Herbert’s paper made waves among the remnants of the Fundamental Fysiks Group. Although the group’s weekly meetings had wound down a few months earlier, former members, including Herbert, Jack Sarfatti, Saul-Paul Sirag, Henry Stapp, and Philippe Eberhard, hashed out Herbert’s design in June 1979. Stapp immediately challenged the idea, building on Eberhard’s argument that statistical averages should wash out any superluminal effects. Sarfatti countered that Herbert’s QUICK scheme evaded that problem. Herbert’s imagined device seemed to promise an immediately distinguishable effect for each individual photon: the half-wave plate at B either rotated or it didn’t.33
Thanks to Einhorn, others began to grapple with Herbert’s QUICK paper, far from the California crew. The paper made its way to GianCarlo Ghirardi, an Italian physicist working at the International Centre for Theoretical Physics—the selfsame center at which Sarfatti had spent much of his sabbatical in 1973–74. Ghirardi had been interested in the interpretation of quantum theory for some time, dating all the way back to his graduate work in the late 1950s. But like so many of his generation (John Bell included), Ghirardi had learned to keep those interests on the sideline. He worked on “more fashionable” topics, “fields in which you might hope to get a permanent position,” as he put it recently. The strategy worked. By the mid-1970s he had secured a double position in Trieste, both at the International Centre for Theoretical Physics and as physics department chair at the neighboring university. At last he felt safe turning his attention squarely to foundational matters.34 It didn’t take long before his plate was full. Around the time he received Herbert’s paper, Ghirardi attended a meeting at nearby Udine, in northeastern Italy, and heard a presentation by his Italian colleague Franco Selleri.35 Selleri and his group in Bari, Italy, had been among the earliest and most active researchers on Bell’s theorem anywhere in the world. (Selleri, a contemporary of Ghirardi’s, had likewise turned to the topic relatively late in his career. He later explained that the university at Bari was new when he was hired and had no other theoretical physicists on staff, so Selleri had more leeway to pursue idiosyncratic or unpopular interests.)36 Selleri and company had hatched a scheme for superluminal signaling remarkably similar to Herbert’s. With a Trieste collaborator, Ghirardi dug into Herbert’s and Selleri’s proposals and isolated a fatal flaw.37
The main weakness, Ghirardi realized, was that Herbert had worked in a kind of semiclassical approximation. He had tacitly assumed that quantum mechanics applied only to the pair of photons, and not to the apparatus itself. Any half-wave plate that functioned as Herbert’s scheme required—flipping individual R photons into L photons and L photons into R photons, while letting H and V photons pass through unaffected—would run afoul of a fundamental quantum limit, akin to Heisenberg’s uncertainty principle. The original half-wave plate, devised back in 1936, had worked because the experimenter sent zillions of photons at the half-wave plate at a time. To get the same results at the single-photon level, Herbert’s half-wave plate would need to be infinitely massive. But then it would be too heavy to rotate whenever an R or L photon zoomed past.38
Ghirardi and his colleague wrote up their analysis in November 1979, and followed up with a more general demonstration a month later. Together with Eberhard’s paper from 1978, the papers by Ghirardi have been hailed as the earliest rigorous proofs that Bell-styled nonlocality could peacefully coexist with Einstein’s relativity.39 Though they reached similar results, however, Ghirardi and his coauthors struck a rather different tone in their conclusions than Eberhard had done in his earlier article. Where Eberhard had argued that it would be “futile or counterproductive” to discourage further work on the topic, Ghirardi saw things differently: “To conclude, we have considered [it] worthwhile to illustrate explicitly the general proof of the impossibility of superluminal transmission, even though it is quite elementary, to stop useless debates on this subject.”40
On this last point, Ghirardi fell somewhat short of his goal. In fact, his demonstration of the weakness of Herbert’s device only spurred on the quest. From the objections by Stapp and Eberhard to Jack Sarfatti’s original proposal, Nick Herbert realized the importance of exploiting individual quantum events, rather than statistical averages. From Ghirardi’s intervention, Herbert came to appreciate the importance of amplifying the tiny distinctions between various quantum states, to evade fundamental limits on signaling. The rules of the game were set. Herbert got back to work.
Within a year he had devised an alternate design. Like his QUICK proposal, his new scheme relied on distinguishing H, V, R, and L photons. Again the experimenter at detector A could encode messages to the faraway experimenter at B by choosing to measure linear or circular polarization. But this time, the signaling did not rely on experimenter B manipulating finicky, single photons. Rather, the photon en route to B first passed through a laser gain tube—the amplifying mechanism at the heart of real-world lasers, which had been around for nearly twenty years by that time. “Laser,” after all, is an acronym for “light amplification by stimulated emission of radiation.” In Herbert’s scheme, the incoming photon would do the stimulating; the laser gain tube would take care of the amplification. He cooked up a new acronym for his latest design: “FLASH,” for “First Laser-Amplified Superluminal Hookup.”41
Herbert latched on to the idea of using a laser because laser light is special. The light that comes out of a laser gain tube is not just amplified—lasers are not just bigger, brighter lightbulbs—but coherent. That is, in principle the light that exits the laser is perfectly in phase with the incoming radiation: its electric field oscillates in the same way, at the same rate, and in the same direction. That is just another way of saying that the polarization of the amplified, outbound signal is exactly the same as the polarization of the incoming light. Send in a weak signal of horizontally polarized light, and get out an intense beam of horizontally polarized light. A leading textbook at the time emphasized the phase coherence of laser output as the laser’s defining quality.42
Armed with this new gadget, Herbert laid out his FLASH design. Experimenter A would make his measurement first; that would force the twin photon, headed toward experimenter B, into a state of either linear or circular polarization. Then photon B would enter the amplifying tube. Out would come a burst of identical copies: a beam of laser light consisting of photons all in the same, still-unmeasured state of polarization as photon B. (Herbert, who had logged all those hours working at the photocopy-machine company, called this the “perfect photon xeroxing provided by the laser effect.”)43 Next the experimenter at B could make some quick measurements on the light beam, no more intrusive than the kinds of manipulations that physicists had managed to perform since the 1930s. In particular, the experimenter at B could send the light through a beam splitter, such as a half-silvered mirror. That would bounce half of the incoming light toward one set of detectors (call them station 1) while allowing the rest to pass through toward a second set of detectors (station 2). Station 1 was equipped to measure linear polarization; station 2 measured circular polarization. In an instant experimenter B would know, from the output of these two sets of detectors, whether experimenter A had set his own device to measure linear or circular polarization. For example, if experimenter A chose to measure linear polarization and got the result H, then the laser gain tube on the other side would spit out a burst of photons all in the state V. Assume for simplicity that the tube released 100 photons, each with V polarization (although real lasers by that time could amplify incoming signals by a factor of several million). Thanks to the beam splitter, 50 of those photons would get diverted toward station 1 (set to measure linear polarization), while 50 sailed toward station 2 (set to measure circular polarization). At station 1, Herbert wrote, the results would be clear: 0 photons in state H, and 50 photons in state V. Station 2 would show 25 photons in state R and 25 photons in state L.44 (Fig. 9.3.)
Excited about the latest design, Herbert wrote up a new paper in January 19
81. He submitted a copy to the journal Foundations of Physics, the relatively new journal that welcomed speculative papers on philosophical topics; Herbert had published a generalization of Bell’s theorem in the same journal a few years earlier.45 He also prepared a preprint version to circulate on his own. As it happened, Herbert finished his FLASH paper the very month that Ira Einhorn hopped bail and fled the country, just days before his murder trial was set to begin. With Einhorn out of the picture, Herbert had to rely on his own informal network to spread the word. Back in 1979, he had distributed the QUICK paper under the aegis of his C-Life Institute. This time Herbert circulated the FLASH paper as a preprint of the “Notional Science Foundation,” another play on his deep interest in the intersections of physics and consciousness.46 (Both the “Institute” and the “Foundation,” of course, shared the same address: Herbert’s post office box in Boulder Creek, California.)
FIGURE 9.3. Nick Herbert’s FLASH design. A source emitted entangled pairs of photons. The experimenter at A measured either linear or circular polarization, after which the photon heading toward experimenter B entered a laser gain tube. The burst of identical copies would then be split by the beam splitter, half going toward detectors to measure linear polarization and half to be measured for circular polarization. (Illustration by Alex Wellerstein, based on Herbert [1982], 1174.)
Had he done it? Could FLASH really communicate messages faster than light? Right after mailing out his preprints, Herbert convened another of his Esalen workshops on Bell’s theorem and the nature of reality. Most of the familiar faces from the Fundamental Fysiks Group were there, including Henry Stapp, John Clauser, Saul-Paul Sirag, and Elizabeth Rauscher. Herbert presented his latest scheme. As his annual report to his Esalen sponsors indicated, his proposal “was described, discussed heatedly, but not refuted.”47 Soon after that Berkeley’s Philippe Eberhard sent Herbert detailed comments. Eberhard thought he had isolated a flaw, but Herbert countered that challenge, too. “Does this mean,” Herbert wrote back—you can almost see the schoolboy grin on his face—“that I can now count on your support for raising investment money for commercial exploitation of FTL [faster-than-light] communication?”48 After several years of intense effort, it looked like the quest for what Jack Sarfatti repeatedly called “the Holy Grail” had come to fruition.49
The gears of academic peer review grind slowly, but grind they do. The same week that Herbert responded to Eberhard’s critique, in March 1981, the editor of Foundations of Physics sent FLASH out to referees.50 One reviewer later recalled that he knew the instant he received Herbert’s FLASH paper that it must be incorrect, because of its incompatibility with relativity. On the other hand, he couldn’t find any error. Whatever the flaw might be, the reviewer reasoned, it must be a nontrivial one, some juicy nugget whose careful elucidation would nudge forward the community’s understanding. Still convinced that the paper was “obviously wrong,” the reviewer wrote back to the journal’s editor recommending publication: the provocative paper was sure to spur further development of the field.51
The journal editor had also asked GianCarlo Ghirardi to review Herbert’s paper. Ghirardi was an obvious choice, especially after his careful debunking of Herbert’s QUICK proposal just a few years earlier. Like the first reviewer, Ghirardi struggled with Herbert’s latest paper—“I must confess that it took some time for me to spot the mistake,” he explained recently. But within a few weeks he had found what he was looking for, and he wrote up a brief report for the editor. Unlike the first reviewer, Ghirardi recommended a flat-out rejection.52
The problem, Ghirardi explained, was that no laser gain tube—not even an ideal one, the kind imagined in thought experiments—could function the way Herbert’s plan required. The flaw didn’t come from any limiting feature of a given laser design; it came from quantum theory itself. Quantum mechanics is what physicists call a “linear” theory. That means that when you add together two different solutions to the governing equations, the result is also a solution. Indeed, linearity is what lies behind many of the conceptual sticking points in quantum theory, such as Schrödinger’s cat. In one solution, the cat lies dead in its box. In another equally valid solution, the cat purrs merrily, awaiting its release. And in a third solution—made possible thanks to the linearity of quantum theory—the cat is caught in some suspended state, neither dead nor alive. Likewise with the famed double-slit experiment. When both slits are open the interference pattern emerges because the photon’s wavefunction is the sum of two possibilities: photon traveling through top slit plus photon traveling through bottom slit.
That same linearity applied to the photon polarization states in Herbert’s FLASH proposal. The states R and L could be broken down into linear combinations of the states H and V, and vice versa. For Herbert’s scheme to work, the laser gain tube needed to behave like a “photon xeroxer,” as Herbert had put it, making duplicate copies of whichever polarization state entered the tube. Suppose the initial photon were in the state R. Even before the photon entered the laser tube, it would exist as a combination of equal parts H and V. If the experimenter had chosen to measure the photon’s linear polarization at that point—before it entered the laser gain tube—she would have had an equal probability of finding it in state H or in state V, that is, all in one state or all in the other, with equal odds. A laser tube that really could amplify any of the polarization states equally would produce 100 duplicate copies of that same combination: a state that had a fifty-fifty chance of containing all 100 photons in the H state, or all 100 photons in the V state. It would never produce 50 H photons and 50 V photons at the same time, as Herbert had assumed. “In other words,” Ghirardi explained in his brief report, “it is impossible that for all 4 states of polarization…the laser gain tube acts simply as a duplicator producing N photons of the same type.”53
Without the ability to make perfect copies of an arbitrary polarization state, Herbert’s FLASH scheme fell apart. The experimenter at B would always see random patterns coming from the detectors at station 1 and station 2: on any given run, each station would find all of its signal in just one state (all H or all V at station 1, and all R or all L at station 2). Any intended message from experimenter A would be hidden within the random strings, just as in Eberhard’s original critique. To Ghirardi, therefore, the decision seemed obvious. “The basic device of the suggested experiment violates the linear nature of quantum mechanics and therefore all the proposal is incorrect. For these reasons I consider that the above paper does not deserve publication in Foundations of Physics.”54
Ghirardi thought he had sealed the deal, and in less than a single page, no less. The journal editor, however, received two reports from leading experts that came to opposite conclusions: the first reviewer recommended immediate acceptance, while Ghirardi urged summary rejection. Faced with the conflicting recommendations, the editor did what most journal editors would do: he asked Herbert to revise his paper in light of the reviews and resubmit it for further consideration. (A kind of editorial version of Schrödinger’s cat. Editing an academic journal appears to be a linear process, too.) Herbert went back to his draft and beefed up his discussion of how the sought-for signal might be gleaned from real-life noise, perhaps by sending an initial triggering pulse and then scrutinizing various coincidences between detectors.55 He seems not to have caught the real thrust of Ghirardi’s critique at this time—not terribly surprising, given how terse Ghirardi’s report had been. Nonetheless, Herbert hammered out a new draft and dutifully mailed it back to Foundations of Physics. It arrived on January 15, 1982, almost exactly a year after he had submitted his original version. He mailed out preprints of the new version, too, including a copy sent directly to John Bell.56
Imagine Ghirardi’s surprise when his Trieste colleague, Tullio Weber, received a request to referee the new version of Herbert’s paper. (Weber had coauthored with Ghirardi the earlier critical analyses of Herbert’s QUICK scheme.) Weber made quick work of the task, essentiall
y repeating the argument from Ghirardi’s previous report, this time condensing the entire analysis to a single paragraph. Despite the brevity, Weber tried to make clear the essential point. “In my opinion,” he closed, the design of a thought experiment “must not present intrinsic contradiction with the theory it pretends to test.”57
Independently of Ghirardi and Weber, Henry Stapp had also caught a glimpse of how the linearity of quantum theory might threaten Herbert’s FLASH scheme. Just as Weber was sending in his referee report, Stapp argued vehemently with Herbert at their annual Esalen workshop. But, Herbert shot back, Stapp’s argument seemed too strong: it also appeared to rule out the ordinary operation of lasers, which amplify incoming signals all the time. They tussled round and round in Esalen’s “big house” without reaching any definite conclusion.58
How the Hippies Saved Physics: Science, Counterculture, and the Quantum Revival Page 25
