Physical and Biological Science as Iterative Cycles of Thought and Action

The episode that we recount here happened to Greenwood 25 years ago. He has retold the story often enough that, in the way of narratives, his recollection of it is as much tied up with the retellings as with the original episode. He did not document the episode with anything other than lecture notes because only on reflection over the years did the larger meanings become clearer. Still, Greenwood feels that he is being true to the episode that he participated in.

At the time this occurred, Greenwood was the chair of the Biology and Society Major at Cornell University. A program for students in their first 4 university years in the U.S. higher education system, this multidisciplinary, multicollege major was designed to link the basic biological and physical sciences with the social sciences and the humanities. It provided opportunities for students with a strong interest in the basic sciences to explore the social sciences and the humanities systematically. Greenwood was responsible for the core, upper-level courses that included an overview of the relationship between biology and society as well as discussions of science and scientific method.

Having taught this course several times, Greenwood discovered that good, advanced undergraduates with strong backgrounds in mathematics, physics, chemistry, and biology had very little in the way of concrete, behavioral under­standing of the scientific method. They were sophisticated enough at using the appropriate language to describe the rules of the scientific method, but they did not understand the scientific method as a form of knowledge-generating and reflective behavior. Instead, they used their notions about the scientific method mainly as a way of advocating scientific values about truth, objectiv­ity, and replicability.

On reflection, Greenwood realized that it was not really surprising because, by their third year at the university, most students had done only rote science work in the introductory courses they had taken. They had very little experience of science as a form of discovery and interpretation in a laboratory setting.

Although this situation was understandable, Greenwood found it unac­ceptable for the Biology and Society major. Many of the majors were preparing for careers in medicine or in other branches of health care in which their understanding of the scientific method as a form of behavior would have direct consequences for thousands of patients. He cast around to find some way to deal with the problem. He knew that, despite his good relations with the students, as a cultural anthropologist, his views about the scientific method would have little credibility to them. He thus decided to invite a Nobel Prize­winning chemist from Cornell to come to the class and lecture on the scientific method. He made this choice partially because he knew the scientist and partly because this professor was known to be an extraordinarily good and commit­ted teacher of science.⁶

The lecture given by the chemist lasted the standard 50 minutes, and the students were on the edge of their seats throughout. The prestige of this indi­vidual, combined with his congenial and down-to-earth manner, made the lesson effective for most present. It was clear at the outset that the students expected a very abstract and theoretical lecture from this eminent scientific intellectual. They apparently equated great science with great abstractions, very general laws, and big theories. What they got was something different. The chemist chose to describe his activities as a scientist and to bring the students into his world through a behavioral perspective, particularly through the per­spective of the principal investigator in charge of a scientific research project.

He began by pointing out to the students that the first issue in any scien­tific inquiry is to generate a problem to study. He explained that this is not a simple process. It was evident from the students’ reaction that they had not been asked previously to think about how scientists come up with questions to ask, probably because students are generally given a set of predigested ques­tions to address in their class work. The chemist pointed out that there are many problems in the world and many more suggested in the scientific litera­ture. Some of these are interesting to the researchers in question; some are not. ^^at is interesting, he argued, is partly a matter of personal preferences and histories. Also, some problems require equipment and funding that are not available; others touch on elements of previous experience that make them attractive or unattractive. Occasionally, an anomaly picked up through obser­vation generates a questioning process and a review of the literature that even­tually causes a group of people to decide it has a problem worth studying.

It was already clear at this point that the students were hearing ideas new to them. Most had not considered the matrix of ideas, experiences, organiza­tional structures, and histories that provide the context in which scientists ask questions. Yet the chemist’s statements accord well with studies carried out in the philosophy of science (Kuhn, 1962) and the social studies of science (Barnes, 1977; Barnes & Shapin, 1979; Latour & Woolgar, 1979; Rabinow, 1996; Traweek, 1992; Zabusky, 1995). There are few convincing accounts of the sci­entific problem generation process. The exception is a study by Paul Rabinow (1996) that addresses this issue effectively in relation to one particular discov­ery in recombinant DNA work. This subject is now a central concern of the field of science and technology studies (for example, Hess, 2001).

Problem selection tends to be bracketed under the headings of“individual creativity,” “genius,” and so on, converting science into a story of individual heroes that, we note, is a story with a hierarchical and authoritarian moral to it, a story of a few leaders and many followers. The lecturer pointed out that this process turns on the creativity ofan i ndividual or team in thinking up and defining problems well enough so they can be studied. The individual and team operate in a social context locally, through the scientific literature, and through their ongoing contacts around the world that place problems in a complex social, intellectual, and spatial web.

The chemist then asked the class how anyone could know that a selected problem is worth studying. Again the students were puzzled. He pointed out that there are many rational tests of the consequences arising from particular subjects, but none guarantees that the problem itself is worth the effort. ^^ether or not a researcher or a team becomes committed to the study of a problem is a matter of individual preference, intuition, insight, and the avail­ability of the required resources, including money. It often is also the result of a chain of previous work in which this particular activity forms a link.

Having defined a problem and decided it is important enough to pursue, the next issue for researchers is to figure out how to study it. The group must ask itself what would be potentially relevant data for the study of the selected problem. The professor problematized this deliberately by showing that it is often not obvious what data might be relevant for a particular problem. In his view, much valuable effort often goes into trying to decide what data could bear some reasonable relationship to the problem and other researchers would find convincing.

Again the students were surprised. The ambiguity of what constitutes data, the amount of social processing that goes on in a scientific research team, and the dependence of local researchers on their wider networks and on the limitations of local equipment and funding were all dimensions of science that their introductory science courses had not revealed to them. They had been given a view of scientific method primarily as an individual encounter with a world of facts and individualistic formulations of hypotheses, research strate­gies, experiments, and reports. That is, they had been given the heroic, radical individualist view of science, and they were listening to a scientific hero who was giving them an antiheroic narrative of science, yet one that was filled with a profound respect for the activity of scientists.

They seemed particularly bewildered by the notion that the data also are determined, to an extent, by the kind of equipment available at the research site. ^^at is at hand plays some role in what data are thought to be relevant and the way data might be collected. Greenwood could see the students were uncomfortable with this, as if it was a form of cheating because of the ideal­ization of scientific processes they were familiar with.

The chemist also emphasized the large number of decisions about how to document the information being collected and organizing the activity among a team of researchers to make it efficient and reasonable. The notions that a Nobel Prize chemist would have to be a team leader, an accomplished grant writer, and a social actor skilled in organizing and motivating groups were surprising to the class. That compromises would be made to design an activity that would not cause the research group to run out of resources before the data collection was completed was also new. Of course, this is not the students’ fault, because few had ever faced the need to write grants, collect resources, and conduct experiments within a budget.

Having emphasized the intellectual and social embeddedness of all the elements in the scientific process, the chemist then argued that it is difficult to decide when data collection is complete. He pointed out that deciding how much data are enough often is a pragmatic matter, not always justifiable in abstract terms. It may be a decision based on fatigue; the exhaustion of finan­cial, physical, or temporal resources; or the sense that there are enough data to say something others will believe about the problem in question. The students realized that this was a much more indeterminate view of the closure of the data collection phase of a scientific process than they had expected.

At this point, the chemist moved on to the second phase of hypothesis or question formulation. He pointed out that, although the activity is initially guided by a sense of a particular problem and possibly by a hypothesis, once a body of data has been collected and is examined, the issue becomes how to account for the array, or the distribution, or the structure of the data at hand. In the physical and natural sciences, this part of the process often is a group activity. A variety of hypotheses is often formulated by a brainstorming activ­ity through interaction influenced by a reading of the literature, flurries of e-mail, interpersonal and interunit relationships, and other interactions.

The chemist then asked the class members how they would know when they had formulated enough hypotheses. The students were mystified, because hypothesis formulation as a form of behavior is apparently not often discussed in science courses. The chemist’s sober answer was that hypothesis formulation is over when you cannot think of any more hypotheses or when you are too tired to go on. The students initially thought he was joking, but it became clear that he was not. He wanted the students to understand that science is not an activity that takes place in some idealized metaphysical space with perfect information, infinite resources to spend, and perfectly rational human beings in attendance. Science is a form of human activity that combines a set of prag­matic compromises between all the elements present at any given moment.

Beyond the pragmatics of the situation, the chemist also wanted to make a deeper point. We believe he was arguing that there is no rational way to know when one has formed enough alternative kinds of explanations for the array of data in question. The world is more complex than our apprehension of it can be, and thus we will always be approaching this complexity through a series of imperfect compromises. Being trained in a particular institution with a partic­ular group of scientists is likely to have a powerful effect on judgments about how many hypotheses are sufficient. The appetites for complexity and other characteristics of these groups will, probably, socialize a young scientist to a particular standard. However this occurs, the chemist was pointing out that one can never know that all the relevant, possible ways of accounting for the data have been formulated. Science, as powerful as it is, is not a means for tran­scending the human condition.

Having completely perplexed his audience, the chemist then moved on to the next step: the process of testing questions or hypotheses against the data. Doing prestructured experiments with finite solutions in laboratory exercises did not prepare the students very well for what he said. In the students’ expe­rience, all the puzzles had specific answers, and they would receive grades for solving the puzzles with a specific set of resources and in a limited amount of time. They knew the answers were there, and they simply had to uncover them. These scientific training practices did not prepare them for the chemist’s much less determinate view.

He pointed out that translating hypotheses into testable propositions and matching data to hypotheses are complex, ambiguous, and creative activities. Chains of assumptions and definitions are required to link data and hypothe­ses, and these chains have to be built so they are capable of convincing others that the reasoning and research process gone through is sensible and, therefore, that the results are acceptable. Doing this in laboratory situations is often a group process with rapid brainstorming and much trial and error, eminently social activities.⁷

Once the group has inventoried all the questions or hypotheses it can think of against the data collected and organized, the lecturer said that the best possible outcome is that the group has not invalidated all the explanations that it initially developed. The hope is that it would have at least one left. Quite often, this does not happen, and the group must return to the process of hypothesis formulation because none of the hypotheses is left standing. Alternatively, the data may not provide the basis for choosing among alterna­tive explanations, and the experiment has to be redesigned.

At this point, the students were relieved because this began to sound like the sort of science that they could identify with. At the end of the process, the group has a validated explanation. But the chemist was not through. He explained that not having invalidated all the hypotheses did not mean that the remaining hypotheses had been proved true.

In making this argument, he was not being perverse. Having pointed out that the initial process of hypothesis formulation is indeterminate, in the sense that there always exists the possibility of hypotheses that the group did not think of and that financial and human resources are finite, he was being con­sistent. If a single hypothesis were left after the testing procedures were com­plete, one could only say that, of the hypotheses thought of, at least one had not yet been invalidated. It might be invalidated in the future, but other better hypotheses might be formulated to account for the data in the future. Thus, the remaining explanation could not be said to be correct. It is simply the only one left of those thought of.

The 50 minutes were over. Greenwood’s class seemed stunned, though appreciative. Rather than making the usual quick and noisy exodus, they wan­dered out of the room silently. The chemist had given a master class, but more important, he had conveyed a view of science as a form of human action involving complexity, ambiguity, creativity, group dynamics, and many prag­matic concessions to the limitations imposed by the time and resources avail­able. Rather than diminishing or demystifying science, this view helps us understand that science is a way of behaving, a way of acting in relation to the nonhuman and human worlds that has resulted in remarkable improvements in our understanding of how those worlds work and our ability to change the state of those worlds. Good scientific practice centers on constant cycles of thought and action.

Something the chemist did not mention at any point was prediction. Although it was dear that a good explanation could be used to generate a prior idea of the way the data should be arrayed if the explanation were to hold, he did not stress prediction itself as a core element in science. Rather, he empha­sized explanation. Yet, commonsense views of science almost always equate science with prediction. We believe the chemist was right to deemphasize prediction as a fundamental criterion for science.

Scientists seek to explain arrays of data. Predicting the expected array of data under given conditions is a powerful way of testing explanations, but the goal is having an explanation that makes sense. Another way prediction is pro­ductively used is when engineers, in attempting to solve an important problem, design an apparatus or system that they “predict” will solve the problem. Here the prediction is the vision they have of the ideal outcome that guides their developmental process.

Prediction is a tool to be used in this effort, and its use varies a great deal with the conditions. Under some conditions, prediction, in the ordinary sense, is out of the question, as in the historical studies of evolution.⁸ Under other conditions, predictions take the form of statistical generalizations about huge populations and cannot accurately capture what is happening in particular segments of those populations. In other situations, predictive activity takes the form of intervening in the phenomenon under study to change its state in a desired direction. This is precisely what the experimental method in science does and what AR aims to achieve in the social world.

The chemist’s view matches closely with our experiences of scientists and engineers at work. It puts them, as human actors, at the center of the combined social-research activity that is science. He made it dear that scientists and engineers are not the enactors of some abstract, perfect, determinate system. The chemist conveyed to the students that scientific method is a form of social behavior, a form that is not foolproof, but one that uses human capabilities to pose questions and attempts to examine those questions through rational but fully social inquiry. He stressed the need to recognize the significant gaps and imperfections in any process of this sort, and he affirmed that human judg­ment, creativity, and social interaction are an intrinsic part of the process.

Repeatedly, he emphasized that science is a collective activity carried out by members of research teams within a larger scientific community. The larger community provides the literature on which the research is built to some degree, as well as the resources used to carry out the research. The research team and the laboratory form a complex, dynamic social system of people act­ing on phenomena and sharing their thoughts within the pragmatic limita­tions set by the availability of key resources and the dynamics of the human relationships involved.

Source: Greenwood Davydd J., Levin Morten (2006), Introduction to Action Research: Social Research for Social Change, SAGE Publications, Inc; 2nd edition.