course notes.

44/64

Completeness via representation of

orthoposets
Definition
An orthoposet is a tuple (P, , 0, 0 ) such that
poset is a reflexive, transitive, and antisymmetric
relation on the set P.
zero 0 p for all p P.
antitone If x y , then y 0 x 0 .
involutive x 00 = x.
inconsistency If x y and x y 0 , then x = 0.
A Key Point
Orthoposets need not have a meet or join operation.

45/64

Orthoposets: two examples

Example
For all sets X we have an orthoposet (P(X ), , , 0 ), where
a0 = X \ a for all subsets a of X .
Example
p1
ppp
p
p
ppp
ppp
p
p
0
p NpN
NNN p >>
NNN >>
NNN >
NNN>>
N

N<NN
<< NNN
<< NN
<< NNN
NNN
<
q
q0
pp

p
p

 ppp
pppp

ppp

(x 0 )0 = x, 00 = 1, 10 = 0.

46/64

Orthoposets: two examples

Example
For all sets X we have an orthoposet (P(X ), , , 0 ), where
a0 = X \ a for all subsets a of X .
Example
p1
ppp
ppp
p
p
pp
ppp
0
p NpN
NNN p >>
NNN >>
NNN >
NNN>>
N

N<NN
<< NNN
<< NN
<< NNN
NNN
<
q
q0

ppp
p

p
 p
 pp
 ppp
ppp

(x 0 )0 = x, 00 = 1, 10 = 0.
The idea
boolean algebra
propositional logic

orthoposet
logic of All, Some and

The details concerning completeness are somewhat different,

and the whole thing would take about 10 minutes.

46/64

Orthoposets from the logic

Let be any set of sentences in the fragment.
Let V be the set of variables.
We already know the preorder :
X Y

iff

` All X are Y.

(so Some plays no role)

We have an induced equivalence relation ,
and we take V to be the quotient V/.
If there is some X such that X X 0 , then set 0 to be [X ].
We finally define [X ]0 = [X 0 ].
If there is no X such that X X 0 , we add fresh elements 0 and 1
to V/.
It is not hard to check that we have an orthoposet V .

47/64

Orthoposets from logic, concretely

Let =
{All B are A, All B 0 are A, All C 0 are B, All C are B 0 , All C are D}.
Then
[A] = {A}
[A0 ] = {A0 }
0
[B] = {B, C } [B 0 ] = {B 0 , C }
[C ] = {B 0 , C } [C 0 ] = {B, C 0 }
[D] = {D}
[D 0 ] = {D 0 }
Here is a picture of the orthoposet V :
s
sss
s
s
s

[A] K

KKK
KKK
K

[C 0 ] = [B]

[D]

[D 0 ] K

KK
KK
KK

[B 0 ] = [C ]

[A0 ]

s
sss
sss

48/64

Points of orthoposets

A point of a orthoposet P = (P, , 0, 0 ) is a subset S P with the

following properties:
up-closed If p S and p q, then q S.
complete For all p, either p S or p 0 S.
pairwise compatible For all p, q S, p 6 q 0 .

49/64

Points are sets

Look back at the Chinese lantern.
There are four points here: the sets marked , , , and :
N;NN
;; NNN
; NNN

1N
pp
p
;;N;NNNN
p
p


; NNN
ppp

p pNN p 0
q p q0
NNN ==
p


NNN==  ppppp
N pp

1
pp
p

p
p


ppp

p pNN p 0
NNN ==
NNN==
N

1N
pp
p
;;N;NNNN
p
p


; NNN
ppp

p pNN p 0
q p q0
NNN ==
p

NNN== ppppp

N pp

1N
pp
p
;;N;NNNN
p
p


; NNN
ppp

p NpN p 0
q p q0
NNN ==
p

NNN== ppppp

N pp

q
pp
pppp

ppp

50/64

What are the points?

s
sss
s
s
s

[A] K

KKK
KKK
K

[C 0 ] = [B]

[D]

[D 0 ] K

KK
KK
KK

[B 0 ] = [C ]

[A0 ]

s
sss
sss

51/64

What are the points?

There are three points.

s
sss
s
s
s

[A] K

KKK
KKK
K

[C 0 ] = [B]

[D]

[D 0 ] K

KK
KK
KK

[B 0 ] = [C ]

[A0 ]

s
sss
s
s
s

S = {[D 0 ], [B], [A]}, T = {[B 0 ], [D], [A]}, U = {[B], [D], [A]}.

51/64

Points need not be filters

Let X = {1, 2, 3}, and let P(X ) be the power set orthoposet.
Then S is a point, where
S

{{1, 2}, {1, 3}, {2, 3}, {1, 2, 3}}.

It is easy to check that the points on this P(X ) are exactly S as

above and the three principal ultrafilters.
S shows that a point of a boolean algebra need not be a filter.

52/64

The Extension Lemma for pairwise consistent

sets
Lemma
Let S P be pairwise consistent: (p, q S)p 6 q 0 .
Then for all x P, either S {x} or S {x 0 } is again pairwise
consistent.
Proof.
Suppose not. Then x and x 0 figure in to problems both times.
There is some p S such that p x 0 .
There is some q S such that q x 00 = x.
And now: q x p 0 . Ooops!

53/64

The Extension Lemma for pairwise consistent

sets

Lemma
Let S P be pairwise consistent: (p, q S)p 6 q 0 .
Then for all x P, either S {x} or S {x 0 } is again pairwise
consistent.
Lemma
If p 6 q, then {p, q 0 } is pairwise consistent.
Thus there is a point S containing p but not q.

53/64

Pairwise consistent sets extend to points

Lemma
For a subset S0 of an orthoposet P = (P, , 0, 0 ),
the following are equivalent:
1

S0 is a subset of a point S in P.

S0 is pairwise compatible.

Proof.
Clearly (1) = (2).
For the more important direction, use Zorns Lemma to get S S0
which is pairwise compatible, and maximal with this property.
For all p, either p or p 0 belongs to S. [By maximality.]
Check easily that S is up-closed:
If p S, p q, but q
/ S, then q 0 S.
And now p q = (q 0 )0 , so S is not pairwise compatible.

54/64

Representation Theorem
the point of points
Let P = (P, , 0 ) be an orthoposet.
Let points(P) be the set of points of P.
We have an orthoposet
(P(points(P)), , , 0 )
Let m : P P(points(P)) be given by
m(p)

{S : p S}.

Theorem
m is a strict morphism of orthoposets:
m(0) = ,
m(p 0 ) = (m(p))0 ,
and p q iff m(p) m(q).

55/64

Representation Theorem
the point of points
Let P = (P, , 0 ) be an orthoposet.
Let points(P) be the set of points of P.
We have an orthoposet
(P(points(P)), , , 0 )
Let m : P P(points(P)) be given by
m(p)

{S : p S}.

Theorem
m is a strict morphism of orthoposets:
m(0) = ,
m(p 0 ) = (m(p))0 ,
and p q iff m(p) m(q).
Corollary
Every orthoposet is isomorphic to a sub-orthoposet of a power set
orthoposet.

55/64

How the representation works

q 1 9MMM
M
qq0q
q
p MMp
q q q0
MMM qq
q

q 1 9MMM
M
qq0q
q
p MMp
q q q0
MMM qq
q

q 1 9MMM
M
qq0q
q
p MMp
q q q0
MMM qq
q

q 1 9MMM
M
qq0q
q
p MMp
q q q0
MMM qq
q

{, , , }

jjjj ss
jjjj sssss
j
j
j
jjj
sss
jjjj
{, } UU
{, }
UUUU
L
UUUU LLLL
UUUU LL
UUUU LL
UUULUL

KKKTTTT
KKK TTTTT
TTTT
KKK
TTTT
K
T
{, }
i {, }
iiii
rr
i
i
r
i
r iiii
rr
ririiiiii
r
ririi

56/64

Another one
S = {[D 0 ], [B], [A]}, T = {[B 0 ], [D], [A]}, U = {[B], [D], [A]}.
rrr
rrr

[A] LL

LLL
LL

[C 0 ] = [B]

[D]

[D 0 ] L

LLL
LLL

[B 0 ] = [C ]

[A0 ]

rrr
rrr

m([A]) = {S, T , U} m([A0 ]) =

m([B]) = {S, U}
m([B 0 ]) = {T }
m([D]) = {T , U}
m([D 0 ]) = {S}
This pretty much solves our earlier problem of getting a model of
where [[B]] 6 [[D]].
But how?

57/64

Sources the Representation Theorem

N. Zierler and M. Schlessinger

Boolean embeddings of orthomodular sets and quantum logic.
Duke Mathematical Journal 32 (1965), 251262.
F. Katrnoska
On the representation of orthocomplemented posets.
Comment. Math. Univ. Carolinae 23 (1982), 489498.
C. S. Calude, P. H. Hertling, K. Svozil
Embedding quantum universes into classical ones.
Foundations of Physics, 29, 3 (1999), 349-379.

58/64

The canonical model

Lemma
Let be consistent in L(all, some, 0 ).
There is a canonical model M = (M, [[ ]]) such that
1

M |= .

If M |= All X are Y, then ` All X are Y.

Proof.
Let V be the syntactic orthoposet for . Let M = points(V ).
The interpretation [[ ]] : V P(M) is given by
V

/ V

/ P(points(V )) = P(M)

Key point If contains Some U are V, need a point including

{[U], [V]}.
If none exists, then wlog U V 0 . But then is inconsistent.

59/64

Half of completeness

Lemma
Let be consistent in `.
There is a canonical model M = (M, [[ ]]) such that
1

M |= .

If M |= All X are Y, then ` All X are Y.

This gives half of completeness:

If |= All X are Y ,
then M |= All X are Y ,
and so ` All X are Y .
For Some sentences, we need a little more.

60/64

If is consistent and |= Some X are Y, then

` Some X are Y

Lemma (Ian Pratt-Hartmann 2007)

There is some existential sentence in , say Some A are B, such
that
all {Some A are B} |= Some X are Y .

61/64

If is consistent and |= Some X are Y, then

` Some X are Y
Fix A and B as in the lemma.
Consider the model M = M(Vall ) of points on Vall . M |= all .
Consider {[A], [B], [X 0 ]}.
If this set were a subset of a point S, then consider {S} as a
one-point submodel of M.
In the submodel, all {Some A are B} would hold,
and yet Some X are Y would fail, since [[X ]] = .
Therefore {[A], [B], [X 0 ]} is not pairwise compatible.
There are six cases:
A A0 A B 0
A X B B0
B X X0 X
Only two are significant.

61/64

If is consistent and |= Some X are Y, then

` Some X are Y
Next, consider {A, B, Y 0 }.
The same analysis gives two other cases: A Y and B Y .
Putting these together with the other two gives four pairs.
The case when A X and B Y is representative:
..
..
All A are X Some B are A
..
..
Some B are X
All B are Y
Some X are B
Some X are Y
The other cases are similar. This completes the proof.

61/64

Beyond first-order logic: cardinality

Note: we only consider finite models

Read (X , Y ) as there are at least as many X s as Y s.

All Y are X
(X , Y )

(X , Y )

(Y , Z )

(X , Z )

All Y are X (Y , X )
All X are Y
Some Y are Y (X , Y )
Some X are X

No Y are Y
(X , Y )

The point here is that by working with a weak basic system,

we can go beyond the expressive power of first-order logic.

62/64

Picture
ng

ri
-Tu
rch
u
Ch

FOL

modal
FO 2

FOmon

monadic FOL
2 variable fragment

no

a
Pe
reg
-F
e
S

adds full N-negation

We have discussed these

S
S

63/64

Intersective adjectives
[[red x ]] = [[x ]] [[red]]

(n, n)

(red x, x)
(n, p)
(I)
(n, n)

(n, p) (p, q)
(B)
(n, q)

(T)

(Adj1 )

(n, red x) (n, y )

(Adj2 )
(n, red y )

(n, q) (q, p)
(D)
(p, n)

(x, red y )
(Adj3 )
(red x, red y )

(red x, blue y ) (red x, green z)

(Adj4 )
(red x, blue z)
(red x, blue y ) (red x, green z)
(Adj5 )
(blue x, green y )

64/64