Chapter 2: Models and Paradigms

Religion in an Age of Science
by Ian Barbour

Chapter 2: Models and Paradigms

This chapter examines some parallels between the methods of science and those of religion. It develops the Methodological Parallels position discussed under Dialogue in the previous chapter. Whereas advocates of Independence see only the differences between science and religion, advocates of Dialogue usually point to similarities. The differences will not be ignored here, though they will be more specifically explored in chapter 3.

We begin with a comparison of the general structures of scientific and religious thought. Then the role of conceptual models in both fields is analyzed. A summary of the debate over the role of paradigms in science follows, and some possible parallels in religion are presented. In the final section, we consider the balance between tentativeness and commitment in each field.

I. The Structures of Science and Religion

We look first at the relation between the two basic components of science: data and theory. It is then suggested that in religion the data are religious experience, story, and ritual, and that religious beliefs have some functions similar to those of scientific theories. The distinctive features of religious story and ritual are also discussed.1

1. Theory and Data in Science

The fundamental components of modern science are: (1) particular observations and experimental data, and (2) general concepts and theories. How are theories related to data? Since Bacon and Mill, the inductive view has held that the scientist starts with observations and formulates theories by generalizing the patterns in the data (this would be represented by an arrow upward from data to theories in figure 1). But this view is inadequate because theories involve novel concepts and hypotheses not found in the data, and they often refer to entities and relationships that are not directly observable.

Observation/Data………Imagination/Analogies/Models………..Concepts/Theories………Theories influence observation

Fig. 1 The Structure of Science

There is, then, no direct upward line of logical reasoning from data to theories in the diagram, but only the indirect line at the left, representing acts of creative imagination for which no rules can be given. Often a new concept or relationship is first thought of by analogy with a more familiar concept or relationship, but with a novel modification or adaptation. Frequently the analogy is systematically developed as a conceptual model of a postulated entity that cannot be directly observed. The model leads to the formulation of a generalized and abstract theory. For example, the billiard ball model of a gas postulated invisible gas particles that were imagined to collide and bounce off each other like billiard balls. From the model, the kinetic theory of gases was developed.

To be scientifically useful, a theory must be tested experimentally. A theory leads us to expect some observations and not others. This is the hypothetico-deductive view of science, represented by the downward arrow from theory to observation. The context of discovery (left-hand loop) differs from the context of justification (downward arrow). If a theory or hypothesis is valid, then particular observational patterns are expected, though the reasoning process always involves a variety of background assumptions, auxiliary hypotheses, and rules of correspondence linking theoretical and observational terms. In the case of the kinetic theory of gases, we can calculate the change in the momentum of the hypothetical particles when they strike the walls of the containing vessel. If we assume perfectly elastic collisions and particles of negligible size, we can derive Boyle’s Law relating the observed pressure and volume of a gas sample. The corroboration of such deductions leads us at least tentatively to accept a theory.2

This hypothetico-deductive view dominated philosophy of science in the 1950s and early 1960s. It assumed that data are describable in a theory-free observation language and that alternative theories are tested against these fixed, objective data. Even though agreement with data does not verify a theory (since there may be other theories that would also agree), it was claimed by Karl Popper and others that disagreement with data will conclusively falsify a theory. But studies in the history of science cast doubt on this claim.

In some cases, discordant data were brought into harmony with a theoretical prediction by the introduction of ad hoc auxiliary hypotheses. Early opponents of Copernican astronomy said the hypothesis that the earth moves around the sun must be false because there is no visible annual change in the apparent position of near stars relative to distant stars. But Copernicus dismissed this discrepancy by introducing the hypothesis (for which there was then no independent evidence) that all the stars are very distant compared to the size of the solar system. In other historical cases a theory was retained without modification and the discordant data were simply set to one side as an unexplained anomaly. Newton in his Principia admitted that the observed motion of the apogee (the most distant point) of the moon’s elliptical orbit in successive revolutions was twice that predicted by his theory. For sixty years the disagreement, which far exceeded the limits of experimental error, could not be accounted for, yet it was never taken to disprove the theory.

We can never test a theory alone, but only as part of a network of theories. If a theory fits poorly with the data at one point, other parts of the network can usually be adjusted to improve the fit. Theories with terms far from the observational boundaries are not uniquely determined by the data.3 Normally, a group of background theories is simply assumed and treated as unproblematic while attention is directed to a new or controversial theory. In many scientific disputes, the contending parties agree on most of these background assumptions, and so they can agree on the kinds of experimental data that both sides will accept as a crucial test for adjudicating between rival theories. But in some cases two theories of broad scope involve differing ways of interpreting the data, or they are correlated with differing bodies of data or differing types of explanation, and no simple experimental adjudication is possible.

Moreover, all data are theory-laden. There simply is no theory-free observational language. Theories influence observations in many ways (as shown in the right-hand loop in the diagram). The selection of phenomena to study and the choice of variables considered significant to measure are theory-dependent. The form of the questions we ask determines the kind of answers we receive. Theories are reflected in our assumptions about the operation of our equipment and in the language in which observations are reported.4 This account differs sharply from the empiricist account, in which the edifice of knowledge is built on the secure foundation of unchanging facts.

In addition, the object observed may be altered by the process of observation itself. We will see that this is particularly problematic in the microworld of quantum physics and in the complex networks of ecosystems. We are not detached observers separate from observed objects; we are participant observers who are part of an interactive system.

Thomas Kuhn has argued that scientific data are strongly dependent on dominant paradigms. A paradigm, as we have seen, is a cluster of conceptual and methodological presuppositions embodied in an exemplary body of scientific work, such as Newtonian mechanics in the eighteenth century or relativity and quantum physics in the twentieth century. A paradigm implicitly defines for a given scientific community the kinds of questions that may fruitfully be asked and the types of explanations to be sought. Through standard examples, students learn what kinds of entities exist in the world and what methods are suitable for studying them. A paradigm shift is "a scientific revolution," "a radical transformation of the scientific imagination," which is not unequivocally determined by experimental data or by the normal criteria of research. Accepted paradigms are thus more resistant to change and more difficult to overthrow than are particular theories. Paradigms are the products of particular historical communities.5 Here we. see a contextualism, a historicism, and a relativism contrasting with the formalism and the empiricism of Popper’s account.

There are four criteria for assessing theories in normal scientific research:

1. Agreement with Data. This is the most important criterion, though it never provides proof that a theory is true. For other theories not yet developed may fit the data as well or better. Theories are always underdetermined by data. Nor does disagreement with data prove a theory false, since ad hoc modifications or unexplained anomalies can be tolerated for an indefinite period. However, agreement with data and predictive success -- especially the prediction of novel phenomena not previously anticipated -- constitute impressive support for a theory.

2. Coherence. A theory should be consistent with other accepted theories and, if possible, conceptually interconnected with them. Scientists also value the internal coherence and simplicity of a theory (simplicity of formal structure, smallest number of independent or ad hoc assumptions, aesthetic elegance, transformational symmetry, and so forth).

3. Scope. Theories can be judged by their comprehensiveness or generality. A theory is valued if it unifies previously disparate domains, if it is supported by a variety of kinds of evidence, or if it is applicable to wide ranges of the relevant variables.

4. Fertility. A theory is evaluated not just by its past accomplishments but by its current ability and future promise in providing the framework for an ongoing research program. Is the theory fruitful in encouraging further theoretical elaboration, in generating new hypotheses, and in suggesting new experiments? Attention is directed here to the continuing research activity of a scientific community rather than to the finished product of their work.

Western thought has included three main views of truth, and each emphasizes particular criteria from the list above. The correspondence view says that a proposition is true if it corresponds to reality. This is the common-sense understanding of truth. The statement "it is raining" is true if in fact it is raining. This is the position adopted by classical realism, and it seems to fit the empirical side of science as specified by the first criterion: theories must agree with data. But we have said that there are no theory-free data with which a theory can be compared. Many theories postulate unobservable entities only indirectly related to observable data. We have no direct access to reality to compare it with our theories.

The coherence view says that a set of propositions is true if it is comprehensive and internally coherent. In mathematics one has a system of logically related statements, none of which can be judged alone apart from the others. In science there are empirical statements, but they turn out to be complex interpretations expressed in propositions, so one ends up judging the coherence of propositions. This view has been adopted by rationalists and philosophical idealists, and it seems to fit the theoretical side of science. We have said that a single theory can never be evaluated in isolation, but only as part of a network of theories. Our second and third criteria were coherence and scope. But this position is also problematic, since there may be more than one internally coherent set of theories in a given domain. Moreover, judgments of agreement with data differ in character from judgments of internal coherence and cannot be assimilated to the latter. In addition, reality seems to be more paradoxical and less logical than the rationalists assume.

The pragmatic view says that a proposition is true if it works in practice. We should judge by the consequences. Is an idea fruitful and suggestive? Is it useful in satisfying individual and social needs and interests? Ideas and theories are guides to action in particular contexts. Instrumentalists and linguistic analysts usually dismiss questions of truth, and they talk only about the diverse functions of language. But they often adopt a pragmatic view of scientific language. There is a pragmatic element in Kuhn’s thesis that scientific inquiry is problem-solving in a particular historical context and within a particular paradigm community. This side of science is reflected in our fourth criterion: fertility. But taken alone this criterion is inadequate; whether an idea "works" or is "useful" remains vague unless these concepts are further specified by other criteria. Even to ask the question "could a false idea have useful consequences?" shows that we distinguish between the meanings of truth and usefulness.

My own conclusion is that the meaning of truth is correspondence with reality. But because reality is inaccessible to us, the criteria of truth must include all four of the criteria mentioned above. The criteria taken together include the valid insights in all these views of truth. One or another of the criteria may be more important than the others at a particular stage of scientific inquiry. Because correspondence is taken as the definition of truth, this is a form of realism, but it is a critical realism because a combination of criteria are used. I will be advocating such a critical realism throughout this volume.

In sum, science does not lead to certainty. Its conclusions are always incomplete, tentative, and subject to revision. Theories change in time, and we should expect current theories to be modified or overthrown, as previous ones have been. But science does offer reliable procedures for testing and evaluating theories by a complex set of criteria. We will later examine the role of individual judgment and the traditions of particular scientific communities in the application of these criteria.

2. Belief and Experience In Religion

The basic structure of religion is similar to that of science in some respects, though it differs at several crucial points. The data for a religious community consist of the distinctive experiences of individuals and the stories and rituals of a religious tradition. Let us start by considering religious experience, which is always interpreted by a set of concepts and beliefs. These concepts and beliefs are not the product of logical reasoning from the data; they result from acts of creative imagination in which, as in the scientific case, analogies and models are prominent (figure 2). Models are also drawn from the stories of a tradition and express the structural elements that recur in dynamic form in narratives. Models, in turn, lead to abstract concepts and articulated beliefs that are systematically formalized as theological doctrines.


Religious experience/Story and ritual……..Imagination/Analogies/Models………Concepts/Beliefs………..Beliefs influence experience and interpretation

Fig. 2 The Structure of Religion

The experiential testing of religious beliefs is problematic (so the downward arrow is shown as a dashed line), though we will find that there are criteria for judging the adequacy of beliefs. Moreover, there are no uninterpreted experiences, as there are no theory-free data in science. Religious beliefs influence experience and the interpretation of traditional stories and rituals (the loop on the right of the diagram) -- an even stronger influence than that of scientific theories on data. Here, too, paradigms are extraordinarily resistant to change, and when paradigm shifts do occur a whole network of conceptual and methodological assumptions is altered. We will examine in turn each of these features of religious life and thought.

Six distinctive types of religious experience recur in a variety of traditions around the world.6

1. Numinous Experience of the Holy. Persons in many cultures have described a sense of awe and reverence, mystery and wonder, holiness and sacredness. Participants may experience a sense of otherness, confrontation, and encounter, or of being grasped and laid hold of. Here individuals typically express awareness of their dependence, finitude, limitation, and contingency. The experience is often interpreted in terms of a personal model of the divine. This pattern is found in both Western and Eastern [Asian] religions but is more prominent in the West. It emphasizes a strong contrast between the finitude of the human and the transcendence of the divine.

2. Mystical Experience of Unity. Mystics in many traditions have spoken of the experience of the unity of all things, found in the depth of the individual soul and in the world of nature. Unity is achieved in the discipline of meditation and is characterized by joy, harmony, serenity, and peace. In its extreme form the unity may be described as selflessness and loss of individuality and the joy as bliss or rapture. The experience is often correlated with impersonal models of the divine, especially in Eastern traditions, though it occurs in the West with both personal and impersonal models. Here the unity rather than the separation of the human and divine is emphasized. The numinous and the mystical seem to be the most common types of religious experience around the world.

3. Transformative Experience of Reorientation. In the lives of some individuals, acknowledgment of guilt has been followed by the experience of forgiveness. Others have described a transition from brokenness and estrangement to wholeness and reconciliation. Some experience a healing of internal divisions or a restoration of relationship with other persons. Such reorientation and renewal, whether sudden or gradual, may lead to self-acceptance, liberation from self-centeredness, openness to new possibilities in one’s life, a greater sensitivity to other persons, or perhaps dedication to a style of life based on radical trust and love. Such transformative experiences are prominent in the Christian tradition, but parallels are found in many traditions.

4. Courage in Facing Suffering and Death. Suffering, death, and transiency are universal human experiences, and responses to them are found in virtually every religious tradition. Meaninglessness is overcome when people view human existence in a wider context of meaning, beyond the life of the individual. Attitudes toward suffering and death are affected when trust replaces anxiety (in the West), or when detachment replaces the attachment that gives suffering and death their power over us (in the East). Such experiences can, of course, be described in psychological terms, but in religious traditions they are understood in relation to a view of ultimate reality beyond the individual.

5. Moral Experience of Obligation. Many people have felt moral demands overriding their own inclinations. Though the voice of conscience is in part the product of social conditioning, it may also lead persons to express judgment on their culture or moral outrage in the face of evil, even at the risk of death. Judgments of good and evil, right and wrong, are made in the light of one’s view of the nature of ultimate reality. Moral demands may be understood as the will of a God of justice and love or as a requirement for harmony with the cosmic process. In the West, prophetic protest against social injustice has been viewed as a response to God’s purposes.

6. Experience of Order and Creativity in the World. At the intellectual level, the presence of order and creativity in nature has served as the basis for inferring a divine source of order, beauty, and novelty (as in the classical argument from design). At the experiential level, people have responded to the world with reverence and appreciation, with gratitude for the gift of life, and with wonder that nature has a rational order intelligible to our minds. In the numinous tradition this is expressed as a dependence on a Creator who is the ground of order and creativity. In the mystical tradition it is more often articulated as dependence on a creative force immanent within nature.

Such experiences sometimes appear private and individual, but they occur in the context of a community. Experience is always affected by prior expectations and beliefs. The founders of new traditions started with inherited cultural assumptions, even if they challenged some of those assumptions. After their distinctive experiences, they evoked powerful responses among their followers. In subsequent generations, the experiences of individuals were subjected to a process of sorting and selecting within the ongoing community. The group affirmed some forms of experience and not others, and it set limits on acceptable beliefs -- though these limits have changed historically and may allow for considerable reformulation. Most traditions ‘have included prophetic figures who criticized accepted ideas and practices, while those in priestly roles were more often dedicated to continuity and the preservation of the past. There have been periods of codification and institutionalization, and periods of reformation and change.

If the task of the theologian is systematic reflection on the life and thought of the religious community, this will include critical assessment according to particular criteria. I suggest that assessment of beliefs within a paradigm community can be undertaken with the same criteria listed above for scientific theories, though the criteria will have to be applied somewhat differently. (The questions of assessing the paradigms themselves and judging among religious traditions are taken up in the next chapter.)

1. Agreement with Data. Religious beliefs must provide a faithful rendition of the areas of experience that are taken by the community to be especially significant. I have argued that the primary data are individual religious experience and communal story and ritual. Here the data are much more theory-laden than in the case of science. We will have to examine the influence of beliefs on experience and on the interpretation of story and ritual.

2. Coherence. Consistency with other accepted beliefs ensures the continuity of a paradigm tradition. The intersubjective judgment of the community provides protection against individualism and arbitrariness. But there is room for reformulation and reinterpretation, and the ideas of religious communities have indeed undergone considerable change throughout history. There are also close internal relationships among a ~t of religious beliefs.

3. Scope. Religious beliefs can be extended to interpret other kinds of human experience beyond the primary data, particularly other aspects of our personal and social lives. In a scientific age, they must also at least be consistent with the findings of science. Religious beliefs can contribute to a coherent world view and a comprehensive metaphysics.

4. Fertility. In the case of science, theories are judged partly by their promise for encouraging an ongoing research program, which is the central activity of science. Because religion involves a greater diversity of activities and serves some functions quite different from those of science, fertility here has many dimensions. At the personal level, religious beliefs can be judged by their power to effect personal transformation and the integration of personality. What are their effects on human character? Do they have the capacity to inspire and sustain compassion and create love? Are they relevant to urgent issues of our age, for example, environmental destruction and nuclear war? Judgments on such questions will of course be paradigm-dependent, but they are an important part of the evaluation of religion as a way of life. These questions are explored later in this chapter.

3. Story and Ritual in Christianity

In addition to religious experience, religious tradition includes a second form of data, namely a set of stories and rituals. Traditions are transmitted primarily through stories and their reenactment in rituals, rather than through abstract concepts and doctrinal beliefs. Religious stories were initially the products of experiences and events, interpreted imaginatively (an activity belonging on the left side of the diagram in figure 2). But the stories were later recorded in scriptures and became part of the data to which people responded in subsequent generations. Many scholars of religion use the term myths to refer to the central narratives of a religious tradition, insisting that the term does not imply any judgment either for or against the narratives’ historicity or validity. However, in popular usage, a myth refers to a fictional and untrue tale, so I have come to prefer the term story, since the status of a story is clearly left open.

The central religious stories are taken to manifest the character of the cosmic order and our relationship to it. They are significant in personal and communal life because they endorse particular ways of ordering experience, and they provide exemplary patterns for human actions. Such stories inform us about ourselves; our self-identity as individuals and as communities is in part constituted by these narratives. They are recalled in liturgy and acted out in ritual. Past events become present (re-presented) in symbolic reenactment. Creation stories found in most cultures portray the essential structures of reality and the cosmic context for human existence. Other stories exhibit a saving power in human life that can overcome some of its flaws or distortions (variously seen as sin, ignorance, or attachment). The power to transform life and restore relationships may be expressed in a personal redeemer or in a law or discipline to be followed.7

It is important to look at particular religions rather than at religion in general; I will primarily consider the Christian tradition, though I will give some examples from other traditions. Christianity re-presents three central stories.

1. The Creation of the World. The opening chapters of Genesis set human life in a framework of significance and meaning. They portray a world that is good, orderly, and coherent. They picture a God who is free, transcendent, and purposeful. These theological affirmations are conveyed through a dramatic narrative, which assumes a prescientific cosmology. In chapter 5 we will consider the interpretation of this story in a scientific age. We will note the connections of the creation story with human experience, theological doctrine, ritual practice, and ethical action. For example, the liturgies of ancient psalms and modern hymns and prayers give recurrent voice to gratitude for the created order. We will see also how a view of creation affects attitudes toward nature and the ways in which we treat the environment.

2. The Covenant with Israel. The Exodus stories of liberation from captivity in Egypt and the giving of the covenant at Sinai are central in Judaism but are also significant in Christian identity. Here the community’s existence is understood as response to a God who is liberator and redeemer as well as creator. It is not surprising that liberation theology among oppressed groups today (blacks, women, Third World nations) has given prominence to the exodus theme. Rituals such as Passover and the liturgies expressing gratitude for the Torah lie at the heart of Judaism and have influenced Christian worship and ethics. Most biblical scholars today hold that many of the details of the Law come from later centuries, but they trace the distinctive features of ethical monotheism and the concept of covenant to the time of Moses. The stories, that is, originated in historical events, but as they appear in scripture they involve centuries of elaboration and interpretation.

3. The Life of Christ. The most important stories for the Christian community recount the life, teachings, death, and resurrection of Christ. These narratives, historically based but inescapably involving interpretation, are central to individual and communal religious identity. The most prominent ritual (Eucharist or Lord’s Supper) and festivals (Christmas and Easter) celebrate and re-present crucial portions of this story. The early Christians wrote of their experience of liberation from anxiety and the fear of death and their empowerment to new patterns of life, which for them was connected with the person of Christ and the continuing activity of God as Holy Spirit. The story continues in the community’s response to the life of Christ (recorded in the book of Acts, Paul’s letters, and subsequent Christian literature).

Each of the major world religions has its own central stories. Hinduism, for instance, tells creation stories portraying the cosmic order as a context for human life. The most popular Hindu scripture, the Bhagavad Gita, recounts the story of Arjuna’s dialogue with Krishna (in the form of a charioteer) on the eve of battle. In the course of the dialogue, the three classical patterns of Hindu religious life are set forth: the way of works (carrying out one’s social duties and home rituals without excessive attachment to them); the way of knowledge (disciplined meditation seeking unity with the all-inclusive Brahman); and, finally, the way of devotion (loving devotion to a personal deity, such as the compassionate Krishna). The Gita includes examples of numinous experience (Arjuna’s awe-inspiring vision of the power of the god Vishnu) and examples of mystical experience (liberation from the illusion of selfhood through the peace of participation in the Infinite which pervades all things). These two strands come together when the personal deity, Krishna, is recognized as one of the manifestations or faces of Brahman, the impersonal Absolute.8

The data of religion, then, are the characteristic experiences, stories, and rituals of particular religious communities. Often the early memories of formative experiences and events are recorded in scriptures, to which members of the community respond in later generations, adding new layers of experience and ritual, Systematic concepts, beliefs, and doctrines are elaborated and reformulated to interpret these primary religious phenomena.

II. The Role of Models

Within these general structures of experience and interpretation, the role of models is particularly interesting both in science and in religion.

1. Models in Science

We have seen that in science there is no direct route by logical reasoning from data to theory. Theories arise in acts of creative imagination in which models often play a role. Here we are talking about conceptual or theoretical models, not experimental or scale models constructed in the laboratory, nor logical or mathematical models, which are abstract and purely formal relationships. Theoretical models usually take the form of imagined mechanisms or processes postulated in a new domain by analogy with familiar mechanisms or processes.

Three general characteristics may be noted in theoretical models:9

1. Models are analogical. A scientist working in a new domain may posit entities having some of the properties of a familiar entity (the positive analogy) and some properties unlike those of the familiar entity (the negative analogy). The Bohr model of the atom, in which "planetary" electrons revolve in orbits around a central nucleus, resembled the familiar solar system in some of its dynamic properties, but the key assumption that only certain orbits are allowed (quantization) had no classical analogue at all. The model aided the formulation of the mathematical equations for the theory (for example, the equations for the energy levels of the electrons). It also suggested how theoretical terms characterizing entities not directly observable might be related to observable variables (for example, how the transition of an electron between two orbits might be related to the frequency of the light emitted).

2. Models contribute to the extension of theories, Some claim that a model is a temporarily useful psychological aid that can be discarded once the equations of the theory are formulated. But this ignores the fact that it is often the model rather than the theory that suggests its application to new phenomena or new domains. It was the billiard ball model that suggested how the kinetic theory of gases might be applied to gas diffusion, viscosity, and heat conduction. Moreover, the model was crucial to the modification of the theory. Gases under high pressure depart significantly from Boyle’s Law. This could be accounted for with a revised model (elastic spheres with finite volume and attractive forces), which departs from the simple billiard ball model, but which would not have occurred to anyone without the earlier model. The suggestiveness and open-endedness of models provide a continuing source of possible applications, extensions, and modifications of theories.

3. Models are intelligible as units. Models provide a mental picture whose unity can be more readily understood than that of a set of abstract equations. A model can be grasped as a whole, giving a vivid summary of complex relationships, which is useful in extending and applying the theory as well as in teaching it. Images are creative expressions of imagination in the sciences as well as in the humanities. The intuitive intelligibility of a model is, of course, no guarantee of its validity. Deductions from the theory to which the model leads must be tested carefully against the data, and more often than not the proposed model must be amended or discarded. Models are used to generate promising theories to test by the diverse criteria outlined earlier.

In the quantum theory that has replaced the Bohr model, mechanical models are given up and there are severe limitations on the use of visualizable models. Nevertheless, two basic models, the wave model and the particle model, underlie the formalisms of quantum theory and suggest ways of correlating theory and experiment. These two basic models cannot be satisfactorily unified (the wave/particle paradox), even though a unified set of equations can be provided in the abstract theory. From the theory we can predict only the probability that a measurement in the atomic or subatomic world will have a particular value; we cannot predict exact values for a measurement. The models are more than a temporary expedient, for they continue to contribute to the interpretation of the mathematical formalism and to the modification of the theory and its extension to new domains.

Some of the novel characteristics of quantum physics will be discussed in a later chapter. Here we note only that complementary models are used despite their problematic status. Bohr formulated the Complementarity Principle, recognizing that "a complete elucidation of one and the same object may require diverse points of view which defy a unique description."10 He acknowledged the interaction between subject and object and the importance of the particular experimental arrangement. But he also stressed the conceptual limitations of human understanding. We must choose between causal or spatiotemporal descriptions, between wave and particle models, between accurate knowledge of momentum or of position. We have successive and incomplete perspectives that cannot be neatly unified.

Such models and theories clearly cannot be taken as literal descriptions of entities in the world, as classical realism assumed. At the opposite extreme, instrumentalism holds that models and theories are calculating devices whose only function is to allow the correlation and prediction of observations. Instrumentalism sees them as heuristic fictions, useful only as intellectual instruments for organizing research and for controlling the world. According to instrumentalists, models and theories do not describe or refer to real entities in the world.

I have elsewhere defended the intermediate position of critical realism.11 On this view, models and theories are abstract symbol systems, which inadequately and selectively represent particular aspects of the world for specific purposes. This view preserves the scientist’s realistic intent while recognizing that models and theories are imaginative human constructs. Models, on this reading, are to be taken seriously but not literally; they are neither literal pictures nor useful fictions but limited and inadequate ways of imagining what is not observable. They make tentative ontological claims that there are entities in the world something like those postulated in the models.

Opponents of realism argue that successive scientific theories are not convergent, cumulative, or progressive. New theories often exhibit radical changes in conceptual framework rather than refinements that preserve and add to earlier concepts. The history of science is said to be littered with theories that were successful and fruitful in their day, but that were later totally rejected rather than being modified -- including Ptolemaic astronomy, phlogiston chemistry, catastrophic geology, Lamarckian evolution, caloric heat theory, and ether theories in physics.12

But recent years have seen a revival of interest in realism. Many books and articles on the subject have appeared in the last few years.13 For example, some have pointed out that new theories exhibit continuity as well as discontinuity in relation to the theories they replace. Usually some of the concepts in the old theory and much of the data accumulated under its guidance are carried over into the new context. Sometimes the laws of the old theory are actually included in the new theory as limiting cases. Thus the laws of classical mechanics are limiting cases of relativistic laws at low velocities, though the fundamental concepts have been radically redefined. Later theories typically provide a better empirical fit and extend to wider domains, so that one can indeed speak of progress according to the criteria listed earlier.

We have greater confidence in the existence of a theoretical entity, such as the electron, if it is linked to many different kinds of phenomena explored in diverse types of experiment. With a new theory, scientists believe they have a better understanding of the structure of the world, not just a more accurate formula for correlating observations. Theoretical concepts are tentative and revisable, but they are taken to characterize and refer to the world. Unless a theory is at least partially true, how can we account for its success in predicting entirely new phenomena with types of observation radically different from those that led to the theory? Science, in short, is at the same time a process of discovery and a venture in human imagination.

The basic assumption of realism is that existence is prior to theorizing. Constraints on our theorizing arise from structures and relationships already existing in nature. Scientific discoveries are often quite unexpected. Humility before the given is appropriate; we learn from nature in order to set limits on our imagination. While the history of science exhibits no simple convergence or "successive approximation," it does include a body of well-attested theory and data, most of which can be considered trustworthy, even though any part of it is revisable. Can anyone doubt, for example, that we know more about the human body than we did five hundred years ago, even though there is still much to be known, and some of our current ideas may be rejected?

Ernan McMullin defends a critical realist view of models, especially those postulating hidden structures. He holds that "a good model gives us insight into real structures, and that the long term success of a theory, in most cases, gives reason to believe that something like the theoretical entities of that theory actually exist."14 A good model, he says, is not a dispensable temporary expedient but a fruitful and open-ended source of continuing ideas for possible extensions and modifications. Like a poetic metaphor, it offers tentative suggestions for exploring a new domain. A structural model may change as research progresses, McMullin observes, but it also exhibits substantial continuity as the original model is extended. One of his examples is the model of continental drift, which proved inconsistent with geological data but which itself suggested the tectonic plate model -- a model supported by more recent evidence concerning mid-ocean rifts and earthquake zones.

Most scientists are incurably realist, but their confidence in the status of models and theoretical entities varies among fields and in different historical periods. Models of larger scale and more familiar types of structure tend to be viewed more realistically. A geologist is not likely to doubt the existence of tectonic plates or prehistoric dinosaurs, though neither can be directly observed. In 1866, Mendel postulated hypothetical "units of hereditary transmission." which were later identified as genes in chromosomes and more recently as long segments of DNA. As we move further from familiar objects, instruments greatly extend our powers of direct or indirect observation.

When we get to the strange subatomic world, common sense fails us and we cannot visualize what is going on. Quarks behave like nothing familiar to us, and their quantum numbers (arbitrarily named strangeness, charm, top, bottom, and color) specify abstract rules for the ways they combine and interact. Even here, I will propose later, our theories are an attempt to represent reality, though microreality is not like the everyday world and ordinary language is inadequate to describe it.

2. Models in Religion

Religious models, we have said, lead to beliefs that correlate patterns in human experience. In particular, models of the divine are crucial in the interpretation of religious experience. They represent in images the characteristics and relationships portrayed in narrative form in stories. But models are less conceptually articulated and less systematically developed than beliefs and doctrines, which take the form of propositional statements rather than narratives or images.

Like scientific models, religious models are analogical. Religious language often uses imaginative metaphors, symbols, and parables, all of which express analogies. The most frequently used and systematically developed analogies are incorporated in models, such as the model of God as Father. Religious models, too, are extensible. A model originating in religious experience and key historical events is extended to interpret other areas of individual and communal experience, and it may be modified in the process. Religious models are also unitary; they are grasped as a whole with vividness and immediacy. 15

As in the scientific case, I defend a critical realism that takes religious models seriously but not literally. They are neither literal descriptions of reality nor useful fictions, but human constructs that help us interpret experience by imagining what cannot be observed. The biblical prohibition of graven images or "any likeness" (Exodus 20:4) is both a rejection of idolatry and an acknowledgment that God cannot be adequately represented in visual imagery. The sense of awe and mystery associated with numinous experience is an additional safeguard against literalism. But we do not have to go to the opposite extreme and take religious models as psychologically useful fictions whose only function is to express and evoke distinctive ethical attitudes, as some instrumentalists hold.16

Janet Soskice has advocated a form of critical realism concerning models in both science and religion. In both cases, she suggests, there were originating experiences and events in which a model was first introduced, and there was a subsequent linguistic community and interpretive tradition which perpetuated it. "The sacred literature thus records the experience of the past and provides the descriptive language by which new experience may be interpreted."17 Particular models are emphasized if they illuminate similar experiences in the later history of the community. The models supported by the experience of many generations find continued literary expression and are used in liturgy and devotional practices.

Soskice also claims that the continuity of the linguistic community guarantees a continuity of reference for models in both science and religion (for example, reference to "electrons" or "God"), even though the descriptive terms used are revisable and change over time. I find her portrayal of the interaction of experiences and interpretive linguistic traditions very illuminating. But I suggest that our acceptance of the referential character of religious language must rest on contemporary evaluation by the criteria outlined above, rather than on linguistic continuity. There has been a continuous interpretive tradition in astrology for several thousand years, but I do not believe that the connections it makes between the planets and human life patterns are referential. Theologians have the tasks of analysis and reformulation as well as passing on a tradition.

Frank Brown has raised some questions about the relation between metaphoric and conceptual thought in theological reflection that are relevant to the discussion of models. 18 He starts from the prominence of metaphor in scripture. Should theologians translate such metaphors into concepts and doctrines that can be systematized and analyzed? No, says Brown, because metaphors cannot be fully expressed in concepts; their implications are open-ended and contextual. Moreover, metaphors will always be valuable in enabling us to re-describe our own experience and in their power to transform our personal lives. Concepts are abstract, but metaphoric symbols are experientially rich and are thus central in ritual and worship. Brown concludes that we have to move back and forth between metaphoric and conceptual modes of thought. I suggest that models can facilitate this dialectic, since they are more fully developed than metaphors and yet they are less abstract than concepts.

Religious models have additional functions without parallel in science, especially in expressing and evoking distinctive attitudes. We have said that religion is a way of life with practical as well as theoretical goals. The life-orienting and emotional power of religious models and their ability to affect value commitments should not be ignored. Models are crucial in the personal transformation and reorientation sought in most religious traditions. Some linguistic analysts and instrumentalists hold that religious language has only these noncognitive functions. I argue, in reply, that such noncognitive functions cannot stand alone because they presuppose cognitive beliefs. Religious traditions do endorse particular attitudes and ways of life, but they also make claims about reality.19

In science, models are always ancillary to theories. In religion, however, the models themselves are as important as conceptual beliefs, partly because of their close association with the stories prominent in religious life. Christian worship is based on those stories of creation, the covenant, and especially the life of Christ. The individual participates in communal ritual and liturgy that reenact and refer to portions of these stories. Narratives in dramatic form are more personally involving and evocative than models, which are relatively static, though models are less abstract than concepts. Moreover, biblical stories can often be correlated with our own life stories, which are also narrative in form. Nevertheless, the movement from stories to models to concepts and beliefs is a necessary part of the theological task of critical reflection.

3. Personal and Impersonal Models

The atomic and subatomic world cannot be directly observed, and its behavior suggests that it is very unlike the world of everyday objects. We have seen that it cannot be represented by any single model, but it can be partially understood through theories formulated with complementary models, such as wave and particle. In religion, too, we are dealing with a reality that cannot be directly observed and that is beyond our capacity to visualize. Here, too, we can admit our conceptual limitations and accept the role of complementary models.

Ninian Smart has traced among world religions the two basic types of religious experience described in the previous section: numinous encounter and mystical union. The first received its classic description in Rudolf Otto’s Idea of the Holy. Its characteristics are a sense of awe and reverence, mystery and wonder, holiness and sacredness. Typical examples are Isaiah’s vision in the temple, the call of Paul or Muhammad, or the theophany of Krishna to Arjuna in the Bhagavad Gita. Human responses to experiencing the numinous include worship, humility, and obedience.20

Smart shows that numinous experience is usually interpreted in personal models. Worshipers think of God as distinct and separate from themselves. The overwhelming character of the experience suggests an exalted view of the divine and an emphasis on transcendence, with a corresponding human self-abasement and recognition of finitude or sinfulness. The sense of being grasped and laid hold of unexpectedly seems to point to a divine initiative independent of human control. The gulf between God and humanity may seem to be so great that it can be bridged only by revelation from God’s side or by a divine savior. Winston King speaks of "the gap between worshiping man and the worshiped Ultimate," and he describes its symbolization in the rituals of personalistic theism, such as sacrifice, prayer, and devotional liturgy and practice.21

The second type of experience is mystical union, which does seem to have common features in different cultures, despite their diversity. Among these, we have seen, are intensity, immediacy, unitary consciousness, unexpectedness, joy, and serenity. The realization of unity can lead to liberation from self-centeredness. All dichotomies (human/divine, subject/object, time/eternity) seem to be overcome in identity with the One beyond time and space. Mysticism is expressed in meditation, contemplation, and an inner quest for enlightenment, rather than in communal worship or ritual.

The mystic is cautious in the use of models and may say that the object of the experience cannot be described. The via negativa asserts only what the divine is not. But the writings of mystics do make extensive use of analogies and models. Sometimes union with the divine is said to be like the most intense union of two lovers. In other cases, ultimate reality is thought of as a Self identical in essence with the individual self, or a world Soul with which one’s own soul is merged. More often mystical experience is interpreted with impersonal models. The self is absorbed in the pantheistic All, the impersonal Absolute, or the divine Ground. The distinction between subject and object is overcome in an all-embracing unity beyond all personal forms. The self loses its individuality "as a raindrop loses its separate identity in the ocean."

Smart demonstrates that although Western traditions have been predominantly numinous and Eastern traditions predominantly mystical, all the major world religions have included both types of experience and both types of models.22 Judaism, Christianity, and Islam all have influential mystical writings, along with more common patterns of worship of the holy as personal; in these writings, the gap between God and humanity is narrowed, but it is never obliterated in total identity. Early Buddhism followed more mystical meditative disciplines, but Mahayana Buddhism includes the numinous strand in the worship of the Eternal Buddha and the Bodhisattvas (especially in Amida Buddhism). In Hinduism, the bhakti path of devotion to personal deities has accompanied the jnana path of meditation, unitive awareness, and an impersonal Absolute. Ramanuja developed the more personal side of Hinduism, whereas Shankara developed the more impersonal side. Modern followers of the jnana path say that their view should be called nondualism rather than monism, since ultimate reality cannot be described in positive terms.

It seems to me appropriate, then, to speak of personal and impersonal religious models as complementary. People who use personal models are often the first to insist that they are inadequate and that God is not literally a person. Sometimes it is said that God is more than a person, and the more is usually spelled out in predominantly impersonal terms (divine Ground, creative power, and so forth). And those who rely mainly on impersonal models may speak of love and grace, or they may hold that the impersonal Absolute is approached through devotion to its personal manifestations. All models are partial and inadequate representations of what is beyond our ordinary categories of thought. Religious models often are analogies of relationships rather than attributes of the divine in isolation. Moreover, differing human temperaments may be more congenial to some patterns of experience and some kinds of models than to others.

The relative priority of personal and impersonal models is not, of course, a minor matter. Only with a personal God can there be decisive divine initiative. The ontological and epistemological distance between the divine and the human is a correlate of ideas of historical revelation, grace, and redemption. Western traditions have found more room for human individuality (which in extreme forms becomes individualism) and social activism, whereas the oriental quest for inner peace has more often led to quietism, though often accompanied by exemplary compassion and respect for all forms of life.

Because models function within a total network of ideas and attitudes, I do not suggest that the Hindu Brahman and the Christian God, or other models from different religious traditions, be considered complementary. But we could consider the use of personal and impersonal models within one paradigm community as complementary, paralleling the use of wave and particle models within quantum physics. Nevertheless, recognizing the diversity of models in our own tradition might help us appreciate the models of other traditions, which would be an important contribution in a religiously pluralistic world. Complementarity encourages us to view models neither as literal pictures nor as useful fictions, but as partial symbolic representations of what cannot be directly observed.

4. Christian Models

The writings of the theologian Sallie McFague provide a good example of the exploration of models in Christian thought. In Metaphorical Theology she starts from Paul Ricoeur’s insights on the importance of metaphor in religious language. A metaphor asserts similarity but it denies identity. In a metaphor, one term "both is and is not" like its analogue. Recognition of the limitations of religious language prevents idolatry of any one formulation, which is the temptation of literalism.23

McFague then discusses models in science and religion, drawing extensively on my earlier writing on this topic. She considers a model to be a systematic and relatively permanent metaphor. A model is more emotionally rich and less abstract than a concept, but it is more precise than a metaphor. Religious models arise from human experience, especially the experience of healing, renewal, and reorientation of patterns of life. Models order our experience, and their implications are systematically developed in doctrines. Whereas Ricoeur says that the purpose of theological interpretation is to return us to experience, McFague gives more emphasis to conceptual clarity and comprehensive ordering. Against naive realism on the one hand and instrumentalism on the other, she defends critical realism in both science and religion. Models are tentative, partial, open-ended, and paradigm-dependent. The dominant paradigm of a tradition sets limits on acceptable models.

McFague defends the use of a multiplicity of models within a paradigm community -- a greater multiplicity than is typical of science. Such multiplicity guards against the temptations of idolatry, absolutism, and literalism, which, appear when one model is dominant. Multiplicity is also appropriate because in both science and religion we are modeling relationships, patterns, and processes rather than separate entities or things-in-themselves. Religious models are analogies for our experience of relating to God, which takes a variety of forms that are not mutually exclusive. God can be related to us in both a fatherly and a motherly way -- and a rich diversity of other ways.

In her more recent book, Models of God, McFague discusses criteria for evaluating Christian models. She mentions general criteria such as comprehensiveness, internal consistency, and potential for dealing with anomalies. Another criterion is continuity with earlier expressions of the Christian paradigm. Scripture is important as the earliest witness to the experience of the transforming power of God and the earliest interpretations of the life and death of Christ as the transformative event. Moral fruitfulness serves as an additional criterion, and she gives special attention to relevance for the crises of "an ecological and nuclear age."24

As she turns to specific models, McFague criticizes the monarchical model, which has been dominant historically. God as King or Ruler is related to the world externally, not intrinsically. Here God controls by domination, acting on the world rather than through it, which undermines human responsibility. McFague’s first proposed alternative is to consider the world as God’s body. This goes to the opposite extreme in stressing immanence rather than transcendence. However, she does not indicate how this model is any more compatible with human freedom and responsibility than the monarchical model. The model of the world as God’s body suggests that the language of scientific laws and the language of divine intentions may be alternative ways of describing cosmic history.

The second half of the book examines in detail three personal models: God as Mother, Lover, and Friend. Each represents not the power of domination but the power of love in a particular form, described classically as agape, eros, and philia, respectively. The three models express in turn God’s activity as Creator, Savior, and Sustainer, and together they illuminate many of the themes of traditional theology. Thus God as Mother (or as Father if understood parentally rather than patriarchially) can draw on the experiences of the mystery of human birth and the nurturing of life. The model suggests an ethic of care and of justice. A mother’s concern for present and future life can be broadened to "universal parenthood," which includes care not only for the needs of present and future human generations but for the life of other species.

Similarly, the model of God as Friend points to a reciprocal mutual bond and also to a common vision that demands our action as coworkers. God suffers with us and works with us to extend the inclusive, holistic, and nonhierarchical vision of the fulfillment of all beings. I find these models very helpful in thinking about God’s relation to humanity and humanity’s relation to nature, but they seem less helpful in thinking about God’s relation to nature. McFague indicates at several points that she is sympathetic to process theology, but she does not explore the ways in which a process metaphysics might facilitate the conceptual articulation of the relationships suggested by the models. We will consider these and other specific Christian models in chapter 9.

III. The Role of Paradigms

In addition to parallels in the structures of scientific and religious inquiry, and in the role of imaginative models, there are some interesting similarities in the role of paradigms in the two fields. There are also, of course, some important differences that must be explored. We will look successively at paradigms in science, in religion in general, and then in Christian thought.

1. Paradigms in Science

Thomas Kuhn defined paradigms as "standard examples of scientific work that embody a set of conceptual and methodological assumptions. In the postscript to the second edition of his book he distinguished several features that he had previously treated together: a research tradition, the key historical examples through which the tradition is transmitted, and the metaphysical assumptions implicit in the fundamental concepts of the tradition. The key examples, such as Newton’s work in mechanics, implicitly define for subsequent generations the type of explanations that should be sought. They mold assumptions as to what kinds of entity there are in the world, what methods of inquiry are suitable for studying them, and what counts as data. A paradigm provides an ongoing research community with a framework for "normal science." Science education is an initiation into the habits of thought presented in standards texts and into the practices of established scientists.

Kuhn describes a major paradigm shift as a scientific revolution. A growing list of anomalies and ad hoc modifications within an existing paradigm produces a sense of crisis. Instead of simply acquiring further data or modifying theories within the existing framework, some scientists look for a new framework, which may involve a questioning of fundamental assumptions. Within the new paradigm, new kinds of data are relevant and the old data are reinterpreted and seen in a new way. The choice between the new and the old is not made by the normal criteria of research, Kuhn maintains. Adherents of rival paradigms will try to persuade each other. "Though each may hope to convert the other to his way of seeing his science and its problems, neither may hope to prove his case."25 Kuhn analyzes several historical "revolutions" in some detail. For example, he describes the radical change in concepts and assumptions that occurred when quantum physics and relativity replaced classical physics. Three features of Kuhn’s account are of particular interest.26

1. All data are paradigm-dependent. We noted earlier that there is no observation-language independent of theoretical assumptions. All data are theory-laden, and theories are paradigm-laden. The features of the world considered most important within one paradigm may be incidental in another. Kuhn claimed initially that paradigms are "incommensurable" (that is, they cannot be directly compared with each other). However, his later writings acknowledged that usually a core of observation statements exists on which the protagonists of rival paradigms can agree, a level of description that they can share. These common data are not free of theoretical assumptions, but some assumptions can be shared even by adherents of rival paradigms. If data were totally paradigm-dependent, they would be irrelevant to the choice of paradigms, which has not been the case historically.

2. Paradigms are resistant to falsification. Comprehensive theories, and the even broader paradigms in which they are embedded, are very difficult to overthrow. Discordant data, as we have seen, can usually be reconciled by modifying auxiliary assumptions or introducing special ad hoc hypotheses, or they can be set aside as unexplained anomalies. Paradigms are not rejected because there is contradictory evidence; they are replaced when there is a more promising alternative. Research can proceed when the theories of a paradigm do not fit all the data, but systematic research cannot proceed in the absence of a paradigm. Commitment to a research tradition and tenacity in developing its potentialities and extending its scope are scientifically fruitful, But observations do exert some control over a paradigm, and an accumulation of ad hoc hypotheses and unexplained anomalies can undermine confidence in it. Without persistent concern for fidelity to the data, science would be an arbitrary and subjective human construction.

3. There are no rules for paradigm choice. A paradigm change is a revolution, achieved more by "persuasion" and "conversion" than by logical argument. Kuhn initially maintained that criteria for choice are themselves paradigm-dependent. In response to his critics, he said that the decision to choose a certain paradigm is not arbitrary or irrational, since reasons can be given for the choice. He acknowledged that there are values common to all scientists and shared criteria of simplicity, coherence, and supporting evidence; but the way the criteria are applied and their relative weight are matters of personal judgment, not rules to be followed. The decision is like that of a judge weighing evidence in a difficult case, not like a computer performing a calculation. There is no court of appeal higher than the judgment of the scientific community itself. The presence of shared values and criteria allows communication and facilitates the eventual emergence of scientific consensus.27 Kuhn thus qualified his more extreme claims.

In recent decades there has emerged what Harold Brown calls "the new philosophy of science." Brown describes the move from empiricism to a more historical view of science as itself a paradigm shift in the philosophy of science. He describes the contributions of Hanson, Toulmin, Polanyi, and others, along with Kuhn, in the emergence of this new view that draws heavily on the history of science. Brown gives this summary:

Our central theme has been that it is ongoing research, rather than established results, that constitutes the lire blood of science. Science consists of a sequence or research projects, structured by accepted presuppositions which determine what observations are to be made, how they are to be interpreted, what phenomena are problematic, and how these problems are to be dealt with.28

Brown gives examples of "normal science," in which work was conducted within an accepted framework, and he describes several scientific revolutions that involved alternative presuppositions and "fundamental changes in the way we think about reality." But he maintains that a revolution shows continuity as well as discontinuity:

For the most part, old concepts are retained in altered form, and old observations are retained with new meanings. The continuity provides the basis for rational debate between alternative fundamental theories. . . . Thus the thesis that a scientific revolution requires a restructuring experience akin to a gestalt shift is compatible with the continuity of science and the rationality of scientific debate.29

Brown takes up the charge that the new view makes science appear subjective, irrational, and historically relative. To be sure, science does not fulfill the empiricists’ definition of objectivity as reliance on strict empirical verification or falsification, nor its definition of rationality as the application of impersonal rules. But science does conform to more appropriate definitions of objectivity and rationality. Objectivity should be identified with intersubjective testability and informed judgment in the community of qualified scientists. It is rational to accept a paradigm if it solves important problems and provides a guide to further research. Brown holds that "crucial decisions as to how a conflict between theory and observation is to be resolved, or how a proposed new theory is to be evaluated, are not made by the application of mechanical rules, but by reasoned judgments on the part of scientists and through debate within the scientific community."30

We can summarize our conclusions about scientific paradigms in three sentences. The first half of each sentence represents a subjective and historically relative feature of science that was neglected in the earlier empiricist accounts. The second half of each sentence represents a reformulation of the objective, empirical, and rational features of science that prevent it from being arbitrary or purely subjective:

1. All data are paradigm-dependent, but there are data on which adherents of rival paradigms can agree.

2. Paradigms are resistant to falsification by data, but data does cumulatively affect the acceptability of a paradigm.

3. There are no rules for paradigm choice, but there are shared criteria for judgment in evaluating paradigms.

Compared to empiricist accounts, then, Kuhn gives a much larger role to historical and cultural factors. He insists that a theory is judged within a network of theories and against a background of assumptions, in terms of its success in solving problems in a particular historical context. Kuhn is a contextualist, in contrast to the earlier formalists, but I do not think that this makes him a subjectivist or an unqualified relativist, for in his view the data do provide empirical constraints, and the presence of shared criteria does represent a defensible form of rationality.

2. Paradigms in Religion

As in the scientific case, a religious tradition transmits a broad set of metaphysical and methodological assumptions that we can call a paradigm. As in science, traditions in religion are passed on by particular communities, partly through respected historical texts and key examples. Here, too, new members enter a tradition by being initiated into the assumptions and practices of the community, and they normally work within its accepted framework of thought, which we can call "normal religion," corresponding to "normal science."

As in science, normal criteria are difficult to apply to major historical "revolutions" or to the choice between competing paradigms. Let us focus first on the relation of paradigm choice to religious experience, returning later to the role of story and ritual and their transmission through scriptures. Each of the three subjective and historically relative features of scientific paradigms listed above is even more evident in the case of religion. Each of the corresponding objective, empirical, and rational features of religion is more problematic. The questions raised are discussed in this chapter and the following one.

1. Religious experience is paradigm-dependent. But are some experiences common to the adherents of rival paradigms? Religious experience seems to be so strongly molded by the believer’s interpretive framework that a skeptic might claim that the experience is entirely the product of prior expectations. Religious experiences are not as publicly accessible as scientific data are, even though both are theory-laden. Yet there are common features of experience within a religious community which exert some control on the subjectivity of individual beliefs. And there do seem to be some characteristics of religious experience in diverse traditions that point beyond cultural relativism, and that make communication between traditions possible.

2. Religious paradigms are highly resistant to falsification. But does cumulative experience influence paradigm choice at all? Discordant data, we have said, does not lead directly to the overthrow of a paradigm. Instead, ad hoc modifications are introduced, or the data are set aside as an anomaly. Yet people may eventually modify or abandon their most fundamental religious beliefs in the light of their experience, especially if they see a promising alternative interpretive framework.

3. There are no rules for paradigm choice in religion. But are there shared criteria for evaluating religious paradigms? Some criteria were proposed above for evaluating beliefs within a dominant paradigm. Can these be applied to the choice between paradigms? Are the criteria themselves totally paradigm-dependent? I will suggest that there are indeed criteria transcending paradigm communities, though their application is a matter of individual judgment in more problematic ways than in the case of science.

Is it appropriate to speak of paradigms in Eastern religious traditions? A few years ago a conference was held in Hawaii entitled, "Paradigm Shifts in and Buddhism and Christianity." As background reading for the conference, participants received copies of the sections of my earlier book dealing with paradigms. In one of the plenary addresses, Frederick Streng maintained that my analysis was applicable to Christianity but not to Buddhism. Religious traditions, he said, have shared attitudes and conceptual structures, which are always closely tied to the experience of personal transformation and reorientation. Religion is above all a "strategy for living," a "means of ultimate transformation." Religious conversion is a change in awareness and in mode of living. Discussion of paradigms, Streng said, makes us look at systems of belief and doctrine, which are indeed important in Christianity. But Buddhism is more concerned about the transformation of consciousness to a less ego-centered awareness, and it urges nonattachment to doctrinal expressions and changing intellectual forms. It offers spiritual practices to achieve enlightened consciousness and to release us from the attachments that cause our suffering.31

A possible response to Streng’s objections would be to give greater emphasis to the primacy of religious experience in my proposed scheme and to downplay the role of concepts and beliefs. Yet surely it is legitimate to hold that Buddhism includes a network of characteristic concepts and beliefs, including the doctrine of "no-self," which imply ontological claims as well as existential commitments. Moreover, major historic changes have taken place in Buddhist thought as well as practice, such as the emergence of Mahayana from Theravada Buddhism, which other participants in this conference described as paradigm shifts. Buddhism may urge nonattachment to doctrinal forms, but it does not seem to have dismissed them entirely.

3. Paradigms in Christianity

A conference was held in Germany in 1983 on the topic "The New Paradigm for Theology." In one of the preparatory essays, the theologian Hans Küng applied the concept of paradigm change to the history of Christian thought. His paper cites five major historical paradigms: Greek Alexandrian, Latin Augustinian, Medieval Thomistic, Reformation, and Modern-Critical. Each paradigm provided a framework for normal work and cumulative growth (comparable to "normal science"), in which the scope of the paradigm was extended and major changes were resisted. As in the scientific case, Küng shows, each new paradigm arose in a period of crisis and uncertainty -- for example, the challenge of gnosticism in the Hellenistic world or the rise of science and biblical criticism in the modern period. In each case, conversion to the new paradigm involved subjective factors and personal decisions as well as rational argument. These paradigm shifts involved both continuity and discontinuity.32

Küng brings out some illuminating similarities, but he also notes some distinctive features of paradigm shifts in Christian thought. The centrality of the scriptural witness to Christ is without parallel in science. "The biblical message, ‘ not scripture itself, is the enduring norm. Each new paradigm arose from a fresh experience of the original message, as well as from institutional crises and external challenges. The gospel thus contributed to both continuity and change. Moreover, there is always a personal dimension to the decision of faith, along with the more intellectual task of showing that a new paradigm is both responsive to the Christian message and relevant to the present world of experience and contemporary knowledge. Küng says that we can acknowledge the distinctive features of religion and yet find the comparison with scientific paradigms helpful in understanding processes of change in the history of a religious tradition.

In another paper from this conference, Stephan Pfürtner shows that it is illuminating to consider Luther’s idea of justification by faith as a new paradigm. It led to the reconstruction of prior beliefs and the reinterpretation of previous data in a new framework of thought. Justification by faith affected almost all other doctrines. Concerning the doctrine of God, for example, Aquinas had combined revelation in Christ with Aristotelian philosophical categories to describe "God in himself" as actus purus. Luther relied on revelation in Christ and the experience of justification in speaking only of "God for us." By stressing the direct relation of each person to God, he also allowed greater scope for individual conscience. While he did not himself advocate religious liberty, he helped to set in motion historical forces leading in that direction justification was not a new idea, but by giving it a central position Luther developed new interpretations of law and gospel, church and state, and the priesthood of all believers. Pfürtner also shows how an acknowledgment of both continuity and discontinuity in paradigm changes can contribute to the Protestant-Catholic dialogue today.33

Many of the conference papers are devoted to the search for new paradigms today. Most of the participants accept Kühn’s thesis that paradigms are influenced by historical, social, and cultural factors, though some hold that the concept of paradigm is too vague to be useful. Many recognize that theology is in a time of crisis that calls for significant change. Factors mentioned as responsible for this crisis include secularization; religious pluralism; historical consciousness; the exploitation of women, races, and developing nations; the ambiguity of science and technology; the destruction of the environment; and the threat of nuclear war. Most participants insist that theology today requires the interpretation of the gospel within concrete historical and social contexts. Küng concludes that we should look for a plurality of Christian paradigms rather than expecting a consensus around any one.

This leads me to ask: How large a group is a paradigm community, and how does one determine its boundaries? When should one consider a historical change to be an evolutionary modification within a paradigm, and when should one consider it a revolutionary paradigm shift? Thomas Kühn’s earlier writing reserved the term scientific revolution for the rare instances when a sweeping change took place in a whole network of assumptions and concepts. Critics felt that he had drawn too sharp a line between normal science and revolutionary science, leaving out changes of intermediate scale. Kühn’s later writing referred to more modest "micro-revolutions" and said that a paradigm community could be as small as twenty-five persons in a subdiscipline. But the conceptual structure and assumptions of most subdisciplines are in large measure shared with other adjacent subdisciplines. If a small group has a really distinctive paradigm it will usually die out, or the ideas will be extended to other subdisciplines and perhaps to the discipline as a whole.

In religion, too, there are communities and sub-communities, and there are large and small historical changes. I suggest that the concept of paradigm shift is most helpful in understanding historical change if we use the term for relatively rare comprehensive conceptual changes. Clearly, the emergence of early Christianity from Judaism represents such a paradigm shift, for despite the continuities, people experienced far-reaching discontinuities in belief and practice. By the time of Paul’s letters, it was evident that Christianity could not be a sect within Judaism or a movement to reform Judaism, and individuals had to choose one paradigm community or the other, focusing on either Christ or the Torah. The discontinuities in the Protestant Reformation were perhaps not as radical, but major changes took place in doctrine and practice as well as in institutional organization.

Would it be illuminating to consider all of Christianity as one paradigm and refer to "the Christian paradigm"? One could then speak of a "paradigm shift" when an individual converted to another religious tradition (or atheism) and joined another paradigm community. The parallels with science would be stretched, for there seem to be few shared data or criteria common to diverse traditions, to which appeal could be made in giving reasons for choice among them. Should we seek such shared data and criteria in a global age, or can the assessment of beliefs be carried out only within a well-defined religious tradition? We will return to the problem of religious pluralism in the next chapter.

IV. Tentativeness and Commitment

In the popular stereotype, the scientist’s theories are tentative hypotheses that are continually criticized and revised, while religious beliefs are unchanging dogmas that the faithful accept without question. The scientist is seen as open-minded, the theologian as closed-minded. Is not faith a matter of unconditional commitment? Are not Christian beliefs attributed to divine revelation rather than human discovery? Have we perhaps lost sight of the distinctive features of religious faith by tracing some limited parallels with science?

1. Tradition and Criticism

Let us ask first how the scientific and religious communities each balance the importance of an ongoing tradition against the value of criticism and change. When major historical changes take place, does continuity or discontinuity predominate?

Whereas Popper identifies rationality and objectivity in science with adherence to explicit rules, Kuhn and Polanyi maintain that the locus of authority is the scientific community itself. Decisions rest on the informed judgment of the community. Shared values and criteria underlie such judgment, but the application and weighting of the criteria are not governed by logic or rules. Kuhn claims that authoritative tradition transmitted by the dominant paradigm provides the framework for thought and action in "normal science." This is a historical and social view of the process of inquiry in which the ongoing community is emphasized.34

As there is no private science, so also there is no private religion. In both cases, the initiate joins a particular community and adopts its modes of thought and action. Even the contemplative mystic is influenced by the tradition in which he or she has lived. Paradigms in religion, as in science, are acquired by example and practice, not by following formal rules. Individual insights are tested against the experience of others, as well as in one’s own life. Here, too, the historical and social context affects all modes of thought and action.

Kuhn pictures normal science as conservative and controlled by tradition. Working within the prevailing paradigm is an efficient way of solving the distinctive problems it raises. Exploring its potentialities and extending its range provide a focus for research. Within that tradition, a person benefits from the work of others, and there is cumulative progress. According to Kuhn, paradigm shifts are relatively rare and occur only when an accumulation of anomalies has produced a real crisis. One cannot speak of progress across the transitions; Kuhn describes paradigm changes in the political metaphor of revolution, which emphasizes discontinuity and the overthrow of the established order.

Kuhn’s critics reply that even in scientific revolutions the old data are preserved (though reinterpreted) and the new concepts and theories can be related to the old (though displacing them). Moreover, shared values and criteria of judgment persist across the change. Most scientists are familiar with other scientific disciplines and subfields, which provide continuity when their own area of specialization is in transition. A scientist has a higher loyalty to the wider scientific community and its values, which goes beyond loyalty to a particular paradigm. The critics urge us to view science as evolutionary and subject to continual reformation, rather than as bound by tradition except during revolutions. Nevertheless, historical studies have tended to support the view that theories are not evaluated separately but as part of networks of assumptions which sometimes change together rather radically.35

Normal theology does indeed show the dominance of tradition. The theologian is concerned to develop the potentialities of a particular paradigm. This provides focus and encourages communication and cumulation. But the process can include considerable reinterpretation, reformulation, and innovation. Scripture is unchanging, but ways of understanding and appropriating it have changed greatly, especially since the rise of historical-critical methods. Theology, we have said, is critical reflection on the life and thought of the religious community, and this implies the revisability of ideas. The Protestant Reformation was not a once-for-all revolution, but rather a vision of a church that is semper reformanda, always reforming. Cardinal Newman defended the development of ideas and the evolution of doctrine within the basic continuity of the Catholic tradition 36

Theological revolutions, such as the Protestant Reformation, or the emergence of Mahayana from Theravada Buddhism, do involve extensive and fundamental changes. Yet here, too, there are impressive continuities amid the discontinuities. There is a common loyalty to the founding leader, common scriptures, and a shared early history. In an ecumenical age, Catholic and Protestant thinkers read each others’ writings and affect each other, as do Buddhists of diverse schools. Feminist theologians criticize the gender biases of Christian thought and propose major reconstruction of traditional doctrines, yet in most cases they affirm a large portion of a common heritage. The theologian, however, does not seem to have a loyalty to an overarching and universal religious community, with shared criteria and values comparable to those shared by all scientists. In a global age, could such wider loyalties be encouraged, without undermining the distinctiveness of each religious tradition?

2. Central and Peripheral Beliefs

Popper maintains that scientific theories are held with great tentativeness and that basic assumptions should be continually questioned and criticized. Kuhn, by contrast, says that there is normally great tenacity in commitment to a prevailing paradigm, which is questioned only in rare times of crisis. Imre Lakatos proposes an intermediate position in which there is commitment to a "hard core" of central ideas that are preserved by making adjustments in a "protective belt" of more tentative auxiliary hypotheses. In place of competing individual theories (Popper) or successive paradigms (Kuhn), Lakatos pictures research programs, which sometimes compete over a protracted period of time. He does not accept the formal criteria for the acceptability of theories proposed by Popper, but he offers more definite and rational criteria than Kuhn acknowledges.

Lakatos maintains that a research program is constituted by a hard core of ideas that is deliberately exempted from falsification so that its positive potentialities can be systematically developed and explored. Anomalies are accommodated by changes in the auxiliary hypotheses, which can be sacrificed if necessary. This strategy calls for commitment in sticking with central ideas, without being distracted from them, as long as the program is "progressive" in predicting "novel facts" (which may refer to new phenomena or to already known facts that had previously been considered irrelevant). A program should be abandoned when it is stalled and not growing for a considerable period and when there is a promising alternative. The old program is not falsified but rather is displaced as a research strategy. However, a degenerating program can stage a comeback if it is reinvigorated by an imaginative new auxiliary hypothesis, as Lakatos shows in several examples. He believes his scheme describes the best scientific practice and prescribes how scientific programs should be evaluated, namely by comparing their progress as strategies for research over a period of time.37

We can apply Lakatos’s analysis to religious communities, which also make a central core of ideas immune to falsification and protect them by adjusting peripheral beliefs. Commitment to a core program allows it to be systematically explored without continual distraction. Rival programs may compete over long periods. The component beliefs are not verified or falsified separately in isolation; they are parts of an ongoing program that can be compared to other programs. Here progress is presumably not judged by the power to predict totally new phenomena, but by the ability to account for known data not previously considered. When anomalies arise -- from historical events, from new experience, or perhaps from new discoveries in science -- adjustment would be made in auxiliary hypotheses before core beliefs were abandoned.38

Ancient Israel held a central belief in the existence of a God of power and justice. An important but less central assumption was that God punishes wrongdoers. I suggest that we could see efforts to deal with the anomaly of undeserved suffering as attempts to preserve the central core by modifying auxiliary hypotheses. In the book of Job, the protagonist is told by his friends that he must have sinned in secret to deserve such suffering. But Job maintains both his innocence and the existence of God, at the cost of the hypothesis that all suffering is deserved.

Israel faced the same anomaly on a national scale in its long exile in Babylon. Some people saw the exile as God’s punishment for Israel’s failure to observe the Torah rigorously, and they counseled stricter observance. Others developed new ways of understanding God’s action in history, which allowed for undeserved suffering (including the vicarious suffering or suffering servant motif in Isaiah 53 and elsewhere). But even the latter "auxiliary hypothesis" is put in question by the magnitude of evil and suffering in the Nazi holocaust. For some people this historical event required reformulation of concepts of God’s power. For a few it led to abandoning theism itself. The holocaust is an anomaly that is only partly resolved within the traditional beliefs of both the Jewish and Christian communities.

Nancey Murphy proposes using Lakatos’s methodology in Christian theology. The primary data would be the practices of the Christian community, including its devotional experience and its use of scripture. The idea of a plurality of competing theological research programs can both illuminate past history and offer a possible pattern for current theological inquiry. As one example, Murphy traces three forms of the doctrine of atonement, in which Christ’s death is understood as a victory over the forces of evil or as a satisfaction of God’s justice or as a demonstration of God’s love. The first program was largely replaced by the other two historically, but it could be revived today with a new auxiliary hypothesis in which the forces of evil are reinterpreted in social and political terms. Murphy gives other examples in Catholic modernism, Swedish "motif-research," and Puritan and Anabaptist efforts to establish criteria for judging the authenticity of religious experience.39

How broad a set of ideas should be thought of as a theological program? An interpretation of a single doctrine, such as one view of the atonement, is perhaps too limited to consider as a "core belief" to which enduring commitment is given. Perhaps a school of Christian thought, such as neo-orthodoxy, Thomism, or process theology, can fruitfully be portrayed as a program. Alternatively, in the context of religious pluralism, one might think of Christianity as a program whose core is belief in a personal God and the centrality of Jesus Christ -- with all other beliefs as auxiliary hypotheses that can be modified to maintain that core. Gary Gutting goes even further in proposing that belief in the existence of a personal God constitutes the Lakatos core to which decisive assent should be given, but this seems to me too broad to define an identifiable religious community. 40

I will suggest in chapter 9 that process theology can be viewed as a theological program in which the "hard core" of the Christian tradition is taken to be belief in God as creative love, revealed in Christ, while divine omnipotence is treated as an "auxiliary hypothesis" that can be modified to allow for the data of human freedom, evil and suffering, and evolutionary history.

We must keep in mind, however, that a program of the type Lakatos proposes can apply to certain theological tasks but not to others. Phillip Clayton has considered Lakatos in relation to four tasks of theology. (1) Historical Criticism. Here the theologian uses the historical disciplines to ask what actually happened in the events narrated in the stories of the tradition. (2) Philosophical Reflection. Here the criteria, according to Clayton, are conceptual coherence, comprehensiveness, and adequacy to the data of human experience, including moral and mystical experience. (3) Interpretation of Texts. Clayton holds that this occurs in "programs of interpretation," which are more like the work of the literary critic than that of the scientist. There are no predictions or falsifications, and judgments are based on aesthetic criteria, personal meaningfulness, and the reader’s response. (4) Programs for Living. Religious traditions offer guidance for living, with ethical and affective as well as cognitive dimensions, and their adequacy in these areas must be elucidated. Clayton concludes that Lakatos’s methodology might be extended to the first two tasks, though it is difficult to identify programs that would fulfill the criteria for "progressiveness." But he suggests that the method is inapplicable to the last two tasks.41

Lakatos’s programs, then, are very similar to Kuhn’s paradigms, but they offer two advantages as ways of analyzing both science and religion. First, they allow one to distinguish between the central core to which a group is committed and the peripheral beliefs that are more readily modified or abandoned -- though Lakatos recognizes that the distinction is not absolute and can change historically. Second, rival programs can coexist during protracted periods, allowing for greater pluralism. We are to look at the fruitfulness of a program in a community over a period of time, rather than evaluating a fixed set of ideas at any moment in abstraction from the ongoing life of the community.

3. Revelation, Faith, and Reason

Even if peripheral beliefs are tentative and revisable, are not the core beliefs of a religious community held with absolute and unconditional commitment? Job may have given up the idea that suffering is always deserved, but his basic faith in God was unshaken. No evidence could count against it: "Though he slay me, yet will I trust in him" Job 13:15 KJV). St. Paul was confident that "neither death nor life nor anything else in all creation will be able to separate us from the love of God which is in Christ Jesus our Lord" (Rom. 8:39). In chapter 1 we noted the existentialist thesis that faith is a matter of passionate personal commitment and decision, far removed from the dispassionate weighing of hypotheses. We also referred to the neo-orthodox theme that faith’s confidence rests on revelation, which was the result of divine initiative rather than of human discovery. Can our account do justice to the importance of faith and revelation in the Christian tradition?

Basil Mitchell contrasts the tentative hypotheses of science with unconditional commitment in religion. But he goes on to qualify the contrast from both sides. He describes the tenacity of a scientist’s commitment to a Kuhnian paradigm. He also insists that ultimate religious commitment is to God and not to Christianity or any other system of belief. And here the cumulative weight of evidence is decisive. All religious ideas are open to revision, according to Mitchell. There must be grounds for accepting a claim of divine revelation in history, even if revelation shows us possibilities that we could not have anticipated. Mitchell says that knowledge of God in religious experience is also not self-authenticating, for there is no uninterpreted experience, and any particular interpretation involves claims that must be judged more plausible than the alternatives. There is thus a continuing dialectic between commitment and reflection, or between faith and reason.42

In the biblical view, faith is personal trust, confidence, and loyalty. Like faith in a friend or faith in a doctor, it is not "blind faith," for it is closely tied to experience. But it does entail risk and vulnerability in the absence of logical proof. If faith were the acceptance of revealed propositions it would be incompatible with doubt. But if faith means trust and loyalty, it is compatible with considerable doubt about particular beliefs. Doubt frees us from illusions of having captured God in a creed. It calls into question every religious symbol. Self-criticism is called for if we acknowledge that no church, book, or creed is infallible and no formulation is irrevocable. The claim to finality by any historical institution or theological system must be questioned if we are to avoid absolutizing the relative.

James Fowler has identified a series of stages of faith on the basis of extensive interviews with hundreds of people of all ages. Paralleling the work of Jean Piaget on cognitive development, Erik Erikson on stages of life, and Lawrence Kohlberg on moral development, Fowler describes six stages of faith: (1) An Intuitive-Projective stage takes place in early childhood, characterized by imagination and dependence on parents. (2) A Mythic-Literal stage follows in later childhood, during which myths are interpreted literally and other adults are significant. (3) In the Synthetic-Conventional stage of adolescence, beliefs are formulated in conformity to peers. Some individuals remain at this stage, dependent on external authority. (4) In the Individuative-Reflective stage, persons question, doubt, and assume responsibility for their own commitments. The locus of authority is internal, and they have a stronger sense of individual identity. (5) In the Conjunctive stage of mature faith, persons integrate tradition and doubt, recognizing the symbolic character of religious language. They show respect for other traditions along with commitment to their own tradition. (6) In the Universalizing stage, reached only by rare individuals, persons exhibit a greater inclusiveness and a more radical living out of convictions. Here there is a greater depth of religious experience, a vision of a transformed world, and a love that reaches out to others.43

Fowler’s ordering of stages of faith is in part empirical. He finds, for example, that in these life histories there is a direction of development, and that the stages rarely occur in the reverse order. But the ordering also reflects value judgments as to which are "higher" or more "mature." I agree with Fowler’s theological assumptions concerning the nature of revelation and religious authority, but I do not think these assumptions can be derived from his data. The fourth stage (Individuative-Reflective), and the fifth (Conjunctive) are clearly more consistent with the goals I have been presenting than is the conventional third stage. The Universalizing sixth stage represents an ideal for us to seek, even if few people attain it.

Religious faith does demand a more total personal involvement than occurs in science, as the existentialists maintain. Religious questions are of ultimate concern, since the meaning of one’s existence is at stake. Religion asks about the final objects of a person’s devotion and loyalty. Too detached an attitude may cut a person off from the very kinds of experience that are religiously most significant. But such religious commitment can be combined with critical reflection. Commitment without inquiry tends toward fanaticism or narrow dogmatism. Reflection alone without commitment tends to become trivial speculation unrelated to real life. Perhaps personal involvement must alternate with reflection, since worship and critical inquiry do not occur simultaneously.

Divine revelation and human response are always inextricably interwoven. Revelation is incomplete until it has been received by individuals, and individuals always live within interpretive communities. The God-given encounter was experienced, interpreted, and reported by fallible human beings. In the history of Israel, crucial events were revelatory only when interpreted in the light of the prophet’s experience of God. God acts in the lives of individuals and communities, especially in the life of Christ, we have said, but the records of these events reflect particular personal and cultural perspectives. There is no uninterpreted revelation.

Moreover, revelation is recognized by its ability to illuminate present experience. Revelation helps us to understand our lives as individuals and as a community today.44 Special events in the past enable us to see what is present at other times but may have been ignored. The cross reveals God’s universal love, everywhere expressed but not everywhere acknowledged. The power of reconciliation in Christ’s life is the power of reconciliation in all life.45 Revelation leads to a new relation to God in the present; thus it is inseparable from reorientation and reconciliation. It is not a system of divine propositions completed in the past but an invitation to new experience of God today. So revelation and experience, like faith and reason, are not mutually exclusive.

To sum up, there are many parallels between science and religion: the interaction of data and theory (or experience and interpretation); the historical character of the interpretive community; the use of models; and the influence of paradigms or programs. In both fields there are no proofs, but there can be good reasons for the judgments rendered by the paradigm community. There are also important differences between science and religion, but some of them turn out to be differences in emphasis or degree rather than the absolute contrasts sometimes imagined. We have traced a number of polarities in which the first term was more prominent in science and the second in religion, but both were found to be present in both fields: objectivity and subjectivity; rationality and personal judgment; universality and historical conditioning; criticism and tradition; and tentativeness and commitment. But some features of religion seem to be without parallel in science: the role of story and ritual; the noncognitive functions of religious models in evoking attitudes and encouraging personal transformation; the type of personal involvement characteristic of religious faith; and the idea of revelation in historical events. Some additional comparisons are explored in the next chapter before we draw overall conclusions.



1. Several sections of this chapter are revisions or summaries of portions of two earlier books: Ian G. Barbour, Issues in Science and Religion (Englewood Cliffs, NJ: Prentice-Hall, 1966) and Myths, Models, and Paradigms (New York: Harper & Row, 1974). The original passages are identified in footnotes.

2. Carl G. Hempel, Philosophy of Natural Science (Englewood Cliffs, NJ: Prentice-Hall, 1966); Karl R. Popper, The Logic of Scientific Discovery (London: Hutchinson’s University Library, 1956).

3. W. V. Quine, "Two Dogmas of Empiricism," in his From a Logical Point of View, 2d ed. (New York: Harper Torchbooks, 1963).

4. N. R. Hanson, Patterns of Discovery (Cambridge: Cambridge University Press, 1958); Michael Polanyi, Personal Knowledge. (Chicago: University of Chicago Press, 1958).

5. Thomas S. Kuhn, The Structure of Scientific Revolutions, 2d ed. (Chicago: University of Chicago Press, 1970).

6. See, for example, Frederick J. Streng, Understanding Religious Life; Ninian Smart, Worldviews (New York: Charles Scribner’s Sons, 1983).

7. Mircea Eliade, The Sacred and the Profane, trans. W. Trask (New York: Harcourt, Brace & World, 1959).

8. Bhagavad Gita, trans. Swami Prabhavananda and Christopher Isherwood (New York: New American Library, 1972); David Kinsley, Hinduism (Englewood Cliffs, NJ: Prentice-Hall, 1982).

9, In Myths, Models, and Paradigms, chap. 3, 1 discussed writings on scientific models by Mary Hesse, Max Black, Richard Braithwaite, Peter Achinstein, and others. See also W. H. Leatherdale, The Role of Analogy, Model and Metaphor in Science (New York: American Elsevier, 1974).

10. Niels Bohr, Atomic Theory and the Description of Nature (Cambridge: Cambridge University Press, 1934), p. 96.

11. See Barbour, Issues in Science and Religion, pp. 162 -74; also Myths, Models, and Paradigms, pp. 34-38.

12. Larry Laudan, "A Confutation of Convergent Realism," in Scientific Realism, ed. Jarret Leplin (Berkeley and Los Angeles: University of California Press, 1984).

13. Laudan, "Convergent Realism"; W. H. Newton-Smith, The Rationality of Science (London: Routledge & Kegan Paul, 1981); Ian Hacking, Representing and Intervening (Cambridge: Cambridge University Press, 1983); Michael Devitt, Realism and Truth (Princeton: Princeton University Press, 1984); James T. Cushing, C. F. Delaney, and Gary Gutting, eds., Science and Reality (Notre Dame: University of Notre Dame Press, 1984); Ron Harré, Varieties of Realism (Oxford: Basil Blackwell, 1986); and Hilary Putnam, The Many Faces of Realism (LaSalle, IL: Open Court, 1987).

14. Ernan McMullin, "A Case for Scientific Realism," in Scientific Realism, ed. Leplin, p. 39.

15. In Myths, Models, and Paradigms, chap. 4, I discuss the writings of Ian Ramsey and Frederick Ferré on models in religion, and I develop a theory of religious models. There is some discussion of models in Earl MacCormac, Metaphor and Myth in Science and Religion (Durham, NC: Duke University Press, 1976).

16. Richard Braithwaite, An Empiricist’s View of the Nature of Religious Belief (Cambridge: Cambridge University Press, 1955); see William H. Austin, The Relevance of Natural Science to Theology (London: Macmillan, 1976), chap. 3.

17. Janet Soskice, Metaphor and Religious Language (Oxford: Clarendon Press, 1985).

18. Frank Brown, "Transfiguration: Poetic Metaphor and Theological Reflection," Journal of Religion 62 (1982): 39- 56; also his Transfiguration: Poetic Metaphor and the Language of Religious Belief (Chapel Hill: University of North Carolina Press, 1983).

19. Barbour, Myths, Models, and Paradigms, pp. 56-60.

20. Ninian Smart, The Concept of Worship (London: Macmillan, 1972) and Worldviews, chap. 3.

21. Winston King, Introduction to Religion: A Phenomenological Approach (New York: Harper & Row, 1968), p. 165.

22. Ninian Smart, Reasons and Faiths (London: Routledge & Kegan Paul, 1958).

23. Sallie McFague, Metaphorical Theology: Models of God in Religious Language (Philadelphia: Fortress Press, 1982).

24. Sallie McFague, Models of God. Theology for an Ecological, Nuclear Age (Philadelphia: Fortress Press, 1987).

25. Thomas Kuhn, Structure of Scientific Revolutions, p. 147.

26. See Barbour, Myths, Models, and Paradigms, chap. 6.

27. See also Polanyi, Personal Knowledge.

28. Harold Brown, Perception, Theory and Commitment: The New Philosophy of Science (Chicago: University of Chicago Press, 1977).

29. Ibid., p. 167.

30. Ibid.

31. Frederick Streng, "Lens and Insight: Paradigm Changes and Different Kinds of Religious Consciousness"(Plenary address to Second Conference on East-West Religions in Encounter, "Paradigm Shifts in Buddhism and Christianity," Hawaii Loa College, Oahu, Hawaii, Jan. 4,1984).

32. Hans Küng, "Paradigm Change in Theology," in Paradigm Change in Theology, eds. Hans Küng and David Tracy (Edinburgh: T. & T. Clark, 1989).

33. Stephan Pfürtner, "The Paradigms of Thomas Aquinas and Martin Luther: Did Luther’s Message of Justification Mean a Paradigm Shift?" in Paradigm Change in Theology, eds. Küng and Tracy.

34. Kuhn, Structure of Scientific Revolutions; Polanyi, Personal Knowledge; W. D. King, "Reason, Tradition, and the Progressiveness of Science," in Paradigms and Revolutions, ed. Gary Gutting (Notre Dame: University of Notre Dame Press, 1980).

35. Mark Blaug, "Kuhn versus Lakatos, or Paradigms versus Research Programs in the History of Economics," in Paradigms and Revolutions, ed. Gutting.

36. Richard Vernon, "Politics as Metaphor: Cardinal Newman and Professor Kuhn," in Paradigms and Revolutions, ed. Gutting.

37. Imre Lakatos, "Falsification and the Methodology of Scientific Research Programmes," in Criticism and the Growth of Knowledge, eds. I. Lakatos and A. Musgrave (Cambridge: Cambridge University Press, 1970). Also Lakatos, Philosophical Papers, vol. 1, eds. John Worall and Gregory Currie (Cambridge: Cambridge University Press, 1978).

38. See William Austin, "Religious Commitment and the Logical Status of Doctrines," Religious Studies 9 (1973): 39-48.

39. Nancey Murphy, "Revisionist Philosophy of Science and Theological Method" (Paper delivered at the Pacific Coast Theological Society, Spring 1983); Theology in the Age of Probable Reasoning (Ithaca, NY: Cornell University Press, 1990). "Acceptability Criteria for Work in Theology and Science," Zygon 22 (1987): 279-97.

40. Gary Gutting, Religious Belief and Religious Skepticism (Notre Dame: University of Notre Dame Press, 1982), chap. 5.

41. Phillip Clayton, paper delivered to American Academy of Religion, Nov. 1988. See also his Explanation from Physics to Theology (New Haven: Yale University Press, 1989).

42. Basil Mitchell, The Justification of Religious Belief (London: Macmillan, 1973), chaps. 5-8.

43. James Fowler, Stages of Faith (San Francisco: Harper & Row, 1981).

44. H. Richard Niebuhr, The Meaning of Revelation (New York: Macmillan, 1941).

45. Paul Tillich, Systematic Theology (Chicago: University of Chicago Press, 1957), 2:165-68.