Robinson: Okay, I'm going to turn the program over today to Jane Eastman, who's going to moderate and head the session this morning on analytical strategies. Thank you very much for coming.
Eastman: Thank you, Ken. We're going to start with a series of four longer presentations dealing with different methods of lithic analysis. The first is given by Dr. Andrea Lee Novick, an archaeologist with the North Carolina Department of Transportation.
Novick: Good morning. Anytime you do an archaeological project there are two fundamental issues that are involved. These are: research questions and the funding structure of the work. We'd like to think that any type of archaeological research is geared towards research questions. How we want to answer these is really related to how we'll structure our analysis. First I'd like to talk a little bit about interpretations, followed by analysis of flakes by type, lithic material, size, some discussion of refitting, thermal alteration, and trajectories of tool production.
One of the most fundamental research questions we ask is how can debitage and flake stone tools help us interpret past behavior at sites? How we address these questions really depends on each of us. For example, say we have 200 sherds that are all blue and white. There are some folks who would just say we have 200 sherds that are blue and white. And they'd just leave it like that. There are others who would go into differences between Delft and blue transfer printed wares, blue shell-edged pearlware, and other things. We have to remember that there are some folks who are interested in the details of lithics, just like there are those who are interested in the specifics of ceramics. The detailed study of flake stone tools and debitage can provide a rich body of information from which inferences about prehistoric human behavior can be proposed.
The aboriginal quarry is a unique type of prehistoric site associated with stone tool production. There are other types of extraction sites as well associated with the mining clay, galena, and mica, but lithic quarries for stone tool manufacturing are the most prevalent. Quarry sites in North Carolina include sites as Morrow Mountain [Slide 1], which we visited yesterday, the Talbot Site that Billy [Oliver] alluded to and other sites. In South Carolina there's been work on chert quarries as well as orthoquartzite workshop localities. In combination with general archaeological and ethnographic literature, this information provides a baseline for current research.
Some archaeologists have conducted ethno-archaeological research in areas such as New Guinea and Australia, and discovered that men make special trips to acquire lithic material for stone tool manufacture. These men camp at or near the quarry site and conduct some of their stone tool manufacturing around the source. This type of behavior would be manifest in the archaeological record by workshops near the outcrops, along with evidence of overnight stays such as hearths. Such evidence may indicate the site was simply a special use site or the activity of acquiring stone was part of an embedded strategy that involved many other activities.
Beyond the mere description of stone tools, lithic materials can be examined in a variety of contexts. Griffin and other archaeologists in the 1940s and 1950s proposed interpretations about site differences and site use. Later Binford and Binford advocated examining site diversity and variability. These models have received increased attention with cultural resource management work. Models such as Binford's forager/collector model have been applied to studies in North and South Carolina.
Research questions may be asked of both surface collections and excavated assemblages. Not all research questions are suitable to both types of artifact collections. An investigator's approach to recording surface collections and excavating is directly related to research questions. Research questions are directly related to the site involved in the study. Some quarries have a high density of debitage on the surface while others cannot be surface collected and there's a lot of excavation involved. So you need to be clear about where your datum point is and how and where your artifacts are going to be collected, especially if they're going to be piece-plotted. The number and size of sample units or collection areas are related again to your research questions. For example, if you go out onto a quarry site and you're not really familiar with them [quarry sites], and you start making large collections, and you start excavating units, if you haven't budgeted for these, you'll end up with thousands and thousands of flakes. And this becomes a problem because then you have to decide how you're going to analyze them. It can be a problem if you haven't carefully thought out your approach to conducting the field work.
Although numerous lithic studies have been published in the past 30 years, many archaeologists remain unfamiliar with this literature. Students at the University of Michigan such as Binford and White published detailed studies of lithic analyses back in the 1950s and 1960s. Crabtree and Butler published an important study in 1971. Crabtree published a detailed dictionary of lithic terms that is still very much relevant to studies today. These works provide the fundamentals for modeling reduction strategies, including definitions for the familiar terms such as primary decortication, secondary decortication, interior flakes, bifacial thinning flakes, pressure flakes, and other terms. Lots of archaeologists have never had a class in lithic analysis, just as, for example, let's say a lot of archaeologists have not had a class in petrology, industrial archaeology or plantation archaeology. But these basic published resources are out there and need to be consulted.
It can be argued that examining a variety of variables in a lithic study will provide more interpretive information than a mere statement that "a number of lithic artifacts were recovered." Again, it depends on the questions, but a replicable analysis method is advocated here. An approach that includes definitions of flake type [Slide 2], lithic material, and size provides information that can be used for interpreting behaviors. Inferences about these prehistoric activities and behaviors are only as good as the information that is collected and analyzed. Consequently different types of collection strategies and analytical methods can be useful for different interpretations. But one needs to be explicit about the strategies and methods used.
A variety of reduction strategies have been identified at quarry sites discovered in the Carolinas. These include reduction of cobbles [Slide 3] by the bipolar technique, reduction of tabular material from an outcrop, such as a flow [Slide 4] or a dike. This is a major outcrop at Glass Buttes, Oregon [Slide 5]. Another method for reducing cobbles is by freehand percussion. Cobbles generally occur in streams but can occur as deposits in alluvial material and sometimes as float in soil. Because cobbles have endured extreme weathering cycles, they tend to be higher quality lithic materials. Cobbles that are of low quality tend to breakdown during this erosion cycle. The selection of a cobble can be identified in the archaeological record by the cortical flakes found on sites. This can be evidenced by cobbles that may have only one or two flakes removed [Slide 6] to examine the quality of the stone. In some cases, the lithic raw material is very small. It is not unusual to find that small pebbles that fit easily between your thumb and index were used to manufacture tools.
The decision of aboriginal peoples to use a bipolar reduction strategy [Slide 7] versus a freehand percussion reduction strategy [Slide 8] remains elusive to us. However, we can identify the differences in these two reduction strategies in the archaeological record. Bipolar debitage is linear in shape and, in some cases, is shaped like small segments of a tangerine or an orange. This contrasts with freehand percussion flakes. In some cases secondary flakes from freehand percussion strategy can look like bipolar flakes. That's why it's critical that the person doing the analyses understand the differences between the types of reduction strategies and how they can be identified in the archaeological record. The key is examining the point of percussion as well as the distal end of the flake. With the bipolar technique there is evidence of direct impact on both ends of the flake, while in freehand percussion there's no evidence for impact on the distal end of the flake. If a cobble's being reduced using a bifacial technique, then one would expect to see a bifacial edge on the flake preform. Often the end product of such a strategy would be a biface that would be carried away from the site. In a bipolar reduction strategy, what remains is a bipolar core with its distinctive flake scars. The chert cobbles that I showed you earlier have facets that are used as natural platforms as in the reduction process.
The reduction of tabular material from an outcrop, such as a flow or dike, would generally result in a distinctive alternative flake scar pattern [Slide 9]. This happens before any detailed biface reduction. Alternate flakes are identified in the archaeological record by the tabular edges and their flat platforms [Slide 10]. These are produced by reducing one side of the tabular piece of the material and alternating it with the other side of the tabular material. Here [Slide 11] the platform is just what remains after you strike the flake. Noncortical flakes are removed and these are usually large in size and they don't exhibit the curvature that you associate with a bifacial flake. Their platform doesn't have a bifacial edge to it either. Bifacial flakes can be present in an archaeological site due to a variety of activities, ranging from early stage reduction to thinning of curated bifaces [Slide 12].
The analyses of the flake and the platform are critical to interpretations of what the presence of the flake means with respect to the activities conducted at the site in the past. Pressure flakes result from manufacturing as well as use. Consequently, the analysis of the dorsal and ventral sides of the flake, as well as the platform preparation are important. Wear from a haft may be visible on the ventral surface of the flake and wear may also be present on the platform.
Pressure flakes [Slide 13], which are recovered only by small screening and flotation samples, are only rarely discovered on surface sites in the Eastern Woodlands. These can, however, provide a lot of information from which interpretations about site use can be proposed. For example, pressure flakes with wear on the surface and platforms indicate the sharpening of tools. Thus, they represent the type of gear that aboriginal peoples carried from site to site. These types of flakes would be anticipated at small hunting camps where bifaces and hafted bifaces were used, or on workshop quarry sites where these types of tools would have been discarded, or on long term habitation sites where retooling would be a regular activity. If tools were carried away from a site (curated), the debitage would be the only evidence of such curated tools. Alternatively if tools are discarded at a site, their evidence would be the pressure flakes recovered in combination with tools.
Another flake class is the notch flake. These represent the manufacture and use of hafted bifaces. Depending on the reduction strategy, a notch flake may represent the final stages of hafted biface production or an earlier manufacturing stage. Blades are generally twice as long as the width and on some sites we find blade cores.
Another diagnostic flake is an outre passe [Slide 14]. It has a bifacial edge which is a platform and the outre passe passes all the way around to the other side of the biface and takes off part of the opposite margin of the tool that's been bifacially flaked.
Tool breakage [Slide 15] is another important consideration when doing analysis. This is useful for telling what is actually being manufactured or repaired at the site. In the 1970s, when I was working at the University of South Carolina, Stanley South, Al Goodyear, and others thought that it was important to try to identify lithic materials. And so I visited [Peter] Cooper at Catawba College and Joffre Coe [Slide 16] at the University of North Carolina and we looked at different types of materials we had recovered in South Carolina. And I remember very vividly Dr. Coe pointing to a lot of banana boxes filled with debitage from Hardaway and he said that I could just take as much as I wanted. And I said, "Well, it just doesn't seem right to do it." But he pressed a few flakes into my hand and I put them in one of my boxes and they're probably still down there at the Institute [of Archaeology, University of South Carolina]. So we had this collection, and in 1978, I learned that Dr. [Bob] Butler was coming to South Carolina to a geological conference. So I telephoned him and asked him if he'd be willing to come to the Institute and take a look at our materials and tell us what he thought of them. He graciously consented to come. He looked at all of our materials and said what he thought about our classification and the outcome of this work was the article I published in South Carolina Antiquities.
The flow-banded rhyolite [Slide 17] is the igneous rock we saw yesterday. The band thickness varies and it may be only a few millimeters. These bands range from straight to undulating and these are formed as the rhyolite flowed along the ground surface. During the weathering cycle these turn a buff color, making them more visible. Porphorytic rhyolite is dark to light gray and exhibits well-formed phenocrysts or mineral crystals of quartz, feldspar, or plagioclase. The presence of these phenocrysts actually determines the use of the term "porphorytic."
Felsic tuffs often weather. These range from buff to tan or from gray to white [Slide 18]. The weathered surface is chalky and can be easily scratched. Welded vitric tuff at first glance looks like a green chert, however, the flake scars are not as crisp and clean as chert flake scars and the difference is readily apparent using a microscope. The material is extraordinarily fine-grained and can have some thin veins of quartz through it. These flakes, just like chert flakes, are translucent, particularly along the margins and the distal ends.
Basalt is another igneous rock that appears as debitage in chipped stone tool assemblages. Often we think of these only as being used for ground stone tool production but some fine-grained basalts are used for flake stone tools. These can range from the dark gray to a weathered purple color. The volcanic slate or argillite is a laminated, soft grayish-green color. It occurs throughout the northeastern United States as well as the southeast. Some of these have a purple or red banding on weathered surfaces and it really represents the clay from which the stone was lithified. Different types of muds and clays can result in different colors of these particular materials.
Quartzite is another rock that we tend to think of as being used for hammerstones, but in some cases, depending on the grain size, could actually be used for making flake stone tools. Cherts [Slide 19] are common in certain parts of the Carolinas and in Tennessee [Slide 20]. Orthoquartzite is a more fine-grained material. White quartz [Slide 21] is common to most of us and this is the result usually of water that was present in the hydrothermal veins when quartz was formed. Quartz can also be recovered as a gray color or other colors. Crystal quartz is one of the more easily worked quartz materials. This is because there are so few impurities.
When analyzing flakes, it's important to consider the information that we are trying to gather from the flake. Size is often helpful. If we're going to try to size a flake, we can use a series of grids or series of baskets or trays that have a particular size of mesh through which these flakes are passed. It's important to realize that if a flake tilts to its narrowest axis, the flake size will appear smaller than it actually is. But if you don't have a lot of time or effort, that's usually the way to go if you have to examine a large group of flakes. But if you have all the time in the world, you can pick up each flake individually and size it on a grid or by using calipers. If you're looking at each of these flakes, you can look at its platform and orient it and actually get the exact size of a flake.
Refitting is also useful. You can sometimes refit the flakes to the tool. As you can see here [Slide 22], three flakes have been refitted to this biface. This was recovered from my dissertation site, Gatecliff Shelter. During Billy's [Oliver] discussion yesterday, he said that we need to have more analysis, better control, larger samples, and at Gatecliff Shelter, the entire site was excavated. The shelter was altogether not much bigger than this auditorium. Most often we don't have the luxury of excavating an entire site. But when we do excavate an entire site, and use small mesh screen or water screen, it's possible to refit these tools and actually know exactly what was left at the site.
We generally expect to find broken tools in the area where they broke, that is, in their area of manufacture, or where they were used around hearths, or where they were redeposited in dump areas. The spatial distribution of tools in relationship to hearth structures and features at the site is crucial for interpretation.
Once you examine the raw material, the size, and the flake type, it's possible to make inferences about tool production trajectories. I have used these in an analysis of lithic materials recovered in the Uwharrie National Forest in 1987. Working with Rodney [Snedeker] and Mike [Harmon], I helped on an overview for the Uwharrie National Forest, and we used these to look at a variety of sites that had been excavated and surface collected by Peter Cooper [Catawba College]. Only by examining the tools and the debitage in combination is it possible to make inferences about exactly what was going on at these types of sites.
I'm just going to talk about six major trajectories here. These include:
The concept of trajectories [Slide 23] is quite old and has roots in the technological stage descriptions of Holmes who published his work in the 1890s. Some tools, such as expedient tools, what Binford called "situational gear," and what Gould described as "just flakes struck from cobbles that were used for knives" are manufactured at the site location. Other types of tools such as curated cores, bifaces and projectile points are what Binford termed "personal gear" and go through a series of stages during their life cycles.
Sometimes tools pass through all stages at one location such as a residential base camp. Other tools pass through different stages of reduction as they are used at a variety of sites in different locations. For example, large bifaces are produced at quarry locations. These are then thinned at different workshop localities and finally transported to a residential base camp. Thus, depending on the site context ad situational context, one tool may fill a variety of needs. A goal of this type of analysis is to discover artifacts indicative as attributes of particular site types. These can be used to define site types and lead to well-grounded significance assessments in the realm of CRM.
It is important to determine how materials and tools were manufactured, used, and transported at and between sites in the region. This requires some assumptions from previous research. For example, the presence of numerous large cortical flakes usually indicate that rocks were reduced on the site. Ethnographic observations indicate large unworked rocks were rarely transported far. It seems unlikely that large intact cobbles were carried far in the Piedmont. It is essential that all lithic material types and combinations of artifacts and debitage be analyzed. To analyze artifacts only by class or material is to lose much information.
Trajectory one begins with on-site reduction of cobbles or pebbles or any nodule form of stone. Usable cobble fragments, cortical flakes, interior flakes, and bifacial flakes are produced. Some debitage may be considered flake blanks, which can be worked into other tools. This first type of trajectory is characteristic of both traditional quarry sites and quarry localities. Binford uses location to describe special purpose sites. A quarry locality as defined here (based on a variety of citations) differs from a quarry site. A quarry locality rather than being an outcrop per se consists of a cluster of cobbles or nodules such as might be found in a creek. These can be collected for use. Since these sites are more casual compared to outcrop quarries, they are less obvious in the archaeological record. An outcrop quarry is often associated with piles of debitage or waste flakes, while evidence at a casual quarry locality may consist only of a few flakes, test nodules, and cobbles. This has been referred to as procurement without quarry production. These sites generally contain flakes struck off by direct freehand percussion.
A second trajectory begins with on-site bipolar reduction of cobbles or pebbles or any nodular material. Cobble fragments, cortical bipolar flakes, interior bipolar flakes and shatter are produced. Some debitage may be suitable as flake blanks that are later worked into other tools. This trajectory is characteristic of traditional quarry sites or any place where small pebbles could be carried.
Trajectory three begins with on-site reduction of pieces of stone derived from large massive outcrops or tabular outcrops or bands of stone. Shatter, cortical flakes, interior flakes, and bifacial flakes are produced. Some debitage may be suitable for flake blanks that are to be worked into other tools.
Another trajectory is production of bifaces from curated flakes, the flakes having been made earlier at a different location. Since these flakes are curated, no large cortical debitage is anticipated. However, cortical flake blanks may be curated for tool production.
Curated bifacial cores for tool production leave a different record. Since a curated bifacial core is used as a source of lithic material, virtually no cortical debitage is anticipated. To obtain adequate flake blanks, tool edges may be prepared resulting in small preparation flakes and biface thinning flakes. Such bifacial thinning flakes are generally large enough to exhibit dorsal ridges characteristic. These are often too thin or too curved in longitudinal cross-section to serve as flake blanks. Evidence of this trajectory shows up at locations, field camps, and residential base camps.
Curated tools, especially hafted bifaces, were probably transported to sites as personal gear. Since these are complete functional tools, they cannot serve as sources of raw material unless they are totally reworked. Reworked tools are not generally included in archaeologist's discussions of production trajectories, except maybe for making hafted scrapers or other tools from broken tools. Broken tools can provide raw material for other tools. The bipolar technique, direct freehand percussion or pressure flaking can be used in these instances.
Thermal alteration [Slide 24] of tools is important when we're trying to make inferences about human behavior. Some cherts are thermally altered and they change from a yellowish color to the red color. We often use experimental research to be able to explain why thermal alteration was used. Here [Slide 25] a large pit has been excavated, then filled with layers of soil and stone, then covered with soil and then the wood is laid on top [Slide 26] and burned. Heat is applied progressively. When the heat is too intense, you can see it makes the material potlidded [Slide 27] and fractured, making it unusable. Again, this tool is from Gatecliff and if you can retrieve all parts of the tool, you are able to refit it. If you are working on some really large site, it's likely that you're not going to be able to find all parts of the tools.
This slide [Slide 28] shows a piece of quartzite that's been thermally altered and you can see that this heat-treated piece (on the left) is more finely-grained than the control sample that's not been heat-treated. This blade was actually removed from the core prior to thermal alteration. You can see the change in color here. The importance of being able to tell where thermal alteration occurs in the reduction strategy is particularly important when you want to make inferences about stone tool manufacture. This particular piece is a hafted biface that was actually flaked [Slide 29]. You can see the remnant flake scars here and up here. This piece was thermally altered and then reflaked again. You can see the pressure flakes have been taken off, leaving this fresh, white surface and the remnant thermally altered surface.
If our goal is to explain what happens on a site, one of our final products might be a chart or discussion explaining what type of reduction strategy was present on the site. We also would be interested in trying to look at the different types of raw material that were brought on to the site [Slide 30] and trying to explain what distance certain materials were carried. Quartz [Slide 31] is often common at sites and we usually expect to find it occurring naturally near the site. If we're at a site that's farther away from the source, we expect fewer flakes of a particular material. We'd expect smaller pressure flakes from materials that are also farther away. Those of you familiar with the Haw River report [Chatham County, NC] [Slide 32], this is John Cable's illustration of tool-life through time. It's possible to try and replicate these types of tools. This process begins by selecting the appropriate flake [Slide 33]. Here we have the replicated Kirk preform [Slide 34] [Slide 35] and the debitage that's actually made during its manufacture [Slide 37], based on what would (on right) and would not (on left) pass through a ¼" screen. Here [Slide 36] again we have the completed point and debitage from its manufacture that we would recover (on right) and not recover (on left) using ¼" screen recovery. We tried this with a Hardaway point [Slide 38] also.
It's important to remember that the kind of picture you paint based on any kind of analysis is only going to be as good as your materials. If you're interested and trying to get as much information as possible, and painting the most accurate picture when you're finished, you'll be interested in trying to gather as much information as you can. If you're not, you'll just end up with a statement that says, "There were 200 lithics collected from this site." So it's up to you. Thank you.