The Fabric of Reality
From Simulism
Author: David Deutsch
Summary:
The Fabric of Reality is a highly ambitious book, and claims to be about explanation. It begins by discussing the series of scientific, philosophical and technological advances which have taken us to the brink of being able to create a "Theory of Everything". For Deutsch, this "Theory of Everything" is not meant in its usual sense, that of a fundamental scientific theory which merges all the fundamental forces (i.e a theory of quantum gravity) providing a mathematical basis for physics, but rather a theory which has profound explanatory power, and addresses questions about existence, being and the fundamental nature of reality.
At the start of this investigation Deutsch examines four areas of human endeavour: quantum physics, epistemology, computation and evolution. He seeks to find profound and diverse connections between each of these areas, and from these connections reveals a fabric of reality in which time, mathematics and our usual interpretation of reality are all shown to be illusory.
For a detailed discussion of the ideas within the book, see this Commentary on Deutsch's Fabric of Reality
[edit] Chapter Summaries
[edit] Chapter 1 - The Theory of Everything
Scientific knowledge is shown to be primarily about explanations, rather than simply predictions. The latter are important in conducting tests which allow us to discriminate between competing hypotheses that have already passed the test of being good explanations. New theories supersede old ones and knowledge becomes broader, encompassing more areas which we can explain, and deeper, in that our theories are more general and explain more. In this contest, depth appears to be winning out, and Deutsch argues that we are moving towards a state where one person could, in principle, understand everything. This is because our deepest theories are converging, and have so many integrated connections that they can only be truly understood when considered in concert with one another. If we do this, it will provide us with a unified theory about the very fabric of reality, a "Theory of Everything". Deutsch argues that this will be far more than a theory about particle physics, because the Fabric of Reality does not simply consist of those reductionist elements which constitute space, matter and time, but include as necessary ingredients, life, thought, knowledge and computation. The Theory of Everything, according to Deutsch, will therefore be based on Quantum Physics, Epistemology, Computation Theory and The Theory of Evolution.
[edit] Chapter 2 - Shadows
This chapter discusses a number of experiments involving interference patterns which occur when light is shone through a series of slits. These experiments lie at the heart of a philosophical debate in quantum mechanics about what the physical world is like at its most fundamental level. The traditional explanation for these interference patterns is that light is wavelike in nature and that the wave interferes with itself to cause illumination (where the peaks of the wave coincide), and shadows (where the peaks of one wave are 'cancelled out' by the troughs of another). However, as the experiments produce the same result even when the intensity of light is reduced to the emission of single photon, Deutsch claims that such explanations cannot, in principle, be a valid description of reality, as it is not possible for a single particle-like object to interfere with itself; therefore we need an alternative approach. He proceeds to argue that the photons are behaving 'as if' they were being interfered with by invisible counterparts - 'shadow photons' . In doing this Deutsch follows the argument of Hugh Everett in the 'many worlds' interpretation of quantum mechanics, that these photons actually exist, but in alternate universes, and that what is actually happening in the experiment is that photons in these different universes are interfering with each other to produce the observable patterns of light and shadow. We are left with the conclusion that the world as we see it through our senses is only part of all that exists, and that physical reality consists of not one universe, but a seres of parallel universes, making up the multiverse.
[edit] Chapter 3 - Problem Solving
This chapter discusses the nature of scientific reasoning, arguing that in fundamental areas of science, observations of ever smaller, more subtle effects are pushing us towards momentous conclusions about the nature of reality. Deutsch argues strongly in this chapter against logical positivism and inductivism. He claims that positivism, merely confims that a theory works; it does not help in offering explanations or insights into why it functions, and as such is impotent. Inductivism, he claims, is a sham - no one actually induces anything from a series of observations; theory is formed by modifying existing theories and looking for explanations which make sense, are logically consistent and accord with other prevailing theories. Deutsch claims that inductive reasoning is therefore invalid, as it is impossible to extrapolate from observations unless one already has an explanatory framework for them. In this case, the prediction is made on the basis of the theory, not the observations. This refutation of inductivism depends on recognizing that science is a process not of deriving predictions from observations, but of finding explanations. We seek new explanations when we have a problem with our existing ones, for example where our predictions are not confirmed by observation. This then leads to a problem solving process, in which new theories begin as unjustified conjectures, which are then criticised and compared to criteria intrinsic to the problem. Those that fail at this stage are abandoned. Surviving theories become the prevailing paradigm, and are used to make predictions. Provided that the predictions are borne out by observations, such theories can remain the paradigm for many years. In time, these theories may may themselves become problematic - for example improved observations may lead to questions of accuracy and detail; new theories in other areas may be based on fundamentally different principles which contradict with the theoretical basis; this leads us to seek better explanations. Deutsch notes that the whole process resembles the process of biological evolution.
[edit] Chapter 4 - Criteria for Reality
In this chapter Deutsch addresses the problems of how we decide what we accept as reality, and what are the criteria for doing so. He looks at two related cases - that of the the displacement of the Geocentric Theory by the Heliocentric Theory, and that of the refutation of solipsism. In the first case, he argues that Galileo came into conflict with the Inquisition, not because they refused to accept that the Heliocentric Theory of the universe gave more accurate results, but that they believed that they had fundamental theological reasons for believing the Geocentric theory was correct view of reality, and that the orbits of the planets were merely arranged 'as if' the heliocentric theory was correct. In the case of solipsism, the view that only one mind exists in the universe (presumably yours), Deutsch argues that in order for this to be valid, the universe must be arranged to behave 'as if' there are other minds and other objects. Therefore although solipsism and related doctrines are logically self-consistent, they can be refuted by taking them seriously as explanations. In doing this they prove to contain indefensible over-elaborations of realism. Clearly here Deutsch is using a version of Occam's Razor, not to decide between competing hypotheses, but to manufacture a criterion for an explanation of reality. Real entities, Deutsch argues, should behave in a complex and autonomous manner, and it is this we can take as a criterion for reality. Bowell relates in The Life of Johnson, how Dr Johnson attempted to refute the doctrine of solipsism by kicking a large rock, and saying: "I refute it thus!". The point is, that the rock was complex and autonomous. You kick it, and it hurts. It kicks back, and therefore it exists - any claim that it does not leads to a convoluted reasoning which is an over-elaboration which would result in the claim being rejected. Deutsch notes that scientific reasoning which uses observation to distinguish between competing explanations can give us genuine knowledge about reality.
[edit] Chapter 5 - Virtual Reality
Here Deutsch begins his discussion of what he calls 'image generators', where an image can be any specifiable sensory input - taste, smell, vision, touch or sound. Image generators are physical objects which affect our senses; a virtual reality generator is a specific version of these which attempts to convince the user that the sensory inputs provided are something other than what they actually are. Deutsch uses the 'kick back' criterion above to show that such versions of reality actually form a real experience which allow the user to interact with a simulated environment. There is no contradiction here. The environment 'is' real; it is just that its true form may not actually coincide precisely with what is perceived. Deutsch goes on to discuss the repertoire of a virtual reality generator, dividing it into Logically Possible Experiences such as piloting an aircraft, and Logically Impossible Experiences such as factorizing a prime number. Logically possible experiences may be subdivided further into those which are physically possible and those which are physically impossible. In addition, we can distinguish between internal and external experiences - internal being related to feelings and mental states, rather than physical sensations. In the discussion of all these categories, Deutsch concludes that any virtual reality generator, running any programme in its repertoire is rendering some physically possible environment (even if that is merely another VR generator rendering a particular environment). Note that it may be rendering other things as well, including many physically impossible environments; however Deutsch goes on to point out that only those physically impossible environments which are indistinguishable from physically possible environments can in fact be rendered. A case in point is 'faster than light travel'; if a VR generator rendered this, our perception of this would be that we were in a fictional scenario, and we would 'suspend disbelief' for the duration of the experience, just as we do when watching Star Trek. We exercise our imaginations in doing this, and to Deutsch, imagination is yet another version of virtual reality. In fact all forms of reasoning and thinking can be reduced to VR; indeed a scientific hypothesis is merely a VR rendering of the real world. The fact that VR is possible is an important fact about the very fabric of reality.
[edit] Chapter 6 - Universality and the Limits to Computation
In this Chapter, Deutsch discusses how the limits to computability imposed by Godel's Incompleteness Theorem affects the Virtual Reality rendering process. In order to do this, Deutsch invents the notion of a CantGoTu environment (named after Cantor, Godel, and Turing), using Cantor's diagonal argument to construct an 'impossible' Virtual Reality which a physical VR generator would not be able to generate. The way that this works is to imagine that all VR environments renderable by such a generator can be enumerated, and that we label them VR1, VR2, etc. Slicing time up into discrete chunks we can create an environment which is unlike VR1 in the first timeslice, unlike VR2 in the second timeslice and so on. This environment is not in the list, and so it cannot be generated by the VR generator. Deutsch then goes on to discuss a universal VR generator, which as a physical device would not be able to render all possible environments, but would be able to render those environments which can be rendered by all other physical VR generators. He argues that 'an environment which can be rendered' corresponds to a set of mathematical questions whose answers can be calculated, and discusses various forms of the Turing Principle, which in its inital form refers to the fact that it is possible to build a universal computer which can be programmed to execute any computation that any other machine can do. Attempts to capture the process of virtual reality rendering provides us with a version which states: "It is possible to build a virtual-reality generator, whose repertoire includes every physically possible environment". In other words, a single, buildable physical object can mimic all the behaviours and responses of any other physically possible process or object. This, it is claimed, is what makes reality comprehensible.
[edit] Chapter 7 - A Conversation about Justification
This chapter takes the form of a dailogue between David (Deutsch) and a crypto-inductivist, whom Deutsch defines as "Someone who believes that the invalidity of inductive justification is a problem for the foundations of science". In Deutsch's words, although such a person accepts the Popperian view that induction per se is not a valid part of the scientific process, nonetheless, they believe that there is a gap between observing and hypothesising which needs to be filled, if not by the traditional notion of induction, then by something else which is similar. Deutsch sets his dialogue at the top of the Eiffel tower, and the dialogue itself concerns whether or not should one of the protagonists jump he or she will 'float to the ground'. As the dialogue proceeds, Deutsch attempts to demolish those arguments which appear to rely on induction, such as given a series of observations that X will occur, making the prediction that X will continue to occur under the same circumstances. He does this by demonstrating that in all cases, the fact that X has occurred leads us to hypothesise about the nature of X, and why it occurs. A new theory is created, not on the basis of extrapolation from a series of observations, but by an extension of existing theories, and the insight into the situation it provides. In doing this we are rendering reality in our own minds. As a result of this rendering, we can make deductions from the understandings created; this is how induction appears to work, but it is an illusion; in order for the rendering to be accepted as a viable model of reality, there needs to be a process of argument, justification and testing. We are effectively producing a new model of reality rendered in a different manner, in order to understand reality better. Hence the problem of induction is removed, and replaced with one of justification. Deutsch argues in the latter stages that logical reasoning itself is a physical process, and that it is an interesting fact that the physical universe admits processes which create knowledge about itself, and it is the Turing Priciple formulated in the last chapter which allows us to form such accurate theories about reality.
[edit] Chapter 8 - The Significance of Life
In this chapter, Deutsch turns to the subject of life itself, noting that scientific progress has appeared to refute the idea that life is a fundamental phenomenon of nature. He claims that modern biology has to some extent confirmed this, by explaining living processes in terms of molecular replicators (i.e. genes), whose behaviours are governed by the same laws of physics that apply to inanimate matter. However, Deutsch makes the fascinating claim that life is in fact a type of virtual reality generator, aimed at creating an organism which is a physically manufactured environment for its genes. These genes embody knowledge about the environment in the same way that a VR program embodies knowledge of the world that is being simulated, and genes are susceptible to the Turing principle described above, and it is this fundamental principle of physics with which life is associated. One of the avenues which Deutsch goes on to explore is how differences in gene segments which are knowledge-bearing and non-knowledge bearing might differ across nearby universes in the multiverse. Deutsch concludes that DNA segments which are knowledge-bearing would need to be consistent across the multiverse in order to create stability of form; junk DNA on the other hand could be very different in nearby universes, without causing very much difference to the rendered organism. When viewed in this manner, life becomes a consistent and significant process tying together many universes across the multiverse, and is a feature occurring on the largest scales of both space and time. Conversely, Deutsch argues, any consistent feature which exists across the multiverse must embody some form of knowledge about the environment. As the future unfolds, life will therefore determine the future behaviour of stars and galaxies; Deutsch claims this is because at the largest scale, regular structures can exist across universes only where knowledge-bearing matter such as brains or DNA gene-segments exist.
[edit] Chapter 9 - Quantum Computers
In this chapter, Deutsch argues that the laws of physics, via the Turing Principle permit computers to render every physically possible environment without using impracticably large resources. Universal computation, as defined by the principle, is therefore not merely possible, it is also tractable. One method of doing this, would be to harness the power of the multiverse, via quantum phenomena which allow interactions to occur across vast numbers of parallel universes simultaneously. However, in doing so, it might mean that quantum phenomena as a physical process may not actually be able to be rendered efficiently within a single universe. However, Deutsch reasons that a strong form of univerality still holds as quantum computational processes can effciently render every physically possible quantum environment, even when vast numbers of universes are interacting. Deutsch is adamant that such devices are not science-fiction. Quantum computers have been shown (via Shor's and Grover's algorithms) to be capable of efficiently solving mathematical problems such as factorization which are classically intractable, and capable of implementing cryptographiuc methods which are classically impossible. The way in which this occurs is that quantum processes effectively deny the 'classical' nature of a single 'bit' of information as a 1 or 0; in some sense treating the bit as a probablility distribution or a fuzzification of the information, across the multiverse. Deutsch concludes that quantum computation is therefore a qualitatively different way of harnessing nature.
[edit] Chapter 10 - The Nature of Mathematics
Here Deutsch discusses the status of mathmatical entities and mathematical reasoning. He concludes that mathematical entities are complex and autonomous, in that independent investigators examining mathematical structures can carry out detailed forensic examinations on them, and arrive at similar conclusions; in other words, they have substance and meaning. In this way, mathematical entities are part of the fabric of reality. There exist logically necessary truths about these entities, and these comprise the subject matter of mathematics. However, as a result of Godel's theorems, such truths cannot be known with certainty, and what we take as mathematical proof does not in fact confer certainty on the conclusions. Mathematics is axiomatic, and dependent upon logic; if the axioms are not an exact match for the reality they seek to model, or if our logical process themselves cannot be shown to be consistently valid, then the entire mathematical enterprise rests on shaky foundations. Deutsch therefore argues that particular forms of proof depend on the truth of our theories about the behaviours of the objects upon which and with which we carry out those proofs. This means that all mathematical knowledge is, in essence, ultimately derivative, depending only on our knowledge of physics. Mathematics can therefore in some sense be regarded as a rendering of reality, and the set of mathematical truths which are comprehensible is precisely those which can be rendered in virtual reality. These truths form an infinitessimal minority aof all possible mathematical truths; the rest are incomprehensible. This can be argued as follows: There are 10^1000 one-thousand digit numbers. Only a small fraction of these can ever be written down, as if we started writing one per unit of Planck time at every point of the known universe (at the Planck scale), it would take approximately 10^500 years to enumerate them all. Our current most optimistic estimates for the future of the universe as a 'usable' environment is about 10^100 years at maximum. This means that even at the most extreme, we could only ever enumerate 1/10^400 of all such mathematical objects; the rest must forever remain unknown, and therefore incomprehensible. However, Deutsch claims that the incomprehensible mathematical entities, for example the CantGoTu environments, also exist, because they appear inextricably liked to our explanations of the comprehensible ones.
[edit] Chapter 11 - Time: The First Quantum Concept
In the first part of this chapter, Deutsch addresses the problem of how we view the flow of time. The classical model, that of a moving object represented as a series of snapshots of 'moments' , moving along a timeline, is shown to be deficient. It is argued that there is no single 'present moment' in the model, and nothing is actually moving; consciousness exists at all moments. Deutsch further contends that we do not actually expereince time passing; what we do experience are differences between our present perceptions and our present memories of past perceptions. We interpret these as evidence that the universe changes with time, and conclude, incorrectly, that our consciousness actually moves through time. Deutsch goes on to consider a spacetime which is fixed and immutable, but one in which the 'moments' are randomly shuffled, and concludes that it would not matter to the inhabitants, who would anyway make sense out of it, but it would be inexplicable to any one observing from 'outside'. Such a shuffled spacetime could in principle be reconstituted by invoking the laws of physics to make predictions about the positions of objects, selecting the appropriate 'moment' to succeed any particular frame. However, Deutsch argues, we cannot view this as simple causality of X causing Y, since in the multiverse, we have a large number of co-existing and interacting spacetimes, in which some of the spacetimes have event X followed by event Y, but in some universes, event X and event Y exist independently of one another. A causal statement such as 'if X, then Y', assumes that X may or may not exist; however in our universe, it is one or the other, it is only when we view the multiverse, that such counterfactuals actually occur. Deutsch shows that we can retain a quantum version of causality only by viewing it statistically, in terms of the proportion of universes where X and Y co-exist. This now leaves us with a problem: in the multiverse, other times are now simply seen as special cases of 'other universes', and other times are distinguished from other universes only in that other times are especially closely related to our 'now' by the laws of physics. In this reading of time, we can see that predictions can be made, and that in some of the multiverses, where X is a stable feature, predictions that Y will occur will be actualised in some of these universes, whereas in others they will not. This means that to an observer within a particular universe, the future would appear to be open, as in some universes 'nearby', events can unfold in different ways. To conclude: time is therefore not a sequence of moments, nor does it flow, but certain events can be seen to exhibit a kind of causality, with the future open, and the past fixed. All moments are physically real; the whole of the multiverse is physically real - but nothing else is.
[edit] Chapter 12 - Time Travel
Here Deutsch speculates on the possibility of travel backwards and forwards through time. Future-directed time travel is not seen as problematic, as there are many physically possible ways (such as hibernation, relativistic time-dilation) in which this might be achieved. In considering past-directed time travel, Deutsch examines whether VR rendering could be used to simulate time travel. Ignoring the inherent difficulties in determining with sufficent accuracy how to render historical events, such virtual-reality renderings would actually be susceptible to the same paradoxes that actual time travel would invoke, for example the grandfather paradox, in which a time traveller kills his own grandfather, preventing himself from being born. In addition, such VR renderings might have to cope with the user deciding to travel back repeatedly to the same moment in time, and thus rendering a copy of the user in the VR setting, while simultaneously responding to the user. Deutsch claims that the way out of such paradoxes is that a time traveller, could appear to change history so that an event occurs which did not previously occur in his timeline, but that in fact he affects the timeline in another universe, one in which the copy of himself does not travel back into the past and perform the equivalent feat. This also resolves the greatest paradox of all, that of the time traveller who returns to the past with the instructions on how to build the first time machine - effectively creating knowledge from nothing. There is a parallel evolutionary paradox here where a time traveller deposits a fragment of life on the ancient shores of some prehistoric ocean in a time before life itself had been created. These situations are paradoxical, becuse we normally expect that knowledge grows and evolves over time, exactly as life itself does. The creation of 'something from nothing' is not normally within our experience. Deutsch claims that activities such as growing and evolving can be seen in terms of computational processes, and speculates that if time travel between parallel universes could occur, it would be possible to trade information between universes. This would allow, for example, highly sophisticated virtual reality generators to be built which would have access to resources and information across a large number of universes, and enable computations to be performed with much greater accuracy that could be achieved on any computer within a single universe. As we have seen in earlier chapters, living beings can be viewed as VR renderings of genetic code; a transfer of genetic information between parallel universes could similarly change the course of evolution in a significant manner.
[edit] Chapter 13 - The Four Strands
Deutsch begins this chapter by discussing Kuhn's notion of a paradigm shift, and while he accepts the general principle, he takes issue with the notion that science clings onto outdated ideas, simply becuse they happen to be the prevailing paradigm. However, in the case of the four main theories outlined in the book, in each case, these ideas have been accepted as the mainstream view, but have not really filtered through into the consciousnesses of scientists. Quantum physicists, Deutsch claims, cling to the Copenhagen interpretation through pragmatic instrumentalism, i.e. because the theory works, they do not question the foundations of why it works. Similarly, while philosophers and scientists have appeared to accept Popper's epistemological analysis as the paradigm, it does not actually constitute the world view of many people. Similarly, while Darwinian explanations of evolution are accepted by all but a handful of people, objections to Darwinism as an explanation of the phenomena of the biosphere are numerous and widespread. Currently it is Dawkins' version of Darwinism that is the prevailing paradigm, but, Deutsch notes, many biologists and philosophers are still haunted by the feeling that there is some fundamental gap in the explanation. Finally the Turing principle has hardly ever been seriously doubted (Roger Pensose being a notable exception) yet the proposition that artificial intelligence is possible in principle is by no means taken for granted. In essence, Deutsch argues, there are gaps in each of these theories; however these gaps can be filled by taking the theories in unison; the theories are complementary to each another, and when viewed in this light form a unified view of the fabric of reality.
[edit] Chapter 14 - The Ends of The Universe
In the final chapter, Deutsch attempts to bring together the various themes previously explored within the book. He stresses that the fabric of reality which he is attempting to describe is not that of fundamental physics alone, but must include the Turing Principle just as much, say, as the conservation of energy and the laws of thermodynamics. However, this principle, which asserts that it is possible to build a machine to perform every physically possible calculation, has an emergent character, arising from the properties of high-level complex entities, and not from subatomic processes. Similarly, in our understanding of epistemology and evolution, we would should expect that knowledge and adaptations which occur as structures spanning large numbers of universes, will be expressed as laws relating to the structure of the multiverse as a whole. There is therefore a fundamental issue in taking any one of the four theories independently, which is that each one is essentially reductionist in nature; in addition, three of the four strands would appear to rule out human beings from the fundamental levels of explanation. In this context Deutsch goes on to discuss Frank Tipler's Omega Point theory, the notion that in the final minutes before the 'Big Crunch', the universe's oscillations can provide sufficient energy and sufficient computational power to do just about anything: to create an intelligence which is virtually omnisicient, to run an infinite number of simulations or to create an omnipresent entity. Deutsch reasons that at very late stages of the universe, the omega point could, undoubtedly embody inconceivable wisdom and creativity; however it will also contain rival conjectures, errors, criticisms and refutations. At the very instant the universe ends, everything that is comprehensible may have been understood, but at every finite point prior to this, our knowledge will be incomplete and riddled with errors. At this moment in time, Deutsch concludes, the current best hope for moving towards this ultimate goal of achieving complete understanding is to hold a unified world view based on a synthesis of Everett's multiverse, Popperian epistemology, the Darwin-Dawkins theory of evolution, and Turing's theory of universal computation. This is not the ultimate theory, merely one which wlll allow us to determine a better method of viewing the world, and on which we can base future, more sophisticated theories which allow us better to understand the fabric of reality.
