Vous êtes sur la page 1sur 18
Gentner, De, Jesforekt, M. (1993). The shift from metaphor to’ analogy in'western science. Tn A. Ortony (Ed.), Metaphor and Thought (2nd ed) (pp. 447480). Cambridge, England: Canbridge University Pro: 20 ‘The shift from metaphor to analogy in Western science DEDRE GENTNER AND MICHAEL JEZIORSKI ‘Analogy and metaphor are ccnral to scientific thought. They figure in ‘dscovery asin Rutherford’s analogy of the sla system fo the atom or Faraday’ use of lies of magnetized ion filings to reason about electric elds (Nersessian, 1984 Tweney, 1983), They are also used in teaching: novice ate tld 10 thik of electricity as analogous to water owing through pipes (Gentner & Geniner, 1983) of of a chemical procs a ‘nslogous 10 ball oliag down a fll (Van Lehn & 1S. Brown, 1980). Yet for al its usefulness, analopcal thinking i aever formally taught 10 us, We Seem to thnk of it as 4 natural human skill, and of its use in science as a taighforward extension of is use in commonsense reason- ‘ing. For example, Willam James belived that “men, taken historically, reason by analogy long before they have learned to veason by abstract character” (ames, 1890, vl. Ip. 363). All his poins to an appealing Incution: that a faculty for analogical reasoning i an innate part of hue ‘man cognition, and thatthe concept ofa sound, inferential sefl anal yze the way in which history of ‘Wester seientie thought, tracing thei use backard from the present le. We begin by laying out the curtent framework for analogical reason ing, followed by two examples that conform othe modern aesthetic, hose ‘ot Sad Carnot (1796-1832) and Robert Boyle (1627-1691). We go 0810 onsder a very different way of using analogy and metaphor in science, that practice by the alchemists (about 300-1600 4.0). Based on these cramples, we conce that there are important differences inthe Kinds of iy DEDRE GENTNER AND MICHAEL JEZIORSK! intries hat wor fl to warn nferences about the werd and in he sent that mere draw from ermparsons. tn short hee vio epcn giant change 0 what has Soe 8 eee we of anaony and metaphor. Ser ee pough a appreciation of simian Gnching os ee ay) i amon rely univer man cops eee hin see of aman Opinion om ow ram the ret cnc uve varied. Many great hikes have simply Fa ere acrononced tat "2 piesoper shold abstain aed ett hough wig to permit metaphor 2 ot ae tral ngage shoud ot be wed in agumentation ment tad el jowever tat the pereption oF series Bewee To el be a sure of speci nap) Atte oposite Pole sara Ct lace embraced metaphor and analogy wih be healer WG eces wae toth quant and quan, They le eager rr of metaphors a they imbued them wih reat power eed a ere (8H) put owned by thst anager than Toe wee nay te moter ew, a represented by Bove and i ac horilsmiaity bu observes firm constants om ca ae bring ea be summed p sls “And me ee Insc roe afi, But you wish them respecte. bet do 0 nt ety 195, 5). vt foes onthe eration it to day a (othe aches plurals se ofall sors of meta wes tus the mone autere mer fos on ria na ae rh ging oo what we ake fo be the caren! eon Euhetics for analog reasoning, A teamenrk fr analogy and sary ante viewed x aking tight simian. In rsesing Aan aly fas on een kinds of commenaliis 30 azo ae tse etn sae ier tn eet del tact are lings ae of ck an ps ta ke a fay el hes SOUS Se ian to enrate produ. Tis ocesen commen ee acon what mkes analogy laminating Analoay Sirucure-mapping and ideal anlogicel competence ory of human procesing of analogy and simi the descriptive constrains that characterize the ‘Saity (Genter, 1982, 1983, 1989), and nen understanding 2 similarity compat ree The central a that an analogy 88 sractresmapping is 8th iy. Iam wo capture bol Interpretation of analogy ad the processes humans engage ton (Markman & Gentine, From metaphor to analogy in science 8 mapping of knowledge from one domain (he base) into another (ihe ta {rt such tha system of relations that holds among the base objects a olds among the target object. In interpreting an analogy, people seek 10 put the objects ofthe base in one-to-one correspondence with the objets ‘rane target so a to obtain the maxima seuctral match. The correspond: ine objets in the base and target ned nt resemble each other, rather, ‘bjon correspondences are determined by Hike oles inthe matching rela float struces, Thus; an analogy tsa way of aligning and focusing on ‘elatinal commonalities independently ofthe objec in which those rel tone are embedded. Central tothe mapping process i the principle of Ipuematciy: people prefer t0 map systems of predicates governed by igherorder relations ith inferential import rather than to map isolated predicates, The systematic principle elec tcitpreferene for coher- nee and inferential power in interpreting analogy "Comer, for example, Rutherford's analogy between the solar system and the hydrogen atom. A peso hearing it for the first time (assuming Same prior Knowledge about the roar system) must find a st of relations ‘Sommon tothe ase and the target that can Be consistently mapped and hat as decp (ve as systematic) as posible. Here the deepest common Felational system the central force causal system cause [ano [arreacts (un, plant), monesasve (sun, planet). tevotves around (planet, sun) Isolated relations, such as Horren-ra un, pane) that donot belong to th connected system, ae dsteparded. The descriptions of individual 0 jews fe, vELvow (ain)} are abo disregarded. The object correspon Sncos arrived at re those dictated by the system of relational matches: ‘sun — nucleo, and planet electron, “Athough thee are some diffrence in emphasis there isa fair amount cof convergence om the Kinds of sroctural principles discussed above (Burein TOR3; Hesse, 1966, Hofstadter, 181: Holyoak & Thagard, 189; Keane, 1985, 1988, Reed, 1987; Romethart & Norman, 1981; Winston, 198) 1942). Tere is widespread xpreement onthe basic elements of one: lovone mappings of objects wth carryover of predicates, and many re ‘Rarchers ta sme for of systematic to constrain the interpretation of Shaogy (although there are exceptions; see Anderson. 1981). Theres aso Cmpinkal support forthe peycholopial predictions of sructuremapping ‘eny, Three findings are of particular relevance her. First, adults tend 10 Include relations and omit atributesia thei iterpretations of analogy. nd {ey judge analogies as more apt and more sound if they share systematic {elaional structure (Gentoer, 1988; Genter & Clement, 1988; Gentner & ander, 1988; Gentner & Rattermann, 1991). Second, adults (and eit Gren) ate more accurate in analogical transfer when there #2 systematic ‘elaonal striate in the base domain that canbe used to guide the map- oe 555 DEDRE GENTNER AND MICHAEL JEZIORSKI “Table 20.1, Modem principles of enological reasoning ‘Ohjcs av plied in ope tone correspondence predicts mane. 2a ystems ar rsered ad objet Sosripions 1 Src conten Sadana comet ‘hina fo (tepaaed 9 tame annng vais rains epretations, the one with he Sprematy, i ne rete degres common bg ore ee are wpe ea ta tins Only commonalities srnges satay. Fur Me runtonsbtwenn the bse adapt = fr example, pens = dont cont 0 he a. aerate Te lana work ob mapped soul te carly ei eter he domain. When two bse aT se. they sould ESchconey aeabeent stm OG Satine Ta oo phenomena ae analogs does oti © pores nth on ees hh ping (Gentner & Toupin, 1986; Markman & Genter, in res: Ros Te (rind, adults shed vo make Hew predictions fom an analogy base ‘ear predictons on common relational tucture. They are moc kel 0 rites anew fat the arget when the corresponding atin he base aor connceted oa comanon srvture (Clement & Gotner 191) “Analogical soundness, The foregoing discussion suggests & se of tit con- deta moder scientists we in analogical reasoning, We believe there Se eh pimps. as given im Table 20.1 The first tex principles, a el soescney, rlaomel focus, nd systema, have already been Fee “The our principle, no extraneous relations, express he point aeons about commonalities, Discovering ther celatonships be eaithe ase and target doesnot prove the analogy. For example the {fatihar the aun and plante are made yp of atoms doesnot strengthen he stom/solar system analogy Pre ned unalote’ principle reflects the sense that analogic com srueted by mapping fom several base domains info the same arg are wre douad nt best cise auch mined comparison edt ak cob cee neronder structure, and inthe worst case they contain contradictory ‘rlppngs avin these examples quoted nthe New Yorker: Sree sting on launching pad xing is musts, Tee 8 and the Mil East are on pall but non-convergent pat Traferetial reasoning we prefer tha he relational system mapped into “A UueGS drawn trom asinge bape domain. There are exceptions in caer LUE tterent analogies are used to capture separable aspects ov subs= wre the tnpet (Burstein, 198; Colas & Genter, 1987). But such (eohupe analogies require rm rls of intersection to avoid inconsistent From metaphor to analogy in sience 451 mappings (sce Coukon, Feltovieh, & Spiro, 1986, fora medical example). Unruly Toss site te consensual rues of sound thinking." Finally, analogy now causation, I our eutentcopntve practice, the presence ofan anaogy between two situations hs no bearing on whether re isa caval relation between the two situations. Conversely evidence ‘fa causal relation between two analogous domains us no bearing on how Similar or analogous they are. This pot can be confusing, since common ‘Hunt relations do contribute 10 the goodness of analogy, For example, {Sven two possibly analogous stations A and B, the analogy between A nu B strengthened by acing like eausal elatons vo both terms. Thus, Snalogy (2) is etter than analogy (I). (1) tds pastes Sam. Samhita tee. a (2) lds pushed Sam, Flipper pushed Shamu Sam hit tee, Shama hita buoy. ‘Adding common causal relations makes analogy (2) superior to (1) for fhrve Feasons (a) adding common features increases the goodness of a match (Tversky, 1977; (0) more specially. adding common featres sha, tre conneced 19 common systems mnereases the goodness Of atch more {han does adding other commonalities, since interconnected elements sup- port each other inthe evaluation process, and () sill more specifically, Sing common higher-order containing relauons (sucks the esa el tion) ereases the goodness ofa match more than adding ater connected. Commonalits, singe the coherence and sjtematiciy of the analogy thereby increased (Clement & Gentse, 199K; Frtus & Gentne, 1989), ‘But although ang bike causal relations wih wo situations stength- ‘ens an analogy, adding causa relations Berween tem docs not. Ths, aa ‘ty @) between A and B fe not better than analogy (1) between the same two event sets, eventhough analogy (3) has an aditonal causal relation between the analogs: i n 8 {@) Ida pushed Sam. CAUSE Flipper pushed Shamu. Samhita tee ‘Shima hits buoy “Tha is the analogy does not improve if we are tld that Sam's iting & tice caused Shamu tit the buoy That A causes B may be an interesting ‘neti, but itJoes not make A and B moc similar or more analogous. Inout current cognitive sesthetic, adding common eawsal relations teach Of the domains increases the goodness Of an analogy. but adding causal ‘Sonnections between the unalog does no.” 452. DEDRE GENTNER AND MICHAEL JEZIORSKI “Analogy in reasoning, Te constants on analogical reasoning are closely AaeP he process of making new inferences. As mentioned above, Shalogicalinferenecs are typically made by a proces of stem completion Mle gome degree of meh has Been established. We have modeled this jrocen in computer simulation eaed SME (the stroctre-mapping en ie) (Faikenhaincr, Forbus, & Genter, 1989; Forbus & Gentner, 19% Pittes & Oblinger, 199: Gentner Falkenbainer, & Skorstad 1988). The ‘ute fst makes ll possible Tocl matches between like components and Mes Skempe to link these int structorally consistent ystems of matches. ‘The Largest and deepest global interpretation wns. Given such 2 comion ‘Jutcarsa now candice inferences generated ia predicate belongs tothe BEE Stem bats counterpart does no yet appear the target system, “That the portly matching systems complete in the target. "Candidate inferences are only conjectres. The sx principles of analog cal seasoning are concerned with whether the analogy Istrcturally sound sae ath whether it inferences are factually correct. Verifying the factual Tatu of the analogy fa separate process. Soundness principles simpli {he tsk, howene, because they specify what most be tue inorder forthe Matogy to old, Ins sjtem of interconnected matches, even one signif TEN dtconfemtion can invalidate a whole analogy Metaphor and other similarity matches, Other kinds of similarity matches Ete stingushed inthis framework. Whereas analogies map relational SMhetare independently of object descriptions, mereappearance marches ‘map aspects of ebject descriptions without regard for relational sractare. cal inforiy maches map bot relational stuctre and object descrip: tions We view metphor asa rather bread category. encompasing analogy ied ere appearance, a wel 3 variety of other Kinds of matches. On iis low, asogy special css of metaphor. one based on a purely TRlsuonal match: The large-scale communication metaphor analyzed by Reddy (hit volume), as well ae other conceptual metaphors analyzed by {ako an Jonson (1980) and by Lakot (this volume), are examples of ‘Sutematiselauonal metaphors, thats, metaphors that could also qualify SP fulogicn There are alzoalebutional metaphors ~ mere-appearance Matches, based on shared object descriptions ~for example, “her arms sepa vans." well 3s metaphors based on micas of object and signa commonalities, Further, there ae metaphor that cannot be an heed nthe simple tere we have used 80 fa. fr example, “the voice of Pear eyes deeper than al roses” (ee cummings); "On a star of faith ea the diling bead 'As the food and flames ofthe snow” (Dylan poms), These metaphors are not bound by the eaefo-one mapping etraint and can include mixtures of several bases, a wells themati Shu metonymc relation (Gentner, 1982: Genter, Falkenfner, & Skor- stad, 1988). From metaphor analogy in scence istorical wes of analogy Despite the plurality f posible match types, the guidelines for use of Jnalogy in scientific discowery and reasoning are quite selective. The “Mrengih ot an analogy in licensing sientie prediction rests. on the degree Stayetematie structural match Between the two domains. WE now ask Sthether ester sientss have always adhered 0 these principles, We wl, ‘Corser evidence that the ascendancy of analogy over metaphor in scien ie reasoning was not always the ease. We begin with Carnot,» faith recent example, and progress in reverse chronological order. Sadi Carot “The French scientist Sad Carnot (1796-1832) was one of the pincers of modern thermodynamics, He described the Carnot cycle fr het engines. Sli taught es am ideal energy conversion system, and lid the foundation forthe Inter discovery ofthe equivalence of heat and work. In his treatise fo heat, Carnot presented an analogy between heat and water tha lifes Fis postion nd generated new questions 15 use of analogy is prototype {aloft rules of rigor descnibed shove and can stand as an example ofthe fmodern use of analogy. In 1824, Carnot published Rflesions sur a pus: ance moti da feu (Reflections onthe Move Powe of Fire). In this bok he deserts the fuetioning ofa hypothetical engine that an convert heat ergy 10 work. This engine consis of a elinger filed with gas and filed Stns ection piston that can move rely ise the eylnder. During Tourstage eye, the gas ise expanded by contact with a heal source (isotherm expansion) and allowed fo continue ating after the source 1 Femeved (adabatic expansion) ‘Te gas then compressed by transmis ‘So of hc to elder body (sothermal compression) and the volume Further decreases after removal of the cold body (adiabatic compresion). Festoing the oriinal conditions of the system. During this period. the ‘rine hs absorbed certain amount of heat and converted ito mechan al work throug the movement of the piston, The operation of this al {engine koown ae the Carot eye, vas an important theoretical continu tion tothe carly development of thermodynamics. Tn the Reflevons, Carat utllzed an analogy Between water falling ‘through «waterfall and calorie heat) “fling trough a heat engine, The Snalogy between heat and Mud was ot ew. Indeed, the dominant theory Stent a he time was the ear theory, which defined heat asa weight less lid that shared the properties of ordinary mater. Like other mater Caloric was considered conserved quantity that could be neither eeated hor destroyed. Carn’ contribution wap n0t the idea of viewing Ret as fuid but rather the thoroughness of his development of the heattwater 45 DEDRE GENTNER AND MICHAEL JEZIORSKI res rom the analogy ~ the extent o which he applied explicit causal wenter domain to the heat domain: 1. Acoiing to established picpe a he pesent ime, we can compare with sofa asuro) the motive power of eat fa of 3 waterfall. Each as 3 [matin that ne camot cee, whatever may be, on he onc hand, he mache ‘Thch casted upon bythe mater, ad whatever, on the ther and he substance ‘Ered upon by the Ret ‘The mative power of weal depends on is ight ad on the quantity of thei the sti power of beat depen ao on the quan of cao wed. Shon wat maybe termed om what act we wil el he eg of fll tat {Osby he ference of temperate he Bodies between the ihe a ower “nhc water he motive powers exaly proportional othe ilerence of level betmcn the higher and toner reservar. due fll of eal the more [pet undoubtedly intense with he ference of temperate between he Walt ide cod bodies but we dono om whether i proyrtna oth ier fee. We do got know, for example, whether the fall fer fom 100 t0 50 ‘pes fishes oreo ee motive power than he fl of hs same cal from Sto sera in asin wich we propose to examine beret. (Carol, 197. 15; mmbers and paragagh eat se ssrted for conenens: the origi Page ontious) In section 1, Carnot introduces the analogy between the motive power of heat andthe motive power of water and establishes and notes simple yet important pralel just asthe amount of power produced by a given fall of rater is limited, the power atsinable fom a sven transfer of heat is Timited. In section 2, Carnot estaihes further corespondences and a shored higher-order principe. He compares the difference in temperate tetween two connected Bodies 10 the height of the fallin waterfall CCammot uses this correspondence ina proposed higher-order relation: he ‘Suerts that in each eat, the power produced bythe system depends on both the amount ofthe substance (Water or caloric) that “falls” and the tance ofthe “drop” between levels “This qualitative combination ~ the fact that power depends on both the Aiterence in level and the amount of substance” involved ~ farther sharp fens the analogy. Figure 20.1 shows the common relational strctore that folds for water and heat, Figure 2.2 ses forth the corresponding terms and assertions ‘So far the enterprise as been one of matching structures between heat and water In Section 3, new candidate inference isransferred from water to beat that is the analogy f used 10 suggest a new hypothesis. Carmot totes & higher-order relation in the domain of water (the fst that the power produced by a given fll of water is dicelly proportional to the fiflerence between levels). He asks whether the same relation exists for heat engines; that i, whether the power produced by 2 given “tll” From metaphor analogy in science 485 ok “SR CaS Goat oe > ee Fer gure 201 Cano’ sly: the common relational strctre for wate and et NTNER AND MICHAEL JEZIORSKE 4% DEDREG! eFERENCe etre) BirenENCE emp emo) Faw nh Rat Powenn ter Max POWER ND Se Pome 5. wate aOPONEN Pt OFFERENCE fer, ne) Sir GQpoweR inn, DEFERENCE fonpe amp) 6 weer POWER O, anc) 7 ne. AND @}POWER ) OFFERENCE vl et ‘xO(OWEFh om) est AND AOPOWER hy DEFERENCE fpr. emp sOPOWER on) Figure 212, Terms and propositions dri fam Carma’ watrhet ana {ny flor a0 Jeotesquaatve portional (Fobas. 1984). 20 Baie: Felina sa monotone fonction of Bt does ot speci the are othe Tancion or whether ther vitaes maya ae A. colori emains constant, regardless of he temperature at which that fal ‘Guinot s development of his analogy ndings rm the modern scene we of suloy. 1 mess te sx principles of gpots analog ‘SSsoning sessed eer. Carmos piv the objets nthe two domain reanane eomeapondence nso doin. he deparde objec atte Slicer He was et conceracd with whether coneponding components Shaved srlce quails, bu with ahicving common ystems eaten incre, Havngenplstdstigher ode rlaonal sem conmonto the {wo domain, Cao war abe to expo that sytem fo map arn futer poms rom the base to the target Btwccn-domai lations such 8 karen be ht were vod was ny Hn of mixed aay In “More, Cara's we of analogy conforms to modern sent practice Robert Boyle We now move back another century and a hal to the English scientist Rober! Bove (1627-1691. Boyle is considered among the founders of modern tomy. Hes best know for his work onthe dea gas law, bat he also comtrtuted to other domains, such asthe theory of acids and From metaphor analogy in science > 857 alkalies, Probably his most aflcntiat work was the Scetical Chymis in Swhich he eticzed both the Greck division of mater into four elements andthe later division into tree principles. Appearing snomy meus in 16 fd again in 1679 wah additions. t "Gd more than any oer work of the Eensury to arouse a rly eritieal spirit of Slee login chemical hh ing” (Stillman. 1950p. 395) Boyle wax a prolife writer. anerested in ‘hilosophy and religion as well asthe sciences, He was aso\a prolife {naloizer. He often put forth several examples or analogs foreach pine ‘le Be wanted to prove, ‘A characters example of Boyle's us of examples and analogies occurs in is book Of he great effets of even languid and unheeded local motion, published in 169. His porpore in his book was to demonstrate the Imper {ance of “ocal motion. the motions of many tiny partes Boyle wanted {establish thatthe combined effects ofthe motion of many tiny particles — each invisile and insignificant in Wsel can cause large sale changes. He Saw such effects as unifying principle serose domaine such sight, sound, fe, ad Maids. Although some of hs points now seem to need no defense. ‘ach was not the casein his time. To marshal sffitent evidence for his Conjecture Rosle cited examples from one domain after anther Boy's examples appear to funtion in two ways. Fist they serve as instances of local motion andi effects ~ tht fa nstances of principle that canbe efcetively applied to several domain. The more numerous and ‘arid the instances the more faith we ean presumably have in the print Ble. Second, the examples serve as analogies that can be aligned with one nother to yield common structualabseacions. By jostaponng separate Instances of fea! motion, Boyle Ted his reader to focus onthe common ‘ual system, The following excerpeiustrate his style of analogizing (Chap. 1) Omarva. I. Men andere the motion of Bode 100 sl 1 be ‘ule or serena Names ih em at 1 oye prams ht ns men think ofthe prices abodes a ike gai of dst ich athowsh ee. camot penetae the Roe he al At Fea "hee gain cart ae the aie | Bus we may heer thoughts, we welcome, th he Comps me ps ‘fae. by thr minatenem ised and oenties bythe re aed Bierce it the ioncrmos recess f he bady they Invade, nd ue them {eves taller at eat maitece of the mite pare where tet Bey ons For hich rate. tough we suppe each ingle flim parte tothe very minute: yet ice we may supe ever so oct be made po Faris that nem to, ad the mame of invading pes oe mot mich Inrioe tha fhe ade nes. teat wo be eneedingy eet it! need scm intel, tht malitade of title Coram motion (whine mation, ‘ma. fr ght me Kw. e very stat) shold be abe to hive comsderabe 46" DEDRE GENTNER AND MICHAEL JEZIORSKI oycration wpm panies eter yucscet, oF hat ve mio Kd koe 1 be ere lcac Which may omapsbe the eter conved hy he ep om cramp SP Faampl of he anit {you tor an Ant wel socks! with Anes ule down, yo may sme oe ecpvceps mango tea rh aoa hose es eee ef ogee weil ot esl a neo aw afer: bi feta tem anne ees ad ee oan en, oh Heres setand any tants some uping tse (1 ave teh sie ne) how gully the heap of gs le place. when alos Tefen ftir eal ma a heen he vaing id des te in and cary cane esi profi bates vopcaion may eile Beane pte in Aunn oy oraesit el ade of eae eke men pcm oe betwee th branches mba ceed snd ins psig om thon ae tear abe et te maul bend ie oe exe Bop end Yes a pei inet opm donb treks sm fhe sais a the. sees (te pound nd wale vers of he aes ht see ettmiy nan nepogs des by sent pow aie Enc onan aes the ston SEimplof grand amber dasooing S Etang omen compound noting 6 vamp of ame vig etal at ep nua a Pa s(n Lupe you wi hk fr he ore sins at my ts) tough you wl ess grat tha he Mame Of Spit of ena ur a away bra bible agen af ouch Efi wily TIAL Thaour any semble Hest would of teste ileal ay wee Cnn he tad oe, ht of Cope a Jee ie cs ihe me ore Spit of nc yo wl ul eae SoG det encarta vehement gon fhe mie Sea Helmet ha mre them mus have ben ect pated Orem in ncn prt, we spe abt = Pee) see ail mati pentane the tanto ur a the me ad ty ur exeriment peas that ad nd (Thesher we sepose thio Se Tioga of Spt of win, abe lint a ee erly upon be ares me 7 Eampe of mime spr movie annals 1 Euomple of rope omracting fom humiiy (oye, 100, pp. 27-39) Boyle begins by noting that laymen may find it implausible tha esl moun could have larg scale cee. Laymen. e observes, consider stich TMovon similar tothe ineffectual motion of ust in air By analogy with ‘Gos i pares ae very sll then although they cam be moved easly, there mavements are inconsequential. Having sct fv the lay inition ~ {hat tal motion is inetectve ~ Boyle then dcfends the opposite postion From meuphorw analogy in since 439 by hilfercntiating the analogy furthes. The inetfectivenes of dust particles, helaims the result of ther aloe to penetrate ther bodies and thereby foaffew those Bodies. He suggests that there ate other kinds of partes involved in local motion that are small enough to diffe through sod ‘objects, and that Hs this penettation that allows them to create large cts. (Thats, he arguce that dan isnot small enough!) He then proceeds {fo prseat instances ofthis kind of local motion. Inparagraph2, Boyle compares the aii of smal particestomove large smanes otha of ants to move their eggs. Although each ants much smaller ante mass of es, the sbity ofeach ano penetrate”the mass nd move one egg eases the ene mass of egg o be displaced. This example ‘onforms well othe peincples of analogizng (Table 20.1) There are cleat ‘e-toone corespondences, based not on characteristics of individual Jets bat on relations between the objets, as shown in Figure 203. For ‘Example, Hoyle docs nt suggest tha the corpuscles involved noc motion Slike ats ~ they ate nt tsi organisms, they do not haves legs, ands forth~ nor does he suggest that particles of matter are white oF sot or otherwise eggike. The only eequited matches ae forthe eave sizeof the fan the exe mass, andthe cpg. The important commonality 3 strsctarat tne: namely tat very Lange numbers ean compensate fra very great size dinadvantage, provided that penetration ofthe lager by the smaller can ‘occur Under these circumstances, many small boiesin motion can cary of, ‘ch larger body "The remining sections provide several additional analogous examples of the eects of local motion. For example, in paragraph 3. Boyle ets the example of wind passing through (penetrating) a tree, blowing off leaves nd breaking branches. In paragraph 6, Boyle presents the effects offre on 2 kof blae as an instanceof loa motion. He perceives fire ak comprred ‘of many small parties and explains the heating of metal in tums ofthe Invasion of igneous parties into the metal, with the result thatthe corps: ‘es of metal themselves hecome "feel agitated” andthe blade becomes hot The remaining two paragraphs, which describe "animal spins” and the contraction of tape. respectively, make analogous points. Bose ob Serves that although animal spi may’ Be minute enovgh to be inisibe they are capable of propelling large animals such as elephants. He de Scabcs seeing hemp shrink in moist weather, ad tates that the “ayeous Sand other humid particles, swimming inthe a, entering the pores ofthe hemp in great numbers, were able to make it shrink, though a weight of fit. sity ar even more pounds of lead were tied atthe end 10 hinder its ontraction" Table 22 shows the corespondences aross Boye’ set of examples. Boye syle of anslogiving is very diferent from Carmo’. Rather than ‘éveling on One pai of examples, carefully explain the erica common Felton sirvture, he uses + rapid succession of analogies and examples 40 DEDRE GENTNER AND MICHAEL JEZIORSKI ier pare) serps ore Rope “Analogs a cm Fiore 20.3, Boyle's analogy: the common relational sist or as moving ‘Table 202, Am overview of Boyle's sels of analogies concerning es and wind owing ees to demonstrate a central principle. The implicit message i that if all these henomenaeecer, the model that summarizes them must be plausible ce, not Simtanty between these examples they relate to one another by Vue of Fea bimom abstraction, They can be compared with one anotber {0 reveal an abetract model of local motion. 146) DEDRE GENTNER AND MICHAEL JEZIORSKI Despite these differences, Boye’ use of analogy falls roughly within the moder standards discussed in Table 201. In cach of his analogies, the ‘bject ae placed in one-to-one correspondence. Object atributes a ‘carded sin the wind analogy Indeed the sheer varity of the examples ‘ietualyguasantes that any specie objet characteristics wil cancel out fare about common relations stems, The complenity ofthe analogies i ot pret ~ they are no so deep SS Carts for example bu thin part due tote state of knowledge ‘ofthe subject matter. Boyles pont was to establish that the motion of ‘many smal particles ean combine to produce powerful sible effects, and further, thatthe requirement for this o oceur is that the smaller particles penetrate the larger matter Boyle demonstrates that this system of com ‘monalties hols throughout these examples. Final inspite ofthe large umber of examples, there are no mixed anaogics oF between analog Felations; each example stands on it own as a separate instantiation ofthe ‘relational structure, Carnot and Boyle: A summary [At fit lance, Boyle and Carnot seem to difer rather sharply in thee use ‘of analogy. Carnot uses one analogy, explaining it precisely and then going. ‘on to we the principles in further inferencing. Boyle, in contrast, offers ‘whole family of analogies, one after the other. It could be maintained that this sustained analysis marks Carmot as a more modern analogize than Boyle Yet despite thie syst dflerences, both Boyle and Carnot ate cestentally modern in their view of what consitutes a sound analogy. ‘The alchemists ‘We have moved back in ime frm Carnot (1795-1112) 9 Boye (1627~ 1601). So far the comparisons we have considered conform othe modern sotion of sound analogy, Now we move Sack sl further, tothe work of {he alchemists, and analy the forms of similarity they used in making tir predictions. The alcbemists were enthusiastic in their embrace of Similarity, But a me shal Se, their sense of how to use simian difered ‘markedly fom the modern sense.” “Alchemy grew ou ofthe fusion of Egyptian chemistry with Greek theory in about 3001. continued in Persia after about SO. and entered the European sphere aguin aftr the fist crsades atthe end ofthe eleventh ‘century no Irwast dominant force in weer science, or prescienee, until the seventeenth century 4.0. Alhough there were many variants, there trere some common themes * Based on certain works of Pato and Aris tote, alchemical thought postulated a pinata source ofall earthy mat ter This Fint Mater was manifested in small number of primary Prom metaphor to analogy in scence 03 ss. we oi am ad wae Figure 204 Schematico the dct of four elements ad our Quai. elements ~ ite, aif, water, and earth ~ each of which combined to of the ‘primary qualities, ho, cold, wet, and dry. Fre was hot and dy, water was {01d and Wet, and Soon (sce Figure 20.4). A Iransmutation could occur sf the proportions ofthe qualities changed: for example, fie (ht and dy) ‘could be change into earth specially ito seh) by losing heat o become old and dry. The alchemists were particularly interested in transmutations ‘of metals, copoly the transmutation of base metals nto gold often with the help ‘of & hypothesized catalyst known a5 the Philosopher's Sone (Redgrove. 1922) Besides bringing wealth achieving such a ransmuation would validate he theory. "The alchemists were pecrless in their enthusiasm for analogy and meta [hor Their comparisons were numerous and striking. Metals were offen ‘eld to consist of two components: mercury, which was fer, active, and ‘male, and sulphur, which was watery, passive, and female. Ths the cof ation of two meals could he viewed as mariage (Taylor, 1949. This ‘male-female ison was extremely nfvenil with the ation of third Principle. t formed the ia prima of mereut. sulphur, and Slt, which Paracas and other sntcenth-centary alchemists held to underlie ll mat ler, Metals were aso compared to heavenly bodies oF 10 mythological Figures, s dncussed below. Sill other metaphors were taken from animals land plants. The eagle wus used to convey vlatlty, Sal ammonia, for example, was called aqua corlestis (heavenly eagle). In another metaphor, the raven, the wun, and the eagle stood fr earth water, and ar (Cros land, 1978, p. 16), 461 DEDRE GENTNER AND MICHAEL JEZIORSKI [A conta comparison was & macrocosm-microcosm analogy (or meta hor) by which man (the microcosm) was likened tothe satral word For example, it was said that copper, ike # human being has 2 spr, a Soul ands body. with the spit being the tincture (Crosland. 1978. p 13) ina related vein, metal were compared with human states of health, “Thus, eld corresponded to'a man in perfect heath and silver "leprous old” (Crosland, 1978, p. 15). In another analogy between the heavenly Snd the earthly plancs, some alchemists counted twelve processes neces Sanyo produce old from base meta (caleiation, solution, separation, Conjunction, puttefaction, et), corresponding tothe twelve signs ofthe oda Others counted seven processes, corresponding 10 the seven days ‘treatin and tothe seven planets, each of which was held to generate its spel metal n the earth (Cavendish, 1967, p19). The importance of the mierocosm-macrocoem metaphor stemmed partly om the fat that Sichemy took as is domain the spinal wold as well 35 the physial Sword A cenisal Beli was thatthe purcation of base metals into gokd Sos analogous tothe sprtual paifeaion of man (Redgrove, 192). This “nslogy could be run i ether direction, 2 that "some men pursued the ‘cnewal and gloiication of matter, guiding themscives by this analog fibers the tenewal and glorifeation of man, using the sare analogy” (CTayor, 1988, p. 14). "The alchemist willingness to heed similarities ofall kinds derived in part rom their Beit that all things above and below are connected, and {ht simrity and metaphor are guides to those connections. This dopma tron coified inthe medieval doctrine of signatures, the sense that “I is Through similitudes that the otherwise occult parenthood between things i Imansted end every sublunar body bears the 1acesof tht parenthood Impressed oni ea signatute” (Eco, 1990, p24). Vickers (198) suggests thatthe alchemists invested analogy with extraordinary importance, even ‘equating analogy with Wnty. ‘What were the rules that governed the alchemists! use of analogy and Imotaphor? We begin with family of comparsons that wed asthe base Alomain the egg othe seed, and asthe target domain ether (or both) the Principles of mater ofthe components of human being, The eae “The egg was used widely in alchemical analogies. Taken a5 2 whole, the ee ould symbolize the limitless of the universe, or infinity ise. The Philosophers Stone was often called an epg (Cavendish, 1967; Stillman, 1940), The egg could alo be divided into components. For example. Sil: rman (1950) notes that the shell, skin, white, and yolk of the ege were Thought to be analogous to the four metals involved in transmutation: Tead. and iron (although no pairings were specied hetween From metaphor o analogy in science 6s ‘Table 203, Aichemicol analogies ofthe Ege icons aralony Furthermore: Domain TheEga Components Elemenis Malefemale Primary sie Sect erie Patsophers Sone Nemberof Gy a Gl o ‘ements Comespondenes Whee Sout Sulphur Male Bee You Spine Merry Malefemae AuWater Stet Body Sai” Female Earth Caren) the components and the metas). Several additonal corespondences are !pparent inthe following excep. from the manuscript of St- Mark's the tenth or eleventh century (copied in 1478, translated from Berthoet’s {1887} Colecion dex Ancient Alchemistes Gres).* The “egg” described is fn fact the Philosopher's Stone: Nomenclature of the Ep, Ths the mystery of the a ‘Tita ben snd tht he ee competed ofthe four clements, case tis the image of te word and cvs nl he oa elements Hailed ako the “Shoe ch suse the moon tour, "sone which sot aston.” "siome of thecal" and “brain faster 2. The shel ofthe ep van clement i earth, cold and dy thas been called ‘append. The hte othe fe the water vie, he yew of {he epee euperne [elt tee portion re 2. These bas bance th cd a sel he skin whit and ts yellow {he fs ts oy prt the ous aqueous the beath a he a (tina, 18s pp. 170TH notation bracket aed) “Ths brief excerpt illustrates the Isl of analopzing daplayed by any alchemists, First the cpa compared 0 several diferent analogs. The use ‘of multiple analogs would nt in isl diferemtate this passage from the Work of Boyle, However, what i distinctive here i thatthe number of ‘components involved inthe correspondence varies trom analog to analog, {8 shown in Figure 205. (See ao Table 20 3.) The fst paragraph maps the “egg fist onto the four elements and then onto a series of single ets (e.g, "the stone which nota stone." the “bran of alabaster"). In paragraphs 3 and 3. the components ofthe epg ate successively compared {o the four elemens of sneent Greck philosophy (eam, water. at, and fre) the layers of & sed, and the aspects of & human being. These Imuhiple analogies are quite diferent from those of Boyle; the alchemist ‘We DEDRE GENTNER AND MICHAEL JEZIORSKI a Fu ciewens Fgwe 05, Object sorrespondencsin the ce nao. oes not attempt 10 delineate a common structure that hols cress the several systems (It would probably nt be possible vo do so.) ‘But mone stking ference arises when we consider the ise of one tine mupping, Figure 20S shows the object cortespondences used the shove st of analogies. tis apparent that achieving one-to-one corespon ‘ence » not of primary concern For example, as Figure 2051 shows, the fobjcet coetespondenees forthe analogy between the egg and the four ele ‘ments of matter are such tha the element af ar must ether be omit ot ‘ke placed in correspondence witha previously sed elemento the ce. Yelling a mapping of four abject ont fe. As Figure 2.5b shws the ‘mapping from the egg tothe four divisions ofthe seed (or snpcts of & human being) sko nt aneto-ane, since bath the white and the yellow pats ofthe eg correspond othe flesh, Thus Figure 20S shows 2 5» 4 ‘rapping, whetess Figure 20a shows 44 5 mapping. When sn sly Sickh tao oF more competing mappings, the modern practice ta choose fetwoen them oF to note that thete ar alternative interpretations I ens taste alchemists resolved the tension by combining both interpretations ino a fased whole. This flute to preserve a one-to-one correspondence Aiferciats his reasoning sharply from that of modern scents “The alchemists invested metaphor with great importance. The heli that ‘metipioes reveal essemial categories, 3s dscused by Glucksherg and Kejar (190), ay taken Io extieme lengths. For example, an atactve spect ofthe fg was that twas recognized a8 something via as symbolic ‘of beginning. Any system that could be related tothe eng was imbued From metaphor to analogy in cence 61 witha similar significance. When some alchemists shifted from the Greek Theory of four elements tothe theory thal thee “principles” ~ usually defines sulphur, mercury, and salt (oF scm) ~ composed ll mater tra all possible To use the exe analogy ‘As on cag composed of hee tgs th sl the white, ad the ot, our ‘Phiccopiea! Egy composed of» ay, sl ad apr. Yt i tinh OU one hig lame ewer ges ty wet ad unity in ity = sr Me fry. tad Arsenic (Drees, quoted sn Huon Sones, 96, p. bac ae ss Here the alchemist suggests a series of parallel analogies among the ee the Philosophers Stonec man, and matter and gives the objet correspon ‘ences among the (now three) parts ofthe epg, the three aspects of man land the three principles of matter Othe alchemists extended the analogy, ‘mapping the three arts ofthe cpg onto the male-female principles snd the four primary qualities, xs shown in Table 203 (Cavendish, 1967, p. 163). Anologies wth symbols ‘Another striking aspect of alchemical snaloizing was a willingness to se Similarities between symbols and thei referents. Vickers (984) suggests {hat the alchemists were influenced by the occult tration in which the ‘word or symbol does not merely stand for ts referent, but identical i ‘sence. This nt led t belie inthe causal powers of words and ‘ther symbols. For example, Kriegsmann (1665) offered an analysis of the properties ofthe three principles = sulphur mercury and salt in terms of their symbology (Crosind, 1978, p. 253) Given the suppostons that a aight line denoted earth, a triangle fre. semicircle ar, ands eice Wate, he attempted 10 analyze the symbol or suphur, mercury, and salt Into the four base elementary symbols. Even at late as 1727, Boerhaave ‘analyz copper by the following chai of though. First, he noted that 3 ‘cl indicates perfection, whereas 4 ross denotes something sharp and ‘orosive It follows that pods symbolized sy ctl, and singe copper i Symboized by cic plas eros, cam be scen tobe “intimately OSI” bt combined with some stu, shit. an corrosive material That thi system was btieved to he predictive. not merely convenicaly ionic, am be seen i this pasage by Boerhaave concerning ion (symbolized by a cle plas an outward arrow) "hah 0 sims Gl; but ht st has with ea de he sharp and oro hooph wth ut hal te degree Acnny te formes yo set ‘Hatt hsb ll he gn tat empress ht quay Indced aoa the trier opinion ofthe slp hat the Aum rram or Pionphorum dae pe ‘once in ronan tat here tertore we mun sek or metline Mens, {ind notin Golda (Crosland. 1978, p21), 468 DEDRE GENTNER AND MICHAEL JEZIORSKt ‘That metaphors between symbols and material objects were el to be informative a marked difference fom the preset aesthetic Kis elated to the doctrine of signatures the sense tht similarity vitally guarantces Siznicant connections. As Ea (199), p. 24) puts Hn his desrption of Renaisance hermelicim, “Te an inp ot oly ha he sear ane Known hough the sar Dat ‘es hat Irom sanity ss eveying an Be conected with eveything Paractous Paracelsus (Theophrastus Bombastus von Hohenheim, 1493-1541) wax a leading chemist of the ssteeth century and a vigorous proponent ofthe value of empirical observation as opposed to received dogma, Despite this Father modern spin, his ise of the analogy remains distinctly diferent {tor madern usage. Here, he describes how gold and silver ean be made: ‘Some ome maya, wht then, he short and eat way whereby Sl an Lana for fe male! The awe this fier you have made hewn. che pr Sh, wih os Me to ton ovr the earth, pace om ale anes so tha the Drtion a Lune maybe the smallest Lt ltr at heaven or Sitar es entiely Srorpeorcd. Then al howe planets wil remain dead sith the old corre fren: having mean obined anther new perfec and incr bay ‘Thotod ibe sito ener, Fomit thes planets span receives Moy ame {nd teem betes Take ht Bod fom the ead earth. Keep es Se ad il reyou howe the At clest anette. Ifyou onl unerstand tt wl Teeter ht shold te Kept conesed apd ot made poi. (quote i alle Here So and Lun (the sun and moon, respectively) signily god and ve. and othr mals nthe eepe ate represented bythe othe planes. accor ing oa widely used system of alchemical analogies (sce helow).Paracesus «docs no dtl the objet corespondences between the to domains. nor {doe he explain how an action i one domain paral action inthe fother The mappings and the theoretical has for the procedure are lft trstted. Indeed, if not always clear which actual metas are eine ferred wo, For example, o what do “earth” and “all thos planets el Dace heaven. or the phere of Saturn” refer ttn? Iso, isthe ia "sprit of heaven derived from the process azo tin? This ast seems implausible, Since the gu is to produce old andor silver: yet if the nal "spinit of heaven is gl or sver then what about the intial “heaven”? Here. and ko inthe pusages below, the alchemists go beyond similarity and into 2 wider set af what Gibbs ihe volume). calls “topes,” including many i ‘ances of metonymy. “This passe, though it exemplies the different rules of analopizing From mewphor to analogy in scence 30 Table 24, The alchemical sytem of corespondences among planes. metal, and color Sn Gold Gold yeow Mercury Oviksver Gray seta Mars” fon ‘Sion tad Based on Cavendish 967. p25 among the alchemists, also rises questions cone these differences, Paracelsus makes it lear that clarity isnot his intention, “The secretive nature ofthe enterprise, the fact that it was Flt necessary hide rere from the common public and from competitors, could have led to the ambiguity ofthe writing. sit possible that his ambiguity shielded a ‘et of informative analogies? To answer this question, we must look more ‘dosly at the system of comparisons that supported this reasoning. The system of correspondences Since the goal of many alchemists wast trasfrmn base metals into higher metals (old oF ser). metals held an important pace in alchemical anal ‘fy and metaphor AS illustrated above, metals hgured in analogies with the principles of matter and withthe sipets of human beings, and the ‘tansmutation of ase metals into gold Was fet 10 be analogous tthe Spiritual purification of man. A forthe set of corespondences existed between metals planes, and colors. as shown in Table 204." (Tie table ad much of the surrounding explication ae used on Cavendish's discus: sion Cavendish, 1967p. 261) y "The first thing thal Sekes us aout this stem isthe importance of surface similarity ~ that i, common color ~ in determining the correspon ences. The Sun, the metal gold, andthe colo gold are linked By a com- ton color as are the Moon, the metal sive. and the color white. But an ‘equally striking aspect ofthis system i hat this commonality f= not uni {ormly maintained. The basso the comparison shit rom one part ofthe system to another. For example, unlike the two tras just mentioned, the Joptertinyblue wad snot ently based on common color. Instead, bine {the color of the sky) s matched 1 Jupiter hecause Jupter wa lord ofthe Sky. And although the match between Jupiter andin may be a color match, a #4M— DEDRE GENTNER AND MICHAEL JEZIORSKL ‘sed on the planet's silvery appearance. may also have been bused on a fncient belie tht the sky was made oi, Ths theo snare that Tire inthe correspondences changes from one row o enum 0 another “The st of etions that inka the ows of this table was remarkably ch and diverse, a ilstrated in this session by Cavendish (1967, p27) Lead the dates shea the et, ws natu sige oSatorn,the ‘Greet an sweat moving planet, which radges hes hugh om path ‘ound the su In he cosa Stor the lathes plane thes he fuer of ie, an he lord of det The analogy between death a ght was ‘town very carly lacks the colour of nigh andthe our ineaaly aad “The chain of eomncetions between Saturn and black isa cae in poi Satur s the lord of death, death similar o night, te col of night sky ‘black, and bsehness symboits death Thus chain ss made Between the Planet Saturn and the color lack. This rich metonymic chan i quite Sifferent from the simple “color of X and ¥ 82° relation that holds for ‘Sunigoldellow and Moonfsvertuhite. The heterogeneity of matches tht ould apply within a single tabular system contrasts sharply with the mod fmmaestheic Th preference for siuctral consistency and stmt n ‘modern anaivgy would dictate that Wendl relations should had across the system: thi, we would expect 1 ind Mooa:whit :Sun:yellow ::Jpitersblue::Saturn-lack {nthe alchemical system there sno such requirement: no Wo ows need have the same sc of relational inks “This example slistrate farther point of dierence: the alchemists’ sys tem of sorrespnncnces voles the mo extraneous elms principle that eros coancctis of all Kinds enter into the analogics. Fr instance bac, lead and Satur are all linked through the chain descrbd owe but the match between ad and Saturn was improved by the fact tha oth are slow and heavy Saturn movesslowly init omit und was thetlore though {5 musive (heavy): lead was known 0 be a dense (heavy) meta, which ‘woul! presmably move slowly. This complex web of sinatitics wast improve the ste though it could nt be applied uniform "As another smtince of the prolife and Reterogencous nature of this ‘elatinal stem, comider the match between Mats and re. Cavendinh (0967, p27) mats tha is Based on several chains of ssciations: MAS looks ed; Maes sth gd of wa, wa iy socatd with Mans ad bod is red faces ae pate red in wa and Mats shel ae violent ‘energy and aetiy and reds he color symbolizing energy. These multiple Imetonymic pats svengthened the analogical connection between Mars ind ed “Tis discussion ofthe alchemists” system of correspondences iustates From metaphor to analogy in sience n some marked differences inthe rules of the game, The alchemists were not moved by the madera "no extraneous relations” rue They aecepiod mined metaphors and fused intrprstatins of single mctapa. Tn Ihe current thetic. once a parle set of relations established. adding local tela ins that hold only for afew eases docs a improve the analogy. But fr alchemists, more Was always belter. A ich set ntrtlationship, however iosycratic, was felt strengthen the similarity bond Comparison berween the alchemists and modern sini “The alchemists embraced similarity i al ts forms in reasoning aout the ratural wold. Yer the examples we have considered how marked devia tion from the current tye of analogical reatoning is summarized in Table 20.1. Are there then historical differences in analogical reasoning? Before

Vous aimerez peut-être aussi