This excerpt is provided, in screen-viewable form, for personal use only by

members of MIT CogNet.

Unauthorized use or dissemination of this information is expressly forbidden.

If you have any questions about this material, please contact

cognetadmin@cognet.mit.edu
Chapter 3
Predicate Lo Rit

3.1 A Formal Language for Predicate Logic

Comment. Sentential logic allows us to study the

logical relations among sentencesthat hold because of
their sb"ucture, insofar as that sb"ucture is detennined
by the presence of connectives. But sentential logic
' '
cannot handle the similarity between Kareem is tall
' '
and ' Akeem is tall , not to mention Someone is
tall ' - these would be represented as P, Q , and R , as if
they had nothing in common . We now introduce a new
language that accommodates this further sb"ucture.

vocabulary . TheVOCABULARYOF PREDICATE

Definition
LOGICconsists of
. SENTENCE LEn ERS,
. CONNECTIVES ,
. NAMES,
. VARIABLES ,
. PREDICATE LEn ERS,
. themENTIT YSYMBOL,
. QUA N' CD' I KRS, and
. PARENTHESES .

Sentence letters, connectives, and parentheses are

adopted from the language of sentential logic .
 Chapter3

names Definition. A NAME is a symbol from the following

list:
a, b , c , d, . ...
ai ' bl ' CI ' dl ' ~ ' b2 '

variables Definition. A VARIABLE is a symbol from the

following list:
0 , V, W , X , y , z , V ' . . ..
1' I WI ' Xl ' YI ' ZI ' 2'

Comment. Names and variables are used to refer to

objects in much the same way as names and certain
kinds of pronouns in English . Section 3.2 deals with
translation between English and the language defined
in this section.

Comment. Where there is no possibility of confusion we shall

sometimesuse lowercaseletters other than those listed above as
names.

I -place . A I -PLACE PREDICATE

Definition LETTERis
predicate any symbolfrom the following list:
letter AI , ... , ZI , AJ , ... , ZJ, ... .

2-place A 2-PLACE
PREDICATE
LETr ER is any symbol
from
the list
following:
2
A,...,Z2 ,...,z5,....
,A5
n - place In general, an n -PLACE PREDICATE LETTER is
any symbolfrom the list
An, ... , Zn,~ , ... , ZO...
 3
Chapter

many-place Comment . Predicateletters with more than one place

are referred to as MANY -PLACE PREDICATE
L E' nE R S. Predicate letters will sometimes be
referredto as ' predicates'for short.

Comment. In practice the superscripts can and will be

omitted . Any of the capital letters may appear as
sentence letters or predicate letters. It is usually
possible to tell how a letter is being used in a wff by
looking at the number of names or variables immediately
following it . A capital letter with no names
or variables is a sentence letter , one followed by one
name or variable is a I -place predicate, and so on.
'
Also , the letters ' R ' and ' 5 are sometimes reserved for
2-place predicates.

identity .The
Definition '=' isthemEN'rI TYSYM
symbol -
symbol BOL.
Comment. The identity symbol is usedto representthe
relationship of numerical identity, such as, for
example, that Mark Twain is identical to (i.e., one and
the sameas) SamuelLanghorneClemens.

metavariables Comment . The Greek letters a , P, 1, etc. are used as

METAVARIABLES for the names and variables of
predicatelogic.
 Chapter3

" '
universal Definition. A UNIVERSAL QUANrD IER is any
quantifier symbol of the form
Va
wherea is a variable.

Comment . Universal quantifiers correspondto the

' '
English word every .

existential '' ~
Definition. An EXISTENTIAL QUAN I DIER is
quantifIer any symbol of the form
3a
wherea is a variable.

Comment .Existentialquantifiers correspond to the

Englishword ' some'.

expression Definition. An EXPRESSION OF PREDICATE

LOGIC is anysequence
of itemsfrom the vocabulary
of predicatelogic.

wffs Definition.A WELL -FORMEDFORMULA of predicate

logic is any expressionin accordance
with the
followingsevenrules:

(1) Sentenceletters are wtTs.

(2) An n -place predicate letter followedI by n namesis a

wff .
Chapter3

(3) Expressions of the forln a=Ji where a and Ji are

names are wffs .

Comment. Although the placement of the identity

symbol superficially resemblesthat of a connective, it
is in fact a special two- place predicate. For historical
reasonsalone it is placedbetweena. and Ii rather than
in front of them.

atomic [Definition . Wffs of the form specifiedin rules 1- 3

sentence are the ATOMIC SENTENCES of predicate logic.
Thoseconforming to rule 3 are also known as mENTITY
STATEMENTS.]

Comment. We adopt the practice of omitting superscripts

from predicateletters.

(4) Negations
, conjunctions, disjunctions, conditionals,
and biconditionals of wft"s are wft"s.

Comment . The formation rules of chapter 1 are

subsumedby this clause.

(5) If (j) is a wff , then the result of replacing at least one

occurrence of a name in (j) by a new variable a (i.e.,
a not in (j) and prefixing 'v'a is a wff .

universal wff [Definition. Such wffs are called UNIVERSALLY

'~ ~
QUAN I D lED wffs, or UNIVERSAL wffs.]
 Chapter3

If cj) is a wff , then the result of replacing at least one

occurrence of a name in cj) by a new variable a (i.e.,
a not in cj) and prefixing 3a is a wff .

existential wff [Definition. Suchwffs are calledEXISTENTIALLY

~
QUANTDIED wffs, or EXISTENTIAL wffs.]

(7) Nothingelseis awIf .

Examples.
Wffs of this languageinclude the following :
Fa v Fb) ~ Gab)
3yFy
Vx ( Fx ~ Gx)
VxVy (Rxy ~ Ryx)
(3xFx H VxGx)
- 3x(Fx H - VyGy)
(3xFx ~ P)
Vx3y Fyxb
- a=b
Vx x=x
VxVy (x=y ~ y=x)
 3
Chapter

Exercise3.1.1 Which of the following expressionsare wffs? If an expression

is a wff , say whetherit is an atomic sentence ,
a negation, a conditional, a conjunction, a disjunction,
a biconditional, a universal, or an existential. ( Note:
Any wff must fall into exactly one of thesecategories.)

Fz
..
11* VxGac
...*
111 VxGcax
iv * 3x' Vy(Gxy & Gyx)
v* ' Vx(Gxy H 3yHy)
vi * 3x(Ax ~ ' VxFxx)
vii * 'v'x'v'y(Fxy ~ 'v'z(Hxyz & Jz
viii * 'v'xFxx H 'v'x'v'yFxy
ix * - 'v'x- 3z(Hz v Jx)
x* Ga ~ 'v'x- (Ha v Fxx)
xi * p ~ Gab
xii * - (P & - 3xFx)
xiii * 'v'x( Fx) & P
xiv * 3y (Fyyy & P)
xv * 'v'xyz(Fzx H Hxyz)
xvi * b=b
xvii * (a=a)
xviii * P=c
xix * Fa=Fa
xx * Vz(Fz ~ a=b)
xxi * Vx (x=x)
xxii * 3x(Fx=Gx)
xxiii * - Vx (Fx & 3y x=y)
xxiv * (- a=b H - Vx (Fxa & Pbx
xxv * Vx3y (- x=y ~ y=- x)
 Chapter3

quantifier Comment. When a wff contains an uninterrupted sequence

convention of quantifiers of the same type , existential or
universal , it is often convenient to omit repetitions of 3
or 'v' .

Examples.
The expression
' Vxyz(Fxy & Gyz H Hzx)
will be readas shorthandfor
' Vx' Vy' Vz
( Fxy & Gyz H Hzx).

The expression
3xyVzw( Fxyz & Gwx - + - Hzx)
is to be readas
3x3y VzVw ( Fxyz & Gwx - + - Hzx) .

non-identity We introduce the special symbol ~ that may be used to

abbreviate statements of the fonD - a = (3 thus: a ~ (3.
It will be useful to bear in mind that sentences of this
fonD are negations, not atomic .

Comment. As with the parenthesis-dropping conventions

introduced in chapter 1, the formulas allowed by
the conventions here are not strictly well -formed . They
are merely acceptable abbreviations for wffs .
 3
Chapter

openfonnula Definition . An OPEN FORMULA is the result of

replacing at least one occurrence of a name in a wff by
a new variable (one not already occurring in the wft ) .

Comment. Open formulas are not wffs and hence never

appear as sentences in proofs . The notion of an open
formula is used to present the rules of proof for
predicate logic .

Examples.
Fx is an openformula.
It occursaspart of the wff ' VxFx.

Fxy is an openformula.
It occursas part of the open formula 3yFxy, which in
turn is part of the wff Vx3y Fxy.

scope Definition. The SCOPE of a quantifierin a given

formulais theshortestopenformulato theright of the
quantifier.

Examples.
In the wff
(VxFx & 3y(Fy - + Gy ,
thescopeof Vx is theexpression
Fx
andthescopeof 3y is theexpression
( Fy- + Gy).
 3
Chapter

In thewff
3y(Fy & Vz(Gz v - Rzy ,
thescopeof 3y is
(Fy & Vz(Gz v - Rzy ,
but the scopeof "V'z is
(Gz v - Rzy).

wider and Definition. A quantifier whose scopecontains another

narrower other quantifier is said to have WIDER SCOPE than
scope the second. The secondis said to have NARROWER
SCOPE than the first.

bound Definition. A variable, a , that is in the scope of a

variable quantifier for that variable (i.eVa or 3a ) is called a
BOUND VARIABLE . A variablethat is not boundby
a quantifier is saidto be UNBOUND or FREE.

3.1.2* Identify all the open fonnulas appearingin exercise

Exercise
3.1. If an openfonnula appearsin an expressionthat is
not well-fonned, give an exampleof a wff in which it
might appear.

Exercise3.1. 3 In the following sentences

, determinethe scopesof all
quantifiers.
.
1 ' Vx(Px - + ' VzRxz)
..
11 - ' VxPxH ' Vx' VzRxz
Chapter3

...
m VxPx- + Vz- VxRxz
.
IV Vz(Px - + VxRxz)
V Vx3yFyxb
.
VI 3y(Fy & Vz(Gz v - Rzy
..
VII VxVy(Fxy - + Vz(Hxyz& Jz
...
VIII VxVy(Rxy - + Ryx)
.
IX 3z3x(Fxz- + VyGyxa)
X 3x(x=a - + VyGyaa)

3.2 Translation of English to Quantified Wffs

translation Definition. A ttanslation schemefor the languageof

scheme predicatelogic consistsof a pairing of predicateletters
with English predicate phrases and of names of
predicatelogic with namesin English. We also include
metavariables with the predicates and associated
phrasesto indicatethe appropriateorder for namesand
variables.

Example.
According to the ttanslationscheme
LaP : a likes p
a: Abigail ,
the sentence
' '
Abigail likes everything
is translatedas
"v'xLax.
 Chapter3

Comment. It is possible to give several non-equivalent

translation schemes for sentences of English , depending
on how many places are assigned to the predicates.

Example.
Using the translationscheme
Fa.: a. is the fatherof Mary
a: John,
F is specified as a I -place predicate. Using this
scheme,the sentence
Johnis Mary' s father
is translatedas
Fa.

Using the b"anslationscheme

Fap : a is the father of p
a: John
b: Mary,
F is specified as a 2-place predicate with the first
position (occupiedby a ) correspondingto the subject
of the phrase ' is the father of ' and the second(occu-
pied by P) correspondingto its object. Using this
scheme,the sentence
Johnis Mary' s father
is b"anslatedas
Fab.
Chapter3

Comment. The choice of whether to represent English

phrases with one-place or many-place predicates is
dependent on the degree of structure that must be
included in order for an argument to be analyzed
adequately. In general, more detail is better than less
detail , since arguments may be labeled invalid
erroneously if insufficient detail is represented.

universals Variantswhoselogical form is

'v'xFx
include the following :
Everythingis F.
All things areF.

Variantswhoselogical form is
'Q'x(Fx - + Gx)
include the following :
Every F is a G.
All Fs areGs.
If it ' s an F, it ' saG .
Everythingthat is F is G.
Anything that is an F is a G.
Any F is G.
If somethingis an F, it is a G.
Only Gs areFs.
 Chapter3

There are several variants having the fonn

'v'x(Fx ~ - Gx)
including these:
No Fs are Gs.
Not a singleF is G.
Fs are neverGs.
Every F is not G.

existentials Variantswith the form

3xFx
include the following :
Somethingis F.
Thereexistsan F.
Thereis at leastone F.

Variantshaving the form

3x( Fx & Ox)
include the following :
SomeFs areOs.
At leastone F is O.
Thereexistsan F that is G.

Comment. Notice the difference between ttanslating

' ' ' '
Every F is G (equivalently All Fs are Gs ) and
' Some Fs are ' . In
G the first case, an arrow is used in
the scope of a universal quantifier . In the second, an
ampersand is the appropriate connective in the scope
of the existential quantifier .
 3
Chapter

Comment. When translating sentences of English

without the use of the identity symbol, the distinction
between ' Some F is G' ( ' At least one F is G' ) and
' Some Fs are G' ' At least two Fs are G' cannot be
( )
represented. We comment on the translation of ' at least
n' below.

identity Variantswith the form

a=f3
include the following :
a is p.
a is (numerically) identical to p.
a is the same(entity) as p.
a and p areone and the same.
a is the very sameindividual as p.

quantities Numericalquantitiescanbe expressedusingthe quanti-

fiers in conjunctionwith the identity symbol.

at least n The existential quantifier expresses ' at least one' .

Other numericalquantitiescan be expressedby asserting
the existenceof non-identical objects. Thus, for
example:
3xy x~y At leasttwo
3xyz x~y & x~ ) & y~ ) At leastthree
The sentence' There are at least two dogs' may
translated3x3y Dx & Dy ) & x~y) .
 3
Chapter

exactly n There are exactly n objects if there are at least n, and

all objectsare identical to one or other of thosen. For
example:
3x'v'y x=y Exactly one
3xy (x~y & 'v'z(z=x v z=y Exactly two
3xyz (x~y & x~z) & y~ )
& 'v'w w=x v w=y) v w=z Exactly three

at most II Thereare at most n objectsif thereare exactly zero, or

exactly one, etc., up to exactly n objects. For example,
' Thereareat most two '
dogs may be b' anslatedas:
- 3xDx v (3x (Dx & Vy (Dy ~ y=x v
3xy (Dx & Dy) & x' *Y) & Vz(Dz ~ z=x v z=y
This is equivalent to saying that there are not three
distinct dogs, i.e.:
- 3xyz Dx & (Dy & Dz & (x* y & (x* z & y* z )

Comment. There are many subtleties in the ttanslation

of English quantifier phrases into the language of
predicate logic . Such phrases often introduce ambiguity
into the expressions of English . The exercises
below illusttate some of the subtleties and ambiguities .
Chapter3

Exercise3.2 Give translation schemesand translatethe following

sentencesof English into the language of predicate
logic. If a sentenceis ambiguous, give all the reasonable
translationsof it .

( 1- 22: Translate using one-place predicates only .)

1* All dogs are mammals.
2* Some sharks are ovoviviparous .
3* No fishes are endothermic.
4* Not all fishes are pelagic .
5 *
Reptiles and amphibians are not endothermic.
6* Some primates and rodents are arboreal.
7* Only lagomorphs gnaw.
8 *
Among spiders, only tarantulas and black widows are
poisonous.
9* All and only marsupials have pouches.
10 * No fish have wings unless they belong to the family
Exocoetidae.
11* Some organisms are chordates and some organisms are
molluscs , but nothing is both a chordate and a mollusc .
12* None but phylogenists are intelligent .
13* Animals behave normally if not watched.
14* Animals behave normally only if not watched.
15* Some sharks are pelagic fish , but not all pelagic fish
are sharks.
16* If Shamu is a whale and all whales are mammals, then
Shamu is a mammal.
17* No sparrow builds a nest unless it has a mate.
18* No organism that is edentulous is a predator.
19* All predators are not herbivorous.
20 * Not all predators are carnivorous.
 Chapter3

.
21* A mammalwith wmgs
22 * A mammalwith wIngs flying .
(23 29: Try thesefirst with one-placepredicates,then with many-
-
placepredicates.)
23* Shamucan do everytrick.
24 * Shamucan do any trick.
25 * Shamucannotdo everytrick.
26* Shamucannotdo any trick.
27 * If any whalecan do a trick , Shamucan.
28 * If everywhale can do a trick , Shamucan.
29 * If any whalecan do a trick , any whale can do a trick.
( 30- 57: Translationswith many-placepredicates.)
30 * Godzilla ate Bambi.
31 * SomethingateBambi.
32 * Godzilla ate something.
33* Bambi ateeverything.
34 *
EverythingateBambi.
35 * Somethingate something.
36 * Somethingateeverything.
37* Everythingate something.
38 * Everythingateeverything.
39 * Everythingate itself.
40 * Somethingateitself.
41* Nothing ate itself.
42 * Somethingate nothing.
43 * Everyonesaidsomethingto everyone.
44 * Everyonesaid somethingto someone.
45* Everyonesaidnothing to someone.
46 * No one saidanythingto anyone.
47 * There is a reptile smaller than a cat but larger than a
dog.
 3
Chapter

48* Somefishesswim slowerthan humans.

49* Somefishesare smallerthan everymammal.
50* Somewhaleseat only fast-moving fishes.
51* Somewhalesdo not eat any fast-moving fishes.
52* If anythingeatsfast-moving fishes, sharksdo.
53* ' '
Jaguarstails are longer thanocelots tails.
54* If an organismis symbiotic with a clown fish then it is
a seaanemone.
55* The phalangesof birds are homologousto the phalanges
of humanswhereasthe eyes of octopi are analogous
but not homologousto the eyesof mammalsand
birds.
56* Somewhaleseat more thanall fishes.
57 * There is a monkey who grooms all and only those
monkeyswho do not groom themselves .
(Translationsinvolving the identity symbol.)
58* Exactly one cheetahexists.
59 * Thereis only one Paris.
60* Bambi ateat leasttwo trees.
61* Bambi ateeverythingexcepthimself.
62 *
Every dog hasexactly one tail.
63 G.odzilla ate Bambi, and somethingelseate Godzilla.
64 Bambi was not eaten by Godzilla but by something
else.
65 Godzilla atenothing but Bambi.
66 Godzilla ateeverythingexceptBambi.
67 Only Bambi is afraid of Godzilla.
68 Nothing but Godzilla likes Bambi.
69 Thereis a fish that' s bigger than all the others.
70 Nobody likes somebodywho eats everything except
Bambi.
 Chapter3

3.3 Primitive Rule. . of Proof

Comment . We introducesix newprimitiverulesof

proof: universalelimination, universalintroduction,
existentialintroduction
, elimination
existential , identity
introduction, andidentityelimination. To allow
succinctstatements of the f1r8tfour of these, the
notionsof universalization , and
, existentialization
instancearedefined .

universa - Definition. A UNIVERSALIZATION of a sentence

lization with respectto a given nameoccurring in the sentence
is obtainedby the following two steps:

( 1) Replace all occurrences of the name in the sentenceby

a variable a , where a does not already occur in the
sentence .

(2) Prefix'v'a to theopenfonnularesultingfrom step1.

Examples .
Universalizationsof
~
(Fa Ga)
include
'v'x(Fx ~ Gx)
and'v'y(Fy ~ Gy).
 3
Chapter

Universalizationsof
Faa
include
'v'xFxx
andVyFyy.

Definition. An EXISTENTIALIZATION of a sentence

lization with respectto a given name occurring in the
sentenceis obtainedby the following two steps:

( 1) Replace at least one occurrence of the name in the sentence

by a variable a, where a. does not already occur
in the sentence .

(2) Prefix 3a. to the openfonnula resulting; from step 1.

Comment. Notice the difference between step 1 in the

definition of universalization and step 1 in the
definition of existentialization . Universalization requires
replacement of all occurrences of the name by
with a .

Examples.
Existentializations
( Fa- + Oa)
include
3x ( Fx - + Ox),
3x( Fa- + Ox),
and3y ( Fy - + Oa) .
 Chapter

Existentializationsof
Faa
include
3xFxx,
3xFax,
and3yFya.

i DStance Definition . An INSTANCE of a universally or existentially

quantified sentence is the result of the following
two steps :

( 1) Removethe initial quantifier.

(2) In the open formula resulting from step 1, uniformly

replace all occurrencesof the unbound
variableby a name.

Comment. This is called INSTANTIATING the sen-

teRce.Thenameis calledtheINSTANTIAL NAME.

Examples.
The sentence
'v'xFx
hasinstances
Fa, Fb, Fc, etc.

The sentence
3x(Fx & Gx)
hasinstances
( Fa& Ga), ( Fb & Gb) , ( Fc & Gc), etc.
Chapter3

The sentence
3xVy(Fxy ~ Gy)
hasinstances
Vy( Fay~ Gy), Vy(Fby~ Gy), etc.

Exercise3.3.1* Pair wffs and from the list of sentences

below. lY appear in several pairs.

.
I VxFax
11 3x(Fxa & VyGyxa)
.. .
111 3xFax
.
IV Fab
V 3yVxFyx
VI 3zx(Fxz& VyGyxa)
VII VxyFxy
...
VII1 VxFxa
.
IX 3zx( Fxz& VyGyxz)
X Fha & VyGyba

Comment . The primitive rules of proof for predicate

logic include all the primitive rules from chapter 1.
There are also introduction and elimination rules for
the two quantifiersand for the identity symbol. Two of
the new rules haveconditionsthat must be met for the
applicationof the rules to be correct.
 Chapter3

universal -e1im Given a universally quantified sentence (at line m),

conclude any instance of it .

Condition : None.
Annotation : mOVE
Assumption set: sameas line m.
Also known as: Universal Instantiation .

Examples.
(a)
1 ( 1) 'v'xFx A
1 (2) Fa I ' VE
1 (3) Ph IVE

(b)
1 (1) VyRyy A
1 (2) Rbb 1 'v'E

Given a sentence(at line m) containing at least one

intro occurrenceof a name, conclude a universalizationof
the sentencewith respectto that name.

Condition: The name in question must not

'
occur in any assumptions in m s
assumptionset.
Annotation: m VI
Assumption -set: sameas line m.
Also known as: Universal Generalization.
Chapter3

.
Examples
a
()
1 (1) VxFx A
1 (2) Fb 1VE
1 (3) VxFx 2VI
1 (4) VyFy 2VI

(b)
1 'v'x(Fx~ Ox)
---
A
1 Fa~ Oa I ' VE
3 'v'xFx A
3 Fa 3 VE
1,3 Oa 2,4 - +E
1,3 'v'xOx 5 VI

Exampleof violation of the VI condition.

(c)
1 (1) 'v'x(Fx- + Gx) A
2 (2) Fa A
1 (3) Fa- + Ga 1 VE
1,2 (4) Ga 2,3 - +E
wrong ! 1,2 (5) 'v'xGx 4VI

Comment. Ordinarily we cannot conclude VxFx

'
merely from Fa- the fact that one thing is F doesn t
guarantee that everything is F ! The condition on VI
ensures that we do not make this mistake. H the
sentence Fa is b"ue, and furthennore would still be true
no Inaner what the name denotes, then clearly everything
is F , so we can conclude VxFx . When the condition
on VI is met, then we are in such a situation : if we
 Chapter3

prove Fa from assumptionsthat do not contain the

name a and hence say nothing in particular about its
referent, then we could just as well have used a
different name, sayb, and provedPh. In fact, when the
condition on VI is met, any proof of Fa can be turned
into a proof of Ph just by replacing any involved
occurrencesof the name a by the name b. This is
sufficient to guaranteethat everything is F; hence, we
can concludeVxFx.

existential- Given a sentence (at line m) containing at leastone

intro occurrenceof a name, concludean existentializationof
that sentencewith respectto that name.

Condition: None.
.
Annotation: m 31
-set
: same
Assumption m.
asline
Also known as : ExistentialGeneralization.

Examples.
(a)
m
~
1
<

(b)
---

1 Vx(Fx~ Gx) A
2 Fa A
1 Fa~ Ga 1VE
1,2 Ga 2,3 ~ E
1,2 Fa& Ga 2,4 &1
1,2 3x(Fx& Gx) 531
 3
Chapter

(c)
1 (1) VxFax
: A
1 (2) 3yVxFyx 131

existential- Given a sentence(at line m) and an assumption(at line

eUm i ) that is an instanceof some existentially quantified
sentencethat is present(at line k), concludethe given
sentenceagain.

Condition: The instantial name at line i must

not appearin the sentenceat line k
or in the sentenceat line m. Also, it
must not appear in any of the assumptions
belonging to the assumption
set at line m, other than the
instancei itself.
Annotation : k , m3E (I)
Assumption set: all assumptionsat line m other than
i , and all assumptionsat line k.

Examples.
(a)
1 (1) 3xFx A
2 (2) Fa A
2 (3) Fav Oa 2vl
2 (4) 3x(Fxv Ox) 331
1 (5) 3x(Fxv Ox) 1,4 3E(2)
 3
Chapter

(b)
1 (1) 3x(Fxx- + P) A
2 (2) Faa- + P A
3 (3) 'v'xFxx A
3 (4) Faa 3 VE
2,3 (5) P 2,4 -+E
1,3 (6) P 1,5 3E(2)

Examples violationof3Econdition.
(a)
1 (1) 3xFx A
2 (2) Fa A

f(~~
3 (3) Ga A
2,3 (4) Fa& Ga
2,3 (5) 3x(Fx& Ox)
r-

1,3 (6) 3x(Fx& Ox) 1,53E

(b)
1 (1) 3xFx
2 (2) Fa
1 (3) Fa 1,2 3E

(c)
1 (1) 3xFax A
2 (2) Faa A
2 (3) 3xFxx 231
wrong ! 1 (4) 3xFxx 1,3 3E(2)
Chapter3

Comment . If all we know is that somethingis F, we are

not entitled to reasonas if we know what it is that is F.
As in the case of VI , a use of 3E that meets the
conditionsaboveand usesa certaininstantialnamecan
be turned into a proof of the sameconclusionfrom the
same assumptionsbut using any different instantial
name. This showsthat the conclusiondoesnot rest on
any assumptionsabout the actual identity of the thing
that is said to exist. That is, if we apply 3E to 3xFx by
discharging the assumedinstance Fa, the conditions
ensurethat we do not mistakenly use any information
about the referent of ' a' in particular . After all , 3xFx
'
saysonly that somethingis Fit doesnt tell us which
individual is F.

identity -intro Concludeany sentenceof the form a=a.

Condition
: None.
Annotation: =1
Assumptionset: Empty.

.
Example
c =c =1

Comment. An identity statementof the form a=a, like

a theorem, requiresno assumptionsto justify its assertion
.
 Chapter3

identity -elim Given a sentence cI> ( at line m) containing a name a,

and another sentence (at line n) that is an identity
statement containing a and another name (3, conclude a
sentence that is the result of replacing at least one
occurrence of a in cI>with (3.

Condition : None.
Annotation : m, 1I =E
Assumption set : The union of the assumption sets at
lines m and n.
Also known as : Leibniz ' s Law , Substitutivity of
Identity

Examples.
- =-

( 1) Fa
(2) a=b
(3) Ph 1,2 =E

Fa& Ga
- - -

A
2 b=:;a A
-.~II~

1,2 Fb & Ga
1,2 Fb & Gb

(c)
1
- - ~

Vx ( Fxa - + x = a) A
2 Fha A
1 Fha - + b=a I ' VE
1,2 b=a 2,3- +E
1,2 Vx ( Fxb - + x= b) 1,4 =E
Chapter3

Comment. The role of identity elimination is not regardedas valid

in all contexts. For instance, if Frank believesthat Mark Twain is a
novelist then, even thoughTwain= Clemens, it doesnot follow that
he believes Samuel Langhorne Clemens is a novelist (if , for
" " " "
example, he has heard the name Twain but never Clemens).
, contextswhere the role fails, such as belief
For historical reasons
reports, are called intensional contexts in contrast to the
extensionalcontextsprovidedby the ordinary predicateswhich the
languagedevelopedin this chapteris intendedto represent.

Exercise3.3.2 Provethe following sequents , using the primitive rules

of predicatelogic. You may also usederived sentential
rules.

S87* 3x(Gx & - Fx), ~x(Gx ~ Hx) ~ 3x(Hx & - Fx)

S88* 3x(Gx & Fx), ~x(Fx ~ - Hx) ~ 3x- Hx
S89* ~x(Gx ~ - Fx), ~x(- Fx ~ - Hx) ~ ~x(Gx ~ - Hx)
S90 3x(Fx & Ga), ~x(Fx ~ Hx) ~ Ga& 3x(Fx & Hx)
S91* ~x(Gx ~ 3y(Fy & Hy ~ ~x- Fx ~ - 3zGz
S92* ~x(Gx ~ Hx&Jx), ~x(Fx v - Jx~ Gx) ~ ~x(Fx~ Hx)
S93* ~x(Gx & Kx H Hx), - 3x(Fx & Gx) ~ ~x- (Fx & Hx)
S94 ~x(Gx~ Hx), 3x Fx&Gx) & Mx) ~ 3x(Fx& (Hx&Mx
S95* ~x(...(jxv- Hx), ~x Jx ~ Fx) ~ Hx) ~ - 3x(Fx & Gx)
S96* - 3x(- Gx & Hx), ~x(Fx ~ - Hx) ~ ~x(Fxv...(jx~ - Hx)
S97 ~x- (Gx & Hx), 3x(Fx & Gx) ~ 3x(Fx & - Hx)
S98 3x(Fx & - Hx), - 3x(Fx & - Gx) ~ - ~x(Gx ~ Hx)
S99 ~x(Hx ~ Hx & Gx), 3x(- Gx & Fx) ~ 3x(Fx & - Hx)
SlOO ~x(Hx ~ - Gx), - 3x(Fx & - Gx) ~ ~x- (Fx & Hx)
SlOl * ~x(Fx H GX) ~ ~xFx H ~xGx
SlO2* 3xFx ~ Vy(Oy~ Hy), 3x.Jx~ 3xGx~ 3x(Fx&Jx) ~ 3zHz
SlO3 3xFxv 3xGx, ~x(Fx ~ Gx) ~ 3xGx
 Chapter3

SI04 'v'x(Fx ~ - Gx) ... - 3x(Fx & Gx)

SIO5* 'v'x(Fx v Hx ~ Gx & Kx), - 'v'x(Kx & Gx) ... 3x- Hx
SI06 'v'x(Fx & Gx ~ Hx), Ga& 'v'xFx ... Fa& Ha
SIO7* 'v'x(Fx H 'v'yGy) ... 'v'xFx v 'v'x- Fx
SIO7 'v'y(Rl~ (3x<:Jx~ Oy ,'v'x(Gx ~ Hx), 'v'x(- Jx~ ....JIx
)
- - -
... 3x Jx ~ Fav 'v'x Gx
SI09* 'v'x(Dx ~ Fx) ... 'v'z(Dz ~ ('v'y(Fy ~ Gy) ~ Gz
SIlO 3xFxH 'v'y(FyVOy~ Hy), 3xHx, - 'v'z- Fz'" 3x(Fx&Hx)
SIII * 'v'xFx'" - 3xGxH - (3x(Fx & Gx) & 'v'y(Gy ~ Fy

S112* "V'x(3yFyx ~ "V'zFxz) ~ "V'yx(Fyx ~ Fxy)

S113 3x(Fx & "V'yOxy), "V'xy(Gxy~ Oyx) ~ 3x (Fx & "V'yOyx)
S114 3x- "V'y(Oxy ~ Oyx) ~ 3x3y(Oxy & - Oyx)
S115 "V'x(Gx~ "V'y(Fy ~ Hxy , 3x(Fx & "V'z- Hxz) ~ - "V'xOx
S116 "V'xy(Fxy ~ Oxy) ~ "V'x( Fxx ~ 3y(Oxy & Fyx
S117* "V'xy(Fxy ~ - Fyx) ~ - 3xFxx
S118 "V'x3y(Fxy & - Fyx) ~ 3x- "V'yFxy
S119 "V'y(3x- Fxy ~ - Fyy) ~ "V'x(Fxx ~ "V'yFyx)
S120 3xFxx ~ "V'xyFxy ~ "V'x(Fxx ~ "V'yFxy)
S121 a=b ~ b=a
S122 a=b & b=c ~ a=c
S123* a=b, ~ ~ a~
S124 Fa & "V'x(Fx ~ x=a), 3x( Fx & Ox) ~ Oa
S125* "V'x x=x ~ 3xFx, "V'x(- Fx v Ox) ~ 3x(Fx & Ox)
S126 "V'x(Fx ~ Ox), "V'x(Ox ~ Hx), Fa & - Hb ~ a~b
S127* 3x Fx & "V'y( Fy ~ y=x & Ox), - Oa ~ - Fa
S128 3x"V'y - Fxy ~ x=y) & Ox) ~ "V'x(- Ox- +3y(y~x & Fyx
S129 3x(Px & ("V'y(Py ~ y=x) & Qx , 3x- (- Px v - Fx)
~ 3x( Fx & Qx)
S130* "V'x3yOyx, "V'xy(Oxy ~ - Oyx) ~ - 3y"V'x(x~y ~ Oyx)
 3
Chapter

3.4 Sequents, Theorems, and Derived Rules of Proof

, using the primitive rules

Exercise3.4.1 Provethe following sequents
from chapter 3 and any of the primitive or derived
rules from chapter1.

S150* - ~xPx+ 3x- Px QuantifierExchange

S151 - 3xPx~~ ~x- Px QE
S152 - ~x- Px + 3xPx QE
S153 - 3x- Px ~~ ~XPX QE
S154 ~x(Px & Qx) ~~ ~XPX& ~ xQx Confinement
S155* ~x(Px ~ Q) ~~ 3xPx~ Q Conf
S156* ~XPXv ~ xQx ~ ~x(Px v Qx) Conf
S157* 3xy(Px & Qy) ~~ 3xPx& 3xQx Conf
S158 3x(Px v Qx) ~~ 3xPxv 3xQx Conf
S159 3x(Px ~ Q) ~~ ~xPx~ Q Conf
Sl60* p ~ 3xQx ~~ 3x(P ~ Qx) Conf
S161 P ~ ~ xQx ~~ ~x(P ~ Qx) Conf

QE Comment. The important quantifier -exchange rules establish

(derived rule ) that an initial tilde can always be moved to the
right of an adjacent quantifier , changing the quantifier
from a universal to an existential (or vice versa) . Also ,
a tilde that immediately follows an initial quantifier
can be moved to the front of the sentence provided ,
again, that the quantifier is changed as just described.
Although the above versions of the rules (S150- S153)
involve quantifications of a simple formula , it is easily
recognizable that the proofs of these sequents do not
 Chapter3

dependon the simplicity of the quantifiedfonnula.

~
QUANrrD IER EXCHANGE(QE) maythusbe used
asa derivedruleof proofasbelow.

.
Example

---
1 3x- (Fx & Ox) A
2 3xOx~ Vx(Fx & Ox) A
1 - Vx(Fx& Ox) 1QE
1,2 - 3xOx 2,3 MTr
1,2 Vx- Ox 4QE

Exercise3.4.2 Prove the following sequents , using any of the

primitiveor derivedrulesestablished
sofar.

T40 I- ' Vx(Fx - + Ox) - + (' VxFx- + ' VxOx)

T41 I- ' Vx(Fx - + Ox) - + (3xFx - + 3xOx)
T42 I- 3x(Fx v Ox) H 3xFxv 3xOx
T43 I- ' Vx(Fx & Ox) H ' VxFx& ' VxOx
T44 I- 3x(Fx & Ox) - + 3xFx& 3xOx
T45 I- ' VxFxv ' VxOx- + ' Vx(Fx v Ox)
T46 I- (3xFx- + 3xOx) - + 3x(Fx- + Ox)
T47 I- (' VxFx- + ' VxOx ) - + 3x(Fx - + Ox)
T48 -
~ ' Vx(Fx H GX) v (' VxFxH ' VxGX )
T49 -
~ ' Vx(Fx H GX) v (3xFxH 3xGx)
T50 ~ - ' Vx(P & Fx) H (P - + - ' VxFx)
T51 ~ - 3x(P & Fx) H (P- + - 3xFx)
T52 ~ ' Vx(P v Fx) H (- P - + ' VxFx)
T53 ~ 3x(P v Fx) H (- P - + 3xFx)
T54 ~ ' Vx(Fx - + P) H (3xFx- + P)
T55 ~ - 3x(Fx - + P) H - (' VxFx- + P)
Chapter3

T56 .- Vx ( Fx H P) - + (VxFx H P)
T57 .- Vx (Fx H P) - + (3xFx H P)
T58 .- (3xFx H P) - + 3x(Fx H P)
T59* .- (VxFx H P) - + 3x(Fx H P)
T60 .- Vx3y x=y
T61 .- Vx ( Fx H 3y(x=y & Fy
T62 .- Vx ( Fx H VY(X=y - + Fy
T63 .- Vxy ( Rxy H X=y) - + VxRxx

prenex fonD Comment

. The quantifier- exchangerules and the confinementrules
(S154- S151) indicate that any sentencemay be convertedinto an
equivalentsentencein which no connectiveis outsidethe scopeof
any quantifier in the formula. Such a sentence
, called a PRENEX
sentence
, has all its quantifiers in a row at the beginnin~ of the
sentence
.

Eyem. Q3.4.3 For each of the following , find a prenex equivalent and prove the
.
equivalence

1* v x(Px-+- VzRxz)
2* 3y(Fy&Vz( Hyz& Jz
3* 3xFxa-+- VyGyaa
4* - VxFx-+- 3xHx
5* - 3x( 3yFyx-+- - VzGzx)

Find prenexequivalentsfor the other non-prenex sentences in this

chapterand the next.
This excerpt is provided, in screen-viewable form, for personal use only by
members of MIT CogNet.

Unauthorized use or dissemination of this information is expressly forbidden.

If you have any questions about this material, please contact

cognetadmin@cognet.mit.edu