802 On the Analytical and Syvthetkal Modes, &c. 



to give a substance, if I niav so express myself, to tho?e 

 fugitive deceptions, by creating new njodcs to design llicni, 

 or by combining, in a different manner, words already 

 Jinown; and we afterwards blindly treat upon these hypo' 

 thetical abstractions : but as they are exposed to the founda- 

 tion, there comes a time when the absurdity of the conse- 

 quences derived from them shows their want of solidity. We 

 then seek a new path, and in general find one formerly 

 trodden, in which we again lose ourselves. 



It is probable we commenced by perceiving, in our sensa^ 

 tions, the origin of our ideas; but being obliged to clas?, 

 divide, distinguish, and abstract the dsffereiit circumstances 

 which the acquired ideas presented, we lost ourselves in the 

 catagories, and all the abstractions which are attendant upon 

 them. The discoveries in physics, by giving a real .subject 

 for reasoning, opened the eves to the abuse which had been 

 made of it. The road traced by Newton in the third book 

 of the Principia, could not be restrained solely to the objects 

 to which it had been applied. The eclat of the discoveries 

 which he made by following it, excited in those who cul- 

 tivated the scii'.nces an emulation which soon produced 

 improvements in metaphvsics. It is to be granted that they 

 have gained much bv tlus revolution ; but there may possi- 

 bly come a time when its progress will be stopped; and by 

 comparing what it has lost on one hand and gained cm the 

 other, we shall know that this alone, amongst all the sci- 

 ences, is susceptible of a limited progress, and that there 

 exists in the theory of the operations of the understanding 

 a point which cannot be gone beyond. 



Let Us turn, therefore, to the physical sciences, which 

 promise numerous and useful discoveries, all the activity of 

 our understandings ; and the theory of probabilities, by be- 

 coming familiar to all those who cultivate the moral sci- 

 ences, will give solid bases to those parts of our knowledge 

 which are not capable of being ramilied to a small number 

 <>f abstract notions and complete ideas. 



XXXV. AffinitT/ 



