It is a pleasure for me to write a paper on an important issue in Anatolian archaeology for my dear friend and colleague Aykut Cinaroğlu, whom I first met in 1984 when he had a Fellowship at the Metropolitan Museum of Art (later in Ankara I met often his wonderful wife and daughters and their sweet dog, Tarçin).

INTRODUCTION

This paper continues the discussion presented in Muscarella 2003 (see also Muscarella 2005/2006: 395, and note 4) concerning the date of the destruction level (DL) at Gordion that terminated the Early Phrygian (EP) period there. I argued that the destruction occurred some time close to 700 B.C. (+/-), not in the late 9th century B C. as maintained in publications and public lectures by the Gordion Team excavators since 2001. The aims of the present paper are to augment some of the issues I raised previously and to present additional and relevant information, thus to expand the data available in the published record. One of the stimuli that generated this review is the growing number of scholars who have uncritically (to me, without reflection) accepted the 9th century B.C. destruction date, and thereby simultaneously embraced the consequent profound historical and archaeological implications for first millennium B.

�

C. Aegean and Anatolian archaeology and history. For example, Prayon (2004: 611) states that the New Chronology has a “weitreichenden Konsequenz für die historisch-politischen wie auch künstlerischen Entwicklungenun Zusammenhänge,” which he then proceeds to document; Prayon and Wittke (2004: 122-23) note that “das bisherige Bild der phrygischen Kultur und des Phrygischen Reiches grundlegend verändern….” See also Kelp (2004: 286, 293); Strobel (2004: 259, 265-68); Genz (2004: 221, 224); Dusinberre (2005: 4, 10, 220-22); Crielaard (2007: 223); and Summers (2006: 2).

In January 2001, a laboratory (Heidelberg) informed the Gordion Team1 that based on C-14 analysis, the EP citadel of Gordion had been destroyed ca. 830-807/800 B.C. (for details see Muscarella 2003: 225-6, 250). The report was immediately and unhesitatingly accepted: for here, was an “objective scientific” fact presented by a scientist working in a scientific laboratory, and thus the previously maintained, for decades, archaeologically argued and thus “subjective” dating of the destruction, ca. 700 B. C., was rejected. The New Chronology, as it came to be designated, was declared a fait accompli, one vigorously upheld by the Gordion Team. The laboratory report was first publicly announced at the Fifth Anatolian Iron Age Conference in Van in August 2001 (where, verbally, I first challenged it; for a modified version of the Van announcement see DeVries et al 2005). In the same year a brief statement reporting the New Chronology was published (Manning et al 2001: 2534; see below). Two

Osc

ar W

hite

Mus

care

lla -

Aga

in G

ordi

on’s

Earl

y Ph

rygi

an D

estr

uctio

n D

ate:

ca.

700

+/-

B.C

.

AGAIN GORDION’S EARLY PHRYGIAN DESTRUCTION DATE:

ca. 700 +/- B.C.

Oscar White Muscarella

1 The Gordion Team consists primarily of a quartet, Mary Voigt, KennethSams,KeithDeVries,andPeterKunihlom.

years later, a more expanded announcement on the revised chronology, here provided with brief C-14 specifics (in Table 1) was published (DeVries et al 2003); this publication overlapped with Muscarella 2003. To date (2007) my 2003 rejection of the presented C-14 date has been supported in print by only one individual, Keenan (2004)—based independently on his C-14 interpretation of the analysis (see also Porter 2005: 66 and note 2). The Gordion Team has to date never acknowledged the controversial nature of the++ new dating.

In Muscarella 2005/2006: 395, note 4, I reported briefly (here expanded) that I had been informed (in February 2005) that the Gordion Team had been notified that more recent work on analysis had caused them to lower the originally presented C-14 chronological range by 40 to 60 years. Such a change effectively lowers the destruction date, in this calculation, to a time close to the mid-8th century B.C. I made two enquiries about this information via e-mail to two Gordion Team members (August 2005). One did not answer my question, except to claim that a work was in progress on Gordion chronological issues. The other never responded. These unprofessional equivocations and non-responses tell us that something is wrong within the Gordion Team—and concomitantly suggest that the information I received in 2005 is correct. And if correct, then the three-years of silence on the expanded date range is a serious offense to scholarship.

DENDROCHRONOLOGY

The starting point for this dramatic back dating of the destruction at Gordion was generated solely from the C-14 laboratory report. But quite soon thereafter dendrochronological evidence from structural timbers recovered in the DL Terrace Building (TB) 2A was soon introduced as both further and independent evidence supporting the 9th century date (Manning et al 2001: 2534). Here four scientists (a Cornell Team) report (in note 28 citing as the source M. Voigt “personal communication”), that “The last preserved ring from construction timbers in Terrace Building 2A…is now dated ca. 883 +4/-7 B.C.,” and which terminal ring-date “approximates construction of this building, which was destroyed…between 830 and 800 B.C.”(italics are mine; see also Voigt 2005: 30-1; DeVries

�

2005: 37 declared with ease that the destruction occurred “not later than 805 BC”!). The crucial message casually introduced here (by another university laboratory) is that an unmeasurable number of missing rings from burnt timbers is irrelevant in chronological and archaeological investigations; they are not to be questioned by archaeologists seeking to determine the absolute date a structure was constructed. This interpretative leitmotif continued over the years.

Earlier, Voigt (1994: 273) had provided an accurate archaeological analysis of the absence of chronological value of the very same TB 2A roof beams. She noted that “Kuniholm states” (personal communication?) that “the latest ring on these timbers is 908 +/- 37 BC” (on this specific date see CC3 below). But she correctly saw fit to record that no bark was preserved on the samples, the number of missing rings is unknown, the beams were burnt and also suffered excavation-process damage, and, further, that the “beams represent re-used or stored timbers, cut long before this structure was built.” This is a precise and neatly stated archaeological interpretation: but fully ignored by the Cornell laboratory (in Manning et al 2001). It should not go unnoticed that Voigt’s 1994 disinterested interpretation of the TB 2A timbers was made before the revelation of the New Chronology—and the later 2001 chronological claim was made after that revelation (see also Kuniholm on CC 3, below; for others who raise the issue of wood re-use as a fundamental component in dendro interpretation see James (1991: 323), Keenan (2006: 11), Mielke (2006: 83), and (yes) Manning et al (2007: 4). Thus, over the years the date of the preserved rings went from 908 to 883 +/- B.C.—with no discussions regarding the reasons for the changes.

DeVries et al (2003: 1), and Voigt (2005: 30-1) suggested the possibility of wood re-use for DL structural timbers. Contrary to the two other dates previously reported for TB 2A—908 and 883 B.C., Voigt now writes that “the latest ring [is] dated ca. 861 B.C.,” again with no references to previously published dates or reasons for the modification.2 The use of Manning et al in 2001 of TB 2A timbers in support of

Osc

ar W

hite

Mus

care

lla -

Aga

in G

ordi

on’s

Earl

y Ph

rygi

an D

estr

uctio

n D

ate:

ca.

700

+/-

B.C

.

2 Personalcommunicationscanbedeceptiveandmisleadingandessentiallyundocumented.Forexample,arecentemailmessagetome(October3,2007)fromaninterestedpartyclaimsthatKuniholm(inapersonalcommunicationtohim)believes“thelastringofTB2ais850B.C.”Ithinkitbestnottociteitinmytext.

Prof. Dr. Aykut Çınaroğlu’na Armağan

a specific chronological determination is therefore puzzling. For aside from the wood-condition problem along with the correct archaeological conclusions articulated earlier by Voigt, the new ring determination ignores, indeed contradicts, previously published dendrochronological evidence and interpretation acquired from another EP/DL structure, CC3. Wood recovered here manifestly demonstrated that burnt beams cannot be used to furnish construction dates of structures at Gordion. I quote Kuniholm (1988: 8), “It is clear [sic] that wood was re-used at Gordion, sometimes many centuries after it was cut”—a conclusion he determined from three pieces of wood here that “were cut about four centuries earlier” in the 12th century B.C., 400 years before eight other wood pieces from the same structure.3 Concerning the later-cut CC3 timbers, Kuniholm (1988: 8) informs us that “....the CC3 Master, composed of eight samples,” [8 separate timbers?] ends in MMTRD 1826—sixty two years after MMT [tumulus MM: Midas Mound].” That is, they were cut 62 years after the cutting date of the outer logs surrounding MM: MMTRD 1764 represents the terminal ring that grew just before the tomb’s logs were felled (ibid: 5). All the final rings of the tumulus MM logs have not been preserved (even though the bark is extant) but Kuniholm “thinks” he has the accurate date. The cutting date of these logs is presented as 740 +4/-7 B.C. by Manning et al (2001: 2534); previously it had been dated to 718 +1/-1 B.C. (see DeVries 2005: 43). To date the ca. 740 B.C. determination has not been revised (although given past prouncements it may very well be.4

�

The 740 MM date is more or less close to that suggested by some scholars based on artifact analysis and ancient records (Muscarella 1982: 9, idem 2003: 230-1, and note 28; Caner 1983: 9, 201). Missed by me and others, including the Gordion Team, is that Kuniholm 1988: 6-7 noted in a chart (but not in the text) that six other structures in various strata at Gordion contained wood re-used centuries after the log’s cutting date.5

For the record, here are some chronological implications of Kuniholm’s 1988 report on CC3, working from his designated cutting date for tumulus MM, and subtracting 62 years. One arrives at these possible dates for the last preserved ring (not cutting) of wood beams in the DL, 682, 678, or 671 B.C. But (not employing Gordion Team dendro analyses), these dates surely only indicate that EP Gordion was destroyed sometime after these years in the early 7th century B.C. (recognized by James 1991: 323-4—but not by me in 2003).

However, the information and conclusions published about CC3 in 1988 have now been thrown into the bulging file of altered and rejected data labeled “Old Chronology.” In an email communication to me (June 29, 2007),6 Kuniholm rejected and modified his previously published conclusions, and changed both the published data and his analysis of the 400-year tree ring spread along with its chronological implications. He wrote that the samples were “charcoal,” (in another email publication, July 12, 2007, they were “horribly burned”), and further, that in 1988 he was then--nota bene— “operating under the old assumptions about tumulus MM and the DL, and in retrospect was trying to force a fit where there really wasn’t any” (italics mine; this sentence speaks volumes about dendrochronological analysis and conclusion-formation within the Gordion Team). In addition, Kuniholm writes that he has since “found missing rings” that make “the old ‘fit’ look nothing more than a random resemblance.” The result is that “The fit with everything else at Gordion is now good. The end date for CC3 is MTRD 1595 or 909 +4/-7 BC” (italics are mine).7 Thus, the previously reported 400-year gap for the CC# wood has now been officially narrowed; the latest beams were actually cut (“end date”) only a century before the New Chronology DL date, (and two hundred years before 700 B. C.). And responding to my earlier email question about the

Osc

ar W

hite

Mus

care

lla -

Aga

in G

ordi

on’s

Earl

y Ph

rygi

an D

estr

uctio

n D

ate:

ca.

700

+/-

B.C

.

3 SeeJames1991:323-4.SeealsoMuscarella2003:227:hereImistakenlythought that the CC3wood derived from the post-DL,Middle Phrygianperiod.In2003Iwasthinkingthenonlyaboutre-useperse,andthereforemissedthechronologicalimplicationsofthelater-datedtimbers.

4 Indeed, one cannot object tomodifications of chronology based onmoreaccurate readingsof tree rings,but thechanges shouldbeexplained, and,eachtimeitshouldbeindicatedthatthedateisindeedtentative.Inonerecentexample,KuniholmdatedthecuttingyearoflogsatAyanisto655-651B.C.,causingimportantandunexpectedrevisionstoroyaldateyears.Yearslaterheannouncedthatthedatewassomewhereinthe670s:Manningetal2001:2534,andA.Çilinğiroglu,inpress.

5 And although he mentions timber re-use here, none of the other issuesconcerningthewoodsamplesfromTB2A--thattheywereroofbeams,thelackofbark,andmissingrings—arediscussed.

6 IfollowheretheGordionTeam’scommonuseinpublicationsofinvokinga “personal communication” to introduce new, hitherto unpublished,information.

7 Isitacoincidencethatthe909B.C.dateistheverysame(minusoneyear)given forTB2A, above—ordidVoigt orKuniholmconfuseTB2AwithCC3?

revised dating of the earlier samples from CC3, Kuniholm replied (in the same July 12, 2007 communication) that they “must be re-used Bronze Age logs.” Translation: the older samples’ date has not changed—they were centuries-old, re-used timbers from the Bronze Age, unlike the later beams, now interpreted to be fresh-cut later—all neatly provided by a scientific laboratory. The concepts of missing rings and re-use are considered meaningful only when they fit into a preconceived scenario.

I repeat here what I wrote in 2003 (page 227), that re-use of wood at Gordion over long periods of time absolutely precludes any chronological use of terminal rings as reflecting “cutting dates” for wood recovered in Gordion structures. The Gordion Team ignores this fact. One of the significant problems in this matter is that they have unreflectively and unequivocally accepted, without independent analysis and verification, every conclusion issued from the Cornell University laboratory. Given Kuniholm’s (2007) candid (and latest) explanation of his ad hoc methodology used in arriving at his 1988 interpretation, it is appropriate to ask: are scholars now (again) obligated to believe that this revised “scientific” review does not equally operate from more recent assumptions about the DL, and that it too is not another ad hoc attempt to “force a fit,” this time with the dogma of the New Chronology?

In the final analysis archaeologists cannot invoke the currently proclaimed 909 B.C. date (last preserved tree-ring) to establish the time of the destruction of CC 3 and the DL—because of the possibility of wood re-use, and because the latest preserved ring says nothing about the date of constructing a building.8 Indeed, although preached and widely accepted otherwise, it is a false dichotomy to claim that dendrochronology is an objective science to be privileged over “subjective” archaeology. It is therefore significant (and surprising) to find this correct position precisely stated recently by the Cornell Laboratory staff--for the first time: “Dendrochronology is not an exact science….” (Manning 2007: 3).

Until the published (and unpublished) reports cited above are reinvestigated and published in detail, giving all information related to the condition of the wood involved, the presence or absence of bark, the evidence and reasons for both the original and subsequent date changes, and the

�

methodologies employed, including recognizing re-use of wood, no archaeologist can use the TB 2A and CC 3 timbers for any chronological determinations—most certainly not relating to the DL date (see also below). All the Cornell laboratory’s claims remain in abeyance--as does also my tentative 7th century dating (above) based solely on the 1988 report. My reading of Manning et al 2007: 1, 5 is that the Cornell laboratory “slightly revises past practice” and in the future will review and critique (“reanalysis”) the reports its staff has published, and will conduct a re-study of all the published Gordion wood reports.9

Strobel (2004: 266-7, 271, and 275, note 4) with ease dismisses my arguments against the Gordion Team’s use of dendrochronology. As the “Fälldatum [sic] der Hölzer” for TB 2A he proclaims it to be 861/883 B.C., and for CC3 it is 912-919 B. C.; missing rings, wood re-use, or revisions of chronologies over time are to him mere minor, pedantic issues of no archaeological concern. Genz (2004: 221, 226) cites the dendrochronological together with the C-14 data as evidence for the 9th century dating of the DL at Gordion.

In 2004 Kuniholm (page 1) accelerated the use of dendrochronology to affirm the New Chronology, here again invoking the last-ring solecism by citing the rings of four logs (Voigt in Yildirim and Gates 2007: 31 says it is one log) recovered from the base of the Early Phrygian fortification wall as data that reinforces the New Chronology. Because on these logs “the last existing ring (no bark preserved) is 862 BC,” they are therefore taken to indicate that the EP settlement and its fortification wall “appear to be a ninth-century affair” (italics are mine). Not mentioned is the significant different conclusion he forcibly presented in 1988, or the conclusions of Voigt in 1994 and 2005, or his ignoring the problem of using

Osc

ar W

hite

Mus

care

lla -

Aga

in G

ordi

on’s

Earl

y Ph

rygi

an D

estr

uctio

n D

ate:

ca.

700

+/-

B.C

.

8 ConcerningKuniholm’scontinuousmodificationsofandchangestodataandconclusions over the years regarding C-14/dendrochronological dating—oftenwithoutreferencetopreviousclaims,seeMielke2006:82-3andnote17, 90,with references tomissing rings;Keenan 2006: 5. James (2002)recordstheseriesofdatespreviouslypublishedbyKuniholmforthetumulusMMburial:547,757,718B.C.;healsocritiquesthepresentdateof740B. C. Forthepracticeofwoodre-useinAnatoliafromearliesttimesseeM.andA.Özdoğan1998:589.

9 I came upon Manning 2007 after I had essentially finished writing thispaper, courtesyofD.Keenan.Aside from informingus (subtly—Ihad toread it several times tounderstand it)of the review tobeundertaken, thereportprovidesacleardiscussionaboutdendromethodologyandthemanyproblemsinvolvedinanalyses;seealsoMielke2006:77-84.

Prof. Dr. Aykut Çınaroğlu’na Armağan

bark-less wood for absolute dating (see Manning et al 2007: 5), and also the possibility of re-use of earlier wood. Voigt (in Yildirim and Gates 2007: 31), ignoring her earlier cautious stance, also cites the wall’s log as evidence to support the New Chronology—another example of the Gordion Team automatically accepting a “scientific” decree.10 All I discuss here has now been manifestly confirmed, for recent dendro examination at Gordion has proved (yet again!) that timbers cannot be used as chronological markers. In the most recent Cornell University Weiner Laboratory report (December 2007: 3) is the statement by Sturt Manning that: “The date for the last preserved ring” of juniper logs from Building A, from the MP (i.e. post DL!) citadel, is 991 +4/-7 B.C. But unfortunately his report refrained from confronting the fact that the MP period is dated by the Gordion Team post DL, to the late 9th early 8th century B.C. (but to me late 8th at the earliest, or the early 7th century B.C.), and that either way we have here evidence of reused beams; and that a beam with its last ring dating almost a century earlier than those from the EP period buildings TB 2A and CC3 (above), is not employed here using the same methodology to date the MP period to the 10th century B.C.!

Dendrochronological dating of tumuli has suffered the same blinkered mis-treatment: in several instances we are commanded to accept as an absolute chronological determinant the last preserved ring of a tomb’s timber construction. Thus Strobel (2004: 271, 276 and note 10) reveals to archaeologists that the plundered and destroyed Kayran Mevkii tumulus (a Gordion tumulus) was constructed in the 9th century B.C. (862 +7/-3)—because of a (not fully published) dendrochronological dating of burnt logs--with no bark; and no tomb artifacts were preserved. In the same sentence (p. 276) we are also commanded to know that there exists other Phrygian tumuli constructed in the 9th century B. C., tumulus W (see below), and also the, to me, 8th century B. C. Mamaderesi tumulus (another Gordion tumulus), which date is known to him “auf die Funde;” in a footnote (10) he

�

mentions the 80 fibulae recovered. The chronology of the tumuli as proclaimed by Strobel, is of major importance to his general Early Chronology project, and he simply announces that the 80 fibulae were made in the 9th century B.C. (pace Muscarella 2003: 231). This chronology can be based only on an a priori determination about the DL (and consequent tumulus construction) chronology: i.e., a circular argument. The Mamaderesi fibulae include Types XII, 2, 9, 13, 13A, 14, and 14A, only one of which type occurs in tumulus W (9th century to Strobel), and there are no 7A examples here, although they are quite common in W. A number of the Mamaderesi fibula types were recovered in manifestly 8th century B. C. tumuli, such as (even to Strobel) MM, which contained XII, 7, 9, 11, 13, and 14 examples. The late 8th century or slightly later tumulus S-1 ((viz. DeVries 2005: 39-40, notes 4-6, 43; Kohler 1995: 115-140) also has two XII, 14A buckles (ibid. 127-8, pl. 66), a late form (compare Strobel 2004: 276 and note 10). An appropriate question raises itself based on artifact analysis: why should archaeologists not assert that tumulus MM must also be dated to the 9th century B. C., inasmuch as it shares with the DL and Mamaderesi tumulus XII, 9, 13, and 14 fibulae forms? (see also the discussion of the clay foundation deposit below).

For the record: that tumulus MM did not contain the burial of king Midas—because of its construction date determined by dendro (more or less) and artifact analysis--and that it belonged to an earlier ruler, probably Midas’ father, has been argued for a long time: by R. S. Young after its excavation, and subsequently by a good number of other scholars, including me, for decades (see Muscarella 1982: 9, idem 1995: 99, note 13, idem 2003: 231, note 28). This has been ignored by the Gordion Team in lectures and writing: viz. Kohler 1995: 192, 228; Manning et al 2002: 2534; DeVries 2005: 42; DeVries et al 2003: 2; DeVries et al 2005: 45. Voigt (in Yildirim and Gates 2007: 311) writes as if we now know for the first time ever that Midas is not buried in MMT. And Strobel (2004: 276) indicts me for continuing to maintain there is a relationship between MM and King Midas and his family. This specious charge requires no response, but I will state simply what has often been recognized as obvious: I believe along with other archaeologists, that MM is manifestly a royal burial based on its labor intensive size and construction, its locus, and its artifact deposition

Osc

ar W

hite

Mus

care

lla -

Aga

in G

ordi

on’s

Earl

y Ph

rygi

an D

estr

uctio

n D

ate:

ca.

700

+/-

B.C

.

10 Itisrelevantheretoremindscholarsofthe(ifcorrect!)spectaculardiscoverybyKuniholmofacentraljoistwithinamodernhouseontheBlackSeathatwascut6200yearsago(Aegean Dendrochronology Project December 2001 Progress Report: 8).Had this house beenburneddown50years ago andrecentlyinvestigated,nodoubtsomedendrochronologistswouldhavedatedthehousetotheearlyNeolithicperiod.

and chronology coincide with/are quite close to that of the historical King Midas/Mita. Its chronology strongly suggests however, that Midas himself is not buried there.

The Gordion Team (DeVries et al 2005: 46) asserts its position loud and clear, “The radiocarbon and dendrochronological evidence provides a firm and consistent absolute chronology for ninth [Early Phrygian] and eighth century [Middle Phrygian] Gordion” (italics are mine). (What is consistent s their circular-citation of each other’s publications--but very few of other, relevant publications). For analytic criticisms and rejections of the “scientific” consistent claims blithely made by DeVries et al 2005, see Keenan 2004, idem 2006, James 1991: 322, idem 2002. Artifacts are not mentioned as evidence in the above quoted claim, but are mentioned in the article itself (p. 45), with the claim that they “independently indicate a late 9th century” New Chronology date; see also DeVries et al 2003: 2), Voigt 2005: 31, and idem, in Yildirim and Gates 2007: 311. These artifacts include fibulae, horse trappings, vessel attachments, ivory horse pieces, attachments, pottery, and orthostates, all of which are chronologically evaluated in Muscarella 2003: 237-245, although not confronted by the Gordion Team. On an (attempted) objective basis, i.e. not a priori using the old date of ca. 700 B.C. for the DL as a guide, I argued that the DL artifacts are 8th, not 9th century productions. I have not changed my mind, and those who disagree should confront each of my evaluations. Strobel (2004: 278), consistent in asserting, but not demonstrating a 9th century date of each DL artifact, does not.

A viable argument in the present framework would be to state up front that consequent to any scholar’s acceptance of the late 9th century C 14 chronology for the DL, it therefore must follow that its artifacts are dated to that time. This has been accomplished formally by Prayon and Wittke (2004: 123), who date one of the ivories from the DL clearly “aufgrund ihrer Fundsituation nach der neueren Chronologie wohl noch in das 9. Jh. zu datieren…;” also Prayon 2004: 614; Genz (2004: 225-6) says we must now date certain types of central Anatolian pottery and fibulae to the 9th century B. C. because of the revised Gordion chronology. If one accepts the C-14 New Chronology, these will be seen as “correct” and necessary assessments: although the evidence actually indicates otherwise.

�

I focus now on a summary of the fibulae recovered in the DL, and add some scholarly commentary. I also add an amplification concerning the loci at Gordion of fibulae and another artifact, socketed arrowheads. The latter has not hitherto been discussed, but is a major component of evidence that must be used to determine the date of the DL at Gordion.

FIBULAE

The Phrygian fibulae, commonly known as Type XII, recovered in the DL are dated both by comparison with those from the tumulus burials and from outside comparanda, and provide a general chronology covering the second half of the 8th century for the DL examples.11

As I reported in 2003 (229-234), the DL fibulae types include two Type XII, 5; 22 XII, 7A; one XII, 9; one XII, 13 or 14;and one XII, 14, total of 27 Phrygian fibulae. The DL also contained 14 imported foreign Aegean and Near Eastern fibulae, dating to the 8th –7th century B. C. and therefore relevant to DL chronology investigations (Muscarella 2004: 233-36), pace Voigt (2005: 31), who states that “no Greek material had ever been found in the….Early Phrygian Destruction Level. ”The Aegean fibulae parallels are not “sehr vage” as Strobel asserts (2004: 276; Maya Vassileva will eventually publish them in detail). Aegean fibulae, along with socketed arrowheads, were also recovered in the fill over tumulus B (Kohler 1995: 21-2, pl. 11). The s arrowheads demonstrate a late 8th or 7th, century date for the construction of the tumulus, as well as a nomadic people’s presence at Gordon (at some undetermined specific time; see below).

The Phrygian fibulae from several of the tumuli parallel the DL fibulae. The most common type from the burials, Type XII, 7A, occurs in tumulus W, KIII, KIV, S and G; XII, 5 in KIII; XII, 9 in KIII, KIV, MM, S-1, and Mamaderesi; and Type XII, 14 in KIV, MM, S-1, and Mamaderesi. In the Ankara tumulus excavated by Sevim Buluç, Type XII, 7A, 9 and 14 forms also occur together (Muscarella 2003: 233). All these tumuli are dated to the second half of the 8th century. For a review of

Osc

ar W

hite

Mus

care

lla -

Aga

in G

ordi

on’s

Earl

y Ph

rygi

an D

estr

uctio

n D

ate:

ca.

700

+/-

B.C

.

11 I use standard terminology here and not Caner’s 1983 unnecessary newterminology, intended to replace Blinkenberg,Muscarella, Boehmer, etc.,andpaceStrobel2004:276,note5.

Prof. Dr. Aykut Çınaroğlu’na Armağan

earlier scholars’ determinations of the relative construction dates of the tumuli and the conclusions that MM, P, KIII, and KIV are relatively chronologically close to each other, W is earlier, and all date to the 8th century B.C. DeVries (2005: 43) dates tumulus MM before Mamaderesi. Strobel (2004: 267, 276-7) dates tumulus KIII and P close to the time of the DL, and for tumulus W makes another throwaway assertion of a date: “bestimmte Fundstücke belegen ein Datum im späten 9. Jh. V. Chr.”, but which Fundstücke are not shared with us. Wittke (2004: 257, note 361; 258, 268) dates tumulus W to 850 BC, and KIII to the first half of the 9th century B.C. Crielaard (2007: 223) dates W to ca. 850 B. C. “on the basis of recent radiocarbon dating,” which in fact does not exist for this tomb (although he may have meant the recent DL carbon dating). All these chronologies follow manifestly from the authors’ acceptance of the New Chronology—for not one provides detailed artifact analyses. Tumulus W is indeed the earliest, in my opinion dated ca. 750 +/- B.C. (Muscarella 1982: 8, idem 1995: 97, and idem 2003: 229; also Kohler 1995:191-2).

Post 2001 the Gordion Team also supports a 9th century B. C. date for tumulus W (verbal communications) primarily because of the presence therein of XII, 7A fibulae, which to them are chronologically paralleled by their presence in the DL—and therefore 9th century B. C. If this is the case, then KIII, KIV, and G are also to be dated to the 9th century. But what of the other later DL fibula forms, Types XII, 7, 9, and 14, all of which occur in other tumuli, especially MM, which tumulus all the involved scholars date more or less to ca. 740 B.C.?

For a discussion of the chronological relationship of the Gordion tumuli based on artifact analyses, the comparisons among them, and conclusions that they were not constructed far apart in time one another within the 2nd half of the 8th century B.C., see Muscarella 1982: 8, idem 1989: 337-342, and idem 2003: 229-30. Whether these conclusions are correct or not, the analyses were based on an analysis of the finds and not on a preconceived date for the DL. Strobel (2004: 276-7) disagrees with this chronology, but without meaningful discussion of the parallels brought forth. In the final analysis, scholars will have to decide which chronological interpretation in fact “drehen sich immer wieder im Kreise.” The vital point that I repeat here

�

is that all the Phrygian fibulae forms recovered in the DL are also present in the tumuli: all of which are accepted by all parties to have been constructed in the 8th century B. C. Moreover, no Phrygian fibulae recovered in the West (none are Type XII, 7A) have been dated by pre- 8th century B. C. (Muscarella 1989: 338-9, and note 21).

Numerous other artifacts recovered in the DL have yet to be published, and thus remain to be analyzed for contributions to the chronological issues (Maya Vassileva is presently engaged in publishing the metal remains).

CLAY FOUNDATION DEPOSIT

Another locus at Gordion for the fibulae remains to be evaluated, one that reinforces the Old Chronology. Only recently did I become aware that there are more fibulae available for chronological discussions than I had hitherto realized, that for some inexplicable reason I had missed. I argue that these fibulae must be brought into the discussion about the date of the DL, and that they reinforce my original artifact analyses and dating.

Mary Voigt’s excavations (Voigt and Henrickson 2000: 52) produced a significant fundamental adjustment in our knowledge and understanding of the nature of the cultural and chronological sequence at Gordion following the destruction. She determined that the citadel was resettled above the DL, not 150 years later (as posited by Rodney Young), but quite soon after, “with little or no gap” (Muscarella 2003: 227-8). In DeVries et al 2003: 2 (for which Voigt was a co-author) the rebuilding time was modified to “a generation after the destruction.” Later, however, Voigt (2005: 32) reinforced her original conclusion (ignoring her 2003 position), claiming that the reconstruction “began immediately” after the EP destruction.

If this stratigraphical/chronological adjustment holds up it seems the labor-intensive and well-planned rebuilding process began soon after the destruction of the EP citadel, although whether within a year, or a few years, or a decade (more or less) remains unknown. This well-organized rebuilding was accomplished first by leveling the DL debris and some walls, then covering most of the DL with a massive clay deposit two meters thick, within which were set the stone

Osc

ar W

hite

Mus

care

lla -

Aga

in G

ordi

on’s

Earl

y Ph

rygi

an D

estr

uctio

n D

ate:

ca.

700

+/-

B.C

.

rubble foundations for the structures forming the settlement that was built directly on top of the clay (see Voigt 1994: 273, pls. 25.5, 6.2; 2005: 33, fig. 3-7A, B). Closely copying the plan of the EP citadel—which is another indication that rebuilding occurred soon after the destruction (Voigt 2005: figs. 3-3, 3-4)--the post-clay settlement is designated Middle Phrygian.

It is worth considering here, at least parenthetically, whether this massive clay deposition was solely an engineering/architectural necessity, to cover the DL with a thick foundation in order to rebuild securely the settlement above. But one may legitimately wonder if the clay deposit may also have been a ritual burial of the destroyed city of King Midas (who committed suicide when the city was destroyed!), or a burial in honor of the Phrygian/Gordion Goddess Matar, a burial directly linked to the rebuilding--on the same plan as the (sacred?) buried city below. There is a tradition of the ritual burial of sacred structures in Anatolia going back to the pre-pottery Neolithic period, as neatly documented by Özdoğan and Özdoğan (1998: 589-592; see also Schmidt 2001: 46). As the Özdoğans noted, the leveling of buried structures and preservation of what remains of walls are followed by a rebuilding “immediately over the earlier structure” (for a deliberate covering deposit at Nuzi see Bjorkman 1999: 106). That is, events precisely as happened at Gordion. Bjorkman (1999) has cited later, historical examples of ritual burials of sacred structures at a number of Mesopotamian sites. A still later example of such burial is that of Nush-i Jan, a 7th century temple site in Iran (Muscarella 1988: 207, 208, note 2). I suggest that the possibility of ritual burial be considered as a viable interpretation for Gordion’s transition from EP to MP—even if it cannot be demonstrated as historical reality at present.

A number of artifacts were recovered from the clay foundation a number of artifacts were recovered; the clay was clean but not quite sterile (as claimed by Strobel 2004: 266-7). Some clay-derived artifacts clearly date from a much earlier period (Dusinberre 2005: 10); I excavated there a beautifully hand-cut, apparently Neolithic, obsidian blade (not published). Other finds include Phrygian terracotta and metal artifacts. Those of us digging at the site believed that these artifacts had been lost or discarded, then accidentally picked up in the buckets by Middle Phrygian workmen from

�

the clay/mud sources (that there were at least two sources is revealed by deposits of different colors, orange and gray). The Phrygians involved in the bucket brigade may also have dropped some contemporary objects (fibulae?) accidentally. Artifacts from the clay could also have derived from the DL debris, displaced in the leveling process (on this see M. and A. Özdoğan 1998: 589). A significant example was a broken stone bearing an inscription recovered from one of the MP structure foundations built into the clay (Muscarella 2003: 246). This stone most probably derived from the DL itself, picked up along with many other available EP wall stones that were obviously in full view and readily available to the re-builders, or it may have come from an off-site source. In either event, the stone’s placement in the foundations set in the clay occurred (according to all scholars concerned) soon after the destruction. And, as it was set into the foundations, it predates the construction of the MP structures.

Among the fibulae recovered in the clay deposit were one Type XII, 7, two XII, 9, three XII, 14 forms (Figure 1, for one example: B 1304; hitherto unpublished). It is possible that some of these are earlier than the DL, picked up in the clay sources (see Caner 1982: 2, note 15 for this source recognized in tumuli formation fill), but it is worth noting that all have parallels with those deposited in the late 8th century B. C. tumuli, MM, S-1, and Mamaderesi being the outstanding examples. Types XII, 9 and 14 also occur in the DL. And Types XII, 7 and 9 are also paralleled on reliefs at Khorsabad and Ivriz, both of which document a late 8th century B. C. floruit for these forms—the XII, 9 at Ivriz has a double pin mechanism and lock-plate, exactly as found on examples from tumulus MM (Muscarella 1982: 7-8; Caner 1983: 9, 173, 201, Taf, 62). Strobel (2004: 277, note 15) casually (and unsuccessfully) dismisses the chronological value of the Ivriz relief. Found in the MP South Cellar were fibulae of Types XII, 13, 9, and 14, the latter two surely relatively late forms, and paralleling those from tumulus S-1, later than those from tumulus MM.12

Osc

ar W

hite

Mus

care

lla -

Aga

in G

ordi

on’s

Earl

y Ph

rygi

an D

estr

uctio

n D

ate:

ca.

700

+/-

B.C

.

12 DeVries 2005 includes a modern reconstruction of the South Cellar’sstratigraphy(thesection,Figure4-2,wasmadeinPhiladelphia,notGordion),and it presents confusing dating and loci of the Greek pottery excavatedthere.ItisdisturbingthatDeVriesmadenomentioninthisarticleofhisandK.Sams’claimsmadeyearsearlieraboutthelociandassociatedproblemsrelatingtotheGreeksherdsunderdiscussion(asifforthefirsttime).Forinfact, inearlierpublications,DeVriesclaimedthatthepotteryderivedfrom

Prof. Dr. Aykut Çınaroğlu’na Armağan

ARROWHEADS

Another group of artifacts recovered in the clay include bronze socketed arrowheads (hitherto unpublished). When I recorded the corpus of socketed arrow heads excavated from various geographical areas in the Near East (Derin and Muscarella 2001: 195, and Muscarella 2006: 156) I reported that no such arrows “could be attributed to the destruction level” at Gordion. Inexplicably, I was not then thinking of reviewing the evidence from the clay --an extension of the DL; that thought occurred to me later, hence the present review.

One example (B1196; Figure 2, left) was recovered by Rodney Young in “the clay and rubble wall beds of Bldg M…. dug right down to the top of the burned fill overlying Megaron 3….From the clay: 5769 B1175. Bronze wing with fragments of gold foil [;] 5903 B1196: bronze arrow point.” (R. S. Young Field Book, Vol. 78: 14). B1196 is a socketed trilobate arrowhead. Another (7876 B1503; Figure 2, right) was found below a white floor covering the clay: “.…work continued digging clay. Beneath the white floor…was found a well-preserved bronze arrowhead, c 4 cm long. 3 flanged” i. e. a trilobate. (R. S. Young Field Book, Vol. 114: 50-51).

There is a third example to be considered (89 SF# 16 Operation 11: a bilobate). This arrowhead was discovered in 1989 by M. Voigt, who generously gave me the following information (verbal 1992, and in recent email communication): “The arrowhead from OP 11 came from a drain that is part of the construction of YHSS 6B or (early) Phrygian…. The stub [on which the arrow was recovered] is under a drain that was built as part of YHSS 6A, the construction phase that was mostly standing when the fire [DL] took place…. The problem is that RSY dug down to the level of the PAP/6B and then below it to the wall stub in 1963 and both lay exposed till we cleared it in 1989….a lot of people went right across that area for many years of excavation so you can’t exclude the possibility that the arrowhead was dropped there [post

�

1963]. We found an Attic black figured sherd on the exposed surface [which?] as well. Since the arrowhead was tucked between stones, it could be in situ but if it is a critical part of any argument, the possibility of it being dropped cannot be excluded.” Indeed, I accept this conclusion as objectively viable. I also add that the arrow could have been dumped—along with the fibulae and the two arrowheads mentioned above-- with the clay deposit. This arrowhead is best left in abeyance—neither included nor excluded in the discussion. I record it as evidence of a possible third example of a nomadic arrowhead from the DL.

The interesting and most important fact about the arrowheads found in the clay is this: not a single excavated bronze socketed arrowhead is known from excavations anywhere in the Near East that pre-dates the 7th century B.C. Some surely may have been present since the very late 8th century B.C., but up to the present none this early has been identified. This chronological situation has been fully documented by Derin and Muscarella 2001: 197, passim, and Muscarella 2006: 157-8. Thus, the arrowheads in the clay deposit manifestly document a post-late 8th century B. C. date for the DL. Indeed, socketed arrowheads traditionally have been culturally associated with intrusive nomadic peoples (first recorded in the late 8th century B.C.—the earliest date is 715 B.C., for the Cimmerians) but they were soon thereafter employed by other cultures (Derin and Muscarella 2001:197-203; Muscarella 2006: 157-8). However, their presence at Gordion in a post- late 8th century B. C. clay deposit context surely suggests a Cimmerian presence, as I and others have argued for some time, pace the Gordion Team’s rejection (Muscarella 2003: 247-249; Muscarella 2005/2006: 395. For possible Cimmerian presence at Norşuntepe and Boğazköy, both because of artifacts recovered including socketed arrowheads, And in northern geographical locations, see Derin and Muscarella 2001: 199-200; for a probable nomadic burial at Gordion in tumulus KY, where two horses were buried, but no arrows (or Phrygian fibulae!) were recovered, see page 195).

Fibulae do not lie (fib); to this (growing) list of truth speakers we now add bronze socketed arrowheads. Both occur within the clay deposit/platform overlying the DL, constructed soon after the destruction and almost simultaneously with the erection of the MP settlement. Both

Osc

ar W

hite

Mus

care

lla -

Aga

in G

ordi

on’s

Earl

y Ph

rygi

an D

estr

uctio

n D

ate:

ca.

700

+/-

B.C

.pre-destructioncontexts“althoughnoneisfromastratigraphicallyinstructivecontext,”andSamsclaimedthatthefragmentshadbeenrecovered“scatteredand dislocated” within the city and elsewhere! I discussed these issuesinMuscarella2003:241-3,withreferences.ItisnotcitedbyDeVries.ForchronologicalproblemsatBoğazköyregardingimportedGreekpottery,seeMuscarella2005/2006:394.IsuggestthatwehavenotheardthelastabouttheSouthCellarstratigraphyorofthelocioftheGreeksherds.

the fibulae and the arrowheads date the DL to ca. 700 +/- B.C., a time just preceding the beginning of the MP period. The New Chronology decreed for the DL must therefore be rejected because: the C-14 data have been accepted ad hoc without confirmation from a second laboratory’s analysis and discussion of statistics, (see also above); the offered dendrochronological conclusions and their interpretations

�0

are flawed; and archaeological analysis of the artifacts preserved in the DL and the clay do not support the New Chronology, they contravene it. Fibulae may indeed be cited as an axiomatisches Argument, they most certainly are one of archaeology’s best chronological indicators, a classic Leitfossil, as are socketed arrowheads. Such realia cannot casually be ignored or manipulated.

Osc

ar W

hite

Mus

care

lla -

Aga

in G

ordi

on’s

Earl

y Ph

rygi

an D

estr

uctio

n D

ate:

ca.

700

+/-

B.C

.

Prof. Dr. Aykut Çınaroğlu’na Armağan

��

Osc

ar W

hite

Mus

care

lla -

Aga

in G

ordi

on’s

Earl

y Ph

rygi

an D

estr

uctio

n D

ate:

ca.

700

+/-

B.C

.

BIBLIOGRAPHY:

J. Bjorkman 1999: “How to Bury a Temple….,” in Studies in the Civilizations of Nuzi and the

Hurrians, vol. 10, ed. D. I. Owen and G. Wilhelm: 103-122.

E. Caner 1983: Fibeln in Anatolia, München.

J. P. Crielaard 2007: review of Wittke 2004, in Bibliotheca Orientalis LXIV, 1/2: 220-226.

Z. Derin and O. White Muscarella 2001: “Iron and Bronze Arrows,” in A. Çilingiroğlu and M.

Salvini, Ayanis 1, Istituto per gli Studi Micenei ed Egeo-Anataolici, Rome

K. DeVries 2005: “Greek Pottery and Gordion Chronology,” in Lisa Kealhofer ed., The

Archaeology of Midas and the Phrygians Recent Work at Gordion, Philadelphia: 37-55.

K. DeVries, P. Kuniholm, G. K. Sams, M. M. Voigt June 2003: “New dates for Iron Age

Gordion,” Antiquity 77 (Internet: http://antiquity.acuk/ProjGall/devries.html.

K. DeVries, G. K. Sams, M.M. Voigt 2005: “Gordion re-dating,” in A. Çilingiroğlu and G.

Darbeyshire, eds., Anatolian Iron Ages 5: 45-6.

E. Dusisnberre 2005: Gordion Seals and Sealings, University of Pennsylvania Museum.

H. Genz 2004: “Erste Ansätze zu einer Chronologie der frühen Eisenzeit in Zentralanatolien,” in

Die Aussenwirkung des späthethitischen Kulturraumes, M. Novák et al, eds. Ugarit-Verlag: 219-228.

P. James 1991: Centuries of Darkness, London.

-----------2002, “The Dendrochronological Debate,” Minerva XIII, 4: 18.

D. J. Keenan 2004: Radiocarbon Dates from Iron Age

��

Gordion are Confounded,” Ancient West and East, 31 1: 100-103.

---------------: “Anatolian tree-ring studies are untrustworthy,” Informath (Internet): 1-20

U. Kelp 2004: “Der Einfluss des späthethitischen Kulturraumes auf Orthostaten in Gordion,” in Die Aussenwirkung des späthethitischen Kulturraumes, M. Novák et al, eds. Ugarit-Verlag: 284-295.

E. Kohler 1995: The Lesser Phrygian Tumuli, Part I The Inhumations, University Museum, The University of Pennsylvania.

P. Kuniholm 1988: “Dendrochronology and radiocarbon Dates for Gordion and Other Phrygian

Sites,“ SOURCE, Notes in the History of Art VII, 3/4: 5-8.

--------------- 2004: “Aegean Dendrochronology Project December 2004 Progress Report,” Cornell University.

S. W. Manning, B. Kromer, P. Kuniholm, M. W. Newton 2001: “Anatolian Tree Rings and a New Chronology for the East Mediterranean Bronze-Ages,” Science 294: 2532-2535.

S. W. Manning, C. Griggs, Maryanne Newton, Tomasz Wazny 2007: “Summary and Discussion of Procedures at the Malcolm and Carolyn Weiner Laboratory for Aegean and Near Eastern Dendrochronology, Cornell Tree-Ring Laboratory,” Internet: www.arts.cornell.edu/dendro/summary%20of%20procedures.pdf.

Mielke, Dirk Paul 2006: “Dendrochronologie und hethitische Archäologie—einige kritische Anmerkungen, “ BYZANS 4: 77-94.

O. White Muscarella 1982: “‘King Midas’ Tumulus at Gordion,” The Quarterly Review of Archaeology, December: 7-10.

------------------------ 1989: “King Midas of Phrygia and the Greeks,” in Anatolia and the Ancient Near East, Studies in Honor of Tashin Öçgüç, K. Emre et al eds., Ankara: 333-344.

Osc

ar W

hite

Mus

care

lla -

Aga

in G

ordi

on’s

Earl

y Ph

rygi

an D

estr

uctio

n D

ate:

ca.

700

+/-

B.C

.

Prof. Dr. Aykut Çınaroğlu’na Armağan

------------------------ 1995: “The Iron Age Background to the Formation of the Phrygian State,” Bulletin of the American Schools of Oriental Research 229/300: 91-101.

------------------------ 2003: “The Date of the Destruction of the early Phrygian Period at Gordion,” Ancient West and East 2/2: 225-252.

------------------------ 2005/6: review of H. Genz, Boğazköy-Hattuşa XXI: Büyüykaya I. Die Keramik der Eisenzeit…., in Archiv für Orientforschung 51: 393-396.

------------------------ 2006: “Bronze Socketed Arrowheads and Ethnic Attribution,” in The Golden deer of Eurasia, J. Aruz, et al, eds. Metropolitan Museum of Art: 154-159.

M. and A. Özdoğan 1998: “Buildings and the Cult of Buildings,” in Light on Top of the Black Hill, Studies presented to Halet Çambel, Istanbul: 581-593.

R. Porter 2005, “Carbon Dating, The Situation in 2005, ” Journal of the Ancient Chronology Forum: 63- 66.

F. Prayon 2004: “Zum Problem von Kultsräten und Kultbildern der Anatolische Muttergöttin in 8.Jh. V. Chr.,” in Anadolu’da Doğu, H. Işkan and G. Işın eds., Istanbul 2004: 611-618.

F. Prayon and A. M. Wittke 2004: “Die Aussenwirkung des späthethitischen Kulturraumes auf Zentral- und Westanatolien und Zpperen,” in Die Aussenwirkung des späthethitischen

��

Kulturraumes, M. Novák et al, eds. Ugarit-Verlag: 121-126.

K. Schmidt 2000: “Göbelki Tepe, Southeastern Turkey A Preliminary Report on the 1995-1999

Excavations,” Paléorient 26: 45-54.

K. Strobel 2004: “Neue Fragen zur Chronologie Gordions und Anatoliens im 1. Jahrtausend v,

Chr.,” in Die Aussenwirkung des späthethitischen Kulturraumes, M. Novák et al, eds. Ugarit-Verlag, 2004: 259-84.

G. Summers 2006: review of Dusinberre 2005 in Bryn Mawr Classical Review, 03.27: 1-5 (Internet).

M. M. Voigt 1994: “Excavations at Gordion 1988-89,” in Anatolian Iron Ages 3, A. Çilingiroğlu and D. H. French eds., British Institute of Archaeology, Ankara: 265-293.

------------- 2005: “Old Problems and New Solutions, Recent excavations at Gordion” in Lisa Kealhofer ed., The Archaeology of Midas and the Phrygians Recent Work at Gordion, Philadelphia: 22-35.

M. Voigt and R. Henrickson 2000: “Formation of the Phrygian State: the Early Iron Age at Gordion,” Anatolian Studies 50: 37-54.

Anne-Maria Wittke 2004: Mušker und Phryger, Wiesbaden.

B. Yildirim and M. H. Gates 2007: “Archaeology in Turkey: 2004-2005,” American Journal ofArchaeology 111: +++

Osc

ar W

hite

Mus

care

lla -

Aga

in G

ordi

on’s

Earl

y Ph

rygi

an D

estr

uctio

n D

ate:

ca.

700

+/-

B.C

.