Shifting priorities apparent in Middle and Late Woodland ceramics from Nova Scotia

Abstract

In the Maine–Maritimes Region, the Late Woodland (1350–500 BP) Period is thought to have been accompanied by a decrease in ceramic quality because of less-skilled potters. Although ceramics made during the Late Woodland tend to physically degrade easier than earlier ceramics because of coarser pastes and less well-joined coils, the reasons for the change in manufacturing practices have not been explored. Using the ceramic assemblage from the Gaspereau Lake Reservoir Site Complex in King’s County, Nova Scotia, Canada, this study used simple statistical techniques to suggest that potters increasingly used more expedient manufacture through time. These practices would have enabled potters to turn out pots under tighter deadlines to support large-scale gatherings that probably became more prevalent during the Late Woodland Period.

Keywords

Maine–Maritimes Region ceramic technology Mi’kmaq coiled and paddled method technological evolution

Introduction

Archaeologists have repeatedly made the assertion that ceramics can be used to infer information about cultures beyond the traditional inferences of ethnic boundaries and chronology. Nevertheless, the majority of studies using ceramics continue to glean a limited range of information from them. In order for these circumstances to improve, there is a need for ceramic literature that demonstrates methodologies useful to analysts in other contexts and regions or that can be applied by generalists. Work by Shepard (1956), Matson (1965), Skibo and Schiffer (1995, 2001), Bronitsky and Hamer (1986), Braun (1986), Rice (2005), and Arnold (1985) and many others have laid the groundwork for understanding ceramics as a fairly direct response to environmental, social, and economic requirements and contexts that can be understood by looking at the ceramics themselves. Increasingly, ceramic analysts are looking for ways to translate this knowledge into methodologies that can be applied to a broader range of archaeological contexts.

One of the difficulties of understanding ceramic change in the context of the society that made them is determining which changes are significant and which are incidental. Without an in-depth knowledge of ceramic technology as well as of the region being studied, ceramic change can appear illogical or simpler than it actually is. For this reason, methodologies for understanding ceramic change through time are important for developing explanations using sound logical inferences. This study presents a method for recognizing a move from smaller-scale, more individualistic pottery manufacture to larger-scale, faster production of pottery using simple statistical analysis to chronologically delimit groups of temper percent and coil breaks.

Aboriginal pottery in the Maine–Maritimes Region evolved throughout the Woodland Period (3000–500 BP), but explanations for the observed changes have not been well explored. The most common reason cited for change has been the declining proficiency of Woodland potters later in time (e.g., Petersen and Sanger, 1991), an explanation that lacks empirical evidence and is based on assumptions about what constitutes a high-quality vessel. These assumptions are that thin walls, harder bodies, finer pastes, and more carefully applied decorations are indications of higher skill (Bourque, 2001: 79; Davis, 1991: 97; MacIntyre, 1988: 326; Petersen and Sanger, 1991: 123; Rutherford, 1991: 105).¹ These assumptions come from Western traditions of pottery manufacture—with which most Americanist archaeologists are most familiar—that have developed for storage and serving vessels but have no antecedent or identifiable tradition of ceramic cooking technology. As a result, there is a tendency to see porous and soft clay bodies, thick walls, and coarse temper as undesirable qualities that potters would avoid if they could (Frink and Harry, 2008; Skibo, 2013: 52) and, therefore, as an indication of low skill. However, examination of ceramics in various contexts of changing mobility, subsistence activities and social organization have shown that the ideal ceramic is not necessarily the most aesthetically pleasing, abrasion-resistant and long-lasting in a depositional environment, nor the most durable against impact (e.g., Frink and Harry, 2008; Skibo, 2013: 52).

A change occurred in Maine–Maritimes Region ceramics that has been acknowledged by researchers but has not been satisfactorily explained. Late Woodland (1350–500 BP) pots are frequently crumblier, thicker walled and exhibit more frequent coil breaks than their Middle Woodland counterparts (2200–1350 BP) (e.g., Allen, 1981; Bourgeois, 1999; Foulkes, 1981; Kristmanson, 1992; Petersen and Sanger, 1991; Sheldon, 1988), attributes that have given archaeologists the impression that these pots were built by less-skilled workers. Yet little attention has been given to the important changes that occurred in manufacturing techniques; of particular note are the increase in temper percentage (nonclay particles added to the clay body) and the decrease in paddling between the Middle and the Late Woodland periods. Almost no attention has been paid to what these changes signify in the activities of the people who made them. This lack of attention has caused archaeologists to overlook important insights about how society changed from the Middle to the Late Woodland, since ceramic manufacturing practices have been shown to follow technological trajectories that are closely tied to subsistence base (Arnold, 1985, 2008; Chilton, 1998; Claassen, 2002; Hoard et al., 1995; O’Brien et al., 1994; Osborn, 1988), economic activities (Arnold, 1985; Deal, 1998; Myers, 2006; van der Leeuw, 1977), and kinship organization (Gosselain, 2000; Graves, 1985; Wright, 1972: 92). The activities of people during the Late Woodland can be better understood through a close examination of how ceramic manufacturing practices changed.

Region-wide changes in ceramics indicate patterning whose responsible mechanisms can and should be investigated. Attributes result from the chain of choices made by potters (Budden and Sofaer, 2009; Dobres, 2001; Michelaki, 2008; Sillar and Tite, 2000) based on a complicated set of needs, ideas, and influences that can be partially unravelled by “reverse engineering” a ceramic assemblage (Peirce, 2005). In this study, I investigated the reasons for the widespread shift in manufacturing practices at the transition to the Late Woodland Period and tested the assumptions about poor quality ceramics during this period. I looked at the two main attributes that cause the crumbliness in Late Woodland pottery noted by many researchers—temper percent and coil breaks—and used statistical analysis and chronological evidence that would suggest why they changed. I investigated the largest ceramic assemblage from the Maine–Maritimes Region, which comes from the Gaspereau Lake Reservoir (GLR) Site Complex in Nova Scotia (Figure 1).

Figure 1.

Location of the GLR site complex in the Maine–Maritimes Region.

Temper and ceramic skill

Archaeological interest in temper has usually concerned manufacturing location and ceramic trade (e.g., Dickinson, 2001; Owen et al., 2014; Peacock et al., 2007; Roper et al., 2010), mechanical function (e.g., Feathers, 2006; Garcia, 1989; Kilikoglou et al., 2007; Schiffer and Skibo, 1987), thermal shock resistance (e.g., Braun, 1986; Bronitsky and Hamer, 1986; Schiffer et al., 1994) and comparison of different kinds of temper (e.g., Arnold, 1985; Feathers, 2006, 2008; Herbert, 2008; Hugill and Dickson, 1988; Skibo et al., 1989). Temper constitutes an important area of pottery analysis because a large amount of information is contained in the choices of minerals, processing methods, sources, and variability within a particular context. Temper attributes are not random—meaning they indicate technological choice (Sillar and Tite, 2000; Skibo and Schiffer, 2001)—but at the same time are constrained by mechanical, geological, and geographical properties to a greater degree than, for instance, decoration or form (Arnold, 1985). Using the wrong minerals or materials (or using them in the wrong way) can cause ceramics to fail (break or disintegrate) during manufacture or use, so working with temper involves a complicated set of negotiations and trade-offs that affect the final product.

Because temper indicates a specialized skillset and knowledge specific to the temper being used, changes in attributes—even seemingly inconsequential ones such as temper percent or a change in mineral type—mean a change in skillset. This is true whether the change arises from technological, resource-related, or ideological needs.² Three things are important to note about such a change in skillset. First, the degree of the change depends on the difference in how the new temper attributes affect the workability and functionality of pottery. For example, the problem of using shell temper—that firing between 600°C and 900°C causes the clay body to spall and disintegrate (O’Brien et al., 1994: 285)—required that potters build their skill and knowledge to practice appropriate firing techniques (Feathers, 2006: 114–115). Firing is a tricky process regardless of temper type, and firing with shell temper would have required an extra degree of control, probably requiring years or even generations before the process was satisfactorily reliable. Firing techniques developed for grit temper simply do not work with shell temper. Because firing represents perhaps the most risky stage of pottery manufacture in terms of both labor and time, changing firing practices represents a large deviation from an existing skillset, requiring a concomitantly large investment of time and labor equal to the number of pots made and ruined in unsuccessful firings. Changes in, for instance, paddling practices may represent far less investment, as unsuccessfully formed pots could still be recycled (Crown, 2007: 205).

Second, a change in skillset would indicate that potters had a reason to develop such a skillset because acquiring new skills can be hard and there is evidence that potters will avoid changing their practices if they can (Sillar and Tite, 2000: 10). Longacre et al. (2000: 277) suggested to potters from the Philippines that they skip the laborious and expensive step of coating their pots with an iron-rich slip before smudging them to turn them black. They pointed out that not even the potters themselves could tell the difference between pots with and without the slip; however, the potters were not interested in changing their process. Arnold (1985: 220) pointed out that “the decision to innovate is economically illogical. In order for innovations to be successful, they must be saleable.” Because of this, he found that only potters of the highest status or the lowest status will innovate, the former to maintain a reputation as distinctive, and the latter, out of desperation, will copy trends. A change in pottery practices, therefore, is indicative of a strong pressure or incentive (but see Crown, 2007 for situations in which innovation may be encouraged during learning).

Finally, depending on the circumstances, potters may retain the earlier skillset, making them more versatile, and this versatility should be reflected in the ceramic assemblage. As a rule, versatility resulting from multiple skillsets would need to show both skillful application (i.e., not missing integral technological attributes) and patterning (i.e., attribute states or clusters repeated in enough specimens to infer meaningful repetition). For example, Sassaman (1998; also O’Brien et al., 1994: 276) reported that the earliest pottery in the Southeast of North America was fibre-tempered or soapstone-tempered, turning to sand, grit, and grog temper around 3000 years ago (Skibo, 2013: 42). Feathers (2009: 130) and others noted that, in this same region, shell temper began to appear as early as 1700 BP, at least 600 years earlier than the conventional understanding of 1000 BP as the advent of shell temper. Feathers also notes that other temper types persisted later than archaeologists have usually thought. It is now clear to archaeologists that different firings would have been required for different temper types (e.g., Feathers, 2006; Herbert, 2008; Hoard et al., 1995; O’Brien et al., 1994); this means that potters would have had to know multiple techniques for firing. Many sites in the Maine–Maritimes Region contained both shell- and grit-tempered ceramics that were contemporaneous, meaning that, collectively, potters knew how to manufacture both kinds of pottery.

These kinds of increased versatility can be evaluated for whether they indicate multiple skillsets by asking whether the variability would have required different technological know-how. In the cases listed above, the answer would be yes, whereas in a case of decorative variability, potters need not have developed entirely new skills or knowledge bases to employ different techniques. A fine line separates technological versatility from a naturally variable attribute, and so an argument must always be made to claim potters were more versatile during one period than another.

Background to the problem: Ceramics in the Maine–Maritimes Region

In the Maine–Maritimes Region, pre-European ceramic manufacture is thought to have been made in a more-or-less unchanging technological tradition from their invention or adoption ca. 3000 years ago to their decline and discontinuation sometime before or around European contact. Ceramic chronologies are based primarily on decorations (Petersen and Sanger, 1991) but also on temper and, to some extent, morphology (Stapelfeldt, 2009). Although many researchers have observed differences in paste, coloration, number of coil breaks, and friability of ceramics between different periods, no assessment of the reasons for these changes has been put forward.

Pre-European ceramics are estimated to have been manufactured in the region since ca. 3000 years BP up to ca. 500 BP or at the time of European contact (Petersen and Sanger, 1991).³ Ceramics can be roughly divided into Early, Middle, and Late Woodland styles based on decorative, forming, and paste attributes. Early Woodland (3000–2200 BP) ceramics are characterized by fabric-impressed surfaces and by coarse, grit-tempered pastes formed into conoidal jars with pointed bottoms and wide mouths. Middle Woodland (2200–1350 BP) ceramics feature delicate pseudo-scallop shell (PSS) and dentate decorations, two variations of linear-edge tools stamped on the surfaces of pots in rows and chevrons. Pastes of Middle Woodland ceramics tend to be harder and redder than those of the Early Woodland, and walls tend to be thinner; the conoidal jar shape continued from the previous period.

Ceramics of the Late Woodland Period (1300–500 BP) were decorated with cord marks made by wrapping cord on a stick or paddle edge or, more rarely, by impression with fabric. Pastes used during this period were coarser and softer, often were tempered with shell instead of grit, and walls tended to be thicker and surface color grayer and less oxidized. A more nuanced ceramic sequence was proposed by Petersen and Sanger (1991) that divided ceramic manufacture into seven ceramic periods, which are summarized in Figure 2.

Figure 2.

Summary of Petersen and Sanger’s (1991) ceramic sequence for the Maine–Maritimes Region.

Although broad temporal changes in style are recognizable in ceramics from the Maine–Maritimes Region, the changes have often been considered superficially or else their connection to social changes has remained uninvestigated. However, many researchers have acknowledged that ceramic change in the Maine–Maritimes Region seems to be connected with ceramic changes in the wider region of the Northeast, characterized as horizons. Horizon styles are not generally thought to relate to social dynamics such as technological innovation, scale of production, subsistence patterns, settlement, or economic activities; however, the widespread adoption of ceramic attributes across ethnic and geographical boundaries repeatedly through time indicates far-ranging networks through which groups participated in macro-practices and traditions.

Changes in ceramic attributes through time have not been considered technological in nature; instead, they have often been considered as indications of general skill levels during different periods. Petersen and Sanger (1991: 130), authors of the most thorough assay of ceramic change in the Maine–Maritimes Region, noted “technological continuities” from each period into the next, and characterized all the periods on an informal scale of “technological proficiency” based on paste hardness, degree of paste crumbliness, number of coil breaks, and wall thickness (Petersen and Sanger, 1991: 123). In their view, shell temper “was largely a stylistic, not strictly functional (i.e., technological or utilitarian) change from the earlier usage of various forms of grit” (Petersen and Sanger, 1991: 139). Unflattering assessments of declining ceramic quality after the early Middle Woodland—that “much of the change is, technologically speaking, for the worse” (Bourque, 2001: 79)—have also been made by other researchers (e.g., Davis, 1991: 97; Kristmanson, 1992; Rutherford, 1991: 105; Sanger, 1979). Unfortunately, most researchers have not proposed the cause of this supposed peak and later decline.

Temper attributes in coil-built pottery

Attempting to recreate Woodland-style grit-tempered pottery will quickly show that the percent of temper is constrained by workability—that is, the more temper, the less workable the clay becomes (Rice, 2005: 60–63; Skibo, 2013: 52). The crumbliness of Late Woodland pottery, therefore, is best understood not as an unconscious degrading of skill or a serendipitous result of experimentation but rather as an intentional push against constraints to increase temper percentages, something potters would have done only if the increased effort were worth doing for some advantageous reason.

Temper in coil-built pottery represents a negotiation of trade-offs: though essential for cooking vessels, it reduces the workability of the clay paste, exposes the pot to cracking during heating events, increases susceptibility to abrasion, and significantly increases clay preparation time because of the lengthy and labor-intensive process of crushing materials such as granite (Bronitsky and Hamer, 1986; Herbert, 2008; Skibo and Schiffer, 1995). The paste’s workability is the most immediate barrier to using temper because if the potter is unable to form the pot (e.g., the pot collapses under its own weight or cracks during forming), then the process of pottery manufacture is entirely impossible.

A strategy for coping with this reduced workability is paddling, which strengthens and thins the walls, smooths and shapes the pot more precisely, and obscures coil joins (Rye, 1981). In the finished product, paddling allows for better thermal shock resistance, a much wider range of shapes, and more symmetrical forms. Although paddling and coil-building are strongly associated in the ethnographic literature (Woolsey, 2017: 88–90), paddling has been understudied and under-recorded in archaeological assemblages. Since paddling adds clear benefits to pottery, a decrease in paddling through time is significant because of what is lost in strength, thermal shock resistance, and durability.

Paddling in ceramic manufacture

Forming pots by means of coiling and paddling is an effective pottery manufacturing method that has been used cross-culturally and during lengthy traditions of pottery manufacture (Arnold, 1985: 206–207; Costin, 2011; Druc, 2013: 494; Gosselain, 2016; Graves, 1985: 12; Miller, 1985: 38–39; Neupert, 2007; Wallaert-Pêtre, 2001: 477). However, in small-scale societies, ceramics are also frequently made by other techniques that indicate different priorities and skillsets of potters. For instance, the vertical half-mold method is considered by Arnold (1985: 202–205) to require less skill than the coiling and paddling method, such that it is sometimes practiced in situations where potters do not have time or resources to learn more in-depth methods or where they need to cut corners to increase output (Arnold, 1985: 220–221). Paddling is therefore significant in understanding the priorities of potters not only because it is not strictly necessary to the production sequence (meaning it is a way of adding value) but also because it is risky in the sense that pots can be ruined during the process of paddling.

The degree of paddling during the forming process is apprehensible by studying the number and smoothness of coil breaks on sherds and by the direction and coarseness of the wall lamellae (the foliated look of ceramics in the broken wall cross-section). These forming attributes are not usually reported in archaeological reports and articles although they often indicate chronological significance.

The increased coil breaks and crumbliness through time noted by other researchers is hypothesized in this paper to indicate that potters tried to increase output and use more expedient methods for ceramic production during the Late Woodland Period. Although reduced paddling could mean lower skill if it is the only observed change, the other, equally important change—increased temper—needs to be considered alongside paddling. The question of skill level is approached by looking at indications of well-developed motor skills that would have been present to a greater degree in highly skilled practitioners and to a lesser degree in potters who made pottery sporadically, as well as attributes that indicate knowledge passed on from teachers (Costin, 1991: 336).

Statistical attribute analysis can detect change through time or add another line of evidence if change is suspected. Since the relationship of temper to paddling is not necessarily straightforward, comparing attribute states to chronological categories can detect the strength of association and correlation, which are not necessarily evident from distributions or descriptive statistics such as range and mean values. Additionally, statistical analysis can suggest which attributes should be considered salient to understanding change by noting how closely related they are to other attributes that are known to be significant (such as chronologically sensitive decoration types). Finally, statistical analysis can suggest pottery standardization by highlighting significantly large categories. If an argument can be made that such a category would have resulted from skill or knowledge, then it constitutes a line of evidence for standardization and/or skill. Statistics are therefore a powerful tool for understanding latent mechanisms of change as well as validating qualitative assessments of significance.

The Gaspereau Lake Reservoir Site Complex

The GLR ceramic assemblage (>18,000 sherds) is the largest ceramic assemblage from the Maine–Maritimes Region. The GLR Site Complex is located on Gaspereau Lake in Kings County, Nova Scotia, Canada. Nova Scotia is situated in the Maine–Maritimes Region within northeastern North America. The GLR assemblage is characterized by large numbers of debitage of high-quality lithic material from the Minas Basin, unusually large numbers of prestige and trade goods such as copper, red ochre, and shark’s teeth, and at least five stone-lined hearths (Sanders et al., 2014) that may have been earth ovens (Woolsey, 2017: 21). The total artifact assemblage is over 300,000 artifacts ranging from as early as the Palaeoindian Period to the early Historic Period.

The GLR Site Complex is a group of 21 sites clustered around the northeastern edge of Gaspereau Lake where it feeds into the Gaspereau River. The total number of sites in the immediate area is 33, 31 of which are pre-European (Sanders et al., 2014: 3), with the largest being the End of Dyke Site situated on the shoreline and just north of the Gaspereau River outlet as well as the reservoir created by the dam. The area is known to be unusually rich in food resources, including seasonal runs of gaspereaux (Alosa pseudoharengus), Atlantic salmon (Salmo salar), and smelt (Osmerus mordax) (Sanders et al., 2014: 350). The Gaspereau River can be navigated beginning ca. 1 km downstream from Gaspereau Lake, providing a route to the coast, where it empties out into the Minas Basin on the Bay of Fundy. Another significant cluster of sites (the largest of which is the Melanson Site) is located along the Gaspereau River (Nash et al., 1991; Sanders et al., 2014: 350), and an important quarry for toolstone is located on the coast near the mouth of the Gaspereau River (Deal, 1988). This toolstone, variously referred to as Minas Basin chert (Gilbert et al., 2006), Scots Bay Agates (Sanger, 1991: 55), and Scots Bay chalcedony (Deal, 2005), is common in sites on the eastern shore of Nova Scotia and occurs in sites all over the Maine–Maritimes Region (e.g., Black, 2004; MacDonald, 1994; Sanger, 1991). The GLR Site Complex represents several millennia of occupation (Erskine, 1972; Laybolt, 1999; Nash and Stewart, 1990; Sanders et al., 2014) in which these various rich resources and important localities were taken advantage of. As such, it probably constituted an important node in trade networks and hosting of regular gatherings as was known to have occurred throughout the Historic Period at the nearby First Nation community of L’sitkuk Bear River ca. 100 km southwest from Gaspereau Lake (Ricker, 1998).

Materials and methods

In this study, I investigated the skill level of potters at Gaspereau Lake through time. Having attempted to recreate Woodland-style pottery, I found the inference that coarser pastes meant lower skill to be untenable, as the coarser the paste I tried to work with, the harder became the shaping and forming processes. I hypothesized that the change to coarser pastes was not an accidental degrading of skill but rather a response to historical events. I therefore examined forming attributes to evaluate skill evident in motor habits (Costin, 1991: 33–36) and tempering attributes to determine how temper changes relate to other attributes through time.

If temper and paddling practices changed through time, there should be a statistically significant difference in forming attributes (such as coil break patterns) depending on chronological markers (such as decorative types). Further, if paddling and tempering practices changed because of decreased skill, then pottery amounts should go down through time because novices are unable to perform pottery-making tasks as quickly as more experienced potters (Crown, 2014: 72; also, see Yamamoto and Fujiyami 2008 for an analysis of clay kneading skill acquisition). Additionally, temper and paddling attributes should show signs of decreased knowledge and motor skills (Kamp, 2001: 430). Decreased motor skills would include more asymmetrical forms, increased variation in wall thickness, and smaller vessels (since larger vessels take more skill to build). On the other hand, increased skill and knowledge could be indicated by a larger number of vessel classes or sizes, attributes that would make parts of the production sequence more difficult, or increased homogeneity or consistency across some attributes as potters develop knowledge of what works and stick with it, passing it down through learning lineages. Finally, if increased skill and increased numbers are evident, the cause should be apparent by examining what value is added by the changes in the ceramics through time and what value has been compromised (Skibo and Schiffer, 2001: 146).

Ten dates were acquired on ceramics from the GLR ceramic assemblage, covering a span of ca. 800 years from ca. 1550 BP to ca. 700 BP, with the majority of dates clustering between 1550 and 1050 BP. Three periods were defined based on clusters of dates between 1550 and 1150 BP; the last two dates contain too little context to define or characterize periods. Within the three defined periods, temper percentages, lamellar character, and frequency of coil breaks change significantly.

Statistical tests of association and significance were performed to investigate the association among coil breaks, lamellar character, and temper percentages through time. Chi-square tests among categories were conducted to discover likelihood of association. Student’s t tests on independent means (two-tailed) were used to test whether differences existed between classes. A Pearson correlation co-efficient was calculated to test the strength of association among some attributes that could be measured numerically. A one-way analysis of variance (ANOVA) test was performed to measure variance between groups and to test whether this variance is statistically significant. Instances of missing data were dealt with by removing the entire specimen from the test; this meant that each test dealt with different sample sizes depending on whether the records contained data in the fields being tested. A significance level of p ≤ 0.05 was considered statistically significant, although lower p values close to this level were also considered in some cases. Analysis was performed using SPSS.

Chronological categories were constructed using two kinds of information: (1) decorative types that have been observed to have chronological significance (see above, also Petersen and Sanger, 1991) and (2) classes of GLR ceramics that were accelerator mass spectrometry (AMS) dated. Decorative types have been divided into ordinal chronological categories (relative positions in time are known but numerical, or date range, values are only partially known) and are summarized in Table 1. AMS-dated categories are summarized in Table 2 and consist of categories of ceramics delineated by manufacturing attributes with associated AMS dates. Statistical tests used one or the other kind of chronological information (Figure 3).

Table 1.

Ordinal chronological categories based on decorative type.

Decoration name	Fabric impression	Pseudo-scallop shell	Dentate	Cord-wrapped stick
Elements of decorations
Period name	Early Woodland	Earlier Middle Woodland	Middle Woodland to Late Woodland	Late Woodland
Date range	3000?–2200? BP	2200?–1650? BP	2200?–1000? BP	1350?–500 BP

Fabric impression is a theoretical category coming before PSS but cannot be used in this analysis because there are so few specimens in this category and because fabric impression has also been noted in the Late Woodland.

Table 2.

Periods developed using the AMS sequence and manufacturing and decorative attributes.

Pre-Period 1	Period 1	Period 2	Period 3
?	1550–1350 CalBP	1400–1300 CalBP	1300–1150 CalBP
PSS and dentate decorations, thin walls and few coil breaks	Dentate decorations and early cord-marked pottery; poorly developed coil breaks and some well-developed coil breaks	Cord-marked pottery with increased grit temper and increased iron oxide; both rough and pronounced coil breaks	Cord-marked pottery with increased organic temper and less temper overall; pronounced coil breaks

CalBP: calibrated years before the present.

Figure 3.

AMS sequence for ten vessel lots from the GLR assemblage. Black blocks represent the 1 Sigma range. White blocks represent the 2 Sigma range. Three periods, from 1550 to 1150 BP, have been defined based on clustering in the sequence.

The sample

One hundred eighty-one vessel lots were defined using sherds from the two greatest concentrations of ceramics on the End of Dyke Site.⁴ Out of a total of 14,601 sherds from the End of Dyke Site, the ceramics so far analyzed represent roughly 20% of the assemblage, and the remainder of the sherds theoretically could belong to over 700 more vessel lots.

Vessel lots were divisible into tentative chronological categories based on their decorative tools. The chronological significance of decorative tool types has been noted by Petersen and Sanger (1991) and other researchers (e.g., Allen, 1981, 2005; Bourgeois, 1999; Bourque, 1995; Deal, 1986; Foulkes, 1981; Kristmanson, 1992; Nash, 1977; Nash and Stewart, 1986, 1990; Petersen, 1997; Sheldon, 1988) who have repeatedly made the following observations: (1) fabric impression is sometimes indicative of manufacture during the Early Woodland Period; (2) PSS decorations are usually indicative of manufacture during the earlier Middle Woodland (2150–1650 BP); (3) dentate decorations occur during the period spanning the Middle (2150–1350 BP) to the Late Woodland (1350–500 BP) (but become increasingly rare after the Middle Woodland); and (4) cord marks, also called cord-wrapped stick (CWS), usually indicate manufacture during the Late Woodland. These categories are constructed with the caveat that a single vessel lot can tell little about chronological context, but that, taken along with the whole assemblage and calibrated using AMS dating on carbonized encrustations, decorations can be used to indicate likely chronological significance.⁵ The distribution of the sample by decorative type is shown in Table 3.

Table 3.

Distribution of decorative types in the assemblage.

	Fabric impression	PSS	Dentate	CWS	Other/none	Total
Vessels	3	22	43	97	16	181
Sherds	23	217	430	1867	148	2685

Other/none refers to undecorated ceramics or ceramics with a decoration that is less chronologically sensitive.

PSS: pseudo-scallop shell; CWS: cord-wrapped stick.

As Table 3 indicates, the majority of vessel lots are decorated with cord marks, while the smallest decorative category is fabric impression. I interpret this to mean that the Late Woodland is most strongly represented in the assemblage, while the Early Woodland Period is least represented. This is consistent with many other sites in the Maine–Maritimes Region, except that the large number of vessel lots decorated with cord marks is significantly more pronounced than in other assemblages. The distribution across decorative categories suggests that ceramics increased in number through time and, further, that a large number of the cord-marked ceramics are associated with the earlier part of the Late Woodland (1350–950 BP) as indicated by an AMS sequence of 10 dates (Figure 4).

Figure 4.

Chronological categories used in statistical analysis.

Results

Statistical analysis of attributes showed changes in manufacturing practices through a period of ca. 500 years at the transition to the Late Woodland Period. The GLR ceramics were built predominantly using the coiling and paddling method. Evidence for this comes from frequent occurrences of coil breaks, occasional anvil marks on interior surfaces, lamellar wall character, and lamellar splitting (Rye, 1981: 84–85). Ninety-seven vessel lots in the sample have coil breaks on at least one sherd, whereas 84 vessel lots have no reported coil breaks on any sherds. This ratio represents roughly equal numbers of vessels with and without coil breaks (54% and 46%, respectively). Lamellae are ubiquitous with only a few specimens having no or minimal lamellar character. Table 4 shows the distribution of coil breaks across decorative categories for vessel lots for which this information exists.

Table 4.

Distribution of coil breaks across decorative categories. Note that this distribution excludes vessel lots that do not fall into one of the decorative categories.

	PSS	Dentate	Cord marks	Total
Coil breaks	5	14	72	91
No coil breaks	16	35	30	81
Total	21	49	102	172

PSS: pseudo-scallop shell.

Two chi-square tests show that paddling decreased through time from the Middle to the Late Woodland Period. A chi-square test found a high probability that the presence of coil breaks depends on the decorative class to which the vessel lot belongs (χ² = 31.6, p < 0.001, n = 172). Table 4 shows that a much larger number of cord-marked vessel lots exhibit coil breaks than lack them, while for both dentate- and PSS-decorated vessels, coil breaks are much less frequent.⁶ Similarly, a test of lamellar character and decorative class showed that lamellar orientation is likely to be dependent on decorative type (χ² = 24.6767, p < 0.001, n = 104; Table 5). Vertical lamellae are more likely to occur earlier in the sequence (associated with PSS), while oblique and U-shaped lamellae occur later (associated with CWS). This is consistent with what other researchers have observed in other assemblages from the Maine–Maritimes Region.

Table 5.

Distribution of decorative categories by lamellar character.

	PSS	Dentate	Cord marks	Totals
Vertical	6	19	11	36
Oblique	2	10	43	56
U-shaped	1	1	11	13
Total	9	30	65	104

PSS: pseudo-scallop shell.

Temper percentages

Four combinations of temper types were identified. Grit temper, composed of crushed granite, was ubiquitous with only a few exceptions. Combined with grit temper, in various amounts and combinations, was iron oxide and organic temper. The iron oxide was a particulate of a poor grade of the type usually used for paint stones. It may have been the product of “hematization” that was reported to be occurring at Gaspereau Lake as pegamtitic feldspars weathered to alteration products including iron-rich minerals such as hematite (MacDonald and Ham, 1992; O’Reilly et al., 1982: 64). Organic material in the pre-fired paste was evident from pore spaces, which were usually angular but sometimes tubular, probably indicating combinations of grasses and other organic material such as cattail fluff. Four categories were found to be chronologically significant (Figure 6): (1) no organic or iron oxide (grit only); (2) iron oxide and grit; (3) organic and grit; and (4) iron oxide, organic, and grit.

An examination of temper percentages through time shows that temper increased in association with CWS decorations (Table 6). A chi-square test of paste coarseness (based on percentage of temper and size of particles) compared with decorative type resulted in a strong likelihood that paste texture is related to decorative type (χ² = 20.1106, p = 0.001, n = 96). The results of the test need to be viewed with caution because of the small numbers in some categories, and so a one-way ANOVA test of temper percentages for the three decorative types was conducted. ANOVA, or analysis of variance, looks at the amount of variance between three or more groups by comparing the means of each group to determine whether they are statistically different from each other. The analysis showed that the difference among these groups is statistically significant (f = 5.45, p = 0.005, n = 140). A one-way ANOVA was also conducted on temper percents for the AMS-defined periods, resulting in even greater variance (f = 11.54, p < 0.0001, n = 91). Table 6 shows that temper amounts increased in association with cord-marked pottery, although significant variability is also associated with this category (Figure 5). As a result, a Pearson correlation test showed only a minimal correlation (R = 0.186, R² = 0.0346, n = 65). This increased variability associated with cord-mark decorations is clarified somewhat by the AMS sequence, and shows that, after 1350 BP, mean temper percentage went down while temper type variability went up (Figure 6). The mean temper percent in this period is affected by vessel lots with organic temper, which tend to have smaller amounts of temper (ca. 10%). Mean temper percentage reached its peak between 1450 and 1300 BP.

Table 6.

Distribution of paste textures compared with decoration tools.

	Fine	Medium–fine	Medium	Medium–coarse	Coarse	Total
CWS	11	4	20	7	17	59
Dentate	9	3	2	4	1	19
PSS	6	0	6	0	1	13
Total	26	7	28	11	19	91

PSS: pseudo-scallop shell; CWS: cord-wrapped stick.

Figure 5.

Distribution of temper percentages in the three decorative groups. CWS: cord-wrapped stick.

Figure 6.

Temper variability in three periods.

Tempering practices followed a trajectory of increasing amounts from sometime as early as 2200 years ago to ca. 1300 years ago, after which temper percentages dropped and diversity of tempering materials increased for a period for ca. 100 years (Figure 7). However, many pots during the last period are coarsely tempered (>40%). Therefore, it appears that coarsely tempered pottery was still important during this period but was one of a variety of paste types used.

Figure 7.

Mean temper percentages of six chronological categories constructed from decorative types and AMS-dated ceramic classes.

Decreased paddling

In contrast with the variable temper percents through time, paddling appears to have steadily decreased from Period 1 to Period 3. Tables 7 and 8 show that, from Period 1 to Period 3, lamellar character moved from vertical to oblique and U-shaped, while the degree of coil joining moved from well joined (no breaks) to moderately joined (rough coil breaks) to poorly joined (smooth coil breaks). The fact that a gradation of coil break and lamellar attributes is observable through time indicates that the shift was not an abrupt break with a previous tradition. Rather, the shift was gradual spanning several hundred years. This indicates that the manufacturing tradition evolved in situ as a response to changing pottery requirements. The connection with an earlier manufacturing practice—one in which pots were extensively paddled—is evident in a minority of vessel lots from later periods that are well paddled, thin, and carefully decorated.

Table 7.

Distribution of coil break types among the three periods defined by AMS date clusters.

Period	Period 1		Period 2		Period 3		Total
Years (CalBP)	1550–1400		1400–1300		1300–1150
Number (percentage)	n	%	n	%	n	%	n	%
Coil break	37	100	14	44	12	37	63	62
None
Type
Rough	0	0	9	28	5	16	14	14
Smooth	0	0	9	28	15	47	24	24
Total	37	100	32	100	32	100	101	100

Coil break types are assigned to vessel lots on the basis of whether coil breaks were visible during analysis, and whether they were poorly developed (rough) or well developed (smooth).

CalBP: calibrated years before the present.

Table 8.

Distribution of lamellar orientation across the time periods that were defined by clustering of dates in the AMS sequence.

Period	Period 1		Period 2		Period 3		Total
Years (CalBP)	1550–1400		1400–1300		1300–1150
Number (percentage)	n	%	n	%	n	%	N	%
Unknown (not recorded)	15	41	13	40	16	50	175	22
Vertical	14	38	4	14	1	3	19	26
Vertical and oblique	5	13	3	9	1	3	9	12
Oblique	3	8	9	28	8	25	20	27
Oblique and U-shaped	0	0	2	6	5	16	7	10
U-shaped	0	0	1	3	1	3	2	3
Total	37	100	32	100	32	100	232	100

Note that Period 1 is artificially large because it contains all examples of PSS decorations that likely came before Period 1 (no earlier than 2100 CalBP) (Petersen and Sanger, 1991).

CalBP: calibrated years before the present; PSS: pseudo-scallop shell.

Periods 2 and 3 show differential responses to expediency requirements. Although temper percentages clearly increased in both the second and third periods over the first, large amounts of temper were not as prevalent in Period 3 as in Period 2. In contrast, paddling steadily declined from Period 1 to Period 3. The drop-off in paddling likely began as a response to increased grit temper but continued to decline as a strategy to increase output.

Evidence of motor skills

The skill of the potters at the transition to the Late Woodland is evident in several ways. Although pottery is less thoroughly paddled after 1550 BP, pottery continued to show even wall thickness, even though walls are thicker overall later in time. A one-way ANOVA of the means of standard deviations of neck thicknesses divided by decorative type did not show a statistically significant difference even though the neck thicknesses were different across time periods (f = 0.39, p = 0.68, n = 115; Table 9).

Table 9.

Comparison of means of standard deviations of neck thicknesses within vessel lots divided by decorative types.

	PSS	Dentate	CWS
Mean	0.0825	0.097	0.0917
SD	0.0422	0.0544	0.0484
n	12	33	70

PSS: pseudo-scallop shell; CWS: cord-wrapped stick; SD: standard deviation.

Coils were joined together using scoring marks, which indicates knowledge about pottery manufacture beyond the basics. Furthermore, though coil breaks show weaknesses in the pottery wall, lamellar directions and flanges along coil breaks tend to be consistent across vessel lots, indicating that the potters consistently used the same smoothing actions using the thumb and fingers. This consistency would have been due to the development of motor skills; it is lacking in experimental vessels that I have built, as I have not developed the same degree of motor habit. Another way in which skill is apparent is that the channeling on the interiors of vessels is almost invariably evenly deep and often the gestures can be recognized across vessels, probably indicating multiple works by the same individual. Another indication of skill is the lack of cracks around the rim. These develop as pots dry out too quickly during forming and their absence indicates two things: that potters were able to make pots very quickly and that potters were able to paddle even very coarsely tempered pots with enough skill and gentleness to avoid these neck cracks that have invariably been a problem for me while making experimental pots.

The shift in pottery practices from before 1550 BP to 1150 BP shows a move from highly elaborate and expressive pottery to more standardized and slightly less durably built pottery and finally to more expedient and variable pottery types that were made for accomplishing different goals. The shift shows that potters not only were moving increasingly towards expedient strategies in pottery manufacture, but they were also diversifying their skillsets. During Period 3, some pottery was made with large amounts of grit temper, making paddling much harder, while other pottery was made with organic temper, which was probably important for the building of larger vessels. Some vessels were made with large amounts of grit temper but crushed to a much smaller particle size overall, and these pots tend to be the largest,⁷ averaging a neck diameter of 34 cm, as well as some of the most coarsely tempered, averaging 40%. Organic-tempered pots tend to be similarly large. Additionally, the percentages of various kinds of temper vary considerably. This mixing-and-matching strategy suggests that potters were using different kinds of temper to add specific characteristics to their pots depending on what they desired at the time of manufacture. This is different from earlier pottery that exhibits different percentages of temper but mostly the same temper type (grit) with small amounts of iron oxide. This is also different from other sites in New Brunswick and Nova Scotia, where temper tends to fall into types (grit, shell, and organic) much easier. The increasing variability suggests that potters were capable not only of handling different kinds of temper but also knew how to take advantage of different kinds of temper to accomplish different goals. Whereas potters in other parts of the region knew how to increase production with coarser pastes, I suggest that potters at Gaspereau Lake were particularly skilled in handling different paste types as a result of the large gatherings that probably occurred at Gaspereau Lake owing to the rich resources in the area.

Discussion

During the Late Woodland Period, temper percents increased overall, with some vessel lots exhibiting a remarkable amount of temper—sometimes more than half the percentage of the overall paste. It is this increase that explains the crumbliness, which in turn accounts for the increase in coil breaks and thicker walls of Late Woodland pottery. Increased temper would have required significant skill to work with since it is accompanied by decreased plasticity and the likelihood of pots drying out too quickly during forming increases (Herbert, 2008: 278). This raises an important and unexplored question: what was the impetus for increasing temper percentages given the significant trade-off of poor workability and the fact that earlier pottery appears to have been a more-than-satisfactory product? The best explanation for the change is manufacturing expediency: large amounts of temper increase the thermal shock resistance of pottery, especially during the initial firing, allowing potters to “cut corners” on drying time and to turn out more pots faster.

Because the shift to increased grit temper would have required that potters develop an expanded skillset, the shift suggests that incentives for learning new skills came from outside the realm of pottery manufacture. Determining what the incentives would have been requires examining the specific benefits and drawbacks of the trade-offs to pottery makers and consumers. In the discussion below, I present an argument that the changes in ceramic attributes through time resulted from an increased desire for expedient manufacture during the Late Woodland because of larger-scale gatherings.

Increased temper and firing regimes

The increase in temper percent would have served an expedient role by making firings more successful and wasters fewer in number. Although increased temper most obviously increases thermal shock resistance during use, an even greater advantage would be that vessels would be more likely to survive firing if they were coarsely tempered (Gosselain, 1992: 257). A third advantage is shorter drying times for grit- and organic-tempered pottery. These two tempering strategies—high percentage of grit temper and organic matter as temper—are known to confer the advantage of faster production, particularly by reducing the necessary drying period, so reduction in drying time was likely a high priority. Reduced drying times and lowered risk during firing are strategies for greater production efficiency, attributes that are found in contexts where production is intensifying (Costin, 1991: 17, 37–39; 2001: 280) or where very short turnaround time is required (Skibo, 2013: 41–43; Skibo et al., 1989). Both these concerns arise from preparing for large-scale gatherings.

Although some period of drying is required before pots are fired, the large percentage of grit temper or the use of organic temper would cut down on the strict necessity that pots be bone-dry. Normally, any water left in the vessel wall is heated to steam and released, sometimes explosively, from pockets within the pot wall (Rice, 2005; Rye, 1981: 24). Pore spaces may not be large or numerous enough to allow the steam to escape before it blasts its way through, but coarsely tempered vessels have many more pockets and much thinner solid clay walls to go through, so steam can escape more gently.

The pottery-production schedule can last quite long if weather does not cooperate to make pots as dry as possible (Arnold, 1985: 211–212; 2008: xxv), and even in optimal conditions, much time and care (shaded structures, pots covered and periodically turned) are required to allow pots slow drying time. Eliminating or reducing this time and effort could potentially save a potter weeks of drying time. It would also increase the number of firings possible, which would increase the number of pots that could be made. The combined reduction of wasted materials and increase in time available, allowing for more firings less dependent on weather, would be attractive to potters under pressure to manufacture more pots in shorter times.

Decreased wall compression and abrasion resistance

The trend of increasing temper would have had significant drawbacks for the longevity of the pottery. The pots are more likely to abrade and fracture; thus, they wear out faster. These drawbacks would mainly affect the end user of pottery over time, as compared to the long use lives of these pots’ Middle Woodland counterparts (evidenced in other assemblages by plentiful carbonized encrustations, use-related striations on interiors and abrasions on bases). Potters have been observed in ethnographic situations to be under pressure to ensure that their pots last in cases where pottery producers sell their products at markets or fairs and where reputation and return business are important in earning a living (Wayessa, 2011: 321). It stands to reason, then, that in situations where no market exists and where pottery is produced for a specific event or time of year (such as the runs of certain fish), these concerns would not be as influential in potters’ decision-making (e.g., Frink and Harry, 2008).

Adaptation to changes in social activities

Some trends seen in the GLR assemblage have been noted in other assemblages, including increased grit temper percentages and larger numbers of coil breaks later in time. This indicates that potters at Gaspereau Lake were responding to the same pressures and incentives that were being felt across the region. Pottery found at other sites therefore also shows signs of attempting to increase output and decrease time spent. Because this change has been noted at sites across the region, it cannot be thought of as an idiosyncratic pottery tradition that developed by accident for unfathomable reasons. Rather, it must be understood as a region-wide adaptation whose trade-offs—decreased workability and impaired abrasion resistance—did not deter potters from these practices.

Pottery from the GLR assemblage shows a marked lack of use wear, despite the large numbers of pots: only 10 yielded significant carbonized encrustations and almost none had use-related abrasion. Coupled with the large number of ceramics produced later in time at Gaspereau Lake, these attributes suggest that potters at the transition to the Late Woodland were preparing large amounts of pottery for consumers (who may or may not have been from the same group as the producers) during special events. The pottery was optimized for such events—in particular, for the volume demands placed on the producers—and not for long-term daily use.

The GLR assemblage reflects the changes taking place at the transition to the Late Woodland Period, a time of increasing trade (Bourque, 1994, 2001; Leonard, 1996: ii–iii, 45), interaction with groups to the north, south, and west (Wolf, 2013: 13; Wright, 1994), and intensification of resources. Trade was an important part of the activities conducted at Gaspereau Lake, evident from the large amounts of high-quality lithics, wealth objects such as copper, and numerous iron oxide particles and paint stones (Sanders et al., 2014). Trade on the scale evident at Gaspereau Lake was likely accompanied or initiated by large-scale gatherings. The increase in pottery manufacture at the transition to the Late Woodland and the strategies for increasing output under tight deadlines support the evidence for trade-oriented gatherings at Gaspereau Lake.

The Gaspereau River headwaters were a logical place for large groups to gather because the seasonal runs of fish would have been able to support many people and would have been a draw for people from all around. These seasonal runs may have been the focal point for aggrandizing events, weddings, political meetings, and so on and likely served this role since the first times that people have inhabited the Maine–Maritimes Region. However, the increase in trade goods across the region during the Late Woodland Period and the significant increase in pottery at Gaspereau Lake at the transition to the Late Woodland suggest that gatherings increased in number, size, or both after 1550 BP. Pottery was probably always important in supporting these events, as cooking pots for preparing large-scale meals, as gifts to newlyweds and dignitaries, and as funerary offerings (Claassen, 2002: 532; Gosselain, 2015: 55; Hayden, 1995, 2001). Demand for pots may also have increased at the transition to the Late Woodland because of their use for fish processing and oil rendering. In this case, not only would pottery have to be made quickly and in large numbers, but the pots made for these events would not have needed to last past the duration of the event.

Conclusion

Paddling and tempering practices during the Woodland Period in the Maine–Maritimes Region show a shift through time in potters’ priorities. While the fully understood picture of Woodland society is still far away, the GLR ceramic assemblage shows that an important part of the picture is a move towards expediency and increased production at the transition to the Late Woodland Period. This move towards expediency is observable cross-regionally, evidenced by an increase in coil breaks and higher temper percents in many sites, but the shift appears to be particularly pronounced at Gaspereau Lake, considering the large amount of pottery that was manufactured and the various kinds of tempering strategies that are discernible. Far from indicating a region-wide decrease in pottery manufacturing skill, Late Woodland ceramics—as revealed in higher resolution by the GLR assemblage than has been achieved previously—were manufactured with less concern for highly intricate decorations and thin, hard walls and with more concern for increasing output. This was accomplished by “cutting corners,” an ability belonging only to potters who have the knowledge base to understand what can be traded off or tweaked to better meet specific goals. Late Woodland potters were willing to accept poor workability (or the necessity to increase their forming and paddling skills) in order to make more coarsely tempered pots that would withstand the initial firing to a better extent and allow them to make pots with a shorter drying time. This indicates that at least one of their goals was faster turnaround within tighter deadlines.

Footnotes

Data Availability Statement

All data acquired on the Gaspereau Lake Reservoir Ceramic Assemblage is stored in a database that can be obtained by contacting the author or the Nova Scotia Museum of Natural History.

Acknowledgments

I am grateful for the following people and organizations for their efforts on my and my work’s behalf: Aubrey Cannon gave in-depth critiques and suggestions on the work presented here. Mike Deal, Andy Roddick, James Skibo, Katie Cottreau-Robbins, and David Black also contributed important insights and encouragement. I am indebted to Matte Robinson for his careful read-throughs and commentaries. The University of New Brunswick contributed laboratory space and facilitated the transport and storage of artifacts during this research. No permit was required for conducting this research.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The work was funded by the Social Sciences and Humanities Council in the form of a Vanier Canada Research Scholarship and by the Nova Scotia Museum of Natural History’s Archaeology Research Grant. Radiocarbon dating funding was partially contributed by the Archaeological Services Branch of the Province of New Brunswick.

Notes

Author Biography

Cora A Woolsey received her PhD in Anthropology from McMaster University. Her current position is postdoctoral fellow at the University of New Brunswick. Her research interests include Precontact ceramics, technological traditions and design principles of ancient and modern cultures.

References

Allen

(1981) The Oxbow Site: Chronology and prehistory in northeastern New Brunswick. New Brunswick Manuscripts in Archaeology 1. Fredericton, New Brunswick: Archaeological Branch Historical Resources Administration.

Allen

(2005) The Oxbow Site 1984: Metepenagiag Mi’kmaq First Nation, Miramichi, New Brunswick. New Brunswick Manuscripts in Archaeology 39. Fredericton, New Brunswick: Archaeological Services, Heritage Branch, Province of New Brunswick.

Arnold

(1985) Ceramic Ecology and Cultural Process. Cambridge: Cambridge University.

Arnold

(2008) Social Change and the Evolution of Ceramic Production and Distribution in a Maya Community. Denver: University Press of Colorado.

Black

(2004) Ponapsqey. Stone materials. In: Blair S (ed) Wolastoqiyik Ajemseg: The People of the Beautiful River at Jemseg, Volume 2: Archaeological Results, Jemseg Crossing Archaeology Project. New Brunswick Manuscripts in Archaeology 36E. Fredericton, New Brunswick: Archaeological Services, Heritage Branch, Province of New Brunswick, pp. 91–116.

Bourgeois

(1999) A regional pre-contact sequence for the Saint John River Valley. MA Thesis, University of New Brunswick, Canada.

Bourque

(1994) Evidence for prehistoric exchange on the Maritime Peninsula. In: Baugh T and Ericson J (eds) Prehistoric Exchange Systems in North America. New York: Springer, pp. 23–46.

Bourque

(1995) Diversity and Complexity in Prehistoric Maritime Societies: A Gulf of Maine Perspective. New York: Plenum.

Bourque

(2001) Twelve Thousand Years: American Indians in Maine. Lincoln: University of Nebraska.

10.

Braun

(1986) Pots as tools. In: Moore J and Keene A (eds) Archaeological Hammers and Theories. New York: Academic, pp. 108–134.

11.

Bronitsky

Hamer

(1986) Experimental ceramic technology: The effects of various tempering materials on impact and thermal-shock resistance. American Antiquity 51(1): 89–101.

12.

Budden

Sofaer

(2009) Non-discursive knowledge and the construction of identity: Potters, potting and performance at the Bronze Age Tell of Szàzhalombatta, Hungary. Cambridge Archaeological Journal 19: 203–220.

13.

Chilton

(1998) The cultural origins of technical choice: Unraveling Iroquoian and Algonkian ceramic traditions in the Northeast. In: Stark M (ed) The Archaeologies of Social Boundaries. Washington: Smithsonian, pp. 132–160.

14.

Claassen

(2002) Mothers’ workload and children’s labour during the Woodland Period. In: Nelson S and Rosen-Ayalon M (ed) Pursuit of Gender. Gender and Archaeology Series, Vol. 1. Walnut Creek: Altamira, pp. 225–234.

15.

Costin

(1991) Craft specialization: Issues in defining, documenting, and explaining the organization of production. Journal of Archaeological Method and Theory 3: 1–56.

16.

Costin

(2001) Craft production systems. In: Feinman G and Price D (eds) Archaeology at the Millenium: A Sourcebook. New York: Kluwer, pp. 273–327.

17.

Costin

(2011) Hybrid objects, hybrid social identities: Style and social structure in the Late Horizon Andes. In: Amundsen-Meyer L, Engel N and Pickering S (eds) Identity Crisis: Archaeological Perspectives on Social Identity. Proceedings of the 42nd (2010) Annual Chacmool Archaeology Conference. Calgary, Alberta: University of Calgary, pp. 211–225.

18.

Crown

(2007) Learning About Learning. In: Skibo J, Graves M and Stark M (eds) Archaeological Anthropology: Perspectives on Method and Theory. Tucson: University of Arizona, pp. 198–217.

19.

Crown P (2014) The archaeology of crafts learning: becoming a potter in the puebloan southwest. Annual Review in Anthropology 73: 71–88.

20.

Davis

(1991) The Ceramic Period of Nova Scotia. In: Deal M and Blair S (eds) Prehistoric Archaeology in the Maritime Provinces: Past and Present Research. Reports in Archaeology No. 8. Fredericton, New Brunswick: Council of Maritime Premiers, pp. 85–100.

21.

Deal

(1986) Late archaic and ceramic period utilization of the Mud Lake Stream Site, Southwestern New Brunswick. Man in the Northeast 32: 67–94.

22.

Deal

(1988) Western Minas Basin Project 1988: Preliminary Report. Halifax: Department of Education and the Nova Scotia Museum.

23.

Deal

(1998) Pottery Ethnoarchaeology in the Central Maya Highlands. Salt Lake City: University of Utah.

24.

Deal

(2005) Vignette: Distribution and utilization of Scots Bay Chalcedony. Available at: http://www.ucs.mun.ca/∼mdeal/Anth3291/DavidsonCove.htm (accessed on 06 December 2015).

25.

Dickinson

(2001) Petrography and geologic provenance of sand tempers in prehistoric potsherds from Fiji and Vanuatu, South Pacific. Geoarchaeology: An International Journal 16(3): 275–322.

26.

Dobres

(2001) Meaning in the making: Agency and the social embodiment of technology and art. In: Schiffer

(ed) Anthropological Perspectives on Technology. Albuquerque: University of New Mexico, pp. 47–76.

27.

Druc

(2013) What is local? Looking at ceramic production in the Peruvian Highlands and beyond. Journal of Anthropological Research 69: 485–513.

28.

Erskine

(1972) The archaeology of some Nova Scotian Indian Campsites. Proceedings of the Nova Scotia Institute of Science 27: 1–10.

29.

Feathers

(2006) Explaining shell-tempered pottery in prehistoric Eastern North America. Journal of Archaeological Method and Theory 13(2): 90–133.

30.

Feathers

(2008) Origins and spread of shell-tempered ceramics in the Eastern Woodlands: Conceptual and methodological frameworks for analysis. Southeastern Archaeology 27(2): 286–293.

31.

Feathers

(2009) Problems of ceramic chronology in the southeast: Does shell-tempered pottery appear earlier than we think? American Antiquity 74(1): 113–142.

32.

Foulkes

(1981) Fulton island: A stratified site in the Saint John River Valley of New Brunswick. MA Thesis, Trent University, Canada.

33.

Frink

Harry

(2008) The beauty of “ugly” Eskimo cooking pots. American Antiquity 37(1): 103–118.

34.

Garcia

(1989) Surface treatment and strength of low-fired ceramic bodies: An experimental study . MA Thesis, University of Arizona, USA.

35.

Gilbert

Gamblin

Black

(2006) The usual suspects: Exotic toolstones in Quoddy region archaeological assemblages. Available at: http://www.unb.ca/fredericton/arts/departments/anthropology/pdfs/dwblack/exoticlithics.pdf (accessed 06 December 2015).

36.

Gosselain

(1992) Bonfire of the enquiries. Pottery firing temperatures in archaeology: What for? Journal of Archaeological Science 19: 243–259.

37.

Gosselain

(2000) Materializing identities: An African study. Journal of Archaeological Method and Theory 7(3): 187–217.

38.

Gosselain

(2015) Roads, markets, migrants. The historical trajectory of a male Hausa pottery tradition in Southern Niger. In: Gauß W, Klebinder-Gauß G and von Rüden C (eds) The Distribution of Technological Knowledge in the Production of Ancient Mediterranean Pottery. Proceedings of the International Conference at the Austrian Archaeological Institute at Athens 23rd–25th November 2012. Vienna: Österreichisches Archäologisches Institut, pp. 277–296.

39.

Gosselain

(2016) The world is like a beanstalk: Historicizing potting practice and social relations in the Niger River area. In: Roddick A and Stahl A (eds) Knowledge in Motion: Constellations of Learning across Time and Place. Tuscon: University of Arizona, pp. 36–66.

40.

Graves

(1985) Ceramic design variation within a Kalinga village: Temporal and spatial processes. In: Nelson

(ed) Decoding Prehistoric Ceramics. Carbondale: Southern Illinois University, pp. 9–34.

41.

Hayden

(1995) The emergence of prestige technologies and pottery. In: Barnett

Hoopes

(eds) The Emergence of Pottery. Washington: Smithsonian, pp. 257–266.

42.

Hayden

(2001) Feasts: Archaeological and Ethnographic Perspectives on Food, Politics, and Power. Washington: Smithsonian Institution.

43.

Herbert

(2008) The history and practice of shell tempering in the Middle Atlantic: A useful balance. Southeastern Archaeology 27(2): 265–285.

44.

Hoard

O’Brien

Khorasgany

et al . (1995) A materials science approach to understanding limestone-tempered pottery from the Midwestern United States. Journal of Archaeological Science 22: 823–832.

45.

Hugill P and Dickson DB (eds) (1988) The Transfer and Transformation of Ideas in Material Culture. College Station: Texas A&M University.

46.

Jeffra

(2008) Hair and potters: An experimental look at temper. World Archaeology 40: 156–161.

47.

Kamp K (2001) Prehistoric children working and playing: a southwestern case study in learning ceramics. Journal of Anthropological Research 57(4): 427–450.

48.

Kilikoglou V, Vekinis G, Maniatis Y, and Day PM (2007) Mechanical performance of quartz-tempered ceramics: part I, strength and toughness. Archaeometery 40(2): 261–179.

49.

Kooiman

(2016) Woodland pottery function, cooking, and diet in the Upper Great Lakes of North America. Midcontinental Journal of Archaeology 41(3): 207–230.

50.

Kristmanson

(1992) The ceramic sequence for Southwestern Nova Scotia: A refinement of the Petersen/Sanger model . MA Thesis, Department of Anthropology, Memorial University of Newfoundland, Canada.

51.

Laybolt

(1999) Prehistoric settlement and subsistence patterns at Gaspereau Lake, Kings County, Nova Scotia . MA Thesis, Memorial University of Newfoundland, Canada.

52.

Leonard

(1996) Mi’kmaq culture during the Late Woodland and early historic periods . PhD Thesis, University of Toronto, Canada.

53.

Longacre

Xia

Yang

(2000) I want to buy a black pot. Journal of Archaeological Method and Theory 7(4): 273–293.

54.

MacDonald

(1994) Exploring patterns of prehistoric lithic material use in the insular Quoddy Region, Charlotte County, New Brunswick . MA Thesis, University of New Brunswick, USA.

55.

MacDonald

Ham

(1992) Geological map of Gaspereau Lake (NTS Sheet 21A/15 and Part of 21H/02), Nova Scotia, Scale 1:50 000. Map ME 1992-1. Halifax: Department of Natural Resources, Mineral Resources Branch.

56.

MacIntyre

(1988) Appendix H: Results of petrographc analysis. In: Sheldon H (ed) The Late Prehistory as Viewed from the Brown Site. Nova Scotia Museum Curatorial Report 61. Halifax: Nova Scotia Museum, pp. 307–329.

57.

Matson

(1965) Ceramics and Man. Chicago: Aldine.

58.

Michelaki

(2008) Making pots and potters in the Bronze Age Maros Villages of Kiszombor-Új-Élet and Klárafalva-Hajdova. Cambridge Archaeological Journal 18(3): 355–380.

59.

Miller

(1985) Artefacts as Categories: A Study of Ceramic Variability in Central India. Cambridge: Cambridge University.

60.

Myers

(2006) Hominy technology and the emergence of Mississippian societies. In: Staller

Tykot

Benz

(eds) Histories of Maize: Multidisciplinary Approaches to the Prehistory, Linguistics, Biogeography, Domestication and Evolution of Maize. New York, NY: Academic, pp. 511–520.

61.

Nash

(1977) Prehistoric and cultural ecology: Cape Breton Island, Nova Scotia. In: Preston R (ed) Papers from the Fourth Annual Congress. Mercury Series Paper No. 40. Ottawa: National Museum of Man, pp. 131–156.

62.

Nash

Stewart

(1986) Mi’kmaq: Economics and Evolution/Faunal Remains from the Delorey Island Site (BjCj-9) of Nova Scotia. Curatorial Report Number 57. Halifax, Nova Scotia Museum.

63.

Nash

Stewart

(1990) Melanson: A Large Micmac Village in Kings County, Nova Scotia. Curatorial Report Number 67. Halifax: Nova Scotia Museum.

64.

Nash

Stewart

Deal

(1991) Melanson: A central place in Southwestern Nova Scotia. In: Deal M and Blair S (eds) Prehistoric Archaeology in the Maritime Provinces: Past and Present Research. Reports in Archaeology No. 8. Fredericton, New Brunswick: Council of Maritime Premiers, pp. 213–220.

65.

Neupert

(2007) Contingency theory and the organizational behavior of traditional pottery production. In: Skibo

Graves

Stark

(eds) Archaeological Anthropology: Perspectives on Method and Theory. Tucson: University of Arizona, pp. 139–162.

66.

O’Brien

Holland

Hoard

et al . (1994) Evolutionary implications of design and performance characteristics of prehistoric pottery. Journal of Archaeological Method and Theory 1(3): 259–304.

67.

O’Reilly

Farley

Charest

(1982) Metasomatic-Hydrothermal Mineral Deposits of the New Ross, Mahone Bay Area, Nova Scotia. Paper 82-2. Halifax: Nova Scotia Department of Natural Resources.

68.

Osborn

(1988) Limitations of the diffusionist approach: Evolutionary ecology and shell-tempered ceramics. In: Hugill

Dickson

(eds) The Transfer and Transformation of Ideas in Material Culture. College Station: Texas A&M University, pp. 23–44.

69.

Owen

Forfa

Greenough

(2014) Geochemical and mineralogical constraints on the provenance of L’sitkuk Bear River Pottery from the Annapolis Basin, Nova Scotia. Canadian Journal of Archaeology 38: 76–105.

70.

Peacock

Neff

Rafferty

et al . (2007) Using Laser Ablation–Inductively Coupled Plasma–Mass Spectrometry (LA-ICP-MS) to source shell in shell-tempered pottery: A pilot study from North Mississippi. Southeastern Archaeology 26(2): 319–329.

71.

Peirce

(2005) Reverse engineering the ceramic cooking pot. Journal of Archaeological Method and Theory 12(2): 117–157.

72.

Petersen

(1988) The pseudo scallop shell horizon style in North American prehistory. In: Paper Presented at the 20th Annual Meeting of the Canadian Archaeological Association, Whistler, Canada.

73.

Petersen

(1997) A prehistoric Native American ceramic vessel from Lake Champlain. Journal of Vermont Archaeology 2: 85–90.

74.

Petersen

Sanger

(1991) An aboriginal ceramic sequence for Maine and the Maritime Provinces. In: Deal M and Blair S (eds) Prehistoric Archaeology in the Maritime Provinces: Past and Present Research. Reports in Archaeology No. 8. Fredericton, New Brunswick: Council of Maritime Premiers, pp. 113–170.

75.

Rice

(2005) Pottery Analysis: A Sourcebook. Chicago: University of Chicago.

76.

Ricker

(1998) L’sitkuk: The Story of the Bear River Mi’kmaw Community. 2nd ed. Roseway: Lockport.

77.

Roper

Josephs

Beck

(2010) Determining provenance of shell-tempered pottery from the central plains using petrography and oxidation analysis. American Antiquity 75(1): 134–157.

78.

Rutherford

(1991) The ceramic period in New Brunswick. In: Deal M and Blair S (eds) Prehistoric Archaeology in the Maritime Provinces: Past and Present Research. Reports in Archaeology No. 8. Fredericton, New Brunswick: Council of Maritime Premiers, pp. 101–112.

79.

Rye

(1981) Ceramic Technology: Principles and Reconstruction. Washington: Taraxacum.

80.

Sanders

Finnie

Green

et al . (2014) End of Dyke Site Mitigation 2012, Gaspereau Lake Reservoir, Kings County, Nova Scotia: Final Report. Halifax: Cultural Resource Management Group Limited.

81.

Sanger

(1979) The ceramic period in Maine. In: Sanger

(ed) Discovering Maine’s Archaeological Heritage. Augusta: Maine Historic Preservation Commission, pp. 99–115.

82.

Sanger

(1991) Five thousand years of contact between Maine and Nova Scotia. Maine Archaeological Society Bulletin 31(2): 55–61.

83.

Sassaman

(1998) The distribution, timing, and technology of early pottery in the Southeast. Revista de Arqueología Americana 14: 101–134.

84.

Schiffer

Skibo

(1987) Theory and experiment in the study of technological change. Current Anthropology 28(5): 595–622.

85.

Schiffer

Skibo

Boelke

et al . (1994) New perspectives on experimental archaeology: Surface treatments and thermal response of the clay cooking pot. American Antiquity 59(2): 197–217.

86.

Sheldon

(1988) The Late Prehistory as Viewed from the Brown Site. Nova Scotia Museum Curatorial Report 61. Halifax: Nova Scotia Museum.

87.

Shepard

(1956) Ceramics for the Archaeologist. Carnegie Institution of Washington, Publication 609. Washington: Carnegie Institution.

88.

Sillar

Tite

(2000) The challenge of ‘technological choices’ for materials science approaches in archaeology. Archaeometry 42(1): 2–20.

89.

Skibo

(2013) Understanding Pottery Function. Manuals in Archaeological Method, Theory and Technique. New York: Springer.

90.

Skibo

Schiffer

(1995) The clay cooking pot: An exploration of women’s technology. In: Skibo

Walker

Nielsen

(eds) Expanding Archaeology. Salt Lake City: University of Utah, pp. 80–91.

91.

Skibo

Schiffer

(2001) Understanding artifact variability and change: A behavioural framework. In: Schiffer

(ed) Anthropological Perspectives on Technology. Albuquerque: University of New Mexico, pp. 139–150.

92.

Skibo

Schiffer

Reid

(1989) Organic-tempered pottery: An experimental study. American Antiquity 54(1): 122–161.

93.

Stapelfeldt

(2009) A Form and Function Study of Precontact Pottery from Atlantic Canada . MA Thesis, Memorial University of Newfoundland, Canada.

94.

van der Leeuw S (1977) Towards a study of the economics of pottery making. Ex Horreo 4: 68–76.

95.

Wallaert-Pêtre

(2001) Learning how to make the right pots: Apprenticeship strategies and material culture, a case study in handmade pottery from Cameroon. Journal of Anthropological Research 57(4): 471–493.

96.

Wayessa

(2011) The technical style of Wallaga pottery making: An ethnoarchaeological study of Oromo potters in Southwest Highland Ethiopia. African Archaeological Review 28: 301–326.

97.

Wolf

(2013) Pre-contact pottery and the woodland occupations of Vermont and the Northeast: A closer look at the significance of the McNeil Site (VT-CH-93). Journal of Vermont Archaeology 13: 1–18.

98.

Woolsey

(2017) A Historical Approach to Shifting Technologies of Ceramic Manufacture at Gaspereau Lake, Kings County, Nova Scotia . PhD Dissertation, McMaster University, Canada.

99.

Wright

(1972) Ontario Prehistory: An Eleven-Thousand-Year Archaeological Outline. Ottawa: National Museum of Man.

100.

Wright

(1994) The prehistoric transportation of goods in the St. Lawrence River Basin. In: Baugh

Ericson

(eds) Prehistoric Exchange Systems in North America. New York: Springer, pp. 47–72.