Radiocarbon Dating
In recent years a new dating method has been developed which is based on
scientific techniques and is therefore quite independent of
archaeologically and historically based chronologies. This new technique
is radiocarbon dating which works in conjunction with its associated
dendrochronological calibration. As far as the lay public are concerned,
the impression has been given that this scientific method of dating the
past is supportive of the conventional [Egyptian] chronology [Rohl proposes
a shorter Egyptian chronology in this book]. However, this is not the
case. Needless to say, in this book I have produced a chronology which is
in direct conflict with modern calibrated high precision radiocarbon dates.
As a result, I am obliged to give a short account of the reasons why I,
and others, currently reject calibrated C-14 as a dating method.1 (I
assume here that the basic principles of radiocarbon dating are understood
by the reader.)
It is not generally appreciated but a number of historians of a
conventional persuasion are similarly troubled by the radiocarbon dates
applied to archaeological samples. Some academics are more aware of the
problems than others because these problems impinge directly on their
personal field of study, They often resort to extreme measures to deal
with the dating conflicts which arise, whilst others perceive only a small
scale problem and think that they can live with the implications. In
reality, the constraints imposed by a radiocarbon-based chronology are such
that (a) few historians would willingly abandon a history-based scheme in
favor of one based on C-14 and (b) consequently the historian tends to be
selective in his use of C-14 dates. As noted by R. Hedges:
... material is often submitted for dating in the spirit of adding a
scientific precision to the archaeologists' pre-existing beliefs.2
The Problem for Radiocarbon Dating in Early Times
In the late 1970's, when C-14 dates were corrected by use of the
bristlecone pine calibration curves, a 'problem' soon became apparent for
the third millennium BC and earlier times. In 1976, R.D. Long looked at
the published radiocarbon dates and found them generally older than the
historical chronology. He suggested that the calibration of C-14 dates
(based on American trees) was not applicable to Egypt - a conclusion which
now seems less tenable in the face of the publication of the new European
oak chronologies. The following year J. Callaway and J. W. Weinstein
produced a chronology for Early Bronze Age Palestine based on the published
C-14 dates.3 This, however, was very unsatisfactory because they had to
select out 25 of the 55 dates, a 45% rejection rate which they themselves
describe as 'disappointing' - a splendid example of the arbitrary selection
of dates by historian! In 1979, J. Mellaart suggested a stretched
chronology to accommodate the new radiocarbon dates.4 He was vociferously
criticised by his fellow archaeologists and was forced to withdraw his
proposal.5 The problems created by radiocarbon dating remain to this day:
for example, Israeli archaeologist A. Mazar refuses to use C-14 dating for
Palestinian archaeological ramains of the fourth and third millinnea BC.6
The scale of the problem is best illustrated by reference to the charts
reproduced on pages 385 and 387. [You will have to get the book.]
In 1987, F. Hole summarised the problems for the period 8000 to 6000 BP
(Before the Present) bequeathed to the Mesopotamian archaeology by the
radiocarbon method:
In view of the prevailing lack of sites ascribed to this phase, not to
mention the lack of stratigiraphic evidence for a substantial gap of time.
we can hardly be confortable with the situation: a chronological
'faultline' on either side of which exist abundant sites, but across which
the evidence is scarce to non-existent. While calibration of dates has
resolved some problems, it has created this giant enigma. The
possibilities are as follows; (1) the calibration curve is wrong, (2) the
date ascribed to written history is wrong, (3) archaeological
stratification has been grossly misinterpreted, (4) the region was
depopulated for a thousand years, or (5) the ceramics in use during this
period are not diagnostic. Since none of these seem tenable, the enigma
remains.7
In 1987, H Haas et al., using modern high-precision methods, discovered
that Old kingdom C-14 dates from Egypt were an average 374 years higher
than the historical dates established for the associated archaeology given
in the Cambride Ancient History.8 Clearly, the conventional historically
derived dating for the third and fouth millennium BC is at odds with the
dates obtained from the calibrated radiocarbon method.
The Problem for Radiocarbon Dating in the 18th Dynasty
What is much less well known is that the C-14 problem also affects Egypt's
18th Dyansty. We have towards the end of this dynasty a set of calibrated
radiocarbon dates from Tell el-Armarna which tie in acceptably well with
the conventional chronology.9 On the other hand, at the beginning of the
dynasty, we have the eruption of Thera, whose ash straddles the Late Minoan
IA period (in Aegean archaeology terms). For many years archaeologists had
tied LM IA into the early 18th Dynasty on the basis of their ceramic
chronology.10 This dated the eruption of the reign of Ahmose or later.
The date for the eruption, established by the archaeologists, has recently
received dramatic confirmation in M. Bietak's discovery of pumice within a
stratified context at Tell ed-Daba (Ezbet Helmi) which spans the period
from Ahmose to Thutmose III (1529-1425 BC).11 However, C-14 dates for
short-lived materials from the Theran eruption span the period 1760-1540 BC
with the great majority falling earlier in that period. As a result, in
1989, the Third International Thera Congress favored an eruption date
between circa 1680 and 1670 BC. Of course, this was before the discovery
of pumice in a datable Egyptian context. In the conventional chronology,
the earliest Ahmose could have reigned according to Egyptian dating is
circa 1550 BC which is at least 120 years later than the date of the
eruption established by the radiocarbon method. If the eruption took place
sometime later than the regn of Ahmose, the chronological chasm between
Egyptian archaeology and radiocarbn dating would be even greater.
The problem for historians is how to live with even a 120-year discrepancy.
Can we expand the history of the 18th Dynasty by over a centruy in order
to accommodate the C-14 dates establised at either end of the period
(eruption of Thera to the Amarna period) -- the equivalent to the reigns of
six pharaohs of whose monuments, inscriptions, officials, etc. we knew
nothing?
considerable improvement on the various dating methods used by
archaeologists and historians. As leading dendrochronologist, M. Baillie,
writes:
In terms of absolute chronology the science of dendrochronology is
superior to the disciplines of archaeologists and historians. The method
is based on a natural clock securely anchored to the present day. In
comparison, archaeology and ancient history are relative dating schemes
with no uncontested fixed points before the mid 1st millinnium BC.
Dendrochrology thus represents the ultimate chronological yardstick.12
This is an extreme position, based on years of painstaking work, building
up various tree-ring chronologies in the British Isles and continental
Europe which appear to replicate each other. Baillie's absolute certainty
permits him to argue that if there should be a significant difference
between radiocarbon and historical dates, it is the latter which need to be
revised to match the former. This is not an argument which readily
commends itself to historians.
Indeed, as we have seen, the difference between radiocarbon dates and
historical dates from the third millennium BC and earlier is of such a
magnitude that historians cannot afford to view the matter this way.
Similarly, the consequences of using C-14 dating to fix an absolute date
for both the start and end of the 18th Dynasty are also unacceptable: the
historian is required to find an extra 120 years of Egyptian 'history' for
which there is absolutely no archaeological evidence. Clearly, it would be
inconsistent to use C-14 to fix an absolute date for just one end of this
dynasty: it is therefore an 'all or nothing situation' in which radiocarbon
dating has to be embraced (with all its consequences) or rejected -- even
where the dates it yields seem to agree with the chronology.
The Nature of the Problem
Reference to the radiocarbon calibration curve supplied by the science of
dendrochronology shows that there is close agreement between the standard
decay curve of C-14 and the dendro-calibration for the last two and a half
millennia, but at around 500 BC the two begin to diverge. As we work
backwards, starting near the end-date for the TIP [Third Intermediary
Period in Egyptian Chronology], we see the calibrated dates gradually
getting older than the standard decay curve dates -- so that by 4000 BP the
gap is 250 years and by 7000 BP it has risen to 630 years. It has thus
been recognised (ever since Long's work) that the problem for historians
derives from the claibration curve. Mazar's refusal to use calibrated
dates amounts to a rejection of the dendrochronology curve. Until such
time as this problem can be sorted out, I would only be prepared to
advocate the adoption of uncalibrated dates in support of a relative, but
not an absolute, chronology.
We cannot know for sure why the various dendrochronology calibration curves
cause this problem. Unlike the dendrochronologies of the sequoia and
bristlecone pine where it was possible to read the rings of individual
trees over a period of a few thousand years, the European and recently
developed Turkish dendrochronologies have had to be constructed from many
shorter lived trees, whose rings have had to be 'wiggle-matched.' By
cross-matching sequences of narrow and wide growth-rings from different
logs the trees can be overlapped -- thus extending the chronology backwards
through time. This is a straightforward technique and should be relatively
easy to implement. However, a number of difficulties have recently come to
light. Some published chronologies have had to be withdrawn from time to
time: for example, the Sweet Track chronology from SW England which had to
be 'remeasured' after it did not conform to the already published
dendrochronology of Northern Ireland (Belfast);13 and the South German
sequence whose detailed construction was abandoned in favor of one based on
the belfast chronology -- even though its authors had originally been
convinced of its accuracy.14
Another notable weakness in the construction of the European oak
chronologies is the use of statistics. In 1991, J. Lasken raised the
problem of inflated t-values.15 A t-value is given to a wiggle-match on
the basis of a statistical analysis of the correspondence between two wood
samples. This statistical assessment is done by computer which assigns
high t-values (3 and above) to good wiggle-matches and low t-values (below
3) to those with poor correspondence between the ring patterns. In 1986,
D. Yamaguchi recognised that trees tend to auto-correlate -- that is they
possess the ability to cross-match with each other in several places within
the tree-ring sequence. He took a douglas fir log known to date between AD
1482 and 1668 and demonstrated that it could cross-match with other
tree-ring sequences to give t-values of around 5 at AD 1504 (for the low
end of the ring age), 7 at AD 1647 and 4.5 at AD 1763. Indeed he found 113
significant candidate wiggle-matches throughout the whole of the AD
tree-ring sequence.16 It is therefore interesting to note that a number of
the crucial dendrochronology sequences -- for example the Garry Bog 2 (GB2)
to Southwark sequences which connect the Belfast absolute chronology (i.e.
the AD sequence) to the 'floating' Belfast long chronology (i.e. the BC
sequence), and ultimately used to redate the South German chronology, have
t-values of around 4. These t-values are considerably lower than those
obtained for some of the historically incorrect dates produced by
Yamaguchi's experiment. Thus one would be justified in asking if the
crucial cross-links which connect up the floating sequences of the Belfast
and German chronologies are based on incorrect wiggle-matches which have
resulted from the phenomenon of auto-correlation. This prompts a second
very basic question: as noted by Lasken, should we expect tree-ring-growth
patterns to produce genuine correspondences at the same historical dates
when the climates (and in particular the micro-climates) of Ireland,
England and Germany are so different?
As a final example of the problems which radicarbon dating and
dendrochronology are currently going through I refer to a recent set of
remarkable statistics published by P. Kuniholm - the man responsible for
producing the new Turkish dendrochronology. He has been trying to date
wood from a city gateway at Tille Hoyuk.17 The problem he found was that
the t-value wiggle-match computer test produced not one but three results
-- 1258, 1140 and 981 BC. Each had a t-value greater than 4 which,
according to the dendrochronologists, makes all three 99.9% certain to be
correct! As R.M. Porter noted in 1994, Kuniholm ended up rejecting the
date with the highest t-value of 5 (981 BC) in favor of the 1140 BC,
presumably because the latter better fitted the historical expections for
the destruction of the gate.18 Kuniholm's difficulties seem to be
consisent with the problems resulting from the auto-correlation phenomenon
and raises the question of whether this phenomenon may significantly affect
the construction of all the other dendrochronologies currently being used
to date archaeological artifacts.
NOTES:
1. This appendix is based on a JACF paper written by B. Newgrosh in 1992: I
am grateful to Dr. Newgrosh for his kind permission to republish the
arguments here in an abbreviated form. See also J. Lasken, 1991, and 1992.
2. R. Hedges, 1981, p. 700.
3. J. Callaway & J. Weinstein, 1977, pp. 1-16.
4. J. Mellaart, 1979, pp. 6-18.
5. For the criticism see J. Weinstein, 1980, pp. 21-24 and B.J. Kemp, 1980,
pp. 25-28; for the withdrawal see J. Mellart, 1980, pp. 225-227.
6. A. Mazar, 1990, p. 28.
7. F. Hole, 1987, p. 562.
8. B. Kemp, 1984, p. 185; I.M. Shaw, 1985, pp. 312-13.
10. See P. Warren, 1991.
11. See editorial note and photograph in J.E. Lasken, 1992, p. 75.
12. M.G.L. Baillie, 1991, pp. 26-27.
13. New Scientist, 16/06/1990, p. 30.
14. See J. Lasken, 1991 and B. Newgrosh, 1992.
15. J. Lasken, 1991, p. 30.
16. D.K. Yamaguchi, 1986, p. 48.
17. P. Kuniholm, 1993, pp.179-90.
18. R.M. Porter, 1994, p. 20.
Bibliography:
Baillie, M.G.L. -- 1991: 'Dendrochronology and Thera: The Scientific Case'
in JACF 4, pp. 15-28.
Callaway, J. & Weinstein, J. -- 1977: 'Radiocarbon Dating of Palestine in
the EB Age' in BASOR 225, pp. 1-16.
Hedges, R. -- 1981: An article reporting on the First Symposium on
Carbon-14 and Archaeology, Groningen, 24-28 August, in Nature 29.10.81.
Hole, F. -- 1987: 'Issues in Near Eastern Chronology' in O. Aurenche et al,
(eds.): Chronologies in the Near Ease, British Archaeological Report S379,
part 2, p. 562.
Kemp, B.J. -- 1980: 'Egyptian Radiocarbon: A Reply to James Mellaart' in
Antiquity 54, pp. 25-28.
Kuniholm, P. -- 1993: Appendix in G. Summers: Tille Huyuk 4, pp. 179-90.
Lasken, J. -- 1991:'Should the European Oak Dendrochronologies be
Re-examined?' in C & C Review XIII, pp. 28-32.
Lasken, J. -- 1992:'The Radiocarbon Evidence From Thera: An Alternative
Interpretation' in JACF 5, pp. 68-76.
Mazar, A. -- 1990: Archaeology of the Land of the Bible: 10000-586 BCE (New
York).
Mellaart, J. -- 1979: 'Egyptian and Near Eastern Chronology: A Dilemma?' in
Antiquity 53, pp. 6-18.
Mellaart, J. -- 1980: 'James Mellaart Replies to his Critics' in Antiquity
54, pp. 225-27.
Newgrosh, B. -- 1992:'Living with Radiocarbon Dates: A Response to Mike
Maillie' in JACF 5, pp. 74-75.
Porter, R.M. -- 1994: 'Tree Rings from Tille Hoyuk' in C & C Workshop
1994:1, p. 20.
Warren, P. -- 1991: 'The Minoan Civilization of Crete and the Volcano of
Thera' in JACF 4, pp. 29-39.
Weinstein,J. -- 1980: 'Palestinian Radiocarbon Dating: A Reply to James
Mellaart' in Antiguity 54, pp. 21-24.
Yamaguchi, D.K. -- 1986: Interpretation of Cross Correlation Between
Tree-Ring Series' in Tree Ring Bulletin 46, pp. 47-54.