pthagard@uwaterloo.ca

Since Plato, most philosophers have drawn a sharp line between reason and emotion, assuming that emotions interfere with rationality and have nothing to contribute to good reasoning. In his dialogue the Phaedrus, Plato compared the rational part of the soul to a charioteer who must control his steeds, which correspond to the emotional parts of the soul (Plato 1961, p. 499). Today, scientists are often taken as the paragons of rationality, and scientific thought is generally assumed to be independent of emotional thinking.

But current research in cognitive science is increasingly challenging the view that emotions and reason are antagonistic to each other. Evidence is accumulating in cognitive psychology and neuroscience that emotions and rational thinking are closely intertwined (see, for example: Damasio 1994, Kahneman 1999, and Panksepp 1999). My aim in this paper is to extend that work and describe the role of the emotions in scientific thinking. If even scientific thinking is legitimately emotional, then the traditional division between reason and emotion becomes totally unsupportable.

My paper begins with a historical case study. In his famous book, The Double Helix, James Watson presented a review of the discovery of the structure of DNA. Unlike the typical dry and scientific biography or autobiography, Watson provided a richer view of the personalities involved in one of the most important discoveries of the 20th century. I will present a survey and analysis of the emotions mentioned by Watson and use it and quotations from other scientists to explain the role of the emotions in scientific cognition. My account will describe the essential contributions of emotions in all three of the contexts in which scientific work is done: investigation, discovery, and justification. Initially, emotions such as curiosity, interest, and wonder play a crucial role in the pursuit of scientific ideas. Moreover, when investigation is successful and leads to discoveries, emotions such as excitement and pleasure arise. Even in the third context, justification, emotions are a crucial part of the process of recognizing a theory as one that deserves to be accepted. Good theories are acknowledged for their beauty and elegance, which are aesthetic values that are accompanied by emotional reactions.

In 1951, James Watson was a young American postdoctoral fellow at Cambridge University. He met Francis Crick, a British graduate student who was already in his mid- 30s. Both had a strong interest in genetics, and they began to work together to identify the structure of the DNA molecule and to determine its role in the operations of genes. The intellectual history of their work has been thoroughly presented by Olby (1974) and Judson (1979). My concern is with emotional aspects of the thinking of Watson and Crick and of scientific thinking in general; these aspects have been largely ignored by historians, philosophers, and even psychologists. The primary question is: What role did emotions play in the thinking of Watson and Crick that led to the discovery of the structure of DNA and the acceptance of their model?

In order to answer this question, I read carefully through The Double Helix, looking for words that refer to emotions. Watson described not only the ideas and the hypotheses that were involved in the discovery of the structure of DNA, but also the emotions that accompanied the development of the new ideas. In Watson's (1969) short book, whose paperback edition has only 143 pages, I counted a total of 235 emotion words. Of the 235 emotions referred to, more than half (125) were attributed by Watson solely to himself. Another 35 were attributed to his collaborator Francis Crick, and 13 emotions were attributed by Watson to both of them. There were also 60 attributions of emotions to other researchers, including various scientists in Cambridge and London. I coded the emotions as having either positive valence (e.g. happiness) or negative valence (e.g. sadness), and found that more than half of the emotions (135) had positive valence.

In order to identify the kinds of emotions mentioned by Watson, I coded the emotion words in terms of what psychologists call basic emotions, which are ones taken to be culturally universal among human beings (Ekman 1992). The usual list of basic emotions includes happiness, sadness, anger, fear, disgust, and sometimes surprise. To cover classes of emotion words that occurred frequently in the book but did not fall under the 6 basic emotions, I added 3 additional categories: interest, hope, and beauty. Figure one displays the frequency with which these categories of emotions appear in Watson's narrative. Happiness was the most frequently mentioned emotion, occurring 65 times with many different words referring to such positive emotional states as excitement, pleasure, joy, fun, delight, and relief. The next most frequently mentioned classes of emotions were: interest, with 43 emotion words referring to states such as wonder and enthusiasm; and fear, with 38 emotion words referring to states such as worry and anxiety.

The point of this study, however, was not simply to enumerate the emotion words used by Watson in his story. To identify the role that emotions played in the thinking of Watson and Crick, I coded the emotion words as occurring in three different contexts of the process of inquiry: investigation , discovery, and justification. Most scientific work occurs in the context of investigation, when scientists are engaged in the long and often difficult attempt to establish empirical facts and to develop theories that explain them. Much preparatory experimental and theoretical work is usually required before scientists are able to move into the context of discovery, in which new theoretical ideas and important new empirical results are produced. Finally, scientists enter the context of justification, when new hypotheses and empirical results are evaluated with respect to alternative explanations and the entire body of scientific ideas.

Figure 1. Frequency of classes of emotion words in Watson (1969).

The distinction between the contexts of discovery and justification is due to Reichenbach (1938). I use it here not to demarcate the psychological and subjective from the philosophical and rational, as Reichenbach intended, but merely to mark different stages in the process of inquiry, all of which I think are of both psychological and philosophical interest. I further divide Reichenbach’s context of discovery into contexts of investigation and discovery, in order to indicate that much work is often required before actual discoveries are made. In scientific practice, the contexts of investigation, discovery, and justification often blend into each other, so they are best viewed as rough stages of scientific inquiry rather than as absolutely distinct.

Most of Watson's emotion words (163) occurred in the context of investigation. 15 words occurred in the context of discovery, 29 in the context of justification, and 28 emotion words incurred in other more personal contexts that had nothing to do with the development of scientific ideas. In the rest of this paper, I will provide a much more detailed account of how the many different kinds of emotions contribute to scientific thinking in the three contexts. I will offer not only a correlational account of what emotions tend to occur in what contexts, but also a causal account of how emotions produce and are produced by the cognitive operations that occur in the different stages of scientific inquiry.

Discussions of scientific rationality usually address questions in the context of justification, such as when is it rational to replace a theory by a competing one. But scientists engage in a great deal of decision making that precedes questions of justification. Students who are interested in pursuing a scientific career must make decisions that answer questions such as the following: What science (e.g. physics or biology) should I concentrate on? What area of the science (e.g. high-energy physics or molecular biology) should I focus on? What particular research topics should I pursue? Which questions should I attempt to answer? What methods and tools should I use in attempting to answer those questions?

On the traditional view of rational decision making familiar from economics, scientists should attempt to answer these questions by calculating how to maximize their expected utility. This involves taking into account scientific goals such as truth and understanding, and possibly also taking into account personal goals such as fame and financial gain, as well as considering the probabilities that particular courses of research will lead to the satisfaction of these goals. Some philosophers (e.g. Goldman 1999 and Kitcher 1993) describe scientists as choosing projects based on maximizing epistemic goals such as truth.

I would like to propose an alternative view of scientific decision making in the context of investigation. Decisions concerning what topics to research are much more frequently based on emotions rather than on rational calculations. It is rarely possible for scientists to predict with any accuracy what choice of research areas, topics, and questions will pay off with respect to understanding, truth, or personal success. The range of possible actions is usually ill-specified, and the probabilities of success of different strategies are rarely known. Because rational calculation of maximal utility is effectively impossible, it is appropriate that scientists rely on cognitive emotions such as interest and curiosity to shape the direction of their inquiries.

Watson’s narrative makes it clear that he and Crick were heavily motivated by interest. Watson left Copenhagen where his postdoctoral fellowship was supposed to be held because he found the research being done there boring: the biochemist he was working with there "did not stimulate me in the slightest" (Watson 1969, p. 23). In contrast, he reacted to questions concerning the physical structure of biologically important molecules such as DNA with emotions such as excitement: "It was Wilkins who first excited me about X-ray work on DNA" (p. 22); "Suddenly I was excited about chemistry" (p. 28). Crick similarly remarks how he and Watson "passionately wanted to know the details of the structure" (Crick 1988, p. 70).

Once interest and curiosity direct scientists to pursue answers to particular questions, other emotions such as happiness and hope can help motivate them to perform the often laborious investigations that are required to produce results. It was clearly important to Watson and Crick that they often became excited that they were on the right track. Watson (1969, p. 99) wrote that "On a few walks our enthusiasm would build up to the point that we fiddled with the models when we got back to our office." Both took delight in getting glimpses of what the structure of DNA might be. They were strongly motivated by the hope that they might make a major discovery. Hope is more than a belief that an event might happen – it also the emotional desire and joyful anticipation that it will happen. I counted 26 occurrences where Watson mentioned the hopes of himself or others for scientific advance.

In addition to positive emotions such as interest and happiness, scientists are also influenced by negative emotions such as sadness, fear, and anger. Sadness enters the context of investigation when research projects do not work out as expected. Watson and Crick experienced emotions such as dismay when their work faltered. Such emotions need not be entirely negative in their effects, however, because sadness about the failure of one course of action can motivate a scientist to pursue an alternative and ultimately more successful line of research.

Fear can also be a motivating emotion. Watson and Crick were very worried that the eminent chemist Linus Pauling would discover the structure of DNA before they did, and they also feared that the London researchers, Rosalind Franklin and Maurice Wilkins, would beat them. Watson wrote that when he heard that Pauling had proposed a structure, "my stomach sank in apprehension at learning that all was lost" (p. 102). Other worries and anxieties arose from setbacks experienced by Watson and Crick themselves. Watson was initially very excited by a proposal that was shown by a crystallographer to be unworkable, but tried to salvage his hypothesis: "Thoroughly worried, I went back to my desk hoping to salvage the like-with-like idea" (p. 122).

The other negative emotion that Watson mentions frequently in the context of investigation is anger. To himself, Watson ascribes only weak forms of anger, such as annoyance and frustration, but he describes Crick as experiencing fury and outrage when his senior professors wrote a paper that failed to acknowledge his contributions. According to Oatley (1992), people experience anger when accomplishment of their goals is blocked by people or events. Most of the anger-related episodes mentioned by Watson are directed at people, but some concern facts, as when Watson and Crick both become annoyed at the complexity of DNA bonds. I do not get the impression from Watson’s chronicle that anger was ever a motivating force in the scientific work he described, but we can see it as an effect of undesirable interactions with other people and the world.

A general model of the role of emotions in the context of investigation of scientific investigations is shown in figure 2. Interest, wonder, curiosity and the avoidance of boredom are key inputs to the process of selecting of scientific questions to investigate. I have made the causal arrows between emotions and the process of question generation bidirectional in order to indicate that emotions are outputs as well as inputs to the generation of questions. Coming up with a good question can increase curiosity and interest, and produce happiness as well. Once questions have been generated, the cognitive processes involved in trying to generate answers to them can also interact with emotions such as interest and happiness, as well as with negative emotions such as fear. For convenience, I have organized positive emotions on the left side of figure 2, and negative emotions on the right side. The causal arrows connecting processes of question generation, question answering, and answer evaluation are also bidirectional, indicating interconnections rather than a linear operation. For example, the attempt to answer a question can generate subordinate questions whose answers are relevant to the original question. For a discussion of the process of question generation in science, see Thagard (1999).

The emotions that Watson attributed to himself and his colleagues are commonly found in other scientists, as evident in a serious of interviews conducted by the biologist Lewis Wolpert for BBC Radio (Wolpert and Richards 1997). Wolpert elicited frank descriptions of leading researchers’ intense emotional involvement in their work. For example, the distinguished biologist Gerald Edelman stated: "Curiosity drives me. I believe that there is a group of scientists who are kind of voyeurs, they have the splendid feeling, almost a lustful feeling, of excitement when a secret of nature is revealed ... and I would certainly place myself in that group." (Wolpert and Richards 1997, p. 137) Similarly, the eminent physicist Carlo Rubbia said of scientists: "We're driven by an impulse which is one of curiosity, which is one of the basic instincts that a man has. So we are essentially driven not by … how can I say … not by the success, but by a sort of passion, namely the desire of understanding better, to possess, if you like, a bigger part of the truth" (Wolpert and Richards, 1997, p. 197).

According to Kubovy (1999, p. 147), to be curious is to get pleasure from learning something that you did not previously know. He contends that curiosity has its roots in animal behavior, having evolved from the need to search for food. Many mammals prefer richer environments over less complex ones. Humans can be curious about a very wide range of subjects, from the trivial to the sublime, and scientists direct their intense drive to learn things that are unknown not just to them, but to people in general. Loewenstein (1994) defends a more cognitive account of curiosity as the noticing of knowledge gaps, but underplays the emotional side of curiosity. There are many knowledge gaps that scientists may notice, but only some of them arouse an emotional interest that spurs them to the efforts required to generate answers.

The discovery of the structure of DNA occurred after Maurice Wilkins showed Watson new X-ray pictures that Rosalind Franklin had taken of a three-dimensional form of DNA. Watson reports: "The instant I saw the picture my mouth fell open and my pulse began to race" (p. 107). This picture moved Watson from the context of investigation into the context of discovery, in which a plausible hypothesis concerning the structure of DNA could be generated. While drawing, Watson got the idea that each DNA molecule might consist of two chains, and he became very excited about the possibility and its biological implications. Here is a passage that describes his resulting mental state; I have highlighted in boldface the positive emotion words and in italics the negative emotion words.

Watson’s initial idea about the chemical structure of DNA turned out to be wrong, but it put him on the track that quickly led to the final model that Watson and Crick published.

Most of the emotions mentioned by Watson in his discussion of discoveries fall under the basic emotion of happiness; these include excitement, pleasure and delight. The chemist who helped to invent the birth control pill, Carl Djerassi, compares making discoveries to sexual pleasure. "I'm absolutely convinced that the pleasure of a real scientific insight - it doesn't have to be a great discovery - is like an orgasm" (Wolpert and Richards 1997, p. 12). Gopnik (1998) also compares explanations to orgasms, hypothesizing that explanation is to cognition as orgasm is to reproduction.

The pleasure of making discoveries is effusively described by Gerard Edelman:

Carlo Rubbia reports that "The act of discovery, the act of being confronted with a new phenomenon, is a very passionate and very exciting moment in everyone's life. It's the reward for many, many years of effort and, also, of failures" (Wolpert and Richards 1997, p. 197).

François Jacob (1988, p. 196-197) describes his first taste of the joy of discovery as follows: "I had seen myself launched into research. Into discovery. And, above all, I had grasped the process. I had tasted the pleasure." Later, when Jacob was developing the ideas about the genetic regulatory mechanisms in the synthesis of proteins that later won him a Nobel Prize, had an even more intense emotional reaction: "These hypotheses, still rough, still vaguely outlined, poorly formulated, stir within me. Barely have I emerged that I feel invaded by an intense joy, a savage pleasure. A sense of strength, as well, of power" (Jacob 1988, p. 298). Still later, Jacob describes the great joy that came with experimental confirmation of his hypotheses. Scheffler (1991, p. 10) discusses the joy of verification, when scientists take pleasure when their predictions are found to be true. Of course, predictions sometimes turn out to be false, producing disappointment and even gloom.

It is evident from Watson and the other scientists that I have quoted that discovery can be an intensely pleasurable experience. In figure 2, I have shown surprise and happiness as both arising from the successful generation of answers to pursued questions. The prospect of being able to experience such emotions is one of the prime motivating factors behind scientific efforts. Richard Feynman proclaimed that his work was not motivated by a desire for frame or prizes such as the Nobel that he eventually received, but by the joy of discovery: "The prize is the pleasure of finding a thing out, the kick of the discovery, the observation that other people use it [my work]-those are the real things, the others are unreal to me (Feynman 1999, p. 12). Discoveries are usually pleasant surprises, but an expanded version of figure 2 would also show unpleasant surprises that come about, for example, when experiments yield data that are contrary to expectations. Disappointment and sadness can also arise when the evaluation stage leads to the conclusion that one’s preferred answers are inferior or inadequate.

Kubovy (1999) discusses virtuosity, the pleasure we have when we are doing something well. He thinks that humans and also animals such as monkeys and dolphins enjoy working and take pleasure in mastering new skills. Scientists can achieve virtuosity in many diverse tasks such as designing experiments, interpreting their results, and developing plausible theories to explain the experimental results. According to physicist Murray Gell-mann, "Understanding things, seeing connections, finding explanations, discovering beautiful, simple principles that work is very, very satisfying" (Wolpert and Richards 1997, p. 165).

Although many would concede that the processes of investigation and discovery are substantially emotional, it is much more radical to suggest that even the context of justification has a strong emotional component. In Watson’s writings, the main sign of an emotional component to justification is found in his frequent description of the elegance and beauty of the model of DNA that he and Crick produced. Watson wrote:

Thus one of the causes of the conviction that they had the right structure was its aesthetic, emotional appeal. Other scientists also had a strong emotional reaction to the new model of DNA. Jacob (1988, p. 271) wrote of the Watson and Crick model: "This structure was of such simplicity, such perfection, such harmony, such beauty even, and biological advances flowed from it with such rigor and clarity, that one could not believe it untrue."

In a similar vein, the distinguished microbiologist Leroy Hood described how he takes pleasure in coming up with elegant theories:

Many other scientists have identified beauty and elegance as distinguishing marks of theories that should be accepted (McAllister 1996).

From the perspective of traditional philosophy of science, or even the perspective of traditional cognitive psychology that separates the cognitive and the emotional, the fact that scientists find some theories emotionally appealing is irrelevant to their justification. My own previous work is no exception: I have defended the view that scientific theories are accepted or rejected on the basis of their explanatory coherence with empirical data and other theories (Thagard 1992, 1999). But my new theory of emotional coherence shows how cognitive coherence judgments can generate emotional judgments (Thagard 2000, ch. 6). I will now briefly review the theory of emotional coherence and indicate how it can be applied to explain emotional scientific judgments concerning the acceptance of theories.

My coherence theory of inference can be summarized in the following theses (Thagard and Shelley forthcoming; see also Thagard 2000):

1. All inference is coherence-based. So-called rules of inference such as modus ponens do not by themselves license inferences, because their conclusions may contradict other accepted information. The only rule of inference is: Accept a conclusion if its acceptance maximizes coherence.

2. Coherence is a matter of constraint satisfaction, and can be computed by connectionist and other algorithms.

3. There are six kinds of coherence: analogical, conceptual, explanatory, deductive, perceptual, and deliberative.

4. Coherence is not just a matter of accepting or rejecting a conclusion, but can also involve attaching a positive or negative emotional assessment to a proposition, object, concept, or other representation.

On this account, a theory is justified if inferring it maximizes coherence, but assessment can also involve an emotional judgment Theories consists of hypotheses which are comprised of concepts. According to the theory of emotional coherence, these representations not only have a cognitive status of being accepted or rejected, they also have an emotional status of being liked or disliked. In keeping with Bower’s (1981) account of the psychology of emotions, I call its emotional status the valence of the representation. A representation receives its valence as the result of its connections with other representations. The valence of a theory will flow from the valences of the hypotheses that constitute it, as well as from the overall coherence that it generates.

Overall coherence requires a judgment about how everything fits together. Such judgments can be made by the computational model HOTCO ("hot coherence"), that not only simulates the spread of valences among representations but also simulates how the coherence of the representations with each other can generate a "metacoherence" inference which is associated with happiness. I will not repeat the computational details (Thagard 2000), but merely want here to give the general flavor of how the model works. A proposition is highly coherent with other propositions if its acceptance helps to maximize the satisfaction of constraints, such as the constraint that if a hypothesis explains a piece of evidence, then the hypothesis and the evidence should be either accepted together or rejected together. If a group of propositions can be accepted together in such a way that the individual propositions each tend to satisfy a high proportion of the constraints on them, the overall system of proposition gets a high metacoherence rating

Figure 3 diagrams how judgments of coherence and other emotions can arise. Units are artificial neurons that represent propositions, and their assessment by neural network algorithms for maximizing constraint satisfaction leads to them either being accepted or rejected. In addition, the extent to which the constraints on a unit are satisfied affects how much it activates a unit representing a judgment of coherence, which produces happiness. For example, if the acceptance of a hypothesis makes possible satisfaction of numerous constraints that tie the hypotheses with the evidence it explains, then a unit representing the hypothesis will strongly activate the coherence node and the thence the happiness node. At the other end of the emotional distribution, incoherence tends to produce unhappiness or even fear-related emotions such as anxiety.

This sketch of the theory and computational model of emotional coherence shows how it is possible for a mind simultaneously to make cognitive judgments and emotional reactions. Watson and Crick’s hypothesis concerning the structure of DNA was highly coherent with the available evidence and with the goals of biologists to understand the genetic basis of life. This high coherence generated not only a judgment that the hypothesis should be accepted, but also an aesthetic, emotional attitude that the hypothesis was beautiful. Coherence made Watson, Crick, Jacob and others very happy.

Discovering the structure of DNA was a major scientific breakthrough, but it was not a major case of conceptual change. The discovery brought about additions to the concepts of a gene and DNA, but did not require revisions of previous components of the concepts. New hypotheses were adopted, but there was no rejection of previously believed hypotheses. In contrast, scientific revolutions involve both major alteration of concepts and revision of previously held hypotheses (Thagard 1992). I will now argue that conceptual changes also often involve emotional change.

Consider, for example, Darwin’s theory of evolution by natural selection. When he proposed it in the Origin of Species in 1859, some scientists such as Huxley quickly realized the strength of the theory and adopted it. But others continued to maintain the traditional view that species arose by Divine creation, and viewed Darwin’s theory as not only false but repugnant. Even today, there are creationists who view the theory of evolution as not only false but evil, since they see it as undermining religious beliefs and values. Most biologists, in contrast, feel that Darwin’s theory is profound and elegant. Thus for a creationist to become a Darwinian, or for a Darwinian to become a creationist, thinkers must undergo emotional as well as cognitive changes. In addition to rejecting some hypotheses already held and accepting new ones, thinkers who are changing theories are also changing their emotional attitudes towards propositions, concepts, and even people: there are different emotions associated with Darwin-as-scientific-hero and Darwin-as-heretic.

Consider also the more recent historical case of the initial rejection and eventual acceptance of the bacterial theory of ulcers (Thagard 1999). When Marshall and Warren (1984) proposed that most peptic ulcers are caused by the newly discovered bacterium now known as Helicobacter pylori, their hypothesis was initially viewed as not only implausible, but even crazy! Marshall’s persistent advocacy of the unpopular hypothesis led people to view him as brash and irresponsible. But annoyance turned to respect as evidence mounted that eradicating the bacteria brought many people a permanent cure of their ulcers. Not only did attitudes change toward Marshall but also towards the bacterial theory of ulcer and towards the concept of ulcer-causing bacteria. What was ridiculed became appreciated as an important part of medical theory and practice.

I propose that these changes come about through the process of achieving emotional coherence. At the same time that decisions concerning what to believe are made by a process that maximizes explanatory coherence (fitting hypotheses with data and each other), attitudes are formed by a process that spreads valences through the representational system in accord with existing valences and emotional constraints. For example, scientists who come to appreciate the explanatory power of Darwin’s evolutionary hypotheses will come to believe the and to attach a positive emotional valence to them, to the concept of evolution, and to Darwin. On the other hand, scientists (if there are any) who adopt religious fundamentalism and decide that the Bible is to be revered will reject the hypothesis and concept of evolution and view it as odious. Our beliefs are revised by a holistic process of deciding what fits with what, and our emotions are similarly adjusted.

If conceptual change in science is often emotional change, what about science education? Often, science education will not require emotional change because students do not have any emotional attachments to prescientific ideas. For example, learning to use Newton’s laws of motion to understand the behavior of projectiles does not seem to require abandoning any emotional attachment on the part of the students to Aristotelian ideas. But learning in biology and the social sciences may involve much more emotional adjustment. Biology students learning about genetics and evolution may have to abandon beliefs that they not only hold but value about the nature of the universe. Similarly, social and political beliefs may be factual, but they are closely intertwined with strongly emotional values. A microneconomist who is convinced that laissez-faire capitalism is the best economic system differs in beliefs and also emphatically in values from a macroenomist who favors socialism. Hence educators should be alert for cases where resistance to conceptual change derives from emotional as well as cognitive factors. Normally, emotional change is more the province of psychotherapists than science teachers, but teaching emotionally laden scientific issues such as the theory of evolution may require therapeutic techniques to identify valences as well as beliefs.

But should it not be the job of science educators to help students aspire to what scientists are supposed to be: objective and nonemotional reasoners? Even if emotions are associated with scientific thinking, perhaps the aim of scientific education, from grade school through the Ph.D., should be to inculcate in students the necessity of eradicating subjective emotions in order to improve their scientific thinking by insulating it from bias. It is not hard to find cases where emotions have distorted science, from Nazis rejecting Einstein’s theory of relativity because they hated Jews to researchers fabricating data because they wanted to be famous.

While recognizing that there are many such cases in which emotions have undermined scientific objectivity and truth-seeking, I think it would be a mistake to demand that scientists strive to eliminate emotions from the cognitive processes enable them to do science. First, severing the cognitive from the emotional is probably impossible. As I mentioned in the introduction, there is abundant evidence from psychology and neuroscience that cognitive and emotional processes are intimately intertwined. The ethical principle that ought implies can applies: we cannot insist that a person’s thinking should be emotion-free when it is biologically impossible for people to think that way. Mitroff (1974, p. 99) found in his study of Apollo moon scientists that they held strongly emotional views, especially of other researchers, but concluded that "the existence of such strong emotions does not prevents scientists from doing excellent scientific work."

Second, even if scientists could become free of the sway of emotions, it is likely that science would suffer as a result. The discussion above of emotions in the contexts of investigation and discovery showed that positive emotions such as happiness and even negative emotions such as anxiety play an important role in motivating and energizing science. Given that science is often difficult and frustrating, why would anyone do it if it were not also sometimes exciting and satisfying? According to Mitroff (1974, p. 248), "to eliminate strong emotions and intense commitments may be to eliminate some of science’s most vital sustaining forces." Even in the context of justification, there is a useful cognitive contribution of feelings of elegance and beauty, because they are attached to theories that excel in explanatory coherence, which, I would argue is what scientific theory is intended most immediately to maximize.

So I would emphatically not urge that scientists aim to become robotically unemotional. Despite dramatic increases in computer speed and intelligent algorithms, computers currently are no threat to replace scientists. A key aspect of intelligence that is rarely recognized is the ability to set new goals, not just as ways of accomplishing other existing goals, but as novel ends in themselves. There is no inkling in the field of artificial intelligence of how computers could acquire the ability to set goals for themselves, which is probably just as well because there is no guarantee that the goals of such artificially ultra-intelligent computers would coincide with the goals of human beings. In humans, the development of non-instrumental goals is intimately tied with emotions such as interest, wonder, and curiosity, which are just the goals that spur scientists on. A scientist without passion would be at best a mediocre scientist.

Bower, G. H. (1981). Mood and memory. American Psychologist, 36, 129-148.

Crick, F. (1988). What mad pursuit: A personal view of scientific discovery. New York: Basic Books.

Damasio, A. R. (1994). Descartes' error. New York: G. P. Putnam's Sons.

Ekman, P. (1992). An argument for basic emotions. Cognition and emotion, 6, 169-200.

Feynman, R. (1999). The pleasure of finding things out. ADD: Perseus Books.

Goldman, A. (1999). Knowledge in a social world. Oxford: Oxford University Press.

Gopnik, A. (1998). Explanation as orgasm. Minds and machines, 8, 101-118.

Jacob, F. (1988). The statue within (F. Philip, Trans.). New York: Basic Books.

Judson, H. F. (1979). The eighth day of creation: Makers of the revolution in biology. New York: Simon and Schuster.

Kahneman, D. (1999). Objective happiness. In D. Kahneman, E. Diener, & N. Schwarz (Eds.), Well-being: Foundations of hedonic psychology (pp. 3-25). New York: Russell Sage Foundation.

Kitcher, P. (1993). The advancement of science. Oxford: Oxford University Press.

Kubovy, M. (1999). On the pleasures of the mind. In D. Kahneman, E. Diener, & N. Schwarz (Eds.), Well-being: Foundations of hedonic psychology (pp. 134-154). New York: Russell Sage Foundation.

Loewenstein, G. (1994). The psychology of curiosity: A review and reinterpretation. Psychological Bulletin, 116, 75-98.

McAllister, J. W. (1996). Beauty and revolution in science. Ithaca: Cornell University Press.

Mitroff, I. I. (1974). The subjective side of science. Amsterdam: Elsevier.

Oatley, K. (1992). Best laid schemes: The psychology of emotions. Cambridge: Cambridge University Press.

Olby, R. (1974). The path to the double helix. London: Macmillan.

Panksepp, J. (1998). Affective neuroscience: The foundations of human and animal emotions. Oxford: Oxford University Press.

Plato. (1961). The collected dialogues. Princeton: Princeton University Press.

Reichenbach, H. (1938). Experience and prediction. Chicago: University of Chicago Press.

Scheffler, I. (1991). In praise of the cognitive emotions. New York: Routledge.

Thagard, P. (1992). Conceptual revolutions. Princeton: Princeton University Press.

Thagard, P. (2000). Coherence in thought and action. Cambridge, MA: MIT Press, fall publication.

Thagard, P., & Shelley, C. P. (forthcoming). Emotional analogies and analogical inference. In D. Gentner, K. H. Holyoak, & B. N. Kokinov (Eds.), The analogical mind: Perspectives from cognitive science . Cambridge, MA: MIT Press.

Watson, J. D. (1969). The double helix. New York: New American Library.

Wolpert, L., & Richards, A. (1997). Passionate minds: The inner world of scientists. Oxford: Oxford University Press.