The clear distinction often has tactical defects, but logically it remains: in practical affairs, man speaks the language of applied science or that of rhetoric. If he plays the scientist, he follows certain rules of expression, such as "no prescription without diagnosis."

An engineer does not build a bridge without surveying the site; a physician does not prescribe medicine without examining the patient; a social scientist does not advise greater economy without investigating the problem for which economy is to be the remedy.

Does not probably had better read should not, for political scientists. They are prone to prescribe first and investigate later. A professor will say "delegate work" without reference to what work and whose work; he will say "more people should vote" without reference to the community at issue; he will urge the "merit system" on every type of government. Several explanations of this behavior may be advanced.

One is that social science is so advanced that it does not need the aid of empirical study of the unique case for which the cure is needed. But save in a few areas, there is no evidence of this perfection.

Another is that social is strong in general principles, weak on specifics. It knows that most delinquency comes from the environment, but not the source of specific cases of delinquency; therefore it will make sense in general but not in the particular. This is nonsense, since its big failures are as common as its small ones.

A third is that social science language makes it merely seem as if diagnosis has been omitted. Sometimes a man knows the case at hand but speaks of it in the abstract, as when, knowing that a bank in trouble needs more honest officers to get more depositors, he declares that all bankers need to be honest.

A fourth is that social scientists may not or cannot speak in a complicated, quantified language and therefore do not express different probabilities. They say, "City manager government is good for Pawtucket," without knowing anything about Pawtucket, and really mean, Sixty per cent of all towns could use the manager plan well," or "The manager plan is good for anyone who is rich and only interested in money and I use his preference system."

Still another is that the scientist is deluded concerning one or one or more of the above. That is, he doesn't know how bad the condition of his applied science is.

Also, social scientists may regard themselves as rhetoricians, guardian angels, as Authorities, as powers, as circus players, etc., and as such are all alike. They consider an applied-science approach a silly thing. They express themselves in the guise of scientists to achieve other goals.

Furthermore, they may be so regarded by their audiences and cannot resist the pressure to unscientific. "If the ladies' society wants an endorsement from an Authority, why not provide it and make the girls happy?" In any event, he may be sure his advice will not be heeded and he need not be responsible. He will not be paid $1,000 or $10,000, either, to study the particular problem before giving advice on it.

Finally, he may dream of a thoroughly correct applied science, but cannot hope to find the data to guide his generalizations. World government, war and peace, capitalism and federal aid to education are big problems that invite wild formulae.

Under these circumstances, every policy scientist (and every applied social scientist) must commit many an offense against science in his lifetime. But to know the reasons therefore suggests how to relive the habit: reject popularity; specify, and identify with, one's clients; apply the rules and language of science; state the indeterminacy of the counsel; and, by all means, try to examine the patient before prescribing the cure.