1  Introduction

Most physicists today believe that a local, deterministic theory of quantum phenomena is impossible. Bell's theorem[1], backed up by various experiments[2][3] of the Einstein-Podolsky-Rosen[4] (EPR) variety, is generally taken as proving this conclusion. Most notably, the experiment of Aspect[5] is viewed as a clear observation of nonlocal behavior.

It will be demonstrated in this paper that the inability to explain quantum phenomena in a local, deterministic manner is not the product of a nonlocal, noncausal universe. Rather, the inability stems from a single basic physical error made in the early days of quantum theory, which error produced all of the myriad contradictions which make up the `weirdness' of current quantum mechanics. When corrected, a very simple, causal, local theory immediately appears.

Mathematically the new theory is, at the level of Schroedinger waves and their matrix elements, identical in most respects to current quantum mechanics. The underlying foundations of the theory are greatly simplified, but the mathematical expressions for the matrix elements for all single-particle processes remain unchanged.

It might be thought that if the mathematics is the same then the theory must be the same. The nonlocality and indeterminism of current theory are generally viewed as being part and parcel of the mathematics. That this is not the case-that the various `weirdnesses' are the product of unidentified (and incorrect) physical assumptions and are not inherent in the mathematics-will be demonstrated by producing an actual theory which is local and deterministic, but without changing (most of) the (matrix element level) mathematics.

The proposed theory is not a `hidden-variables' theory, in the sense of a theory which accepts present quantum mechanics and then adds new variables. Rather, current theory is modified by correcting the basic physical error, with the result that one can understand the mathematics in a local, deterministic manner without the addition of any new variables. In the new form, however, it will be shown that one can account for the unpredictability-as opposed to indeterminism-of particle behavior as being the product of `hidden' variables which have exact-if unknown-values at all times.

2  Preliminary Experimental Evidence

To begin presenting the evidence for the proposed theory, and for the error in present theory, I will take a new look at a number of the key experiments confirming quantum behavior. Some of these experiments were not available when quantum mechanics was first developed. Had they been available, some very different theoretical conclusions might have been drawn. So, how might the experimental evidence available today be interpreted using the principles generally accepted in the pre-quantum era?

By `pre-quantum' principles I do not mean `classical' physics per se. Much about the proposed theory will be very non-classical. Rather I mean simply that the experiments are to be interpreted based on the view that real objects have a single identity-that a wave is a wave and a particle is a particle, that a particle is located in one place at one time and not many places simultaneously, etc. Facts are facts. Cause and effect is strictly obeyed.

And, in particular, nonlocality is unacceptable in this view of things. Nonlocality implies that distant events can affect one another instantaneously by no physical means. But no effect can be produced by no means. Nor can any effect be propagated over a distance instantaneously, which would imply an absence of means, as well as contradicting the well established fact that physical effects cannot propagate with a velocity greater than c, the velocity of light.

Consider first the standard double-slit experiment. This is the basic experiment confirming the wave-like behavior of particles. One sees a wave-like pattern on the screen, but only as the result of numerous individual particle events. Each particle is observed to arrive at only one point on the screen.

If one tries to observe particles at any point before or after the slits, or as a particle passes through a slit, one always observes only particles-with a single location-never a wave. Nonetheless a wave-like pattern appears on the screen (assuming no attempt to observe the particles before they arrive at the screen). So clearly both waves and particles are present. One sees waves and one sees particles, so one has both waves and particles.

Yet, as is well known, all attempts to interpret the experiment using separate waves and particles have failed.

To begin to see why, consider the following additional fact about the experiment. Suppose one tries to explain the pattern on the screen with a hypothetical set of particle trajectories. The maxima on the screen might then be explained by particles following the trajectories shown in Figure 1.

However, if the screen is moved to position B, clearly the particles from each slit, and still following those same trajectories, will no longer arrive at the same points on the screen; the particles from one slit will fall somewhere between the points of impact of the particles from the other slit. The pattern would then be washed out. And yet a similar wave pattern is observed at all screen distances.

If the particles are assumed to be particles, and if they follow straight lines between the slits and the screen, there is only one conclusion that can be drawn: the trajectories depend on the screen position. If one moves the screen the particles follow different trajectories.

But this could only happen if something is moving from the screen to the oncoming particles to affect their motion. Without some real, physical process to explain the screen dependence, one would be left with the need for a nonlocal interaction to account for the dependence. If one rejects nonlocality, then this experiment constitutes direct observational evidence of the reverse motion of something from the screen to the particles.

Perhaps the reader is so used to the usual quantum mechanical description of this phenomenon, in which the particles follow no trajectory in particular, that it isn't immediately apparent that the trajectories are screen-location-dependent. But unless one is already wedded to the usual quantum picture, what one has here is a direct experimental observation of the fact that something moves from the screen to the particles. There is no other local manner in which one can explain what is observed.

One might try to invent a theory such as Bohm's[6] in which a potential of some kind exists in the region behind the slits. This potential would not depend on the screen position, but rather only on the slits. The particles would then follow curved paths of some kind, but paths which do not change as the screen is moved. However, as proved indirectly by Bell's theorem, it is impossible to accomplish this with a local potential. And Bohm's `quantum potential' is explicitly nonlocal.

Something has to move from the screen to the particles.

Consider next the emission of photons by an excited atom in a resonant micro-cavity. Experiment[7] confirms that the atom can emit a photon only if the appropriate state is available in the cavity-as this is described in current theory. If the cavity is `mis-tuned,' the atom cannot radiate.

Clearly the cavity affects the emission process. But this can only happen if something travels from the cavity to the atom to affect its emission. Otherwise there must be a nonlocal interaction with the cavity walls or the photon must first be emitted and then un-emitted if the necessary state is found to be unavailable. Neither alternative makes any sense. Again we see the reverse motion of something.

Perhaps the best example confirming the reverse motion is provided by EPR experiments. Consider, to be specific, the experiments with two photons and measurements of their polarization[2]. Bell's theorem[1], generalized to this experiment[8], proves-although it is not usually interpreted in this manner-that whatever variables might describe the photon on one side of the experiment, there must be a dependence on the orientation of the polarizer on the other side. Otherwise one cannot explain the observed correlations between the polarizations of the two photons. This must be true for any description of the photon based on parameters which have exact values. But, given our `pre-quantal' outlook, all parameters must have exact values, and nonlocality is unacceptable. Hence what we have is a direct proof that the photon state depends on the orientation of the opposite polarizer.

It might seem strange to interpret Bell's theorem as proving the fact of this dependence. Bell's reasoning was premised on the absence of any such dependence. He then proved that, given this lack of dependence, there is no local manner in which one can account for what is observed, assuming a description of the photons based on parameters with exact values. But far from proving nonlocality and/or the absence of parameters with exact values, as this is frequently interpreted, the fact that nonlocality and the absence of exact parameters are unacceptable means instead that the theorem is a reductio ad absurdum of its major premise. The photon state must depend on the polarizers, because if not one is forced to accept nonlocality and/or the absence of parameters with exact values describing the photons. But both of these latter conclusions are absurd.

But, again, this dependence could only occur if something travels from the polarizers to the photons and/or the photon source to cause the dependence. Again we see evidence-in this instance proof-that something moves in reverse.

But what about Aspect's experiment[5] with `double-delayed-choice?' It would seem that even with something travelling from the polarizers toward the photons there would be no way to account for his result locally. When both polarizers are rotated in the delayed manner, there isn't time for a signal from one polarizer to reach the photon on the other side before it reaches its polarizer.

As will be demonstrated in Section 4, there is a shortcoming in Aspect's experiment, arising from the repetitive switching back and forth between the same two polarizer states on each side of the apparatus. For the parameters chosen by Aspect, something travelling in reverse, from polarizers to photons, will, in fact, explain his result in a local manner. A proposal for a definitive experiment which corrects the shortcoming will be made in Section 4. A local explanation for the single-delayed-choice experiments will also be presented.

All instances of non-commuting observables constitute similar evidence of reverse motion. The state of a particle depends on the measuring device one uses to observe it. If one uses one measuring device, what is observed contradicts anything that might have been observed with another (`non-commuting', so to speak) device. Indeed, the observation with one device forces one to conclude that the particle had no state in particular for the other device-an absurd conclusion, if, again, we are maintaining that facts are facts.

But the state of a particle can only depend on the measuring device if something moves from the device to the oncoming particle to affect its state. There must be a real physical basis for the dependence.

So, if looked at `pre-quantally,' the collective experiments yield much evidence of something moving in reverse-from a detector or measuring device to the particles that will be observed by that device. There is no other local manner in which one can explain what is observed.

But, at the same time, we have the mountainous evidence that the equations of current quantum mechanics work. When applied to any known physical system, those equations yield what is observed in the laboratory. Yet nothing moves in reverse in current theory. If some new entity is to be added, moving in reverse, how can one still explain the fact that the current mathematics works so well?

The essence of the proposed answer is very simple: It is the quantum wave itself that moves in reverse. With a reverse wave all that changes mathematically is the sign of the momentum exponent in the exponential describing the wave.¹ But that exponential is squared in absolute value when deriving an observable result anyway; so a change to its sign changes nothing. And, by reciprocity[9], the matrix element for any scattering of a reversed wave is identical, but for a possible phase factor, with the forward scattering. If one can make a theory work with reverse waves, it should yield the same mathematics and yet still provide the `something' that moves in reverse.

It must be the quantum waves that move in reverse. We know by direct observation that the particles move forward; and we know-with near certainty, if not complete certainty-from the mathematics of the current theory that nothing other than the particles and the quantum waves is involved. It is difficult to imagine how some third thing might be involved, and yet still permit one to recover the current mathematics. So the only thing left to move in reverse is the quantum wave itself.

Clearly reverse waves imply a radically different theory. No longer are the waves somehow the particles. Rather, the waves are present in the environment already, and the particles then follow those waves. But enough evidence supports such a picture to warrant its consideration.

The reason for the failure of all previous attempts at a theory with separate waves and particles-indeed, of all previous attempts to account for quantum behavior in a local, deterministic manner-is that the waves were always assumed to move forward, with (or as) the particles. Because the waves actually move in reverse-as will be even more fully demonstrated in what follows-and carry with them `information' regarding the environment into which a particle is moving, all such forward-waves theories were necessarily nonlocal. The physical effects caused by the `information' carried by the reverse waves could only be accounted for through one kind or another of nonlocal interaction. This, I submit, is the real physical basis for Bell's theorem.

The basic error in current theory is that the waves are moving in the wrong direction.

3  Outline Of The Theory

Consider how the reverse waves might explain the double-slit experiment. Imagine that every point on the screen is continually emitting waves, having the same properties as the usual quantum waves. The waves in all directions from a single point on the screen are mutually coherent; but the waves from different points are mutually incoherent. The waves from a given point penetrate back through the slits, and the two wavelets leaving the slits toward the particle source then interfere with one another. Clearly, by reciprocity, the waves from a point on the screen corresponding to a `light' fringe (many particles reaching this point) would interfere constructively at the particle source. Waves originating from a `dark' fringe (no particles) would interfere destructively. Waves from an intermediate point would suffer partial destructive interference. Indeed, by reciprocity, the intensity of the wave reaching the source from a point on the screen is identical to the intensity of the usual forward moving quantum wave reaching that point on the screen, assuming identical intensity upon emission.

Suppose further that a particle is emitted by the source only in response to the stimulation of these waves, with the probability of emission being proportional to the intensity-the absolute value of the amplitude squared-of the waves. Suppose further that the particles, once emitted in response to the wave from a particular point on the screen, are causally determined to follow that wave to that point on the screen.

So particles reach the `light' fringes because the coherent intensity of the waves from those screen locations is a maximum at the particle source, and many particles are created in response to those waves. No particles reach the dark fringes because the waves from those fringes suffer destructive interference after penetrating the slits, and no particles are generated.

At every point on the screen, the number of particles arriving is proportional to the intensity, at the particle source, of the wave emitted by that point. But, again, that intensity is the same as the intensity at that point on the screen of the usual, forward moving wave, assuming equal intensity upon emission. And that latter intensity gives, in current theory, the probability that a particle is observed. So, if one can explain how the screen emits the waves, and how the particles then follow them, this picture would account for what is observed on the screen in exact mathematical detail.

The wave is present at all times, and not only when the particle is emitted. There is thus no problem in explaining why the wave is present when the particle `needs' it. And the wave doesn't have to `carry' the particle in any sense. The particle simply follows the direction from which the wave is coming (by a process which will be described in detail in Sections 8 and 9), following it back to its source, which it reaches with probability 1. No nonlocal interaction between an extended wave and the particle is required to understand how the particle follows the wave. The theory is both local and deterministic. Waves are waves and particles are particles, and both have an exact state at all times.

Because the particle follows the wave, the physics of the particle motion is determined entirely by the wave-which is why the wave computation determines the ultimate trajectory or set of trajectories. But it is the intensity of the wave at the source-the amplitude squared-that gives the probability that the particle is created so as to follow that particular wave. Thus we see in trivial physical fashion why one computes wave amplitudes and then squares to get the probability of the particle process. Rather than yielding the probability that a particle is somehow `created' at the screen out of the wave, the square rather gives the probability that a particle is created at the source so as to follow that particular wave in the first place. The square occurs at the source of the particle, not at the detector.

The particle might follow any one of many paths `through' the wave; but any one particle follows only one path. It makes no difference to the cross section-the probability that a particle reaches the particular point on the screen-which path the particle takes. The cross section is already determined at the particle source by the intensity of the wave reaching the source. All that is necessary to reproduce what is observed is that the particle, once emitted in response to the particular wave, reach the source of the wave at the screen by some path.

The particle travels through only one slit. The wave goes through both slits. But the wave goes through first, setting up the interferences, before the particle arrives.

With this picture one doesn't even need any `measurement theory' to understand what happens when the particle reaches the screen. The squaring of the wave takes place at the particle source;-and this makes perfect sense: one would expect the probability of particle emission to be proportional to the intensity of the stimulating wave. At the screen one simply sees the particle with probability 1. There is no wave function `collapse'; the wave is there all the time.

As discussed by Feynman[10], the double-slit experiment captures the entire essence of the `problem' of quantum mechanics. A theory which can account for this experiment in a local, deterministic manner should be able to account for all quantum phenomena in a similar manner.

Even if the picture sketched here is found to be incorrect, in the light of evidence from other experiments, the picture nonetheless provides, in principle, an explanation for the double-slit which is both local and deterministic. But according to the currently accepted view, this should be impossible. Clearly, by explicit example, it is possible. The various alleged proofs to the contrary all make, implicitly or explicitly, physical assumptions-in particular, the assumed forward motion of the waves. With changed assumptions, the proofs are no longer applicable.

The prescription outlined here for the double-slit experiment works immediately for any experiment in which particles are emitted by a source, penetrate through or scatter from a system of some kind, and are then observed at a detector. Each point on the detector emits waves, just as with the screen above. Those waves penetrate back through the system. By reciprocity-which applies generally to any kind of system-the intensity of the wave at the particle source is the same as the intensity of the usual, forward moving quantum wave at the same point on the detector, assuming equal emission intensity. So if particles are created at the source in proportion to the intensity of the reverse wave, and if those particles then follow that wave to the detector with probability 1, the probability of seeing the particle at the given point on the detector is exactly the same as in current theory.

So we see with complete generality why one adds amplitudes and squares to get a probability. We also see why it is the wave computation that yields the particle trajectories.

If no detector is present at a point along the path of the particles, then no reverse waves are emitted at that point. Rather, the waves originate from another detector or object further `downstream'-downstream, that is, from the point of view of the particles. So the dynamics of the wave at the given location when the detector is absent is determined by the first power of the wave function; the first power is what appears in the Schroedinger equation. But if one inserts a detector, the reverse waves now originate from that detector. So the probability of detecting a particle is given by the square-at the source-of the wave. But the square at the source of the reverse wave is the same as the square at the detector of the usual quantum wave. So we see why the first power determines what is present when one doesn't look, but the square gives what one sees when one does look.

And, just as for the double-slit, if a particle might take more than one path between the source and the detector, any interference is explained by the fact that the reverse waves take all paths. Each particle takes only one path. It is never necessary to have a particle in more than one location at a single time. Instead of particles being two places at once, one simply has two waves.

Perhaps the best feature of this picture is, again, that it requires no special measurement theory. When a particle arrives at a detector it is simply observed with probability 1 (assuming perfect detector efficiency). There is no wave function `collapse,' no transition from microscopic to macroscopic, or what-have-you. The wave needn't (nonlocally) disappear in order to prevent the generation of two or more particles by a single-particle wave, as in current theory. The wave simply remains, stimulating further particles from the source as long as the experiment lasts. It makes no difference that the wave leaves the system in directions other than the source, because there are no sources in other directions. Or if there were other sources, one would expect other particles.

We thus immediately have a theory which will account-locally and deterministically-for all single-particle experiments in which the apparatus through which the particle moves is static.² This includes the vast majority of quantum experiments.

And this has been achieved with virtually no change to the mathematics, and, in particular, with no additional variables, `hidden' or otherwise. The cross section for any process is determined by the reverse wave matrix elements, which are equal to the forward wave matrix elements but for a possible phase factor. All that is necessary to make perfect sense out of the current mathematics is simply to reverse the direction of the waves.

The `Schroedinger's cat' paradox[11] receives a trivial resolution in this theory. If a particle is going to change from one state to another, as in the decay of a radioactive nucleus, the waves for both states exist and interact throughout the process. There is never any need to assume that the particle itself is in both states simultaneously. The decisive moment-when the square is computed-is not at the point of observation, but rather at the point of emission. The act of observation itself-looking at the dial (or in this case the cat) as opposed to inserting the detector-plays no role.

The reverse wave picture immediately explains the phenomena associated with non-commuting observables. One can't measure two such observables because one can't simultaneously set up the reverse waves corresponding to both. The apparatus and/or detectors which would yield the waves for one variable destroy the waves corresponding to the other.

In current quantum mechanics, the value of a measured parameter cannot be viewed as existing prior to a measurement. In some way the act of measurement puts the system into the state measured. The reverse waves theory shows that this latter idea is, if anything, an understatement. The act of measurement affects the very creation of the particle in the first place.³ The particle which comes into existence at the source is determined in its state in part by the reverse wave, which wave depends in turn on the experimental apparatus employed. But now the sequence of events by which the act of measurement affects the particle makes sense. In current theory the affect of the measurement on the particle occurs at the detector. A mysterious, noncausal and nonlocal jump into the state determined by the act of measurement must take place, and prior to this the particle has (in general) no value in particular of the measured parameter. With the reverse waves the effect occurs at the source. Waves exist corresponding to all possible values of the parameter, and so the emitted particle might have any one of these values. But once created with a particular value of the parameter, the particle maintains that value at all times. There is thus no contradiction between the conclusion that the act of measurement affects the value measured, while simultaneously maintaining that the value of the parameter exists prior to the measurement-prior, that is, to the actual detection of a particle at the detector. There is no unknowable `jump' upon detection.

Consider the experiment with the atom in the resonant micro-cavity. The explanation as to how the cavity affects the emission is that the wave is emitted by the cavity. All different frequencies of waves are emitted by the cavity walls, but only those corresponding to resonance of the cavity will interfere constructively and remain in the cavity. Other frequencies suffer destructive interference as they reflect back and forth in the cavity. A photon will only be emitted if the wave of the proper frequency is present to stimulate the emission.

Notice that if the cavity is `tuned' to the atom, so that a photon can be emitted, then the `available state' into which the photon is emitted-as described in current theory-is, of course, mathematically identical to the photon wave itself when it is emitted. What the reverse wave theory proposes is that these available state waves are in fact real reverse waves emitted by the cavity. So, instead of having the wave emitted as the photon, rather the photon is simply a particle emitted in response to the already existing `available state' wave. The particle photon then follows that wave. The only change is that the wave moves in reverse, from the cavity walls to the particle photon source.

Mathematically, again, this is identical to current theory. The matrix element for the emission is identical. All that we have done is to say that the exponential factor in the matrix element corresponding to the emitted photon wave (in current theory) is instead the available state, reverse wave which stimulates the emission. We have simply changed the physical interpretation of the same mathematical expression. But now the causality of the process makes sense. The `weirdness' has been eliminated, but with no change to the mathematics.

Indeed, all particle emission, as described by current theory, requires the availability of a final state. But how can the mere availability of a state affect the emission process? In order for the available state to affect something, that state must be something itself. The mere `place' where something might go isn't anything in itself. So the very fact that the quantum description works, and requires an available state, serves as evidence that those available states are in fact something in their own right-something real-which are present in the environment before a particle is emitted.

And those available states, in order that they be able to affect an emitting system, must move toward that system-that is, in the direction opposite to that in which the particle will move when it is emitted. Any phenomenon involving an available state thus constitutes further evidence of the reverse motion of something.

Because all final state particles in any interaction require the availability of a final state, and because all initial particles were themselves final particles in some previous interaction, this prescription should work generally for all particle processes. Whatever the available final state is in current theory, simply reverse its direction, say that that wave stimulates the emission of the particle, and that the particle then follows that already existing wave. It will be shown in Section 8 that this prescription works in general to explain Feynman diagrams in a straightforward, pictorial manner.

It is clear qualitatively how the reverse waves picture will explain EPR experiments, or at least those without delayed choice. The reverse waves penetrate the polarizers before they arrive at the particle source, and thus carry with them `information' regarding the polarizer orientations. The particle photons are then created in a state that reflects the polarizer orientations at the outset. Bell's major premise is violated: the variables that describe the photons do depend on both polarizer orientations.⁴

I have spoken of the reverse waves as being `emitted' by the detector. This is not strictly correct. Particles can of course be emitted by a source into free space, and not simply in the direction of a physical detector. So if particle emission requires an available, reverse wave, the reverse waves corresponding to free-particle states must exist also. We must have reverse waves corresponding to all possible particle states. That is, waves exist corresponding to a complete set of quantum states. All such wave states must be filled by a wave, whether or not a physical detector is present to emit them. The postulate, then, is that a complete set of waves exists at all times.

It may seems far-fetched to postulate the existence of such a complete set of waves, moving in all directions, with all frequencies, corresponding to all kinds of particles. However, this picture is not substantially different from the usual classical picture of a lighted room. Electromagnetic waves of all frequencies-albeit with amplitudes that vary with frequency-and moving in all directions, fill the room. This is, in essence, the picture I propose for the reverse waves.

In the theory that I will develop through the remainder of this paper, these waves exist independently, in addition to the elementary particles. I will argue that they are primary constituents of reality on the same level as the elementary particles. In particular, they are not waves in any kind of medium. For this reason I will from now on call them `elementary waves.'

The elementary waves are real waves. They are not simply a mathematical fiction allowing one to obtain the correct answer for the particle process, or the like. They exist as real objects.

Because the waves are not waves in a medium, they do not propagate according to the usual dynamics of waves. In fact, as will be described more fully later, the description of their propagation is much simpler than that of the usual waves. They actually propagate much like a simple flux of material, with the material carrying a wave `implanted' in it, so to speak. However, the product looks exactly like a wave propagating according to the usual field equations.

Detectors and other particulate objects do not actually emit these waves. The waves are present continually, and with constant intensity. All that detectors-particles or combinations of particles in general-do is to establish mutual coherence among the waves leaving their vicinity. An organization is imposed on the already existing waves. It is the mutual coherence that then leads to the observed interference effects. I will continue to refer to detectors as `emitting' the waves, but this must be understood in this sense.

The quantity of wave material along any direction in space never changes, even in a `scattering' of the wave. All that happens at a scattering vertex is that the coherence of the incident wave flux becomes rearranged due to the interaction with the other waves at that vertex. When two waves interact, one wave might impose its coherence on the other. This gives the appearance that the second wave is the product of a scattering of the first wave; but in fact no actual scattering occurs.

The processes by which the coherence is imposed by a detector will be discussed in Section 9. But clearly the wave processes involved must correspond to inelastic particle processes. It is only by inelastic processes that we observe particles. So at the detector, wave processes occur looking exactly like the wave process at the detector in current theory-but in reverse. When a particle arrives at the detector while following the resulting wave, the particle continues to follow the wave as it scatters; the particle `mimics' the wave process in reverse. But it is specifically the inelastic processes that are relevant to a detector.

While the total wave intensity in any single-particle state is a constant, the wave can be divided into separately coherent `pieces.' A wave state can act as if it were empty by having its `pieces' arranged to be mutually coherent but out of phase with one another. This is what occurs in the resonant micro-cavity for the `mis-tuned' states.

The theory requires that the separate, mutually incoherent `pieces' of a single wave state act independently from one another-so one adds intensities at a particle source, not amplitudes. Also, `pieces' can be mutually coherent while still having different phases-one adds amplitudes, not intensities. Hence waves that are mutually coherent must be able to `recognize' one another, and waves that are not mutually coherent must also be able to recognize this fact. How this occurs will be discussed in Section 15.

All of the dynamics of particles are determined by the waves. The particle itself needn't carry any of the `classical' dynamical quantities generally attributed to particles: mass, momentum, energy, etc. All of these properties describe only the waves, with the particle then acting accordingly. Particles need carry only those parameters required for them to recognize and follow their wave. Of course it is by virtue of these latter parameters that a particle will follow only waves of particular characteristics; so in this sense one might say that the particle has mass or momentum or what-have-you. But the actual numerical quantity is carried by the wave. This, of course, accords directly with the mathematical description of `wave-particle' dynamics in current quantum theory. I will continue to describe a particle as having momentum or energy or etc., but this must be understood simply as meaning that it is following a wave with these characteristics.

Particles are emitted in response to waves of particular frequency/momentum. The behavior of the particle then reflects exactly the momentum of the wave. There is no `uncertainty' in the emission process. The process does not follow a `classical' model, in which the source `measures' the frequency of the wave and then emits a particle of appropriate momentum. In this latter model the momentum of the particle would be uncertain, given the finite time period during which the source would `measure' the wave's frequency. No such uncertainties are involved here.

Because the waves exist in their own right, there is no need to somehow obtain the laws of the waves from those of the particles, as is done in the usual canonical quantization procedure. It is from the observed behavior of particles that one determines the fact that the waves exist and what their properties are; but once one knows their properties, one simply says that the waves exist. There is no need to explain their properties from something else. Canonical quantization becomes entirely superfluous in this theory.

The full mathematics of the waves will be developed in Sections 7-9. However, with the above partial picture one can deduce a few more quantitative results of some consequence.

4  Einstein-Podolsky-Rosen Experiments

The elementary waves theory yields a quantitative resolution of the EPR paradox[4]. Consider again the experiments with photons, and in particular the experiment of Freedman and Clauser[2], pictured in Figure 2.

An atom decays twice in a J = 0 ® J = 1 ®J = 0 cascade, emitting two correlated photons in opposite directions, which then traverse polarizers and, if not absorbed in the polarizers, strike detectors. If the two polarizers are oriented an angle q apart, quantum mechanics predicts a cos² q dependence for observing coincidences (assuming perfect polarizers and detectors), a dependence thought not to be explainable in a local, deterministic manner[8].

In the elementary waves theory, waves are `emitted' by both the detector and the polarizer on both sides of the experiment. The detector emits waves of all polarizations. However, as these waves penetrate the polarizer, half are absorbed and the other half become polarized parallel to the polarizer's axis of transmission. In addition, the polarizer itself emits waves. Just as it absorbs photons that are polarized perpendicular to its axis of transmission, it emits only such waves.⁵ So two waves arrive at the photon source: the polarized wave from the detector and the perpendicularly polarized wave from the polarizer.⁶

When the waves arrive at the photon source, they stimulate the emission of photons. The two waves-one from the polarizer and one from the detector-are not mutually coherent because they arise from different sources; so they act independently in stimulating the emissions. When a photon is emitted in response to one of the two polarized waves, it follows that wave to its source-either the polarizer or the detector-with probability 1. No waves coming from the direction of the polarizer are present other than these two, so only photons following one of these two waves are emitted toward the polarizer.

Suppose the two polarizers are oriented an angle q apart, and the decaying atom is stimulated to emit the first photon in response to the wave that traverses its polarizer from the detector-which photon will thus itself traverse the polarizer and be detected. Now the atom wants to emit a second photon in the opposite direction with the same polarization as the first. But there is no stimulating photon elementary wave with this polarization, only one an angle q apart, coming from the detector on the other side, and another an angle q+ 90^° apart coming from the polarizer. Each of these waves might stimulate the emission of the second photon by the atom, but with a diminished probability. The amplitude of the first wave, relative to the needed polarization, is cosq, so the probability-proportional to the intensity of the stimulating wave-goes as cos² q. This gives the probability that the second photon will be emitted in response to the wave that traverses the other polarizer-and hence the probability that the particle photon will do the same, and be detected. So the probability of coincidence is exactly the result predicted by quantum mechanics. The stimulating photon elementary wave at angle q+ 90^° will, with probability sin² q, create a photon which is then absorbed in the polarizer.

The key to making sense out of the cos² q dependence is that the square occurs at the source-which in turn results from the reverse motion of the waves.

To be strictly accurate, it isn't the case that the atom emits two photons in separate processes. As will be explained more fully in Section 12, the cascade is actually a single quantum process, for which a single overall amplitude needs to be computed. An overall wave interaction, involving both photon elementary waves, occurs before either particle photon is emitted. Current theory obscures this point, because it makes the particle into the wave. Thus the electron waves corresponding to the middle and lower level in the emitting atom don't exist until the jump into those levels occurs. There is no way that the interactions corresponding to the emission of the second photon can begin until the first photon has been emitted. But in the elementary waves theory the waves for all three levels exist at all times, and the interactions corresponding to the cascade thus also exist at all times. Whatever cascade occurs, the corresponding overall wave interactions were taking place prior to the emission of the first photon. If the photon waves have polarizations that are orientated an angle q apart, then the amplitude goes as cosq. Hence the probability of the two particle process goes as cos² q. The above two-step description, though inaccurate, is offered here to help in visualizing the origin of the cos²q factor.

Current theory is actually inconsistent on this point. If the two steps in the cascade were actually independent, then the resulting photon waves would not be mutually coherent/entangled. To be entangled, as per current theory, one must have a single amplitude. But there is no mechanism for this if the particle is the wave. The mathematics can be made to `work,' but the theory is inconsistent.

Notice, however, that the photon waves on either side of the photon source in the above elementary waves explanation for EPR are not in any way `entangled.' Each wave is simply a plane wave (approximately) with phase determined solely by the detector or polarizer from which it originated. The effects of the two waves at the source are, one might say, `entangled'-that is, the emission of each photon is affected by both waves-but not the waves themselves.

Entangled wave functions are necessary in current theory because of the forward motion of the waves. The actual `entanglement' occurs at the photon source, as just indicated. But that entanglement must, mathematically, be present at the location where the square is performed. With forward moving waves the squares occur at the polarizers and detectors, not at the source. So in order to make the forward wave theory work, the waves must be entangled-with subsequent `collapse'-in order to carry the entanglement from the source to the detectors. With reverse waves no wave entanglement is necessary. Each wave is simply an independent, single particle wave. As will be demonstrated in Section 11, quantum statistics in general can be accounted for without wave entanglements.

Wave entanglements are generally viewed as being essential to the description of identical particle phenomena and to the entire structure of quantum mechanics; so it may strike the reader as absurd to try to account for multi-particle effects without them. But certainly the above EPR experiment is one instance where the effects in question are manifested. And, using reverse waves, as just demonstrated, the correct result is obtained with no entanglements. It really is only the erroneous forward wave motion that gives rise to them.

The elimination of wave entanglements constitutes the only major change to the mathematical formalism (at the matrix element level) of quantum mechanics that is required by the elementary waves theory. But clearly this change represents a major simplification. In general, no multi-particle states are necessary in the theory.

Such multi-particle states are, or course, nonlocal in their behavior. One would thus expect them to disappear in a local theory.

Actually, the independence of the elementary waves from different detectors or from different points on a detector is a general property of the theory, even for single particle phemonema, as indicated above. In current theory, the wave arriving at various points of a detector, even for a single particle wave, must be treated as a single coherent wave. It is the self-interference of this single wave from the source that produces the various quantum wave effects. That wave then `collapses' (nonlocally) when the particle is observed. In the elementary waves theory the interference occurs in the reverse direction; the wave from each point on the detector interferes with itself at the source. There is no need for the waves from separate points on the detector to interfere with one another. The processes connecting the source with different detection points are-for single particles-entirely independent.

The two elementary waves which actually stimulate the two photons in this example do not have parallel polarization for a general angle q. This might appear to contradict the finding, using the present theory, that the two photons are emitted with the same polarization. However, if the polarizers are parallel, one will, in the elementary waves theory, always see both photons or neither photon. The probability that one photon will be stimulated by the wave along the polarizer axis, and hence be observed, while the other photon will be stimulated by the perpendicular wave from the other side, and then not be observed, is cos² 90^° or zero. When the two polarizers are oblique, the waves stimulating the emitted photons are oblique; but this doesn't contradict what is actually observed experimentally. Whatever angle one uses, the probability of coincidence is exactly that predicted by current quantum theory.

Furthermore, as explained in the previous section, there is no need in this theory to assign any `spin' to the particle itself. All of the spin behavior is captured by the waves-which, again, is exactly what the mathematics of current quantum mechanics says. The waves act as current theory describes, and the particle then `blindly' follows. Spin, thereby, acquires a simple, pictorial explanation.

It is necessary, however, to explain delayed choice situations, and, in particular, the experiment in which one polarizer, initially oblique to the other, is rotated back into alignment with the other polarizer after the photon pair is emitted. If the wave `spins' are actually oblique, for oblique angles between the polarizers, and we then rotate the polarizers into alignment with each other while the particles are in flight, does the theory still predict the correct answer? Indeed, does it predict the correct answer for delayed choice situations in general?

Consider a photon in flight from the source toward its polarizer. The polarizer is rotated, destroying the original elementary waves and creating new ones. The new waves arrive at the photon while it is somewhere between the source and the polarizer. But a particle must always follow an existing wave-the wave by which it was generated; it is that wave that determines the dynamics. If that wave disappears, the photon must jump into `coherence' with one of the new waves; it can't remain in its original state because the corresponding wave is gone.⁷

The jump of a particle into a new state while out in space, not interacting with any other (local) particles, might appear strange and/or arbitrary at this point. But when the process by which a particle follows its wave is described in more detail (Sections 8 and 9), it will become clear that this is necessary, and fits directly into the overall theory. The only observable effect of the jump is the subsequent interaction with the newly orientated polarizer. But, as we will see in a moment, the theory being offered predicts exactly the same results of that interaction as the current theory (for single delayed choice)-which of course agrees with what is observed experimentally.

Furthermore, current quantum theory actually predicts exactly the same phenomenon, although it is not pictured as such. Consider a `wave-particle' in a definite state in a box. If the box is changed, the `wave-particle' immediately jumps into a superposition of the newly available states. In Section 12 it will be shown that the mathematics describing the jumping process in current quantum theory is identical to that describing the jump in the elementary waves theory.

And remember here, again, that the particle photon itself does not carry a spin. Only the waves carry the spin; the particle then acts accordingly. So there is no need to conserve any angular momentum of the particle photon when the `jump' process occurs. Conservation of angular momentum is required only for the waves.

The detailed physics of a `jump' is rather complex, but follows the pattern of the theory established up to this point. The `jump' of a photon involves the annihilation of the initial photon and the creation of a new one. The annihilation of the initial photon is equivalent to the creation of an anti-photon-that is, another photon-moving in the opposite direction. But in the elementary waves theory all particles are created in response to waves. To be consistent, this would have to include the effective (anti-)photon involved in the jump. Because the (anti-)photon moves in the opposite direction, it is emitted, in effect, in response to a wave coming from the direction of the photon source-a wave, that is, which is moving along with the initial photon.⁸ The new photon, which continues on to the polarizer, is emitted in response to the new waves coming from the polarizer. In effect, a pair of photons is created in response to the waves coming from opposite directions.

But a similar process occurs upon the initial emission of the photon at the source. An electron in an atom scatters and emits the photon. The electron is a point-like particle following a wave, as is the photon, so the emission occurs at a single vertex.⁹ At that vertex, the scattering electron looks to the photon exactly as if another (anti-)photon had been created. The scattering electron can emit a photon, so it is the equivalent, electromagnetically, of an (anti-)photon. So what one has, in effect, is again the creation of a photon pair, with the (anti-)photon corresponding to the electron scattering.

But-again as with all particles in this theory-the effective (anti-)photon is emitted in response to a wave. Because the (anti-)photon is absorbed by the electron, that wave must come from the electron.¹⁰ This (anti-)photon wave captures the spin orientation of the emitting atom-of the scattering electron. It is the angle between the polarization of this (anti-)photon wave and the photon wave coming from the polarizer that gives one the cos²q.¹¹

Furthermore, because the photon and the effective (anti-)photon move in opposite directions, the (anti-)photon wave from the electron moves in the same direction as the photon. That is, it travels with the photon. So it is this very same wave that is present when the photon `jumps' later on. Hence, the `jump' occurs exactly as it would have occurred had the new waves from the rotated polarizer arrived at the photon source before the initial emission. The photon pair process at the jump is exactly the same-in response to exactly the same waves-as that which would have occurred at the source had there been no delay. The result is exactly as if no delay had occurred.

In general, all particles will be accompanied by the wave which affected the (effective or actual) anti-particle involved in their emission, for reasons to be explained in Section 12. So if one changes the wave being followed by that particle in a delayed manner, and a jump to a new state occurs, the same pair process takes place that would have taken place had the new waves arrived before the particle's initial creation. The result is exactly as if there had been no delay.¹²

Notice, then, that it is not necessary in general for a wave to make the entire trip from detector to source in order to understand quantum processes. If the wave changes while the particle is in midflight, the particle jumps into exactly the state it would have been in had the change occurred before the particle's creation at the source. With this fact one can understand how the elementary waves theory explains dynamic, changing systems as well as the static systems treated in Section 3.

There is one circumstance, however, in which the predictions of the elementary waves theory differ from those of standard quantum mechanics: double-delayed-choice experiments, in which both polarizers are independently rotated after a photon pair is emitted. When this occurs, each photon jumps into a new state with a probability which depends on the original orientation of the opposite polarizer, not its new orientation. The respective anti-photon wave involved in each jump will reflect the initial wave from the polarizer on the opposite side, and not the new wave that appears after the rotation. So one will no longer obtain the quantum mechanical cos² q form.

It might be thought, then, that the experiment of Aspect[5] refutes the elementary waves theory. However, Aspect did not simply change each polarization once in a delayed manner. In his experiment each polarization was switched rapidly back and forth between two particular polarizations, using an optical commutator. Furthermore, the distance between each commutator and the photon source was chosen as twice the distance D that light can travel in the time that the commutator remained in one condition[5]. As a photon travels from the source to the commutator, in the elementary waves theory, it will experience changes in its elementary wave due to the commutator switching. But a quick check shows that, because of the above factor of 2 in the distance, when the photon arrives at the commutator, the commutator will always be in the same condition that it was in when it transmitted the wave that stimulated the initial emission of that photon. So even though the photon might have jumped back and forth between the two different wave sets as the commutator switched, it will end up in the same state at the commutator that it was in when it was emitted, and the commutator will be in the same state that it was in when the wave was transmitted. The net result will be exactly as if the commutator had never changed. The factor of 2 nullifies the effect that was to have been observed in the experiment.

In order to serve as a test of the elementary waves theory, the distance from commutator to source would have to be a half-integral multiple of the distance D. If the distance were, say, 2 ¹/₂ times D-that is, if the separation between the two commutators were 5 times D-the experiment would be a valid test. I predict that if the experiment is repeated with the half-integral separations, it will not reproduce the present quantum predictions.

In all respects other than double-delayed-choice, this explanation of photon EPR exactly reproduces the predictions of quantum mechanics. Some of the mathematics is different, due to the fact that no `entangled' wave functions are required; but otherwise the mathematics of the waves is identical. The theory is local and deterministic-both the waves and the resulting particles follow local, deterministic laws. No nonlocal `collapse' of an entangled wave at the polarizers is involved. The key is the fact that the square occurs at the source, at which point the decaying system has information regarding the orientation of both polarizers.

There is some unpredictability-as opposed to indeterminism-in this theory, in that we do not know in advance which wave the source will respond to in emitting a particular particle photon. However, unlike the situation in the usual theory, here the unpredictability can be described as resulting from a random process following an ordinary probability distribution. All of the wave states exist as real waves. The source then simply has a constant probability of responding to the intensity of each incident wave. The randomness thus reflects lack of knowledge of the value of some parameters in the source, rather than representing a fundamental indeterminism.

The `hidden variables' must, mathematically, come into play as part of the event at which the squaring of the wave is performed. In current theory this is at the detector. But in fact the hidden variables are in the source, not the detector.

The explanation of EPR experiments using particles other than photons, or of experiments involving parameters other than spin, exactly parallels the case for photons.

Whether or not the elementary waves theory is correct, this theory of EPR experiments clearly constitutes a counter-example to the conclusions usually drawn from Bell's analysis[1]. It is indeed possible to explain EPR experiments with a local, deterministic theory. Bell's theorem, coupled with the experiments confirming the associated quantum mechanical predictions, does not `refute reality,' as is so frequently claimed.

5  The Uncertainty Principle

The elementary waves theory explains the appearance of `wave-packet' phenomena, and hence gives a physical explanation for the uncertainty principle. Perhaps the best way to picture this is with the experiment of Kaiser et. al.[13], illustrated in Figure 3.

This experiment employs a Werner-type[4], crystal neutron interferometer, with a bismuth (Bi) sample in one arm to delay the beam. An analyzer crystal is also placed in front of one of the detectors in an exit beam to select a narrow wave band from the wider band-width that is otherwise accommodated by the interferometer. If the Bi sample is made large enough, interference, in the absence of an analyzer crystal, disappears. The explanation given by the current theory is that the wave-packet is not long enough to maintain the coherence-the coherence length is too short. Interference disappears because, with the delay due to the bismuth, the packets travelling on the two arms of the analyzer no longer overlap at the final crystal plate.

However, with the analyzer crystal in the exit beam one can narrow the observed band-width further. And, as if by magic, the interference returns, even with the larger Bi sample in place; varying the width of the Bi sample now makes the beam reflected by the analyzer crystal come on and off. This is interpreted by Kaiser as implying that the subsequent action-after traversing the interferometer-of narrowing the band-width affects the prior band-width of the wave-packet and hence its coherence length-one of many examples of reverse-temporal causality in current quantum mechanics-which, of course, makes no sense.

In the Kaiser experiment, what is actually happening is that each detector is emitting elementary waves back through the system at each frequency in the full band-width that the interferometer will accommodate. With no Bi sample inserted, the interference-now at the first plate-of the waves is such that all frequencies within the band-width exhibit the same interference. Thus all of the waves from a particular detector will go one way-for a perfectly aligned interferometer-as they leave the analyzer, either toward the particle source or along the other direction. What we have is exactly the usual quantum interferometer, but with the waves moving in reverse. Thus all particles from the source will go one way in the end-toward the detector which emitted the waves that reached the neutron source and that the neutrons are thus following.¹³ As one inserts a little Bi, all frequencies still exhibit the same interference. But with enough Bi, different parts of the band-width begin to exhibit different interference. The delay due to the bismuth creates a different phase shift depending on the wavelength of the wave. Some parts thus go one way and some the other, and the interference is washed out. Waves from both detectors arrive at the source, and so particles then arrive at both detectors.

However, if one inserts the analyzer crystal to single out a narrow band of the frequencies, the interference is found to still be there. That is, all of the elementary waves emitted by the detector behind the analyzer crystal and then selected by the crystal will interfere in a common manner when they reach the first plate of the interferometer. The bandwidth selected by the analyzer crystal is now too narrow for the bismuth to produce phase shifts which differ enough from one another to produce a significant effect. Hence one will either see particles or not, depending on the particular frequency band that has been selected by the analyzer crystal.

One can imagine performing the Kaiser experiment with a large, fixed Bi sample, but with a variable analyzer crystal. As one swept across the wide band-width with the narrow-range analyzer, the observed particle beam leaving the analyzer would go on and off. But these peaks and valleys are mixed together in the exit beams when the analyzer is removed-which is why one sees no apparent interference.

Quantitatively the relationship between Dx and Dp is the same as in current theory. As an approximation, describe the band-width accepted by the interferometer as having a width Dl centered on wavelength l, with all frequencies within this width having equal amplitude. Then the interference will be completely wiped out when the bismuth causes the waves at one end of the bandwidth to shift by 2 p relative to those at the opposite end. For each wavelength l by which the wave is delayed by the bismuth, the shift of the two extreme waves relative to one another will be Dl. So to get the full shift of l one needs a number n of wavelengths given by

But the momentum is given by

So one sees how the wider band-width gives a shorter apparent coherence length, and hence the appearance of a shorter wave-packet, and vice versa. The accuracy of the `knowledge' of the frequency/momentum is thus inversely proportional to the accuracy of one's `knowledge' of the position-which is the uncertainty principle.

But this entire `uncertainty principle' way of looking at things is necessary only in a theory which holds that the particle is the wave. With the elementary wave picture it is clear that there is no actual uncertainty at all. Indeed, there are no wave-packets at all. Every individual wave frequency acts independently from all others, and every particle follows its own individual wave.

Remember that in present quantum theory one can treat a general scattering process either by individual frequency waves or by wave-packets. The results are identical. There is no need to have any `glue' to stick the various frequencies together in a packet. All frequencies act with complete independence.

What forces one to assign a fundamental uncertainty to particles in current theory is the forward motion of the waves. By assuming that the wave goes from source to detector, and that the wave is the particle, one is forced to conclude that the particle exists in multiple states simultaneously in order to explain phenomena involving `widths.' But with the correct direction of motion one can understand the phenomena of `widths' without the need for any uncertainty in any parameter-without the need to assume that the particle itself was in all of the states in the width simultaneously. Only the waves were in all of the states, not the particle. And the existence of waves in all the states merely means that there was more than one wave involved, not that a single wave was in multiple states. Each wave is in one state at one time, as is the particle.

The exact value of the particle momentum is unpredictable. We don't know which wave will lead to the emission of a particle at which time, and hence don't know in advance the value of the parameters describing a particular particle. But this is now due solely to ignorance of the value of parameters in the emitting system and not to any fundamental uncertainty.

There must indeed be such parameters in the emitting system to explain why it reacts to one wave rather than another, as indicated earlier. And these parameters are additional to those in standard quantum mechanics. They thus do constitute `hidden variables' in the usual sense. But it is clear now that they create no conceptual difficulties.

All `unpredictability,' as distinguished from `uncertainty' in the usual quantum mechanical sense, is now explained as resulting from lack of knowledge of the values of parameters in the particle source. Hence there is no need to conclude that there is any lack of strict determinism. The `uncertainty principle' is thereby explained.

Further investigation will be necessary to determine the nature of the parameters involved. However, the fact that some such parameters can in principle account for the unpredictability of quantum phenomena has been demonstrated.

As an aside I must say that the notion of `hidden' variables of any kind is a misnomer. If a variable really were hidden, this would imply that it had no observable consequences, in which case one would never know of its existence-it would play no role in any theory. Indeed, a proper empiricism dictates that any such `variable' would be entirely meaningless. If a variable has any observable consequence, then, by that very fact, it is not hidden. The `hidden variables' in the source above clearly do have observable consequences: the emission of one particle rather than another, and at a particular time. They are therefore clearly not hidden. A more correct designation would be `more indirectly observed variable.'¹⁴

`Tunnelling,' usually though of as an expression of the uncertainty principle, is simply explained by the elementary waves picture. The dynamics of particle motion is determined by the waves; and the waves obey the same laws as in current theory. In current theory, the waves `tunnel.' So the elementary waves also tunnel-and the particles follow those waves. No uncertainty in the particle state is involved. This picture of tunnelling will become clearer after the details of the process by which a particle follows its wave are presented in Sections 8 and 9.

6  Special Relativity

The elementary waves theory provides a simple, physical explanation for the fact that light travels at the same velocity c relative to all observers, and thus serves to explain the Lorentz transformation.

According to the theory, all particles obey a dynamics in which they follow a wave coming from the `detector.' This is true of particle photons also, as seen in Section 4. Whenever we see a photon, our eye becomes the `detector.' What we see is the particle photon, not the wave. It is the particle that imparts any energy or momentum to the retina, thus producing a visual effect.¹⁵ The same is true for any other object or `detector' that absorbs a particle photon. But if the dynamics of the particle photon is determined by a wave that comes from the observer, then it is the observer's frame that determines the velocity. The constancy of c relative to the observer is thereby explained.

The elementary waves are not actually emitted by the observer, as indicated earlier. The observer merely rearranges the organization of the passing wave. It must be assumed, then, that the organization is imposed in such a manner as to reflect the frame of reference of the `emitting' particle. The particle photon which responds to that organization will then travel at velocity c relative to the `emitting' particle. I will discuss this further at the end of the next section. For the moment, simply imagine that the waves actually are emitted by the observer, with the observer's frame thereby determining the dynamics. I will show that the actual situation is equivalent to this.

This explanation of the constancy of c, as will be shown below, does not require that a single wave travel the entire distance from an observer to the source of any photon seen by that observer-a proposition that would clearly be absurd for, among other things, intergallactic light. This need be true only for light observed locally-that is, for those distances at which our basic, directly perceivable units of length and time are established. The behavior of particle photons over long distances will be shown to be exactly the same as if a single wave made the entire trip.

Light, then, doesn't simply move from object to observer, or from observer to object; it does both. Nor is it simply a wave or simply a particle. It consists of a wave from observer to object, and a particle from object to observer. However, the fact that a wave travels from observer to object does not make this an `extramissive'[14] light theory-one in which light travels from observer to object. The light that is observed is the particle photons, which travel (`intromissively') from object to observer.

`Relativistic' phenomena can thus be understood without the requirement that space be a physical object of some kind that stretches and shrinks as we change frames of reference. What changes when one changes frames is only the light used to observe objects.

However, the fact that space does not change does not mean that one can dispense with Lorentz transformations and use simply Galilean transformations along with the change in the light. To see why Lorentz transformations are still necessary, and why this does not conflict with the claim that space is unchanging, consider the following example.

Imagine for a moment that space-time were Galilean, and consider the experiment pictured in Figure 4.

Two lamps in the same frame of reference flash at the same instant as observed in that frame. An observer at the midpoint, and also in the same frame, will observe the light from both lamps at the same instant. The light moves at velocity c relative to the observer.

A second observer is in a spaceship moving rapidly in the direction from one lamp to the other. The timing of the ship's motion is such that the light from the lamp behind the spaceship arrives at the ship just as it passes the first observer. But that light is moving with velocity c relative to the spaceship (the photons observed by the spaceship are following waves from the spaceship) and thus at velocity c + v relative to the first observer-this, again, in our imagined Galilean universe. Light from the second lamp, similarly, moves at velocity c - v relative to the first observer. Clearly, then, the light from the second lamp will not arrive at the spaceship at the same time as that from the first lamp; the light travels equal distances but at different velocities. But to the spaceship both light signals move with velocity c, and the distance travelled is the same (or would be if both signals reached the space ship at the midpoint). So if the lamps fire simultaneously as viewed by the spaceship, the two flashes would be observed simultaneously. Because they are not, we must conclude that, to the spaceship, the flashes do not occur simultaneously, even though they are simultaneous to the first observer.

It is not simply the case that the lights appear to flash at different times. Even with a correction for the time of flight of the photons, the actual flashes of the lamps occur at different times as viewed from the spaceship.

We are thus forced to conclude that simultaneity is relative-even within this initially Galilean framework. But if simultaneity is relative, then so is length, as this is usually defined. If an object is moving, by its length we mean the distance, in the observer's frame, between the positions of the two ends of the moving object observed simultaneously. So given the relativity of simultaneity, we see that length will be relative also.

Indeed, given the constancy of c-regardless of the physical reason for it-one can deduce the full Lorentz transformation.¹⁶ The steps are directly parallel to current standard derivations. I will merely refer to two of them briefly in order to identify a few points of difference.

In one standard textbook derivation[15], a flash of light is observed from all directions in two frames of reference, one moving relative to the other. Coordinate systems are defined in the two frames so that the two origins coincide with one another and with the light source at the moment of the flash. In both frames the light is observed to travel out from the origin in a spherical pattern, due to the constancy of the velocity of light relative to all observers. The Lorentz transformation is then derived as the transformation necessary to produce the light seen by observers in one frame from that seen in the other.

But according to the elementary waves theory, exactly these two spherical pulses is what would be seen by observers in the two frames. Imagine an array of observers in each frame, placed around the origin, but intersperced so as not to block one another. The light seen by each observer will move with velocity c relative to that observer, because it is that observer's own elementary waves that will determine the velocity of the light he sees. The light will thus be seen by both arrays of observers as moving in a spherical pattern with velocity c. So the light seen by one array of observers is exactly what one would obtain by applying a Lorentz transformation to the light seen by the other array. The elementary waves theory thus predicts exactly the relationship captured by the Lorentz transformation.

In the standard derivation it is assumed that the light seen by both observers is physically the same light-the same photons. Space and time are then distorted in order to account for the fact that both observers see a spherical pulse. In the elementary waves theory observers in both frames still see a spherical pulse. But this is because the light is different, not because of a deformation of space-time. The two observers in two different frames do not see the same photons (this, again, for local observations where our units of space and time are established.)

In this derivation it might appear as if what we have with the elementary waves is simply Galilean space with a change to the light. So it would be instructive for the reader to follow through another standard textbook derivation of the transformation, namely, that of Panofsky and Phillips[16]. Every aspect of that derivation remains the same except for one change. The transformation for a time interval is derived by considering light that travels from a source to a mirror where it reflects and then returns to the source, this as observed first in the source's frame and then in a frame moving in a direction perpendicular to the light's direction of propagation. In this derivation, the mirror used when the light is observed from the moving frame must be fixed in that moving frame, not in the frame of the light source. The light must consist entirely of light as it would be observed in the moving frame; so the mirror must be in that frame in order to emit the corresponding elementary waves. However, given that we obtain the Lorentz transformation anyway, the result is the same either way; the light arrives at the moving observer at the same instant regardless of which mirror is used.

It is clear in this latter derivation that lengths do in fact change when one changes frames. So, even though the space itself does not change, one nonetheless must use a Lorentz transformation to relate what is observed in one frame to what is observed in another. If this seems to be a contradiction, remember that a coordinate system is not the same thing as space. A coordinate system is a real object or imagined real object in space. An axis is the equivalent of a real ruler. A length measurement, as with all measurements, is not a measurement against some absolute standard, whatever that might mean. It is rather a comparison between two extended objects, one of which is taken as a unit. When one measures objects by comparison with a coordinate system, one is similarly comparing two objects. But if objects appear differently when moving, due to the change in the light used to observe them, the same will be true of coordinate systems. The coordinate system used in one frame, if viewed from another moving frame, will not look the same as the coordinates that one would use in that moving frame.

Given the fact of Lorentz transformations, all of the consequences of that transformation occur in the elementary waves theory exactly as in current theory. Moving objects appear shorter, time intervals in moving systems appear dilated, and etc. However, none of these apparent changes require any change to the objects themselves. Only their appearance changes, due to the change to the light.

What we call the length of an object when viewed from a moving frame-the distance between the end-points observed simultaneously-is physically not the same thing as the length in the rest (or any other relatively moving) frame. Because simultaneity is relative, if one wants to get the same physical quantity in the moving frame one would have to use non-simultaneous times. It is only if one mistakenly holds that the `length' in the moving frame is the same physical quantity as the length in the rest frame that one will think that a moving object has shrunk. A moving object doesn't shrink.

The invariant quantity-the actual, objective nature of the object observed-is exactly what current theory says: the invariant interval. That interval is not simply mathematically equal in all frames, it is physically the same thing. The interval appears to change physically, because the `mix' of space and time is different in different frames. But this is entirely due to the change in the light, not to a change in the nature of the interval. All observers see the same reality.

The very definition of length, for a moving object, involves time-simultaneity-as indicated above. And the very definition of time involves length. Time is the measure of motion. It is by comparing motions-over distances, or lengths-that we arrive at a concept of time. So it should come as no surprise that the two concepts end up being `mixed' together as in the Lorentz transformation. It is specifically the motion of two observers relative to one another that affects the means of observation. But motion means length over time. The surprise, then, would be if there were no `mixing.' But the fact that lengths and times change under transformation does not mean that an object itself changes.

What we have traditionally called length and time are inextricably tied up with the nature of our (principal) means of observing objects: light. The nature of light as a particle following a wave from the observer dictates that simultaneity is relative. This in turn forces us to use Lorentz transformations, even in a space that is unchanging. Lengths and times thus become `mixed.'

It is thus clear that there is no contradiction involved in the fact that two observers in relative motion each see objects as being shorter in the other observer's frame. The apparent `shrinking' effect is reciprocal. Similarly for time dilation. The `twin paradox' in its various forms is thereby resolved.

The elementary waves theory of `relativistic' phenomena is an objective theory of those phemonena. Reality is the same for all observers. It is not the case that `everything is relative.'

What, after all, do we mean when we speak of what exists objectively, independent of our means of observation. It means that, whatever means of observation we use, we subtract its effect from what we see in order to determine what was due to the object itself aside from the method of observation. Ordinarily we think that, for visual observations of position, this can be accomplished simply by taking into account the velocity of light. We notice when we see the light pulse, we take into account the time it took the light to travel, and we then determine where the actual emission occurred and when. But this ordinary means of removing the effect of the light is actually premised on a Galilean view of things. This doesn't actually remove all of the effects when we observe a moving object. To completely remove the effects of the light requires-exactly as we just showed above-that we do a Lorentz transformation. The use of different light doesn't just mean that the velocity changes. Also apparent distances change, time intervals change, simultaneity changes, etc.

The difference between what two observers see, as the result of using different light, is exactly described by a Lorentz transformation. So that, exactly, is what we must perform to remove the effect of the use of different light, thus insuring that what remains is physically the same for the two observers. Because what remains is physically the same, it must act the same. Hence all physical laws must be `covariant.' Covariance, then, simply means that one has removed the effects of the means of observation.

Space, after all, is nothing. Space is merely the place where real objects can be located. What is real are the objects, not the space. We arrive at our concept of space by abstraction from real objects. So space as such, aside from the objects located in that space, can be neither Galilean nor Lorentzian, nor have any other special properties. Nothingness can't have properties. If we assign any properties to space, what we mean is that these are properties that would be possessed by any object that might be located in space. If all objects transform in a Lorentzian manner, one might then say that space-time is Lorentzian. But this must not be understood as implying any modifications to the space as such. Nothingness can't be modified.

It will be demonstrated in Section 14 that general relativity also can be understood without attributing `curvature' or other properties to space as such.

The elementary waves theory is `automatically' relativistic-it is already relativistic as it stands. It is not necessary to add relativity to a non-relativistic theory. Had relativistic phenomena not yet been discovered, the elementary waves theory would have predicted them. I offer this as the single most significant piece of evidence supporting the theory. The same theory which explains quantum phenomena, immediately-with no further assumptions-predicts and explains special relativity.

Quantum mechanics and relativity are, indeed, one and the same theory. This explains the `intimacy' between quantum mechanics and relativity that was discovered when quantum mechanics was made relativistic.

With the insight that it is something moving from the observer that produces relativistic effects-namely the elementary waves-we see what is not obvious from the current presentations of relativity theory: the theory is-in those current formulations-a thoroughly nonlocal theory. If observed from a moving frame, an object is shorter. It doesn't just look shorter, it actually is shorter. So if one gets up and moves across the room, the fact of one's motion causes every (initially stationary) object in the universe, to its farthest reaches, immediately to shrink. It would be hard to imagine a more nonlocal theory.¹⁷

Indeed, turning this argument around, the fact of relativistic phenomena is the single largest piece of evidence that something must be travelling from the observer/detector to the particle photons. Without this there is no local means of understanding how objects change-or appear to change-when one moves. This, then, must be added to the list of evidences of reverse motion in Section 2.

Objects do indeed appear to change when one moves. But facts are facts; facts don't change because one looks at them differently. So one knows for certain that it is the means of observation that changes when one moves, not the objects observed. But motion of the observer can affect the means of observation only if the means involves something travelling from the observer.

Therefore, rather than starting with quantum phenomena and applying the `pre-quantal' philosophy, one might just as well have started with relativistic phenomena and applied the same philosophy-a philosophy that one might then call also `pre-relativistic.' From this one would deduce the fact of the reverse motion of the waves; and then from that fact one would explain quantum mechanics. Relativistic phenomena alone provide a sufficient basis to deduce the elementary waves theory, at least for photons-provided one maintains the view that facts are facts.

7  `Relativistic' Transformation Of The Waves

Consider a particle following its elementary wave as in Figure 5.

The particle moves to the right, the wave to the left. The energy-momentum of the particle is related to the wavelength and frequency of the wave in the usual manner.

Suppose we transform to a system moving to the left as shown. The particle will be moving faster in the new system, and hence should have a larger energy-momentum. In order for the theory to be invariant, the wave must similarly transform to a wave of higher energy-momentum. Otherwise a particle of one energy-momentum would appear in the transformed system to be following a wave of the wrong wavelength. But this can only happen if the wavefronts are moving to the right, with the particle. We thus seem to have a contradiction: the wavefronts move to the right, but the wave moves to the left.

However, the wave is present at all times. The effect of the particle or particles `emitting' the wave is not to generate the wave-there is no oscillation of the source as in usual wave emission-but rather is to establish coherence in the already existing wave. Furthermore, the phase velocity of a particle wave is given by c²/v, which is an unphysical velocity anyway. So it can't be the case that the wavefronts carry the wave signal, as would be the case for a wave in a medium.¹⁸ The coherence signal and the wavefronts must propagate independently.

Remember that the usual resolution of the problem of unphysical phase velocities-using group velocity for a packet-is no longer applicable in this theory. Here there are no packets, as demonstrated in Section 5.

But if the wavefronts don't carry the signal, then there is no reason why the wavefronts might not move in either direction relative to the signal propagation.

As an analogy, consider an infinitely flexible, stretched string which can move in either direction along its length, but which at the same time can oscillate in a direction perpendicular to the string, this independently at every point along the string. With appropriate coordination of the oscillations at each point, one can make the string look like a wave at any instant of time, and the wavefronts can be made to travel in either direction at any velocity, including velocities greater than c. But such wavefronts would not carry any information or signal, due to the infinite flexibility-due, that is, to the causal independence of the motion of each point of the string. Any signal is carried by the string itself as it moves along its length.

The elementary wave objects are like the moving string. They are not waves in a medium. The elementary waves are the medium-they are the `material' filling otherwise empty space. They move with velocity c (as will be shown in a moment), and the phase velocity can be in either direction relative to this actual velocity. There is no propagation of a signal through the wave. Rather, the wave object itself moves with velocity c, and thereby carries whatever coherence has been implanted on it. The coherence velocity-the velocity with which the coherence implanted on the wave travels-is the actual velocity c of the wave object.

As strange as the notion of phase velocities being in reverse might seem, we will see in Section 9 that this is essential to the understanding of Feynman diagrams.

An elementary wave with phase velocity in the opposite direction from the velocity of the wave object I will call a `positive phase velocity wave,' or a `positive phase wave' for short. If the phase velocity is in the same direction as the wave object it is a `negative phase wave.' The wave in Figure 5 is thus a positive phase velocity wave. Even though the wave object moves to the left, the phase velocity is to the right, with the particle.

Negative phase waves must not be confused with `negative frequency waves.' The latter appear in the elementary waves theory just as in current theory. The negative frequency waves are in fact positive frequency anti-particle waves. Both particle and anti-particle waves can have positive or negative phase velocity.

By having the phase velocity in the direction of motion of the particle, we achieve invariance for that velocity; the wave and the particle transform correctly together. However, the overall picture is still not invariant. The phase velocity of the wave will transform correctly, but we still must transform the wave or `coherence velocity.' That velocity is opposed to the motion of the particle-this fact is the essence of the entire elementary waves theory. There is only one way that the overall picture can be invariant: if the coherence velocity of the wave is c, the velocity of light. Then it is c in all frames, and the overall picture is invariant.

To summarize, a plane elementary wave is like a flux of material with velocity c, along any flux line of which has been implanted a wave (which varies with time in a manner which will become clearer in a moment). For a single coherent plane wave, the wave on every flux line looks the same-has the same phase. The wavefronts will appear to move with a phase velocity that is greater than c, either positive or negative. But, as with the string above, there is no actual propagation of the wave along the material. Nothing actually moves with a velocity greater than c.

If a detector continually emits such a wave, with positive phase, then once the wave has been set up between the detector and some particle source, the resulting wave object along any line between the two looks exactly like the usual forward moving quantum wave along that same line.¹⁹ So, as indicated in Section 2, the sign of the exponential describing the wave actually needn't be reversed. The wave looks mathematically identical to current quantum waves even though its propagation is reversed.

I am reluctant, however, to refer to a wave `material,' as if it were something aside from the waves. There is no evidence of such a material. Indeed, if the wave objects are genuinely elementary, then it is meaningless to refer to a (more elementary?) material out of which they are composed. One can only say for sure that the wave objects exist.

Elementary waves are waves only in the sense that they add and subtract as waves when they are mutually coherent. That is, they so add and subtract insofar as they act to stimulate the emission of any particles. No actual cancellation of waves occurs; all `pieces' of every wave are present at all times. It is only the effects of a wave that cancel when its `pieces' are mutually coherent (and out of phase). This is unlike current wave theory, but is actually necessary in a theory where the waves are real things. The real waves don't go out of existence when they interfere; only their effects disappear.

What we end up concluding, then, is that space is filled with waves of all frequencies and wavelengths, all of which move with velocity of (coherence) propagation equal to c. Particle photons follow the photon waves (in reverse) with velocity c. Given the Lorentzian nature of space-time-where this is to be understood in the sense indicated at the end of the previous section-this `medium' of waves appears the same in all frames of reference. A given wave will appear to have a different frequency in another frame; but another wave will take its place in the new frame. What we have, then, is an `aether' of sorts, but one that is Lorentzian in nature. Rather than having a material medium through which the waves propagate, with the medium thereby fixing a preferred frame of reference, the waves themselves are the medium. They move with velocity c in all frames, so there is no preferred frame.

The existence of a medium through which the waves propagate would clearly contradict this entire picture. One must view the waves as constituting the medium, and thus as being elementary. The fact that space-time is observed to be Lorentzian-that is, that objects in space time transform in a Lorentzian manner-is the primary evidence that the waves are indeed elementary.

Given that the phase velocities are not signal velocities, it is necessary to show that the wave, viewed as a geometrical object spread out over space, will transform correctly. That is, applying a Lorentz transformation to the space-time coordinates of all parts of the object should produce a new object with the appropriate wavelength-the wavelength corresponding to the appropriate momentum particle. While this follows also in current theory, it is not generally spelled out in treatments of this subject.

Consider as an example the wave corresponding to a stationary particle. Its wavelength is infinite. It oscillates with the same phase over all space with a frequency m c²/h . Consider how this would appear to an observer moving in the - x direction with velocity v. A particle at rest in the first frame will now move with velocity v in the + x direction. Because of its motion, clocks that were synchronous in the rest frame become asynchronous. This means that the phase of the wave motion will now appear to be different at different points in space; the oscillations now take the form of a travelling wave.

At a distance L in front of the moving observer, as measured in the observer's frame, clocks will appear to be ahead by an amount[16]

In general, changes to the phase velocity simply reflect changes to simultaneity resulting from a change in the relative velocity of the observer's frame. So in this manner also we see that the phase velocity cannot correspond to a signal of any kind. In the `rest frame' of a wave-that is, the rest frame of the particle that might follow that wave-there is a common phase at all points along the wave. The `travelling wave' time dependence which occurs in a moving frame results entirely from changes to simultaneity.

A similar analysis can be performed for transformations perpendicular to the motion of the wave. Again the effect is one of a change to simultaneity (as well as a contraction of length in the direction of motion). The result shows that if a particle is moving perpendicular to the wavefronts in one frame of reference, it will move perpendicular to the wavefronts as viewed from any other frame of reference.²⁰ In general, then, the picture of the wave objects as propagating in the direction opposite to the particle, but with the wavefronts moving with the particle with phase velocity c²/v , and with the particle moving in a direction perpendicular to the wavefronts (by a mechanism to be explained in Section 9)-that picture transforms correctly between frames.²¹

Returning to an earlier point: As a photon travels, the wave that it is following will frequently be disrupted due to motion of the wave source or to the intervention of other objects between the photon and the wave source. So the photon will have to `jump' waves, as described in Section 4. However, given that all elementary waves travel with velocity c, we see that the photon's velocity will not be affected by the jump. Its velocity when following the new wave will be the same as when following the old. And the direction of motion will also not change during a jump.<sup>22</sup> So photons travelling over any distance will always travel with velocity c in a straight line (aside from gravitational effects)-which of course is what is observed. A single wave need make the full trip only for `local' light signals-this in order to explain the Lorentzian nature of space-time.

Furthermore, as demonstrated in Section 4, whenever a jump occurs, the state of the photon after the jump is exactly the same as it would have been had the new waves travelled the entire distance to the photon source before the emission. What one sees when a photon arrives is thus exactly what would have been seen had the wave made the entire trip. Only the state of motion of the observer at the instant the photon is observed will affect what is seen. This is genuinely an `intromissive' theory.

The final picture of the waves that we have arrived at might appear to contradict the initial physical explanation for the constancy of c. Every observer, it was argued, sees particle photons as moving with velocity c because they follow a wave from that observer. But now we have concluded that all waves move with velocity c, as do all photons. So how is the wave from one observer any different from the wave from any other observer? And if they are not different, how does the original explanation work?

The problem here is the assumption that it is the velocity of the wave that determines the velocity of the photon following it. An essentially Galilean model of the situation is assumed, in which a wave will move with a fixed velocity relative to the source; so if the source is moving the wave will move with a different velocity, as will the particle photon which then follows that wave. But clearly that kind of picture won't work here; all waves move with the same velocity.

What one is forced to conclude is that it is the `organization' imposed on a wave that determines the velocity of a particle following it. Somehow the organization reflects the frame of the particle which `emitted' the wave, and the particle following the wave then moves accordingly.

The constancy of c requires that a photon's velocity be causally determined by the frame of the observer. Since the velocity of the waves is not unique to a particular frame, it cannot be the velocity that causes the frame dependence. It must be that the organization itself is frame dependent.

This is connected with a point made at the beginning of the last section. The waves are not actually emitted by a particle, but rather are only rearranged as they pass by. So if the velocity of the wave were what determined the velocity of a particle photon following it, that particle velocity would not be c relative to the wave `emitter,' but rather would be c relative to a frame determined by the wave aside from the `emitter.' We would then have a Galilean and not a Lorentzian theory. But if the moving particle responds instead to the organization of the wave, and if that organization in turn reflects the frame of the wave `emitter,' then the explanation for the constancy of c still works. The `emitter' really does emit the `stuff' that determines the velocity of the particle: the organization.

The velocity of all photons will still be c anyway, regardless of which wave a photon follows. But the fact that the velocity of a photon is always c does not mean that the causal connection between the photon velocity and the organization of the wave disappears. Rather, it is only because of the existence of that causal relationship that space-time is Lorentzian, and therefore that the velocity is always c regardless of the wave source. From a causal point of view, the fact that waves from different sources are not distinguished by their velocity is, in effect, a coincidence.

So, for example, in the first derivation of the Lorentz transformation described in the previous section, the two sets of photons observed by the two arrays of observers actually both move with velocity c relative to both sets of observers. But this does not mean that they are interchangeable. The photons are distinguished by the fact that they are following waves with a different organization. And it is only because of this fact that one obtains Lorentz transformations, which-by `coincidence'-dictate that both sets of particle photons move with the same velocity c relative to both arrays of observers-a fact that was certainly not apparent when one began the derivation.²³

And there is no contradiction involved in saying that both sets of photons move with velocity c relative to both arrays of observers, now that it is clear that velocity itself, involving length and time, will be affected by the means of observation. Velocity is length over time. But both length and time appear to change when we change frames. So what we call the velocity in one frame is physically not the same thing as velocity in another frame. It is only by considering the invariant interval that we would have the same physical quantity in both frames. But the interval for a signal moving with velocity c is zero. So it will be zero in all frames.

A frame-dependent organization is, of course, not a new idea. The electromagnetic field of a charged particle is different depending on the velocity of the charge relative to the observer. By measuring the electric and magnetic fields at a point, both in direction and magnitude, one can determine the velocity of the charge emitting that field. The electromagnetic field, as will be demonstrated in Section 8, is actually itself simply an elementary wave, with particle photons then following the wave. The electromagnetic `organization' of the elementary waves reflects the emitter's frame.

The fact that the organization reflects the emitter's frame is actually essential to the ability of mutually coherent waves to distinguish between themselves and other, mutually incoherent waves. I will return to this point in Section 15.

8  Feynman Diagrams I: Qualitative
Interpretation

Scattering experiments can be described using the picture for a general experiment given in Section 3. Elementary waves are `emitted' by the detector, scatter off the target, and arrive at the source where they stimulate the emission of particles. The particles then follow the waves to the detector. The square of the wave amplitude at the source gives the cross section. The cross section will be exactly that of current theory because, by reciprocity, the wave scatters with the same matrix element.

Because the wave scattering problem is in essence the same as that in current theory, one can perform the usual analysis and express the scattering in terms of matrix elements between plane waves. Furthermore, it is not necessary to the theory that the same coherent wave make the entire trip from detector to source. Indeed, for most particle scatterings outside the laboratory, the environment will be continually changing, and hence the wave being followed by any particle will be continually changing. For a general scattering, a particle will be travelling in the direction of a target while following one wave or another in free space. As it approaches the target it will jump into coherence with one of the waves scattering off the target. This wave will have originated from another plane wave incident on the target from some other direction. The particle follows that wave, and will then be travelling in that other direction after the scattering. If a detector is located in that direction, eventually the particle will experience the waves coming from the detector, will jump into coherence with one of them, and then be detected.

In this picture, the square of the amplitude-the square, that is, which determines the cross section-doesn't occur at a physical particle source, but rather takes place when the jump occurs into coherence with the wave from the target. The `source,' in effect, is the oncoming particle beam.

Feynman diagrams, in this theory, actually picture accurately what is going on in the scattering, both when the waves scatter and when the particles subsequently scatter. The wave scattering description is very similar to current theory, with the waves taking all of the various (configuration space) Feynman paths. Just as for any other quantum system, there is no problem with the waves taking multiple paths; one simply has multiple waves. Each particle, when it scatters, will take only one path.

However, the elementary waves theory requires some essential changes to the interpretation of the diagrams. To illustrate these changes, consider electrons scattering off target muons in first order (Figure 6).

In the first place, the photon propagator waves which scatter the electron waves only become organized in the presence of a particle muon. In the absence of the muon, only disorganized photon waves exist, which produce various `renormalization' processes (see below), but not the wave processes corresponding to an electron-muon particle scattering. So at least the target muon particle must be present. This is unlike the usual theory in which the diagrams describe interacting wave fields.²⁴

Furthermore, the electron wave scattering off the muon particle must occur in the absence of the particle electron. This is the whole idea of the elementary waves theory: the electron wave scatters toward the electron particle source, and the cross section is determined by the wave intensity at the source. So we must understand how Feynman diagrams describe the scattering of the electron waves from the particle muon, this in the absence of the particle electron. Then we must also understand how the very same diagrams describe the actual scattering of the particle electron when it arrives at the muon. I will describe the theory qualitatively at first, and then show the detailed correspondence to the mathematics of current theory.

The sequence of events is as follows: First the muon `emits' the photon propagator waves. The muon is following some muon elementary wave, and might scatter into any of the other available states-that is, real elementary waves-around it. Even though it doesn't yet scatter, it emits the corresponding photon propagator wave. One can represent this as in Figure 7.

This is a general property of particles: they `emit' all of those waves corresponding to scattering processes in which they might participate-this in reaction to the incident waves corresponding to the other particles in that scattering process. The intensity of the emitted photon waves is given by an expression similar to that arising from the usual fermion `current,' e [`(y)]g<sup>m</sup> y. Notice, however, that all potential photon propagators are being emitted, and not just the one corresponding to a particular particle scattering that will occur in the future. In present quantum theory the `current' means e [`(y)] g^m y for the specific final state relevant to the scattering. But one doesn't know which final state the muon will have until one knows which propagator the electron has responded to. But the electron can't respond until the propagator has been emitted, and the propagator can't be emitted until the muon scatters, ... There is a clear lack of causality in this formulation. The problem disappears when the wave and the particle photon are separate things.

In the figures the photon propagator is shown as a dashed line connecting particular points. This is merely to indicate that one is concerned with the effect of that propagator specifically at the vertex in question. Actually it propagates spherically in all directions as usual.

However, there is another important physical difference here from the usual theory, as indicated earlier. Elementary waves do not actually scatter from one direction to another. The wave flux in one direction continues in that direction indefinitely. At a vertex, or at a particle which interacts with the wave, the effect is one of rearranging the organization of the wave. The vertex, in effect, leaves a `shadow' on the passing wave flux-a line along which the coherence is reorganized. A spherical propagator is actually composed of waves that were incident on the originating point of that propagator from all directions, which waves are then reorganized at the vertex. A `shadow' is left on each passing wave. The sum of all the `shadows' looks like a wave propagating spherically out from the vertex, because the vertex imposes a common coherence on each shadow line. But actually the wave along any single line out from the vertex is independent of the waves along the lines in all other directions. The use of lines in the Feynman diagrams is thus not merely symbolic, but rather pictures the physics that is actually going on.

Notice here again, then, that there is no propagation of waves according to the usual wave dynamics, requiring a field equation and etc. Even spherical waves are entirely a product of simple, straight-line flux propagation.

When a particle or a vertex leaves a `shadow,' it isn't clear whether the wave in the shadow is fully or only partially organized. All that one can say is that the cross sectional area of the particle or vertex, multiplied by the degree of organization imposed, is such that the `emitted' propagator is equal in intensity to that in current theory. Stated differently, the `charge' of a particle is proportional to this product.²⁵

The photon propagator contains both positive and negative frequency waves as usual. That is, it contains particle waves moving in one direction, and anti-particle waves in the other.²⁶ For photons the anti-particle is still a photon, of course. However, I will refer to the photons following the negative frequency waves as (anti-)photons in order that the analogous situation with mesons or other particles is clear.²⁷

Notice that the negative frequency waves are essential to the picture. Only the target muon is present when the electron waves scatter, so we must understand how the muon can emit waves corresponding to photons that will travel both to and from the muon. The positive frequency waves correspond to photons that will move toward the muon, the negative to photons that will move away.²⁸

Electron waves passing in the vicinity of the muon will be scattered by the photon propagator waves, again, just as in current theory. The amount of any electron wave that is scattered at a vertex-that is, the degree of mutual coherence established between the incident wave and any other wave at the vertex-is given by the same vertex function e [`(y)]g<sup>m</sup> y. The overall scattering will look very similar to the usual `current-current' form.

But there is a further important physical difference. The electron elementary wave objects propagate with velocity c, and not with the velocity v of the particle electron or the phase velocity of the waves. So the timing of the vertex interactions of the wave is not the same as in the usual theory. However, the variable t in the usual theory does not correspond to the actual motion of the particle or particles. That time corresponds to the phase motion of the waves which move along the various Feynman paths. It is the phase velocity that relates to the variable t which occurs in the current mathematical expressions for the diagrams. Once the vertex interconnections of any diagram have been set up by the elementary wave objects-propagating between vertices with velocity c-the resulting phase wave object will look identical to the wave object pictured in the usual theory by the same diagram. Even the direction of motion of the wavefronts will look like the forward moving waves of current theory, as explained above. A `snapshot' of the waves along the lines of a particular diagram with particular vertices will thus look identical to the object described by the same diagram in current theory. Integrating over all possible vertex locations will then yield the same result. The different timing of the vertex interactions changes nothing.

When two waves interact at a vertex, each wave `scatters' the other by the same amount; the usual vertex expression for scattering from state 1 to state 2 is simply the complex conjugate of the scattering from 2 to 1. So the net amount of coherence doesn't change at a vertex. Whatever coherence is imposed on wave 2 by wave 1, the initial coherence of 1 is reduced by the same amount as 2 imposes its coherence on 1. There is `conservation of coherence.'

The particular electron wave from the detector that is in question scatters from all points around the muon in the direction of the electron source (and in all other directions also, of course). The electron, then, has some probability of being emitted in `coherence' with that particular wave, with a probability determined by the square of the coherent amplitude at the source.

When the electron arrives at the target muon, two things can happen-this regardless of the particular location of the vertex at which the particle electron interacts. The combination of the photon waves from the muon, plus the electron wave from the detector-with which the electron is coherent-can induce the electron to emit a particle photon, and hence scatter. The particle photon then follows its wave to the muon and causes it to scatter. Or, second, the (anti-)photon wave being emitted by the electron-just as the muon emitted its photon propagator waves-can stimulate the emission of a particle (anti-)photon from the muon, causing the muon to scatter, following which the particle (anti-)photon travels to the electron and causes it to scatter.

The (anti-)photon must follow a wave from the electron; otherwise it will not arrive at the electron and cause it to scatter. However, that (anti-)photon wave isn't present until the electron is present. This might make it appear as if the corresponding electron wave scattering cannot occur at the muon before the particle electron is present. However, again, this is the role of the negative frequency photon wave from the muon. The negative frequency photon wave from the muon produces the electron wave scattering that corresponds to the electron particle scattering caused by the (anti-)photon exchange.

Remember that it doesn't matter which path the electron takes anyway, just so it arrives at the detector by one means or another. Only the wave scattering is involved in determining the cross section (at the source). But, in fact, as just illustrated, the particle scattering can always mimic any wave scattering.

Notice, however, that when the photon is emitted by the muon, that occurs as the result of the interaction between an (anti-)photon wave from the electron, the muon wave being followed by the particle muon, and the muon wave into which the muon will scatter. The electron wave itself does not directly participate. However, the electron wave must somehow participate, because it is that wave that dictates the behavior of the electron. A particle photon unrelated to the electron wave would scatter the electron out of coherence with its wave. The only way in which the electron wave could participate would be if the (anti-)photon wave from the electron is produced by a potential scattering of the electron that will leave it coherent with its wave.

But this makes perfect sense in the above picture. The electron emits propagator waves corresponding to scattering processes in which it can participate. But once the electron has become coherent with a particular wave, it can only participate in scatterings dictated by that wave. The electron won't respond to incident electron waves with which it is not coherent, and hence won't emit the corresponding (anti-)photon waves. So only particle (anti-)photons which leave the electron coherent with its wave will arrive at the electron, because only the corresponding (anti-)photon waves from the electron are present to stimulate the emission of those (anti-)photons from the muon.

If a wave is disorganized, a particle following that wave will, in effect, follow all pieces of the wave. It can then potentially scatter into the state of any incident particle wave. So it `emits' all of the corresponding photon propagator waves. But once the particle has become coherent with a particular coherent, organized piece of the wave state, it can only respond to that coherent wave. If the coherent wave scatters at a vertex, then the particle, when it arrives at that vertex, can respond to the incident wave from the new direction-but not to any (mutually incoherent) particle waves from any other direction. So it emits only the corresponding propagator waves.

In a two particle scattering, one particle is always the `leader.' The leader is that particle which first responds to-`jumps' into coherence with-a coherent wave scattered by the other particle. From that point on the entire scattering process is dictated by the leader's wave, because any exchanged quantum, moving in either direction between the scattering particles, is dictated by that wave. The `follower' particle then scatters accordingly. In the above example the muon is the `follower,' the electron the `leader.' Momentum is always conserved in a scattering, of course, because any exchanged particle affects the momentum equally at both ends of its trip.

The above description extends to diagrams of all orders. The waves scatter by all possible Feynman diagrams simultaneously, and by all arrangements of vertices for any one (topologically distinct) diagram. The Feynman picture here is equivalent to a series of successive approximations, as in current theory. But the `heuristic' derivation of that picture[17], in which waves scatter at individual points, and move on straight lines until scattering