PAf,I

ll r

SECTION

Ct

Classes

AlttCAL tOGt(

ald

C,ategorical Propositions

Like Moliere's M.

The theory of deductiontainrs to explain the relatious of pr ernises and conclusion in valid argunrents. li'aims also to provide techniqtres for the appmisal
of deductive arguments, that is, for discriminating betvr:err ralid and inralid cleductions. To accomplish this, nvo great bodies oftheory hart becn developed.
The first of these is called dassical or Aristotelian lagic, aker the grcat Greek
philosopher who initiatcd this study'. The second is callecl nodnn or modnn s;ttbolitlogic, developed mainll,during the rwentieth ce ,tury. Classical logic will be

lardoin who
found thot he had
long been speoking prose, I found

that I hod long

been forming

propositions. I said

to rnysel{,

"Yes,

the topic of this and the follorving rwo chapters (Chaplers 6, 7, and 8); urodern
symbolic logic rvill be the topic gf Chapters g. 10, and I l.

farm propositiwrs

whm mytongue
does more than

wag- I tonn them

out oi terms. I say

fv arrnnorraI
i . Lgv rvv,

DDnDrrcrrrn
vJr r tvrtJ r(

something about
something.

6.1 The Theory of Deduction

6.2 Clarserand ategorical Propositions
5.3 The Foor tfinds of Categorical Propositionr
6.4 Quality, Quantity, and Distribrrtion
6.5 The Traditional Square of Opposition
6.6 Further lmmediate lnferences
6.7 [xistential lmport and the lnterpretation

Therefore ! ought

to be oble, in
serious tolk, to

pinpoint those two

ports oi my

proposition.

ought to know
exadlywhat
tclking

lom

obul

ond

exoctly whot I om
saying obant

e.s

of (ategorlcal Propositions
Symbolism and Diagramr for
Categorical Propositions

6.2

Deductlve
Argement
trrn argument that
daims to establish iB

concluion conclusively one d$e two

dasses of argumentr.

Valid Argument
A deductive argument in which, if all

ttre premises are

true, the conclusion
must be'.rue.:

176

and Categorical propositisnr

classical Iogic deals mainly with argumenrs that are based on the relations. of
ofobjecs to one another. By a'class we mean a collecdon ofall objects har"
ing some speci{ied characteristic in common.r Everyone can see immediately
that two classes can be related in at least the following three ways:

-Arthurfuton

Glasses

classes

it

Luce

Anstode (384-322s.c.) wasoneof thetorveringintellecisof thear:cientworld.

.{fter studf ing for nr'enry )ars in Plato's Academ; he becarne tur.or to fexander
the Great; later he founded his onn school, the L1'ceurn, tvhert he conrr,iLruted
substantiall,v to nearly everv field of human knowledge. His great treadses on reasoning were gathered together after his dearj, and came tc be called the Oryaaon,
meaning literally rhe "instrument " the fundamenal rool of knou,ledge.
The lvord logic did not acquire ic modem meaning until the second ceut'ry
e-o., but the subject matter of logic was long turderstood to be the mattem treared in Aristotle's seminal organon.Aristotelian logic has been the foundation of
rational analpis for thousands of years. oler the course of those centuries it has
been very greatly refined: its notation has been much improved, is principles
ha'e been carefully formulated, is intricate srrucrure has been completed.'ihis
great system ofclassical loglc, set forth in this anc the next two chapters, remains
an intellectual tool of enormous powe4 as beautiful as it is penetritug.

6.I The Theory of Deduction

l.

llb turn now to the detailed analpisofarguments. Preceding chapters hare dealt
with the l*guug. in which arguments are formulated; in this and succeeding

?.

chapters we explore and explain the relations benupen the premises of an argumcnt and its conclusion.
AII of Part Two of this book is devoted to deductive arguments. A deductjve
argument is one whose premises are claimed to provide conclusive grounds for
the truth ofits conclusion. If thatclaim is correct, that is, if the premises of the argument really do assure the ruth of ir conclusion with necessity, that deductive
argument is ralid" Every deductfue arjument ei&er does n'hat ii claims, or does

no! therefore

every deductive argument is

impossible for

ib premises to be rue without its conclusion being true also.

eitherralid, or invalid. If it

is

valid it

AII of one class may be included in all of another crass. Thus the class of all
dogs is wlolly included (or wholly contained) in the class of al| mammak.
Some, but not all, of the members of

o,e class may be incl*ded in another

Thus the clxs of all athtetzs is partially included (or partially contained) in the class of allfanalts.
class.

3.

Two classes may hal'e no members in common. Thus the class of all triangles
and the class ofall circlesmay be said to exclude one another.

These three reladons may be applied to classes, or categories, of evcry

sort.
a deductive argument, we present propositions that itate thr relations

In
be-

tween one category and some other category The propositrons

with which

(Ir.ss aleeet<4
The collection of all

is

lThe

concept
was briefly introduced in chapter 4, in expraining the intension
-ofclasses
terms, and definitions
based on intension.

obies

of

that have

some specifgd characteristic in common.

177

178

Ctrapter6 Categoricalpropositions

The FourKinds ofCategorical

such arguments are formurated are therefore caled categorical propositions.

Cltegorical propositions are the fundamental elemenrsihe nritaing

lto.L,
of argument, in the classical account of deductive logic.
Consider the argument

-.

No athletes are vegetarians.

All football players are athletes.
Therefore no football players are vegetarians.

This argument contain$ three categoricar propositions. we may dispute

the
tu& of its premises, of course. But the r.Ltio* of the classes'expressed in

these propoaitions
lerds an argumenr that is cenainty varid; if those premises
ar true. tlrat conclusion rarrJ, be true. And it is plain that
each of the premises
is indeed catgorical; thatis\ wh praniv
or btia, thal some cta,ts

d,rdd in saau otlw

$frns,

S is

pit

f; in uthde

irt

Propositions 179

Categorical propositions are often represented wirh diagrams, using

nuo interlocking circles to smnd for the two classes invohed. These are
called Vmn diagrans, narned alter the English logician and rnathematician,John Venn (1824-1923), who invenred them. later we will explore
these diagrams more fully, and we will find that such diagrams ari exceedingly helpful in appraising ihe validity of deductive argument* For
the present we use these diagrams only to exhibitgraphically the sense of
each categorical proposition.
We label one circle S, for subject clas, and the other circle p, for predicate class. The diagram for &e A proposition, which xserts tlrat all iis p,
will have that portion of S which is outside of P shaded our, indicating

that there are no members of S that are not members of p. So tfo

proposition is diagrammed thus:

or A
tn this illustrative arg'ment the
three categorical propositions are about the class ofal athretes,
&i crass ofa[
vegetarians, and the class ofall football playen.
aitical first step in deveroping theory of deducrion based on classes.
therefore, is that of identif,ing the kinds of categoricar proposirions,
and exploring the relations among them.
dnss

- !e

6,3 The Four Kinds of (ategoricat propositions

There can

be

four and only four ki ndsd randad{ont catqmfu

ar pnpositions. Hue

fsllsw examples cf each of the four kinds:

Catcgorkd
Propoddon
A propositioa,
used in deductine
arguments, *lat
asserts a relationship between one
category and some
other category.

Unlverrat

i.
L
3.
d

AIi poiiCcians are

No politicians are

lian. ) e,ojed Ti
liars. -) prcor;cat

cqle
.r" o* tl"i cnat;t*l / 641tp2,.
^

Some politicians are liars.

Some poliricians

S ro F

@are

ttt'

Univelsal negative propositions. The second example above, .No politilian" is a proposition in which it is denied, universally, that any
member of the class of politicians is a member of the class of lian. It as.
serts that the subiect class, S, is wholly excluded from the predicate class,
P. Schematicaily, categorical propositions of this kind can be written:
cians_are

No Sis P

\rfs asy .xrmine, dose!, each of these tinds in hrrn.

l.

P/oo6'i'da'" ''

A PtW*t*lrr
Univcrd eemarirc pmpodfirc. In these we asse rt that ttu whatc of w
d,o ie iruludd a aataindin Mottw cr*si"A[ politicians are lian is an

exarrple; it asserts thateverymember ofone crass, the crass ofpoliticians,

is a member of inother crass, the crass of liars. Any univenar
aftirmative
proposition can be written schematicallv as:

where again .tand Prepresent the subject and the predicate terms. Thb
kind of proposition dad.*s the rel aion of hclusionbetrscen the aru erua
and denies it uniaasatly.It tells rx that no members of Sare memben of
P. Propositions in this standard forrn are called rmiversat negarive propositions. They are also catled E propositions.
The diagram for the E proposition will exhibit this munral exclusion by
having the overlapping portion ofthe nvo circles representing the clasees
Sand P, shaded out So the E proposition is diagrammed thus:

Affirmatlve

Propofilor
(A Proposltiorr)
Propositions that

where the letten Sand prepresent the

assert that the

whole of one chss

tively. such a proposition

is included or

contained in
another chs;
All 5is A

Universal
Negative
Proposkions

All Sis P

wlieaand ptdimr,

rerrns respec-

(C

dirit

th^tthe reration of class iarluura hords

between the rwo classes, and sap that the inclusion
is complete, or znr
oenal AII members of Sarc said to be ateo memben of p. piopositions in
this standard form are called rmiversal amrmatrc proposrtons.
Th.y rrr

also called A pro. positions.

Proporitiom)

Propositions that
assert Ont the

wtroh of one dass

is excluded fiorn the
whole
No Sis P

class; No

another
S

is P.

180

Chapter

Categodral Prcpositions

3.

The Four Kinds ofCategorical Propositions

l8r

aropxrtrong

Partiorlar affirmative pmpcitions. The third example above, "Some

of the class of all politicians are members of the class of all liars. But it does not afiirm this of
politicians universally. Only some particular politician or politicians are
said to be lian. This proposition does not affrrm or deny an).thing about

which sap that at least one.member of the class designated by the subject
term Sis excluded from the whole of the clas designated by the predicate

. politicians are liars," aflirms that some members

rcim P. The denial is not universal. Propcsitions in this sundard fonn are
called particular negatir,e propmitions. They are also called 0 propositions.
The diagram for the O proposition will indicate tlut there is at least one
memberof Sthat is notamember ofPby havingan rplaced in that region of S
&atis ouside of P So the O proposition is diagrammetl thus:

the class of all politicians; it makes no pronouncements about that entire

class. Nor does it say that some politicians are not liars, although in some
contexts it may be taken to suggest that. The literal and exact interpreation of this proposition is the assertion thatthe class of politicians and the
dass of Iian iaue sorru mmfu or nanfuis in cwmm. That is what we undrrstand this standard form proposition tc meao.
"Some" is an indefinim term. Does it mean 'at least one" or "at least
h^,o" or "at least several"? Or how many? Context might afiect our understailding of the term as it is used in everyday speech, hut logicians, for the
sake of definiteness, interpret "some" to mean 'a, lzasi ont." A particular
afirmative proposition may be written schematically as:

Some Sis not P

Some Sis P
says that at least one member ofthe class designated by the subject
term S is also a rnember of the class designated by the predicxe tena P.
The proposition alfirmsthatthe relation of class indusion holds, but does
not afirm itof thefint class univervllybutonlypartialty, that is, itis afrrmed
of some @tiatlarmembe4, or members, of &e first class. Propositions in this
standard form are called partiorhr aftmatirr propositims. Th"y
arq
"..
called I propositions.
The &agram for the I proposition will indicate that there is atleast one

which

member of

that is also a member of P by having an r placed in that region

in which the two circles overlap So the I proposition is diagrammed thus:

The examples that have been used in this section employ classes that are
simplynamed: politicians, lian, vegetarians, athletes, and so on. Butsubjectand
predicate terms in standard-fornr proposidons can be more complicated than

for example, the proposition "All candidates for the position are
penons ofhonor and integrity" has the phrase'candidates for the position" as
itssubject term, and the phrase "persons'of honor and integrity" as its predicate
tcrm. Subject and predicate tnns can become more intricate still, but in each
these. Thrs,

of the four standard forms a relation is expresed betrueen a subject class and a
predicate class. These four-A E, I, and
propositionr-.rl th. building
blocks of deductive argumens

. Once put before us &is analpis of categorical propositions appean to be

sinple and straightfonuard. But tlre discovery of thCfunaamenat iote of rhese
r

Propositions, and the exhibition of their relations to one anothet was a great
IteP in the qxtematic development of logic. It was on.. ofAristotle's
permient
contributions to human knowledge.2,Its apparent simplicity is deceptive. On
this

Some Sis

Partkrlar
Afflrmative
Proporltlons

(l Proposttlons)
Proposiuom that
assert that two
classes have sorne

mernbercmen$ers
in cornrnorg
So{ne S b P.

O ercrcthbotf
4.

Parriorlr

negative

cians are not liars," like the

The fourth o<ample above, "Some politii

thid,

foundation-classcs of oQjects and the iilations ,*ong thoru classes-logithe course of centuries, a highly-sophisticated spte-m
br the analph of deductive argumenr. This qnrem, i-hose subtlety and penetration mark it as one of the greaiest of intellecnral
achievements, we nowgoon
to explore in the following three srcps:
cians have erected, oyer

does not refer to politicians universally, but

Proposltlons
(O Propositlonr)

only to sonnmember or members of that class; it is @iannx But, unlike the

third example, it does not effirm the inc,lusion of sorne member or memben

o1&g fi$t rlas in

icallywriren

'Jre second

das; this is precirlywhat is drdd. It

Propositions that
$at at ha6t

is schemar

assrt
is a modem fuistorle. Readers may take delight in the beauties of
nrutode's logical discoveries" as reconstructed
in fiction for very-clever children by Matthew
See Matthew Upman, Ilarrl Stottbnuicr\ Irlscovq (Upper Montclair, NJ: institure
|rPman.
tor the Advancement
''
of irhilosophy ior Children, 1982).

$$,"--r

as:

Some Sis not

Particular
Negative

'

one memberofa

das isscfuded from

the urhde

anotfier

clars; Some Sis not P.

182

Qtufitl, Quantity' and

Chapter6 CategoricalPropositions

miles high are very delicate devices that cost many thousands

categorical
W. rho* what inferences miy be drawn,linctly from these
be mav
"i,
can
see'
will
we
reasoning,
p*p"tiriont. A good deal o[deductive
propositions
O
hnd
I,
E,
of
grasp
a
thorough
A
than
Lreh *irf, no rn1..
and their interconnections.

*5.

C-

Some members of families that are rich and famous are not Persons

of either wealth or distinction.

Some paintings prodgced by artists who are universally recognized as masters are not works of genuine merit that either are
or deserve to be preserved in museums and made available to

that are commonIn the next chapte4, we explain ryllogrims-the arguments

We explore the
propositions'
categorical
standard-form
using
d
Iy.o*t*.t
characis
uniquely
fonn
LaIm of syllogis*r] itt t"hi.t every valid a{gument
for detechniques
powerful
develop
we
its own name. And
terized and

of

dollars to manufacture.

ott

B.

183

All satellites that are currently in orbis less than ten thousand

of standard-form
tn &e remainder of this chapte6, we examine the features
relations to one antheir
explaining
deeply,
more
propositions
carcgorical

,{

Distribution

the public.

$Ln

7.

termining the ralidity (or inralidi$) of syllogisms'

and the lanIn the chapter following thar' we integnte syllogistic reasoning
of reasonlimitations
some
We
identi$
life'
evcryday
guage of argument in
i"gi*"d ri this ioundation, L'utwe also glimpie the penet'ration and wide
tt at this foundation makes posible'
"p-pticatitity

All drivers of automobiles thu are not sirf are desPeradoes who
threaten the lives of their fcllous.
Some politicians who could not be elected to the most

tions are appointed officials in our govemmerrt

minor posi-

tday.

Some drugs that are very elfective when properly administered are

not safe remedies that all medicine cabinets should contain.

*10.

No people who have not themselves done creative work in the arts
are responsible critics on whosejudgment we can rely.

Standard Form Categorical Propositlons

PnorocrnoxFonn

Nelauo1Yrr

All SisP

A-Univenal

NcSisP

E-UniversdNegative

Some Sis

Some $is not

Ercrclccr

avtd.bh

on elogk CD-POi?
Sec0on 6.3:

l,

'1,5,6,7,49

?, 3,

Affirmatile

l-ParticularAffirmative

Particular

Negative
.

E)rllan.e
AII lawYers
are wcalthy peoPle.

No criminals are

Ertry sandard-fbrm categorical proposition either affinns, or tienies, some class

relation, as we have seen. If lhe prcpositim atrros some class inc-lusion, whether
oomplete or partial, its qudity is affirnative. So tle A propocition, "All Sis P,"

Some insecs are

not Pests.

like
some historians are extremely gifted writers whose works read
first-rate novelsNo athletes who have ever accepted pay for participating in sporu
are amatfurs.

are cardidates for blue nbbons

,' 3. No dogs that are without pedigrees
by the American Kennel CIub'
shorvs sponsored

Quality
An attribute of
every categorical

md the I proposition, "Some Sis P" are both affirmative in qulity' Thcir letter
propositioq
names,Aand I, are thought to come hom the Latin wond, "lflrrrq" meaning, determined by
'l affirm.1 If the propocftion deiliec drm iodurdolr' whpther cornpletc or partial. . whether the
iB
$ality is negative. So the E proposition, "No S is R" and ttre 0 pnipositiou proposition affirms
Some S is not P," are both negatire in qudity. Their letter names, E and O, are or denies some
form of clars
&ought to come from the Latin word, "nEgO," meaning "I deny." Every cate- inclusion.
gorical proposition has one quality or the othe! afrrmative or negative.
Quantlty
B.

following propositions.

in offiIal dog

Quality

Some chemicals
are PoNons"

the
Identif the subject and predicate terms in, and name the form of' each of

2.

A"

good citizens.

'ExE-c,stt
*!.

6.4 Quality, Quantity, and Distribution

Quantity

trery sandar&form categorical proposition has some chss as is su\lecr If the

popositionrefers to allmembers of the dass designatedbyits wbjee-t term, its
luanlity is uiversal. So the A proposition, lAIl S is P," aod the E proposition,
\o S is P," are both univenal in quantity. If tbe p,ropositioa refere ria! to some
eembers of the ctass 6";*,"d by ie srbject term, ie quanEty is Particular. So
he I proposition, "Some S is R" and the O proposition, 'Some S is not R" are
both particular in quantity.

An attribute of
every categorical

proposition,
determined by
whether the
proposition refers
to all members
("universal') or onlY
some member

("particular') of the
subiect clas.

1U

Qtaliry, Quantit, and Distribtrtion

Chapter6 Categorical Proposidons

In the E proposition (for example, "No athletes arc legetarians"): The subject
term, "athletes," is distributed, because the whole class of athletes is said to be
excluded from the class ofregetarians. But it is also the case that, in asserting
that the rvhole class ofathletes is excluded from the class ofvegetarians, it is a*
serted as well that the whole class of regetarians is excluded frorn the cla^s of
athletes. Of each and every vegetarian, the proposition sars that he or she is not
athlete. Unlike an A proposidon, therefore, an E proposition rcfers to all
an
.
members of the class designated by its predicate term, and therefore distributes
ir predicate term also. E Wositions (uniuersal negaliax) distribute both their subject
and tfuir prdimte tems

The quantitv ofa standard-form categorical proposition is revealed by the

&ord viith which it begins, either "atl," "no," or'some." "AIi" and "no" indicate
that the ploposition is universal; "some" indicates that the proposition is narticular. The word "no" serves also, in the case of the E proposition, to indicate iu
negative quality, ,ls rve have seeir.

Since every standard-form categorical proposition must be either alfirmath'e or negatire, and must be either universal or particular, the four names
uniquel;, describe each one of the four standard forms by indicating its quantity and its quality: universal affirmative (A), particular aflirmative (I), univenal
negative (E), particular negative (O).

In the I proposition (for example, 'Some soldiers are cowards"): No assertion is

made about all soldiers in this proposition, and no assertion is made about all
cowards either. It sap nothing about each and el'ery soldier, and norhing about
each and every couard. Neither class is lvholly inchrded, or wholly excluded
from the other. /ra I propositions (particulnr afJirmatiua) both subiect anil @italt
tfims are not distributzd

Grneral lchrma of Standard-fom Catcgorlcal ProPotltlont

Beween the srrbject and predicate terms oferery standard-form categorical propa
sition occurs some form of the lerb "to be." This verb (accompanied by "not" in the
case of the O proposition) serrles to connect the suQject and predicate terms and is
called the copula.tflntngthe four propositions sfhematicall,v, as we did above (All S
Sis P, etc.) only the ruords "is" and "is not" appear; but (depending on

In the O pmposition (for example, "Some horses are not thoroughbreds'):

is P, Some

Nothing is said about all horses. The proposition refers to some members of the
clus designated by the subject term; it sap *rat of this part of the class of horses
that it is excluded from the class of all thoroughbreds. But they are exduded

context) ot-her forms of the rerb "to be" may be appropriate. We may change the
tense (for erample, "Some Roman empercrs were monsten" or 'Some soldiers will
not be heroes'), or change to the plural form ofthe verb (for example, "All squares
are rectangles.") kr thesc ev.amples, "were" and "are" and fuill not be" serve as cop
ulas. But the general skeleton ofa standani{orm categorical proposition afi,r"ap
consiss ofjust four pars first the quantifier, then the subject term, next the coprl
la and finally the predicate term: Thc schema may be.written as:

from the uhoh of the latter class. Given the particular horses referred to, the
proposition sa;r that each and every member of the class of thoroughbreds is zol
one of those particular horses. When something is said to be excluded frcm a
clas, the whole of the class is referred to, just as, when a person is excluded from
a country, all parts of that country are forbidden to that penon. In O
WositiDtlt
(furtiatlarntgatiuet) tfu subjeA ttnn ls nnt disiributai, krt tha@i{atltennit d;shibutaL

Quantifier (subject term) copuia (predicate term).

D. Distrlbutiron
Categoriel propositiors are regarirled

as

being about clases, the classes of o$ecb

designated by ttre subjea and predicate tcrms. We have seen that a proposition
may refer to classes in difierent wap; it may re{er to all members of a class or refer

to only ror4, memben of that clas. Thus the proposition "All senators are citizens" refers to, or is about, allxnators;tut it does not refer to all citizens. That

Dlstilbrtlm
A characterization
cf whether terms

in a categorical
proposition refer to
all memben of the
class designated by
that term.

proposition does not affirm that every citizen is a senator, but it doesn't deny that
either. Every A proposition is thrx seen to refer. to all members of the dass desig'
nated by its subject terrn, $ but doe-s not refer to all memben of the dass designated by its predicate term, P.
To characterize the wals in which tenns can occur in categorical propositions, we introduce the technical term distribution. A proposition distributes r
term if it refers to all members of the dass designated by that term. In A E, L
and O propositions, the terns,that are distributed vary as follows:

In the Apropoftion (for example, 'AIl senaton are citizens"): In this proposttion, "senators" is disuibuted, but "citizens" is nor In A pnpositions (uniuersal at'
frmathta) tfu rubjea wm is distributed,, bur tlu Wdicari bn is undistributed. i
,

Understanding distributicn, we will see that universal proposirions, both affirmative and negative, distribute their subject terms, while particular propo
sitions, whether affirmative.or negative, do not distribute their subject terms.
Thus the quantiE of any standard-form categorical proposition determines
whether its sunjea term is distributed or undistributed. We likewise see that
affirmative propositions, whether universal or particutar, do not distribute
their predicate terms, while negative propositions, both universal and particular, do distribute their predicate rerms. Thus the quality of astandard-form
categorical proposition determines whether its predicate term is distributed
or undistributed.
. In sum: The A proposition distributes only its subject term; the E propositlon distributes both its subject and predicate terms; the I proposition distributes neither its subject nor its predicare term; the O proposirion distributes
only its predicate tenn.
Which terms are distributed bywhich standani-iorm categorical propositions
will.become very important when we turn ro the eraluation of-syltogiims. tle fol
towing diagram presents all these distributions graphically, and may be usefirl in
nelping one to remember which propositions distribute which of their terms

185

,ffi

Chapter

The Traditional Square of Oppocition

Categorical ProPi:sitions

No harmfirl pesticides or chemical defoliants were genuine contributions to the long-range agricultural goals of the nation.

subject ternr distributed

predicare I

terrn

,{: All Sis P.

E No

Sis P.

Some advocates of
predicate

term
distributed

unaistrluutea
I: Some Sis P.

187

O: Some Sis not P.

m{or political,

social, and economic rcforms are

not responsible people who have a stake in maintaining the status quo.
*10.

All new labor-saving devices are major threats to the trade union
movement.

subjea term undisuibuted

able that presents all the critical informatior

about each ofthe four sandard{orm categorical propositions:

We conclude this section with a

5.5 The Tradltional Square of Opposition

The foregoing analysis ofcategorical propositions enables us to exhibit the relations among those propositions, which in turn give solid grounds for a great deal
of the reasoning we do in everyday life. We need one more technical term:
opposition. Standard-form categorical propositions having the same subject
ternrq and the same predicate terrns may (obviously) differ from each other in
quality, or in quantity, or in b<lth. Any such kind of differing has been traditionally called Wsitittl This term is used even when there is no apparent disagreement between the propositions. The rarious kinds of opposition (we will see) are
correlated with some very important truth relations.

A. Gontradictorier

Erxircr

eveliabic

on eLoglcCD-ROM
S.GSoo

6.+ t, 2,

1,5,6"7,8,9

l,

--ExERcrsEs
Name the quality and quantity of each of the following propositions, and state
whether their suQiect and predicate terms are distributed or undisributed.

*1.
2.

3.

Some p,residential candidates will bc sadly disappointed people.

Atl thosc who died in Nazi concetrtration camp,e were victims of a

cnrel and irrational tyrannySome recendy identified unstable elemenb were not entirely acci-

denal discoveries.

4,
*5.

6.

miliarfindustrial complex are mild mannered people to wtrom violence is abhorrenu

Some memben of ttre

No leader of the feminist movement is a m4ior busines executive.

AII hard{ine advocates oflaw and order at any cost are people who
will be remembered, if at a[, only for having hiled to undentand

&e major social prssures of thc ttrentffirst cen$ry.

?.

Some recent nrlingp of thc Supreme Courtwere politically

motivated decisions thatflouted the entirc history of American
legal practice.

tuo propositions

are contradictories ifone is the denial or negation of the other;

that is, if they canrrc,i botii be true and cannot borh be false. Two standard-form
categorical propositions having the same subject and predicate terms.but differing from each other in ioat quantity and quality are contradictories.
Thus &e A proposition "All judges are lawyers" aqd the O proposition,
"Somejudges are nor lawyers,'are clearlycontradictories. Theyare opposed in
bo&qudity (one affirms, the otherdenies) and quantity (one refen to all, and
the other to some). Of the pair, exactly one is true, and exacdy one is false.
They cannot both be true; they cannot both be false.
Similarly, the E proposition "No politicians are idealists" and the I proposition uSome politicians are idealisr" are opposed in both quantity and quality,
and they too are contradictories.
In sum: A and O propositions are contradictories: "All S is P " is centradicted by "Some S is not P." E and I propositions are also contradictories: "No S is
P" is conradicted by "Some Sis P."

Oppotitlon
Any logical relation
among the kinds
of categorical
propositions (4, t,
l, and O) exhibited
on the Square of

Opposition.

Contradlctorles
Two propositions

that cannot both be

true and cannot
both be false.

(ontrarler

B. Contraries

ifthey cannot both be true; that is,

if the truth of one entails the falsity of the other. Thus, "Texas will win the conr
ing game with Oklahoma," and "Oklahorna will win thi coming game with
Texas," are contraries. If either of these propositions (referring to the same

Two propositions

Two propositions are said to be contraries

that cannot both be

true; if one is true,
the other must be
fabe. they can both
be false.

188

'I.\
-

Chapter6 GategoricalPropositions

ga$e, of course) is true, then the other must be false. But these two proposi
tions are not contradictories, because the game could be a draw and then ioth
would be false. Contraries cannot both be true, but, unlike contradictories,
they may both be false.
The traditional account of categorical propositions held that univenal
propositions (A and E) having the same subject and predicate terms, but differing in quality {one affirming, the other denlng) were conrraries. Thus it was
said that an A proposirion, "AIl poers are dreamers,' and its corresponding E
proposition, "No poets _are dreamers," cannot both be true-but they can both
be &Ise and may be regaided as contraries",
One difficulty with this Arirtotelian account arises if either the A proposi
6on or the E propoaition is necessarily true-that is, if either is a logicil or
mathematical tnith, suchpsrAll quaresare rectangtes," or "No squares are cirdes-" In such a case, the clafuil that the A and the E proposition ire contraries
qmnot be correct, because a necesarill- true proposition cannot possibiy ix
false, and so cannot have a contrary, because iorrb"ri.r are two pioporition,
that can both be false. Propositions that are neither necessarily true nor nece$
sarilyhlse are said to be @ntinp* so the reply to this dilficulty is &at the pn:sent interpretation assumes (not unreasonably) that the propositions in
question are contingent, in which case the claim thatA and E piopositions having the same subject and predicate terms a.re contraries may be correct. For the
remainder of &is chapter, we therefore make the assumption that the proposi.
tions inyolved are cortingent

D. SubalternaUon
When two propositions have the same subject and the same predicate terms, an<i

4gree in quality (both afiirming or both denlng) but differ in quantity (one un!
veml, the other particular) they are called corapondingpropositions. This is also
a form of"opposition" as that term has traditionally been used. Thus the A prope
sition 'All spiders are eight-legged animals" has a corresponding I proposition,
"Some spiden are eight-legged animals." Likewise the E proposition 'No whales
are fishes" has a corresponding O proposition, "Some whales are not fuhes" This
opposition betrueen a universal proposition and is corresponding particular
proposition is known assrbalternation- la anysuch pairofcorrespondingpro.positions, the univenal proposition is called the "superaltern,, and the particutar

the "subaltern.o

Insribaltemation (on the classical analpis), the superaltem implie the truth
of the subaltern. Thus, from the unhenal afirrnatile "AIl birds have featherq?
the corresponding particular allirmativ.e, "Some birds have feathers,. rms
held to follow. From the universal negative "No whales are fishes," t}re corresponding particular, "Some whales are not fishes,' was likewise held to follow
But of course the implicarion does not hold from the particular to the univer-

from the subaltern to the superakern. From the proposicion "Some aniit is qbvious that we cannot infer that "All animals are cats."
And it would be absurd to infer, from "Some animals are not cats" that "No
sal,

mals are cam,"

animals are cats."

tubalternatlon
G.

Subcontrarles

[. The Square of Opposition

Two propositions are said to be subcontraries if they cannot both be false, aF

though theymayboth be true.
The traditional account held that particular propositions (I and O) having
subject and predicate tntrs, but differing in quality {one affrming
9*
the other denfng) arp subcontraries- It was said that the I proposition, "Some
'liamonds are precious stones,'and the O proposition, .Some diarnonds are
not precbus stones," could both be true-but they could not both be false, and
therefore must be regarded as subconFaries.
A difficulrT similar to rhe one nored above arises here too. If eitrer the I or
the O propcrsition is necessarily hlse (for example, "Some squarcs are circles,

o-.

Subcontrarls
Two propositions

thatcannotboth be

The opposition

betwecil a univenal
proposition (the

There are th,.rs fourwap in uihich propmitions may be

"opposed"-x ontrodidoria
Md u wb md neerattatxThese are represented rith an inr
porrant and widely'used diagram called rhe Squae of Oppositioa which is re
wttdri6,

subconb6ri6,

produced as Figure

&l:

daskallogL0re
unircrsal proposition

im$esfrettlth
(All Sis

P) d +contraris ----> g

lftberirtotclian interprcation hassome troubling

soction 6.7.

of iB comsponding
pailiailar prryo,sitftrn.

lquarc of
A diagram showing
the logical relationstrips among the
four types of cate
gorical propositions
(4, E, l, and O). Ihe
traditional Square of
Opposition differs
from the modem

!abe; if onek false

the otrer must be

true.Iheyon both

(No Sis p)

Opporltlon

or "some squares utre not rectangles'), it cannot have a subiontrary, because

subcontraries are two propositions that can both be true. But ifbottr the I and
the O are contingent propositions, they can both be true, and as we noted in
'connection
with contrariesjust above, we shall assume for the remainder of this
chapter that they are contingent

b bue.

superaltem) and its

conesponding
particrtlar prooosition
(ttc*baltem). ln

Square

consequences that will bc discussed in

ftGuRE

6t

of

Opposition in
important ways.

190

Chapter6

Funherlmmediatelnferences' 191

CategoricalProPositions

Relations exhibited by this square of opposition were believed to provide

tExt-",=as

fxertlses rvallable
oo eloglc (D-ROM

truth or falsehood of the remaining

propositions in each of the following set$ if we assume the first proposition in
the set to be true? (2)'What can be affirmed ifwe assume the first proposition

le<tlon 5.5:

the logical basis for ralidating certzin elementalv forms of argument. To explain tiese, we must first distinglish betlr'een immediate inferences and mediate
infercnces.
When we drarv a conclusion from one or more premises some inference
must te involved. That inference is said to be mediate when more than one
premise is relied upon (as in a syllogism) because the .conclusion is drawn from
the fint premise rirrough the mediation of the second. But, where a conclusion
is drawn from only one premise there is no such mediation, and the inference

(l) What

*I.

A number of very useful immediate infierences may be readily drawn from

information embedded in the tradirional square of opposition. Here are

2.

some examples:

is our premise, then (according to.the Square of

Opposition) one can validly infer that the corresponding O proposiaoi Una is, the O proposition with the same subjeci and predicate

Ifan A proposition

3.

terms) is false.

lfanA proposition

is

our premise, then the corresponding I proposition

is true.

4-

is our premise, iu corresponding E proposition, which

be false.
must
it,
contradicts

If an I proposition

be inferred immediatebr. There are a considerabie nurni:er of

immediate inferences based on the traditional square of opposition; we list
them here:

imm.date
lnfcrence

oti,.o

Iis truq

O is false.

Abeing given as true:

f, beingglrcn as true:

A is hlse; I is false; O

I being giran as true:

E is

O being given

as

A being given

as

is

true.

hlse;A and O are undetermined.

A is fake; E and I are undetermined.

false:

O is true; E and I are undetermined.

llcdate

O being given

I being given

E is false;

true:

E being given as false:

as false:
as false:

I
A

is

true;Aand O are undetermined.

is false; E is

true; O is true.

A is true; E is false;

is true.a

lnfcrarcc

penrise; treconclu
s*rr b drittrn ftom

thefrtpremise
dxough the mediation

d te

second.

{May immediate inferences be yelidly drawn when it is given that one of the standardform categorical propositions is undetermined? ProfessorJoseoh Gilberthas argued,
that rhey may. But Profcssor Christopher Viger points out-r,e
..iU..
*o.itly-rhat the limirations of one's kngwled-gg^abogt a given proposition
cauoot bc a reliable ground for determining the truth or falsity of other propositions

*a
".
i"*'tfri*
rdated to

"doi,

it

2"

t,4

AII succesfirl executirresare intelligent people.

No succesfrrl executives are intelligent people.
Some nrcc*sfirl executirrcs are intelligentpeople.
Some successful erecutiwsare notintelligentpeople.

No animalswidr hornsare carnirores.

Some animals witt hornsare carnivores.
Some animals.with horns are not carnivores.
All animdls witi homs are camivores
Some uranium isotopes are highly unstable substances,
Some uranium isotopesare nothighlyunstable subsances.
All uranium isoopes are highly unsable subsances.
No uranium isotopes are highly unsable subshnces.
Some college professon are notentertainingtecturers.
AII college profesors are entertaining lecnren:

Nocollegeprofessonareentertaininglecturers.
professon are entertaining lecnrers.

, Some college

i*

An inference drauln
dire$from only
one pemise.

An inferencedrawn
tonr more than one

a.
b.
c
d. '
a
b.
c.
d"
a.
b.
&
d.
a.
b.
c.
d.

Given the truth, or the falsehood, of any one of the four standard-form cate
gorical propositigns, it will be seen that the truth or hlsehood of some o1 all ol

ir.

l,

in each set to be false?

issaidtobe inw,diab.

thr

can be affirmed about the

6.6 further lmmedlate lnferences

Three other

imporant kinds of immeiliat inference are not directly associated

Qpposilo* Cnnnsiott, eoas*n, alotd Cantratositiul'fiirv-

witlr tbe
Quare of
we

norrexplain:

A. Gonverslon

lmrcrsion is an inference that p rccedsby ;ntcrctnnglngttu rub.i* and pwnitatz ums

ofthe proposition. T.{o men are angels .oo,*a o;NJangels are men," and these
Propositions maybe ulidlyinferred from one another. Similar$ "Some women are
write$" and "Some writen are women" are logically equirralent, and by conrenion
either can be ralidly inferred from the other. Convenion is perfectlyralid for all E
pmpositions and for all I propositions. One sandard-form categorical proposition
u therefore said to be the corr,wre of another, when we derive it
simply inter-

cle$ng

the subjea and pradicate teuns of thatodrer propooition The propm

uon from wtrich it is defi,ed is called the coruntmd" Thts "l.Io idedists are
politicians" is the converse of "llo politicians are idealisn " which is is conrrcrtend.

Converslon
An inference formed
by intercharyjng
Sre subiect and
prcdrcah termsda
categorkal proposition. Not all corwersions are valid.

192

Chapter

Further Immediate Infercnces

The conversion ofan O proposition is not, in general, valid. The O propo

sition, "Some animals are not dogs" is plainly true; is converse is the proposi
tion "Some dogs are not animals," which is plainly false. An 0 proposition and
its converse are not, in general, logically equivalent.
The A proposition presents a sPecial problem here. Of course the converse
of an A proposition does not in general follow from its convertend. From "All
dogs are animals" we certainly may not infer that "All animals are dogs.'
Traditional logic recognized this, of coume, but asserted, nevertheless' that
something lile conversion was ralid for A propositions- On the traditjonal
dogs
ftuare of Opposition, one could validly infer from the A proposition' LAI]
ari animals," is subaltern I propositioh, "Some dogs are animals." The A
proposition sap something about all members of the subject clas (dogs)' the !
propos,tion makes a more limited daim, about only some of the members of
rhai class. In general, it was held that one could infer 'Some S is P" from "All S
is P." And, as we saw above, an I proposition may be converted ralidly; if some
dogs are animal$, then some animals are dogs.

So, if we are given the A proposition that "All dogs are animals," we fint
infer "Some dogs are animals" by subalternation, and from that subaltem we
can by conversion validly infer that'Some animals are dogs." Hence, by a combination of subalternation and conversion, we advance validly from "AIl S is P'

to 'Some P is S." This pattern of inference, called conversion by limitation (or

snatniorr po anldms) proceeds Uy intercUanging subject and prcdicate tems
and changing the quantity of the proposition from rmiversal to particular. This
type of convenion will be considered further in the next section.
In all conversions, the ccnverse of a given proposition contains exacdy thc
same sqhject anti predicate'Irns as ihe convertend, their order being rerrered,
urd alwala has the same quality (of amrmation or denial). A complete plcture -of
this irrmediate inferer.. rt tt"ditiorrlbundentoodis given by the following able:

Valld ConYerions

Corrmrmro
P.
It No Sis P.
I: Somc Sis P.
0: Some Sis notP.

A All Sis

Coxrryns
I: Some Pis S (by limitation)
E: No Pis S
I Some Pis S
(conversion

notralid)

the collection of all things that hale the characteristic of being human; its dass
defining characteristic is the anribute of being human. The classdefining characteristic need not be a "simple" attribute; any attribute may determine a class. For
example, the complex attribute of being left-handed and red-headed'and a snrdent
determines a class-the class of all left-handed, red-headed students.
Every class has, associated with it, a complementarv class, or compbntnt,
which is the collection ofall things that do not belong to rhe original class. The
'complement of the class of all people is the class of all things thar are not
WU
ple. The classdefining characteristic of that complementary clas is the {negagve) attribute of not being a person. The complemenr of thc class of all people
contains no people, but it contains everything else: shoes and ships and sealing
wax and cabbages--but no kingt, since kings are people. It is often conlrnient
to speak of the complement of the class of all persons as the 'class of all non-

of the class designated by rhe rerrn S is then designated by the term roni-i we mav speak ol' the term non-S as being the
complement of the term S.
The word compbmentis thus used in two senses. In one sense it is the complement of a class, in the other it is the complement of a rerm. These are dil
ferent but very closelv connected. One ternr is the (term) complement of another
just in case the lirst term designates the (class) complement of the class desigpersons" The complement

nated by the second term.

Note that a class is the (class) complement of its own complemenr Likewise,
is. the (term) complement of its own complement. A son of "double negative'rule is involved here, to avoid strings of "non's" prefixed to a rcrm. Thus, the
complement of the term \oter" is "nonvoter;" but the complement of "nonvoter"
should be uritten simply as "votef rather than as "nonnonvoter."
One must be careful not to mistake contrary terms for complemenary
ttrms. "Cowardn and "hero" are conEaries, for no penon can be both a coward
'and a hero. But we must not identiry "cowards"
wi'th "nonheroes" because not
eleryone, and certainly not everything, need be one or the other. Likewise, the
complement of the term 'winner" is not "loser" but "nonwinner," for althougtr
Dot everything, or even everyone, is either a winner or a lose6 absolutely eveqr
thing is either a winner or a nonwinner,
a

term

Obyerslon

Obversion is an immediate inference easy to explain once the concept of a

term complement is understood. To obvert a proposition we change its quality
(affinnative to negative or negative to affrrmative) and replace the predicate
hrm with its complement But the subject term remains unchanged, and so

B. (lasser and Class ComPtements

(omplenrmt
The collection

.*pl"io odrer typqof immediate infererice r+e must exmdne more dmely ihe
concep of a "chs;and erplain wlnt b meant by the qlcmec of a das''{otr

To

of a Clars

ofd

193

Categorical Propositions

thing that do not

"lass,

belong to thatdass.

that

dl otiects haling a cerain common arritnrE

b
refer O,4s the "classffning ciraracterielic." the class of all hurnans

we have sai4 is the collection of

r,ve may

$metimes we reason using what is called ttre relitive complement of a chs, its comple
&ent within some other elass. For example, wi&in tlre class of "children of minc' &ere is a
ohlas, "daughiers of mine," whose relltive complement is another subclass, "children of
llline rtfio are not daughtem' or "sons of mine.' But obversions, and other immediarc infererlces, rely on the absolute complement ofclasses, as defined abore.

Obverslon
An inference formed
by changing the
quality of a proposi-

tion and replacing

the predicate tem
by its complement.
Obversion isvalid
for arry standardform categorical
proposition.

194

Furtherlmmediatelnferences 195

Chapter6 Caqwical ProPoitions

doesthequantityof&epropositionbeingobverted.Forexample,theApropo
"Ali residents are-vo;rs" has as its obverse the E proposition "No resisition

and either
dents are nonvoters," These truo are logically equiyalent propositions

other'

ralidly inferred from the

maybe
-Obversion
is a r"alid immediate inference when applie$

to an, standard-

form categorical ProPosition:

r
r
r

The E proposition "No umpires are partisans" has as is obverse the logical
ly equitralent proposition "All umpires are nonPartisans'"
The I proposition 'some metals are conductors" has as its obverse the

proposition "Some metals are not nonconductors'"

The o proposition'some nationswere not hlli8erents" has as its obverse
*re I proposirion 'Some nations were nonbelligerents'"

Theproposition serving as premise for the obversion is called rhe &atrtnLthe

conclusiou of the infereln.. is c alled the ofume"Every standard-form categorical
form of
proposition is logically equivalent to its obverse, so obversion is a valid
obtain
To
propositions.
categorical
sandard-iorm
all
for
immediate inferince
the obverse ofany proposition we learre the quantity (universal or particular)
and
and the subject t..m,,n.h"',g"d; we change the quality ofthe proposition
a
gives
table
following
The
with
complement.
by
term
replace thJpredicate

complete picture of all valid opversions:

Obuerlons
OxvBsE

OBrBrtrp

ENoSisnon-P.

,A.-A!8isP.

&NoSisP.

rtAIl

ESome'tisP.

0: Some Sis not non'P.

O: Somo Sisnoif..

I: Some Sis non-P.

Sisnon-P.

\bnitatim-andthis

a propositionwidt

the comphment of
its predicate teflrt
and replacing the
prcdicateterm by
its

equiralent propositions, as will be evident upon reflection. Contraposition is

plainly a ralid form of immediate inference when applied to A propositions. It
really introduces nothing neq because we can get from any A proposition to is
' contrapositive by first obverting it, next appllng contersion, and then applyrng
obversion again. Beginning with "All Sis P," we obvert it to olrtain "No Sis nonP," which converts validly to "No noh-Pis S," whose obverse is "All non-Pis nonS," The coatrapositive of any A proposition is the obverse of the converse of the
ohierse of that proposition.
Contraposition is a lalid form of immediate inference when applied to O
propositions also, although is conclusion may be awkward ro express. The conrapositive of the O proposition "Some studehts are not idealists' is the somewhat cumbersome O proposition "Some nonidealists are not nonstuden8,"
which is Iogically equirralent to its premise. This also can be shown to be the
outcome of first obverting, then converting. then obverting again. "Some S is
not P" obverts to "Some Sis non-P" which converts to "Some non-Pis S' which
obverts to "Some non-Pis not non-S'
For I propositions, howerae4 contraposition is not, in general, a ralid form
ofinference. The tnre I proposition "Some citizens are nonlegislators" has as its
contrapositive the false proposition "Some legislaton are noncitizens." The reason for this inr"alidity becomes evident when we try to derive the contmpositive
ofthe I proposition by sr:ccessively obverting, converting, and obverting. The
obverse of the onginal I proposition "Some ,Sis P" is the O proposition "Some
S is not non-P." b:rt (as rve saw earli'er) rhe converse of an O propositign does
not generally foliow va!id!;'frnm it
In the case of E propositions, the contrapositive does not follow ralidly from the
original, as can be seen when, if we begin with the tnre proposition "No wrestlers
are weaklings,'we would get, as is conrmpositive, the obviously hlse propodtion
t'lo nom,rrcakiings are nonwrestlen." The reason for this irmlidity we will see,
egain, ifrre attempt to derirr it
S successire ohersion, conrrersion, and obrrersion
begin with the E propoeition, lrlo Sis P" andobvert i', $ obtain theApropo
{we
sitbn "All Sis non-P"-which in general cannot be lidly- corrveftd *rqtti h;Itu
,ion Ifwe do then converr it by limiation to obtain 'Some non-Pis
S" we can ohert
.lhb to obtarn "Some non-Pis nol non-S" This ourcome r,e may call trc mtrapitite

(ontraPosltloil
An inference
fonned by replacfuq
the iubiectterm

the comphment

For example, the connapositive of the A oroposition ".All members are votA proposition "All nonvoters are nonmembers." These are logically

ers" is the

$rbiectteilL

Not a[ confapositions arevdkl.

D. Contraposltlon

fint
A third type of immediate inference, contraposition, can be reduced to the
tro,.on inion and obvenion. To form the contrapositive of a given proposition,
we re
we replace its subject term with the complement of is predicate term, and
the
Neither
term'
of
cornplement
plarits predicaic trrn with the
its lubjecl
so^the
is
ctnnged,
propoo'ition
q,ratityoortt. q.r-rity of the originat
p-p"Stlon is an A proposition, the contrapositive of an O propo*
ti*

clnra{

i* I

tion

is an

O proposition, and

so

forth-

too we will consider further in the next section.

Contraposition by limitation, in which we infer an O proposition from an E
proposition (for example, we infer "Some non-Pis not non-S" from "No Sis P"),
Ilu tle same peculiarity as conversion by li;;ritation, on which it depends.
Decause a particular proposition is inf'erred
from a universal proposition, the
tesulting contrapositive cannot have the same
meaning, and cannot be logicdly
ftluivalent to the proposition that was the original premise. On thi other hand,
Ille contrapositive of an A proposition is an
A proposition, and the contraposiutt ofan O proposition is an 0 proposition, and in each ofthese cases.the contrapositive an{ the premise from which it is derived
are equirralenl

196

Chapter

Further lmmediate Inferences

Categorical hopoaitions

Contraposition is thus seen to be ralid only when applied to A and 0 propo

sitions. it is not valid at all for I propositions, and is ralid for E propositions only
by limitation. The complete picture is exhibiteci in the following table:

197

lmmedlate lnferences:
Conversion, Obversion, Contraposition
Cotvyrniloal

'

Contraporitlon

PTEtraSE
A Atl Sis P.
F- No Sis P
I: Some Sis P.
O: Some Sis

Couwnrrxo.

not-P.

Commse

P.
L
No.lisP.
E
SomeSisP. I
Some S is not PAll Sis

C.ourr"enosrrrvu

A All non-Pis non-S

0: Some non-Pis not non-S. (by limitation)
(contraposition notvalid)
O: Some non-Pis not non-S.

"All animals are cats,' the tru proposition "Some animals are cats" follows by
subalternation. And from the false proposition "All parents are students," conversioo by limitation yields the true proposition 'Some students are parents.'
Therefore, if a propcsitiun is given to be hlse, and the question is raised aboqt
the truth orfalsehood of some Alw,relatd,proposition, the recommendid pro
cedure is to begin drawing immddiate inferences from either (a) the contradictory of the proposition known to be false, or (b) from the problematic
proposition iself. For the contradictory of a false proposition must be true, and
all ralid inferences from that will also be uue propositions. And if we follow the
other coune and are able to show that the problematic proposition implies the
proposition that is given to be false, we know that it must iself be false. Here fo}
lolvs a able in which the forms of immediate inference: conversion, obversion,
and.contraposition, are fully displayed:

Orrmsr

A All Sis P.

E:

I: Some Sis P.

o:

Some Sis notP.

No Sis non-P.

A All Sis non-P.

0: SomeSisnotnon-P.
I: Some Sis non-P.

Cowrneposmoiv

problematic proposition---or its contradictory or contrary-follow validly

from the one given as true? To answer we proceed as follows: From what we
are given, 'Ail surgeons are physicians," we can validly infer its contrapositive,
"All nonphpicians are nonsurgeons." From this, using conversion by limiation (valid according to the traditional view) we can derive "Some ncnsurgeons are nonphysicians." But this is the contradictorl of the proposition in
question ("No nonsurgeons are nonphpicians"), which is thus no longer

from the faise premise

OswK-mND
E: No Sis P.

Questions about the relations benreen propositions can often be an-

whose premises are false carharre a true conclusion. Thus,

NoPisS
SomePisS.
(conrcrsion not valid)

Owr*flau

swered by exploring the various immediate inferences that can be drawn

from one or the other of them. For example: given that the proposition iAIl
surgeons are phpicians" is true, what can we know about the truth or falsehood of the proposition "No nonsurgeons are nonphysicians?" Does this

problematic but known to be false.

' tn the rrcr1 first chapter of *ris book we noted that a valid argument whose
premises are true rlrus, have a true conclusion, but also that a wlid argument

SomePis S(bytimitation)

htrflsE

Corvrrposrrnry,

& All non-Pis non-S.

0: Sorne non-Pis not non-S. (by limitation)

E:

All Sis,i?.
NoSisP.

O:

Some Sis

i:

Some Sis P.

notP.

O:

{contraposition not rlalid)

Some non-Pis not non-S

H
ExuRcrsrs
L state the converses

of the forlowing propositions, and indicate which of

,them are equivalent to the given p.opoiltionr.

' *1.

'

2.
3'
4.
*5.

No people who are coruiderate ofothers are reckless drivers who

pay no attention to traffic regulations.

All graduates ofWestPoint are commissioned officers in the U.S. Army.

SomeEuropeancarsareoverpricedandunderpoweredautornobiles.
No reptiles are warrrblooded animals.
Some professional wrestlers are elderly persons who are incapable
of doing an honest day's work.

kertlset lvalhblc
oo.N.ogk CD-XOM

tecflon 6.6, l:
t-,2, r,4

198 Chapter6
frcrd*rava$rbh IL
on elogic CILf,Oil

7,2,r,1

Extr,citas.vdhHc
on clogk CD-f,ot|

.*5.

Some clergyare notabstainers.

No geniuses are conformists.

Atl objects suitable for boat anchors are objects teighing at least
fifteen pounds.

State the contraPositives of rhe

fo$odng propositions and indicate

*1.

r,3,1

2.

Some soldiers are not officers.

3.

All scholars are nohdegenerats.

4.

All things weighing

less than 50 pounds are

Erertlrer avallabh
on cfoglc

(D-noll

*10.

VL

ry.

sctloo 5'6 lV:

1,Z,r,t

n5.

Some noncitizens are not nonresidents.

2. No socialists are nonPacifists

3. All nonsocialists are nonpacifiss'
4. No nonpacifists are sccialists'
*5. No nonsocialistsarenonpacifists
6. All nonpacifists are norxocialbrs
7. NopacifistsarenonsocialisG.
8. Some socialists are notPacifists
L All pacifistsaresocialists.
*10.

Some nonpacifists are socialists.

on efoglc CD-IOf,l
Sedlon 6.6,

t,zr,1

ll!

v.

scientists are philosophens" is true, what may be inferred about

or falsehood of the foilowing propositions? That is, which
truth
'he
could be known to be true, which known to be false; and which would

If "No

be undetermined?

*1, No nonphilosophenare scienti$s"

L Some nonphilosophers art nrt nonscientists.
3. AII nonscientists are nonphihophers.

tcctlon6.4 Yl:

1,2,r,4

All martyrs were nonsaints.

l'10.

Some martyrs were rionsaints.

l.

some sarnLs were, not martlrs,

12. Some martyrs were notsaints.

13. No saints were nonmartyrs.
14. No n6nsainb were marty{s.

itr.

Some martlrs rrere trot Donsaints.

I'II. If "some merchants are not pirates"

is true, what may be inferred about

the truth or falsehood of the follouing propositions? That is, nihich
could be known to be true, which known to be false, and which would
be undetermined?

Excrrlt$fldLb0.

oodogk(IROil

6. Some nonmartyrs were sains.

7. Some saints were not nonmartyrs,
8. No martyrswere saints.
9. Some nonsaints Were martyrs.

is true, nhat may be inferrtd abotrt the tmdr

or frlsehood of the following proposition$ That is, which could be knom
to be true, r*rich known to be false, audwhich r"ould be undetermined?

Some nonpacifisB are not lroos@ialisb.

trccdrcs avdl$le

2. All saints were nonmartyrs.

3. Some martyrs were saints.
4. No saints were marryrs.

If "All socialise are pacifists"

r'1.

No philosophers are nonscientists.

If "Some saints were martyrs" is true, what may be infeued about tlre truth
or filsehood of the folloring propositions? That is, wtrich could be knonn
to be true, which known to be fal*e, and which would be undetermind?

objecs not more than

four feet high.

*5.

No scientisrc are nonphilosophen.

No nonscientists are nonphilosophers.

. *1. All saints were martyni.

All joumalists are pcssimiss

199

6. All philosophersare.scientists.
7. Some nonphilosophers are scientists.
8. All nonphilosophen are nonscientjsts.
9. Some rientiss are not philosophers.

Some college athletes are Professionals.

No organic compounds are metals.

which of them are equiralent to the giren proporitions.

tcctlor6.4 lll:
1,

4.

State the ob-r,erses of the following propositions.

*1.
2.
3.
4.
*5.
.

tcctlon 5.6, tl:

Furtlrcr Irnacdiatc hfcrcnces

ProPocitions

1. No pirates are merchants.

2, No merchants are nonpirates.
3. Some merchants are nonpirates.
4. All nonmerchants are pirates.

+5.

6.
7.

Some nonmerchanB are nonpirates.

All merchants are pirates.

No nonrnerchants are pirates.

IrcrtbcryrffSlc

ca dogrk (D-nOftI
Jccthn 6.6, Vll:

7,2"r,1

2{)0

Ctrapter

Carcgorical Propositions
Existential Import and the Interpretation of Categorical propositions

8.

No pirates are nonmerchants.

9.

All nonpirates are nonmerchants.

+10.

Sorne nonpirates are not nonmerchanB.

A*ttotle ,. Boole on rntGrP'rfig crteqrorrcnr

pro@n!.

lt.

Some nonpirates are.merchan8.

12.

No nonpirates are merchants.

13.

Some pirates are' merchans.

t4.

No merchants are nonpirates.

inteqpreadons of categorical propositions: *re

T.
Aristotelian, u'hich is naditional, and die BJolean,
*hi.t l, *od._.
In the interpreratjon of the ancient Greek philosopher Arttotle,

No merchans are pirarcs.

,ft. Te.gl"-u**nal proposition ( All

*15.

*o.

ll*

ly

,t

leprnchauns ru-, ti
go.r,
hats,' or 'Nofrqgs are poisonous') implies the truth
of is correspor

prf.rt* pnposition
Fq
has'or 'Some frogs

('Some Lprecharms

r+,ear

lirle green

are not poisonoru'j.

In contrast, George Boole, a lgth century English mathematician, argued that we cannor infer rhe truth oithe
particular proposition from the truth ofits corresponding universal
proposition, because (as both sides
agree) every particular proposition

6.7 Exisientiai import and-the lnterpretation

of Categorical Propositions
Categorical propositions are the building blocks

of arguments, and our

aim

throughout is to analyze and evaluate arguments. To do this we must be able to di

agram and symbolize the A E, I, and O propositions. But before we can do that
we must confiont and resolve a deep logical problem-one that has been a
source of controversy for literally thousands of years. In this section we explain
this problem, and we will provide a resolution upon which a coherent analysis of
q,llogisms may be developed.

The issues here,

as we

ical propositions-named after George Boole (1815-1864), an

tion may be safely bypassed.

To understand the problem, and the Boolean outcome with which we emerge,

The modern

interpretation of
categorical
propositions, in
which universal

propsitions (A
f) are not
asi;urned to rdfer
to dasss that
and

have nremben.

if

the corresponding particular piopositioir,

rechauns wear llttle green hats'*o,ria
some leprechauns do, and that woutrt

th.n ,All lep

p.r it us to infer that

imply that there really

in the modern or Boolean intrryrretation, a universal

propoi.
(rn A or an E oroposirion) **ib.
r.rnderstood to assert onlv
{on
such a thingas a leprechaun, it wean a fittfe
So,

*l and
-1P.:::,*
'If tliere is such
; hat,'

"

tt ing

tog, it i"not polsono*.,

8re;

English

mathematician whose contributions to logical theory played a key role in

the later development of the modern compurer. So if the outcome of thc
following discussion--+ummarized in the final nro paragraphs of this section on pages 204-20F-is fully grasped, the intervening pages of this sec

lntctprctrtlon

asserts the existence of its subject

some frogs are not poisonous there must be
at least
one frog. But if the universil proposition
permits us to infer
class;

are leprechauns!

shall see, are far from simple. But the analysis of

syllogisms in succeeding chapters does not require that the complications of
this controversy be mastered. It does require that the interpretation of categorical propositions that emerges frbm the resolution of the controversy be
understood. This is commonly called the Boolean interpretation of categor-

godeaa

201

it must be seen that some propositions have existentiai import; and some do
not. A proposition is said to have evistential import if it typically b
uttered to assert the existence of objects of some kind. Why should this seern
ingly abstruse matter be of concem to the student of logic? Because the correctness of the reasoning in many arguments is directly affected by whether the
propositions of which those arguments are built do, or do not, have existential

o propositions, which surery do have existentiat imporr Thus

H:
I 'titme
soldiers are heroes" sap that there exists at least
one sol_
;:-.
qcrwrlo ts a hero.
And the O plopcytion, "Some dogs are not companions"
sals

I* l* td
l.f*lon

dBt ttrere exists

at reast one dog that is

a rr...t

,ot a companion. particular propositions,

f*3,f:^ry.::lllTnry

"iLtone

member.6

import We must arrive at

a clear and consistent interpretation of categoricar

propositions in order to determine with con{idence what may be righdy ilr
ferred from thbm, and to guard against incorrect inferences that are sometimf,s
drawn from them.
':

that the ols.s aesisnated by their subject

Erlns (Ior example,
soldiers and dogs) are nor empty_*re ilas
of ,oldi.n, ardtre
ctas of dogs (if
the examples given iere are t r.l,'frl

liffi ffi r,:l::l*x'tr$-.

fftho*';**::i'..l*l*q;i1iili,'x,x':"jffi
,:l:, uvt w.u.uc,.uur{ p cemnry rue that therr
contain or imply
impi,
udil[f,ll.il5T:
^:::.
;::lill::]:Jl'y T..m?t their writings
jilfi"ffi Hl..T',1fl ii#.l',.*Hfir.J#rffiffIffi ..XTil::Hffi l[lffi

Existcntial Importand the Interpretation of Cate6orical

N2

Chapter

Catcgodcal Propositions

"r"

rpU."u..eight-leggedanimals.Andsimilarly,wesaidthatanOproposition
6lnws rafiafifrom its corresponding E proposition' But if I and O proposi-

corresponding
tions have existential impoG and they follow ralidly from their

e-arpropositions,thenAandEpropositionsmust'allohaveexistentialirr

p"" U;;

with exisLntial impon could not be derived ralidly

"proposition
Lom another that did not have such importJ

and O
This consequenci creates a very serious problem' 1{e know thatA
are contradictories. "All
ofopposition,
square
traditional
the
on
orooositions,
b-'.r.p"rf fngfish" is contradicted by "Some Danes do not speak lnglish:
must be false, nor
Contradictories lannot both be true, since one of the pair
y'corresponding
But
true'
be
mus[
pair.
of
the
one
since
fatse
be
U",f,
tft.y
in the para
concluded
w-e
a and o propositions do have existential imPort, as
TheA
To
illustrate:
false!
coul.ilbe
then both contradictories
g."pt J*i
"b.*,
"eU inhabitants of Man are blond" and its corresponding O propa
i.opotitio"
iiti* "So*. iirhabitants of Mars are not blond" are contradictories; if they have
&at there
existential import-that is, if we were to interpret them as aserting

.*

czinhabiansofMars_thenbo&thesepropositionsarefalseifManhasno

of
a, of course, we know ttrat tr4ars has no inhabitants; the clas
just
grven
example
in
the
in inhabiants is empty, so both of the propositions
are false. But if they could both be false , they conld nnt bc contradictorics! .
of
Somethingseerns to have gonewrongwirh rheraditiona Serul
Qnrytuol
us thatA and
in cases ort}likind. If the traditional square is correctwhen it tells
e:
[fropotitiont t"my implv their corresponding I and O propositions' e".".

rJrU"r[,n

it not correci whin it tells us thlt corresponding A anti O propositions

in holc
ai.. contradictories. And in that case, the square would also be misaken
are
subcontraries'
O
propositions
ing that the corresponding I aod
S$rare of Opposition be
What is to be done? C,an the t
yes, it can, but the price would be "Atio"af
high. we could rehabiliate the traditionat
'Opp*ition
by introducing ihe notion of a prauffosiri'on'Much.ea{
"f
(in
5.5) thaisome complex questions are properly alr
section
.r, *" obu.*.a
answer to a prior question has been-presuppo{
ifthe
or
"no'only
"yes"
smered
"no
Tid,vou spend the money you stole?" can be reasonably answered "yes" or
rev
to
Now'
granted'
be
money
some
you
stole
p.oupposition that

*rrri

to"'{

d*.
o.rry

if *r"

,*;;;;
f* m*

fo!
to show that the existential import of A and Elropositions must
of
Opposid'*
the
traditiond
on
O
f
ani
propositionq
Quare
,n" of
(railitionally assumed) mlidity 0I.
of th" ,l p-pouiri*, rc ib"ta rnow it g relying on lte
;f ,ir; f prtposition, rve co,ld show it by relying on dr
*"*rS6" &-mation; in ,h.
the sarr
validitv of conuaposiionf limitation. The result is alw'4,s
if
I
and O propositions
Opposirion,
rhe-traditionaisquare
On
rU.,.,
o ,t r.ofi.a
"f
",
;p"rq A and E proPositioru must have existential import also'

I",*:-Y

Gffi;rt"",*ed)
;r. ;;drl

of Opposition, we might insist that a/propositions-that is, the

four staodard-form categorical propositionsd E, I, and G-presuppose (in the
rnse indicated above) that the classes to which'they refer do have members
are not empty. That is, questions abbut the truth or falsehood of propositions,
and about the logical relations holding among them, are admissible and may be
rqasonably answered (on this interpreation) only if we presuppose that they
never refer to empty classes. In this rray, we may save all of the relationship set
forth in the traditional Square of Opposition: A and E will remain contraries,
and O will remain subcontraries, subalterns ryill follow ralidly from their superalternq and A and 0 will remain contridictories, as will I and E. To achieve this
result, hovreve4 we must pay by making the blanket presupposition that all classes designated by our rcrms (and the complements of these dasses) do have
membcrs, are not empt)r.s
Wcll, why not dojust that? This existential presupposition is both neces.sary
artd sufficient to rescue Aristotelian logic. It is, moreorreq a presupposirion in
frrll accord with the ordinary use of modern languager like English in very many
cases. Ifyou are told, "AIl the apples in the barrel are Delicious," and {ind when
you look into the barrel that it is empty, what would you say? You would probably not say that the claim was false, or true, but, would instead point out that
drere az no apples in the barrel. You would thus be explaining that the speaker
had made a mistake, that in this case the existential presupposition (that there
exist appies in the barrel) was faise. And the fact that we would respond in this
corrective fashion shows thatwe do undentand, and do generally accept, the
existential presupposition of propositions ordinarily uttereti.
Unfornrnately, howeve4 this blanket existential presupposition, introduced to
rescue the traditional Square of Opposition, imposes intellecnral penalties too
heavyto bear.There areverygoodreasons notto do it Here are three zuch reasons.
First, this rescue
t}e traditional relations among A E, I, and O
propositions, but only at the .cost of reducing their power to formulate assercue the Square

(as no one
But if this is so, if I and o propositions have existential import
from the
arises
The
problem
the
problem?
lies
wherein
deny)
would wish to
we said that an I propo
Earlier
very
awkward'
are
which
iact,
this
of
wxcrymca
,iirri foUo* validly from is corresponding A proposition by subalternation'
eight-tegged animals," we infer validly that some
rn t ir, from "All ,pid.r.

p."r.ri,

tions that we may need to formulate.

designated has members, w uitl rwucr

If we iovariably presuppose that the class

bc abb n fmnulole tta yopwition thaf dmi6

that it Las wnbs.rAnd such denials may sometimes be rrery important and must
zurely be made intelligible.
Second, even ordinary usage of language is not in complete accord with thls
.
ttlanket presupposidoa. &nAilna uhat we say dus rwt wppse that there arc naths

in tlu classtsw me tatking abrrltlltyou say, for example, "AIi trespassen will be pro*
ecuted," far from presupposing that the class of trespassers has membe rg you will

ordinuiiy be intending to ensure that the clas will become and remain empty!

ct'ass

It

Wi"b. urgues rhat fubtotelian logic does not require the assumption that rhe
designared by the complement of the subject term be nonempty. See nExistential

'Phitlip

+sumpdons for fuistotelian Logic,'Jounrd, of Prtilowphicei R*rnch 16 (199G-1991): 321-28:

rutAristotclian logic cerrainly does require the assumption that at least the classes designatby the other three terms (dre zubject term, the predicate term, and the complement of
$
roe
PrcdicatE term) are not empty-and &is exisrcntial assumption gives rise to all the di&
cUties noted in the remarls o follow.

Propositioru 203

204

Chapter

Third, in science, and in other theoretical spheres, w olm wish b nason with'
nohing any prauppsAions Aottt eittcrce Newton's First Law of Mo$on, for example, asera that certain thingp are true about bodies that are not acted on by any

sections

extemal forces: that theyperserrcre in res! orintheirstraight-line motion. Thc law

may be true; a phpicist may wish to expres and defend it without wanting to presuppo$e that there actually are any bodies that are not acted on by external forces.
Objections of &is kind make the blanket existential presupposition unacceptable for modern logicians. The fuistotelian.interpreation of categorical
p,ropositions long thought to !9 correct must be abandoned, and a more mod'

6. In like

and 11.3. For thepresentitwill sumce to

say that,

in the Boolean

manner,

in the Boolean interpretation, corresponding

and

7.

ir

corresponding

D--ir za

geftnal\ ?6rA This is because, plainly, one may not Elidly infer a proposition
that has exi$ential import from one that does nol

8.

The Bootrean inrcrpretation pazrenmsome immediarc inferences: wan:noafor

E andfml pnposifioru is preserved; antrapositionforA annfor0 fnfusitiws
is,

presrl*fi ofunior fo

arry

@ionis

prcrr,rvd"But coilrsion ryfimiadon,

and conn-aposition by limitation, are not generallyralid.

9. The traditional

also remains true in this intelpretation th* tlu whtmal prxpositials, A

andEn ue tlu antrdictoia of tlu partiafur papuitiona O adl. That is, the
proposition 'All men art mottal' does contradict the proposition "Some
men are not mortal," and the proposition "No gods are rnortal" does contradict the proposition "Some gods are morta!."

In short, the blanket existential presupposition is rejected by modern logicians.

It is a mistake, we hold, to assuncthataclass has members if it is not asserted explicitly that it does. Any argument that relies on this mistaken assumption is said

'

It

pnpwitions arc

in@

as houing

tln

inporL Even when the

Square of Opposition,

in the

Boolean interpretation, is

tr-ansformed in the following general way: Rclations alangtlw siilaoJtlu sryan

ue undmtt, but tlu d,iagonal contradittory nlatian^s rmain infmce.

to commit the fallaq of existential assumption, or more briefly, the existmtial

fallacy. Wth this Boolean interpretation dearly in mind, we are now in a pos!

tion to set forth a powerfiri s;ntem

standard-form categorical syllogisms.

Bdtan intaprtation, unhtmal

w aistailial

whkh the

argurnent relies on
the illegitlmate assumption that a dass
has memben when
there is no expticit
asertion that it does.

In the Boolean interpreation , rubaltmation-:rnfening an I propositioir from

is corresponding A and an O proposition from

In some respects, the traditional interpretation is not upsel I anl O Wsitiorts

cvntinlu to hauc existntial inporl on tlu futzon hter@ation, so the proposition "Sbme Sis P" is false if the class Sis empty, and the proposition "Some
S is not-P" is likervise false if the clas Sis empty.

3. All this is entirely coherent because, h

f,xlstentlal Fallacy
A fallacy in

propositions, because they do have existential imporg can both be false if the
subject class is empty. So cffiesponllngt arulO pnpsilirrru arc nd s.ubcontrarb.

ern inrcrpretation emplopd.

In modern logic it is not assumed &at 0re clases to which categorical
propositions refer alwap hara members. The modern interpretation explicitlv
rejecting this assumption is called, as we no'ted earlier, Boolean.e We adopt the
Boolean interpretation ofrategorical propositions in all ttrat follows. This has
important logical consequences. Therefore we set forth now what this Boolean
interpretation of categorical propositions entails:

2.

ll.2

interpretation, "All unicorns have wings" is taken to assert that "If there is a
unicorn, *ren it has wings" and'No unicorns have wings" is,taken to assert
that "If there is a unicom, it does not have wings." And both of these "if. . .then'
propositions can be true if indeed there are no unicoms.

a*

l.

205

f,xisrcntial lmport and the Interpretation of Catcgorieal Pro.positions

Categorical Propositions

clas

therefore, the proposition "AIl .lii.P' can be true, as can the propc'
sition "No Sis P" For examplq the propciriors "All uaicorns hale horns'
and "No unicorns harc Hing$" may both be noe errn if there are no uru'
corns" But if thert are no unicoms, thelpropcition "Sorae unicorns hart
horns" is fahe, as is the O proposition "Some tmicornsdo not ha/e wings'"

ior rhe qanboiizing and diagramming oi

is empty,

4.

5.

Sometimes, in ordinary discorrse, urc utter a universal proposition with

which we do intend to asert existen cc. Tht Bcoimt htcrfrarion pefi.ilt kis tt
Ds aUDrasa4 of course, but doing so requires two propositiort, on" existential
in force but particula4 the other universal but not e'ristential in force.

imporhnt changes result from our adoption of the Boolean inter'

preadon. Corftsponding A a'dV pnpuitions can both b tnu and are tlwrelon not
contraia. This may seem paradoxical and will be explained in detail later, io
Some very

Russell, another of the famders of modern symbolic logh, also advanced this aP
proach in e hmous essay enritlcd "The Existenthl lmport of Propcitiolts"-in M;rt4 July
i905, and tlrere refcrred o it as 'Peano's interprcation' of pmpcitiors, after Guis.Fpe
Peano, a great Italian mathemariciao of ttre early trcntieth cenmry

hertrand

r-_
ExeRclsEs
'

Imrtlcclu[&h
G

In the preceding discussion of existential import, it was shqun why, in &e

Boolean interpretation of propqsitiorx adopted in this boo( some inferences
that traditionally were thought to be ralid, mistakenly assume that cerrain classes
hare members; these inferences commit the existential fallacy and are notmlid.
In each of the following arguments, rhis erdstential falhcy is committed; explain the
Point at which, in each argumeng the misaken existential assumption is made.
I

-.'_

_'- -*

Exeuprr:
[
(l) No mathematician is one who has squared the circle.
tl*tfrr" (2) No one who has quared the circle is a ma&ematician;
Uryfou (3) AII who have squared the circle are nonnratherraticiars;
thaefore, (4) Some nonmathematician is one who has squared
the circle.

GtogtlcCDdOit

ftcdoo6.7, !, t,

u,tv,v

zffi

Symbolism and Diagrams for Categorical

Chapar 6 Categoricat Propositions

Mutian

this point is conveT':i by

Step (3) to srcp (4) is invalid. The inference at
acceptable in
limitation (thai is, from d// S ar Pto SaeP & S)' which was
interpretation'
Boolean
the
in
the traditional interpretation but is inrralid

a
relies on an inference from a univenal proposition 1o gartictl
that the classes in
shown
has
dirussion
preceding
the
but
tar proposition,
the
proposition cannot be assumed to have members' while
par
inr"alid
the
Thus
".ril.''.o"t
members'
have
do
classes in a p".ti*t., proposition
cla"q.in
predicate
the
that
inference
the
(4)
permit
(!)
would
to
og. no*
there issomeone who has squared the
{+i i, not.mpty, and thereiore that
one commits the existentid fallacv'
from

Thir

rrp

circle! In inferring {4)

t3),

To say that the class designated by Sdoes have members is to deny that Sis
empty. To assert that there are S's is to deny the proposition rymbolized.by,S=
sign'
0. We symbolize that denial by drawing a slanting line through the equals
Thus the inequality S + 0 sa)ts that there are S's, by denying that Sis empty.
Standard-form categorical propositions refer to nvo classeq so the equations
that represent them are somefvhat more complicated. Where each of tr,vo classes
is alreidy designated by a symbol, the class of all things that belong to both of
them can be representec byjrxtaposing the symbols for the two original dasses.
For example, if the letter Sdesignates the class of all satires and the letter Pdesigilars rhe cla.ss of all poemq then the clas of all thingp *nt are both satires and
poems is represented by the symbol SP, wtrich thus designates ttre clas of all satirL po"* (or poetic satires). The common part or common membership of ao
called the pmduct or intenection of the tno classes. The pmdzadtm
the class of all things &at b"Jong to bodl of &em. The product of the
clas of all Americans and the class of all composen is the class of all American
composers. (One must be on one's guard against certain oddities of the English
language here. For exampte, the product ofthe clas ofall Spaniards and the
clas ofall dancers is not the class ofall Spanish dancers, for a Spanish dancer is

classes is

classes is

tr

0)
therdore,

(2)

therefore,

(3)

l'lo citizen

one

is

wtp hasrrceeded in rccom$istring the imposible;

is
No one who has succeeded in accomplistring Sre impossible
a citizen;
All who have succeeded in accomplishing the impossible

are noncitizens;

thuefore, (4)

impossible are
Some who have succeeded in accomplishing the

noncitizem;

thuefore,

m.
thudore.

thuefue,

IV.

(5)

Some noncitizen is one who has succeeded in accomplishino

the imPossible.

0)

No acrobat

is

one vvho can lift himself

his own bootstmps;

(2)
(3)

0)

lt

ete

is bue

SEt l'lo uftome ae arimals found

in the Brwrx Zoo;

in the pronx Zoo;

tterctore, (2) lt is false that All unkoms are aoimals found
thqdqe (3) lt is true frat Sorre unicorm ae not animak found in the Bronx
Zoo. Crom-wtrUr ltfuilo,vs

ry.

0)

tlwelue

68

an acrobat
No one who can lift himself by his own bootstraps is
is not afi
bootst'-aps
his
own
by
himsefi
liit
can
who
Someone
being who can
bat. (From which itfollorys thatthere is at least one
lift himsetf bY his own bootSraPs')

(2\

It is fabe

flaL

Some

tut

nenni*t

here

exists at least one unicom')

are rr,embers

cd$ege sorotiti'es

It is kue thaL Some nrerma'rt are rrot members of college sorodties'

(From which it follour that there exists at hast one mermaid')

Symbolism and Diagrams for(ategorical Propotitiont

depends heavi[ on
$nce tlre Booiean interpreation of caegorical propositions
s1mtol".*P::T*
sne{
a
toha"e
the notion of an empty class, itis conveniint
designate<t oI
ir The zcro qrrnbol,b, it ,..a fot tftis purpose'Ibsay that the class
0' Thus the
and
s
.grrb ti8o between
tse term s has no -"*u"r", ,r. oirc
members'
no
Shas
that
or
equation S= 0 sala that there are no S!

not a dancer who is Spanish, but any person who performs Spanish dances.
Sii,nilarly, with abstract painters, English m4ion, antique dealen, and so on.)
Thii new notation permia us tJ symboiire E and I propositioru as equations
and inequalities. The E proposition 'No S is P" sap that no memben of the dass
,Sare members of the class P; that is, &ere are no things that irlong to bo*t clasg
es. This can be rephrxed by sayrng that the product of the twc classris is empty,
usome
S is P" sqr
which is symbolized by the equation SP= 0. The I ProPosition
ftat at least one memlrer oisis aiso a member of P. This means that the product

notemptyand issymbolized by the inequality SP*0.

To symbolize A and O propo$tions, it is convenient to introduce a nw
method of representing class co*plemerrts. The complement of a class is the
tollection or class ofall things that do not belong to the original dasq as cxof the classes Sand Pis

plained earlier in section 6.6. The complement of the cJes of all soldiers is the
clas of all things ttrat are not soldiers, the class of all nonsoldien. I{here the
leuer S symbolizes the class of all soldien, we symbolize the class of a1l nonsol
dien by 5 (read "S bar"), the symbol for the original class wi& a bar abore
Ihe A proposition "All S is P" sap that all members of the class S are also memben of the class P, that is, that there are no members of the class S that are not
members of P or (by obversion) that "No S is non-P." This, like any other E
proposition, says that the product of the classes designated by its subject and
Piedicate terms is ernpty. It is symbolized by the equation SP= 0' The O propo
sition "Some S is nct P" obverts to the Iogically eguivalent I proposition "Some

it

which is symbolized by the inequality SP* 0.

In their rymbolic formutrations, the interrelations among 'Jre four sundardform
categorical propositions appear very dearty. It is obrviors that the A and O
8is non-P,"

Pr"PT

tionsare contraaicorieswhen theyaresyrnbolizedas SP=0 and Ytt0, anditiseqn}

h obvious that the E and I propositionq SP= 0 and SP* 0 are contradictorie$ Tlre
nrUn Sryae o|Appositiott$aybe represented as strown in IiSnre &2 on page 210'

Propositions 207

208

Chapter

Symbolism and Diagrams for Categorical Propositions

Catqorical Propositions

The A propotitlon: All bananas are fruits'

The I proposltion: Some bananas are frults.

This A proposition asserts that auery member of the class of bananas (the subject class) is also a member of the class otfmits (the predicate class). when a
ierm refers to every member of a class we say that it is tlistributed.In au A
propocition the subject term is always disrributed. Bpt the A proposition does
irot-refer to every member Of the predicate ctass; this illustrative example does
not assrt that all fruits are bananas; it says nothing about every fruit. Ia an A

The word "some" in this I proposition tells us that at least one member of the
class designated by the subject term, "bananas," is also a member of Lhe class
designated by the predicate trm, "fruiu"-but this proposition makes no
claim about the subject class as a rvhole. Therefore, in this illustrative propo,
sition, as in every I proposition, the subject term is not distributeil. Nor doer
this proposition say anything about every member for the class of fruits (we
are only told that tlere is at least one member of the class of bananas in it) so
the predicate is not distributed either. In an I proposition neither the subject
term, aor tbe predicate term is distributed.

proposition tte predicate term is not distributed.

ti

Pndicak Cbss

Subject Cla.s

(h1li0

(Bananas)

fudicatz Class
(Fruit)

Subjxt Clnss
(Bananas)

All Sis P
Some Sis P

The

prcposltlon: No bananas are fruits.

This E proposiiion asserts that aual rnember cf the class of baaants'rs outsi.d.l
the class of fruits. The subject term, bananas, is plainly distributed. But since

are excluded from the dass of fruis, this illustrative proposition

refers tg every member of the ptedicatc class as well, because it plainly says
that zo fruit is a banana. In an E propo*tioo both the subject term aod the
predicete tera are distributed.
Note that the concept of distribution has nothing to do wi*r truth or falsity.
This sample proposition is certainly false'-bul as in every E proposition, both
b.ananas

of its terms are distributed.

The O proporitioft:'Somsbananas are not fnrits.

The word "some" again tells us that this proposition is not about all members
of the class of bananas; the su'bjed term is therefore not distributed. But since
we are told, in &is illustrative proposition, that some bananas.are notfiiuits,
we are told something about the entire predicate class-namely, that tlre entire class of fruits does not have one of those subject bananas among them.
In an O proposition the ptedicate term is distributed but the subject term is
not distributed.

hrdicdc

SuhjntCless

Clots

(F1uit)

(Banana)
No SisP

Subjed Clast

hedirau Aess
(Fruit)

(Bananas)

Some Sis

notP

209

?10

Ctrapter6 Crtegoricathopocitioas

Synbolism and Diagremsfor Categorical

rflE 8oo$Alrtq[rA*f oF oPPotmoN

sF=o

E: SP=0

HGURE 6.3

It represents the class g but says

nothing about ir. To diagram the propoeition rhat s has no members, or that
drere are no s's, we shade all of the interior of rhe circle representing s, indi?Pg i" ft1, *y that it contains nothing and is empty. To diagram the proposition that there are s's, which we interpret as salng ,"hat there is at least one
member of s, we place an r anywhere in the interiorof the circle representing
E, indicating in this way that therc is something inside it, that it is n"t
Thus the nuo propositions. "There are no S's" and "There are S's" ,r.
sented by *re tno diagrams in Figure G4.
This diagram is of a class, not a proposirion.

I:SP{}
$GUNE62

..pr[
,"p*

tymbollc Reprcsenl.tlon of Categorlcal Proposltloni

Swrosc

Fonr Pnorosrrrox Brrrrsmnmor

AII

SisP

lr^
c:^
lluo61

Some

=0

cD- n

SisP

Some Sis

8P

notP

SP+O

SP;*O

Harrumou
The class of thingp that are
both Sand non-Pis empty.
The ctass of things that are
both Sand Pis'empty.
Thc .la$ of thing;s that are
both Sand Pisnotempty.
(SPhas at least one member.)
The class of thingp that are
both Sand non-Pis notempty.
(SF has at least one member.l

The noation shown in &e able is useful, for example, in representing the re
htionship among contradictories in the Boolean square of opposition.
l{hen first explaining the four types ofstandard-form categorical proposi
tiong in section 6,3, we represented the reletions of the classes in those prope
sitions graphically with intersecting circles, Iabeled S and P. Now we carr/ that
process of diagramming categorical propositions somewhat further, enriching
our notation in wdls that uill holitete tle analysis to follow. We begin by repr*
sentinganyehssf ift xs rrnmarkedcirch,labeledwidr the term that designates
that dass. The
"las" S is diagramrned wi& a simple circle, &us:

S=0
;rcrrnE

S*0

64

Note that the circle that diagrams the class swiu also, in effecg diagram
the
dass]forjustas the interiorif the circle representsall membenof

d,;;,h.;-

terior of the circle represents alt memben ofSI

To diagram a saadard-form cargorical propooition, as we notd earlier, tr+o
required. The following diagrarn using intersecring circles (as was exT9o
plained in seaion 6.3) is the skeieton, or nameliort" for
dialamming any srrndard'form categoricai proposition whose subject tenns are

auuireviatea 6y sana

HGUm 6-5

r.

propositions

211

212

Chapter

Symbolism and Diagrams for Categorical

&tegoriral Propositions

This figure diagrams the two classes of Sand Pbut diagrams no proposition con
ceming them. It dbes not affrm that either or both have memben, ncr does it
deny that they hare. fu a matter of fact, therc are more than two classes diagrammed by the two intersecting circles. The part of the circle labeled S that does
not overlap the circle labeled Pdiagrams all S's that arc not P's and can be
thought of as representing the product of the clases S and P. We may label it SPI
The overlappingpartofthe two circles represents the product ofthe clxses Sand
P, and diagrams all things belonging to both of them. It is labeled SP. The part of
the cirde labeled Pthat does not overlap the circle Iabeled S diagrams all P's tlut
ar not S's, and represents the produrt of the class S and P. It is labeled SI
Iinally, the part of the diagram ortemal to bodr circles represents dl *rings that are
neither in S nor in P; it diagrams the fourth class SPso labeled. Wth &ese labels inserted, Figure G5 becomes [i6ure 66.

labeled, but not marked in any other way, they represent cizsa ont1,. Figure 66
illustrates this. It does not represent any proposition. In such a diagrzm, if a circle or part of a circle is left blank, that signifies nothing-neither that there are,
nor that there are not, members ofthe clru represented by that space.
Howeve4 with certain additions, Venn diagrams can be used to represent
popositionsaswell as classes. By shading out some spaces, or by inserting r s in

various parts of the picture, we can accumtely diagram any one of the four
standard-form categorical propositions. Because Venn diagrams (wirh appro
priate markings) represent categorical propositions so fully and so graphically, these diagrams have become one of the most powerful and most widely

kt

for the appraisal of syllogistic argumena.

us consider
how each of the four basic categorical propositions would be represented
using this technique.
To diagram the A proposition "All S is P." symbolized 255p= Q we simply
shade out the part ofthe diagram that represenb the class Sp, rhus indicating it
has no members or is empty. To diagram the E proposition "No Sis P," sl,rnbolused instruments

out that part of the diagram which represens the class

to indicate that it is empty. To diagram the I proposition "Some Sis p," symbolized SP* 0, we insert an x into that part of the diagram that represens the
ized as ,SP= 0, we shade
SP,

This insertion indicates that the class product is not mpty but has at
for the O proposition "Some Sis not P," symbolized
SP* 0, we insert an * into that part of the diagram that represenb the class SF
to indicate thit it is not empry but has at least one member. placed side by side,
diagrams for the four standard-fornr categorical propositions display their dii
ferent meanings very clearly, as shown in Figure G7.
class SP.

FIGURT

least one mernber. Finally,

6{

This diagram can be interpreted in terms ofthe several different classes deteimined by the class of all Spaniards (S) and the class of all painters {P). SPis the
product of these two classes, containing all those things and only those things
irlrai belong io both of them. Ever,v- member of S.Dmusi be a member of both -(
and P; every member must be both a Spaniard and a painter. This product clas
SPis the class of all Spanish painters, which contains, among others, Veldsquez
and C'op. SP is tlre product of the first dass and the complement of the secon4
containing all thqe thingt aad only &669 rhin8s that belong to &e clas Sbut

Vrnn DLagramt
A method of repre-

rntlq

clases and
catego{ical propoJitirxs using orerlap
png circles.

rrot to the .les P- It is t]re c-las ofalt Spnniards who are not painters, dl Spanish
nonpainters, and it will corain neitherllelSsquez nor Goya, but it will include
both the novelist Cervantcs end the dictator Fraaco, among mary othen. SPis
the product of the second class and thc complcment of the fint, and is the clas
of all painten who are not Spaniards. This dass SP of all nonSpanish painten
includes, among others, both the Dutch painter Rembrandt and &e American
painter Georgia O'Keeffe. Finally,SPis the product of the complements of the
nro original classes. It coritains all those things and only those thingr &at art
neither Spaniards nor painters. It is a very lalge clas indeed, containing not
merely English admirals and Swis mountain climben, but such things as the
Mississippi River and Mount Everesr All these dasses are diagrammed in Figurc
66, where the letten Sand Pare intelpreted as in the present paragraph.
Diagmms of this kind are called

I&m dagrms, named

(as noted earlier)

afterJohn Venn, the English loCr.i- who inuoduced this method of repre
senting classes and propositions. When, in srrch diaglams, the several areaszre

iLAll Sis P
sp= o

tr'No$isP
SP=0

I:SomeSisP

SP+O

&SomeSbnotP
SPr*O

HGURT 6-7

lVe have constructed diagmmmatic representations for 'No S is

p"

and

'Some Sis P," and since these are logically

iquivalent to their converses "No p
is.S' and "Some Pis
$" the diagrams"for the latter.hane already been shown. To

$ugt"* the A proposition "Ail P is S," symbol'zed as pS=0 within the same
t'amework, we must sbade out the part of the diagram that represents the
ctass
PS. It should be obvious that the
ct*t fS is the sime rs rtre ciasrsp if not immediately, then by recognizing that every ohject that belongs to
the class ofall
Patnters and the class of all non-Spaniards must (also) belong to the class of
all nonSpaniards and the class of al painters-all painting no'n-spaniards are

Propositions 213

2t4

Ctrapter

Categorical Propoitiotu

Summary 215

nonspanish pairters, and vice versa. And'to diagram the 0 proposition

*Sorne
Pis not $,' symbolized by PS* 0 we insert an r into the pan of the dia
gram that represents the clas PS ( = SP). Diagramsfor these propositions then
appear as sholr,n in Figure 6{.

2.

No peddlers are millionaires.

3.

AII merchants are speculators

4.

Some musicians are not pianists.

*5.

No shopkeepers are members.

6.

Some political leaders of high reputation are scoundrels.

1.

All phpicians licensed to practice in this state are medical school

8.

Some stockbroken rvho advisc their customers about making inrrest-

graduarcs wtro have passed special

rt:All

Pfu S

r5=o

E' No Pis
Ps=

S.

I: Some Pis

H*0

menB are not partners in companies whose securities &cy

G Some Pis not 5-

g. Irtrlnr**r

The Venn diagrams constitute an rcoarirrepresentation of the standard-form

categorical propositions, in which spatial inclusions and exclusions correspond
to the nonspatial inclusions and exclusions ofdasses. They provide an exceP
tionally clear method of notation. They also provide the basis for the simplest
and most direct :nethc<! of tes'ing the validity of categorical syllogisms'.as rvill
be explained in the following chapter.

tc<dro

|--

*10.

No modern paintings are photognphic likenesses of their objects.

ll.

Some student activists arc middle-aged men.and women striving to

'

12. All medierral scholan were pious monks living in monasteries.

13. Some state employees are not prblic*pirited citizens.
14. No magistrates subject to election and recall wilI be punitive tyrants.
*15. Some patients exhibiting all the symptoms of schizophrenia are
manicdepressives.

16.
17.

lE.

ExERCISES

each of the followingprofositiorx as equalities or ineqrralities, repre

seriting each clr.* by the firslfetter of the English term designiting it, *d q,r"
bolizing the proposition by means of aVenn diagnm.

t'

reiapture their lost youth.

6-t r, 2, 3, E:pres

s' 6' 7

reject all useles pleasure are strangers to much

that makes life worth living.

This further adequacy of the tr,vo<ircle diagrams is mentioned because in the

following chapter it will be important to be able to use a given pair of overlap
ping circles with given labels-+ay, ,tand M-to diagram any standard-form car
cgorical proposition containing Sand Mas its terms, regardles of the order in
which they occur in ir

ood.ogfu(D-nOil

rccor

PS* o

flGUNE6A

ExrrdrgrYrhblc

qulifying oraminations.

Some passengen on largejet aiqplanes are not satisfied customrs.

Some priesb are militant advocates ofradicat social change.
Some salwart deienders of the existing or-der are Dot membert 0f
political parties.

. No pipelines hid
*20. AII pornographic
19.

acros foreign territories are safe investments

films are menaccs to civilization and deccnq.

EXAMPLE:

1.

Some sculptors are parnters.

SulrtunRy
Srlllltion:

SPro

This ctrapter has introduced and exptained the basic elements

of clasica!" or
as distinguished from modern symbolic logic. (See

Aristotelian, deductirrc logic,

cecdon 6.1.)

a._---.-

.---.------

In section 6.2 we inuoduced the concept of classes on which traditional logic is built, and.introduced the categorical propositioas tliat exPress relations between classes.

216

Chapter6 CategoricalProPositions
PAnr rr

stcTtoN

A (t^5flc^t

LoGtc
I consider the

In section 6.3 we explained the four basic standard-form

categorical

invention of the

propositions;

form of syllogisms
one of the most

A propositions: universal affi rmative

E propositions: universal negative

beoutiful, ond

I propositions: particular allirmative

O propositions: particular negative

olso one

mode by the

In section 6.4 we discussed yarious features of these standard-form categorical

propositions: their quality (affrmative or negative) and their quantity (universal or panicular)- We also explained why differcnt terms are distributed or
rmdishibuted, in each of the four basic kinds of propositions.
In section 6.5 we explored the kinds of oppositioa arising among the ser"
eral standard-form categorical propositions: what it means for propositions
io be conradictories, or cortraries, or subcotrtrilies, or sub arrd
of one another. We showed how these relations are exhibited on the traditional
'
Square of Opposition, and explained the immediate inferences that can be
drawn from them.
In section 6.6 we examined other kinds of immediate inferences that are
based upon categorical propositrons: conversion, obversion, and contraposition.
In.section 6.7 we explored the controversial isue of existential import,
showing that the tiaditional Square of Opposition can be retained only if we
make a blanket assumptlon that the classes to which the subjects of proposi
tions refer always do have some members-an assumption that modern logi'

cians are unwilling

to make. We then explained the intelpretation of

propositions to be adopted throughout this book" cailed Boolean, which retains much, but not ali, of the traditional Square of Opposition, while rejecung
the blanket assumption of nonempty classes. In this Boolean interpretatioh,
we explained that partiorlar prnpositions (I and O propositions) are inter
preted es having existential imporq while rmiwsal propositionls (A and E
propositions) are intetpreted as not having such iilport The consequences,of
adopting this interpretation of propositions we carefirllydetailed.
In section 6.8 we renrrned to the use of Vm diagrms, using intenecting cir'
cles for the representation of classes. We showed how, with additional markings'
Venn diagrams may also be used to represent catcgorical popositions.
This chapter has provided the tools that will be needed to analyze the
categorical syllogisms ofwhich standard-form propositions are the essential
building blocks

of the

most impoftont,

humon mind.

-Gr:ufrietl

7.1

Standard-Form Categorical Syllogismr

7.2

The Formal Nature of Sy'logirtic Argument

Venn Diagram Technique for

7,t

Leihniz

Testing Syllogirmr
ly!logistic Rules and Syllogistic Fallaciec
7.5 Exporltion df the t5 Valid Formi of the
Categorical lyllogism
7.6 Deduction of the 15 Valid Formr of the
Categorica! Syllogism
7.4

7.1 Standard-Form Categorical Sylloglsm;

I{e are now in a position to use categorical propositions ir.r more extended
reasoning. Arguments that rely upon the A E, I, and 0 propositions commonly have two categorical propositions as prernises. and one categorical
proposition
is,

as

its conclusion. Such arguments are called syfloglsms; arsyilogism

in general, a deductive argument in which a conclusion is inferred from

Sylloglrn
Any deductive
argument in which a

conchxion is infened
from tuo prernises.

Categorical

Sylloglrm

because they

A deductive
argument consisting
of three categot{cal
proposition5 that

the constituent propositions.

propositions.

two premises.

The sflogisms with which we are here concerned are called categmical
are arguments based on the relations of classes, or categories-those relations being expressed by the categorical propositions.with together contain
which we are familiar. More formally, we define a categorical syllogism as a exactly three terms,
deductive argument consisring of three categorical propositions rhat ro- each of which
occur in exactlY two
gtther contain exactly three terms, each of which occurs in exactly two of of the constituent

217