Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Posting-Version: version B 2.10.1 6/24/83; site gargoyle.UChicago.UUCP Path: utzoo!watmath!clyde!burl!hou3c!hocda!twitch!hocad!houxm!ihnp4!gargoyle!carnes From: carnes@gargoyle.UChicago.UUCP (Richard Carnes) Newsgroups: net.origins Subject: What is a scientific theory Message-ID: <213@gargoyle.UChicago.UUCP> Date: Thu, 4-Oct-84 20:58:00 EDT Article-I.D.: gargoyle.213 Posted: Thu Oct 4 20:58:00 1984 Date-Received: Sat, 6-Oct-84 05:36:11 EDT Organization: U. Chicago - Computer Science Lines: 57 [] In response to a recent request of Paul Dubois, here is a definition of "scientific theory", cribbed from an article by Robert Root-Bernstein in _Science_and_Creationism_ , ed. Ashley Montagu, Ox. U. Press (1984) (herewith highly recommended). I'm not claiming that this is the only possible definition, but I think it's worth considering. Root-Bernstein says that a scientific theory must fulfill four sets of criteria: 1. LOGICAL criteria. A theory must be: a) a simple, unifying idea that postulates nothing unnecessary ("Occam's Razor"); b) logically consistent internally; c) logically falsifiable (i.e., cases must exist in which the theory could be imagined to be invalid); d) clearly limited by explicitly stated boundary conditions so that it is clear whether or not any particular data are or are not relevant to the verification or falsification of the theory. 2. EMPIRICAL criteria. A theory must: a) be empirically testable itself or lead to predictions or retrodictions that are testable; b) actually make VERIFIED predictions and/or retrodictions; c) concern reproducible results; d) provide criteria for the interpretation of data as facts, artifacts, anomalies, or as irrelevant. 3. SOCIOLOGICAL criteria. A theory must: a) resolve recognized problems, paradoxes, and/or anomalies irresolvable on the basis of preexisting scientific theories; b) pose a new set of scientific problems upon which scientists may work; c) posit a "paradigm" or problem-solving model by which these new problems may be expected to be resolved; d) provide definitions of concepts or operations beneficial to the problem-solving abilities of other scientists. 4. HISTORICAL criteria. A theory must: a) meet or surpass all of the criteria set by its predecessors or demonstrate that any abandoned criteria are artifactual; b) be able to accrue the epistemological status acquired by previous theories through their history of testing---or, put another way, be able to explain ALL of the data gathered under previous relevant theories in terms either of fact or artifact (no anomalies allowed); c) be consistent with all preexisting ancillary theories that already have established scientific validity. According to Root-Bernstein, evolutionary theory meets all these tests, while creationism flunks them all except for 1(a). --Richard Carnes