For Further Reading

Chapter 3

Religion and the Philosophy of Science

The discipline of the philosophy of science deals, in very general terms, with the philosophical issues associated with or arising from the natural sciences. Examples of these issues include: What is a law of nature? What kind of data can be used to distinguish between real causes and accidental regularities? What kind of evidence do we need before we accept hypotheses? Why do scientists use models and theories which they know are at least partially inaccurate and open to revision?

Some of those issues overlap with the traditional themes of philosophy, raising the question of the extent to which the philosophy of science can be considered to be a distinct discipline in its own right. This point can be appreciated by considering the ‘laws of nature’, which attempt to represent the regularity or ordering which appears to exist within nature. Is this ‘regularity’ really present within nature itself? Or is it something that is imposed upon nature by the human mind? Is this ‘lawlikeness’ discerned within nature, or projected upon nature? This debate about the tendency of the human mind to generate and impose patterns upon observation, which was given particular stimulus during the late eighteenth century by the Scottish philosopher David Hume, is of general philosophical interest; yet it clearly has particular significance for the natural sciences.

Other philosophical discussions have a more specific relation to the natural sciences. For example, suppose a certain experiment is carried out, which suggests that a type of particle exists. This particle cannot itself be observed, but its existence seems to be implied by the behaviour of other aspects of the system. So what is the status of this hypothetical and unobserved particle? Can it really be said to ‘exist’? For some writers, the only things that ‘really exist’ are the experimental observations themselves. The theoretical particle is just a ‘useful fiction’, a helpful way of explaining the phenomena. Others, however, hold that the best explanation of why scientific theories work is that they represent an account or depiction of the way things are – in other words, that some form of realism represents the best explanation of the successes of the natural sciences.

Many questions in the philosophy of science are relevant to religious or theological reflection, as well as offering a conceptual and methodological bridge between science and theology. There are obvious parallels between scientific and theological discussions at multiple points, such as what it means to explain something, whether reality is a free construction of the human mind, and what criteria might be used to evaluate possible accounts or explanations of reality.

The present chapter aims to deal with some of the leading themes in the philosophy of science and explore how they have a wider significance within reflection on religious issues. will then explore the ways in which the philosophy of religion has drawn on insights from the natural sciences.

We begin our discussion by considering the question of whether scientific theories offer us depictions of the nature of the universe, or whether they are simply useful fictions to help us predict what happens within the universe.

Fact and Fiction: Realism and Instrumentalism

The term ‘realism’ is generally used to refer to a group of philosophies which affirm that an external reality exists, and that the human mind is somehow able to copy or represent this. So why has realism been so influential in the natural sciences? For many, the best explanation of the success of the natural sciences is the belief that scientific theories offer real descriptions of the world. ‘If scientific realism, and the theories it draws on, were not correct, there would be no explanation of why the observed world is as if they were correct; that fact would be brute, if not miraculous’ (Michael Devitt).

On this approach, the simplest and most persuasive explanation of what makes theories work is that they offer an account of the way things really are. If the theoretical claims of the natural sciences were not correct, their massive empirical success would appear to be totally coincidental. As the physicist and theologian John Polkinghorne remarked:

The naturally convincing explanation of the success of science is that it is gaining a tightening grasp of an actual reality. The true goal of scientific endeavour is understanding the structure of the physical world, an understanding which is never complete but ever capable of further improvement. The terms of that understanding are dictated by the way things are.

For reasons such as these, natural scientists tend to be realists, at least in the broad sense of that term. It seems to many that the success of the natural sciences show that they have somehow managed to uncover the way things really are, or to lock into something which is fundamental to the structure of the universe. The importance of this point is considerable, not least in that it raises the question of whether theologians wishing to argue for the independent existence of God (rather than as a construct of the human mind) may learn anything from the forms of realism associated with the natural sciences. This present section aims to explore this issue, beginning with an examination of the nature of realism itself, before moving on to consider its alternatives.

Realism

The term ‘realism’ denotes a family of philosophical positions which take the general view that a real world exists, external to the human mind, which the human mind can encounter, understand, and represent, if only in part. The credibility of realism arises directly from the successes of the experimental method, which disclose patterns of observational behaviour which seem to be best accounted for on the basis of a realist point of view. As the philosopher of science Michael Redhead notes:

Physicists, in their unreflective and intuitive attitude to their work, the way they talk and think among themselves, tend to be realists about the entities they deal with, and while being tentative as to what they say about these entities and their exact properties and interrelations, they generally feel that what they are trying to do, and to some degree successfully, is to get a ‘handle on reality’.

Scientific realism is thus, at least in part, an empirical thesis. Its plausibility and confirmation arise from direct engagement with the real world, through repeated observation and experiment. It should not be thought of primarily as a metaphysical claim about how the world is, or ought to be. Rather, it is a focused and limited claim which attempts to explain why it is that certain scientific methods have worked out so well in practice.

Realism, as the philosopher Hilary Putnam and others have argued, is the only explanation of scientific theories and concepts which does not ‘make the success of science a miracle’. Unless the theoretical entities employed by scientific theories actually exist and the theories themselves are at least approximately true accounts of the world at large, the evident success of science (in terms of its applications and predictions) would surely be a miracle. The argument for realism based upon scientific success can be set out like this:

Realism, as noted earlier, refers to a family of philosophies. One form of realism which has received especial attention in the dialogue between science and religion is generally known as ‘critical realism’. A distinction is often drawn between a ‘naïve realism’, which holds that reality impacts directly upon the human mind, without any reflection on the part of the human knower, and a ‘critical realism’, which recognizes that the human mind attempts to express and accommodate that reality as best it can with the tools at its disposal – such as mathematical formulae or mental models. Both of these forms of realism can be contrasted with various forms of non‐realism or anti‐realism, which hold that the human mind freely constructs its ideas without any reference to an alleged external world.

The key feature of a ‘critical realism’ is its recognition that the human mind is active in the process of perception. Far from being a passive recipient of knowledge of the external world, it actively constructs this knowledge using ‘mental maps’, often known as schema. This point was made by the psychologist of religion William James (1842–1910) in 1878, and has been widely accepted since.

The knower is an actor, and co‐efficient of the truth on the one side, whilst on the other he registers the truth which he helps to create. Mental interests, hypotheses, postulates, so far as they are bases for human action – action which to a great extent transforms the world – help to make the truth which they declare.

More recently, the New Testament scholar N. T. Wright (born 1948) offers a helpful account of this approach, which he describes as:

… a way of describing the process of ‘knowing’ that acknowledges the reality of the thing known, as something other than the knower (hence ‘realism’), while also fully acknowledging that the only access we have to this reality lies along the spiralling path of appropriate dialogue or conversation between the knower and the thing known (hence ‘critical’).

This insight does not pose a challenge to the notion that there exists a world, independent of the observer. It is to acknowledge that the knower is involved in the process of knowing, and that this involvement must somehow be expressed within a realist perspective on the world.

But what of alternatives to realism? The two most significant are often regarded as idealism and instrumentalism, both of which we shall now consider.

Idealism

Idealism is an approach to our knowledge of the world which concedes that physical objects exist in the world, while holding that we can have knowledge only of how things appear to us, or are experienced by us, not things as they are in themselves. The most familiar version of this approach is that associated with the great German idealist philosopher Immanuel Kant (1724–1804), who argued that we have to deal with appearances or representations, rather than things in themselves. Kant thus draws a distinction between the world of observation (the ‘phenomena’) and ‘things in themselves’, holding that the latter can never be known directly. The idealist will thus hold that we can have knowledge of the manner in which things appear to us through the ordering activity of the human mind. We cannot, however, have knowledge of mind‐independent realities.

This view is expressed particularly forcefully in the approach often known as ‘phenomenalism’. This holds that we cannot know extra‐mental realities directly, but only through their ‘appearances’ or ‘representations’. Although this view is relatively uncommon within the natural sciences, it has been defended by a number of significant figures, including the noted physicist Ernst Mach. For Mach, the natural sciences concern what is immediately given by the senses. Science concerns only the investigation of the apparent ‘dependence of phenomena on one another’. This led Mach to take a strongly negative view of the atomic hypothesis, arguing that atoms were merely useful fictions or theoretical constructs which cannot be observed. Atoms were thus not ‘real’; they were simply useful fictitious notions which helped observers to understand the relationship between various observed phenomena. The central concern is to ‘preserve the phenomena’ – a phrase first used by Aristotle – which emphasizes the priority of experimental observation over theoretical reflection. In the end, a theory will be judged by the extent to which it is able to accommodate existing observations, whether or not it is able to predict new and unexpected additional observations.

It is helpful to note here that two of the most important scientific theories developed in the last two centuries – Charles Darwin's theory of evolution by natural selection and Albert Einstein's theory of general relativity – were both highly successful in explaining known observations. While Darwin was clear that his theory did not, and could not, predict new observations, Einstein identified new observations that were to be expected, if his theory was correct. These included the phenomenon of gravitational lensing, in which the warping of space‐time due to the gravitational influence of the sun causes light to bend to an extent greater than that predicted by Newtonian mechanics.

Mach, however, apparently influenced by a Kantian philosophical framework, argued that it was impossible to move from the world of phenomena to the world of ‘things in themselves’. We therefore cannot move beyond the world of experience. Nevertheless, Mach was prepared to allow the use of ‘auxiliary concepts’ which serve as bridges linking one observation with another, provided that it is understood that they have no real existence. They are ‘products of thought’ which ‘exist only in our imagination and understanding’.

The point at issue in Mach's discussion is of considerable importance and is often discussed in terms of the technical phrases ‘hypothetical entities’, ‘theoretical terms’, or ‘unobservables’. The basic issue is whether something has to be seen before it can be held to exist. Mach, who argued that the natural sciences were concerned only with reporting experimental observations, held that science was not committed to defending the real and independent existence of ‘unobserved’ or ‘theoretical’ entities which those observations might suggest – such as atoms.

Mach's views were challenged, and eventually refuted, by Albert Einstein. In one of his remarkable scientific papers of 1905, Einstein offered an explanation of the puzzling phenomenon known as ‘Brownian motion’ – the observation that very small particles of matter, when suspended in a liquid, do not remain stationary but move around randomly. Einstein took the view that these particles moved around in irregular patterns because of the molecular movement of the liquid in which they are suspended. Einstein's theoretical analysis made it clear that, while it might not be possible to see atoms or molecules, their real existence could nevertheless be inferred from the properties of particles suspended in liquid. There are important parallels here with Isaac Newton's proposal of gravity. For Newton, gravity was a legitimate scientific inference from an observable phenomenon to the unobservable entity that best explains it. Similarly, atoms were not observed; their existence was inferred. Yet that process of inference was robust and generated further hypotheses that were open to empirical confirmation.

Instrumentalism

Instrumentalism holds that scientific concepts and theories are merely useful instruments whose worth is measured not by whether the concepts and theories are true or false, or how correctly they depict reality, but by how effective they are in correlating and predicting phenomena. They are not true descriptions of an unobservable reality, but merely useful ways of organizing observations. A scientific theory is best understood as a rule, principle, or calculating device for deriving predictions from sets of observational data.

The distinguishing features of instrumentalism can be studied from Ernest Nagel's comments on the kinetic model of gases. This model proposes that the molecules of a gas can be thought of as analogous to inelastic spherical objects, such as billiard balls. Nagel argues that this approach is nothing more than a useful instrument for making sense of observations.

The theory that a gas is a system of rapidly moving molecules is not a description of anything that has been or can be observed. The theory is rather a rule which prescribes a way of symbolically representing, for certain purposes, such matters as the observable pressure and temperature of a gas; and the theory shows among things how, when certain empirical data about a gas are supplied and incorporated into that representation, we can calculate the quantity of heat required for raising the temperature of the gas by some designated number of degrees (i.e., we can calculate the specific heat of the gas).

Scientific concepts, while being clearly grounded in observations of the natural world, are thus not to be identified with, or reduced to, those observations. Similarly, Stephen Toulmin argues that, instead of speaking about the ‘existence’ or ‘reality’ of such entities as electrons, scientists should recognize that such language is not used to refer to a real entity. The issue has to do with how observations are organized, with a view to stimulating further research, or predicting the behaviour of systems in the future.

More recently, the philosopher of science Bas van Fraassen developed ‘constructive empiricism’, which incorporates some instrumentalist themes. He draws a distinction between a realist, who holds that science aims to give a literally true description of what the world is like, and what he calls a ‘constructive empiricist’, who argues that acceptance of a theory does not involve commitment to the truth of that theory, but to the belief that it adequately preserves the phenomena to which it relates:

To be an empiricist is to withhold belief in anything that goes beyond the actual, observable phenomena, and to recognize no objective modality in nature. To develop an empiricist account of science is to depict it as involving a search for truth only about the empirical world, about what is actual and observable … it must invoke throughout a resolute rejection of the demand for an explanation of the regularities in the observable course of nature, by means of truths concerning a reality beyond what is actual and observable, as a demand which plays no role in the scientific enterprise.

To speak of ‘laws of nature’ or theoretical entities such as electrons is thus to introduce an unwarranted and unnecessary metaphysical element into scientific discourse.

Yet historically, most instrumentalist understandings of science have often transmuted into realist understandings with the passing of time. The Copernican (and subsequently Keplerian) theory of the solar system is a case in point. Initially, many scientists and non‐scientists interpreted the Copernican heliocentric theory instrumentally as a calculating device, believing that there were too many problems with Copernicus’ approach to allow it to be seen as ‘real’. Andreas Osiander, in his famous preface to Copernicus’ On the Revolution of Heavenly Bodies (1543), suggested that this theory was a fruitful hypothesis which was useful for astronomical calculations, but did not necessarily correspond to the way things were. Copernicus offered a useful mathematical model which was consistent with observations. Yet although the theory ‘saved the phenomena’, it did not necessarily commit its readers to a heliocentric model of the solar system.

It is the duty of an astronomer to establish the history of the celestial movements through careful and skilful observation, and then to conceive and devise causes of these motions or hypotheses about them. Now since he cannot in any way attain to the true causes, these assumed hypotheses enable those movements to be calculated correctly from the principles of geometry, both for the future and the past. The present author has performed both these duties excellently. For these hypotheses do not need be true or even probable. It is sufficient if they only provide a calculus consistent with these observations.

Yet with growing observational evidence for the heliocentric model of the solar system, an instrumentalist approach subtly changed into its realist counterpart. With the development of Galilean and Newtonian physics and the new observational data that became available through the invention of the telescope, the heliocentric theory began to be interpreted realistically, rather than instrumentally. It was not merely a convenient way of thinking about the solar system, or a convention that enabled certain useful mathematic calculations to be performed. This was the way things were. The solar system really was heliocentric.

Theology and Debates about Realism

So what is the relevance of these debates for theology? Perhaps the most important point to note is that each of these positions in the philosophy of science has its theological counterpart. Non‐realism is well represented by the radical philosopher of religion Don Cupitt, who argues we must ‘abandon ideas of objective and eternal truth, and instead see all truth as a human improvisation.’ Instead of responding to reality, we create whatever we choose to regard as real. Reality is something which we construct, not something to which we respond. ‘We constructed all the world‐views, we made all the theories …. They depend on us, not we on them.’

On the whole, however, theologians who have engaged with the natural sciences tend to be persuaded of the merits of more realist approaches to theology. For example, the Scottish theologian Thomas F. Torrance developed a rigorous form of theological realism, insisting that theology gives an account of the reality of things. Ian Barbour, Arthur Peacocke, and John Polkinghorne all adopt forms of critical realism, ultimately based on William James's emphasis on the active role of the knower in the process of knowing.

Polkinghorne set out his understanding of ‘critical realism’ in some detail in his 1996 Terry Lectures at Yale University, clarifying why he was not simply a realist in general, but a critical realist in particular:

I believe that the advance of science is not just concerned with our ability to manipulate the physical world, but to gain knowledge of its actual nature. In a word, I am a realist. Of course, such knowledge is to a degree partial and corrigible. Our attainment is verisimilitude, not absolute truth. Our method is the creative interpretation of experience, not rigorous deduction from it. Thus I am a critical realist.

The important point to note here is Polkinghorne's recognition that the scientific enterprise involves the active interpretation of our world, not merely passive observation. Polkinghorne thus highlights the importance of the realization that ‘theory and practice are inexplicably intertwined in scientific thought’, as a result of which scientific ‘facts’ are to be understood as having already been interpreted, whether consciously or not. ‘There is an inescapable self‐sustaining circularity in the mutual relationship of theory and experiment.’

Alister McGrath (born 1953) develops a somewhat different form of critical realism, drawing on the insights of the social scientist Roy Bhaskar concerning the stratification of reality, which emphasized that all intellectual disciplines or sciences are under an intrinsic obligation to give an account of reality according to its distinct nature. In his Territories of Human Reason, McGrath explores how diverse scientific and other intellectual disciplines – such as theology – develop their own distinct research methods, developed with their specific object of research in mind.

This point is so important that it needs further consideration. Is there a single ‘scientific method’? Or do individual natural sciences develop their own distinct approaches, based on their specific fields of investigation? And if so, what are the implications for theology?

Explanation, Ontology, and Epistemology: Research Methods and the Investigation of Reality

Does every intellectual discipline use essentially the same method of investigating reality? Or are these methods developed and adapted to cope with the specific areas of investigation? Some writers speak of the ‘scientific method’ – note the use of the singular – so that the natural sciences are all characterized by essentially the same working methods. This approach is found in the writings of the Oxford physicist and scientific populariser Peter Atkins, who argues that the distinctive ‘scientific method’ is capable of illuminating everything in a uniquely reliable manner. Yet his approach fails to take account of the distinct characteristics and objectives of individual sciences, in effect reducing them all to a single ‘mono‐science’ which overlooks their distinct identities, histories, and objects of inquiry. This view of a unique ‘scientific method’ has long been undermined by scholarly studies of the history and practice of the natural sciences, which point to a wide range of methods being developed and deployed, depending on the specific object of investigation. While astronomy, biochemistry, and psychology are all natural sciences, they nevertheless use different research tools in carrying out these investigations.

Earlier in this work, we noted a remark of Werner Heisenberg: ‘We need to remember that what we observe is not nature itself, but nature as it is disclosed by our methods of investigation.’ There is no generalized scientific method which can be applied to all sciences without some modification. While certain general principles may be argued to lie behind the specific approaches found in any given natural science, the nature of the field to be investigated shapes the approach to be adopted. In that each science deals with a different object, it is under an obligation to respond to that object according to its distinctive nature. The methods which are appropriate to the study of one object cannot be abstracted and applied uncritically and universally. Each science develops procedures which it believes to be appropriate to the nature of its own particular field of research. Yet each research method engages and illuminates only part of a bigger picture. Its results may be reliable and precise – but they are nevertheless incomplete.

The suggestion that the natural sciences themselves adopt a plurality of methods and criteria of rationality finds ample support in scientific practice. The biologist Steven Rose, reflecting on the scientific task of engaging and explaining the world within his own field, noted that a plurality of methods was required to engage our world. ‘As a materialist, as all biologists must be, I am committed to the view that we live in a world that is an ontological unity, but I must also accept an epistemological pluralism.’ We cannot reduce all cognitive activity to ‘a single fundamental method’, but must rather make use of a range of conceptual tool‐boxes, adapted to specific tasks and situations, to give as complete an account as possible of our world.

Rose explains his point with a parable of five biologists, representing different subdivisions of that discipline, who notice a frog jump into a pond. Each offers an explanation of this observation from the specific perspective of their own biological subdiscipline. The physiologist explains that the frog's leg muscles were stimulated by impulses from its brain. The biochemist supplements this by pointing out that the frog jumps because of the properties of fibrous proteins. The developmental biologist locates the frog's capacity to jump in the first place in the biological process that gave rise to its nervous system and muscles. The animal behaviourist points out that the frog jumped to escape from a lurking predatory snake. The evolutionary biologist adds that the process of natural selection ensures that only those ancestors of frogs which could detect and evade snakes would be able to survive and breed. Rose's point is simple: all five explanations are part of a bigger picture. All of them are right; they use different research methods to illuminate aspects of a greater whole, which none of them can disclose completely on the basis of its own methods.

Rose's parable helps us to identify the issues that need to be considered in moving from the recognition of multiple perspectives to the development of a unified theoretical account. Each of the five approaches can be treated as offering a specific perspective on the frog's jump. These perspectives reflect their own distinct disciplinary methods and emphases, and do not need to be treated as ‘fictions’, or even as instrumentalist accounts of the phenomenon.

A wide variety of methodologies are deployed across the spectrum of these disciplines. Physics, evolutionary biology, and psychology each have their own vocabularies, methods, and procedures, and engage with nature in their own characteristic ways. Each natural science develops a vocabulary and a working method which is appropriated or adapted to its object. The more complex that object, the more levels of explanation are required. A classic example is the human body, which can be investigated at a series of levels – anatomical, physiological, and psychological – each of which illuminates one aspect of the greater whole, but none of which is adequate by itself to give a full account.

So what are the implications of this ‘discipline‐specific’ understanding of scientific method for theology? During the 1930s, the Swiss Protestant theologian Karl Barth argued vigorously for the recognition of the distinctive sources, norms, and methods of Christian theology. Like other disciplines, it had its own sources and norms. The German philosopher Heinrich Scholz, however, had suggested that theology should be judged by the same criteria as every other discipline. Scholz was influenced here by the Enlightenment idea that there was a single rational method, which applied to all fields of study. In Barth's view, Christian theology was not ‘scientific’ because it conformed to some allegedly universal method, but rather because it used a method that was appropriate to the object under investigation.

Thomas F. Torrance is a good example of a theologian who engaged the natural sciences extensively in developing his own theological positions and who affirmed the uniqueness of theological method. This is probably best seen in Torrance's Theological Science (1969), which asserted the distinctiveness of Christian theology, both in terms of its object of inquiry and its research method:

Theology is the unique science devoted to knowledge of God, differing from other sciences by the uniqueness of its object which can be apprehended only on its own terms and from within the actual situation it has created in our existence in making itself known.

For Torrance, both Christian theology and the natural sciences are responses to a reality that lies beyond them, and which are shaped by the reality that is to be apprehended. All intellectual disciplines or sciences are under an intrinsic obligation to give an account of reality ‘according to its distinct nature’. A similar point is made by John Polkinghorne, who affirms that ‘there is no universal epistemology, but rather entities are knowable only through ways that conform to their idiosyncratic nature.’

For Torrance, this means that both scientists and theologians are under an obligation to ‘think only in accordance with the nature of the given’. The object which is to be investigated must be allowed a voice in this process of inquiry. The distinctive characteristic of a science is to give an accurate and objective account of things in a manner that is appropriate to the reality being investigated. Both theology and the natural sciences are therefore to be understood as a posteriori activities which respond to ‘the given’ rather than as a priori speculation based on philosophical first principles. In the case of the natural sciences, this ‘given’ is the world of nature; in the case of theological science, it is God's self‐revelation in Christ.

A Case Study in Explanation: Nancey Murphy on ‘Non‐Reductive Physicalism’

How do Christians understand human nature? What are the essential features of a Christian anthropology? The American philosopher Nancey Murphy has contributed significantly to what she terms the ‘debate over the “ontological constituents” of human beings’. The traditional Christian answer to this question, which received its definitive statements in the Middle Ages, is to draw a distinction between ‘body’ and ‘soul’ (Latin: anima). Humans, it was argued, were distinguished from all other animals and inanimate objects by the possession of this spiritual entity. This approach was held to be justified on biblical grounds, in that the New Testament generally speaks of ‘body and soul’, and occasionally of ‘body, soul, and spirit’. References to the ‘body’ were generally understood by medieval writers to refer to the physical and material parts of humanity, whereas the ‘soul’ was understood as an immaterial and eternal spiritual entity, which merely resided within the human body.

There are two questions requiring further discussion here. First, is this really how we ought to interpret the biblical anthropological statements? Many scholars in the twentieth century pointed out that the notion of an immaterial soul was a secular Greek concept, rather than a biblical notion. The Hebraic vision of humanity was that of a single entity, an inseparable psychosomatic unit, with many facets or aspects. The Old Testament conceives of humanity ‘as an animated body and not as an incarnate soul’ (H. Wheeler Robinson). Second, what challenges are posed to this traditional view by modern neurosciences, which offer no place for such notions as a ‘soul’? How are we to conceive human nature, in the light of both recent trends in biblical interpretation and developments in neuroscience?

Murphy's work engages both these questions, especially the second. She follows the British New Testament scholar James D. G. Dunn in holding that the biblical authors were not concerned to catalogue the metaphysical components of human beings, such as body, soul, spirit, or mind. Their interest was primarily in relationships, and especially a person's relationship with God. Murphy insists on the need for a physicalist account of humanity, which does not invoke or presuppose spiritual or immaterial components. She correctly notes, for example, that biblical terms for aspects of human existence came to be translated by, and eventually elided with, Greek philosophical terms, thus coming to be understood as referring to constituents of humanity. Murphy considers these questions against the backdrop of a broader discussion of ‘physicalism’ in mainstream philosophy, which often presents a physicalist approach to mental causation and mental events as the default position.

So does this lead to a reductionist account of human nature? Murphy rightly points out that this question needs a careful response, given that many are reluctant to accept purely physicalist accounts of the human person, in that these often seem to deny the existence, meaning, or value of those aspects of human life that are seen as particularly significant. Reductionist accounts of human nature seem to call into question many traditional concerns and beliefs about the dignity and theological position of the human person. Murphy helpfully distinguishes a number of senses in which the word ‘reductionist’ is used:

Murphy uses the phrase ‘nonreductive physicalism’ to designate accepting ontological reductionism, while rejecting causal reductionism and reductive materialism. Murphy's position thus involves reclaiming the biblical idea of humanity as an inseparable psychosomatic unit, which is clearly consistent with the modern neuroscientific consensus. Yet her most significant achievement is to show how this ‘nonreductive physicalism’ can avoid the reductionist pitfalls that its critics might anticipate. This involves the development of two ideas: supervenience and ‘downward causation’ (also known as ‘top‐down causation’ or ‘whole‐part causation’).

The notion of supervenience was introduced by 1970 by Donald Davidson (1917–2003) to describe the relation between mental and physical characteristics. Since it is widely regarded as implausible that ideas, minds, and so forth do not exist at all, physicalists often assert instead that ideas and minds ‘supervene’ on material objects. Murphy adopts this idea to show how the behaviour of any given higher‐order system may be influenced strongly, but not completely determined, by the behaviour of its lower‐order components. Neither the freedom of the human mind nor the human will are thus abolished by their physical natures and contexts.

Murphy also appeals to the notion of ‘downward causation’ to defeat reductionism. The importance of this approach has been noted by other writers in the field of science and religion, including Arthur Peacocke and John Polkinghorne. This approach involves challenging the mechanical model of causation which holds that the lower levels of a system determine its higher‐level properties, so that behaviour at higher levels is in some sense ‘explained away’ by the lower‐level systems. On this approach, consciousness is explained by physics. Yet this can easily be challenged. Even if the lower levels of a system were deterministic, the behaviour of the system as a whole is shaped by the configuration of its individual components. The case of biological evolution is an excellent example, where the relation of organisms to their environments plays a significant role in natural selection, which cannot be predicted on the basis of a mechanical model of ‘upward causation’.

What Does it Mean to Explain Something?

Human beings long to make sense of things – to identify patterns in the rich fabric of nature, to offer explanations for what happens around them, and to reflect on the meaning of their lives. Knowing that something has happened, or that something exists, is not the same as understanding why it happened, or why it exists. There is a significant gap between knowing that and knowing why. Both the scientific and religious communities aim to make sense of what is observed. They set out to wrestle with the ambiguities of experience, in order to offer the ‘best explanations’ for what is observed. This is not to say that either science or religion may be reduced to such interpretations but is simply to note that both possess an explanatory dimension.

Human beings clearly consider it important to be able to explain our world – to offer an account, however incomplete, of the interconnections of events and forces in our world that allow us to understand why certain things happen, or why they happen in a certain way. The belief that there is some reasonable explanation for what we observe within our world and experience within ourselves seems to be a universal human intuition. The task of finding such explanations – to make sense of things – is an integral aspect of the human engagement with reality.

The natural sciences are widely understood to offer evidence‐based explanations for what we observe within our universe. Science is about uncovering the basic intelligibility of nature, aiming to identify the deeper structures and broader patterns which lie behind events and entities in the natural world. The philosopher of science Peter Dear's summary of this point would command wide assent within the scientific community:

The hallmark of natural philosophy is its stress on intelligibility: it takes natural phenomena and tries to account for them in ways that not only hold together logically, but also rest on ideas and assumptions that seem right, that make sense; ideas that seem natural.

Some would argue that the intelligibility of nature itself requires explanation. Why is the human mind able to discern the deep rationality of our universe, when there appears to be no good reason for being able to do so? The great German physicist Max Planck, for example, remarked that ‘science cannot solve the ultimate mystery of nature’, in that ‘we ourselves are part of nature and therefore part of the mystery that we are trying to solve’.

John Polkinghorne developed this point further, arguing that Christianity provides an intellectual framework which makes sense of our ability to understand our universe. ‘Theology can render this discovery intelligible, through its understanding that the Mind of the Creator is the source of the wonderful order of the world.’ Theology thus positions scientific insights within a broader and deeper context of intelligibility, providing a framework that holds together both objective and subjective accounts of reality. Science ‘needs to be considered in the wider and deeper context of intelligibility that a belief in God affords.’

Interest in what it means to offer a scientific explanation of observations can be traced back to the Renaissance. However, it became a serious philosophical issue in the early nineteenth century, especially through the work of the English philosopher William Whewell (1794–1866). In his Philosophy of the Inductive Sciences, Whewell set out an account of what a scientist does when trying to make sense of a set of observations. For Whewell, observation involves what he terms ‘unconscious inference’, meaning that we unconsciously or automatically interpret what we observe in terms of a set of ideas. In doing this, Whewell suggests, we add something to this process of observation – namely, some kind of organizing principle which we ‘superimpose’ on the empirical observations, so that they can be seen as interconnected. A good theory is able to ‘colligate’ observations that might otherwise be seen as disconnected, just as a string holds together a group of pearls in a necklace. For Whewell, a theory both identifies and illuminates the ‘true bond of Unity by which the phenomena are held together.’ The theory thus explains empirical observations by ‘superinducing’ upon them a way of thinking which allows their interconnectedness to be grasped.

Ontic and Epistemic Approaches to Explanation

In the late twentieth century, two main approaches to scientific explanation emerged, which are now generally designated as ‘ontic’ and ‘epistemic’. This distinction is due to the philosopher of science Wesley Salmon, who argued that ontic accounts of explanation hold that explanations involve identifying structures and processes within the world which are responsible for the production of the phenomena that are to be explained. Epistemic accounts hold that explanation is concerned with our making phenomena understandable, predictable, or intelligible by setting them in an informing context. In effect, we create or generate a mental framework which organizes empirical observations and fits them into a coherent pattern. Salmon thus contrasts ontic explanations that are grounded externally to the observer, located in the structure of the world, with epistemic approaches that see explanation as a human cognitive achievement, based on conceptual structures generated by the human mind. While many approaches to explanation actually mingle ontic and epistemic elements, it is nevertheless helpful to distinguish these two broad approaches.

The simplest ontic explanations are causal. If A causes B, then A explains B. Salmon points out that this approach to scientific explanation is based on the assumption that ‘underlying causal mechanisms hold the key to our understanding of the world.’ Causal processes and causal laws provide the mechanisms by which the world operates. To understand why certain things happen, we need to see how they are produced by these mechanisms. To explain an event may thus mean providing information about its causal history. It is important to note here that an event may have multiple causes. The complex causal chain of an apple falling to earth might include, for example, the force of gravity, a decaying branch on the apple tree, or a wayward bird which happened to alight on the apple at the time. Causal sequences often involve multiple factors, making it difficult to assign a single cause to an event.

Epistemic approaches to explanation, however, are based on the belief that science achieves explanation by offering a unified picture of the world. A scientific explanation provides a unified account of a range of different phenomena. To understand any given phenomenon is to see how it fits together with other phenomena within a unified whole, discerning the fundamental unity that underlies the apparent diversity of the phenomena themselves. Philip Kitcher is one of many philosophers of science to emphasize the importance of discerning common patterns within nature as the basis of explanation:

Understanding the phenomena is not simply a matter of reducing the ‘fundamental incomprehensibilities’ but of seeing connections, common patterns, in what initially appeared to be different situations.

Many locate the fundamental source of explanatory power lying in ontology – an understanding of the fundamental order of things. Epistemic explanation works best when its conceptual frameworks are considered to correspond to, or be grounded in, the structures of the world. For this reason, many philosophers of science hold that there is an irreducible ontic element to the process of explanation. The French philosopher of science Pierre Duhem (1861–1916) argued that to explain something ‘is to strip reality of the appearances covering it like veils, in order to see this naked reality face to face.’ It is by discovering the ‘big picture’ that its individual elements are able to be both known and understood; explanation is about the location of an event or observation within this deeper context.

So how do we decide what scientific theory provides the best explanation of what we observe? It has long been recognized that any group of observations can be explained in a number of ways. This naturally raises the question of what criteria we might use to decide which of a group of possible explanations is the most reliable. In the next section, we shall look at the complex question of theory choice in science. But before we turn to consider this question, we need to ask whether religion has any explanatory role.

Religion and Explanation

Does religion explain anything? Alvin Plantinga is one of a number of philosophers of religion who suggest that its explanatory potential is not of primary importance for Christianity.

Suppose theistic belief is explanatorily idle: why should that compromise it, or suggest that it has low epistemic status? If theistic belief is not proposed as an explanatory hypothesis in the first place, why should its being explanatorily idle, if indeed it is, be held against it?

Similarly, the philosopher of religion Dewi Z. Phillips (1934–2006) also marginalizes the explanatory aspects of belief in God. Religion does not require explanation, nor does it offer explanations. Phillips here follows the philosopher Ludwig Wittgenstein in playing down any explanatory role for faith.

Others, however, argue that religion does have an explanatory or sense‐making role, and that this is essential in understanding both its appeal and its function. The philosopher Keith Yandell is a representative of this approach:

A religion is a conceptual system that provides an interpretation of the world and the place of human beings in it, bases an account of how life should be lived given that interpretation, and expresses this interpretation and lifestyle in a set of rituals, institutions and practices.

Similarly, Richard Swinburne argued that God is the best explanation for the complex patterns of phenomena that we observe in the natural world. ‘I am postulating a God to explain what science explains; I do not deny that science explains, but I postulate God to explain why science explains.’ The existence of God may, he argues, be inferred from what is observed in the world. And once this idea of God is accepted, it is found to offer an explanation for what we experience around us.

The physicist and theologian John Polkinghorne holds that religion has a particularly important role in dealing with ‘metaquestions’ that science raises, but which point us beyond what science by itself can presume to speak about. Why is the physical universe rationally transparent to us so that we can discern its pattern and structure? Why is it that some of the most beautiful patterns proposed by pure mathematicians are actually found to occur in the structure of the physical world? How are we to account for the ability of mathematics to model so accurately the fundamental structures of the universe? Science gladly exploits the rational transparency of the universe but is unable to explain why the universe is like this. Christianity provides a theistic framework that explains what otherwise has to be regarded as a miracle or a fortunate accident.

The great medieval theologian Thomas Aquinas argued that there was an explicit connection between the existence of God and the human capacity to make sense of our world. God can be considered to be an explanatory agent, whose existence and nature provides a retrospective explanation of various aspects of our experience of the world – such as the ordering of the world, or our sense of goodness or beauty. In a famous 1948 debate between Bertrand Russell and Frederick Copleston, Russell ruled out the need for any explanation of the universe. The universe is there, and nothing can be added to the brute fact of its existence. Aquinas, in contrast, takes the view that it is reasonable to seek an explanation of why the world exists, and why it has its distinct characteristics. The universe is seen to require an explanation in terms of a relationship to something other than itself – that is, God. A similar point is made by the philosopher Thomas Nagel, who argues that the existence of our universe requires a larger explanatory context than scientific laws, in that such a limited explanation ‘would still have to refer to features of some larger reality that contained or gave rise to it.’

Aquinas's approach to explanation in the second of his ‘Five Ways’ is basically causal. This argument can be set out in four stages:

Aquinas sees this argument as offering a causal explanation of what is observed and experienced within the world. Although Aquinas's approach is ontic rather than epistemic, the context within which it is set presupposes some form of conceptual integration of both ontic and epistemic modes of explanation, even if this is not developed at this point in the Summa Theologiae.

This approach has been developed further by many modern philosophers, who have argued for the existence of a transcendent necessary being that is the ground for an ultimate explanation of why particular contingent beings exist and undergo particular events. Theism in general (and Christianity in particular) is argued to articulate a theoretical framework that make possible ultimate explanation of reality. Some have argued for developing a case for the existence of such a transcendent being on the basis of its capacity to explain the existence and character of the world, and then sought to correlate this with the Christian God; others have developed approaches which seek to demonstrate the explanatory capacity of the Christian understanding of God, seeing such a capacity as an indicator of such a God's existence. Yet this specific approach to explanation also includes a critical element, in that it poses the question of how a purely naturalist or materialist interpretation of our world can account for the appearance, through the operation of the laws of physics and chemistry, of conscious beings such as ourselves, who prove to be capable of discovering those laws and understanding the universe that they govern.

Thus far, we have considered general theistic approaches to explanation. But what of specifically Christian approaches? It has often been noted that many leading scientists of the Renaissance saw theology as offering an imaginative template that enabled them to make sense of the world. Many theologians regarded the notion of humanity bearing the ‘image of God’ has to have important epistemic outcomes, including a propensity or capacity to discern God within creation. William Whewell's inductive scientific method reflected his belief that the ‘Fundamental Ideas’ which we use to organize our sciences resemble the ideas used by God in the creation of the physical universe. God has created our minds such that they contain these ideas (or their ‘germs’), so that ‘they can and must agree with the world’.

Much the same idea was set out earlier, in a more mathematical form, by the astronomer Johann Kepler. In his Harmonies of the World (1619), Kepler argued that the theological ‘given’ that human beings were constituted after the form of the ‘image of God’ predisposed them to think mathematically, and thus grasp the structure of the created order:

In that geometry is part of the divine mind from the origins of time, even from before the origins of time (for what is there in God that is not also from God?), it has provided God with the patterns for the creation of the world, and has been transferred to humanity with the image of God.

Kepler's work is of interest for many reasons, not least the fact that it is generally regarded as the last major scientific work to use the musical notion of harmony in scientific reflection on the nature of the world.

Philip Clayton on Explanation in Religion

The American philosopher of religion and philosopher of science Philip Clayton, presently Ingraham Professor of Theology at Claremont School of Theology and Professor of Philosophy and Religion at Claremont Graduate University, California, has given careful thought to the relation of explanatory practices and conventions in science and religion. Clayton's first major published work was a study of explanation in science and religion. Explanation from Physics to Theology: An Essay in Rationality and Religion (1986) is widely regarded as having made a powerful case for retaining the notion of ‘explanation’ as religiously meaningful.

Clayton countered the trend to treat religion as lacking any explanatory potential, which was then prevalent within the philosophy of religion. This view is found in the writings of Ludwig Wittgenstein (1889–1951), especially his caustic remarks on Sir James Frazer's Golden Bough, and has had a significant impact on the philosophy of religion. As noted above, an excellent example of this ‘religion without explanation’ approach is found in the writings of Dewi Z. Phillips.

Responding to this challenge, Clayton pointed out that religious belief systems could indeed be said to offer ‘explanations’, when this term was properly defined. At this stage, explanation was still primarily understood in causal terms. For example, philosopher of science Wesley Salmon argued that ‘to give a scientific explanation is to show how events and statistical regularities fit into the causal network of the world.’ Yet Clayton rightly pointed out that causal notions of explanation were being displaced by their coherentist counterparts. In other words, an ‘explanation’ could be understood in terms of the provision of an intellectual framework which proved capable of the maximum accommodation of observations. Clayton's approach to explanation is reflected in his emphasis upon the theological importance of the notion of ‘inference to the best explanation’, which does not necessarily depend upon a causal account of explanation. The critical question is which of a group of potential explanations of a given set of observations seem to offer the best ‘fit’.

While noting that religious explanations often focused primarily on the way in which individuals make sense of their experience, Clayton pointed out that religious intuitions are not limited to the domain of what might be called specifically religious experience. Indeed, religious explanation has a capacity to make sense of experience as a whole. Thus the believer or mystic senses (or ‘sees’) that things fit together, that there is an underlying coherence to our world.

How Do We Decide What is the Best Explanation?

Recent years have seen a growing interest within the philosophy of science in the idea of ‘inference to the best explanation’. This represents a decisive move away from older positivist understandings of the scientific method, still occasionally encountered in popular accounts of the relation of science and religion, which holds that science is able to – and therefore ought to – offer evidentially and inferentially infallible evidence for its theories. This positivist approach, found at many points in the writings of Richard Dawkins, is now realized to be deeply problematic. It is particularly important to note that scientific data are capable of being interpreted in many ways, each of which has some evidential support. In contrast, positivism tended to argue that there was a single unambiguous interpretation of the evidence, which any right‐minded observer would discover. Since there are many explanations, the question of how the best such explanation is to be identified becomes of acute importance.

A further problem that needs to be noted here is the entanglement of observation and theory. For example, the current estimate of the age of the universe is roughly 13.8 billion years. But how do we know this, in the absence of any continuous chronological monitoring of this history? In 1919, most scientists thought that the universe was of indefinite or infinite age; in 1929, based on an early determination of Hubble's constant, it was believed to be 2 billion years old; now it is believed to be 13.8 billion years old. Current estimates of the age of the universe are based on observations that are interpreted within the premises and parameters of what is known as the ‘Lambda‐CDM model’. The observations themselves tell us nothing directly about the age of the universe; a theoretical framework is needed to interpret them.

The point here is that science does not read off the age of the universe directly; it interprets certain observations within the framework of the Lambda‐CDM model to yield the age of the universe. The speeds and distances of receding galaxies are not observed directly but are inferred on the basis of the premises and parameters of additional physical theories – such as the correlation between velocity and the Doppler red‐shift of the spectra of stars. Like all scientific models the Lambda‐CDM model is provisional and can be expected to undergo change and modification over time. Yet we cannot predict what those changes will be, nor what their impact will be on estimates of the age of the universe.

To begin our reflections on these questions, we will consider the distinction between generating new theories and testing these new theories. This distinction is often expressed in terms of a contrast between a ‘logic of discovery’ and a ‘logic of justification’.

‘Logic of Discovery’ and ‘Logic of Justification’

How do new theories emerge? How, for example, did Charles Darwin come up with the idea of evolution through natural selection? Or, to give a famous example from the field of chemistry, how did the German organic chemist August Kekulé realize that molecules of benzene had a cyclical structure? Kekulé answered this question in 1890, telling how the idea came to him while he was drowsing in front of his fire at home. He had a vision of a snake chasing its own tail, which suggested to him that benzene had a ring structure. And having proposed such a structure for benzene, Kekulé checked it out against the experimental evidence, and found that it seemed to account for it satisfactorily.

The way in which Kekulé came up with the idea that benzene had a cyclical structure is now seen as a classic example of a ‘logic of discovery’, the process of developing new hypotheses for consideration, which is distinguished from the subsequent ‘logic of justification’ in which these hypotheses are checked out against the available evidence and any predictions they make. It is now widely accepted that the ‘logic of discovery’ is essentially imaginative or creative, involving making connections which others have failed to see. The ‘logic of justification’, however, is fundamentally rational and critical, aiming to subject this new theory to rigorous critical examination. The manner in which a theory is derived is not of ultimate importance in determining whether it is right; what really matters is how well that theory can account for existing evidence and perhaps predict new and unknown findings.

An early anticipation of this approach to a ‘logic of discovery’ can be found in the American pragmatist philosopher Charles Peirce (1839–1914), son of a noted Harvard astronomer and himself a scientific practitioner. For Peirce, scientific thinking can be described as a specific form of ‘abductive inference’, which can be set out as follows:

So how do we arrive at the hypothesis A, which explains C? Peirce argues that observing how scientists actually work shows that new theories and hypothesis are generated in many ways, including processes which Peirce describes as ‘inspiration’ and ‘imagination’. The way in which Kekulé imagined benzene as having a ring structure fits easily within Peirce's description of this logic of discovery.

A related approach is found in the American philosopher of science N. R. Hanson's reflections on the advance of scientific knowledge. Hanson (1924–1967) argued there were three common features in what he called ‘the logic of scientific discovery’:

Like Peirce, Hanson thus identifies astonishing or surprising observations as a fundamental motivation in the enterprise of scientific discovery. Is there a theoretical standpoint from which these observations would not be astonishing but would be expected?

A good example of this lies in Albert Einstein's explanation of a puzzling observation relating to the planet Mercury which could not be accounted for by existing theories – such as Newtonian mechanics. The perihelion of Mercury (the point at which it was closest to the sun) was discovered to move by a tiny – but observable – amount each year. There was no obvious reason for this, although the French mathematician Urbain Le Verrier argued in 1859 that it could be explained if there was a hitherto unknown planet of roughly half the mass of Mercury, positioned closer to the sun. Le Verrier named this hypothetical planet ‘Vulcan’. It was never found, despite several false reports of observations of this planet throughout the 1860s.

In November 1915, Einstein announced that the advancing perihelion of Mercury was explained precisely and persuasively by his new general theory of relativity. This arose on account of the planet's motion through a gravitational field that was warped by the Sun's enormous mass. Although this effect could be observed for all the planets, it would be most noticeable in the case of Mercury as Mercury is the closest planet to the sun.

Einstein's general theory of relativity provided a new theoretical framework which was able to accommodate a known and baffling observation. Yet Einstein also predicted that the same basic pattern could be observed with all the remaining planets, even though it would be virtually impossible to detect this using the instrumentation available in the 1910s. His theory also made some novel predictions – most notably, phenomenon now known as ‘gravitational lensing’, in which the warping of space‐time due to the gravitational influence of the sun causes light to bend. Although Newtonian mechanics predicted a certain degree of bending of light by gravity, Einstein's theory made it clear that this was supplemented by the warping of space‐time. In the end, this enhanced bending of light by gravity was observed during a solar eclipse in 1919, providing a spectacular confirmation of Einstein's theory.

Inference to the Best Explanation

The approach now generally known as ‘inference to the best explanation’ recognizes that multiple explanations might be offered for any given set of observations and sets out to identify criteria by which the best such explanation might be identified and justified. Yet the best theory may not be a true theory – it may simply be the best approach available at this specific moment in history. So what are these criteria? A number of ways of evaluating theories or explanations have been put forward by philosophers of science. In what follows, we shall consider three widely used criteria.

A further issue of debate concerns the ranking of these criteria. Which is the most important? And what is the scientific basis of this choice? In practice, these criteria and others like them are seen as signs, not as proofs, of theoretical reliability. There is, however, a further issue that needs to be noted. The method of ‘inference to the best explanation’ may help us work out which of a group of possible explanations is the ‘best’ – but it does not follow that the ‘best’ of these explanations is actually true. It is simply better than its rivals. The philosopher Gilbert Harman, who is widely credited with introducing the idea of ‘inference to the best explanation’, nevertheless considered that there were good reasons for believing that the best explanation was likely to be true:

In making this inference, one infers from the fact that a certain hypothesis would explain the evidence, to the truth of that hypothesis. In general, there will be several hypotheses which might explain the evidence, so one must be able to reject all such alternative hypotheses before one is warranted in making the inference. Thus one infers, from the premise that a given hypothesis would provide a better explanation for the evidence than would any other hypothesis, to the conclusion that the given hypothesis is true.

This discussion of theory choice is somewhat abstract. In what follows, we shall look at a case study which helps clarify some of the issues. How did Charles Darwin come to the conclusion that his theory of evolution by natural selection was preferable to the alternative hypotheses of his age?

A Case Study: Darwin and Natural Selection

Charles Darwin's appeal to the novel concept of natural selection as the ‘best explanation’ of an accumulated body of observations concerning natural history is widely cited as a successful application of the inductive process now known as ‘inference to the best explanation’. For Darwin, as we noted earlier (43–4), four features of the natural world in particular seemed to require particularly close attention, in the light of problems and shortcomings with existing explanations, especially the idea of ‘special creation’ offered by earlier religious apologists such as William Paley (1743–1805). Paley argued that the complexity of the biological realm was best explained by the idea of special divine creation. Such complexity, he argued, could not have come about accidentally, and was to be seen as evidence of design.

While Paley's theory offered explanations of these four observations, they seemed cumbersome and forced. Darwin believed that there had to be a better explanation than that offered by Paley for these four observations. These observations are:

As noted, all these aspects of the natural order could be explained on the basis of William Paley's theory. Yet the explanations that Paley and his followers offered seemed cumbersome and contrived. What was originally a relatively neat and elegant theory began to crumble under the weight of accumulated difficulties and tensions. Darwin believed that there had to be a better explanation, which would account for these observations more satisfactorily than the alternatives which were then available.

Darwin was quite clear that his theory of natural selection was not the only possible explanation of the biological data. He did, however, believe that it possessed greater explanatory power than its rivals, such as the doctrine of independent acts of special creation. ‘Light has been shown on several facts, which on the belief of independent acts of creation are utterly obscure.’ Darwin's theory had many weaknesses and loose ends. Nevertheless, he was convinced that these were difficulties which could be tolerated on account of the clear explanatory superiority of his approach. Yet even though Darwin did not believe that he had adequately dealt with all the problems which required resolution, he was confident that his explanation was the best available. In the sixth edition of the Origin of Species, he responded to some theoretical objections to his approach as follows.

It can hardly be supposed that a false theory would explain, in so satisfactory a manner as does the theory of natural selection, the several large classes of facts above specified. It has recently been objected that this is an unsafe method of arguing; but it is a method used in judging the common events of life, and has often been used by the greatest natural philosophers.

While recognizing that his theory of natural selection lacked rigorous proof, Darwin clearly believed that it could be defended on the basis of criteria of acceptance and justification that were already widely used in the natural sciences, and that its explanatory capacity was itself a reliable guide to its truth.

Theory Choice and Religion

So are there parallels within religion to these questions about theory choice? In recent years, Christian philosophers and apologists have become increasingly interested in inductive approaches to the rationality of faith, especially in relation to the question of the existence of God. The philosopher Richard Swinburne, for example, argued that the existence of God is to be seen as the best explanation of what is observed within the world, when seen as part of a larger cumulative case. For Swinburne, the existence of the universe can be made comprehensible if we suppose that it is brought about by God.

Swinburne sets his approach within a wider framework, grounded in the core belief that the existence of a universe needs explaining, rather than merely being accepted as a ‘brute fact’ (Bertrand Russell). For Russell, there can be no further explanation for its existence or its fundamental features other than the statement ‘it is there’. So what are the possible explanations, and which of them is the best? Swinburne suggests that there are basically two main rival theories which need to be considered as possible explanations: the view that science can provide a natural explanation for the existence of this universe, or the theistic view that the universe and its phenomena exist because of the intentional causal activity of a personal being, known as ‘God’.

Swinburne thus sets out to identify possible explanations of the universe, and to determine which of these is ‘best’. In making this decision, Swinburne does not see himself as having to prove the existence of God. His task is rather to show that the existence of God, however unlikely this might appear as an independent hypothesis, is better at explaining our nexus of observations and experiences than its alternatives – such as a materialist naturalism. A priori, theism might perhaps seem very unlikely; yet, Swinburne argues, it is far more likely than its explanatory rivals.

So what criteria does Swinburne use in evaluating rival explanations of the existence of the universe? In developing this kind of inductive cosmological argument, however, Swinburne appeals to the criterion of simplicity in deciding between competing hypotheses concerning the existence of the universe, arguing that ‘science requires us to postulate the simplest explanation of the data’. The simpler a theory, the more likely it is to be right. His argument is open to challenge. However, Swinburne's approach to the rationality of theism is an important indication of the way in which scientific criteria of theory choice have found their way into religious discussions.

Let us now refocus on the question of how beliefs – whether scientific theories or religious ideas – may be verified. In what follows, we shall look at the questions which arise with any attempt to develop true statements about our world. In what ways can we assess these? Two particularly significant approaches to this question emerged during the twentieth century: verificationism, which held that the natural sciences were capable of stating their ideas in forms capable of being confirmed from experience, and falsificationism, which held that the natural sciences were able to state their ideas in such a way that defective approaches could easily be shown to be false, even if it turned out to be rather more difficult to confirm valid theories than verificationists had once thought.

The background to this important debate is to be found within the Vienna Circle, one of the most significant philosophical movements to arise in the twentieth century, which had its origins in the Austrian capital city. We begin by considering this highly influential movement and its impact on ‘Logical Positivism’.

Verification: Logical Positivism

The ‘Vienna Circle’ is generally regarded as the group of philosophers, physicists, mathematicians, sociologists, and economists who gathered around the philosopher Moritz Schlick (1882–1936) during the period 1924–1936. One of the core statements of the group was that beliefs must be justified on the basis of experience. This belief is grounded in the writings of David Hume, and is clearly empirical in tone. For this reason, the members of the group tended to place a particularly high estimation on the methods and norms of the natural sciences (which were seen as the most empirical of human disciplines) and a correspondingly low estimation of metaphysics (which was seen as an attempt to disengage with experience). Indeed, one of the more significant achievements of the Vienna Circle was to cause the word ‘metaphysics’ to have strongly negative connotations.

For the Vienna Circle, statements which did not directly relate to the real world were of no value. Every proposition must be capable of being stated in a manner which relates directly to the real world of experience. The overall programme which was proposed by the Vienna Circle had two basic parts, as follows:

The most significant attempt to carry this programme through is to be seen in the works of Rudolph Carnap (1891–1970), particularly his 1928 work The Logical Construction of the World. In this work, Carnap set out to show how the world could be derived from experience by logical construction. It was, as he put it, an attempt at the ‘reduction of “reality” to the “given”’, which applied the methods of logic to statements derived from experience. The only two sources of knowledge are thus sense perception and the analytical principles of logic. Statements are derived from and justified on the basis of sense perception and related to each other and their constituent terms by logic.

Carnap set out what is now generally known as the ‘principle of verification’. Only statements which are capable of being verified can be considered to be meaningful. The natural sciences are thus to be seen as privileged in any theory of knowledge. Philosophy is best seen as a tool for clarifying what has been established on the basis of sense perception. Philosophy, according to Carnap, consists in the ‘logical analysis of the statements and concepts of empirical science’. Carnap thus asserted that religious statements were unscientific. Sentences which make statements about ‘God’ or ‘the transcendent’ are meaningless, in that there is nothing that is given in human experience that can verify them.

These views were popularized in the English language world by A. J. Ayer (1910–1989), especially in his famous book Language, Truth and Logic (1936). Although the Second World War interfered with the process of its reception and evaluation, this single work is widely regarded as setting the philosophical agenda for the two decades following the end of the war. Ayers’ vigorous and radical application of the verification principle held that metaphysical statements (which included religious beliefs) were ‘meaningless’. Ayer allowed that religious statements might provide indirect information concerning the state of mind of the person making such a statement. They could not, however, be considered as making meaningful statements concerning the external world.

So how did theologians respond to this challenge? One popular approach was the notion of ‘eschatological verification’, which was widely discussed during the period 1955–1965. This can be regarded as a direct response to the issues raised by the demand for verification as a condition for meaningfulness. (The term ‘eschatological’ derives from the Greek phrase ta eschata, ‘the last things’, and relates to issues such as the Christian hope and heaven.) The Oxford philosopher Ian M. Crombie remarked that the experience on the basis of which religious statements could be verified was simply not accessible at present – but that it would be available after death.

However, this issue has since receded in importance since the 1960s, not least on account of an awareness of the severe limitations placed upon the verification principle proposed by logical positivism. To illustrate some of these difficulties, we may consider the following statement: ‘There were six geese sitting on the front lawn of Buckingham Palace at 5.15 p.m. on 18 June 1865.’ This statement is clearly meaningful, in that it asserts something which could have been verified. But we are not in a position to confirm this statement now. A similar difficulty arises in relation to other statements concerning the past. For someone such as Ayer, these statements must be considered to be neither true nor false, in that they do not relate to the external world. Yet this clearly runs contrary to our basic intuition that such statements do make meaningful (and potentially verifiable) affirmations.

A further issue concerned the status of unobservable theoretical entities – such as subatomic particles – whose existence is inferred, rather than observed. In a 1938 paper entitled ‘Procedures of Empirical Science’, the American physicist Victor F. Lenzen (1890–1975) argued that certain entities had to be inferred from experimental observation, even though they could not themselves be observed. For example, the behaviour of oil droplets in an electric field leads one to infer the existence of electrons as negatively charged particles of a certain mass. These electrons cannot be seen (and hence cannot be ‘verified’) – yet their existence must be seen as a reasonable inference from the observational evidence. Lenzen's comments highlighted a problem with verificationism, at least in the original forms of the verification principle.

Verificationism, then, has serious limits. It is therefore instructive to note a rival approach that developed in response to some of the perceived difficulties with verificationism. This rival approach is generally known as ‘falsificationism’, and will be considered in the following section.

Falsification: Karl Popper

The Austrian philosopher Karl Popper (1902–1994) argued that the development of scientific knowledge was an evolutionary process in which a number of competing conjectures, or tentative theories, are systematically subjected to the most rigorous attempts at falsification possible. This process of error elimination, he suggested, was analogous to the process of natural selection in evolutionary biology. For Popper, scientific knowledge advances through the interplay between tentative theories (conjectures) and error elimination (refutation).

Popper felt that the verification principle associated with the Vienna Circle was too rigid and ended up excluding many valid scientific statements. ‘My criticism of the verifiability criterion has always been this: against the intention of its defenders, it did not exclude obvious metaphysical statements; but it did exclude the most important and interesting of all scientific statements, that is to say, the scientific theories, the universal laws of nature.’ Yet Popper was also convinced that verificationism was open to criticism for another reason. It ended up by allowing a number of ‘pseudo‐sciences’, such as psychoanalysis and astrology, to pass themselves off as being ‘scientific’ when they were nothing of the sort.

So what was wrong with psychoanalysis? Popper's criticisms of psychoanalysis were mainly directed against the ideas of Alfred Adler, which were influential in Vienna in that time. Adlerianism seemed able to explain just about anything. Why is someone is honest? The answer lies in early‐childhood events. Why is someone dishonest? The answer lies in early‐childhood events. The Adlerians were never wrong; indeed, they seemed incapable of being wrong about anything, in that they were able to make any given observation fit their theories perfectly. Everything in the world was to them evidence for their theory; nothing seemed to count as evidence against it.

Popper proposed the criterion of falsifiability as a way of distinguishing real science from pseudo‐science – now generally known as the ‘problem of demarcation’. At some point around 1920, Popper recalls reading a popular scientific account of Einstein's general theory of relativity. He was particularly impressed by Einstein's precise statements of what would be required to demonstrate that his theory was incorrect. Einstein declared that ‘if the red shift of the spectral lines due to the gravitational potential does not exist, then the general theory of relativity will be untenable.’ Einstein was looking for something which might falsify his theory. For Popper, this represented a totally different attitude and outlook from that he associated with pseudo‐sciences such as astrology. Those committed to these ideologies simply looked for evidence which could confirm their ideas.

While following the Vienna Circle in insisting that a theoretical system must be capable of being tested against observation of the world, Popper argued that scientific theories had to be formulated in such a way that they could be shown to be wrong. There was a need for a ‘criterion of demarcation’ between the genuine natural sciences and those which claimed a scientific status for what was essentially a pseudo‐science.

I shall certainly admit a system as empirical or scientific only if it is capable of being tested by experience. These considerations suggest that not the verifiability but the falsifiability of a system is to be taken as the criterion of demarcation …. It must be possible for an empirical scientific system to be refuted by experience.

Where logical positivism stressed the need for stating the conditions under which a theoretical statement could be verified, Popper held that the emphasis must fall upon being able to state the conditions under which the system could be falsified.

Popper's approach had considerable influence within the philosophy of religion during the 1950s and 1960s and is especially linked with what has come to be known as the ‘falsification’ debate. In his influential 1950 essay ‘Theology and Falsification’, the philosopher Anthony Flew argued that religious statements cannot be regarded as meaningful, in that nothing drawn from experience can be regarded as falsifying them. Yet this debate has now faded, as the difficulties associated with Popper's approach have become clearer. Popper's attempt to set up a meaningful falsification criterion turns out to be rather more difficult than he had hoped.

Popper argued that experiments could falsify a theory. However, the French philosopher of science Pierre Duhem had earlier argued that it was, as a matter of fact, impossible to devise a ‘critical experiment’, in that there would always be a significant degree of uncertainty as to whether the experiment required a theory to be abandoned in its totality, or whether the difficulty lay in only one of its hypotheses, or even an auxiliary hypothesis, which was not of fundamental importance to the theory itself. Popper's approach appeared to ignore the strongly theory‐laden nature of experimental observation, which rendered his critique considerably less potent than he might have hoped.

So what is meant by the assertion that observation is ‘theory‐laden’? The basic idea is that we see and interpret the world through pre‐existing mental maps which are brought into play as we observe the world. We think we are seeing the world as it actually is, without realizing that we are actually looking at it – and making sense of it – through a kind of mental map that tells us what we are seeing. The process of observation is at the same time a process of interpretation. The great nineteenth‐century philosopher of science William Whewell made this point when he declared that there is ‘a mask of theory over the whole face of nature’. Observation and interpretation are interconnected in an inescapable circularity. ‘A fact under one aspect is a theory under another.’ For this reason, the distinction between the ‘factual’ and ‘theoretical’ was problematic, in that it rested on unacknowledged epistemological precommitments on the part of the observer.

Whewell's idea of the ‘theory‐laden’ nature of observation has been developed more recently by N. R. Hanson, who insisted that we do not merely ‘see’ nature; we understand it in a certain way, seeing it as something. The allegedly ‘objective’ process of observation is not neutral and unbiased, but is in reality a theory‐laden process which involves the observer using implicit conceptual schemas, even though these are open to challenge and change. When we observe nature we are wearing a set of theoretical ‘spectacles’, an implicit set of assumptions or expectations which create a degree of perceptual bias, and thus we tend to overlook or disregard evidence because it does not fit with our existing mental schemes.

For example, it is now known that many observations of the planet Uranus were made before its ‘discovery’ by William Herschel in 1781; while Uranus was indeed ‘seen’ by earlier observers, its true status as a planet was not recognized. It was not seen as a new planet, but was simply seen as yet another star. Its low magnitude (Uranus is only just visible to the naked eye) and its slow period of rotation around the sun meant that it was never recognized as a planet until Herschel noticed its distinct planetary disc using a new telescope with considerable light‐gathering power.

This brings us back to Duhem, who noted that any theory that is proposed to make sense of observations will be made up of a number of hypotheses, some of which may be of central importance, others of which are subsidiary. Duhem's point is that a theory consists of a complex network of interlocking hypotheses, some of which were central and others of which were peripheral. So, if something which is predicted by the theory does not correspond with experimentation, which of the assumptions is wrong? A central hypothesis? If so, the theory would have to be abandoned. Or one of the peripheral assumptions? If so, the theory simply needs modification.

According to Duhem, the physicist simply is not in a position to submit an isolated hypothesis to experimental test. An experiment in physics can never condemn an isolated hypothesis but only indicate that there is a problem with a group of hypotheses. The physicist cannot subject an individual hypothesis within this group to an experimental test, in that the experiment can only indicate that one hypothesis within a larger group of hypotheses requires revision. The experiment tests a group of hypotheses, and does not itself indicate which of the hypotheses requires modification. Duhem argued that the notion of a ‘crucial experiment’ needed to be treated with considerable caution.

Consider a well‐known historical case study. Following the discovery of the planet Uranus in March 1781, it was found that the observed motion of the new planet was not what was predicted by Newtonian mechanics. Popper thus concluded that this had to be seen as a clear case of the falsification of Newton's gravitational theory by observation. Others at the time, however, held that this observation called for the modification, not the rejection, of Newton's theory. It was assumed that no planets lay beyond Uranus. But what if there was a planet beyond Uranus, so that the observed orbital perturbation of Uranus might reflect the gravitational influence of this hypothetical trans‐Uranic planet? Calculations of the location of this possible planet by mathematicians in England and France led to the discovery of the trans‐Uranic planet Neptune in 1846. Popper's approach was unable to account for the well‐known and widely accepted scientific practice of theory modification in response to observations.

Yet this raises another question, of considerable interest in itself and to the general field of science and religion. How does the scientific community decide that an existing theory is inadequate, and needs modification – or possibly rejection – in favour of an alternative? In the next section, we shall consider the relevance of the views of the American philosopher of science Thomas S. Kuhn for these important questions.

Theory Change in Science: Thomas S. Kuhn

In his The Structure of Scientific Revolutions (1962), Thomas S. Kuhn (1922–1996) argued that the prevailing view of the nature of scientific progress was that radically new theories arise gradually through verification or falsification. This ‘gradual progress’ model is found in many works, including Karl Popper's Logic of Scientific Discovery. Kuhn disagreed, arguing that the historical evidence suggested that the transition from one scientific ‘paradigm’ to another is not gradual, but rather takes the form of periods of relative theoretical stability with occasional radical shifts in understanding, which he termed ‘paradigm shifts’.

On the basis of his historical studies of the development of the natural sciences, Kuhn argued that a given paradigm comes to be accepted as normative on account of its past explanatory success. Once a given paradigm has been accepted, a period of what Kuhn terms ‘normal science’ follows. During this period, the paradigm which resulted from this earlier success is treated as unproblematic and is generally unchallenged. Empirical evidence that seems to be inconsistent with this paradigm is either treated as an anomaly or seen as something that can be accommodated within this existing approach. This evidence may pose difficulties for the prevailing paradigm but is not seen as requiring the paradigm to be abandoned. The anomaly is regarded as something for which a solution is anticipated within the context of that paradigm, even if at present the precise nature of that solution remains unclear. Ad hoc modifications are proposed to the existing paradigm – as in the case of Ptolemaic astronomy, in which disparity between theory and observation could be accounted for through the addition of additional epicycles to the system.

But what happens if a series of anomalies build up and achieve a cumulative force which calls the paradigm into question? Or if a single anomaly is recognized as being of such significance that the challenge which it poses can no longer be overlooked? Kuhn argues that, in such situations, a crisis emerges within the paradigm. Sections of the scientific community realize that the paradigm is at breaking point, and that something new and more satisfying needs to be found. A ‘scientific revolution’ takes place when the scientific community realizes that a tipping point has been reached, and that the old model must be abandoned in favour of something new.

Kuhn contrasts this revolutionary approach with an essentially evolutionary model, where the latter sees a steady progression in scientific understanding through a gradual accumulation of data and understanding. Where other historians of science spoke of ‘scientific progress’, Kuhn preferred the imagery of a revolution, in which a major change in assumptions took place over a short period of time. Kuhn argued that a complex network of issues lie behind the decision to abandon one paradigm and accept another, and that these cannot be explained solely on the basis of scientific considerations. Highly subjective issues are involved. Kuhn compares a ‘paradigm shift’ to a ‘conversion’. The adoption of a new paradigm is accompanied by a sudden and intuitive alteration of perception such as a ‘gestalt switch’ – a popular psychological image of the late 1950s. Kuhn remarks that ‘no ordinary sense of the term “interpretation” fits these flashes of intuition through which a new paradigm is born.’

At this point, we should note that Kuhn's use of the term ‘paradigm’ is not entirely consistent. He uses the term in two broad senses. Generally, the term refers to the broad group of common assumptions which unites a particular group of scientists. Yet it can sometimes be used in a more specific and restricted sense to refer to a past scientific explanatory success, which seems to offer a framework which can be treated as normative, and is hence treated as exemplary or normative thereafter – until something finally causes that paradigm to be abandoned.

Kuhn's emphasis on the subjective reasons for paradigm shifts has led some of his critics to suggest that his account of scientific development rests on little more than ‘mob psychology’. This is probably rather unfair. Kuhn has simply noted the sociological aspects of opinion change within scientific communities.

The transition between competing paradigms cannot be made one step at a time, forced by logic and neutral experience …. It must occur all at once (though not necessarily in an instant) or not at all …. In these matters neither proof nor error is at issue. The transfer of allegiance from paradigm to paradigm is a conversion experience that cannot be forced.

A new theory thus allows us to observe things in a new way, displacing an older way of interpreting and envisaging the world. Kuhn thus highlights the ‘theory‐laden’ nature of observation, and points out its implications for the transition from Ptolemaic to Copernican theories of the solar system.

Before it occurred, the sun and moon were planets, the earth was not. After it, the earth was a planet, like Mars and Jupiter; the sun was a star; and the moon was a new sort of body, a satellite.

Kuhn's point is that the phenomena were unchanged; they were, however, interpreted in a new way, in that the sun was now seen as a star. Observation is thus not a neutral process, but is shaped by assumptions, explicit or implicit, about what is being observed. Observation is about seeing as, not just seeing. One observer might see the sun rise and set; another might see the earth turning on its axis, leading to the apparent motion of the sun across the heavens. Both, however, are looking at the same natural phenomenon.

Kuhn's analysis has importance for religious belief, and two of his central themes may be explored to illustrate his relevance. First, Kuhn's concept of ‘paradigm shifts’ is helpful in attempting to understand the major intellectual shifts which have taken place in the history of religious thought. As we have noted, religious thinking is influenced, at least to some extent, by the cultural and philosophical presuppositions of the day. Radical shifts in these background assumptions can thus be of major importance, as the development of Christian theology has demonstrated. The Reformation and the Enlightenment are epochs in Christian thought which can be seen as representing ‘paradigm shifts’ in an understanding of how theology should be done. Existing understandings of background presuppositions, norms, and methods are often radically altered – and occasionally abandoned altogether – in the transition from one paradigm to another.

A good example of such a radical change in a theological paradigm can be seen in the late twentieth‐century discussion of the question of whether God suffers. Throughout the early Christian era and Middle Ages, it was generally assumed that God could not suffer. There are examples of writers who speak of God suffering; these, however, are few and far between. Jesus Christ suffered on the cross; yet this suffering was held to be something that related to Christ's human, but not his divine, nature. God knew about human beings experiencing suffering, and was sympathetic to their pain. Yet God was seen as being above suffering. The reasons for this theological consensus remain unclear. Some scholars see this as an expression of a philosophical view of the perfection of God. Philo of Alexandria vigorously defended the impassibility of God: ‘What greater impiety could there be than to suppose that the Unchangeable changes?’

This consensus was shattered in the 1970s, apparently in response to a growing belief that the levels of suffering in the twentieth century made it apologetically impossible to treat God as being above or beyond suffering. In A Theology of the Pain of God (1946), the Japanese writer Kazoh Kitamori argued that God's love was rooted in the pain of the world. In his Crucified God (1972), Jürgen Moltmann argued that a God who cannot suffer is a deficient, not a perfect, God. Stressing that God cannot be forced to change or undergo suffering, Moltmann declares that God willed to suffer. The suffering of God is the direct consequence of God's decision to suffer and willingness to suffer. The nature of love is such that it involves the lover participating in the sufferings of the beloved. Such was Moltmann's influence that a tipping point was reached within western Protestantism. A ‘new orthodoxy’ emerged, holding that God suffered. Others, however, were resistant to the idea of a suffering God, seeing this as both unnecessary and improper.

One final point must be made in assessing Kuhn's approach to ‘scientific revolutions’. What explanation may be offered for progress in scientific theorizing, as opposed to change in scientific theorizing? The term ‘progress’ clearly implies a judgement – that these changes are for the better. So does a scientific revolution necessarily lead to a better grasp of the truth about nature? Kuhn rejects realism as an explanation of the successes of scientific research, and thus does not see an increasing convergence between ‘theory’ and ‘reality’ as an explanation of scientific progress. Kuhn argues that nothing is lost in rejecting a realist account of scientific development. Yet how can we meaningfully talk about scientific ‘progress’, unless there is some means of knowing that science is proceeding in the right direction, rather than taking a false turn which will need to be corrected in future? It seems clear there is need for further discussion at this point.

This chapter has considered some areas of potential engagement and interaction between religion and the philosophy of science. So what about traffic in the other direction? In we shall move on to consider some areas of potential engagement and interaction between the natural sciences and the philosophy of religion.

For Further Reading

General Themes

Realism and Instrumentalism

Explanation in Science and Religion

Non‐reductive Physicalism and Explanation

Verification and Falsification

↑ Back to top