Abstract
Archaeological palimpsests are depositional units where the remains of various human occupations have been mixed for hundreds to hundreds of thousands of years. They create various sets of analytical scales that archaeologists must deal with routinely. In this paper, I argue that sociocultural processes derived from a communities of practice framework — scaffolding, guided participation, and continuity through shared activities — can be used by archaeologists to describe a palimpsest’s lithic assemblage, to differentiate its patterns, and to interpret their meaning. These processes must first be remapped onto an ecological approach to skill before they can be expanded onto new sets of scales, however. I ground my work at the site of La Martre (Quebec, Canada), a nexus of fifteen marine terraces spread over 500,000 m2. Slow depositional processes and plowing have mixed its lithic remains to create a 10,000-year-wide depositional unit with poor chronological and spatial control. Fieldwork conducted between 1995 and 1999 sampled 0.03% of its total surface. Most of its 2111 tools and 207,506 flakes were uncovered in its 40-cm-thick plowzone. I build methodological tools — dispersion surfaces, skill combinatorics, and extended skilled reduction sequences — to describe a small subset (N=93) from one of La Martre’s loci (16-west). I describe ten extended skilled reduction sequences showing various degrees of skill and knapping methods. Concepts of scaffolding, guided participation and continuity through shared activities are then used to interpret these patterns.
Similar content being viewed by others
Avoid common mistakes on your manuscript.
Introduction
Archaeological palimpsests are depositional units formed by the mixing of multiple episodes of occupation over hundreds to thousands and hundreds of thousands of years (Bailey, 2007; Lucas, 2005). They are the most common form of depositional material that archaeologists must work with. How they can reshape epistemic, methodological, and interpretive practices has remained little addressed, however. Time perspectivism grapples with some of these issues. It is an epistemic approach according to which a conceptual framework should be used at an appropriate scale of analysis (Bailey 1983, Bailey, 2007, Binford, 1981, Dibble et al., 2017, Holdaway & Wandsnider, 2008, Lake, 1996, Murray, 1997, 1999, 2002, 2013, Perreault, 2019, Rezek et al., 2020, Robb & Pauketat, 2013, Shott, 1998, 2010, Stahl, 1993, Vaquero et al., 2012). Scale refers to both the extent and the resolution of an archaeological context. Extent refers to the timespan (e.g., hours, days, years) of a phenomenon, while resolution (e.g., high, coarse) refers to the unit of measurement available for describing said phenomenon. Describing a short-lived phenomenon requires high-resolution units of measurement, while longer-lived ones can be addressed with longer, coarser units. Different conceptual frameworks must be used to guide the analysis of phenomena according to the scales that they unfold at. Depositional processes create various scales of analysis where the sociocultural dynamics that make up social anthropological frameworks need not always be relevant. Artefacts caught up in an archaeological palimpsest need not have had any other connection besides their spatial proximity.
How depositional constraints are addressed is fundamental to the ongoing development of archaeology, and matching a depositional scale to a conceptual framework certainly makes for sound methodological and interpretive practices. But this is also an epistemic choice that stems from underlying assumptions regarding (i) how sociocultural processes can expand beyond the frameworks and scales they were initially built in; (ii) how scales are created to become units of understanding for past practices; and (iii) how archaeologists map events to perceive patterns. In this paper, I argue that processes derived from a communities of practice framework — guided participation, scaffolding, continuity through shared activities — can be used to differentiate patterns where heavy mixing coupled with high artifactual density have destroyed depositional units of control. These processes can help reframe past (e.g., stone knapping) and present (e.g., description) practices. They can provide an alternative starting ground for working in palimpsests. They can be used to inform the methodological tools used to describe practices as well as to create new sets of scales. In fact, these processes must be used to create the conditions for archaeologists to even perceive anything, provided that they are first released from the social anthropological frameworks and scales they were initially built in.
I ground this research within the plowed fields of La Martre (Quebec, Canada). It has been a special place (sensu Bamforth, 2007) in northeastern North American history for thousands of years due to the affordable setting that it provides for continuous and repeated occupation episodes (Chalifoux, 1999; Dumais, 2000; Lothrop et al., 2016). Mid-twentieth century plowing activities disrupted spatial and chronological control except for its approximately 10,000 cal BP occupation baseline based on terrace emergence, macrofossil dating in nearby lakes, and Late Paleoindian (11,600 to 9000 cal BP [Chapdelaine, 2020, Lothrop et al., 2016]) Plano projectile points production. Most discarded lithic remains accumulated in its 40-cm-thick topsoil over its years of continuous occupation and over its fifteen marine terraces that cover a total surface of approximately 500,000 m2. Sampling conducted between 1995 and 1999 covered approximately 0.03% of La Martre’s surface, and uncovered 2111 tools and 207,506 flakes. Plowing might not create the worst palimpsests that archaeologists must deal with. Discrete structures might be preserved better than in thousands-year-old fluvial deposits spread along downstream riverbanks for example (Bailey, 2007; Dunnell & Simek, 1995; Stern, 1994). But plowing activities at La Martre merely aggravated a lack of contextual control. Slow sedimentation rates coupled with a high density of lithic materials produced mostly from bifacial stone knapping in locally available chert (Burke, 2010; Chalifoux, 1999; Kolhatkar, 2020) have created a 10,000-year-wide unit of practice and mixed hundreds of thousands of short knapping episodes — bifaces, projectile points, unifacial tools, flakes — with no definite occupation ceiling.
I begin with a brief overview of prior works at La Martre and in Late Paleoindian northeastern North America. Second, I expand on the methodological constraints that the palimpsest poses. I show that La Martre reveals deeper epistemic assumptions regarding how practices are framed, how sociocultural processes derived from communities of practice framework can help inform these practices, and how an ecology of skill is key to expanding these processes to broader sets of scales. Third, I build a methodology based on dispersion surfaces, skill combinatorics, and extended skilled reduction sequences to organize my descriptions according to these sociocultural processes. Fourth, I sample 93 bifaces from a smaller locus — 16-west — to work out the application of these methodological tools. Fifth, I use processes of guided participation, continuity through shared activities, and scaffolding to make sense of 16-west’s new set of patterns.
Prior works at La Martre and in Late Paleoindian northeastern North America
La Martre is located on the northern part of the Gaspe Peninsula, in the eastern fringes of Quebec (Canada) (Fig. 1). It is a constellation of fifteen marine terraces — dubbed “stations” by its investigators — located along the actual St-Lawrence and La Martre rivers. These terraces emerged following the Goldthwait sea’s various regression episodes, between 13,000 and approximately 4000 cal BP (Dionne & Coll, 1995). Archaeological work has focused on upper terraces (stations 12, 15, and 16) due to their Late Paleoindian affiliation and an emphasis on Quebec’s first populations research hypotheses (Chalifoux & Tremblay, 1998). This affiliation is based on Late Paleoindian, Sainte-Anne/Varney projectile points (Fig. 2A), dated elsewhere between 10,800 and 9000 cal BP (Chapdelaine, 1994; Chapdelaine, 2020; Lothrop et al., 2016). Other Late Paleoindian points — Pseudo-Agate Basin points (Fig. 2B) — were recovered here as well, but their chronological framing is even less certain, tentatively located between 11,600 and 10,800 cal BP (Bradley et al., 2008; Lothrop et al., 2016) and which conflicts with a 10,400 cal BP occupation baseline derived from macrofossil deposits in the region’s lakes (Asnong & Richard, 2003; Richard et al., 1997). No absolute dates could be obtained on site.
La Martre in a northeastern North American setting. Lower left: red square indicates northeastern North America’s limits, and red arrow shows the location of La Martre. Upper left: location of 16-west (in red) relatively to stations 15 and 16. Right: La Martre’s stations (in red: stations mentioned in the text). Map background: Google Earth, January 12, 2023
Late Paleoindian “Plano” projectile points found at La Martre, locus 16-west. A Sainte-Anne-Varney; B pseudo-Agate Basin. Sainte-Anne/Varney outline redrawn from Bradley et al., 2008. Drawing inserted in outline is from biface no 16-480. Pseudo-Agate Basin outline redrawn from Lothrop et al., 2016. Drawing inserted in outline is from biface no 16-881. Dashed lines show missing part. Photographs and drawings by the author
La Martre is part of a larger nexus of sites tentatively attributed to the Late Paleoindian cultural period based on site altitude following the emergence of their marine terraces, and on the recovery of Plano projectile point (Benmouyal, 1987; Chapdelaine, 1994; Dumais, 2000; Lothrop et al., 2016; Pintal, 2006). Late Paleoindian northeastern North America spans 2500 years, it encompasses four subregions (eastern Great Lakes, New England Maritimes, mid-Atlantic, middle-upper Ohio Valley), five different biomes (from open tundras to closed forests), and 42 known archaeological sites (Dyke, 2005; Lothrop et al., 2016). Various transformations seem to occur around that time. Environmental shifts, demographic intensification, and stylistic drifts by local populations descending from Early (13,000–12,200 cal BP) and Middle (12,200–11,600 cal BP) Paleoindian groups might have led to the further diversification of projectile points as well as increased regionalization (Benmouyal 1987, Chapdelaine & Richard, 2017, Jackson, 2004, Ellis et al., 2011, Lothrop et al., 2016, Newby et al., 2005). An eastward migration from the Plains following the opening of previously glaciated landscapes along the Canadian Shield’s southern flank might have brought in additional unfluted projectile points modeled after western templates (Chapdelaine, 1996). Similar projectile point types found in a variety of biomes and over an extensive area suggest that complex cross-regional sociocultural dynamics might have been at play (Anderson 1995, Fiedel, 2014, Petersen, 2004).
The interplay between these and other events probably varies regionally. Dynamics in the Great Lakes’ closed forests might have been very different from those at play in Eastern St-Lawrence’s periglacial and open tundra biomes (Meltzer, 1988). The Gaspe Peninsula provided its inhabitants with affordable and predictable environmental settings — a nearby Goldthwait seashore and other local water streams, flat and well-drained sandy surfaces, stable protein resources owing to the nearby seashore as well as caribou populations attracted to a close by residual ice cap (Pelletier & Robinson, 2005; Richard, 2007), and massive Cap-Chat chert outcrops (Burke, 2007, 2010; Kolhatkar, 2006). The sheer size of La Martre suggests that it could have been a place of special importance within an already significant nexus of Quebecois Late Paleoindian sites. But the sampling of the various terraces also suggests that the locality could have been used as a workshop for intensive knapping for much longer. Lithic remains have been found on all terraces in high to very high densities. Even though test-pitting and excavations could only sample a very small percentage (0,03%) of La Martre’s total estimated surface of 500,000 m2, 2111 bifacial and unifacial tools as well as 207,506 flakes were recovered (Kolhatkar, 2020). Various groups would have come back here either to refresh their hunting gears (Dibble et al., 2017), to teach their young how to knap without endangering the rest of the group with fatal mistakes and wasting precious materials (Ferguson, 2003; Hiscock, 2014; Milne, 2005; Pigeot, 1990), to take advantage of its nearby Cap-Chat chert outcrops (Burke, 2010), and to test some advanced lithic procedures, since Cap-Chat chert was of sufficiently good quality to withstand sometimes impressive lithic craftwork (Kolhatkar, 2022b), all the while taking advantage of La Martre’s other affordable features for shorter occupation episodes or longer, collective stays.
Yet, for all this available material, research on Late Paleoindian northeastern North America has remained sparse (Chapdelaine, 2020, Jackson, 2004, Lothrop et al., 2016, Petersen et al. 2000). Most analyses have focused on earlier Paleoindian groups’ settlement patterns, technological organization, and subsistence (e.g., Ellis et al., 2011; Ellis & Deller, 2000; Ellis & Lothrop, 1989; Lothrop et al., 2016; Meltzer, 1988; Spiess et al., 1998; Storck, 1997). Archaeological knowledge available for the Late Paleoindian period further dwindles as one proceeds north across the southern fringes of the Quebec Province (Chapdelaine, 2020; Chapdelaine & Richard, 2017) towards the Gaspe Peninsula (Benmouyal 1987, Chalifoux, 1999, Chapdelaine, 1994, 1996, Dumais, 2000, Pintal, 2006). What little research has been conducted there has been framed within a historical cultural approach. Archaeological analyses have focused mostly on typochronological affiliations between Plains and Eastern Paleoindian groups based on their projectile points to explore migration hypotheses (Chapdelaine, 1996). Analyses have focused on bifacial and unifacial tool types, raw material provenance, and activities within a site’s various areas (Benmouyal 1987, Chalifoux & Tremblay, 1998, Chapdelaine, 1994). Additional research hypotheses have been briefly explored to try and account for Late Paleoindian groups’ ways of life as guidelines for future research (Chalifoux, 1999; Chalifoux & Tremblay, 1998; Dumais, 2000). Little is known regarding this period’s lithic technology, as neither detailed chaînes opératoires nor reduction sequences have been described. This problem extends to the whole of northeastern North America’s Paleoindian, except for a few detailed technological analyses of projectile points manufacture on Early Paleoindian Great Lakes sites (Deller & Ellis, 1992; Ellis & Deller, 2000).
In other words, there is little regionally and culturally specific knowledge that can help frame analyses at La Martre. This is even more problematic given the site’s mixing and density. Indeed, on the one hand, myriads of lithic remains produced in an affordable and safe environmental setting provide archaeologists with high resolution knapping episodes that would have been discarded quickly without the need for extensive reworking. On the other hand, such lithic remains got caught in mixed depositional contexts owing to slow sedimentation rates and later, twentieth century plowing activities over many of its terraces. Most lithic remains have been found in La Martre’s 40-cm-thick plowed topsoil and are still found to this day. This does not allow for clear-cut, spatial, and chronological control through its homogeneous horizontal and vertical spreads of lithic remains. Certainly, previous research has shown that plowing activities do not completely destroy an archaeological site (Dunnell & Simek, 1995; Odell & Cowan, 1987). Its discrete structures of artifacts endure, although they are spread to degrees that are not always easy to control. Vertical distribution might also be preserved to some degree, either within a plowzone’s minimax (Dunnell & Simek, 1995), or within unplowed lower strata. An archaeological B horizon layer can sometimes be found below the plow zone at La Martre (Chalifoux & Tremblay, 1998). But this does not address the fact that prior to plowing, occupation episodes were never sealed from one another due to slow sedimentation rates. Upper terraces were accessible for thousands of years, in addition to lower terraces that emerged later. Plowing aggravated rather than created the La Martre palimpsest.
An ecology of palimpsests
Methodological and epistemic constraints at La Martre
A 10,000-year-wide depositional unit of analysis made of high-resolution events must be worked with. In the absence of clear spatial control, lithic technology can be used to describe stone knapping practices and help structure La Martre’s assemblage. However, there are methodological problems to this. No knapping methods have been described for Late Paleoindian northeastern North America, let alone for the Gaspe Peninsula, that might provide with a technological groundwork to build upon. At the same time, palimpsests could provide with important and different data regarding past practices that sites with better controlled depositional contexts might not hold (Bailey, 2007; Binford, 1981). Hence, it might not be advisable to simply look at La Martre for patterns that would have been described in other settings, but rather to try and make do with the specifics of its materials.
However, how one should work outside of well-controlled depositional layers is hampered by the fact that methods used to describe knapping methods are built on normative assumptions regarding past practices. This is especially the case with technologists’ main tool for analyzing past manufacturing practices, the chaîne opératoire approach. This approach is defined as both a theoretical and a methodological framework (e.g., Audouze & Karlin, 2017; Bar-Yosef & Van Peer, 2009; Pelegrin, 1995; Pigeot, 2011; Ploux et al., 1992). Its theoretical component stems from André Leroi-Gourhan’s (1973) works on techniques, with which he sought to define a group’s ethnicity on the ground of its milieu intérieur, or inner world. In this world, the milieu technique, or technical world, is considered as more stable than other cultural components (e.g., language, beliefs) because it is constrained by a milieu extérieur’s, or environment’s adaptive requirements, even if environmental constraints can leave various degrés de fait, or degrees of liberty, for people to play with. Knowledge is stored in the technical world as chaînes opératoires, or cognitive templates, of which there are two types — machinale and exceptionnelle (Leroi-Gourhan, 1964). The chaînes opératoires machinales are of greater importance because they are at once learned and stored in one’s unconscious mind, removed from active manipulation until they are revealed to one’s attention when procedures do not go as planned.
The chaîne opératoire is also the methodology used to identify the cognitive templates that make up this technical world. These cognitive templates embrace the technological and economical components of a group’s organization to inform archaeologists on its broader social and cognitive structures (Pigeot, 2011). Because these templates are understood as chains, they must be built by (i) linking the gestures that produced an assemblage in a successive fashion towards (ii) an end goal, or intention, and through (iii) as many intermediate goals as needed to achieve this goal. This temporal, or step-by-step, description allows a technologist to reconstruct an object’s technical context and to compare objects that are comparable (Ploux et al., 1992) because they have reached the same stage of their life — e.g., production, use, recycling, discard, and as many substages as each stage might hold (Inizan et al., 1995:15). Production is key in identifying cognitive templates. It must be described by distinguishing between knowledge (mental representations of ideal shapes as well as of a procedure’s material result that are learned and shared by all members of a group) and know-how (the ability to evaluate the situation at hand as well as to program a sequence of gestures from this evaluation to reproduce an ideal shape) (Pelegrin, 1985, 1991, 1995).
Refitting, either physical or mental — the ordering of flake scars that produced a tool or a core rather than the flakes themselves — is an important method for enacting a chaîne opératoire’s temporal description (e.g., Bar-Yosef & Van Peer, 2009; Inizan et al., 1995). They help (i) identify various technical properties such as hierarchy between faces, removal sequence patterns, the type of hammer employed and importantly, (ii) show how these properties are ordered in time from blank to specific finished (tools) or exhausted (cores) products (Inizan et al., 1995). Refitting is then used to identify technological structures: the technological choices (Lemonnier, 1993, 2010[1983]), knapping methods (Tixier, 1967) and méthodes majeures (or main methods [Perlès, 1991]) that frame an assemblage’s lithic practices. In the case of bifacial knapping, roughout and finishing stages are used to help organize this refitting — roughout refers to the creation of two convex faces along a center plane, while finishing refers to the production of a symmetrical outline along a longitudinal axis (Inizan et al., 1995:44, after Roche & Texier, 1991). Finally, an a posteriori selection of variables is preferred to an a priori one (Soressi & Geneste, 2011). In contrast with the latter that considers all attributes equally, an a posteriori approach states that variations can be almost infinite between objects, and that an analyst’s judgment is also central to drawing relevant variables and attributes. This a posteriori judgment must be constrained within well-controlled contexts such as depositional layers that provide subsets with archaeological significance.
La Martre poses major constraints on the theoretical and methodological components of this framework. On the theoretical side, the chaîne opératoire was developed from the underlying assumption that a specific, ethnic group would be a technologist’s unit of analysis. When applied in an archaeological context, it requires that an assemblage is not a random aggregate but a “methodically interconnected association of artifacts” (Bar-Yosef & Van Peer, 2009:105). As time perspectivists and others have argued (e.g., Bailey, 2007; Murray, 1999; Rezek et al., 2020; Shott, 1998, 2010), the latter does not hold in palimpsests, and we must rather start from the “randomness” assumption. This then puts into question the links that weave the various methodological concepts of the chaîne opératoire. An intention is defined by assuming that ideal representations underlie a group’s technical world, but as we have seen, a palimpsest does not afford for this underlying assumption. Knapping sequences used to compare bifaces at similar steps cannot be identified, except for early (cortex rich preforms) and late (projectile points) manufacturing stages. This leaves most objects in a poorly defined in-between that cannot be tied to a specific starting point or to an intention. Because there is no common basis for comparison, each object becomes potentially unique; the more so as additional data is gathered through mental refitting without clear means for reconnecting data extracted from single bifaces to a broader whole. Most bifaces uncovered at La Martre are broken (see below), which also complicates further comparisons.
Reduction stages can certainly be used to some extent to help describe knapping sequences, because some mechanical properties can be generalized from research on other forms of projectile point manufacture (Andrefsky, 2005; Bradley, 1993; Callahan, 1979; Deller & Ellis, 1992; Smallwood, 2010). But the qualitative and morphometric criteria used to classify bifaces are defined prior to an analysis. In contrast with a chaîne opératoire approach, they are not allowed to grow from the technical contexts where an analysis is conducted (Soressi & Geneste, 2011). In addition, a reduction stage is part of a reduction typology where one category has meaning in relation to the others (Ellen, 2006; Hill & Evans, 1972). These categories pre-order the development of an object from start to finish, here again upstream from one’s analysis (Kolhatkar, 2022b; Shott, 2017). Finally, stages must be enacted by knappers if they are to be recognized by archaeologists as such, even with broader definitions such as Roche and Texier’s (1991). This also excludes lesser craftspeople from a descriptive framework. In other words, this one size-fits-all approach narrows the broader array of lithic expressions that La Martre’s settings might have afforded and that knappers might have enacted for various reasons. This limited descriptive architecture frames lithic technologists’ selection process upstream from their analyses.
In sum, comparison between bifaces is compromised by the lack of underlying clear depositional structures. One must start from a looser lithic dispersion that I have defined elsewhere as “the spread of various shapes that have been discarded at similar or various steps of their development by stone knappers” (Kolhatkar, 2022b:354). This methodological problem becomes epistemic: how is one to build knowledge in such a dispersion? How is one to work outside of the ready-given structures that are routinely used and expected, either from prior works in technology, or from a site’s fine depositional history? What might such “knowledge” be about?
To begin answering these questions, we can turn to the works of anthropologist Gregory Bateson (1972, 1979). His ecological epistemology works along the relation between data (recording procedures and the resulting descriptions) and fundamentals (propositions which are generally true and that can inform descriptions). It asks what fundamentals constitute a phenomenon, what and how data should be mapped onto such fundamentals, as well as how processes might evolve from this mapping (Bateson, 1979). It foregrounds conceptualization as a creative practice that must follow the pathways afforded by one’s materials. It seats at the crossroad between epistemic (how we know the world, how we create knowledge) and ontological (what is the world made of) inquiries. Both work recursively: what one thinks the world is made of constrains the questions one will ask about it, and how one knows the world will reinforce what one thinks it is made of (Bateson, 1972:314). To escape from this recursive dynamic, errors are mandatory, but so are the corrective processes that will occur to account for these errors (ibid, 292–301). Data and fundamentals must be fit together in such a way that data are mapped onto the processes that pertain in the fundamental body of knowledge. However, if the data and the fundamentals do not fit, either the data is wrong, or another fundamental body of knowledge must be at play.
We have seen that we have a problem with even mapping the data from La Martre. It could simply be that building knapping sequences and reduction stages is not possible here. But we must also inquire about the fundamental knowledge that we are trying to map data on when we use such tools. As we have seen, knapping sequences and reduction stage cannot be built from the onset because they require that some organization is already present. This prior level of organization is provided by depositional structures. A technological structure can then be built from well-controlled depositional contexts, or it can be used to order a mixed assemblage if it has been described elsewhere in better-controlled contexts. These structures are needed to position knappers in time and space, according to which they can be affiliated to a chronocultural taxon formed by shared norms. Knowledge can then be added to this taxon. Taxons are important because structures derived from archaeological materials must have sociocultural significance (e.g., Harding, 2005; Ihuel & Pelegrin, 2008; Robb & Pauketat, 2013; Shanks & Tilley, 1987; Soressi & Geneste, 2011). In the absence of a depositional organization necessary to ground one’s normative assumptions, how practices are framed — whether they are related to stone knapping or not — must be revaluated (see also Dibble et al., 2017; Kolhatkar, 2022a, 2022b; Rezek et al., 2020; Shott, 2010). Fundamental knowledge used to reframe practices must let structures emerge, be maintained, and disappear from lithic dispersions. It must also show how alternate archaeological structures might still have sociocultural significance.
Communities of practice: key processes
A communities of practice framework provides an important step in that direction. It has been of increasing importance in archaeology for understanding past sociocultural dynamics. This approach was first expanded from Lev Vygotsky’s (1930) works on proximate development by Jean Lave, Étienne Wenger, and Barbara Rogoff in various contemporary cultural and educational settings (Lave, 1988; Lave, 1993; Lave & Wenger, 1991; Rogoff, 2003; Rogoff & Lave, 2000; Wenger, 2000; Wenger, 1998). It contrasts with classical approaches that frame learning through fixed stages of cognitive development based on occidental standards (Piaget, 1977). It forefronts the sociocultural contexts that people shape as they engage in shared activities.
To explain how such contexts are created, a communities of practice framework focuses on the development of individuals and communities as two meshed but distinct levels of practice (Rogoff, 2003; Roux, 2020). An individual grows up by being confronted by their peers to sets of technical problems that make up a community’s daily life (Bril, 2002a, 2002b). This confrontation may be guided to various degrees, as an infant is either left on their own, led through steps that ease the learning process, or prohibited from attempting certain tasks before a certain age or outside of certain settings (e.g., the safer setting afforded by a nearby chert quarry [Ferguson, 2003]). Regardless of how this confrontation occurs, an individual must learn how to resolve these problems within the social scaffolding provided by their peers — that is, to understand the problems posed by a task, to work out its solutions, and to harmonize their actions with that of others to work efficiently with them. Such scaffolding also benefits more specialized craftspeople. Becoming an expert takes time from other, subsistence related activities. Some form of social complementarity must afford for that focus and justify one’s investment (e.g., Ferguson, 2003; Hiscock, 2014; Milne, 2005; Pigeot, 1990; Stout, 2002). A scaffolding affords for individuals to ensure cultural continuity — and change — through time and space.
Problems are not culturally specific. They can be shared across various cultural horizons, allowing for answers that may be more specific to a community and afford for cross-cultural comparisons (Rogoff, 2003). But problems and solutions are not strictly co-dependent either. Solutions may vary as people adapt and expand their toolkits to address problems that can arise in various contexts. Likewise, individual ways of addressing problems introduce much variability into collective practices. Individuals must learn by themselves how to harmonize and calibrate their actions over short- and long-term tasks because mere observation is not enough. One must learn to identify and resolve the motor problems involved in a task through trial and error, (Bernstein, 1996; Bril et al., 2005). Indeed, an action works at various interlocking levels to produce a visible and effective gesture (Bernstein, 1996; Biryukova & Bril, 2002; Keller & Keller, 1996; Latash et al., 1996). It results from the harmonization of thousands of muscular contractions that must be tied to a goal through repeated practice. A goal works at various scales, from an elementary unit of action (e.g., detaching a flake) to sub-objectives and general strategies put in place to produce a shape (Bril et al., 2005). Low (actions) and high (goals) orders of action work retroactively as low-order calibrations achieve higher-order stabilization while the latter orients the enactment of the former (Bateson, 1972, 1979) in what Bernstein calls “repetition without repetition” (Bernstein, 1996). Hence, a goal emerges from the situation at hand, as a craftsperson must continuously evaluate their next step (Smitsman et al., 2005). Intended effects and unintentional consequences result of this. Variations become part of a production process and its resulting, “finished” shape (e.g., a projectile point) whose unforeseen consequences might send one’s work along a trajectory far removed from a model they might have been attempting to reproduce (Bril et al., 2005; Ingold, 2000, 2011; Keller & Keller, 1996; Latash et al., 1996; Pelegrin, 1985; Smitsman et al., 2005).
Beyond the genealogical model of practice
Guided participation, social scaffolding, and continuity through shared activities are three fundamental processes of a communities of practice framework. One must work at multiple levels of practice to address them: collective and individual levels; technical problems and solutions levels. This framework eschews a normative understanding of practices where norms are a given that people merely reproduce (see also Ingold, 2000). Anthropologists can address how norms are created, maintained, or lost. They have worked from the short temporal contexts provided by their bodily presence with a community’s participants, as much as with the long-term cultural memory that their shared activities enact daily (Rogoff, 2003).
This framework poses one major limitation for archaeologists, however. It emphasizes co-presence: people need to work together to create said contexts and shapes, even though they summon short- and long-term knowledge inherited from their peers over generations. From a time-perspectivist perspective, such a framework could only be used in the most fine-grained occupation floors with palethnographical scales of analysis (e.g., Karlin & Julien, 2019; Pigeot, 1990). Consequently, archaeologists work at inferring the processes that make up communities of practice framework within a broader range of scales. Some have suggested working with specific steps of a chaîne opératoire to identify technical traditions (Roux, 2020). Steps that cannot be copied from a finished product would require such co-presence to be reproduced (see also Gosselain, 2000). Others have made use of the variability seen in levels of skill to explore landscapes of learning within a shared chronocultural horizon (e.g., Cannon, 2011; Goldstein, 2019; Hiscock, 2014; Milne, 2005, 2011, 2013; Wendrich, 2013). However, we have seen that there are no fine-grained occupation floors at La Martre. The palimpsest cannot be affiliated to a closed chronocultural horizon, or even to a specific chronological slice of time. In addition, any steps could have been copied by later knappers. Preforms scattered at various stages of their manufacturing process could have been picked up by anyone, just as technologists might do. Finally, La Martre could have been visited by one community, or several at different moments or for big gathering events, whose traces would have been mixed up due to its slow depositional processes.
More generally however, the fact that neither technical traditions nor landscapes of learning can be used at La Martre shows that these frameworks need an underlying closed structure as a way of connecting artifacts and assemblages at various scales of practice. As we have seen, this closed structure would be provided by the continuity that has endured within a community’s deep-seated technical world with which this community can adapt to its environment. This stable technical knowledge is key to a group’s identity as it is transmitted from one generation to the next (Leroi-Gourhan, 1973). In turn, extending these normative assumptions over the longer and often unknown scales of archaeological materials reveals the genealogical model (sensu Ingold, 2000) that underlies normative assumptions. This model essentializes knowledge. It remains unchanged by the contexts of present activity where it is merely expressed, because it is transmitted from one generation to the next along a diachronic axis neatly separated from a synchronic one (Ingold, 2000:134-139, Ingold, 2007:113-119). Normative structures become synchronic beads that must be slipped onto a diachronic string to follow sociocultural evolution. This framework ensures that knowledge is transmitted diachronically in unchanged form across synchronic episodes of interaction to produce unchanging traditions and create a well-identified subject for archaeologists’ narratives.
As anthropologist Tim Ingold has argued (Ingold, 2000; Ingold, 2007), this model runs so deep within the history of Western thought that it has become a natural way of ordering the world. Yet, we have seen above how the dual dynamics of practice produce loosely integrated wholes. There is no reason why co-presence should be used to tie back communities of practice framework to normative assumptions. The chronological control afforded by well-bounded depositional contexts seems necessary because of the underlying genealogical model used to frame practices. Depositional contexts disrupt this model when they open a wide chronological horizon where we simply do not know “whom” we are talking about. But this also means that it is this model for understanding practices and sociocultural development that must be changed, rather than eschewing sociocultural processes altogether when a chronological scale of analysis is too coarse (Perreault, 2019; Rezek et al., 2020). An emphasis on ongoing engagement, continuous generation, and growth rather than transmission along lines of descent is needed to uphold the potential for change held by a communities of practice framework (Bateson, 1972, Bergson, 2001[1907], Deleuze & Guattari, 1980, Ingold, 2000, 2007, 2011). It could release processes of guided participation, social scaffolding, and continuity through shared activities beyond communities of practice to allow new patterns to grow in palimpsests. It could provide an alternate starting point to genealogy or chronology for mapping practices. It could be used to create other types of sociocultural structures within units of analysis such as La Martre. Co-presence, communities of practice, and technical traditions would only be types of sociocultural contexts and conceptual frameworks that practices produce.
The social life of skill
Skill can provide archaeologists with the thread needed to remap the dynamics between an individual and a collective and between technical problems and solutions onto practices in a palimpsest; to draw patterns in its assemblage that have sociocultural meaning; and to expand the production of sociocultural contexts to new scales. But we need a broader understanding of skill, or what I have called an ecological approach to skill. Elsewhere, I have already fleshed out two of its main properties (i) immanence and (ii) duration (Kolhatkar, 2016, 2022a, 2022b).
First, an ecological approach to skill starts from the assumption that without skill there would be no context to control, no artifact to analyze, no chronocultural, or theoretical framework to build and no assemblage, palimpsest or aggregate to explore. Skill is immanent to any practice and traces left behind by people. It generated lithic dispersions long before depositional processes did, as various knappers with different sets of skills produced many tools, flakes, or any other lithic shapes. In addition, knappers had to follow certain key knapping principles with which they could generate a vast array of shapes and methods, regardless of their specific cultural horizon. Hence, this framework does not need finer depositional control over lithic practices prior to its implementation, nor is it bound to prior genealogical lines of descent. Rather, skill creates the condition for depositional contexts to become objects of archaeological analysis. Where genealogical, chronological, and normative threads were disrupted, skill endures. It is the main generative process for structuring practices.
Second, skill differs from the simple notion of practice in that some things are easier or more difficult to do than others, that one must learn how to enact them through repeated practice, and that this takes time. A skilled practitioner is not merely an expert, but someone who acts and who as a result, attests of their abilities, or lack thereof to do things. Improvement, absence of improvement, and drop in improvement can be used to generate a sense of duration where other temporal markers have been lost. A chronological, linear mapping through the seconds, minutes, or years that it takes to achieve a task and improve (or not) is one such specific tool. It can be used to measure the progression of one’s skill. Measurements help this evaluation but that is because measurement and evaluation are tied to one’s betterment (or not) at a specific task. Measurements come second in an attempt at translating one’s skilled growth and allow other temporalities to be enacted and explored. It follows that a scale’s extent does not necessarily have to be chronological (e.g., minutes, years, centuries). Other units of measurement can be used, in so far as they spread along skilled pathways (e.g., fashioning a projectile point from early crude attempts to a highly regular shape) through myriads of calibrations and fluctuations.
From the above discussion on sociocultural processes, this ecology can be updated with a third key property: (iii) differentiation. As processes of guided participation, social scaffolding, and continuity through shared activities have shown, people enact the motor solutions to the recurrent problems that define their community’s way of life (e.g., living with stone knapping). Hence, a skilled practitioner works at dual levels of practice — individual and collective; problems and solutions. Skill is a pathway for differentiation precisely because practitioners work at dual levels of skilled practice. This differentiation is tied to the similarities which practitioners depart from even as they try to enact them. Lithic dispersions are generated through this dual level of skilled practice. Patterns in a dispersion must also be grown from this dual analysis if they are to have sociocultural meaning. Differences-with-similarities must be archaeologists’ unit of analysis. The finer sociocultural dynamics that underlie these patterns will vary according to the scale of a palimpsest. Inversely, patterns can grow at all scales because they result from a process of differentiation that is immanent to all skilled practices, and that enacts its own duration. That such patterns exceed the boundaries of our models and expectations should not necessarily invalidate them. Rather, this might simply point at the limitations of our existing models, at new shapes that need further theorizing, and at alternative subjects for the stories archaeologists can tell.
Methodology
Skill in stone knapping and description
We may now address the methodological implications of these epistemic proposals to enact these sociocultural processes in a poorly controlled depositional process, to organize our descriptions of stone knapping, and to reframe how long- and short-term goals might be meshed to analyze bifacial manufacture. Skill is a methodological as much as it is an epistemic baseline. This basis allows for comparisons to be drawn between shapes that might have resulted from various causes — e.g., knapping sequences, reduction stages, technological organizations, function, intentions, and post-depositional alterations, in addition to discrepancies between results envisioned (knowledge) and obtained (know-how) due to variable abilities and to which skill tends to be restricted. But skill is not one cause amongst others, because causes for variability and their interplay need to be enacted through the skill without which there would be nothing to analyze. For the same reasons, skill cannot be isolated inside a specific set of variables or classifications. Neither is it limited to identifying levels of skill. Skill redefines the baselines used for comparison. Much of what is presented in this first subsection has already been addressed elsewhere in greater details (Kolhatkar, 2022a:267-280, Kolhatkar, 2022b).
First, technological analyses must work from discards rather than intentions. Discard is meant in a broad sense here: that of a common denominator, regardless of knappers’ intention. It must not be framed in opposition to intention. Discards are necessary to first produce the technical contexts where intentions might emerge at the short, middle, and long scales of action mentioned above. Intentions will vary, notably between knappers with various levels of skill. Every knapper discards their production at some point however, whether knappers had the skill to enact their short and long-term goals, or they attempted to do so but could not because their lesser skill did not allow them to. Discards, then, emphasize the problems encountered by knappers before they entered the archaeological record at various steps in the knapping process and became visible to archaeologists. They can be used to work from the short scales of action visible to technologists without requiring that long-term goals be defined from the onset because, as we have seen, long-term goals emerge through practice. This last point is important. Discards show problems that were being worked out up to the moment of discard (e.g., caused by a major fracture). They also produce incomplete knapping sequences that open a field of possibilities for the technologist who wonders what next step could have happened if the biface had not broken. This field is the same as that which knappers must ponder in the course of their work. They might have envisioned a long-term goal before beginning their work, but the various procedures that they must be able to carry out during their knapping will also be part of their ongoing knapping and its result. Discards in palimpsests do not provide technologists with a specific vantage point over knappers, because calibrations and mistakes can send knapping along unforeseen trajectories. Prior technological structures cannot be used here to anticipate knappers’ intentions. The question then becomes that of extending one’s sight beyond the short-term goal of detaching a flake.
Second, looser classifications allowing for comparisons across a broader set of skills must be used. In the case at hand, “biface” refers to anything that was knapped on two faces along a center plane (Boëda 2001, Shott, 2017), including types of objects that might traditionally fall in other categories (e.g., drills, bifacial cores, bipolar cores, pièces esquillées) because we cannot assume that lesser knappers were not trying to produce a projectile point while ending up producing a bipolar core. These knappers needed to learn the basics of stone knapping through trial-and-error. Likewise, fashioning a biface to produce a projectile point does not exclude using some of its larger thinning flakes, muddling the strict distinction between cores and tools in the process (Bamforth & Becker, 2000; Kelly, 1988; Kuhn, 2007).
Third, a grid is used to record width and thickness values every 3 cm along a longitudinal axis that crosscuts a biface along its facial symmetry plane. Experimentation has shown width, thickness, and particularly width-by-thickness (W/T) ratio to be of central importance to stone knapping (Callahan, 1979). One must learn to master their relationship, as higher W/T ratio is harder to enact. This does not mean that high W/T ratios are required for all kinds of bifaces (e.g., pièces esquillées, bipolar cores), but simply that this ratio is a baseline for broader comparison. How this ratio fluctuates is another matter. In addition, subsuming various shapes under three main mean values clears the ground by removing much of the variability produced during manufacture — various knapping procedures, breaks, and available length. These values can be used to provide a broad distribution and to order bifaces before finer variations are explored by considering shape and other technological variables (see below).
Fourth, a scatter diagram is built from mean width and mean thickness values to produce a reduction continuum and follow W/T ratio fluctuations. This continuum structures reading in a leftward and downward manner, following stone knapping’s general subtractive work. Bifaces can be positioned relative to one another according to their available volume of knappable material. This available volume is another important and broad basis for comparison. The beginning or ending of the continuum is not defined according to the start or ending of a knapping sequence as defined by chaîne opératoire or reduction stage methodologies (e.g., from a blank to a finished projectile point). It does not make use of long-term intentions that got mixed up in a palimpsest or that might require very high levels of skill in some cases (e.g., Plano projectile points). It allows for comparing bifaces according to the situation at hand within a general subtractive goal. This general orientation leaves room for knappers to work in, as there are many ways to proceed from the upper right to the lower left corner of the continuum, according to the methods and techniques (Inizan et al., 1995) used, and the skill enacting these techniques. It positions short-term flake detaching goals within a broader framework that endures regardless of mixing, because whatever their long-term intention might be, knappers need to reduce materials to achieve it.
Fifth, subsets can be differentiated using levels of technical difficulty, raw materials, blank types, and end products. These subsets have been described in more details elsewhere (Kolhatkar, 2022a, 2022b). End products include cores, to emphasize the fact that they are first bifaces shaped along a common reduction process even if intentions might have differed (see above). In addition, I have drawn on the literature of stone knapping in experimental and ethnoarchaeological settings (Bril et al., 2005; Callahan, 1979; Pelegrin, 1985; Stout, 2002; Waldorf, 1993; Whittaker, 1994) to suggest four levels of technical difficulty — easy, slightly difficult, difficult, and very difficult. These levels are ordered relative to W/T ratios and further refined using cross-section, surface, and retouch configurations. A very difficult biface must show a W/T ratio higher than 4, lenticular cross-section, invasive to covering flakes produced in a regular scar sequence, and, if retouch can still be found, it must be highly regular as well. In contrast, an easy biface will not exceed a W/T ratio of 3, its cross-section will be irregular, its thinning marginal, and its retouch, if still apparent, will be equally irregular. A slightly difficult biface shows a W/T ratio below 3, an irregular cross-section, marginal to invasive thinning, as well as marginal and irregular retouch if present. Finally, a difficult biface has a W/T ratio between 3 and 4, a lenticular cross-section, invasive to covering flakes, and regular retouch if present (see Kolhatkar, 2022a:271-275 for additional details).
These broad difficulty level categories make for a similarly broad comparison baseline in an assemblage where broken bifaces at various stages of their manufacture reduces the number of variables that one can use systematically. Levels of difficulty are horizons of attainment made of tasks that can only be achieved through repeated practice. Importantly, simply attempting difficult and very difficult tasks needs repeated trial-and-error. This means that novices cannot try their hand at very difficult procedures. In addition, easy tasks can be worked by various knappers, and they will not allow for distinguishing between lesser and better knappers. More experienced knappers will not fail easy tasks, meaning that easy bifaces tend to be discarded by lesser knappers. Inversely, as only experienced knappers can attempt tasks of higher difficulty, their discard attests of better knappers at work.
Dispersion surfaces
With these broad comparative baselines in hand, we can work at differentiating stone knapping practices further. Three additional methodological concepts must be used: surfaces of dispersion, skill combinatorics, and extended skilled reduction sequences (Fig. 3).
Methodological plate showing dispersion surfaces (1 to 5) and how surfaces are drawn progressively to construct extended skilled reduction sequences (6). Surfaces are ordered according to increasing resolution and decreasing extent of analysis. (1) reduction continuum; (2) metric sectors; (3) skilled reduction sequences; (4) stages; (5) partial knapping sequences; (6) extended skilled reduction sequences. Drawing by the author
A dispersion surface is a dual level of description where sets of shared problems are populated by the various solutions explored by knappers. These problems can be identified at various scales. Contrary to the definition of a scale used in a time-perspectivist epistemology, in a surface there is no distinction between extent (a phenomenon) and resolution (the description of a phenomenon), that would then need to be matched. It blurs the distinction between what a phenomenon should be according to one’s preconceptions and how it unfolds along one’s descriptions. A phenomenon exists only in so far as it can be described. In that case, extent refers to the size of the sample used to describe a phenomenon, from an assemblage of hundreds of bifaces to a biface unique from any other. Resolution refers to the type of connections that can be made between and within objects. For example, bifaces can be connected along broad trends — size, technical difficulty levels, blank types, and specific knapping problems. This eschews the conflict between the high-resolution provided by a short-lived stone knapping event, and a palimpsest’s coarse chronological and depositional resolution. Rather, high-resolution, short-lived knapping procedures can be used to describe the similarly high-resolution phenomena that they collectively serve to create along the process of skill.
From this definition of a dispersion surface, five types of surfaces can be defined, corresponding to various interconnected scales of analysis: (i) the reduction continuum, which encompasses the whole extent of an assemblage’s stone knapping practices, depending on what an archaeologist puts in it (see above); (ii) metric sectors; (iii) skilled reduction sequences; (iv) stages; and (v) partial knapping sequences.
Metric sectors are defined in metric terms. Metrics provide with the broadest common denominator. On the one hand, they remove some differences that might have made a difference (e.g., qualitative attributes). On the other hand, they also make a difference because knappers need the skill to, for example, achieve high W/T ratios, while width is the easiest property to shape and can be similar across various levels of skills and W/T ratios. Knapper with varying levels of skill are allowed some (metric) similarity, even as other properties of their shapes (e.g., scar regularity, cross-section, W/T ratio) will differ because they are not yet good enough to attempt or produce them. Hence, metrics allow for variously skilled knappers to conjoin in shared surfaces of varying scales. Width and thickness must be considered separately because they are disjointed by one’s skill at enacting a certain W/T ratio range.
Three methodological points must be considered in defining metric sectors. First, we work with the subtractive property of stone knapping. Width and thickness must be used to frame change in discard rates through time, where “time” is enacted by material subtraction. This rate is better defined by considering width and thickness along separate line graphs, and by identifying inflection points along these lines. Second, we work through knappers’ discards. Line graphs will be skewed towards what knappers were trying to achieve or achieved but nonetheless discarded. Abrupt changes towards high counts of discard are of specific interest in choosing between possible inflection points because they show important points of convergence between knappers. Third, since metric sectors help focus on important points of convergence, the number of inflection points used to define them should be kept at a minimum (1 or 2 by line graph) in order to provide lithic analysis with the most inclusive surfaces possible. This also justified by the fact that mean width, thickness and W/T values are used. Each surface must leave some room for shape variation, while a surface too narrow might be heavily influenced by single biface value dispersion. Inflection points are then combined on the reduction continuum to create various ranges of W/T ratios.
Skilled reduction sequences are levels of technical difficulty spread within the reduction continuum. They cross-cut various metric sectors. As we have seen, reduction sequences help mesh isolated shapes with a technical context to compare objects that are comparable. They tend to be defined with fixed stages however, which produces ideal reduction sequences and makes them prone to essentialism. Enskilling reduction sequences is a step away from this. It increases one’s ability to capture a broader range of stone knapping expressions and their resulting shapes.
Stages point at more specific technical problems that appear within metric sectors and along skilled reduction sequences. Shared problems between bifaces circumscribe subsets that become stages by virtue of them being defined inside a reduction continuum’s subtractive orientation. Attributes help define problems that knappers would have needed to resolve if they were to continue removing materials. Stages are not given in advance of an analysis but are instead defined using Soressi and Geneste’s (2011) a posteriori approach (see above). To define stages, one zooms in progressively, starting from the reduction continuum, then to a specific metric sector, then proceeding to a specific skilled reduction sequence, and finally to any shared variables (or motor technical problems) between bifaces to identify a stage. Hence, an analyst’s judgment is constrained within previous and broader dispersion surfaces. An attribute is defined as an emerging property at the group level rather than at an individual level. In addition, higher levels of difficulty can justify that the definition of a stage be more fine-grained. Better knappers have more control over their biface’s proportions, and thus define and refine the similarities that make up a stage. The qualitative variables used to define levels of technical difficulty (W/T ratio, cross-section shape, thinning, and retouch) are used to specify shared problems more precisely than the coarser metrics of a metric sector. Additional qualitative or finer metric properties can also be used to account for these shared properties: scar patterning and sequencing, platform preparation, cortex, plan view, specific thickness following a flake removal, and other variables used in a technological approach (Inizan et al., 1995) but that cannot always be found on all sampled objects and along a whole reduction continuum (e.g., retouch is almost useless at very early stages, or platform preparation is removed almost instantly by thinning).
The finer details of bifaces’ partial knapping sequences can be explored at this point through mental refitting, and illustrated through diacritical schemes (e.g., Boëda 2001, Inizan et al., 1995, Pelegrin, 2019). Scars are ordered in time according to their relative position to one another. Their general orientation and the presence of a negative bulb are also recorded. Scars are then grouped according to the technical goal that they achieve — e.g., various generations of thinning, edging, retouch of ridges — and ordered as technical steps using scar numbering, orientation, and morphological result (e.g., thinning a median section with covering or comedial flaking). Steps appear as gray layers with various tones according to their order (e.g., darker layers are older). Information redundant with layers (e.g., final retouch scar direction and sequencing) is not indicated to simplify diacritical schemes, except if negative bulbs are apparent or if scar orientation is very variable. These readings help connect various stages according to the problems encountered and the solutions that would have been needed to proceed. Inversely, a knapping sequence’s high reading resolution is constrained within a stage and along problems to be resolved later in the reduction continuum.
Skill combinatorics and extended skilled reduction sequences
Connections between dispersion surfaces can be explored further through skill combinatorics. Combinatorics are fixed-systematic ways of varying (sensu Descola, 2005). They work at the level of the individual body. Skilled combinatorics are meant to capture a fixed set of knapping scenarios as they unfold between a knapper’s fingers: continuity, change, and simultaneity. Continuity means that a difficulty level stays constant through at least part of a reduction sequence; change means that difficulty levels vary along a reduction sequence (e.g., tasks become more difficult or easier); simultaneity means that as one attempts a knapping sequence once more, they have managed to switch to a different reduction sequence and enact its stages, whereas they once could not.
Skilled combinatorics must be used with the specifics of surfaces of dispersions to follow the unfolding of these knapping scenarios. They connect stages found in skilled reduction sequences and create extended skilled reduction sequences. An extended skilled reduction sequence can encompass various levels of difficulty, as tasks vary along a stone knapping process (e.g., blanks must be edged first, a relatively easy task, before they can be further thinned). Technological arguments ground the use of these combinatorics: subtraction, but also the kinds of operation that working further might call for (e.g., using a central ridge might call for end-thinning; extensive pressure retouch needs a finely crafted ridge network). In addition, connections vary according to the level of difficulty one is working with. More precise technological arguments can be used in higher levels to connect stages because knappers can enact precise tasks, while lesser tasks might limit one to connect stages along metric arguments only. Better knappers are also better able to follow a knapping reduction through than novices who might create dead ends very quickly (e.g., very low W/T ratios that forbid further transformation).
Once these extended skilled reduction sequences have been constructed, they can be used together as a network to help follow the resolution of problems through the various metric sectors of a reduction continuum. They show various ways of resolving problems. This points at another important property of this descriptive architecture. We do not simply follow the transformation of blanks through various stages, but also of various bodily developments within a lithic duration. Indeed, a biface discarded at the state it is analyzed at captures a certain knapper at a certain moment of their life, where they have reached a certain level of skill, a certain step in their knapping process and importantly, a step constrained by the level of difficulty that their skill affords. It follows that multiple discards freeze these various moments to create a temporality made of various bodily developments. Simultaneity and change combinatorics point at both a change in the way one must handle their blank, and, thus, the change that a knapper must be able to enact to proceed. It also points at a skillset that knappers will be able to or will have to develop in the future, but that are not accessible to them for the time being when they have had to discard a biface at an earlier stage of their skilled reduction sequence (change) or at a skilled reduction sequence outside of their reach (simultaneity). They simply could not yet enact the required changes (e.g., switching from an edging stage to an advance thinning stage). Inversely, low level tasks can inform us on the early stages of a highly skilled reduction sequence because they are easier to imitate. Continuity attests to a knapper who has attained a certain horizon of difficulty. If it is a lesser difficulty level, it might point to someone who was also unable to change their way of handling a biface as they reached a certain place in the reduction continuum before discarding their biface. They might be on the right track if they were able to proceed through some more advanced skilled reduction sequence that others could not even attempt. If it is a higher difficulty level, we can see how long this level was held (e.g., a very difficult skilled reduction sequence is spread all over the reduction continuum).
Coding
The methodology presented above serves to locate bifaces within dispersion surfaces of increasing resolution and decreasing extent, from a broad W/T distribution and general width and thickness measurements, down to individual knapping gestures and finer morphometric and morphological properties. Dispersion surfaces constrain an observer’s subjectivity within subsets of decreasing size where no depositional structure can be used, while also better framing how a biface might relate to others. Arguments are drawn from prior knowledge on the skilled practice of stone knapping to justify boundaries between a class of surface (e.g., between metric sectors). These surfaces also serve to acknowledge that a subjectivity cannot be removed from the object under study and from the patterns that are created to become an object of thought and inquiry. They do not try to negate this subjective process of analysis, as it is equally central to the selection of differences between bifaces. Dispersion surfaces mesh objective knowledge with subjective assessments that in turn create new objective knowledge about an assemblage’s lithic practices.
Each biface can be coded according to its position in these surfaces of dispersion and how skill combinatorics connect these surfaces. This coding serves to follow the progressive differentiation of bifaces (Fig. 3). They might share a similar metric sector but might already bifurcate along different skilled reduction trajectories. This coding also helps follow how an analyst’s judgment is constrained inside a dispersion surface in the absence of depositional control. I have coded bifaces according to their metric sector, level of technical difficulty, reduction stage, and skilled reduction sequence, such as “MS2/D/5/A” where “MS2” signals metric sector 2; “D” signals a difficult skilled reduction sequence; 5 signals reduction stage 5 within an extended skilled reduction sequence; and A signals a segment of extended skilled reduction sequence.
These two last codes need additional explanation. Once a “stage” has been defined as a subset of bifaces, it must be connected to other stages — where they come from, and where they might lead — across the reduction continuum, its metric sectors, and its skilled reduction sequences, before a stage number can be attributed. Combinatorics are key because they define three possible connecting scenarios, which must be selected using knapping arguments (see above). Stages can be connected across a similar skilled reduction sequence (e.g., very difficult to very difficult — “continuity” combinatoric), or across different skilled reduction sequences (e.g., difficult to very difficult — “change” combinatoric). They can also coexist between different skilled reduction sequences, or finer branching can also be observed within a common skilled reduction sequence (e.g., MS2/VD/5/A and MS2/VD/5/C coexist as different branching options in a common metric sector, along a similarly Very Difficult skilled reduction sequence, and at similar stage — “simultaneity” combinatoric). IND will appear in some biface coding (e.g., MS1/D/3/IND) to mean that stages might lead anywhere. This happens in the upper right side of the reduction continuum where bifaces show much more available materials to subtract. Stages are thus connected progressively from the reduction continuum’s upper right to its lower left corner, and end products linked to possible blanks. Complete extended skilled reduction sequences can then be rebuilt from these segments. A new extended skilled reduction sequence is created each time a blank appears in the reduction continuum.
Sampling
I focus my analysis on a small locus — 16-west. It is a 350m2-wide area that was circumscribed in previous works conducted over the 45,000m2-wide areas of stations 15 and 16 (Kolhatkar, 2022a). It showed (i) a particularly high level of skill when compared to other excavated loci; (ii) a peripheral location on station 16’s northwestern flank; (iii) lesser vertical mixing; (iv) consistency with a general northward orientation of stations 15 and 16’s spatial organization 16. In addition, it contains both Sainte-Anne/Varney and pseudo-Agate Basin projectile points (11,600 to 9000 cal BP) with a complete range of bifacial preforms from various levels of technical difficulty, even if higher levels are predominant (see below). It might provide a better depositional setting than the whole of stations 15 and 16, as well as a narrower time frame of approximately 2500 years. But what is of interest here is the more manageable subset of knapping practices that this locus provides for exploring the methodological tools presented above, as well as their interpretive value. Prior analyzes of stations 15 and 16 also suggest that some level of organization was still present, that these stations and its smaller loci might not simply be random aggregates of artifacts, and that additional levels of organization might be investigated using the dispersion surfaces presented above. However, this organization still does not mean that a single ethnic group occupied La Martre over the years.
Out of a total of 164 bifacial tools, 60 unifacial tools, and 30,253 flakes, sampling was limited to 93 bifaces that were complete enough (two opposable edges must allow for recording at least two width measurements) and complex enough (they need to show at least three removal scars to attest to human manipulation). Flakes were not considered at this point because they show less complex technological information than bifaces, and physical refitting could not be conducted due to mixing. Bifaces were found all over 16-west during 0.5 × 0.5m test-pitting, visual surveys, and broader excavation works (Fig. 4). An excavation unit (area 1) 21.75m2 large and various test pits account for the 30.5m2 sampling of an estimated 350 m2 for the whole of 16-west. Bifaces come in equal part from the plowed layer (N=48) and its underlying B horizon (N=45). Artifacts can be found all over 16-west, with a slight concentration in area 1’s northwest corner. Bifaces with fresh and weathered breaks were included because they still contained valuable technological information. Out of 93 bifaces, eight are complete, 81 have weathered breaks, three have fresh breaks, and one has both fresh and weathered breaks. Out of 85 broken bifaces, 62 have one break, and 23 have two breaks.
Test pits and excavation area (16-west)
Results
Overview
Width and thickness distribution curves show two very distinct patterns (Figs. 5 and 6). They can be read right to left, following stone knapping’s subtractive process. Width distribution is close to normal, with a slight skew to the right. Two major points of inflection (50 and 30 mm) can be identified as the distribution peaks abruptly towards its highest value (43.5 mm). A smaller peak (16.5 mm) also stands out following the 30 mm point of inflection. To the right of the 50-mm inflection point, the distribution rises smoothly. Thickness distribution is not normal, with a much stronger skew to the right, and a higher kurtosis value than the width curve. There is only one major inflection point here, that follows the distribution’s gradual right tail and precedes a very abrupt rise towards the thickness’s highest value (8.4 mm).
Mean width distribution curve (skewness: .544, kurtosis: .147) (16-west)
Mean thickness distribution curve (skewness: 1.563, kurtosis: 2.6666) (16-west)
These three major inflection points (width: 30 and 50 mm; thickness: 15 mm) suggest various rhythms in the discard of bifaces. They can be transferred onto the reduction continuum to define five metric sectors (Fig. 7). Number of breaks per biface in each metric sector suggests two things (Fig. 8). First, metric sector boundaries are not an artifact of the fragment portions of the bifaces recorded. Number of breaks is the main cause in biface outline variability — bifaces become either triangular or square (see also plates in the remainder of the paper). Each metric sector contains both, which also shows why sectors should not be defined too narrowly. Second, breakage is the most common cause for discard, and it happens across all the reduction continuum. Hence, the breath of these sectors’ width and thickness defines the horizon of problems they pose to knappers who wish to enter and exit them. Inflection points mark metric value conditions for entry and exit. Many knappers can enter metric sector 4, but many do not leave it either. Sample is sparser for metric sectors in the reduction continuum’s upper right corner than it is towards the lower left. As tasks become more difficult, accidents become more frequent and discards are more numerous as a result (especially for higher levels of difficulty), or, for lesser levels, low W/T ratios quickly limit one’s field of action and prevent from further knapping.
Metric sectors in the reduction continuum (16-west)
Number of breaks by metric sector (16-west)
All bifaces are knapped in Cap-Chat chert, a homogenous material with no discernable variations in raw material quality. Very few blanks could be found (Table 1). They are spread all along the reduction continuum, suggesting various starting points for knapping procedures. End products comprise 14 Sainte-Anne/Varney, two pseudo-Agate Basin projectile points, two drills, two bifacial cores and one bipolar core (Table 1). As I have already argued, these end products remain bifaces first, because end products do not mark an endpoint in the knapping process. Large bifacial cores leave plenty of room for proceeding through stone knapping, and even projectile points might be recycled to produce drills. All emerge from a same general subtractive process, regardless of their specific chaînes opératoires, as long as sufficient materials and other technical properties (e.g., ridge networks) are present. Very difficult bifaces are more numerous. Very difficult, difficult, and slightly difficult levels crosscut the whole reduction continuum as distinct skilled reduction sequences, but easy bifaces are very few and appear punctually (Fig. 9 and Table 1).
Skilled reduction sequences in the reduction continuum (16-west)
Ten extended skilled reduction sequences and 24 stages could be defined (Table 2). Extended skilled reduction sequences have an unequal number of stages. These sequences all unfold from discernable blanks. Stages show variable W/T ratios through metric sectors (Fig. 10). W/T ratio values and qualitative attributes for each stage are indicated in Supplementary Materials 1. Sectors four and five are the densest, suggesting both a more inclusive horizon for knappers to work at and through in various ways. Table 2 orders stages by extended skilled reduction sequence and metric sectors. In the remainder of this section, I detail each stage by metric sector.
Boxplot of stages ordered by metric sector (16-west)
Metric sector 1
This sector consists of 10 bifaces, larger than 50 mm and thicker than 15 mm, as well as two tabular blanks and two bifacial cores (Figs. 11, 12, and 13, Table 1). All difficulty levels can be found here in equal proportions (Table 1). Tasks were less difficult here, even when considering “very difficult” bifaces, and will not have left many discards. As the sample is very dispersed and low for each difficulty level, no clear pattern emerges. We are very early on in the reduction process. Big blanks provide with plenty of room for all kinds of future procedures and will appear in different extended skilled reduction sequences (Table 2).
Stages from metric sector 1 (16-west)
Bifaces from metric sector 1, ordered by stage and catalog number (16-west; photographs by the author)
Diacritical schemes of bifaces from metric sector 1, ordered by stage and catalog number (16-west; drawings by the author)
MS1/E/1/IND
Two very large tabular blanks point at this surface as one of 16-west’s various starting horizons (Figs.12 and 13). They show two different ways of working out a blank at very early stages: 16-984 attests of a symmetrical, lenticular working of its faces, while 16-973 shows that one face is worked out completely before the obverse one is more fully shaped, making for a planoconvex cross-section and a massive ridge that undercuts one whole face. This might be due in part to 16-973’s greater starting thickness. Such large bifaces suggest that all reduction trajectories could develop from these two types of blanks, depending on how knappers are able to change their biface’s width relative to its thickness.
MS1/D/2/IND
Difficult bifaces could have developed at this stage. It is only visible thanks to a large, somewhat thick, and difficult biface (Fig. 12). Knapping is covering on one face, invasive on the obverse one, and work proceeds from beveled edges to produce a generally biconvex cross-section. This proximal or distal part could suggest a massive preform that is patterned much like 16-973’s asymmetrical work.
MS1/VD/3/IND
Two very difficult bifaces could have developed from difficult, stage 2 procedures, owing to their size and to the presence of cortex suggestive of tabular blanks (Figs. 12 and 13). Their lenticular cross-section is obtained by a covering knapping on both faces that proceeds in both a lateral and centripetal fashion from a beveled platform located beneath the center plane and removes a face’s central ridge.
MS1/SD/2/IND
An alternate trajectory can proceed from 16-973 or 16-984 towards narrower and thicker preforms, with slightly difficult bifaces that show a planoconvex cross-section, covering flakes on one face, and invasive flakes on the obverse one which create a massive, central ridge (Figs. 12 and 13).
MS1/D/3/IND
Three difficult bifaces suggest that the ridge obtained at MS1/SD/2/IND would have been dealt with end-thinning (Figs. 12 and 13). 16-481 shows a planoconvex cross section and a flexion break indicative of a failed end-thinning (Crabtree, 1972). 16-831 attests of a partially successful end-thinning attempt. It resulted in a massive step that would have been very difficult to correct, considering the overall convexity of the other face. Its break might have been provoked by an attempt at correcting either face. 16-572 shows a square base with a distal thinning scar that remained marginal before the break occurred.
Metric sector 2
This sector extends beyond the 15-mm thickness threshold while remaining to the right of the 50 mm width inflection point (Figs. 14, 15, and 16, Table 1). No blank or other kinds of end products can be observed, and only very difficult pieces (N=14) are present. If these bifaces indeed emerged from sector 1, they did so thanks to a very sharp turn in their width-by-thickness development, suggestive of knappers’ means to pull away efficiently and quickly from a blank’s natural properties. Two extended skilled reduction sequences can be followed here (Table 2).
Stages from metric sector 2 (16-west)
Bifaces from metric sector 2, ordered by stage and catalog number (16-west; photographs by the author)
Diacritical schemes of bifaces from metric sector 2, ordered by stage and catalog number (16-west; drawings by the author)
MS2/VD/4/A
The main sequence is illustrated by nine bifaces (Figs. 15 and 16). They show lateral thinning, covering on one face, very invasive to almost covering with a slight central ridge on the obverse one. Thicker bifaces with two beveled edges can also be found. Knapping procedures are thus generally similar to those described for this sequence’s first three stages, suggesting that work proceeds in a very repetitive fashion while allowing for metric breaks, until adequate volumetric properties have been met and new operations can be implemented. A complex ridge network is already emerging on some of these bifaces, suggesting that this technique works all four dimensions of a biface simultaneously (surface, cross-section, platform, W/T ratio). The spread of this sequence suggests either that these bifaces follow additional, smaller steps within a same general stage by repeating the same knapping procedures, or that they reached this stage from blanks of slightly varying starting size. Regardless of which, from this point onward, thickness is held constant, but width is reduced until the main 50 mm width inflection point is reached.
MS2/VD/4/B
A second set of five bifaces differs from the first one (Figs. 15 and 16). They are larger and as thin, or even thinner. Covering flakes were removed from this early stage on one face, sometimes both, from all directions around the biface. Transversal flakes allow one to produce flatter faces and to remove much thickness as they can travel farther, but they are at the same time more difficult to enact (Chauchat & Pelegrin, 2005). The main problem here is that thickness is being reduced too fast relative to the biface’s width. This might lead a knapper in a dead end (Waldorf, 1993). That is, if producing Plano projectile points was their main objective. But considering that knappers show a very high control of their knapping, and that some parts of these bifaces are even thinner than what their mean thickness suggests, they were good enough to have known this. Thus, we must consider that knappers were in fact exploring the limits of their raw materials and pushing against their very own limits to produce some of the most impressive bifaces in La Martre’s assemblage.
Metric sector 3
Some of sector 1’s bifaces might have developed towards sector 3’s narrow and thick shapes and carried on a low W/T ratio beyond the 50-mm width flexion (Figs. 10, 17, and 18). Seven slightly difficult bifaces make up this sector (Table 1). One of them is a flake blank (Fig. 18 — 16-853), another could have been used as a bipolar core (Figs. 18 — 16-1004). Two reduction sequences can be suggested that could not have been carried through to the following metric sector due to too low W/T ratios (Fig. 10, Table 2).
Stages from metric sector 3 (16-west)
Bifaces from metric sector 3, ordered by stage and catalog number (16-west; photographs by the author)
MS3/SD/1/B
16-853 (Fig. 18) marks a new initial stage because it is a barely worked blank. It also suggests that other types of blanks could be used. It is a very large and thick flake with an extremely thick edge. This would make for a very difficult blank to start with, however. Both faces are thinned, and the edge acquires a relatively central and regular shape. A covering operation that thins the biface’s median section is noticeable as well, but thinning did not get very far before the biface got discarded in still workable (unbroken) form.
MS3/SD/3/A
This stage would proceed from a prior MS1/E/1/IND and MS1/SD/2/IND stage, but run in parallel to MS1/D/3/IND. Indeed, rather than using end-thinning, thinning would have remained lateral here, and knappers would not have been able to change their W/T ratio. Some of these bifaces show some retouch that started developing a ridge network (Fig. 18 — 16-516, 16-622). 16-1005 (Fig. 18) shows distal thinning on a very narrow biface, although this did not help change its W/T ratio.
Metric sector 4
This sector is comprised of 33 bifaces located below the 15-mm thickness and left of the 50 mm width inflection points (Figs. 19, 20, 21, 22, and 23, Table 1). Very difficult bifaces are the most numerous here. There is one flake blank here as well, but no identifiable end products. Six extended skilled reduction sequences crosscut this sector (Table 2). These sequences attest to various W/T ratio (Fig. 10) and technological developments from earlier stages located either in metric sectors 1 or 2, or to new entries thanks to the use of smaller blanks. Contrary to metric sector 2, this sector is more inclusive of various levels of skills.
Stages from metric sector 4 (16-west)
Bifaces from metric sector 4, ordered by stage and catalog number (16-west; photographs by the author) (1/2)
Bifaces from metric sector 4, ordered by stage and catalog number (16-west; photographs by the author) (2/2)
Diacritical schemes of bifaces from metric sector 4, ordered by stage and catalog number (16-west; drawings by the author) (1/2)
Diacritical schemes of bifaces from metric sector 4, ordered by stage and catalog number (16-west; drawings by the author) (2/2)
MS4/VD/5/A
These bifaces (Figs. 20 and 22) show a general, asymmetrical manufacturing technique: one face is covered with fine, lateral, covering scars; the obverse face is more broadly knapped, with larger scars and sometimes more profound negative bulbs, and flakes have been detached either laterally, distally, or diagonally. All cross-sections are lenticular. They extend the features already encountered at the prior, MS2/VD/4/A stage.
MS4/D/4/A
These narrower and thicker bifaces (Figs. 20 and 22) extend from sector 1’s MS1/AD/2/IND and MS1/D/3/IND, where this kind of problem was worked at through distal thinning. The stages in sector 4 would proceed from successful operations conducted in sector 1 (Table 2). Indeed, end-thinning operations can be observed on some of the bifaces from this stage, with a combination of lateral thinning. In addition, work that creates regular ridges seems focused on one face only. It is also noteworthy that thick edges are protruding. As experimentation has shown (Callahan, 1979), a thick edge can indeed provide with a stronger platform for thinning one’s biface. However, this technical solution eventually becomes a problem that unbalances the volumetric properties of one’s work.
MS4/VD/5/C
This stage (Figs. 21 and 22) would proceed from the above, MS4/D/4/A. Difficulty increases here because regularity must be worked in a tighter — narrower and thicker — space. Thus, while work is covering on at least one face, and creates a lenticular cross-section like what could be observed on MS4/VD/5/A, it also seems to be carrying from its prior stages a thick edge that could not be satisfactorily corrected. This goes in line with a lesser regularity of the ridge network when compared to bifaces from stage MS4/VD/5/A. An overall greater thickness (Fig. 10), caused by more accidents that failed to subtract materials, to straighten the edges or the cross-section, and to remove a thick edge, is another important difference with MS4/VD/5/A. Work that proceeds through metric sector 2 rather than proceeding directly from metric sector 1 such as here, would not have produced such a problem, in part due to very regular volumetric properties being initiated earlier on.
MS4/VD/5/D
A small sample of bifaces appears to be larger and even thinner (Fig. 21 and 23) than most bifaces when considering their true thickness values rather than their mean width, all the while using the same asymmetrical knapping of both faces. These very difficult bifaces of an even higher level of difficulty could suggest the same as surface 2’s alternative sequence (MS2/VD/4/B), that is, an exploration of one’s limitations: one’s skills, materials, along with the limits that a technique can afford, but with added difficulty when trying to progress through the remainder of the reduction continuum.
MS4/E/1/A and MS4/SD/1/A
These two segments can be regrouped because they show a similar late entry into the continuum (Fig. 20). One slightly knapped large flake (Fig. 20 — 16-612) could be found that shows the same long and narrow morphological plan view properties as some of the other difficult and very difficult bifaces from this sector. Although it is interesting to note that blanks of similar properties might have been selected, it also seems very unlikely that good knappers started from such a flake, owing to the lack of room that it provides for creating the slightly convex surfaces and ridge network necessary to proceed as far as possible in the reduction continuum.
MS4/VD/6/A, MS4/VD/6/C, and MS4/D/5/A
Although we are dealing here with three alternate segments, I will consider them together because they show a similar, general knapping technique (Figs. 20, 21, and 23). Indeed, all show lenticular cross-sections and relatively narrow outlines. They probably proceeded from earlier surface 4’s stages. The main knapping strategy that could be identified consists of comedial knapping (sensu Bradley, 1993) with regularly spaced ridges on one face and covering flakes on the observed face (Figs. 21 and 23 — 16-587). Steps on a very regular biface (Figs. 21 and 23 — 16-720) would suggest that the knapper tried to use that method but only managed to remove an invasive flake. Another method consists of invasive flaking on both faces, although a high number of hinged scars might have prevented from using a more covering flaking (Figs. 20 and 22 — 16-472). The difference between these stages resides in their overall regularity and thinness, which is greater in stage MS4/VD/6/A, and lower in the other two. This is in line with these bifaces’ prior stages, which would have taken them through various metric sectors and opportunities for development and regularization (Table 2). It might result for example in producing bifaces that are regular on one face only, due to convexities or material irregularities that were not corrected earlier on (Figs. 21 and 23 — 16-705; Fig. 20 and 22 — 16-472). Another important difference between them is the use of retouch on very difficult bifaces to calibrate protruding ridges on the edges. While this point in the reduction continuum might be reached in various ways, it might not necessarily produce bifaces fit for further work. Proceeding without advanced retouch might result in thicker and less standardized projectile points.
Metric sector 5
This sector is located left of the 30-mm width and below the 15-mm thickness inflection point. It is made up of 29 bifaces, with very difficult bifaces being predominant (Figs. 24, 25, and 26, Table 1). Plano projectile points appear here (N=16), along with two drills (Fig. 25 — 16-400, 16-697) and one flake blank (Fig. 25 — 16-400). Six reduction sequences can be followed through this sector (Table 2).
Stages from metric sector 5 (16-west)
Bifaces from metric sector 5, ordered by stage and catalog number (16-west; photographs by the author)
Diacritical schemes of bifaces from metric sector 5, ordered by stage and catalog number (16-west; drawings by the author)
MS5/VD/7/A
Retouch initiated at this reduction sequence’s prior stage takes an increasing importance here (Figs. 25 and 26), in that it shapes very advanced preforms for Sainte-Anne/Varney projectile points with fine parallel, pressure retouch, and flat cross-sections. Bifaces show an asymmetrical retouch pattern, where one face has wider scars than the obverse one, along with a slightly planoconvex cross-section.
MS5/VD/8/A
Plano projectile points are appearing in this sector in the guise of both Sainte-Anne/Varney and pseudo-Agate Basin types (Figs. 25 and 26). They show lenticular and symmetrical cross-sections with comedial pressure retouch on both faces. Retouch by pressure is applied in a regular fashion on Sainte-Anne/Varney projectile points. Slight variations can be noted among the latter points. The most usual technique produces a slightly asymmetrical cross-section: one face is flatter than the other thanks to covering flakes from a slightly beveled platform; the obverse face is slightly convex with comedial knapping (Figs. 25 and 26 — 16-480). The other observable technique shows lenticular and symmetrical cross-sections with comedial retouch by pressure on both faces (Fig. 25 and 26 — 16-881). Another procedure produces two very flat faces, thanks to large and covering flaking on both faces, before pressure retouch is applied to both faces while creating a slightly less regular ridge network (Figs. 25 and 26 — 16-774).
MS5/D/9/B
This stage could stem from various procedures from which drill blades can be produced (Fig. 25, Table 2). They can be shaped either from flakes (MS4/E/1/A and MS5/E/1/B), or from recycling projectile points (MS5/VD/8/A and MS5/VD/8/C).
MS5/D/6/A
Thinning to reach this shape could have proceeded from an earlier MS4/D/5/A stage (Fig. 25, Table 2). However, ridges were too far apart, and the face was not convex enough to provide with the pressure retouch necessary to proceed with knapping.
MS5/SD/2/A
Other slightly difficult bifaces seem a little bit more advanced because their starting blank cannot be identified (Fig. 25). Thick flakes from earlier MS4/E/1/A or MS4/SD/1/A stages could have been used to produce these shapes (Table 2).
MS5/E/1/B
Slightly knapped flakes suggest another short-circuit later into the reduction continuum to produce drills (Fig. 25, Table 2).
Discussion
Much variability can be observed all along the reduction continuum. The 10 extended skilled reduction sequences are rhythmed by various stages, W/T ratio developments, as well as blanks and end products. The question is how and why these extended reduction sequences emerged in the first place. Answering this requires that (i) we tone down the deterministic role that material constraints can have in archaeological explanations, and that (ii) we consider the broader environmental and social context that knappers lived in thanks to the interpretive power of guided participation, continuity through shared activities, and social scaffolding concepts. These processes can be addressed to flesh out the patterns uncovered at 16-west because the dual levels of practice that they entail were also part of the descriptive architecture used. As a reminder, guided participation refers to a relation between lesser and better knappers. Continuity through shared activities points at similarities that extend over time (e.g., lithic, chronological) in spite of differences. Scaffolding is the social structure that affords for novice, expert, and intermediate knappers to learn, improve, and test some critical material thresholds without risking endangering themselves or their group (from raw materials availability to the time diverted from other subsistence related activities). I will address them in turn.
Guided participation
Metric sector 1 attests of failures at very early stages, most probably by novice knappers given how easy such tasks are, and that no incipient fracture plane might have caught advanced knappers off guard. These discards show us how more advanced knappers might have initiated their knapping, because while they will seldom fail at this early stage, they will nonetheless be emulated by learners. Second, there might be little difference between novices, advanced or even expert knappers at this point, except for the fact that lesser skilled knappers failed at more difficult tasks. Third, sequence divergence at such an early metric sector suggests that knappers could pull free relatively quickly from a blank’s natural properties. Fourth, narrowing a biface too early did not bring an immediate dead end. Rather, it simply set the conditions for future developments.
These developments already mean that extensive knapping knowledge could have widened the range of blanks which knappers could choose from at the nearby quarry. Knappers changed the initial properties of the blank to create a new situation and constraints of their own doing that they had to adapt to in order to further work their biface. In Jacques Pelegrin’s words (Pelegrin, 1985), knowledge was at least available as novices saw what more knowledgeable others did, but know-how had to be acquired to successfully enact certain procedures. Indeed, knappers of various proficiency need not necessarily knap alone. To the contrary, quarries and nearby workshops such as 16-west and the whole of La Martre are ideal playgrounds for beginners to learn and improve their skill as they are less likely to endanger their group by wasting precious raw material. Furthermore, such skill is best acquired and honed under the supervision of a more knowledgeable other whose counsel, advice and cues not only educate the attention of the pupil but accelerate it by narrowing one’s mistakes and routes taken to accomplish certain goals. Skilled and less skilled knappers pattern profiles are then expected to make up such places because they necessarily interacted.
Various extended skilled reduction sequences suggest two interaction scenarios. On the one hand, this interaction can take place from one stage to the next on the same biface, such as with extended skilled reduction sequence 3. Removing protruding central ridges with end-thinning flakes is a difficult procedure that contrasts with the less skilled lateral removals found on bifaces in metric sector 3. These lateral removals could have been used to calibrate a biface and facilitate the following end-thinning move. This strategy could be more adapted to blanks that were either naturally narrower and thicker, or made so by novices. It would have then been given back to more skilled knappers for this more difficult procedure. It keeps a steady width and thickness balance with less impressive but safer technical strategies than extended skilled reduction sequence 1. It is more inclusive of a wider array of skill levels.
On the other hand, one can also keep to themselves. They must eventually learn to enact these more difficult tasks on their own, practising and improving certain procedures by working through various levels of technical difficulty. Trial-and-error starts early on, with metric sector 1, where invasive to covering flakes were attempted on thick initial blanks rather than systematic lateral preparation for end-thinning (MS1/SD/2/IND, MS1/D/2/IND, MS1/D/3/IND, MS1/VD/3/IND). These attempts sometimes failed, producing steps or hinged endings, exacerbating the size of the central ridge in relation to the periphery of the biface, and narrowing it without thinning it properly.
Some knappers managed to exit metric sector 1 and carry on further, as stages found in metric sectors 2 and 4 attest. But while skill levels are mostly very high in metric sector 2, various levels and extended skilled reduction sequences overlap in sector 4, meaning they were enacted by knappers with various skill sets. These overlaps can take various forms, from broad shared metric (width only) properties without clear technological similarities, to closer metric (W/T ratios) and qualitative (scar patterning, cross-section) similarities. We have seen that the same general knapping method overlapped in metric sectors 4 over difficult and very difficult bifaces (MS4/VD/5/A, MS4/VD/5/C, MS4/VD/6/A, MS4/VD/6/C), while differences in execution created various surface and edge (ir)regularities as well as various W/T ratios. The difference between difficult and very difficult levels of technical difficulty is not actualized by a “special technique” that would be specific to higher experts, but by a higher regularity and control of a general technique that allows one to produce more regular and thinner bifaces. Such regularity also stems from preceding stages found in metric sector 2. Less advanced knappers could start practicing some advanced procedures such as covering flakes early on, but with less success because they did not yet control a long enough segment of their reduction sequence. True experts can hold higher execution levels for a longer while, a feat that must be acquired through repeated practice.
Continuity through shared activities
One removal at a time, knappers forego initial blank properties and set their own material constraints towards a more standardized shape. Certain mechanical requirements needed to produce specific shapes may have proved to be much more powerful constraints than initial blanks were. They might be found anywhere in the reduction continuum without being restricted to the reproduction of a specific projectile point type through time. Hence, continuity through shared activities can certainly encompass projectile points production. But what matters most, and that surfaces of dispersion help reveal, are the similarities that emerge along the way and that recur despite the many differences that knappers enact.
Two kinds of similarities can be found at 16-west. The first and more straightforward one is that extended skilled reduction sequence 1 provides with a sequence that was maintained continuously high across different metric sectors. It also encompasses more bifaces than other sequences. Experts’ technical gestures are better controlled and calibrated, from the kind of blank selected for certain goals, to the kind of projectile point achieved, and to the continuously high level maintained through a whole knapping sequence. They are better at standardizing their production from one attempt to the next. In addition, since extended skilled reduction sequences crosscut the reduction continuum through various places, sequences that maintain a continuously very difficult level of difficulty were not “necessary” to work at bifacial knapping. The fact that unnecessarily complex procedures were nonetheless enacted further underlines their special status as a standard sequence for the production of Plano projectile points at 16-west. This sequence would provide a template to guide others, whether they shared the same place at the same time or worked through some of the discards that experts left behind in the same way archaeologists do.
Another, maybe more important kind of similarity refers to the tasks that endured regardless of the knapping objective that might have motivated knappers. This is all the more important because we are dealing here with an archaeological palimpsest. Knappers could have occupied the same place at various moments and brought different needs and backgrounds with them. Similarities might have been tied by its knappers to a specific Plano projectile point template. They could have been consciously attempting to reproduce such a template. But whatever objective they were trying to achieve, they had to proceed towards it, they failed, and their resulting discards crystallized similarities that endured through repeated attempts. This similarity can thus be distinguished from the original intention that justified it to organize knapping practices according to what does not change through time.
In that case, producing Plano projectile points would certainly have required a specific set of width, thickness, cross-section shape and flake scar criteria to be achieved. At the same time, such points also come in various sizes that can leave some room for various kinds of knapping procedures. All knappers, however, would have had to align their work below the 50 × 15 mm inflection point — the most important sector in 16-west’s general reduction sequence. Metric sector 4 could be identified as a similarity because it was enacted along knappers’ various attempts at enacting it and failing in it. The coexistence of various extended skilled reduction sequences shows various understandings of common principles. Such a knapping objective took its toll on manufacturing procedures, as the many discards attest to. The very dispersion of thickness values and extended skilled reduction sequences in this sector shows creating these properties may have been attempted by most but could not be successfully accomplished by everyone. Rather, different knappers reached this width along various thickness values and technical strategies, and often broke their biface while thinning it closer to 8,4 mm peak thickness values. Lesser skilled knappers could achieve an appropriate width and outline for their biface, as it is one of the easiest knapping aspects one can master. Appropriate thickness with width and possibly precise scar patterns were another matter, however. Along this stricter constraint, we can see various enactments, from attempts, improvisations even, to further thin an already too narrow biface, to the skilled achievement that producing Plano projectile points means.
Scaffolding
Experts should neither be abstracted from their social context nor idealized. They no more blindly replicate their knowledge than lesser knappers do. Skilled knappers are freer from material constraints and can express certain preferences by transforming materials better than novices. Highly standardized tasks show many differences. Finer grained knapping differences (flake patterns, cross-section, and edge shape) found all along their very difficult stages could be understood as preferred ways of doing envisioned and imposed from the very start. In addition, there is some variability within available Plano projectile points. There are different types (Sainte-Anne/Varney and pseudo-Agate Basin), different sizes (thickness), and different technological properties (asymmetrical vs. symmetrical knapping). Manufacturing patterns seen from the very start of the reduction continuum could attest to precise and distinct ways of doing (e.g., enacting asymmetrical vs. symmetrical knapping early on). Sainte-Anne/Varney and pseudo-Agate Basin projectile points would not have needed entirely different knapping procedures. Some very skilled knappers even thinned their bifaces heavily while keeping them wide, although this was not a necessary condition from a strictly technofunctional viewpoint (MS2/VD/4/B and MS4/VD/5/D). They do not provide with useful preforms for obtaining Plano projectile points, and they pose a new level of difficulty within an already very difficult horizon of tasks. And yet they can be found, albeit in small numbers.
These patterns can be understood as personal preferences that expert knappers indeed have the skill to explore and produce. But we can push the meaning of such practices a bit further by replacing such preferences within the broader context that made them possible. La Martre is an affordable setting, thanks to its nearby chert outcrops; flat, expansive, and well-drained sandy terraces; easy access to sea and river resources and axes of communication; a nearby ice cap that would have attracted caribou populations. This affords for repeated occupations as much as seasonal collective aggregation. Whether such performances were part of a teaching strategy, were enacted for playing purposes, or were enacted to reassert one’s status as an expert craftsperson, the important thing is that such performances could have been sighted by many, while being less desirable farther from the quarry where such operations could also be riskier. Consequently, following Hiscock’s (2014) proposal, we should consider whether this was the public demonstration of skilled procedures that require a high investment in time, effort and resources and requires the social scaffolding to make this investment possible (other people take care of subsistence related activities) and worthwhile (other people can see these performances and value them). Stone knapping prowess would have had to be meshed in a sociocultural framework that valued it.
This scenario can be considered in two ways. First, while little is known about Late Paleoindian knapping practices, knapping prowess is well known from cache sites dating to the Early Paleoindian period (Kilby, 2008). People also preferred high quality lithic raw materials to such an extent that they are often found far from their source — a pattern that does not conform with regional raw material decline to the source found in later periods (Burke, 2006; Ellis & Lothrop, 1989; Lothrop et al., 2016; Meltzer, 1988, 1989). Knapping patterns from 16-west could extend the importance of stone knapping to the Late Paleoindian period. These patterns could also be used to tie La Martre to other, better-known sites from this phase in a stronger way than projectile point typology alone allows for.
Second, if knappers are free to produce various types of Plano projectile points until very late in the reduction continuum, then the need to produce the Plano projectile point subtypes must be addressed as well. As a reminder, the apparition of unfluted projectile points is correlated with the end of the Young Dryas and the Early Holocene (Lothrop et al., 2016; Newby et al., 2005). A general warming trend may have led to a reorganization of the regional vegetation and prey species populations requiring technological adaptation. Correlation may not imply direct causation here, however. It remains unclear whether choosing such a point (and more generally, any kind of point) was severely constrained by very specific adaptive needs and changing environmental conditions, or if it was itself one choice out of many functionally equivalent possibilities (Ellis, 1997). However, the expansion of Plano projectile points covers a vast territory made of various biomes and two separate regional clusters of sites (Lothrop et al., 2016). This would imply other non-strictly adaptive factors to understand how people coped with such varying environmental conditions to pool adaptive risk (Schortman, 1989). It could hint at this point type serving as an emblemic style (sensu Wiessner, 1983, 1984) of regional integration for the various groups living in northeastern North America – the more so, if this emblemic style was reproduced in spite of various personal and local differences in production templates. These scenarios, however, hit the limits of what can be addressed from 16-west alone.
Conclusion
In this paper, I have shown how sociocultural processes drawn from a communities of practice framework — guided participation, continuity through shared activities, and affordance — might be used in the archaeological palimpsest of La Martre to analyze its lithic practices. I argued that such dynamics do not need to be tied to a high-resolution depositional context. This limitation stems from normative, genealogical, and chronological expectations regarding past (e.g., stone knapping) and present (e.g., description) practices. An ecological approach to skill was used to release these processes from such expectations and generate an alternate baseline for building knowledge and for comparing artifacts in a palimpsest. In turn, these processes informed this ecology to update it with a dual level of description necessary to producing structures that have sociocultural significance despite the mixed setting they grew from. I have worked through the methodological implications of this ecology at a small locus from La Martre — 16-west — where dispersion surfaces and skill combinatorics allowed for the development of extended skilled reduction sequences. I then used guided participation, continuity through shared activities, and affordance to explore the meaning of 16-west’s patterns further.
Palimpsests confront archaeologists to their underlying epistemological and ontological assumptions regarding how a scale is defined, how time should be mapped, how practices are framed, and how collectives are constituted. The conceptual (theoretical and methodological) frameworks that rationalize these assumptions generate errors when they are used to translate palimpsests back into intelligible and familiar terms. Palimpsests require that we return to fundamentals that are broad and true enough to include them in our narratives — before conceptual bifurcations rationalized their exclusion.
In that case, skill forefronts tasks regardless of any chronological, spatial, and cultural distance between stone knappers. Similar problems bring knappers together, and different answers to these problems can bring them even closer because they reveal similarities even more clearly. This idea underlies the work conducted to push through the constraints posed by a dense and mixed assemblage and reveal patterns with sociocultural significance, while showing that some sets of assumptions were not necessary to include La Martre’s knappers into an archaeological narrative. These constraints are not limited to the La Martre palimpsest, however. Similar problems are posed to archaeologists in other depositional settings and across various scales. Northeastern North America’s Paleoindian is one such gargantuan palimpsest. An ecology of skill could help reorganize it and explore its sociocultural dynamics further by drawing on its wide range of depositional scales.
Data Availability
Lithic assemblage (La Martre, DhDm-1) lent by the Laboratoire et Réserve d'Archéologie de Québec, and housed at Université de Montréal, Département d’Anthropologie (C-3103) for the duration of this study.
References
Anderson, D. G. (1995). Paleoindian Interaction Networks in the Eastern Woodlands. In M. S. Nassaney & K. E. Sassaman (Eds.), Native American Interaction: Multiscalar Analyses and Interpretations in the Eastern Woodlands (pp. 1–26). Knoxville: University of Tennessee.
Andrefsky, W. (2005). Lithics: Macroscopic approaches to analysis. Cambridge University Press.
Asnong, H., & Richard, P. J. H. (2003). La Végétation et le Climat Postglaciaires du Centre et de l'Est de la Gaspésie, au Québec. Géographie physique et Quaternaire, 57(1), 37–63.
Audouze, F., & Karlin, C. (2017). 70 years of "Chaîne Opératoire": What French prehistorians have done with it [La Chaîne Opératoire a 70 Ans : Qu'en Ont Fait Les Préhistoriens Français]. Journal of Lithic Studies, 4(2), 5–73.
Bailey, G. (1983). Concepts of Time in Quaternary Prehistory. Annual Review of Anthropology, 12(1), 165–92.
Bailey, G. (2007). Time perspectives, palimpsests and the archaeology of time. Journal of Anthropological Archaeology, 26(2), 198–223.
Bamforth, D. B. (2007). The Allen site: A Paleoindian camp in Southwestern Nebraska. University of New Mexico Press.
Bamforth, D. B., & Becker, M. (2000). Core/biface ratios, mobility, refitting, and artifact use-lives: A Paleoindian example. Plains anthropologist, 45(173), 273–290.
Bar-Yosef, O., & Van Peer, P. (2009). The Chaîne Opératoire approach in Middle Paleolithic archaeology. Current Anthropology, 50(1), 103–131.
Bateson, G. (1972). Steps to an Ecology of Mind. Chandler Publishing Company.
Bateson, G. (1979). Mind and nature - A necessary unity. Bantam Books.
Benmouyal, J. (1987). Des Paléoindiens aux Iroquoiens En Gaspésie: 6000 Ans D'histoire. MAC.
Bergson, H. L'Évolution Créatrice. Quadrige. 9e edition. Paris, France: PUF, 2001 (1907).
Bernstein, N. A. "On dexterity and its development." In Dexterity and Its Development, edited by Mark L. Latash and Michael T. Turvey, 1-235. Mahwah, N.J.: L. Erlbaum Associates, 1996.
Binford, L. R. (1981). Behavioral archaeology and the "Pompeii Premise". Journal of Anthropological Research, 37(3), 195–208.
Biryukova, E., & Bril, B. (2002). Bernstein Et Le Geste Technique. Technologies, Idéologies, Pratiques: Revue d'Anthropologie des Connaissances, XIV(2), 49–68.
Boëda, E. (2001). Détermination Des Unités Techno-Fonctionnelles De Pièces Bifaciales Provenant De La Couche Acheuléenne C’3 Base Du Site De Barbas I. In D. Cliquet (Ed.), Actes De La Table-Ronde Internationale Organisée à Caen, Les Industries À Outils Bifaciaux Du Paléolithique Moyen D’europe Occidentale (pp. 51–75). Liège: ERAUL.
Bradley, B. A. (1993). Paleo-Indian flaked stone Technology in the North American high plains. In O. Soffer & N. D. Praslov (Eds.), From Kostenki to Clovis: Upper Paleolithic-Paleo-Indian Adaptations (pp. 251–262). Plenum Press.
Bradley, J. W., Spiess, A., Boisvert, R., & Boudreau, J. (2008). What’s the point?: Modal forms and attributes of Paleoindian bifaces in the New England-maritimes region. Archaeology of Eastern North America, 36, 119–172.
Bril, B. "L'apprentissage De Gestes Techniques: Ordre De Contraintes Et Variations Culturelles." Technologies, Idéologies, Pratiques: Revue d'Anthropologie des Connaissances XIV, Le geste technique, no. 2 (2002a): 113-49.
Bril, B. (2002b). Apprentissage Et Contexte. Intellectica, 2(35), 251–268.
Bril, B., Roux, V., & Dietrich, G. (2005). Stone knapping: Khambat (India), a unique opportunity? In V. Roux & B. Bril (Eds.), Stone Knapping - the Necessary Conditions for a Uniquely Hominin Behaviour (pp. 53–71). McDonald Institute for Archaeological Research, University of Cambridge.
Burke, A. L. "Paleoindian Ranges in Northeastern North America based on lithic raw materials sourcing." In Notions De Territoire Et De Mobilité. Exemples De L'europe Et Des Premières Nations En Amérique Du Nord Avant Le Contact Européen., edited by C. Bressy, Ariane Burke and H. Martin. Actes De Sessions Présentées Au Xe Congrès Annuel De L'association Européenne Des Archéologues (Eaa), 77-90. Liège: ERAUL, 2006.
Burke, A. L. (2007). Quarry source areas and the organization of stone tool technology: A view from Quebec. Archaeology of Eastern North America, 35, 63–80.
Burke, A. L. (2010). Extraction, reduction, and production at a Late Paleoindian Chert quarry in Eastern Quebec. In M. Brewer-Laporta, A. L. Burke, & D. Field (Eds.), Ancient Mines and Quarries - a Trans-Atlantic Perspective (pp. 85–96). Oxford Books.
Callahan, E. (1979). The basics of biface knapping in the eastern tradition: A manual for flintknappers and lithic analysists. Archaeology of Eastern North America, 7(1), 1–180.
Cannon, A. (2011). Introduction. In A. Cannon (Ed.), Structured worlds: The archaeology of hunter-gatherer thought and action (pp. 1–9). Routledge.
Chalifoux, E. (1999). Late Paleoindian occupation in a coastal environment. A perspective from La Martre, Gaspé Peninsula, Québec. Northeast Anthropology, (57), 69–79.
Chalifoux, E., & Tremblay, R. (1998). Entre les Monts Chic-Chocs et la Mer de Goldthwait: l'Adaptation des Premiers Occupants de la Vallée de La Martre, Synthèse Archéologique. Université de Montréal, MCC.
Chapdelaine, C., ed. Il y a 8000 Ans À Rimouski… Paléoécologie et Archéologie d’un Site de la Culture Plano, Paléo-Québec 22. Montréal: Recherches Amérindiennes au Québec, 1994.
Chapdelaine, C. (1996). Réflexion sur l’Ancienneté du Peuplement Initial du Québec à Partir de Nouveaux Indices Matériels du Paléoindien Récent de la Région de Rimouski, Québec. Géographie physique et Quaternaire, 50(3), 271–286.
Chapdelaine, C. "Conclusion. La Place du Site Kruger 2 au Paléoindien Récent dans le Nord-Est de l'Amérique du Nord." Chap. 13 In Kruger 2: Un Site Du Paléoindien Récent À Brompton, edited by C. Chapdelaine and É. Graillon. Paléo-Québec, 275-93. Montréal: Recherches amérindiennes au Québec, 2020.
Chapdelaine, C., & Richard, P. J. H. (2017). Middle and Late Paleoindian adaptation to the landscapes of Southeastern Québec. PaleoAmerica, 3(4), 299–312.
Chauchat, C., & Pelegrin, J. (2005). Projectile Point Technology and Economy : A Case Study from Paiján, North Coastal Peru. Center for the Study of the First Americans.
Crabtree, D. E. An introduction to flintworking. Occasional Papers of the Idaho State University Museum. Vol. 28, 1972.
Deleuze, G., & Guattari, F. (1980). Mille Plateaux. Les Éditions de Minuit.
Deller, D. B., & Ellis, C. J. (1992). Thedford Ii, a Paleo-Indian site in the Ausable River watershed of Southwestern Ontario. Museum of Anthropology, University of Michigan.
Descola, P. Par-Delà Nature Et Culture. Bibliothèque Des Sciences Humaines. Paris: Gallimard, 2005.
Dibble, H. L., Holdaway, S., Lin, S. C., Braun, D. P., Douglass, M. J., Iovita, R., McPherron, S. P., Olszewski, D. I., & Sandgathe, D. (2017). Major fallacies surrounding stone artifacts and assemblages. Journal of Archaeological Method and Theory, 24, 813–851.
Dionne, J.-C., & Coll, D. (1995). Le Niveau Marin Relatif Dans La Région De Matane (Québec), De La Déglaciation À Nos Jours. Géographie physique et Quaternaire, 49(3), 363–380.
Dumais, P. (2000). The La Martre and Mitis Late Paleoindian Sites : A reflection on the peopling of Southeastern Québec. Archaeology of Eastern North America, 28, 81–112.
Dunnell, R. C., & Simek, J. F. (1995). Artifact size and plowzone processes. Journal of Field Archaeology, 22(3), 305–319.
Dyke, A. S. (2005). Late quaternary vegetation history of Northern North America based on pollen, macrofossil, and faunal remains. Géographie Physique et Quaternaire, 59(2-3), 211–262.
Ellen, R. (2006). The categorical impulse - Essays in the anthropology of classifying behaviour. Berghahn Books.
Ellis, C. (1997). Factors influencing the use of stone projectile tips - An ethnographic perspective. In H. Knecht (Ed.), Projectile Technology. Plenum Press.
Ellis, C. J., Carr, D. H., & Loebel, T. J. (2011). The younger Dryas and Late Pleistocene peoples of the great lakes region. Quaternary International, 242, 534–545.
Ellis, C. J., and D. B. Deller. An Early Paleo-Indian Site near Parkhill, Ontario. Mercury Series. Hull: Canadian Museum of Civilization, 2000.
Ellis, C. J., & Lothrop, J. C. (1989). Eastern Paleoindian lithic resource use. Investigations in American Archaeology. Westview Press.
Ferguson, J. R. (2003). An experimental test of the conservation of raw material in flintknapping skill acquisition. Lithic technology, 28(2), 113–131.
Fiedel, S. J. (2014). Abrupt changes of climate and of point styles along the Atlantic seaboard of Eastern North America: how were the connected? Archaeology of Eastern North America, 42, 77–100.
Goldstein, S. T. (2019). Knowledge transmission through the lens of lithic production: a case study from the Pastoral Neolithic of Southern Kenya. Journal of Archaeological Method and Theory, 26, 679–713.
Gosselain, O. P. (2000). Materializing identities: An African perspective. Journal of Archaeological Method and Theory, 7(3), 187–217.
Harding, J. (2005). Rethinking the Great Divide: Long-term structural history and the temporality of event. Norwegian Archaeological Review, 38(2), 88–101.
Hill, J. N., & Evans, R. K. (1972). A model for classification and typology. In D. Clarke (Ed.), Models in Archaeology (pp. 231–273). Methuen.
Hiscock, P. (2014). Learning in lithic landscapes: A reconsideration of the hominid “Toolmaking” niche. Biological Theory, 9, 27–41.
Holdaway, S., & Wandsnider, L. (2008). Time in archaeology : An introduction. In S. Holdaway & L. Wandsnider (Eds.), Time in Archaeology: Time Perspectivism Revisited (pp. 1–12). Utah University Press.
Ihuel, E., and J. Pelegrin. "Du Jura Au Poitou En Passant Par Le Grand-Pressigny: Une Méthode De Taille Et Des Poignards Particuliers Vers 3000 Av. J.-C.". In Les Industries Lithiques Taillées des IVe et IIIe Millénaires en Europe Occidentale, edited by M. -H. Dias-Meirinho, V. Léa, K. Gernigon, P. Fouéré, F. Briois and M. Bailly, 135-82. Oxford: BAR, 2008.
Ingold, T. The Perception of the Environment: Essays in Livelihood, Dwelling and Skill. : Routledge, 2000.
Ingold, T. (2007). Lines: A Brief History. Routledge.
Ingold, T. (2011). Being alive: Essays on movement, knowledge and description. Routledge.
Inizan, M.-L., Reduron, M., Roche, H., & Tixier, J. (1995). Technologie De La Pierre Taillée. CNRS.
Jackson, L. J. (2004). Changing our views of Late Palaeo-Indian in Southern Ontario. In L. J. Jackson & A. Hinshelwood (Eds.), The Late Palaeo-Indian Great Lakes: Geologial and Archaeological Investigations of Late Pleistocene and Early Holocene Environments (pp. 25–56). Canadian Museum of Civilization.
Karlin, C., & Julien, M. (2019). An Autumn at Pincevent (Seine-Et-Marne, France): refitting for an ethnographic approach of a Magdalenian settlement. Archaeological and Anthrropological Sciences, 11, 4437–4465.
Keller, C. M., & Keller, J. D. (1996). Cognition and tool use - The blacksmith at work. Cambridge University Press.
Kelly, R. L. (1988). The three sides of a biface. American Antiquity, 53(4), 717–734.
Kilby, J. D. (2008). An investigation of Clovis Caches: content, function, and technological organization. PhD, University of New Mexico.
Kolhatkar, M. Les Carrières Préhistoriques Du Québec. MCCCF (Montréal: 2006).
Kolhatkar, M. (2016). À La Martre, Le Temps S’écoule Le Long De Lignes De Pierre. In P. Dejarnac & C. Sondages (Eds.), Du Silex Au Gobelet En Plastique (pp. 111–134). Fedora.
Kolhatkar, M. (2020). Comment Raconter Des Histoires Lithiques Dans Les Labours De La Martre (Québec, Canada). Thèse de doctorat, Université de Montréal, Département d’Anthropologie.
Kolhatkar, M. (2022a). Skill in Stone Knapping: An Ecological Approach. Journal of Archaeological Method and Theory, 29, 251–304.
Kolhatkar, M. "Analyzing skill in lithic dispersions using width, thickness and width-by-thickness ratio." Lithic Technology (2022b).
Kuhn, S. L. (2007). Cores, tools, and the priorities of lithic analysis. In S. P. McPherron (Ed.), Tools Versus Cores - Alternative Approaches to Stone Tool Analysis (pp. 267–276). Cambridge Scholars Publishing.
Lake, M. "Archaeological inference and the explanation of hominid evolution." In The Archaeology of Human Ancestry: Power, Sex, and Tradition, edited by Stephen Shennan and D.G. Steele. London: Routledge, 1996.
Latash, M. L., Turvey, M. T., & Bernstein, N. A. (1996). Dexterity and its development. Resources for Ecological Psychology. L. Erlbaum Associates.
Lave, J. Cognition in practice: mind, mathematics and culture in everyday life. Cambridge; New York: Cambridge University Press, 1988.
Lave, J. (1993). Understanding Practice: Perspectives on Activity and Context. Learning in Doing. Cambridge University Press.
Lave, J., & Wenger, E. (1991). Situated learning: legitimate peripheral participation. Learning in Doing. Cambridge University Press.
Lemonnier, P. (1993). Technological Choices: Transformation in Material Cultures since the Neolithic. Routledge.
Lemonnier, P. (2010). L’Étude des Systèmes Techniques, une Urgence en Technologie Culturelle. Techniques et Culture, (54-55), 46–67.
Leroi-Gourhan, A. (1964). Le Geste Et La Parole (Vol. 2). Albin Michel.
Leroi-Gourhan, A. (1973). Milieu Et Techniques. Albin Michel.
Lothrop, J., Lowery, D., Spiess, A., & Ellis, C. (2016). Early human settlement of Northeastern North America. PaleoAmerica, 2(3), 192–251.
Lucas, G. (2005). The Archaeology of Time. Routledge.
Meltzer, D. J. (1988). Late Pleistocene Human Adaptations in Eastern North America. Journal of World Prehistory, 2, 1–52.
Meltzer, D. J. (1989). Was stone exchanged among Eastern North American Paleoindians. In C. J. Ellis & J. C. Lothrop (Eds.), Eastern Paleoindian Lithic Resource Use (pp. 11–39). Westview Press.
Milne, S. B. (2005). Palaeo-Eskimo Novice Flintknapping in the Eastern Canadian Arctic. Journal of Field Archaeology, 30(3), 329–345.
Milne, S. B. "Landscape Learning and Lithic Technology. Seasonal Mobility, Enculturation and Tool Apprenticeship among the Early Palaeo-Eskimos." In Structured Worlds: The Archaeology of Hunter-Gatherer Thought and Action, edited by A. Cannon, 95-115. London; New York: Routledge, 2011.
Milne, S. B. (2013). Lithic raw material availability and Palaeo-Eskimo novice flintknapping. In W. Wendrich (Ed.), Archaeology and Apprenticeship: Body Knowledge, Identity, and Communities of Practice. University of Arizona Press.
Murray, T. (1997). Dynamic Modelling and New Social Theory of the Mid - to Long Term. In S. E. Van der Leeuw & J. McGlade (Eds.), Time, Process and Structured Transformation in Archaeology (pp. 449–463). Routledge.
Murray, T. (1999). A Return to the "Pompeii Premise". In T. Murray (Ed.), Time and Archaeology (pp. 8–27). Routledge.
Murray, T. "Evaluating Evolutionary Archaeology." World Archaeology 34, Archaeology and evolutionary ecology, no. 1 (2002): 47-59.
Murray, T. "Why the history of archaeology is essential to theoretical archaeology." Complutum 24, no. 2 (2013): 21-31.
Newby, P., Bradley, J., Spiess, A., Shuman, B., & Leduc, P. (2005). A Paleoindian response to Younger Dryas climate change. Quaternary Science Reviews, 24, 141–154.
Odell, G. H., & Cowan, F. (1987). Estimating tillage effects on artifact distributions. American Antiquity, 52(3), 456–484.
Pelegrin, J. "Réflexions sur le Comportement Technique." In La Signification Culturelle des Industries Lithiques, edited by M. Otte. Studia Praehistorica Belgica 4, 72-91. Oxford: BAR, 1985.
Pelegrin, J. (1991). Les Savoir-Faire: Une Très Longue Histoire. Terrains, 16.
Pelegrin, J. Technologie Lithique : Le Châtelperronien de Roc-De-Combe (Lot) et de la Côte (Dordogne). Cahiers Du Quaternaire. Vol. 20, Paris: Éditions du CNRS, 1995.
Pelegrin, J. "Les Pointes de Volgu : jusqu'aux Limites de la Taille Bifaciale." Chap. 4 In Les Silex Solutréens de Volgu (Rigny-Sur-Arroux, Saône-Et-Loire, France) : Un Sommet dans l'Art de la "Pierre Taillée", edited by J.-P. Thévenot, 75-118. Dijon: Revue Archéologique de l'Est, 2019.
Pelletier, B. G., & Robinson, B. S. (2005). Tundra, ice and a pleistocene cape on the Gulf of Maine: a case of Paleoindian Transhumance. Archaeology of Eastern North America, 33, 163–176.
Perlès, C. "Économie des Matières Premières et Économie du Débitage: Deux Conceptions Opposées?". In 25 Ans D’études Technologiques En Préhistoire: Bilan Et Perspectives, edited by Collectif, 35-46. AP-DCA, Antibes, 1991.
Perreault, C. (2019). The quality of the archaeological record. The University of Chicago Press.
Petersen, J. B. "Foreword: West Athens Hill, the Paleoindian Period, and Robert E. Funk in Northeastern Perspective." In An Ice Age Quarry-Workshop, the West Athens hill site revisited, edited by R. Funk, XI-XLIX. New York: New York State Museum Bulletin, 2004.
Petersen, J. B., Bartone, R. N., & Cox, B. J. (2000). The Varney Farm Site and the Late Paleoindian Period in Northeastern North America. Archaeology of Eastern North America, 28, 113–40.
Piaget, J. (1977). La Construction du Réel chez l'Enfant. Actualités Pédagogiques et Psychologiques. Delachaux & Niestlé.
Pigeot, N. (1990). Technical and Social Actors: Flintknapping Specialists and Apprentices at Magdalenian Etiolles. Archaeological Review from Cambridge, 9(1), 126–141.
Pigeot, N. "Chaînes Opératoires : Contexte Théorique Et Potentiel Cognitif." In L’archéologie Cognitive, edited by R. Treuil, 149-171. : Éditions de la maison des sciences de l'homme, 2011.
Pintal, J.-Y. (2006). Le Site de Price et les Modes d’Établissement du Paléoindien Récent dans la Région de la Rivière Mitis. Archéologiques, (19).
Ploux, S., Karlin, C., & Bodu, P. (1992). D'une Chaîne à l'Autre: Normes et Variations dans le Débitage Laminaire Magdalénien. Techniques et Culture, (17-18), 81–114.
Rezek, Z., Holdaway, S. J., Olszewski, D. I., Lin, S. C., Douglass, M. J., McPherron, S. P., Iovita, R., Braun, D. R., & Sandgathe, D. (2020). Aggregates, formational emergence, and the focus on practice in stone artifact archaeology. Journal of Archaeological Method and Theory, 27, 887–928.
Richard, P. J. H. (2007). Le Paysage Tardiglaciaire du "Grand Méganticois": État des Connaissances. In C. Chapdelaine (Ed.), Entre Lacs et Montagnes au Méganticois, 12 000 Ans d'Histoire Amérindienne (pp. 23–45). Recherches Amérindiennes au Québec.
Richard, P. J. H., Veillette, J. J., Larouche, A. C., Hétu, B., Gray, J. T., & Gangloff, P. (1997). Chronologie de la Déglaciation en Gaspésie: Nouvelles Données et Implications. Géographie Physique et Quaternaire, 51(2), 163–184.
Robb, J. E., and T. R. Pauketat. "From moments to millennia." In Big Histories, Human Lives: Tackling Problems of Scale in Archaeology, edited by J. E. Robb and T. R. Pauketat. School for Advanced Research Advanced Seminar Series, 3-33. Santa Fe, New Mexico: School for Advanced Research Press, 2013.
Roche, H., and P.-J. Texier. "La Notion De Complexité Dans Un Ensemble Lithique. Application Aux Séries Acheuléennes D'isenya (Kenya)." In 25 Ans D'études Technologiques En Préhistoire, edited by Collectif, 99-108. Juan-les-Pins: Éditions APDCA, 1991.
Rogoff, B. (2003). The cultural nature of human development. Oxford University Press.
Rogoff, B., & Lave, J. (2000). Everyday cognition: its development in social context. Replica Books.
Roux, V. "Chaîne Opératoire, Technological networks and sociological interpretations." Cuadernos de Prehistoria y Arqueología de la Universidad de Granada 30, special issue "Social ceramics: Anthropological approaches to technology in Archeology" (2020): 15-34.
Schortman, E. M. (1989). Interregional interaction in prehistory: The need for a new perspective. American Antiquity, 54(1), 52–65.
Shanks, M., & Tilley, C. (1987). Re-Constructing Archaeology: Theory and Practice. Cambridge University Press.
Shott, M. J. (1998). Status and Role of Formation Theory in Contemporary Archaeological Practice. Journal of Archaeological Research, 6(4), 299–329.
Shott, M. J. (2010). Size dependence in assemblage measures: Essentialism, materialism, and "She" analysis in archaeology. American Antiquity, 75(4), 886–906.
Shott, M. J. (2017). Stage and Continuum Approaches in Prehistoric Biface Production: A North American Perspective. PLoS ONE, 12(3).
Smallwood, A. M. (2010). Clovis biface technology at the Topper Site, South Carolina: evidence for variation and technological flexibility. Journal of Archaeological Science, 37, 2413–2425.
Smitsman, A. W., Cox, R. F. A., & Bongers, R. M. (2005). Action dynamics in tool use. In V. Roux, B. Bril, & Mcdonald Institute Monographs (Eds.), Stone Knapping: The Necessary Conditions for a Uniquely Hominin Behaviour (pp. 129–146). McDonald Institute for Archaeological Research.
Soressi, M., and J.-M. Geneste. "The history and efficacy of the Chaîne Opératoire approach to lithic analysis: studying techniques to reveal past societies in an evolutionary perspective." PaleoAnthropology, no. Special Issue: Reduction Sequence, Chaîne Opératoire, and Other Methods: The Epistemologies of Different Approaches to Lithic Analysis (2011): 334-50.
Spiess, A., Wilson, D., & Bradley, J. W. (1998). Paleoindian occupation in the New England maritimes region: beyond cultural ecology. Archaeology of Eastern North America, 26, 201–264.
Stahl, A. B. (1993). Concepts of time and approaches to analogical reasoning in historical perspective. American Antiquity, 58(2), 235–260.
Stern, N. (1994). The implication of time-averaging for reconstructing the land-use patterns of early tool-using hominids. Journal of Human Evolution, 27, 89–105.
Storck, P. L. (1997). The fisher site, archaeological, geological and paleobotanical studies at an Early Paleo-Indian Site in Southern Ontario, Canada. Memoir 30. Museum of Anthropology, University of Michigan.
Stout, D. (2002). Skill and cognition in stone tool production: An ethnographic case study from Irian Jaya. Current Anthropology, 43(5), 693–722.
Tixier, J. (1967). Procédés D’analyse Et Questions De Terminologie Dans L’étude Des Ensembles Industriels Du Paléolithique Récent Et De L’épipaléolithique En Afrique Du Nord-Ouest. In W. W. Bischop & J. D. Clark (Eds.), Background to Evolution in Africa (pp. 771–820). The University of Chicago Press.
Vaquero, M., Gema Chacón, M., Dolores García-Antón, M., Gómez de Soler, B., Martínez, K., & Cuartero, F. (2012). Time and space in the formation of lithic assemblages: The example of Abric Romaní Level J. Quaternary International, 247, 162–181.
Vygotsky, L. S. (1930). Mind and society. Harvard University Press.
Waldorf, D. C. (1993). The art of flint knapping. Mound Builder Books.
Wendrich, W. (Ed.). (2013). Archaeology and apprenticeship: body, knowledge, identity, and communities of practice. University of Arizona Press.
Wenger, E. (1998). Communities of practice: Learning, meaning, and identity. Learning in Doing. Cambridge University Press.
Wenger, E. (2000). Communities of Practice and Social Learning Systems. Organization, 7(2), 225–246.
Whittaker, J. C. (1994). Flintknapping: Making and understanding stone tools. University of Texas Press.
Wiessner, P. (1983). Style and Social Information in Kalahari San Projectile Points. American Antiquity, 48, 253–276.
Wiessner, P. (1984). Reconsidering the behavioural basis for style: A case study among the Kalahari San. Journal of Anthropological Archaeology, 3, 190–234.
Acknowledgements
I would like to thank Adrian L. Burke, under whose supervision the doctoral thesis that served as the basis for this manuscript was conducted. I would also like to thank three anonymous reviewers and the editor for their thoughtful comments on earlier versions of this manuscript.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Competing interests
This author declares no competing interests.
Additional information
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Supplementary information
ESM 1
Supplementary Materials 1: Qualitative and metric variables used to define stages, ordered by metric sector and level of technical difficulty (16-west). (PDF 78 kb)
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Kolhatkar, M. The Social Life of Palimpsests: Skill, Bifacial Stone Knapping, and Differentiation in the Plowed Fields of La Martre. J Archaeol Method Theory 31, 946–1005 (2024). https://doi.org/10.1007/s10816-023-09629-2
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10816-023-09629-2