There are other methods of discovering and learning knowledge about
nature … but science is the only method that results in
the acquisition of reliable knowledge.
Steven D. Schafersman
An Introduction to Science: Scientific Thinking and the Scientific
Method
But no faith can be more misleading than an unquestioned
personal conviction
that the apparent testimony of one's own eyes must provide a purely
objective
account, scarcely requiring any validation beyond the claim itself.
Utterly
unbiased observation must rank as a primary myth and shibboleth of
science,
for we can only see what fits into our mental space, and all
description
includes interpretation as well as sensory reporting. Moreover, our mental
spaces
embody a complex architecture built of social constraint, historical
circumstance,
and psychological hope--as well as nature's factuality, seen through a
glass darkly.
Stephen Jay Gould
"The sharp-eyed Lynx, outfoxed by nature"
Natural History, June 98, 107(5) 23.
Abstract:
In the late 1950s, Edwin Land proposed a new theory of color vision, but a
large segment of the scientific community has failed to embrace it. The
reason for this failure is considered in Kuhnian terms. Implications for
Christians contemplating the interface between their faith and natural
science are addressed.
Introduction and Background
In order to appreciate the arguments and conclusions presented below, you must be acquainted with three things: 1. The generally accepted scientific explanation of how we see color; 2. Thomas Kuhn’s explanation of how the natural sciences make progress; and, 3. Edwin Land’s experiments with black-and-white photography performed in the late 1950s. Since many readers will not be familiar with all three of these, they are presented briefly below. Further, all of this is predicated on your having a clear understanding of how scientists typically describe "the scientific method." /1/
Classic Color Vision Theory
Anyone who remembers high school or college physics can tell you why an apple looks red and leaves appear green. We were taught the "classic color vision theory" (CCVT) that developed as the result of the work of Isaac Newton (1642-1727), John Dalton (1766-1844), Thomas Young (1773-1829), Hermann von Helmholtz (1821-1894), James Clerk Maxwell (1831-1879), and a few other scientists unknown to most people outside the field. /2/
Natural light (also called white light) is just part of a larger spectrum of electromagnetic radiation (EMR) that includes radio waves, microwaves, x rays, gamma rays, infrared, and ultraviolet radiation./3/ Light is that portion of the EMR spectrum with wavelengths between 740 and 390 nm./4/ White light is made up of red, orange, yellow, green, blue, and violet portions, each corresponding to different ranges of wavelengths (see Fig. 1).
According to the CCVT, a red apple absorbs all of the wavelengths of light except the red wavelength, which it reflects, and a green leaf absorbs all of the wavelengths of light except the green wavelength, which it reflects (see Fig. 2). This is not only the explanation taught in high school, it is also typically offered in college physics textbooks. Many people—even scientists—assume that the CCVT is the only satisfactory explanation of color vision./5/ Though not referred to as a "law," this CCVT is so well established that it is treated as "known."
In order to appreciate the problems in evaluating color vision theory, you need to understand how progress is made in the natural sciences.
How Science Progresses
Until about 1962, most students of science thought science progressed through a smooth continuum of one researcher building on the work of previous scientists. Then Thomas Kuhn’s Structure of Scientific Revolutions was published. Kuhn presented philosophers of science with a new understanding of how science progresses by suggesting that significant progress in science is not accomplished through evolutionary steps /6/—a generally accepted view of that day—but via revolutionary breakthroughs. /7/
In this seminal work, Kuhn recites a litany of historical examples to demonstrate that the major advances in the natural sciences did not typically result from researchers building on the work of their predecessors, but on their developing radically new and competing theories. Kuhn suggested that science progressed when phenomena were observed that could not be accounted for by the prevailing theory. These anomalies produced a crisis; the more anomalies, the greater the crisis. If and when the crisis became large enough, someone would offer a new theory—an explanation that accounted for everything the prevailing theory explained, but also explained the new observations. Kuhn used the term "paradigm" to refer both to the new theory and to the experiment or group of experiments that demonstrated the inadequacy of the old view and the superiority of the new. Discussions between researchers embracing the competing paradigms were often difficult, if not impossible. Frequently people on opposing sides of the issues used the same terms so differently that their views were virtually incommensurable. But, as the new paradigm gained acceptance, those who failed to embrace it were eventually "read out of the literature" (prestigious journals would not publish their work), and they could not compete successfully for research funds. /8/ Thus the new replaced the old, and the paradigm shift was finished.
The process of scholars in that discipline (or sub-discipline) discarding the old theory and embracing the new, Kuhn called a paradigm shift. This shift resolved the crisis. Typically not all researchers in that field would make the paradigm shift—they could not discard the old paradigm. Did they have a vested interest in seeing the old paradigm prevail?
Here are a few examples of some better-known scientific paradigm shifts.
Sixteenth century cosmology posited a universe in which the earth was stationary and the sun, moon, and planets revolved around the earth. In the 1500s this geocentric model was challenged and eventually replaced by another theory. Polish astronomer Nicolaus Copernicus (1473-1543) argued that the earth rotates on its axis and revolves around the sun. /9/ Although this heliocentric theory was not completely correct and was later modified, it did provoke a paradigm shift.
About this same time Europeans generally thought the world was flat. The explorations of people like Christopher Columbus (1451-1506) and Ferdinand Magellan (1480-1521) exploded that myth, but it took a few generations for people to make a shift in their thinking.
Alfred Wegener’s The Origin of Continents and Oceans (1915) provides a twentieth century example. For about 50 years the scientific community ignored or rejected the continental drift ideas of this German meteorologist. Then in the 1960s evidence mounted that continental plates exist, rupture, drift apart, and collide with each other. /10/ This increased evidence finally produced a paradigm shift and Wegener’s ideas were embraced (though modified).
Medicine provides a couple of more modern examples. Before the mid-1970s, a person diagnosed with diverticulitis /11/ was typically prescribed a low-fiber diet. Research during the early 1970s indicated that this low fiber treatment exacerbated the problem, and thus the prescription for those suffering with diverticulitis was reversed from a low-fiber to a high-fiber diet./12/
A similar process took place in 1983 when Dr. David Larson reported that he had discovered that many ulcers are caused by a bacterium (Helicobacter pylori) and could be cured—not just treated—with antibiotic therapy. /13/ For several months (perhaps even years) following the announcement of this discovery and its publication in medical journals, many physicians continued to prescribe traditional therapies that neutralized or reduced production of stomach acid. /14/ Antibiotic therapy for ulcers was a revolutionary change in ulcer treatment.
Revolutionary changes like these always take time to gain wide acceptance, since they are invariably initially met with resistance in scientific circles, often with strong and prolonged opposition. What does it take to accomplish a major shift in thinking or practice within the scientific community? This article is an attempt to shed light on the answer to this question by providing a contemporary example of a new and better scientific theory that has failed to displace an older theory. /15/ This will be discussed in light of Kuhn’s insights. (Doubtless the reader will at times be dismayed over specific information that is not included herein. For that I apologize. I try to provide enough information for you to find any additional information you desire.)
Now that you understand the CCVT and appreciate Kuhn’s proposal that progress in the natural sciences is revolutionary, let’s consider a body of research that challenges the CCVT.
Land’s Theory-Challenging Experiment
In the late 1950s, physicist and inventor Edwin Land published the results of some novel experiments—results that the CCVT could not explain. Land used black-and-white transparency film /16/ to photograph a highly colored still life scene. /17/ (See Fig. 3.) As he photographed this scene, he held a piece of red cellophane in front of the camera lens (call this picture the red-record). Then he photographed this scene again, but this time he held a piece of green cellophane in front of the camera lens (call this picture the green-record). /18/
When these two transparencies were projected, they appeared to be very similar. One might be a little darker or lighter in a particular area, but they were just B&W slides. Land then put the red-record in one projector and the green-record in another projector and superimposed the two on a screen (see Fig. 4). Then he placed the red cellophane in front of the projector with the red-record and the green cellophane in front of the projector with the green-record. /19/ What did he see on the screen?
At this point, rather than tell you what Land saw, let’s consider what the CCVT would predict. Mixing colored lights is not the same as mixing pigments./20/ Imagine two projectors with no slide in either of them, but casting two superimposed beams of white light on the projection screen. Any lighting technician who has ever operated theater spotlights can tell you that if you hold a red filter in front of one and a green filter in front of the other, the superimposed area on the screen will appear yellow. Thus the CCVT would predict the image of the two superimposed B&W slides (described in the previous paragraph) with their respective red and green filters would produce a black-and-yellow image (though with various shades of black and gray and various shades of yellow).
But that is not what Land saw. Instead he saw a full color image—and the colors were true to life! /21/ How could B&W slides have encoded in them the information for color? Since the CCVT could not explain this phenomenon, Land developed what he called the "Retinex Theory" of color vision. /22/
A full explanation of Land’s Retinex Theory is beyond the scope of this article, and readers interested in knowing more are directed to Land’s articles listed in the bibliography. Suffice it to say here that Land proposed that the information people require to see color is not in the wavelength of light reflected itself, but that color information is encoded in the ratio of the longer and shorter wavelengths of light reflected by an object./23/ He wrote, "It appears…that colors…arise not from the choice of wavelength but from the interplay of longer and shorter wavelengths over the entire scene." /24/ He continues, "It turns out that there must be a certain minimum separation between the long-record wavelength and the short." /25/
Though specialists in color vision readily admit to the validity of Land’s Retinex Theory, the CCVT persists not only in textbooks, but also among many professional scientists. Montgomery summarizes the challenge to the CCVT in these words:
So the trichromatic theory is right—as far as it goes. The three types of cones together identify the wavelength of the light coming from each point in the field of view, as Young and Maxwell predicted. But that alone can’t constitute our perception of color. As Land’s 1955 projection experiments showed, there is no point-for-point correspondence between color and wavelength. If there were, the colors of objects would not be constant; they would change continually as the illumination changed. /26/
Why has Land’s theory not found its way into the general knowledge of the scientifically literate? Why is it not routinely taught in science classrooms? You would probably have to ask many practicing physicists before finding one even familiar with Land’s work. Does this indicate his Retinex Theory wasn’t "scientific" or that Land’s work was fatally flawed or not "good science"? Is this the reason that his work has not precipitated a paradigm shift?
Why Has Land’s Work Failed to Cause a "Paradigm Shift"?
Did Land’s research contain a fatal flaw? Let’s list possible causes why his theories have not been widely embraced by the scientific community and then respond to these suggested reasons.
1. Land lacked the appropriate educational or scientific credentials. I do not seriously entertain the suggestion that Land’s work was not embraced by the scientific community because he did not have his "union card"—a Ph.D. in a field of science. Several of the paragraphs that follow more than substantiate my rejection of this suggestion.
2. His work was not "good enough" to be published in respected scientific journals. Although Land had not earned a graduate degree in the sciences—he did not have even a bachelor’s degree!—the most prestigious, peer-reviewed scientific journals published his work including Proceedings of the National Academy of Science (USA), Proceedings of the Royal Institution of Great Britain, Nature, Vision Research, Science, The Journal of the American Medical Association, American Scientist, and Journal of the Optical Society of America (see bibliography). He also twice had major articles in the Scientific American, a magazine that is highly regarded by scientists—though it is less scholarly and more popular. So we cannot conclude that his work has not achieved acceptance because it was not disseminated in respected scientific journals. The experts in the field deemed it worthy of publication.
3. His work was not built on the work of other, respected scientists, and no one could duplicate his work. Land’s experiments, as noted earlier, were built on the work of another scientist (Maxwell), and other scientists have duplicated Land’s work. /27/ For example, Robert Szabo, a graduate student at Cleveland State University, duplicated Land’s experiments./28/ Again, we cannot conclude that the Retinex Theory has failed to gain acceptance because it was not built on the accepted work of other scientists or that others were unable to reproduce Land’s work.
4. No practical applications have been found for his work. Through the years many researchers in the field have recognized the significance of his work. (Years later, Mark Ruzon—then a third-year Ph.D. student in the Computer Science Department at Stanford University—described Land as the scientist "who, in proposing [the] retinex theory in the 1950s, revolutionized the entire field in a single stroke." /29/) But the ultimate affirmation of Land’s research and theory is in the various ways other scientists have applied it, like NASA’s use of Land’s theory to develop a method to enhance photographs. /30/ In describing their work, these NASA researchers said, "Our starting point was the last retinex concept proposed by Edwin Land…." /31/ On their web site, Zia-ur Rahman and his colleagues at NASA provide a partial bibliography of scientific articles describing some of the applications of the Retinex Theory. /32/ So the reason for Land’s work being ignored is not because no one has successfully applied it to real life problems.
5. No other experimental research supports his theory—maybe there is no way to definitively answer the question. Why doesn’t someone just measure the intensity of the different wavelengths of light coming from a colored surface? This should settle the issue once and for all. Montgomery reports that
Land built a photometer (a light-measuring device) that tests this theory. For each of the three wavelength bands, the photometer measures the intensity coming from a small spot in the visual field and from a large surrounding area. It then computes a weighted ratio, giving more importance to the spot’s immediate surroundings than to more distant parts of the field. On the basis of the three ratios, it assigns a color to the spot. Sure enough, the machine-calculated color closely matches the perception of human observers. /33/
What do these results mean? Montgomery continues:
That suggests, although it does not prove, that human beings make similar calculations. But if so, where and how? Land has never focused on that question; he is a not a biologist. /34/
But Montgomery reports that other researchers have also found biological evidence to support the Retinex Theory.
For Land’s theory to be true, the visual system had to include a special kind of nerve cell. The cell had to receive information from a patch of cones on the retina and then compare the light coming from a central spot in the cones’ visual field to the light coming from the spot’s surround. Such cells exist. In 1968 Nigel Daw first found them in goldfish, which also have trichromatic color vision. /35/
As biologists continued their research, they next found these specialized nerve cells in macaque monkeys, then in other primates, and finally in humans. /36/ Thus we cannot conclude that this theory has failed to gain wide acceptance because other, necessary, supporting evidence has not been discovered.
6. The amount of evidence for the CCVT is greater than for the Retinex Theory. Just the opposite: the weight of the evidence seems to favor the new paradigm. The Retinex Theory explains everything the CCVT explains and more.
7. His theory is too narrow. We also cannot conclude that the Retinex Theory is useful only in explaining this one set of experiments—all else is more than adequately explained by the CCVT. As a matter of fact, what is probably the first observation that could not be explained by the CCVT was made in 1672 by Otto von Guericke. He noticed that in the early morning hours, as light comes through a window and other light comes from a candle, an object (e.g., a hand) casts a blue-green shadow (see Fig. 5). The CCVT would predict only a colorless, gray shadow, but Land’s Retinex Theory provides an excellent explanation for the colored shadow. /37/ The CCVT is arguably narrower than the Retinex Theory.
Then what is the answer? If Land’s work had all of the characteristics of excellent science (and it does), then why does an examination of college physics texts produce no work that presents Land’s Retinex Theory? Why will many college physics graduates know nothing of the Retinex Theory? /38/ Why do encyclopedias continue to present the CCVT and ignore the Retinex Theory? (For instance, in its article on color, the Microsoft Encarta 98 Encyclopedia seems to present the CCVT in its animation, but mentions Land’s work and theory in the accompanying article. /39/) Why is it that Land’s theory is taught only in some graduate courses and not in high school or college physics courses? /40/ Why is it that only a handful of specialists are familiar with it? Land’s experiments were "successful," but they failed to provoke a paradigm shift. Why? Kuhn’s explanation of how science advances can inform our thinking here.
A Kuhnian Explanation
Kuhn argued that when observations are made that cannot be explained by the prevailing theory, these anomalies produce a crisis. When and if the crisis is great enough—and when a new theory emerges that cannot only explain these anomalies, but also continue to explain everything the prevailing theory explains—a paradigm shift occurs. /41/
Using Kuhn’s argument, we might conclude that, first, there are not enough anomalies that the classical theory cannot explain (even though that particular CCVT explanation might not be true). Without a doubt, the number of anomalies it takes to produce a crisis of sufficient size to precipitate a paradigm shift varies from one situation to another. There is no "magic number," no necessary and sufficient quantity. It is easy to conclude that if no paradigm shift occurs, then the crisis must not have been of sufficient magnitude—but doesn’t this beg the question?
Second, even though the new paradigm may explain "real-world" (vs. the laboratory) color vision better, high school students and college science majors typically study light and color perception only in the lab. Thus, the CCVT "works" for them. In the places and ways that they study and use it, they avoid the anomalies. And this has permitted the two paradigms to coexist for at least 40 years.
Finally, perhaps our question of why there has not been a paradigm shift can in part be answered by an awareness of who scientists are and how they understand and use the scientific method (praxis).
Implications for Understanding Science
Do scientists have an accurate concept of "the scientific method"?
What significance does all this have for someone studying philosophy of science or the interface between science and faith? In Kuhnian terms, Land’s work serves as an illustration of a new paradigm failing to replace an older one. The fact that a newer better theory can fail to replace an older, poorer theory gives insight into the process of science.
Most scientists embrace a number of premises as characteristic of the scientific method—premises that are not necessarily true. Scientists like to think that they approach their work with no presuppositions and that they accept nothing on faith, but validate everything for themselves; but this is not true. /42/ Scientists claim that their method requires that they hold their conclusions quite tentatively—always ready to quickly discard them and embrace another when the empirical evidence warrants. The examples described above help to illustrate that at times scientists violate these principles. Sometimes some of them are unable or unwilling to abandon a theory even when the evidence clearly favors another explanation.
Finally, scientists like to claim that the method of science is always "self-correcting." That is, even if practitioners of the scientific method arrive at a wrong conclusion, they—or their successors—will ultimately correct their mistakes. It has been 40 years since Land’s work was first published. During this time its observations have been published in the "right journals," corroborated by other researchers, and applied to practical situations. How long should it take for scientists to experience a paradigm shift and include his work and theory in their "color vision grid"?
Are scientists really any different from other people?
Perhaps the failure of Land’s work to provoke a paradigm shift illustrates that science is not always method-driven, but at times is people-driven. (Of course, scientists like to imagine that their work is totally the former.) Perhaps teachers/professors find the older paradigm easier to teach; certainly it is simpler, and it is so logical. Ockham’s Razor might coax some people to favor the CCVT over the Retinex Theory, but this principle should be considered and included in the decision making process only when everything else is equal—and this is not the case here. /43/
Regardless of how objective and "ideal" the scientific method may be, the scientists who use it are all human and subject to all of the frailties of the human species. This is substantiated by the fact that we periodically read news accounts of scientists who have acted dishonestly or with bias. /44/ In a similar way, scientists can be overly reluctant to embrace a new, better explanation because of time and energy already invested in an older theory and the work involved in restructuring her "grid" in order to accommodate the new theory.
Another common misconception closely related to the first is that scientists are not biased, but form their conclusions solely on the basis of the empirical data. This example illustrates that scientists do not always take the time to examine new data in order to evaluate new theories.
A crisis of sufficient magnitude has not arisen—indeed, many perceive no crisis at all! Most scientists find that the CCVT still "works," so there is no great impetus to abandon it and embrace a better theory.
Summary and Conclusion
The failure of Land’s experiments and Retinex Theory to provoke a paradigm shift illustrates the nature of scientific progress and the way scientists work. Not only do non-scientists typically not understand the revolutionary nature of the scientific process, but professional scientists often also share this ignorance. Many—perhaps most—practicing scientists do not recognize the philosophical foundations of their work.
The epigraph on the first page of this essay quotes Steven Schafersman’s opinion that science is the only method for the acquisition of reliable knowledge about nature. An examination of the history of scientific discovery reveals that often scientific knowledge was advanced, not by the “scientific method,” but by other “means” (at times, serendipity). Naturalists would quickly declare that until people using the scientific method confirmed these discoveries, it was not “reliable knowledge.” Further, many natural scientists see the physical realm as either all that exists, or all that we can acquire reliable knowledge about. Realizing the limitations of the scientific method and its practitioners should help Christians realize that our faith is not at the mercy of people who posit that the physical realm is all that exists or all that we can know exists. Such a statement is a faith statement; it can neither be proven nor refuted by the scientific method or any other method.
Unquestionably, since science is a process, we are often unable to determine when the process has reached “the end,” that is, “truth,” and when it is just a work in progress. In virtually every scientific discipline, we cannot know when to expect the next paradigm shift—and we have no way of guessing in how many areas a shift should have already occurred, but hasn’t yet. With this in mind, we Christians are wise to be less concerned with harmonizing our faith with the latest scientific thinking and more concerned with keeping our faith Bible-based and Christ-centered.
Supporting Materials
The bibliographies that follow are just a small portion of articles, books, and Internet sites that have information about ideas discussed in this essay. Some of the web sites have bibliographies citing publications either supporting the ideas presented here or demonstrating applications of these ideas.
Dr. Steve Badger 730 South Duke Springfield, MO 65802 417.865.2815 x8327
