Parallel issues about a late Lapita site on the island of Watom in need of clarification and additional

Get Complete Project Material File(s) Now! »

An evaluation through deconstruction of 1970s motif database

Simon Best (2002:9) raised issues that I feel need to be further addressed with respect to the initial databases for motifs from Solomon Island Lapita sites SE-SZ-8, SE-RF-2 and SE-RF-6 (Figure 2). The issues relate to both presence/absence and to frequency counts used in various analyses. Essentially, his claims reduce to two concerns: (a) The published results describing the number of motifs on the decorated sherds recovered from the investigation of these three early Lapita sites are a dubious representation of their inhabitant’s human behaviour, given the enormous disparity (in percentage terms) of the limited area of each site subjected to excavation. Instead, the numbers of listed motifs are for the most part affected by the small areas sampled and the great variation exhibited by the number of sherds recovered from any one site. (b) Therefore, the number of motifs recorded for each site during their analyses in 1973 is deemed to be directly proportional to the actual areas excavated at that time, and that factor is the one which in large part controls their frequency. As a consequence, changes in motif frequencies between sites do not reflect chronological differences in the age of the site’s contents, whether of potsherds or of other items, along the lines claimed by different authors using that data. In short any employment of the findings of previous investigators is wholly unjustified.

Basic lists for the 100 coded lapita-style motifs providing number frequency and percentages used in all those computational analyses appearing in parts A to G.

The database of the motifs identified and described by Donovan (1973) in due course became the basic descriptive information on which nearly all subsequent analyses were conducted during the rest of the 1970s and 1980s. However, neither the text and tables of Donovan’s research essay, nor her accompanying catalogue of illustrations were ever edited or prepared for publication. Nor did various emendations or expansions to it appear except in outline form (Green 1990: Appendices 1 to 5). Moreover, Anson (1983, 1990) converted a great number of the rule-related variations on a Donovan/Mead motif, designated as alloforms of the initial coded design, into a fine-grained corpus of dentate-stamped motif designs that consecutively numbered 516 in total. He subsequently, using both the older and more recently recovered dentate-decorated sherds from the Reber-Rakival Lapita site on Watom Island, expanded that total to 527 illustrated and numbered dentate-stamped motif designs (Anson 2000: Table 1). Tables 7 and 8 which follow, published in full for the first time, provide the basic numerical information on the frequency with which each of the 100 motifs occurs in the three sites of SE-SZ-8, SE-RF-2 and SE-RF-6. The percentages for their occurrences are then calculated, first to a base of the total number of motifs identified in any one of the three sites, and then to a base of the total number of decorated sherds from each site.

Sherds bearing motifs whose presence and frequency is affected by area excavated

One predictable outcome of an effect of sample size is that a far higher than expected number of very low frequency or uncommon motifs would be expected in site SE-RF-2, where the excavated area was 72m². Fewer low frequency or uncommon motifs would be expected for SE-SZ-8 where the area was 51m². And very few such motifs would be expected in SE-RF-6, where the sample derived from just 20m². This prediction relies on the same density of items per m² in each site as set out earlier in this essay. It also relies on approximately the same thickness of the cultural deposit from which the sub-samples derive. The evidence supporting these proposals is for sherds bearing motifs that are not encountered in any other site, and are also in very low numbers in the site in which they occur, especially when expressed in % terms. This can be seen in Table 9.

Motifs discarded due to infrequent occurrence

This part includes those motifs which occur on sherds in more than one site—usually all three—but which may not be useful in detecting changes through time, if the relative age order of sites is SE-SZ-8, SE-RF-2, SE-RF-6, i.e. from the earliest to the latest sub-sample. This is an order postulated principally on the basis of a suite of older and current 14C dates indicative of an age order where SE-RF-6 is clearly the youngest and SE-RF-2 is older, and SE-SZ-8 near contemporary but perhaps slightly older that SE-RF-2. (a) In this analysis certain motifs are not recommended for further use in chronological analyses, either because (1) their presence in each site falls below 1.2% (Table 10), or (2) they were present in some sites and their failure to appear in others is probably due to sampling error (Table 11), or (3) their number in SE-SZ-8 and SE-RF-2 is too small to predict their expected frequency in SE-RF-6 (Table 12). [NOTE: in this Part B—and in Parts C through F—square bracketed outcomes are those calculated from a total decorated sherd base set out in Table 8. In all cases the number of examples appears first, followed by the percentage in brackets.

Motifs exhibiting no trend between the three sites

A number of motifs are found at raw number frequencies, which might be thought to indicate some trend over time. But when converted to percentages, they fit a neutral pattern of relatively constant occurrence from early to late. This means the observed outcomes do not really deviate significantly from those to be expected, but can be explained most simply as normal variations due to chance in the sampling process. For purposes of caution, these can be separated into two clusters, each of which is set out separately in Table 14. The first cluster of motifs includes those where occurrences in any of the three sites are at 1.2% or less, and it would be unwise to place much confidence in the outcome. The other cluster, in contrast, is one where there are significant numbers of sherds, and therefore reasonably informative percentages, allowing one to discount sampling error with more assurance.

Table of Contents :

• Background to this essay
• Parallel issues about a late Lapita site on the island of Watom in need of clarification and additional
• investigation
• Issues associated with the interpretation of early decorated pottery in Lapita sites of the Reef and Santa Cruz Islands
  • Publication
  • An evaluation through deconstruction of the 1970s Lapita motif database
  • Table 1 – Comparisons of survey and excavation strategies on three Lapita sites in
  • Normalised number counts for the three Reef/Santa Cruz Island decorated Lapita-style ceramic assemblages
  • Table 2 – Reef/Santa Cruz Lapita site area data
  • Table 3 – Excavated specimens of decorated body sherds, obsidian pieces and chert items in three Lapita sites in the Reef/Santa Cruz Island group
  • Table 4 – Reef/Santa Cruz motif counts for body sherds as given in Anson (1983) and the same data normalised to a base of 100m
  • Table 5 – Rim sherd counts and MNI assembled from various tables in Parker (1981) and the same data normalised to a base of 100m
  • Table 6 – Relative proportions of dentate and incised body sherds, recalculated using data presented by Donovan (1973)
  • Objectives and Predictions employed in the analytical section to follow evaluating motif occurrence
  • Basic lists for the 100 coded Lapita-style motifs providing number frequency and percentages used in all those computational analyses appearing in parts A to G
  • Table 7 – Base for percentage calculation: total motifs identified and subsumed under an overall motif designation
  • Table 8 – Another base for percentage calculation: total number of decorated sherds recovered from each site
  • Part A. Sherds bearing motifs whose presence and frequency is affected by area excavated
  • Table 9 – Motifs uncommon in, and not in the other two sites, or seemingly common in one site, though not appearing in the other two sites
  • Part B. Motifs discarded due to infrequent occurrence
  • Table 10 – Motifs present in each site, but at too low a frequency
  • Table 11 – Motifs present between 0.33 and 1.0% in SE-SZ-8 and SE-RF-6, but whether to be expected in SE-RF-2 is uncertain
  • Table 12 – Motifs present in SE-RF-2 and SE-RF-6, but whether or not in SE-SZ-8 is uncertain
  • Part C. Motifs over-represented in Site SE-RF
  • Table 13 – Motifs over-represented in site SE-RF-2 due to the far greater area excavated Part D. Motifs exhibiting no trend between the three sites
  • Table 14 – Motifs exhibiting fairly constant occurrence over 500 years’ time
  • Part E. Motif numbers not in conformity with the assumed temporal trend
  • Table 15 – An unexplained outcome of motif analysis – perhaps due to a recording failure
  • Table 16 – Yet another unexplained outcome of motif analysis – perhaps due to low numbers in predictive base
  • Part F. Motifs exhibiting a trend between the three sites in support of a chronological order
  • Table 17 – Highest frequency of motifs by percentage in Site SE-SZ-8, with declining frequency in SE-RF-2 and/or SE-RF
  • Table 18 – Lowest frequency of motifs by percentage in Site SE-SZ-8, with rising frequency in SE-RF-2 and/or SE-RF
  • Table 19 – Motifs present in SE-SZ-8 and SE-RF-2, but not represented in SE-RF
  • Part G. Motifs which have significance beyond that of establishing a serial order of change
  • Conclusion
  • Acknowledgements
  • References

GET THE COMPLETE PROJECT
An evaluation of sample adequacy for the Lapita-style ceramic assemblages from three sites located in the Reef/Santa Cruz group, Outer Eastern Islands of the Solomons.