turnsouttobeeitherhopelessormeaningless。Milldoesnotdirectlyassertthatall’ontological’speculationsaremerelycobwebsofthebrain。Buthetriestoshowthat，whatevertheymaybe，theyarestrictlyirrelevantinreasoning。Allmetaphysiciansareexpectedtogranthimcertainpostulates。
Theseoncegranted，hewillbeabletoaccountforthewholestructureofknowledge。’Intuitions，’transcendentalspeculations，andontologywillthenbedeprivedofthewholeconditionsunderwhichtheythrive。Idonotnowassert，hevirtuallysays，thatyourdoctrineiswrong，butIshallshowthatitisthrownaway。Itisapretenceofexplainingsomethingwhichliesaltogetherbeyondthelimitsofrealknowledge，andthereforeadmitsofnoexplanation。
Millstartsfromtheclassificationgiveninoldlogicaltextbooks，towhich，differentasarehisconclusions，heattachedaveryhighvalue。10*Theschoolmenhadbytheirelaborateacutenessestablishedawholesystemoflogicaldistinctionsanddefinitionswhicharebothimportantandaccurate，howeversteriletheinquiriesinwhichtheywereused。
Themachinerywasexcellent，thoughitscontriversforgotthatamillcannotgrindoutflourifyouputinnograin。Millbeginsaccordinglybyclassifyingthevariouskindsofwordsinthelightaffordedbypreviouslogicalsystems。
Heistogiveatheoryofproof。Thatwhichistobeprovedisaproposition；andapropositiondealswithnames，andmoreoverwiththenamesof’things，’notmerelywiththenames’ofourideasofthings。’11*That，insomesense，reasoninghastodowiththingsisofcoursehisessentialprinciple；andtheproblemconsequentlyarises，Whatareempirical’things’？12*