Errors for Unit J5

James L. Walker – March 2014

Introduction
Form
Record
Judgment

Introduction

The general approach to errors is described in the Identification of errors and accidents section of the Mozan sitewide book. The approach is to preserve each recorded data element that is based on observation or measurement These data are occasionally incomplete, in conflict, or in error. We specifically address these factors when they significantly impact a constituent. However, for the most part errors are small and have little cumulative effect. We discuss the recurring problems below for clarity.

Form

In this excavation unit all members personally documented their observations and measurements. But for the supervisor, almost all were non-native English speakers. In particular the supervisor required the principal assistant to write the daily journal after a brief training period. The content was reviewed daily but left substantially uncorrected. As the season progressed, there was considerable improvement in gramatical quality. Each team member wrote short descriptions of the items that were excavated but time did not permit a gramatical review.

Record

1.Excavation

Time pressures and inexperience resulted in two major sets of errors during excavation:

a. The first was in volumetric location by relays. This is accomplished by recording accurate observations relative to survey points which are later resolved by a computer program in the laboratory. A mistake in any of these processes results in an erroneous location which typically cannot be re-established because the target has often been removed.

b. The second is the failure to establish the volumetric coordinates for a q-lot. This is done by relaying one corner and measuring the length, width and depth relative to it. Occasionally one or more of these coordinates are missed during the rapid pace of excavation.

2.Ceramics Processing

There were two sets of ceramics processing errors:

a. For a few shape sherds the recorded diameters noted on the drawing and in the description vary ( some by substantial amounts). Most, if not all of these errors arise because the measurements are made separately by the ceramics analyst and the draftsman. The sherds often represent a small percentage of the rim or base of a complete vessel. Also it is not always clear what the orientation of the rim should be while making the diameter measurement (the difference between a bowl rim and a jar rim is one example).

b. Rain from a fast-moving thunderstorm washed the identification marks off a number of sherds in the q400 series that were needed to date the first accumulations over the stone escarpment and the soil escarpment above the stone one. The data were recovered by assigning new numbers in the q900 series (usually reserved for sherds recovered from multiple features by sifting the output of the shafat) to each of the damaged sherds. None had been drawn prior to this event so the impact was minimal.

3. Objects Processing

Through the years procedures were refined to incorporate the advantages of digital photography into the record system. It became possible to photograph each item. In early seasons only important items were photographed, so there was a considerable backlog which began to be cleared in 2010. Unfortunately this was interrupted by the war which continues as of this writing. In this last season storage was also being re-organized so there are some items whose location is problematical.

4. Staff Turnover

The majority of the excavation staff are young undergraduate and graduate students from universites in the US, Europe, and Syria. We endeavor to provide a comprehensive exposure to all aspects of field archaeology during the excavation season. They have a lot to learn in a short period of time and often only return for a season or two before continuing their education elsewhere. As a result, work is slowed, errors occur and are not recognized or corrected despite review.

Judgment

1. Continuation of 2006 J1 excavations

There was a probe in 2006 by J1 excavators to determine the westward extent of the revetment wall, J1f72. In the east part of the unit they exposed the full south face of EDIII wall (now f189) and at the western extent they exposed the southwest face of a Mittani overlay (now f3). J5 incorporated this portion of J1 and added unexcavated loci to the north and west to determine if there was a formal access route from the Akkadian palace to the temple mound. Excavation in J5 began before the J1 excavation was analyzed (2009) so there was an incomplete understanding about the stratigraphy of the revetment wall and its escarpments.

2. Middle Assyrian Horizon

Excavation north and west of the previously exposed revetment wall began at the surface in loci k23 and k24. It was begun with the large pick and continued until the monumental Mittani staircase, f21, was reached. At a later point in time the adjacent locus k13 was more slowly excavated and small quantities of Middle Assyrian ceramics were recovered. If we had anticipated that this previously unknown occupation phase existed, we may have proceeded more slowly in k23 and k24 to determine its full extent.

3. Missing Section

At the close of the 2008 excavation season, a line of mud bricks, f196, running N-S and covered by Mittani deposits was exposed at the northwestern corner of the unit. In the 2009 season, we exposed the full extent of the western face of the revetment wall, f41, possibly from the Ninevite V period. These two features were separated by a distance of several meters and formed a channel for running water at various times. At the revetment wall end was attached a crude dam and settling basin. The dam and settling basin were removed to look for a northern extension of a stone escarpment to the revetment wall. We anticipated that excavations in a future season would expose the east face of the bricks and clarify whether they formed a wall which would have anchored the west end of the dam and formed the eastern edge of a channel. Warfare interrupted excavations. We could have clarified several of these issues had we preserved a section running E-W in the vicinity of the dam. Unfortunately this was not done.

3. Collapse of Portion of Late Mittani Wall

As excavations progressed in 2009, we found that the mock revetment wall f3 had been built atop and followed an earlier Ninevite V wall f41 (see illustration). At its southern point it diverged to the south, overhanging the earlier wall. In lieu of removing the overhanging portion we preserved it by leaving it supported by a soil pillar f0294 (see illustration), intending to remove it in a following season. Unfortunately hostilities intervenened and during heavy rains in the late fall of 2018, the soil in the pillar saturated and the section of the f3 wall collapsed. Furthermore, the water possibly eroded part of the EDIII wall behind it (see photograph). The resident Mozan workforce under the direction of the on-site archaeologist, Amer Ahmed, stabilized the damaged wall section. In the spring and summer of 2019 they designed and built a more permanent solution which is documented in Amer’s written report.