メタアナリシス的全体論

― コリンズに応えて ―

抄録

Scientific experiments are often in disputes. Their qualities are challenged. Doubts are casted if their results are proper or fake. Under such polemical circumstances, how did scientists reach agreements on the qualities of experiments and their proper outcomes? By appealing to his idea of ‘experimenters' regress’, Harry Collins rejected an accepted answer to this question, and gave his own. This paper proposes a counterexample to his views that is taken from the ‘measurement networking’ for standard values of fundamental physical constants. Examinations of a sort of meta-analysis; i.e. statistical technique that is conventionally used in the networking, reveal that, in contrast to Collins' answer, the agreements at issues can be attained only in a holistic way and simultaneously within the network.