Pre-Contact Pottery Analysis: 2017 Paper

Two years ago, I did an analysis of 53 rim sherds from an Indigenous Late Woodland site in southeast Michigan - 20WN21. For some reason, only a select few have actually read the paper that I wrote for the study. Since I'm significantly expanding the initial analysis next week, I decided to post the whole paper on here.
Quite frankly, this is the first ceramic analysis that I have ever done - and it sparked an obsession with ceramics and ceramic theory in archaeology. Whenever I take a look back on my initial research notes I smile, knowing how much I have grown in my own learning. This analysis also prepared me for graduate school in ways that I had never known until now. Doing the work for this study shaped me into an actual ceramic analyst. A ceramic glow-up, if you will.

I'm expanding the study - or, I guess you could say that I'm actually redoing the whole thing. After doing more reading, I decided to employ a couple of statistical techniques to complete temporal and spatial analyses at the site (actually, that's subject to change. There's a lot more to it, but I'll explain in a later post). Additionally, I'm going to use body sherds instead of only focusing on rim sherds. In June and August of 2018, myself and a few other volunteers helped excavate the site, and we found quite a few more sherds - therefore, I'm including those. Aside from adding more sherds from 20WN21, my superiors at the Mannik & Smith Group are letting me use ceramic data from sites that they previously investigated in northwestern Ohio (thanks, Bob and Phil, for letting me do that!).
I want to tip my hat, so to speak, at Dr. Brad, my undergraduate faculty advisor at Eastern Michigan. Realistically, it's because of his help and patience that I was able to do this study to begin with; he's also letting me come back and do more analysis, even though I graduated in the fall. He's willing to just set me loose in the anthropology lab, which to me, is a sign of trust - something for which I am genuinely grateful. Also, without the existence of the Eastern Michigan University Archaeology Field School (EMUAFS), this project would have obviously never happened. This gesture extends to Jeff Sommer, for editing an earlier, more awful version of this paper. Also, he willingly discussed ceramic themes with me, and let me gawk over pre-contact ceramics during my time volunteering at the Castle Museum. 
I'm not sure how the tables will convert into the web version, so bear with me.

 Regional Interaction and Pottery at aN INDIGENOUS Late Woodland Site in Southeast Michigan

Julia R. Joblinski

Eastern Michigan University

ABSTRACT

One facet of this project is to address the regional interaction of pre-contact Indigenous populations during the Late Woodland period in southeast Michigan, which has not been well-researched. This ceramic analysis utilizes stylistic attributes (technical and visual styles) to answer questions about the assemblage of rim sherds recovered at site 20WN21 – a Late Woodland site in Wayne County, Michigan. Another facet of this ceramic analysis is to challenge current discourses among archaeologists, such as the use of type systems, by using attribute analyses as alternatives to utilizing formal type systems. The results demonstrate that studying stylistic attributes, such as visual and technical styles, can be useful when measuring Late Woodland regional interaction with ceramics.

Introduction

            This paper discusses the analysis of 53 pre-contact Indigenous ceramic rim sherds, which were recovered by the Eastern Michigan University Archaeological Field School during excavations of site 20WN21 – a Late Woodland site in southeast Michigan. The purpose of this analysis is to address whether the pottery sample from Site 20WN21 resembles the ceramic types of the Western Basin Tradition, or of other regional traditions in Michigan and/or Ohio. Based on the pottery attributes, what can be said about the regional interactions at Site 20WN21? The analysis focuses on the stylistic attributes (i.e. technical and visual styles) of the rim sherds, while also using and evaluating existing type systems.   

            Since this analysis focuses on stylistic attributes of pre-contact ceramics, the descriptions below focus on each regional tradition’s stylistic attributes as well. The ceramic assemblage from site 20WN21 consists primarily of rim and body sherds. The rim sherds lack shoulders and body portions; similarly, the body sherds lack shoulder and base portions. Due to their small size, the sherds provide limited information for forming formal types. For this reason, formal type systems are not utilized during the actual analysis process. Rather, I propose an analysis of visual and technical styles as an alternative to the use of type systems. Chivis (2016) defines these concepts: “technical style” refers to ceramic morphology and construction (2016: 56), and “visual style” refers to the surface treatment and decoration of vessels (2016: 54). Since vessel morphology and decoration can act as means to convey cultural messages, this alternative is useful for addressing Indigenous regional interaction when typology is inefficient.

Differences in ceramic style can be interpreted as indicators of cultural and ethnic identity (Hart and Engelbrecht 2011: 323). Pottery styles are also reflections of an artisan’s own preference, although subconscious cultural identity may be projected through those pottery styles (Chivis 2016). Therefore, pre-contact pottery is particularly useful for studying ancient regional interaction. Excavations and analyses of sites all over the state of Michigan have revealed that different pottery wares from distinct traditions were exchanged during the Late Woodland period. For example, in the Saginaw Valley, wares from the Juntunen Tradition were recovered in association with Western Basin Tradition and Sandusky Tradition wares. In southeast Michigan, archaeologists have discovered pottery that appear to be characteristic of both the Western Basin Tradition and the Sandusky Tradition. However, interpretations of regional interaction in the southeast region of Michigan are still, at best, unclear. Therefore, two hypotheses are to be tested in this analysis. According to the literature, pottery of the Western Basin Tradition are most commonly found in southeast Michigan. Additionally, as stated by Brashler and Holman, territories became more strict after AD 1000. Therefore, the first hypothesis is that the Western Basin Tradition should represent most of the rim sherds at 20WN21. The second hypothesis is that the presence of Sandusky Tradition pottery should be stronger than pottery indigenous to the northern Lower Peninsula of Michigan. The Sandusky ceramic tradition is thought to have originated in northwestern Ohio (Stothers 1999), so that region’s proximity to groups in southeast Michigan may be reflected in the assemblage from 20WN21. This study was conducted by using an attribute analysis, particularly employing analysis of both visual and technical styles as defined by Chivis (2016). Additionally, this analysis challenges a current discourse among archaeologists (the use of type classifications), by using attribute analysis as an alternative. The regional pottery traditions mentioned here are described in the sections below.

Issues in Typology and Taxonomy, and Implications for Analysis

            Typology, taxonomy, and classification form the core of archaeological studies; however, they are sometimes misused. They more or less provide “shorthand for description” for the archaeologist, and a “means of expressing time-space relationships in material culture” (Chilton 1999: 44). Type is a kind of artifact class, which must always be mutually exclusive, as they are used as sorting categories. Types are created with three basic purposes in mind: to describe, to compare, and to analyze, both historically and with interpretation (Adams and Adams 1991: 157; 159; 366). Many would argue that typologies have become a sort of archaeological fetish, and that some typologies serve no purpose; however, sometimes the “usefulness” is in the eye of the beholder (Adams and Adams 1991: 157). Ceramic wares are often divided into types; each type has stylistic attributes (decoration, temper, vessel shape, etc.) that may be assigned to a specific phase and/or cultural group. According to Adams and Adams (1991: 365), taxonomy refers to “a particular kind of classification having a specifically hierarchic feature; that is, a classification in which smaller and more specific classes, or taxa, are grouped into larger and more general ones”. Examples of taxonomy include the “Sandusky Tradition”, or “Wayne Ware”, just to name two.

The history of American archaeology is filled with near-constant debate about the validity of typologies and how useful or arbitrary they are in practice. Such disputes are between “natural” and “artificial” classification, attribute clustering and object clustering, “lumping” and “splitting”, and objectivity and subjectivity (Adams and Adams 1991: 265). These disputes represent a “Typological Debate”, and some would argue that this debate continues today. However, current theory supports the idea that types are merely tools. They are the means to an end and are not the end themselves (Rice 1984: 254). The historic Typological Debate has brought us to an understanding that typologies are both artificial and natural, and that typologies have certain degrees of both objectivity and subjectivity. Therefore, typologies are made for specific research purposes of the archaeologist (Adams and Adams 1991: 48; 279). Although many have argued about the best methods of type classification, many have also failed to question their practicality (Adams and Adams 1991: 235). Due to the nature of the assemblage presented here, I propose a stylistic attribute analysis as an alternative to typological analysis.

            The analysis of the rim sherd assemblage from Site 20WN21, as presented here, employs a mix of both visual style analysis and technical style analysis. “Visual style” refers to surface treatment and decoration of pottery, which often act as means to convey cultural messages, and express cultural identity. Visual styles include the different modes of decoration (e.g. incising, linear punctate, cordmarking). “Technical style” refers to the variables related to vessel morphology, such as lip and rim form, and body shape, amongst others. Although technical styles are less pronounced than visual styles, they reflect the personal style of the artisan, and would still project subconscious aspects of the artisan’s cultural identity (Chivis 2016: 54-56). Attribute analyses are particularly useful when formal type systems cannot be observed on pot sherds. For example, when they are recovered in the field, pottery vessels are seldom intact, due to both pre- and post-depositional events. It can be difficult or even impossible to assign small sherds to a specific type, since significant design attributes that are related to formal types are not usually present.  

Terms such as variable and attribute are used heavily in this paper and should be defined for clarification. A variable is a characteristic, used to describe artifacts (the decoration on a pot, for example). Variables are considered in the descriptions or definitions of types, as they vary from one entity to the next (Adams and Adams 1991: 370). Lastly, an attribute is a characteristic of a variable (e.g. the kind of decoration on a pot, such as oblique incising) (Adams and Adams 1991: 331). Concerning the regional traditions discussed in this study, the constructs of typology and taxonomy are entrenched in the literature. Therefore, emphasis is placed upon the vessel forms and surface treatments that were most significant in each phase of each regional ceramic tradition. It is understood, however, that without the use of typology in past research, there would be no basis of classification between these regional traditions. As already stated, typologies are incredibly useful means of classification. However, in this analysis, they are not the most efficient or useful tools.

This study uses a culture-historical framework, which is based upon taxonomic nomenclature that have already been established by researchers in the past. For example, Stothers’s definitions of the Western Basin Tradition and Sandusky Tradition are used, in addition to McPherron’s initial analyses on pottery from the Juntunen site. Stothers’s take on the Western Basin Tradition is largely ignored by archaeologists in Michigan, in favor of Fitting’s original definition of southeast Michigan pottery traditions, namely the Wayne and Younge Traditions. Neither school of thought is completely absolute, as they should never be. The argument between the two schools of thought is a hallmark example of the shortcomings of forming rigid taxonomies; in both cases, archaeologists formed equally valid theories based upon archaeological evidence. For one reason or another, each became a “gospel”, and they became canonized in the literature. However, although Stothers’s culture-historical framework is not completely without fault, the Western Basin taxonomical nomenclature are used in this study for the sake of simplicity.

The Western Basin Tradition

At Late Woodland sites in southeast Michigan, Western Basin Tradition ceramics are most commonly found. This regional tradition is believed to have originated in the sand points and stream estuaries of northwestern Lake Erie and Lake St. Clair (Stothers and Abel 2002: 74). The Western Basin Tradition consists of four known sequential phases: the Gibraltar phase (AD 500-750); the Riviere au Vase phase (AD 750-1000); the Younge phase (AD 1000-1200); and the Springwells phase (AD 1200-1300) (Stothers 1999: 197; Monaghan and Lovis 2005: 92). Each sequential phase represents a shift in decoration style, vessel shape, and vessel size. Research shows that the Western Basin Tradition represents one homogenous cultural group; this tradition has been defined as a “regionally distinct prehistoric cultural tradition” (Bechtel and Stothers 1993: 96). Western Basin Tradition ceramics have also been found in regions such as northwestern Ohio, northeastern Indiana, southcentral Michigan and the Saginaw Valley, and southwestern Ontario (Bechtel and Stothers 1993: 96; Fitting 1965: 151; Brashler and Holman 1985: 145). This suggests that populations from southeast Michigan were often interacting with the other cultures around them through trade and seasonal movement (Brashler and Holman 1999: 214). Table 1, below, compares the most common attributes of each phase of the Western Basin Tradition.

Gibraltar Phase (AD 500-750) ceramics tend to be fairly plain in decoration. Vessels are often cordmarked or cord-impressed, sometimes with slightly more elaborate tool-impressed patterns around the lip; rim interiors are typically undecorated. Gibraltar Phase vessels tend to be globular in shape, with round bases. Rims are usually vertical, sometimes decorated with cross-hatching (Fitting 1975: 152). Temper consisted of sand and crushed granitic material. Contextual evidence suggests that the populations that made Gibraltar phase ceramics originally resided in southern Ontario (Stothers 1999: 196).

            Riviere au Vase Phase (AD 750-1000) vessels are more elongated in shape than their earlier, more globular predecessors. Vessel bases are round or subconical in shape. Rims are either straight-flaring or slightly outcurved, and castellations are common (Bechtel and Stothers 1993: 98). Stothers and Bechtel (1993: 96) have noted that almost all Riviere Ware vessels “have at least one horizontal band of oblique impressions on the interior of the rim, they frequently have two bands on exterior rim surfaces, and they commonly exhibit transverse impressions on the lip”. Lips can also be flattened and/or folded (Stothers and Abel 2002: 77). Plait and triangle motifs are typical among Riviere au Vase vessels, and only the paddle-and-anvil technique is used to make the vessels themselves. Diverse types of incising and impressions are exhibited (Bechtel and Stothers 1993: 98). Temper consists of crushed granitic material and sand (Fitting 1965: 154).

            Younge Phase (AD 1000-1200) vessels are slightly elongated in shape, with vertical or constricted necks (Bechtel and Stothers 1993: 98). Castellations occur, though less often than in the Riviere au Vase Phase; rim interiors are also left undecorated. Lips are flat, thickened, or folded; they are often decorated with cord-wrapped stick impressions, oblique tool impressions, and tool incising (Stothers 1977: 21; Bechtel and Stothers 1993: 98). Collared vessels increased in popularity during the Younge Phase. Multiple bands of oblique stamping and plain cord-roughened exterior body decorations are rather popular during this phase. No interior decorations have been recorded (Bechtel and Stothers 1993: 99). According to Bechtel and Stothers, tool impressions are the hallmark of the Younge phase (1993: 99).

            The Springwells Phase (AD 1200-1300) vessels are comparatively larger in size and capacity than Riviere or Younge Phase vessels, and are either elongated or slightly elongated in shape. Exterior decoration was usually done with cord-roughening techniques, and large scalloped castellations often occurred. The most common decoration motifs include oblique stamped impressions, and one or more horizontal bands on the collar. Lips are rounded or pointed, and are decorated with tool impressions. Rim interiors are left undecorated; rims are either vertical or slightly outcurved (Bechtel and Stothers 1993: 99). Net-impressing was also been used as a mode of decoration (Fitting 1965: 155).

Table 1: Phases of the Western Basin Tradition
Phase	Exterior body decoration	Collar presence and decoration	Rim form	Exterior rim decoration	Interior rim decoration	Castellations	Lip shape	Lip decoration
Gibraltar phase: AD 500-750	Cordmarking; cord-impressed	N/A	Vertical or outcurved	Cordmarking; horizontal cord impressions	Undecorated	N/A	Rounded	Cord and tool impressions
Riviere au Vase phase: AD 750-1000	Cordmarking; dentate stamp; tool impression; cord-wrapped stick	N/A	Straight-flaring; slightly outcurved	Two or more bands of oblique impressions; transverse impressions; plait and triangle motifs	One or more horizontal bands of oblique impressions	Castellations commonly occur	Flat; folded	Transverse tool impressions
Younge phase: AD 1000-1200	Multiple bands of inclined stamping; cord-roughened	Collars increasing in popularity	Slightly outcurved	Oblique tool impressions; cord-wrapped paddle or stick impressions	Undecorated	Castellations occur infrequently	Flat; thickened; folded; everted	Cord-wrapped stick impressions; tool incising; oblique tool impressions and incising
Springwells phase: AD 1200-1300	Cord-roughening; smoothed-over cord impressions;  net impressed	One or more horizontal bands of tool impressions; oblique tool impressions; cord roughening; net impressions; stab-and-drag  incisions	Vertical; slightly outcurved	Transverse tool impressions; non-transverse tool impressions	Undecorated	Large, scalloped castellations are popular	Rounded or pointed; folded	Transverse tool impressions; tool impressions

Mackinac Straits/Sault Ste Marie Wares

            As stated above, Late Woodland ceramic wares characteristic of the Mackinac Straits and Sault Ste Marie region of the northern Lower Peninsula of Michigan have been found at sites in the Saginaw Valley. These regional wares are situated into three sequential phases: the Mackinac phase (AD 800-1000), the Bois Blanc phase (AD 1000-1200), and the Juntunen phase (AD 1200-1300). Cord-impressed varieties of Mackinac phase pottery may have appeared at the Younge site in southeastern Michigan, suggesting some regional interaction (McPherron 1967: 91). Table 2 (below) compares the most common attributes of each phase from this region.

            Mackinac Phase pottery (AD 800-1000) are loosely defined, but some characteristics can be noted. The Mackinac Phase is characterized by vessels with outcurved rims, often cord-decorated and subconical bases. Mackinac phase pottery typically lacked collars and castellations. Rim and lip interiors are largely left undecorated, but cord-wrapped tool impressions have been found on them (Monaghan and Lovis 2005: 88; McPherron 1967: 89). Lips are usually thickened and are occasionally everted (Fitting 1975: 183; McPherron 1967: 86).  Bodies are decorated with single or double rows of tool impressions, and/or cord-wrapped paddle (fabric-wrapped paddle also appeared to be popular); more than one horizontal bands of single cord impressions also occur (McPherron 1967: 90). It should be noted, however, that the majority of Mackinac phase pottery are undecorated (McPherron 1967: 88).

            Bois Blanc phase vessels (AD 1000-1200) are globular in shape, with vertical rims and rim and shoulder appliqués. The primary mode of decoration is cord-wrapped tool impression. Pottery from this phase exhibit “braced” rims, i.e. rims have been thickened either by being folded over themselves or by adding an extra strip of clay (McPherron 1967: 104). Shoulders are less pronounced on Bois Blanc Phase pottery than on Mackinac phase pottery (McPherron 1967: 105), and rims often bear low castellations (Monaghan and Lovis 2005: 88). Lips are sometimes decorated using a cord-wrapped stick, or with cord impressions. Lips are usually pointed, and are occasionally everted (Fitting 1975: 183; McPherron 1967: 104). Closely-spaced tool impressions, overlapping punctuate, and brushing often decorate the vessel body. Rims are typically left undecorated and are generally vertical (Monaghan and Lovis 2005: 88). However, some rims are decorated by “beading”, or with deep tool impressions on the rim. Interestingly, however, when a beaded rim is present, castellations are never present (McPherron 1967: 108).

            Juntunen phase ceramics (AD 1200-1500) are known to have geometric decorative motifs (e.g. chevron-like patterns), and plaits of stab-and-drag and punctuate designs. Cord-wrapped object impressions sharply decrease during this cultural phase (McPherron 1967: 111). Large castellations and well-defined collaring occur frequently. Rims flare slightly outward, and rim interiors are sometimes cordmarked (Monaghan and Lovis 2005: 88). Lips are rounded or flat, and lip exteriors are smooth and plain. Unlike the Bois Blanc phase, most lips remain unthickened (McPherron 1967: 112). Drag-and-jab impressions (e.g., dragging a tool across the clay, while simultaneously making deeper tool impressions) immediately below the lip are seen on Juntunen phase pottery (Clark 1993: 3; McPherron 1967: 113).

Table 2: Phases of Pottery Wares from the Northern Lower Peninsula
Phase	Exterior body decoration	Collar presence and decoration	Rim form	Exterior rim decoration	Interior rim decoration	Castellations	Lip shape	Lip decoration
Mackinac phase: AD 800-1000	Cord-impressed; cordmarked	N/A	Outcurved to slightly outcurved	Cordmarked; plain	Undecorated	N/A	Flat; everted	Tool-impressed; undulating
Bois Blanc phase: AD 1000-1200	Cord-impressed in bands; overlapping tool impressions; brushing	N/A	Vertical; thickened	Appliqués; beading	Undecorated	Low castellations	Pointed or everted	Cord-wrapped stick; cord impressed
Juntunen phase: AD 1200-1500	Geometric tool-impressed designs	Geometric and chevron-like motifs	Slightly outcurved	Linear tool impressions	Cordmarked; drag-and-jab impressions; linear impressed	Castellations occur frequently	Rounded	Smooth/plain

Northern Ohio and the Sandusky Tradition

            Ceramic wares characteristic of the Western Basin Tradition have been discovered in northern Ohio, suggesting that Western Basin populations interacted with populations indigenous to Ohio (Bechtel and Stothers 1993: 96; Stothers and Schneider 1999). Alternatively, Sandusky Tradition wares, which are indigenous to northern Ohio, have been found in southern Michigan and the Saginaw Valley (Brashler and Holman 1985: 146; Stothers and Graves 1983: 124). The Sandusky Tradition is split into five sequential phases: the Green Creek phase (AD 500-1000); the Eiden phase (AD 1000-1250); the Wolf phase (AD 1250-1450); the Fort Meigs phase (AD 1450-1550); and the Indian Hills phase (AD 1550-1650). Table 3 lists and compares the most common attributes of each phase of the Sandusky Tradition.

            Green Creek phase ceramics (AD 500-1000) are also loosely defined, but again, some characteristics can be noted. Green Creek pots are typically small and globular in shape. Rims are incurved or vertical; rim exteriors are decorated with vertical cord impressions. Rim exteriors are also decorated with checkered or stamped motifs towards the end of the phase. Vessel bodies are often decorated using a cord-wrapped paddle (Stothers and Abel 2002: 78). Lips are flat and smooth (Abel 1999: 209). Plainly decorated Green Creek phase pottery can look undeniably like Gibraltar phase pottery.

            Eiden phase ceramics (AD 1000-1250) are larger and more bag-like in shape than their Green Creek phase predecessors. High rims and subconical bases are common (Stothers and Abel 2002: 85). No interior decoration among Eiden phase pottery has been recorded. Body decorations increasingly became smoothed-over towards the end of the Eiden phase and during the onset of the Wolf phase (Stothers and Abel 2002: 87). Castellations occur infrequently, but when they do, they are typically rounded (Abel 1999: 220). Festoon patterns, executed by incising or linear tool impression (“Parker Festooned” is one variety), became popular during the Eiden phase (Abel 1999: 218). Zone incising and zone linear impressions also began to emerge during this phase. “Zone” decoration refers to incising or linear tool impressions, which form a shape (usually angular), with tool impressions and/or dentate stamping inside of the “zone”. Some zones are left undecorated (Abel 1999: 210, 228).

            Wolf phase (AD 1250-1450) vessels almost always have one to three bands of “festoon” (chevron) patterns on the rim. Usually, there are one to two bands of horizontal design above or below the festoons on the rim (Abel 1999: 208). The formal type “Parker Festooned” has been used to refer to ceramics temporally situated in the Eiden and Wolf phases (Abel 1999: 207; Stothers and Graves 1983: 121). Castellations are either angled or rounded. Lips are rounded and are usually left plain (Abel 1999: 222). The use of shell as temper, instead of granitic grit, begins to emerge at the end of this temporal phase. Shell would continue to be used in the Sandusky Tradition through the end of the Indian Hills phase (Stothers and Graves 1983: 116).

             Fort Meigs (AD 1450-1550) vessels are characterized by squat, globular bodies; bodies are often roughened by fabric-wrapped paddles. Orifices are often as wide as the shoulders. Rims are thickened, and lips are occasionally folded. Rim and lip exteriors are often decorated with tool impressions; rim interiors are often left undecorated. Lugs and strap handles adorn Fort Meigs wares, along with notched appliqués (Abel 1999: 208). Notches appear on the lip, directly above the handles (Abel 1999: 237). As noted above, shell was predominantly used as temper during the later phases of the Sandusky Tradition (Stothers and Graves 1983: 116; Abel 1999: 208). Festoons are still popular during this phase (Abel 1999: 218).

            The Indian Hills phase (AD 1550-1650) is characterized by a single horizontal band of vertical or oblique linear or dentate stamps, usually just below the lip. The stamping was typically done using the edges of shells. Occasionally, festoon patterns would appear below the stamping (Abel 1999: 208). All recorded sherds from the Indian Hills phase are tempered with shell (Stothers and Graves 1983: 116). Lips on Indian Hills phase vessels are usually rounded and plain. Sometimes, when handles are present, notches appear on the lip directly above the handles (Abel 1999: 242). Festoon decoration is abandoned at the onset of the Indian Hills Phase (Abel 1999: 218).

Table 3: Phases of the Sandusky Tradition
Phase	Exterior body decoration	Collar presence and decoration	Rim form	Exterior rim decoration	Interior rim decoration	Castellations	Lip shape	Lip decoration
Green Creek phase: AD 500-1000	Cordmarking	N/A	Vertical or incurved	Vertical cordmarking; cord-wrapped paddle	Smoothed	N/A	Flat	Smooth
Eiden phase: AD 1000-1250	Typically plain; bands of horizontal cord impressions	N/A	Straight-flared; vertical; slightly outcurved	Oblique tool impressions; festoons; zone impressions or incising	Undecorated	Infrequent; rounded	Flat	Tool-impressed
Wolf phase: AD 1250-1450	Typically plain; tool-impressions occur in horizontal bands	Occasional bands of tool-impressions or cord-impressions	Straight-flaring; sometimes outcurved	Oblique impressions; festoon pattern; horizontal bands of tool impressions above or below festoons; zone impressions or incising	Cordmarking; smoothed-over cord impressions	Angled; round	Rounded	Predominantly plain
Fort Meigs phase: AD 1450-1550	Typically plain; some cordmarking; lugs and handles	Largely lacking collars; some are cord-impressed	Outcurved to slightly outcurved; folded	Appliqués; thickened rims; smoothed	Smoothed-over cordmarking	Occur infrequently	Flat	Tool-impressed; notches above handles
Indian Hills: AD 1550-1650	Oblique dentate stamping	N/A	Slightly outcurved; sometimes tapered	Notched appliqués; oblique and vertical tool impressions	Cordmarked; sometimes plain	N/A	Rounded	Plain; notches above handles

The Saginaw Valley

Although no regional ceramic “tradition” originated out of the Saginaw Valley during the Late Woodland period, it was a rich locus for Indigenous groups. Ceramic remains of Early, Middle, and Late Woodland populations have been uncovered in the Valley, and there is evidence that ceramics were manufactured there (Fischer 1972). Ceramics characteristic of surrounding regions in Michigan have been found in the Saginaw Valley, suggesting that distinct regional groups of people often interacted with one another in this area on a frequent basis (Brashler and Holman 1985: 145). In the early Late Woodland Period, seasonal subsistence cycles and the trading of Norwood and Bayport chert allowed regional cultures to maintain peaceful relations with one another, perhaps also having close marriage or kin relations. This interpretation is reinforced by stylistic similarities between regionally distinct ceramic wares found in the Saginaw Valley (Brashler and Holman 1999: 220). After AD 1000, however, regional ceramic styles become less similar in style, suggesting that boundaries between groups became more distinct. Nonetheless, the presence of regionally distinct ceramic wares indicates that more than one cultural group simultaneously used the Saginaw Valley in some way or another (Brashler and Holman 1999: 220). Further reinforcing this, pottery resembling that of the Western Basin Tradition has been uncovered in the Saginaw Valley, as were ceramic wares characteristic of Michigan’s northern Lower Peninsula (Sault Ste Marie, Mackinac Straits region), such as Mackinac phase, Bois Blanc phase, and Juntunen phase wares (Brashler and Holman 145-147).

Methods

            Located in Oakwoods Metropark, site 20WN21 is situated on a terrace overlooking the Huron River and adjacent floodplain and is believed to have been a temporary settlement (Ensor 2009: 209). The Eastern Michigan University Archaeology Field School (EMUAFS) conducted excavations and shovel test surveys in 2007. During these surveys, EMUAFS uncovered a total of 875 pre-contact artifacts; these artifacts included pottery sherds, stone tools and lithic debitage, small bits of ground stone, and fire-cracked rock. Charcoal fragments from a hearth at the site were radiocarbon dated to AD 1030-1220 (Ensor 2009: 207; 210). According to a study conducted by Danielle Julien, the lithic assemblage from 20WN21 consisted of Stoney Creek chert, Bayport chert, and Coshocton and Onondaga cherts (2016: 10). Stoney Creek chert is found locally, in southeast Michigan; Bayport chert was sourced from the Saginaw Bay region of Michigan; Coshocton (also known as Upper Mercer) chert originates from central Ohio; and Onondaga chert was sourced in New York and Ontario (Julien 2016: 6; Monaghan and Lovis 2005: 80). As stated previously, Western Basin Tradition ceramics are found in southeastern Michigan, the Saginaw Valley, northern Ohio, and Ontario (Monaghan and Lovis 2005: 92; Brashler and Holman 1985: 149; Fitting 1975: 150). Therefore, there seems to be a correlation between the regional origins of chert types found at site 20WN21 and the regions in which Western Basin Tradition ceramics are found.

The EMUAFS pottery collection was recovered from excavation and site surveys. Shovel tests were dug in either 15- or 20-centimeter-deep intervals and were approximately 40 centimeters in diameter. The soil from each interval was screened through six-millimeter screens. Shovel tests were usually placed 20 meters apart. Test units and excavation units were excavated either by 10-centimeter arbitrary levels or by unnatural sediment changes (cultural levels). Sediment from each level was screened through ¼-inch screens. In each case, artifacts were collected from the screens and were placed in bags, which had each artifact’s proveniences written on them. At the end of each field day, the artifacts were brought back to Eastern Michigan University’s Anthropology Lab. Students of the field school catalogued, washed, and re-bagged the artifacts for storage, curation, and future analysis.

            A total of 53 rim sherds were analyzed in this study. As stated previously, variables such as rim form and decoration, and lip form and decoration, can vary distinctly between the temporal phases of each regional tradition. Therefore, only the rim sherds collected from Site 20WN21 were examined in this study. The following variables of each rim sherd were recorded on a form: catalog number; rim form; lip form; shape of castellations (when present); folded lip (when present); exterior body decoration (when present); exterior and interior rim decoration; shoulder decoration (when present); and lip decoration. Temper was not recorded, unless a sherd exhibited a different temper material other than the usual granitic grit and sand. The attributes for these variables were abbreviated; these abbreviations acted as “codes”, which made the recording process quicker and easier. For example, linear oblique incising was abbreviated as “INLO”, vertical cord impression was abbreviated as “CIV”, and so forth. This analysis design somewhat resembles Chilton’s analysis of Late Woodland sherds, which were recovered from sites in Massachusetts and New York (1999: 46). She used attribute analysis of vessel forms (which she termed “an attribute analysis of technical choice”) as a means to address her own questions, and she proved that it could be done without using quantified cultural-historical “types” (Chilton 1999: 46). Additionally, it can be impossible to assign typology to sherds, since certain stylistic attributes are simply not present (Orton et al. 1993: 80). Many of the sherds in this study were small enough in which body shape and body decoration were not present; therefore, types are not assigned or considered in this study. As indicated in the literature cited above, it is possible to recognize differences from region-to-region, and phase-to-phase, based upon the visual and technical styles of rim sherds.

Results

            Some forms of decoration (cordmarking, for example) are similar among the regional temporal phases, so that had to be taken into account. Table 4 displays the number of rim sherds that exhibit decoration representing each regional temporal phase. Some regional phase attributes overlap, as illustrated in the table. Two rim sherds (catalog numbers 181 and 253), which exhibit outcurved rims and rounded lips, are not included due to their lack of decoration; they could represent an undecorated variety of three different temporal phases, so they are therefore deemed indeterminate. Of the 53 rim sherds analyzed, 62% represent only the Western Basin Tradition; 1.8% represent only the northern Lower Peninsula of Michigan; and 5.6% only represent the Sandusky Tradition. 13.2% represent a mixture between the Western Basin Tradition and the northern Lower Peninsula; 9.4% represent a mixture of the Western Basin Tradition and the Sandusky Tradition; and, interestingly, 1.8% represent a mix between the Sandusky Tradition and the northern Lower Peninsula (Tables 5 and 6). As illustrated by Table 4, no sherds have Gibraltar, Mackinac, Fort Meigs, or Indian Hills phase attributes.

Table 4: Quantities of Regional Attributes
	Gibraltar phase	Riviere au Vase phase	Younge phase	Springwells phase	Mackinac phase	Bois Blanc phase	Juntunen phase	Green Creek phase	Eiden phase	Wolf phase	Fort Meigs phase	Indian Hills phase
Gibraltar phase
Riviere au Vase phase				1
Younge phase			6	12		3		1		1
Springwells phase		1	10	3		5	1	2		1
Mackinac phase
Bois Blanc phase			1
Juntunen phase							1*
Green Creek phase									1
Eiden phase								1
Wolf phase						1		1
Fort Meigs phase
Indian Hills phase

Table 5: Percentage of true attributes derived from different regions represented at 20WN21
Tradition/Region	Percentage
Western Basin Tradition	62%
Sandusky Tradition	5.6%
Northern Lower Peninsula	1.8%

Table 6: Percentage of sherds combining attributes from different traditions at 20WN21
Region/Tradition	Percentage
Western Basin/northern Lower Peninsula	13.2%
Western Basin/Sandusky	9.4%
Sandusky/northern Lower Peninsula	1.8%

Six sherds exhibit attributes characteristic of only the Younge phase (AD 1000-1200), and three rim sherds exhibit only attributes from the Springwells phase (1200-1300). Twelve sherds have mostly attributes of the Younge phase, but have fewer attributes of the Springwells phase. Alternatively, ten sherds exhibit traits mostly of the Springwells phase, with fewer traits characteristic of the Younge phase. A mixture of these attributes suggests that the pottery being used were of transitional varieties. This may be supported by the radiocarbon dates from charcoal fragments at the site (AD 1030-1220). However, two rim sherds have attributes characteristic of the Riviere au Phase (AD 750-1000); this suggests that either 20WN21 was used earlier than previously thought, or that the attribute analysis may not be the best option for placing sherds into temporal phases. As already stated, 62% of the rim sherd assemblage from 20WN21 is represented by the Western Basin Tradition. Overall, these results confirm my first hypothesis.

As to confirm my second hypothesis, there is a stronger presence of attributes from the Sandusky Tradition (5.6%) than from the northern Lower Peninsula (1.8%). However, attributes of the Younge phase and the Springwells phase were seen combined with Sandusky Tradition attributes (Green Creek and Wolf phases; 9.4%). This is somewhat unexpected, since it is assumed that boundaries became more rigid after AD 1000 (Brashler and Holman 1999: 213). As stated previously, however, much interaction between northern Ohio and southeast Michigan populations has been recorded (Brashler and Holman 1985: 146; Stothers and Graves 1983: 124). According to Julien’s study, the presence of Coshocton chert at 20WN21 already proves that interaction between northern Ohio and populations in southeast Michigan occurred (2016: 6, 10). Therefore, this analysis supports Julien’s study of chert sources. Although the “blending” of regional attributes on the same pots could be interpreted as the result of regional interaction, further testing needs to be done in order to confirm it. One sherd has attributes only from the Juntunen phase (AD 1200-1500), although its placement on the table is tentative. Three sherds exhibit characteristics of both the Younge and Bois Blanc (AD 1000-1200) phases. One sherd exhibits Springwells phase attributes, combined with Juntunen (AD 1200-1300) phase attributes[1].

Conclusions

Most of the rim sherds from Site 20WN21 resemble pottery of the Western Basin Tradition (62%). A significant number of rim sherds represent the Sandusky Tradition (5.6%), which is continuing evidence of interaction between people of southeast Michigan and northwest Ohio during the Late Woodland period. This association at the site is complimented by the presence of Coshocton chert, which, as Julien explained in her lithic analysis, is sourced in Ohio (2016: 6). A less significant amount of pottery from 20WN21 exhibit attributes characteristic of the northern Lower Peninsula (1.8%), which suggests some form of regional interaction. This is supported by the presence of Bayport chert, which was evidently distributed in southeast Michigan, the Saginaw Valley, and the northern Lower Peninsula of Michigan during the Late Woodland period.

There appears to be a blending of regional attributes, which was not expected at the beginning of this analysis. Previous research suggested that regional and cultural boundaries became more strict around AD 1000, marked by the shift from interregional homogeneity in pottery decoration to increased regional distinctions between regional pottery styles (Brashler and Holman 1999: 220). However, combined attributes of the Western Basin Tradition and the northern Lower Peninsula (13.2%), and combined attributes from the Sandusky and Western Basin Traditions (9.4%), were observed on rim sherds from 20WN21. These suggest that populations from southeast Michigan and surrounding regions were continuing to form close ties with each other after AD 1000. Some regionally-specific manufacturing techniques may have been passed along through the generations, if marriage and kinship practices between regional populations continued. The presence of non-local chert types at 20WN21 (Coshocton, Bayport, and Onondaga chert) (Julien 2016: 10) also supports this interpretation. However, the “blending of attributes” may be interpreted differently. The literature cited previously in this paper discusses the most common visual and technical styles of the phases in each regional pottery tradition; few phases have attributes that overlap, which can cause confusion to the analyst. Additionally, potters were able to express their own artistic ability through their craft, which makes the proposed phases non-absolute. To assess the “larger picture” regarding pottery and Indigenous regional interaction, body sherds and pottery assemblages from other sites excavated by the EMUAFS should be analyzed. 

Overall, this study shows that classification by type is not necessary in pottery analysis, especially when studying a small assemblage of rim sherds. These results suggest that an attribute analysis of visual and technical style can help one interpret which region/phase a sherd once belonged to, although this method should be tested and expanded. Lastly, this analysis reinforces the interpretation that not only seasonal migration, but economic exchange and marriage and kinship ties were driving forces of regional interaction in southeast Michigan.

WORKS CITED

Abel, Timothy J.
1999. A Temporal and Spatial Analysis of the Parker Festooned Ceramic Type. Midcontinental                                                                       Journal of Archaeology 24(2): 201-256. The University of Iowa, Iowa City.

Adams, William Y. and Ernest W. Adams.
1991. Archaeological Typology and Practical Reality. Cambridge University Press, London.

Arnold, Dean E.
1984. Social Interaction and Ceramic Design: Community-Wide Correlations in Quinua, Peru. In Pots and Potters: Current Approaches in Ceramic Archaeology, edited by Prudence M. Rice, pp. 113-131. Institute of Archaeology, University of California, Los Angeles.

Bechtel, Susan K. and David M. Stothers.
1993. New Perspectives on the Settlement-Subsistence System of the Late Woodland Western Basin Tradition, ca. 500-1300 A.D. North American Archaeologist 14(2): 95-122. Baywood Publishing Co., Amityville.

Brashler, Janet G. and Margaret B. Holman.
1985. Late Woodland Continuity and Change in the Saginaw Valley of Michigan. Arctic Anthropology 22: 141-152. University of Wisconsin Press.

Brashler, Janet G. and Margaret B. Holman
1999. Economics, Material Culture, and Trade in the Late Woodland Lower Peninsula of Michigan. In Retrieving Michigan’s Buried Past: the Archaeology of the Great Lakes State, edited by John R. Halsey, pp. 212-220. Cranbrook Institute of Science, Bloomfield Hills.

Chilton, Elizabeth S.
1999. One Size Fits All: Typology and Alternatives for Ceramic Research. In Material Meanings, edited by Elizabeth S. Chilton, pp. 44-60. The University of Utah Press, Salt Lake City.

Chivis, Jeff.
2016. The Introduction of Havana-Hopewell in West Michigan and Northwest Indiana: an Integrative Approach to the Identification of Communities, Interaction Networks, and Mobility Patterns. PhD Dissertation. Department of Anthropology, Michigan State University, Lansing.

Clark, Caven P.
1993. Mott Sauna Beach Excavation and Site Survey at Isle Royale National Park. Midwest Archaeological Center Technical Report 27. National Park Service, Midwest Archaeological Center, Lincoln.

Ensor, Bradley E.
2009. Chipped Stone Debitage, Material Selection, and Expedient Tool Production in Southeast Michigan. Midcontinental Journal of Archaeology 34(2): 199-222.

Fitting, James E.
1965. Late Woodland Cultures of southeastern Michigan. Anthropological Papers: Museum of Anthropology, University of Michigan 24: 151-159. University of Michigan Press, Ann Arbor.

Fitting, James E.
1975. The Archaeology of Michigan: A Guide to the Prehistory of the Great Lakes Region. Cranbrook Institute of Science, Bloomfield Hills.

Hart, John P. and William Engelbrecht.
2011. Northern Iroquoian Ethnic Evolution: A Social Network Analysis. Journal of Archaeological Method and Theory 19(2): 322-349.

Julien, Danielle.
2016. Prehistoric Exchange along the Huron River in Southeastern Michigan. Eastern Michigan College of Arts and Sciences, Ypsilanti.

McPherron, Alan.
1967. The Juntunen Site and the Late Woodland Prehistory of the Upper Great Lakes Area. Museum of Anthropology: Anthropological Papers, 30. The University of Michigan, Ann Arbor.

Monaghan, William G. and William A. Lovis.
2005. Modeling Archaeological Site Burial in Southern Michigan: A Geoarchaeological Synthesis. Michigan State University Press, Lansing.  

Orton, Clive, Paul Tylers, and Alan Vince.
1993. Pottery in Archaeology. Cambridge University Press, Cambridge.

Rice, Prudence M.
1984. Overview and Prospect. In Pots and Potters: Current Approaches in Ceramic Archaeology, edited by Prudence M. Rice, pp. 245-255. Institute of Archaeology, University of California, Los Angeles.

Stothers, David M.
1977. The Emergence and Development of the Younge and Ontario Iroquios Traditions. Ontario Archaeology, 25: 21-30.

Stothers, David M.
1994. The “Wayne” Taxonomic Construct in Michigan Prehistory. North American Archaeologist, 15(1): 1-15. Baywood Publishing Co., Amityville.

Stothers, David M.
1999. Late Woodland Models for Cultural Development In Southern Michigan. In Retrieving Michigan’s Buried Past: the Archaeology of the Great Lakes State, edited by John R. Halsey, pp. 194-202. Cranbrook Institute of Science, Bloomfield Hills.

Stothers, David M. and Andrew M. Schneider
1999. Cooperation in the Name of Preservation: Southeastern Michigan Site Excavation and Analyses by the River Raisin Chapter – MAS and University of Toledo.

Stothers, David M. and James R. Graves.
1983. Cultural Continuity and Changes in the Western Basin, Ontario Iroquois, and Sandusky Traditions – A 1982 Perspective. Archaeology of Eastern North America 11: 109-142. Eastern States Archaeological Federation.

Stothers, David M. and Timothy J. Abel
2002. The Early Late Woodland in the Southwestern Lake Erie Littoral Region. In Northeast Subsistence-Settlement Change – A.D. 700-1300, edited by John P. Hart and Christina B. Rieth, pp. 78-92. The University of the State of New York, New York.  

---

[1]  McPherron (1967: 107) has stated that varieties of Bois Blanc phase pottery has been discovered “over a relatively great distance, with very little stylistic variation, at least from the north end of Lake Michigan to the Saginaw Bay area, and always in place places readily available by water transport”. McPherron also stated that a sherd, which was discovered at a site near Flat Rock, Michigan (located in southeast Michigan), had attributes that are undeniably similar to the popular braced-rim Bois Blanc style (1967: 276). He also suggested that cord-impressed varieties of Mackinac phase pottery was present at the Younge site in southeast Michigan (1967: 91), which alludes to earlier association between Western Basin peoples and peoples from the northern Lower Peninsula. Therefore, it is not surprising that pottery styles resembling the Bois Blanc phase make an appearance again in southeast Michigan. The presence of pottery attributes from the Bois Blanc phase is complimented by the presence of Bayport chert at 20WN21, which was exchanged in the northern Lower Peninsula and in the Saginaw Valley (Julien 2016: 6; Ensor 2009: 212).