PALEORIENT Vol. 3 1975-1976-1977
A PROPOSED SYSTEM
FOR COMPUTER ANALYSIS
OF IDENTIFIABLE FAUNAL MATERIAL
FROM ARCHAEOLOGICAL SITES
R. REDDING
J. WHEELER PIRES-FERREIRA
M.A. ZEDER
Introduction
With the increasing quantity and high quality of archeological faunal collections available for study, a widely used and standardized system for recording data is necessary. The difficulities encountered in handling large quantities of faunal remains can best be overcome with a system of attribute coding for computer analysis. Such a concept was suggested to Richard Redding by John McArdle in 1970. At that time McArdle showed Redding his approach, and they discussed the variables that should be included and problems of coding the information obtained for each element. Since that time there has been no direct communication between McArdle and our group. McArdle's system, presented elsewhere in this volume, and our system have both changed, becoming, in fact, more similar. Recently, McArdle and our group have learned of systems developed by O. LaBianca, R. Meadow, and H.-P. Uerpmann. The convergence of all of these systems demonstrates the underlying need for such a system. All those working on such approaches should now meet and resolve the differences between the various systems so that the advantage of a uniform approach to recording and handling archeological faunal data can be realized.
The system we have developed is presented below. It includes a method for coding the attributes of identifiable elements and a computer program for preliminary sorting of coded data.
The system
In the system each identifiable bone is numerically coded for twenty nine discrete variables. This encoded information is then punched on a standard computer card, utilizing seventy two spaces (fields). One card carries the coded information for one identifiable bone.
Appendix I contains a computer print-out of the program designed for the sorting and translating of this numerical code with comment cards which list the variables utilized, define the values for each variable, describe the use of each variable, and explain the operation of the program. Each line of the appendix is numbered of the right hand side. Lines 1000 to 365,000 describe the variables and their use. Included is the numerical values for each state of each variable and the corresponding verbalized abbrevation displayed in the print-out.
Several general points concerning the coding method deserve emphasis.
1. The code 01 (or 001) has the value "indeterminate" for all variables. Only those elements that could or might exhibit a state of a variable being coded but do not (possibly because they are damaged or the information provided is ambiguous) are coded as indeterminate.
2. The code 99 (or 999) reads as "not applicable" for all variables. Not applicable is used to describe those elements that could not provide, even if complete, the information required by the variable being considered.
3. In the variables used to describe fragmentation, variables 10-13, a value may be read in more than one manner depending upon the type of element it is being used to describe. For example, when variable 12, lateral/medial, is being utilized on vertebrae the value 03 should be interpreted as "the right half while if it is utilized on limb elements the value 03 should be interpreted as "lateral". A list of the equivalencies for each element type for these variables is presented in table 1.
4. Variables 15-18, dealing wu4h cultural modification of the bone, may be altered to be applicable to each site and/or worker's interest. Any substitute series of values should be reducible to three categories of values; yes, no, and possible.
5. We have found variables 25 and 26, referring to tooth-rooting and tooth-wear, to be useful. They may
