The Ethical Dilemma in the Context of Research Museums
Ideally, the decision to collect materials should imply the decision to curate them. A long-standing ethic of professional archeology in the United States, implied in the ethics code of the Society of Professional Archeologists (SOPA 1982), is to curate all materials resulting from a professional investigation. Adherence to this standard already is difficult since resources for curation are limited. In the context of a museum with a mission to preserve research collections, it is poor practice to acquire collections having limited research potential. The ethics of the museum profession require critical collecting and justification of acquisition decisions (New York State Association of Museums 1974, International Council of Museums 1987, National Park Service 1990).
Because the context of much archeological research has changed as a result of the necessity of managing irreplaceable cultural resources, it may be time to evaluate the "goodness of fit" of present ethics to the current situation (Salwen 1981; Dunnell 1984). Formulating plans for controlled growth of curated collections may mean that not all professionally collected materials will be curated (Salwen 1981). The proposed amendment to the Federal curation regulations,4 which provides deaccessioning guidelines, implies that this will be the case under certain conditions. If it is necessary to bend to practicality in deciding what can be curated in perpetuity, then there must be a well-articulated basis for making these decisions. As Milner (1987) has pointed out, "archeologists today...are faced with the considerable problem of ensuring that future researchers inherit a satisfactory database." Perhaps this goal is the most compelling ethic from a scientific point of view. There also is a need to balance this worthy goal with the equally compelling need to preserve materials for their value to the nation's cultural heritage and education.
Potential and Curation
In those institutions that do have curation of research collections as a fundamental responsibility, and thus are the appropriate institutions for curating most archeological collections, questions concerning research value arise when collections are considered for acquisition. When such criteria are applied to archeological collections, questions of research potential revolve around three general considerations: (1) project designs or plans; (2) quality of recording and recovery; and (3) redundancy of information. Without the benefit of an overall context or plan for making decisions about what to curate many of these questions are difficult to address.
The designs of field projects are important considerations in assessing the future research potential of archeological collections, because these plans structure the ways in which information and materials are collected. In the past individual archeologists essentially controlled which sites they investigated and under what circumstances. This is not the case with most cultural resource management projects. Decisions concerning contracted fieldwork often are compromises (Green 1984:ix-xi). The correspondence between the archeological record represented by collections and the intact sites is not always the most compelling force behind these decisions.
For the purposes of this discussion it is important to distinguish between the ways in which the scientific or research value of collections is assessed and the ways in which site eligibility for the National Register of Historic Places and other legal requirements of Section 106 of the National Historic Preservation Act are determined. Even though archeological research problems and cultural resource management problems clearly are linked, the goals of the two endeavors are not necessarily the same (Dunnell 1984). Although the agendas of either endeavor can lead to creation of collections suitable for long-term research, the long-term research value of collections must be evaluated separately from the eligibility of sites for the National Register and other aspects of the Section 106 process.
National Register eligibility for archeological sites generally is based on research significance, but there are relative "degrees" of significance and ongoing debate as to what constitutes National Register significance. It probably is fair to say that, in general, documenting National Register significance is much easier for large, complex sites or sites of great antiquity than for sites with limited assemblages and no clear chronological information, such as upland lithic scatters. In fact, basic research information often can be acquired from the latter type of site during a comprehensive site evaluation program that determines that the site is not eligible for the National Register. This does not mean that the collections obtained from such sites have no research significance. The same argument can be made for collections resulting from many survey projects. On the other hand, depending on the scope of a specific project, it is entirely possible for collections made from a National Register site to have little or no research value. For example, collections from highly disturbed areas on such sites would have little or no contextual integrity, and thus limited research value. If the future research potential is found to be extremely limited or nonexistent, then there is no justification on a scientific basis for curating collections for long-term research, no matter under what context the collections were made.
Basic descriptive and contextual data must accompany a collection if it is to have future research potential. A major factor in assessing adequacy of documentation is whether a systematic collecting methodology was used (Brown 1981). Even in the context of modern professional investigations, there is a wide range of performance in basic recording and recovery techniques. Although documentation may be legally acceptable, the minimum level of adequacy may or may not ensure long-term research value, especially those collections that lack associational data or any consideration of sampling biases. Curation decisions in these cases may depend on the type of archeological resource from which the collections came. Other non-scientific considerations such as humanistic or educational values also may influence curation decisions for those collections or materials found to be lacking in scientific value.
While a certain level of redundancy in collected information is necessary for research purposes, excessive redundancy in collected materials may exist for some types of sites and projects. Sites with large and highly redundant sets of materials, e.g., quarry sites and brickyards, pose questions of trade-offs between large samples and costs of facility space (Salwen 1981:570-57 1; Jelks 1990). What constitutes a sufficient sample of material from these types of sites? Regional variation in the archeological record must be considered since redundancy at the regional level, e.g., regions with many quarry sites, may allow conservatism in sample size at the site level. Consideration of sample size and composition leads to a second key factor in ensuring a satisfactory database for future research - the overall representation of the archeological record in curated collections.
Collections as Samples of the Archeological Record
This same argument can be extended to curated collections. Preservation of such representative samples in collections would insure persistence of the inherent variation in the archeological record and thus preserve the ability of future archeologists to study variation and its causes. The goal of managing collections growth would then be to move toward curating more representative samples of archeological resources for specific areas. A central point to be made here concerning the representative sample concept is that if a broad range of archeological studies of cultural characteristics and processes through time and across space is to continue, collections from one type of site/resource must not be curated to the exclusion of other kinds of sites, and that various kinds of sites, features, and materials must be represented in sufficient, yet practical, quantities to allow meaningful comparisons. Because the nature of the sample collected by cultural resource management projects is driven largely by development needs, the representation problem for curated collections becomes one of relating accumulating individual collections to a long-range plan based on long-term research needs.
Defining the composition of a representative sample is fraught with difficulties, but is not entirely impossible. Categories must be used to structure the sample, and any sampling plan will need to consider various categorical levels, such as kinds of sites, contexts, (e.g., features,) and artifacts/samples. Definition of representative samples of site types has been accomplished successfully in certain regions, as Glassow (1977:414) demonstrates. He also makes the important point that the criteria for defining categories must transcend as much as possible the restrictions of current archeological concerns and biases. The problem of assuring representativeness when blind sampling is necessary is discussed by Dunnell (1984:71-72). He suggests that a spatial frame of reference can provide an independent control.
A complementary problem inherent in obtaining representative samples is determining the proportions of the categories. This can present difficulties when samples must be drawn from numerous, broadly distributed resources as well as rare or localized occurrences (Dunnell 1984:72). For example, in the hierarchical settlement systems of complex societies, large sites with rich deposits form a small proportion of the sites in a region. Yet, given the diversity of materials and data potential of these sites, a strong argument could be made for curating collections from a larger proportion of such sites than other types of sites. There also may be multiple reasons, i.e., scientific, humanistic, educational, and artistic, to curate aesthetically and culturally significant artifacts that often are found at such sites. Such artifacts merit special consideration for scientific reasons because of the wide variety of information that can be derived from them (Brown 1982:181-183).
An even larger problem in defining samples is the inability
to predict the kinds of biases that future researchers will discover,
even in the collections resulting from today's most carefully designed
data collection strategies. While it is not possible to predict the
future, it is possible to evaluate the past. To not correct existing
recognizable biases in the curated data bank is counterproductive. Such
biases severely limit the kinds of research problems that presently
can be addressed with collections, and these limitations are apt to
become more profound over time. As was discussed above, periodic review
and evaluation of collections is necessary to identify and correct biases
so that collections management becomes responsive to new information
and changing research needs. An important component of a review and
evaluation process is comparison of the curated sample with the overall
composition of the archeological record.