GENERAL SCIENCE



The Basics : Theories



How are they developed ?



Once you understand how the scientific method works, the next step in understanding science is understanding the nature of scientific theories. If scientific observations are the fuel which power scientific discoveries, then scientific theories are the engine. Theories are what allow scientists to organize and understand earlier observations, then predict and create future observations. Scientific theories can be described by characteristics which they share in common and which differentiate them from unscientific theories. Informally, we can say that the criteria of scientific theories can be grouped into a few basic principles: scientific theories should be consistent, parsimonious, correctable, empirically testable/verifiable, useful, and progressive.

A key factor to keep in mind is that the term theory is used by scientists in a manner different from common usage. For most contexts, a theory is just a vague and fuzzy idea about how things work - in fact, one which has a low probability of being true. This is where we get the complaint that something in science is only a theory and hence shouldn't be given a great deal of credibility. For scientists, however, a theory is a conceptual structure which is used to explain existing facts and predict new ones. According to Robert Root-Bernstein in his essay, On Defining a Scientific Theory: Creationism Considered, to be considered a scientific theory by most scientists and philosophers of science, a theory must meet most, if not all, of certain logical, empirical, sociological and historical criteria. These groupings are important because most people tend to focus only on the logical and empirical criteria - it is easy to forget that science also encompasses important social and historical aspects which play a role in how theories are evaluated.

Logical Criteria

A scientific theory must be

  • a simple unifying idea that postulates nothing unnecessary (Occam's Razor)
  • logically consistent
  • logically falsifiable (i.e., cases must exist in which the theory can be imagined to be invalid)
  • clearly limited by explicit boundary conditions so that it is clear whether or not particular data are or are not relevant to verification or falsification The above logical criteria are among the most cited when it comes to discussions about the nature of scientific theories and how science differs from nonscience or pseudoscience. When a theory includes unnecessary ideas or is inconsistent, it becomes difficult to see how it actually explains anything. When a theory is not falsifiable, it is impossible to tell if it is true or not, and thus it won't be possible to correct it via experimentation. Finally, when there aren't any clear boundaries, we'll never know if particular observations count for or against it.

Empirical Criteria

A scientific theory must

  • be empirically testable or lead to predictions or retrodictions that are
  • actually make verified predictions and/or retrodictions
  • involve reproducible results
  • provide criteria for the interpretation of data as factual, artifactual, anomalous or irrelevant

The above criteria are not cited quite as often as the logical criteria, but they come close. A theory which cannot be tested empirically is useless for researchers. A theory which has not made any actually verified predictions might prove useful in the future when its predictions are verified, but not currently. A theory which cannot provide retrodictions (to utilize present information or ideas to infer or explain a past event or state of affairs - e.g., to retrodict past eclipses as opposed to predicting future eclipses) may also be useful in the future, but not currently. If a theory's results cannot be reproduced, it is impossible to determine if those results were ever actually valid (rather than the result of error or fraud).

Finally, a scientific theory must allow us to better understand the nature of the data we have - after all, not all data are valid for a particular theory. Some data may be interpreted as factual (that is, they fall within the boundary conditions specified by the theory and verify its predictions or retrodictions); some may be artifactual (that is, the result of secondary or accidental influences lying outside the boundaries set for the validity of the theory); some are anomalous (that is, demonstrably valid within the bounds of the theory, but also at odds with predictions or retrodictions made by the theory); some are irreproducable and thus invalid; and some are irrelevant since they address the theory not at all.

Sociological Criteria

A scientific theory must

  • resolve recognized problems, paradoxes, and/or anomalies irresolvable on the basis of preexisting theories
  • pose a new set of scientific problems upon which scientists may work
  • suggest a paradigm or problem solving model to help resolve these new problems
  • provide definitions of concepts or operations which will help other scientists solve problems

The above criteria are more often recognized by critics than supporters of science. This is unfortunate because they are not actually problems - on the contrary, they underscore the fact that science is generally something ultimately done by a community of researchers and that many scientific problems are problems discovered by the work of a community.

Thus, a scientific theory must address some problem which the scientific community needs resolved. It need not be an old problem; in fact the articulation of the theory can itself be the means of informing people about the problem. However, if there is no actual problem and we are not looking at an issue which is already covered by other theories, then how can this new theory qualify as scientific? Furthermore, this new theory must offer a means of resolving the problem - simply pointing it out isn't enough.

Historical Criteria

A scientific theory must

  • meet or surpass all of the criteria set by its predecessors or demonstrate that any abandoned criteria are artifactual
  • be able to explain all of the data gathered under previous relevant theories in terms either of fact or artifact (no anomalies allowed)
  • be consistent with all preexisting ancillary theories that already have established scientific validity

These historical criteria are closely related to the sociological ones. A scientific theory does not need to simply solve some problem, but it needs to do so in a way which is superior to other, competing theories which people have offered in the past. In addition, a good theory needs to be able to explain more data than the competition - scientists prefer fewer theories which explain more things rather than many theories, each of which explains very little. All of this is what ensures that scientific theories increase in their explanatory power.

Legal Criteria

Root-Bernstein does not actually list any legal criteria for scientific theories - and ideally, there wouldn't be. Science is a matter of research, not a matter of law. However, in 1981, there was a trial in Arkansas over an equal treatment law that made creationism a required subject which had to be taught whenever evolution was taught. This law was challenged and overturned, as were similar laws elsewhere. Ultimately the U.S. Supreme Court made the determination that such laws were unconstitutional.

The Arkansas trial is frequently highlighted because it was the first of its kind, and also because of the very comprehensive judgment that came out of it. The trial had several high profile, well known scientists involved in giving evidence, including descriptions of science, so the outcome of the trial has some relevance. In his ruling, Judge Overton found that science has four essential features

  • It is guided by natural laws, and is explanatory by references to natural laws
  • Science is testable against the empirical world
  • Its conclusions are tentative, not the final word
  • It is falsifiable

Summary

Overall, the criteria for scientific theories can be loosely be summarized by a few basic principles. Scientific theories are

  • Consistent (internally and externally)
  • Parsimonious (sparing in proposed entities or explanations)
  • Useful (describes and explains observed phenomena)
  • Empirically Testable and Falsifiable
  • Based upon Controlled, Repeated Experiments
  • Correctable and Dynamic (changes are made as new data is discovered)
  • Progressive (achieves all that previous theories have and more)
  • Tentative (admits that it might not be correct rather than asserting certainty)

It is true that the exact nature of science is open for debate, and some if not all of these criteria could be questioned. However, in practice, the above criteria are a pretty good description of characteristics that one would expect a theory to posses for it to be considered a scientific theory. Lacking one or two might not mean that a theory isn't scientific, but only if there are very, very good reasons; lacking most or all, however, will certainly disqualify an idea from being genuinely scientific.