The Top 30 Problems with the Big Bang

Scientific Paper
Title	The Top 30 Problems with the Big Bang
Read in full	Link to paper
Author(s)	Tom Van Flandern
Keywords	Big Bang, 20 problems, static universe, microwave background, quasar redshifts, dark matter, tired light
Published	2002
Journal	Apeiron
Volume	9
Number	2
No. of pages	19
Pages	72-90

Read the full paper here

Abstract

Earlier, we presented a simple list of the top ten problems with the Big Bang. Since that publication, we have had many requests for citations and additional details, which we provide here. We also respond to a few rebuttal arguments to the earlier list. Then we supplement the list based on the last four years of developments — with another 20 problems for the theory.

Overview

This is the best known of Tom Van Flandern's cosmological papers and probably the single most widely circulated dissident critique of Big Bang cosmology. It expands a "top ten" list Van Flandern had published in the Meta Research Bulletin in 1997, supplying the literature citations that readers had asked for, answering the counter-arguments raised on Ned Wright's well-known rebuttal web page, and then adding twenty further objections drawn from observations of the preceding four years. The result is a densely referenced catalogue rather than a derivation: sixty-odd numbered references support thirty distinct complaints.

Van Flandern's organising thesis is methodological rather than technical. His charge is that the Big Bang "no longer makes testable predictions wherein proponents agree that a failure would falsify the hypothesis", and is instead "continually amended to account for all new, unexpected discoveries". A model, he insists, "has value only when it can predict new things that differentiate the model from chance and from other models before the new things are discovered", and explanations of new results should follow from the basic theory with at most an adjustable parameter or two, "not from add-on bits of new theory". He names four alternatives he regards as at least equally consistent with the data — Quasi-Steady-State cosmology, Plasma Cosmology, his own Meta Model, and Halton Arp's variable-mass cosmology — and asks why none of them is even comparatively discussed by most astronomers.

The original ten problems

The first section restates the 1997 list with citations.

Static models fit better. Static universe models match most observations with no adjustable parameters, whereas the Big Bang matches the critical observations only with fitted parameters, one of which — the deceleration parameter — "requires mutually exclusive values to match different tests." Occam's razor then favours the static model.
The microwave background as the temperature of space. Eddington's 1926 calculation of the minimum temperature a body in space would cool to, immersed in distant starlight, gave 3 K with no adjustable parameters (later refined to 2.8 K). Van Flandern argues the same reasoning extends to intergalactic space heated by galaxy light, giving a "fog" that explains the microwave radiation and its blackbody shape. He adds an absorption argument from radio galaxies: if long wavelengths are increasingly absorbed by the intergalactic medium, microwaves from beyond all the galaxies could not reach us undisturbed or stay uniform. He notes that Big Bang predictions of the background temperature ranged from 5 K to 50 K — Gamow's revised 50 K coming only two years before the discovery — and that none qualified as a success.
Light-element abundances. Hoyle predicted the abundances of everything heavier than lithium within the Steady State model; the Big Bang "co-opted those results" and each light-element prediction "requires at least one adjustable parameter unique to that element". Remove the freedom and no genuine prediction remains.
Too much large-scale structure. At measured galaxy speeds, assembling superclusters and walls and clearing the voids takes roughly the age of the universe only if initial directions of motion are specially chosen; built up by gravitational acceleration alone it would take "in excess of 100 billion years."
Quasar luminosities. A quasar at z = 1 should be ~100 times fainter than one at z = 0.1 by the inverse-square law, yet on average they are comparably bright — requiring quasars to evolve "in just this magical way". In Arp's picture, most large quasar redshifts are largely intrinsic, and the magnitude–distance relation is a simple inverse-square law.
Globular cluster ages. The Hubble age (12 ± 2 Gyr) and the oldest globular clusters (16 ± 2 Gyr) still do not have overlapping error bars, and a newer bias-free technique lowering the Hubble age to 10 Gyr makes it worse.
Local streaming motions. Group flow of galaxies relative to the microwave radiation is seen on scales of at least 130 Mpc, with streaming on both sides of us out to 60–70 Mpc in a consistent direction and no backside infall to a "Great Attractor". Either interpretation — real flow, or the microwave radiation itself in motion relative to us — is trouble.
Non-baryonic dark matter. Over 90% of the universe must be "something we have never detected", sprinkled in increasing amounts at every scale, whereas Milgrom's modified-dynamics model gives a one-parameter account that works at all scales.
No evolution in the Hubble Deep Field. The earliest quasars and galaxies show substantial metal content rather than the primitive composition required, and ordinary galaxies are found in what should be the "dark age".
The flatness fine-tuning. The density ratio must have differed from unity by no more than a part in 10⁵⁹. Inflation having failed, the Big Bang has added the cosmological constant — itself fine-tuned against a theoretical expectation of order 10¹²⁰ — and quintessence or dark energy, "the ultimate fudge factor".

Replies to Wright

Van Flandern devotes several pages to Ned Wright's defences. On Eddington he answers that the temperature-of-space argument extends to intergalactic space, and that opaqueness is needed for a blackbody spectrum only in a finite universe — an infinite one can reach thermodynamic equilibrium through the light-carrying medium itself at scales far below quantum particles. On Lerner's structure-formation timescale, he holds that Wright's shorter figure requires fine-tuned initial velocities directed out of the voids.

The tired light replies are the most substantive. To the objection that no known interaction degrades photon energy without blurring images, he answers that his own class of models postulates entities many orders of magnitude smaller than photons, all substance being infinitely divisible. To the objection that tired light fails the Tolman surface-brightness test, he argues that a realistic tired-light model must lose energy transversely as well as longitudinally, since light is a transverse wave, giving a (1 + z)⁻² intensity dependence "in good agreement with most observations without any adjustable parameters", against the Big Bang's (1 + z)⁻⁴. On supernova light-curve stretching he concedes the effect cannot be measured directly, since the time under the light curve depends on intrinsic brightness, and argues that indirect indicators such as rise time alone do not unambiguously favour either model.

The additional twenty

The supplementary list is presented as bullet points, each with citations. Among them: thirteen evenly spaced wall-like galaxy features found in pencil-beam surveys out to more than 1 Gpc, against the Big Bang's requirement of uniform mixing above ~20 Mpc; cosmic rays above 60 × 10¹⁸ eV, beyond the theoretical limit for propagation more than 20–50 Mpc through the microwave background; the unfound matter–antimatter asymmetry; the rarity of the Gunn–Peterson trough and the lensed galaxy seen at z = 6.56, before the supposed reionization epoch; an excess of QSOs around foreground clusters too strong for lensing amplification; the charge that continuous creation of new space filled with zero-point energy violates the first law of thermodynamics; inhomogeneity persisting in redshift surveys to 200 Mpc and beyond; the shear instability of elliptical galaxies formed by mergers; mean Faraday rotation lower near z = 2 than near z = 1; a claim that the Boomerang acoustic-peak ratio favours MOND over dark matter; quantization of galaxy and quasar redshifts; the decline in quasar number density above z ≈ 2.5–3; a non-evolving intergalactic medium temperature of ~20,000 K; a required time-variation of the fine structure constant; the near-perfect power-law two-point correlation function that n-body simulations do not reproduce; super-solar quasar metallicities at z > 4; undetected damped Lyman-alpha absorbers with surprisingly uniform abundances; the mixed luminosity evolution of brightest cluster galaxies; the unsolved origin of globular clusters at early epochs; and a tenfold excess of faint blue galaxies at magnitude 28.

He closes: "Perhaps never in the history of science has so much quality evidence accumulated against a model so widely accepted within a field."

Assessment

The paper's value is as a well-referenced index of genuine open problems, and several of its entries were real and are acknowledged as such. The globular-cluster age conflict was a live crisis in the 1990s. The quasar luminosity-function decline, the origin of globular clusters, the two-point correlation function and n-body mismatch, and the faint blue galaxy excess were all standing puzzles at the time of writing. Van Flandern's methodological point is also worth stating plainly: a theory whose free parameters are increased whenever an observation surprises it does lose the capacity to be falsified, and readers who have watched dark matter and dark energy enter the model as unmeasured components are entitled to ask what would count as a failure. His willingness to name four rival cosmologies and to insist that they be compared rather than ignored is a fair scientific demand.

Against this, much of the list has not survived the two decades since. Several items were factually overtaken almost immediately. The globular-cluster discrepancy closed from both ends — Hipparcos and later Gaia parallaxes revised cluster distances and ages downward, and the Hubble constant tightened — so the error bars now overlap comfortably. The claim that Boomerang's acoustic peaks favour MOND over dark matter was reversed by the higher-precision data that followed: the third peak, weak in the 2000 data Van Flandern relied on, is unambiguously present in WMAP and Planck at the amplitude the dark-matter fit requires, and it is the observation MOND without a dark component does not reproduce. The Gunn–Peterson trough he treats as falsified by a single lensed galaxy at z = 6.56 was subsequently confirmed in a large sample of z > 6 quasars. Redshift quantization has not held up in the large modern surveys (2dF, SDSS) that were designed to test it.

The most serious weakness is in the tired-light replies, which repeatedly answer a specific quantitative objection with a qualitative possibility. The blurring objection is met by postulating entities "many orders of magnitude smaller than photons" whose properties are not specified and from which no scattering calculation is offered; the blackbody objection is met the same way. The supernova time-dilation reply is the clearest case. Van Flandern argues that light-curve stretching cannot be measured because intrinsic brightness varies — but the decisive test does not rely on the width of the light curve at all: it uses the spectral ageing of Type Ia supernovae, where the sequence of spectral features observed at redshift z runs slow by exactly (1 + z), a measurement made independently of luminosity and confirmed out to z ≈ 1. That result is a direct prediction of expansion and has no tired-light counterpart. Similarly, the Tolman surface-brightness test he claims for a (1 + z)⁻² tired-light law has since been carried out on large galaxy samples and returns the (1 + z)⁻⁴ behaviour.

There is also an unevenness of standard running through the paper. Objections to the Big Bang are held to a strict criterion — a prediction that requires a fitted parameter is not a prediction — while the alternatives are credited with successes that themselves rest on unconstrained assumptions, and the "no adjustable parameters" claim made for Eddington's 2.8 K rests on a coincidence of order of magnitude rather than on a calculation that reproduces the observed spectrum, isotropy to one part in 10⁵, or the acoustic structure. The Meta Model is invoked several times as the source of a better explanation without being set out. And the format of the paper — thirty bullet points, most a paragraph long — means that no single objection is developed far enough for a reader to judge it independently of the citations.

Read today the paper is best treated as a snapshot of the observational situation in 2002 from a determinedly dissident vantage, valuable for the questions it collects and for the discipline of its methodological demand, but with a substantial number of its individual items now settled against it.