9  Hypotheses - Good and Bad

(Modified on 2026-08-11)
582 words

Discovery vs. Justification

Traditionally, philosophers of science have made a distinction between the context of discovery and the context of justification (Popper 1959; Reichenbach 1938). According to this view, scientists shouldn’t care where an idea comes from. The important issue is whether the idea is justifiable or not—that the idea survives experimental testing. A hypothesis might come from a careful theoretical calculation, a religious text, a drug-induced vision, or last night’s dream. It simply doesn’t matter where a theory, conjecture, idea, or hypothesis comes from.

A classic historical example of the irrelevance of the context of discovery is Friedrich Kekulé’s discovery of the chemical structure of benzene. Throughout the first half of the 19th century, benzene had thwarted chemists’ best efforts to understand its structure. One night, Kekulé had a dream of a snake swallowing its tail. Inspired by his dream, he began a series of experiments that ultimately established that benzene is structured as a circular ring consisting of six carbon atoms. From the scientific perspective, it doesn’t matter that Kekulé got the idea from a dream. He could have been inspired by reading his horoscope, for all scientists care. The important point is that he carried out a series of experiments that ultimately justified the idea.

Traditionally, empiricists are taught not to care where hypotheses come from. Good research centers on the testing of hypotheses. In conventional philosophy of science, what counts is the context of justification, not the context of discovery. In more recent years, scholars have become aware that one shouldn’t have such a cavalier attitude towards hypotheses.

Better and Worse Hypotheses

Recall that a hypothesis is a testable conjecture, derived from a theory. It is a claim or prediction that can, in principle, be compared to existing or future observations. Not all hypotheses are equally good. One might claim, for example, that the moon is made of green cheese. A multi-billion-dollar mission might be carried out, sending a remote vehicle to the moon in order to test this hypothesis. Clearly, this would be a waste of time, effort, and money. Compare the green-cheese hypothesis to one that geophysicists consider more interesting: that the moon was once part of the earth, and that it broke away in a catastrophic collision with another large celestial object. This is a hypothesis that geophysicists regard as more worthwhile.

Since testing hypotheses is time-consuming and expensive, researchers ought to be picky about which hypotheses they test. Ideally, researchers should focus on the most important hypotheses first. There are at least two issues involved here. First, how do we prioritize some hypotheses over others. Second, how do we encourage the creative process by which researchers come up with good hypotheses?

Prioritizing Hypotheses

Consider first the evaluative question: How do we evaluate the potential worth of different hypotheses? The following list of five evaluative criteria is intended to be suggestive rather than exhaustive:

  • Plausible: Good hypotheses should be plausible. There is little point testing hypotheses (like the green-cheese hypothesis) that we a priori consider unlikely.
  • Beneficial: Good hypotheses are morally or aesthetically pertinent. Depending on the results of the research, the hypothesis may have repercussions for human health or wellbeing.
  • Critical: Good hypotheses help distinguish between two or more competing theories. A so-called “critical experiment” is an experiment that pits two theories directly against each other. If theory A is right, then the results should be +X; if theory B is right, then the results should be -X.
  • Interesting: Good hypotheses are compelling in some way. Depending on the results of the research, the hypothesis may be transformative in how people think about themselves and the world.
  • Fundamental: In many cases, nearly everyone relies on an assumption that has never been tested. Depending on the results of the research, if a commonplace assumption is found to be wrong, this would potentially have major repercussions.

Creating New Hypotheses

So how might a researcher generate new hypotheses? In particular, how might a researcher generate new hypotheses of the highest value? As with all creative enterprises, discovery is an imperfect art. One can only offer suggestions, such as the following tips:

  1. Keep abreast of contemporary research.
  2. Converse regularly with other scholars.
  3. Read widely. Include materials beyond your discipline. Take notes.
  4. Seek out new personal experiences, such as through travel.
  5. Get in the habit of thinking cross-culturally and historically. Try to imagine what people in other times and places would want to say to you. Read old literature with an eye to possible forgotten ideas.
  6. Develop curiosity about things other people do that you don’t understand. Listen broadly to the experiences of others.
  7. Keep a research journal. Maintain an intellectual life.

We’ve left the most important advice for developing good hypotheses for last:

  1. Make use of exploratory research methods.