Answering Mermin’s Challenge

John, Robert,

part of what i asked for is an answer to whether or not Robert's theory is falsifiable. In my opinion there are only 3 options: yes, no, "we do not know".

If the answer is "no", then maybe a detailed description of the theories assumptions can alter the picture (if some clever person comes the way).

If the answer is "yes", that would be great, because the theory can be tested or already has been tested in the past.

If the answer is "we do not know", again a detailed description of the - hitherto developed assumptions - may help some person to find a way to test the theory.

By "testing" i do not mean a computer program, but an experimental test with real physical "quanta" (or "partices" or whatever you like to name it).

Concerning all your questions John, i would say i have the opinion that causality as we know it (timely ordered etc.) breaks down at a certain level within the microcosm. What makes our world stable then must lie beyond our familiar space-time (the latter's assumptions being deeply engraved into our ways of thinking). This is another reason why i consider the falsifiability of Robert's theory interesting, because if some predictions of that theory could be verified, this would obviously speak against my rather "transcendental" view of space-time and all the rest.

Stefan

If the particles produced as a pair have the same orientation of magnetic moment and are then exposed to the same orientation of magnetic field , the same response can be expected from each member of the pair. They remain correlated- shown as 1/2 and 1/2 of the same orientation possibilities. if anticorrelated they remain anticorrelated.)

If instead a pair is exposed to different orientations of field they will respond to the field they individually experience. As only 'spin up" and 'spin down" results are "measured" 1/4 of each kind of pairing is what should be expected for lost correlation-as if random.

Rather than there being two types of the particle with different spin characteristic there is one. Orientation producing the resulting bit is nurtured during exposure to the magnetic field. There is no equivalent isolated bit prior to the execution of the "measurement" protocol.

The Bell's statistics assume the spins to be inherent characteristics, measured not generated by the apparatus.

Dear Georgina,

trying to trace your argument.

Particles are produced as a pair. By production, they always have the same orientation of a pre-measurement property. Then they are exposed to the same physical forces. This then results in the same response. This is the case for the angles 0 degree as well as 180 degree.

If we have a relative angle of 90 degree, it seems that the "correlation" is random, but this is only so because we have no way to know the original orientation of the particle pair at the time of its production. But according to this scheme, we at least assume that both particles have both the same inherent orientations of some inherent property.

So far we have at least figured out that at the angles of 0, 90 and 180 degrees, there could be a physical mechanism to explain the results.

But how do results for the intermediate angles come about? In your scheme the individual particle's inherent properties simply respond to different angles of the individual measurement devices - relative to a particle. That's a local scheme of course - every particle's response is only determined by its original orientation at the time of production and its relative angle to the local measurement device. Since the influence of the measurement device is thought to be non-linear when it is not aligned at the same orientation as the particle's original orientation (at the time of pair production), one then could eventually explain this non-linearity of statistical outcomes (bell curve) by the relative angle between measurement device (aka magnetic field) and the particle's magnetic moment (or something like that) - what is exclusively only then the case when both magnets are neither oriented 0 or 180 degrees relative to each other - because if they are orientated 0 or 180 degrees relative to each other, in these cases the responses are always linear due to the production of perfectly correlated pairs.

So if the non-linearity in question (bell curve) stems from the non-linear forces of a measurement device's magnetic field relative to a starting orientation of a particle (and its magnetic component), one could expect that even for the strong correlations at 0 and 180 degrees, the particle pairs could well be *initially* orientated such that the magnetic field of the measurement devices exhibit the same amount (and directions!) of forces to the particles as it would be the case for the intermediate relative angles (the angles not being 0 or 180 degrees). If that would be the case, the statistics of course would be altered from the well known bell curve to something other. In your scheme I see no reason so far why the source should only produce pairs that have only two options to be orientated in space: either "up" relative to the source, or "down" relative to the source (this kind of explanation would use a preferred reference frame). But let's figure it out:

Can this explanation via a preferred reference frame be tested? Yes, according to your scheme there is a difference in the amount and directions of the forces of the measurement devices (and therefore there must also be a difference in the resulting statistics) that act on the particle's for different relative orientations.

So to test the hypothesis that the non-linearity in question (bell curve) stems from different forces (amount and directions) of the measurement magnets, one theoretically could make a series of measurements with perfectly aligned magnets, but with different orientations relative to the source. This may at first glance appear to be nonsense, since measuring always only with the help of relative magnet angles of 0 or 180 degrees (but both changed from time to time relative to the *source*) will not alter the perfect correlations (the latter means that for 0 degree there is predictable "anti-correlation, for 180 degree there is predictable "correlation").

But nonetheless I think there are consequences for your scheme here. Assume we orient both magnets relative to the source such that we define that angle "0". Keep in mind that both magnets are nonetheless still always orientated at, say, 180 degrees relative to each other as the starting point of our measurements. With these settings then we make the full circle of bell measurements.

After that, we only change the relative angle in space of the magnets to the source by a certain degree (the magnets nonetheless are still oriented 180 degrees relative to each other) and again take this as the starting point to make the full circle of bell measurements.

This goes on till we arrive at an angle of 180 degrees for the relative orientation of source with magnets (our starting point was "0" degrees).

Finally we compare the measurement results of all the different settings. According to your scheme there must be significant differences for the different runs of full circle bell measurements. And this conclusion is independent of whether you assume a preferred reference frame relative to the source or not.

For the case that the source only has two options to align the perfectly correlated pairs in space (relative to itself, the source), changing the whole measurement settings angle relative to the source (as just described) should result in dramatic differences compared to the well known bell curve. This should be so because your scheme aims to be local and is dependent on the relative angle (amount and direction of forces) of the magnets to explain the peculiar non-linearities apart from the trivial angles 0 and 180.

For the case that the source does orient the correlated pairs randomly over say, 360 degrees (relative to the source itself), the comparison of the results of the above mentioned different full circle runs should also result in some dramatic deviations from the well known bell curve - at least for some settings.

If I am right, the question now is whether or not that sophisticated series of experiments has already been done or not. Apart from that I would bet that once done, it would not alter the well known bell curve in any case - what would be in conflict with your local explanation. But my guess isn't really important for what nature does.

Please write back what you think about all of this.

Greetings,

Stefan

Stefan, that is a long reply. I'll read it carefully tomorrow.

As I understand the issue- there are three different angles of orientation of the magnetic fields of the magnets and a choice of same orientation for "measuring" each of the pair OR choice of different orientations for "measuring" each of what was a matched pair but will now not give a certainly same state outcome. The two different environments causes them to be uncorrelated. Though by probability still giving the same/matched state outcome 1/4 of the time.

You ask why only up and down [outcomes}? Because that is the options allowed by the apparatus. The bits, ups and downs are not particles but tell something about the particle's prior behaviour; that led to the outcome.

Dear Georgina,

"Though by probability still giving the same/matched state outcome 1/4 of the time. "

No no, the ¼ you speak of are only the case if the two measurement magnets have a relative angle of 90 degrees (or 270 degrees) to each other. In all other cases the probabilities behave according to the bell curve, not linear. Look at that Bell-curve (roughly speaking the orange line):

https://fqxi.org/data/forum-attachments/MceachernMissedDetections.JPG

Hi Stefan. I think we are talking about different experiments.

Just to clarify as I can't edit -I should have said - Though by probability still giving each same/matched state outcome 1/4 of the time.

You ask why just spin up or spin down particles are produced. I don't think thy are oriented as such but spin up , spin down are outcomes from the Stern Gerlach analyzer. I think the magnetic moment is a dipole . So if one end is given a designation ,say N. Then the pole opposite will be S. Which is up or down does not make two different kinds of particle. If electrons vibrate along the magnetic moment (speculation) then according to orientation it can be in phase with a magnet, leading to attraction .Or out of phase, giving repulsion.

Treating a pair of particles in the same way after production maintains their relation to each other. Treat them differently, the relation is broken and the outcome appears to be random. Given enough runs there should by probability be 1/4 of each of the 4 possible outcomes. Showing that spin up and spin down are not fived types like blue or green socks. But can change according to the environment encountered. Bell's inequalities apply to fixed characteristics, like sock colour. The magnetic moment is a fixed characteristic of the particle alone but orientation of it, giving the interaction producing the outcome isn't.

I only know about 3 different orientation of the SG apparatus, X,Y and Z being used. When you talk about a full circle of rotation relative to the source, I think that is something completely different. Maybe to do with whether or not the particles are intercepted.

Let me clarify what I meant by "The magnetic moment is a fixed characteristic of the particle alone but orientation of it, giving the interaction producing the outcome isn't."

Having a magnetic moment is a part of the nature of the particle. Interaction of that nature with the environmental exposure gives the orientation, which will lead to the output state (bit) from the SG apparatus.

The apparatus us not measuring the orientation of the particles on production from the source. That does not need to be known. If the correlated (or anticorrelated) pair are treated the same the outcome states (bits) will be correlated (or anticorrelated); whatever same orientation of apparatus is chosen.

Hi Georgina,

I am talking about the SG entanglement experiment. I strongly assume you too refer to that experiment so I take that as a given.

"Just to clarify as I can't edit -I should have said - Though by probability still giving each same/matched state outcome 1/4 of the time. "

At a relative angle of 90 degree between the two SG magnets, the chance to find the possible combinations up/up, down/down, up/down, down/up are each ¼. So if you define "same/matched" as up/up or down/down, the chances to find such correlated pairs are ½. ½ is then equal to a random coin tossing with a chance of 50:50 to obtain a matched pair. So at 90 degrees it seems to be a matter of coin tossing to obtain what in the case of 0 degrees was guaranteed, namely a correlation that could be a hint that both particles initially had been given opposite orientations by the source.

The picture I attached is a plot of the experimental results of the SG entanglement experiment. It shows the degree of correlations, starting with a relative angle of the two magnets of 0 degrees. Since the two "spins" are always generated by the source such that one is up and the other is down (we just do not know in advance which one is up or down), at a relative magnet angle of 0 degrees, there is always "anti-correlation" (not a good term because it suggests that there is nothing correlated here. But as you may know, that anti-correlation is such stable that one concludes due to energy-conservation of spin that the source only can produce up/down pairs, no up/up or down/down pairs).

So at a relative angle of 0 degrees, this correlation doesn't appear randomly, but regularly. At 90 degrees it appears to be random. At 180 degrees there is again always the correlation, either the results are up/up or they are down/down. Only the signs here have a 50:50 chance to change and could be labelled random, but that could be explained by not knowing in advance the initial orientations for each particle. Up to this point all this is pretty local.

Now, apart form these special angles, in an SG experiment other angles are measured as well. The results are also depicted in the plot. The angles form 0 to 360 are depicted horizontally from left to right. The orange line depicts the experimental results whereas the red line depicts the predictions of Bell's inequality.

The question now is how the orange line comes about for a hidden variable theory (or more generally speaking, how it comes about in any theory that aims to explain the orange line). In your scheme the magnetic moment is changed by interaction with the measurement apparatus. More precise, the magnetic moment's orientation is changed. For different relative angles of the two magnets, the chance to find an up/up or down/down pair increases from 0 to 180 degrees. Further it decreases from beyond 180 degrees to 360 degrees. But it does not increase or decrease linearly. Obviously, the probabilities (frequencies) of producing up/up and down/down pairs at the magnets is a function of the relative angles of these magnets. That's a perfectly rational explanation since we assume that the produced (by the source) and pre-measured pairs are always strongly correlated (up/down or down/up) until the correlation is eventually altered by some measurement settings.

Now the question for me (in my first long reply to explain what I consider an explanatory problem) was why at an angle of, say, 0 degrees, the results are always "anti-correlated" and not in the area of the orange line that differs from the red line. If I imagine a particle pair being produced by the source, according to your scheme there must be a certain relative angle of the initial magnetic moments relative to the orientation of the two magnets in space. The magnets are both perfectly aligned (0 degrees of relative angle between them), so whatever happens to both particles at their respective magnets, the result will always be "anti-correlation" (means up/down or down/up). BUT, since in your scheme all the measurement results are a function of the relative angle between the initial magnetic moment and magnets, for obtaining the results depicted in the plot the source must create the initial magnetic moments such that the up-down axis is a fixed parameter in the laboratory space for all particle pairs - or equivalently, it should be a fixed feature for the pair production at the source. Because otherwise the initial orientations relative to the magnets would be randomly distributed over 360 degrees (for the whole experimental series) and this would counteract the function of magnetic angle(s) relative to the incoming particle's orientations of magnetic moment. Because of that I wrote in my first long reply

"For the case that the source only has two options to align the perfectly correlated pairs in space (relative to itself, the source), changing the whole measurement settings angle relative to the source (as just described) should result in dramatic differences compared to the well known bell curve. "

Do you see the paradox? If the source produces the pair's up/down features always along the same axis in space relative to itself and during flight these orientations remain fixed in space, why should measurements at a relative angle of 0 degrees at the magnets always result in perfect "anti-correlation" if these two magnets aren't deliberately oriented such that they fully match the pair's initial orientations? You may say that this is irrelevant since both magnets are perfectly aligned. But on the other hand you explain any deviation from perfect "correlation" (or likewise "anti-correlation") by a deviation of the relative angle between particle's magnetic moment and magnet. Why should this angle initially be such that for the case of our 0 degrees, no deviations are measured?

Hope to have clarified my rather long first reply. Sorry for that rather longer reply :-) - I tried to explain my case as detailed as possible.

I would be happy if you would again reply to see if I understood your scheme correctly.

Greetings

Stefan

Stefan, I'm addressing Mermin's experiment puzzle written about at the start of the article. It doesn't mention angles of 0, 90,and 180 degrees. Are these difference between angles of rotation of the SG devices- like x,y,z but different angles?

RE. "Though by probability still giving each same/matched state outcome 1/4 of the time. " I added 'each' to make clear for up/up and for down/ down. As you say 1/2 for both. I don't know where the red and orange lines you mention come into it. However no hidden variable accounts for the violation of Bell's inequalities. The inequalities only apply to fixed characteristics, like sock colours. Orientation of magnetic moment is variable.

I think it is important that the bits output are not the same as orientation of particles. There could be /is a selection of orientations giving up result for example-lets say some variation in up-ness. We can not know if the output bit term matches the orientation of the particles magnetic moment at production.

Hi Georgina,

Mermins puzzle is a description of the SG entanglement experiment, so we are talking about the same experiment. Only that Mermin handles the whole experiment as a kind of black box and invites us to figure out what the inner workings of the box are to produce the reported results. Moreover, he challenges us with his suggestion that there is no way to classically, mechanically explain the results.

The angles 0, 90 and 180 degrees always refer to the magnets' orientations relative to each other. As you can imagine, one can turn one magnet around its axis by 90 degrees, leaving the other magnet unchanged. This then would result in a relative angle of 90 degree between these two magnets.

"RE. "Though by probability still giving each same/matched state outcome 1/4 of the time. " I added 'each' to make clear for up/up and for down/ down. As you say 1/2 for both."

The 1/4 are exactly the case when the relative angle is 90 degrees.

But this is not the same as Mermin's case b) - read Mermin's paper here:

https://www.informationphilosopher.com/solutions/scientists/mermin/Mermin_short.pdf

Mermin's case b) is for the case of a relative angle of 45 degrees (or alternatively 315 degrees).

Mermin's puzzle confronts us with the assumption that each particle has a set of instructions that tells the particle how to react when switch 1, 2 or 3 is on. The switches symbolize the 3 possible measurements axis' x, y and z in the SG experiment.

His case a) is equivalent to the 180 degree case of the SG experiment: both spins either are up/up or down/down.

Mermin's conundrum is (although he doesn't mention a Bell curve or orange or red lines) that fixed instruction sets should produce the red line depicted in the plot i gave you (or the blue line of the plot here: https://en.wikipedia.org/wiki/Bell%27s_theorem#/media/File:Bell.svg).

But Mermin's devices produce the orange line (or blue line for the wikipedia picture) - and that is not compatible with a fixed set of instructions.

You are right that Mermin's conundrum - at first sight - only applies to the assumption that there are fixed instruction set in play. Take the wording "instruction set" as equivalent with "physical law". The particle has a magnetic moment with a certain orientation. the physical law is how this orientation is altered when a well defined influence acts on that magnetic moment directly (or alternatively on some other property indirectly which in turn again acts directly on the magnetic moment). So your scheme aims to explain the Mermin-cases a) and b) by instruction sets (means physical laws that dictate how the magnetic moment should be altered when influenced by a certain amount of force from the magnet).

The essence of Bell's considerations is that a theory without fixed instruction sets is physically impossible - except for the strange world of quantum mechanics. Whatever mechanical, physical forces along the chain of events produce the Bell curve, it must be instructions how to react when a particle's property x in a state of orientation y relative to a magnet encounters a force z of some environment (the magnet). I think you would agree on that, otherwise we are left with randomness in your scheme.

Since your scheme explains things by a pre-existing magnetic moment that can be changed via a measurement, it is in conflict with the assumption that the particle does *not* have a definite orientation prior to the mesurement.

What is this conflict? the conflict arises about whether or not particles have well defined spin states (up or down) in all 3 directions prior to a measurement. It is possible to *imagine* a scheme like yours where these spin states aren't detected but altered via measurement. But nonetheless your scheme says that until altered via such a measurement, the spin orientations remain unaltered. This does mean that they are unaltered between their production at the source until they are "measured" for the first time.

The question now is in which direction around the axis of flight these spin orientations leave the source. If these orientations are evenly distributed around 360 degrees in a statistical sense (but surely pairwise always correlated), how can a fixed set of two magnets in a row (means they are oriented identically in space) always give anti-correlated results (means up/down or down/up, but never up/up or down/down)? To assume that the source prefers only one direction around the axis of flight is evenly absurd. So both possibilities are likewise absurd and that is another reason why the orthodox interpretation of QM says that the particle's spin orientations aren't predetermined by the source until measured for the first time.

Hope that clears some misunderstandings.

Greetings,

Stefan

John,

"Remove that dipole and apply one at a different angle of incidence, and the electron forgets the first measurement, and aligns orthogonally with the second."

If that dipole sits in the outer shell of the silver atom at a definite place, the problem i spoke of in my last reply remains. If that dipole's vectors are smeared out somewhat uniformly over the whole silver atom, how can the well known deviations (violations of the Bell inequality) from equi-partitioned probabilities then come about?

Stefan, thanks for clarifying that we are talking about the same experiment. "Mermin's puzzle confronts us with the assumption that each particle has a set of instructions that tells the particle how to react when switch 1, 2 or 3 is on. The switches symbolize the 3 possible measurements axis' x, y and z in the SG experiment." Stefan.

"I don't think I'm arguing for instruction sets. The particle does not have the capacity to carry them. The apparatus encounters the particles however they are oriented relative to it. The orientation of the magnetic moment does not change unless it encounters an environment that makes it do so. That can be known from experiments where particles are collected after a fist run through the apparatus and retested with the same orientation of analyzer.. Which produces same spin outcome as previous test. A different orientation of analyzer-random spin outcome. The field encountered in the SG apparatus is inhomogeneous and so each individual particle will have its unique experience according to its orientation and position of entry and trajectory though the apparatus. There are not set instructions as to how it must behave but ad hoc (not generalizable, as it happens) response. This fits with an 'open' unwritten material future rather than the outcomes already within space-time. What is generalizable is- treat the 'entangled' particles to the same environmental influence, they will respond in the same way. Treat them differently correlation is lost,

Hi Georgina,

thanks for your reply.

The term "instruction sets" is just an alternative shorthand to say that all what happens physically in those experiments is governed by some classical laws of physics - known or unknown. So in this view, all the components of those experiments and all their laws of motion and interaction *are* the complete "instruction set" that determines every outcome. Of course no particle or other component of these experiments has some kind of list to look at for how to react to specific encounters.

"The apparatus encounters the particles however they are oriented relative to it."

Yes, this would be the view of classical physics.

"There are not set instructions as to how it must behave but ad hoc (not generalizable, as it happens) response. This fits with an 'open' unwritten material future rather than the outcomes already within space-time."

This is of course an ad hoc explanation. I think it falls under the category about Bell said it would be local, but unrealistic. You know Einstein insisted on a theory where all elements of the physical mechanics that lead to the results have a counterpart in the theory. Remember "God does not play dice". Surely Einstein's view was a deterministic one, all outcomes then had to be predetermined in space-time.

If there are no "set instructions" (aka physical laws) as to how it must behave, then we are left with bottomless randomness that somewhat manages to be generalizable, because it follows the Bell curve. Even without explicitely naming that kind of randomness as generically non-local, implicitely it seems to be just that - because how can some local random events (by random i mean here events out of the blue, ad hoc and without classical or other laws dictating the events) produce the regularity of the Bell curve?

The term "realistic", in my opinion, is always used to characterize a physical system as behaving classically, means mechanically, with the ususal pictures like forces, properties, interactions. Boiled down i think "realistic" and "classical" are another shorthand for "cause and effect" as we as humans are used to. The question for me is then, is it possible that there could be unphysical (immaterial) causes that can have physical (material) effects?

Stefan, I think it is not quite right to think of the laws of physics as instructions that must be obeyed. Instead I think they are a distillation of what happens, from observation. Like Kepler's laws. There are no instructions telling the planets what to do. But from observation of what they do a pattern can be found.

The results are not what would be expected for a fixed property which seems to be the classical assumption. If change of orientation. potentially altering output state, can happen that is like blue socks turning pink- and Bell's inequalities don't apply

Hi Georgina,

thanks again for the reply. I agree sofar as blue socks can turn pink. Mermin's challenge is to explain how and why they do it. Or in other words, what goes on in Mermin's boxes to produce the results. Classically, blue socks *must* turn pink under some specific influences. Your attempt seems to be that they can, but they must not. What decides then that it nonetheless happens, that is the question nobody could answer me yet.

Stefan, classically socks are either blue or pink they can not change.[ that of course is an analogy for any fixed classical characteristic; excluding laundry accidents!]. I'm proposing the idea that the orientation of thee magnetic moment is not like that but responds to the environment encountered. What the outcome will be is not preordained or pre-written but develops as the relationship of particle and magnets evolves. A pair can either undergo the same environmental 'journey' or undergo different 'journeys'. Same journey they produce same state outcomes. Different journeys-different or same state outcome, ie, not necessarily the same.

Why 0 degrees and 180 degrees give matched outcome all the time.

lets call the phase of the particles vibration and that of the magnets electrons I for in and O for out. (Into the magnet/out of the magnet. Imagine too the magnetic moment as a magnet.

Lets have output ports that divide the electrons exiting the machine. Instead off calling them up and down lets call them R and G ,R taking those closest to N pole facing in to center of apparatus magnet and G those closest to S pole facing center of apparatus magnet. When the device is inverted relative to the other, of course, the ports are too. Which is why the names up and down would be confusing. The following lines refer to phases of; top magnet (first), test particle, and bottom magnet.

Alice's apparatus, attraction to top, repulsion by bottom R outcome

1 0 1 0 1

0 1 0 1 0

0 1 0 1 0

Bob's apparatus (160 degrees cf. Alice's) Repulsed by top attracted by bottom. Also R outcome as ports inverted with apparatus. This is for correlated particles

0 1 0 1 0

0 1 0 1 0

1 0 1 0 1

If anti correlated ( meaning a pair of opposite orientation of phase) particles is used the anticorrelation is preserved as can be seen by drawing out more phase interaction diagrams, and thinking carefully about what I and O mean on each line.

The particles are responding to their local environment-relationship of the apparatus to Earth's gravity doesn't matter.

Pink and Blue,

socks don't change themselves.

Neither do the poles of a magnet. It is purely a matter of convention that North and South are platted from the archaic traditions of early civilization and the mysterious 'lodestone' before the earth's magnetic field was known. So in practice a compass needle points N but that's the south end of the magnetized needle and customarily painted red, though Blue is commonly attributed to N. And, by convention, the direction of magnetic field lines of flux (really arbitrary isobars of the same level of intensity) are commonly shown with arrows as 'moving' from the North pole and looping around (usually diagrammed as upwardly) past parallel to loop again around to the south pole. Still, you'll find many diagrams that label a magnet end 'N' colored red. And there is no detectable direction other than that isobaric shape. Also by convention if you look 'down' onto the north pole, rotation is Clockwise and is really due to the fact that most people are right-handed and twisting a screwdriver that direction has greater strength than CCW. So in diagrams with N Up, the direction of the horizontal arrow showing is pointing leftward, and UP is Negative torque. Two common bar magnets oriented on a plane at right angles will display an attraction of the S end of one magnet from the right angle plane of the S end of the other, all the way to the N loop, and vice-versa. Equilibrium is 'superposition' made physically evident in the macroscopic realm. Nothing mystical about it, no spooky action. Just real easy to get confused, Like the same equilibrium displayed by opposite electric charge.

So Mermin's device holds not hidden variables. An electron is like a 2sphere, there is no cowlick, the hairs on that coconut all stand on end! The orthogonal relationship is freely gimballed, it doesn't matter if the electron is rotating or any surface discrete region is circling an equatorial plane. There are an infinite possible number of possible equatorial planes. It is the shape of the external magnet group producing a non homogenous field that becomes less intense towards one element which influences the flight of the electron. And ON AVERAGE the electron's freely gimballed propensity to be oriented either UP or DOWN will be equally distributed. Superposition of both electric charge and magnetic moment has no preference and persists throughout. Half of 'em will go one way, and the other half will just as likely go the other, they need not be all uniform, just basically the same. But the electron does not need change at all. jrc

The simplest electric motor has no preferred direction of rotation that isn't designed into it.