Modes of Explanation

Affordances for Action and Prediction

M. Lissack,A. Graber

• Éditeur: Palgrave Macmillan
• Année: 2014
• ISBN:
  9781137403865
• Sujet: Economics
• Sous-sujet: Econometrics

Modes of Explanation is the first book in decades to attempt to bring these conflicting approaches together and to offer a compelling narrative to explore how the paradox of 'explanation' can converge.

AFTERWORD 1

The Scientific Attitude Toward Explanation

Lee McIntyre

The problem with explanation in the sciences is not that we do not have the right concept of what it means to “explain” or that we misunderstand the domain of science. The problem is that many of those who profess to seek an explanation do not actually seek knowledge, but instead confirmation for what they already believe.

There is a lot of ignorance in the world, but this ignorance cannot always be overcome merely by “explaining.” First, we had better be sure that we are explaining things to someone who really wants to understand. This is not a trivial matter. Many people are convinced that they already understand (on extra-scientific grounds) how things work and turn to science only for the prestige of its confirmatory power over their ideology. That is, they pursue explanation, but they do not want to understand. Explanation is relative to a question, but if someone thinks that they already understand then there is actually no question.

What is missing, therefore, is not the right definition of explanation or of science. Science is not a method or a subject matter or even a way of knowing. It is an attitude. Philosophers cannot “define” their way out of the problem of ideology or willful ignorance. For what needs changing is not our conception of scientific explanation, but instead our attitude toward it.

The scientific attitude is one where you set out to do two things: (1) learn from empirical evidence and (2) change your theories based on what you have learned. It is appalling how much of what passes for science does not do these things. Until recently, I would put in this camp a good deal of what came out of economics. The simplifying assumptions and focus on perfect rationality just did not jibe with actual experience and were not expected to. Over the last decade, however, the advent of behavioral economics has changed this situation. Now we understand that the study of human behavior also can be an experimental discipline. Just because we are complex thinking beings does not mean that our behavior is not caused and that we cannot gather evidence in order to understand it.

Perhaps the best demonstration of the need for a scientific attitude is to examine a case where it is completely lacking: the debate over intelligent design. It is sometimes said that intelligent design is creationism in a cheap tuxedo. What does this mean? It means that at its heart intelligent design does not seek scientific explanation. Its motivation instead is to gather evidence to support the ideological belief that the creation story in the Bible is correct and that evolution is an incomplete—if not outright false—account of human origins. This is to say that when intelligent design theorists approach the question of whether we can explain human origins, they are not actually seeking an answer; they think they already have it. They know at the outset where they want their “scientific” inquiry to end up. This is why intelligent design is not science (and why in my judgment recent court cases have been correct in keeping it out of the public school science classroom). For it is missing something essential to science, which is the scientific attitude. The scientific attitude requires an openness to new evidence and a willingness to revise based on what is discovered, which is completely missing in creationism. That great philosopher Bill Clinton perhaps put it best: “The problem with ideology is that it gives you the answer before you look at the evidence.”

I am afraid that all of this means, however, that the labored distinction between Science 1 and Science 2 (which seems to be at the heart of this book) is a false choice. There are not different senses of science—some of which deal with simple, ordered phenomena and some of which deal with complex, emergent ones—because science is not a subject matter. Neither is it a method. Instead, science is an attitude that can be applied to any field of study. You cannot break things off into different senses of science based on some alleged a priori distinction about what you are studying. If you really want to understand what is essential about scientific explanation, you must look instead at the way in which scientists approach and learn from evidence. This is why I am a skeptic about the Science 1 versus Science 2 distinction. There should be no different senses of science. There is only good science and bad science. All science should aim at the same goal, which is the discovery and understanding of the causes at work behind the phenomena about which we are puzzled. In some cases, these causes are contingent on human decision making. This does not make them any less discoverable through using the scientific attitude, else why does behavioral economics work at all? We cannot defy the causes of human behavior just because we know about them. Even where free will exists, we encounter the stunning regularity of human action. Is this not enough to motivate us to come to the study of social science with a genuine sense of wonder about why we keep behaving in the ways that we do?

And I am afraid that there is another false dichotomy at work here: the one between realism and constructivism. This is surely a heated debate, but it has left out one of the most important antirealist players, which is a position that I will call—for want of a better word—“nominalism.” This is the idea that we are realist enough to suppose that we do not create the world merely through our language, but we are antirealist enough to understand that we do create the language that we use to describe and understand the world. And this makes all the difference to our explanations. Indeed, language is the backbone of scientific explanation. We do not come to the study of the world as naked perceivers. We have categories and we make our explanations relative to these. Does this mean that if our theories work, we have a blueprint (or something close to it) for how the world actually works? What naïve hubris to suppose that we have somehow stumbled our way into a true conception of nature just because we have gotten some data points to fit. But science does not require this; all that it requires is that we have models that work. Yet, there may be an infinite number of other models that work just as well, which we have not yet discovered, and admitting this is part of scientific understanding. This is the problem with taking the success of science as any sort of evidence for realism. It is well known that theory is underdetermined by evidence. There are many possible “true” explanations for one and the same phenomenon. Scientific explanation therefore cannot guarantee that it has discovered the truth, but only that it has given us a fitting answer to a question that was formed relative to a particular description of nature. However, this returns us to the original point: explanation is always relative to a question. It is based on a point of view. This does not mean that we are creating the world; instead, we are intended to understand that any linguistic or theory-bound explanation of the world will always be incomplete. In formulating scientific explanations we are seeking to grasp a reality that—due to the limitations of our language—we will never completely understand.

Yet, science is marvelous. And it can work equally well for simple, well-ordered phenomena as it does for complex, emergent, reflexive phenomena. Can scientific explanation occur at many different levels of description? Yes—just look at the difference between a physicist’s conception of a molecule and a chemist’s. Can science work in explaining human behavior, when the person is trying his or her best to be unpredictable? Yes—just look at the latest models of cognitive irrationality in behavioral economics. Science need not be reductive to explain, but this does not mean that we need to cleave our understanding of science in two. We can be antirealists and anti-reductionists and still embrace scientific explanation, once we understand what scientific explanation is truly about. Scientific explanation can proceed at many different levels of description; there are a multitude of scientific explanations even for one and the same reality.

Science is about a commitment to following the evidence, in search of causation, wherever it may lead. This means that we must be willing to redescribe the world when our scientific vocabulary is failing us and to change our theories altogether when they are not explaining what it is that we seek to know. Why can science do this? Because science is not defined by its subject matter or its method: neither is it defined by its commitment to reductionism, realism, or any preconceived notions of simplicity or complexity. Science is instead defined by its attitude toward evidence: toward being genuinely interested in having something explained, because we are secure enough to understand that even though we may have a preferred vocabulary, and perhaps even some ideological prejudices, we do not understand something until it has been scientifically scrutinized.

Nearly everything I have said so far about the scientific attitude can be summed up by a physicist who largely deplored the role of philosophy in understanding science. In his own effort to capture what was special about science, Richard Feynman once said to an audience of undergraduates: “It doesn’t matter how beautiful your theory is, it doesn’t matter how smart you are. If it doesn’t agree with experiment, it’s wrong.”

The ability to admit that we may be wrong . . .  that our understanding is incomplete . . .  that our ideology may reflect only wishful thinking . . .  that our descriptions may capture only a piece of the true nature of reality—that is the scientific attitude toward explanation.

AFTERWORD 2

Explanation Revisited

Jan Faye

Explanations spring up everywhere where people are. And for good reasons. Explanation is one of the main sources of human understanding. We may even say that the purpose or function of an explanation is to communicate understanding. The explainer is involved in an intentional act in which he or she expresses his or her understanding of a particular issue to others, perhaps to address an explicit question raised by an interlocutor. How this understanding is received by the other person is governed by the communicative situation in which the explainer and explainee take part, their background knowledge, the subject being entertained, and their cognitive and personal interests.

Explanations take many forms and address many types of questions. One can find explanations that appeal to laws, causes, functions, structures, analogues, intentions, interpretations, or what have you; and what is being explained are effects, traces, outcomes, movements, behavior, actions, appearances, and meanings. Explanations do not only address how- and why-questions, other explanation-seeking questions could be, for instance, what-questions. It all depends on the topic of explanation, what the explainee wants to understand, and the explainer´s ability to provide a convincing answer. For some, this may seem both plain and obvious, but only recently almost all philosophers took a different view. Their focus was on the appropriate logical form of explanation or the appropriate content of an explanation. Now more philosophers realize that explanations are inseparable from the way we use them and the context which gives rise to them. This is also how I see it. In my opinion, explanation arises in a rhetorical situation that contains an exigency (an epistemic problem), an audience, and some constraining factors.

In her contribution, Nancy Nersessian argues that scientists focus more on understanding than explanation, whereas philosophers are more occupied with explanation than understanding. She believes that there are historical reasons for that, and I think she is right. But I also think that she misses the point that explanation is not an act of reason that can be separated from the communicative practice of science. Understanding, scientific or otherwise, has to be communicated to others and the only way this can be done is by publishing explanations in journals, textbooks, talks, handouts, power point presentations, etc. An explanation need not have a fixed structure or a definite content. In practice, no scientists do Hempelian-style explanation or anything like that, nor do they have to appeal to causal mechanisms or the like. It is the scientists’ intentions that determine whether or not a description counts as an explanation. Much of scientists’ understanding comes by reading explanations given by others in journals, textbooks, emails, or by listening to explanation in talks, interpersonal communication, teaching, etc. In those contexts, the word “explain” or “explanation” may seldom be used. Instead, the context together with the argumentative or narrative structure of the text or the use of words associated with proofs, processes, mechanisms, functions, structures, intentions, meanings, etc. in the text or in the speech act gives the audience the impression that the description functions as an explanation due to the intentions of the communicator. So when philosophers explicitly use words like “explain” and “explanation” it may very well be in settings where they attempt to give a philosophical reconstruction of science.

However, explanation is not the only way by which one may achieve understanding which I take to be organization of beliefs. Understanding also comes with interpretation in which one relates meaning to the phenomenon under investigation. In these cases, interpretation is a process of constructing meaning, not a process of explaining meaning. This is what happens, I think, in the examples Nersessian mentioned. Scientists construct models as representations of a process or a function and such constructions may be top-down from some highly theoretical principles or bottom-up from experimental data and physical or computational simulation. How understanding is related to both explanation and interpretation is something I have elaborated in detail in my most recent book (Faye, 2014).

The editors of the present book, Michael Lissack and Abraham Graber, draw a distinction between Science 1 and Science 2 coinciding with the natural sciences on the one hand and the social sciences and the humanities on the other. Indeed others have seen a similar distinction. Gustav Droysen and Wilhelm Dilthey described it as a distinction between the disciplines aiming at explanation and those seeking understanding. William Windelband and Heinrich Rickert later characterized the distinction as one between the nomothetic and idiographic disciplines. But humanistic theories like structuralism, generative grammar, semiotics, formal semantics, behaviorism, and cognitive science prove these characterizations dead wrong—relics of a bygone age. Science 2 explains as much as Science 1 and the latter involves understanding as much as the former. I think it is fair to say that the functional role of explanation is to provide understanding and that the natural sciences use interpretation as much as social science and the humanities. But how you define your research object depends on your research interest. It is not given by nature herself whether you should see an object nomotetically as only one of a kind or idiographically as a unique individual.

Lissack and Graber take Science 1 to target at objectivity, truth, universal laws, invariance, and context-free descriptions by the use of models of representing in contrast to Science 2 which they see attempts to understand contingencies, individuals, meaning, and how human considerations, deliberations, and decisions are context-bound. They associate Science 1 with a worldview of scientific realism, according to which the world consists of unobservable and mind-independent entities, very much similar to Kant’s things-in-themselves. Science 2 does not share this worldview; rather the worldview underlying it can be identified with pragmatic constructivism. It works with meaning and contexts, with how context influences our thinking and bestows meaning to our actions. This position considers the world in which we live and act as a result of useful constructions. Like all generalizations, these worldviews only hold at a certain level of abstraction. Unfortunately, this has become the ideology of the sciences as well as the humanities and has been nurtured by philosophers, scholars, and scientists ever since.

If we look at the scientific practices in Science I and Science 2, respectively, I think it is obvious that the distinction generates a false dichotomy. The result of Science 1 is not less dependent on interest and context than Science 2, and Science 2 needs not be less objective and truth-tracking than Science 1. Several of the many quotations in chapter 3 attest to this claim. Context, interest, and contingency do not preclude truth and objectivity, nor does the opposite. I prefer what may be called constructive realism that has room for both atoms and human intentions, including as real those social institutions that follow from such assumptions. I do not claim that the contrast between Science 1 and Science 2 does not make sense. In some sense, it is important. It tells us something about which explanation we usually will select, since which types of explanation we pick as relevant is determined by how we understand the topic of the explanation. Here, Science 1 and Science 2 deal with topics with different ontological characteristics. The objects of Science 1 are real but not intentionally constituted, whereas those of Science 2 are real but intentionally invented. The acclaimed ontological differences make some explanatory genres appropriate and others inappropriate. In short, the distinction reflects that we often exhibit contrasting cognitive interests in Science 1 and Science 2 because we believe their research objects come into the world by different causes. A good explanation is a linguistic response to an explanation-seeking question that has been adjusted to the explanatory situation, including the nature of the topic, the cognitive interests involved, and the audience’s expectation. This point brings me to another important issue.

Explanations and methods are quite distinct. In my opinion, explanations reflect our cognitive interest, which undoubtedly varies from understanding the atmosphere of the planet Venus to Titian’s Venus of Urbino. Most of working scientists and scholars think of these phenomena as belonging to different ontological categories. However, the general methods of science by which we can possibly warrant our explanation are the same for Science 1 and Science 2. Claims within Science 2 must be justified by empirical evidence as much as claims within Science 1. Such a requirement is due to the fact that the practice-oriented methods as well as the most sophisticated methods have their origin in the cognitive evolution of Homo sapiens. Ancestors who were not adapted to using cognitive mechanisms such as induction and deduction on sensory information did not have a chance of survival. Lions took them. Even very sophisticated methods of science have to answer our innate cognitive mechanism for collecting, interpreting, and treating information. So Science 1 and Science 2 are united, in the sense all disciplines have to stand up to what we call the empirical methods; comparing their explanations with supporting evidence and data is the only legitimate way scientists and scholars can make sure that their explanations are true. This is how our biology has given us the ability to stay tuned to the world and select between all our free thoughts, imaginations, and speculations. (For a further vindication of a methodological unification of Science 1 and Science 2, I confer the reader to an earlier book of mine (Faye 2010), which gives a naturalistic reconstruction of the humanities.)

This takes me to my comments on the case study on creationism and intelligent design (ID). Most philosophers have seen explanation as a statement that has to be true in order to count as an explanation. But I do not share this belief. Indeed, we all want our explanations to be true, but it is not a defining characteristic of an explanation that it has t...

Table des matières

• Cover
• Title
• Context
• Case Study
• Examining the Case
• Dialogue
• Conclusion
• Afterword 1: The Scientific Attitude Toward Explanation
• Afterword 2: Explanation Revisited
• Afterword 3: Is The World Completely Intelligible? A Very Short Course
• Afterword 4: Explanation and Pluralism
• Reprise
• References
• Suggestions for Further Reading
• Notes on Contributors
• Index