Download as pdf or txt
Download as pdf or txt
You are on page 1of 42

What You Need From Propositional Logic

2013 James Gray

Table of Contents
Chapter 1: What is Logic?..........................................................................................................................3 Formal logic......................................................................................................................................3 In ormal logic...................................................................................................................................! What"s the #i erence $et%een logic an# epistemology?..................................................................& What is the essence o logic?............................................................................................................& Chapter 2: Why logic is important.............................................................................................................' What is a goo# arg(ment?................................................................................................................' What characteristics #o goo# arg(ments ha)e?................................................................................' Why is goo# arg(mentation important?.........................................................................................10 Chapter 3: What is propositional logic?...................................................................................................11 Chapter !: *ranslation..............................................................................................................................12 What are logical connecti)es?........................................................................................................12 +cheme o a$$re)iation..................................................................................................................1& ,rg(ments.......................................................................................................................................1& Chapter -: *r(th *a$les............................................................................................................................1' Logical connecti)es........................................................................................................................1' , comple. tr(th ta$le......................................................................................................................22 *a(tologies......................................................................................................................................22 +el /contra#ictions..........................................................................................................................23 Consistent statements......................................................................................................................23 01(i)alent statements.....................................................................................................................2! 2ali# arg(ments..............................................................................................................................2Chapter &: 3a4ing yo(r o%n tr(th ta$les................................................................................................25 +tep 1: *ranslate statements o or#inary lang(age.........................................................................25 +tep 2: 6rea4 all comple. statements into smaller parts................................................................2' +tep 3: 7etermine ho% many col(mns are re1(ire#.......................................................................28 +tep !: 7etermine ho% many ro%s are re1(ire#............................................................................28 +tep -: 7etermine the tr(th )al(es o statement letters..................................................................30 +tep &: 7etermine the tr(th )al(es o comple. statements.............................................................31 Chapter 5: 9at(ral #e#(ction...................................................................................................................3! :(les o in erence...........................................................................................................................3! ;o% to constr(ct a proo ................................................................................................................3Chapter ': Con#itional < In#irect =roo ..................................................................................................38 Con#itional proo ............................................................................................................................38 In#irect proo .................................................................................................................................!1

Chapter 1: What is Logic?


Logic is a #omain o philosophy concerne# %ith rational criteria that applies to arg(mentation. Logic incl(#es a st(#y o arg(mentation %ithin nat(ral lang(age> consistent reasoning> )ali# arg(mentation> an# errors in reasoning. It is #i)i#e# into t%o main #omains: Formal an# in ormal logic.

Formal logic
Formal logic is the tra#itional #omain o logic in %estern philosophy. It is a #omain that co)ers logical orm> consistency> )ali# arg(mentation> an# logical systems. Logical form Logical orm allo%s (s to sym$oli?e statements $y stripping statements o their content. For e.ample> consi#er the statement @i it %ill rain to#ay> then the roa#s %ill $ecome slippery.A *he logical orm o this statement %o(l# $e presente# in propositional logic as @i ,> then 6.A In that case @,A stan#s or @it %ill rain to#ayA an# @6A stan#s or @the roa#s %ill $e slippery.A Logical connecti)es are 4ept> s(ch as @i Bthen>A @an#>A @or>A an# @not.A Logicians #on"t (s(ally %rite statements as @i ,> then 6.A Instea#> they (s(ally (se a sym$ol or logical connecti)es> s(ch as @C.A We can state @i ,> then 6A as @, C 6.A Consistency *%o statements are consistent i it"s possi$le or them $oth to $e tr(e at the same time. For e.ample> the statement @i it %ill rain to#ay> then the roa#s %ill $e slipperyA is consistent %ith the statement @it %ill not rain to#ay.A Logic pro)i#es (s %ith a %ay to #etermine %hen statements are consistent> %hich is important to (s $eca(se all tr(e statements a$o(t the %orl# are consistent. D*%o tr(e statements can ne)er orm a contra#iction. For e.ample> @,liens li)e on another planetA an# @aliens #on"t li)e on another planetA orm a contra#iction> so one o the statements is alse.E We 4no% that t%o statements are consistent as long as they can all $e tr(e at the same time> an# contra#ictory %hen they can"t. Whene)er t%o propositions contra#ict> one proposition can $e sym$oli?e# as @,A an# the other can $e sym$oli?e# as @not/,.A For e.ample> @it %ill rain to#ayA contra#icts @it %ill not rain to#ay.A +ome statements are also sel /contra#ictory> s(ch as @one person e.ists an# no people e.ist.A 3any sel /contra#ictions can $e sym$oli?e# as @, an# not/,.A *hese statements are al%ays alse. *a(tological statements are al%ays tr(e> s(ch as @either the 3oon re)ol)es aro(n# the 0arth or the 3oon #oesn"t re)ol)e aro(n# the 0arth.A 3any ta(tologies can $e sym$oli?e# as @, or not/,.A

Valid argumentation , )ali# arg(ment has an arg(ment orm that co(l# ne)er ha)e tr(e premises an# a alse concl(sion at the same time. For e.ample> @I it %ill rain to#ay> then the roa#s %ill $e slippery. It %ill rain to#ay. *here ore> the roa#s %ill $e slipperyA is )ali# $eca(se it has the arg(ment orm @I ,> then 6. ,. *here ore> 6.A ,ll arg(ments %ith this orm are )ali#. Logic gi)es (s the tools to #etermine %hen an arg(ment is logically )ali#. I a #e#(cti)e arg(ment is not logically )ali#> then it #oes not pro)i#e (s %ith a goo# reason to agree %ith the concl(sion. I the premises are tr(e> then the concl(sion co(l# still $e alse. ,n e.ample o an in)ali# arg(ment is @,t least one person e.ists. I at least one person e.ists> then at least one mammal e.ists. *here ore> no mammals e.ist.A ,ltho(gh the premises are tr(e> the concl(sion is alse. *his arg(ment #oes not #o %hat arg(ments are s(ppose# to #oFpro)i#e (s %ith a goo# reason to thin4 the concl(sion is tr(e. Logical systems Logical systems ha)e D1E a ormal lang(age that allo%s (s to sym$oli?e statements o nat(ral lang(age> D2E a.ioms> an# D3E r(les o in erence. 1. , ormal lang(age is a %ay %e can present the orm o o(r statements in)ol)ing logical connecti)es. 2. ,.ioms are r(les> s(ch as the r(le that states that contradictions can't exist. 3. :(les o in erence are r(les that state %hat premises can $e (se# to )ali#ly in er )ario(s concl(sions. For e.ample> a r(le 4no%n as @modus ponensA states that %e can (se @,A an# @i ,> then 6A as premises to )ali#ly in er that @6.A Logical systems are nee#e# in or#er or (s to $est #etermine %hen statements are consistent or %hen arg(ments are )ali#.

Informal logic
In ormal logic is #omain that co)ers the application o rational arg(mentation %ithin nat(ral lang(age Fho% people act(ally tal4. What %e call @critical thin4ingA is o ten sai# to in)ol)e in ormal logic> an# critical thin4ing classes generally oc(s on in ormal logic. In ormal logic is mainly a$o(t rational arg(mentation> the #istinction $et%een in#(cti)e an# #e#(cti)e reasoning> arg(ment i#enti ication> premise an# concl(sion i#enti ication> hi##en ass(mption i#enti ication> an# error i#enti ication. Rational argumentation ,rg(ments are a series o t%o or more statements incl(#ing premises Ds(pporting statementsE an# concl(sions Dstatements that are s(ppose# to $e G(sti ie# $y the premisesE. For e.ample> @,ll h(man $eings that ha# li)e# in the #istant past ha# #ie#. *here ore> all h(man $eings are pro$a$ly mortal.A *he i#ea o rational arg(mentation is that it is s(ppose# to gi)e (s a goo# reason to $elie)e the !

concl(sion is tr(e. I an arg(ment is goo# eno(gh> then %e sho(l# $elie)e the concl(sion is tr(e. I an arg(ment is rationally pers(asi)e eno(gh> then it %o(l# $e irrational to thin4 the concl(sion is alse. For e.ample> consi#er the arg(ment @,ll o$Gects that %ere #roppe# near the s(r ace o the 0arth ell. *here ore> all o$Gects that are #roppe# near the s(r ace o the 0arth %ill pro$a$ly all.A *his arg(ment gi)es (s a goo# reason to $elie)e the the concl(sion to $e tr(e> an# it %o(l# seem to $e irrational to thin4 it"s alse. The distinction between deductive and inductive reasoning 7e#(cti)e arg(ments are meant to $e )ali#. I the premises are tr(e> then the concl(sion is s(ppose# to $e ine)ita$le. In#(cti)e arg(ments are not meant to $e )ali#. I the premises o an in#(cti)e arg(ment are tr(e> then the concl(sion is s(ppose# to $e li4ely tr(e. I an in#(cti)e arg(ment is strong an# the premises are tr(e> then it is (nli4ely or the concl(sion to $e alse. ,n e.ample o a )ali# #e#(cti)e arg(ment %as gi)en a$o)e %hen )ali# arg(ments %ere #isc(sse#. Let"s ass(me that @i it %ill rain to#ay> then the roa#s %ill $e %etA an# that @it %ill rain to#ay.A In that case %e ha)e no choice $(t to agree that @the roa#s %ill $e %et.A ,n e.ample o a strong in#(cti)e arg(ment %as gi)en in the arg(ment in)ol)ing #ropping o$Gects. It is (nli4ely that #roppe# o$Gects %ill not all in the (t(re ass(ming that they al%ays ell in the past. Argument identification Hno%ing %hat arg(ments are an# %hy people (se them helps (s 4no% %hen people gi)e arg(ments in e)ery#ay con)ersation. It can also $e help (l to 4no% the #i erence $et%een arg(ments an# other similar things. For e.ample> arg(ments are not mere assertions. , person %ho gi)es a mere assertion is telling yo( %hat she $elie)es to $e tr(e> $(t a person %ho gi)es an arg(ment tells yo( %hy she $elie)es %e sho(l# agree that a concl(sion is tr(e. Premise and conclusion identification Hno%ing %hat premises an# concl(sions are helps (s 4no% ho% to 4no% %hich are %hich in e)ery#ay con)ersation. For e.ample> a person can say @the #eath penalty is %rong $eca(se it 4ills people.A In this case the premise is @the #eath penalty 4ills peopleA an# the concl(sion is @the #eath penalty is %rong.A Hidden assumption identification Hno%ing that an arg(ment is meant to $e rationally pers(asi)e can help (s reali?e %hen hi##en ass(mptions are re1(ire# $y an arg(ment. For e.ample> the arg(ment that @the #eath penalty is %rong $eca(se it 4ills peopleA re1(ires the hi##en ass(mption that @it"s al%ays %rong to 4ill people.A Witho(t that ass(mption the arg(ment %ill not $e rationally pers(asi)e. I it"s not al%ays %rong to 4ill people> then perhaps the #eath penalty is not %rong a ter all. Error identification Hno%ing a$o(t se)eral errors o reasoning Di.e. allaciesE can help (s 4no% %hen people ha)e errors o reasoning in arg(ments they present in e)ery#ay con)ersation. For e.ample> the arg(ment @my rien# -

Joe ne)er #ie#> so no person %ill #ie in the (t(reA contains an error. *he pro$lem %ith this arg(ment is the (nG(sti ie# ass(mption that %e can 4no% %hat %ill happen to e)eryone in the (t(re $ase# on %hat happene# to a single person gi)en a limite# amo(nt o time. *his type o error is 4no%n as the @hasty generali?ationA allacy.

What's the difference between logic and epistemology?


0pistemology is the philosophical st(#y o 4no%le#ge> G(sti ication> an# rationality. It as4s 1(estions> s(ch as the ollo%ing: 1. 2. 3. !. -. What is 4no%le#ge? Is 4no%le#ge possi$le? What are the %ays %e can rationally G(sti y o(r $elie s? When it is irrational or a person to ha)e a $elie ? When sho(l# a person agree that a $elie is tr(e?

*hese iss(es are highly relate# to logic> an# many philosophers ha)e e1(ate# @logicA %ith @epistemology.A For e.ample> the +toic philosophers incl(#e# epistemology in their #omain o @logic.A I $elie)e that logic sho(l# no% $e consi#ere# to $e part o the #omain o epistemology. ;o%e)er> or e#(cational p(rposes it is consi#ere# to $e a separate s($Gect an# it"s not ta(ght in epistemology classes. Logic classes #eal %ith arg(ment orm an# certain rational criteria that applies to arg(mentation> $(t epistemology classes generally #eal %ith some%hat a$stract 1(estions> as %ere liste# a$o)e. =erhaps one o the most important iss(es that logic #eals %ith m(ch less than epistemology is G(sti icationF logic ten#s not to tell (s %hen premises are G(sti ie# an# ho% %ell G(sti ie# they are> $(t epistemology attempts to tell (s %hen premises are G(sti ie#> an# %hen a premise is G(sti ie# eno(gh to rationally re1(ire (s to $elie)e it"s tr(e. Why #o logic an# epistemology classes teach #i erent things? =erhaps $eca(se philosophers %ho ha)e an interest in epistemology ha)e historically not care# as m(ch a$o(t logic an# )ice )ersa. 6(t %hy %o(l# philosophers %ho care a$o(t epistemology not care as m(ch a$o(t logic? =erhaps $eca(se logic ten#s to $e concerne# %ith iss(es that can $e ans%ere# %ith a m(ch higher #egree o certainty. We 4no% %hat arg(ments are. We 4no% that goo# arg(ments m(st apply certain rational criteria. We can #etermine %hen arg(ments are )ali# or in)ali#. We can #etermine that many arg(ments ha)e hi##en premises or )ario(s errors. ;o%e)er> %e can"t #etermine the nat(re o 4no%le#ge> G(sti ication> an# rationality %ith that #egree o certainty. It is more contro)ersial %hen a $elie is G(sti ie# an# at %hat point a $elie is G(sti ie# eno(gh to rationally re1(ire (s to $elie)e it"s tr(e.

What is the essence of logic?


I #on"t thin4 that logic has an essence. It"s a #omain concerne# %ith certain rational criteria in)ol)e# %ith arg(mentation> $(t not all criteria. 0pistemology also co)ers relate# iss(es. What %e consi#er to &

$e logic or epistemology mainly has to #o %ith a history o philosophers Dan# mathematiciansE %ho la$el themsel)es as @logiciansA or @epistemologistsA an# teach classes in the correspon#ing #omains. *hese terms are (se# merely $eca(se they are con)enient to (s. ;o%e)er> I thin4 %e can say that logic is a #omain o epistemology that has a restricte# oc(s> an# that oc(s is mainly restricte# to iss(es that %e thin4 %e can ans%er %ith a m(ch higher #egree o certainty than (s(al. Logic an# mathematics are no% o ten ta4en to $e part o the same #omain> an# $oth generally o er (s %ith a #egree o certainty higher than the nat(ral sciences. Whene)er scienti ic in#ings con lict logic> %e are m(ch more li4ely to thin4 that o(r scienti ic in#ings are alse than that o(r (n#erstan#ing o logic is alse. *he same can not $e sai# o epistemology in general. *here are e.amples o epistemological iss(es that #o seem to in)ol)e a great #eal o certainty. I thin4 %e sho(l# $e con i#ent that %e sho(l# $elie)e that @1I1J2A an# that it"s irrational to $elie)e that @1I1J3.A 0pistemology tells (s %hat %e sho(l# $elie)e in that sense. ;o%e)er> there is also a great #eal o (ncertainty that is (s(ally in)ol)e# %ith epistemology. ,ns%ers to the $ig 1(estions in epistemology are still )ery contro)ersial.

Chapter 2: Why logic is important


Why is logic e#(cation important? *he main 1(estion here is %hat the real point o logic e#(cation is. *he real point o logic is not to teach people ho% to $e logic pro essors> or to increase test scores> or to impress potential employers. =hilosophers an# mathematicians %ere )ery intereste# in (n#erstan#ing logic long $e ore it %as ta(ght in (ni)ersities precisely $eca(se o ho% important it is. Why is logic so important? *he ans%er is that logic helps (s $etter (n#erstan# goo# arg(mentsFit helps (s #i erentiate $et%een goo# an# $a# reasons to $elie)e something. We sho(l# %ant to ha)e %ell/ G(sti ie# $elie s. We %ant to 4no% %hat %e sho(l# $elie)e. Kn#erstan#ing goo# arg(mentation helps (s (n#erstan# %hen %e sho(l# $elie)e something> an# (n#erstan#ing logic helps (s (n#erstan# goo# arg(mentation. ;o% e.actly #oes logic help (s (n#erstan# goo# arg(mentation? *here are many necessary characteristics that goo# arg(ments m(st ha)e> an# logic tells (s %hat some o those characteristics are. Logic also helps (s $etter (n#erstan# concepts that are rele)ant to goo# arg(mentation.

What is a good arg ment?


Goo# arg(ments are goo# reasons to $elie)e something is li4ely tr(e. I %e 4no% o a goo# arg(ment to $elie)e something> then %e sho(l# $elie)e it. For e.ample: 1. ,ll #ogs are mammals. 2. I all #ogs are mammals> then all #ogs are li)ing organisms. 3. *here ore> all #ogs are li)ing organisms. =eople sho(l# agree that @all #ogs are li)ing organisms.A We 4no% the premises are tr(e Dthat @all #ogs are mammalsA an# @i all #ogs are mammals> then all #ogs are li)ing organismsAE. We 4no% that i the premises are tr(e> then the concl(sion has to $e tr(e as %ell. *he premises can"t $e tr(e an# the concl(sion alse at the same time $eca(se the arg(ment is logically )ali#.

What characteristics do good arg ments ha!e?


Kltimately goo# arg(ments m(st ha)e s( iciently G(sti ie# premises> an# the premises sho(l# $e appropriately rele)ant to the concl(sion. 0)en so> there is m(ch to $e sai# a$o(t this criteria. 3any o the necessary characteristics o goo# arg(ments are co)ere# $y the )ario(s iss(es #isc(sse# in logic classesFlogical orm> logical )ali#ity> the #istinction $et%een in#(cti)e an# #e#(cti)e reasoning> arg(ment interpretation> an# in ormal allacies. 0.amples a$o(t %hat )ario(s general logical iss(es can teach (s a$o(t goo# arg(mentation incl(#es the ollo%ing: 1. Logical form L Kn#erstan#ing logical orm is o paramo(nt importance to (n#erstan#ing goo# #e#(cti)e arg(mentation> $(t or#inary lang(age ma4es it )ery #i ic(lt to #isc(ss logical orm. It is m(ch easier to (n#erstan# logical orm an# ho% it relates to goo# arg(mentation a ter '

2.

3.

!.

-.

learning a$o(t logical orm in a logic class. For e.ample> @i +ocrates is a man> then he is mortalA has the logical orm @i ,> then 6.A Logical validity L Logically )ali# #e#(cti)e arg(ments ha)e premises that g(arantee the tr(th o the concl(sion Dass(ming they are tr(eE. ,n in)ali# #e#(cti)e arg(ment gi)es (s no reason to thin4 the concl(sion is tr(e. In that case the premises can $e tr(e an# the concl(sion can $e alse at the same time. The distinction between inductive and deductive reasoning L ,ll goo# #e#(cti)e arg(ments are logically )ali#> $(t goo# in#(cti)e arg(ments aren"t. In#(cti)e arg(ments are not meant to $e )ali# $eca(se the premises are only s(ppose# to ma4e the concl(sion pro$a$le. , goo# in#(cti)e arg(ment is (nli4ely to ha)e tr(e premises an# a alse concl(sion> $(t it can happen. For e.ample> the e)i#ence scientists (se to s(pport scienti ic theories is inductive an# it is possi$le that the theories are act(ally alse Dimper ectly acc(rateE. *he pre#ictions ma#e $y scientists co(l# al%ays t(rn o(t to $e alse> $(t they are li4ely to $e tr(e %hen they are %ell/ G(sti ie#. 0)en #roppe# o$Gects co(l# ail to all in the (t(re. 0)en so> %e sho(l# agree that #roppe# o$Gects %ill all in the (t(re any%ay. Argument interpretation L It is important to (lly (n#erstan# people"s arg(ments an# to 4no% ho% to clari y their arg(ments. It"s important that %e 4no% %hat e.actly the premises an# concl(sions are. +ometimes (n#erstan#ing an arg(ment also re1(ires (s to i#enti y (nstate# ass(mptions an# some creati)ity co(l# $e re1(ire#. It is impossi$le to properly #e$ate %ith someone %ho #oesn"t (n#erstan# yo(r arg(ments. ,n arg(ment can"t $e properly re (te# (nless it is (n#erstoo# %ell. nformal fallacies L In ormal allacies are errors in reasoning other than ha)ing an in)ali# arg(ment orm. Interpreting arg(ments (ncharita$ly is one common e.ample calle# the @stra% man allacy.A

*here are inn(mera$le speci ic e.amples a$o(t ho% each general iss(e can apply to goo# arg(mentation. Mne e.ample is that people o ten re (te the concl(sion o an arg(ment $y arg(ing against a premise. ;o%e)er> it is possi$le or a #i erent arg(ment to $e gi)en or any gi)en concl(sion. :e (ting a premise o an arg(ment #oes not sim(ltaneo(sly re (te the concl(sion. Consi#er the ollo%ing arg(ment: 1. I the =resi#ent o the Knite# +tates is a li?ar#> then the =resi#ent is a mammal. 2. *he =resi#ent is a li?ar#. 3. *here ore> the =resi#ent is a mammal. In this case $oth premises are alse> $(t the concl(sion is tr(e. +tating that the premises are alse gi)es (s no reason to thin4 the concl(sion is alse. +(ch an arg(ment co(l# loo4 li4e the ollo%ing: 1. @*he =resi#ent is a li?ar#A is alse. 2. *here ore> %e sho(l# reGect that @the =resi#ent is a mammal.A *his arg(ment is clearly in)ali#. *he premise is tr(e> $(t the concl(sion is alse. *his e.ample sho%s ho% (n#erstan#ing )ali#ity can help (s (n#erstan# %hy certain arg(ments ail to $e goo# arg(ments. ,ct(al people #o arg(e this %ay no% an# then> so learning a$o(t it in a logic class co(l# act(ally help people come to a reali?ation that they might not thin4 a$o(t other%ise. 3any e.amples %ere gi)en concerning ho% logic classes can help (s $etter (n#erstan# goo# 8

arg(mentation. ;o%e)er> there are potentially other characteristics o goo# arg(ments that are not #isc(sse# in logic classes. *he philosophical #omain calle# @epistemologyA concerns the nat(re o 4no%le#ge> rationality> an# G(sti ie# $elie . *here are certain #etails a$o(t %hat co(nts as s( icient G(sti ication an# rational tho(ght that logic #oes not co)er> $(t is co)ere# $y epistemology instea#. 0pistemology is the #omain concerning ho% open/min#e# %e sho(l# $e to a)oi# $eing close/min#e#> ho% s4eptical %e sho(l# $e to a)oi# $eing g(lli$le> an# ho% m(ch e)i#ence a $elie re1(ires in or#er to $e s( iciently G(sti ie#. ,lso> logicians #on"t tell (s i any $elie s are sel /e)i#ent> i int(ition is e)er a goo# reason to $elie)e something> or %hen %e can rationally ass(me a premise to $e tr(e %itho(t arg(ment. *hose are iss(es concerning epistemology.

Why is good arg mentation important?


Logic alone can"t tell (s %hy goo# arg(mentation is important. It"s a philosophical 1(estion. ,ns%ers incl(#e the ollo%ing: 1. We %ant to 4no% %hat"s tr(e. Goo# arg(ments can tell (s %hat is li4ely tr(e. 2. We o ten can"t $elie)e %hat %e sho(l# $elie)e (nless %e (n#erstan# goo# a(gmentation. Kn#erstan#ing goo# arg(mentation helps (s 4no% %hat %e sho(l# $elie)e> an# it helps (s pro)e to other people %hat they sho(l# $elie)e. 3oreo)er> %e sho(l# $elie)e certain things because %e 4no% a$o(t goo# arg(ments. 6elie)ing %hat is tr(e at ran#om is not appropriate. We sho(l# $elie)e %hat is li4ely tr(e $ase# on a goo# reasoning process. 3. M(r $elie s can moti)ate (s to $eha)e in certain %ays> an# alse $elie s are more li4ely to moti)ate (s to $eha)e in inappropriate %ays. For e.ample> %e o(n# o(t that lea# is poisono(s an# %e try to ma4e s(re chil#ren"s toys no longer contain lea# or that reason. =eople elt ree to p(t lea# in chil#ren"s toys (ntil their $elie s %ere correcte#. !. 3any people %ant to manip(late (s to $elie)e certain things> an# (n#erstan#ing goo# arg(mentation can help (s spot the a(lty arg(ments that are (se# to manip(late (s. For e.ample> charlatans %ant to sell (s pro#(cts that #on"t act(ally %or4. 3any me#ical pro#(cts are sol# that #on"t act(ally #o %hat they are sai# to #o. =eople %aste their money on those pro#(cts %hen they get #(pe#.

10

Chapter ": What is propositional logic?


=ropositional logic is a sym$olic lang(age that lac4s content> an# (ses letters to sym$oli?e propositions. For e.ample> consi#er the statement @the s4y is re# or $l(e.A We co(l# replace @the s4y is re#A %ith @,A an# @the s4y is $l(eA %ith @6.A In that case %e %o(l# en# (p %ith @, or 6.A =ropositional logic is (se# to e)al(ate the )ali#ity o #e#(cti)e arg(ments an# the consistency o statements. +tatements that are inconsistent are sai# to orm @contra#ictions.A For e.ample> the statement @the s4y is $l(eA contra#icts the statement @the s4y is not $l(e.A =ropositional logic consists o logical connecti)es> r(les o in erence> an# a.ioms. *hese %ill $e #isc(sse# in more #etail in later chapters. ;o%e)er> it is important to 4no% %hat @statementsA an# @propositionsA are> so they %ill $e #isc(sse# in greater #etail here. !hat are statements" +tatements are sentences that are tr(e or alse. For e.ample> @all mammals are animalsA is a statement %ritten in 0nglish. In this case the statement is tr(e. !hat are propositions" =ropositions consist in the meaning o statements. We co(l# say that propositions are the @concept(al contentA o statements. For e.ample> @the s4y is $l(eA is a statement %ritten in 0nglish that re ers to the same proposition as @le ciel est $le(>A %hich is a statement %ritten in French. ,nother e.ample o t%o statements that re er to the same proposition is @all %hales are mammalsA an# @i something is a %hale> then it"s a mammal.A

11

Chapter #: Translation
"*ranslation" re ers to the act o con)erting statements o nat(ral lang(age to statements o a sym$olic logical system. In this case I %ill #isc(ss ho% to con)ert statements o 0nglish into statements o propositional logic. *ranslation re1(ires (s to 4no% logical connecti)es (se# in propositional logic> an# %ays %e (se logical connecti)es in 0nglish. 9ot all propositions (se logical connecti)es. In or#er to translate s(ch a proposition> replace it %ith a capital letter. For e.ample @1I1J2A is a proposition that can $e translate# as @,.A

What are logical connecti!es?


Logical connecti)es are %ays o saying %hen propositions are tr(e. For e.ample @orA is a %ay o saying at least one o t%o propositions are tr(e. @*he =resi#ent o the K+, is a man or a #ogA means that either the proposition @the =resi#ent o the K+, is a manA is tr(e or the proposition @the =resi#ent o the K+, is a #ogA is tr(e. We can then replace the propositions %ith capital letters to get @, or 6.A ;o%e)er> the %or# @orA is also replace# in sym$olic logic %ith the sym$ol @>A so %e %o(l# %rite it o(t as @, 6.A *here are i)e connecti)es (se# in propositional logic: @notA DNE> @an#A DE> @orA DE> @impliesA DCE> an# @i an# only i A DOE: #egation *he @negationA connecti)e is @N>A %hich means @is alse.A We also state negation $y saying @is not tr(eA or sometimes simply as @not.A For e.ample> @the =resi#ent o the K+, is not a #ogA sho(l# $e ta4en to mean @it is alse that the =resi#ent o the K+, is a #og.A It can $e translate# into propositional logic as @N,.A In this case @,A stan#s or @ the =resi#ent o the K+, is a #og.A 3ore e.amples o translating 0nglish statements that (se negation: 1. It"s alse that only t%o people e.ist. P@,A stan#s or @only t%o people e.ist.A *ranslation: @N,.AQ 2. 9ot all #ogs are ish. P*his means @it"s alse that all #ogs are ish.A @6A stan#s or @all #ogs are ish.A *ranslation: @N6.AQ 3. It is not the case that +amantha %ent to the store. P*his means @it"s alse that +amantha %ent to the store. @CA stan#s or @+amantha %ent to the store.A *ranslation: @NC.AQ !. 0)ol(tion is alse. P*his %ill $e ta4en to mean @the central claims o the theory o e)ol(tion are not consistent %ith o(r o$ser)ations.A @7A stan#s or @the central claims o the theory o e)ol(tion are consistent %ith o(r o$ser)ations.A *ranslation: @N7.AQ -. =eople are %rong %ho thin4 that nothing e.its. P*his means @it"s alse that nothing e.ists.A @0A stan#s or @nothing e.ists.A *ranslation @N0.AQ &. It"s a$s(r# to thin4 that @2I2J-.A P*his implies that @2I2J-A is alse. @FA stan#s or @2I2J-.A *ranslation: @NF.AQ 12

Con$unction *he conG(nction connecti)e is @>A %hich means @an#.A We also state conG(nctions $y saying @$(t.A For e.ample> @the =resi#ent o the K+, has an =h7> $(t #isli4es chocolateA sho(l# $e ta4en to mean @the =resi#ent o the K+, has a =h7> an# the =resi#ent o the K+, #isli4es chocolate.A *his can $e translate# as @, 6.A 6oth @,A an# @6A m(st represent statements that are tr(e or alse> so @6A m(st represent @the =resi#ent o the K+, #isli4es chocolateA rather than @#isli4es chocolate.A 3ore e.amples o translating 0nglish statements that (se conG(nctions are the ollo%ing: 1. Jill is tall an# Je is short. P@,A stan#s or @Jill is tallA an# @6A stan#s or @Je is short.A *ranslation: @, 6.AQ 2. Jenni er is a h(man an# a mammal. P@CA stan#s or @Jenni er is a h(manA an# @7A stan#s or @Jenni er is a mammal.A *ranslation: @C 7.AQ 3. ,t least t%o people e.ist> $(t no (nicorns e.ist. P@0A stan#s or @at least t%o people e.istA an# @FA stan#s or @no (nicorns e.ist.A *ranslation: @0 F.AQ !. It is tr(e that @1I1J2A an# it is tr(e that @2I2J!.A P@GA stan#s or @1I1J2A an# @;A stan#s or @2I2J!.A *ranslation: @G ;.AQ -. 0)ol(tion is tr(e an# %e sho(l# $elie)e it. P@IA stan#s or @the central claims o e)ol(tion are consistent %ith o(r o$ser)ationsA an# @JA stan#s or @%e sho(l# $elie)e the central claims o e)ol(tion.A *ranslation: @I J.AQ %is$unction *he #isG(nction connecti)e is @A an# it means @or.A We also state this $y saying @an#BorA or @(nless.A It means that one or $oth o t%o propositions are tr(e. For e.ample> @J(lia %ent to the store (nless she %ent to schoolA can $e translate# as @, 6.A In this case @,A stan#s or @J(lia %ent to the storeA an# @6A stan#s or @J(lia %ent to school.A +ometimes #isG(nctions are meant to $e e.cl(si)eFonly one o the t%o propositions are tr(e. For e.ample> @the =resi#ent is a man or a %omanA co(l# imply that the =resi#ent is a man or a %oman> $(t not $oth a man an# a %oman. In that case %e can translate the #isG(nction as saying @either , or 6> $(t not $oth/,/an#/6.A *his can $e %ritten in sym$olic orm as @D, 6E ND, 6E. D@,A stan#s or @the =resi#ent is a manA an# @6A stan#s or @the =resi#ent is a %oman.AE We (se parentheses to gro(p propositions together that are intricately tie#. *he logical connecti)e that is not in any parentheses is the @main connecti)e.A 3ore e.amples o translating 0nglish statements that (se #isG(nctions are the ollo%ing: 1. 0ither the roa# lea#s $ac4 to to%n or the roa# lea#s to a arm. P@,A stan#s or @the roa# lea#s $ac4 to to%nA an# @6A stan#s or @the roa# lea#s to a arm.A *ranslation: , 6Q 2. 0ither Jac4 or Jill is tall. P@C stan#s or @Jac4 is tallA an# @7A stan#s or @Jill is tall.A *ranslation: @C 7.AQ 3. ;(mans are mammals (nless they"re reptiles. P@0 stan#s or @h(mans are mammalsA an# FA stan#s or @h(mans are reptiles.A *ranslation: @0 F.AQ !. 0ither e)ol(tion is tr(e or creationism is tr(e. P@GA %ill stan# or @the central claims o e)ol(tion are consistent %ith o(r o$ser)ationsA an# @;A %ill stan# or @the central claims o 13

creationism are consistent %ith o(r o$ser)ations.A *ranslation: @G ;.AQ Conditional *he con#itional is @CA an# means @implies.A We also state con#itionals $y saying @i ... thenA or @only i .A For e.ample> @i all men are mortal> then +ocrates is mortal.A We can translate this to $e @, C 6.A In that case @,A stan#s or @all men are mortalA an# @6A stan#s or @+ocrates is mortal.A *he con#itional means that the irst part is alse or the secon# part is tr(e. =eople o ten tal4 a$o(t necessary an# s( icient con#itions> an# they can $e translate# to $e propositions (sing con#itionals. @, is necessary or 6A means @6 C ,.A @, is s( icient or $eA means @, C 6.A For e.ample> %e co(l# say that @the e.istence o the +(n is s( icient or the e.istence o light>A or that @the e.istence o atoms are necessary or the e.istence o h(mans.A ,lso note that people sometimes say @, i 6>A s(ch as @light e.ists i the +(n e.ists.A In this case %e ha)e to s%itch the irst an# secon# parts o the con#itional to translate it into logical orm. @,A can stan# or @light e.istsA an# @6A can stan# or @the +(n e.ists.A *he statement %ill then $e translate# as @6 C ,.A 3ore e.amples o translating 0nglish statements that (se material con#itionals are the ollo%ing: 1. I Lisa 4ic4e# John> then John got h(rt. P@,A stan#s or @Lisa 4ic4e# John.A @6A stan#s or @John got h(rt.A *ranslation: @, C 6.AQ 2. George li4es mil4 i he li4es mil4sha4es. P@CA stan#s or @George li4es mil4.A @7A stan#s or @George li4es mil4sha4es.A *ranslation: @7 C C.AQ 3. 6ir#s are animals only i $ir#s are li)ing organisms. P@0A stan#s or @all $ir#s are animalsA an# @FA stan#s or @all $ir#s are li)ing organisms.A *ranslation: @0 C F.AQ !. *he act that all h(mans are mammals implies that all h(mans are animals. P@GA stan#s or @all h(mans are mammalsA an# @;A stan#s or @all h(mans are animals.A *ranslation: @G C ;.AQ -. 6eing h(man is s( icient to $e a mammal. P@IA means @something is a h(man.A @JA means @something is a mammal.A *ranslation: @I C J.AQ &. 6eing a mammals is necessary or $eing a h(man. P@IA means @something is a h(manA an# @JA means @something is a mammal.A *ranslation: @J C I.AQ E&uivalence 01(i)alence is @OA an# it means @i an# only i .A +ometimes %e state it $y saying @G(st in case.A It is (se# to state that t%o propositions ha)e the same tr(th )al(eFthey"re $oth tr(e> or they"re $oth alse. For e.ample> @+ocrates is a person i an# only i he"s a rational animalA co(l# $e translate# as @, O 6.A In that case @,A stan#s or @+ocrates is a personA an# @6A stan#s or @+ocrates is a rational animal.A 0ither +ocrates is a person an# a rational animal> or he is neither a person nor a rational animal. 3ore e.amples o translating 0nglish statements that (se material e1(i)alence are the ollo%ing: 1. Jessica %ill %in the game i an# only i 6o$ loses the game. P@,A stan#s or @Jessica %ill %in the game.A @6A stan#s or @6o$ loses the game.A *ranslation: @, O 6.AQ 1!

2. Water is in the glass G(st in case ;2M is in the glass. P@CA stan#s or @%ater is in the glass.A @7A stan#s or @ ;2M is in the glass.A *ranslation: @C O 7.AQ 3. Li?ar#s are animals i an# only i li?ar#s are creat(res. P@0A stan#s or @li?ar#s are animals.A @FA stan#s or @li?ar#s are creat(res.A *ranslation: @0 O F.AQ 'ultiple connectives 3any statements (se m(ltiple connecti)es. =ropositional logic re1(ires that %e 4eep all o the logical connecti)es possi$le. @9either... norA is a common e.ample. @;(mans are neither plants nor reptilesA means @it is not the case that h(mans are plants> an# it is not the case that h(mans are reptiles.A *his can $e translate# as @N, N6.A In this case @,A stan#s or @all h(mans are plantsA an# @6A stan#s or @all h(mans are reptiles.A ,nother e.ample is @(nless.A ,ltho(gh I translate# @(nlessA to mean @orA a$o)e> there are other %ays people can (se the %or# an# the conte.t sho(l# $e consi#ere#. It has $een s(ggeste# that %e sho(l# sometimes translate @, (nless 6A as @N, N6A or @ 6 C N,.A When m(ltiple connecti)es are (se#> %e o ten nee# to 4no% %hich connecti)e is the @main connecti)e.A I#enti ying the main connecti)e is nee#e# %hen other connecti)es are gro(pe# together insi#e parentheses. For e.ample> @either e)ol(tion is consistent %ith o(r o$ser)ations an# a species can change o)er time> or creationism is consistent %ith o(r o$ser)ations.A In this case a #isG(nction is the main connecti)e an# the conG(nction is gro(pe# together insi#e parentheses. @,A stan#s or @e)ol(tion is consistent %ith o(r o$ser)ations> @6A stan#s or @a species can change o)er time>A an# @CA stan#s or @creationism is consistent %ith o(r o$ser)ations.A We can then translate the statement as @D, 6E C.A It is important to ma4e it clear %hen t%o propositions contra#ict one another. We co(l# nee# to translate a contra#ictory statement s(ch as @no h(mans are reptiles> an# some h(mans are reptiles.A We co(l# translate @no h(mans are reptilesA as @,>A an# @some h(mans are reptilesA as @6>A $(t that %o(l#n"t ma4e it clear that $oth propositions orm a contra#iction. For that reason> %e sho(l# ma4e s(re to ma4e it clear that one is the negation o the other. We co(l# simply translate this statement as saying @, N,.A In that case @some h(mans are reptilesA is ta4en to $e @N,A Dthat it"s alse that no humans are reptilesE. 3ore e.amples o translating 0nglish statements that (se m(ltiple connecti)es incl(#e the ollo%ing: 1. I +ophia is not a #og> then +ophia is a h(man. P@,A stan#s or @+ophia is a #og.A @6A stan#s or @+ophia is a h(man.A *ranslation: @N, C 6.AQ 2. ;ansel %ill %in the game i an# only i Gretel #oesn"t %in. P@CA stan#s or @;ansel %ill %in the game.A @7A stan#s or @Gretel %ill %in the game.A *ranslation: @C O N7AQ 3. 9either 7emocrats nor :ep($licans %ill stop corporate %el are. P@0A stan#s or @7emocrats %ill stop corporate %el are.A @FA stan#s or @:ep($licans %ill stop corporate %el are.A *ranslation: @N0 NF.AQ !. 0ither the =resi#ent o the K+, is a reptile> or he"s $oth a mammal an# an animal. P@GA stan#s or @the =resi#ent o the K+, is a reptile.A @;A stan#s or @the =resi#ent o the K+, is a mammal.A @IA stan#s or @the =resi#ent o the K+, is an animal.A *ranslation: @G D; IE.AQ -. 0ither all li?ar#s are mammals or some li?ar#s are not mammals. P@JA stan#s or @all li?ar#s are mammals.A *ranslation: @J NJ.AQ 1-

$cheme of abbre!iation
When translating> it"s important to state %hat each letter stan#s or. *he scheme o a$$re)iation is %hat tells (s %hat each letter represents> an# m(ltiple statements can $e translate# (sing the same scheme o a$$re)iation. *ranslation re1(ires that each letter represents the same thing %ithin the scheme o a$$re)iation. Consi#er the ollo%ing three statements: 1. Jac4 is tall or Jill is short. 2. Jac4 is tall i an# only i Jill is not short. 3. I Jill is short> then Jac4 is not tall. We co(l# (se the ollo%ing scheme o a$$re)iation or these three statements: ,: Jac4 is tall. 6: Jill is short. We co(l# then (se this scheme o a$$re)iation to translate all three statements in the ollo%ing %ay: 1. , 6 2. , O N6 3. 6 C N,

%rg ments
,rg(ments are a series o t%o or more statements %hen any n(m$er o statements is meant to $e a reason to $elie)e another. *he statement that is $eing s(pporte# is the @concl(sionA an# the other statements are the @premises.A ,rg(ments are o ten (se# or pers(asionFpeople %ho agree %ith the premises are li4ely to agree %ith the concl(sion o a logically )ali# arg(ment. +ometimes arg(ments are state# as a single statement. For e.ample> @+ocrates is a manR i +ocrates is a man> then he"s mortalR there ore> +ocrates is mortal.A When translating arg(ments> %e %ant to separate each premise an# concl(sion. We also %ant to sho% the ma.imal n(m$er o premises. In this case %e %ant to separate the premises an# concl(sion into three separate statements: 1. +ocrates is a man. 2. I +ocrates is a man> then +ocrates is mortal. 3. *here ore> +ocrates is mortal.

1&

, translation an# scheme o a$$re)iation or this arg(ment is the ollo%ing: 1. , 2. , C 6 3. 6 ,: +ocrates is a man. 6: +ocrates is mortal. 9otice that @A is the sym$ol (se# to mean @there ore.A It in#icates that the inal statement is a concl(sion. DS(ite o ten> a line is (se# to separate premises rom concl(sions as %ell.E Mne reason that translation is important is $eca(se %e %ant to 4no% %hen arg(ments state# in nat(ral lang(age are logically )ali#> an# %e can (se propositional logic to 4no% %hen an arg(ment is logically )ali#. *he arg(ment gi)en a$o)e (ses an arg(ment orm calle# @modus ponens.A 0)ery arg(ment %ith this orm is logically )ali#. To( can replace @,A an# @6A %ith any t%o propositions> an# the arg(ment %ill $e logically )ali#. For e.ample> @,A co(l# stan# or @all #ogs are mammalsA an# @6A co(l# stan# or @all mammals are animals.A *he arg(ment %o(l# then $e @,ll #ogs are mammals. I all #ogs are mammals> then all #ogs are animals. *here ore> all #ogs are animals.A It is important that #e#(cti)e arg(ments are logically )ali#Fi %e 4no% a logically )ali# #e#(cti)e arg(ment has tr(e premises> then %e sho(l# agree %ith the concl(sion. ;o%e)er> not all logically )ali# arg(ments are goo# arg(ments. Mne o the most important reasons is that logical )ali#ity only concerns the logical orm o the arg(ment. , logically )ali# arg(ment can ha)e alse premises.

15

Chapter &: Tr th Tables


*r(th ta$les are )is(al ai#s to help (s #etermine all the tr(th )al(e possi$ilities o )ario(s statements. *r(th ta$les are (se# to #e ine logical connecti)es> an# to help (s i#enti y )ario(s #istinctions Ds(ch as ta(tologies> sel /contra#ictions> consistent statements> e1(i)alent statements> an# )ali# arg(mentsE.

Logical connecti!es
*he i)e connecti)es (se# in propositional logic are the ollo%ing: @an#A DE> @notA DNE> @orA DE> @impliesA DCE> an# @i an# only i A DOE. 0ach o the logical connecti)es has a precise #e inition> %hich is pro)i#e# $y a tr(th ta$le: Con$unction p * * F F 1 F F p1 F F

* * * F

Lo%er case letters represent @pre#icate constants.A *hese lo%er case letters stan# or any possi$le statement> s(ch as @roc4s e.istA or @i roc4s e.ist> then $ananas are pin4.A *he irst ro% contains )ario(s statements D@p>A @1>A an# @p 1AE. @p 1A ro(ghly translates to mean @$oth p an# 1.A For e.ample> @pA can mean @roc4s e.istA an# @1A can mean @$ananas e.ist.A In that case @p 1A means @roc4s an# $ananas e.ist.A *he ollo%ing image tells (s ho% to rea# the tr(th ta$le:

1'

*here is a col(mn D)ertical areaE (n#er each statement> %hich contains e)ery possi$le tr(th )al(e. *he col(mn (n#er @pA has @*> *> F> FA Dtr(e> tr(e> alse> alseE. *he col(mn (n#er @1A is @*> F> *> FA Dtr(e> alse> tr(e> alseE. *he col(mn (n#er @p 1A contains @*> F> F> FA Dtr(e> alse> alse> alseE. 0)ery ro% Dhori?ontal areaE $eneath the statements contains e)ery com$ination o tr(th )al(es. *he irst ro% o tr(th )al(es states that @p>A @1>A an# @p 1A are all tr(e. *he secon# ro% states that @pA is tr(e> @1A is alse> an# @p 1A is alse. *he thir# states that @pA is alse> @1A is tr(e> an# @p 1A is alse. *he o(rth states that @p>A @1A an# @p 1A are all alse. We can replace @pA an# @1A %ith statements o the 0nglish lang(age to clari y ho% the tr(th ta$le %or4s. @pA co(l# stan# or @li e (se# to e.ist on 3arsA an# @1A co(l# stan# or @li e %ill e.ist on 3ars in the (t(re.A We #on"t c(rrently 4no% i either o those statements are tr(e> $(t %e can tal4 a$o(t all the possi$ilities. Li e (se# to e.ist on 3ars. Li e %ill e.ist on 3ars in the (t(re. Li e (se# to e.ist on 3ars an# li e %ill e.ist on 3ars in the (t(re. * * F F * F * F * F F F

0ach ro% states the ollo%ing possi$ilities: Row (: It"s tr(e that @li e (se# to e.ist on 3ars.A It"s tr(e that @li e %ill e.ist on 3ars in the (t(re.A In that case it"s also tr(e that @Li e (se# to e.ist on 3ars> an# that li e %ill e.ist on 3ars in the (t(re.A Row ): It"s tr(e that @li e (se# to e.ist on 3ars.A It"s alse that @li e %ill e.ist on 3ars in the (t(re.A In that case it"s also alse that @Li e (se# to e.ist on 3ars> an# that li e %ill e.ist on 3ars in the (t(re.A Row *: It"s alse that @li e (se# to e.ist on 3ars.A It"s tr(e that @li e %ill e.ist on 3ars in the (t(re.A In that case it"s also alse that @Li e (se# to e.ist on 3ars> an# that li e %ill e.ist on 3ars in the (t(re.A Row +: It"s alse that @li e (se# to e.ist on 3ars.A It"s alse that @li e %ill e.ist on 3ars in the (t(re.A In that case it"s also alse that @Li e (se# to e.ist on 3ars> an# that li e %ill e.ist on 3ars in the (t(re.A

*he tr(th ta$le ma4es it clear that @p 1A is only tr(e %hen $oth @pA is tr(e an# @1A is tr(e. For e.ample> @h(mans are mammals an# they are animalsA is tr(e $eca(se @h(mans are mammalsA is tr(e an# @h(mans are animalsA is tr(e. Consi#er %hat happens i one o these statements is alse. @pA can mean @h(mans are reptilesA an# @1A can mean @h(mans are animals.A In that case %e %ill ha)e the statement @h(mans are reptiles an# they"re animals.A *hat statement is alse $eca(se one part o the statement is alse.

18

#egation p * F Np F *

:o(ghly spea4ing> @pA is any possi$le statement an# @NpA means @it"s not the case that p.A 0ach $o. on the top ro% contains a logical statement. DIn this case @pA an# @Np.AE 0ach $o. $elo% a statement tells (s the possi$le tr(th )al(es o that statement. @pA can $e tr(e or alse> an# @NpA can $e alse or tr(e. 0ach ro% o $o.es $elo% the logical statements contains the possi$le com$inations o tr(th )al(es o the statements a$o)e. *he irst ro% #o%n says @pA is tr(e an# @NpA is alse. Whene)er @pA is tr(e> @NpA %ill $e alse. For e.ample> @pA can stan# or @roc4s e.ist.A In that case the statement is tr(e> an# @NpA is alse $eca(se it stan#s or @it"s not the case that roc4s e.ist.A *he inal ro% says @pA is alse an# @NpA is tr(e. Whene)er @pA is alse> @NpA %ill $e tr(e. For e.ample> @pA co(l# stan# or @1I1J3>A %hich is alse. In that case @NpA is tr(e $eca(se it means @it"s not the case that 1I1J3.A *r(th ta$les pro)i#e e)ery possi$le com$ination o tr(th )al(es that logical statements can ha)e. *he only t%o tr(th )al(es nee#e# here are tr(e an# alse> so there are only t%o ro%s $eneath the logical statements. %is$unction p * * F F 1 F F p1 * F

* * * *

@p 1A ro(ghly translates as @either p or 1.A For e.ample> @pA can $e @#ogs are mammalsA an# @1A can $e @#ogs are reptiles.A In that case @p or 1A %ill $e @#ogs are mammals or #ogs are reptiles.A *he tr(th ta$le in#icates that e)ery @p 1A statement is tr(e (nless $oth @pA an# @1A are alse> %hich is sho%n on the inal ro% #o%n. For e.ample @pA can $e @#ogs are reptilesA an# @1A can $e @#ogs are li?ar#s.A In that case @p 1A stan#s or @either #ogs are reptiles or they"re li?ar#s.A *hat statement is alse.

20

Conditional p * * F F 1 F F pC1 F *

* * * *

@p C 1A ro(ghly translates as @i p> then 1.A For e.ample> @i h(mans are mammals> then h(mans are animals.A *he tr(th ta$le in#icates that @p C 1A is tr(e (nless @pA is tr(e an# @1A is alse. @pA can $e @the =resi#ent o the K+, is a h(manA an# @pA can $e @the =resi#ent o the K+, is a reptile.A In that case @p C 1A %ill mean @i the =resi#ent o the K+, is a h(man> then the =resi#ent o the K+, is a reptile.A *hat statement is alse. We can also consi#er a tr(e statement %here @pA is alse an# @1A is alse. For e.ample> @pA can $e @the =resi#ent o the K+, is a li?ar#A an# @1A can $e @the =resi#ent o the K+, is a reptile.A In that case @p C 1A %ill $e @i the =resi#ent o the K+, is a li?ar#> then the =resi#ent o the K+, is a reptile.A *hat statement is tr(e. Finally> let"s consi#er a con#itional statement %here @pA is alse an# @1A is tr(e. @pA can stan# or @the =resi#ent o the K+, is a li?ar#A an# @1A can stan# or @the =resi#ent o the K+, is an animal.A In that case the statement is @i the =resi#ent o the K+, is a li?ar#> then the =resi#ent o the K+, is an animal.A *hat statement is tr(e. E&uivalence p * * F F 1 F F pO1 F *

* * * F

@p O 1A ro(ghly translates as @p i an# only i 1.A For e.ample> @pA can stan# or @1I2J3A an# @1A can stan# or @2I1J3.A In that case @p O 1A stan#s or @1I2J3 i an# only i 2I1J3.A *he ta$le a$o)e ma4es it clear that @p O 1A is only tr(e %hen @pA an# @1A ha)e the same tr(th )al(es. *hey m(st $oth $e tr(e or alse. I not> the statement is alse. Consi#er %hen @pA stan#s or @#ogs are animalsA an# @1A stan#s or @#ogs are reptiles.A In that case @p O 1A stan#s or @#ogs are animals i an# only i #ogs are reptiles.A *hat statement is alse.

21

% comple' tr th table
We can (se the a$o)e tr(th ta$les to create a more comple. tr(th ta$le. For e.ample> @+ocrates is a h(man an# he"s not a )ampire.A We can translate this into logical orm as @, N6.A DWe (se capital letters $eca(se they stan# or something speci ic.E In this case @,A stan#s or @+ocrates is a h(manA an# @6A stan#s or @+ocrates is not a )ampire.A *he tr(th ta$le or this is the ollo%ing: , 6 N6 , N6 * * F F * F F F * * * F F * F F

We ha)e @N6A on the tr(th ta$le $eca(se simpler statements m(st $e resol)e# $e ore %e can in# the tr(th )al(es o more complicate# statements Dthat is in less parenthesesE. @N6A is containe# in @, N6.A *he tr(th )al(e or @N6A is the opposite o the tr(th )ale or @6>A so %e G(st %rite in the opposite )al(es there. *he tr(th )al(e or @, N6A %ill $e tr(e %hene)er $oth @,A an# @N6A are tr(e. *here"s only one place on the tr(th ta$le %here @, N6A is tr(e. We can highlight %here @, N6A is tr(e $eca(se it"s the same ro% %here @,A an# @N6A are tr(e: , 6 N6 , N6 * * F F * F F F * * * F F * F F

Ta tologies
, ta(tology is a statement that"s al%ays tr(e $eca(se o it"s logical orm. For e.ample> @there are li e orms on other planets or there are no li e orms on other planets.A *hat statement has the orm @= N=.A We can i#enti y a ta(tology $y loo4ing at the tr(th ta$le $eca(se all the possi$le tr(th )al(es o a ta(tology are tr(e. When loo4ing at all the a$o)e tr(th ta$les> yo( %ill notice that none o the 22

statements are ta(tologies $eca(se there"s al%ays one possi$ility o each o the a$o)e statements to $e alse. , tr(th ta$le o a ta(tology: = N= = N= * F F * * *

@= N=A is tr(e %hene)er at least on o those statements is tr(e. *hat"s %hy it"s al%ays tr(eF Whene)er = is alse> N= is tr(e an# )ise )ersa.

$elf(contradictions
, sel /contra#iction is a statement that"s al%ays alse. For e.ample> @there are li e orms on other planets an# there are no li e orms on other planets.A *hat statement has the orm @, N,.A We can i#enti y a sel /contra#iction on a tr(th ta$le $y seeing %hen a statement is al%ays alse. 9one o the a$o)e tr(th ta$les contain sel /contra#ictions $eca(se none o those statements are al%ays alse. , tr(th ta$le o a sel /contra#iction: = N= = N= * F F * F F

@= N=A is alse %hene)er either @=A or @N=A is alse. Mne o those simple statements is al%ays alse.

Consistent statements
+tatements are logically consistent as long as they can all $e tr(e at the same time> an# contra#ictory Dor inconsistentE %hene)er they can"t $e. @=A an# @SA are consistent $eca(se it"s possi$le they are $oth tr(e> $(t @=A an# @N=A are inconsistent $eca(se it"s not possi$le that they"re $oth tr(e. Consi#er the ollo%ing t%o statements: N= C S ND= SE

23

We can ma4e a tr(th ta$le or them: = S N= N= C S = S ND= SE * * F * F F F * * F F * * * * F * * * F F F F *

@N=A m(st $e gi)en tr(th )al(es $e ore %e can in# the tr(th )al(es o @N= C SA $eca(se it is part o that more comple. statement. We nee# to in# the tr(th )al(es or @= SA $e ore @ND= SEA $eca(se it"s also part o that more comple. statement. @N= C SA %ill only $e alse %hen @N=A is tr(e an# @SA is alse. *hat only happens on the $ottom ro%. @ND= SEA %ill only $e tr(e %hen @= SA is alse. *hat only happens on the $ottom ro% as %ell. *he tr(th ta$le a$o)e sho%s that @N= C SA an# @ND= SEA are contra#ictory statements $eca(se they"re ne)er $oth tr(e at the same time.

)* i!alent statements
01(i)alent statements al%ays ha)e the same tr(th )al(es. For e.ample> @all h(mans are mammals an# all h(mans are animalsA is logically e1(i)alent to @all h(mans are animals an# all h(mans are mammals.A @= SA is logically e1(i)alent to @S =.A Consi#er the ollo%ing t%o statements: ND= SE N= NS We can ma4e a tr(th ta$le or them: = S = S ND= SE N= NS N= NS * * * * F F F * F F F F F * * * F F * * F * F * F * * *

*his tr(th ta$le sho%s %hy they are e1(i)alentFthey al%ays ha)e the same tr(th )al(es. Whene)er @ND= SEA is alse @N= NSA is also alse> an# they are $oth al%ays tr(e at the same time.

2!

+alid arg ments


,n arg(ment is logically )ali# %hene)er it"s impossi$le or all the premises to $e tr(e an# the concl(sion alse at the same time. *hey are logically in)ali# %hene)er it is possi$le or all the premises to $e tr(e an# the concl(sion to $e alse at the same time. Argument ( Consi#er the ollo%ing arg(ment: 1. +ocrates is a h(man an# +ocrates isn"t a )ampire. 2. I +ocrates is a h(man an# +ocrates isn"t a )ampire> then +ocrates is mortal. 3. *here ore> +ocrates is mortal. We can translate this arg(ment into the ollo%ing logical statements: , N6 D, N6E C C C 0ach letter stan#s or a speci ic statement: ,: +ocrates is a h(man. 6: +ocrates is a )ampire. C: +ocrates is mortal. *he tr(th ta$le or this arg(ment is the ollo%ing: , 6 C N6 , N6 D, N6E C C C * * * F * * F F * F * * * F F * F * * F F * F F F F * * F F F * F F * * F F F F * * * F * * * * * F * F * F * F

*here"s only one spot on the ta$le %here $oth premises are tr(e> an# the concl(sion is also tr(e. D+ee the highlighte# area.E *here ore> this arg(ment is logically )ali#.

2-

Argument ) Consi#er the ollo%ing arg(ment: 1. +ocrates is an animal. 2. I +ocrates is a mammal> then +ocrates is an animal. 3. *here ore> +ocrates is a mammal. We can translate this arg(ment into logical orm: , 6C, 6 *he tr(th ta$le or this arg(ment is the ollo%ing: , 6 6 C, 6 * * * * F * F * F F F * * F * F

@6 C ,A is only alse %hen @6A is tr(e an# @,A is alse. *hat"s only on the secon# ro% rom the $ottom. *his tr(th ta$le pro)es that the arg(ment is in)ali# $eca(se there"s a ro% o tr(e premises an# a alse concl(sion.

2&

Chapter ,: -a.ing yo r own tr th tables


We can create o(r o%n tr(th ta$les (sing ollo%ing steps: 1. 2. 3. !. -. &. *ranslate statements o or#inary lang(age. 6rea4 all comple. statements into smaller parts. 7etermine ho% many col(mns are re1(ire#. 7etermine ho% many ro%s are re1(ire#. 7etermine the tr(th )al(es o statement letters. 7etermine the tr(th )al(es o comple. statements.

I %ill ill(strate ho% to ollo% these steps $y (sing an e.ample.

$tep 1: Translate statements of ordinary lang age/


We can ma4e a tr(th ta$le o one or more statements. ,rg(ments can $e presente# on a tr(th ta$le $y ha)ing all the statements rom the arg(ment. We %ill ma4e a tr(th ta$le o an arg(ment> $(t that re1(ires that %e irst translate an arg(ment o or#inary lang(age into propositional logic. Consi#er the ollo%ing arg(ment %ritten in 0nglish: I 3ary %ill either go to the store or stay at home> then 3ar4 %ill go to the store an# $(y mil4. 3ary %ill not go to the store or stay at home. *here ore> it"s not the case that 3ar4 %ill go to the store an# $(y mil4. We can i#enti y the premises an# concl(sion: 1. =remise 1: I 3ary %ill either go to the store or stay at home> then 3ar4 %ill go to the store. 2. =remise 2: 3ary %ill not go to the store or stay at home. 3. Concl(sion: It"s not the case that 3ar4 %ill go to the store. We can then replace each statement %ith a propositional )aria$le an# translate these statements into propositional logic: D, 6E C C ND, 6E NC ,: 3ary %ill go to the store. 6: 3ary %ill stay at home. C: 3ar4 %ill go to the store

25

$tep 2: 0rea. all comple' statements into smaller parts/


We can $rea4 statements into smaller parts $y remo)ing logical connecti)es. 0ach logical connecti)e m(st $e remo)e# one at a time (ntil only statement letters are le t. Whene)er a logical connecti)e is remo)e#> it sho(l# $e the one %ith the least amo(nt o parentheses aro(n# it. ,irst statement Loo4 at the irst propositional statement o the arg(ment: D, 6E C C First remo)e the con#itional DCE to create the ollo%ing simpler statement an# statement letter: , 6> C We still nee# to $rea4 @, 6A into smaller parts. 9e.t remo)e the #isG(nction DE to get the ollo%ing t%o statement letters: ,> 6 -econd statement Loo4 at the secon# propositional statement o the arg(ment: ND, 6E First %e nee# to remo)e the negation DNE to get the ollo%ing simpler statement: , 6 We still nee# to $rea4 @, 6A into smaller parts. We nee# to remo)e the #isG(nction to get the ollo%ing statement letters: ,> 6 Third statement Loo4 at the inal statement o the arg(ment: NC We nee# to remo)e the negation to get the ollo%ing statement letter: C

2'

.verview , list o all statements incl(#ing the smaller parts incl(#e the ollo%ing: 1. 2. 3. !. -. &. 5. D, 6E C C C , 6 , 6 ND, 6E NC

$tep ": 1etermine how many col mns are re* ired/
Col(mns are the )ertical areas o the tr(th ta$le. We nee# one col(mn or each statement o the arg(ment an# the smaller parts o those statements. I yo( loo4 at the list o all the statements an# smaller parts> yo( %ill see that there are nine statements that the ta$le %ill nee#. *he top ro% o the tr(th ta$le %ill contain these statements an# loo4 li4e the ollo%ing: , 6 C ,6 D, 6E C C ND, 6E NC

$tep #: 1etermine how many rows are re* ired/


:o%s are the hori?ontal areas o the tr(th ta$le. , n(m$er o ro%s is re1(ire# in or#er to #etermine e)ery possi$le com$ination o tr(th )al(es> %hich #epen#s on the n(m$er o statement letters (se#: Mne statement letter: 2 *%o statement letters: ! *hree statement letters: ' Fo(r statement letters: 1& Fi)e statement letters: 32 *he arg(ment (ses three statement letters D,> 6> CE> so %e nee# ' ro%s o tr(th )al(es.

28

*he tr(th ta$le %ill no% loo4 li4e the ollo%ing: , 6 C , 6 D, 6E C C ND, 6E

NC

$tep &: 1etermine the tr th !al es of statement letters/


*he last statement letter alternates rom $eing tr(e an# alse e)ery other ro% starting %ith tr(e D*> F> *> F...E. *he secon# to last statement alternates rom $eing tr(e an# alse e)ery t%o ro%s D*> *> F> F> etc.E *he thir# rom last statement alternates rom $eing tr(e an# alse e)ery o(r ro%s D*> *> *> *> F> F> F> F> etc.E *he irst statement o each ta$le sho(l# $e tr(e in the irst hal o the ro%s an# alse on the $ottom hal . We can no% #etermine %hen ,> 6> an# C are tr(e or alse: , 6 C , 6 * * * * * F * F * * F F F * * F * F F F * F F F D, 6E C C ND, 6E

NC

30

$tep ,: 1etermine the tr th !al es of comple' statements/


*here are tr(th ta$les that #etermine %hen a statement %ith a logical connecti)e is tr(e or alse. We can s(mmari?e these ta$les as saying the ollo%ing: ConG(nction DE L @p 1A is only tr(e %hen p is tr(e an# 1 is tr(e. 9egation DNE L @NpA has the opposite tr(th )al(e o @p.A 7isG(nction DE L @p 1A is only alse %hen $oth p an# 1 are alse. Con#itional DCE L @p C 1A is only alse %hen p is tr(e an# 1 is alse. 01(i)alence DOE L @p O 1A is only tr(e %hen p an# 1 ha)e the same tr(th )al(e.

*he irst statement %ith a connecti)e that %e nee# to #etermine the tr(th )al(es or is @, 6.A It %ill only $e alse %hen $oth , an# 6 are alse: , 6 C , 6 * * * * * * F * * F * * * F F * F * * * F * F * F F * F F F F F *he secon# statement %e nee# the tr(th )al(es or is @D, 6E C C.A It %ill only $e alse %hen @ , 6A is tr(e an# @CA is alse $eca(se it"s a con#itional: , 6 C , 6 * * * * * * F * * F * * * F F * F * * * F * F * F F * F F F F F D, 6E C C * * * F * F * * ND, 6E NC D, 6E C C ND, 6E NC

*he thir# statement %e nee# to in# tr(th )al(es or is @ND, 6E.A It %ill ha)e the opposite tr(th )al(es 31

as @, 6A $eca(se it"s a negation: , 6 C , 6 * * * * * * F * * F * * * F F * F * * * F * F * F F * F F F F F D, 6E C C * * * F * F * * ND, 6E F F F F F F * *

NC

*he o(rth an# inal statement %e nee# to in# tr(th )al(es or is @NC.A It %ill ha)e the opposite tr(th )al(es as @CA $eca(se it"s a negation: , 6 C , 6 * * * * * * F * * F * * * F F * F * * * F * F * F F * F F F F F D, 6E C C * * * F * F * * ND, 6E F F F F F F * * NC F * F * F * F *

*he reason %e (s(ally %ant to ma4e a tr(th ta$le o an arg(ment is to in# o(t i it"s logically )ali#. It is logically )ali# i it"s impossi$le or it to ha)e tr(e premises an# a alse concl(sion at the same time. It is logically in)ali# i it is possi$le or it to ha)e tr(e premises an# a alse concl(sion at the same time. We can loo4 at the tr(th ta$le to in# o(t i it"s )ali#. Loo4 at all the ro%s %here $oth premises are tr(e an# see i the concl(sion is e)er alse on those ro%s. I so> it"s in)ali#. Mther%ise it"s )ali#. *he premises are the ollo%ing: D, 6E C C ND, 6E *he concl(sion is the ollo%ing: 32

NC *he premises are only tr(e on the same ro% on the $ottom t%o ro%s> an# the concl(sion is alse on one o those ro%s. *here ore> the arg(ment is logically in)ali#: F F * F F F F F T T T T , T

33

Chapter 2: 3at ral ded ction


9at(ral #e#(ction is (se# to gi)e proo s o )ali#ity $y sho%ing all the steps in reasoning re1(ire#. In this case nat(ral #e#(ction (ses r(les o in erence to allo% (s to reach concl(sions rom statements o propositional logic.

4 les of inference
:(les o in erence incl(#e r(les o implication D)ali# arg(ment ormsE an# r(les o replacement Dstatements %ith e1(i)alent logical ormE. *he r(les are the ollo%ing: Rules of implication
'odus ponens /'P0 'odus tollens /'T0 Hypothetical syllogism /H-0 %is$unctive syllogism /%-0

pC1 p 1

pC1 N1 Np

pC1 1Cr r C s

p1 Np 1

Constructive %ilemma /C%0

-implification /-imp0

Con$unction /Con$0

Addition /Add0

pC1 rCs pr 1 s

p1 p

p 1 p 1

p p 1

0ach o these r(les states a )ali# arg(ment orm. For e.ample> modus ponens states that the ollo%ing arg(ment orm is )ali#: 1. p C 1 2. p 3. 1 @A merely states that the inal statement is a concl(sion. Whene)er %e 4no% @p C 1A an# @pA are tr(e> %e can concl(#e @1.A 0ach o these lo%er/case letters can stan# or any statement> no matter ho% comple.. 0ach letter m(st stan# or the same statement %hen concerning the same arg(ment. For e.ample> @pA can stan# or @all h(mans are mammalsA an# @1A can stan# or @all h(mans are li)ing organismsA or an entire arg(ment. In that case %e can #e)elop the ollo%ing )ali# arg(ment in the 3!

0nglish lang(age (sing modus ponens: 1. I all h(mans are mammals> then all h(mans are li)ing organisms. 2. ,ll h(mans are mammals. 3. *here ore> all h(mans are li)ing organisms. Rules of replacement
%e'organ1s Rule /%'0 Commutativity /Com0 NDp 1E :: DNp N1E NDp 1E :: DNp N1E Dp 1E :: D1 pE Dp 1E :: D1 pE Associativity /Assoc0 Pp D1 rEQ :: PDp 1E rQ Pp D1 rEQ :: PDp 1E rQ %istribution /%ist0 Pp D1 rEQ :: PDp 1E Dp rEQ Pp D1 rEQ :: PDp 1E Dp rEQ %ouble negation /%#0 p :: NNp

Transposition /Trans0 Dp C 1E :: DN1 C NpE

'aterial implication / mpl0 Dp C 1E :: DNp 1E

'aterial e&uivalence /E&uiv0 Dp O 1E :: Dp C 1E D1 C pE Dp O 1E :: Dp 1E DNp N1E

E2portation /E2p0 Tautology /Taut0 PDp 1E C rQ :: Pp C D1 C rEQ p :: Dp pE p :: Dp pE

0ach o the r(les o replacement state that certain statements are logically e1(i)alent. To( can )ali#ly concl(#e either part o a r(le o replacement rom the other partFthe irst part o the @::A sym$ol can $e concl(#e# rom the secon# part> an# )ise )ersa. For e.ample> #o($le negation states that @pA an# @NNpA are logically e1(i)alent. I yo( 4no% @p>A then yo( can concl(#e @NNpA an# )ice )ersa. @pA can stan# or @all h(mans are mammals.A In that case %e can concl(#e that @it"s not the case that "all h(mans are mammals" is alse.A +ome r(les o replacement act(ally ha)e more than one type o e1(i)alence. For e.ample> 7e3organ"s r(le has t%o #i erent %ays it can $e (se#.

5ow to constr ct a proof


=ro)ing )ali#ity (sing nat(ral #e#(ction re1(ires the ollo%ing steps: 1. Fin# the logical orm o an arg(ment. 2. Write the logical orm o the premises. 3. Write the logical orm o the concl(sion on the same line as the last premise a ter the @BA sym$ol. !. Kse those premises an# the r(les o in erence to reach the logical orm o the arg(ment"s concl(sion.

3-

I %ill present t%o e.amples o proo s. E2ample ( Consi#er the ollo%ing arg(ment: 1. 2. 3. !. ,ll h(mans are li?ar#s. I all h(mans are li?ar#s> then all h(mans are reptiles. I all h(mans are reptiles> then they are col#/$loo#e#. *here ore> all h(mans are col#/$loo#e#.

*he logical orm o this arg(ment is the ollo%ing: 1. 2. 3. !. = =CS SC: :

0ach letter is capitali?e# an# represents a speci ic statement in 0nglish: =: ,ll h(mans are li?ar#s. S: ,ll h(mans are reptiles. :: ,ll h(mans are col#/$loo#e#. *he proo that the arg(ment is )ali# loo4s li4e the ollo%ing: 1. = 2. = C S 3. S C : !. S -. : B: 1> 2> 3= 3> !> 3=

*he premises are %ritten on lines 1> 2> an# 3. *he concl(sion is %ritten a ter the inal premise on line 3. Line ! concl(#es @SA rom lines 1 an# 2 (sing modus ponens. @1> 2> 3=A is %ritten on the right/han# si#e to ma4e that clear. @=R = C SR SA is )ali# $eca(se it (ses modus ponens. DIt has the same orm as @pR p C 1R 1.AE Line - concl(#es @:A rom lines 3 an# ! (sing modus ponens. @3> !> 3=A is %ritten on the right/han# si#e to ma4e that clear. @S C :R S. :A is )ali# $eca(se it also (ses modus ponens. DIt has the same orm as @pR p C 1R 1.AE Line - has the concl(sion o the original arg(ment> so %e ha)e pro)en that the arg(ment is )ali#. We can #eri)e the concl(sion rom the premises an# r(les o in erence.

3&

E2ample ) Consi#er the ollo%ing arg(ment: 1. ,ll h(mans are mammals. 2. I all #ogs are %arm/$loo#e# an# ha)e tho(ghts> then it is not the case that all #ogs are reptiles or insects. 3. I all h(mans are col#/$loo#e#> then it is not the case that all h(mans are mammals. !. I it is not the case that all h(mans are col#/$loo#e#> then no h(mans are li?ar#s. -. *here ore> no h(mans are li?ar#s. *he logical orm o this arg(ment is the ollo%ing: 1. 2. 3. !. , PD6 CE C ND7 0EQ F C N, NF C G G

*he letters stan# or the ollo%ing: ,: ,ll h(mans are mammals. 6: ,ll #ogs are %arm/$loo#e#. C: ,ll #ogs ha)e tho(ghts. 7: ,ll #ogs are reptiles. 0: ,ll #ogs are insects. F: ,ll h(mans are col#/$loo#e#. G: 9o h(mans are li?ar#s. *he proo that this arg(ment is )ali# is the ollo%ing: 1. , PD6 CE C ND7 0EQ 2. F C N, 3. NF C G !. , -. NN, &. NF 5. G BG 1> +imp !> 79 2> -> 3* 3> &> 3=

Lines 1> 2> an# 3 contain the premises. *he concl(sion is also %ritten on line 3. Line ! concl(#es @,A $y simpli ication (sing line 1. @, PD6 CE C ND7 0EQR ,A is a )ali# arg(ment $eca(se it has the same orm as simpli ication. @, PD6 CE C ND7 0EQR ,A has the same orm as @p 1R p.A

35

Line - concl(#es @NN,A $y (sing line ! an# #o($le negation. @,R NN,A is a )ali# arg(ment $eca(se they are e1(i)alent. @, :: NN,A has the same orm as @p :: NNp.A Line & concl(#es @NFA $y (sing lines 2 an# -> an# modus tollens. @F C N,R NN,. NFA is a )ali# arg(ment $eca(se it has the same orm as modus tollens. @F C N,R NN,R NFA has the same orm as @p C 1R N1.R Np.A Line 5 concl(#es @GA $y (sing lines 3 an# &> an# modus ponens. @NF C GR NFR GA is a )ali# arg(ment $eca(se it has the same orm as modus ponens. @NF C GR NFR GA has the same orm as @p C 1R pR 1.A @GA is the concl(sion o the arg(ment> so the arg(ment is )ali#. We %ere a$le to #e#(ce the concl(sion rom the premises an# r(les o in erence.

3'

Chapter 6: Conditional 7 Indirect 8roof


*he straight or%ar# %ay to constr(ct proo s (sing nat(ral #e#(ction is calle# the @#irect metho#.A 0)ery line o that type o proo is )ali#ly #e#(ce# rom the premises an# r(les o in erence. 0)ery line o s(ch a proo co(l# $e consi#ere# to $e tr(e as long as %e consi#er the premises to $e tr(e. ;o%e)er> there are t%o other strategies: *he con#itional proo an# the in#irect proo . 6oth o these types o proo s intro#(ce an a##itional premise that is ass(me# to $e tr(e @ or the sa4e o arg(ment.A

Conditional proof
I a proo contains a con#itional statement as a premise> concl(sion> or as a pro)en statement $ase# on the premises> then %e can a## an a##itional premise a ter%ar#Fthe irst part o the con#itional. For e.ample> @,A is the irst part o the con#itional @, C 6.A Let"s ass(me @,A is a##e# as an @ass(mption or con#itional proo .A In that case the letters @,C=A are p(t on the right/han# si#e o the line it"s a##e# on. *his line can $e a##e# to the proo at any time> $(t it is o ten a##e# right a ter the other premises are state#. *he ass(me# premise is then (se# to #eri)e a con#itional statement. *hen once a conditional is #eri)e# (sing the ass(me# premise> %e ha)e a con#itional proo an# the inal line o the proo has @C=A on the right/han# si#e. ,ll lines (sing the ass(mption are also cite#. ;o%e)er> the con#itional proo is not necessarily the entirety o the proo Fmore might ha)e to $e pro)en to #e#(ce the concl(sion o the original arg(ment that %e %ant to pro)e to $e )ali#. *he lines o con#itional proo s that are (se# to #eri)e the con#itional concl(sion can"t $e (se# $y any other part o the proo . *hese lines re1(ire a tentati)e ass(mption an# are not )ali#ly #e#(ce# $y the premises o the original arg(ment. *he ass(mption itsel is not )ali#ly #e#(ce# rom the original arg(ment> an# the rest o the proo re1(ires that ass(mption> other than the concl(sion o the con#itional proo . E2ample Consi#er the ollo%ing arg(ment: 1. I all h(mans are mortal> then no h(mans are go#s an# no h(mans are angels. 2. I no h(mans are go#s> then no h(man is omnipotent an# no h(man is omniscient. 3. *here ore> either all h(mans are mortal or it"s not the case that no h(man is omnipotent. *his arg(ment has the ollo%ing orm in propositional logic: 1. , C D6 CE 2. 6 C D7 0E 3. N, 7

38

,: ,ll h(mans are mortal. 6: 9o h(mans are go#s. C: 9o h(mans are angels. 7: 9o h(man is omnipotent. 0: 9o h(man is omniscient. *he ollo%ing is the proo that this arg(ment is )ali# D(sing con#itional proo sE: 1. , C D6 CE 2. 6 C D7 0E *3 A +3 4 C 63 4 &. , C 6 73 4 83 % E 93 % 10. 6 C 7 11. , C 7 12. N, 7 B N, 7 ACP (5 *5 'P +5 -imp 3> !> -> C= ACP )5 75 'P 85 -imp 5> '> 8 C= 5> 10> ;+ 11> Impl

Lines 3> !> an# - contain a single con#itional proo . *he ass(mption or con#itional proo D,E is allo%e# $eca(se line 1 contains a con#itional> an# the irst part o the con#itional is @,.A Lines 5> '> an# 8 contain another con#itional proo . *he ass(mption or con#itional proo D6E is allo%e# $eca(se line 2 contains a con#itional> an# the irst part o the con#itional is @6.A 6eca(se lines 3/- an# 5/8 contain con#itional proo s> those lines are not act(ally )ali#ly #e#(ce# rom the original premises. I they %ere> line - %o(l# ha)e alrea#y pro)en @6>A so the secon# con#itional proo %o(l# ha)e $een (nnecessary. *hese lines are merely tentati)ely ass(me# to $e )ali#ly #e#(ce# or the sa4e o arg(ment. *he concl(sions o each con#itional proo are pro)en> an# they are (se# %ith a hypothetical syllogism to pro)e @, C 7.A We 4no% @, C 6R 6 C 7R , C 7A is a )ali# arg(ment $eca(se it has the same orm as the hypothetical syllogism Dp C 1R 1 C rR p C rAE. Finally> 4eep in min# that the concl(sion o o(r arg(ment can also G(sti y the (se o an ass(mption or a con#itional proo as long as it"s a con#itional statement. For e.ample> i an arg(ment concl(#es @, C D6 CE>A then %e can ha)e @,A as an ass(mption or a con#itional proo .

!0

Indirect proof
In#irect proo s are also 4no%n as a @reductio ad absurdumA Di.e. @re#(ction to the a$s(r#AE. In#irect proo s can $e (se# to pro)e any )ali# arg(ment is )ali#. In#irect proo s ha)e three a##itional steps: 1. *here"s an a##itional premiseFa statement that is tentati)ely ass(me# to $e tr(e. *his is the @ass(mption or in#irect proo A an# @,I=A is %ritten on the right/han# si#e o the line o the ass(mption. *his statement is one %e %ill act(ally hope to pro)e to $e alse. 2. *he ass(mption is (se# to #eri)e a contra#iction Dp NpE. *he contra#iction m(st appear on a line> an# it is o ten e.plicitly #eri)e# on a single line (sing the r(le o conG(nction. 3. Mnce the contra#iction is #eri)e#> the negation o the ass(mption is pro)en. @I=A is %ritten on the right/han# si#e o that line along %ith the n(m$ers o all lines that (se the ass(me# premise. Mnce again> the lines o an in#irect proo re1(ire tentati)e ass(mptions> an# they are not pro)en to $e )ali#ly #e#(ce# rom the original premises. *hey can"t $e (se# $y other parts o the proo or that reason. Mnly the inal line o an in#irect proo is act(ally pro)en Dan# it is )ali#ly #e#(ce# rom the original premisesE. 9ote that in#irect proo s o ten ass(me the negation o the concl(sion. Mnce it is pro)en that ass(ming the negation o the concl(sion lea#s to a contra#iction> the concl(sion is act(ally pro)en to $e tr(e. For e.ample> %e can ass(me that the concl(sion o modus ponens is alse. In that case %e ass(me @p C 1>A @pA an# @N1A to $e tr(e. 6(t @p C 1A an# @N1A pro)es @NpA to $e tr(e )ia modus tollens. We no% 4no% that the arg(ment is )ali# $eca(se ass(ming the concl(sion is alse lea#s to a contra#iction D@p NpAE. E2ample Consi#er the ollo%ing arg(ment: 1. 0ither 4illing people is sometimes %rong or al%ays %rong. 2. I the 4illing people is al%ays %rong> then 4illing people %hen necessary or sel /#e ense is %rong> $(t 4illing people %hen necessary or sel /#e ense is not %rong. 3. I 4illing people is sometimes %rong> then not all homici#e is m(r#er. !. *here ore> sometimes homici#e is not m(r#er. *he logical orm o this arg(ment is the ollo%ing: 1. 2. 3. !. =S = C D: N:E SC+ +

=: Hilling people is sometimes %rong. S: Hilling people is al%ays %rong. :: Hilling people %hen necessary or sel /#e ense is %rong. +: +ometimes homici#e is not m(r#er.

!1

, proo that this arg(ment is )ali# (sing an in#irect proo is the ollo%ing: 1. = S 2. = C D: N:E 3. S C + !. S = 63 :; <3 P 73 R :R '. S 8. + B+ 1> Com AP +5 65 %)5 <5 'P -> &> 5> I= 3> '> 3=

*he in#irect proo occ(rs on lines -> &> an# 5. It %as necessary to sho% that @SA is tr(e in or#er to (se line 3 DS C +E %ith modus ponens to reach the concl(sion D+E. 9ote that the lines o the in#irect proo are not act(ally ta4en to $e )ali#ly #e#(ce# in the long r(n. *hey are only tentati)ely ass(me# to $e )ali#ly #e#(ce#. *he ass(mption is act(ally the negation o %hat is pro)en.

!2

You might also like