March i6, 191 i] 



NATURE 



95 



I 



with their source only by a distinct mental effort. 'J'he 

 wonders of practical science have been recited so often 

 that their reiteration has become tedious, and we no longer 

 feel strongly impelled to felicitate mankind on the parlour 

 match, the telephone, and the antitoxins, although we 

 indulge at present in an unsubdued, excited anticipation 

 of wonders to come, especially in the domain of medicine. 

 Are we not all on the watch for the announcement of the 

 cure for cancer, and vaguely for other new and astound- 

 ing reliefs from disease ? Such concentration of interest 

 upon novel practical results is not wholly favourable to 

 science. 



It is true that a large amount of investigation is going 

 on which aims to secure immediate practical results. In 

 chemistry and medicine especially, the activity in the work 

 of applied science is very great. This condition gives a 

 powerful fresh reason for defending pure, abstruse science. 

 Applied science always has been, is now, and probably 

 always will be, distinctly subsidiary to pure science. The 

 final justification of all scientific research is undoubtedly 

 the power it creates for the use of mankind, but the 

 power must be created before it can be used. A little study 

 of the history of science should suffice to convince any 

 reasonable mind that the command we possess to-day over 

 nature is due to the labours of men who have almost 

 invariably pursued knowledge with a pure devotion un- 

 contaminated by any worship of usefulness. These devoted 

 idealists have gathered the varied mighty harvests by 

 which all men have profited, but the debt of gratitude to 

 them is unpaid. 



The pursuit of abstruse science needs to be encouraged. 

 It is insufficiently esteemed. This doctrine ought to be 

 emphasised on all suitable occasions, but especially before 

 the section of ex{>erimental medicine. The people cry for 

 relief from sickness, and their demand for prompt, useful 

 discoveries is so urgent that there is danger in it, since it 

 tempts medical investigators away from the fundamental 

 inquiries, which, answered, will give great results, and 

 seduces them to work exclusively at secondary problems, 

 from the solution of which quicker, but smaller, results 

 may be expected. Pure science is broad ; it embraces all. 

 Applied science is a congery of fragments, of isolated 

 problems, which lack cohesion and are without any 

 necessary connection with one another. It is easy to 

 understand why students of applied science have seldom 

 made great discoveries. 



In fact, scientific knowledge will not be compelled. We 

 have to take what knowledge we can get, and by no 

 means can we get always what knowledge we want. 

 Pure science adapts its undertakings to these rigid con- 

 ditions, and works where the opportunity is best — not so 

 applied science. " 



Compared with the growth of science, the shiftings of 

 Governments are minor events. Until it is clearly realised 

 that the gravest crime of the French Revolution was not 

 the execution of the king, but the execution of Lavoisier, 

 there is no right measure of values, for Lavoisier was one 

 of the three or four greatest men France has produced. 



Since pure science has been pre-eminent in the past, not 

 only in furnishing useful knowledge, but also as a chief 

 foundation of human progress, and is likely long to con- 

 tinue equally pre-eminent, it is well worth while to study 

 the general principles by which original research is guided. 

 No previous definite study of these principles is known to 

 me, although I have searched not a little to find one. All 

 that I have been able to discover are treatises on logic, 

 the reading of which, most active investigators would, I 

 fear, find tedious and unprofitable rather than helpful and 

 [inspiring. 



It is my belief that the logical work of scientific men 

 lis usually well done, and is the part of their work which 

 is least faulty. The diflficulties and the majority of failures 

 are due, it seems to me, to two chief causes, the first, in- 

 adequate determination of the premises, the second, 

 exaggerated confidence in the conclusions. If I am right, 

 the method of science is the result of the effort to get rid 

 of these two causes of error. 



We must recognise in starting that the expression " the 

 method of science" means more than "logic," being far 

 more comprehensive when rightlv defined. We cannot 

 alter the fundamental conditions of knowledge, for we are 

 still unable to add new senses or improve the brain — ' 

 NO. 2159, VOL. 86] 



although eugenics dreams of a future with such possibili- 

 ties — nor can we change the nature of the phenomena. 

 The same fundamental resources are available for daily 

 life and for science. We must be clear in our minds on 

 this point, in order to comprehend that the fundamental 

 distinction of the scientific method is its accuracy. Such 

 being the case, a broad examination of the method of 

 science reduces itself to the study of the general principles 

 of securing accuracy. 



If you will examine frankly your own opinions and 

 those of your acquaintance, you will, it may be presumed, 

 quickly acknowledge that many, perhaps most, of the 

 opinions are not of scientific accuracy. On the contrary, 

 they are, to a large e.xtent, mental habits and the result 

 of the summation and averaging of impressions. We get 

 along in ordinary life satisfactorily enough with opinions 

 thus formed by summation. Most human opinions, even 

 when they are merely imitative, originate in this way, 

 and are correspondingly untrustworthy. If we seek to 

 explain the fallibility of ordinary opinions and testimony, 

 must we not attribute it to the absence of the detailed 

 evidence and the .consequent impossibility of verifying the 

 testimony ? 



We are thus led to recognise the preservation of the 

 evidence as the fundamental ' characteristic of scientific 

 work, by which it differs radically from the practice of 

 ordinary life. I venture, accordingly, to define the method 

 of science as the art of making durable, trustworthy 

 records of natural phenomena. The definition may seem 

 at first narrow and insufTicient, but I hope to convince 

 you that it is so comprehensive as to be not only adequate, 

 but also almost complete. 



All science is constructed out of the personal knowledge 

 of individual men. Science is merely the collated record 

 of what single individuals have discovered. According!} , 

 we must consider, first, the way in which the individual 

 knowledges are recorded and collated. The process begins, 

 of course, with the publications of the special scientific 

 memoir in which the investigator records his original 

 observations and makes known his conclusions. 



It is interesting to note that our present standards for 

 original memoirs have developed gradually. In Harvey's 

 essay on the circulation of the blood, published in 1628, 

 there are no precise data as to his observations. The 

 author does not think it necessary to specify how he has 

 laid bare the heart or how often he has repeated his 

 observations. His descriptions of the beating heart are 

 vividly realistic. He writes with conviction and authority. 

 The reader is compelled to believe him. Harvey, however, 

 does not provide information to facilitate repetition of his 

 work — he offers little aid towards the verification of his 

 results. 



In a contemporary article we expect a presentation of 

 all the data necessary to render subsequent verification by 

 other observers possible. We further expect clear informa- 

 tion as to the amount of material on which the observa- 

 tions were made, or the number of experiments on which 

 the work is based. In other words, a modern investigator 

 will hardly receive consideration for his researches unless 

 he furnishes every aid he can to facilitate criticising and 

 testing his results. This severe standard has been onlv 

 gradually evolved, but is now stringently enforced in all 

 departments of science, and is the response in our practice 

 to our need of eliminating the purely personal factor. It 

 would be advantageous if scientific authors, generally, 

 viewed the obligation of providing for verification as an 

 even more serious duty than it is esteemed at present. It 

 might, indeed, be a wholesome practice to demand that 

 every scientific article should contain a special section or 

 paragraph on the means of verifying the result, for verifi- 

 cation by Fachgenossen is second in importance only to 

 discovery in the progress of science. 



The conditions of scientific progress have changed 

 greatly, though verv frndually. Two hundred years ago 

 the number of active investigators was small. This year 

 there are at least ten thousand men of substantial ability 

 carrying on original researches, consequently each theme 

 is being worked at by several men, and the final out- 

 come is the consequence of collaboration, which is non.- 

 the less actual and effectual because it is unorganised, and 

 is usually not formally desi£?nated as collaboration. 



These conditions have rendered great men somewhat less 



