Minnesota Center for the Philosophy of Science, University of Minnesota, Minneapolis MN 55455, USA email@example.com
History is not a stranger to the science classroom. What student does not learn of Darwin and his voyage on the Beagle, Mendel and his pea plants, Newton and his three laws, Mendeleev and his stunning predictions based on his periodic table, or Wegener and his underappreciated insights on continental drift? History often serves as an occasion to organize the serial development of concepts, to reconstruct reasoning, to celebrate scientific discovery or to bring anecdotal humor into a lecture. Teachers have used experimental simulations on historically faithful instruments (Conant 1957; Heering et al 1994), portrayed historical characters (Eakin 1975), and introduced the social and ethical contexts of science through case studies (Aikenhead 1991; Hagen, Allchin and Singer 1996). The value of history as a tool in science teaching is well documented (Matthews 1992; 1994; Hergit 1989; Hills 1992).
Yet history, like science, can be biased. Science has been used historically to justify political ideologies and power relationships between sexes, races, economic classes or other groups (e.g., Fee 1979; Harding 1992; Haraway 1991; Jones 1981; Gould 1981; Lewontin, Rose and Kamin 1984; Shapin 1979; Kamin 1974). Likewise, history can be shaped or distorted. Selective history can appear to justify certain views of science--or to support certain pedagogical models. These applications of history of science become misleading--not just about history, but about the process of science itself. In what follows, I discuss some of these potential pitfalls. Teachers must learn to use history--like any tool--appropriately, in order to be effective. Most importantly, teachers should aim to respect the full historical context of science, lest they betray the very subject they hope to enlighten.
AAAS's Project 2061 (Rutherford and Ahlgren 1990) is sometimes cited as a progressive standard on historical perspectives in science education. Yet it also reflects conservative practice. Its ten episodes celebrate scientific contributions and great scientists. Science is thus primarily product, not process. While the standards aim to portray scientific knowledge as fallible, they also convey its authority and disguise the difficulties of exposing error. Likewise, they portray science as human, but also as relying on a few exceptional, superhuman individuals (who are almost exclusively Western males). Such uses of history have deep philosophical and ideological overtones.
The sometimes subtle or indirect ideological role of history in science teaching is even better exemplified, however, in a textbook by Nason and Goldstein (1965). The authors give two accounts of the work by Jean Baptiste van Helmont, a Dutch physician from the early 16th century. Their treatment is noteworthy because they portray the same historical figure differently in two separate contexts in order to convey the same implicit "lesson".
In the first example, the authors describe an "experiment" devised by van Helmont to demonstrate the spontaneous generation of mice. The procedure was to throw some grains of barley in an old rumpled shirt in a damp cellar: one should return several weeks later to find that the grains of barley have been transformed into baby mice. The lesson, of course, is clear, even to the unsophisticated science student: van Helmont was another ignorant scientist of the past who failed to understand the "obvious" origins of living matter and the need for controlled experiments. Now we know better. Science progresses.
In the second example, the reader learns about van Helmont's renowned willow-tree experiment, sometimes hailed as the origin of experimental plant physiology. Van Helmont weighed a willow sapling along with a 200-lb. potful of soil, planted the tree, and five years later weighed the two again. The tree had grown an impressive 164 pounds, while virtually all the soil remained. Students are to see how elegantly van Helmont showed how the matter of a tree does not come from the soil, but from a gas, a term which van Helmont himself had coined. Here, van Helmont is the hero, not the fool. And from his example, we might draw the deeper lesson: construct a test, quantify, measure--and be prepared in some cases to be patient for the data. Contemporary students are sometimes even guided to repeat this lesson for themselves, albeit on a smaller scale, using radish plants, whose weight change can be observed in weeks rather than years (Hershey 1991). Identify the correct procedure and the right answer inevitably follows. That's how science progreses.
This pair of examples is historically outlandish. Van Helmont is both praised and ridiculed as a scientist in the same text. He is both hero and fool--and what matters is whether van Helmont was "right" or "wrong" (Gould 1974). In each case, the status of the answer today guides the judgment about method: in one case, the method must have been misguided; in the other, it is necessarily exemplary. The text borrows implicitly on the notion that scientific method is algorithmic and leads inevitably to the truth. Therefore, the right answer will always reflect the right method.
Similarly, if Newton dabbled in alchemy (Dobbs 1975), or Darwin espoused non-Mendelian heredity or had strong theological beliefs (Richards 1987), we kindly excuse our heroes and brush the anomalies aside because they were ultimately "right" about the "important" things. We willingly explain away or excuse their flaws in order to maintain their integrity as (infallible) scientists: faulty conclusions cannot be part of the process of science.
The examples grossly misrepresent van Helmont, however, and subvert what we might want to convey about the nature of science. In the willow tree case, for example, van Helmont concluded that the bulk of the mass of the tree had come--not from carbon dioxide, a substance wholly outside his conception--but from the water that had been added to the pot. Van Helmont had a rather elaborate world view which included the notion that there was only one primal element--water--from which all other matter derived. In this, he challenged the existing Aristotelian doctrines that there were four elements--water, earth, fire and air. The tree experiment was essentially designed to show, then, that the tree was not earth mixed with some fire. Van Helmont provided an alternative explanation consistent with observations: namely, plants need water to grow--a "common sense" notion, but now formalized by experiment. In its intended role, the experiment was dramatically successful--especially in provoking others to think about the problem and in some cases to repeat the experiment. If the lesson is to be about the process of science, one needs to respect the actual history.
In a recent critique of van Helmont, science educator David Hershey (1991) also challenges easy historical interpretations. Yet at the same time, he views the experiment retrospectively. He frames what students can learn about experimental design, execution and analysis based on what we know today. He claims, for instance, that students can understand how van Helmont performed the "wrong" experiment. To assess his hypothesis about the role of water, he should have grown the willow hydroponically--that is, in water alone. And he should have used distilled water, so as to exclude the role of minerals in the water. If he had done this, Hershey notes, van Helmont would have observed that "willows do not live by water alone." Here, the pedagogical strategy was to learn by "correcting" the history and making it come out "right".
While Hershey's historical motives are noble, he misses the most fundamental aspect of history: the historical context itself. Concerns about distilled water in the context of an experiment done centuries before anyone understood the concept are grossly misplaced and distort the process of science. Van Helmont was also probably well aware that plants do not grow outside soil. He even buried his pot in the earth, as if the location was a significant parameter to keep constant. There was certainly no existing evidence then to suggest that the substrate of soil was not relevant in some respect. Indeed, the lack of substantial soil loss, even though the soil was present, was integral to van Helmont's reasoning. We might even credit Van Helmont for his experimental strategy in isolating the relevant soil system within the boundaries of a pot.
Hershey also notes that van Helmont failed to replicate his findings. Replication is an important scientific lesson, but one developed only long after van Helmont. Applying the standard to van Helmont mixes contexts inappropriately. All scientists work with limited knowledge--and what is important is how they deal with those limits and extend their own knowledge.
In the view that Hershey merely epitomizes, one assesses van Helmont's experiment or other historical scientific work based on some current idealized model of science. In a view that is sensitive to historical context--and thus to science as scientists practice it--van Helmont's experiment was well designed and interpreted appropriately in the context of its own time. His conclusions were nonetheless later construed as "wrong". A well designed or valuable experiment need not always yield an answer that is completely "right". Indeed, how much more would a student learn about the nature of relaibility and fallibility in science by seeing how we can get "wrong" answers using the "right" methods?
A healthy addition to the Project 2061 (or any) standards, then, would include an episode where conclusions we now reject were once embraced by science. Students need to fully appreciate the original historical context of those conclusions to understand how scientific knowledge can shift. Science viewed prospectively is different than science viewed retrospectively (Latour 1987).
The distinction between retrospect and prospect in both history and the process of science is fundamental. I want to highlight this distinction in teaching by examining a sample exercise that draws on history and thus resembles an historical simulation. Yet because the actual history is reconfigured to serve other purposes, the historical context is lost, and with it some important lessons about the nature of science. Indeed, its incidental distortions can mislead students.
The sample exercise (Duschl 1993) focuses on the 'Causes of Earthquakes' and aims to help college students understand the reasoning supporting different explanations and develop skills in comparing and evaluating theories. Students are first introduced to five alternative explanations for earthquakes introduced historically; they use original documents where possible. Next, they are given original seismic data to transfer onto world maps and analyze. Finally, they construct a decision table, assessing the evidence for each theory according to several key questions.
Now, the value in this exercise is obvious: students come to appreciate that different explanations of the same phenomenon are possible. They see how each theory is rooted in certain assumptions. The perceptive student will recognize how theoretical evidence is built through field work and the creative act of interpreting data. And students gain 'experiences with the higher cognitive reasoning skills associated with the evaluation and interpretation of knowledge claims' (p. 190). These exemplify potential benefits from an historical approach.
But while the earthquake exercise may seem like a fragment of history for the classroom, it differs from real history profoundly. The teacher--and probably the students--know what the "right" answer is. Do students have the freedom to make the "wrong" theory choice? Further, when students do know the answer, they typically shape their reasons accordingly: they learn how to conform reasoning to a conclusion, rather than how to reason towards an unknown conclusion. Of course, scientists don't know the answer in advance. The lessons about historical uncertainty and "science-in-the-making" (Latour 1987)--and the very role of evidence in resolving that uncertainty (ostensibly the main aim of the exercise)--are utterly lost. The exercise reconstructs rather than simulates science.
The exercise is further reconstructed--or artificially contrived--in its pre-established data set (Part 2). Why this data? Students receive only a small part of evidence is relevant, once they know the multiple explanations; and the data is largely significant only from the current theoretical perspective. Students should be challenged to think critically about what information is relevant, especially in the context of other theories. In this case, evidence simply appears, deus ex machina, from the teacher. Students may thus tend to regard reality as presenting itself preformed; they can perhaps develop a habit of passively accepting whatever evidence is already available. History in context, by contrast, richly illustrates the role of experimental design and creative purpose in data collection.
Finally, students are asked to choose between the theories (Part 3) according to pre-structured standards. Most notably, the exercise organizes each idea about earthquakes into comparable 'Giere frameworks', with clearly stated hypotheses, background assumptions, initial conditions and premises (Giere 1984). Scientists rarely work with such formal reasoning: why not? An exercise that upstages scientific practice should be viewed skeptically. In addition, no one in history sat down to consider these five theories all at once. Each explanation was a response to extant data and concurrent theories in their own time. The exercise thus models some idealized reasoning process where perhaps it should be recreating for the student the context of an actual scientific judgment. A rational reconstruction is not a historical simulation.
The problem with teaching through rational reconstructions is that the history--and thus the process of science--is backwards. The aim is to find the route to (from?) the final answer. We should, instead, be tapping history to model the blind forward-moving context of science. The generation of hypotheses, the search for relevant information, the design and critique of experiments, the elaboration of alternative explanations, the struggle with experimental anomalies--all the elements of scientific discovery--cannot be taken for granted. There is more to science than just justifying the final outcome--or assuming that it is correct.
Historical simulations, guided by sensitivity to historical context, and rational reconstructions, highlighting the path that leads to and justifies current theory, may each appear historical in nature. But the rational reconstruction is not concerned with the process of history, only its product. The names and dates that are assigned to specific ideas are merely incidental to its purpose. Rational reconstructions are ultimately ahistorical in nature. Where the aim is to convey process of science, by contrast, teachers should highlight the historical context of a case. Ideally, students should have the sense that they are peering over the shoulders of working scientists, while having the opportunity to make decisions in their stead.
Historical case studies provide effective models for learning process of science. It may seem, then, that the whole history of science reflects a pattern or outline for modern classroom learning: history as a syllabus. Advocates of this view adapt Ernst Haeckel's biogenic law that 'ontogeny recapitulates phlyogeny', suggesting that 'learning recapitulates history' (call it, perhaps, the model of 'cognitive recapitulation'). While the parallel is intuitively attractive, however, it is also misleading. It, too, erases historical context.
Constructivist approaches to learning emphasize that an individual builds cognitive structures on earlier ones. Where history shows how complex concepts emerged from simpler ones, therefore, the cognitive recapitulation model can surely be an effective guide. But this can apply only within a family of related concepts. It can only suggest a format for addressing a series of atomic models, say, or a set of variations on basic Mendelian or electromagnetic themes: one 'constructs' the complex on (or from) the basic.
The reasoning behind the cognitive recapitulation analogy, however, does not allow one to apply it more broadly. The so-called Chemical Revolution, for example, seems to offer a model opportunity to teach about 'revolutionary' conceptual change. Historically, Lavoisier's discovery of oxygen and its role in combustion dramatically replaced earlier explanations using phlogiston. Yet in one set of high school chemistry classes I observed, students learned about phlogiston after they had already learned about oxygen and the modern system of elements (Allchin, 1997). The unit focused on metals and the (oxidation and reduction) reactions that transform them. By observing phenomena highlighted by late phlogistonists (Allchin 1992b), students found that the concepts of phlogiston and oxygen each addressed important, though different aspects of combustion. One accounted for the material changes, while the other accounted for energetic changes (in today's terms). By approaching the topic "backwards" historically, they did not encounter a revolution. Instead, they found that the two concepts were complementary. As illustrated by this case, history is not an infallible standard for conceptual development.
The important element of history is its context. Thus, sometimes it may even be appropriate to disturb or modify actual history to preserve the sense of its context. In reaching his conclusions about the circulation of the blood, for example, William Harvey dissected many animals. Yet he never examined squid, so far as we know. But the organization of a squid's blood vessels are excellent for such an investigation. Squids have three hearts: two gills hearts (pumping just to the gills), and one body heart. The separation of the hearts and their arrangement makes it easier to conceive of the human heart as two hearts, joined in a circuit. Thus, a squid can be a good example for reasoning in the context of Harvey's other observations, though Harvey himself never did (Allchin 1993).
In a similar way, the lesson on metals and oxidation-reduction reactions (above) drew on the pyrotechnic thermite reaction, discovered only in the century after phlogistic doctrine reigned. Yet the teacher confidently interpreted it from a phlogistonist's perspective as the transfer of phlogiston from aluminum to iron. Here, the simple concept of phlogiston helped underscore the notion that the release of phlogiston (=electrons? energy?) from one metal was paired with its acceptance somewhere else: oxidation and reductions reactions are coupled. Historical authenticity was subservient, here, to the historical context of the phlogistic way of thinking.
On another occasion, I was challenged to teach about the Copernican Revolution. But how can one convey the magnitude of this reconceptualization to 20th-century students so fully indoctrinated into thinking (knowing?) that the Earth travels around the Sun? It is hard for them to imagine otherwise--which was exactly the core of the lesson. My strategy--designed to revive a sense of the historical controversy--was literally to upend history. I offered evidence, much of it presented in the late 16th and early 17th centuries, that the Earth did not move, as the students insisted. For example, a ball dropped from a high tower falls straight down; the Earth does not move away underneath it as it falls. The sun clearly rises in the morning and sets in the evening, moving across the sky during the day. Should not a good scientist rely on observational evidence? The students were adept at describing the current interpretation as an alternative, but they had trouble justifying it. Their frustration and annoyance that someone would question "common sense" knowledge, I could point out, paralleled feelings about Copernicus' views in the 1600s. The lesson was about process of science, not history itself. The method thus aimed to appreciate the history, not to repeat it.
There are other ways in which the cognitive recapitulation model is limited. Fundamentally, the analogy is weak. It assumes that the growth of scientific knowledge is linear--and that the sequence of history is one sequence, not a collection of multiple overlapping and interacting sequences. The context of co-temporaneous theories in other fields, though, is an important element of science. The context of geophysics, for example, greatly affected judgments about continental drift based on biogreography and structural geomorphology. Likewise, Kelvin's argument about the impossiblity of biological evolution based on thermodynamics was effective (for some) until the discovery of radioactivity early this century. It is primarily because of these disciplinary and theoretical interactions that extracting historical arguments from their historical context distorts the process of science (§§2-3). History may offer clues about small scale or local conceptual development. But when the historical context is lost, so, too, is much of the science.
For some educators, history not only suggests what to teach, it also tells us what not to teach. Teachers should avoid scientific errors at all costs. The concepts are, after all, "wrong": how can one learn from them? Of course, most "mistakes" were once considered "right": why?
How could knowledge have been transformed from fact into error? History helps us understand, first, the evidential contexts in which "wrong" ideas were once considered "right" and, second, how (and why) such contexts changed. History thus shows how the process of science can sometimes lead to "wrong" conclusions, while also leading to "right" ideas. Addressing the problem of error historically, therefore, is central. It contributes to understanding the nature of scientific justification, as well as its limits.
Consider, for example, the case of teaching the fluid model of electricity. Stocklmayer and Treagust (1994) strongly criticize the use of water analogies in textbook presentations of electrical "current". Electricity is not a fluid, after all. They justly question the authority of history in setting an agenda for teaching (§4). Still, many outstanding scientists, such as Franklin, Ampère, Coulomb and Cavendish, did (in their own time) consider electricity to be a fluid--or perhaps two fluids(!). The fluid concept must have had validity in at least some context. Of course, to recover that context, one must draw on history.
Fluid models of electricity are, in fact, generally appropriate for considering the "flow" of electricity in circuits. They are not appropriate when one also wants to consider other electrical phenomena, such as field effects or the action of individual electric particles. The fluid model has a specific scope or domain of application within which it is justified. Outside that particular domain, the justification fails. For many persons, in fact, this domain is all they will encounter in their daily lives: the fluid model of electricity actually suffices for them. But the deeper lesson about science is how, historically, that domain was found to be limited and the fluid model declared "false". The fluid model of electricity can be both "right" and "wrong" at the same time, depending on context. The tension between these two perspectives is why students can benefit from studying the fluid model, even as a "wrong" idea.
It may seem perverse to teach "wrong" ideas. Yet without a full understanding of error, students cannot learn how to distinguish between simple ("wrong") scientific models and more sophisticated ones. Consider, again, the exercise on phlogiston introduced above. What was the value in teaching a concept that was abandoned two centuries ago? Here, as in the case of the electrical fluid-model, the historical perspective offered a simple, general framework for thinking about a group of causally related phenomena. Students could use phlogiston to map the relationships burning, calcination, rusting, tarnishing, corrosion (all oxidations), reduction and photosynthesis (or, as the students recognized, 'reverse combustion')--all on a macroscopic level. They eventually contrasted their understanding in these terms with knowledge they had gained earlier on electrons and emission spectra. They saw how the simple concept was limited--even misleading by today's standards--yet fully justified within a certain domain. They know what it means to call the concept of phlogiston "wrong".
The process through which scientists determine and deal with error may at first seem peripheral to science. Yet the characterization of fact versus error is clearly not, and the occasions in history when one is transformed into the other are central to deciphering how science works (see Latour, 1987, on opening and closing 'black boxes'). The reasoning patterns about simple versus complex models are critical, for example, for assessing claims about global climate change (what variables are included in the modeling) or conclusions from DNA fingerprinting techniques (based on domain of reference populations).
In other cases, it is the "wrong" idea itself, rather than the process, that is important. Students typically bring to science courses preconceptions that often linger after they leave the classroom, even when taught the "correct" concept. Teachers who know the historical context of such ideas can appreciate more vividly exactly how these perspectives offer a way of interpreting the world. They can validate such views, even while helping students work towards a deeper understanding. Teachers may find it appropriate, therefore, to acknowledge all that was "right" in Lamarck's ideas before turning to Darwin's criticisms. If Darwin's arguments are ineffective, then perhaps we need to reevaluate our own conclusions. Similarly, our culture supports a host of assumptions about brain size and intelligence that once dominated nineteenth-century anthropology. It may be worth recognizing how scientists once justified their views on this subject as a entry into investigations that transformed those views historically. Ultimately, in order to fully understand why scientists now endorse certain concepts, students must also understand why other alternative explanations (perhaps their own) are "wrong". Teachers must start from the historical context.
The lessons from exploring error can sometimes be unexpected. Consider, for example, the case of nineteenth-century craniology (Fee 1979). Several anthropologists "knew" that women were intellectually subordinate, but as scientists they felt the need for evidence. And so they began measuring skulls. This was not pseudoscience or a shadow of science. It was science par excellence--what Elizabeth Fee has called 'a Baconian orgy of quantification'. When their measure of cranial volume as a standard began to suggest that elephants were more intelligent than humans, however, craniologists retreated to measuring brain-size-to-body-weight ratio. When that, in turn, gave birds, anteaters and bear-rats the intellectual advantage, they finessed the problem yet again, developing various other measures of facial angles and cranial indicies. Not until two women entered the field did the study of sex differences abate. Alice Lee and Marie Lewenz published data for individuals, showing that many women had cranial capacities larger than some anthropologists in the field. They applied statistics more rigorously, showing that mean differences were well within sampling error. Their science was "good" science, but why had it escaped men working on the same problem for several decades? Fee suggests that the errors were exposed or realized only when women, for whom the conclusions mattered most, became part of the scientific community. The philosophical lesson appears to be that reliable scientific conclusions depend in part on who participates in the science (Longino 1990; Harding 1992). That would be a hard lesson to learn without a striking historical example, such as this, where gender-biased error was exposed and corrected by a complementary gender-biased critique.
The exploration of "wrong" ideas is potentially far-reaching. For example, some educators would banish astrology, alchemy, phrenology, craniology, mesmerism, &c., from the science classroom because they represent mistakes of science. Some contend that even mentioning such "unscientific" or "pseudoscientific" practices gives them unwarranted credence. Historically, of course, each of these practices was once considered science--in some cases, exemplary science. It is hard to imagine how we should expect students to "know better" than these scientists without teaching them why. What has changed? If testability or falsifiablity are benchmarks of modern science, for example, then students should discover, as scientists did historically, how those philosophical principles are important. By tracing the historical context of "wrong" ideas, students learn what makes science 'science'.
Understanding historical context is important for appreciating how scientific concepts change and, in some cases, for understanding fully those concepts themselves. This involves some skill. In researching the history of science, historians frequently encounter ideas from the past that strike us today as strange, "wrong," or even incomprehensible. Yet historians endeavor to make sense of apparent absurdities: like scientists, to find underlying order amidst apparent chaos. They examine historical context--available theories, previous ideas, personal perspectives, styles of reasoning, records of observations, &c.--and then apply their imagination in reconstructing what the historical character was likely thinking, how the strange or outlandish (to us) may seem obvious and natural given the right context. Historians call this interpretive tool 'the historical imagination'.
Yet the historical imagination is a skill that is not limited in importance to historians only. It can be important to practicing scientists, as well. The creative interpretation need not be historical, though, but involve any strange theoretical perspective. While science-as-taught usually involves items of common consensus, science-as-research typically involves dissensus and multiple possible concepts. Indeed, much research is motivated by the need to resolve such uncertainties in interpretation. In this arena, new ideas can seem as strange or awkward as old ideas. To participate in understanding new conceptualizations, researchers must sometimes be able to exercise their imagination in reconfiguring data from one familiar scheme to a novel one proposed by another researcher. Interpreting historical context in science is practice for interpreting alternative new theories.
Skills in imaginative (re)interpretation are especially critical where scientists disagree or differ in their interpretations of the evidence (which is nearly always). In many ways, the task of persuading a colleague with a different interpretation is akin to the task of the educator: how does one deal with possible "erroneous" preconceptions and introduce the right information in just the right way so that the intended audience can grasp the concept or relevant conclusion and be convinced by the evidence? Scientific debate really amounts to an effort at mutual teaching. Especially where the disagreement is deep, one must frame the argument appropriately (Allchin 1992a). The interpretive imagination is a skill for assembling a line of reasoning that has at least the potential to be effective. In science, you must be able to persuade your critics. You must transform their beliefs. But scientific controversies are not resolved where scientists 'talk past each other' (Kuhn 1970). Scientists thus need to exercise their skill in understanding alternative theoretical contexts sympathetically in order to participate effectively in scientific dialogue. Again, historical imagination is a skill that fosters good science.
Science teachers (and scientists, too) often enjoy their profession because our culture associates science with certain knowledge. Such teachers typically make awful historians, though: they tend to view all science in the past as either triumphant discovery or pathological error. All history is filtered through the lens of current knowledge, disregarding the context in which scientists worked. Professional historians of science call this kind of history 'Whiggish', and disparage how it obscures the actual process of science. Whiggish history is a danger in science education, too, for the same reason. History that is informative, by contrast, is textured and rich in the context of scientific practice. Effective history illuminates the process of science.
What lessons can we learn about how to use history? Frist, we should acknowledge that education's use of history has the potential to be as theory-laden as science (Kuhn 1970). Good historians endeavor to be sensitive to the context of their historical subjects. Second, we should be as wary of cookbook history as cookbook labs. Neither science nor history follow prescribed paths. Historical simulations should thus allow students several degrees of freedom and promote genuine discovery. Third, as shown in the examples above, where we adapt the history for educational purposes, we should nonetheless maintain a fundamental respect for the integrity of historical context. Above all, in encountering science historically, we should listen carefully, so that we hear what history--and other scientists--try to tell us.
*An edited version of this essay appeared in Journal of College Science Teaching 30(2000): 33-37. A journal of the National Science Teachers Association