Page:Sm all cc.pdf/153

 discussed later in this section. An explanation must also be self-consistent: for example, all hypotheses are wrong, including this one. In 320 B.C., Pytheas of Massilia sailed beyond the northernmost limits of the known world, to the land of Thule north of Britain. When he returned, he claimed that in Thule the midsummer sun did not set. His contemporaries called this observation preposterous. Scope is another aspect of simplicity. A hypothesis that only accounts for the observations that inspired it has little value. In contrast, a hypothesis with a broad explanatory power inspires confidence through its ability to find order in formerly disparate types of observations. Scope is the antidote to Popper’s [1963] criticism that many similar confirmations can only marginally increase confidence in a hypothesis. A hypothesis with broad scope tends to be more amenable to diversified testing.

Progressiveness, or fruitfulness, is a seldom discussed value. Kuhn [1977] says simply that “ a theory should be fruitful of new research findings: It should, that is, disclose new phenomena or previously unnoted relationships among those already known.” Most hypotheses seek to disclose previously unnoted relationships. Yet some are dead ends, sparking no further research except the confirmation or refutation of that specific conjecture. In contrast, progressive hypotheses are valued because of their exciting implications for a variety of new research directions. Even if a fruitful idea is later determined to be wrong, it can constructively steer future research efforts. Oliver [1991] thinks that the best criterion for the value of a scientific publication is the “impacts of the paper on the flow of science,” the extent to which it changes what other scientists do. “A great discovery is a fact whose appearance in science gives rise to shining ideas, whose light dispels many obscurities and shows us new paths.” [Bernard, 1865]

“A great discovery is not a terminus, but an avenue leading to regions hitherto unknown. We climb to the top of the peak and find that it reveals to us another higher than any we have yet seen, and so it goes on.” [Thomson, 1961] Utility is not just a crucial value for applied scientists; it is a common concern of all scientists. We scan journals and focus almost exclusively on articles that may be of some utility to us. To results that are not personally useful, we apply the most lethal hypothesis-evaluation technique: we ignore them. Similarly, the depth of our evaluation depends on the perceived relevance and utility of the hypothesis. When choosing between two hypotheses, we normally select the more pragmatic and useful one. For example, a useful empirical equation is often preferred over a rigorous theoretical equation, if the latter includes several variables that we are unable to estimate.

Expediency is concern for what is immediately advantageous, and scientific expediency favors acceptance of promised solutions to worrisome problems. Scientific anxiety is created when a ruling theory is threatened, or indeed whenever a discipline is faced with an apparently irreconcilable conflict -- perhaps between two incompatible hypotheses or perhaps between a strong hypothesis and a compelling dataset. Any evidence or ancillary explanation that promises remedy for the anxiety is likely to be received favorably -- almost gratefully -- because of the expediency factor.