580 



SCIENCE 



[N. S. Vol. XXXIII. No. 850 



two groups. The metliod of investigation 

 correspondingly changes since social science 

 can not readily get at causes by experiment 

 but must begin with observation of results 

 and work back towards causes by indirect 

 methods. This method has lately been re- 

 named pragmatism and involves a judgment 

 of causes through their effects. Consequences 

 are open to observation; causes are not. A 

 cause must therefore be judged by its observed 

 effects. The leading exponent of this method 

 was the late Professor James. I do not wish 

 to defend his arguments, but to call attention 

 to his method. His observations — those on 

 which truth depends — are psychic phenomena. 

 We may therefore call this method psychic 

 pragmatism, for the satisfaction that the per- 

 ception of truth affords becomes its test. There 

 is, however, another method that goes out 

 from consequences just as Professor James 

 does but which uses objective social tests in- 

 stead of psychic tests. Social pragmatism 

 uses marks to visualize wholes, but the marks 

 are the objective social consequences of acts 

 which can be measured and verified. Social 

 consequences can be measured and through 

 their observation a steady advance is possible 

 in ways that will put social observation on a 

 par with physical measurements as a means of 

 developing science. 



If it can be agreed upon that observation 

 and experiment furnish the only basis upon 

 which investigations can rest the next sub- 

 ject of importance is the canons of reasoning. 

 Dr. Minot assumes that the reasoning of sci- 

 entists is seldom defective and that their main 

 errors are those of measurement. Scientists 

 however, are as liable to errors of logic as 

 other people and scientific method can not dis- 

 regard the laws of thought. They are the very 

 essence of good thinking and must therefore 

 be formulated. 



The first and primary rule is that only ob- 

 servations and experiments can be used as 

 premises in deductions. This seems an inno- 

 cent rule, but it involves more than at first 

 sight is apparent. All experiments are local 

 and specific in their conditions. There is no 

 such thing as a general experiment and hence 



premises derived from them are local definite 

 facts. The same is even more true of observa- 

 tions. They are made by individuals and no 

 one can extend his observations over more 

 than a local field. All valid data are there- 

 fore local and specific. Reasoning consists in 

 extending the scope of these premises to other 

 and broader fields. It follows from this that 

 there is no difference in kind between an in- 

 duction and a deduction. This distinction is 

 due to the well-deserved disrepute into which 

 the dogmatic assertions of theologians and 

 metaphysicians fell. Scientists wanted there- 

 fore to get a peculiar mode of reasoning that 

 would avoid these evils. In this they have 

 failed. So many sciences have become de- 

 ductive that the reasoning of scientists differ 

 in no essential respect from that of any other 

 group. The differences are in the premises 

 and in the verifications, not in the reasoning. 

 There is but one method of reasoning. Its 

 rules apply to all thought and to all subjects. 

 The same end that the distinction between 

 inductive and deductive thought is attained 

 by the second rule of good thinking. No 

 generalization should be used as a premise in 

 reasoning. A generalization is a result of 

 previous thought and is only an approximate 

 truth. Every new chain of reasoning must 

 go back to the original data in the form of 

 observations and experiments and be based on 

 them together with the new data obtained 

 since the original generalization was made.- 

 The chief violations of this rule are in social 

 science, but scientists are not free from this 

 error. They, like other people, form them- 

 selves into social groups and thus acquire 

 dogmas and prejudices that induce them to 

 use the generalizations of their group as 

 premises when they should confine themselves 

 to their data in the form of observations and 

 experiments. Sound reasoning always goes 

 from the local to the general. Universals are 

 made either by some social group imputing 

 value to a premise that serves their practical 

 needs or they are loose generalizations based 

 on imperfect data. Whatever their source, 

 they are unsafe premises and lead to wide- 

 spread popular errors. 



