Anth. 473L/ 573L: Archaeological Measurement and Laboratory Analysis

Spring 2003

 

Exercise 4b: Measuring Reductive Variation in Archaeological Collections

(More Debitage and Cores)

Due April 7 With Exercise 4a
15 POINTS

 

Exercise 4B is the continuation of Exercise 4A. In exercise 4A, you performed an item analysis, examining and coding for attributes of each piece of debitage (Shott 1994; Steffen et al 1998). The item was the synthetic unit; attributes were the analytic units. In preparation for completing this exercise, we asked you to construct a paradigm of the attributes you measured across items and examine the distribution of classes. Associated with that construction should have been a series of tables or figures.  In this part of the exercise, you will make further observations of the debitage sample and also make observations on cores.

 

For this exercise, you will need several copies of the debitage coding sheet (more will be provided in the lab), one copy of the core coding sheet, and your copy of the debitage protocol.

 

To reiterate the goals of this 2-part exercise: First, we want to give you some hands-on familiarity with archaeological chipped stone residue.  In doing so, the exercise highlights some of the basic problems encountered in examining chipped stone collections, e.g., reliably assessing flake dimensions, and reliably identifying and separating removals from cores. Second, we want to introduce paradigmatic classifications as research tools.  That is, the exercise provides an example of how paradigms can be used as analytical devices to partition variation in archaeological collections of objects.  Third, we want you to describe and evaluate variation in core reduction within and among different raw materials. There is considerable diversity in the size, type, and mode of reduction of materials on each tray. By making a series of observations, you will be able to describe some of that variation.  After completing this exercise, you should be able to:  (1) distinguish among cores, flakes and angular debris, (2) define and assess the basic dimensions of cores and flakes, and (3) construct paradigmatic classifications relevant to your research interests (4) build preliminary descriptions of variation in chipped stone residues by raw material.

 

 

BASIC ORGANIZATION OF THIS EXERCISE:

 

The procedure for 4B has two parts.  For flakes and angular debris, you will continue the item analysis by coding for one of the two variables you did not complete (exterior flake scars or platform).

 

You will also work on an item-by-item basis with cores (because, after all, cores and removals comprise the totality of reduction). With cores, you will choose what variables you want to measure, but the measurements must be related to the goal of describing reduction.  Because you only have a week to complete this exercise, it is important to make all of your measurements during your lab period.

 

 

STEP ONE:  Completion of Coding for debitage

 

Now that you have greater familiarity with the diversity of debitage, you should begin by examining your previous decisions. Were some of the items that you coded as angular debris possibly flakes?  Did you correctly identify the raw materials?  Then decide which of the two additional variables will likely provide you with more information about the process of reduction in your sample.  Perhaps much of the debitage is missing platforms.  If that is the case, then coding for platforms will likely not be very productive.  Perhaps you want to try to apply Baumler's grid to the complete flakes in your sample and to orient the direction of previous flake removals. (Although this was not suggested, you can do it if you choose.  It is likely to be very time-consuming.) This information and the location of the platform could, in turn, inform on core form.  Alternatively, you could simply record the number of flake scars on the exterior surface of complete flakes.  Complete the coding for whatever variable you choose.

 

 

STEP TWO: Coding cores

 

Make observations on the cores in your collection.  You will decide what variables are important to record, and record them using the core coding form. Because you are deciding, your first step is to create a protocol.  Decide on the variables, and use numbers to identify attribute states for each variable. (Write this down!) To describe the cores,  you should focus on size, material, on platforms and direction of removals -- see discussions in Baumler (1988) and Kuhn (1995) for further discussion of how these variables are relevant to measuring variation in reduction. Because size is difficult to measure, weight is always a useful proxy for size.  You might note the number of platforms and their relationships to one another, e.g., opposed platforms.  Also note whether the cores exhibit unidirectional, bidirectional or multidirectional removals. We suggest you code for no more than three variables, including weight.

 

 

STEP FOUR: Data entry and analysis

 

Put the information from your coding forms into a database/ spreadsheet/ statistical program (Excel should work well).  You should already have a file from Exercise 4A.  The first step is to enter the new observations on the debitage.  Then, on separate worksheet, enter your observations on cores. Because there are many fewer cores than flakes, this table will be much smaller.  Manipulating these observations will be the basis for examining membership in your paradigmatic classes. Below we suggest a number of questions that can be used to query your data base and to construct appropriate tables and charts.  To address the questions, you should construct appropriate tables and figures (these are your data!!)

 

 

Debitage Analysis, Manipulating Paradigms, and Constructing Data

 

Create a second paradigm for your flakes using all five variables (excluding weight), and compare the results of paradigms 1 and 2  in terms of the number of classes, and how your debitage sample is distributed across the classes.  N.B. You will not be able to draw a paradigm with five variables as a table (at least not in any simple way in two dimensions).  Put the information into your computer program and look at the distribution of classes.  Calculate the number of classes in the paradigm, count how many classes you have represented in your sample, and then calculate a ratio of filled to empty classes.  Once you have constructed the “master paradigm”, you can begin to collapse variables in order to address specific questions (e.g. you can look specifically at the relationship between material type and weight or material type and number of external surface flake scars.)  Construct tables and charts that convey this information.  These tables and figures are the basis of your descriptions.  Consider whether the addition of another variable provided more information about the process of reduction.  Did the addition create too many classes in your “master paradigm” such that there are very few items in each class?  Consider various ways of collapsing 2 variables into one or removing variables in order to more usefully display the variation you examined. 

 

When working with debitage it is useful to consider both counts and weights because they contain different kinds of information.  Obsidian, for instance may be very abundant, but weight might be 1/100 of basalt debitage which is represented by fewer but larger pieces.  Examine the distribution of items across the classes you create, considering proportions (percentages) as well as frequencies.  Which debitage classes have the most members?  Which raw material type is represented by the most numbers of flakes? Does this variation change if you consider both counts and weights?  If it does change, why does it change? Does the proportion of flakes match the proportion of angular debris by raw material type? Going further, does debitage vary by size and reduction characteristics?  In other words, is there any tendency for kinds of debitage to vary by raw material?

 

Although the focus of this analysis is on how classes partitioned variation, it might be useful to consider variation in individual attributes. For example, determine the proportions of cortical and noncortical flakes or the proportions of flakes and angular debris for each raw material. In other words, you have two ways to proceed with the analysis: by class and by single variables.  Examining the latter can be extremely useful for elucidating differences among raw materials.

 

 

Core Analysis

 

First, there are many fewer cores in your sample than debitage. This fact means that your core classes will have many fewer items.  Nonetheless, we want you to characterize the diversity of cores through paradigmatic classes.  The first step is to characterize your cores in terms of classes.  Perhaps you have 3 raw material types represented by cores.  If this is the case, then a simple histogram of weight units by raw material may be an effective way to characterize cores.  Degree of reduction may be reflected in the differential presence of cortex remaining on the nuclei.  Does core size inform on the use life of the material?

 

To maximize the information present on cores, you need to link cores and debitage together.  Here, controlling for raw material is essential.  For a given material type, does size of core relate in a general way to size of flakes?  Does the number of core faces relate to number and direction of exterior flake scars on flakes? 

 

 

DISCUSSION OF RESULTS

 

Excluding tables, figures and coding forms, your write-up should not exceed 8 pages, and please note which collection you examined.  As usual, your discussion should include both analytic and synthetic components (follow the recommended lab write-up structure).

 

 

Part One: Evaluate the unit structures

 

First, discuss how the paradigms sorted the observed variation for both cores and flakes. The paradigms should be discussed separately; summarize them.  How well did the paradigms perform? How did your classification of cores sort the variation across raw materials--at this point you should enumerate which variables you used to construct your classification and give some kind of justification for using them. Second, evaluate both paradigms of debitage in terms of how they partitioned the variation, particularly across raw materials. If, for instance, each class contained a small number of members, then the paradigm was too fine-grained for the sample. If you collapsed some variables, be sure to describe this and justify your decisions. Conversely, if a small number of classes contain most of the specimens, then the paradigm was too coarse-grained for the sample. Discuss the advantages and disadvantages of paradigmatic classifications for measuring variation in core reduction.  Did examining the classes allow you to document and examine variation across raw materials?  Lastly, discuss what you learned about variation in reduction from sorting the collection using paradigmatic classifications.

 

 

Part Two (Measuring Reduction)

 

While Part One focuses on evaluating the utility of paradigmatic classifications, Part Two is concerned with assessing variation in core reduction. Here the analytic units are the classes you created and variation in core reduction is the synthetic unit.  First, discuss the documented differences (or lack thereof) in reduction across raw materials.  Second, did weight indicate differences among the raw materials, and did it allow you to evaluate differences observed relative to formal (nominal) dimensions? Third, did your classification partition the variation in terms relevant to the goals of your analysis?  That is, did your classification allow you to document differences among raw materials deemed significant for reduction?  Finally, did examination of the cores assist you in understanding differences in reduction across the raw materials; in other words, how do the cores and flakes articulate with and inform on one another?

 

In evaluating your results, consider the reliability and validity of your units.  Are the variables used in the analysis valid measures of reduction? It might be appropriate to consider here Steffen's (1998) discussion of the difference between technological and functional units. What reliability issues did you encounter in classifying the debitage?  Once again, evaluate your units, to the best of your ability. In 1971 in Systematics in Prehistory, Dunnell stated: "If the units [classes] are the product of a properly executed paradigmatic classification--i.e., all possible meanings that any correlations the units might have are known--they are overtly built into the units. The application of the units in a practical problem constitutes the testing of the hypotheses made in the classification." (1971:76). Did you find this to be the case? Steffen et al. (1998) provide an excellent example of a paradigm that shows the difference between cores and tools.

 

 

 


 

 

                                                            CORE CODING FORM

Name___________                                                                                        Tray______

 

SPEC NO.