ORGANIZED SCIENCE 301 



The referees will submit to the editor reports embodying specific 

 criticisms and a general recommendation to publish or reject the 

 manuscript. Most often the editor then returns to the author his 

 manuscript, the referees' comments (ordinarily unsigned), and any 

 comments the editor may himself wish to add. The author is expected 

 either to accept and act upon the criticisms or else to show how, in 

 his opinion, they are unjustified. The editor will then himself review 

 the situation. In difficult cases he may call for reports from additional 

 referees (including, perhaps, some suggested by the author), or for 

 further changes or justifications to be made by the author. When at 

 last reached, the final decision is then less a matter of individual fiat 

 than the consensus of an expert group. 



This is the pattern of scientific administration generally. The 

 award of a doctoral degree will ordinarily depend on the judgment 

 of an examining committee drawn from the local academic commu- 

 nity, but consultation with an outside authority is not uncommon. 

 Appointments to important posts are made similarly, on the recom- 

 mendations of local committees— which will, however, base their 

 decisions on nominations and testimonials solicited from a much 

 wider group of those competent to appraise the candidates and their 

 works. Committees concerned with award of prizes, fellowships, re- 

 search grants, and so forth function in much the same way. In no 

 one case is there a general consultation of the entire electorate: the 

 "voters" are always, for the purpose in hand, a relatively small elite 

 group of the specially qualified. But the sum of all administrative 

 judgments is, effectively, constituted from some opinions of practi- 

 cally everybody competent to judge anything— and, as Polanyi notes, 

 these judgments generally command remarkably wide approval. 



Two scientists acting unknown to each other as referees for the publi- 

 cation of one paper usually agree about its approximate value. Two 

 referees reporting independently on an application for a higher degree 

 rarely diverge greatly. Hundreds of published scientific papers pass 

 review of thousands of scientific readers before any of them finds 

 reason to protest against the insufficient standard of a paper. 



This impressive feat of communal administration is of course 

 founded on a communal consensus of belief no less impressive. Scien- 

 tists take it quite for granted that they should agree. Polanyi properly 

 stresses that: 



