How Science Changes

HOW SCIENCE CHANGES

Copyright 1998 by Emerson Thomas McMullen

Abstract

This article presents a broader historical and philosophical framework than previous studies involving scientific discoveries, change, and advance (1). Science is differentiated from technology and its interactions with technology are briefly discussed. The various types of discoveries and their interrelationships are modeled. Scientific change and advance are defined. Also, the important role of serendipity is discussed. A `reject-replace' methodology is introduced as a first-order description of conceptual scientific discoveries. Historical examples include William Harvey, Alexander Fleming, Nicholas Copernicus, Michael Faraday, James Clerk Maxwell, Antoine Laurent Lavoisier, J Harlan Bretz, and the debates over fossils and spontaneous generation.

Introduction

As explained in another article, William Harvey followed a reject-replace-accept pattern in his conceptual discovery of the blood's circulation (2). Keys to Harvey's discovery were the one-way function of the venous valves, their comparatively large number, and their purpose in the body's design. The one-way function of the venous valves and how this contradicted accepted theory were known to earlier researchers, but only Harvey seemed to be willing to take action. He rejected explanations by Galen and his teacher Fabricius, and sought a replacement. After a two-year intense program of dissection, experimentation, and asking "why," Harvey finally found a better explanation (the `replace' step). He then used his approach, including using the idea of purpose, to justify his discovery to others (the `accept' step). Because of Harvey, a scientific revolution occurred in physiology.

Harvey's pattern of discovery at first glance seems to follow Thomas Kuhn's ideas about scientific revolutions. However, when examined in detail, Kuhnian analysis starts to break down. An obvious follow-up question is whether Harvey's pattern of discovery was anomalous, or were other discoveries made this way? This paper demonstrates that Harvey's pattern is not anomalous, but is one of several ways of advancing science. This result calls into question Kuhn's analysis of scientific advance. Therefore this paper also presents a wider scope for scientific discovery, change, and advance than has been described before.

Science has both theoretical and knowledge components. The theoretical component of science consists of explanatory ideas about nature, while the knowledge component is an accepted body of information about nature. Change in science can occur by adding either to the knowledge component by means of new research results or to the theoretical component by generating new ideas which others judge to be better explanations than the old ones. A scientific advance is a change which significantly adds to the body of scientific knowledge, or else provides a much better explanation than before. Given this definition, it is clear that science can advance over a broad front, and also interact with technology.

Science and Technology

Basically, science pertains to some aspect of nature while technology involves tools (3). Tools are human creations which can be physical or mental. An example of a physical tool is a microscope, while a mental tool would be a computer's software. Generally, technology underlies all modern scientific research. Galileo's telescopic discovery of four moons of Jupiter is an example of a scientific advance arising from the use of a technology.

Mental tools also can be an industrial process such as the steps involved in brewing beer or producing gasoline. Of course these processes include physical tools like fermentation tanks and cracking towers, but the steps in the processes constitute the mental tools. The combination could be considered the industrial technique. Scientists use techniques in investigating nature. Examples involve purification, calibration, and the experimental method. These tools are also important in their role as aids to scientific testing and verification. As physical and mental tools, techniques also can be important in making scientific discoveries.

The Types of Discoveries

The types of scientific discoveries are: physical, mental, conceptual, and combinations of these. Physical discoveries involve physical entities - an example is Louis Pasteur's discovery of optical isomers. Pascal's sudden solution of the mathematics of the cycloid - a problem which Galileo said was not solvable - is an example of a mental discovery, as is Fermat's proof of the sine law of refraction. Conceptual discoveries, like Harvey's discovery of the blood's circulation, are ideas about nature. Their interrelationships with the other types of discovery are shown as follows:

Conceptual discoveries involve new ideas about nature and thus, being new interpretations of the body of scientific knowledge, inherently promote change. Some conceptual discoveries also result in scientific advance, such as Friedrick August Kekule's conceiving of the ring structure of benzene, an insight which opened up a whole new realm of chemistry.

Physical and mental discoveries add to scientific knowledge but do not necessarily change any interpretation of that knowledge, while conceptual discoveries always do. If physical and mental discoveries do eventually lead to a reinterpretation of scientific knowledge, it will be by a combination with a

conceptual discovery. An example of a physical/conceptual combination is Walter Alvarez's discovery of an overabundance of iridium in the clay at the Cretaceous-Tertiary (K-T) boundary. This physical anomaly eventually led to the hypothesis that the impact of a large extraterrestrial object caused the mass extinction of the dinosaurs (4).

Physical Discoveries

A great many past physical discoveries, in one way or another, have led to scientific advances. Unfortunately for philosophers of science, many of these physical discoveries were serendipitous or incidental. Thus these discoveries neither lend themselves to in-depth analysis, nor to much philosophical discussion.

Igor Stravinsky once wrote that, "just as appetite comes by eating, so work brings inspiration, if inspiration is not discernible at the beginning" (5). This could be paraphrased for scientists by saying that discoveries occur during their work. This paraphrase could seem to be an allusion to Kuhn's "normal science" (research conducted within a paradigm), but it is not the same thing. One difference between the two is that not only could anomalies occur, but also significant serendipitous discoveries happen during scientific research. Serendipitous discoveries of anomalies could also happen, such as mentioned above with the finding of an overabundance of iridium at the K-T boundary. Another difference is that, if Stravinsky is right, scientific research itself, rather than being bound by a paradigm, is a source of inspiration. In other words, the process of scientific research opens the scientist's mind to possibilities beyond the existing paradigm. This seems to be the case for Alexander Fleming, who was alert to the unexpected and open to discoveries which challenge existing paradigms (6).

The above makes a scientist's attitude toward the occurrence of the unexpected important. Nobel laureate Paul Flory pointed out the value of the prepared mind in making discoveries: "Significant inventions are not mere accidents. . . .. Unless the mind is thoroughly charged beforehand, the proverbial spark of genius, if it should manifest itself, probably will find nothing to ignite" (7). The story of penicillin offers an example of serendipitous physical discoveries involving prepared minds.

Penicillin

Having worked with wounded soldiers in World War I, Alexander Fleming realized the limitations of using carbolic acid as an antiseptic. Carbolic acid killed white blood cells, thus interfering with the body's natural defense against bacteria. In 1921 Fleming very serendipitously discovered a substance that killed bacteria, (8) but did not harm white blood cells (9). Unfortunately the antibiotic, which he called lysozyme, killed only relatively harmless bacteria, and so his unexpected discovery had little practical significance in and of itself. However, it did prepare Fleming's mind for penicillin.

In 1928, Fleming was doing routine laboratory research on influenza. He was grumbling to a colleague about the amount of work left behind when the colleague transferred out. As evidence, Fleming picked up a plate from a group of Petri dishes that were going to be cleaned. Then he happened to notice that a mold had contaminated the staphylococcus culture growing in the dish. This unknown mold had a clear area around it similar to what he had seen earlier with the lysozyme discovery. Therefore, instead of handing the plate over to his colleague, he rescued it from the wash basin and investigated this new find (10). Fleming had not been systematically searching for antibiotics, but, because of his work on lysozyme, he immediately realized the possible importance of his discovery. He began studying the mold's antibiotic product, which he called penicillin. He found that penicillin killed harmful bacteria without harming white blood cells. This was what he knew was needed. However, Fleming was not chemist enough to purify penicillin adequately, so his crude preparations had limited results.

Feeling the pressure of what would become World War II, biochemists Ernst Chain and Howard Florey decided to investigate penicillin for human use. They produced enough of it to begin the safety testing process. Once again serendipity reigned. They happened to use mice instead of the usual guinea pigs, and received positive results. Fortunately for the development of penicillin, guinea pigs were not available for testing in wartime British laboratories; penicillin is toxic to them. If they had used guinea pigs, the negative result could have stopped the whole program (11). Happily, they moved on to human tests which, although shaky at the start, were encouraging overall.

The next hurdle was mass producing penicillin for the war. Researchers at the United States Department of Agriculture's Northern Regional Research Laboratory in Peoria, Illinois began work on this problem. It happened that corn-steep liquor, a byproduct in the manufacture of cornstarch, was abundant there in the corn belt. They found that adding it to the culture medium increased yields twelve-fold. Also serendipitously, a local woman, Mary Hunt, spotted a mold with "a pretty golden look" on an overripe cantaloupe at a Peoria fruit market. It was Penicillium chrysogenum which further increased yields. (The mold Fleming had discovered was Penicillium notatum.)

Enough penicillin had been produced by the end of the war that it began to be sold freely on the civilian market. This was because of P. chrysogenum, corn-steep liquor, new manufacturing techniques, and relatively selfless wartime cooperation among pharmaceutical companies as well as government scientists. Later, an accidental contamination during the start-up of penicillin production in an old Austrian brewery led to the discovery of an acid-resistant form of penicillin which could be taken orally.

Fleming, Chain, and Florey received the Nobel Prize in 1945 in recognition of the significance of the discovery and development of penicillin. Fleming was aware of the serendipitous nature of this scientific advance: "The story of penicillin has a certain romance in it and helps illustrate the amount of chance, or fortune, or fate, or destiny, call it what you will, in anybody's career" (12). In a 1944 address on penicillin to the Tenth Annual Chemurgic Conference, Albert Elder pointed out the importance of necessity and timing: "Were it not for the war, it is most likely that it [Penicillin] would still be little more than a laboratory curiosity" (13).

Some might want to classify the story of the discovery and development of penicillin under Kuhn's "normal science." However, there was nothing "normal" in the chain of events that gave us a powerful, life-saving antibiotic. Rather, the penicillin story is one that could not be repeated today, not only because of serendipity, necessity, and timing, but also because of more restrictive government rules for drug testing.

Conceptual Discoveries

Bearing in mind the great importance of physical discoveries which occur as scientific knowledge is increased, let us now turn to the theoretical aspect of science and discuss conceptual discoveries. This is the area where philosophers of science have concentrated their efforts and consequently why there exists a narrow framework for discovery and advance. This article too, will now focus on conceptual discoveries, but with the caveat that both serendipity and the knowledge aspects of science have a very important place in scientific change and advance, as we saw with the above history of penicillin.

As mentioned earlier, the first step in a conceptual discovery is a researcher's rejection of the received view. The reason(s) for rejection can vary. It(they) could be a result of Stravinsky's inspiration, a Kuhnian accumulation of anomalies, a Popperian falsification, (14) or a combination of factors. However, rejection alone is fruitless. This step is neither a scientific advance, nor even a change.

The second step in a conceptual discovery is conceiving of a replacement for the rejected idea. The way the replacement is conceived also varies: It could be invented freely from the imagination, or suggested by an analogy with something else, or in some other way (15). Having a replacement idea, however, does not necessarily lead to change. At this point there is competition with the old theory.

The third step in a conceptual discovery is the scientific community's acceptance of the replacement idea. This last phase creates a competition with the old theory and, in some cases, rivalry with some other hypothesis as well (16). William Harvey illustrates this Reject-Replace-Accept Process of conceptual discovery, but he is not alone.

Copernicus

At the time of the construction of the Alfonsine Tables (1275), Alfonso the Wise, the scholarly King of Castile and Leon, learned about Ptolemaic cosmogony. He remarked that if he had been present at the creation, he would have wanted to say something to God about the mechanism involved. He reasoned that if Ptolemy's system were real, it could have been more efficiently designed.

In his De revolutionibus orbium Coelestrium (1543), Nicholas Copernicus follows the same line of thinking as King Alfonso. In the Preface Copernicus criticizes those mathematicians who, in effect, take beautifully-shaped pieces: hands, feet, and other body parts, and then haphazardly reassemble them into something that is more of a monster than a man. This is not the universe of symmetry that God, "the Best and Most Orderly Workman" would have created. There are other reasons, but this is the core of Copernicus' rejection of the old cosmogony. Order, harmony, symmetry, and simplicity led him to hypothesize that the earth orbits the sun.

Copernicus' arguments for acceptance of his heliostatic idea include a comparison of the orbital times of the moon, sun, planets, and stars. In Ptolemy's geostatic system, the moon goes around the earth in the lunar month. Mercury, Venus, the sun, and the rest of the planets move progressively slower as the distance from the earth increases. Then this pattern changes dramatically. Farthest out from the earth, where one would expect the slowest movement or no motion at all, the gigantic sphere of the fixed stars makes a complete revolution in twenty-four hours. To Copernicus, this pattern seemed unharmonious and lacking in symmetry.

In Copernicus' system, each planet goes around the sun in progressively slower times the further it is from the sun, and then the distant stars are motionless, a pattern which made much more sense to him. After listing the times of the planets as they go around the sun, he concludes: "In this arrangement, therefore, we discover a marvelous symmetry of the universe, and an established harmonious linkage between the motion of the spheres and their size, such as can be found in no other way. . . ." (17). In summary, Copernicus rejected the old paradigm, not because of a new discovery, but because of his ideas about harmony and symmetry conflicted with the accepted view. His new conception of the universe better fit these ideas.

Fossils

There was a time when most naturalists thought that forces in the ground formed fossils. Fossils were `sports of nature.' While a few in the past had argued for the organic origin of fossils, the debate did not become serious until the seventeenth century. Robert Hooke explained why these forces (which he called a `plastic virtue') did not form fossil objects in the earth.

For it seems to me quite contrary to the infinite prudence of Nature, which is observable in all its works and produc tions, to design everything to a determinate end, and for the attaining of that end, makes use of such ways as are (as far as the knowledge of man has yet been able to reach) altogether consonant, and most agreeable to man's reason, and of no way or means that does not contradict, or is con trary to humane Ratiocination; whence it has a long time been a general observation and maxime, that Nature does nothing in vain; It seems, I say, contrary to that great Wisdom of Nature, that these prettily shap'd bodies should have all those curious Figures and contrivances (which many of them are adorn'd and contriv'd with) generated or wrought by a Plastic virtue, for no higher end then only to exhibite such a form. . . . (18).

This argument, based on purposeful design in nature, was also used by John Ray.

From his work as a naturalist, Ray realized how improbable it was for fossils and living organisms to exhibit so many similarities. For example, if fossil bivalve shells were hinged like living bivalves, it was nearly impossible to deny that the hinge served the same purpose in both cases. Therefore, the fossil had to have had an organic origin. The same line of reasoning using efficient purpose could be applied to other fossil objects, especially those with living counterparts.

With the above arguments and many others, scientists finally resolved the debate in the eighteenth century. This also resulted in scientific advance since it opened the door for paleontology.

Faraday - Maxwell

While conducting his many worth-while experiments, Michael Faraday gained several tremendous insights that challenged the views of his day. Concerning the nature of electricity, he imagined that it was a vibration or a force that was transmitted through tensions in the conductor. He conceived that a magnetic field is composed of lines of force, and that magnetic and electrical actions are not transmitted instantaneously. This concept was rejected by most mathematical physicists because they thought the attraction and repulsion of electrical charges occurred as action at a distance. Most importantly, Faraday described light "radiation as a high species of vibration in the lines of force." Concerning the latter hypothesis, he boldly stated that "It endeavours to dismiss the aether but not the vibrations" (19).

Faraday was not a mathematician and could not develop his insights any further, but James Clerk Maxwell did. In the Preface to his landmark Treatise on Electricity and Magnetism (1873), Maxwell wrote that his major task was to mathematize Faraday's conceptual discoveries. In a highly creative work, Maxwell developed four partial differential equations describing the electromagnetic field (20). Known as Maxwell's equations, they are one of the greatest advancements of nineteenth-century physics. Here Maxwell provides us an example of scientific advance based on Faraday's conceptual discoveries.

Combinations

The most intriguing combinations are those involving conceptual discoveries. In these cases the above Reject-Replace-Accept process yields interesting analyses.

Spontaneous Generation

For centuries, natural philosophers thought living beings spontaneously formed from matter, but this idea eventually came under attack. The resulting controversy lasted even longer than that over fossils. As with the fossil debate, this controversy also heated up in the seventeenth century. One example is John Ray, who used purpose to attack spontaneous generation. He asked what accounts for the two different sexes, male and female, if spontaneous generation actually occurs. Sexes and the sex act are anomalies in the spontaneous generation paradigm: "Why should there be implanted in each sex such a vehement and inexpugnable Appetite of Copulation?" (21) In the nineteenth century, Louis Pasteur finally resolved this debate with a series of convincing experiments. (These experiments are what make this advance a combination which involves a conceptual discovery.)

Lavoisier

Antoine-Laurent Lavoisier rejected the reigning phlogiston theory that Georg Ernst Stahl had developed. This rejection occurred when he found in 1772 that "sulpher, in burning, far from losing weight, on the contrary gains weight." This result is the opposite of that predicted by the phlogiston theory. He was repeating experiments by Robert Boyle, who had already observed this phenomenon. Unfortunately, Boyle had arrived at the wrong interpretation.

Lavoisier doggedly sought a new hypothesis, but was initially unsuccessful. Finally, in 1774, he correctly realized the role of oxygen in combustion. His breakthrough came when Joseph Priestly communicated to him his discovery of "dephlogisticated air" (oxygen). Only then did it all come together in Lavoisier's mind and he had his replacement theory of combustion.

In a 1775 memoir, Lavoisier announced his "new theory of combustion." He argued that it agreed more with the facts than did the phlogiston theory.

Furthermore, I repeat, in attacking here Stahl's doctrine my object is not to substitute a rigorously demonstrated theory but solely a hypothesis which appears to me more probable, more conformable to the laws of nature, and which appears to me to contain fewer forced explanations and fewer contradictions (22).

In a 1783 memoir, he continued this argument based on simplicity, saying that "all the phenomena of combustion and calcination may be explained in a far simpler and easier manner without phlogiston than with it." Then he resignedly wrote: "I do not expect that my ideas will be adopted all at once." He went on to explain that human nature tends toward one viewpoint, and the longer one holds a certain point of view, the harder it will be for that person to change. His hope was in "the young people who are commencing to study science without prejudice" (23). Here he turned out to be too pessimistic. Nearly everyone accepted his replacement theory because of his logical and clear presentation in Traite elementaire de chimie (1789). This change was such an important scientific advance for chemistry that it is often labeled revolutionary (24).

Bretz

After World War One, J Harlen Bretz, a professor at the University of Chicago, examined the deep coulees, denuded soil, and abandoned waterfalls of the Columbia plateau in Washington State. He decided these formations could not be explained by the normal geological processes of ponding and stream erosion. He finally hypothesized that they had to have been formed by a gigantic flood that swept through the Eastern part of Washington State. (While Stephen Jay Gould argues that Bretz was a "rigid empiricist of the naive indictivist [sic] school," (25) Bretz's approach was more sophisticated than this. In his seminal 1923 paper, he gives us evidence that he considered alternatives when he says that no other hypothesis will explain what he observed (26). So Bretz did consider competing hypotheses and compared them to his.) In one sense, Bretz's theory fits the conceptual category, but in another sense it does not; he was the only one who believed it. It languished for decades and did not gain acceptance until the physical discoveries of giant ripple marks and the source of the floodwaters, now called glacial Lake Missoula. In this case there was no scientific advance until physical discoveries were made which could only be explained by Bretz's theory.

Discussion

The ideas presented here about scientific discovery, change, and advance are based in historical studies. Thus, they are certainly descriptive, but they are also prescriptive in that they tell how scientific research ought to be conducted. This descriptive/predictive aspect will become more discernible in the following discussion.

In the examples of conceptual discovery, or any combination involving a conceptual discovery, the data and evidence that form the basis for rejection often were in existence for years, decades, or even centuries. There was no Kuhnian "crisis" (due to the accumulation of anomalies) in the scientific community as a whole or even in parts of it. Copernicus made no new discovery; he reinterpreted data known for over a millennium and a half. The abundance of the valves in the veins had been known for decades before Harvey; only he decided this was important. Lavoisier had made the same combustion experiments and observed the same results nearly a century after Boyle did. The catastrophic nature of the geological formations of the Columbia Plateau were there for any geologist to see, but only Bretz `saw.' The lesson here is that, as well as searching for new knowledge, scientists must be alert to re-interpreting old data.

Further, the rejection of a theory comes before the conception of a new hypothesis and is itself a most difficult step. As noted above, Lavoisier understood how our minds can be restricted by thought-constructs. It is a creative act to break out of this kind of mental cage. Even when investigators do realize anomalies exist concerning a theory, or the theory's falsification has happened, they often continue to hold onto that particular theory, even though it is now suspect. Some persons are comfortable in scientifically ambiguous situations, but others are not. Those uncomfortable with ambiguity seek, are driven even, to resolve the problems with a better explanation. Science should not have to rely solely on this type of individual. From a prescriptive standpoint, all scientists, not just some, should be looking for, and seeking to resolve, anomalies.

However, as mentioned earlier, simply rejecting a theory gains little. The scientists we have studied all realized that a new hypothesis must be posed and that it must be defended. This is not easily done. It took Harvey at least two years to conceive of his replacement theory. Lavoisier had struggled for two years looking for a new hypothesis, and even then it took the physical discovery of oxygen to trigger his new conception of combustion. We should expect that rejecting the old theory may be the easiest step, and that conceiving a new hypothesis could be much more difficult, consequently consuming much time and energy.

Even when a scientist conceives of a new hypothesis, it may exist side-by-side in the scientific community for a long time with the theory it eventually replaces. This was true for both our biological examples, fossils and spontaneous generation. There was no sudden conversion or 'gestalt' process. Generations of scientists passed before general acceptance occurred. Further, there may be more than just a two-way competition between hypotheses. The physical discovery of the location of a nova and a comet caused Tycho Brahe to reject Aristotelian cosmogony. He developed a geo-heliocentric system that challenged both Copernicus' heliostatic and Ptolemy's geostatic ones. So we should not be surprised if there are several newer hypotheses competing for acceptance with the old theory, and also with each other. As a matter of fact, we should be uncomfortable if there were not several theories being hotly debated and argued over in each field. How can science progress if researchers allow one theory to dominate a field?

Incommensurability

In the examples given, Harvey, Hooke, and Ray employed teleological thinking in their science. They were challenging Galen and Aristotle, who also used purpose in their science. However, even though there was revolutionary change, the use of purpose as a scientific approach continued throughout the Scientific Revolution (27). Thus, concerning change due to conceptual discoveries, there are certain core assumptions that do carry over from the old theory to the new. So there is no Kuhnian incommensurability between an old theory and its replacement. Imre Lakatos had thought this, (28) as did Gerald Holton. Holton wrote that "once formed, the thematic commitment of a scientist typically is remarkably long lived." Thus, one would expect that "thematic oppositions persist during `normal science' and themata persist through revolutionary periods" (29).

The Logic of Scientific Discovery and Justification

Is there a logic of scientific discovery? This has been a much-discussed question and the consensus answer seems to be "No." Certainly the serendipitous nature of many physical discoveries indicates that there is no overall logic of scientific discovery.

What happens if we narrow the above question and change it to "Is there a pattern of conceptual discovery?" At the overall, general level, the answer is "Yes." The pattern outlined in this paper is that first there is rejection of the old theory, then a search for a replacement theory, and finally arguments for acceptance of the new theory. Conversely, at the detail level, there is no pattern or logic of conceptual discovery because psychological and sociological factors often are involved in the rejection and/or replacement steps. However, the acceptance step of the process presented here does have a logic of justification in that evidence is presented and rational arguments are made for the new hypothesis.

Conclusion

The above discussion of the Reject-Replace-Accept-Process has been in a history of science context, but it can be applied also to the philosophy of science, specifically to that of Kuhn. There have been many important criticisms of Kuhn's explanation of scientific revolutions, but according to the above analysis, rejection is not enough. This explains why, in spite of these criticisms, Kuhn's thinking and nomenclature have permeated intellectual discourse. Taken as a whole, the ideas presented here challenge Kuhn's explanation of scientific change. Examined at the detail level, there are a few things that are the same, anomalies for instance, but there are many other things that are different. Crisis, conversion, incommensurability, and the definition of normal science, are examples of the differences. We are now in the competition part of the acceptance phase. Following Lavoisier's thinking, let us compare the examples here, or from any other source for scientific change, and see which explanation best provides a framework for what happens in science and which explanation is more inclusive.

Acknowledgments: I thank Charles Crouch and Barbara Horan for their comments.

References

1. Kuhn T: "The Structure of Scientific Revolutions." Chicago: University of Chicago Press, 1962/1970. Kitcher P: "The Advancement of Science." New York: Oxford University Press, 1993, esp. Chapter Three. Solomon M: Legend naturalism and scientific progress: an essay on Philip Kitcher's "The Advancement of Science." Stud Hist Phil Sci 26: 205-218, 1995.

2. McMullen ET: New insights on William Harvey's Discovery. GA J Sci 53: 101-115, 1995. Also see McMullen ET: Anatomy of a physiological discovery: William Harvey and the circulation of the blood. J Royal Soc Med 8: 491-8, 1995.

3. Basalla G: "The Evolution of Technology." New York: Cambridge University Press, 1988.

4. Alvarez LW, Alvarez A, Asaro F and Michel HV: Extraterrestrial cause for Creataceous-Tertiary extinction. Science 208: 1095-1108, 1980.

5. Madigan CO and Elwood A: A gallery of quotes about inspiration. In Brainstorms and Thunderbolts. New York: Macmillan Publishing Co., 1983, p286.

6. Kuhn's concept of normal science is that which engages "most scientists throughout their careers." It is a "mopping-up operation" of extending facts within a paradigm. Kuhn, op. cit., p24. My concept of what most scientists do is an extending of facts all right, but their thinking is not confined as rigidly by present theory as Kuhn seems to think. They realize the contingincy of their theoretical world. On the contrary, Kuhn argues that scientific research is "an attempt to force nature into the preformed and relatively inflexible box that the paradigm supplies." I disagree. Fleming was not doing that. He was not trying to fit his results into some theory. Further, my own experience in working in three different scientific research laboratories is that there was a thrill in finding something new, period. Whether or not the discovery fit current theory was not a prime factor. If anything, the discovery was more challenging if it did not fit the prevailing paradigm because then I was forced to double-check and rethink everything.

7. Quoted in Roberts RM: "Serendipity." New York: John Wiley and Sons, Inc., 1989, px.

8. Fleming A: On a remarkable bacteriolytic element found in tissues and secretions. Proceedings Royal Soc London Series B 93: 306-317, 1922.

9. The odds against this discovery are discussed in Macfarlane G: "Alexander Fleming, the Man and the Myth." Cambridge, MA: Harvard University Press, 1984, pp100-103.

10. Ibid, pp120-121.

11. Sheehan JC: "The Enchanted Ring, the Untold Story of Penicillin." Cambridge, MA: The M.I.T. Press, 1982, p32.

12. Roberts RM: op. cit., p164.

13. Elder AL: Penicillin. Scientific Monthly 58: 406, 1944.

14. Popper KR: "The Logic of Scientific Discovery." New York: Basic Books, Inc., 1959, pp78-92 and 112-135.

15. Carl Hempel also mentions "happy guesses," but his context is different from mine. See "Philosophy of Natural Science." Englewood Cliffs, NJ: Prentice Hall, Inc., 1966, p15.

16. Imre Lakatos discusses a "three-cornered fight," but again in a different context than here. See Falsification and the methodology of scientific research programmes. In Criticism and the Growth of Knowledge (Lakatos and Musgrave, Eds) New York: Cambridge University Press, 1970/1985, p115.

17. Copernicus N: "On the Revolution" (Rosen, Trans.) Baltimore, MD: The Johns Hopkins University Press, 1978/1992, pp21-22.

18. Hooke R: "Micrographia: or Some Physiological Descriptions of Minute Bodies Made by Magnifying Glasses with Observations and Inquiries Thereupon." London, 1665; New York: Dover Publications reprint, 1961, p112.

19. Faraday M: Thoughts on ray vibrations. London, Edinburgh and Dublin Phil Mag and J Sci, Third Series, May 1846, p348.

20. Maxwell's use of analogies and mental models are discussed in Nersessian N: How do scientists think? In Cognitive Models of Science (Giere, Ed) Minneapolis, MN: University of Minnesota Press, 1992, pp3-44.

21. Ray J: "The Wisdom of God Manifested in the Works of theCreation," The Seventh Edition, Corrected, London, 1717, p115.

22. Lavoisier AL: Memoir on combustion in general, September 5, 1775. In A Source Book in Chemistry, 1400-1900 (Leicester and Klickstein, Eds) Cambridge, MA: Harvard University Press, p173.

23. Lavoisier AL: Reflections on phlogiston, serving to develop the theory of combustion and calcination. Ibid, pp173-174.

24. There are many in-depth analyses of Lavoisier and the other examples cited. For example, see Thagard, P: "Conceptual Revolutions." Princeton: Princeton University Press, 1992. I have only discussed Lavoisier and the others as illustrations of the overall framework presented here.

25. Gould SJ: Toward the vindication of punctuational change. In Catastrophes and Earth History (Berggren and van Cowering, Eds) Princeton: Princeton University Press, 1984, p20.

26. Bretz JH: The channeled scablands of the Columbia Plateau. J Geol 21: 621, 1923.

27. McMullen ET: "A Barren Virgin? Teleology in the Scientific Revolution." Ph.D. Thesis, Indiana University, 1989. "William Harvey and the Use of Purpose in the Scientific Revolution: Cosmos by Chance, or Universe by Design?," University Press of America, 1998.

28. Lakatos I: "The Methodology of Scientific Research Programs." New York: Cambridge University Press, 1978, pp49-50, 68, 88.

29. Holton G: "The Scientific Imagination." New York: Cambridge University Press, 1978/1988, p23.