INTRODUCTORY KEMABKS. 13 



credit which they pre-eminently deserve for having devoted 

 their lives to purely intellectual pursuits, and for having 

 thought, seldom frivolously, often profoundly, nothing can be 

 more difficult than to seize and apprehend the ideas of those 

 who reasoned from abstraction to abstraction who, although, 

 as we now believe, they must have depended upon observa- 

 tion for their first inductions, afterwards raised upon them 

 such a complex superstructure of syllogistic deductions, that, 

 without following the same paths, and tracing the same sinu- 

 osities which led them to their conclusions, such conclusions 

 are to us unintelligible. To think as another thought, we 

 must be placed in the same situation as he was placed : the 

 errors of commentators generally arise from their reasoning 

 upon the arguments of their text, either in blind obedience to 

 its dicta, without considering the circumstances under which 

 they were uttered, or in viewing the images presented to the 

 original writer from a different point to that from which he 

 viewed them. Experimental philosophy keeps in check the 

 errors both of a priori reasoning and of commentators, and, 

 at all events, prevents their becoming cumulative ; though 

 the theories or explanations of a fact be different, the fact 

 remains the same. It is, moreover, itself the exponent of its 

 discoverer's thought : the observation of known phenomena 

 has led him to elicit from nature the new phenomena : and, 

 though he may be wrong in his deductions from this after its 

 discovery, the reasonings which conducted him to it are them 

 selves valuable, and, having led from known to unknown 

 truths, can seldom be uninstructive. 



Very different views existed amongst the ancients as to the 

 aims to be pursued by physical investigation, and as to the objects 

 likely to be attained by it. I do not here mean the moral ob- 

 jects, such as the attainment of the summum lonum, &c. 

 but the acquisitions in knowledge which such investiga- 

 tions were likely to confer. Utility was one object in view, 

 and this was to some extent attained by the progress mde in 



