After washing and drying, sherds from all excavations were first examined and sorted in the field into categories of plain body sherds and of “diagnostic” sherds, the latter including rims, carinations, bases, sherds with any form of decoration or any other unusual characteristic. Due to the prohibitive expense of shipping large quantities of sherds to the laboratory, only selected lots of plain body sherds were retained (representative lots were deposited in the Papua New Guinea National Museum). The remainder of the plain body sherds were discarded in Mussau. All diagnostic sherds were sent to the laboratory in Seattle (and later moved to Berkeley) for further study.
Kirch’s initial approach to the analysis of the Mussau ceramics was that of a sherd-based attribute analysis, in which systematic observations of individual sherds were entered into a database, with each sherd providing a single record. The attributes selected for recording were based on Kirch’s prior experience with ceramic assemblages from Niuatoputapu (Kirch 1988a) and Tikopia (Kirch and Yen 1982), also following methods advocated by Shepard (1963), Rye (1981), and Bennett (1974). In addition to basic provenience data (site, excavation unit, level), the following attributes with their possible attribute states were recorded:
Sherd Type: 1, Body sherd; 2, Rim/neck, with lip intact; 3, Rim/neck, lip missing; 4, Carination; 5, Base; 6, Rim/neck + carination; 7, Rim/neck + body; 8, Rim/neck + body + base; 9, Pedestal or ring foot.
Vessel Form: 0, indeterminate; 1, simple bowl; 2, carinated bowl; 3, bowl attached to pedestal; 4, bowl, either 1 or 2 (above); 5, flat-bottom dish (flaring sides); 6, large carinated jar; 7, narrow-necked jar; 8, cylinder stand; 9, “drum”; 10, large jar, otherwise indeterminate;
11, lid.
Temper Type: temper was recorded as a four-digit numeric field, with most to least frequent temper grains noted in rank order: 0, indeterminate; 1, calcareous sand; 2, black sand (dark minerals, e.g., pyroxene); 3, translucent sand (light minerals, e.g., quartz); 4, shell fragments (not rounded); 5, olivine sand; 6, pumice fragments.
Firing Core: 0, indeterminate; 1, pronounced; 2, present but not pronounced; 3, absent (wholly oxidized).
Main Motif Decorative Technique: This was recorded as a five-digit numeric field, with each category indicated by 0 for absent, 1 for present. Category 1: fine dentate stamping; Category 2: coarse dentate stamping; Category 3: Incising; Category 4: Carving or cutouts;
Category 5: Shell impressed.
Lip Edge Decorative Technique: This was recorded as a six-digit numeric field, with each category indicated by 0 for absent, 1 for present. Category 1: fine dentate stamping; Category 2: coarse dentate stamping; Category 3: Incising; Category 4: Carving or cutouts; Category 5: Notching (tool notched); Category 6: Crenate (finger pinched).
Inner Vessel Edge Decorative Technique (numeric, 4 chars): Recorded as a four-digit numeric field, with each category indicated by 0 for absent, 1 for present. Category 1: fine dentate stamping; Category 2: coarse dentate stamping; Category3: Incising; Category 4:
Carving or cutouts.
Raised Bands (presence/absence): I, indeterminate; Y, present; N, absent. Lime-Infilling (presence/absence): I, indeterminate; Y, present; N, absent.
Exterior Surface Treatment: 0, indeterminate; 1, eroded; 2, burnished or smoothed; 3, red-slipped.
Rim Form: 0, indeterminate; 1, vertical; 2, everted; 3, inverted; 4, sharply out-turned (90º angle).
Lip Form: 0, indeterminate; 1, flat; 2, outward beveled; 3, inward beveled; 4, rounded; 5, double-beveled; 6, expanded; 7, grooved; pointed.
Rim Diameter: recorded in centimeters, to nearest cm.
Lip Thickness: recorded in millimeters, to nearest mm.
Vessel Wall Thickness: recorded in millimeters, to nearest mm.
Lip Edge Motif: Recorded using the catalog numbers in Anson (1983).
Exterior Defining Motif: Recorded using the Anson catalog number of the motif defining the upper part of the main body decoration (zone marker).
Main Exterior Motif: Recorded using the Anson catalog number of the motif; 999 if not present in the Anson catalog.
In the initial years of laboratory analysis following the completion of fieldwork in 1988, a total of 2,297 sherds from sites ECA (924 sherds), ECB (830 sherds), and EHB (523 sherds) were analyzed according to this protocol. The data were at first recorded using Borlan’s Paradox database system, and were later migrated to Microsoft Excel and Access platforms.
In 2010, Kirch invited Chiu—who was then engaged in establishing the Lapita Pottery Online Database (LPOD) at the Academia Sinica in Taiwan—to collaborate on the analysis of the Mussau pottery. The core of the LPOD involves a standardized recording procedure for Lapita assemblages throughout the southwestern Pacific, with the goal of ultimately allowing for systematic comparisons between site assemblages (http://lapita.rchss.sinica.edu.tw/web/). Beginning in the summer of 2010, and for several subsequent summers up through 2015, Chiu and her assistants used the LODP system in Kirch’s Berkeley laboratory to record 6,976 sherds from sites ECA, ECB, EHB, and EKQ. This effort included not only the recording of an expanded suite of attributes, but also photographing of all sherds. These data and photographs have been incorporated into the LPOD.
During initial analysis, Kirch recognized that many sherds from ECA in particular could be refitted, and began to rejoin sherds to form partially reconstructed vessels. Chiu and her assistants picked up where this effort had left off, so that by 2017 a substantial number of vessels had been reconstructed, at least sufficiently to determine vessel diameter and hence to prepare drawings showing the reconstructed vessel forms with their decorative motifs. This work allowed Kirch to prepare a Mussau Ceramic Catalog of 269 such reconstructed vessels, which he compiled during the 2017–2018 academic year. The catalog allowed for a second, complementary approach to the ceramic analysis, in which the reconstructed vessels rather than individual sherds form the basis for determining frequency trends. In reporting on the various Mussau ceramic assemblages in this chapter, we have drawn upon both the sherd-based and the vessel-based data sets as appropriate.
The recording of Lapita pottery decoration, which can be quite complex as it involves combinations of multiple motifs, has posed a challenge since the initial efforts of
Mead and others (1975) to document the relatively simple Lapita pottery of Fiji. In spite of efforts to extend the Mead system to Western and Far Western Lapita assemblages
(Donovan 1973; Sharp 1988), following Anson (1983) most researchers have adopted simple numbered lists or indexes of motifs. For the LPOD project, Chiu compares the motifs on a specimen against all other motif lists and associated graphics in order to determine the uniqueness of the motif in question. Whenever possible, a drawn motif is also compared against the original sherd that contains such a motif to verify not only the accuracy of the line drawing, but also what techniques were employed to execute it. Rather than lumping motifs made of different orientations of a given design element into a single motif ID, or combing two or even three bands of motifs that are separated with lines of void spaces into one motif, in the LPOD these cases are assigned a new motif identification number and treated separately.
Once a new decorative motif has been recognized, a unique motif ID number is assigned. Usually the initials of the excavator are used to create new motif IDs for that particular assemblage (thus motifs uniquely identified in the Mussau assemblage are numbered PK1 to PKn). When a motif is too fragmented to reconstruct its original pattern, it is assigned an ID starting with “TFG” (Too Fragmented), in order to separate it from the more complete motifs that may contain grammatical information. Attributes such as what techniques were used to create it, what design elements and design units were used to compose it, what construction rules had been applied to generate it, what larger motif “theme” or family it belongs to, and where it appears on a ceramic vessel, are all coded in the LPOD. We have used the motif ID numbers as recorded in the LPOD in our descriptions in the Mussau Ceramic Catalog. A total of 472 new motifs have been identified in the Mussau ceramic assemblage.
Whether a motif may be assigned to a particular motif “theme” depends on whether it appears on the central decorative panel on the surface of a vessel (Chiu and Sand 2005; Sand 2007, 1996). If so, the next step is to determine whether it is an alloform of a more general decorative pattern shared with similarly constructed motifs. In all, 32 motif themes have been recognized so far in the LPOD, and some of these themes can be further divided into subcategories. For example, the face motif theme is further divided into “undetermined,” “triangular face,” “long-nose face with headdress,” “simple face,” “odd face,” and “double face,” a total of six subcategories. Each of these subcategories has its particular construction grammar while maintaining the basic shared underlying concept of a face in its various designs. Different subcategories of a given motif theme seem to have acquired rather different emphases among island groups; for example, long-nose face motifs are far more popular in the Reefs/Santa Cruz Lapita ceramic assemblages, while simple faces are the most abundant subcategory in the New Caledonian Lapita sites (Chiu 2007). Similar observations hold for other motif themes as well (Chiu 2015).