This article, along with the corresponding article on Forestation, was prepared in 1994 and updated
in 1998 for inclusion in the Archaeology in Anatolia Encyclopedia, a compendium that, for financial reasons,
was never published. While we remain hopeful that someday the article may get into press, we have included
it in our web bibliography because some people, somewhere, might find it useful in the here and now.
Please note! Absolutely no effort was made to bring this article up to date before publishing on the web (2004), and scientific and
technological methods for establishing time-control in archaeology have lept into the future since the original writing in 1994.
The three most important scientific chronological techniques employed in the archaeological study of Anatolia which permit the tightest time-controls are radiocarbon, dendrochronology, and thermoluminescence. A fourth, but indirect, technique is palynology or the study of pollen profiles. Other dating techniques, including those used for much earlier geological periods--some of which are often less relevant for the last ten thousand years or are still in the process of development--include varve dating, archaeomagnetic dating, obsidian hydration, potassium-argon dating, fission-track dating, electron spin resonance dating, and amino acid racemization.
Scientific chronologies are relatively new phenomena in Anatolian archaeology, and it is probably fair to say that the Anatolian applications of most of the dating techniques are still in their infancy. Radiocarbon with a four decade history since it was 'discovered' in 1950 is the most well-known of these techniques in Anatolia, yet only rarely is it possible to perform either radiocarbon determinations or derive dendrochronological dates on material from excavations before 1950 because almost all such material was discarded by early excavators.
Radiocarbon, the "oldest" of the dating techniques from the Anatolian point of view, depends on the measurement of the steady decay of the radioactive isotope (14C) of carbon as it transmutes into a stable isotope of nitrogen (14N), losing half its radioactive nuclei every 5730 years. First hailed by some archaeologists as the final solution to all their dating questions, the technique has caused almost as much controversy as it has solved problems. The problem began with the attempt to find the true half-life of 14C. Libby's initial estimate of a 5568-year half-life was first modified by 3% to a 5730-year half-life. Then the radiocarbon time scale underwent even greater change as it was calibrated by absolutely-dated tree-ring chronologies, first of the bristlecone pine, then of the European oak. This calibration reflects the recognition that, although the decay of radioactive carbon is constant, its production throughout time is not.
Archaeologists who have taken the trouble to understand the radiocarbon method have tended to use it with faith. Others, particularly those uncomfortable with statistics, have discarded it as hopelessly complex. And for some researchers, the radiocarbon curve is too "flat" in some key historical periods to be of any real use. Two of these periods in particular are just at the beginning of the Early Bronze Age around 3000 B.C. and again for the years around 850-400 B.C. In practice, this means that two of the most important periods of time in Anatolian archaeology--the beginning of the Bronze Age and the entire pre-Hellenistic Iron Age (leading to the rise of the Classical period) --are represented, respectively, by Before Present (BP) dates of 4150 and 2450. These apparently specific BP dates are not as precise as they might seem. Because of the "flatness" of the curve, they translate roughly into a range of 2950 to 2600 B.C. for the beginning of the Early Bronze Age and a range of 800 to 400 B.C. for Geometric/Archaic/Classical, all of which is far too imprecise for practical use.
Radiocarbon dating is an expensive technique, and many excavators, daunted by the cost of submitting a set of samples for determination, send in only one or two samples, with subsequent loss of precision in their dates. A set of ten or more would be far preferable. Accuracy to within 5% error can be reduced to 2% by the use of accelerator mass spectrometry (AMS) dating, but the latter technique is even more expensive. Wiggle-matching, which is the radiocarbon dating of selected dendrochronologically dated tree-ring sequences, can reduce the error margin to around 1% or less.
The general outline of early Anatolian chronology as dated by radiocarbon, set forth in the recent summary by Mellink in the third edition of Chronologies in Old World Archaeology in 1992 (COWA 3), is probably roughly correct, at least for the relative sequences, even though the research is already nine years out of date. The tables list only 210 radiocarbon dates from 48 subsets of sites covering 6500 years of prehistory from about 8000 B.C. to about 1500 B.C. (uncalibrated). Fifty-six of these dates have error margins between ±100 years and ±570 years. Many sites are represented by a single determination or at best two, or three, or four, and all the dates need now to be reconsidered and recalculated with the newest calibration program in Stuiver, et al., (1993), and they need to be combined with many others that have now been published elsewhere or are to be published. "Basal" Mersin, for example, is represented by one radiocarbon date with an error margin of ±300 years. The true potential of radiocarbon for Anatolian chronology is best seen in Korfmann et al., (1987) for Demircihüyük, and in Korfmann and Kromer (1994) for Troy and Beşiktepe, where literally hundreds of radiocarbon determinations have now been made and integrated with the stratigraphic sequences from which the samples were collected, and where statistically significant comment can at last be made. S. W. Manning's long-awaited book (1995) resolves some of the unanswered questions left by COWA 3.
Dendrochronology, the identification of tree-ring growth sequences that are unique in time, is the only archaeometric technique where determination of absolute dates accurate to the year is either theoretically or practically possible. Schweingruber provides a thorough treatment of recent progress in dendrochronological and dendroclimatological work world-wide. The method is limited only by the scarcity of well-preserved, long-lived wood or charcoal samples in some periods. Dendrochronology is also the least expensive of the archaeometric techniques.
Over 6500 years of tree-ring chronologies covering much of the period back to about 7000 B.C. have been developed at Cornell University over the past 22 years for Anatolia and surrounding regions. The goal is to have an unbroken tree-ring chronology from the present back to the Neolithic period against which archaeologists, anthropologists, and art historians may date finds of wood or charcoal with a theoretically possible precision of one year.
Progress in building a long chronology for Anatolia and environs as of spring 2002 is shown in the accompanying chart. About 6500 years of tree-ring chronologies are spread out over the last 9000 years, covering a region bounded by Georgia in the East, the mountains of northern Lebanon in the south, all of Turkey, Cyprus, Greece, parts of Bulgaria and Yugoslavia, and southern Italy in the west. Cedar and juniper found in Egypt but imported from Lebanon are expected to crossdate with the Anatolian chronologies.
Absolute chronologies for several genera of trees extend back from the present to A.D. 360. The millennium from roughly A.D. 500 to 500 B.C. is the most problematic of the last 4200 years because so few samples have been found, although the B.C. portion has been partially filled in by the recent construction of a 513-year ring-sequence from boxwood timbers in the Comacchio (Ferrara) shipwreck. The Comacchio sequence, with which a number of other ring-chronologies crossdate, is dated to the last decades B.C. by three and a half tons of lead ingots stamped with the name AGRIPPA. A 1598-year continuous ring-sequence from 627 B.C. to 2224 B.C. is the longest chronology as of 1998. In the absence of a connecting link to modern, absolutely-dated tree-ring chronologies, this chronology is placed by wiggle-matching of selected decade-long wood samples, and is accurate to within ±37 years. A 503-year Early Bronze Age sequence, also wiggle-matched, probably overlaps with the beginning of the 1598-year chronology by about 35 years but cannot yet be dendrochronologically crossdated with it because of the short overlap. For the prehistoric period the most notable advances in the last few years have been the development of a 298-year sequence for Late uruk Arslantepe VIA ending at 3374 ± 30 B.C. and a 570-year chronology for the Neolithic site of Çatal Höyük ending at 6449 +20/-39 B.C. If the chronologies could be completed down to the present, no plus-or-minus factor would be needed for even the earliest of these chronologies.
When a clay pot is fired to 500 degrees Celsius, all electrons in the clay flash off, and the clay is "clean." As time goes by, the pot is bombarded by environmental radiation, and new electrons are gradually trapped in the lattice or matrix of the clay. By firing the pot again to 500 degrees and measuring, with a photovoltaic cell, the accumulated electrons that flash off, the scientist can determine how much time has elapsed since the pot was last fired to this temperature. Since potsherds exist by the millions in almost every level since the invention of pottery in the Neolithic, as do fired tiles, baked bricks, and terracotta figurines, thermoluminescence dating, in theory, should be one of the most practicable of the archaeometric techniques. In practice, however, there are enough limitations and complications to make anyone but the most dauntless chemist give up in despair: the chemistry of the clay and its inclusions, intricate measurement problems, problems such as anomalous fading, prompt luminescence, spurious thermoluminescence, and questions of the stability of the thermoluminescence record are only a few. One of its most devoted practitioners admits that even under the best of circumstances thermoluminescence allows an estimate at the 68% confidence level that is 7% to 10% in error. For some classes of pottery TL dating is in effect impossible, and no figures can be cited. For the tight chronological distinctions archaeologists would like to make in Anatolia for the past 8000 years, thermoluminescence is unsatisfactorily imprecise.
Pollen analysis is not properly classifiable as an independent, self- standing chronological technique because it relies on radiocarbon for its absolute dates. Workers are attempting to establish a regional vegetational history over the last 20,000 years including the matching and synchronization of complex pollen diagrams. Since the palynologist matches undated diagrams against others which are radiocarbon-dated, chronology is very much a part of any pollen study (van Zeist, et al., 1968, 1975). Imprecision is an important problem in palynology, with dates usually cited to the nearest millennium, only occasionally to the nearest half-millennium, and even less often to the century. The method nevertheless has considerable potential, especially since AMS dating is now available for milligram-sized organic particles, and long cores can be divided into chronological subsets with far greater precision than was allowed by the old one-or-two dates per core that characterized the early pollen work.
Varve dating depends on the accurate counting and measuring of annually deposited layers of silt. Proper application of the method requires an undisturbed sampling spot where sea or lake currents have not interfered with the depositions. A 2000-year varve sequence exists for the north shore of the Black Sea, but the method has not yet received much attention in Anatolia. Lake Van is one theoretically likely spot for such study.
Archaeomagnetism, where one measures how much the alignment of a burned object's magnetic field (thermoremanent magnetism) differs from the earth's current magnetic pole and then calculates how much time has elapsed since the last burning, is another theoretically likely dating method whose error margins are still insufficiently refined for useful application in Anatolia. Proper application requires extraordinarily precise measurements in the field, both of the structure under study as well as of structures of archaeologically known date. This last provision is a cause for concern since many so-called "archaeologically-known" structures are not all that well-dated themselves, and the danger of a circular argument is ever-present.
The methods listed in the first paragraph above, as well as others, are still in various stages of development (Aitken, 1990). At an international archaeometry meeting one specialist in amino acid racemization referred to the process as "an art-form rather than a dating technique." Eventually, however, these methods should produce useful chronological information.