Inconstant Constants

Science changes our understanding of the universe all the time. We move from vacuum tubes to integrated circuits to quantum computers. Hypotheses are proposed and then tested. But for the last hundred years or more, there's been an important assumption that all scientists share: there's nothing different about one patch of space compared to another patch of space. There are certain physical constants that should always be the same no matter where you measure them: Earth, Mars, Alpha Centauri, or the Andromeda galaxy. We call these values the fundamental constants, and they include things like: the speed of light, pi, the value of an electron's charge, the Planck constant, and the electrical permittivity of free space. So far that's always been a sound assumption. We can't measure things on Betelgeuse yet, but measurements taken on different places on and near Earth, and at different times, have so far always come out the same. In the last fifteen years however, some intriguing evidence has come to light that may end up undermining that assumption and changing a lot about how we understand the universe.

Most science fiction writers tend to play in the sort of standard constants-are-constant universe that we all grew up in. They may have a 'warp drive' powered by unobtanium, but the speed of light that is being broken is probably always assumed to be 299,792,458 meters per second (in a vacuum), everywhere and everywhen. Still, a few authors have considered a universe in which the laws of physics change from region to region. Vernor Vinge's A Fire Upon the Deep may be the most famous example of this, where physics as we know it is a by-product of being in a 'Slow Zone' — in other zones superluminal travel is possible, as well as other highly advanced technologies. Alastair Reynolds does something similar on a more local scale in Terminal World, where the ancient remains of some sort of starport projects a field that divides up local space into advanced regions, where AI and high-tech are possible, devolving down to less advanced regions, finally degenerating into Horse Town, where nothing more complicated than a flour mill is possible.

In Greg Bear's Eon it's an alien artifact that has people traveling from one set of multiverse physics to the next; I've always remembered fondly the 'fundamental constants meter' that one character developed to let them know when pi or the gravitational constant might have changed. Greg Egan, as one might expect, has played with this concept several times: mathematically in short stories "Luminous" and "Dark Integers"; on the quantum scale in Schild's Ladder; and most elaborately in the forthcoming Clockwork Rocket, kicking off a trilogy set in a universe where the speed of light is not constant thanks to one sign change in the equations governing relativity. These are all interesting thought-experiments, and more often than not make for some interesting story settings. And until recently there's been no good reason to think of them as any more reflective of our universe's structure than your average secondary world fantasy setting.

Now, however, we have some fascinating astronomical and terrestrial evidence, and there is a possibility that a particular fundamental constant of the universe, one labeled alpha, might be variable over time and space. One thing I love about this story is how it highlights the amazing flexibility of our understanding of the universe, as well as the small-c conservative nature of the scientific method. The evidence we have is intriguing, and theoretical physicists have already been examining our models of the universe to see what impacts a variable alpha would have. However, no one is claiming that the final answers are in yet — the wheels of physics move even slower than those of justice, but they do gratifyingly keep moving.

Alpha, the main suspect here, is called the "fine structure constant." It is a number made up of other fundamental numbers. If you combine such values as the charge of an electron, the Plank constant, the speed of light, the electric permittivity of free space, and pi, all together, you get a unitless number that we call alpha. It's valuable since it will be the same no matter what unit system you use to measure its components (i.e. whether you measure distances in kilometers or miles).

This may feel like a fairly abstract exercise in mathematical convenience, but alpha turns out to have real physical relevance. Specifically, it governs the extent to which two electrons repel each other, including relativistic effects. This has many ramifications, but the observable effect is how electrons jump between energy levels in the atom, causing specific spectral lines when you look at light through a spectrometer. The 'gross structure' of spectral lines has to do with electrons jumping between energy levels. The 'fine structure' is caused by the same phenomena, but with corrections for relativistic effects — thanks to alpha.

The latest word is that alpha may not have the same value in all times and all places. What is the evidence behind that claim? It comes from two sources, astronomical and terrestrial. The astronomical data starts in 1994, and comes from astronomers looking at quasars. The important thing about quasars is that they are very bright (among the brightest objects in the universe) and very far away. This means that they are very old and moving very fast. (In the 1920's, Edwin Hubble discovered that objects farther away from us have higher relative velocities, due to the expansion of the universe). They are ideal subjects to observe because of their brightness (we can see them), and also because they are moving so fast relative to us that small relativistic effects, such as the ones governed by alpha, are magnified.

The astronomers looked for quasars, but also for gas clouds that lay directly between a quasar and the Earth. Since alpha is most easily measured by looking at spectral lines, a gas cloud provides a perfect set of 'absorption spectra' for astronomers to examine. As the quasar light passes through the diffuse gas cloud it inevitably knocks a few atoms around, causing electrons to change levels and re-emit new, altered, photons. This process leaves dark lines in an otherwise continuous spectral rainbow, and these are the lines that allow us to measure alpha.

The first study involved data taken between 1994 and 1996, and had 128 quasar/gas cloud pairs. It showed statistically significant changes in alpha of almost one part per million over the last 6 to 12 billion years. That doesn't sound like much, but alpha doesn't need to change much to have big effects. John K. Webb at the University of New South Wales in Australia led this research effort. His team understood how controversial the findings were, so they did not publish until 1999 — 'extraordinary claims require extraordinary evidence.' They spent the extra time testing the data for any alternate explanation they could think of (e.g. a spectroscope being miscalibrated, data processing errors, different elemental ratios than expected, etc). When they felt that they had eliminated all reasonable sources of error, they published their paper ("A Search for Time Variation of the Fine Structure Constant").

The study made quite a splash, so another team went out to seek additional evidence. It is a testament to the relatively glacial pace of astronomical studies that the first data was collected in 1994, the first paper was published in 1999, and only in 2010 has a new wave of data and papers been produced ("Evidence for Spatial Variation of the Fine Structure Constant"). Telescope time is not cheap and often must be scheduled years in advance. The second team was not able to replicate the first team's results. Team 1 (Webb's team) believes that Team 2's analysis is fundamentally flawed — and that finding does not seem to be under dispute.

Team 1 then took the second set of data and found something that confirmed their original hypothesis, but in a surprising way: at first, it looked like the new data showed alpha changing in the opposite direction — that it was larger in the past instead of smaller. After much head scratching, they realized that the original data was taken with the Keck telescope in the Northern hemisphere, and the new data was taken with the ESO Very Large Telescope in the Southern hemisphere. Alpha may be changing in space as well as (or 'in addition to' or 'instead of') in time. It must be said, however, that Team 1 using new data to reconfirm Team 1's original hypothesis, despite differing findings from Team 2, is not the preferred progress of the scientific method. It will be much more convincing when Team 2, or 3, or 4, also confirms Team 1's findings.

Another point in Team 1's favor is that the astronomical data is backed up by a phenomenon closer to home. Alpha impacts the rate of some nuclear reactions, so digging up nuclear isotopes from the past may also give us some indication of whether alpha has changed with time. The Oklo mine in Gabon (Africa) is a natural nuclear fission reactor — a uranium deposit where at least some self-sustaining nuclear reactions have occurred. Data from Oklo published in 2004 indicates that alpha may have changed its value over the past two billion years. However, this data hasn't yet been replicated — so far the Oklo mine is a unique phenomenon on Earth.

Needless to say, findings as controversial as these have their share of opponents, and Dr. Webb readily admits that the data his team has gathered so far, while tantalizing, is nowhere near conclusive. Some critics suggest that there may be other explanations for the opposite directional findings of the two telescopes — and also that it would be a mighty coincidence if the plus/minus axis of alpha's change lined up with the north/south alignment of the Earth. Others have pointed to possible contradictions between a varying alpha and other observational data. Everyone agrees that more observations and data are needed. There's already a proposal out to measure alpha as it would have appeared near the beginning of the universe, by imaging the cosmic microwave background radiation. However, the instrumentation is not yet able to achieve fine enough resolution to be useful.

I mentioned above that alpha is a composite number, made up of relationships between other fundamental physical constants. If alpha is variable, which of the constants involved might be changing? Unfortunately, none of the experiments done so far can make that distinction. Nothing about nuclear fission reactions or absorption spectra from gas clouds could tell us if it's the speed of light, the charge of an electron, the electrical permittivity of free space, or some other component that may be changing over time and space. Obviously, the effects of this variability would be very different depending on which constant or constants turn out to be flaky. Different experimental data will be needed to tease apart all of the potential suspects.

Interestingly, if alpha does turn out to be changing, this wouldn't necessarily mean a revolutionary upset of physics-as-we-know-it. It would upset the Standard Model of physics, but that is already under any number of challenges. It is possible that a variable alpha could fit neatly within a string-theory based model, which is one of the reasons the researchers have been willing to entertain the possibility that we're looking at a real change, and not simply measurement errors.

What does all this mean for us? Well, if electrons interacted differently, if alpha varied by only 5% from its current value, life-as-we-know-it wouldn't exist. From J.D. Barrow's Cosmology, Life, and the Anthropic Principle:

"For instance, were α to change by 4%, stellar fusion would not produce carbon, so that carbon-based life would be impossible. If α were > 0.1, stellar fusion would be impossible and no place in the universe would be warm enough for life."

So it's definitely a big deal. On the other tentacle, even if alpha is changing as much as Webb et. al., estimate, there is no place in the universe where it would be 4% different from our value — it's not changing that rapidly over space. (The astronomical data show potential changes of one part in a hundred thousand or less over 6-12 billion years, and the Oklo data suggests a change over 2 billion years of 4.5 parts in 100 million.)

Finally, there's this conclusion to Webb et. al.'s 2010 paper:

"Qualitatively, our results suggest a violation of the Einstein Equivalence Principle, and could infer a very large or infinite universe, within which our 'local' Hubble volume represents a tiny fraction, with correspondingly small variations in the physical constants."

Now that made me sit up and take notice. It echoes a statement that Dr. Webb made in an April 2003 article in Physics World ("Are the laws of nature changing with time?"):

"For instance, the equivalence principle — one of the cornerstones of relativity — states that in freely falling reference frames, the outcome of any non-gravitational experiment is independent of when and where it is carried out."

Changing alpha would violate this principle.

Could our universe be infinite? Could it be stranger than we expect? We're not sure yet, but at least we have one avenue to investigate and find out.

Tell a friend, share this on:

ISSUE 59, August 2011





Karen Burnham

Karen Burnham is an electrical engineer working at NASA's Johnson Space Center (all opinions expressed in this piece are her own, and no reflection of NASA's official policy). She is also a book reviewer writing for Locus, Strange Horizons, SFSignal, and Cascadia Subduction Zone. She writes non-fiction about science when she can, and recently started writing for


Also by this Author



Amazon Kindle

Amazon Print Edition


Weightless EPUB/MOBI