How to Describe Physical Reality?
| Scientific Paper | |
|---|---|
| Title | How to Describe Physical Reality? |
| Read in full | Link to paper |
| Author(s) | Jean-Claude Pecker |
| Keywords | cosmology, Big Bang, tired light, Olbers' paradox, anthropic principle |
| Published | 1988 |
| Journal | Apeiron |
| Volume | 1 |
| Number | 2 |
| No. of pages | 18 |
| Pages | 1-12 |
Read the full paper here
Abstract
Because cosmology touches our ideas about what lies "beyond" as much as about what went "before," it will always raise controversy and even cause bitterness. Although I am a classical astrophysicist, I reckon that even in the solar spectrum, some elements indicate some general principle of the universe. At the same time, another universe than that of the astrophysicists is offered by mathematicians who conceive all sorts of objects and geometries. But there is too great a temptation to consider these constructions "real," as soon as they are plausible.
Overview
This is Jean-Claude Pecker's commentary on Martin Rees's paper at the 1986 Nobel Symposium in Stockholm, Possible worlds in Arts and Science. Pecker — a professional astrophysicist of the Collège de France, not an outsider — accepts Rees's account of the observations as "balanced, subtle and precise" and then attacks the interpretive scaffolding built on top of it. His targets are three: the selection of which observed facts get labelled "of cosmological significance," the standing of the Big Bang as an inference rather than a datum, and the anthropic principle as a form of explanation.
The essay's method is not to propose a rival cosmology but to argue that the standard one is underdetermined by its evidence. Pecker's recurring move is to show that a fact advertised as confirming the Big Bang was also predicted, independently and earlier or contemporaneously, by a rival model — and therefore confirms nothing in particular. His conclusion is a warning against "any dogmatic attitude in the field," and a distinction between a working hypothesis and a theory: "A working hypothesis is not a theory but only a starting point."
The argument
Which facts are "cosmological"?
Pecker lists the three canonical facts: the Hubble redshift-distance relation, with H0 somewhere between 50 and 100 km/s per Mpc; the 2.7 K Cosmic Microwave Background; and the light-element abundances. He grants each. But he notes that even the solar spectrum shows "displacement of solar spectral lines with respect to their ideal laboratory vacuum, and gravitation-free wavelengths," and asks why such facts are excluded from the cosmological set. "The selection of those observed facts which could be labeled 'of cosmological significance' is somewhat arbitrary. Could not another set of choices have led to confirm some other cosmology?"
The hierarchical universe and Olbers' paradox
His first counter-example is Olbers' Paradox. The Big Bang resolves the dark night sky by imposing a horizon: integration to infinity is illegitimate because there were no stars before a certain epoch. Pecker objects that inflation, introduced to explain the isotropy of the background, pushes that horizon back to an epoch of very high sky brightness and encompasses far more mass, weakening the resolution.
The alternative he prefers is Fournier d'Albe's and Charlier's hierarchical universe. If the density averaged over a sphere of radius R falls faster than R-1, the brightness integral does not diverge; in Mandelbrot's language the mass distribution must be a fractal of index N < 2 rather than the N = 3 of a uniformly filled volume. And this is what is observed: de Vaucouleurs (1970) found the observable universe hierarchized with a fractal index of 1.3, satisfying Charlier's condition over some 25 orders of magnitude in distance, from white dwarfs to the Lick stellar counts. Here, Pecker says, is a fact predicted by the Steady State Theory and verified forty years later — though he immediately adds, in the essay's characteristic even-handedness, "we must still admit that the Big Bang explains the facts just as well!"
He adds Seeliger's nineteenth-century paradox as a constraint in the other direction: if the observable mass were infinite, tides from the most remote parts would swamp lunar tides and tear the Earth apart. That nothing of the sort happens is a real constraint, and one that bears on inflationary models.
Static cosmology and tired light
Einstein's original static solution required the Cosmological Constant and is unstable unless empty — but Pecker observes that the hierarchical law, if universally valid, gives an average density of zero and a universe flat on large scales, so "average emptiness" is not fatal. Such a universe can accommodate a background radiation as a long-term matter-radiation equilibrium.
The difficulty is Hubble's law. Either mass is continuously replaced (Fred Hoyle's continuous creation) or "the redshift is not at all produced by motion but by some other as yet undiscovered cause." Pecker's preference is the second: Tired Light models, which he concedes "never lasted long because of a lack of detailed physical mechanisms that could be checked in laboratory," but which he thinks are being revived by LaViolette's subquantum kinetics and by his own work with Jean-Pierre Vigier on photon interactions with a Dirac vacuum. He notes that LaViolette's re-examination of the Hubble-Tolman tests — redshift versus angular size, redshift versus magnitude, galaxy and radio source counts — "finds that they indicate tired-light mechanisms more than an expanding model," while acknowledging that within an expanding frame the same tests can be read as evolution.
He also gives space to Segal's chronogeometry, which predicts a quadratic rather than linear redshift-distance law and explains the background as light from distant galaxies interacting with the traversed medium; and to Dirac's varying-constants cosmology, which he dismisses on the ground that e, h, c and me are observed constant out to large redshift and the fine structure constant "is constant in the totality of the observed universe."
The value of proofs
This is the essay's sharpest section. Gamow, Alpher and Herman predicted a fossil radiation from a hot Big Bang in 1948-54, with numerically wrong estimates. But at the same time Finlay-Freundlich (1953-54) published four papers and Max Born two more developing an undefined tired-light mechanism, and they too predicted a background of about 2 K — Born adding a footnote that the new radio telescopes might detect it. Pecker's conclusion is uncompromising: "two completely different theories predicted beforehand background radiation before it was detected; hence, background radiation cannot be a bona fide test, no matter what errors might be detected in the analyses leading to the prediction!" Predictive power, by itself, is not an argument. Facts that several theories can account for "prove nothing."
Difficulties with the Big Bang
He then turns to the age problem as it stood in 1988: the Sandage-Tammann school giving 18-20 billion years, the de Vaucouleurs school about 10 billion, with globular clusters at 15-17 billion. If de Vaucouleurs is right, "the Friedmann models cannot be considered valid anymore, and the cosmological constant must be reintroduced." Add inflation for isotropy, and "the simple model has moved progressively into a very complicated one, as one ad hoc assumption after another was added." Occam "is left to cry in his grave over the need for 'too many epicycles'." He quotes Hannes Alfvén calling the Big Bang at most a "wonderful myth" and says he is "tempted to agree," and cites Narlikar on the implausibility of a singular point.
Flatland, extra dimensions, and the anthropic principle
The last third turns on mathematical over-reach. Pecker allows that multi-dimensional geometry may be a useful tool for particle description — Kaluza-Klein, and the Green-Schwarz superstring models where particles become vibrations along strings — but insists that constructibility is not existence: "That mathematicians can fantasize about these situations does not imply that they can even exist in reality." His Flatland analogy is turned against its usual use: there is no planar flatland embedded in our space; even the flattest animal is three-dimensional in its sight and sensation.
On the anthropic principle he is blunter. Rees now concedes it "cannot claim to be a scientific explanation in the proper sense," and Pecker agrees, comparing it to Bernardin de Saint-Pierre's claim that cantaloupes are segmented so families can share them. He rejects Carr and Rees's hierarchy relations (man = (planet × atom)1/2) on the ground that each term is a continuum — "in the animal realm from the flea to the whale" — and cannot be fed into a dimensionless ratio. He is willing to speculate about other universes only if they interact with ours; if not, "I could even claim I could not care less!"
Assessment
The essay's central epistemological point is genuinely good and remains underappreciated: a prediction confirms a theory only to the extent that rival theories did not also make it. The Finlay-Freundlich and Born papers are real, and their ~2 K estimate does complicate the standard textbook story in which the CMB uniquely vindicated Gamow. Pecker states the point cleanly and applies it symmetrically — he says the same about the hierarchical universe, which he prefers. That symmetry, and his refusal to claim a decisive result for his own side, is what distinguishes the piece from polemic.
The difficulties are also real. The argument from prediction is a statement about 1988's evidential situation, and the CMB has since been characterized far beyond a temperature: the blackbody spectrum measured by COBE/FIRAS to a few parts in 105, and the acoustic peak structure of the angular power spectrum, are not quantities any of the tired-light or equilibrium proposals Pecker cites predicted or can currently reproduce. He does not, and in 1988 could not, address these. Likewise, the fractal index of 1.3 that he leans on has not survived: later large redshift surveys find the hierarchy flattening toward homogeneity above roughly 100 Mpc, which removes the Charlier resolution of Olbers' paradox he offers. And the age discrepancy he treats as fatal to the Friedmann models has since narrowed from the other end, as the Hubble constant and globular-cluster ages were both revised.
The most serious gap is the one Pecker himself names and does not fill. He concedes that tired-light models "never lasted long because of a lack of detailed physical mechanisms that could be checked in laboratory," and then endorses them anyway on the strength of two programmes that were, at the time, promissory. Any tired-light mechanism must scatter photons without blurring distant images and without a wavelength-dependent residue, and it must reproduce the (1 + z) stretching of Type Ia supernova light curves — a measurement that postdates this essay but that a static interpretation of redshift still has to answer. Pecker's own reservations about the state of the evidence would apply with full force here.
What the paper does establish, and what a reader should take from it, is a methodological demand rather than a cosmological one: that the set of "cosmologically significant" facts is chosen and not given, that agreement between a theory and a fact both of which were available at the time of construction is weak evidence, and that mathematical possibility is not physical reality. Coming from a mainstream astrophysicist responding to a mainstream cosmologist at a Nobel Symposium, it carries a weight that the same argument from an outsider would not.