BookI
1
LETusnowdiscusssophisticrefutations，i。e。whatappeartoberefutationsbutarereallyfallaciesinstead。Wewillbegininthenaturalorderwiththefirst。
Thatsomereasoningsaregenuine，whileothersseemtobesobutarenot，isevident。Thishappenswitharguments，asalsoelsewhere，throughacertainlikenessbetweenthegenuineandthesham。Forphysicallysomepeopleareinavigorouscondition，whileothersmerelyseemtobesobyblowingandriggingthemselvesoutasthetribesmendotheirvictimsforsacrifice；andsomepeoplearebeautifulthankstotheirbeauty，whileothersseemtobeso，bydintofembellishingthemselves。Soitis，too，withinanimatethings；
forofthese，too，somearereallysilverandothersgold，whileothersarenotandmerelyseemtobesuchtooursense；e。g。thingsmadeoflithargeandtinseemtobeofsilver，whilethosemadeofyellowmetallookgolden。Inthesamewaybothreasoningandrefutationaresometimesgenuine，sometimesnot，thoughinexperiencemaymakethemappearso：forinexperiencedpeopleobtainonly，asitwere，adistantviewofthesethings。Forreasoningrestsoncertainstatementssuchthattheyinvolvenecessarilytheassertionofsomethingotherthanwhathasbeenstated，throughwhathasbeenstated：refutationisreasoninginvolvingthecontradictoryofthegivenconclusion。Nowsomeofthemdonotreallyachievethis，thoughtheyseemtodosoforanumberofreasons；andofthesethemostprolificandusualdomainistheargumentthatturnsuponnamesonly。Itisimpossibleinadiscussiontobringintheactualthingsdiscussed：weusetheirnamesassymbolsinsteadofthem；andthereforewesupposethatwhatfollowsinthenames，followsinthethingsaswell，justaspeoplewhocalculatesupposeinregardtotheircounters。Butthetwocases（namesandthings）arenotalike。
Fornamesarefiniteandsoisthesum—totalofformulae，whilethingsareinfiniteinnumber。Inevitably，then，thesameformulae，andasinglename，haveanumberofmeanings。Accordinglyjustas，incounting，thosewhoarenotcleverinmanipulatingtheircountersaretakeninbytheexperts，inthesamewayinargumentstoothosewhoarenotwellacquaintedwiththeforceofnamesmisreasonbothintheirowndiscussionsandwhentheylistentoothers。Forthisreason，then，andforotherstobementionedlater，thereexistsbothreasoningandrefutationthatisapparentbutnotreal。Nowforsomepeopleitisbetterworthwhiletoseemtobewise，thantobewisewithoutseemingtobe（fortheartofthesophististhesemblanceofwisdomwithoutthereality，andthesophistisonewhomakesmoneyfromanapparentbutunrealwisdom）；forthem，then，itisclearlyessentialalsotoseemtoaccomplishthetaskofawisemanratherthantoaccomplishitwithoutseemingtodoso。Toreduceittoasinglepointofcontrastitisthebusinessofonewhoknowsathing，himselftoavoidfallaciesinthesubjectswhichheknowsandtobeabletoshowupthemanwhomakesthem；andoftheseaccomplishmentstheonedependsonthefacultytorenderananswer，andtheotheruponthesecuringofone。Those，then，whowouldbesophistsareboundtostudytheclassofargumentsaforesaid：foritisworththeirwhile：forafacultyofthiskindwillmakeamanseemtobewise，andthisisthepurposetheyhappentohaveinview。
Clearly，then，thereexistsaclassofargumentsofthiskind，anditisatthiskindofabilitythatthoseaimwhomwecallsophists。
Letusnowgoontodiscusshowmanykindsthereareofsophisticalarguments，andhowmanyinnumberaretheelementsofwhichthisfacultyiscomposed，andhowmanybranchestherehappentobeofthisinquiry，andtheotherfactorsthatcontributetothisart。
2
Ofargumentsindialogueformtherearefourclasses：
Didactic，Dialectical，Examination—arguments，andContentiousarguments。Didacticargumentsarethosethatreasonfromtheprinciplesappropriatetoeachsubjectandnotfromtheopinionsheldbytheanswerer（forthelearnershouldtakethingsontrust）：
dialecticalargumentsarethosethatreasonfrompremissesgenerallyaccepted，tothecontradictoryofagiventhesis：
examination—argumentsarethosethatreasonfrompremisseswhichareacceptedbytheanswererandwhichanyonewhopretendstopossessknowledgeofthesubjectisboundtoknow—inwhatmanner，hasbeendefinedinanothertreatise：contentiousargumentsarethosethatreasonorappeartoreasontoaconclusionfrompremissesthatappeartobegenerallyacceptedbutarenotso。Thesubject，then，ofdemonstrativeargumentshasbeendiscussedintheAnalytics，whilethatofdialecticargumentsandexamination—argumentshasbeendiscussedelsewhere：letusnowproceedtospeakoftheargumentsusedincompetitionsandcontests。
3
Firstwemustgraspthenumberofaimsentertainedbythosewhoargueascompetitorsandrivalstothedeath。Thesearefiveinnumber，refutation，fallacy，paradox，solecism，andfifthlytoreducetheopponentinthediscussiontobabbling—i。e。toconstrainhimtorepeathimselfanumberoftimes：oritistoproducetheappearanceofeachofthesethingswithoutthereality。Fortheychooseifpossibleplainlytorefutetheotherparty，orasthesecondbesttoshowthatheiscommittingsomefallacy，orasathirdbesttoleadhimintoparadox，orfourthlytoreducehimtosolecism，i。e。
tomaketheanswerer，inconsequenceoftheargument，touseanungrammaticalexpression；or，asalastresort，tomakehimrepeathimself。
4
Therearetwostylesofrefutation：forsomedependonthelanguageused，whilesomeareindependentoflanguage。Thosewaysofproducingthefalseappearanceofanargumentwhichdependonlanguagearesixinnumber：theyareambiguity，amphiboly，combination，divisionofwords，accent，formofexpression。Ofthiswemayassureourselvesbothbyinduction，andbysyllogisticproofbasedonthis—anditmaybeonotherassumptionsaswell—thatthisisthenumberofwaysinwhichwemightfalltomeanthesamethingbythesamenamesorexpressions。Argumentssuchasthefollowingdependuponambiguity。’Thoselearnwhoknow：foritisthosewhoknowtheirletterswholearnthelettersdictatedtothem’。Forto’learn’isambiguous；itsignifiesboth’tounderstand’bytheuseofknowledge，andalso’toacquireknowledge’。Again，’Evilsaregood：
forwhatneedstobeisgood，andevilsmustneedsbe’。For’whatneedstobe’hasadoublemeaning：itmeanswhatisinevitable，asoftenisthecasewithevils，too（forevilofsomekindisinevitable），whileontheotherhandwesayofgoodthingsaswellthatthey’needtobe’。Moreover，’Thesamemanisbothseatedandstandingandheisbothsickandinhealth：foritishewhostoodupwhoisstanding，andhewhoisrecoveringwhoisinhealth：butitistheseatedmanwhostoodup，andthesickmanwhowasrecovering’。For’Thesickmandoessoandso’，or’hassoandsodonetohim’isnotsingleinmeaning：sometimesitmeans’themanwhoissickorisseatednow’，sometimes’themanwhowassickformerly’。
Ofcourse，themanwhowasrecoveringwasthesickman，whoreallywassickatthetime：butthemanwhoisinhealthisnotsickatthesametime：heis’thesickman’inthesensenotthatheissicknow，butthathewassickformerly。Examplessuchasthefollowingdependuponamphiboly：’Iwishthatyoutheenemymaycapture’。Alsothethesis，’Theremustbeknowledgeofwhatoneknows’：foritispossiblebythisphrasetomeanthatknowledgebelongstoboththeknowerandtheknown。Also，’Theremustbesightofwhatonesees：oneseesthepillar：ergothepillarhassight’。Also，’Whatyouprofessto—be，thatyouprofessto—be：youprofessastoneto—be：ergoyouprofess—to—beastone’。Also，’Speakingofthesilentispossible’：
for’speakingofthesilent’alsohasadoublemeaning：itmaymeanthatthespeakerissilentorthatthethingsofwhichhespeaksareso。Therearethreevarietiesoftheseambiguitiesandamphibolies：
（1）Wheneithertheexpressionorthenamehasstrictlymorethanonemeaning，e。g。aetosandthe’dog’；（2）whenbycustomweusethemso；（3）whenwordsthathaveasimplesensetakenalonehavemorethanonemeaningincombination；e。g。’knowingletters’。Foreachword，both’knowing’and’letters’，possiblyhasasinglemeaning：butbothtogetherhavemorethanone—eitherthatthelettersthemselveshaveknowledgeorthatsomeoneelsehasitofthem。
Amphibolyandambiguity，then，dependonthesemodesofspeech。Uponthecombinationofwordstheredependinstancessuchasthefollowing：
’Amancanwalkwhilesitting，andcanwritewhilenotwriting’。Forthemeaningisnotthesameifonedividesthewordsandifonecombinestheminsayingthat’itispossibletowalk—while—sitting’
andwritewhilenotwriting]。Thesameappliestothelatterphrase，too，ifonecombinesthewords’towrite—while—not—writing’：forthenitmeansthathehasthepowertowriteandnottowriteatonce；
whereasifonedoesnotcombinethem，itmeansthatwhenheisnotwritinghehasthepowertowrite。Also，’Henowifhehaslearnthisletters’。Moreover，thereisthesayingthat’Onesinglethingifyoucancarryacrowdyoucancarrytoo’。
Upondivisiondependthepropositionsthat5is2and3，andodd，andthatthegreaterisequal：foritisthatamountandmorebesides。
Forthesamephrasewouldnotbethoughtalwaystohavethesamemeaningwhendividedandwhencombined，e。g。’Imadetheeaslaveonceafreeman’，and’God—likeAchillesleftfiftyahundredmen’。
Anargumentdependinguponaccentitisnoteasytoconstructinunwrittendiscussion；inwrittendiscussionsandinpoetryitiseasier。Thus（e。g。）somepeopleemendHomeragainstthosewhocriticizeasunnaturalhisexpressiontomenoukataputhetaiombro。Fortheysolvethedifficultybyachangeofaccent，pronouncingtheouwithanacuteraccent。Also，inthepassageaboutAgamemnon’sdream，theysaythatZeusdidnothimselfsay’Wegranthimthefulfilmentofhisprayer’，butthathebadethedreamgrantit。Instancessuchasthese，then，turnupontheaccentuation。
Otherscomeaboutowingtotheformofexpressionused，whenwhatisreallydifferentisexpressedinthesameform，e。g。amasculinethingbyafemininetermination，orafemininethingbyamasculine，oraneuterbyeitheramasculineorafeminine；or，again，whenaqualityisexpressedbyaterminationpropertoquantityorviceversa，orwhatisactivebyapassiveword，orastatebyanactiveword，andsoforthwiththeotherdivisionspreviously’laiddown。Foritispossibletouseanexpressiontodenotewhatdoesnotbelongtotheclassofactionsatallasthoughitdidsobelong。Thus（e。g。）
’flourishing’isawordwhichintheformofitsexpressionislike’cutting’or’building’：yettheonedenotesacertainquality—i。e。
acertaincondition—whiletheotherdenotesacertainaction。Inthesamemanneralsointheotherinstances。
Refutations，then，thatdependuponlanguagearedrawnfromthesecommon—placerules。Offallacies，ontheotherhand，thatareindependentoflanguagetherearesevenkinds：
（1）thatwhichdependsuponAccident：
（2）theuseofanexpressionabsolutelyornotabsolutelybutwithsomequalificationofrespectorplace，ortime，orrelation：
（3）thatwhichdependsuponignoranceofwhat’refutation’is：
（4）thatwhichdependsupontheconsequent：
（5）thatwhichdependsuponassumingtheoriginalconclusion：
（6）statingascausewhatisnotthecause：
（7）themakingofmorethanonequestionintoone。
5
Fallacies，then，thatdependonAccidentoccurwheneveranyattributeisclaimedtobelonginlikemannertoathingandtoitsaccident。Forsincethesamethinghasmanyaccidentsthereisnonecessitythatallthesameattributesshouldbelongtoallofathing’spredicatesandtotheirsubjectaswell。Thus（e。g。），’IfCoriscusbedifferentfrom\"man\"，heisdifferentfromhimself：forheisaman’：or’IfhebedifferentfromSocrates，andSocratesbeaman，then’，theysay，’hehasadmittedthatCoriscusisdifferentfromaman，becauseitsohappens（accidit）thatthepersonfromwhomhesaidthathe（Coriscus）isdifferentisaman’。
Thosethatdependonwhetheranexpressionisusedabsolutelyorinacertainrespectandnotstrictly，occurwheneveranexpressionusedinaparticularsenseistakenasthoughitwereusedabsolutely，e。g。intheargument’Ifwhatisnotistheobjectofanopinion，thenwhatisnotis’：foritisnotthesamething’tobex’and’tobe’
absolutely。Oragain，’Whatis，isnot，ifitisnotaparticularkindofbeing，e。g。ifitisnotaman。’Foritisnotthesamething’nottobex’and’nottobe’atall：itlooksasifitwere，becauseoftheclosenessoftheexpression，i。e。because’tobex’
isbutlittledifferentfrom’tobe’，and’nottobex’from’nottobe’。Likewisealsowithanyargumentthatturnsuponthepointwhetheranexpressionisusedinacertainrespectorusedabsolutely。Thuse。g。’SupposeanIndiantobeblackallover，butwhiteinrespectofhisteeth；thenheisbothwhiteandnotwhite。’Orifbothcharactersbelonginaparticularrespect，then，theysay，’contraryattributesbelongatthesametime’。Thiskindofthingisinsomecaseseasilyseenbyanyone，e。g。supposeamanweretosecurethestatementthattheEthiopianisblack，andwerethentoaskwhetherheiswhiteinrespectofhisteeth；andthen，ifhebewhiteinthatrespect，weretosupposeattheconclusionofhisquestionsthatthereforehehadproveddialecticallythathewasbothwhiteandnotwhite。Butinsomecasesitoftenpassesundetected，viz。inallcaseswhere，wheneverastatementismadeofsomethinginacertainrespect，itwouldbegenerallythoughtthattheabsolutestatementfollowsaswell；andalsoinallcaseswhereitisnoteasytoseewhichoftheattributesoughttoberenderedstrictly。Asituationofthiskindarises，whereboththeoppositeattributesbelongalike：forthenthereisgeneralsupportfortheviewthatonemustagreeabsolutelytotheassertionofboth，orofneither：e。g。ifathingishalfwhiteandhalfblack，isitwhiteorblack？
Otherfallaciesoccurbecausetheterms’proof’or’refutation’havenotbeendefined，andbecausesomethingisleftoutintheirdefinition。Fortorefuteistocontradictoneandthesameattribute—notmerelythename，butthereality—andanamethatisnotmerelysynonymousbutthesamename—andtoconfuteitfromthepropositionsgranted，necessarily，withoutincludinginthereckoningtheoriginalpointtobeproved，inthesamerespectandrelationandmannerandtimeinwhichitwasasserted。A’falseassertion’aboutanythinghastobedefinedinthesameway。Somepeople，however，omitsomeoneofthesaidconditionsandgiveamerelyapparentrefutation，showing（e。g。）thatthesamethingisbothdoubleandnotdouble：fortwoisdoubleofone，butnotdoubleofthree。Or，itmaybe，theyshowthatitisbothdoubleandnotdoubleofthesamething，butnotthatitissointhesamerespect：
foritisdoubleinlengthbutnotdoubleinbreadth。Or，itmaybe，theyshowittobebothdoubleandnotdoubleofthesamethingandinthesamerespectandmanner，butnotthatitissoatthesametime：
andthereforetheirrefutationismerelyapparent。Onemight，withsomeviolence，bringthisfallacyintothegroupoffallaciesdependentonlanguageaswell。
Thosethatdependontheassumptionoftheoriginalpointtobeproved，occurinthesameway，andinasmanyways，asitispossibletobegtheoriginalpoint；theyappeartorefutebecausemenlackthepowertokeeptheireyesatonceuponwhatisthesameandwhatisdifferent。
Therefutationwhichdependsupontheconsequentarisesbecausepeoplesupposethattherelationofconsequenceisconvertible。Forwhenever，supposeAis，Bnecessarilyis，theythensupposealsothatifBis，Anecessarilyis。Thisisalsothesourceofthedeceptionsthatattendopinionsbasedonsense—perception。Forpeopleoftensupposebiletobehoneybecausehoneyisattendedbyayellowcolour：also，sinceafterrainthegroundiswetinconsequence，wesupposethatifthegroundiswet，ithasbeenraining；whereasthatdoesnotnecessarilyfollow。Inrhetoricproofsfromsignsarebasedonconsequences。Forwhenrhetoricianswishtoshowthatamanisanadulterer，theytakeholdofsomeconsequenceofanadulterouslife，viz。thatthemanissmartlydressed，orthatheisobservedtowanderaboutatnight。Thereare，however，manypeopleofwhomthesethingsaretrue，whilethechargeinquestionisuntrue。Ithappenslikethisalsoinrealreasoning；
e。g。Melissus’argument，thattheuniverseiseternal，assumesthattheuniversehasnotcometobe（forfromwhatisnotnothingcouldpossiblycometobe）andthatwhathascometobehasdonesofromafirstbeginning。If，therefore，theuniversehasnotcometobe，ithasnofirstbeginning，andisthereforeeternal。Butthisdoesnotnecessarilyfollow：forevenifwhathascometobealwayshasafirstbeginning，itdoesnotalsofollowthatwhathasafirstbeginninghascometobe；anymorethanitfollowsthatifamaninafeverbehot，amanwhoishotmustbeinafever。
Therefutationwhichdependsupontreatingascausewhatisnotacause，occurswheneverwhatisnotacauseisinsertedintheargument，asthoughtherefutationdependeduponit。Thiskindofthinghappensinargumentsthatreasonadimpossible：forintheseweareboundtodemolishoneofthepremisses。If，then，thefalsecausebereckonedinamongthequestionsthatarenecessarytoestablishtheresultingimpossibility，itwilloftenbethoughtthattherefutationdependsuponit，e。g。intheproofthatthe’soul’
and’life’arenotthesame：forifcoming—to—bebecontrarytoperishing，thenaparticularformofperishingwillhaveaparticularformofcoming—to—beasitscontrary：nowdeathisaparticularformofperishingandiscontrarytolife：life，therefore，isacomingto—be，andtoliveistocome—to—be。Butthisisimpossible：accordingly，the’soul’and’life’arenotthesame。Nowthisisnotproved：fortheimpossibilityresultsallthesame，evenifonedoesnotsaythatlifeisthesameasthesoul，butmerelysaysthatlifeiscontrarytodeath，whichisaformofperishing，andthatperishinghas’coming—to—be’asitscontrary。Argumentsofthatkind，then，thoughnotinconclusiveabsolutely，areinconclusiveinrelationtotheproposedconclusion。Alsoeventhequestionersthemselvesoftenfailquiteasmuchtoseeapointofthatkind。
Such，then，aretheargumentsthatdependupontheconsequentanduponfalsecause。Thosethatdependuponthemakingoftwoquestionsintooneoccurwheneverthepluralityisundetectedandasingleanswerisreturnedasiftoasinglequestion。Now，insomecases，itiseasytoseethatthereismorethanone，andthatananswerisnottobegiven，e。g。’Doestheearthconsistofsea，orthesky？’Butinsomecasesitislesseasy，andthenpeopletreatthequestionasone，andeitherconfesstheirdefeatbyfailingtoanswerthequestion，orareexposedtoanapparentrefutation。Thus’IsAandisBaman？’’Yes。’’ThenifanyonehitsAandB，hewillstrikeaman’（singular），’notmen’（plural）。Oragain，wherepartisgoodandpartbad，’isthewholegoodorbad？’Forwhicheverhesays，itispossiblethathemightbethoughttoexposehimselftoanapparentrefutationortomakeanapparentlyfalsestatement：fortosaythatsomethingisgoodwhichisnotgood，ornotgoodwhichisgood，istomakeafalsestatement。Sometimes，however，additionalpremissesmayactuallygiverisetoagenuinerefutation；e。g。supposeamanweretograntthatthedescriptions’white’and’naked’and’blind’
applytoonethingandtoanumberofthingsinalikesense。Forif’blind’describesathingthatcannotseethoughnaturedesignedittosee，itwillalsodescribethingsthatcannotseethoughnaturedesignedthemtodoso。Whenever，then，onethingcanseewhileanothercannot，theywilleitherbothbeabletoseeorelsebothbeblind；whichisimpossible。
6
Therightway，then，iseithertodivideapparentproofsandrefutationsasabove，orelsetoreferthemalltoignoranceofwhat’refutation’is，andmakethatourstarting—point：foritispossibletoanalysealltheaforesaidmodesoffallacyintobreachesofthedefinitionofarefutation。Inthefirstplace，wemayseeiftheyareinconclusive：fortheconclusionoughttoresultfromthepremisseslaiddown，soastocompelusnecessarilytostateitandnotmerelytoseemtocompelus。Nextweshouldalsotakethedefinitionbitbybit，andtrythefallacythereby。Forofthefallaciesthatconsistinlanguage，somedependuponadoublemeaning，e。g。ambiguityofwordsandofphrases，andthefallacyoflikeverbalforms（forwehabituallyspeakofeverythingasthoughitwereaparticularsubstance）—whilefallaciesofcombinationanddivisionandaccentarisebecausethephraseinquestionorthetermasalteredisnotthesameaswasintended。Eventhis，however，shouldbethesame，justasthethingsignifiedshouldbeaswell，ifarefutationorproofistobeeffected；e。g。ifthepointconcernsadoublet，thenyoushoulddrawtheconclusionofa’doublet’，notofa’cloak’。Fortheformerconclusionalsowouldbetrue，butithasnotbeenproved；weneedafurtherquestiontoshowthat’doublet’meansthesamething，inordertosatisfyanyonewhoaskswhyyouthinkyourpointproved。
FallaciesthatdependonAccidentareclearcasesofignoratioelenchiwhenonce’proof’hasbeendefined。Forthesamedefinitionoughttoholdgoodof’refutation’too，exceptthatamentionof’thecontradictory’ishereadded：forarefutationisaproofofthecontradictory。If，then，thereisnoproofasregardsanaccidentofanything，thereisnorefutation。Forsupposing，whenA
andBare，Cmustnecessarilybe，andCiswhite，thereisnonecessityforittobewhiteonaccountofthesyllogism。So，ifthetrianglehasitsanglesequaltotworight—angles，andithappenstobeafigure，orthesimplestelementorstartingpoint，itisnotbecauseitisafigureorastartingpointorsimplestelementthatithasthischaracter。Forthedemonstrationprovesthepointaboutitnotquafigureorquasimplestelement，butquatriangle。Likewisealsoinothercases。If，then，refutationisaproof，anargumentwhicharguedperaccidenscouldnotbearefutation。Itis，however，justinthisthattheexpertsandmenofsciencegenerallysufferrefutationatthehandoftheunscientific：forthelattermeetthescientistswithreasoningsconstitutedperaccidens；andthescientistsforlackofthepowertodrawdistinctionseithersay’Yes’
totheirquestions，orelsepeoplesupposethemtohavesaid’Yes’，althoughtheyhavenot。
Thosethatdependuponwhethersomethingissaidinacertainrespectonlyorsaidabsolutely，areclearcasesofignoratioelenchibecausetheaffirmationandthedenialarenotconcernedwiththesamepoint。Forof’whiteinacertainrespect’thenegationis’notwhiteinacertainrespect’，whileof’whiteabsolutely’itis’notwhite，absolutely’。If，then，amantreatstheadmissionthatathingis’whiteinacertainrespect’asthoughitweresaidtobewhiteabsolutely，hedoesnoteffectarefutation，butmerelyappearstodosoowingtoignoranceofwhatrefutationis。
Theclearestcasesofall，however，arethosethatwerepreviouslydescribed’asdependinguponthedefinitionofa’refutation’：andthisisalsowhytheywerecalledbythatname。Fortheappearanceofarefutationisproducedbecauseoftheomissioninthedefinition，andifwedividefallaciesintheabovemanner，weoughttoset’Defectivedefinition’asacommonmarkuponthemall。
Thosethatdependupontheassumptionoftheoriginalpointanduponstatingasthecausewhatisnotthecause，areclearlyshowntobecasesofignoratioelenchithroughthedefinitionthereof。Fortheconclusionoughttocomeabout’becausethesethingsareso’，andthisdoesnothappenwherethepremissesarenotcausesofit：andagainitshouldcomeaboutwithouttakingintoaccounttheoriginalpoint，andthisisnotthecasewiththoseargumentswhichdependuponbeggingtheoriginalpoint。
Thosethatdependupontheassumptionoftheoriginalpointanduponstatingasthecausewhatisnotthecause，areclearlyshowntobecasesofignoratioelenchithroughthedefinitionthereof。Fortheconclusionoughttocomeabout’becausethesethingsareso’，andthisdoesnothappenwherethepremissesarenotcausesofit：andagainitshouldcomeaboutwithouttakingintoaccounttheoriginalpoint，andthisisnotthecasewiththoseargumentswhichdependuponbeggingtheoriginalpoint。
ThosethatdependupontheconsequentareabranchofAccident：
fortheconsequentisanaccident，onlyitdiffersfromtheaccidentinthis，thatyoumaysecureanadmissionoftheaccidentinthecaseofonethingonly（e。g。theidentityofayellowthingandhoneyandofawhitethingandswan），whereastheconsequentalwaysinvolvesmorethanonething：forweclaimthatthingsthatarethesameasoneandthesamethingarealsothesameasoneanother，andthisisthegroundofarefutationdependentontheconsequent。Itis，however，notalwaystrue，e。g。supposethatandBarethesameasC
peraccidens；forboth’snow’andthe’swan’arethesameassomethingwhite’。Oragain，asinMelissus’argument，amanassumesthatto’havebeengenerated’andto’haveabeginning’arethesamething，orto’becomeequal’andto’assumethesamemagnitude’。Forbecausewhathasbeengeneratedhasabeginning，heclaimsalsothatwhathasabeginninghasbeengenerated，andarguesasthoughbothwhathasbeengeneratedandwhatisfinitewerethesamebecauseeachhasabeginning。Likewisealsointhecaseofthingsthataremadeequalheassumesthatifthingsthatassumeoneandthesamemagnitudebecomeequal，thenalsothingsthatbecomeequalassumeonemagnitude：
i。e。heassumestheconsequent。Inasmuch，then，asarefutationdependingonaccidentconsistsinignoranceofwhatarefutationis，clearlysoalsodoesarefutationdependingontheconsequent。Weshallhavefurthertoexaminethisinanotherwayaswell。
Thosefallaciesthatdependuponthemakingofseveralquestionsintooneconsistinourfailuretodissectthedefinitionof’proposition’。Forapropositionisasinglestatementaboutasinglething。Forthesamedefinitionappliesto’onesinglethingonly’andtothe’thing’，simply，e。g。to’man’andto’onesinglemanonly’andlikewisealsoinothercases。If，then，a’singleproposition’beonewhichclaimsasinglethingofasinglething，a’proposition’，simply，willalsobetheputtingofaquestionofthatkind。Nowsinceaproofstartsfrompropositionsandrefutationisaproof，refutation，too，willstartfrompropositions。If，then，apropositionisasinglestatementaboutasinglething，itisobviousthatthisfallacytooconsistsinignoranceofwhatarefutationis：forinitwhatisnotapropositionappearstobeone。If，then，theanswererhasreturnedananswerasthoughtoasinglequestion，therewillbearefutation；whileifhehasreturnedonenotreallybutapparently，therewillbeanapparentrefutationofhisthesis。Allthetypesoffallacy，then，fallunderignoranceofwhatarefutationis，someofthembecausethecontradiction，whichisthedistinctivemarkofarefutation，ismerelyapparent，andtherestfailingtoconformtothedefinitionofaproof。
7
Thedeceptioncomesaboutinthecaseofargumentsthatdependonambiguityofwordsandofphrasesbecauseweareunabletodividetheambiguousterm（forsometermsitisnoteasytodivide，e。g。
’unity’，’being’，and’sameness’），whileinthosethatdependoncombinationanddivision，itisbecausewesupposethatitmakesnodifferencewhetherthephrasebecombinedordivided，asisindeedthecasewithmostphrases。Likewisealsowiththosethatdependonaccent：fortheloweringorraisingofthevoiceuponaphraseisthoughtnottoalteritsmeaning—withanyphrase，ornotwithmany。
Withthosethatdependontheofexpressionitisbecauseofthelikenessofexpression。Foritishardtodistinguishwhatkindofthingsaresignifiedbythesameandwhatbydifferentkindsofexpression：foramanwhocandothisispracticallynextdoortotheunderstandingofthetruth。Aspecialreasonwhyamanisliabletobehurriedintoassenttothefallacyisthatwesupposeeverypredicateofeverythingtobeanindividualthing，andweunderstanditasbeingonewiththething：andwethereforetreatitasasubstance：foritistothatwhichisonewithathingorsubstance，asalsotosubstanceitself，that’individually’and’being’aredeemedtobelonginthefullestsense。Forthisreason，too，thistypeoffallacyistoberankedamongthosethatdependonlanguage；inthefirstplace，becausethedeceptioniseffectedthemorereadilywhenweareinquiringintoaproblemincompanywithothersthanwhenwedosobyourselves（foraninquirywithanotherpersoniscarriedonbymeansofspeech，whereasaninquirybyoneselfiscarriedonquiteasmuchbymeansoftheobjectitself）；secondlyamanisliabletobedeceived，evenwheninquiringbyhimself，whenhetakesspeechasthebasisofhisinquiry：moreoverthedeceptionarisesoutofthelikeness（oftwodifferentthings），andthelikenessarisesoutofthelanguage。WiththosefallaciesthatdependuponAccident，deceptioncomesaboutbecausewecannotdistinguishthesamenessandothernessofterms，i。e。theirunityandmultiplicity，orwhatkindsofpredicatehaveallthesameaccidentsastheirsubject。LikewisealsowiththosethatdependontheConsequent：fortheconsequentisabranchofAccident。Moreover，inmanycasesappearancespointtothis—andtheclaimismadethatifisinseparablefromB，soalsoisB
fromWiththosethatdependuponanimperfectioninthedefinitionofarefutation，andwiththosethatdependuponthedifferencebetweenaqualifiedandanabsolutestatement，thedeceptionconsistsinthesmallnessofthedifferenceinvolved；forwetreatthelimitationtotheparticularthingorrespectormannerortimeasaddingnothingtothemeaning，andsograntthestatementuniversally。
Likewisealsointhecaseofthosethatassumetheoriginalpoint，andthoseoffalsecause，andallthattreatanumberofquestionsasone：
forinallofthemthedeceptionliesinthesmallnessofthedifference：forourfailuretobequiteexactinourdefinitionof’premiss’andof’proof’isduetotheaforesaidreason。
8
Sinceweknowonhowmanypointsapparentsyllogismsdepend，weknowalsoonhowmanysophisticalsyllogismsandrefutationsmaydepend。ByasophisticalrefutationandsyllogismImeannotonlyasyllogismorrefutationwhichappearstobevalidbutisnot，butalsoonewhich，thoughitisvalid，onlyappearstobeappropriatetothethinginquestion。Thesearethosewhichfailtorefuteandprovepeopletobeignorantaccordingtothenatureofthethinginquestion，whichwasthefunctionoftheartofexamination。Nowtheartofexaminingisabranchofdialectic：andthismayproveafalseconclusionbecauseoftheignoranceoftheanswerer。Sophisticrefutationsontheotherhand，eventhoughtheyprovethecontradictoryofhisthesis，donotmakeclearwhetherheisignorant：forsophistsentanglethescientistaswellwiththesearguments。
Thatweknowthembythesamelineofinquiryisclear：forthesameconsiderationswhichmakeitappeartoanaudiencethatthepointsrequiredfortheproofwereaskedinthequestionsandthattheconclusionwasproved，wouldmaketheanswererthinksoaswell，sothatfalseproofwilloccurthroughallorsomeofthesemeans：forwhatamanhasnotbeenaskedbutthinkshehasgranted，hewouldalsograntifhewereasked。Ofcourse，insomecasesthemomentweaddthemissingquestion，wealsoshowupitsfalsity，e。g。infallaciesthatdependonlanguageandonsolecism。Ifthen，fallaciousproofsofthecontradictoryofathesisdependontheirappearingtorefute，itisclearthattheconsiderationsonwhichbothproofsoffalseconclusionsandanapparentrefutationdependmustbethesameinnumber。Nowanapparentrefutationdependsupontheelementsinvolvedinagenuineone：forthefailureofoneorotherofthesemustmaketherefutationmerelyapparent，e。g。thatwhichdependsonthefailureoftheconclusiontofollowfromtheargument（theargumentadimpossible）andthatwhichtreatstwoquestionsasoneandsodependsuponaflawinthepremiss，andthatwhichdependsonthesubstitutionofanaccidentforanessentialattribute，and—abranchofthelast—thatwhichdependsupontheconsequent：moreover，theconclusionmayfollownotinfactbutonlyverbally：then，insteadofprovingthecontradictoryuniversallyandinthesamerespectandrelationandmanner，thefallacymaybedependentonsomelimitofextentorononeorotherofthesequalifications：moreover，thereistheassumptionoftheoriginalpointtobeproved，inviolationoftheclause’withoutreckoningintheoriginalpoint’。Thusweshouldhavethenumberofconsiderationsonwhichthefallaciousproofsdepend：fortheycouldnotdependonmore，butallwilldependonthepointsaforesaid。
Asophisticalrefutationisarefutationnotabsolutelybutrelativelytosomeone：andsoisaproof，inthesameway。Forunlessthatwhichdependsuponambiguityassumesthattheambiguoustermhasasinglemeaning，andthatwhichdependsonlikeverbalformsassumesthatsubstanceistheonlycategory，andtherestinthesameway，therewillbeneitherrefutationsnorproofs，eitherabsolutelyorrelativelytotheanswerer：whereasiftheydoassumethesethings，theywillstand，relativelytotheanswerer；butabsolutelytheywillnotstand：fortheyhavenotsecuredastatementthatdoeshaveasinglemeaning，butonlyonethatappearstohave，andthatonlyfromthisparticularman。
9
Thenumberofconsiderationsonwhichdependtherefutationsofthosewhoarerefuted，weoughtnottotrytograspwithoutaknowledgeofeverythingthatis。This，however，isnottheprovinceofanyspecialstudy：forpossiblythesciencesareinfiniteinnumber，sothatobviouslydemonstrationsmaybeinfinitetoo。Nowrefutationsmaybetrueaswellasfalse：forwheneveritispossibletodemonstratesomething，itisalsopossibletorefutethemanwhomaintainsthecontradictoryofthetruth；e。g。ifamanhasstatedthatthediagonaliscommensuratewiththesideofthesquare，onemightrefutehimbydemonstratingthatitisincommensurate。Accordingly，toexhaustallpossiblerefutationsweshallhavetohavescientificknowledgeofeverything：forsomerefutationsdependupontheprinciplesthatruleingeometryandtheconclusionsthatfollowfromthese，othersuponthosethatruleinmedicine，andothersuponthoseoftheothersciences。Forthematterofthat，thefalserefutationslikewisebelongtothenumberoftheinfinite：foraccordingtoeveryartthereisfalseproof，e。g。
accordingtogeometrythereisfalsegeometricalproof，andaccordingtomedicinethereisfalsemedicalproof。By’accordingtotheart’，Imean’accordingtotheprinciplesofit’。Clearly，then，itisnotofallrefutations，butonlyofthosethatdependupondialecticthatweneedtograspthecommon—placerules：forthesestandinacommonrelationtoeveryartandfaculty。Andasregardstherefutationthatisaccordingtooneorotheroftheparticularsciencesitisthetaskofthatparticularscientisttoexaminewhetheritismerelyapparentwithoutbeingreal，and，ifitbereal，whatisthereasonforit：whereasitisthebusinessofdialecticianssotoexaminetherefutationthatproceedsfromthecommonfirstprinciplesthatfallundernoparticularspecialstudy。
Forifwegraspthestartingpointsoftheacceptedproofsonanysubjectwhateverwegraspthoseoftherefutationscurrentonthatsubject。Forarefutationistheproofofthecontradictoryofagiventhesis，sothateitheroneortwoproofsofthecontradictoryconstitutearefutation。Wegrasp，then，thenumberofconsiderationsonwhichallsuchdepend：if，however，wegraspthis，wealsograsptheirsolutionsaswell；fortheobjectionstothesearethesolutionsofthem。Wealsograspthenumberofconsiderationsonwhichthoserefutationsdepend，thataremerelyapparent—apparent，I
mean，nottoeverybody，buttopeopleofacertainstamp；foritisanindefinitetaskifoneistoinquirehowmanyaretheconsiderationsthatmakethemapparenttothemaninthestreet。Accordinglyitisclearthatthedialectician’sbusinessistobeabletograsponhowmanyconsiderationsdependstheformation，throughthecommonfirstprinciples，ofarefutationthatiseitherrealorapparent，i。e。
eitherdialecticalorapparentlydialectical，orsuitableforanexamination。
10
Itisnotruedistinctionbetweenargumentswhichsomepeopledrawwhentheysaythatsomeargumentsaredirectedagainsttheexpression，andothersagainstthethoughtexpressed：foritisabsurdtosupposethatsomeargumentsaredirectedagainsttheexpressionandothersagainstthethought，andthattheyarenotthesame。Forwhatisfailuretodirectanargumentagainstthethoughtexceptwhatoccurswheneveramandoesnotinusingtheexpressionthinkittobeusedinhisquestioninthesamesenseinwhichthepersonquestionedgrantedit？Andthisisthesamethingastodirecttheargumentagainsttheexpression。Ontheotherhand，itisdirectedagainstthethoughtwheneveramanusestheexpressioninthesamesensewhichtheanswererhadinmindwhenhegrantedit。Ifnowany（i。e。boththequestionerandthepersonquestioned），indealingwithanexpressionwithmorethanonemeaning，weretosupposeittohaveonemeaning—ase。g。itmaybethat’Being’and’One’havemanymeanings，andyetboththeanswereranswersandthequestionerputshisquestionsupposingittobeone，andtheargumentistotheeffectthat’Allthingsareone’—willthisdiscussionbedirectedanymoreagainsttheexpressionthanagainstthethoughtofthepersonquestioned？If，ontheotherhand，oneofthemsupposestheexpressiontohavemanymeanings，itisclearthatsuchadiscussionwillnotbedirectedagainstthethought。Suchbeingthemeaningsofthephrasesinquestion，theyclearlycannotdescribetwoseparateclassesofargument。For，inthefirstplace，itispossibleforanysuchargumentasbearsmorethanonemeaningtobedirectedagainsttheexpressionandagainstthethought，andnextitispossibleforanyargumentwhatsoever；forthefactofbeingdirectedagainstthethoughtconsistsnotinthenatureoftheargument，butinthespecialattitudeoftheanswerertowardsthepointsheconcedes。Next，allofthemmaybedirectedtotheexpression。For’tobedirectedagainsttheexpression’meansinthisdoctrine’nottobedirectedagainstthethought’。Forifnotallaredirectedagainsteitherexpressionorthought，therewillbecertainotherargumentsdirectedneitheragainsttheexpressionnoragainstthethought，whereastheysaythatallmustbeoneortheother，anddividethemallasdirectedeitheragainsttheexpressionoragainstthethought，whileothers（theysay）therearenone。Butinpointoffactthosethatdependonmereexpressionareonlyabranchofthosesyllogismsthatdependonamultiplicityofmeanings。Fortheabsurdstatementhasactuallybeenmadethatthedescription’dependentonmereexpression’describesalltheargumentsthatdependonlanguage：whereassomeofthesearefallaciesnotbecausetheanswereradoptsaparticularattitudetowardsthem，butbecausetheargumentitselfinvolvestheaskingofaquestionsuchasbearsmorethanonemeaning。
Itis，too，altogetherabsurdtodiscussRefutationwithoutfirstdiscussingProof：forarefutationisaproof，sothatoneoughttodiscussproofaswellbeforedescribingfalserefutation：forarefutationofthatkindisamerelyapparentproofofthecontradictoryofathesis。Accordingly，thereasonofthefalsitywillbeeitherintheprooforinthecontradiction（formentionofthe’contradiction’mustbeadded），whilesometimesitisinboth，iftherefutationbemerelyapparent。Intheargumentthatspeakingofthesilentispossibleitliesinthecontradiction，notintheproof；
intheargumentthatonecangivewhatonedoesnotpossess，itliesinboth；intheproofthatHomer’spoemisafigurethroughitsbeingacycleitliesintheproof。Anargumentthatdoesnotfailineitherrespectisatrueproof。
But，toreturntothepointwhenceourargumentdigressed，aremathematicalreasoningsdirectedagainstthethought，ornot？Andifanyonethinks’triangle’tobeawordwithmanymeanings，andgranteditinsomedifferentsensefromthefigurewhichwasprovedtocontaintworightangles，hasthequestionerheredirectedhisargumentagainstthethoughtoftheformerornot？
Moreover，iftheexpressionbearsmanysenses，whiletheanswererdoesnotunderstandorsupposeittohavethem，surelythequestionerherehasdirectedhisargumentagainsthisthought!Orhowelseoughthetoputhisquestionexceptbysuggestingadistinction—supposeone’squestiontobespeakingofthesilentpossibleornot？’—asfollows，’Istheanswer\"No\"inonesense，but\"Yes\"inanother？’If，then，anyoneweretoanswerthatitwasnotpossibleinanysenseandtheotherweretoarguethatitwas，hasnothisargumentbeendirectedagainstthethoughtoftheanswerer？Yethisargumentissupposedtobeoneofthosethatdependontheexpression。Thereisnot，then，anydefinitekindofargumentsthatisdirectedagainstthethought。Someargumentsare，indeed，directedagainsttheexpression：butthesearenotallevenapparentrefutations，letaloneallrefutations。Fortherearealsoapparentrefutationswhichdonotdependuponlanguage，e。g。thosethatdependuponaccident，andothers。
If，however，anyoneclaimsthatoneshouldactuallydrawthedistinction，andsay，’By\"speakingofthesilent\"Imean，inonesensethisandintheothersensethat’，surelytoclaimthisisinthefirstplaceabsurd（forsometimesthequestionerdoesnotseetheambiguityofhisquestion，andhecannotpossiblydrawadistinctionwhichhedoesnotthinktobethere）：inthesecondplace，whatelsebutthiswilldidacticargumentbe？Foritwillmakemanifestthestateofthecasetoonewhohasneverconsidered，anddoesnotknoworsupposethatthereisanyothermeaningbutone。
Forwhatistheretopreventthesamethingalsohappeningtousincaseswherethereisnodoublemeaning？’Aretheunitsinfourequaltothetwos？Observethatthetwosarecontainedinfourinonesenseinthisway，inanothersenseinthat’。Also，’Istheknowledgeofcontrariesoneornot？Observethatsomecontrariesareknown，whileothersareunknown’。Thusthemanwhomakesthisclaimseemstobeunawareofthedifferencebetweendidacticanddialecticalargument，andofthefactthatwhilehewhoarguesdidacticallyshouldnotaskquestionsbutmakethingsclearhimself，theothershouldmerelyaskquestions。