The Porter Zone

Philosophical musings and more

Category Archives: Science

The illusion of time and space


We are in the habit of assuming that the way the world appears to us is the way that it is.  Folk epistemology (and, rather regrettably, some academic epistemology) tends towards the Platonic notion that when we perceive an object then we, as it were, directly apprehend its true nature in our minds.  And yet any amount of evidence suggests that this is not true.  People who are red-green colour-blind cannot distinguish red from green, but can distinguish shades of green that those of us with trichromatic vision cannot, and none of us can equal the amazing mantis shrimp with their sixteen separate colour receptors.   We perceive not the world as it is, but those parts of it that are mediated to us via our senses.

Thus it is only reasonable to say that the world as we perceive it is born of an interaction between whatever it is that is actually out there and whatever our senses are capable of perceiving.  But we can now go one stage further and note that there is considerable evidence that it is not just the hard-wired circuitry in our heads, but the ideas in our minds that structure the world we see.  In a famous experiment, people simply do not see a man wearing a gorilla suit walking through an office because they know that you don’t get gorillas in offices.  Likewise, people who have as mother tongue a language that does not distinguish blue and green as colours tend to get very confused when those whose mother tongues do distinguish them insist that they are different.

It follows from this that there is the exciting possibility that much of the world as we think we see it, the ontology, the basic structure, is in fact an artefact of our perceptual and psychological systems, and that there is no simple or direct relationship between it and the thing in itself (whatever that is).  The only question is how much can we be certain about?  Well, quantum mechanics suggests that our idea of ‘thing’ is unreliable, as it replaces localised things with global wave-functions that we happen to perceive as localised things, but it leaves the concepts of time and space intact.  In this essay I intend to argue that in fact even these, the apparent bedrock of being are illusory, and that there is no such thing as the flow of time, the arrow of time, or physical space (no matter how many dimensions it has).  Instead there is a complex of instants and patches of space that we assemble because it is in our nature to expect a continuous flow of events.

So, the argument will go as follows.  First I will examine why time and space might be illusions.  Then I will analyse the appearance of continuous time and space and attempt to determine why it arises, and what it is that creates the illusion of an arrow of time.  Finally I will note some interesting correspondences between the model and ideas from modern physics.

Why might time and space be illusions?

This section is intended by way of a taster, setting out some of the reasons for believing that time and space might not be fundamental concepts after all.  It also provides an opportunity to make first use of a style of argument to be used repeatedly in this effort, which essentially involves taking conventional assumptions and standing them on their heads.  Finally, I will look at reasons why abandoning the ideas of large-scale time and space might not be so bad a thing.

To start off, when I speak of time and space I mean large-scale structures with continuous variation, hence the idea that time is a continuous line leading from then to now, and that space is three (or more) dimensional, continuing in all directions.  In other words, it is the Copernican hypothesis that space and time are big and look pretty much the same everywhere.

In place of this, I ask what it means that we sense the passage of time?  Clearly we don’t experience the passage of time itself directly (though see below), so what makes us know that time is passing?  What we actually sense is change; if nothing ever changed, could we have any sense of time passing?  Likewise, if there were no spatial variation in the nature of things, would we have any concept of position or distance?  The answer to both questions has to be no.  Which then means that in fact time and space are mental constructs that we have invented in order to understand change.  In other words, we have taken the standard view that we detect change because we know about time and space, and turned it on its head.

Suppose the concepts of time and space are features of our psychology rather than of the universe.  Suppose the arrow of time, the infamous second law of thermodynamics are in fact consequences of our existence in the sense that it is impossible for them to be false, because by virtue of our nature we can perceive only that that is in accord with them.  What this means is that we are freed from any number of worries; for example the problem in quantum mechanics that observed reality appears to be created by the observer becomes a simple tautology.  By placing ourselves firmly into the world, and confronting the effect that our preconceptions have on what we think we perceive, as opposed to taking the traditional scientific view of treating the world as if we are somehow not part of it, we can see perhaps the first glimpses of the real truth that reality, if there is such a thing, is so far divorced from what we think it is that it is, quite literally, inconceivable.

Time and space as illusions

The sensation of time

Problems with the naive theory of time

We are so used to the sensation of time passing, of an endless ‘tick’ repeating in the background of our lives, that it is very easy to assume that it is a constant, uniform structure, and that it is innate.  Let me put some flesh on these ideas.  When I say that we expect the time sense to be constant I mean that the rate at which time appears to pass should not change, so what is an hour now was an hour yesterday and will be an hour tomorrow.  When I say that it appears to be uniform I mean that we expect there to be common agreement as to the passage of time, so if I think an hour has passed, you will agree and will not think that in fact it was a year or three seconds.  Finally, when I say that it is innate I mean that it seems to us that the passage of time is part of the fabric of reality; it is not something that we create ourselves, but is simply there, waiting for us to experience it.

The first problem with this naive time sense is that it is not at all clear where it comes from.  Do we observe changing events in the outside world (waves crashing, clocks ticking, clouds moving) and deduce the passage of time from them, or is there an immanent sense of the passage of time that we experience directly?  Let us consider these two possibilities.

External sources

The problem with external sources is that there is no obvious external ‘tick’.  Certainly, there are a number of apparent natural rhythms, from the sub-nanosecond vibrations of caesium atoms to the five billion years half-life of uranium atoms, but if we restrict ourselves to the cycles we can perceive directly (the day, the lunar month, the year) there is a problem in that they tend to be quite long.  We are discussing here not unconscious bodily cycles like the circadian rhythm, but the conscious time-sense, which tends to work at a scale much shorter than even the day.  If it were to derive from (say) the diurnal cycle we would have to posit that we have a quite sophisticated internal timer capable of dividing one day into some number of equal units.  Which means that in order to derive our fast ‘tick’ from the slow natural ‘tick’ we need an internal time-sense.

So, can I proceed using irregular external stimuli to drive the time sense?  The problem is now that I run into problems with constancy and uniformity.  So, I cannot guarantee that the stimuli I experience are the same as those you experience, and so the only way to make the time-sense uniform is for us to have some further datum that tells us how fast the observed changing events are moving relative to the uniform ‘tick’.  But this means that once again we need an internal time-sense that has no basis in external stimuli.  As for constancy, the only way to extrapolate a constant ‘tick’ from irregular stimuli is to have a pre-existing concept of what a constant ‘tick’ is, which requires a purely internal time-sense.

Therefore, though the time sense may, as appears to be the case with the circadian rhythm, rely on external data to correct systematic errors, it cannot be purely external.  There must be an internal sense of time that understands the concepts of constancy and uniformity.

Immanent sources

So say that the time-sense is immanent.  Now we run into all kinds of problems, because, as we all know, any internal sense of time that we have is as far from being constant as one can get.  We have all experienced the phenomenon whereby waiting thirty seconds for a computer to switch on can seem like forever, and yet when we are happily absorbed hours can pass in what subjectively feel like moments.  Our time-sense is not constant and also, as we see from the fact that I can be happily absorbed while you are consumed with boredom, it is far from uniform.

So, subjective time is neither constant nor uniform.  Is there, perhaps an objective time-sense that provide a basic ‘tick’ distinct from the extremely variable subjective sense of time, that helps drive the naive time-sense?  There are two points of attack on this: first the origin of the tick and second whether such an objective sense does in fact exist.

Consider the source of the tick.  We have bodily rhythms: heartbeats, breathing, circadian rhythms.  The first two of these are inherently variable and the third is long and has been shown to be very far from regular when it is not regulated by exposure to the external stimulus of the sun’s diurnal cycle.  Thus these sources fail on the same basis as the external time source, in that using them as the source of a ‘tick’ simply begs the existence of a yet more fundamental ‘tick’ used to regularise and sub-divide them.  The only alternative is an immanent time sense which is constant and uniform, and yet is not directly apparent.  That is to say, we sense time not by interpreting other sensible data, but by access to some transcendent source of information that is otherwise entirely undetectable.  This saves the theory of the naive time-sense but at an enormous cost, for such an in-principle undetectable, and therefore unprovable, time-sense smacks equally of theology and desperation.

Consider now the reality of the objective time-sense and, by implication, the naive time sense as a whole.  Do we actually have any evidence for its existence other than an idea that it should exist?  The evidence is tenuous and very susceptible to the turning-on-its-head style of argument.  Is it the fact that events have naturally defined time-stamps indicating when they started, when they ended and how long they lasted, or is it just that as we are used to living in a world of clocks, we expect to be able to impose that structure on the world?  If I say that such and such an event lasted one hour, do you agree because your time-sense tells you that it did, or because you, like me, refer to clocks that say it lasted an hour?  Are precise time-measurements possible because they are real, or do they seem real because they are possible?  The fact that we appear to feel the passage of time that clocks represent is as likely to be a result of our knowing what they claim to represent as it is a result of our having any genuine innate sense of what (say) a minute means.  It seems that we believe in an objective time sense, not because we are aware of it, but because we are led to believe, by our cultural assumption that time is real, that such a sense should exist.  On our own, all we can directly attest is the hopelessly irregular subjective time-sense.

Thus we must abandon the naive theory.

Alternatives to the naive theory

Consider alternatives to the naive time-sense that do not suffer from these problems.  It is clear from the discussion above that this means ditching at least the concepts of constancy and uniformity, for though we could establish some kind of innate time sense, it was highly subjective and irregular.  Once again we look for sources of this time-sense, but this time we will be a bit more focussed.  The question is now do we sense the passage of time because of concatenations of events that we interpret as indicating the passage of time, or is it that there is some genuine innate sense of time that we then apply to events.

Innate subjective time

Revisiting the argument above, internally we have a rather irregular long-period timer, in the form of the circadian rhythm, and a highly subjective time-sense that is a measure less of time than of boredom.  This subjective time-sense can get seriously inaccurate if it is not reset by regular reference to outside sources.  We have all seen this in the way that strong absorption can lead to a complete loss of idea as to what the time is.  Moreover it is well-known that sensory deprivation can lead to a complete failure of the subjective time-sense, in that it seems to simply cease to function.  This is most notable in sleep, where our idea of how long we have slept generally does not match the measured time, and time in dreams is often wildly at variance with the measured duration of the time of dreaming.  Thus while we may have internal timers that are capable of giving very approximate timing information, they are not constant, they are not necessarily consistent with one another, or themselves and they depend on external stimuli to keep them accurate.

So there is no reliable internal sense of time beyond the observation that this seemed to take longer than that, so we have very crude relative duration.  Moving on to more general concepts of tense, we have a very clear sense of ‘now’, represented by the eternal instant in which we live, and it seems fairly well-attested that our memories are organised so that we have a (rather unreliable) sense of ‘before’, ‘after’ and ‘simultaneously with’.  Moreover we have the concept of aspect, in that we can think of events as complete or continuing, and so relate memories based on continuing activity.

Internal subjective time: the arrow of time

One other aspect of the innate time-sense is the apparent ‘arrow of time’, the fact that it seems clear to us that time moves only in one direction.  Discussion of this tends to get bogged down in confusion as to whether the arrow is cause or effect.  To see the issue, consider the following.  We have two facts:

  1. We sense that we are moving always from the past to the future.
  2. We see glasses break and trees fall, rather than seeing trees erect themselves and glasses reconstitute themselves.

Of these, (1) is vacuous.  In fact we sense a constant present and have a changing memory which purports to represent evidence of previous present moments; I discuss this further below.  (2) is more interesting, as it appears to be genuine evidence for a time-related effect that cannot be attributed to psychology, but in fact it is susceptible to the standard turned-on-its-head argument.  Is it the case that the second law of thermodynamics is true, and things do tend towards states of higher entropy, or is it simply that we assume that they do and hence our perceptual machinery forces the world to appear that way?  I discuss this further below.

External subjective time

Turning to external stimuli, we have abandoned regular time sources (which tend to be slow).  Clearly things in the world around us change: leaves fall, waves break, wind blows.  But none of these provide anything like a constant time or consistent source.  It seems that all we can deduce from the world around us is that change is inevitable and that one concatenation of events will generally lead to another.  We can apparently deduce concepts of ordering and simultaneity, and (very crudely) of relative time, and we can relate instants by saying that at those two instants some particular event was still in progress, but one again it is not clear whether we have these concepts (see above) because they exist in the real world, or that we see them in the real world because that is how we organise our memory.

Alternative time-senses

So it seems that I have no internal source of time and all I can deduce from the outside world is that things change.  I see certain structures in the temporal organisation of the world around me, but it is not clear whether these are inherent in reality or merely artefacts arising from the structure of my mind.

The simple time-sense

Let us start by examining the basic temporal sense described above.  What I will call the simple time sense relies on four primitive concepts:

  1. The Present: There is  a specific instant in which we always exist.  All our perceptions exist in the present.  All else is memory.
  2. Ordering: Memory is organised so that we can say that one memory happened before, after or simultaneously with another.
  3. Aspect: Memory is organised so that we can say that different memories are memories of the same event at different times, and said event may continue into the present.
  4. Duration: we can classify events by saying that one was longer or shorter than another.

These concepts are generally represented in languages as tense and aspectual structures, which are syntactical, while more exact concepts of time require idiomatic structures (consider the sheer number of ways of telling the time in English).  This is clear confirmation that while precise time is something grafted on to our basic nature, the simple time sense is, as it were, baked in.

It is very easy to see how the simple time sense might lead one to infer the existence of universal time: from tense and aspect one can easily construct the idea of time as a constant progression forming a line, with events arranged along it and now moving along it, so memories are memories of ‘earlier’ instances of ‘now’.  Again, we have turned the standard view, that our simple time sense and our languages reflect the existence of time in the world, on its head.  Much of our apparent view of the world around us is a result of interaction between highly unreliable sense data and mental categories and concepts, so why should time be exempt? Assuming (as we shall) that it is not, then the obvious question is whether any of the concepts in the simple time sense can be taken as evidence for a temporal structure in the external world, of whether they can all be treated as artifactual.

The endless ‘now’

The fundamental concept is that of the present.  If time is real, ‘now’ represents a particular time-slice through our perception of reality.  But if the concept of time-slice exists only in our minds then instead what we have is an eternal ‘now’ and an ever-growing collection of memories.  We sense that these memories represent other instances of ‘now’ and we organise them into a quasi-linear progression, giving rise to the simple time-sense.  In fact, it is noteworthy that often our memories are not linear, so we can have memories of two instants and yet have no idea of whether one comes before or after the other.  In other words, rather than being a straight line, our natural time-sense seems to be structured more live a river with many sources, all of which converge on ‘now’, for the one fact we can guarantee is that all memory happened before ‘now’.  Or, at least, to turn the observation on its head again, we apply the blanket term ‘before’ to all that we remember; the concept of the remembered past as a fundamental asymmetry is just as much a creation of psychology as time itself.

‘Now’ creates tense, aspect and duration

It seems that this one, guaranteed fact is the basis of the concept of tense.  If I know ‘now’ that some event was ‘before’, and then at some other time I examine my memory of this ‘now’, as part of that memory I will recall that I knew that the event was ‘before’.  Therefore I now have two events, both past, because both in memory, but my memory tells me that one is in the past of the other.  Hence, based on this model, the entire edifice of our apparent sense of temporal ordering can be reduced to building chains of memories.

This is, indeed, very much how our minds work when we try to determine the order in which events occurred, ignoring for the moment such aides memoires as temporal labels attached to memories.  I ignore these because they clearly go beyond any innate time sense that we might have and into the territory of artificial constructs.  Thus the tense concept derives from the combination of ‘now’ and memory.  In other, words, there is no a priori linear organisation of the temporal sense; it is an artefact.

Similarly, we can generate the concept of aspect from memory and the simple apposition of ‘now’ and ‘past’.  Thus far I have considered memories of other instances of ‘now’, but of course we tend to separate out memorable events as memories in their own rights.  Then these memories carry with them information like which instances of ‘now’ they are associated with, which other events they coexist with, and so on and so forth.  This allows us to co-ordinate events aspectually, by saying that this event coexisted with that event and the other event, but (memory tells us) that event and the other event were not part of the same instance of ‘now’, and therefore were not simultaneous.

Finally, duration also arises in this way.  Naturally enough, events seem further way in time from us if we can remember more events between ‘now’ and them.  Indeed, this gives a neat explanation for the fact that time seems to slow down when we are deluged by events, because the constant onrush of new facts to remember creates many instances of ‘now’ to remember, and so a greater apparent distance in memory between us and the recent past (this might also explain the massively speeded up time-sense of dreams).  Likewise, periods where little worth remembering happens feel short, even if we know they are not, as little has been committed to memory.

The simple time-sense revisited

It therefore seems that the simple time-sense, that which we can genuinely point to as being somehow innate, can be reduced to the following basic structure:

  1. We have an immanent sense of a ‘now’ state which is a unified single view of the world provided by our sensory apparatus.
  2. We have memories of events and objects, including references to instances of ‘now’ other than the current one (that is, we can remember experiencing other ‘now’ states).
  3. We build links between memories so if an event was already a memory at the time of the ‘now’ state we are remembering / experiencing, we interpret that as meaning that the event lies ‘before’ the remembered / experienced ‘now’ state.

As we have shown, that is all that is required.  But that means that the sense of time is no such thing.  Rather than there being any sense of time or temporal continuity, there is merely memory and a constantly changing present.  And if the present did not change, there would be no memory, and hence no concept of time.  Therefore time in the sense of a universal external ‘tick’ does not exist; it is an accounting device that arises more or less by default when we attempt to give structure to our memories.

Non-time and the illusion of time

If time is as illusory as I say it is, if in fact our sense of time is an artefact that arises from the linkage of memories, then there are some obvious questions that I need to answer if I am to make my case at all convincing.  The first question, which is relatively easy to dismiss, is why do we have the time sense provided by memory at all?  Why do we not simply live in the eternal present or, like simpler animals, remember only the last few instants?  The answer to that is that we are not simple animals and our survival strategy is predicated on decision-making based on large stores of memory.  Therefore we need memories, and if we are to function well, those memories need to be organised.

The next group of questions are rather more perplexing and deal with the rather deep question of why it is that, if there is no such thing as time, the universe around us seems to behave in such a way as to make us believe that there is.  There are four related questions:

  1. Why do we perceive the world as changing from instant to instant when there is no fundamental external concept of time to make it change?
  2. What is it that selects the succession if instants presented to us?
  3. Why, if there is no deep concept of time, is it that the information we perceive and piece together in memory is so very orderly?
  4. How do we as individuals come to agree with one another as to the apparent order of events?

The fundamental worry underlying all these questions is that in dismissing the concept of time from the psychological realm, replacing it with loose associations of memory, I have failed to note that the physical realm requires some motive agency to drive it forward and create the changing sensory impressions that we perceive.

In fact this is not the case.  There is a very persuasive model for how a timeless world could exist and yet result in our perceiving an apparently changing series of instants.  Moreover it brings an elegant explanation for the arrow of time and consensus history.  Therefore in this section I will start by sketching an answer to questions 1 and 2 and then show how that answer deals with questions 3 and 4.

The appearance of time

The fundamental observation that underlies my attempt to explain away time is this: that we do not just perceive an instant, the eternal now, but that we perceive a single instant.  That is to say, our minds integrate all the information available to them from sensory resources into a single view, a single picture of the world.  Though we can think of multiple things at once, and think of something other than what we perceive, we appear to be rigidly locked into the single point of view of the outside world, so we can only be aware of being in one instant.

Note the critical rider that this is not to say that I cannot experience several instants simultaneously; merely that any stream of consciousness can only inhabit one.  This inspires the following thought: our problem with apparently denying time only to reinvent it as the thing that presents moments to my consciousness arises because we have tacitly assumed that there is a real succession of instants reflecting my awareness of such a succession.  But suppose that instead of this, I actually experience all possible instants simultaneously and that my mind selects from this plethora a sequence of individual instants as its focus of consciousness, this sequence giving rise to the appearance of a stream of consciousness and hence of time.  This is because though I may exist in all moments, I can only be aware of one as my current eternal ‘now’.

This is a rather startling idea, and at first sight it seems to succeed only in shifting the time concept from under one carpet to another, but it will turn out to remove the need for a concept of time entirely.  So let us explore.  The key idea is that all possible instants coexist in some unordered way, unrelated to one another and with no hierarchy or ordering.  Then we are presented not with one moment that the universe selects for us but by all of the moments in this ensemble and that our minds, in a wholly automatic and non-volitional way that I will explain below, select moments from the ensemble to be the single focus of attention, with the result that time begins to appear to flow.   This means that there is no need to say that I am always located in particular instant.  I am located in every instant, and it only seems as if I am only in one.  Therefore the problem of the selection of ‘now’ evaporates.

But it is still not clear how the apparent succession of instants occurs.  Why do I not just stick in one instant forever?  Also, it would be nice to have some idea of how the single focus of attention is selected, given that this is meant to be happening almost automatically, with no deliberation or intention on our part.

Consider any two instants.  We can say how similar they are, and how easy it would be to turn the world-picture of one into the world-picture of the other.  This can be made very precise if we use the right physics (what the right physics is is, however, another question), but all we need is the observation that this means we can say how likely it is that one world-picture can be turned into another.  This likelihood is going to be in some way (see the comments about physics above) a measure of the number of ways of getting from the first world-picture to the second.

Here it is worth taking a brief digression.  Is it not always the case that the number of ways of getting from one world-picture to another is precisely one?  Well, it would be if we had access to perfect information about the structure of the world, but as it happens we do not; our world-pictures present inadequate information about the bits of the world that we can see, and there are large chunks of it about which we neither know nor care.  So the number of ways of getting from one world-picture to another is essentially a measure of the number of ways that the bits of the world that we don’t care about can comport themselves while the bits that we do care about make the required transition.

Back to the main argument.  Given any sequence of instants, we can obtain the probability of that sequence by combining the likelihoods of moving from each instant in the sequence to the next.  Now say I sit in the middle of all this.  The idea I want to propose now is that actually I experience all possible sequences of moments and have memories corresponding to each sequence.  So I have not, in fact, selected one sequence from many.  Rather I experience the entire set of sequences.

So why does one get picked?  Because when I look in my memory I see instants represented based on their probability of occurrence; that is to say, the more ways there are of creating a particular world-picture, the more I see it.  And this, in a simple form of free market, means that I end up seeing, most of the time, memories associated with that sequence of instants that has the maximal probability.  I may also see memories from near-by sequences with almost maximal probability, creating a kind of shagginess about my memories of certain facts, but I will, on the whole, see only the sequence of maximal probability.  And then this selects the instant I see as my ‘now’, because if my memory is already on a particular sequence that means I am going to continue to see that sequence or one of higher probability.

Let me rehearse that argument, as it is crucial.  High probability sequences essentially swamp low probability sequences in memory because they produce far more world-pictures, because there are more ways of achieving them.  This means that when we introspect we see only what happens on or near the sequence of maximal probability.  And this means that, having ended up being pushed into the realm of maximal probability by sheer force of numbers, we end up sticking in it because we have nowhere else to go.

So I do not need any concept of time.  I can be equally present in all instants, though some of them are extremely hard to get to from mainstream instants.  We replace the concept of the passage of time with the concept of looking at the collection of all possible sequences of events with their representation depending on probability.  But note that now the concept of any privileged arrow of time has gone out of the window.  We are not saying that the passage of time selects events.  We are saying that we select sequences of events based on their probability as compared with other sequences as a whole and then we derive time from those sequences.  We have succeeded in turning the concept of time on its head: time does not define memory; memory defines time.

Note that I am not implying that this selection of probable sequences is either hard-wired into the cosmos or is done intentionally by our minds.  Rather, memories compete in our minds in a kind of free market, with the most common (for which read those that arose from the most probable sequences) winning out.  It is not that we choose the most probable path; it is simply that when we look in memory, it is memories of the  most probably path that we are overwhelmingly likely to find.

Orderly information and the arrow of time

Why should the selected high probability sequence of instants be orderly in the sense that it leads to smooth changes and obeys the apparent arrow of time?  The first question is reasonably simple to answer: the closer two instants are to one another, the higher the probability than one succeeds the other.  Therefore sequences where change is discontinuous will be of low probability and so are likely to be selected against.

The arrow of time is quite interesting.  At first sight one might think it is answered, because we have selected a particular path using a non-time-based approach and that forces us into accepting a particular apparent time direction.  But what about the apparent truthfulness of the second law of thermodynamics?  Why is it that the time direction we selected in a purely mechanistic way should end up favouring an overall transition from less to more disordered states of being?

If I am in some instant and I am looking at possible instants to transition to, there will generally be many more ways of transitioning to an instant that is more disordered, simply because by increasing disorder I am reducing the strictures I place on myself in selecting an instant.  Therefore the transition from one instant to another will generally favour transitions that increase disorder.  Now, when I build an entire sequence of instants, it may be that considerations involving the sequence as a whole mean that the most probable overall sequence will occasionally decrease disorder for a little while, but the overall trend will be to increased disorder.  This is exactly what we see, where local events can apparently reverse the flow of entropy over a short time-period, but the overall trend is for entropy to increase.  So the arrow of time too is an artefact.

Consensus history

Finally, where does consensus history come from?  There is a very simple answer: from memory and debate.  The mechanism I have proposed involves no volition; different individuals will see the same events in slightly different ways, and so end up with different sequences of instants.  But as they are in more or less the same circumstances, their sequences of instants will not differ very much.  There will indeed be, as noted above, a rather fuzzy quality to them, in that no two individuals, even in the same circumstance, will necessary suffer the exact same sequence of events, but this fuzziness will be no more than we have come to expect when trying to build history from testimony.  Indeed, experiments with eyewitness testimony make it clear that history, even in the short-term, is very far from being a deterministic repetition of agreed facts.

In addition, there is one volitional factor that we must include in the framework, that is to say the matter of influence: by communication we can influence one another, and so introduce biases into our memories, resulting in shifts in the selected sequence of events.  Thus, in a rather elegant turning on head of the traditional view, it seems that historians do indeed make history as opposed to discover it.

The sensation of space


Inevitably, after deconstructing the concept of time, the next place to look is the concept of space.  This is in many ways a simpler problem, in that though space is as (apparently) all-pervasive as time, there is no equivalent of the arrow of time.  That is to say, there is no concept of necessary movement from there to here which parallels the flow from then to now, and there is no privileged direction which parallels the future-pointing vector provided by the second law of thermodynamics.  Therefore we need only consider the concept of space and localisation within it.

However there are still issues to cover in showing that spatial concepts have no a priori existence.  We will follow the same path as that followed in the discussion of time, starting from an investigation of inner and outer sources for the concept of space, moving on to the innate theory of space that encodes our basic spatial ideas, and then proposing a new model.  The discussion can be more brief than that of time, partly because much of the argument is simply a matter of taking ideas from above and substituting space and spatial concepts for time and temporal concepts.  But the main reason is, rather surprisingly, that in analysing time we have, it turns out, already done most of the heavy lifting required to understand space.  In fact, the probabilistic instant-based model for temporal perception given above is also a model for spatial perception.  This coincidence can be seen as being striking support for our theory.

The naive sense of space

So, we have a naive sense of space, which is perhaps better described as a sense of spatial positioning and size.  That is to say we expect there to be an innate, constant and uniform concept of relative position and size for objects, with constancy and uniformity taking the same meanings as for time, so we expect the measures of relative position or size to apply universally and to be agreed on by all observers (as usual, modulo relativistic considerations).

We cannot expect an innate sense of absolute position or size to have these properties.  Absolute position requires a fixed point to act as the origin from which all distances and positions are measured.  We all have an innate fixed point, that is, ourselves, so we can each of us establish a personal innate and constant sense of absolute position.  But this choice of fixed point is not uniform, for yours differs from mine; a uniform theory of absolute position requires that we agree on some one fixed point as origin.  But no such naturally privileged point exists, so the choice of point must be arbitrary, meaning that the resulting sense of absolute position is not innate.  Similarly, absolute size requires a fixed object to act as the standard scale, and so the same argument applies.  Therefore the most we can expect is relative position and size.

Now looking at possible sources for the sense, much the same argument as was used for time works here.  An externally-sourced sense is impossible, because there is no a priori standard unit, and an internally-sourced sense founders on the fact that first we tend to be hopeless at estimating distance and size, second that given we can disagree with one another even about such crude concepts as ‘large’ and ‘small’, what hope is there of uniformity, and third that our perception of distance and size is hopelessly mired in the problems inherent in our sensory (primarily visual) system which, as anyone familiar with forced perspective knows, is capable of convincing us that any two objects are smaller or larger than one another regardless of their ‘actual’ relative size.  Therefore the naive sense of space is not viable.

Alternatives to the naive theory

From now on I will discuss only the sense of size.  This is not because my arguments do not apply to distance, rather it is to spare the reader from me repeatedly saying ‘and the same applies for distance’.  It is also because the sense of size is clearly more fundamental than that of distance, given that we often deduce distance from apparent size.  Therefore, all of the following arguments relating to the sense of size apply equally, mutatis mutandis, to the sense of distance.

So what is the alternative to the naive sense of size?  As with time, let us investigate the possibilities for a sense of size that is not required to be constant or uniform, but is still innate.  As it turns out, there is no need to look at external and immanent sources separately, as there is a huge problem that applies equally wherever the source originates.

Consider again the case of forced perspective.  It is entirely possible that I can make radically different judgements as to the relative size of two objects based entirely on where I am positioned relative to them.  Moreover, I can be fooled into thinking that tiny models are gigantic by suitable use of perspective.  So whatever the sense of size is, the size it measures is not an intrinsic property of the thing being measured.  Rather, it depends entirely and only on the world-picture that my senses paint in my mind.  As such it is not a feature of external reality or the thing in itself, but is psychological in its origin.  Therefore the sense of size is innate, but is not immanent, in that it does not provide any deep and direct connection between our minds and deep reality.  It is in us and of us.

The simple spatial sense

The sense of size seems to work as follows.  I experience a world-picture, fed to me by my sensory apparatus, and discern things within it.  I compare them and label them as larger or smaller than one another.  Then, chaining these relative sizes together, I arrive at an overall picture of the sizes of the things I see.  In order to do this I make an arbitrary choice of scale factor, which turns relative sizes into absolute sizes.  Support for this hypothesis comes from the fact that the issued with the sense of size discussed above arise from confusions in relative size.  If we dealt directly in absolute size, there would be no forced perspective.

So, the sense of size depends on two processes:

  1. Breaking up the world-picture into chunks corresponding to ‘things’.
  2. Comparing the size of the resulting chunks.

The second process is not particularly enlightening or interesting from a philosophical point of view.  It could be achieved by something so simple as counting the number of active neurones in two patches of the visual cortex.  The first process is much more significant, but before we can consider it, we must discuss the origin of the concept of space itself.

The sense of space

Where does space come from?

We have seen that there is no reason to believe that much of the nature of the space we appear to see is intrinsic to reality, but now we need to address the fundamental question of why do we see things as inhabiting an ambient space at all?  We do we not just see isolated and disconnected things?

The answer is surely inherent in that last statement.  Our senses give us information to use in planning actions.  Thus the concepts of the sequence of instants and sequence of locations join, and just as the sequence of instants, by virtue of being a sequence, creates the impression of a flow of time, the sequence of locations creates the impression of  flow of something.  That is to say, we experience a way of relating the collection of things I see ‘now’ to the collections of things I saw in my memory of the preceding ‘now’.  So space is a mental construct that allows us to picture a sequence of locations at consecutive instants.

To aid in planning, I need to be able to relate the things I see ‘now’ to the things I saw at a previous ‘now’.  That is to say, I need to have a way of relating ‘here’ to my world-picture that makes sense of how that world-picture changes as I progress along a sequence of instants.  There are a number of possibilities for how this can be achieved, but the most obvious are that we model ‘here’ as static and the things as changing, or we model ‘here’ as changing and the things as (largely) static.  Clearly the latter is the case.  The purpose of the model is to facilitate our selection of sequences of instants, so we need to plot a course; this is much easier with a static model.  In addition, keeping the model static has the advantage that all we have to update from instant to instant is where ‘here’ is; there is no need to keep track of the behaviour of all the things we perceive, because they remain (largely) static.

In this case, the inherent limitation of our minds that we can only handle one world-picture at a time means that I end up with a sequence of instances of the model, one per instant. Now, remember that I use this model for planning how I select instants.  That means that I have to be able to conceptualise the sequence of world-pictures resulting from any plausible sequence of instants.

The world as neighbourhoods

I could make this model by being aware of all possible sequences of instants as a collection of sequences.  This is simple, but it has three problems.  First, I experience nothing of the sort.  Second,  to do so would mean I have to know in advance about possible sequences I might select, and yet my knowledge of the model is based purely in ‘now’, I have no foreknowledge of what is to come.  Third, it violates the single world-picture restriction, by expecting me to be conscious of multiple potential futures.  If the spatial model can only exist in ‘now’ then the only information to hand is where ‘here’ is, my knowledge of the things I perceive, and possibilities for how this can change as I transition to the next ‘now’.  In other words, it has to be compatible with the temporal model of probabilistic paths through instants, where the transition probability from instant to instant now takes into account the difference between successive models (which was always inherent in the discussion above, anyway).

So I assign to the model of ‘here now’ a collection of models of ‘there then’ each with an associated probability.  Then all my information about position is built up by joining these nearby collections together in sequences, one per instant.  What changes between successive models is my location.  So to go from the current model to the new model I need to change a small neighbourhood of ‘here’ corresponding to the set of possible ‘theres’ deriving from the most probable succeeding instants.  And I patch these neighbourhoods together, from moment to moment, to construct my trajectory.

Now I can generalise.  I have the concept of a neighbourhood of ‘here’, which is essentially my current ‘here’ and ranked candidates for my next ‘here’.  And so there is an ensemble of possible instants, which collectively represent all possible states of ‘here’.  So as I progress from instant to instant I construct the appearance of a path, in that I change from one ‘here’ to another.  And so I end up patching all these neighbourhoods together.

I do not have to be present in all instants and neighbourhoods.  That is to say, the ensemble of possible instants will contain instants where I am ‘here’ and instants where I am not.  Moreover, each choice I make of moving from ‘here’ now to ‘there’ the instant after now involves not just a selection of ‘there’ with me in it, but a selection of a collection of nearby ‘theres’ with me not in them, in that by selecting going ‘there’ next, I rule out these other ‘theres’ as my destination.  So the mere fact of patching together my trajectory makes locations off that trajectory more or less probable, and then, very naturally, the probability of my trajectory is influenced by the probability of patching together all these neighbouring locations.

But now I can go on patching together neighbourhoods, joining them via their connections, until I end up with some kind of maximal world-structure at that instant, and again the total probability of this structure is what I need to consider, not just the probability of the particular neighbourhood that I inhabit.  So, in fact, I get ‘space’, as a smooth geometrical object, for free.

Once again, we have turned things on their head: instead of space being fundamental as a smooth geometrical object, which can be split into smaller and smaller patches, we see that the fundamental object is the infinitesimal patch centred on ‘here’ and we construct space by gluing these together.

Why is space uniform and Euclidian?

Geometrical objects are still potentially very complex, with curvature, distortions, holes, etc. But we perceive space as being uniform; that is to say that we perceive that things inhabit a three-dimensional space that looks the same everywhere, so that wherever ‘here’ is and wherever we look in it, sizes and lengths stay the same.  This is part of the appearance of constancy that we expect: in other words, if two things look the same size, then they are.

Of course, this is not true.  We have seen that we can only be aware of relative size.  When we see two apparently equally sized things one of which we have trained ourselves to know is larger than the other, we deduce that the larger must be further away, but equally well it could be smaller than we expect and at the same distance.  Or it could be that there is some distortion in the patch of space it inhabits that interferes with the overall scale factor.

So the reason that we see space as uniform is the same as the reason we suffer from the illusion of forced perspective.  It is that we assume that space is uniform, just as we assume that things that look the same size are.  In fact, as the only information we have is relative sizes, and given that we have no intrinsic sense of the geometry of space, it must be the case that we have to build a space which is uniform, for it is only thanks to the assumption of uniformity that it is possible for us to build a model of things’ positions from their relative sizes.  As all we know is relative sizes, to turn these into a model of distances and positions we must assume some rule, and there is only one natural and consistent way of doing this.  Turning assumptions on their heads again: we do not compute distances from sizes the way we do because space is flat; we perceive space as being flat because we compute distances from sizes the way we do.

Why is space three-dimensional?

There is no particular reason why space should be three-dimensional, but then again, there is no particular reason why it should not be.  This may just be an aspect of our mental processing.  If it is indeed intrinsic to our minds that we receive two slightly different sets of visual sense data organised as flat arrays, then three-dimensionality is intrinsic to us, but not necessarily intrinsic to the universe.  For even if the universe were seven-dimensional, we would not be able to see it as anything other than three-dimensional, because that is what our visual senses force upon us (note that those who have the use of only one eye from birth have no concept of a third dimension).

So, as with time, it may be the case that our mental models are consequences of deep physical truths about the world, or it may be the case that these apparent truths are consequences of the way we build our mental models.

The origin of things

The complete model

I said earlier that I had to model our perception of things.  Before I do that, let me summarise the model as it stands at the moment; doing so will help motivate the way we construct things.

  1. The basic entities in the world are instants, to each of which is attached a neighbourhood, which is a collection of putative next instants, ranked by the probability of transition from ‘now’ to that instant.  It provides relations that allow me to connect my current ‘here’ to possible next ‘heres’ (for geometers, this choice of terminology is not accidental).
  2. I inhabit the set of all possible sequences of instants; sequences because of the single point-of-view restriction that my mind imposes.  However, again because of the single point-of-view restriction, I am aware of only one such sequence at a time and, by force of numbers, that sequence will be a sequence of maximal (or close to it) probability.
  3. My sense of time arises from passage along a sequence.  My sense of space arises from the fact that a sequence contains at each instant the object created by patching neighbourhoods together in all possible ways.  Taking the maximum probability sequence gives the maximum probability succession of objects, which are just three-dimensional Euclidian spaces at successive instants.

Note that the construction of space from patched together neighbourhoods uses the same machinery as the construction of time from patched together instants.  Moreover, they are one and the same process: selecting a sequence of instants based on probability automatically patches the consequent instants via their connections to one another, and we then build a mental model to accommodate the resulting structure.

Where do things come from?

So, I am suggesting that the fundamental entities that go to make up what we perceive as space are neighbourhoods, patches of possibility that we translate into individual points and the directions that connect them to other points.  Of course, these neighbourhoods will not all be identical: if they were then there would be no most probably state, so we would perceive nothing.  Rather there can be variation in preferences as to which other neighbourhoods they can connect to and how they can to them, so one can imagine stronger and weaker connections, leading to tightly clumped and more diffuse assemblages and thence beginning to give the semblance of structure or fabric to the resulting ensemble.

Massive bodies curve space-time

Massive bodies curve space-time

This is, I think, the final crucial insight.  We are accustomed to thinking of objects dictating the shape of the space around them, so in General Relativity, we say that a massive body warps space-time (see figure), but, to turn this on its head, what if the truly fundamental objects are the warps and curves in space-time, and the objects are artefacts that we see because we are not equipped to detect space-time curvature directly?  That interpretation is much more in the spirit of Einstein’s theory, and it is exactly what I have just described.  Preferential attachment will cause inhomogeneities in the most probable collection of patched together neighbourhoods, and I am proposing that what we think of as ‘things’ are simply our minds’ way of representing those inhomogeneities to us, in so far as we can detect them at all.

The consequences of this are quite stunning.  As well as space and time, things themselves dissolve into contingent associations of formless and unknowable units of ‘stuff’.  Everything we think we know falls apart, and we are left certain only of Descartes’ founding assertion: I think.

Correspondences with physics

Quantum fields

The reader who is apprised of quantum field theory will be well aware that the description I have given of how a plethora of possible states with probabilistic transitions between them gives rise to apparently stable time and space and things is rather similar to Feynman’s sum over histories model of quantum fields.  This is not surprising.  For those who are not apprised, it is a principle in quantum mechanics that if you want to find the probability of a system starting in state A and ending in state B you have to sum the probabilities of all ways of getting from A to B.  Note, all, not just paths that obey the laws of physics or are plausible, but all possible paths.  And then, as if by magic, it turns out that what pops out as the most probable path is exactly the one that obeys the laws of physics.

This is just what I have proposed above.  Essentially we have a system consisting of a multitude of tiny patches, each corresponding to a single instant, we throw them all together and, by the magic of probability, something surprisingly like space-time emerges.  The reason for this seeming magic is exactly the same as the reason why Feynman’s trick works; less probable states cancel and concatenations of states that behave in a systematic way tend to have higher probability.  But the underlying picture is not one of space-time, just as in physics the underlying picture is not one of electrons like billiard balls, but of a strange, almost formless something about which we can know directly absolutely nothing.

Gravity and black holes

I have already discussed the deep relation between my model of things and General Relativity.  Einstein always considered the fact that it was necessary to introduce matter into his theory ‘by hand’ as a failing; here we see the beginnings of a route to removing that failing.

Other interesting correspondences exist.  According to General Relativity, clocks move more slowly in the presence of a high gravitational field, and according to our understanding of black holes, a high gravitational field corresponds with a high entropy flux, that is to say enhanced change in the nature of things.  But by our model, as time is a percept caused by our need for book-keeping on our memories, we end up being able to argue in reverse that if more happens then we get more memories, so time will inevitably appear to slow down.  The correspondence is tantalising, as are many others, for example the concept of `space-time foam’ current in some attempts to unify gravity with quantum theory is very similar to our concept of reality as being made up of tiny germs of space, each existing individually, with space as an emergent property of the ensemble.

Paradigm shifts, perspective, impressionism, epistemology


What I want to talk about today is Kuhn’s concept of paradigms and paradigm shifts.  This much-misunderstood theory of science is often popularised as claiming that science progresses not by orderly development and deduction from evidence, but rather as a form of beauty contest, with current orthodoxy being that which happens to be most popular.  It need hardly be said that this view is rather popular with those who wish to claim that science is merely another form of culturally-derived dogma, and has no special epistemological status.  However, it is entirely at variance with Kuhn’s position.

Kuhn claimed that the paradigm at any moment is the world-view that scientists use when attempting to understand the universe around them.  At intervals that paradigm shifts and a whole new world-view arises, which supersedes the old one by being better at describing the universe.  But this new world-view requires a new set of conceptual apparatus; in Kuhn’s, again often-misunderstood, words, it is incommensurable with the old world-view, in that there is no easy mapping between its world of ideas and that of the old.  And so its acceptance is not an immediate process, as there is a need for familiarisation with and translation to the new way of seeing things.  Thus the sociological aspect of the theory does not question the nature of scientific truth, but is about how people react when their world-view shifts, forcibly or otherwise.

This set of ideas, particularly the notion of incommensurable paradigms, is rather tricky (which is why it is so widely misunderstood), but it is very reminiscent of  my concept of epistemic barriers, for they too describe world-views which are so different in the way they describe the universe that there is an effective barrier to communication between them.  So my goal in this essay is to explore this relationship with a view to providing a clear statement of what it means to say that there is, has been, or is about to be a paradigm shift.

I am aware that part of the problem with popularisation of Kuhn’s theory is that it is not clear to the lay-person why (say) the shift from pre-relativistic to relativistic physics was such a jolt, and little attempt has been made to explain it.  Therefore, my discussion will take examples from the history of art, specifically two major changes, that is the introduction of perspective and impressionist approaches to colour.  Both of these are paradigm shifts, and both can be described very naturally as sudden expansions of the formal language of art.  So once again we will see a link between paradigm shifts and epistemic barriers.

So this essay is organised as follows: first I will talk about my two examples, giving the art-historic background.  Then moving to looking at them in terms of the flow of ideas, I will draw out the key features of the transitions that they embody in order to arrive at a model for what a paradigm, and a paradigm shift might look like, and what incommensurability might mean.  Having established Kuhn’s terms, I will then look at them in terms of the theory of information flow provided by the concept of the epistemic barrier, with a view to reformulating Kuhn’s terms within its epistemological language and thus making a precise definition possible.

Paradigms shifts in the plastic arts

Example 1: perspective

Egyptian HuntingIf we look at early art, though there are occasional attempts at creating a three-dimensional effect, the artists generally seem intent on communicating an idea that on producing a perfect reproduction of things as they are (or, rather, as we perceive them to be).  The goal was a poetic truth rather than the literal truth of the senses, so art was not a scene, but rather a collection of image that conveyed the ideas that the artist considered important.

For example, consider the fine piece of Egyptian art reproduced here.  At first it looks like a typical hunting scene.  But then, why are the three figures of wildly varying size?  And why is the lady wearing the Egyptian equivalent of evening dress while on a hunting trip?  And what sort of person takes his pet duck with him when he’s going out to hunt – other ducks?  And, of course, the answer to all these questions is that if you even ask any of those questions you are looking it the image with the wrong mind-set.  In this case, we’re lucky in that the image comes (as you can see) with commentary, so we know the artist’s intent.  However, be reassured that this is not an exception; rather the, to us, unconventional conventions it displays are the norm across Egyptian art (and yes, it does say that the duck is a pet).  So, the piece is intended to show how manly and generally spiffing is the man who is the central figure, what a fine man he is, as shown by his vigour in hunting.  It’s his tomb and his afterlife, so he’s bigger than his wife, not because she was a midget, but because he is the subject of the piece, and therefore the most important object in it.  Now she’s all dressed up because obviously you want to look your best in the afterlife, don’t you?  This is an art of idealisation, not representation.  And as for the duck: well, obviously he was fond of it and wanted it to be with him in eternity.  All of which means that the image is not a picture of a real event; it represents the hero’s aspirations for the kind of afterlife he wants to live.  Hunting (in the form depicted here) was, of course, a pastime of the upper classes, so it is very clear that he has no intention of being a mere labourer in the land of the dead; he will live the life of the elite, as an obvious corollary of which he has a sexy trophy wife  and a duck he loves very much.  Note that, by Egyptian ways of thinking, even if his wife was not, in reality, a sex-bomb, the mere fact of her depiction as one within the tomb would make her one in the land of the dead.  So the image is not just a statement of desire; it is a magical machine that makes that desire a reality.  The image contains a wealth of information that our mind-set puts us in danger of losing, leaving only incomprehension.

Alfred WallisAs a second example, consider this piece by Alfred Wallis, the famed naive painter of St Ives, discovered by Ben Nicholson.  Naive art is another field where perspective is used only rarely, but it is a grave mistake to think, as do some of the more patronising among us, that this stems from technical incapability.  In the case of Wallis, as with le Douanier Rousseau, ability was not an issue because perspective was not a part of the world view they presented.  They did not represent things as they were, they presented them as the ideas they embodied.

Looking at the painting, to our eyes, accustomed as they are to art which simulates the third dimension, there is something weirdly wrong with Wallis’ composition.  If we take it at face value, we have a number of normal-sized houses in the midst of which is a positively Brobdingnagian building.  Vertically above this a giant thing that can only be a headland, only we are seeing it from above, which is impossible, juts out into a strange sea, a sea which embraces the roof of the mansion and has waves whose swell is larger than the boat which appears to be suspended in mid air.  Now, I am not satirising Wallis here; rather, I am trying to represent in words just how unsettling his image appears if one approaches it from within a mind set that expects to see correctly worked-out perspective everywhere.  But now we need to consider not how this looks to us, but how it looked to Wallis.  Start with the headland: this is a feature of St Ives known as ‘the Island’; Wallis knew what shape it was, and what mattered was showing it in its true nature.  In this he followed the same principle as the Egyptians and much pre-modern art: what mattered was presenting the essence of the thing rather than the way the thing looked.  The giant house is a real building in St Ives that Wallis often painted; again, as with the Egyptian piece, its size is exaggerated because it is the focus of the piece, and it has to be large so we can see all the fine detail that makes up its essence.  In addition, Wallis has conflated the topography of St Ives so as to bring together a house and a headland that are actually separated by a length of beach, but that is no matter, because the painting is about the Island and the house, and not the beach, so the beach is reduced to a bare minimum.  What matters is that the two key ideas are seen together, associated, as they clearly were, in Wallis’ mind.  And finally, the other houses, less significant, are real houses near the central house, but rearranged (and they are arranged differently in each painting featuring the central house) so as to create specific emphases and structures around the focus.  Therefore, for Wallis the space of the image is not simply a replica or simulacrum of the three dimensional space of St Ives; rather it is a conceptual space, in which he arranges ideas of things, rather than the things themselves, so as to create the overarching idea that constituted his vision for the piece.  In this, his approach is almost identical to that of the Egyptians: art communicates an idea by showing things as they should be, or need to be, rather than as they are.  Space is the servant of the artist, not the master.

Dora MaarAs a final example I want to look at the process of unlearning perspective that occurred in the twentieth century.  This is one of Picasso’s many portraits of Dora Maar, and it is immediately apparent that though perspective is present, in that there are some three dimensional effects, each segment of the image has its own vanishing point, its own rules, its own three dimensional space, and these spaces are related only by the image’s internal logic; there is no over-riding reality, at least not in any photographic sense, though one can (and I will) argue that there is a higher reality at work.

The key to understanding this painting is the realisation that Picasso was not painting a woman badly, or deliberately distorting her.  He did not start from her face and say to himself ‘how can I make a funny shape out of that?’  Instead he saw her head and upper body artistically as a number of geometric units, each of which has three-dimensional form, but existing distinctly from one another.  As an analogy to aid understanding, consider Quine’s famous example of the rabbit.  You and I see a rabbit run past and I say ‘gavagai’.  You naturally assume that when I say ‘gavagai’ I mean ‘rabbit’, but in fact, I might mean by it ‘an assembled collection of specific body parts’, and you would have no way of knowing that my meaning was different to yours, because whenever I say ‘gavagai’ you will say ‘rabbit’ and vice versa.  Likewise, you or I might see a woman as a single three-dimensional object, but to Picasso, looking at her as an artist, she was a toolkit, a menu of individual three-dimensional components that he could turn into an artistic object.  But, unlike the case of the rabbit, we can see into Picasso’s world via the images he made for us, even though they present us with a grave conceptual challenge.

And so, Picasso takes this ensemble of components, each independently having three-dimensional form, and has then attempted to represent them as he seems them in a two-dimensional medium, in the hope of communicating to the rest of us something of his vision.  But now, as each component is an individual, they have their own three-dimensional structure, and their own perspective when rendered into two dimensions, and so there is no unified perspective or view-point.  This means that we seem to see a congeries of units of the woman, each seen from its own view-point and with its own perspective, but that was not necessarily part of Picasso’s intention.  The intension of this painting is the vision of Picasso’s artistic eye, that vision of a form of forms, each a living thing in its own right, and how those forms interact to create what the rest of us see as only one form.  It gives us a glimpse into Picasso’s reality, showing the form behind the outward appearance.  And so, again, it is more than just a woman, it represents the complex of ideas created in Picasso’s mind by gazing on that woman, the execution of which requires a move away from traditional perspective-based art.

Example 2: impressionism

Cezanne Bathers

Another convention of modern Western art has been the use of colour in a ‘naturalistic’ way, representing things as having the colours we see when we look at them, so skies are blue, grass is green and dandelions are yellow.  Like perspective, the whole purpose is to give art a greater realism, so its business is clearly that of representing objects as accurately as possible in just the way that the eye would see them, so as to leave no ambiguity about what it is that the viewer is looking at.  And just as the cubists smashed perspective, aiming to represent the idea of the thing, rather than the outer form of the thing, so the impressionists smashed the tyranny of chocolate-box colour.

As my first example, I want to look at Cezanne’s Bathers.  One thing is obviously missing from this piece: perspective; depth is created by masses of colour lying in superimposed planes rather than by any formalistic use of vanishing points.  As a result, a ‘normal’ sense of depth is missing, instead we feel the image as hanging in a space of its own creation where form is defined by colour rather than any conventional geometry.  And looking again, we see that even forms have severed their link with the objects they represent; the three foreground women have utterly unrealistic shapes, as does the woman standing at the left, but their point is not to create photorealistic images of naked woman.  Anyone could do that.  What Cezanne did was to use paint and colour to show the mass and the depth of bodies as things seen in light, with their exoteric or real form only the starting point of his attempt to portray their sheer physicality.

GainsboroughBut now look at the use of colour, especially in the sky and the trees on the left-hand side.  In the trees we see something quite revolutionary: colour is no longer a decoration or property of the forms depicted, it is the forms.  The forms and shapes of the trees have dissolved, leaving behind only patches of colour, and yet in this totally non-realistic depiction we sense the depth, the mystery of the woods and feel an emotional effect far greater than anything a strictly realist depiction could achieve.  As an experiment, look at the picture of Gainsborough’s Mr and Mrs Andrews which I have deliberately rendered in greyscale.  Something is lost, but the composition and its structure and out knowledge of its forms remain intact, as does the interesting contrast between the general pastoral tone and Mrs Andrews’ manner of comporting herself.  Now try to imagine the Cezanne in greyscale; instead of a complex of emotions and ideas, the greenery would evince only muddle.  So Cezanne has achieved a breakthrough in that now colour and form are one and the same.  And, even more importantly, paint and form are one and the same.  Cezanne’s colours are not applied in naturalistic shapes, they are applied in the shape of his brush.  The medium, the technique, the artist, the idea and the content are merging into one another.  The same, of course, is true of the sky.  Those extraordinary dark blue blobs have no ‘real’ function, but they create a sense of the depth and emotional complexity to be found even in a clear sky that no photographic or ‘accurate’ depiction could.  Art has shifted from simple representation to being a way of communicating ideas that could never put put into words.

MatisseNext I want to look at one of Matisse’s portraits of his wife, noting that we are now squarely into neo-impressionist territory.  As with the previous example, colour has become almost entirely non-functional; it is plausible that Mme Matisse was wearing a green blouse, but it is not at all plausible that the green of the blouse should spill across onto part of her neck.  And looking closer at the greens, on the blouse and the chair-back we see that there are two groups of green – strong and weak – and Matisse moves between them in a way that makes no sense at all in terms of the play of light or shadow, but which creates a kind of depth without actual physical depth, for of perspective there is not a trace.  Also worth noticing is the way that on the left hand side the chair-back is solid, in that there is a mass of white-green between the spokes of the chair, while on the right hand side we see straight through to the mysterious blue background.  So while the left hand part of the chair is  massive block that seems to cohere with the background, the right hand part floats serenely in a mysterious space that seems to have dimensions completely unlike the three we are used to.  This dichotomy is, of course, paralleled in the depiction of Mme Matisse’s body, where the left side is dark and heavy and seems rooted in the chair, while the right side is lighter and floats in front of the chair, it’s lightness emphasised by the massive black triangle of the collar.

And between these two mysterious spaces, one earthbound and unified, the other light and clearly separated into distinct things, lies the face.  Passing over the fact that it is grey, the key thing here is that Matisse has made no effort to actually portray his wife.  What we see is, instead an idealisation of femininity, a Platonic essence of woman, created from a minimal structure of lines and curves, with no detail or redundancy.  The face could never be that of a real woman, and yet in it we see all real women, rendered with a serene beauty that any degree of realism could only spoil.  And so we see this face, reminiscent perhaps of the strangely inhuman sculptures of the Cyclades, unifying weight and lightness, darkness and light, mass and space in one perfect expression of what it is to be human.  No representational painting, let alone do it without use of allegory or imagery: Matisse creates his luminous image purely out of masses of colour, and thus stands at the ne plus ultra of Impressionism, the last point at which any vestige of representation was possible.

Once again, therefore, we see that by passing beyond simple realism the artist can convey sophisticated concepts without the need for labels or subtexts.  Matisse was not painting his wife, but the idea that the image of his wife created in his mind, and doing that required going one step beyond the conceptual realm of the artistic conventions of his day.  To those versed only in those conventions, his work must seem mysterious, even nonsensical, but once one accepts that he has no ambition to realism, and thus that his painting is not a failed photographic portrait but a successful conceptual portrait, meaning and ideas become available that one could, before that acceptance, not even know were there.

Gwen John

As my last example, I want to take Gwen John’s Terrace in Moonlight, Meudon.  This is at the opposite extreme of impressionism to the Matisse, in the here, rather than colour crystallising into forms, forms dissolve into mark on canvas.  We can see the painting in three ways.

First there is, indeed, the view of a terrace, with a mysterious sky against which rise stones on the horizon.  This image is deeply puzzling because the objects we think we see – the tree, the stones, the path trees, seem to exist in isolation from one another.  Rather than being unified to form a field of view, they hang, suspended in no-space and apparently superimposed with no concern for order or realism.  Indeed, they seem to have even something of the Egyptian approach to depicting objects, each seen face on, so as to emphasise its individual nature, and coexisting only by virtue of those relations which the artist decides to imbue them with, rather than those forced on her by mundane reality.  So, on this level, where on the right hand side we see what we could take as two trees, one passing in front of the other.  But that is not what we see: we see a single entity that bifurcates, standing proud above the plane of the wall and ground, while other ‘trees’ lie within that plane.

Second, we see forms, limned by black lines which fade in and out of existence.  So there is a horizon, there are the blobs that may be rocks, the wall and the massive triangle of the central tree.  These seem to begin to make an effort at breaking up the unity that is the mass of colour, and turning it into what we might perceive as recognisable objects, as if the true reality is the colour, and the forms are that which our minds, schooled in thinking of solid three-dimensional objects, try to impose on it by way of structure.  And yet, as I noted, the lines are not uniform.  In places they almost vanish, and at times they become huge swathes of black.  As a result, none of the forms of the imposed order have any deep reality.  They merge into one another in an endlessly shifting symphony of colour where form leaches into form, one moment of which John has captured for us.  Therefore the painting, in spite of being a representation of a wholly static subject, is dynamic in the extreme, being but one glimpse of an ongoing conceptual process.

Third, we see the mass of colour, no longer the subtle gradations of representational art, or even the complexes of colour of Cezanne and Matisse.  Rather we see colour at its most elemental, reduced to brush-strokes, each of which conveys nothing in itself.  So in one instant we see hundreds of isolated, meaningless brush-strokes, but in the next we see them come together to makes those forms that we, with the guidance of John, discover within them.  We see an emergent structure that is not present in its parts, and yet is present in their totality, just as there is nothing in any one cell that can be seen as being human, and yet in their communion humanity is born.  John has emancipated colour, so now her message lies not in the way that colour serves the forms, but lies in individual marks on canvas, each clearly identifiable as a mark on canvas, and the ‘subject’ of the painting is not the ostensible subject, but rather the way that John uses it to create near abstract structures that play off our ideas of reality in order to help us take the plunge into the deeper realities lying in the gaps between the marks.  One might say that whereas the surrealists tried to use photo-realism to depict the results of their  delvings in the unconscious, John shows us the royal road to understanding the unconscious and making it part of ourselves.  Realism is left far behind.

Paradigms, paradigm shifts, incommensurability

What is a paradigm?

Rather than state Kuhn’s definition and then try to force my examples to fit it, what I want to do is explore the artistic examples described above and look at them for something that one might call a paradigm.  This approach has to dual advantage of both validating Kuhn’s approach, by showing that unprejudiced examination of art history can reconstruct it, and critiquing it, by indicating places where it needs to be sharpened or modified.

In the cases I have described there are clearly several views as to how one should look at art and what the function of art is.  For example, you could look at the Matisse portrait and observe that people don’t have grey faces, dismissing it is obviously not being a good portrait.  Likewise you could look at the Wallis piece and complain on any number of grounds: that the scales of things are wrong, that the land is hanging in mid-air about the house, that boats tend to be larger than waves.  And so on for each of the paintings I have discussed.  In each case you would be missing what I have claimed is the true content and intent of the painting, but that’s not what matters here.  I am not claiming divine revelation, so your objections on grounds of irrealism are just as valid as observations about the paintings as are my rhapsodies to irrealism.  You would have a valid, consistent, coherent world-view according to which the only piece of art that I have discussed that is successful is Mr and Mrs Andrews, just as I have mine, in which all of the pieces are successul in their own way.

So it seems there is not one overarching theory of aesthetics: we seem to have three:

  1. Realism.  Art is an attempt to render the world as we perceive it in as precise a way as is possible given the constraints of the medium.  Certain artifices (e.g. perspective) are agreed on as tools that are used to achieve this.  Art is judged based on its ability to do this, as well as by qualities that the artist conveys by their arrangement of the objects depicted.
  2. Formal irrealism.  Art is an attempt to render the ‘thing in itself’, the reality underlying the outward appearance of objects.  The objects are arranged and portrayed in such a way as to represent those of their features that the artist wishes to emphasise, and their relations to one another.  Art is judged based on its ability to connect to the viewer at a visceral level, creating a vision of the thing in itself.
  3. Structural irrealism.  Again, art is an attempt to render the ‘thing in itself’, but now objects lose their individual status, merging together to form larger structures, and are portrayed in such a way as to make those structures emerge.  Art is judged based on its ability to show the wider reality that underlies the thing in itself.

The key point here is that each of these theories is quite sound in itself.  It is not that realism is right and formal irrealism wrong, or vice versa.  Rather, each provides a valid way of looking at and understanding art.  Call such a body of theory a paradigm.

What is a paradigm shift?

So we now know what a paradigm is: it is a way of looking at something, art in our case, but it could be physics or theology or whatever, that makes sense, is internally consistent, and provides a comprehensive model for understanding what you see.  So what happens when the world changes, when a Cezanne starts to paint masses of colour rather than things, or you first discover a Wallis?  The realist paradigm worked pretty well until the final decades of the twentieth century, but then it began to break down as more and more painters started to reach beyond its bounds in the attempt to represent a message that no longer lay in the interaction between the viewer’s cultural apparatus and the way the things depicted were shown relating to one another, but in the things themselves.

Mr and Mrs AndrewsGo back to Mr and Mrs Andrews.  His carefree posture with gun and dog shows that he is trying, perhaps a bit too hard, to be at ease as a member of the landed gentry, but the slight, but definite, dissonance between the countrified setting and her dress raises a question about how successful this is, a question reinforced by the slight, but definite, ‘come hither’ quality in her expression.  Gainsborough quite neatly fulfils his commission and lampoons pretension at the same time.  But none of this is really inherent in the image itself.  The painting tells us a story, but it does this by making us think about things outside of it: our expectations for country gentlemen, for well-mannered wives and so on.

A Battery Shelled

Compare this with Wyndham Lewis’ A Battery Shelled.  Now the painting exists in and of itself: it represents a mysterious world with a crumpled ground in which part-human figures labour while super-sized humans look on while the cosmos exists as mighty structures in the sky, if it is a sky, indifferent to the insects beneath.  Lewis is not interested in showing us some things and then pointing out to us some of their more interesting features.  He has captured a moment and is trying to convey to us the emotional content of that moment directly from his painting.

What then is to be done?  There are three possibilities for dealing with this defection of artists from the established paradigm.

  1. Ignore them.  This is very, very easy to do, but helpful only if the movement away from the established view is brief and a return quickly follows.  All that happens as a result of ignoring them is that those doing the ignoring end up in their own little ghetto, isolated from future developments and unable to comprehend them.  And this is, to be honest, exactly what we have seen in some cultural conservative circles in art, with the likes of the ludicrous Alfred Munnings bravely proclaiming that post-realism was just a passing fad.  It also happens in, say, physics, where there are still cranks who proclaim that Einsteinian relativity is all wrong.
  2. Try to accomodate both points of view.  This seems very tempting, but it is doomed to fail.  The reason is fairly easy to see: what kind of aesthetic theory is there that can simultaneously embrace Gainsborough, Wallis and John, all of whom have their own great merits?  How can one reconcile commentary through accurate depiction with blazing a way to the immanent other?  Any such theory would end up being so general and so attenuated that either it would be useless, or else it would, by force majeure turn into several separate theories masquerading as one, in which case we are in the third scenario.
  3.  Accept that the paradigms are both equally valid.  In this case we accept that Gainsborough, Picasso, Matisse and all the others are great artists, but no one body of theory can be used to approach all of them.  We need a new paradigm to understand the new art, different from the old one.  That does not make the old one invalid, as it works perfectly for realist art, but it fails to work for the new art, and that is where we need a new paradigm.  In fact, if we look at option 1, it is clear that it is just a special case of this option, consisting of those individuals (of whom there will always be some) who refuse to acknowledge the new paradigm.

Looking at the process that has occurred, what happens is something like this.  We started in the realist paradigm and carried along, most of the time able to exercise option 2, extending the paradigm to incorporate new developments.  But there came a point where artists, in their quest to represent their vision, moved so far away from the realist model that option 2 was no longer available: the disparity between what they painted and what the paradigm expected was just too great to be accommodated.  And at this point a paradigm shift happened: the only way to describe the new art sensibly was to create an entire new paradigm that acknowledged that it was separate, and was a new kind of art, different from realist art.  The paradigm shifts when facts on the ground force the acknowledgement that existing descriptions no longer work, and that a whole new way of looking at things is required.

What is incommensurability?

So, let us say a paradigm shift has occurred, so I now have two paradigms (for sake of argument): realism and formal irrealism.  Above I wrote some words about three formal irrealist pieces; the question is, are my words meaningful within the realist paradigm?  Well, obviously their semantic content remains, but the answer has to be no: the criteria I am using to make my judgements simply have no meaning within the realist paradigm.  Viewed from within realism, Wallis’ paining is simply bad, and everything I have to say in its defence is simply meaningless.  Indeed, following up on that point, it’s not just critical language that becomes meaningless when we switch paradigm; the art itself becomes meaningless.  Without the irrealist concern for the thing in itself, the Egyptian tomb painting is incomprehensible.

It’s not hard to see why this happened.  Remember that I said that a paradigm shift happens when the existing paradigm is stretched to the point where it can no longer contain everything within it and yet remain coherent, so all it can do is split asunder.  So a paradigm shift occurs precisely when it is no longer possible to present a single viewpoint within which everything we are looking at is comprehensible.  Which, inevitably, means that after the paradigm shift, one of the paradigms cannot comprehend the other, for if each can comprehend the other then there would have been no reason for the shift to occur.  So a paradigm shift creates a barrier to comprehension in at least one direction.

This makes sense.  If I a come to a gallery with a conception of art based on Gainsborough, and find myself confronted with Picasso’s portrait of Dora Maar, then it will indeed be incomprehensible to me.  Now note that I am not saying that this means I will inevitably dislike it.  No, the point is that it will be something entirely outside of my current conception of art.  That doesn’t mean that I cannot feel the shock of the new and discover the new paradigm within myself, merely that I will have to do so, that or decide that I want no part in it.  What I cannot do is look at it and appreciate it on the same terms that I would appreciate Mr and Mrs Andrews.  So the realist and formal irrealist paradigms are incommensurable.

In this case we have concluded that it is impossible to describe or appreciate art of the new paradigm from within the old paradigm.  The obvious question is, is this always the case?  What (if any) are the rules governing incommensurability?  Well, say we have two paradigms, A, which is the parent, and B, which is the child.  Clearly each can comprehend itself.  Also we have seen that at least either A cannot comprehend B or B cannot comprehend A.  Can we say any more?

It turns out we cannot.  There are three possible cases:

  1. A fails to comprehend BB comprehends A.  The process of forming a new paradigm certainly involves additions to the vocabulary, whether of art or science: new tools, techniques, concepts.  However it is not so clear that it requires one to reject what one already knew before the paradigm shift.  It is entirely possible to view realist art from, say, a cubist perspective, looking beneath the surface to the work of art as a thing.  It can bring fresh insights and fresh ways of executing realist art.
  2. A comprehends B; B fails to comprehend A.  Once the perspective paradigm shift had occurred, earlier art, like our Egyptian example, would be more or less incomprehensible: certainly much of its meaning would have been lost, because its notion of art as symbol of deep reality would have been incommensurable with the realist paradigm.  However, there is no obvious reason why realist art should not be comprehensible from within the older paradigm.
  3. A fails to comprehend BB fails to comprehend A.  In this case that comprehended both of them but which they could not comprehend.  There is some evidence suggesting that such situations can occur; I will discuss some possible examples below.

Therefore it seems that all cases are possible, and so the concept of incommensurability needs to be refined to take account of information flows.

Paradigms and epistemic barriers

Epistemic barriers demarcate paradigms

I discussed the concept of an epistemic barrier at length in another essay, but for convenience I will recapitulate the key notions here.  An epistemic barrier arises between two communities when their conceptual schemes are so much at variance with one another that reliable communication between the two communities becomes an impossibility.  Note that this is nothing to do with the language used to communicate, but is about the concepts that are being communicated.  Indeed, it is entirely possible for an epistemic barrier to exist where no linguistic barrier exists.  Community A might have within its group world-view concepts that simply don’t exist in community B‘s world-view and which therefore could not be communicated to community B in any language.  It is not so much a matter of only being able to know things you have a word for, as of only being able to know things that you are prepared to know about.  That is to say, you need, as it were, a mental pigeon-hole of the right shape to fit an idea if it is to be successfully communicated to you.

Now this is not to say that any new idea creates a barrier.  Obviously many new facts exist that fit entirely within our conceptual framework and do not require a major change in the way we ourselves conceive of knowledge.  As above, the barrier arises precisely when the new ideas are so far from those we are used to that they simply cannot be expressed in terms of those that we are used to, and so cannot be communicated using existing terms.  What is required to comprehend them is a substantive conceptual shift, a deliberative act of extending the world-view.  Our usual, incremental process of assimilating new information will not work.

All of this will look very familiar, for the good reason that this is just a slightly more abstract way of looking at the ideas we’ve gone over above.  The merit of this approach, however is that it lets us study more precisely what happens at the points of intersection, where paradigms abut one another, and how information does and does not flow between them.  Moreover, the definition of an epistemic barrier is, in philosophical terms at least, reasonably precise, in that it can be modelled quite precisely, whereas that of a paradigm is, perforce, somewhat woolly, as it ends up having to assert, more or less, that we know a paradigm shift when we see one.  In fact, as I will show below, this is not necessarily true.

So we can (at last) define a paradigm more or less rigorously, by saying that it is a region in epistemic space, that is to say a collection of ideas and concepts, and its boundaries, that is to say, the points beyond which one cannot go without a change of world-view, are defined by epistemic barriers.  Of course, this definition is complicated by the impossibility of firmly locating a barrier.  For if one could say where the barrier was, that means one could tell precisely how far one could go conceptually, which means one would then know precisely what one couldn’t know, which is impossible.

This systemic vagueness is actually quite reasonable in terms of what we know about paradigms in real life.  For example, Mr and Mrs Andrews is clearly realist, the Picasso portrait is clearly irrealist, but what about this Cezanne landscape?  Is it realist or cubist?  It seems to be one of those liminal cases that could be claimed as either.  In general we can tell which paradigm most art belongs to, but there will be a group of works that could be in one or the other, and it is impossible to come up with a consistent classification.  This well-known vagueness, which applies in all fields, not just art, this difficulty of classifying the liminal cases, is precisely the real-world evidence of the philosophical claim that the location of an epistemic barrier is unknowable.

The anatomy of a paradigm shift

One advantage of the epistemic barrier approach is that instead of seeing a process where paradigm B replaces paradigm A, we can look at paradigms as areas of epistemic space separated from one another by epistemic barriers, seeing them in the round, so a paradigm shift is less about creating something new than about finding a way to explore a region of epistemic space that was previously accessible.

So, the starting condition, before the paradigm shift occurs, is that I am sitting in some region of epistemic space, which is to say that I have a collection of thoughts and concepts that I am able to frame in my mind and communicate with others who are in (roughly) the same place.  Now, as we’re discussing the case in which a paradigm shift occurs, that means that somewhere in my neighbourhood there is an epistemic barrier, though I can’t say precisely where, as described above.  Of course, nobody stays in the same place epistemically; we learn and think and expand our knowledge and the range of concepts within our world-view, and this means that my location in epistemic space shifts.  Now say I happen to shift towards the barrier.  At some unknowable point I will cross the barrier.  This is important, so it’s worth considering in detail.  The epistemic analysis allows us to understand what happens in much greater detail than the simply binary apposition of the paradigm approach.

Initially I am safely within the same world-view as my peers: I am in paradigm A.  Now, as I extend myself, gradually I will find myself entertaining more and more concepts that I cannot share with my peers, and I may start having difficulty understanding them, because my new thought-patterns render theirs increasingly incomprehensible.  So a conversation with one of my erstwhile peers may be like talking on a noisy phone line, with more and more noise occurring, until eventually mutual comprehension breaks down, at which point the paradigm shift has occurred, and I am now in paradigm B.  But the point to note is that there is no clear dividing-line, no clear point at which I can say ‘there we became incommensurable’.  Rather, in one or both directions, communication degrades gracefully from perfect to non-existent.  And likewise, mutual communication with those already on the other side of the barrier now becomes possible.

That’s what it looks like to me.  What does it look like to those I have left behind?  There are two cases, depending on whether the barrier is total or partial.  The distinction is based on whether the barrier applies to all or part of one’s world view, and hence affects all or part of communication with those in paradigm A.  The point here is that in some cases, for example artistic judgement, the fact that I and you are now separated by an epistemic barrier on matters relating to art does not force us to be separated when discussing, say, physics.  The barrier in this case affects only those parts of our common epistemic space that deal with art, and does not impede communication on other subjects, and so it is partial.  Based on day-to-day experience, where most people share a common core epistemic space, but otherwise sit one one side or another of various partial barriers in a mix-and-match kind of way, we would expect this to be perhaps the norm.  And yet consider the apparent total disconnect that seems to exist between majoritarian culture and, say, adherents of all-embracing belief systems that aspire to cultural isolation, or the gap between the majoritarian world-view and that of those suffering from extremely severe mental illness.  Here we seem to be close to finding a total barrier, which precludes any meaningful communication.

So, in the case of a partial barrier, what happens when I cross it is simply that, on topics that the barrier controls, I become incomprehensible, but on other topics you can still understand me.  This raises the interesting possibility that I may, therefore, be able to help you to, as it were, peek around the edge of the barrier, by using whatever analogical or metaphorical language I can to describe the other side.  This is, in a sense, how education works.

In the case of a total barrier, something much stranger happens.  As I cross the barrier, I will, like a station on a wandering radio receiver, tune out until, epistemically speaking, I vanish: all possibility of communication with me from your side of the barrier is at an end.  In this context, note that friends and family of those who adopt exclusionist belief systems or suffer from extreme mental illness often talk of their loved one being ‘taken away’.  Very often this is put down to the ill intentions of those with whom the individual now aligns themselves, and many in the ‘anti-cult’ movement claim that those they oppose deliberately separate their ‘victims’ in order to increase power over them.  In fact, a more plausible explanation may be simply that the ‘taking away’ is just the natural consequence of passing beyond a total epistemic barrier: where no communication is possible, none will occur.

Quantum Mechanics stole my car

Quantum ordinaryness

I have lost count of the number of ways people abuse the word ‘quantum’.  From describing every breakthrough as a ‘quantum leap’, to Terry Pratchett saying that something is very strange, but it’s all probably due to quantum, to one book I read that attempted, rather desperately I thought, to argue that God was the thing that made quantum mechanics so weird.

And here’s the thing.  Quantum theory is used as a lazy excuse to justify more or less everything, and its probabilistic nature is viewed as a deep, deep mystery that we mere mortals cannot hope to fathom (going back to the God = quantum argument for a moment).

Now, as I hope to show, actually quantum mechanics is actually very simple, and its statistical nature makes perfect sense.  You just need to be a bit less anthropocentric in your reasoning than we like to be.

Wave-particle duality and quantum measurement

First let’s look at the so-called paradox of wave-particle duality, as evinced in the dual-slit experiment. Now ontology is determined both by knowledge and one’s perceptual capabilities.  

To see this, let us imagine a thought experiment.  I have three cards, and I show them to two people, one with normal vision, the other red-green colour-blind.  They agree that one of the cards is a different colour, but disagree as to which one it is (this is because, as has recently been discovered, red-green colour-blind people can distinguish shades of green that the rest of us can’t).  So perception shapes ontology.

Another thought experiment.  I look into the sky and I see two surprisingly identical looking stars quite close to one another.  I try to build a theory that might allow creation of groups of identical stars. Then I learn general relativity, and I realise that there is only one star, but there is also another thing, to act as the gravitational lens, that I cannot otherwise see.  So my ontology now includes ‘things I can’t see but that bend light’.  So knowledge shapes ontology.

Back to quantum mechanics.  Pretend for the moment that quantum mechanics is the ultimate reality. We are big, slow-moving, fairly light objects, and the other objects we interact with are the same. We inhabit the macroscopic world, where quantum effects get averaged out. So our perceptions, our philosophies are based on the presupposition that everything behaves like macroscopic objects. Then we see things that don’t fit the conceptual scheme and they, well, don’t fit the conceptual scheme.

Now imagine what would have happened if we had evolved on a neutron star. We’d be very small. We’d be used to enormous gravitational fields, and our senses would almost certainly work at the level of wave functions (or something equivalent). In which case the whole ‘electron in two places at once’ thing would be meaningless to us, because of course an electron is an extended object, not a point.

I made an assumption that quantum mechanics is correct here, but you can see that it isn’t really needed. All that’s needed is the observation that we have a point-object conceptual scheme, whereas (at least at our current level of observational acuity) we are beginning to see things that conflict with that conceptual scheme due to our having extended our sensory range. But we still in our hearts want to stick with the point-object paradigm. Hence the so-called paradox.

Quantum mechanics and probabilities

‘But’, you say, ‘I thought quantum mechanics was probabilistic.  That’s weird.’  Well, no, actually.  Imagine I have a field and I divide it up into lots of squares, and the only information I have about things in the field is whether they are in one of the squares (in other words, I have deliberately limited myself to a coarse perceptive scheme – you’ll see why in a moment).  Now say there’s a cow in the field.  It shows up not as a cow, but as a selection of squares.  But if I try to grab the cow by reaching into  a square, I won’t always succeed, because the cow isn’t square.  I will succeed with a certain probability.  In other words, because I have thrown information away, the cow has abruptly become probabilistic!

Now, in quantum mechanics it’s not that we’re throwing information away, it’s just that our conceptual scheme is so at variance with whatever is going on, that when we try to look at quanta we lose information.  So we end up with the cow syndrome.  The quanta are perfectly ordinary objects, when viewed with an appropriate conceptual scheme and we don’t have, and never will have, that scheme.

I can go one stage further.  I have divided the field into squares.  Say you divide it into chunks of your own devising.  We both register the cow as a set of probabilities associated to chunks.  What relation is there between those two sets of probabilities?  A very complex one.  In fact, there need be no relation.  Say I divide the field into strips running north-south, and you divide it into strips running east-west.  Say the cow is facing north.  Then I see the cow as being in one strip with probability 1 (certainty) while you see it as being spread across several strips.  Now say the cow starts to walk towards me.  You see it change; I do not.  There is no relation between the probabilities I see and the probabilities you see.  In other words, we have not the Heisenberg uncertainty principle, but something far worse: a complete failure of our theories (the division into strips) to interoperate.

(Technical note for QM cognoscenti: the argument above is not an example of hidden variables, because I am not saying there need be anything other than the wave function evolving according to QM’s rules.  My point is that we try to partition the space of wave functions into point particle shaped chunks, and this is not a good partition.)

And for my next trick . . .

Coming up: why I hate ‘Twilight’ and what I intend to do about it . . .