by Nick Herbert

Abstract: Two people looking at light from the moon through a beam splitter have their retinas entangled trillions of times a second. If this entanglement could actually be perceived, I show how this perception could be exploited to send signals faster than light. On the other hand, the lunerator might alter the states of the participants without conscious perception.

April (2006)
Nick Herbert



The lunerator (1) is a simple optical device which purports to quantum-entangle the retinas of two observers. It uses a beam splitter to divide the wavepackets of moonlight into two parts and directs each part to a separate eye. The size (lateral coherence width) of each moonlight wavepacket is smaller than the pupil diameter so this entanglement involves only the two retinas and does not "spill over" to entangle the environment as is the case with the stellerator.

A few commentators, including Dick Shoup from the Boundary Institute, have noted that entanglement is doubtful because one subpacket will surely strike the retina of the observer closest to the beam splitter first giving him or her right of "first refusal" to absorb the photon. Then only later does the other subpacket strike the second observer's retina.

If the near observer has exercised his "seeing option" then the subpacket that strikes the second observer's retina is "empty" and there is no possibility of her seeing this photon.

On the other hand, if the first observer missed his chance, the subpacket that enters the second observer's eye will now be observed for certain. (I have made the simplifying assumption that for each photon only two options exist--absorption at one of the other retina. Scattering inside the eye or absorption in the humors for instance are ignored as rare events compared to retinal absorption.)

For entanglement to occur, the two observers must be simultaneously competing for the same photon, not alternately competing as in the above scenario. This means that, for the lunerator to work as advertised, the retinas of the two observers must be equidistant from the beam splitter to high precision.

How precise? For simultaneous observation the two retinas must be equidistant to within the dimensions of the wavepacket in its direction of travel (longitudinal coherence length). The longitudinal coherence length of unfiltered moonlight is about 1 micron (or one thousandth of a mm), that is, very small indeed.

On an optical bench with mechanical parts an alignment precision of 1 micron is possible but not with living beings. However this simultaneity of photon seeing can be attained temporarily by rocking back and forth near the region of equidistance. If retinal entanglement produces a strong subjective sensation, the position where moonlight photons are truly shared between both parties might be readily detectable as a brief spurt of unusual connectedness.

Thus part of a successful lunerator operation must include a rocking of the head of one of the participants so as to pass back and forth through the critical very narrow optical coincidence region.

Thanks to those of you who pointed out this flaw in the original lunerator design.

However a more serious objection to the lunerator's purported mode of operation has arisen. If the sensation of retinal entanglement could actually be detected--either by one of the participants alone or by measurements made on the pair--such as correlated brain waves--the way would lie open to build a time machine! This possibility either casts doubt on the entire lunerator enterprise or considerably raises the stakes for achieving its success, depending on your opinions concerning the possibility of time machines.

If lunerator entanglement produces measurable results--either subjective sensations or objective changes--then you could build a time machine? How does that work, Nick?

Well, by time machine, I'm not talking about the Alley Oop variety where you can send somebody or some thing backwards in time but the simpler variety where you can just send a signal from the present into the past. Backwards-in-time signaling becomes possible when you combine faster-than-light signals with moving sources and/or mirrors. That's just straight-forward Einsteinian physics. FTL signals + moving reference frames = signaling backwards-in-time.

And I claim that a working lunerator would allow you to signal faster-than-light.

Several years ago I showed (2) that the ability to make a particular distinction between two types of unpolarized light would allow you--in conjunction with a twin-photon EPR (Einstein-Rosen Podolsky) set-up--to send signals FTL. This crucial distinction is that between a beam of plane-unpolarized photons (PUP) and circularly-unpolarized (CUP) photons. It is possible to produce such beams--they form the basis of many quantum cryptography schemes--but paradoxically it is considered impossible to distinguish CUP and PUP light by any conceivable measurement.

Somebody (usually dubbed "Alice" in the cryptographic context") can make such a beam but nobody (usually dubbed "Bob") can measure that she's done so? How's that work?

Well the lady that makes the beam also obtains in the process a random-number table (RNT) that (if Bob possessed it) could show him for every photon how to set his polarizer so that he will get a 100% certain result, hence verify that the beam is--as Alice claims--either CUP or PUP. But without access to Alice's RNT no effort on Bob's part can disclose that information. Thus Alice can prepare either a CUP or a PUP beam--and prove, with the help of her RNT that she has really done so. But Bob, without access to Alice's table cannot physically distinguish these two kinds of beam. Nature, as Einstein famously remarked, is subtle but not malicious.

If you can distinguish CUP from PUP light (without access to the distant lady's random number table) then you can exploit this ability to send faster-than-light signals because in an EPR setup Alice making one kind of beam here (CUP or PUP) INSTANTLY produces the same kind of beam (for Bob) there where "here" and "there" may be light years apart. And the lunerator--if it works--apparently has the potential to detect this distinction without access to the distant lady's secret code book.

Light can be polarized in an infinite variety of ways but only two types of polarization interest us here--plane polarization in the vertical (V) and in the horizontal (H) direction, and circular polarization rotating to the right (R) and to the left (L).

We imagine that we have 4 types of polarizing filters that pass each of these 4 types of polarized light. H and V are complementary pairs. Light that has passed an H filter will be 100% blocked by a V filter.

Likewise light that has passed an R filter will be 100% blocked by an L filter. R and L are complementary filters.

Mathematically R and L light are equal superpositions of H and V light. And contra-wise H and V light are equal superpositions of R and L light. So that when light that has passed an R filter goes into an H filter, 1/2 of it will get thru. The same is true of any combination of filters taken one from the (H,V) menu and one from the (R,L) menu. Of the light that passes the first filter, half will pass thru the second.

If these properties of polaroid filters are unfamiliar to you, just skip ahead to the conclusion.

I now show that if the lunarator works, then it can achieve the impossible--that it can distinguish between a CUP and a PUP light beam.

What does it mean for the lunerator to "work"? It means that, altho each photon can only be detected by one of the participants, that the situation of simultaneously competing for the same photon has a subjective and/or physiological component. In other words, if the lunerator works, you could actually "feel" the entanglement. The ability to tell the difference between photons that are "up for grabs"--that both of you have equal possibility to perceive but that only one of you will actually perceive--and photons whose detection possibility is not shared is what it means for the lunerator to "work".

I'll call this entanglement sensation MERGE. MERGE would be experienced, for instance, when the two retinas are equidistant from the beam splitter. MERGE disappears when one or the other retina gets the photon first or in any other case where the photon wavefunction does not overlap both retinas at the same time.

Now we examine what happens to MERGE when we put some polaroid filters into the lunerator.

First let's imagine that the beam of unpolarized moonlight incident on the beamsplitter can be represented as a random mixture of H and V photons--in other words, it is what I call a PUP beam.

Now place an H filter in front of EYE #1 and a V filter in front of EYE #2. The MERGE will be zero because the participants are no longer competing for the same photon. If the photon is (randomly) H, then only EYE #1 can absorb it. If the photon is (randomly) V, then only EYE #2 can absorb it. In no case are the participants exposed to subpackets of the same photon wavefunction. So their experienced MERGE should be zero.

On the other hand we can imagine that the unpolarized beam consists of a random mixture of R and L photons--what I call CUP light. Since each of these circularly polarized photons is 1/2 H and 1/2 V, both eyes are competing for the same photon with a mutual wavepacket overlap of 1/2. So in this case a half-amplitude MERGE should occur.

So if MERGE is a real phenomenon, it appears to be able to distinguish between CUP and PUP beams. With H and V polarizers in place, PUP produces no MERGE and CUP produces a MERGE of 1/2 (compared to the intensity of the MERGE experience when the two wavelets are completely entangled).

Likewise it's easy to see that if R and L filters are used instead of H and V filters, that CUP light produces no MERGE and PUP light produces a MERGE of half strength.

So to determine whether the beam is of the CUP or PUP variety, one merely switches polarizers between H/V type and R/L type and observes with which arrangement the MERGE experience occurs.

The discovery that if retinal entanglement produces any perceptible effect, that effect could be used to build a time machine is a powerful argument against the existence of MERGE experiences.

Quantum entanglement represents a delicate balance between a classically inexplicable non-local connection and the inability (because of the necessarily random nature of quantum events) to exploit that connection to send superluminal signals. Any scheme that purports to learn more about nature than present quantum measurement theory allows runs the risk of upsetting that balance and allowing nature's hidden superluminality to surface in the real world.

So intimacy machines of the first kind (directly perceivable MERGE experiences due to quantum entanglement) seem blocked by the no FTL signaling rule of special relativity.

However the possibility that physical entanglement can imperceptibly alter the nature of the participants (creating what I have elsewhere called "oblivious links"(3)) so that their subsequent histories are profoundly different than without the entanglement is not excluded (intimacy machines of the second kind). Of course every experience does at least this much--altering us forever--but one could imagine that entangling in the quantum way might move us (necessarily imperceptibly) into experience spaces completely inaccessible by any conceivable classical interaction. One could imagine that lunarators (or similar human entanglement devices) will have the ability to transport us to islands of experience completely isolated from the classical trade routes. Thus instead of immediate experiences of unusually strong connection, the entanglement trip might more closely resemble wandering blindfolded together through a dark cavern and emerging on the other side into a strange new world no two human beings had ever before experienced--a world completely inaccessible by classical means.


(1) Nick Herbert (2006) "Lunerator: a Second Generation Quantum Intimacy Machine" <>

(2) Nick Herbert (1982) "FLASH--a Superluminal Communicator Based Upon a New Kind of Quantum Measurement" Foundations of Physics 12, 1171 (1982)

(3) Nick Herbert (1999) "Werner (Heisenberg) Alone Has Looked Upon Reality Bare" Ars Electronica: Facing the Future, Timothy Druckrey, ed MIT Press (1999) <>