The Universe Around Us: An Integrative View of Science & Cosmology

Chapter 2: Understanding the Universe

Section 2.1: Different views

Section 2.2: The common base

Section 2.3: Scientific investigation

Section 2.4: Key elements of a broad approach

Section 2.5: The basic understanding attained

References for Chapter 2

This chapter considers the basis of the scientific approach to understanding the physical universe, and how this can be broadened into an approach to understanding in general. This provides a good background for what follows..


Section 2.1: Different views

The scientific viewpoint is but one of the various ways humanity has tried in the quest for understanding and the search for a pattern of meaning [7-10]. We cannot attain understanding without such a search because of one of the fundamental aspects of the situation that confronts us: namely, the hidden nature of reality. I assume this reality exists, and can to some extent be discovered. There are of course various viewpoints that deny one or other of these assumptions. While they may be philosophically amusing, they cannot be sustained in a serious search for a satisfying overall view of the nature of the world, precisely because they deny that that quest can have any meaning.

Many of the scientific aspects of the world are far from obvious; we cannot without considerable effort and perspicacity deduce the nature of the chemical elements, the fundamental forces that bind matter together, and so on. We can easily appreciate the majesty of mountains and the beauty of flowers, but only with skilled experimentation determine that they are made of a combination of carbon, nitrogen, oxygen, and other elements. Furthermore when we do succeed in understanding physical aspects of reality, its nature may be quite unexpected (for example, the essence of relativity theory and the nature of quantum mechanics).

Thus one of the prime issues is determining the essence and scope of the hidden nature of reality. The initial attempts to relate to this hidden aspect of nature may be broadly characterised as magical or superstitious approaches: it is assumed either that if one wishes hard enough for something it will happen, or that if one adopts various ritual practices intended to bring about an effect, it will occur, whereas in fact they have no causal relation to it and cannot influence it. The evident inefficacy of this approach, coupled with obvious injustices that have often resulted from its practice (witch-hunts leading to the death of innocent people, for example) have then led to development of two diverging world views.

On the one hand the religious view has evolved serious approaches to moral issues and to possible understandings of ultimate reality, although often in a faulty and misleading manner; and on the other the scientific view has with tremendous success tackled the issue of immediate causes, leading to an extraordinarily effective understanding of how things work in terms of physical cause and effect. Indeed one of the major triumphs of science is precisely the understanding that because the sequence of natural events is governed by regular laws of behaviour, each physical event has causes that can be determined by appropriate investigation; and desired final effects can be achieved by organising an appropriate set of initial conditions. This establishes the firm relation between physical cause and effect.

The achievements of science are undeniable: it has led to discovery of the physical basis of nature and laid the foundations for the huge explosion of technology enabling us to order our lives and control our environment in ways previously unimaginable. In the face of this enormous success, other world views, and in particular the theological one that held sway for many hundreds of year, have been forced to retreat: they have had to modify their truth claims, abandoning to science much previously claimed territory. However there has recently also been a reaction against science in many quarters, on the one hand because of its perceived contribution to environmental degradation and the development of weapons of mass destruction, and on the other because many see it as based on an outlook that dehumanises and denies the value of the individual [2]. In many cases this has led to what is essentially a return to magic, in the form of astrology, arcane effects ascribed to pyramids or crystals, and so on. These are bound to disappoint eventually, as they are not based on real cause and effect relations; however the fact that people turn to them is evidence that many are seeking a world view with elements of humanity and hope not provided by the present dominant scientific one. Sometimes the effort is dressed in pseudo-scientific terms, wanting to inherit some of the mantle of scientific success but avoiding the kind of logic and indeed hard work demanded to make good that hope.

We will here follow the scientific method, adopting a specific set of values that underlie the approach. Broadly speaking, the values adopted here are that we try to find a description that reflects the truth insofar as we are able to achieve that aim, in particular therefore taking seriously the discoveries and viewpoint of modern science. This does not prevent us also acknowledging the worth and significance of the broad range of human activity (including aesthetic and moral choices, personal life decisions and experiences), but investigating them is a separate project.

We can then argue strongly we have determined in a culture-free way some invariant aspect of underlying reality. This does not mean we can describe the ultimate nature of that reality, but rather that we can characterise its effective nature, which as far as we are concerned is absolute; for example, if we let go of an object we hold in our hand, it will fall; if people have no food, they will die.


Section 2.2: The common base

In each area of understanding, whether we make the fact explicit or not, our understanding is based on mental models of reality that will to a greater or lesser degree reflect accurately the nature of some aspect reality. They may be simple ideas, perhaps comprised in a single label (the word "cat", for example, conjures up a whole model of appearance and behaviour), or complex theories (e.g. the theory of relativity, or Jungian psychology). Each such model will have a range of applicability, specifying both the set of phenomena that are its concern, and a (usually restricted) set of conditions within that domain it is supposed to explain. For example Newtonian mechanics explains the motion of physical bodies, provided they do not move at speeds close to the speed of light; a theory of psychology may describe the ordinary behaviour of people at work, but not that of psychopaths.

Theories can never be absolute: indeed the understanding they give will always be partial, because no model can circumscribe within itself the full nature of reality. Thus theories are always subject to revision. Hence the key element underlying the approach proposed is that, whatever the field of application, growth in understanding is based on a creative proposal of theories to explain reality}, but always founded on the combination of skepticism, i.e. the willingness to doubt the current orthodoxy, and testing, i.e. checking that orthodoxy, or any proposed alternatives, against reality in the whole variety of ways that is practical. Indeed this is the foundation of all learning, through the basic learning cycle [11]: in essentials, we

1: set up a hypothesis on the basis of present knowledge;

2: work out its consequences, in the process checking that it is coherent, making logical sense,

and then

3: test its consequences against evidence,

- if possible performing new experiments that can check if it is true or not.

Then we return to step 1, if the agreement is not satisfactory

4: reconsidering the hypothesis and modifying the theory on the basis of the new evidence available

- and on the basis of any new ideas that may have come up.

Of course in reality life is more complicated than this: the hypotheses are set up within a pre-existing framework of understanding that will be based on a cultural and temporal viewpoint; and a theory is made up of a complete set of interlocking hypotheses and assumptions, which are tested as a whole. Thus it will in general not be obvious what it is that needs altering to make a better theory than the present one. Nevertheless the power of observational tests underlies the examination and improvement of theories, whatever their domain of application, and is the basis of all our knowledge of the real nature of the world and the Universe.


Section 2.3: Scientific investigation

The scientific method is nothing other than the basic learning cycle just discussed, but applied in a rather systematic manner; for discussions of how science is carried out, see Chapter 1 of [6] or the Rutgers internet notes "Philosophy of Science - How do we know what we know?". The way it works out in the specific case of biology is nicely described in and for the case of astronomy see.

The power of science as a method of investigation arises from two features it has practised to perfection: firstly, the use of the analytic method, i.e. dividing a system into its parts and understanding how the parts work in isolation from each other; and secondly, the systematic application of quantitative analysis, based on accurate measurement and used in conjunction with measuring instruments of ever increasing accuracy, so that the regularities underlying nature are formulated as mathematical laws. These features have been employed together in formulating theories and devising precise experimental tests of the theories proposed, leading both to the ability to predict to great accuracy the behaviour of simple isolated systems, and also to considerable knowledge about the nature of the constituent parts of matter.

One of the remarkable features that has emerged is the amazing power of mathematics [12,13] in describing the nature and behaviour of matter; it is something of a puzzle as to why this should be so (see Barrow and Davies in [6]). It should be emphasized that this has become particularly apparent in recent times, due to the dramatic improvements in measurement technology: we can now measure times, distances, masses, and other properties of objects to incredible accuracy, due to enormous improvement in imaging and measurement techniques, often through use of entirely new processes (scanning microscopy, Nuclear Magnetic Resonance imaging, laser interferometry, CAT scanners, Charge Coupled Devices (CCD's), and so on).

In searching for scientific laws, we are looking for invariant behaviour common to many systems, providing a unifying explanation of different generic and specific cases. We do so by separating out their behaviour into a universal part, "laws of nature" applicable to all similar systems, and specific information determining the response of a particular system to those laws, usually in the form of "initial conditions" and "boundary conditions" specifying the nature and state of the specific system under investigation. Thus for example general laws of motion describe how any falling object moves; the particular place and speed with which a ball is propelled, together with a specification of wind conditions, determines its particular path and pace of motion. Experimental tests function by varying some of the initial conditions of the system, keeping the rest fixed, and then seeing if the response to these new conditions follows the universal pattern described by the laws we suppose are applicable to that situation. Thus we may release projectiles with various velocities from a tower, and verify if the way they fall complies with our theories of wind-resistance and gravity; the test confirms the theory if they move precisely as predicted (within the experimental error). By carrying out such tests, we have successfully been able to confirm that physical laws that do indeed describe accurately the behaviour of a large variety of physical systems, and enable us to predict their future behaviour with precision.

2.3.1 Different kinds of science

While the characterisation just given describes the nature of a large part of the natural sciences, including the fundamental sciences of physics and chemistry, it is important to realise there are other forms of "hard" science with major differences in their practice.

Firstly, there are the purely logical sciences, specifically logic itself and most of mathematics, which are not susceptible to experimental test or proof (The exception here is recent use of computers to determine the nature of mathematical systems, and in particular to examine the behaviour of chaotic and fractal systems, where some of mathematics has begun to take on aspects of an experimental science). Rather they are based on pure analysis and examination of the satisfactoriness of that analysis. The other sciences build their theories on the basis of the logical sciences (the analysis of physics is based on mathematics, for example).

Next there are the natural sciences, comprising the analytic sciences, as outlined above, and the integrative (or synthetic) sciences: for example ecology, where the emphasis is not on the behaviour of the parts of a system, but rather on the behaviour of a complex system made up of many interacting parts (the behaviour of each of these parts being susceptible to analysis by the analytic method). The applied sciences, such as engineering and computer science, fit into this category, and in many cases remarkable success has been achieved; for example we can now use computer simulations to predict the behaviour of aircraft before they have been made, on the basis of Newtonian dynamics and the theory of fluid flow. As in the case of the analytic approach, these understandings of how complex systems function are open to experimental test (provided we can isolate the system adequately from outside interference). One of the main differences from simple systems is that usually there is no way we can test all possible types of initial configurations of a complex system; we can only check its behaviour under a representative sample of initial conditions, and hope this gives us sufficient insight into its behaviour in the face of all the conditions that will be encountered in reality (for example we test the computers that control aircraft in a way we believe will adequately reflect the whole range of conditions they will encounter in practice; however we cannot test all conditions that might occur).

Clearly the synthetic sciences build on the analytic sciences, in that as we attain a better understanding of the behaviour of the components of a complex system, we are in a better position in our attempts to comprehend the whole. However they cannot be reduced to the analytic sciences: it is precisely the relations between the parts of a human body that enable the whole to function, and this cannot be understood by examining the parts in isolation. This is why physiology is of necessity an integrative science. Furthermore one should note that while the objects of study in the reductive sciences may often be identical to each other (each proton is identical to every other proton, similarly form electrons and neutrinos), in the case of the integrative sciences there will, despite major similarities, always be some differences (each human body is different from each other one; each ecosystem has its own individuality).

Finally there are the historical and geographical sciences, such as geology and astronomy, where we examine the nature, history, and origin of unique systems (a particular mountain range, the Earth, the Local Cluster of galaxies, and so on). Major examples are the investigation of the creation of the Solar System, of the history of continental drift on Earth, and of the evolution of life on Earth. In these cases, we cannot set initial data so as to repeat the situation that occurred in the past: experimental tests in the sense implied above are not possible, because we are concerned with specific events that have only happened once in the Universe. However we can on the one hand look at properties of similar systems or events, hoping they will help us understand this particular one (the issue of course being just how similar or different all these other examples are); and on the other hand we can today make observations of many kinds of data that tell us much about the specific historical event of concern, representing features that would be necessary consequences if our theory is correct (for example we can search for similarities in the DNA patterns of animals we believe to be closely related through evolution, or we can compare ages of different fossils as determined by measurements of radioactive decay).

Thus we can try to predict the results of observations that have not yet been made, on the basis of our current best theories about the specific historical event in question (we may predict that various rocks must be similar in South America and South Africa if continental drift is indeed true, and then go out and verify if this is so or not). We may also be able to measure present behaviour that tends to confirm our ideas about the past because it is the same process going on at the present time (we can for example measure the present rate of change of distances between the continents, and observe mutations presently occurring in population species).

All of this provides corroborative evidence which may be very convincing, but is still not the same as observing the unique course of events that took place in the past, finding out the effect of altering initial conditions at that time, or carrying out experiments that repeat the same course of events. The historical sciences build on the analytic and integrative sciences in that these give guidelines as to the kinds of behaviour to expect, and put strict limits on the kinds of things that could have happened in the past, assuming the fundamental laws were the same then as now. This basic assumption is to a certain extent susceptible to test.

The Sciences:

















Molecular biology



Table 2.1: A classification of the different kinds of sciences, with examples of each listed [Note that applied sciences and social sciences are not shown here]. The Natural Sciences and Historical sciences together comprise the observational sciences.

Table 2.1 shows this classification of sciences (apart from applied sciences and the social sciences), reflecting the relation of each to testing and confirmation. I use the latter word advisedly; it is not possible to verify any theory in the sense of proving without doubt that it is correct, but we can confirm it by providing more and more evidence that supports its correctness (for example no one seriously doubts that Newton's laws of motion adequately describe how a motorcar moves). The use of Bayesian statistics gives such confirmation a solid logical foundation [14-16]; in this approach we always regard knowledge as incomplete, but with each new bit of positive evidence adding to the previous evidence for believing a theory is true.

We can sometimes disprove theories, including historical theories, in a decisive way, when observations clearly contradict some of their predictions (for example, the finding of a human skull that radioactive dating proved came from the era of the dinosaurs would confound the present theory of evolution).

2.3.2 The invariant underlying nature

The point that now needs to be made, in the face of much recent writing emphasizing the sociological basis of scientific activity and suggesting that all scientific understanding is therefore culturally bound and relative (see e.g. [9]), is the efficacy of physical laws. Social and cultural issues do indeed play an important part in shaping science, for they help determine what is regarded as an important issue at any time, and therefore help shape what questions are asked by the working scientist; and to some extent they shape the kinds of theories proposed to provide explanation. However the fundamental point is that provided these questions and theories are then refined and developed appropriately to lead to true scientific tests, the answers obtained (in the logical and natural sciences) are not relative: on the contrary, they reveal some aspect of the working of nature that is universal: it is independent of the time and place where the experiment takes place, and of the cultural and sociological nature of society. Irrespective of these factors, when I release a weight it drops to the ground; water is always made of hydrogen and oxygen; electric waves propagate at the speed of light; and so on. There is a trivial relativisation in terms of the language used to describe reality, for example "oxygen" and "sauerstoff" are the same thing; I assume we are able to handle such issues of multiple representation without getting misled by them. Indeed the functioning of the world around is rigorously determined by the laws of physics and chemistry, whatever we may wish or do, and no matter what our life view may be; inter alia these laws determine the functioning of our own bodies and minds, the physiological nature of that functioning also being completely determinate (we breath oxygen; DNA determines our genetic inheritance; calcium and sodium ions are the basis of signal transmission in our nervous systems; and so on).

This is the triumph of natural science: it has been able to locate an underlying reality that is invariant and universal, despite the differing social and cultural positions from which different scientists operate. The impression that relativity theory has done away with such a reality is incorrect: that theory emphasizes both the relativity of different observer's views of that reality, and the possibility of characterising its nature in an invariant way [31]. We expect motor cars, television sets, refrigerators to work in a specific way that results from their design; and they do so completely reliably (unless defective) precisely because the laws of physics govern their behaviour in a reliable and repeatable way.

While the implications of the natural sciences may be absolutely firm in many cases, the description we use may not be rigidly specified. Indeed there may be different viewpoints and even different mathematical descriptions that describe the same processes equally well, and give the same results. For example, there are different ways of formulating Newton's theory of gravity: in terms of a potential (the field view), in terms of action at a distance (the force view), or in terms of a variational principle (a Lagrangian or Hamiltonian view). In the case of quantum mechanics, apparently different mathematical descriptions (the Heisenberg, Schrödinger, and Feynman descriptions) have been shown to represent the same physical behaviour: they are equivalent descriptions, although this is far from obvious. In these cases we may prefer one or other of the alternative views of the same phenomenon on cultural or aesthetic grounds; this will in no way alter the solid nature of the predictions obtained from them.

Furthermore it is important to realise there are simple effective theories we may employ, even when we know that they are fundamentally wrong, because they may indeed be effective in a specific domain of application. Thus for example even though we know the rotation of the Earth is responsible for the daily appearance and disappearance of the Sun, it is still useful in everyday life to talk about the Sun rising and setting (in essence a completely wrong theory - but still a very practical viewpoint in daily life). Similarly we may apply the Newtonian theory of gravitation to predict the motion of planets in the Solar System unless we need very high accuracy predictions, when we need to use the correct theory (Einstein's theory of gravitation); and we can apply the Galilean theory of gravitation in engineering construction on Earth, even though it derives form the more fundamental Newtonian theory. Thus such effective (non-fundamental) theories are very useful; the real question is a meta-question: when can we usefully apply such theories, and when not?

2.3.3 Criteria of choice for theories

The situation is not so clear cut in the case of the historical sciences (as defined above), such as geology, evolutionary theory, archaeology, and astronomy. Here there may well be a cultural or sociological bias influencing the conclusions we derive (as well as the description we use), for in the face of our inability to perform the experiments we would like to carry out, we have to make assumptions determining what kind of theory we regard as reasonable, and the shape of what we find is biased by these assumptions.

Thus a key issue then is what are appropriate criteria for a satisfactory theory, that can help us choose amongst possible alternative explanations. The primary candidates are,

Criteria for scientific theories:

* simplicity - the Occam's razor idea that one uses the simplest possible theory that can accommodate the facts;

* beauty, on the face of it a very subjective criterion, but there is remarkably good consensus about it in many cases;

* prediction and verifiability - the ability to confirm the theory by a variety of observations or tests; in particular,

- verified predictions of a new kind provide major support for correctness of a theory.

- the converse is the Popper criterion that a good theory should be clearly falsifiable by experimental test;

* overall explanatory power and unity of explanation, in particular congruence with the rest of our current body of knowledge.

These are the key requirements for a "good" theory; the problem is that in general these criteria will not agree, and differing emphasis on which of them is important may lead to different choice of theory (see [10] for a discussion of the case of cosmology).

Nevertheless in many cases we may attain a high degree of certainty because these criteria concur, selecting uniquely as preferred a theory with high explanatory power that also fits into the body of established theory in a satisfying way, particularly when it makes new predictions (such as the existence of anti-matter, the bending of light by the sun, or the transformation of matter into energy) that are then verified.

2.3.4 Reasonable certainty in historical sciences

There will however be historical situations where we are destined to always remain in doubt about the true course of events, because of the fragmentary nature of the evidence available to us; for example many cases in archaeology, and many of the details of evolutionary history. Thus we will attain various degrees of certainty, according to the quantity and quality of evidence available to us.

However in particular cases, there is a vast interlocking array of facts that are explained in a unitary way if we adopt one particular explanation of a complex phenomenon, but remain a set of disconnected features that have common aspects by pure chance if we do not. By marshalling evidence in this way, we can be virtually certain for example that continental drift and evolution did in fact take place, and that carbon dating gives correct estimates of ages of archaeological finds. It is possible to put forward logically consistent alternatives that explain the same historical facts differently, such as the so-called "creationist" view; the problem is their incongruence with the rest of scientific knowledge, together with the small number of features they explain. They provide a consistent scheme, but considered as a scientific scheme of interpretation and understanding, it is one of narrow scope and small vision, with low integrative power. As far as possible we demand consistency with present day scientific theory and understanding, together with the requirement of broad explanatory power. This will in many cases lead to a unique interpretation of specific events that fits in with present scientific theory in a mutually reinforcing way, so that an unrivalled interpretation is both attainable and in effect required: for example, the evolutionary principle not merely becomes an explanation of past events but becomes a central feature and profound organising principle of biological theory, explaining also many events occurring today.

It is interesting to view this whole discussion from the experience gained in law courts, for the examination of evidence as undertaken there is nothing other than applied historical research; and it is taken most seriously because people's freedom and livelihood hang on the outcome of the verdict. The crucial point is that in many cases we believe that, despite all the pitfalls, we are indeed able to arrive at a verdict "beyond all reasonable doubt". The cautionary remark is that sometimes such a verdict is wrong.


Section 2.4: Key elements of a broad approach

Whatever the field of concern, the basic learning cycle discussed above is the way we obtain and extend our understanding. Employing this method consciously will increase its effectiveness. This demands

2.4.1 Setting the scene: Reflection, Observation, and Openness

The start is a reflection on issues, principles, and data, with an openness to possibilities, the readiness to search out grains of truth that may be hidden in alternative viewpoints and a willingness to test them for possible validity. Thus a key question one can ask oneself is, What am I prepared to question, and what am I not prepared to question?

The answer lays down the fundamental parameters within which one is willing to learn. Those issues that one is unprepared to question are those domains where one has chosen to proceed on the basis of preconception and dogmatic assertion, rather then reflective investigation. As an example, Einstein's dramatic progress in understanding space and time came because he was willing to query the nature of space measurements, of time, and of simultaneity - which everyone else did not question because they took them for granted.

The adoption of ideologies of various kinds is a common way of protecting oneself from questioning what one holds dear. The aim of the approach suggested here is that one will try to avoid having a closed mind. This means in particular, emphasizing freedom of information and freedom to support other ideas than the current dogmatism; indeed that the basis of progress and understanding is an enquiring atmosphere [11].

In developing a consistent theory based on this approach, one is engaged in integrating the possibilities considered into a coherent and logical scheme; our logical and creative skills come into play as we fashion a satisfactory whole. A fundamental point here is

2.4.2 Identifying Main Issues and Causal effects

Next is identifying the important issues and concepts, naming them, and characterising them. We can then show how they relate to each other and to the extant wider body of theory and knowledge.

The point is not only to state what is fundamental and what is less important (from a causal viewpoint), but also to consider what may have been left out, perhaps because it is so obvious that it is taken for granted and therefore not taken into consideration.

Having set up a theory (or set of competing theories) our task is to separate the wheat from the chaff by suitable testing. The key point here is

2.4.3 Testing

Next is determining valid methods of testing theories and of assessing the results of the tests. This will vary greatly from area to area. We have to ask, What is acceptable data for this area and how is its quality assessed? We can then review existing data, and if possible, set up new experiments or observations to test the theories.

If some feature regarded as key by some is not admitted as data at all by others, agreement cannot be reached. However often the disagreement is not over the general admissibility of a class of data, but the quality of specific data. These problems apply particularly to historic records.

2.4.4 Evaluation

Finally, having analysed the competing theories and considered the evidence for and against them, we need to use our chosen criteria for good theories to choose between them, asking how good is each theory relative to alternatives, in terms of our criteria?

2.4.3 Real Observations

Two final comments relating to realistic observations:

Firstly, we must realise that in real-life situations, serious counter-evidence to some theory does not often by itself lead to dropping a theory, but rather to investigating various perturbing influences we have not taken into account so far (the wind, a temperature gradient, an electric field, that may have affected the path of a falling object; an unknown person entered the room and interfered with the murder scene before the deed was discovered). The point here is that even if we are investigating identical systems (say electrons), each experiment is in fact an individual experiment that is in detail different from every other experiment. We have to take this difference into account in interpreting the experiment.

Secondly, whenever observations are made in support of a theory, we should always be aware not only of possible distortion of the data but also of the various possible selection effects that might be in action, distorting the appearances of what happened by effectively preventing a whole segment of data getting through (for example, they determine what galaxies we detect in photographic plates of distant regions of the sky, because many are too faint to be detected). Thus a particular question we need to always ask is, What is the information that is not getting through? Furthermore, as emphasized in the famous Sherlock Holmes comment on the strange incident of the dog in the night (the dog did not bark), The absence of a signal may convey vital information. But we must also remember that absence of evidence is not evidence of absence. Just because we can't see it does not mean it is not there (cf. the celebrated case of the preponderance of dark matter in the universe).

Overall, the essential feature advocated here, in line with one of the major movements in methodology of the last half-century, is that our concern ultimately must be an emphasis on process rather than the particular presently available data and knowledge; for if the process is right, then in due course errors in our understanding will be corrected. Thus

* The basic need is a process of learning that continually checks theories and models for incongruities and problems in all possible ways, and corrects the errors found.

This chapter has briefly set out the nature of such a process - which is nothing but the scientific method. Its implications will be developed in the following Chapters in considering the nature of the physical Universe.


Section 2.5: The basic understandings attained

The rest of the book gives summaries of understandings attained by this approach in various important areas. Here we preview what follows by giving the following summary of the main findings of science.

2.5.1 The Main Themes

1: Everything is changing

Contrary to our expectations, we have discovered that the universe is changing, stars evolve, the mountain chains and continents on earth are changing, and new life forms have come into being on earth while others have become extinct.

2: The way things are changing can be understood as the result of a regular relation between cause and effect which can be determined by scientific investigation

The changes do not happen randomly, but result from systematic and reliable causal relationships. We can determine the nature of those causal relationships by careful observation and experiment.

3: The relation between cause and effect can be understood as due to the interplay between (i) an unchanging constant part (physical laws) and (ii) variable circumstances (boundary conditions and initial conditions) that will lead to different outcomes even though the underlying laws of behaviour are the same.

The causal relationships can be characterised as the interplay between chance, the contingent set of circumstances that just happened to be, and necessity, the inviolable set of physical laws that characterise how all matter behaves - both living and non-living.

4: The behaviour of complex systems, including life, is grounded in the behaviour of the fundamental particles that are their ultimate components.

Complex systems can behave the way they do because they are made up of interacting simpler systems. Laws underlying the micro-behaviour, such as energy conservation, result in limits on how macro-bodies can behave, for example the First and Second laws of Thermodynamics (which follow from energy conservation at the micro level). However their structural relations also give them higher level behaviour:

5: Complex systems derive their essential properties from the hierarchically ordered structural relationships between their component parts that are set up by their physical structure.

It is these basic theses that will be explored and illustrated in what follows.

2.5.2 Summary Tables

There are three main kinds of tables in the Chapters that follow.

Firstly, there are tables setting out the hierarchical relationships that exist between components of the various structures around us. Understanding these hierarchies is an essential component of understanding each subject.

Secondly there are tables setting out some of the vast variety of objects that have come into being as a result of these relationships - the specifics of the world as we know it, as discovered by scientific investigation. This enormous variety of existence partly reflects the `chance' component of existence - giraffes happen to be because of the vagaries of evolutionary development. But they also exist because they are allowed to exist by the underlying laws of physics and chemistry.

Thirdly there are tables summarising important invariant underlying principles that have given rise to this variety of existence - that is they summarise essential features of the `necessity' component of existence. The contrast between the relatively few basic principles and the vast variety of resulting structures that is one of the fascinating features facing one in a comprehensive view of the nature of science, such as is attempted here.

Finally there are some tables setting out theoretical options in cosmology.

As noted above, these representations are all under review: they may be improved as comments are received.


References for Chapter 2: Understanding the Universe

Philosophical analysis is discussed in

[7] T Nagel: What does it all mean? A very short introduction to philosophy (Oxford University Press, 1987)

[8] J Hospers: An introduction to philosophical analysis. (Routledge and Keegan Paul, 1959)*

(and see also [1]). Philosophy and science is discussed in

[9] W H Newton-Smith, The Rationality of Science (Routledge, 1991)*

with the special issue of philosophy and cosmology presented in

[10] G F R Ellis: "Major Themes in the Relation of Philosophy and Cosmology", Memoirs Italian Astronomical Society, 62: 553 (1991).

The learning cycle, and organisational structure conducive to learning, is discussed in

[11] G F R Ellis: Organisation and administration in a democratic era} (Book draft, University of Cape Town, 1991).

The nature of mathematics and mathematical descriptions of nature is discussed in

[12] L A Steen (Ed.): Mathematics Today (Vintage Books, 1980)

[13] M Kline: Mathematics: The loss of certainty (Oxford University Press, 1980)*.

Use of Bayesian statistics in analysing theories is presented in

[14] H Jeffreys: The Theory of Probability (Oxford University Press, 1939)*

[15] R T Cox: American Journal of Physics, 14: 1 (1946)*

[16] A J M Garrett: Ockham's Razor. Physics World (May 1991), 39-42.