Subject: The speed of light controversy.
The speed of light controversy.
By Robert Day in the Creationism Digest
From "The Velocity of Light and the Age of the Universe," part 1,
by Barry Setterfield, in the journal Ex Nihilo, vol. 4, no.1, 1981:
"If you propose that the universe and all in it is the product of an
act of creation only 6-7000 years ago, many people ask - "How is it
that objects millions of light years away can be seen? Surely such light
would take millions of years to reach us."
Setterfield, in this and subsequent articles, proposes to set the
scientific world on its ear, stating that "The basic postulate of this
article is that light has slowed down exponentially since the time of
creation. This thought is radical and at first looks outside of
confirmation. However, there are at least 40 observations of the
speed of light since 1675 which support this suggestion." Setterfield
goes on to supply various measurements for c in a table that will
become a major point of contention in the months to come.
The range in dates in the table supplied by S (as I shall call him)
are from 1675 (by Romer) to 1976 (by Bay, Luther and White). At this
point, it is already difficult not to feel a twinge of suspicion since
40 is a strikingly low number of observations and one gets the feeling
that S certainly has the ability to be selective in his choice of
At this point, S's case rests almost entirely on convincing the
reader that he has the one and only reasonable curve to fit his
data, a curve with rather remarkable properties. In determining
this curve, S states that "Exponential curves could not reproduce
the observed rate of changes ... Power curves, polynomials, logarithmic
and hyperbolic functions were all tried with lack of success. There
was only one curve tried that fitted the data points exactly and
reproduced all of the observed features." (Note very carefully this
last statement, "the ONLY curve tried that fitted ... the OBSERVED
features." We will be coming back to this in a later follow-up posting.
It should also be noted that S's contention that polynomials were
inadequate is incorrect as Lagrange interpolation is capable of
exactly fitting n points with a polynomial of degree at most n-1.
This is already not inspiring a great deal of confidence in S's
mathematical ability but, as we shall see, this will be the least of
And just what is the magic curve? S explains, "Its general form
is a log sine curve, with a logarithmic vertical axis, typical of
most energy functions. The general equation is
Log c = A + B (Log sine (T))
where A and B are constants that have the respective values 5.47..
and -1.94.. [just the sort of thing one would expect, no? :-) -RD]"
And just where do these magic values come from?, Again, we have S,
"The time function T expresses the years (t) of time elapsed since
Creation in a scale of degrees. For example, if we take 6000 years
as being the complete range of the curve, these 6000 years take up 90
degrees of the Log sine scale so the transformation to obtain (T)
in degrees is T = t x 90/6000, that is, T = .015 t."
At this point, the knowledgeable reader can be forgiven for groaning
in dismay at such a thoroughly contrived and artificial curve, a curve
whose value at the instant of Creation is conveniently infinity and
whose subsequent date for levelling off is, amazingly enough, one of
the parameters of the formula (6000 years after Creation)! Based on
questionable selection of observations in the first place and aided
and abetted by a hideously contrived curve to fit the data, we have
S concluding that "The best date of origin for the curve as determined
by the observational evidence is 4040 B.C. +- 20 years."
But Setterfield is not yet done fudging his curve, for reasons that
will become obvious later. Having somehow magically produced this
earth-shattering curve, the first thing S does is modify it by lopping
off the head, stating that "The origin of the curve is virtually
asymptotic, but a very good estimate of the actual initial value is given
by the curve at one to 1 1/2 days from the origin. The speed of light
was then about 1.5 x 10^17 km/sec. or roughly 5 x 10^11 faster than it
is now. I will assume [for reasons that completely escape your humble
moderator -RD] that this value held from the time of creation until the
time of the fall, as in my opinion the Creator would not have permitted
it to decay during his initial work."
Sit back and ponder with me now the final behavior of the value of c
as supplied by S. Beginning with a perfectly natural(?) value of
1.5 x 10^17 km/sec at the time of Creation,
the value of c remains constant, crossing the curve supplied by S
some one and a half days after initial Creation and remaining constant
until the end of the Creation week, at which point it suddenly drops
to 6.2 x 10^15 km/sec to match S's curve, following the curve until
it plateaus around 1960. The mind reels at the rationalization that
must have been taking place.
But the best is yet to come. In defending the fit of his curve, S
makes the astonishing statement, "... the predicted curve stands up well
to four types of error analysis. The value of r^2 is 1 to nine figures
indicating a perfect fit to the data," a rather stunning claim since S
supplies, for each observation used, its corresponding predicted value,
NOT A SINGLE ONE OF WHICH MATCHES EXACTLY, demonstrating quite clearly
S's inability to comprehend even introductory undergraduate statistical
A subsequent letter to the editor questioning S's statistical analysis
and asking about X^2 (chi-squared) analysis produced the following
"X^2 is the same as r^2 in the article. This r^2 is the 'Co-efficient of
Determination' which tells how accurately the proposed curve fits the
data. If the fit is perfect the value of r^2 is 1.000000000. The DEC-10
computer at Flinders University decided that the published curve had an
r^2 value of 1.000 to nine significant figures."
Incredibly, S obviously understands the significance of an r^2 value
of 1 while not being able to comprehend how his data could not possibly
have produced such a value. S is also completely at a loss as to the
difference between X^2 and r^2, further demonstrating his appalling
S's rather odd act of decapitating his curve for the first week of Creation
suddenly becomes clear when one reads part 2 of his article, which appeared
in the creationist journal Ex Nihilo, vol. 4, no. 3, 1981, section 4,
"Binding Energy and Mesons", where we find another earth-shaking
discovery by S, namely "In the above sections we have shown mathematically
that the half-lives of the radioactive elements are proportional to 1/c
for all forms of decay." Apparently, S has now solved the problem of
radiometric dating by stating that half-lives early in history were much
shorter than they are now, implying that all radiometric dates will give
an inflated apparent age. According to S,
"... it means that given a very high value of c at the time of Creation
(as outlined in Part I), and the consequent very short half-lives of the
radioactive elements from ((66)) etc., whenever radioactive materials
in rocks throughout the entire universe are interpreted using the assumption
of constant decay rates, they will give results which are erroneously
and enormously too high. This is a necessary consequence of a decreasing
value of c."
And just exactly how inflated will these values be? Given a value of
c during Creation week that was 5 x 10^11 times faster than the current
value, each day of Creation would correspond to 5 x 10^11 days of radiometric
age, with the entire week representing 9.6 x 10^9 years. To tabulate the
effect of time following Creation week, one need only integrate over the
decay curve to get the final cumulative value of 11.5 x 10^9 years, a value
allegedly in good agreement with conventional ideas about the age of the
And it is here that one suddenly sees S's rationale for the decapitation
of his curve -- that is, the selection of the apparently arbitrary value of
1.5 x 10^11 km/sec for c during Creation week; simply so that the integral
will give the correct final answer. Selecting a higher or lower value for
c before having it plummet to catch up with the curve would allow one to
justify any of a wide range of desired final ages for the elements, and
there is no reason to believe S's value is any better than a host of others.
Rather than strengthening his model, S has simply made it appear even MORE
contrived (if this is even possible).
Having cast some doubt on S's competence, it is not a great leap
to cast the same doubt on his integrity, based on testimonials contained
in Ex Nihilo, vol. 6, no. 4, from several individuals whose academic
credentials are somewhat (shall we say?) suspect. Hyping S's
implications for modern science, we find individuals such as Dr. Gerardus Bouw
(a hard-core creationist from Berea, Ohio), Dr. Thomas Barnes (familiar
to most critics of creationist pseudo-science as the purveyor of the Barnes
theory of magnetic field decay) and Dr. Walter Brown.
But it is two other testimonials that draw the eye, from Drs. Barry Tapp
and Peter Cadusch, the authors of previous letters to the journal severely
critiquing S's work. Cadusch is quoted as saying:
"Despite extensive reworking and analysis, these determinations (of c
prior to 1940) cannot be harmonised with today's values,"
while Tapp is quoted as,
"The values of c determined between 1870 and 1940 do show a definite
Dealing with Tapp first, we find in the original letter the statement,
"The values for 'c' determined between 1870 and 1940 do however appear
to show a definite 'decay' patterning." Note that the quote is already
slightly modified, but what S fails to note is the critical tone of Tapp's
letter and Tapp's subsequent qualification "there does appear to be a decay,
given that the data is accurate," followed immediately by the statement,
"the function given by [S] does appear to be contrived, although this does
not necessarily negate the equation."
Cadusch gets even more outrageous treatment, as his original statement says,
"... despite extensive reworking and re-analysis, pre-war determinations
are now mainly of historical interest," which does not have even remotely
the same meaning as S's quote.
After having observed creationist Barry Setterfield's penchant for playing
fast and loose with other people's statements, we come to one of the most
enlightening episodes in the saga; that involving S's attempts at damage
control once his talent for this sort of behaviour started becoming evident.
In vol. 5, no. 3 of the journal Ex Nihilo, S found himself having
to justify how his statistical analysis could have possibly produced
a coefficient of determination value r^2 of 1, particularly when not a
single observation point actually fell on the curve. S dismissed this
objection by stating that "It was subsequently noticed that it [the value
of 1] had been obtained at an incorrect point in the computer programme,
and a check gave the value as r^2 = 0.99+ ...", a value that also turned
out to be incorrect and was subsequently corrected to 0.986. In all, five
different values for r^2 made their way into publication.
The fact that S and colleague Trevor Norman were incapable
of writing a computer program to determine a value of r^2 for a set of
observations certainly does not inspire confidence in their analytical
abilities, but even more frightening is the observation that neither
realized the significance of r^2 = 1, an absolutely unforgivable oversight.
Keep in mind that the basis of S's claim is that the curve in question
(which was eventually rewritten as a cosec^2 formula) is the best (and only)
reasonable fit to the data. Given this, it was of course only a matter of
time until S was supplied with an alternative fit that was just as good if
not better; a curve which, according to S's own criteria from his early
article, "fitted the data points exactly and reproduced the all of the
observed features." Ignoring the fact that the points fit S's original curve
not at all, S rejected the alternative fit by claiming that it now did not
produce the correct value for the integral of the decay curve, thereby
violating his own initial criteria for matching "observed" features by
adding another requirement for the fit which his curve possesses only
because of an arbitrary selection for his "initial" value of c.
The best is yet to come, however. S's case seems to rest heavily on
the 1675 observation of Romer, whose source S explains as follows:
"The recent critical look at Roemer's data and dates [allowing a precise
value to be placed on the orbit diameter] concluded that the best possible
result was still 0.5% above the current value ..."
The reference given is first of all to nothing more than a summary in
"Sky and Telescope"  of a full article in "The Astronomical Journal," 
leading one to wonder why S did not simply refer directly to the original
article. And the summary? It reported the findings of the Virginia
scientists who wrote the original article as,
"We conclude that the velocity of light did not differ by 0.5% in 1668
to 1678 from the current value."
It is obvious that S's description of the article is hideously inaccurate.
This point was brought up by one E. D. Fackerell in EN Tech. J., vol. 1,
pp. 87-94, to which S responded that there were two errors in the original
article by Goldstein et al, completely avoiding the accusation of deliberate
S goes on to show his lack of understanding of confidence intervals by
"The decay in c is very liberally covered by this error margin,"
while elsewhere stating,
"this finally brings us to the Bradley value which [Vivian E.] Bounds
gives as 303,000 +- 6000 km/sec ... this is still well above the current
value of 299,792 km/sec and consequently the determination is not at
variance with the changing c suggestion," ignoring the fact that the
current value of c falls WELL within the error bounds given.
S seems quite at home trying to have his cake and masticate it as well.
There is more, but I feel that this is a sufficient coverage of the topic
to expose most of the fallacies and deliberate fabrications in Setterfield's
work. I would like to acknowledge the invaluable assistance of Dr. Ken
Smith, senior lecturer in math at the University of Queensland, Aust.,
for most of the material and for having followed Setterfield so tenaciously
for the past several years.
 Sky and Telescope, June 1973, p. 353.
 "On the velocity of light three centuries ago," Goldstein, Trasco and
Ogburn, The Astronomical Journal, vol. 78, no. 1, Feb 1973, p. 122.
The Institute for Creation Research puts out a monthly monograph called
"Impact" with its "Acts & Facts" newsletter. The latest (May 1988)
"Impact" is titled "Has the Speed of Light Decayed?" by Gerald Aardsma.
The answer to that question appears to be "no, and Barry Setterfield is
a lousy researcher." While the article claims only to "caution
creationists against a wholesale, uncritical acceptance of the Norman
and Setterfield hypothesis," it's actually a pretty good debunking.
Among the problems Aardsma notes are that:
1. Uncertainty ranges are not given for most of the pre-1850 data
points. Most of those after 1850 have error bars which are too small to
be seen on the scale of the graph, while several between 1850 and 1900
are very large (resulting from indirect methods of measuring c).
Aardsma says, "For a data set consisting of measurements having error
bars of varying lengths, it is not appropriate to give every data point
equal weight as Norman and Setterfield have done. It is standard
practice to weight the data points in inverse proportion to the size of
their error bars." When Aardsma did his own weighted linear least
squares analysis, he calculated the "decay of c" to be 0.0000140 +/-
2. Norman and Setterfield's data contains a data point from 1693 which
is far higher than any of the others, attributed to "uncorrected
observations of the Roemer type, by Cassini." Aardsma quotes Norman and
Setterfield: "Observations by Cassini (1693 and 1736) gave the orbit
radius delay as 7 minutes 5 seconds. Roemer in 1675 gave it as 11
minutes from selected observations. Halley in 1694 noted that Roemer's
1675 figure for the time delay was too large while Cassini's was too
small." (p. 11 in _The Atomic Constants, Light, and Time_.
In Aardsma's words, Norman and Setterfield are using a "reworked or
'corrected' value for Roemer's c determination ... and an uncorrected
value for Cassini. It is peculiar that Norman and Setterfield were
content to use an uncorrected value for Cassini, given the comments by
the eminent and talented Halley, above. It is also unfortunate, since
this single, anomalous point is responsible for most of the apparent 38
km/s/year decay which they report. Furthermore, Roemer's uncorrected c
determination would graph below the line at -24%, more than offsetting
the uncorrected Cassini value."
Firstly, given the recent skepticism with which the ICR views the
Setterfield-Norman work, it seems that their previous glowing reviews may
prove to be skeletons in the closet for the folks at ICR. The previously
referenced creationist journal, Ex Nihilo, which contained most of the
original Setterfield work also contained, in vol. 6, no. 4, a page of
testimonials for S's work, including the two rather controversial ones
by Cadusch and Tapp discussed in previous issues. In addition, there were
statements from both Thomas G. Barnes, formerly of ICR, and Dr. Walter Brown,
now director of the Center for Scientific Creation in Napierville, Illinois.
"I agree with all that has been presented: I positively support it."
And from Brown,
"The theoretical derivation of the same cosec^2 decay function as the
computer curve, but this time from electromagnetic theory alone, makes the
whole proposition of c decay virtually unassailable."
[Ex Nihilo, vol. 6, no. 4, p. 46]
The statement by Barnes is actually not overly damaging -- it can be
interpreted as simply that Barnes views the theory as reasonably
scientific and feels that it is worth pursuing, no big deal.
Brown, however, has left himself much further out on a limb by agreeing
explicitly with the techniques used which have been shown many times to
be mathematically unsound. It would seem that the credibility of both
individuals may be on shaky ground because of this but Brown, if asked,
may fall back on a rather novel technique that he has employed in the
past -- denying having ever supported the idea.
Article 4607 of talk.origins:
From: bill@ut-emx.UUCP (Bill Jefferys)
Subject: Re: lying (again) Bales
Date: 2 Oct 89 22:11:44 GMT
References: <8909181630.AA23269@krypton.jhuapl.edu> <3444@tekfdi.FDI.TEK.COM> <10814@dasys1.UUCP>
Reply-To: bill@emx.UUCP (Bill Jefferys)
Organization: UTexas Computation Center, Austin, Texas
Posted: Mon Oct 2 17:11:44 1989
In article <10814@dasys1.UUCP> aj-mberg@dasys1.UUCP (Micha Berger) writes:
#Let me hedge. Lets not say that the speed of light had to have decreased
#logarithmicly (altho Occam would disagree). Lets try to device an experiment
#to prove whether or not it had. The constancy of the universal constants IS so
#often assumed to be constant, I don't think it has ever been provern that they
There is an extensive literature on the testing of universal
constants for true constancy in time. I recently posted a list
Here it is again. Ref  is probably the most complete.
 Barrow, J. and Tipler, F., _The Anthropic Cosmological
Principle_, Oxford, 1986. pp. 326-327.
 Shlyakhter, A.I., "Direct test of the constancy of
fundamental nuclear constants," _Nature_ v. 264, p. 340
 McCrea, W.H. _et. al._, "The Constants of Physics," in
_Phil. Trans. Roy. Soc. London_, Series A, Vol. 310, pp.
 Teller, E., "On the Change of Physical Constants,"
_Phys. Rev._ Vol. 73, p. 801 (1948).
 Dyson, F., "Time variation of the charge of the proton,"
_Phys. Rev. Lett._ Vol. 19, 1291 (1967).
 Davies, P.C.W., "Time variation of the coupling
constants," _J. Phys._, Series A, Vol. 5, p. 1296
 Wolfe, A.M. _et. al._, "A comprehensive radio study of
the z=0.524 absorption system in AO 0235+164,"
_Astrophys. J._ Vol. 222, p. 752 (1978).
 Tubbs, A.D. and Wolfe, A.M., "Evidence for large-scale
uniformity of physical laws," _Astrophys. J._ Vol. 236,
p. L105 (1980).
 Kolb, E.W. _et. al._, "Time variation of fundamental
constants, primordial nucleosynthesis, and the size of
extra dimensions," _Phys. Rev. D_, Series 3, Vol. 33, p.