Relativity-10: Science and public relations

(For previous posts in this series, see here.)

Scientists want their work to influence the field and so they would like it to gain the widest possible audience. Most of the time, their peers (and funding agencies) are their target audience because they are the only ones who really understand what they do. But when the work also has appeal to the general public because of its practical applicability or its revolutionary implications, then there can arise tensions in how the work is publicized and in the case of the OPERA experiment on faster-than-light neutrinos, there has been considerable unease with how this whole episode was handled with respect to the media.

The usual process when scientists have something new to say is that they write up a paper with their results and send it to a journal. The journal then sends the paper to referees who work in the same field (the number of referees depends on the journal and the discretion of the editor) who provide feedback to the editor. The referees do not usually check the results or repeat the calculations and experiments. What they do is see if the paper makes sense, the methodology is correct, if the authors have taken into account all the relevant factors and provided all the necessary information so that readers know exactly what was done (and how) so that they could repeat and check the results if they are so inclined, and that proper credit has been given for prior related work. Based on this feedback, the editor decides whether to accept the paper, reject it, or send it back to the authors for revisions and/or additional work. Good referees and editors can improve a paper enormously by providing the authors with valuable feedback and useful information and suggestions.

In the sciences, authors also usually simultaneously send out copies of the paper (known as preprints) to colleagues in the field. This serves to give their colleagues advance notice of their work (since the time taken to appear in the journal can often take over a year), to get feedback, and to establish priority for any discovery. All this occurs out of the public eye. Once the paper has been accepted and published by a journal, then it enters the public discussion and the media can publicize it. If the paper has significant implications, the journals may alert the media and give reporters a copy of the paper before it appears in print so that they can research and prepare an article about it, but the reporter is under an embargo to not publish until the journal article actually appears. Some of the more influential journals will refuse to publish an article if the authors release the information to the media before the journal prints it.

In the pre-internet days, and for research results that do not have revolutionary implications, this system worked reasonably well. Due to the cost of mailing, not too many preprints went out so the pre-publication discussions remained within a fairly small circle. With the internet, it became much easier to send out preprints to huge numbers of people at no cost and it was not long before it was realized that it made sense to create a system that could serve as a permanent archive that would allow scientists to post their preprints online so that anyone could gain access to them and search for those results that interested them. Currently the most popular venue for such preprints is arXiv and Wikipedia has a good article about its history and how it works.

The articles that are found on arXiv are preprints and thus have not been peer-reviewed but the system is minimally moderated to keep out rubbish. In general, scientists are concerned about their reputations among their peers and so most are careful to only post articles that they think would meet the standards of quality required if they were submitting to a peer-reviewed journal. Almost all of them do simultaneously submit their articles to such journals. As a result, the papers that appear on arXiv tend to be of pretty good quality. All the papers associated with the faster-than-light OPERA experiment are on arXiv.

A few scientists feel that peer-reviewed print journals are an anachronism and do not bother to try to even get their work into journals, feeling that the quality of the work will speak for itself. They think that if their work is correct and important, the community of scientists will accept it and build on it, while if it is wrong the community will criticize and reject it. Possibly the worst fate is that the community will think it is useless and a waste of time and completely ignore it. It may well be the case that in the future, expensive peer-reviewed print journals will disappear and that this kind of open-source publication will become the norm, with quality being determined by the consensus judgment of the scientific community. We are not there yet.

In the case of the OPERA experiment, the system broke down somewhat for several reasons. The OPERA experiment is very difficult and is a huge enterprise involving many collaborators and lasting over three years, with the paper having over 150 authors. Given the culture of the free sharing of information in science, it is very hard to keep preliminary results under wraps and it was pretty much an open secret that these faster-than-light results had been obtained. But this knowledge stayed within the community. What the OPERA team did was the day after they posted their preprint on arXiv on September 22, they issued a press release announcing their results and promoting a big press conference the next day with media and scientists present.

This rubbed some scientists the wrong way. Scientists can be as publicity hungry as celebrities but there are norms and there is a discreet way of making one’s name known. Holding press conferences or issuing press releases so early in the game, before the scientific community has had time to pass its verdict on the research, is considered bad form and the OPERA team has received some criticisms on this score.

While some of the carping may be due to jealousy, it is also the case that trumpeting that a scientific revolution has occurred can harm the image of science if the claim has to be later retracted. The reliable knowledge that science produces tends to be the consensus verdict of the community, achieved after a lot of behind-the-scenes work has smoothed out the rough edges and corrected mistakes. Bypassing that filtering process and going public too soon can lead to embarrassing reversals and give ammunition to the critics of science that its results cannot be trusted.

Next: Recalling an earlier public relations debacle

What use is half a wing?

Creationists like to challenge the theory of evolution by asking how it can be that things can evolve incrementally since in its early stages the new feature seems to lack its final functionality. They pose questions like “What is the use of half an eye or half a wing?” Of course, scientists have long explained this. They have shown how the eye could have evolved by tiny changes and in fact even right now almost the full spectrum of differential eye development can be seen in existing species.

They have also pointed out that it is a mistake to assume that the final functionality of a feature was the only functionality all along, and that features may have had other functions in the early stages and only later became adapted to its final use.

Carl Zimmer had a nice article earlier this year in National Geographic about the evidence that feathers might have evolved for a different purpose long before flight occurred. More recently, he reports on new research results that add to our knowledge of what purpose those non-flying feathers in primitive wing forms might have served.

Relativity-9: The importance of corroborating evidence in science

(For previous posts in this series, see here.)

In my series on the logic of science, I recounted how philosopher of science Pierre Duhem had pointed out as far back as 1906 that the theories of science are all connected to each other and changes in one area will have unavoidable effects on others that should be discernible. In this case, if neutrinos in the OPERA experiment did in fact travel faster than the speed of light, then we should be able to look at some other effects that should occur and see if they are observed.
[Read more…]

Relativity-8: General relativity

(For previous posts in this series, see here.)

To understand the role of Einstein’s general theory of relativity, recall that the original OPERA experiment claimed that they had detected neutrinos traveling faster than the speed of light. This posed a challenge to what is known as Einstein’s theory of special relativity, proposed in 1905, which said that the relationship between the clock and ruler readings for two observers moving relative to one another would be different from the ones given by the seemingly obvious relationships derived by Galileo centuries earlier. According to Einstein’s theory, it is the speed of light that would be the same for all observers, while clock readings could differ, and that Einstein causality (the temporal ordering of any two events that are causally connected by a signal traveling from one to another) would be preserved for all observers. One inference that followed from Einstein causality is that no causal signal can travel faster than the speed of light, and this was what was seemingly violated by the OPERA experiment.

But Einstein had a later and more general theory that he proposed in 1915, called the general theory of relativity, that included the effects of gravity. He showed that clock readings were not only affected by the speed with which the clock was moving, they were also affected by the size of the gravitational field in which the clock found itself. This is the source of what is referred to as the ‘gravitational red shift’ that enters into cosmology that causes the light emitted by distant stars and galaxies to be shifted towards larger wavelengths as they escape the gravitational field of those objects on their journey to us.

To understand what is going on, recall that when we measure the elapsed time between two events, what we are really doing is measuring the number of clock ticks that occur between the events. According to general relativity, the stronger the gravitational field, the slower the rate at which a clock ticks. The slower the rate at which a clock ticks, the less time that it records as having elapsed between two events.

So, for example, since we know that the Earth’s gravitational field decreases as we go up, this means that if we take two identical clocks, one on the floor and the other on the ceiling, the one on the floor would have fewer ticks between two events than the one on the ceiling, even if both are stationary. So the clock on the floor would ‘run slower’ than the one on the ceiling and hence the time interval measured between two events measured by clocks on the floor will be less than that measured by clocks on the ceiling.

In the OPERA experiment, the time measurements were made using GPS satellites. These are whizzing by at both high speeds (about 4 km/s) and high altitudes (about four Earth radii). Typically, the signals are handed off from one satellite to another as they appear and disappear over the horizon and the transition is almost seamless and produces such small errors that we do not notice it. But the OPERA experiment requires such high precision that they arranged to do the experiment during the transit time of just a single satellite so that even that source of error was eliminated.

Because the rate at which clocks run depends upon the size of the gravitational field, one has to make corrections to allow for the fact that the time readings given by clock readings of the satellites will be different from the time readings given by clocks on the Earth, and so one needs to make extremely subtle corrections to the GPS time stamp to get the correct clock readings on the Earth. This is why much of the attention has focused on this aspect. It is not that the OPERA experimenters overlooked this obvious feature (such general relativistic corrections are routinely made by GPS software in order to make the GPS system function with sufficient accuracy) but whether they have made all the necessary corrections to the extremely high level of precision required by this experiment.

Carlo Contaldi at Imperial College, London has suggested that the clocks at CERN and Gran Sasso were not synchronized properly due to three effects, one of which is the fact that the gravitational field experienced by the satellite is not the same at all points on its path since the Earth is not a perfect sphere. He says that the errors that would be introduced are of the size that could produce the OPERA effect. (You can read Contaldi’s paper here.)

Ronald A. J. van Elburg at the University of Groningen has argued that subtle effects due to the motion of the detectors with respect to the satellite could have shifted the time measurements at each clock on the ground by 32 nanoseconds in the directions required to explain the 60 nanosecond discrepancy. (You can read van Elburg’s paper here and reader Evan sent me a link to a nice explanation of this work.)

The OPERA researchers (and some others) have challenged some of these explanations and said that they will provide a revised paper that explains more clearly all the things they did.

There have been no shortage of ideas and papers pointing out problems and possible alternative explanations for the OPERA results. Sorting and sifting through them all before we arrive at a consensus conclusion will take some time.

Siri and the Turing test

I don’t have an iPhone of any kind but was intrigued by the reports of the latest one that had the voice recognition software known as Siri that seems to have a conversational ability reminiscent of HAL in 2001: A Space Odyssey, as can be seen from this compilation of a conversation.

I am not sure if this is a hoax but the person who put up the video assures skeptics that this is real and says that anyone can test it by getting hold of a Siri-enabled iPhone. I am curious if any blog reader who has it can confirm.

As an aside, I am a bit bothered by Siri referring to the user as ‘Master’. I know it is not a real person but the feudal overtone is jarring.

Taking his claims at face, it seems as if Siri is able to pass at least a low-level Turing test.

When did humans arrive in the Americas?

It used to be thought that they came 13,000 years ago across the then-existing land bridge connecting Siberia and Alaska, during what is known as the ‘Clovis’ period.

A paper published today in the journal Science has measured with high precision (with new techniques) the age of a mastodon fossil bone with a weapon point embedded in it that was found in 1970. It found that it is 13,800 years ago, with an uncertainty of only 20 years, suggesting that humans were here earlier than thought, supporting other evidence that there was human hunter activity here as early as 15,000-16,000 years ago.

A large number of mammals (mastodons, woolly mammoths, sabre-toothed cats, giant sloths, camels) disappeared rapidly around 12,700 years ago and it was thought that this must have been due to rapid climate change as the Ice Age ended, since Clovis hunters were not thought to have been around for that long.

But the new earlier date for humans in the Americas suggests that mammal extinction may have been accelerated by humans hunting them with weapons.

Relativity-7: What could be other reasons for the CERN-Gran Sasso results?

(For previous posts in this series, see here.)

The reactions to the reports of the CERN-Gran Sasso discovery of possibly faster-than-light neutrinos open a window into how science operates, and the differences in the way that the scientific community and the media and the general public react whenever a result emerges that contradicts the firmly held conclusions of a major theory.

The initial reaction within the scientific community is almost always one of skepticism, that some hitherto unknown and undetected effect has skewed the results, while the media and public are much more likely to think that a major revolution has occurred. There are sound reasons for this skepticism. Science would not have been able to advance as much if the community veered off in a new direction every time an unusual event was reported.
[Read more…]

Relativity-6: Measuring time and space more precisely

(For previous posts in this series, see here.)

In the previous post in this series, I said that Einstein’s claim that the speed of light must be the same when measured by all observers irrespective of how they were moving led to the conclusion that the rate at which time elapsed must depend on the state of motion of the observer. But if time is not an invariant entity, then we need to be more precise about how we measure it for observers in relative motion to one another so that we can better determine how their measurements are related.

What we now postulate is that associated with each observer is a grid of rulers that spreads out into all space in all directions. At each point in space are also a clock and a recorder. It is assumed that all the rulers and clocks of all the observers are constructed to be identical to each other, the clocks are properly synchronized, and the recorders never make errors. When an event occurs anywhere at any time, the location and time of that event are those noted by that recorder who happens to be exactly at the location of the event and who notes the ruler and clock readings located at the place at the instant when the event occurred. This rules out the need to make corrections for the time that elapses for the light to travel from the location of the event to the recorder.

If there is another observer who is moving with respect to the first, that person too will have her own set of rulers and clocks and recorders spread out through all space, and the location and time of an event will be that noted by her recorder using her rulers and clocks at the location where the event occurs. This set up seems rather extravagant in its requirement of infinite numbers of rulers and clocks and recorders but of course all these rulers and clocks and recorders are merely hypothetical except for the ones we actually need in any given experiment. The key point to bear in mind is that the location and time of an event for any observer is now unambiguously defined to be that given by that observer’s ruler and clock readings at the location of the event, as noted by the observer’s recorder located right there.

What ‘Einstein causality’ says is that if event A causes event B, then event A must have occurred before event B and this must be true for all observers. If one observer said that one event caused another and thus the two events had a particular ordering in time, all observers would agree on that ordering. Thus causality was assumed to be a universal property.

What we mean by ’causes’ is that event B occurs because of some signal sent by A that reaches B. So when the person at B is shot by the person at A, the signal that caused the event is the bullet that traveled from A to B. Hence the clock reading at event A must be earlier than the clock reading at event B, and this muust be true for every observer’s clocks, irrespective of how that observer is moving, as long as (according to Einsteinian relativity) the observer is moving at a speed less than that of light. The magnitude of the time difference between the two events will vary according to the state of motion of the observer, but the sign will never be reversed. In other words, it will never be the case that any observer’s clocks will say that event B occurred at a clock reading that is earlier than the clock reading of event A.

But according to Einstein’s theory of relativity, this holds only if the signal that causally connects event A to B travels at speeds less than that of light. If event B is caused by a signal that is sent from A at a speed V that is greater than that of light c (as was claimed to be the case with the neutrinos in the CERN-Gran Sasso experiment) then it can be shown (though I will not do so here) that an observer traveling at a speed of c2/V or greater (but still less than the speed of light) will find that the clock reading of when the signal reached B would actually be earlier than the clock reading of when the signal left A. This would be a true case of the effect preceding the cause. The idea that different observers would not be able to agree on the temporal ordering of events that some observers see as causally connected would violate Einstein causality and this is what the faster-than-light neutrino reports, if confirmed, would imply.

Note that this violation of Einstein causality occurs even though the observer is moving at speeds less than that of light. All it requires is that the signal that was sent from A to B to be traveling faster than light.

(If the observer herself can travel faster than the speed of light (which is far less likely to occur in reality than having an elementary particle like a neutrino doing so), then one can have other odd results. For example, if the speed of light is 1 m/s and I could travel at 2 m/s, then one can imagine the following scenario. I could (say) dance for five seconds. The light signals from the beginning of my dance would have traveled 5 meters away by the time my dance ended. If at the end of my five-second dance, I traveled at 2 m/s for 5 seconds, then I would reach a point 10 meters away at the same time as the light that was emitted at the beginning of my dance. So if I look back to where I came from, I could see me doing my own dance as the light from it reaches me. So I would be observing my own past in real time. This would be weird, no doubt, but in some sense would not be that much different from watching home movies of something I did before. It would not be, by itself, a violation of Einstein causality since there is no sense in which the time ordering of causal events has been reversed.)

So the violation of Einstein causality, not the theory of relativity itself, is really what is at stake in the claims that neutrinos traveling at speeds faster than light have been observed. This is still undoubtedly a major development, which is why the community is abuzz and somewhat wary of immediately accepting it is true.

Next: What could be other reasons for the CERN-Gran Sasso results?

Scientific responsibility

Science has a unique role in the growing recognition that it is the source of authoritative and reliable knowledge. But that carries with it a great burden to make sure that the public’s trust is not abused. Via Machines Like Us, I learned about the General Assembly of the International Council for Science (ICSU) issuing a statement last month on “The Principle of Universality (freedom and responsibility) of Science” that spelled out what the responsibilities of scientists are.

The free and responsible practice of science is fundamental to scientific advancement and human and environmental well-being. Such practice, in all its aspects, requires freedom of movement, association, expression and communication for scientists, as well as equitable access to data, information, and other resources for research. It requires responsibility at all levels to carry out and communicate scientific work with integrity, respect, fairness, trustworthiness, and transparency, recognising its benefits and possible harms.

This followed up on the second World Conference on Research Integrity held in Singapore in July 2010 that issued a statement that “emphasizes the need for honesty in all aspects of research, accountability in the conduct of scientific research, professional courtesy and fairness in working with others, and good stewardship of research on behalf of others.”

Scientists have to be vigilant in maintaining these standards.