THE MECHANICAL
FALLACY 107
his canvas orcompounds his
pigments, butwe do
take
delightintheadjustmentofsupporttoload,and
thrust to thrust. Itisno doubtlegitimate toadd
decorativedetail tothesefunctionalelements
;
theymay be enriched
by colour orcarving
;but ourpleasureinthecolourandthecarvingwillbepleasurein painting
or sculpture
;our specificallyarchitec-turalpleasurewillbeinthefunctionsofthestructuralelements
themselves. Itisinthisvividconstructivesignificance ofcolumnsandarches thattheirarchi-tecturalbeauty lies, andnotsimplyin theircolourandshape,assuch,andsofarasthestructuralvaluesareabsent,andthe eyeis merelycharmedbyotherqualities, it is nolonger architectural beautythatweenjoy. Only,thesefunctional elementsmust
bevividly
expressed,and, ifnecessary, expressed withemphasisandexaggeration. Thesupporting
membersmust assure us of theirsupport. Thus, the Doricor the Romanesquemassiveness, whileit wasin asense bad science, was good art;yet its beautywasnone thelessessentiallystructural.Thus, theprintingpressortherailwaystationwillnowappro-
priatelyfall outside our definitionbecause,althoughtruthfullyandperfectlyconstructed,andfitfortheirfunctions, theydo notvividlyenoughexpress.whatthosefunctions are,northeir fitnessforperformingthem. Structurallyperfect, theyare still structu-rallyunbeautiful. Ontheotherhand,
thearchesand