Sunteți pe pagina 1din 49

Aristotle Posterior Analytics translated by G. R. G.

Mure

Book I 1 All instruction given or received by way of argument proceeds from pre e!istent knowledge. "#is becomes evident upon a survey of all t#e species of suc# instruction. "#e mat#ematical sciences and all ot#er speculative disciplines are ac$uired in t#is way% and so are t#e two forms of dialectical reasoning% syllogistic and inductive& for eac# of t#ese latter make use of old knowledge to impart new% t#e syllogism assuming an audience t#at accepts its premisses% induction e!#ibiting t#e universal as implicit in t#e clearly known particular. Again% t#e persuasion e!erted by r#etorical arguments is in principle t#e same% since t#ey use eit#er e!ample% a kind of induction% or ent#ymeme% a form of syllogism. "#e pre e!istent knowledge re$uired is of two kinds. In some cases admission of t#e fact must be assumed% in ot#ers compre#ension of t#e meaning of t#e term used% and sometimes bot# assumptions are essential. "#us% we assume t#at every predicate can be eit#er truly affirmed or truly denied of any sub'ect% and t#at (triangle) means so and so& as regards (unit) we #ave to make t#e double assumption of t#e meaning of t#e word and t#e e!istence of t#e t#ing. "#e reason is t#at t#ese several ob'ects are not e$ually obvious to us. Recognition of a trut# may in some cases contain as factors bot# previous knowledge and also knowledge ac$uired simultaneously wit# t#at recognition * knowledge% t#is latter% of t#e particulars actually falling under t#e universal and t#erein already virtually known. +or e!ample% t#e student knew before#and t#at t#e angles of every triangle are e$ual to two rig#t angles& but it was only at t#e actual moment at w#ic# #e was being led on to recogni,e t#is as true in t#e instance before #im t#at #e came to know (t#is figure inscribed in t#e semicircle) to be a triangle. +or some t#ings -vi,. t#e singulars finally reac#ed w#ic# are not predicable of anyt#ing else as sub'ect. are only learnt in t#is way% i.e. t#ere is #ere no recognition t#roug# a middle of a minor term as sub'ect to a ma'or. Before #e was led on to recognition or before #e actually drew a conclusion% we s#ould per#aps say t#at in a manner #e knew% in a manner not. If #e did not in an un$ualified sense of t#e term know t#e e!istence of t#is triangle% #ow could #e know wit#out $ualification t#at its angles were e$ual to two rig#t angles/ 0o1 clearly #e knows not wit#out $ualification but only in t#e sense t#at #e knows universally. If t#is distinction is not drawn% we are faced wit# t#e dilemma in t#e Meno1 eit#er a man will learn not#ing or w#at #e already knows& for we cannot accept t#e solution w#ic# some people offer. A man is asked% (2o you% or do you not% know t#at every pair is even/) 3e says #e does know it. "#e $uestioner t#en produces a particular pair% of t#e e!istence% and so a fortiori of t#e evenness% of w#ic# #e was unaware. "#e solution w#ic# some people offer is to assert t#at t#ey do not know t#at every pair is even% but only t#at everyt#ing w#ic# t#ey know to be a pair is even1 yet w#at t#ey know to be even is t#at of w#ic# t#ey #ave demonstrated evenness% i.e. w#at t#ey made t#e sub'ect of t#eir premiss% vi,. not merely every triangle or number w#ic# t#ey know to be suc#% but any and every number or triangle wit#out reservation. +or no premiss is ever couc#ed in t#e form (every number w#ic# you know to be suc#)% or (every rectilinear figure w#ic# you know to be suc#)1 t#e 1

predicate is always construed as applicable to any and every instance of t#e t#ing. 4n t#e ot#er #and% I imagine t#ere is not#ing to prevent a man in one sense knowing w#at #e is learning% in anot#er not knowing it. "#e strange t#ing would be% not if in some sense #e knew w#at #e was learning% but if #e were to know it in t#at precise sense and manner in w#ic# #e was learning it. 5 6e suppose ourselves to possess un$ualified scientific knowledge of a t#ing% as opposed to knowing it in t#e accidental way in w#ic# t#e sop#ist knows% w#en we t#ink t#at we know t#e cause on w#ic# t#e fact depends% as t#e cause of t#at fact and of no ot#er% and% furt#er% t#at t#e fact could not be ot#er t#an it is. 0ow t#at scientific knowing is somet#ing of t#is sort is evident * witness bot# t#ose w#o falsely claim it and t#ose w#o actually possess it% since t#e former merely imagine t#emselves to be% w#ile t#e latter are also actually% in t#e condition described. 7onse$uently t#e proper ob'ect of un$ualified scientific knowledge is somet#ing w#ic# cannot be ot#er t#an it is. "#ere may be anot#er manner of knowing as well * t#at will be discussed later. 6#at I now assert is t#at at all events we do know by demonstration. By demonstration I mean a syllogism productive of scientific knowledge% a syllogism% t#at is% t#e grasp of w#ic# is eo ipso suc# knowledge. Assuming t#en t#at my t#esis as to t#e nature of scientific knowing is correct% t#e premisses of demonstrated knowledge must be true% primary% immediate% better known t#an and prior to t#e conclusion% w#ic# is furt#er related to t#em as effect to cause. 8nless t#ese conditions are satisfied% t#e basic trut#s will not be (appropriate) to t#e conclusion. 9yllogism t#ere may indeed be wit#out t#ese conditions% but suc# syllogism% not being productive of scientific knowledge% will not be demonstration. "#e premisses must be true1 for t#at w#ic# is non e!istent cannot be known * we cannot know% e.g. t#at t#e diagonal of a s$uare is commensurate wit# its side. "#e premisses must be primary and indemonstrable& ot#erwise t#ey will re$uire demonstration in order to be known% since to #ave knowledge% if it be not accidental knowledge% of t#ings w#ic# are demonstrable% means precisely to #ave a demonstration of t#em. "#e premisses must be t#e causes of t#e conclusion% better known t#an it% and prior to it& its causes% since we possess scientific knowledge of a t#ing only w#en we know its cause& prior% in order to be causes& antecedently known% t#is antecedent knowledge being not our mere understanding of t#e meaning% but knowledge of t#e fact as well. 0ow (prior) and (better known) are ambiguous terms% for t#ere is a difference between w#at is prior and better known in t#e order of being and w#at is prior and better known to man. I mean t#at ob'ects nearer to sense are prior and better known to man& ob'ects wit#out $ualification prior and better known are t#ose furt#er from sense. 0ow t#e most universal causes are furt#est from sense and particular causes are nearest to sense% and t#ey are t#us e!actly opposed to one anot#er. In saying t#at t#e premisses of demonstrated knowledge must be primary% I mean t#at t#ey must be t#e (appropriate) basic trut#s% for I identify primary premiss and basic trut#. A (basic trut#) in a demonstration is an immediate proposition. An immediate proposition is one w#ic# #as no ot#er proposition prior to it. A proposition is eit#er part of an enunciation% i.e. it predicates a single attribute of a single sub'ect. If a proposition is dialectical% it assumes eit#er part indifferently& if it is demonstrative% it lays down one part to t#e definite e!clusion of t#e ot#er because t#at part is true. "#e term (enunciation) denotes eit#er part of a contradiction indifferently. A contradiction is an opposition w#ic# of its own nature e!cludes a middle. "#e part of a contradiction w#ic# con'oins a predicate wit# a sub'ect is an affirmation& t#e part dis'oining t#em is a negation. I call an immediate basic trut# of syllogism a (t#esis) w#en% t#oug# it is not susceptible of proof by t#e teac#er% yet ignorance of it does not constitute a total 5

bar to progress on t#e part of t#e pupil1 one w#ic# t#e pupil must know if #e is to learn anyt#ing w#atever is an a!iom. I call it an a!iom because t#ere are suc# trut#s and we give t#em t#e name of a!ioms par e!cellence. If a t#esis assumes one part or t#e ot#er of an enunciation% i.e. asserts eit#er t#e e!istence or t#e non e!istence of a sub'ect% it is a #ypot#esis& if it does not so assert% it is a definition. 2efinition is a (t#esis) or a (laying somet#ing down)% since t#e arit#metician lays it down t#at to be a unit is to be $uantitatively indivisible& but it is not a #ypot#esis% for to define w#at a unit is is not t#e same as to affirm its e!istence. 0ow since t#e re$uired ground of our knowledge * i.e. of our conviction * of a fact is t#e possession of suc# a syllogism as we call demonstration% and t#e ground of t#e syllogism is t#e facts constituting its premisses% we must not only know t#e primary premisses * some if not all of t#em * before#and% but know t#em better t#an t#e conclusion1 for t#e cause of an attribute)s in#erence in a sub'ect always itself in#eres in t#e sub'ect more firmly t#an t#at attribute& e.g. t#e cause of our loving anyt#ing is dearer to us t#an t#e ob'ect of our love. 9o since t#e primary premisses are t#e cause of our knowledge * i.e. of our conviction * it follows t#at we know t#em better * t#at is% are more convinced of t#em * t#an t#eir conse$uences% precisely because of our knowledge of t#e latter is t#e effect of our knowledge of t#e premisses. 0ow a man cannot believe in anyt#ing more t#an in t#e t#ings #e knows% unless #e #as eit#er actual knowledge of it or somet#ing better t#an actual knowledge. But we are faced wit# t#is parado! if a student w#ose belief rests on demonstration #as not prior knowledge& a man must believe in some% if not in all% of t#e basic trut#s more t#an in t#e conclusion. Moreover% if a man sets out to ac$uire t#e scientific knowledge t#at comes t#roug# demonstration% #e must not only #ave a better knowledge of t#e basic trut#s and a firmer conviction of t#em t#an of t#e conne!ion w#ic# is being demonstrated1 more t#an t#is% not#ing must be more certain or better known to #im t#an t#ese basic trut#s in t#eir c#aracter as contradicting t#e fundamental premisses w#ic# lead to t#e opposed and erroneous conclusion. +or indeed t#e conviction of pure science must be uns#akable. : 9ome #old t#at% owing to t#e necessity of knowing t#e primary premisses% t#ere is no scientific knowledge. 4t#ers t#ink t#ere is% but t#at all trut#s are demonstrable. 0eit#er doctrine is eit#er true or a necessary deduction from t#e premisses. "#e first sc#ool% assuming t#at t#ere is no way of knowing ot#er t#an by demonstration% maintain t#at an infinite regress is involved% on t#e ground t#at if be#ind t#e prior stands no primary% we could not know t#e posterior t#roug# t#e prior -w#erein t#ey are rig#t% for one cannot traverse an infinite series.1 if on t#e ot#er #and * t#ey say * t#e series terminates and t#ere are primary premisses% yet t#ese are unknowable because incapable of demonstration% w#ic# according to t#em is t#e only form of knowledge. And since t#us one cannot know t#e primary premisses% knowledge of t#e conclusions w#ic# follow from t#em is not pure scientific knowledge nor properly knowing at all% but rests on t#e mere supposition t#at t#e premisses are true. "#e ot#er party agree wit# t#em as regards knowing% #olding t#at it is only possible by demonstration% but t#ey see no difficulty in #olding t#at all trut#s are demonstrated% on t#e ground t#at demonstration may be circular and reciprocal. 4ur own doctrine is t#at not all knowledge is demonstrative1 on t#e contrary% knowledge of t#e immediate premisses is independent of demonstration. -"#e necessity of t#is is obvious& for since we must know t#e prior premisses from w#ic# t#e demonstration is drawn% and since t#e regress must end in immediate trut#s% t#ose trut#s must be indemonstrable.. 9uc#% t#en% is our doctrine%

and in addition we maintain t#at besides scientific knowledge t#ere is its originative source w#ic# enables us to recogni,e t#e definitions. 0ow demonstration must be based on premisses prior to and better known t#an t#e conclusion& and t#e same t#ings cannot simultaneously be bot# prior and posterior to one anot#er1 so circular demonstration is clearly not possible in t#e un$ualified sense of (demonstration)% but only possible if (demonstration) be e!tended to include t#at ot#er met#od of argument w#ic# rests on a distinction between trut#s prior to us and trut#s wit#out $ualification prior% i.e. t#e met#od by w#ic# induction produces knowledge. But if we accept t#is e!tension of its meaning% our definition of un$ualified knowledge will prove faulty& for t#ere seem to be two kinds of it. Per#aps% #owever% t#e second form of demonstration% t#at w#ic# proceeds from trut#s better known to us% is not demonstration in t#e un$ualified sense of t#e term. "#e advocates of circular demonstration are not only faced wit# t#e difficulty we #ave 'ust stated1 in addition t#eir t#eory reduces to t#e mere statement t#at if a t#ing e!ists% t#en it does e!ist * an easy way of proving anyt#ing. "#at t#is is so can be clearly s#own by taking t#ree terms% for to constitute t#e circle it makes no difference w#et#er many terms or few or even only two are taken. "#us by direct proof% if A is% B must be& if B is% 7 must be& t#erefore if A is% 7 must be. 9ince t#en * by t#e circular proof * if A is% B must be% and if B is% A must be% A may be substituted for 7 above. "#en (if B is% A must be);(if B is% 7 must be)% w#ic# above gave t#e conclusion (if A is% 7 must be)1 but 7 and A #ave been identified. 7onse$uently t#e up#olders of circular demonstration are in t#e position of saying t#at if A is% A must be * a simple way of proving anyt#ing. Moreover% even suc# circular demonstration is impossible e!cept in t#e case of attributes t#at imply one anot#er% vi,. (peculiar) properties. 0ow% it #as been s#own t#at t#e positing of one t#ing * be it one term or one premiss * never involves a necessary conse$uent1 two premisses constitute t#e first and smallest foundation for drawing a conclusion at all and t#erefore a fortiori for t#e demonstrative syllogism of science. If% t#en% A is implied in B and 7% and B and 7 are reciprocally implied in one anot#er and in A% it is possible% as #as been s#own in my writings on t#e syllogism% to prove all t#e assumptions on w#ic# t#e original conclusion rested% by circular demonstration in t#e first figure. But it #as also been s#own t#at in t#e ot#er figures eit#er no conclusion is possible% or at least none w#ic# proves bot# t#e original premisses. Propositions t#e terms of w#ic# are not convertible cannot be circularly demonstrated at all% and since convertible terms occur rarely in actual demonstrations% it is clearly frivolous and impossible to say t#at demonstration is reciprocal and t#at t#erefore everyt#ing can be demonstrated. < 9ince t#e ob'ect of pure scientific knowledge cannot be ot#er t#an it is% t#e trut# obtained by demonstrative knowledge will be necessary. And since demonstrative knowledge is only present w#en we #ave a demonstration% it follows t#at demonstration is an inference from necessary premisses. 9o we must consider w#at are t#e premisses of demonstration * i.e. w#at is t#eir c#aracter1 and as a preliminary% let us define w#at we mean by an attribute (true in every instance of its sub'ect)% an (essential) attribute% and a (commensurate and universal) attribute. I call (true in every instance) w#at is truly predicable of all instances * not of one to t#e e!clusion of ot#ers * and at all times% not at t#is or t#at time only& e.g. if animal is truly predicable of every instance of man% t#en if it be true to say (t#is is a man)% (t#is is an animal) is also true% and if t#e one be true now t#e ot#er is true now. A corresponding account #olds if point is in every instance predicable as contained in line. "#ere is evidence for t#is in t#e fact t#at t#e ob'ection we raise against a <

proposition put to us as true in every instance is eit#er an instance in w#ic#% or an occasion on w#ic#% it is not true. =ssential attributes are -1. suc# as belong to t#eir sub'ect as elements in its essential nature -e.g. line t#us belongs to triangle% point to line& for t#e very being or (substance) of triangle and line is composed of t#ese elements% w#ic# are contained in t#e formulae defining triangle and line.1 -5. suc# t#at% w#ile t#ey belong to certain sub'ects% t#e sub'ects to w#ic# t#ey belong are contained in t#e attribute)s own defining formula. "#us straig#t and curved belong to line% odd and even% prime and compound% s$uare and oblong% to number& and also t#e formula defining any one of t#ese attributes contains its sub'ect * e.g. line or number as t#e case may be. =!tending t#is classification to all ot#er attributes% I distinguis# t#ose t#at answer t#e above description as belonging essentially to t#eir respective sub'ects& w#ereas attributes related in neit#er of t#ese two ways to t#eir sub'ects I call accidents or (coincidents)& e.g. musical or w#ite is a (coincident) of animal. +urt#er -a. t#at is essential w#ic# is not predicated of a sub'ect ot#er t#an itself1 e.g. (t#e walking >t#ing?) walks and is w#ite in virtue of being somet#ing else besides& w#ereas substance% in t#e sense of w#atever signifies a (t#is somew#at)% is not w#at it is in virtue of being somet#ing else besides. "#ings% t#en% not predicated of a sub'ect I call essential& t#ings predicated of a sub'ect I call accidental or (coincidental). In anot#er sense again -b. a t#ing conse$uentially connected wit# anyt#ing is essential& one not so connected is (coincidental). An e!ample of t#e latter is (6#ile #e was walking it lig#tened)1 t#e lig#tning was not due to #is walking& it was% we s#ould say% a coincidence. If% on t#e ot#er #and% t#ere is a conse$uential conne!ion% t#e predication is essential& e.g. if a beast dies w#en its t#roat is being cut% t#en its deat# is also essentially connected wit# t#e cutting% because t#e cutting was t#e cause of deat#% not deat# a (coincident) of t#e cutting. 9o far t#en as concerns t#e sp#ere of conne!ions scientifically known in t#e un$ualified sense of t#at term% all attributes w#ic# -wit#in t#at sp#ere. are essential eit#er in t#e sense t#at t#eir sub'ects are contained in t#em% or in t#e sense t#at t#ey are contained in t#eir sub'ects% are necessary as well as conse$uentially connected wit# t#eir sub'ects. +or it is impossible for t#em not to in#ere in t#eir sub'ects eit#er simply or in t#e $ualified sense t#at one or ot#er of a pair of opposites must in#ere in t#e sub'ect& e.g. in line must be eit#er straig#tness or curvature% in number eit#er oddness or evenness. +or wit#in a single identical genus t#e contrary of a given attribute is eit#er its privative or its contradictory& e.g. wit#in number w#at is not odd is even% inasmuc# as wit#in t#is sp#ere even is a necessary conse$uent of not odd. 9o% since any given predicate must be eit#er affirmed or denied of any sub'ect% essential attributes must in#ere in t#eir sub'ects of necessity. "#us% t#en% we #ave establis#ed t#e distinction between t#e attribute w#ic# is (true in every instance) and t#e (essential) attribute. I term (commensurately universal) an attribute w#ic# belongs to every instance of its sub'ect% and to every instance essentially and as suc#& from w#ic# it clearly follows t#at all commensurate universals in#ere necessarily in t#eir sub'ects. "#e essential attribute% and t#e attribute t#at belongs to its sub'ect as suc#% are identical. =.g. point and straig#t belong to line essentially% for t#ey belong to line as suc#& and triangle as suc# #as two rig#t angles% for it is essentially e$ual to two rig#t angles. An attribute belongs commensurately and universally to a sub'ect w#en it can be s#own to belong to any random instance of t#at sub'ect and w#en t#e sub'ect is t#e first t#ing to w#ic# it can be s#own to belong. "#us% e.g. -1. t#e e$uality of its angles to two rig#t angles is not a commensurately universal attribute of figure. +or t#oug# it is possible to s#ow t#at a figure #as its angles e$ual to two rig#t angles% t#is attribute cannot be demonstrated of any figure selected at @

#ap#a,ard% nor in demonstrating does one take a figure at random * a s$uare is a figure but its angles are not e$ual to two rig#t angles. 4n t#e ot#er #and% any isosceles triangle #as its angles e$ual to two rig#t angles% yet isosceles triangle is not t#e primary sub'ect of t#is attribute but triangle is prior. 9o w#atever can be s#own to #ave its angles e$ual to two rig#t angles% or to possess any ot#er attribute% in any random instance of itself and primarily * t#at is t#e first sub'ect to w#ic# t#e predicate in $uestion belongs commensurately and universally% and t#e demonstration% in t#e essential sense% of any predicate is t#e proof of it as belonging to t#is first sub'ect commensurately and universally1 w#ile t#e proof of it as belonging to t#e ot#er sub'ects to w#ic# it attac#es is demonstration only in a secondary and unessential sense. 0or again -5. is e$uality to two rig#t angles a commensurately universal attribute of isosceles& it is of wider application. @ 6e must not fail to observe t#at we often fall into error because our conclusion is not in fact primary and commensurately universal in t#e sense in w#ic# we t#ink we prove it so. 6e make t#is mistake -1. w#en t#e sub'ect is an individual or individuals above w#ic# t#ere is no universal to be found1 -5. w#en t#e sub'ects belong to different species and t#ere is a #ig#er universal% but it #as no name1 -:. w#en t#e sub'ect w#ic# t#e demonstrator takes as a w#ole is really only a part of a larger w#ole& for t#en t#e demonstration will be true of t#e individual instances wit#in t#e part and will #old in every instance of it% yet t#e demonstration will not be true of t#is sub'ect primarily and commensurately and universally. 6#en a demonstration is true of a sub'ect primarily and commensurately and universally% t#at is to be taken to mean t#at it is true of a given sub'ect primarily and as suc#. 7ase -:. may be t#us e!emplified. If a proof were given t#at perpendiculars to t#e same line are parallel% it mig#t be supposed t#at lines t#us perpendicular were t#e proper sub'ect of t#e demonstration because being parallel is true of every instance of t#em. But it is not so% for t#e parallelism depends not on t#ese angles being e$ual to one anot#er because eac# is a rig#t angle% but simply on t#eir being e$ual to one anot#er. An e!ample of -1. would be as follows1 if isosceles were t#e only triangle% it would be t#oug#t to #ave its angles e$ual to two rig#t angles $ua isosceles. An instance of -5. would be t#e law t#at proportionals alternate. Alternation used to be demonstrated separately of numbers% lines% solids% and durations% t#oug# it could #ave been proved of t#em all by a single demonstration. Because t#ere was no single name to denote t#at in w#ic# numbers% lengt#s% durations% and solids are identical% and because t#ey differed specifically from one anot#er% t#is property was proved of eac# of t#em separately. "o day% #owever% t#e proof is commensurately universal% for t#ey do not possess t#is attribute $ua lines or $ua numbers% but $ua manifesting t#is generic c#aracter w#ic# t#ey are postulated as possessing universally. 3ence% even if one prove of eac# kind of triangle t#at its angles are e$ual to two rig#t angles% w#et#er by means of t#e same or different proofs& still% as long as one treats separately e$uilateral% scalene% and isosceles% one does not yet know% e!cept sop#istically% t#at triangle #as its angles e$ual to two rig#t angles% nor does one yet know t#at triangle #as t#is property commensurately and universally% even if t#ere is no ot#er species of triangle but t#ese. +or one does not know t#at triangle as suc# #as t#is property% nor even t#at (all) triangles #ave it * unless (all) means (eac# taken singly)1 if (all) means (as a w#ole class)% t#en% t#oug# t#ere be none in w#ic# one does not recogni,e t#is property% one does not know it of (all triangles). 6#en% t#en% does our knowledge fail of commensurate universality% and w#en it is un$ualified knowledge/ If triangle be identical in essence wit# e$uilateral% i.e. wit# eac# or all e$uilaterals% A

t#en clearly we #ave un$ualified knowledge1 if on t#e ot#er #and it be not% and t#e attribute belongs to e$uilateral $ua triangle& t#en our knowledge fails of commensurate universality. (But)% it will be asked% (does t#is attribute belong to t#e sub'ect of w#ic# it #as been demonstrated $ua triangle or $ua isosceles/ 6#at is t#e point at w#ic# t#e sub'ect. to w#ic# it belongs is primary/ -i.e. to w#at sub'ect can it be demonstrated as belonging commensurately and universally/.) 7learly t#is point is t#e first term in w#ic# it is found to in#ere as t#e elimination of inferior differentiae proceeds. "#us t#e angles of a bra,en isosceles triangle are e$ual to two rig#t angles1 but eliminate bra,en and isosceles and t#e attribute remains. (But) * you may say * )eliminate figure or limit% and t#e attribute vanis#es.) "rue% but figure and limit are not t#e first differentiae w#ose elimination destroys t#e attribute. ("#en w#at is t#e first/) If it is triangle% it will be in virtue of triangle t#at t#e attribute belongs to all t#e ot#er sub'ects of w#ic# it is predicable% and triangle is t#e sub'ect to w#ic# it can be demonstrated as belonging commensurately and universally. A 2emonstrative knowledge must rest on necessary basic trut#s& for t#e ob'ect of scientific knowledge cannot be ot#er t#an it is. 0ow attributes attac#ing essentially to t#eir sub'ects attac# necessarily to t#em1 for essential attributes are eit#er elements in t#e essential nature of t#eir sub'ects% or contain t#eir sub'ects as elements in t#eir own essential nature. -"#e pairs of opposites w#ic# t#e latter class includes are necessary because one member or t#e ot#er necessarily in#eres.. It follows from t#is t#at premisses of t#e demonstrative syllogism must be conne!ions essential in t#e sense e!plained1 for all attributes must in#ere essentially or else be accidental% and accidental attributes are not necessary to t#eir sub'ects. 6e must eit#er state t#e case t#us% or else premise t#at t#e conclusion of demonstration is necessary and t#at a demonstrated conclusion cannot be ot#er t#an it is% and t#en infer t#at t#e conclusion must be developed from necessary premisses. +or t#oug# you may reason from true premisses wit#out demonstrating% yet if your premisses are necessary you will assuredly demonstrate * in suc# necessity you #ave at once a distinctive c#aracter of demonstration. "#at demonstration proceeds from necessary premisses is also indicated by t#e fact t#at t#e ob'ection we raise against a professed demonstration is t#at a premiss of it is not a necessary trut# * w#et#er we t#ink it altoget#er devoid of necessity% or at any rate so far as our opponent)s previous argument goes. "#is s#ows #ow naive it is to suppose one)s basic trut#s rig#tly c#osen if one starts wit# a proposition w#ic# is -1. popularly accepted and -5. true% suc# as t#e sop#ists) assumption t#at to know is t#e same as to possess knowledge. +or -1. popular acceptance or re'ection is no criterion of a basic trut#% w#ic# can only be t#e primary law of t#e genus constituting t#e sub'ect matter of t#e demonstration& and -5. not all trut# is (appropriate). A furt#er proof t#at t#e conclusion must be t#e development of necessary premisses is as follows. 6#ere demonstration is possible% one w#o can give no account w#ic# includes t#e cause #as no scientific knowledge. If% t#en% we suppose a syllogism in w#ic#% t#oug# A necessarily in#eres in 7% yet B% t#e middle term of t#e demonstration% is not necessarily connected wit# A and 7% t#en t#e man w#o argues t#us #as no reasoned knowledge of t#e conclusion% since t#is conclusion does not owe its necessity to t#e middle term& for t#oug# t#e conclusion is necessary% t#e mediating link is a contingent fact. 4r again% if a man is wit#out knowledge now% t#oug# #e still retains t#e steps of t#e argument% t#oug# t#ere is no c#ange in #imself or in t#e fact and no lapse of memory on #is part& t#en neit#er #ad #e knowledge previously. But t#e mediating link% not being necessary% may #ave peris#ed in t#e interval& and if so% t#oug# t#ere be no c#ange in #im B

nor in t#e fact% and t#oug# #e will still retain t#e steps of t#e argument% yet #e #as not knowledge% and t#erefore #ad not knowledge before. =ven if t#e link #as not actually peris#ed but is liable to peris#% t#is situation is possible and mig#t occur. But suc# a condition cannot be knowledge. 6#en t#e conclusion is necessary% t#e middle t#roug# w#ic# it was proved may yet $uite easily be non necessary. Cou can in fact infer t#e necessary even from a non necessary premiss% 'ust as you can infer t#e true from t#e not true. 4n t#e ot#er #and% w#en t#e middle is necessary t#e conclusion must be necessary& 'ust as true premisses always give a true conclusion. "#us% if A is necessarily predicated of B and B of 7% t#en A is necessarily predicated of 7. But w#en t#e conclusion is nonnecessary t#e middle cannot be necessary eit#er. "#us1 let A be predicated non necessarily of 7 but necessarily of B% and let B be a necessary predicate of 7& t#en A too will be a necessary predicate of 7% w#ic# by #ypot#esis it is not. "o sum up% t#en1 demonstrative knowledge must be knowledge of a necessary ne!us% and t#erefore must clearly be obtained t#roug# a necessary middle term& ot#erwise its possessor will know neit#er t#e cause nor t#e fact t#at #is conclusion is a necessary conne!ion. =it#er #e will mistake t#e non necessary for t#e necessary and believe t#e necessity of t#e conclusion wit#out knowing it% or else #e will not even believe it * in w#ic# case #e will be e$ually ignorant% w#et#er #e actually infers t#e mere fact t#roug# middle terms or t#e reasoned fact and from immediate premisses. 4f accidents t#at are not essential according to our definition of essential t#ere is no demonstrative knowledge& for since an accident% in t#e sense in w#ic# I #ere speak of it% may also not in#ere% it is impossible to prove its in#erence as a necessary conclusion. A difficulty% #owever% mig#t be raised as to w#y in dialectic% if t#e conclusion is not a necessary conne!ion% suc# and suc# determinate premisses s#ould be proposed in order to deal wit# suc# and suc# determinate problems. 6ould not t#e result be t#e same if one asked any $uestions w#atever and t#en merely stated one)s conclusion/ "#e solution is t#at determinate $uestions #ave to be put% not because t#e replies to t#em affirm facts w#ic# necessitate facts affirmed by t#e conclusion% but because t#ese answers are propositions w#ic# if t#e answerer affirm% #e must affirm t#e conclusion and affirm it wit# trut# if t#ey are true. 9ince it is 'ust t#ose attributes wit#in every genus w#ic# are essential and possessed by t#eir respective sub'ects as suc# t#at are necessary it is clear t#at bot# t#e conclusions and t#e premisses of demonstrations w#ic# produce scientific knowledge are essential. +or accidents are not necessary1 and% furt#er% since accidents are not necessary one does not necessarily #ave reasoned knowledge of a conclusion drawn from t#em -t#is is so even if t#e accidental premisses are invariable but not essential% as in proofs t#roug# signs& for t#oug# t#e conclusion be actually essential% one will not know it as essential nor know its reason.& but to #ave reasoned knowledge of a conclusion is to know it t#roug# its cause. 6e may conclude t#at t#e middle must be conse$uentially connected wit# t#e minor% and t#e ma'or wit# t#e middle. B It follows t#at we cannot in demonstrating pass from one genus to anot#er. 6e cannot% for instance% prove geometrical trut#s by arit#metic. +or t#ere are t#ree elements in demonstration1 -1. w#at is proved% t#e conclusion * an attribute in#ering essentially in a genus& -5. t#e a!ioms% i.e. a!ioms w#ic# are premisses of demonstration& -:. t#e sub'ect * genus w#ose attributes% i.e. essential properties% are revealed by t#e demonstration. "#e a!ioms w#ic# are premisses of demonstration may be identical in two or more sciences1 but in t#e case of two different genera suc# as arit#metic and geometry you cannot apply arit#metical demonstration to t#e properties of D

magnitudes unless t#e magnitudes in $uestion are numbers. 3ow in certain cases transference is possible I will e!plain later. Arit#metical demonstration and t#e ot#er sciences likewise possess% eac# of t#em% t#eir own genera& so t#at if t#e demonstration is to pass from one sp#ere to anot#er% t#e genus must be eit#er absolutely or to some e!tent t#e same. If t#is is not so% transference is clearly impossible% because t#e e!treme and t#e middle terms must be drawn from t#e same genus1 ot#erwise% as predicated% t#ey will not be essential and will t#us be accidents. "#at is w#y it cannot be proved by geometry t#at opposites fall under one science% nor even t#at t#e product of two cubes is a cube. 0or can t#e t#eorem of any one science be demonstrated by means of anot#er science% unless t#ese t#eorems are related as subordinate to superior -e.g. as optical t#eorems to geometry or #armonic t#eorems to arit#metic.. Geometry again cannot prove of lines any property w#ic# t#ey do not possess $ua lines% i.e. in virtue of t#e fundamental trut#s of t#eir peculiar genus1 it cannot s#ow% for e!ample% t#at t#e straig#t line is t#e most beautiful of lines or t#e contrary of t#e circle& for t#ese $ualities do not belong to lines in virtue of t#eir peculiar genus% but t#roug# some property w#ic# it s#ares wit# ot#er genera. D It is also clear t#at if t#e premisses from w#ic# t#e syllogism proceeds are commensurately universal% t#e conclusion of suc# i.e. in t#e un$ualified sense * must also be eternal. "#erefore no attribute can be demonstrated nor known by strictly scientific knowledge to in#ere in peris#able t#ings. "#e proof can only be accidental% because t#e attribute)s conne!ion wit# its peris#able sub'ect is not commensurately universal but temporary and special. If suc# a demonstration is made% one premiss must be peris#able and not commensurately universal -peris#able because only if it is peris#able will t#e conclusion be peris#able& not commensurately universal% because t#e predicate will be predicable of some instances of t#e sub'ect and not of ot#ers.& so t#at t#e conclusion can only be t#at a fact is true at t#e moment * not commensurately and universally. "#e same is true of definitions% since a definition is eit#er a primary premiss or a conclusion of a demonstration% or else only differs from a demonstration in t#e order of its terms. 2emonstration and science of merely fre$uent occurrences * e.g. of eclipse as #appening to t#e moon * are% as suc#% clearly eternal1 w#ereas so far as t#ey are not eternal t#ey are not fully commensurate. 4t#er sub'ects too #ave properties attac#ing to t#em in t#e same way as eclipse attac#es to t#e moon. E It is clear t#at if t#e conclusion is to s#ow an attribute in#ering as suc#% not#ing can be demonstrated e!cept from its (appropriate) basic trut#s. 7onse$uently a proof even from true% indemonstrable% and immediate premisses does not constitute knowledge. 9uc# proofs are like Bryson)s met#od of s$uaring t#e circle& for t#ey operate by taking as t#eir middle a common c#aracter * a c#aracter% t#erefore% w#ic# t#e sub'ect may s#are wit# anot#er * and conse$uently t#ey apply e$ually to sub'ects different in kind. "#ey t#erefore afford knowledge of an attribute only as in#ering accidentally% not as belonging to its sub'ect as suc#1 ot#erwise t#ey would not #ave been applicable to anot#er genus. 4ur knowledge of any attribute)s conne!ion wit# a sub'ect is accidental unless we know t#at conne!ion t#roug# t#e middle term in virtue of w#ic# it in#eres% and as an inference from basic premisses essential and (appropriate) to t#e sub'ect * unless we know% e.g. t#e property of E

possessing angles e$ual to two rig#t angles as belonging to t#at sub'ect in w#ic# it in#eres essentially% and as inferred from basic premisses essential and (appropriate) to t#at sub'ect1 so t#at if t#at middle term also belongs essentially to t#e minor% t#e middle must belong to t#e same kind as t#e ma'or and minor terms. "#e only e!ceptions to t#is rule are suc# cases as t#eorems in #armonics w#ic# are demonstrable by arit#metic. 9uc# t#eorems are proved by t#e same middle terms as arit#metical properties% but wit# a $ualification * t#e fact falls under a separate science -for t#e sub'ect genus is separate.% but t#e reasoned fact concerns t#e superior science% to w#ic# t#e attributes essentially belong. "#us% even t#ese apparent e!ceptions s#ow t#at no attribute is strictly demonstrable e!cept from its (appropriate) basic trut#s% w#ic#% #owever% in t#e case of t#ese sciences #ave t#e re$uisite identity of c#aracter. It is no less evident t#at t#e peculiar basic trut#s of eac# in#ering attribute are indemonstrable& for basic trut#s from w#ic# t#ey mig#t be deduced would be basic trut#s of all t#at is% and t#e science to w#ic# t#ey belonged would possess universal sovereignty. "#is is so because #e knows better w#ose knowledge is deduced from #ig#er causes% for #is knowledge is from prior premisses w#en it derives from causes t#emselves uncaused1 #ence% if #e knows better t#an ot#ers or best of all% #is knowledge would be science in a #ig#er or t#e #ig#est degree. But% as t#ings are% demonstration is not transferable to anot#er genus% wit# suc# e!ceptions as we #ave mentioned of t#e application of geometrical demonstrations to t#eorems in mec#anics or optics% or of arit#metical demonstrations to t#ose of #armonics. It is #ard to be sure w#et#er one knows or not& for it is #ard to be sure w#et#er one)s knowledge is based on t#e basic trut#s appropriate to eac# attribute * t#e differentia of true knowledge. 6e t#ink we #ave scientific knowledge if we #ave reasoned from true and primary premisses. But t#at is not so1 t#e conclusion must be #omogeneous wit# t#e basic facts of t#e science. 1F I call t#e basic trut#s of every genus t#ose clements in it t#e e!istence of w#ic# cannot be proved. As regards bot# t#ese primary trut#s and t#e attributes dependent on t#em t#e meaning of t#e name is assumed. "#e fact of t#eir e!istence as regards t#e primary trut#s must be assumed& but it #as to be proved of t#e remainder% t#e attributes. "#us we assume t#e meaning alike of unity% straig#t% and triangular& but w#ile as regards unity and magnitude we assume also t#e fact of t#eir e!istence% in t#e case of t#e remainder proof is re$uired. 4f t#e basic trut#s used in t#e demonstrative sciences some are peculiar to eac# science% and some are common% but common only in t#e sense of analogous% being of use only in so far as t#ey fall wit#in t#e genus constituting t#e province of t#e science in $uestion. Peculiar trut#s are% e.g. t#e definitions of line and straig#t& common trut#s are suc# as (take e$uals from e$uals and e$uals remain). 4nly so muc# of t#ese common trut#s is re$uired as falls wit#in t#e genus in $uestion1 for a trut# of t#is kind will #ave t#e same force even if not used generally but applied by t#e geometer only to magnitudes% or by t#e arit#metician only to numbers. Also peculiar to a science are t#e sub'ects t#e e!istence as well as t#e meaning of w#ic# it assumes% and t#e essential attributes of w#ic# it investigates% e.g. in arit#metic units% in geometry points and lines. Bot# t#e e!istence and t#e meaning of t#e sub'ects are assumed by t#ese sciences& but of t#eir essential attributes only t#e meaning is assumed. +or e!ample arit#metic assumes t#e meaning of odd and even% s$uare and cube% geometry t#at of incommensurable% or of deflection or verging of lines% w#ereas t#e e!istence of t#ese attributes is demonstrated by means of t#e a!ioms and from previous conclusions as premisses. Astronomy too proceeds in t#e same way. +or indeed every demonstrative science #as t#ree elements1 -1. 1F

t#at w#ic# it posits% t#e sub'ect genus w#ose essential attributes it e!amines& -5. t#e so called a!ioms% w#ic# are primary premisses of its demonstration& -:. t#e attributes% t#e meaning of w#ic# it assumes. Cet some sciences may very well pass over some of t#ese elements& e.g. we mig#t not e!pressly posit t#e e!istence of t#e genus if its e!istence were obvious -for instance% t#e e!istence of #ot and cold is more evident t#an t#at of number.& or we mig#t omit to assume e!pressly t#e meaning of t#e attributes if it were well understood. In t#e way t#e meaning of a!ioms% suc# as ("ake e$uals from e$uals and e$uals remain)% is well known and so not e!pressly assumed. 0evert#eless in t#e nature of t#e case t#e essential elements of demonstration are t#ree1 t#e sub'ect% t#e attributes% and t#e basic premisses. "#at w#ic# e!presses necessary self grounded fact% and w#ic# we must necessarily believe% is distinct bot# from t#e #ypot#eses of a science and from illegitimate postulate * I say (must believe)% because all syllogism% and t#erefore a fortiori demonstration% is addressed not to t#e spoken word% but to t#e discourse wit#in t#e soul% and t#oug# we can always raise ob'ections to t#e spoken word% to t#e inward discourse we cannot always ob'ect. "#at w#ic# is capable of proof but assumed by t#e teac#er wit#out proof is% if t#e pupil believes and accepts it% #ypot#esis% t#oug# only in a limited sense #ypot#esis * t#at is% relatively to t#e pupil& if t#e pupil #as no opinion or a contrary opinion on t#e matter% t#e same assumption is an illegitimate postulate. "#erein lies t#e distinction between #ypot#esis and illegitimate postulate1 t#e latter is t#e contrary of t#e pupil)s opinion% demonstrable% but assumed and used wit#out demonstration. "#e definition * vi,. t#ose w#ic# are not e!pressed as statements t#at anyt#ing is or is not * are not #ypot#eses1 but it is in t#e premisses of a science t#at its #ypot#eses are contained. 2efinitions re$uire only to be understood% and t#is is not #ypot#esis * unless it be contended t#at t#e pupil)s #earing is also an #ypot#esis re$uired by t#e teac#er. 3ypot#eses% on t#e contrary% postulate facts on t#e being of w#ic# depends t#e being of t#e fact inferred. 0or are t#e geometer)s #ypot#eses false% as some #ave #eld% urging t#at one must not employ false#ood and t#at t#e geometer is uttering false#ood in stating t#at t#e line w#ic# #e draws is a foot long or straig#t% w#en it is actually neit#er. "#e trut# is t#at t#e geometer does not draw any conclusion from t#e being of t#e particular line of w#ic# #e speaks% but from w#at #is diagrams symboli,e. A furt#er distinction is t#at all #ypot#eses and illegitimate postulates are eit#er universal or particular% w#ereas a definition is neit#er. 11 9o demonstration does not necessarily imply t#e being of +orms nor a 4ne beside a Many% but it does necessarily imply t#e possibility of truly predicating one of many& since wit#out t#is possibility we cannot save t#e universal% and if t#e universal goes% t#e middle term goes witb. it% and so demonstration becomes impossible. 6e conclude% t#en% t#at t#ere must be a single identical term une$uivocally predicable of a number of individuals. "#e law t#at it is impossible to affirm and deny simultaneously t#e same predicate of t#e same sub'ect is not e!pressly posited by any demonstration e!cept w#en t#e conclusion also #as to be e!pressed in t#at form& in w#ic# case t#e proof lays down as its ma'or premiss t#at t#e ma'or is truly affirmed of t#e middle but falsely denied. It makes no difference% #owever% if we add to t#e middle% or again to t#e minor term% t#e corresponding negative. +or grant a minor term of w#ic# it is true to predicate man * even if it be also true to predicate not man of it * still grant simply t#at man is animal and not not animal% and t#e conclusion follows1 for it will still be true to say t#at 7allias * even if it be also true to say t#at not 7allias * is animal and not not animal. "#e reason is t#at t#e ma'or term is predicable not only of t#e middle% but of somet#ing ot#er t#an t#e 11

middle as well% being of wider application& so t#at t#e conclusion is not affected even if t#e middle is e!tended to cover t#e original middle term and also w#at is not t#e original middle term. "#e law t#at every predicate can be eit#er truly affirmed or truly denied of every sub'ect is posited by suc# demonstration as uses reductio ad impossibile% and t#en not always universally% but so far as it is re$uisite& wit#in t#e limits% t#at is% of t#e genus * t#e genus% I mean -as I #ave already e!plained.% to w#ic# t#e man of science applies #is demonstrations. In virtue of t#e common elements of demonstration * I mean t#e common a!ioms w#ic# are used as premisses of demonstration% not t#e sub'ects nor t#e attributes demonstrated as belonging to t#em * all t#e sciences #ave communion wit# one anot#er% and in communion wit# t#em all is dialectic and any science w#ic# mig#t attempt a universal proof of a!ioms suc# as t#e law of e!cluded middle% t#e law t#at t#e subtraction of e$uals from e$uals leaves e$ual remainders% or ot#er a!ioms of t#e same kind. 2ialectic #as no definite sp#ere of t#is kind% not being confined to a single genus. 4t#erwise its met#od would not be interrogative& for t#e interrogative met#od is barred to t#e demonstrator% w#o cannot use t#e opposite facts to prove t#e same ne!us. "#is was s#own in my work on t#e syllogism. 15 If a syllogistic $uestion is e$uivalent to a proposition embodying one of t#e two sides of a contradiction% and if eac# science #as its peculiar propositions from w#ic# its peculiar conclusion is developed% t#en t#ere is suc# a t#ing as a distinctively scientific $uestion% and it is t#e interrogative form of t#e premisses from w#ic# t#e (appropriate) conclusion of eac# science is developed. 3ence it is clear t#at not every $uestion will be relevant to geometry% nor to medicine% nor to any ot#er science1 only t#ose $uestions will be geometrical w#ic# form premisses for t#e proof of t#e t#eorems of geometry or of any ot#er science% suc# as optics% w#ic# uses t#e same basic trut#s as geometry. 4f t#e ot#er sciences t#e like is true. 4f t#ese $uestions t#e geometer is bound to give #is account% using t#e basic trut#s of geometry in con'unction wit# #is previous conclusions& of t#e basic trut#s t#e geometer% as suc#% is not bound to give any account. "#e like is true of t#e ot#er sciences. "#ere is a limit% t#en% to t#e $uestions w#ic# we may put to eac# man of science& nor is eac# man of science bound to answer all in$uiries on eac# several sub'ect% but only suc# as fall wit#in t#e defined field of #is own science. If% t#en% in controversy wit# a geometer $ua geometer t#e disputant confines #imself to geometry and proves anyt#ing from geometrical premisses% #e is clearly to be applauded& if #e goes outside t#ese #e will be at fault% and obviously cannot even refute t#e geometer e!cept accidentally. 4ne s#ould t#erefore not discuss geometry among t#ose w#o are not geometers% for in suc# a company an unsound argument will pass unnoticed. "#is is correspondingly true in t#e ot#er sciences. 9ince t#ere are (geometrical) $uestions% does it follow t#at t#ere are also distinctively (ungeometrical) $uestions/ +urt#er% in eac# special science * geometry for instance * w#at kind of error is it t#at may vitiate $uestions% and yet not e!clude t#em from t#at science/ Again% is t#e erroneous conclusion one constructed from premisses opposite to t#e true premisses% or is it formal fallacy t#oug# drawn from geometrical premisses/ 4r% per#aps% t#e erroneous conclusion is due to t#e drawing of premisses from anot#er science& e.g. in a geometrical controversy a musical $uestion is distinctively ungeometrical% w#ereas t#e notion t#at parallels meet is in one sense geometrical% being ungeometrical in a different fas#ion1 t#e reason being t#at (ungeometrical)% like (unr#yt#mical)% is e$uivocal% meaning in t#e one case not geometry at all% in t#e ot#er bad geometry/ It is t#is error% i.e. error based on premisses of t#is kind * (of) t#e science 15

but false * t#at is t#e contrary of science. In mat#ematics t#e formal fallacy is not so common% because it is t#e middle term in w#ic# t#e ambiguity lies% since t#e ma'or is predicated of t#e w#ole of t#e middle and t#e middle of t#e w#ole of t#e minor -t#e predicate of course never #as t#e prefi! (all).& and in mat#ematics one can% so to speak% see t#ese middle terms wit# an intellectual vision% w#ile in dialectic t#e ambiguity may escape detection. =.g. (Is every circle a figure/) A diagram s#ows t#at t#is is so% but t#e minor premiss (Are epics circles/) is s#own by t#e diagram to be false. If a proof #as an inductive minor premiss% one s#ould not bring an (ob'ection) against it. +or since every premiss must be applicable to a number of cases -ot#erwise it will not be true in every instance% w#ic#% since t#e syllogism proceeds from universals% it must be.% t#en assuredly t#e same is true of an (ob'ection)& since premisses and (ob'ections) are so far t#e same t#at anyt#ing w#ic# can be validly advanced as an (ob'ection) must be suc# t#at it could take t#e form of a premiss% eit#er demonstrative or dialectical. 4n t#e ot#er #and% arguments formally illogical do sometimes occur t#roug# taking as middles mere attributes of t#e ma'or and minor terms. An instance of t#is is 7aeneus) proof t#at fire increases in geometrical proportion1 (+ire)% #e argues% (increases rapidly% and so does geometrical proportion). "#ere is no syllogism so% but t#ere is a syllogism if t#e most rapidly increasing proportion is geometrical and t#e most rapidly increasing proportion is attributable to fire in its motion. 9ometimes% no doubt% it is impossible to reason from premisses predicating mere attributes1 but sometimes it is possible% t#oug# t#e possibility is overlooked. If false premisses could never give true conclusions (resolution) would be easy% for premisses and conclusion would in t#at case inevitably reciprocate. I mig#t t#en argue t#us1 let A be an e!isting fact& let t#e e!istence of A imply suc# and suc# facts actually known to me to e!ist% w#ic# we may call B. I can now% since t#ey reciprocate% infer A from B. Reciprocation of premisses and conclusion is more fre$uent in mat#ematics% because mat#ematics takes definitions% but never an accident% for its premisses * a second c#aracteristic distinguis#ing mat#ematical reasoning from dialectical disputations. A science e!pands not by t#e interposition of fres# middle terms% but by t#e apposition of fres# e!treme terms. =.g. A is predicated of B% B of 7% 7 of 2% and so indefinitely. 4r t#e e!pansion may be lateral1 e.g. one ma'or A% may be proved of two minors% 7 and =. "#us let A represent number * a number or number taken indeterminately& B determinate odd number& 7 any particular odd number. 6e can t#en predicate A of 7. 0e!t let 2 represent determinate even number% and = even number. "#en A is predicable of =. 1: Gnowledge of t#e fact differs from knowledge of t#e reasoned fact. "o begin wit#% t#ey differ wit#in t#e same science and in two ways1 -1. w#en t#e premisses of t#e syllogism are not immediate -for t#en t#e pro!imate cause is not contained in t#em * a necessary condition of knowledge of t#e reasoned fact.1 -5. w#en t#e premisses are immediate% but instead of t#e cause t#e better known of t#e two reciprocals is taken as t#e middle& for of two reciprocally predicable terms t#e one w#ic# is not t#e cause may $uite easily be t#e better known and so become t#e middle term of t#e demonstration. "#us -5. -a. you mig#t prove as follows t#at t#e planets are near because t#ey do not twinkle1 let 7 be t#e planets% B not twinkling% A pro!imity. "#en B is predicable of 7& for t#e planets do not twinkle. But A is also predicable of B% since t#at w#ic# does not twinkle is near * we must take t#is trut# as #aving been reac#ed by induction or sense perception. "#erefore A is a necessary predicate of 7& so t#at we #ave demonstrated t#at t#e planets are near. "#is syllogism% t#en% proves not t#e reasoned fact but only t#e fact& since t#ey 1:

are not near because t#ey do not twinkle% but% because t#ey are near% do not twinkle. "#e ma'or and middle of t#e proof% #owever% may be reversed% and t#en t#e demonstration will be of t#e reasoned fact. "#us1 let 7 be t#e planets% B pro!imity% A not twinkling. "#en B is an attribute of 7% and A * not twinkling * of B. 7onse$uently A is predicable of 7% and t#e syllogism proves t#e reasoned fact% since its middle term is t#e pro!imate cause. Anot#er e!ample is t#e inference t#at t#e moon is sp#erical from its manner of wa!ing. "#us1 since t#at w#ic# so wa!es is sp#erical% and since t#e moon so wa!es% clearly t#e moon is sp#erical. Put in t#is form% t#e syllogism turns out to be proof of t#e fact% but if t#e middle and ma'or be reversed it is proof of t#e reasoned fact& since t#e moon is not sp#erical because it wa!es in a certain manner% but wa!es in suc# a manner because it is sp#erical. -Het 7 be t#e moon% B sp#erical% and A wa!ing.. Again -b.% in cases w#ere t#e cause and t#e effect are not reciprocal and t#e effect is t#e better known% t#e fact is demonstrated but not t#e reasoned fact. "#is also occurs -1. w#en t#e middle falls outside t#e ma'or and minor% for #ere too t#e strict cause is not given% and so t#e demonstration is of t#e fact% not of t#e reasoned fact. +or e!ample% t#e $uestion (6#y does not a wall breat#e/) mig#t be answered% (Because it is not an animal)& but t#at answer would not give t#e strict cause% because if not being an animal causes t#e absence of respiration% t#en being an animal s#ould be t#e cause of respiration% according to t#e rule t#at if t#e negation of causes t#e non in#erence of y% t#e affirmation of ! causes t#e in#erence of y& e.g. if t#e disproportion of t#e #ot and cold elements is t#e cause of ill #ealt#% t#eir proportion is t#e cause of #ealt#& and conversely% if t#e assertion of ! causes t#e in#erence of y% t#e negation of ! must cause y)s non in#erence. But in t#e case given t#is conse$uence does not result& for not every animal breat#es. A syllogism wit# t#is kind of cause takes place in t#e second figure. "#us1 let A be animal% B respiration% 7 wall. "#en A is predicable of all B -for all t#at breat#es is animal.% but of no 7& and conse$uently B is predicable of no 7& t#at is% t#e wall does not breat#e. 9uc# causes are like far fetc#ed e!planations% w#ic# precisely consist in making t#e cause too remote% as in Anac#arsis) account of w#y t#e 9cyt#ians #ave no flute players& namely because t#ey #ave no vines. "#us% t#en% do t#e syllogism of t#e fact and t#e syllogism of t#e reasoned fact differ wit#in one science and according to t#e position of t#e middle terms. But t#ere is anot#er way too in w#ic# t#e fact and t#e reasoned fact differ% and t#at is w#en t#ey are investigated respectively by different sciences. "#is occurs in t#e case of problems related to one anot#er as subordinate and superior% as w#en optical problems are subordinated to geometry% mec#anical problems to stereometry% #armonic problems to arit#metic% t#e data of observation to astronomy. -9ome of t#ese sciences bear almost t#e same name& e.g. mat#ematical and nautical astronomy% mat#ematical and acoustical #armonics.. 3ere it is t#e business of t#e empirical observers to know t#e fact% of t#e mat#ematicians to know t#e reasoned fact& for t#e latter are in possession of t#e demonstrations giving t#e causes% and are often ignorant of t#e fact1 'ust as we #ave often a clear insig#t into a universal% but t#roug# lack of observation are ignorant of some of its particular instances. "#ese conne!ions #ave a perceptible e!istence t#oug# t#ey are manifestations of forms. +or t#e mat#ematical sciences concern forms1 t#ey do not demonstrate properties of a substratum% since% even t#oug# t#e geometrical sub'ects are predicable as properties of a perceptible substratum% it is not as t#us predicable t#at t#e mat#ematician demonstrates properties of t#em. As optics is related to geometry% so anot#er science is related to optics% namely t#e t#eory of t#e rainbow. 3ere knowledge of t#e fact is wit#in t#e province of t#e natural p#ilosop#er% knowledge of t#e reasoned fact wit#in t#at of t#e optician% eit#er $ua optician or $ua mat#ematical optician. Many sciences not standing in t#is mutual relation enter into it at points& e.g. medicine and geometry1 it is t#e p#ysician)s business to know t#at circular wounds #eal more slowly% t#e geometer)s to know t#e reason w#y. 1<

1< 4f all t#e figures t#e most scientific is t#e first. "#us% it is t#e ve#icle of t#e demonstrations of all t#e mat#ematical sciences% suc# as arit#metic% geometry% and optics% and practically all of all sciences t#at investigate causes1 for t#e syllogism of t#e reasoned fact is eit#er e!clusively or generally speaking and in most cases in t#is figure * a second proof t#at t#is figure is t#e most scientific& for grasp of a reasoned conclusion is t#e primary condition of knowledge. "#irdly% t#e first is t#e only figure w#ic# enables us to pursue knowledge of t#e essence of a t#ing. In t#e second figure no affirmative conclusion is possible% and knowledge of a t#ing)s essence must be affirmative& w#ile in t#e t#ird figure t#e conclusion can be affirmative% but cannot be universal% and essence must #ave a universal c#aracter1 e.g. man is not two footed animal in any $ualified sense% but universally. +inally% t#e first figure #as no need of t#e ot#ers% w#ile it is by means of t#e first t#at t#e ot#er two figures are developed% and #ave t#eir intervals closepacked until immediate premisses are reac#ed. 7learly% t#erefore% t#e first figure is t#e primary condition of knowledge. 1@ Iust as an attribute A may -as we saw. be atomically connected wit# a sub'ect B% so its disconne!ion may be atomic. I call (atomic) conne!ions or disconne!ions w#ic# involve no intermediate term& since in t#at case t#e conne!ion or disconne!ion will not be mediated by somet#ing ot#er t#an t#e terms t#emselves. It follows t#at if eit#er A or B% or bot# A and B% #ave a genus% t#eir disconne!ion cannot be primary. "#us1 let 7 be t#e genus of A. "#en% if 7 is not t#e genus of B * for A may well #ave a genus w#ic# is not t#e genus of B * t#ere will be a syllogism proving A)s disconne!ion from B t#us1 all A is 7% no B is 7% t#erefore no B is A. 4r if it is B w#ic# #as a genus 2% we #ave all B is 2% no 2 is A% t#erefore no B is A% by syllogism& and t#e proof will be similar if bot# A and B #ave a genus. "#at t#e genus of A need not be t#e genus of B and vice versa% is s#own by t#e e!istence of mutually e!clusive coordinate series of predication. If no term in t#e series A72... is predicable of any term in t#e series B=+...% and if G * a term in t#e former series * is t#e genus of A% clearly G will not be t#e genus of B& since% if it were% t#e series would not be mutually e!clusive. 9o also if B #as a genus% it will not be t#e genus of A. If% on t#e ot#er #and% neit#er A nor B #as a genus and A does not in#ere in B% t#is disconne!ion must be atomic. If t#ere be a middle term% one or ot#er of t#em is bound to #ave a genus% for t#e syllogism will be eit#er in t#e first or t#e second figure. If it is in t#e first% B will #ave a genus * for t#e premiss containing it must be affirmative1 if in t#e second% eit#er A or B indifferently% since syllogism is possible if eit#er is contained in a negative premiss% but not if bot# premisses are negative. 3ence it is clear t#at one t#ing may be atomically disconnected from anot#er% and we #ave stated w#en and #ow t#is is possible.

1@

1A Ignorance * defined not as t#e negation of knowledge but as a positive state of mind * is error produced by inference. -1. Het us first consider propositions asserting a predicate)s immediate conne!ion wit# or disconne!ion from a sub'ect. 3ere% it is true% positive error may befall one in alternative ways& for it may arise w#ere one directly believes a conne!ion or disconne!ion as well as w#ere one)s belief is ac$uired by inference. "#e error% #owever% t#at consists in a direct belief is wit#out complication& but t#e error resulting from inference * w#ic# #ere concerns us * takes many forms. "#us% let A be atomically disconnected from all B1 t#en t#e conclusion inferred t#roug# a middle term 7% t#at all B is A% will be a case of error produced by syllogism. 0ow% two cases are possible. =it#er -a. bot# premisses% or -b. one premiss only% may be false. -a. If neit#er A is an attribute of any 7 nor 7 of any B% w#ereas t#e contrary was posited in bot# cases% bot# premisses will be false. -7 may $uite well be so related to A and B t#at 7 is neit#er subordinate to A nor a universal attribute of B1 for B% since A was said to be primarily disconnected from B% cannot #ave a genus% and A need not necessarily be a universal attribute of all t#ings. 7onse$uently bot# premisses may be false.. 4n t#e ot#er #and% -b. one of t#e premisses may be true% t#oug# not eit#er indifferently but only t#e ma'or A 7 since% B #aving no genus% t#e premiss 7 B will always be false% w#ile A 7 may be true. "#is is t#e case if% for e!ample% A is related atomically to bot# 7 and B& because w#en t#e same term is related atomically to more terms t#an one% neit#er of t#ose terms will belong to t#e ot#er. It is% of course% e$ually t#e case if A 7 is not atomic. =rror of attribution% t#en% occurs t#roug# t#ese causes and in t#is form only * for we found t#at no syllogism of universal attribution was possible in any figure but t#e first. 4n t#e ot#er #and% an error of non attribution may occur eit#er in t#e first or in t#e second figure. Het us t#erefore first e!plain t#e various forms it takes in t#e first figure and t#e c#aracter of t#e premisses in eac# case. -c. It may occur w#en bot# premisses are false& e.g. supposing A atomically connected wit# bot# 7 and B% if it be t#en assumed t#at no 7 is and all B is 7% bot# premisses are false. -d. It is also possible w#en one is false. "#is may be eit#er premiss indifferently. A 7 may be true% 7 B false * A 7 true because A is not an attribute of all t#ings% 7 B false because 7% w#ic# never #as t#e attribute A% cannot be an attribute of B& for if 7 B were true% t#e premiss A 7 would no longer be true% and besides if bot# premisses were true% t#e conclusion would be true. 4r again% 7 B may be true and A 7 false& e.g. if bot# 7 and A contain B as genera% one of t#em must be subordinate to t#e ot#er% so t#at if t#e premiss takes t#e form 0o 7 is A% it will be false. "#is makes it clear t#at w#et#er eit#er or bot# premisses are false% t#e conclusion will e$ually be false. In t#e second figure t#e premisses cannot bot# be w#olly false& for if all B is A% no middle term can be wit# trut# universally affirmed of one e!treme and universally denied of t#e ot#er1 but premisses in w#ic# t#e middle is affirmed of one e!treme and denied of t#e ot#er are t#e necessary condition if one is to get a valid inference at all. "#erefore if% taken in t#is way% t#ey are w#olly false% t#eir contraries conversely s#ould be w#olly true. But t#is is impossible. 4n t#e ot#er #and% t#ere is not#ing to prevent bot# premisses being partially false& e.g. if actually some A is 7 and some B is 7% t#en if it is premised t#at all A is 7 and no B is 7% bot# premisses are false% yet partially% not w#olly% false. "#e same is true if t#e ma'or is made negative instead of t#e minor. 4r one premiss may be w#olly false% and it may be eit#er of t#em. "#us% supposing t#at actually an attribute of all A must also be an attribute of all B% t#en if 7 is yet taken to be a universal attribute of all but universally non attributable to B% 7 A will be true but 7 B false. Again% actually t#at w#ic# is an attribute of no B will not be an attribute of all A eit#er& for if it be 1A

an attribute of all A% it will also be an attribute of all B% w#ic# is contrary to supposition& but if 7 be nevert#eless assumed to be a universal attribute of A% but an attribute of no B% t#en t#e premiss 7 B is true but t#e ma'or is false. "#e case is similar if t#e ma'or is made t#e negative premiss. +or in fact w#at is an attribute of no A will not be an attribute of any B eit#er& and if it be yet assumed t#at 7 is universally non attributable to A% but a universal attribute of B% t#e premiss 7 A is true but t#e minor w#olly false. Again% in fact it is false to assume t#at t#at w#ic# is an attribute of all B is an attribute of no A% for if it be an attribute of all B% it must be an attribute of some A. If t#en 7 is nevert#eless assumed to be an attribute of all B but of no A% 7 B will be true but 7 A false. It is t#us clear t#at in t#e case of atomic propositions erroneous inference will be possible not only w#en bot# premisses are false but also w#en only one is false. 1B In t#e case of attributes not atomically connected wit# or disconnected from t#eir sub'ects% -a. -i. as long as t#e false conclusion is inferred t#roug# t#e (appropriate) middle% only t#e ma'or and not bot# premisses can be false. By (appropriate middle) I mean t#e middle term t#roug# w#ic# t#e contradictory * i.e. t#e true conclusion is inferrible. "#us% let A be attributable to B t#roug# a middle term 71 t#en% since to produce a conclusion t#e premiss 7 B must be taken affirmatively% it is clear t#at t#is premiss must always be true% for its $uality is not c#anged. But t#e ma'or A 7 is false% for it is by a c#ange in t#e $uality of A 7 t#at t#e conclusion becomes its contradictory * i.e. true. 9imilarly -ii. if t#e middle is taken from anot#er series of predication& e.g. suppose 2 to be not only contained wit#in A as a part wit#in its w#ole but also predicable of all B. "#en t#e premiss 2 B must remain unc#anged% but t#e $uality of A 2 must be c#anged& so t#at 2 B is always true% A 2 always false. 9uc# error is practically identical wit# t#at w#ic# is inferred t#roug# t#e (appropriate) middle. 4n t#e ot#er #and% -b. if t#e conclusion is not inferred t#roug# t#e (appropriate) middle * -i. w#en t#e middle is subordinate to A but is predicable of no B% bot# premisses must be false% because if t#ere is to be a conclusion bot# must be posited as asserting t#e contrary of w#at is actually t#e fact% and so posited bot# become false1 e.g. suppose t#at actually all 2 is A but no B is 2& t#en if t#ese premisses are c#anged in $uality% a conclusion will follow and bot# of t#e new premisses will be false. 6#en% #owever% -ii. t#e middle 2 is not subordinate to A% A 2 will be true% 2 B false * A 2 true because A was not subordinate to 2% 2 B false because if it #ad been true% t#e conclusion too would #ave been true& but it is e! #ypot#esi false. 6#en t#e erroneous inference is in t#e second figure% bot# premisses cannot be entirely false& since if B is subordinate to A% t#ere can be no middle predicable of all of one e!treme and of none of t#e ot#er% as was stated before. 4ne premiss% #owever% may be false% and it may be eit#er of t#em. "#us% if 7 is actually an attribute of bot# A and B% but is assumed to be an attribute of A only and not of B% 7 A will be true% 7 B false1 or again if 7 be assumed to be attributable to B but to no A% 7 B will be true% 7 A false. 6e #ave stated w#en and t#roug# w#at kinds of premisses error will result in cases w#ere t#e erroneous conclusion is negative. If t#e conclusion is affirmative% -a. -i. it may be inferred t#roug# t#e (appropriate) middle term. In t#is case bot# premisses cannot be false since% as we said before% 7 B must remain unc#anged if t#ere is to be a conclusion% and conse$uently A 7% t#e $uality of w#ic# is c#anged% will always be false. "#is is e$ually true if -ii. t#e middle is taken from anot#er series of predication% as was stated to be t#e case also wit# regard to negative error&

1B

for 2 B must remain unc#anged% w#ile t#e $uality of A 2 must be converted% and t#e type of error is t#e same as before. -b. "#e middle may be inappropriate. "#en -i. if 2 is subordinate to A% A 2 will be true% but 2 B false& since A may $uite well be predicable of several terms no one of w#ic# can be subordinated to anot#er. If% #owever% -ii. 2 is not subordinate to A% obviously A 2% since it is affirmed% will always be false% w#ile 2 B may be eit#er true or false& for A may very well be an attribute of no 2% w#ereas all B is 2% e.g. no science is animal% all music is science. =$ually well A may be an attribute of no 2% and 2 of no B. It emerges% t#en% t#at if t#e middle term is not subordinate to t#e ma'or% not only bot# premisses but eit#er singly may be false. "#us we #ave made it clear #ow many varieties of erroneous inference are liable to #appen and t#roug# w#at kinds of premisses t#ey occur% in t#e case bot# of immediate and of demonstrable trut#s. 1D It is also clear t#at t#e loss of any one of t#e senses entails t#e loss of a corresponding portion of knowledge% and t#at% since we learn eit#er by induction or by demonstration% t#is knowledge cannot be ac$uired. "#us demonstration develops from universals% induction from particulars& but since it is possible to familiari,e t#e pupil wit# even t#e so called mat#ematical abstractions only t#roug# induction * i.e. only because eac# sub'ect genus possesses% in virtue of a determinate mat#ematical c#aracter% certain properties w#ic# can be treated as separate even t#oug# t#ey do not e!ist in isolation * it is conse$uently impossible to come to grasp universals e!cept t#roug# induction. But induction is impossible for t#ose w#o #ave not sense perception. +or it is sense perception alone w#ic# is ade$uate for grasping t#e particulars1 t#ey cannot be ob'ects of scientific knowledge% because neit#er can universals give us knowledge of t#em wit#out induction% nor can we get it t#roug# induction wit#out sense perception. 1E =very syllogism is effected by means of t#ree terms. 4ne kind of syllogism serves to prove t#at A in#eres in 7 by s#owing t#at A in#eres in B and B in 7& t#e ot#er is negative and one of its premisses asserts one term of anot#er% w#ile t#e ot#er denies one term of anot#er. It is clear% t#en% t#at t#ese are t#e fundamentals and so called #ypot#eses of syllogism. Assume t#em as t#ey #ave been stated% and proof is bound to follow * proof t#at A in#eres in 7 t#roug# B% and again t#at A in#eres in B t#roug# some ot#er middle term% and similarly t#at B in#eres in 7. If our reasoning aims at gaining credence and so is merely dialectical% it is obvious t#at we #ave only to see t#at our inference is based on premisses as credible as possible1 so t#at if a middle term between A and B is credible t#oug# not real% one can reason t#roug# it and complete a dialectical syllogism. If% #owever% one is aiming at trut#% one must be guided by t#e real conne!ions of sub'ects and attributes. "#us1 since t#ere are attributes w#ic# are predicated of a sub'ect essentially or naturally and not coincidentally * not% t#at is% in t#e sense in w#ic# we say ("#at w#ite -t#ing. is a man)% w#ic# is not t#e same mode of predication as w#en we say ("#e man is w#ite)1 t#e man is w#ite not because #e is somet#ing else but because #e is man% but t#e w#ite is man because (being w#ite) coincides wit# (#umanity) wit#in one substratum * t#erefore t#ere are terms suc# as are naturally sub'ects of predicates. 9uppose% t#en% 7 suc# a term not itself attributable to anyt#ing else as to a sub'ect% but t#e pro!imate sub'ect of t#e attribute B * i.e. so t#at B 7 is immediate& suppose furt#er = related immediately to +% and + to B. "#e first $uestion is% must t#is 1D

series terminate% or can it proceed to infinity/ "#e second $uestion is as follows1 9uppose not#ing is essentially predicated of A% but A is predicated primarily of 3 and of no intermediate prior term% and suppose 3 similarly related to G and G to B& t#en must t#is series also terminate% or can it too proceed to infinity/ "#ere is t#is muc# difference between t#e $uestions1 t#e first is% is it possible to start from t#at w#ic# is not itself attributable to anyt#ing else but is t#e sub'ect of attributes% and ascend to infinity/ "#e second is t#e problem w#et#er one can start from t#at w#ic# is a predicate but not itself a sub'ect of predicates% and descend to infinity/ A t#ird $uestion is% if t#e e!treme terms are fi!ed% can t#ere be an infinity of middles/ I mean t#is1 suppose for e!ample t#at A in#eres in 7 and B is intermediate between t#em% but between B and A t#ere are ot#er middles% and between t#ese again fres# middles& can t#ese proceed to infinity or can t#ey not/ "#is is t#e e$uivalent of in$uiring% do demonstrations proceed to infinity% i.e. is everyt#ing demonstrable/ 4r do ultimate sub'ect and primary attribute limit one anot#er/ I #old t#at t#e same $uestions arise wit# regard to negative conclusions and premisses1 vi,. if A is attributable to no B% t#en eit#er t#is predication will be primary% or t#ere will be an intermediate term prior to B to w#ic# a is not attributable * G% let us say% w#ic# is attributable to all B * and t#ere may still be anot#er term 3 prior to G% w#ic# is attributable to all G. "#e same $uestions arise% I say% because in t#ese cases too eit#er t#e series of prior terms to w#ic# a is not attributable is infinite or it terminates. 4ne cannot ask t#e same $uestions in t#e case of reciprocating terms% since w#en sub'ect and predicate are convertible t#ere is neit#er primary nor ultimate sub'ect% seeing t#at all t#e reciprocals $ua sub'ects stand in t#e same relation to one anot#er% w#et#er we say t#at t#e sub'ect #as an infinity of attributes or t#at bot# sub'ects and attributes * and we raised t#e $uestion in bot# cases * are infinite in number. "#ese $uestions t#en cannot be asked * unless% indeed% t#e terms can reciprocate by two different modes% by accidental predication in one relation and natural predication in t#e ot#er. 5F 0ow% it is clear t#at if t#e predications terminate in bot# t#e upward and t#e downward direction -by (upward) I mean t#e ascent to t#e more universal% by (downward) t#e descent to t#e more particular.% t#e middle terms cannot be infinite in number. +or suppose t#at A is predicated of +% and t#at t#e intermediates * call t#em B B) BJ... * are infinite% t#en clearly you mig#t descend from and find one term predicated of anot#er ad infinitum% since you #ave an infinity of terms between you and +& and e$ually% if you ascend from +% t#ere are infinite terms between you and A. It follows t#at if t#ese processes are impossible t#ere cannot be an infinity of intermediates between A and +. 0or is it of any effect to urge t#at some terms of t#e series AB...+ are contiguous so as to e!clude intermediates% w#ile ot#ers cannot be taken into t#e argument at all1 w#ic#ever terms of t#e series B...I take% t#e number of intermediates in t#e direction eit#er of A or of + must be finite or infinite1 w#ere t#e infinite series starts% w#et#er from t#e first term or from a later one% is of no moment% for t#e succeeding terms in any case are infinite in number. 51 +urt#er% if in affirmative demonstration t#e series terminates in bot# directions% clearly it will terminate too in negative demonstration. Het us assume t#at we cannot proceed to infinity eit#er by ascending from t#e ultimate term -by (ultimate term) I mean a term suc# as was% not itself attributable to a sub'ect but itself t#e sub'ect of attributes.% or by descending towards an ultimate 1E

from t#e primary term -by (primary term) I mean a term predicable of a sub'ect but not itself a sub'ect.. If t#is assumption is 'ustified% t#e series will also terminate in t#e case of negation. +or a negative conclusion can be proved in all t#ree figures. In t#e first figure it is proved t#us1 no B is A% all 7 is B. In packing t#e interval B 7 we must reac# immediate propositions * as is always t#e case wit# t#e minor premiss * since B 7 is affirmative. As regards t#e ot#er premiss it is plain t#at if t#e ma'or term is denied of a term 2 prior to B% 2 will #ave to be predicable of all B% and if t#e ma'or is denied of yet anot#er term prior to 2% t#is term must be predicable of all 2. 7onse$uently% since t#e ascending series is finite% t#e descent will also terminate and t#ere will be a sub'ect of w#ic# A is primarily non predicable. In t#e second figure t#e syllogism is% all A is B% no 7 is B%..no 7 is A. If proof of t#is is re$uired% plainly it may be s#own eit#er in t#e first figure as above% in t#e second as #ere% or in t#e t#ird. "#e first figure #as been discussed% and we will proceed to display t#e second% proof by w#ic# will be as follows1 all B is 2% no 7 is 2...% since it is re$uired t#at B s#ould be a sub'ect of w#ic# a predicate is affirmed. 0e!t% since 2 is to be proved not to belong to 7% t#en 2 #as a furt#er predicate w#ic# is denied of 7. "#erefore% since t#e succession of predicates affirmed of an ever #ig#er universal terminates% t#e succession of predicates denied terminates too. "#e t#ird figure s#ows it as follows1 all B is A% some B is not 7. "#erefore some A is not 7. "#is premiss% i.e. 7 B% will be proved eit#er in t#e same figure or in one of t#e two figures discussed above. In t#e first and second figures t#e series terminates. If we use t#e t#ird figure% we s#all take as premisses% all = is B% some = is not 7% and t#is premiss again will be proved by a similar prosyllogism. But since it is assumed t#at t#e series of descending sub'ects also terminates% plainly t#e series of more universal non predicables will terminate also. =ven supposing t#at t#e proof is not confined to one met#od% but employs t#em all and is now in t#e first figure% now in t#e second or t#ird * even so t#e regress will terminate% for t#e met#ods are finite in number% and if finite t#ings are combined in a finite number of ways% t#e result must be finite. "#us it is plain t#at t#e regress of middles terminates in t#e case of negative demonstration% if it does so also in t#e case of affirmative demonstration. "#at in fact t#e regress terminates in bot# t#ese cases may be made clear by t#e following dialectical considerations. 55 In t#e case of predicates constituting t#e essential nature of a t#ing% it clearly terminates% seeing t#at if definition is possible% or in ot#er words% if essential form is knowable% and an infinite series cannot be traversed% predicates constituting a t#ing)s essential nature must be finite in number. But as regards predicates generally we #ave t#e following prefatory remarks to make. -1. 6e can affirm wit#out false#ood (t#e w#ite -t#ing. is walking)% and t#at big -t#ing. is a log)& or again% (t#e log is big)% and (t#e man walks). But t#e affirmation differs in t#e two cases. 6#en I affirm (t#e w#ite is a log)% I mean t#at somet#ing w#ic# #appens to be w#ite is a log * not t#at w#ite is t#e substratum in w#ic# log in#eres% for it was not $ua w#ite or $ua a species of w#ite t#at t#e w#ite -t#ing. came to be a log% and t#e w#ite -t#ing. is conse$uently not a log e!cept incidentally. 4n t#e ot#er #and% w#en I affirm (t#e log is w#ite)% I do not mean t#at somet#ing else% w#ic# #appens also to be a log% is w#ite -as I s#ould if I said (t#e musician is w#ite%) w#ic# would mean (t#e man w#o #appens also to be a musician is w#ite).& on t#e contrary% log is #ere t#e substratum * t#e substratum w#ic# actually came to be w#ite% and did so $ua wood or $ua a species of wood and $ua not#ing else.

5F

If we must lay down a rule% let us entitle t#e latter kind of statement predication% and t#e former not predication at all% or not strict but accidental predication. (6#ite) and (log) will t#us serve as types respectively of predicate and sub'ect. 6e s#all assume% t#en% t#at t#e predicate is invariably predicated strictly and not accidentally of t#e sub'ect% for on suc# predication demonstrations depend for t#eir force. It follows from t#is t#at w#en a single attribute is predicated of a single sub'ect% t#e predicate must affirm of t#e sub'ect eit#er some element constituting its essential nature% or t#at it is in some way $ualified% $uantified% essentially related% active% passive% placed% or dated. -5. Predicates w#ic# signify substance signify t#at t#e sub'ect is identical wit# t#e predicate or wit# a species of t#e predicate. Predicates not signifying substance w#ic# are predicated of a sub'ect not identical wit# t#emselves or wit# a species of t#emselves are accidental or coincidental& e.g. w#ite is a coincident of man% seeing t#at man is not identical wit# w#ite or a species of w#ite% but rat#er wit# animal% since man is identical wit# a species of animal. "#ese predicates w#ic# do not signify substance must be predicates of some ot#er sub'ect% and not#ing can be w#ite w#ic# is not also ot#er t#an w#ite. "#e +orms we can dispense wit#% for t#ey are mere sound wit#out sense& and even if t#ere are suc# t#ings% t#ey are not relevant to our discussion% since demonstrations are concerned wit# predicates suc# as we #ave defined. -:. If A is a $uality of B% B cannot be a $uality of A * a $uality of a $uality. "#erefore A and B cannot be predicated reciprocally of one anot#er in strict predication1 t#ey can be affirmed wit#out false#ood of one anot#er% but not genuinely predicated of eac# ot#er. +or one alternative is t#at t#ey s#ould be substantially predicated of one anot#er% i.e. B would become t#e genus or differentia of A * t#e predicate now become sub'ect. But it #as been s#own t#at in t#ese substantial predications neit#er t#e ascending predicates nor t#e descending sub'ects form an infinite series& e.g. neit#er t#e series% man is biped% biped is animal% Kc.% nor t#e series predicating animal of man% man of 7allias% 7allias of a furt#er. sub'ect as an element of its essential nature% is infinite. +or all suc# substance is definable% and an infinite series cannot be traversed in t#oug#t1 conse$uently neit#er t#e ascent nor t#e descent is infinite% since a substance w#ose predicates were infinite would not be definable. 3ence t#ey will not be predicated eac# as t#e genus of t#e ot#er& for t#is would e$uate a genus wit# one of its own species. 0or -t#e ot#er alternative. can a $uale be reciprocally predicated of a $uale% nor any term belonging to an ad'ectival category of anot#er suc# term% e!cept by accidental predication& for all suc# predicates are coincidents and are predicated of substances. 4n t#e ot#er #and * in proof of t#e impossibility of an infinite ascending series * every predication displays t#e sub'ect as some#ow $ualified or $uantified or as c#aracteri,ed under one of t#e ot#er ad'ectival categories% or else is an element in its substantial nature1 t#ese latter are limited in number% and t#e number of t#e widest kinds under w#ic# predications fall is also limited% for every predication must e!#ibit its sub'ect as some#ow $ualified% $uantified% essentially related% acting or suffering% or in some place or at some time. I assume first t#at predication implies a single sub'ect and a single attribute% and secondly t#at predicates w#ic# are not substantial are not predicated of one anot#er. 6e assume t#is because suc# predicates are all coincidents% and t#oug# some are essential coincidents% ot#ers of a different type% yet we maintain t#at all of t#em alike are predicated of some substratum and t#at a coincident is never a substratum * since we do not class as a coincident anyt#ing w#ic# does not owe its designation to its being somet#ing ot#er t#an itself% but always #old t#at any coincident is predicated of some substratum ot#er t#an itself% and t#at anot#er group of coincidents may #ave a different substratum. 9ub'ect to t#ese assumptions t#en% neit#er t#e ascending nor t#e descending series of predication in w#ic# a single attribute is predicated of a single sub'ect is infinite. +or t#e sub'ects of w#ic# coincidents are predicated are as many as t#e constitutive elements of eac# 51

individual substance% and t#ese we #ave seen are not infinite in number% w#ile in t#e ascending series are contained t#ose constitutive elements wit# t#eir coincidents * bot# of w#ic# are finite. 6e conclude t#at t#ere is a given sub'ect -2. of w#ic# some attribute -7. is primarily predicable& t#at t#ere must be an attribute -B. primarily predicable of t#e first attribute% and t#at t#e series must end wit# a term -A. not predicable of any term prior to t#e last sub'ect of w#ic# it was predicated -B.% and of w#ic# no term prior to it is predicable. "#e argument we #ave given is one of t#e so called proofs& an alternative proof follows. Predicates so related to t#eir sub'ects t#at t#ere are ot#er predicates prior to t#em predicable of t#ose sub'ects are demonstrable& but of demonstrable propositions one cannot #ave somet#ing better t#an knowledge% nor can one know t#em wit#out demonstration. 9econdly% if a conse$uent is only known t#roug# an antecedent -vi,. premisses prior to it. and we neit#er know t#is antecedent nor #ave somet#ing better t#an knowledge of it% t#en we s#all not #ave scientific knowledge of t#e conse$uent. "#erefore% if it is possible t#roug# demonstration to know anyt#ing wit#out $ualification and not merely as dependent on t#e acceptance of certain premisses * i.e. #ypot#etically * t#e series of intermediate predications must terminate. If it does not terminate% and beyond any predicate taken as #ig#er t#an anot#er t#ere remains anot#er still #ig#er% t#en every predicate is demonstrable. 7onse$uently% since t#ese demonstrable predicates are infinite in number and t#erefore cannot be traversed% we s#all not know t#em by demonstration. If% t#erefore% we #ave not somet#ing better t#an knowledge of t#em% we cannot t#roug# demonstration #ave un$ualified but only #ypot#etical science of anyt#ing. As dialectical proofs of our contention t#ese may carry conviction% but an analytic process will s#ow more briefly t#at neit#er t#e ascent nor t#e descent of predication can be infinite in t#e demonstrative sciences w#ic# are t#e ob'ect of our investigation. 2emonstration proves t#e in#erence of essential attributes in t#ings. 0ow attributes may be essential for two reasons1 eit#er because t#ey are elements in t#e essential nature of t#eir sub'ects% or because t#eir sub'ects are elements in t#eir essential nature. An e!ample of t#e latter is odd as an attribute of number * t#oug# it is number)s attribute% yet number itself is an element in t#e definition of odd& of t#e former% multiplicity or t#e indivisible% w#ic# are elements in t#e definition of number. In neit#er kind of attribution can t#e terms be infinite. "#ey are not infinite w#ere eac# is related to t#e term below it as odd is to number% for t#is would mean t#e in#erence in odd of anot#er attribute of odd in w#ose nature odd was an essential element1 but t#en number will be an ultimate sub'ect of t#e w#ole infinite c#ain of attributes% and be an element in t#e definition of eac# of t#em. 3ence% since an infinity of attributes suc# as contain t#eir sub'ect in t#eir definition cannot in#ere in a single t#ing% t#e ascending series is e$ually finite. 0ote% moreover% t#at all suc# attributes must so in#ere in t#e ultimate sub'ect * e.g. its attributes in number and number in t#em * as to be commensurate wit# t#e sub'ect and not of wider e!tent. Attributes w#ic# are essential elements in t#e nature of t#eir sub'ects are e$ually finite1 ot#erwise definition would be impossible. 3ence% if all t#e attributes predicated are essential and t#ese cannot be infinite% t#e ascending series will terminate% and conse$uently t#e descending series too. If t#is is so% it follows t#at t#e intermediates between any two terms are also always limited in number. An immediately obvious conse$uence of t#is is t#at demonstrations necessarily involve basic trut#s% and t#at t#e contention of some * referred to at t#e outset * t#at all trut#s are demonstrable is mistaken. +or if t#ere are basic trut#s% -a. not all trut#s are demonstrable% and -b. an infinite regress is impossible& since if eit#er -a. or -b. were not a fact% it would mean t#at no interval was immediate and indivisible% but t#at all intervals were divisible. "#is is true because a conclusion is demonstrated by t#e interposition% not t#e apposition% of a fres# term. If suc# interposition could continue to infinity t#ere mig#t be an infinite number of terms between any 55

two terms& but t#is is impossible if bot# t#e ascending and descending series of predication terminate& and of t#is fact% w#ic# before was s#own dialectically% analytic proof #as now been given. 5: It is an evident corollary of t#ese conclusions t#at if t#e same attribute A in#eres in two terms 7 and 2 predicable eit#er not at all% or not of all instances% of one anot#er% it does not always belong to t#em in virtue of a common middle term. Isosceles and scalene possess t#e attribute of #aving t#eir angles e$ual to two rig#t angles in virtue of a common middle& for t#ey possess it in so far as t#ey are bot# a certain kind of figure% and not in so far as t#ey differ from one anot#er. But t#is is not always t#e case1 for% were it so% if we take B as t#e common middle in virtue of w#ic# A in#eres in 7 and 2% clearly B would in#ere in 7 and 2 t#roug# a second common middle% and t#is in turn would in#ere in 7 and 2 t#roug# a t#ird% so t#at between two terms an infinity of intermediates would fall * an impossibility. "#us it need not always be in virtue of a common middle term t#at a single attribute in#eres in several sub'ects% since t#ere must be immediate intervals. Cet if t#e attribute to be proved common to two sub'ects is to be one of t#eir essential attributes% t#e middle terms involved must be wit#in one sub'ect genus and be derived from t#e same group of immediate premisses& for we #ave seen t#at processes of proof cannot pass from one genus to anot#er. It is also clear t#at w#en A in#eres in B% t#is can be demonstrated if t#ere is a middle term. +urt#er% t#e (elements) of suc# a conclusion are t#e premisses containing t#e middle in $uestion% and t#ey are identical in number wit# t#e middle terms% seeing t#at t#e immediate propositions * or at least suc# immediate propositions as are universal * are t#e (elements). If% on t#e ot#er #and% t#ere is no middle term% demonstration ceases to be possible1 we are on t#e way to t#e basic trut#s. 9imilarly if A does not in#ere in B% t#is can be demonstrated if t#ere is a middle term or a term prior to B in w#ic# A does not in#ere1 ot#erwise t#ere is no demonstration and a basic trut# is reac#ed. "#ere are% moreover% as many (elements) of t#e demonstrated conclusion as t#ere are middle terms% since it is propositions containing t#ese middle terms t#at are t#e basic premisses on w#ic# t#e demonstration rests& and as t#ere are some indemonstrable basic trut#s asserting t#at (t#is is t#at) or t#at (t#is in#eres in t#at)% so t#ere are ot#ers denying t#at (t#is is t#at) or t#at (t#is in#eres in t#at) * in fact some basic trut#s will affirm and some will deny being. 6#en we are to prove a conclusion% we must take a primary essential predicate * suppose it 7 * of t#e sub'ect B% and t#en suppose A similarly predicable of 7. If we proceed in t#is manner% no proposition or attribute w#ic# falls beyond A is admitted in t#e proof1 t#e interval is constantly condensed until sub'ect and predicate become indivisible% i.e. one. 6e #ave our unit w#en t#e premiss becomes immediate% since t#e immediate premiss alone is a single premiss in t#e un$ualified sense of (single). And as in ot#er sp#eres t#e basic element is simple but not identical in all * in a system of weig#t it is t#e mina% in music t#e $uarter tone% and so on * so in syllogism t#e unit is an immediate premiss% and in t#e knowledge t#at demonstration gives it is an intuition. In syllogisms% t#en% w#ic# prove t#e in#erence of an attribute% not#ing falls outside t#e ma'or term. In t#e case of negative syllogisms on t#e ot#er #and% -1. in t#e first figure not#ing falls outside t#e ma'or term w#ose in#erence is in $uestion& e.g. to prove t#roug# a middle 7 t#at A does not in#ere in B t#e premisses re$uired are% all B is 7% no 7 is A. "#en if it #as to be proved t#at no 7 is A% a middle must be found between and 7& and t#is procedure will never vary.

5:

-5. If we #ave to s#ow t#at = is not 2 by means of t#e premisses% all 2 is 7& no =% or not all =% is 7& t#en t#e middle will never fall beyond =% and = is t#e sub'ect of w#ic# 2 is to be denied in t#e conclusion. -:. In t#e t#ird figure t#e middle will never fall beyond t#e limits of t#e sub'ect and t#e attribute denied of it. 5< 9ince demonstrations may be eit#er commensurately universal or particular% and eit#er affirmative or negative& t#e $uestion arises% w#ic# form is t#e better/ And t#e same $uestion may be put in regard to so called (direct) demonstration and reductio ad impossibile. Het us first e!amine t#e commensurately universal and t#e particular forms% and w#en we #ave cleared up t#is problem proceed to discuss (direct) demonstration and reductio ad impossibile. "#e following considerations mig#t lead some minds to prefer particular demonstration. -1. "#e superior demonstration is t#e demonstration w#ic# gives us greater knowledge -for t#is is t#e ideal of demonstration.% and we #ave greater knowledge of a particular individual w#en we know it in itself t#an w#en we know it t#roug# somet#ing else& e.g. we know 7oriscus t#e musician better w#en we know t#at 7oriscus is musical t#an w#en we know only t#at man is musical% and a like argument #olds in all ot#er cases. But commensurately universal demonstration% instead of proving t#at t#e sub'ect itself actually is !% proves only t#at somet#ing else is ! * e.g. in attempting to prove t#at isosceles is !% it proves not t#at isosceles but only t#at triangle is ! * w#ereas particular demonstration proves t#at t#e sub'ect itself is !. "#e demonstration% t#en% t#at a sub'ect% as suc#% possesses an attribute is superior. If t#is is so% and if t#e particular rat#er t#an t#e commensurately universal forms demonstrates% particular demonstration is superior. -5. "#e universal #as not a separate being over against groups of singulars. 2emonstration nevert#eless creates t#e opinion t#at its function is conditioned by somet#ing like t#is * some separate entity belonging to t#e real world& t#at% for instance% of triangle or of figure or number% over against particular triangles% figures% and numbers. But demonstration w#ic# touc#es t#e real and will not mislead is superior to t#at w#ic# moves among unrealities and is delusory. 0ow commensurately universal demonstration is of t#e latter kind1 if we engage in it we find ourselves reasoning after a fas#ion well illustrated by t#e argument t#at t#e proportionate is w#at answers to t#e definition of some entity w#ic# is neit#er line% number% solid% nor plane% but a proportionate apart from all t#ese. 9ince% t#en% suc# a proof is c#aracteristically commensurate and universal% and less touc#es reality t#an does particular demonstration% and creates a false opinion% it will follow t#at commensurate and universal is inferior to particular demonstration. 6e may retort t#us. -1. "#e first argument applies no more to commensurate and universal t#an to particular demonstration. If e$uality to two rig#t angles is attributable to its sub'ect not $ua isosceles but $ua triangle% #e w#o knows t#at isosceles possesses t#at attribute knows t#e sub'ect as $ua itself possessing t#e attribute% to a less degree t#an #e w#o knows t#at triangle #as t#at attribute. "o sum up t#e w#ole matter1 if a sub'ect is proved to possess $ua triangle an attribute w#ic# it does not in fact possess $ua triangle% t#at is not demonstration1 but if it does possess it $ua triangle t#e rule applies t#at t#e greater knowledge is #is w#o knows t#e sub'ect as possessing its attribute $ua t#at in virtue of w#ic# it actually does possess it. 9ince% t#en% triangle is t#e wider term% and t#ere is one identical definition of triangle * i.e. t#e term is not e$uivocal * and since e$uality to two rig#t angles belongs to all triangles% it is isosceles $ua triangle and not triangle $ua isosceles w#ic# #as its angles so related. It follows t#at #e w#o knows a conne!ion 5<

universally #as greater knowledge of it as it in fact is t#an #e w#o knows t#e particular& and t#e inference is t#at commensurate and universal is superior to particular demonstration. -5. If t#ere is a single identical definition i.e. if t#e commensurate universal is une$uivocal * t#en t#e universal will possess being not less but more t#an some of t#e particulars% inasmuc# as it is universals w#ic# comprise t#e imperis#able% particulars t#at tend to peris#. -:. Because t#e universal #as a single meaning% we are not t#erefore compelled to suppose t#at in t#ese e!amples it #as being as a substance apart from its particulars * any more t#an we need make a similar supposition in t#e ot#er cases of une$uivocal universal predication% vi,. w#ere t#e predicate signifies not substance but $uality% essential relatedness% or action. If suc# a supposition is entertained% t#e blame rests not wit# t#e demonstration but wit# t#e #earer. -<. 2emonstration is syllogism t#at proves t#e cause% i.e. t#e reasoned fact% and it is rat#er t#e commensurate universal t#an t#e particular w#ic# is causative -as may be s#own t#us1 t#at w#ic# possesses an attribute t#roug# its own essential nature is itself t#e cause of t#e in#erence% and t#e commensurate universal is primary& #ence t#e commensurate universal is t#e cause.. 7onse$uently commensurately universal demonstration is superior as more especially proving t#e cause% t#at is t#e reasoned fact. -@. 4ur searc# for t#e reason ceases% and we t#ink t#at we know% w#en t#e coming to be or e!istence of t#e fact before us is not due to t#e coming to be or e!istence of some ot#er fact% for t#e last step of a searc# t#us conducted is eo ipso t#e end and limit of t#e problem. "#us1 (6#y did #e come/) ("o get t#e money * w#erewit# to pay a debt * t#at #e mig#t t#ereby do w#at was rig#t.) 6#en in t#is regress we can no longer find an efficient or final cause% we regard t#e last step of it as t#e end of t#e coming * or being or coming to be * and we regard ourselves as t#en only #aving full knowledge of t#e reason w#y #e came. If% t#en% all causes and reasons are alike in t#is respect% and if t#is is t#e means to full knowledge in t#e case of final causes suc# as we #ave e!emplified% it follows t#at in t#e case of t#e ot#er causes also full knowledge is attained w#en an attribute no longer in#eres because of somet#ing else. "#us% w#en we learn t#at e!terior angles are e$ual to four rig#t angles because t#ey are t#e e!terior angles of an isosceles% t#ere still remains t#e $uestion (6#y #as isosceles t#is attribute/) and its answer (Because it is a triangle% and a triangle #as it because a triangle is a rectilinear figure.) If rectilinear figure possesses t#e property for no furt#er reason% at t#is point we #ave full knowledge * but at t#is point our knowledge #as become commensurately universal% and so we conclude t#at commensurately universal demonstration is superior. -A. "#e more demonstration becomes particular t#e more it sinks into an indeterminate manifold% w#ile universal demonstration tends to t#e simple and determinate. But ob'ects so far as t#ey are an indeterminate manifold are unintelligible% so far as t#ey are determinate% intelligible1 t#ey are t#erefore intelligible rat#er in so far as t#ey are universal t#an in so far as t#ey are particular. +rom t#is it follows t#at universals are more demonstrable1 but since relative and correlative increase concomitantly% of t#e more demonstrable t#ere will be fuller demonstration. 3ence t#e commensurate and universal form% being more truly demonstration% is t#e superior. -B. 2emonstration w#ic# teac#es two t#ings is preferable to demonstration w#ic# teac#es only one. 3e w#o possesses commensurately universal demonstration knows t#e particular as well% but #e w#o possesses particular demonstration does not know t#e universal. 9o t#at t#is is an additional reason for preferring commensurately universal demonstration. And t#ere is yet t#is furt#er argument1 -D. Proof becomes more and more proof of t#e commensurate universal as its middle term approac#es nearer to t#e basic trut#% and not#ing is so near as t#e immediate premiss w#ic# is itself t#e basic trut#. If% t#en% proof from t#e basic trut# is more accurate t#an proof not so 5@

derived% demonstration w#ic# depends more closely on it is more accurate t#an demonstration w#ic# is less closely dependent. But commensurately universal demonstration is c#aracteri,ed by t#is closer dependence% and is t#erefore superior. "#us% if A #ad to be proved to in#ere in 2% and t#e middles were B and 7% B being t#e #ig#er term would render t#e demonstration w#ic# it mediated t#e more universal. 9ome of t#ese arguments% #owever% are dialectical. "#e clearest indication of t#e precedence of commensurately universal demonstration is as follows1 if of two propositions% a prior and a posterior% we #ave a grasp of t#e prior% we #ave a kind of knowledge * a potential grasp * of t#e posterior as well. +or e!ample% if one knows t#at t#e angles of all triangles are e$ual to two rig#t angles% one knows in a sense * potentially * t#at t#e isosceles) angles also are e$ual to two rig#t angles% even if one does not know t#at t#e isosceles is a triangle& but to grasp t#is posterior proposition is by no means to know t#e commensurate universal eit#er potentially or actually. Moreover% commensurately universal demonstration is t#roug# and t#roug# intelligible& particular demonstration issues in sense perception. 5@ "#e preceding arguments constitute our defence of t#e superiority of commensurately universal to particular demonstration. "#at affirmative demonstration e!cels negative may be s#own as follows. -1. 6e may assume t#e superiority ceteris paribus of t#e demonstration w#ic# derives from fewer postulates or #ypot#eses * in s#ort from fewer premisses& for% given t#at all t#ese are e$ually well known% w#ere t#ey are fewer knowledge will be more speedily ac$uired% and t#at is a desideratum. "#e argument implied in our contention t#at demonstration from fewer assumptions is superior may be set out in universal form as follows. Assuming t#at in bot# cases alike t#e middle terms are known% and t#at middles w#ic# are prior are better known t#an suc# as are posterior% we may suppose two demonstrations of t#e in#erence of A in =% t#e one proving it t#roug# t#e middles B% 7 and 2% t#e ot#er t#roug# + and G. "#en A 2 is known to t#e same degree as A = -in t#e second proof.% but A 2 is better known t#an and prior to A = -in t#e first proof.& since A = is proved t#roug# A 2% and t#e ground is more certain t#an t#e conclusion. 3ence demonstration by fewer premisses is ceteris paribus superior. 0ow bot# affirmative and negative demonstration operate t#roug# t#ree terms and two premisses% but w#ereas t#e former assumes only t#at somet#ing is% t#e latter assumes bot# t#at somet#ing is and t#at somet#ing else is not% and t#us operating t#roug# more kinds of premiss is inferior. -5. It #as been proved t#at no conclusion follows if bot# premisses are negative% but t#at one must be negative% t#e ot#er affirmative. 9o we are compelled to lay down t#e following additional rule1 as t#e demonstration e!pands% t#e affirmative premisses must increase in number% but t#ere cannot be more t#an one negative premiss in eac# complete proof. "#us% suppose no B is A% and all 7 is B. "#en if bot# t#e premisses are to be again e!panded% a middle must be interposed. Het us interpose 2 between A and B% and = between B and 7. "#en clearly = is affirmatively related to B and 7% w#ile 2 is affirmatively related to B but negatively to A& for all B is 2% but t#ere must be no 2 w#ic# is A. "#us t#ere proves to be a single negative premiss% A 2. In t#e furt#er prosyllogisms too it is t#e same% because in t#e terms of an affirmative syllogism t#e middle is always related affirmatively to bot# e!tremes& in a negative syllogism it must be negatively related only to one of t#em% and so t#is negation comes to be a single negative premiss% t#e ot#er premisses being affirmative. If% t#en% t#at t#roug# w#ic# a trut# is proved is a better known and

5A

more certain trut#% and if t#e negative proposition is proved t#roug# t#e affirmative and not vice versa% affirmative demonstration% being prior and better known and more certain% will be superior. -:. "#e basic trut# of demonstrative syllogism is t#e universal immediate premiss% and t#e universal premiss asserts in affirmative demonstration and in negative denies1 and t#e affirmative proposition is prior to and better known t#an t#e negative -since affirmation e!plains denial and is prior to denial% 'ust as being is prior to not being.. It follows t#at t#e basic premiss of affirmative demonstration is superior to t#at of negative demonstration% and t#e demonstration w#ic# uses superior basic premisses is superior. -<. Affirmative demonstration is more of t#e nature of a basic form of proof% because it is a sine $ua non of negative demonstration. 5A 9ince affirmative demonstration is superior to negative% it is clearly superior also to reductio ad impossibile. 6e must first make certain w#at is t#e difference between negative demonstration and reductio ad impossibile. Het us suppose t#at no B is A% and t#at all 7 is B1 t#e conclusion necessarily follows t#at no 7 is A. If t#ese premisses are assumed% t#erefore% t#e negative demonstration t#at no 7 is A is direct. Reductio ad impossibile% on t#e ot#er #and% proceeds as follows. 9upposing we are to prove t#at does not in#ere in B% we #ave to assume t#at it does in#ere% and furt#er t#at B in#eres in 7% wit# t#e resulting inference t#at A in#eres in 7. "#is we #ave to suppose a known and admitted impossibility& and we t#en infer t#at A cannot in#ere in B. "#us if t#e in#erence of B in 7 is not $uestioned% A)s in#erence in B is impossible. "#e order of t#e terms is t#e same in bot# proofs1 t#ey differ according to w#ic# of t#e negative propositions is t#e better known% t#e one denying A of B or t#e one denying A of 7. 6#en t#e falsity of t#e conclusion is t#e better known% we use reductio ad impossible& w#en t#e ma'or premiss of t#e syllogism is t#e more obvious% we use direct demonstration. All t#e same t#e proposition denying A of B is% in t#e order of being% prior to t#at denying A of 7& for premisses are prior to t#e conclusion w#ic# follows from t#em% and (no 7 is A) is t#e conclusion% (no B is A) one of its premisses. +or t#e destructive result of reductio ad impossibile is not a proper conclusion% nor are its antecedents proper premisses. 4n t#e contrary1 t#e constituents of syllogism are premisses related to one anot#er as w#ole to part or part to w#ole% w#ereas t#e premisses A 7 and A B are not t#us related to one anot#er. 0ow t#e superior demonstration is t#at w#ic# proceeds from better known and prior premisses% and w#ile bot# t#ese forms depend for credence on t#e not being of somet#ing% yet t#e source of t#e one is prior to t#at of t#e ot#er. "#erefore negative demonstration will #ave an un$ualified superiority to reductio ad impossibile% and affirmative demonstration% being superior to negative% will conse$uently be superior also to reductio ad impossibile. 5B "#e science w#ic# is knowledge at once of t#e fact and of t#e reasoned fact% not of t#e fact by itself wit#out t#e reasoned fact% is t#e more e!act and t#e prior science. A science suc# as arit#metic% w#ic# is not a science of properties $ua in#ering in a substratum% is more e!act t#an and prior to a science like #armonics% w#ic# is a science of pr%operties in#ering in a substratum& and similarly a science like arit#metic% w#ic# is constituted of fewer basic elements% is more e!act t#an and prior to geometry% w#ic# re$uires additional elements. 6#at I

5B

mean by (additional elements) is t#is1 a unit is substance wit#out position% w#ile a point is substance wit# position& t#e latter contains an additional element. 5D A single science is one w#ose domain is a single genus% vi,. all t#e sub'ects constituted out of t#e primary entities of t#e genus * i.e. t#e parts of t#is total sub'ect * and t#eir essential properties. 4ne science differs from anot#er w#en t#eir basic trut#s #ave neit#er a common source nor are derived t#ose of t#e one science from t#ose t#e ot#er. "#is is verified w#en we reac# t#e indemonstrable premisses of a science% for t#ey must be wit#in one genus wit# its conclusions1 and t#is again is verified if t#e conclusions proved by means of t#em fall wit#in one genus * i.e. are #omogeneous. 5E 4ne can #ave several demonstrations of t#e same conne!ion not only by taking from t#e same series of predication middles w#ic# are ot#er t#an t#e immediately co#ering term e.g. by taking 7% 2% and + severally to prove A B * but also by taking a middle from anot#er series. "#us let A be c#ange% 2 alteration of a property% B feeling pleasure% and G rela!ation. 6e can t#en wit#out false#ood predicate 2 of B and A of 2% for #e w#o is pleased suffers alteration of a property% and t#at w#ic# alters a property c#anges. Again% we can predicate A of G wit#out false#ood% and G of B& for to feel pleasure is to rela!% and to rela! is to c#ange. 9o t#e conclusion can be drawn t#roug# middles w#ic# are different% i.e. not in t#e same series * yet not so t#at neit#er of t#ese middles is predicable of t#e ot#er% for t#ey must bot# be attributable to some one sub'ect. A furt#er point wort# investigating is #ow many ways of proving t#e same conclusion can be obtained by varying t#e figure. :F "#ere is no knowledge by demonstration of c#ance con'unctions& for c#ance con'unctions e!ist neit#er by necessity nor as general conne!ions but comprise w#at comes to be as somet#ing distinct from t#ese. 0ow demonstration is concerned only wit# one or ot#er of t#ese two& for all reasoning proceeds from necessary or general premisses% t#e conclusion being necessary if t#e premisses are necessary and general if t#e premisses are general. 7onse$uently% if c#ance con'unctions are neit#er general nor necessary% t#ey are not demonstrable. :1 9cientific knowledge is not possible t#roug# t#e act of perception. =ven if perception as a faculty is of (t#e suc#) and not merely of a (t#is somew#at)% yet one must at any rate actually perceive a (t#is somew#at)% and at a definite present place and time1 but t#at w#ic# is commensurately universal and true in all cases one cannot perceive% since it is not (t#is) and it is not (now)& if it were% it would not be commensurately universal * t#e term we apply to w#at is always and everyw#ere. 9eeing% t#erefore% t#at demonstrations are commensurately universal and universals imperceptible% we clearly cannot obtain scientific knowledge by t#e act of perception1 nay% it is obvious t#at even if it were possible to perceive t#at a triangle #as its angles e$ual to two rig#t angles% we s#ould still be looking for a demonstration * we s#ould not -as some say. possess 5D

knowledge of it& for perception must be of a particular% w#ereas scientific knowledge involves t#e recognition of t#e commensurate universal. 9o if we were on t#e moon% and saw t#e eart# s#utting out t#e sun)s lig#t% we s#ould not know t#e cause of t#e eclipse1 we s#ould perceive t#e present fact of t#e eclipse% but not t#e reasoned fact at all% since t#e act of perception is not of t#e commensurate universal. I do not% of course% deny t#at by watc#ing t#e fre$uent recurrence of t#is event we mig#t% after tracking t#e commensurate universal% possess a demonstration% for t#e commensurate universal is elicited from t#e several groups of singulars. "#e commensurate universal is precious because it makes clear t#e cause& so t#at in t#e case of facts like t#ese w#ic# #ave a cause ot#er t#an t#emselves universal knowledge is more precious t#an sense perceptions and t#an intuition. -As regards primary trut#s t#ere is of course a different account to be given.. 3ence it is clear t#at knowledge of t#ings demonstrable cannot be ac$uired by perception% unless t#e term perception is applied to t#e possession of scientific knowledge t#roug# demonstration. 0evert#eless certain points do arise wit# regard to conne!ions to be proved w#ic# are referred for t#eir e!planation to a failure in sense perception1 t#ere are cases w#en an act of vision would terminate our in$uiry% not because in seeing we s#ould be knowing% but because we s#ould #ave elicited t#e universal from seeing& if% for e!ample% we saw t#e pores in t#e glass and t#e lig#t passing t#roug#% t#e reason of t#e kindling would be clear to us because we s#ould at t#e same time see it in eac# instance and intuit t#at it must be so in all instances. :5 All syllogisms cannot #ave t#e same basic trut#s. "#is may be s#own first of all by t#e following dialectical considerations. -1. 9ome syllogisms are true and some false1 for t#oug# a true inference is possible from false premisses% yet t#is occurs once only * I mean if A for instance% is truly predicable of 7% but B% t#e middle% is false% bot# A B and B 7 being false& nevert#eless% if middles are taken to prove t#ese premisses% t#ey will be false because every conclusion w#ic# is a false#ood #as false premisses% w#ile true conclusions #ave true premisses% and false and true differ in kind. "#en again% -5. false#oods are not all derived from a single identical set of principles1 t#ere are false#oods w#ic# are t#e contraries of one anot#er and cannot coe!ist% e.g. ('ustice is in'ustice)% and ('ustice is cowardice)& (man is #orse)% and (man is o!)& (t#e e$ual is greater)% and (t#e e$ual is less.) +rom establis#ed principles we may argue t#e case as follows% confining ourselves t#erefore to true conclusions. 0ot even all t#ese are inferred from t#e same basic trut#s& many of t#em in fact #ave basic trut#s w#ic# differ generically and are not transferable& units% for instance% w#ic# are wit#out position% cannot take t#e place of points% w#ic# #ave position. "#e transferred terms could only fit in as middle terms or as ma'or or minor terms% or else #ave some of t#e ot#er terms between t#em% ot#ers outside t#em. 0or can any of t#e common a!ioms * suc#% I mean% as t#e law of e!cluded middle * serve as premisses for t#e proof of all conclusions. +or t#e kinds of being are different% and some attributes attac# to $uanta and some to $ualia only& and proof is ac#ieved by means of t#e common a!ioms taken in con'unction wit# t#ese several kinds and t#eir attributes. Again% it is not true t#at t#e basic trut#s are muc# fewer t#an t#e conclusions% for t#e basic trut#s are t#e premisses% and t#e premisses are formed by t#e apposition of a fres# e!treme term or t#e interposition of a fres# middle. Moreover% t#e number of conclusions is indefinite% t#oug# t#e number of middle terms is finite& and lastly some of t#e basic trut#s are necessary% ot#ers variable. Hooking at it in t#is way we see t#at% since t#e number of conclusions is indefinite% t#e basic trut#s cannot be identical or limited in number. If% on t#e ot#er #and% identity is used in anot#er 5E

sense% and it is said% e.g. (t#ese and no ot#er are t#e fundamental trut#s of geometry% t#ese t#e fundamentals of calculation% t#ese again of medicine)& would t#e statement mean anyt#ing e!cept t#at t#e sciences #ave basic trut#s/ "o call t#em identical because t#ey are self identical is absurd% since everyt#ing can be identified wit# everyt#ing in t#at sense of identity. 0or again can t#e contention t#at all conclusions #ave t#e same basic trut#s mean t#at from t#e mass of all possible premisses any conclusion may be drawn. "#at would be e!ceedingly naive% for it is not t#e case in t#e clearly evident mat#ematical sciences% nor is it possible in analysis% since it is t#e immediate premisses w#ic# are t#e basic trut#s% and a fres# conclusion is only formed by t#e addition of a new immediate premiss1 but if it be admitted t#at it is t#ese primary immediate premisses w#ic# are basic trut#s% eac# sub'ect genus will provide one basic trut#. If% #owever% it is not argued t#at from t#e mass of all possible premisses any conclusion may be proved% nor yet admitted t#at basic trut#s differ so as to be generically different for eac# science% it remains to consider t#e possibility t#at% w#ile t#e basic trut#s of all knowledge are wit#in one genus% special premisses are re$uired to prove special conclusions. But t#at t#is cannot be t#e case #as been s#own by our proof t#at t#e basic trut#s of t#ings generically different t#emselves differ generically. +or fundamental trut#s are of two kinds% t#ose w#ic# are premisses of demonstration and t#e sub'ect genus& and t#oug# t#e former are common% t#e latter * number% for instance% and magnitude * are peculiar. :: 9cientific knowledge and its ob'ect differ from opinion and t#e ob'ect of opinion in t#at scientific knowledge is commensurately universal and proceeds by necessary conne!ions% and t#at w#ic# is necessary cannot be ot#erwise. 9o t#oug# t#ere are t#ings w#ic# are true and real and yet can be ot#erwise% scientific knowledge clearly does not concern t#em1 if it did% t#ings w#ic# can be ot#erwise would be incapable of being ot#erwise. 0or are t#ey any concern of rational intuition * by rational intuition I mean an originative source of scientific knowledge * nor of indemonstrable knowledge% w#ic# is t#e grasping of t#e immediate premiss. 9ince t#en rational intuition% science% and opinion% and w#at is revealed by t#ese terms% are t#e only t#ings t#at can be (true)% it follows t#at it is opinion t#at is concerned wit# t#at w#ic# may be true or false% and can be ot#erwise1 opinion in fact is t#e grasp of a premiss w#ic# is immediate but not necessary. "#is view also fits t#e observed facts% for opinion is unstable% and so is t#e kind of being we #ave described as its ob'ect. Besides% w#en a man t#inks a trut# incapable of being ot#erwise #e always t#inks t#at #e knows it% never t#at #e opines it. 3e t#inks t#at #e opines w#en #e t#inks t#at a conne!ion% t#oug# actually so% may $uite easily be ot#erwise& for #e believes t#at suc# is t#e proper ob'ect of opinion% w#ile t#e necessary is t#e ob'ect of knowledge. In w#at sense% t#en% can t#e same t#ing be t#e ob'ect of bot# opinion and knowledge/ And if any one c#ooses to maintain t#at all t#at #e knows #e can also opine% w#y s#ould not opinion be knowledge/ +or #e t#at knows and #e t#at opines will follow t#e same train of t#oug#t t#roug# t#e same middle terms until t#e immediate premisses are reac#ed& because it is possible to opine not only t#e fact but also t#e reasoned fact% and t#e reason is t#e middle term& so t#at% since t#e former knows% #e t#at opines also #as knowledge. "#e trut# per#aps is t#at if a man grasp trut#s t#at cannot be ot#er t#an t#ey are% in t#e way in w#ic# #e grasps t#e definitions t#roug# w#ic# demonstrations take place% #e will #ave not opinion but knowledge1 if on t#e ot#er #and #e appre#ends t#ese attributes as in#ering in t#eir sub'ects% but not in virtue of t#e sub'ects) substance and essential nature possesses opinion and not genuine knowledge& and #is opinion% if obtained t#roug# immediate premisses% will be bot# :F

of t#e fact and of t#e reasoned fact& if not so obtained% of t#e fact alone. "#e ob'ect of opinion and knowledge is not $uite identical& it is only in a sense identical% 'ust as t#e ob'ect of true and false opinion is in a sense identical. "#e sense in w#ic# some maintain t#at true and false opinion can #ave t#e same ob'ect leads t#em to embrace many strange doctrines% particularly t#e doctrine t#at w#at a man opines falsely #e does not opine at all. "#ere are really many senses of (identical)% and in one sense t#e ob'ect of true and false opinion can be t#e same% in anot#er it cannot. "#us% to #ave a true opinion t#at t#e diagonal is commensurate wit# t#e side would be absurd1 but because t#e diagonal wit# w#ic# t#ey are bot# concerned is t#e same% t#e two opinions #ave ob'ects so far t#e same1 on t#e ot#er #and% as regards t#eir essential definable nature t#ese ob'ects differ. "#e identity of t#e ob'ects of knowledge and opinion is similar. Gnowledge is t#e appre#ension of% e.g. t#e attribute (animal) as incapable of being ot#erwise% opinion t#e appre#ension of (animal) as capable of being ot#erwise * e.g. t#e appre#ension t#at animal is an element in t#e essential nature of man is knowledge& t#e appre#ension of animal as predicable of man but not as an element in man)s essential nature is opinion1 man is t#e sub'ect in bot# 'udgements% but t#e mode of in#erence differs. "#is also s#ows t#at one cannot opine and know t#e same t#ing simultaneously& for t#en one would appre#end t#e same t#ing as bot# capable and incapable of being ot#erwise * an impossibility. Gnowledge and opinion of t#e same t#ing can co e!ist in two different people in t#e sense we #ave e!plained% but not simultaneously in t#e same person. "#at would involve a man)s simultaneously appre#ending% e.g. -1. t#at man is essentially animal * i.e. cannot be ot#er t#an animal * and -5. t#at man is not essentially animal% t#at is% we may assume% may be ot#er t#an animal. +urt#er consideration of modes of t#inking and t#eir distribution under t#e #eads of discursive t#oug#t% intuition% science% art% practical wisdom% and metap#ysical t#inking% belongs rat#er partly to natural science% partly to moral p#ilosop#y. :< Luick wit is a faculty of #itting upon t#e middle term instantaneously. It would be e!emplified by a man w#o saw t#at t#e moon #as #er brig#t side always turned towards t#e sun% and $uickly grasped t#e cause of t#is% namely t#at s#e borrows #er lig#t from #im& or observed somebody in conversation wit# a man of wealt# and divined t#at #e was borrowing money% or t#at t#e friends#ip of t#ese people sprang from a common enmity. In all t#ese instances #e #as seen t#e ma'or and minor terms and t#en grasped t#e causes% t#e middle terms. Het A represent (brig#t side turned sunward)% B (lig#ted from t#e sun)% 7 t#e moon. "#en B% (lig#ted from t#e sun) is predicable of 7% t#e moon% and A% (#aving #er brig#t side towards t#e source of #er lig#t)% is predicable of B. 9o A is predicable of 7 t#roug# B.

Book II 1 "#e kinds of $uestion we ask are as many as t#e kinds of t#ings w#ic# we know. "#ey are in fact four1 * -1. w#et#er t#e conne!ion of an attribute wit# a t#ing is a fact% -5. w#at is t#e reason of t#e conne!ion% -:. w#et#er a t#ing e!ists% -<. 6#at is t#e nature of t#e t#ing. "#us% w#en our :1

$uestion concerns a comple! of t#ing and attribute and we ask w#et#er t#e t#ing is t#us or ot#erwise $ualified * w#et#er% e.g. t#e sun suffers eclipse or not * t#en we are asking as to t#e fact of a conne!ion. "#at our in$uiry ceases wit# t#e discovery t#at t#e sun does suffer eclipse is an indication of t#is& and if we know from t#e start t#at t#e sun suffers eclipse% we do not in$uire w#et#er it does so or not. 4n t#e ot#er #and% w#en we know t#e fact we ask t#e reason& as% for e!ample% w#en we know t#at t#e sun is being eclipsed and t#at an eart#$uake is in progress% it is t#e reason of eclipse or eart#$uake into w#ic# we in$uire. 6#ere a comple! is concerned% t#en% t#ose are t#e two $uestions we ask& but for some ob'ects of in$uiry we #ave a different kind of $uestion to ask% suc# as w#et#er t#ere is or is not a centaur or a God. -By (is or is not) I mean (is or is not% wit#out furt#er $ualification)& as opposed to (is or is not >e.g.? w#ite).. 4n t#e ot#er #and% w#en we #ave ascertained t#e t#ing)s e!istence% we in$uire as to its nature% asking% for instance% (w#at% t#en% is God/) or (w#at is man/). 5 "#ese% t#en% are t#e four kinds of $uestion we ask% and it is in t#e answers to t#ese $uestions t#at our knowledge consists. 0ow w#en we ask w#et#er a conne!ion is a fact% or w#et#er a t#ing wit#out $ualification is% we are really asking w#et#er t#e conne!ion or t#e t#ing #as a (middle)& and w#en we #ave ascertained eit#er t#at t#e conne!ion is a fact or t#at t#e t#ing is * i.e. ascertained eit#er t#e partial or t#e un$ualified being of t#e t#ing and are proceeding to ask t#e reason of t#e conne!ion or t#e nature of t#e t#ing% t#en we are asking w#at t#e (middle) is. -By distinguis#ing t#e fact of t#e conne!ion and t#e e!istence of t#e t#ing as respectively t#e partial and t#e un$ualified being of t#e t#ing% I mean t#at if we ask (does t#e moon suffer eclipse/)% or (does t#e moon wa!/)% t#e $uestion concerns a part of t#e t#ing)s being& for w#at we are asking in suc# $uestions is w#et#er a t#ing is t#is or t#at% i.e. #as or #as not t#is or t#at attribute1 w#ereas% if we ask w#et#er t#e moon or nig#t e!ists% t#e $uestion concerns t#e un$ualified being of a t#ing.. 6e conclude t#at in all our in$uiries we are asking eit#er w#et#er t#ere is a (middle) or w#at t#e (middle) is1 for t#e (middle) #ere is precisely t#e cause% and it is t#e cause t#at we seek in all our in$uiries. "#us% (2oes t#e moon suffer eclipse/) means (Is t#ere or is t#ere not a cause producing eclipse of t#e moon/)% and w#en we #ave learnt t#at t#ere is% our ne!t $uestion is% (6#at% t#en% is t#is cause/ for t#e cause t#roug# w#ic# a t#ing is * not is t#is or t#at% i.e. #as t#is or t#at attribute% but wit#out $ualification is * and t#e cause t#roug# w#ic# it is * not is wit#out $ualification% but is t#is or t#at as #aving some essential attribute or some accident * are bot# alike t#e middle). By t#at w#ic# is wit#out $ualification I mean t#e sub'ect% e.g. moon or eart# or sun or triangle& by t#at w#ic# a sub'ect is -in t#e partial sense. I mean a property% e.g. eclipse% e$uality or ine$uality% interposition or non interposition. +or in all t#ese e!amples it is clear t#at t#e nature of t#e t#ing and t#e reason of t#e fact are identical1 t#e $uestion (6#at is eclipse/) and its answer ("#e privation of t#e moon)s lig#t by t#e interposition of t#e eart#) are identical wit# t#e $uestion (6#at is t#e reason of eclipse/) or (6#y does t#e moon suffer eclipse/) and t#e reply (Because of t#e failure of lig#t t#roug# t#e eart#)s s#utting it out). Again% for (6#at is a concord/ A commensurate numerical ratio of a #ig# and a low note)% we may substitute (6#at ratio makes a #ig# and a low note concordant/ "#eir relation according to a commensurate numerical ratio.) (Are t#e #ig# and t#e low note concordant/) is e$uivalent to (Is t#eir ratio commensurate/)& and w#en we find t#at it is commensurate% we ask (6#at% t#en% is t#eir ratio/).

:5

7ases in w#ic# t#e (middle) is sensible s#ow t#at t#e ob'ect of our in$uiry is always t#e (middle)1 we in$uire% because we #ave not perceived it% w#et#er t#ere is or is not a (middle) causing% e.g. an eclipse. 4n t#e ot#er #and% if we were on t#e moon we s#ould not be in$uiring eit#er as to t#e fact or t#e reason% but bot# fact and reason would be obvious simultaneously. +or t#e act of perception would #ave enabled us to know t#e universal too& since% t#e present fact of an eclipse being evident% perception would t#en at t#e same time give us t#e present fact of t#e eart#)s screening t#e sun)s lig#t% and from t#is would arise t#e universal. "#us% as we maintain% to know a t#ing)s nature is to know t#e reason w#y it is& and t#is is e$ually true of t#ings in so far as t#ey are said wit#out $ualification to #e as opposed to being possessed of some attribute% and in so far as t#ey are said to be possessed of some attribute suc# as e$ual to rig#t angles% or greater or less. : It is clear% t#en% t#at all $uestions are a searc# for a (middle). Het us now state #ow essential nature is revealed and in w#at way it can be reduced to demonstration& w#at definition is% and w#at t#ings are definable. And let us first discuss certain difficulties w#ic# t#ese $uestions raise% beginning w#at we #ave to say wit# a point most intimately connected wit# our immediately preceding remarks% namely t#e doubt t#at mig#t be felt as to w#et#er or not it is possible to know t#e same t#ing in t#e same relation% bot# by definition and by demonstration. It mig#t% I mean% be urged t#at definition is #eld to concern essential nature and is in every case universal and affirmative& w#ereas% on t#e ot#er #and% some conclusions are negative and some are not universal& e.g. all in t#e second figure are negative% none in t#e t#ird are universal. And again% not even all affirmative conclusions in t#e first figure are definable% e.g. (every triangle #as its angles e$ual to two rig#t angles). An argument proving t#is difference between demonstration and definition is t#at to #ave scientific knowledge of t#e demonstrable is identical wit# possessing a demonstration of it1 #ence if demonstration of suc# conclusions as t#ese is possible% t#ere clearly cannot also be definition of t#em. If t#ere could% one mig#t know suc# a conclusion also in virtue of its definition wit#out possessing t#e demonstration of it& for t#ere is not#ing to stop our #aving t#e one wit#out t#e ot#er. Induction too will sufficiently convince us of t#is difference& for never yet by defining anyt#ing * essential attribute or accident * did we get knowledge of it. Again% if to define is to ac$uire knowledge of a substance% at any rate suc# attributes are not substances. It is evident% t#en% t#at not everyt#ing demonstrable can be defined. 6#at t#en/ 7an everyt#ing definable be demonstrated% or not/ "#ere is one of our previous arguments w#ic# covers t#is too. 4f a single t#ing $ua single t#ere is a single scientific knowledge. 3ence% since to know t#e demonstrable scientifically is to possess t#e demonstration of it% an impossible conse$uence will follow1 * possession of its definition wit#out its demonstration will give knowledge of t#e demonstrable. Moreover% t#e basic premisses of demonstrations are definitions% and it #as already been s#own t#at t#ese will be found indemonstrable& eit#er t#e basic premisses will be demonstrable and will depend on prior premisses% and t#e regress will be endless& or t#e primary trut#s will be indemonstrable definitions. But if t#e definable and t#e demonstrable are not w#olly t#e same% may t#ey yet be partially t#e same/ 4r is t#at impossible% because t#ere can be no demonstration of t#e definable/ "#ere can be none% because definition is of t#e essential nature or being of somet#ing% and all demonstrations evidently posit and assume t#e essential nature * mat#ematical demonstrations% ::

for e!ample% t#e nature of unity and t#e odd% and all t#e ot#er sciences likewise. Moreover% every demonstration proves a predicate of a sub'ect as attac#ing or as not attac#ing to it% but in definition one t#ing is not predicated of anot#er& we do not% e.g. predicate animal of biped nor biped of animal% nor yet figure of plane * plane not being figure nor figure plane. Again% to prove essential nature is not t#e same as to prove t#e fact of a conne!ion. 0ow definition reveals essential nature% demonstration reveals t#at a given attribute attac#es or does not attac# to a given sub'ect& but different t#ings re$uire different demonstrations * unless t#e one demonstration is related to t#e ot#er as part to w#ole. I add t#is because if all triangles #ave been proved to possess angles e$ual to two rig#t angles% t#en t#is attribute #as been proved to attac# to isosceles& for isosceles is a part of w#ic# all triangles constitute t#e w#ole. But in t#e case before us t#e fact and t#e essential nature are not so related to one anot#er% since t#e one is not a part of t#e ot#er. 9o it emerges t#at not all t#e definable is demonstrable nor all t#e demonstrable definable& and we may draw t#e general conclusion t#at t#ere is no identical ob'ect of w#ic# it is possible to possess bot# a definition and a demonstration. It follows obviously t#at definition and demonstration are neit#er identical nor contained eit#er wit#in t#e ot#er1 if t#ey were% t#eir ob'ects would be related eit#er as identical or as w#ole and part. < 9o muc#% t#en% for t#e first stage of our problem. "#e ne!t step is to raise t#e $uestion w#et#er syllogism * i.e. demonstration * of t#e definable nature is possible or% as our recent argument assumed% impossible. 6e mig#t argue it impossible on t#e following grounds1 * -a. syllogism proves an attribute of a sub'ect t#roug# t#e middle term& on t#e ot#er #and -b. its definable nature is bot# (peculiar) to a sub'ect and predicated of it as belonging to its essence. But in t#at case -1. t#e sub'ect% its definition% and t#e middle term connecting t#em must be reciprocally predicable of one anot#er& for if A is to 7% obviously A is (peculiar) to B and B to 7 * in fact all t#ree terms are (peculiar) to one anot#er1 and furt#er -5. if A in#eres in t#e essence of all B and B is predicated universally of all 7 as belonging to 7)s essence% A also must be predicated of 7 as belonging to its essence. If one does not take t#is relation as t#us duplicated * if% t#at is% A is predicated as being of t#e essence of B% but B is not of t#e essence of t#e sub'ects of w#ic# it is predicated * A will not necessarily be predicated of 7 as belonging to its essence. 9o bot# premisses will predicate essence% and conse$uently B also will be predicated of 7 as its essence. 9ince% t#erefore% bot# premisses do predicate essence * i.e. definable form * 7)s definable form will appear in t#e middle term before t#e conclusion is drawn. 6e may generali,e by supposing t#at it is possible to prove t#e essential nature of man. Het 7 be man% A man)s essential nature * two footed animal% or aug#t else it may be. "#en% if we are to syllogi,e% A must be predicated of all B. But t#is premiss will be mediated by a fres# definition% w#ic# conse$uently will also be t#e essential nature of man. "#erefore t#e argument assumes w#at it #as to prove% since B too is t#e essential nature of man. It is% #owever% t#e case in w#ic# t#ere are only t#e two premisses * i.e. in w#ic# t#e premisses are primary and immediate * w#ic# we oug#t to investigate% because it best illustrates t#e point under discussion. "#us t#ey w#o prove t#e essential nature of soul or man or anyt#ing else t#roug# reciprocating terms beg t#e $uestion. It would be begging t#e $uestion% for e!ample% to contend t#at t#e soul is t#at w#ic# causes its own life% and t#at w#at causes its own life is a self moving number& for one would #ave to postulate t#at t#e soul is a self moving number in t#e sense of being identical wit# it. +or if A is predicable as a mere conse$uent of B and B of 7% A will not on t#at account be t#e :<

definable form of 71 A will merely be w#at it was true to say of 7. =ven if A is predicated of all B inasmuc# as B is identical wit# a species of A% still it will not follow1 being an animal is predicated of being a man * since it is true t#at in all instances to be #uman is to be animal% 'ust as it is also true t#at every man is an animal * but not as identical wit# being man. 6e conclude% t#en% t#at unless one takes bot# t#e premisses as predicating essence% one cannot infer t#at A is t#e definable form and essence of 71 but if one does so take t#em% in assuming B one will #ave assumed% before drawing t#e conclusion% w#at t#e definable form of 7 is& so t#at t#ere #as been no inference% for one #as begged t#e $uestion. @ 0or% as was said in my formal logic% is t#e met#od of division a process of inference at all% since at no point does t#e c#aracteri,ation of t#e sub'ect follow necessarily from t#e premising of certain ot#er facts1 division demonstrates as little as does induction. +or in a genuine demonstration t#e conclusion must not be put as a $uestion nor depend on a concession% but must follow necessarily from its premisses% even if t#e respondent deny it. "#e definer asks (Is man animal or inanimate/) and t#en assumes * #e #as not inferred * t#at man is animal. 0e!t% w#en presented wit# an e!#austive division of animal into terrestrial and a$uatic% #e assumes t#at man is terrestrial. Moreover% t#at man is t#e complete formula% terrestrial animal% does not follow necessarily from t#e premisses1 t#is too is an assumption% and e$ually an assumption w#et#er t#e division comprises many differentiae or few. -Indeed as t#is met#od of division is used by t#ose w#o proceed by it% even trut#s t#at can be inferred actually fail to appear as suc#.. +or w#y s#ould not t#e w#ole of t#is formula be true of man% and yet not e!#ibit #is essential nature or definable form/ Again% w#at guarantee is t#ere against an unessential addition% or against t#e omission of t#e final or of an intermediate determinant of t#e substantial being/ "#e c#ampion of division mig#t #ere urge t#at t#oug# t#ese lapses do occur% yet we can solve t#at difficulty if all t#e attributes we assume are constituents of t#e definable form% and if% postulating t#e genus% we produce by division t#e re$uisite uninterrupted se$uence of terms% and omit not#ing& and t#at indeed we cannot fail to fulfil t#ese conditions if w#at is to be divided falls w#ole into t#e division at eac# stage% and none of it is omitted& and t#at t#is * t#e dividendum * must wit#out furt#er $uestion be -ultimately. incapable of fres# specific division. 0evert#eless% we reply% division does not involve inference& if it gives knowledge% it gives it in anot#er way. 0or is t#ere any absurdity in t#is1 induction% per#aps% is not demonstration any more t#an is division% et it does make evident some trut#. Cet to state a definition reac#ed by division is not to state a conclusion1 as% w#en conclusions are drawn wit#out t#eir appropriate middles% t#e alleged necessity by w#ic# t#e inference follows from t#e premisses is open to a $uestion as to t#e reason for it% so definitions reac#ed by division invite t#e same $uestion. "#us to t#e $uestion (6#at is t#e essential nature of man/) t#e divider replies (Animal% mortal% footed% biped% wingless)& and w#en at eac# step #e is asked (6#y/)% #e will say% and% as #e t#inks% proves by division% t#at all animal is mortal or immortal1 but suc# a formula taken in its entirety is not definition& so t#at even if division does demonstrate its formula% definition at any rate does not turn out to be a conclusion of inference. A 7an we nevert#eless actually demonstrate w#at a t#ing essentially and substantially is% but #ypot#etically% i.e. by premising -1. t#at its definable form is constituted by t#e (peculiar) :@

attributes of its essential nature& -5. t#at suc# and suc# are t#e only attributes of its essential nature% and t#at t#e complete synt#esis of t#em is peculiar to t#e t#ing& and t#us * since in t#is synt#esis consists t#e being of t#e t#ing * obtaining our conclusion/ 4r is t#e trut# t#at% since proof must be t#roug# t#e middle term% t#e definable form is once more assumed in t#is minor premiss too/ +urt#er% 'ust as in syllogi,ing we do not premise w#at syllogistic inference is -since t#e premisses from w#ic# we conclude must be related as w#ole and part.% so t#e definable form must not fall wit#in t#e syllogism but remain outside t#e premisses posited. It is only against a doubt as to its #aving been a syllogistic inference at all t#at we #ave to defend our argument as conforming to t#e definition of syllogism. It is only w#en some one doubts w#et#er t#e conclusion proved is t#e definable form t#at we #ave to defend it as conforming to t#e definition of definable form w#ic# we assumed. 3ence syllogistic inference must be possible even wit#out t#e e!press statement of w#at syllogism is or w#at definable form is. "#e following type of #ypot#etical proof also begs t#e $uestion. If evil is definable as t#e divisible% and t#e definition of a t#ing)s contrary * if it #as one t#e contrary of t#e t#ing)s definition& t#en% if good is t#e contrary of evil and t#e indivisible of t#e divisible% we conclude t#at to be good is essentially to be indivisible. "#e $uestion is begged because definable form is assumed as a premiss% and as a premiss w#ic# is to prove definable form. (But not t#e same definable form)% you may ob'ect. "#at I admit% for in demonstrations also we premise t#at (t#is) is predicable of (t#at)& but in t#is premiss t#e term we assert of t#e minor is neit#er t#e ma'or itself nor a term identical in definition% or convertible% wit# t#e ma'or. Again% bot# proof by division and t#e syllogism 'ust described are open to t#e $uestion w#y man s#ould be animal biped terrestrial and not merely animal and terrestrial% since w#at t#ey premise does not ensure t#at t#e predicates s#all constitute a genuine unity and not merely belong to a single sub'ect as do musical and grammatical w#en predicated of t#e same man. B 3ow t#en by definition s#all we prove substance or essential nature/ 6e cannot s#ow it as a fres# fact necessarily following from t#e assumption of premisses admitted to be facts * t#e met#od of demonstration1 we may not proceed as by induction to establis# a universal on t#e evidence of groups of particulars w#ic# offer no e!ception% because induction proves not w#at t#e essential nature of a t#ing is but t#at it #as or #as not some attribute. "#erefore% since presumably one cannot prove essential nature by an appeal to sense perception or by pointing wit# t#e finger% w#at ot#er met#od remains/ "o put it anot#er way1 #ow s#all we by definition prove essential nature/ 3e w#o knows w#at #uman * or any ot#er * nature is% must know also t#at man e!ists& for no one knows t#e nature of w#at does not e!ist * one can know t#e meaning of t#e p#rase or name (goat stag) but not w#at t#e essential nature of a goat stag is. But furt#er% if definition can prove w#at is t#e essential nature of a t#ing% can it also prove t#at it e!ists/ And #ow will it prove t#em bot# by t#e same process% since definition e!#ibits one single t#ing and demonstration anot#er single t#ing% and w#at #uman nature is and t#e fact t#at man e!ists are not t#e same t#ing/ "#en too we #old t#at it is by demonstration t#at t#e being of everyt#ing must be proved * unless indeed to be were its essence& and% since being is not a genus% it is not t#e essence of anyt#ing. 3ence t#e being of anyt#ing as fact is matter for demonstration& and t#is is t#e actual procedure of t#e sciences% for t#e geometer assumes t#e meaning of t#e word triangle% but t#at it is possessed of some attribute #e proves. 6#at is it% t#en% t#at we s#all prove in defining essential nature/ "riangle/ In t#at case :A

a man will know by definition w#at a t#ing)s nature is wit#out knowing w#et#er it e!ists. But t#at is impossible. Moreover it is clear% if we consider t#e met#ods of defining actually in use% t#at definition does not prove t#at t#e t#ing defined e!ists1 since even if t#ere does actually e!ist somet#ing w#ic# is e$uidistant from a centre% yet w#y s#ould t#e t#ing named in t#e definition e!ist/ 6#y% in ot#er words% s#ould t#is be t#e formula defining circle/ 4ne mig#t e$ually well call it t#e definition of mountain copper. +or definitions do not carry a furt#er guarantee t#at t#e t#ing defined can e!ist or t#at it is w#at t#ey claim to define1 one can always ask w#y. 9ince% t#erefore% to define is to prove eit#er a t#ing)s essential nature or t#e meaning of its name% we may conclude t#at definition% if it in no sense proves essential nature% is a set of words signifying precisely w#at a name signifies. But t#at were a strange conse$uence& for -1. bot# w#at is not substance and w#at does not e!ist at all would be definable% since even non e!istents can be signified by a name1 -5. all sets of words or sentences would be definitions% since any kind of sentence could be given a name& so t#at we s#ould all be talking in definitions% and even t#e Iliad would be a definition1 -:. no demonstration can prove t#at any particular name means any particular t#ing1 neit#er% t#erefore% do definitions% in addition to revealing t#e meaning of a name% also reveal t#at t#e name #as t#is meaning. It appears t#en from t#ese considerations t#at neit#er definition and syllogism nor t#eir ob'ects are identical% and furt#er t#at definition neit#er demonstrates nor proves anyt#ing% and t#at knowledge of essential nature is not to be obtained eit#er by definition or by demonstration. D 6e must now start afres# and consider w#ic# of t#ese conclusions are sound and w#ic# are not% and w#at is t#e nature of definition% and w#et#er essential nature is in any sense demonstrable and definable or in none. 0ow to know its essential nature is% as we said% t#e same as to know t#e cause of a t#ing)s e!istence% and t#e proof of t#is depends on t#e fact t#at a t#ing must #ave a cause. Moreover% t#is cause is eit#er identical wit# t#e essential nature of t#e t#ing or distinct from it& and if its cause is distinct from it% t#e essential nature of t#e t#ing is eit#er demonstrable or indemonstrable. 7onse$uently% if t#e cause is distinct from t#e t#ing)s essential nature and demonstration is possible% t#e cause must be t#e middle term% and% t#e conclusion proved being universal and affirmative% t#e proof is in t#e first figure. 9o t#e met#od 'ust e!amined of proving it t#roug# anot#er essential nature would be one way of proving essential nature% because a conclusion containing essential nature must be inferred t#roug# a middle w#ic# is an essential nature 'ust as a (peculiar) property must be inferred t#roug# a middle w#ic# is a (peculiar) property& so t#at of t#e two definable natures of a single t#ing t#is met#od will prove one and not t#e ot#er. 0ow it was said before t#at t#is met#od could not amount to demonstration of essential nature * it is actually a dialectical proof of it * so let us begin again and e!plain by w#at met#od it can be demonstrated. 6#en we are aware of a fact we seek its reason% and t#oug# sometimes t#e fact and t#e reason dawn on us simultaneously% yet we cannot appre#end t#e reason a moment sooner t#an t#e fact& and clearly in 'ust t#e same way we cannot appre#end a t#ing)s definable form wit#out appre#ending t#at it e!ists% since w#ile we are ignorant w#et#er it e!ists we cannot know its essential nature. Moreover we are aware w#et#er a t#ing e!ists or not sometimes t#roug# appre#ending an element in its c#aracter% and sometimes accidentally% as% for e!ample% w#en we are aware of t#under as a noise in t#e clouds% of eclipse as a privation of lig#t% or of man as some species of animal% or of t#e soul as a self moving t#ing. As often as we #ave accidental :B

knowledge t#at t#e t#ing e!ists% we must be in a w#olly negative state as regards awareness of its essential nature& for we #ave not got genuine knowledge even of its e!istence% and to searc# for a t#ing)s essential nature w#en we are unaware t#at it e!ists is to searc# for not#ing. 4n t#e ot#er #and% w#enever we appre#end an element in t#e t#ing)s c#aracter t#ere is less difficulty. "#us it follows t#at t#e degree of our knowledge of a t#ing)s essential nature is determined by t#e sense in w#ic# we are aware t#at it e!ists. Het us t#en take t#e following as our first instance of being aware of an element in t#e essential nature. Het A be eclipse% 7 t#e moon% B t#e eart#)s acting as a screen. 0ow to ask w#et#er t#e moon is eclipsed or not is to ask w#et#er or not B #as occurred. But t#at is precisely t#e same as asking w#et#er A #as a defining condition& and if t#is condition actually e!ists% we assert t#at A also actually e!ists. 4r again we may ask w#ic# side of a contradiction t#e defining condition necessitates1 does it make t#e angles of a triangle e$ual or not e$ual to two rig#t angles/ 6#en we #ave found t#e answer% if t#e premisses are immediate% we know fact and reason toget#er& if t#ey are not immediate% we know t#e fact wit#out t#e reason% as in t#e following e!ample1 let 7 be t#e moon% A eclipse% B t#e fact t#at t#e moon fails to produce s#adows t#oug# s#e is full and t#oug# no visible body intervenes between us and #er. "#en if B% failure to produce s#adows in spite of t#e absence of an intervening body% is attributable A to 7% and eclipse% is attributable to B% it is clear t#at t#e moon is eclipsed% but t#e reason w#y is not yet clear% and we know t#at eclipse e!ists% but we do not know w#at its essential nature is. But w#en it is clear t#at A is attributable to 7 and we proceed to ask t#e reason of t#is fact% we are in$uiring w#at is t#e nature of B1 is it t#e eart#)s acting as a screen% or t#e moon)s rotation or #er e!tinction/ But B is t#e definition of t#e ot#er term% vi,. in t#ese e!amples% of t#e ma'or term A& for eclipse is constituted by t#e eart# acting as a screen. "#us% -1. (6#at is t#under/) ("#e $uenc#ing of fire in cloud)% and -5. (6#y does it t#under/) (Because fire is $uenc#ed in t#e cloud)% are e$uivalent. Het 7 be cloud% A t#under% B t#e $uenc#ing of fire. "#en B is attributable to 7% cloud% since fire is $uenc#ed in it& and A% noise% is attributable to B& and B is assuredly t#e definition of t#e ma'or term A. If t#ere be a furt#er mediating cause of B% it will be one of t#e remaining partial definitions of A. 6e #ave stated t#en #ow essential nature is discovered and becomes known% and we see t#at% w#ile t#ere is no syllogism * i.e. no demonstrative syllogism * of essential nature% yet it is t#roug# syllogism% vi,. demonstrative syllogism% t#at essential nature is e!#ibited. 9o we conclude t#at neit#er can t#e essential nature of anyt#ing w#ic# #as a cause distinct from itself be known wit#out demonstration% nor can it be demonstrated& and t#is is w#at we contended in our preliminary discussions. E 0ow w#ile some t#ings #ave a cause distinct from t#emselves% ot#ers #ave not. 3ence it is evident t#at t#ere are essential natures w#ic# are immediate% t#at is are basic premisses& and of t#ese not only t#at t#ey are but also w#at t#ey are must be assumed or revealed in some ot#er way. "#is too is t#e actual procedure of t#e arit#metician% w#o assumes bot# t#e nature and t#e e!istence of unit. 4n t#e ot#er #and% it is possible -in t#e manner e!plained. to e!#ibit t#roug# demonstration t#e essential nature of t#ings w#ic# #ave a (middle)% i.e. a cause of t#eir substantial being ot#er t#an t#at being itself& but we do not t#ereby demonstrate it.

:D

1F 9ince definition is said to be t#e statement of a t#ing)s nature% obviously one kind of definition will be a statement of t#e meaning of t#e name% or of an e$uivalent nominal formula. A definition in t#is sense tells you% e.g. t#e meaning of t#e p#rase (triangular c#aracter). 6#en we are aware t#at triangle e!ists% we in$uire t#e reason w#y it e!ists. But it is difficult t#us to learn t#e definition of t#ings t#e e!istence of w#ic# we do not genuinely know * t#e cause of t#is difficulty being% as we said before% t#at we only know accidentally w#et#er or not t#e t#ing e!ists. Moreover% a statement may be a unity in eit#er of two ways% by con'unction% like t#e Iliad% or because it e!#ibits a single predicate as in#ering not accidentally in a single sub'ect. "#at t#en is one way of defining definition. Anot#er kind of definition is a formula e!#ibiting t#e cause of a t#ing)s e!istence. "#us t#e former signifies wit#out proving% but t#e latter will clearly be a $uasi demonstration of essential nature% differing from demonstration in t#e arrangement of its terms. +or t#ere is a difference between stating w#y it t#unders% and stating w#at is t#e essential nature of t#under& since t#e first statement will be (Because fire is $uenc#ed in t#e clouds)% w#ile t#e statement of w#at t#e nature of t#under is will be ("#e noise of fire being $uenc#ed in t#e clouds). "#us t#e same statement takes a different form1 in one form it is continuous demonstration% in t#e ot#er definition. Again% t#under can be defined as noise in t#e clouds% w#ic# is t#e conclusion of t#e demonstration embodying essential nature. 4n t#e ot#er #and t#e definition of immediates is an indemonstrable positing of essential nature. 6e conclude t#en t#at definition is -a. an indemonstrable statement of essential nature% or -b. a syllogism of essential nature differing from demonstration in grammatical form% or -c. t#e conclusion of a demonstration giving essential nature. 4ur discussion #as t#erefore made plain -1. in w#at sense and of w#at t#ings t#e essential nature is demonstrable% and in w#at sense and of w#at t#ings it is not& -5. w#at are t#e various meanings of t#e term definition% and in w#at sense and of w#at t#ings it proves t#e essential nature% and in w#at sense and of w#at t#ings it does not& -:. w#at is t#e relation of definition to demonstration% and #ow far t#e same t#ing is bot# definable and demonstrable and #ow far it is not. 11 6e t#ink we #ave scientific knowledge w#en we know t#e cause% and t#ere are four causes1 -1. t#e definable form% -5. an antecedent w#ic# necessitates a conse$uent% -:. t#e efficient cause% -<. t#e final cause. 3ence eac# of t#ese can be t#e middle term of a proof% for -a. t#oug# t#e inference from antecedent to necessary conse$uent does not #old if only one premiss is assumed * two is t#e minimum * still w#en t#ere are two it #olds on condition t#at t#ey #ave a single common middle term. 9o it is from t#e assumption of t#is single middle term t#at t#e conclusion follows necessarily. "#e following e!ample will also s#ow t#is. 6#y is t#e angle in a semicircle a rig#t angle/ * or from w#at assumption does it follow t#at it is a rig#t angle/ "#us% let A be rig#t angle% B t#e #alf of two rig#t angles% 7 t#e angle in a semicircle. "#en B is t#e cause in virtue of w#ic# A% rig#t angle% is attributable to 7% t#e angle in a semicircle% since B;A and t#e ot#er% vi,. 7%;B% for 7 is #alf of two rig#t angles. "#erefore it is t#e assumption of B% t#e #alf of two rig#t angles% from w#ic# it follows t#at A is attributable to 7% i.e. t#at t#e angle in a semicircle is a rig#t angle. Moreover% B is identical wit# -b. t#e defining form of A% since it is w#at A)s definition signifies. Moreover% t#e formal cause #as already been s#own to be t#e middle. -c. (6#y did t#e At#enians become involved in t#e Persian war/) means (6#at cause originated t#e waging of war against t#e At#enians/) and t#e answer is% (Because t#ey raided 9ardis wit# t#e =retrians)% since t#is originated t#e war. Het A be war% B unprovoked raiding% 7 t#e At#enians. :E

"#en B% unprovoked raiding% is true of 7% t#e At#enians% and A is true of B% since men make war on t#e un'ust aggressor. 9o A% #aving war waged upon t#em% is true of B% t#e initial aggressors% and B is true of 7% t#e At#enians% w#o were t#e aggressors. 3ence #ere too t#e cause * in t#is case t#e efficient cause * is t#e middle term. -d. "#is is no less true w#ere t#e cause is t#e final cause. =.g. w#y does one take a walk after supper/ +or t#e sake of one)s #ealt#. 6#y does a #ouse e!ist/ +or t#e preservation of one)s goods. "#e end in view is in t#e one case #ealt#% in t#e ot#er preservation. "o ask t#e reason w#y one must walk after supper is precisely to ask to w#at end one must do it. Het 7 be walking after supper% B t#e non regurgitation of food% A #ealt#. "#en let walking after supper possess t#e property of preventing food from rising to t#e orifice of t#e stomac#% and let t#is condition be #ealt#y& since it seems t#at B% t#e non regurgitation of food% is attributable to 7% taking a walk% and t#at A% #ealt#% is attributable to B. 6#at% t#en% is t#e cause t#roug# w#ic# A% t#e final cause% in#eres in 7/ It is B% t#e non regurgitation of food& but B is a kind of definition of A% for A will be e!plained by it. 6#y is B t#e cause of A)s belonging to 7/ Because to be in a condition suc# as B is to be in #ealt#. "#e definitions must be transposed% and t#en t#e detail will become clearer. Incidentally% #ere t#e order of coming to be is t#e reverse of w#at it is in proof t#roug# t#e efficient cause1 in t#e efficient order t#e middle term must come to be first% w#ereas in t#e teleological order t#e minor% 7% must first take place% and t#e end in view comes last in time. "#e same t#ing may e!ist for an end and be necessitated as well. +or e!ample% lig#t s#ines t#roug# a lantern -1. because t#at w#ic# consists of relatively small particles necessarily passes t#roug# pores larger t#an t#ose particles * assuming t#at lig#t does issue by penetration * and -5. for an end% namely to save us from stumbling. If t#en% a t#ing can e!ist t#roug# two causes% can it come to be t#roug# two causes * as for instance if t#under be a #iss and a roar necessarily produced by t#e $uenc#ing of fire% and also designed% as t#e Pyt#agoreans say% for a t#reat to terrify t#ose t#at lie in "artarus/ Indeed% t#ere are very many suc# cases% mostly among t#e processes and products of t#e natural world& for nature% in different senses of t#e term (nature)% produces now for an end% now by necessity. 0ecessity too is of two kinds. It may work in accordance wit# a t#ing)s natural tendency% or by constraint and in opposition to it& as% for instance% by necessity a stone is borne bot# upwards and downwards% but not by t#e same necessity. 4f t#e products of man)s intelligence some are never due to c#ance or necessity but always to an end% as for e!ample a #ouse or a statue& ot#ers% suc# as #ealt# or safety% may result from c#ance as well. It is mostly in cases w#ere t#e issue is indeterminate -t#oug# only w#ere t#e production does not originate in c#ance% and t#e end is conse$uently good.% t#at a result is due to an end% and t#is is true alike in nature or in art. By c#ance% on t#e ot#er #and% not#ing comes to be for an end. 15 "#e effect may be still coming to be% or its occurrence may be past or future% yet t#e cause will be t#e same as w#en it is actually e!istent * for it is t#e middle w#ic# is t#e cause * e!cept t#at if t#e effect actually e!ists t#e cause is actually e!istent% if it is coming to be so is t#e cause% if its occurrence is past t#e cause is past% if future t#e cause is future. +or e!ample% t#e moon was eclipsed because t#e eart# intervened% is becoming eclipsed because t#e eart# is in process of intervening% will be eclipsed because t#e eart# will intervene% is eclipsed because t#e eart# intervenes.

<F

"o take a second e!ample1 assuming t#at t#e definition of ice is solidified water% let 7 be water% A solidified% B t#e middle% w#ic# is t#e cause% namely total failure of #eat. "#en B is attributed to 7% and A% solidification% to B1 ice w#en B is occurring% #as formed w#en B #as occurred% and will form w#en B s#all occur. "#is sort of cause% t#en% and its effect come to be simultaneously w#en t#ey are in process of becoming% and e!ist simultaneously w#en t#ey actually e!ist& and t#e same #olds good w#en t#ey are past and w#en t#ey are future. But w#at of cases w#ere t#ey are not simultaneous/ 7an causes and effects different from one anot#er form% as t#ey seem to us to form% a continuous succession% a past effect resulting from a past cause different from itself% a future effect from a future cause different from it% and an effect w#ic# is coming to be from a cause different from and prior to it/ 0ow on t#is t#eory it is from t#e posterior event t#at we reason -and t#is t#oug# t#ese later events actually #ave t#eir source of origin in previous events * a fact w#ic# s#ows t#at also w#en t#e effect is coming to be we still reason from t#e posterior event.% and from t#e event we cannot reason -we cannot argue t#at because an event A #as occurred% t#erefore an event B #as occurred subse$uently to A but still in t#e past * and t#e same #olds good if t#e occurrence is future. * cannot reason because% be t#e time interval definite or indefinite% it will never be possible to infer t#at because it is true to say t#at A occurred% t#erefore it is true to say t#at B% t#e subse$uent event% occurred& for in t#e interval between t#e events% t#oug# A #as already occurred% t#e latter statement will be false. And t#e same argument applies also to future events& i.e. one cannot infer from an event w#ic# occurred in t#e past t#at a future event will occur. "#e reason of t#is is t#at t#e middle must be #omogeneous% past w#en t#e e!tremes are past% future w#en t#ey are future% coming to be w#en t#ey are coming to be% actually e!istent w#en t#ey are actually e!istent& and t#ere cannot be a middle term #omogeneous wit# e!tremes respectively past and future. And it is a furt#er difficulty in t#is t#eory t#at t#e time interval can be neit#er indefinite nor definite% since during it t#e inference will be false. 6e #ave also to in$uire w#at it is t#at #olds events toget#er so t#at t#e coming to be now occurring in actual t#ings follows upon a past event. It is evident% we may suggest% t#at a past event and a present process cannot be (contiguous)% for not even two past events can be (contiguous). +or past events are limits and atomic& so 'ust as points are not (contiguous) neit#er are past events% since bot# are indivisible. +or t#e same reason a past event and a present process cannot be (contiguous)% for t#e process is divisible% t#e event indivisible. "#us t#e relation of present process to past event is analogous to t#at of line to point% since a process contains an infinity of past events. "#ese $uestions% #owever% must receive a more e!plicit treatment in our general t#eory of c#ange. "#e following must suffice as an account of t#e manner in w#ic# t#e middle would be identical wit# t#e cause on t#e supposition t#at coming to be is a series of consecutive events1 for in t#e terms of suc# a series too t#e middle and ma'or terms must form an immediate premiss& e.g. we argue t#at% since 7 #as occurred% t#erefore A occurred1 and 7)s occurrence was posterior% A)s prior& but 7 is t#e source of t#e inference because it is nearer to t#e present moment% and t#e starting point of time is t#e present. 6e ne!t argue t#at% since 2 #as occurred% t#erefore 7 occurred. "#en we conclude t#at% since 2 #as occurred% t#erefore A must #ave occurred& and t#e cause is 7% for since 2 #as occurred 7 must #ave occurred% and since 7 #as occurred A must previously #ave occurred. If we get our middle term in t#is way% will t#e series terminate in an immediate premiss% or since% as we said% no two events are (contiguous)% will a fres# middle term always intervene because t#ere is an infinity of middles/ 0o1 t#oug# no two events are (contiguous)% yet we must start from a premiss consisting of a middle and t#e present event as ma'or. "#e like is true of future events too% since if it is true to say t#at 2 will e!ist% it must be a prior trut# to say t#at A will e!ist% and <1

t#e cause of t#is conclusion is 7& for if 2 will e!ist% 7 will e!ist prior to 2% and if 7 will e!ist% A will e!ist prior to it. And #ere too t#e same infinite divisibility mig#t be urged% since future events are not (contiguous). But #ere too an immediate basic premiss must be assumed. And in t#e world of fact t#is is so1 if a #ouse #as been built% t#en blocks must #ave been $uarried and s#aped. "#e reason is t#at a #ouse #aving been built necessitates a foundation #aving been laid% and if a foundation #as been laid blocks must #ave been s#aped before#and. Again% if a #ouse will be built% blocks will similarly be s#aped before#and& and proof is t#roug# t#e middle in t#e same way% for t#e foundation will e!ist before t#e #ouse. 0ow we observe in 0ature a certain kind of circular process of coming to be& and t#is is possible only if t#e middle and e!treme terms are reciprocal% since conversion is conditioned by reciprocity in t#e terms of t#e proof. "#is * t#e convertibility of conclusions and premisses * #as been proved in our early c#apters% and t#e circular process is an instance of t#is. In actual fact it is e!emplified t#us1 w#en t#e eart# #ad been moistened an e!#alation was bound to rise% and w#en an e!#alation #ad risen cloud was bound to form% and from t#e formation of cloud rain necessarily resulted and by t#e fall of rain t#e eart# was necessarily moistened1 but t#is was t#e starting point% so t#at a circle is completed& for posit any one of t#e terms and anot#er follows from it% and from t#at anot#er% and from t#at again t#e first. 9ome occurrences are universal -for t#ey are% or come to be w#at t#ey are% always and in ever case.& ot#ers again are not always w#at t#ey are but only as a general rule1 for instance% not every man can grow a beard% but it is t#e general rule. In t#e case of suc# conne!ions t#e middle term too must be a general rule. +or if A is predicated universally of B and B of 7% A too must be predicated always and in every instance of 7% since to #old in every instance and always is of t#e nature of t#e universal. But we #ave assumed a conne!ion w#ic# is a general rule& conse$uently t#e middle term B must also be a general rule. 9o conne!ions w#ic# embody a general rule * i.e. w#ic# e!ist or come to be as a general rule * will also derive from immediate basic premisses. 1: 6e #ave already e!plained #ow essential nature is set out in t#e terms of a demonstration% and t#e sense in w#ic# it is or is not demonstrable or definable& so let us now discuss t#e met#od to be adopted in tracing t#e elements predicated as constituting t#e definable form. 0ow of t#e attributes w#ic# in#ere always in eac# several t#ing t#ere are some w#ic# are wider in e!tent t#an it but not wider t#an its genus -by attributes of wider e!tent mean all suc# as are universal attributes of eac# several sub'ect% but in t#eir application are not confined to t#at sub'ect.. w#ile an attribute may in#ere in every triad% yet also in a sub'ect not a triad * as being in#eres in triad but also in sub'ects not numbers at all * odd on t#e ot#er #and is an attribute in#ering in every triad and of wider application -in#ering as it does also in pentad.% but w#ic# does not e!tend beyond t#e genus of triad& for pentad is a number% but not#ing outside number is odd. It is suc# attributes w#ic# we #ave to select% up to t#e e!act point at w#ic# t#ey are severally of wider e!tent t#an t#e sub'ect but collectively coe!tensive wit# it& for t#is synt#esis must be t#e substance of t#e t#ing. +or e!ample every triad possesses t#e attributes number% odd% and prime in bot# senses% i.e. not only as possessing no divisors% but also as not being a sum of numbers. "#is% t#en% is precisely w#at triad is% vi,. a number% odd% and prime in t#e former and also t#e latter sense of t#e term1 for t#ese attributes taken severally apply% t#e first two to all odd numbers% t#e last to t#e dyad also as well as to t#e triad% but% taken collectively% to no ot#er sub'ect. 0ow since we #ave s#own above) t#at attributes predicated as belonging to t#e essential nature are necessary and t#at universals are necessary% and since t#e attributes w#ic# we select as in#ering in triad% or <5

in any ot#er sub'ect w#ose attributes we select in t#is way% are predicated as belonging to its essential nature% triad will t#us possess t#ese attributes necessarily. +urt#er% t#at t#e synt#esis of t#em constitutes t#e substance of triad is s#own by t#e following argument. If it is not identical wit# t#e being of triad% it must be related to triad as a genus named or nameless. It will t#en be of wider e!tent t#an triad * assuming t#at wider potential e!tent is t#e c#aracter of a genus. If on t#e ot#er #and t#is synt#esis is applicable to no sub'ect ot#er t#an t#e individual triads% it will be identical wit# t#e being of triad% because we make t#e furt#er assumption t#at t#e substance of eac# sub'ect is t#e predication of elements in its essential nature down to t#e last differentia c#aracteri,ing t#e individuals. It follows t#at any ot#er synt#esis t#us e!#ibited will likewise be identical wit# t#e being of t#e sub'ect. "#e aut#or of a #and book on a sub'ect t#at is a generic w#ole s#ould divide t#e genus into its first infimae species * number e.g. into triad and dyad * and t#en endeavour to sei,e t#eir definitions by t#e met#od we #ave described * t#e definition% for e!ample% of straig#t line or circle or rig#t angle. After t#at% #aving establis#ed w#at t#e category is to w#ic# t#e subaltern genus belongs * $uantity or $uality% for instance * #e s#ould e!amine t#e properties (peculiar) to t#e species% working t#roug# t#e pro!imate common differentiae. 3e s#ould proceed t#us because t#e attributes of t#e genera compounded of t#e infimae species will be clearly given by t#e definitions of t#e species& since t#e basic element of t#em all is t#e definition% i.e. t#e simple infirma species% and t#e attributes in#ere essentially in t#e simple infimae species% in t#e genera only in virtue of t#ese. 2ivisions according to differentiae are a useful accessory to t#is met#od. 6#at force t#ey #ave as proofs we did% indeed% e!plain above% but t#at merely towards collecting t#e essential nature t#ey may be of use we will proceed to s#ow. "#ey mig#t% indeed% seem to be of no use at all% but rat#er to assume everyt#ing at t#e start and to be no better t#an an initial assumption made wit#out division. But% in fact% t#e order in w#ic# t#e attributes are predicated does make a difference * it matters w#et#er we say animal tame biped% or biped animal tame. +or if every definable t#ing consists of two elements and (animal tame) forms a unity% and again out of t#is and t#e furt#er differentia man -or w#atever else is t#e unity under construction. is constituted% t#en t#e elements we assume #ave necessarily been reac#ed by division. Again% division is t#e only possible met#od of avoiding t#e omission of any element of t#e essential nature. "#us% if t#e primary genus is assumed and we t#en take one of t#e lower divisions% t#e dividendum will not fall w#ole into t#is division1 e.g. it is not all animal w#ic# is eit#er w#ole winged or split winged but all winged animal% for it is winged animal to w#ic# t#is differentiation belongs. "#e primary differentiation of animal is t#at wit#in w#ic# all animal falls. "#e like is true of every ot#er genus% w#et#er outside animal or a subaltern genus of animal& e.g. t#e primary differentiation of bird is t#at wit#in w#ic# falls every bird% of fis# t#at wit#in w#ic# falls every fis#. 9o% if we proceed in t#is way% we can be sure t#at not#ing #as been omitted1 by any ot#er met#od one is bound to omit somet#ing wit#out knowing it. "o define and divide one need not know t#e w#ole of e!istence. Cet some #old it impossible to know t#e differentiae distinguis#ing eac# t#ing from every single ot#er t#ing wit#out knowing every single ot#er t#ing& and one cannot% t#ey say% know eac# t#ing wit#out knowing its differentiae% since everyt#ing is identical wit# t#at from w#ic# it does not differ% and ot#er t#an t#at from w#ic# it differs. 0ow first of all t#is is a fallacy1 not every differentia precludes identity% since many differentiae in#ere in t#ings specifically identical% t#oug# not in t#e substance of t#ese nor essentially. 9econdly% w#en one #as taken one)s differing pair of opposites and assumed t#at t#e two sides e!#aust t#e genus% and t#at t#e sub'ect one seeks to define is present in one or ot#er of t#em% and one #as furt#er verified its presence in one of t#em& t#en it <:

does not matter w#et#er or not one knows all t#e ot#er sub'ects of w#ic# t#e differentiae are also predicated. +or it is obvious t#at w#en by t#is process one reac#es sub'ects incapable of furt#er differentiation one will possess t#e formula defining t#e substance. Moreover% to postulate t#at t#e division e!#austs t#e genus is not illegitimate if t#e opposites e!clude a middle& since if it is t#e differentia of t#at genus% anyt#ing contained in t#e genus must lie on one of t#e two sides. In establis#ing a definition by division one s#ould keep t#ree ob'ects in view1 -1. t#e admission only of elements in t#e definable form% -5. t#e arrangement of t#ese in t#e rig#t order% -:. t#e omission of no suc# elements. "#e first is feasible because one can establis# genus and differentia t#roug# t#e topic of t#e genus% 'ust as one can conclude t#e in#erence of an accident t#roug# t#e topic of t#e accident. "#e rig#t order will be ac#ieved if t#e rig#t term is assumed as primary% and t#is will be ensured if t#e term selected is predicable of all t#e ot#ers but not all t#ey of it& since t#ere must be one suc# term. 3aving assumed t#is we at once proceed in t#e same way wit# t#e lower terms& for our second term will be t#e first of t#e remainder% our t#ird t#e first of t#ose w#ic# follow t#e second in a (contiguous) series% since w#en t#e #ig#er term is e!cluded% t#at term of t#e remainder w#ic# is (contiguous) to it will be primary% and so on. 4ur procedure makes it clear t#at no elements in t#e definable form #ave been omitted1 we #ave taken t#e differentia t#at comes first in t#e order of division% pointing out t#at animal% e.g. is divisible e!#austively into A and B% and t#at t#e sub'ect accepts one of t#e two as its predicate. 0e!t we #ave taken t#e differentia of t#e w#ole t#us reac#ed% and s#own t#at t#e w#ole we finally reac# is not furt#er divisible * i.e. t#at as soon as we #ave taken t#e last differentia to form t#e concrete totality% t#is totality admits of no division into species. +or it is clear t#at t#ere is no superfluous addition% since all t#ese terms we #ave selected are elements in t#e definable form& and not#ing lacking% since any omission would #ave to be a genus or a differentia. 0ow t#e primary term is a genus% and t#is term taken in con'unction wit# its differentiae is a genus1 moreover t#e differentiae are all included% because t#ere is now no furt#er differentia& if t#ere were% t#e final concrete would admit of division into species% w#ic#% we said% is not t#e case. "o resume our account of t#e rig#t met#od of investigation1 6e must start by observing a set of similar * i.e. specifically identical * individuals% and consider w#at element t#ey #ave in common. 6e must t#en apply t#e same process to anot#er set of individuals w#ic# belong to one species and are generically but not specifically identical wit# t#e former set. 6#en we #ave establis#ed w#at t#e common element is in all members of t#is second species% and likewise in members of furt#er species% we s#ould again consider w#et#er t#e results establis#ed possess any identity% and persevere until we reac# a single formula% since t#is will be t#e definition of t#e t#ing. But if we reac# not one formula but two or more% evidently t#e definiendum cannot be one t#ing but must be more t#an one. I may illustrate my meaning as follows. If we were in$uiring w#at t#e essential nature of pride is% we s#ould e!amine instances of proud men we know of to see w#at% as suc#% t#ey #ave in common& e.g. if Alcibiades was proud% or Ac#illes and A'a! were proud% we s#ould find on in$uiring w#at t#ey all #ad in common% t#at it was intolerance of insult& it was t#is w#ic# drove Alcibiades to war% Ac#illes wrat#% and A'a! to suicide. 6e s#ould ne!t e!amine ot#er cases% Hysander% for e!ample% or 9ocrates% and t#en if t#ese #ave in common indifference alike to good and ill fortune% I take t#ese two results and in$uire w#at common element #ave e$uanimity amid t#e vicissitudes of life and impatience of dis#onour. If t#ey #ave none% t#ere will be two genera of pride. Besides% every definition is always universal and commensurate1 t#e p#ysician does not prescribe w#at is #ealt#y for a single eye% but for all eyes or for a determinate species of eye. It is also easier by t#is met#od to define t#e single species t#an t#e universal% and t#at is w#y our procedure s#ould be from t#e several species to t#e universal genera * t#is for t#e furt#er reason too t#at e$uivocation is less readily detected in <<

genera t#an in infimae species. Indeed% perspicuity is essential in definitions% 'ust as inferential movement is t#e minimum re$uired in demonstrations& and we s#all attain perspicuity if we can collect separately t#e definition of eac# species t#roug# t#e group of singulars w#ic# we #ave establis#ed e.g. t#e definition of similarity not un$ualified but restricted to colours and to figures& t#e definition of acuteness% but only of sound * and so proceed to t#e common universal wit# a careful avoidance of e$uivocation. 6e may add t#at if dialectical disputation must not employ metap#ors% clearly metap#ors and metap#orical e!pressions are precluded in definition1 ot#erwise dialectic would involve metap#ors. 1< In order to formulate t#e conne!ions we wis# to prove we #ave to select our analyses and divisions. "#e met#od of selection consists in laying down t#e common genus of all our sub'ects of investigation * if e.g. t#ey are animals% we lay down w#at t#e properties are w#ic# in#ere in every animal. "#ese establis#ed% we ne!t lay down t#e properties essentially connected wit# t#e first of t#e remaining classes * e.g. if t#is first subgenus is bird% t#e essential properties of every bird * and so on% always c#aracteri,ing t#e pro!imate subgenus. "#is will clearly at once enable us to say in virtue of w#at c#aracter t#e subgenera * man% e.g. or #orse * possess t#eir properties. Het A be animal% B t#e properties of every animal% 7 2 = various species of animal. "#en it is clear in virtue of w#at c#aracter B in#eres in 2 * namely A * and t#at it in#eres in 7 and = for t#e same reason1 and t#roug#out t#e remaining subgenera always t#e same rule applies. 6e are now taking our e!amples from t#e traditional class names% but we must not confine ourselves to considering t#ese. 6e must collect any ot#er common c#aracter w#ic# we observe% and t#en consider wit# w#at species it is connected and w#at.properties belong to it. +or e!ample% as t#e common properties of #orned animals we collect t#e possession of a t#ird stomac# and only one row of teet#. "#en since it is clear in virtue of w#at c#aracter t#ey possess t#ese attributes * namely t#eir #orned c#aracter * t#e ne!t $uestion is% to w#at species does t#e possession of #orns attac#/ Cet a furt#er met#od of selection is by analogy1 for we cannot find a single identical name to give to a s$uid)s pounce% a fis#)s spine% and an animal)s bone% alt#oug# t#ese too possess common properties as if t#ere were a single osseous nature. 1@ 9ome conne!ions t#at re$uire proof are identical in t#at t#ey possess an identical (middle) e.g. a w#ole group mig#t be proved t#roug# (reciprocal replacement) * and of t#ese one class are identical in genus% namely all t#ose w#ose difference consists in t#eir concerning different sub'ects or in t#eir mode of manifestation. "#is latter class may be e!emplified by t#e $uestions as to t#e causes respectively of ec#o% of reflection% and of t#e rainbow1 t#e conne!ions to be proved w#ic# t#ese $uestions embody are identical generically% because all t#ree are forms of repercussion& but specifically t#ey are different. 4t#er conne!ions t#at re$uire proof only differ in t#at t#e (middle) of t#e one is subordinate to t#e (middle) of t#e ot#er. +or e!ample1 6#y does t#e 0ile rise towards t#e end of t#e mont#/ Because towards its close t#e mont# is more stormy. 6#y is t#e mont# more stormy towards its close/ Because t#e moon is waning. 3ere t#e one cause is subordinate to t#e ot#er.

<@

1A "#e $uestion mig#t be raised wit# regard to cause and effect w#et#er w#en t#e effect is present t#e cause also is present& w#et#er% for instance% if a plant s#eds its leaves or t#e moon is eclipsed% t#ere is present also t#e cause of t#e eclipse or of t#e fall of t#e leaves * t#e possession of broad leaves% let us say% in t#e latter case% in t#e former t#e eart#)s interposition. +or% one mig#t argue% if t#is cause is not present% t#ese p#enomena will #ave some ot#er cause1 if it is present% its effect will be at once implied by it * t#e eclipse by t#e eart#)s interposition% t#e fall of t#e leaves by t#e possession of broad leaves& but if so% t#ey will be logically coincident and eac# capable of proof t#roug# t#e ot#er. Het me illustrate1 Het A be deciduous c#aracter% B t#e possession of broad leaves% 7 vine. 0ow if A in#eres in B -for every broad leaved plant is deciduous.% and B in 7 -every vine possessing broad leaves.& t#en A in#eres in 7 -every vine is deciduous.% and t#e middle term B is t#e cause. But we can also demonstrate t#at t#e vine #as broad leaves because it is deciduous. "#us% let 2 be broad leaved% = deciduous% + vine. "#en = in#eres in + -since every vine is deciduous.% and 2 in = -for every deciduous plant #as broad leaves.1 t#erefore every vine #as broad leaves% and t#e cause is its deciduous c#aracter. If% #owever% t#ey cannot eac# be t#e cause of t#e ot#er -for cause is prior to effect% and t#e eart#)s interposition is t#e cause of t#e moon)s eclipse and not t#e eclipse of t#e interposition. * if% t#en% demonstration t#roug# t#e cause is of t#e reasoned fact and demonstration not t#roug# t#e cause is of t#e bare fact% one w#o knows it t#roug# t#e eclipse knows t#e fact of t#e eart#)s interposition but not t#e reasoned fact. Moreover% t#at t#e eclipse is not t#e cause of t#e interposition% but t#e interposition of t#e eclipse% is obvious because t#e interposition is an element in t#e definition of eclipse% w#ic# s#ows t#at t#e eclipse is known t#roug# t#e interposition and not vice versa. 4n t#e ot#er #and% can a single effect #ave more t#an one cause/ 4ne mig#t argue as follows1 if t#e same attribute is predicable of more t#an one t#ing as its primary sub'ect% let B be a primary sub'ect in w#ic# A in#eres% and 7 anot#er primary sub'ect of A% and 2 and = primary sub'ects of B and 7 respectively. A will t#en in#ere in 2 and =% and B will be t#e cause of A)s in#erence in 2% 7 of A)s in#erence in =. "#e presence of t#e cause t#us necessitates t#at of t#e effect% but t#e presence of t#e effect necessitates t#e presence not of all t#at may cause it but only of a cause w#ic# yet need not be t#e w#ole cause. 6e may% #owever% suggest t#at if t#e conne!ion to be proved is always universal and commensurate% not only will t#e cause be a w#ole but also t#e effect will be universal and commensurate. +or instance% deciduous c#aracter will belong e!clusively to a sub'ect w#ic# is a w#ole% and% if t#is w#ole #as species% universally and commensurately to t#ose species * i.e. eit#er to all species of plant or to a single species. 9o in t#ese universal and commensurate conne!ions t#e (middle) and its effect must reciprocate% i.e. be convertible. 9upposing% for e!ample% t#at t#e reason w#y trees are deciduous is t#e coagulation of sap% t#en if a tree is deciduous% coagulation must be present% and if coagulation is present * not in any sub'ect but in a tree * t#en t#at tree must be deciduous. 1B 7an t#e cause of an identical effect be not identical in every instance of t#e effect but different/ 4r is t#at impossible/ Per#aps it is impossible if t#e effect is demonstrated as essential and not as in#ering in virtue of a symptom or an accident * because t#e middle is t#en t#e definition of t#e ma'or term * t#oug# possible if t#e demonstration is not essential. 0ow it is possible to consider t#e effect and its sub'ect as an accidental con'unction% t#oug# suc# con'unctions would not be regarded as conne!ions demanding scientific proof. But if t#ey are accepted as suc#% t#e middle will correspond to t#e e!tremes% and be e$uivocal if t#ey are e$uivocal% generically one if t#ey <A

are generically one. "ake t#e $uestion w#y proportionals alternate. "#e cause w#en t#ey are lines% and w#en t#ey are numbers% is bot# different and identical& different in so far as lines are lines and not numbers% identical as involving a given determinate increment. In all proportionals t#is is so. Again% t#e cause of likeness between colour and colour is ot#er t#an t#at between figure and figure& for likeness #ere is e$uivocal% meaning per#aps in t#e latter case e$uality of t#e ratios of t#e sides and e$uality of t#e angles% in t#e case of colours identity of t#e act of perceiving t#em% or somet#ing else of t#e sort. Again% conne!ions re$uiring proof w#ic# are identical by analogy middles also analogous. "#e trut# is t#at cause% effect% and sub'ect are reciprocally predicable in t#e following way. If t#e species are taken severally% t#e effect is wider t#an t#e sub'ect -e.g. t#e possession of e!ternal angles e$ual to four rig#t angles is an attribute wider t#an triangle or are.% but it is coe!tensive wit# t#e species taken collectively -in t#is instance wit# all figures w#ose e!ternal angles are e$ual to four rig#t angles.. And t#e middle likewise reciprocates% for t#e middle is a definition of t#e ma'or& w#ic# is incidentally t#e reason w#y all t#e sciences are built up t#roug# definition. 6e may illustrate as follows. 2eciduous is a universal attribute of vine% and is at t#e same time of wider e!tent t#an vine& and of fig% and is of wider e!tent t#an fig1 but it is not wider t#an but coe!tensive wit# t#e totality of t#e species. "#en if you take t#e middle w#ic# is pro!imate% it is a definition of deciduous. I say t#at% because you will first reac# a middle ne!t t#e sub'ect% and a premiss asserting it of t#e w#ole sub'ect% and after t#at a middle * t#e coagulation of sap or somet#ing of t#e sort * proving t#e conne!ion of t#e first middle wit# t#e ma'or1 but it is t#e coagulation of sap at t#e 'unction of leaf stalk and stem w#ic# defines deciduous. If an e!planation in formal terms of t#e inter relation of cause and effect is demanded% we s#all offer t#e following. Het A be an attribute of all B% and B of every species of 2% but so t#at bot# A and B are wider t#an t#eir respective sub'ects. "#en B will be a universal attribute of eac# species of 2 -since I call suc# an attribute universal even if it is not commensurate% and I call an attribute primary universal if it is commensurate% not wit# eac# species severally but wit# t#eir totality.% and it e!tends beyond eac# of t#em taken separately. "#us% B is t#e cause of A)s in#erence in t#e species of 21 conse$uently A must be of wider e!tent t#an B& ot#erwise w#y s#ould B be t#e cause of A)s in#erence in 2 any more t#an A t#e cause of B)s in#erence in 2/ 0ow if A is an attribute of all t#e species of =% all t#e species of = will be united by possessing some common cause ot#er t#an B1 ot#erwise #ow s#all we be able to say t#at A is predicable of all of w#ic# = is predicable% w#ile = is not predicable of all of w#ic# A can be predicated/ I mean #ow can t#ere fail to be some special cause of A)s in#erence in =% as t#ere was of A)s in#erence in all t#e species of 2/ "#en are t#e species of =% too% united by possessing some common cause/ "#is cause we must look for. Het us call it 7. 6e conclude% t#en% t#at t#e same effect may #ave more t#an one cause% but not in sub'ects specifically identical. +or instance% t#e cause of longevity in $uadrupeds is lack of bile% in birds a dry constitution * or certainly somet#ing different. 1D If immediate premisses are not reac#ed at once% and t#ere is not merely one middle but several middles% i.e. several causes& is t#e cause of t#e property)s in#erence in t#e several species t#e middle w#ic# is pro!imate to t#e primary universal% or t#e middle w#ic# is pro!imate to t#e species/ 7learly t#e cause is t#at nearest to eac# species severally in w#ic# it is manifested% for t#at is t#e cause of t#e sub'ect)s falling under t#e universal. "o illustrate formally1 7 is t#e cause

<B

of B)s in#erence in 2& #ence 7 is t#e cause of A)s in#erence in 2% B of A)s in#erence in 7% w#ile t#e cause of A)s in#erence in B is B itself. 1E As regards syllogism and demonstration% t#e definition of% and t#e conditions re$uired to produce eac# of t#em% are now clear% and wit# t#at also t#e definition of% and t#e conditions re$uired to produce% demonstrative knowledge% since it is t#e same as demonstration. As to t#e basic premisses% #ow t#ey become known and w#at is t#e developed state of knowledge of t#em is made clear by raising some preliminary problems. 6e #ave already said t#at scientific knowledge t#roug# demonstration is impossible unless a man knows t#e primary immediate premisses. But t#ere are $uestions w#ic# mig#t be raised in respect of t#e appre#ension of t#ese immediate premisses1 one mig#t not only ask w#et#er it is of t#e same kind as t#e appre#ension of t#e conclusions% but also w#et#er t#ere is or is not scientific knowledge of bot#& or scientific knowledge of t#e latter% and of t#e former a different kind of knowledge& and% furt#er% w#et#er t#e developed states of knowledge are not innate but come to be in us% or are innate but at first unnoticed. 0ow it is strange if we possess t#em from birt#& for it means t#at we possess appre#ensions more accurate t#an demonstration and fail to notice t#em. If on t#e ot#er #and we ac$uire t#em and do not previously possess t#em% #ow could we appre#end and learn wit#out a basis of pre e!istent knowledge/ +or t#at is impossible% as we used to find in t#e case of demonstration. 9o it emerges t#at neit#er can we possess t#em from birt#% nor can t#ey come to be in us if we are wit#out knowledge of t#em to t#e e!tent of #aving no suc# developed state at all. "#erefore we must possess a capacity of some sort% but not suc# as to rank #ig#er in accuracy t#an t#ese developed states. And t#is at least is an obvious c#aracteristic of all animals% for t#ey possess a congenital discriminative capacity w#ic# is called sense perception. But t#oug# sense perception is innate in all animals% in some t#e sense impression comes to persist% in ot#ers it does not. 9o animals in w#ic# t#is persistence does not come to be #ave eit#er no knowledge at all outside t#e act of perceiving% or no knowledge of ob'ects of w#ic# no impression persists& animals in w#ic# it does come into being #ave perception and can continue to retain t#e sense impression in t#e soul1 and w#en suc# persistence is fre$uently repeated a furt#er distinction at once arises between t#ose w#ic# out of t#e persistence of suc# sense impressions develop a power of systemati,ing t#em and t#ose w#ic# do not. 9o out of sense perception comes to be w#at we call memory% and out of fre$uently repeated memories of t#e same t#ing develops e!perience& for a number of memories constitute a single e!perience. +rom e!perience again * i.e. from t#e universal now stabili,ed in its entirety wit#in t#e soul% t#e one beside t#e many w#ic# is a single identity wit#in t#em all * originate t#e skill of t#e craftsman and t#e knowledge of t#e man of science% skill in t#e sp#ere of coming to be and science in t#e sp#ere of being. 6e conclude t#at t#ese states of knowledge are neit#er innate in a determinate form% nor developed from ot#er #ig#er states of knowledge% but from sense perception. It is like a rout in battle stopped by first one man making a stand and t#en anot#er% until t#e original formation #as been restored. "#e soul is so constituted as to be capable of t#is process. Het us now restate t#e account given already% t#oug# wit# insufficient clearness. 6#en one of a number of logically indiscriminable particulars #as made a stand% t#e earliest universal is present in t#e soul1 for t#oug# t#e act of sense perception is of t#e particular% its content is universal * is man% for e!ample% not t#e man 7allias. A fres# stand is made among t#ese rudimentary universals% and t#e process does not cease until t#e indivisible concepts% t#e true universals% are <D

establis#ed1 e.g. suc# and suc# a species of animal is a step towards t#e genus animal% w#ic# by t#e same process is a step towards a furt#er generali,ation. "#us it is clear t#at we must get to know t#e primary premisses by induction& for t#e met#od by w#ic# even sense perception implants t#e universal is inductive. 0ow of t#e t#inking states by w#ic# we grasp trut#% some are unfailingly true% ot#ers admit of error * opinion% for instance% and calculation% w#ereas scientific knowing and intuition are always true1 furt#er% no ot#er kind of t#oug#t e!cept intuition is more accurate t#an scientific knowledge% w#ereas primary premisses are more knowable t#an demonstrations% and all scientific knowledge is discursive. +rom t#ese considerations it follows t#at t#ere will be no scientific knowledge of t#e primary premisses% and since e!cept intuition not#ing can be truer t#an scientific knowledge% it will be intuition t#at appre#ends t#e primary premisses * a result w#ic# also follows from t#e fact t#at demonstration cannot be t#e originative source of demonstration% nor% conse$uently% scientific knowledge of scientific knowledge.If% t#erefore% it is t#e only ot#er kind of true t#inking e!cept scientific knowing% intuition will be t#e originative source of scientific knowledge. And t#e originative source of science grasps t#e original basic premiss% w#ile science as a w#ole is similarly related as originative source to t#e w#ole body of fact.

<E

S-ar putea să vă placă și