In the last chapter I outlined various approaches that have been taken in analysing quartz assemblages, discussing how a variety of approaches to quartz have been instigated. This chapter outlines the approach I will take in assessing quartz lithic technology in Ireland. The main approach of this thesis is experimental knapping; a series of experimental knapping events have been devised in order to create an experimental assemblage with which to analyse quartz artefacts from two excavated assemblages – Belderrig, Co. Mayo, and Thornhill, Co. Londonderry – which serve as case studies. These case study assemblages, and their excavations, will be properly introduced in their respective chapters. In Section 5.2 I will turn again to the issue of the sociality of technology introduced in the previous chapter, outlining my understanding of this and the implications it has for my methodology. Section 5.3 introduces the general methodology, the chaîne opératoire, used in this thesis and how it relates to a social archaeology perspective. Section 5.4 returns again to the subject of typological and terminological considerations discussed in the previous chapter, while Section 5.5 introduces the petrographic analysis undertaken. Section 5.6 introduces the knapping experiments and the analysis of the case studies; Section 5.7 introduces the quartz recognition experiments undertaken on participants at the Belderrig excavations and the WAC conference in Dublin; Section 5.8 introduces the experiment with burnt quartz; and Section 5.9 introduces the case studies.
"It is time we took more seriously the likelihood that subtle, non-discursive, tacit, and unintended sorts of political interests were practiced by even the most 'primitive' technicians in the course of pursuing their most mundane and simplest (adaptive) technologies...[E]ven in the most cohesive of communities, where the homogeneity of habitus is said to be most pronounced (such as egalitarian hunter-gatherers), everyone will not necessarily and always agree on what must be done nor how to proceed...Life is a terrain full of ambiguity, ambivalence, and contestation; technology is both an arena where they come to the fore and an intersubjective site where they can be negotiated and worked through" (Dobres 2000, 118, 140-1 emphasis in original).
This thesis concerns our understandings of quartz use by early prehistoric communities as part of their lithic technology. As discussed in the previous chapter, my understanding of technology is that it is social in its formulation, application, and implication. Consequently, while placing emphasis on the physical attributes of the quartz – through the experimental knapping, which will provide data on the material characteristics and constraints of quartz – the concurrent consideration is the sociality of technology. As mentioned previously, I do not view the social and technical as two separate entities, nor are they simply two sides of the same coin. Rather, through the enaction of technology people situate themselves in the world in relation to themselves and to others. The implication of this stance is that any analysis of technical acts or technological systems must recognise that material considerations, and material constraints, cannot be analysed first with a dressing of social considerations put on after.
This naturally complicates matters: it would easier to move from the seemingly concrete – the hard facts of the material we have evidence for in the shape of the surviving lithics – to the seemingly abstract – the elusive facts we want evidence for in the shape of what the surviving lithics meant to the prehistoric communities. However, such a train of thought derives from the “seductiveness of the tangible” (Dobres 2000, 66); from the idea that the physical remains are easier to assess than the social systems that created them – Dobres (2000, 106, emphasis in original) put the question thus,
"how does a researcher know where, when, and how to draw that imaginary line between the material and non-material components of technology (such as hunting), if in practice they are inseparably linked through the thoughts and value systems of technicians? For example...the technology of faunal acquisition and processing...often focuses narrowly on the material and hardware ‘side’ of hunting and dismembering activities. In doing so, western sensibilities about what a technology of ‘the hunt’ involves are reaffirmed, but this may bear little resemblance to the (emic) ethnographic or ancient situation".
Along similar lines, Lemonnier (1986) and Petrequin (1993) have discussed the parameters of the material and the social, the former from an anthropological perspective and the latter from an archaeological perspective. Lemonnier (1986, 164), from a study of the Anga in New Guinea, examined the distribution of various technical traits, such as traps, houses, arrows, and the chaîne opératoire of gardening amongst different groups and discussed how the technical choices of a given group can occur due to “purely social dimension[s]”. Addressing an archaeological community wishing to understand the social parameters of technical choices, and subtitling his article “toward an anthropology of technical systems”, Lemonnier (1986, 176, 179) suggested that one of the aims of his paper was to “stress the complexity of social content in material culture”; “[w]e shall conclude...that the logic of technical choices made by the Anga still escapes us and that the classifications resulting in the similar distribution of functionally independent traits have mutual relations which are not all direct...[i]t is rather the relation among terms, indeed relations among relations, which are reflected in the variety of choices made by societies.”
Petrequin (1993, 52, passim) highlighted that considerations of raw material physical characteristics alone will not dictate how it might have been worked, but that technological choices are dependent on the cultural understandings of the “right” way to work material. His research on the production of axes by Neolithic communities in the Jura Mountains suggested that their stone craft was influenced by the traditional, culturally defined, techniques used – when a particular type of hard rock that had been used was no longer as readily available through exchange, the communities began to extract a local, softer rock. Even though this softer rock required a change in technique to form the axe, they continued for some generations to use their culturally defined technique that was suitable for the imported hard rock. Eventually, their technique became modified to suit the new material, and this new technique continued even after the raw material reverted back to the original hard rock. Therefore, while material constraints are important in understanding prehistoric technology, there was more at play in their technological decisions than materiality.
Another level of complexity in the sociality of technology derives from contextualising notions of individuality and personhood. Taylor (1992, 112) comments that the modern idea that a self is something that comes from within a single individual is a peculiar, historically contingent understanding, and a difficult one to see beyond: “who among us can understand our thought being anywhere else but inside, ‘in the mind’? Something in the nature of our experience of ourselves seems to make the current localization almost irresistible, beyond challenge”. Fowler (2004) has outlined various differing anthropological studies of individuality and dividuality, such as the concept of partible people and multiple authorship, and also of permeable people. He elaborates how these differing ways that people conceive of themselves in the world do not stand in clear distinction from the current modern western idea of the individual, but rather
"each person negotiates a tension between dividual and individual characteristics, and, in all societies, personhood emerges from the constant reconciling of one with the other. In some contexts, like modern Europe, individual features are accentuated, while in others, like contemporary Melanesia, dividual features are accentuated – but these are dominant features, not factors which completely repress or override the other" (Fowler 2004, 34 emphasis in original).
Along these lines, Finlay (2003b) has approached the study of Scottish Mesolithic lithic technology from the perspective of the partibility of people and things. Using the concept of multiple authorship, she highlights how the composite nature of making and using microliths epitomises this idea, and that a shift away from a focus on the fixed and bounded nature of people and things allows for the contemplation of the fluidity and ambiguity of social relations.
This thesis, therefore, takes the sociality of technology as axiomatic. What is not axiomatic, however, is the minutiae of how quartz technology was involved in the social reproduction of prehistoric communities. If we accept that social relations were key in the enacting of technology, and that social beings were not fixed or bounded, we must accept that the use and perceptions of quartz technology were also fluid and ambiguous. Consequently, this calls for a careful consideration of the material remains that provide us with our evidence of prehistoric technologies, and a careful consideration of how archaeological practice forms our perceptions. In terms of the ramifications for the methodology of this thesis, therefore, while placing emphasis on the material characteristics of quartz which will be defined through the experimental knapping and geological studies, the framework will maintain a firm social focus, and be concerned with not splitting the material from the social; this entails that a separate chapter cannot be devoted to social implications once the material concerns are outlined. On the contrary, they are implicated in the analysis from the start, and acknowledged as so; the analysis of the sociality of technology cannot be contrived as a check list of attributes to be sought, but rather as a fundamental premise from which one begins. A useful methodology for beginning with such a premise is the concept of the chaîne opératoire.Top of Page
Leroi-Gourhan developed the concept of the chaîne opératoire as a methodology for comprehending and analysing the technical actions undertaken in the manufacture and use of material culture, especially from a social, cognitive, and evolutionary perspective (Leroi-Gourhan 1993; Bleed 2001; Audouze 2002). While Leroi-Gourhan coined the term, he “had crystallized an idea or a set of ideas that were emerging at the time”, such as those of Mauss and Maget (Audouze 2002, 286-8). As well as the French concept of the chaîne opératoire – operational sequences – American and Japanese researchers have analysed technology from the sequence model perspective (Bleed 2001). Bleed (2001) suggests that two broad approaches can be discerned in the sequence models research – the teleological and the evolutionary. The teleological approach treats the technical acts as a product of a predetermined pattern, or mental template, with emphasis on the results; the evolutionary approach treats the technical acts as reactions to situations, with emphasis on the actions and the possible diversity in the sequence (Bleed 2001, 120-1). Bleed (2001, 121) suggests, however, that the distinction between these two approaches “is not a simple one”, and that “sometimes both perspectives are evidenced in the same analysis.” This highlights that the concept of the chaîne opératoire is not necessarily a programmatic device, but open to differing conceptual perspectives.
Dobres (2000, 154-5) comments that while the initial concept of the chaîne opératoire held an explicitly social focus, this has been lost as analyses have “developed a more rigorous scientific basis and analytic program.” Dobres argues that an engendering of the chaîne opératoire is needed, which can allow a move away from normative templates as the focus of attention, to a focus on variation, which brings to the fore the social dynamics of the communities. This is achieved by using a practice framework, which “means, in essence, working from the premise that the social context of site-specific production and use activities played a major role in structuring the choice of productive strategies actually practiced” (Dobres 2000, 158).
Conneller (2000) discussed her use of the chaîne opératoire in investigating the Mesolithic in England. Conneller (2000, 140, passim) stated that her “analysis has focused on the reduction sequences represented in the lithic assemblages recovered from these sites [in the Vale of Pickering]” and suggested that using the chaîne opératoire “engendered certain problems when conceptualising human action in the landscape”; she was critical of the linearity that can be imposed on the interpretation of the chaîne opératoire, and argued that its use can segregate human action from the landscape. Conneller translated the ‘chaîne opératoire’ as the “chains of action” (Conneller 2000, 143) or as a “chain” (Conneller 2000, 148); this may partially explain her uneasiness with the concept – in describing the chaîne opératoire as a ‘chain’, one can get the impression of a solid, linear connection, a continuous chain, such as that of a metal chain. However, a correct (and more common) translation would be to call it an operational sequence – this immediately removes the necessity of contemplating the sequence in such a rigid manner, and allows for how Conneller (2000, 148) considers it to be, “non-linear, numerous, and frequently of multiple authorship”.
The chaîne opératoire is used in this thesis as a methodology with which to assess the technical practices at play in the selected case studies. At base, the chaîne opératoire examines all the actions – or enactions – at play from an artefact’s birth to death; this can also include an artefact’s rebirth as a reused artefact or its rebirth as an archaeological entity (see Woodman et al. 2006), and is not limited to its use to analyse reduction sequences (contra Conneller 2000). Figure 5-1 shows at a basic level the possible chaîne opératoires seen in the archaeological record for lithics, highlighting the myriad of possibilities even at this level of representation; each of the six main groups listed concurrently contain their own multiple chaîne opératoires. Odell (2001, 81) has commented that many adherents to the chaîne opératoire methodology do not use the full battery of methods available, and the results are “warmed-over lithic reduction sequences, with occasional lip service to procurement”. Figure 5-2 highlights the limitations of this thesis: as no use-wear analysis will be conducted, the possible chaîne opératoires available for study are consequently lessened. As stated in the previous chapter, it is beyond the scope of this present project to include a use-wear component, which would in effect be a full PhD project its own right, but it is hoped that this research can provide a coherent classificatory framework for the understanding of the fracture mechanics of the quartz assemblages, which can be added to by future use-wear analysis.
Figure 5-1 Possible chaîne opératoires
Figure 5-2 Chaîne opératoires without use-wear analysisTop of Page
"To dictate definition is to wield cultural power" (Livingstone 1992, 304).
This thesis will attempt, through experimental work, to define a quartz core and debitage product typology in order to devise a framework with which to understand quartz technology. At base, lithic typologies are devised to organise artefacts into mutually exclusive types, and to communicate this organisation to others, often in order to delimit cultures, chronologies, and technologies and are generally understood as heuristic devices (see Klejn 1982; Adams and Adams 1991; Wylie 1992; Lucas 2001; Woodman et al. 2006). As discussed in the previous chapter, there is a general division between typological studies and technological studies, with the latter using ‘types’ as well, and this has been described as debitage typological analysis (e.g. Andrefsky 2001), or typotechnological analysis (e.g. Cornelissen 2003; Ballin 2008).
As a prime function of typologies is to communicate, questions of terminology are important. Terminology in lithic analysis is characterised by the use of various terms that can appear to have ambiguous or multiple meanings, and it is often unclear which meaning is implied in a specific instance, unless the authors append a glossary to their work – in the last chapter I noted the example of categories of ‘formal’ and ‘expedient’ tools. Another prime example is the terminology of ‘debitage’ and ‘debris’; Table 5-1 lists the descriptions given for ‘debitage’ and ‘debris’, sensu Inizan et al. (1999), from six publications (one French, four British, and one Irish) that have all explicitly outlined the terminology of their lithic analysis. Inizan et al. are used as the comparison for three reasons –
1. As this was the first of these published;
2. Woodman et al. (2006) and Ballin (2008) cite this book and were therefore aware of Inizan et al.’s terminology;
3. Wickham-Jones (1990, 57) noted that the terminology of the original French edition of Inizan et al. was used for her publication, and subsequently Finlayson (2001, 61) based their terminology on an adaption of Wickham-Jones (1990), and therefore ultimately on Tixier et al. (1980).
|Term||Inizan et al. (1999, 138)||Wickham-Jones (1990, 58, 73)||Finlayson et al. (2001, 62, 64, 547)||Woodman et al. (2006, 86)||Andrefsky (1998, xxii, 81-2)||Ballin (2000, 10-11)|
|Debitage||T]he intentional knapping of blocks of raw material, in order to obtain products that will either be subsequently shaped or retouched, or directly used without further modification [and] all removals resulting from the knapping of a core||[D]ebris that was not suitable for any further purpose, material discarded immediately upon the end of the knapping exercise. It includes much very small material||[U]sed in a conventional manner and does not imply that such artefacts are indeed waste||[S]mall pieces – flakes, chunks, shatter and chips||Detached pieces that are discarded during the reduction process||[U]nmodified flaked products (various blanks and waste material including chips, flakes, indeterminate fragments and chunks) from the reduction of a core|
|Microdebitage [and] Small fraction debitage||[Artefacts that are] < 10 mm maximum dimension|
|Debitage Products||[A]ll removals resulting from the knapping of a core, i.e. to all flakes in the broader sense of the term: those resulting from preparation, potential tool blanks, and all waste products|
|Debris||[S]hapeless fragments whose mode of fracture cannot be identified, and which cannot be assigned to any category of objects||[A] by-product of knapping: that material which inevitably results from the knapping process but which was not the goal of that process. Some debris may be suitable for use with or without modification||Detached pieces that are discarded during the reduction process|
|Shatter||[Flake shatter] All flake debitage with no recognisable striking platform [Angular shatter] Lack of a recognisable single dorsal or ventral feature|
|‘Chip’||Blades below 5mm in width||[introduced but not defined]||All flakes and indeterminate fragments, the greatest dimension (GD) of which ≤ 10 mm|
|Chunk||[R]emovals with neither platform nor ventral surface. They are generally the unintentional by-products from knapping. They may be large or small||Pieces that have neither platforms nor ventral surfaces. These are generally an accidental knapping product, or the result of pieces shattering because of heat||A] piece with neither platform nor ventral surfaces. Generally an accidental knapping product or product of heat shatter; common in chert and quartz||[L]arger indeterminate pieces...chunks being a sub-group...covering the larger, ‘chunky’ pieces so characteristic of...quartz assemblages|
|Indeterminate Pieces||GD > 10 mm, which cannot be unequivocally identified as either debitage...or cores|
Table 5-1 Variable meanings of 'debitage' and 'debris'
Inizan et al. suggest that ‘debris’ is a piece that cannot be assigned to any other category. Andrefsky maintains that ‘debris’ is the same as ‘debitage’; instead, he uses ‘shatter’, with a distinction between ‘flake shatter’ and ‘angular shatter’ – with ‘flake shatter’ being mesial and distal flake fragments. Ballin has two usages of ‘debris’ – in his outline of terminology cited above, he excludes the term and instead uses the two terms ‘indeterminate pieces’ and ‘indeterminate fragments’, depending on size; but in other articles he uses ‘debris’ but does not define its meaning (e.g. Ballin 2005, 2008). Woodman et al. and Finlayson et al. do not use the term but use ‘chunk’ instead. While Wickham-Jones also calls ‘debris’ ‘chunks’, she describes ‘debris’ as the by-product of knapping which may have been suitable for use, and ‘debitage’ as ‘debris’ that was not suitable for use. Ballin (2000, 10) rightly argues that this use of ‘chunk’ is often inappropriate, as the chunks may not actually be ‘chunky’, but may be thin pieces, and he suggests that ‘chunk’ should be kept for artefacts that are actually sensu stricto ‘chunky’, especially in relation to quartz pieces, which , according to him, often fracture in such a way.
From this we can see that a relatively straightforward concept – debitage – can have substantially different meanings. What is interesting is that Ballin’s (2000) publication was a call for a more accurate, standardised, classification and description of lithics. Yet he nevertheless eschewed a published, workable terminology, i.e. Inizan et al.’s, for his own. Whereas Inizan et al.’s terminology leaves out size, with which the terms could then be sub-divided into, others constrain various terms to size, using the >, <, and ≤ 10 mm as arbitrary dividing points. As noted in Chapter 3, Flenniken’s work showed that unretouched pieces under 10 mm were actually used ‘tools’, and under Ballin’s classification these would be chips – which he describes as “primary refuse” (Ballin 2000, 10). Furthermore, there is an inconsistency in the use of the 10mm division: Finlayson et al. use “< 10 mm” while Ballin uses “≤ 10 mm”.
Even when lithic analysts note in their reports that they are using a chosen terminology such as Inizan et al. or Finlayson et al., they use terms that are undefined, or in a different manner. For example, while stating that they are using the Finlayson et al. (2001) publication’s terminology, Warren (Forthcoming) and Dolan and Warren (2006) use ‘chip’ which is undefined by Finlayson et al., which consequently means that it is not possible to understand what they mean by that category; while stating she is using the Inizan et al. publication, Sternke (n.d.-c) uses the term ‘debitage’ in a manner that is inconsistent with Inizan et al.
The Inizan et al. terminology is the more useful, and straightforward, in its usage, and this thesis will use its terminology. However, the Inizan et al. publication is not without its own peculiarities – for instance, while devoted to the techniques and methods of knapped stone, it makes no mention of the bipolar technique of knapping. The development of a debitage typology for the quartz material will necessitate terminology not included in the Inizan et al. publication, and therefore all terms used will be defined in Appendix A-1.
“Typology generalises...[it] draws its support from what is repetitive and stable and steers clear of what is individual and fugitive” (Klejn 1982, 78-9). In sorting artefacts into types, therefore, using a typological framework rests uneasily with understanding a social technology, which acknowledges fluidity and ambiguity at the core of social, and hence technical, relations. The construction of a quartz debitage typology, however, which will generalise an assemblage into fracture types, is a heuristic device to gain a better understanding of quartz assemblages. From this starting point, one can assess the chaîne opératoires in action in any assemblage, allowing for a detailed picture to be drawn of how the communities of practice (sensu Dobres 2000) varied in their lithic technology.Top of Page
The initial stage in the experimental knapping is the assessment of the quartz from the main case study area, Belderrig, Co. Mayo. The excavated material appeared to be of differing quality, and from a variety of sources such as psammite and metadolerite outcrops, and beach cobbles. A selection of quartz from different sources in the locality of the site was analysed macroscopically and through thin section by Julian Menuge, School of Geology, UCD. This was conducted in order to determine the range of quartz under consideration, to ascertain its structural properties, and possible effects on fracturing.
The quartz collected for the knapping experiments was collected from around the vicinity of Belderrig Harbour. Four locations were chosen for collecting material for the knapping – from metadolerite outcrops, psammite outcrops, beach cobbles, and a vein associated with metadolerite from about 1 km west of the harbour, which has been called Rose Cottage quartz (because of a so-named nearby house). The psammite hammerstones used were collected from the Belderrig shore. Festooned chert collected from an outcrop overlooking Lough Derravaragh, Co. Westmeath was knapped as a comparison to the fracture patterns of the quartz. This is the site of the possible Mesolithic chert quarry investigated by O’Sullivan et al.(2007). The chert was extracted away from the face containing the possible prehistoric extraction, in an area where weathering had loosened the chert into small, workable slabs.Top of Page
Understanding the fracture mechanics of raw materials is a key consideration in analysing lithic technology (Cotterell and Kamminga 1987; Knutsson 1988a; Callahan et al. 1992; Whittaker 1994; Andrefsky 1998; Inizan et al. 1999; Odell 2003). By using glass as a proxy for homogenous, isotropic rock, numerous studies have investigated flake formation and variables contributing to debitage attributes (e.g. Whittaker 1994; Dibble and Pelcin 1995; Pelcin 1997a, 1997b), while other studies have used materials used by prehistoric knappers such as flint (e.g. Pelegrin 2006), obsidian (e.g. Davis and Shea 1998), rhyolite (e.g. Darmark 2006) and quartz (e.g. Flenniken 1981; Knutsson 1988a; Callahan et al. 1992).
Cotterell and Kamminga (1987) define three major flake types based on the type of initiation – conchoidal, bending and wedging (Figure 5-3). The authors commented that an overemphasis on conchoidal flakes in lithic analysis is a legacy of the eolith debates – when first attempting to define humanly struck stone “[a]nything that did not have characteristic conchoidal features was relegated to the status of "chips, spalls or splinters" and effectively ignored in artifact [sic] descriptions and analyses” (Cotterell and Kamminga 1987, 681).
Figure 5-3 Flake initiation - adapted from Cotterell and Kamminga (1987)
For conchoidal flakes, the flake fracture initiation begins at the point of impact and the resultant flake will generally have a visible bulb and compression rings – two signatures used to identify a humanly struck flake. For bending flakes, the initiation begins away from the point of impact, and therefore does not create a bulb and has less apparent compression rings; Cotterell and Kamminga (1987, 690) note however, for bending flakes that “the flake surface created during the transition from initiation to propagation can look superficially like a diffuse bulb and has been mistaken as such by archaeologists”. Transverse breaks in flakes during manufacture, use and after discard are usually formed by bending fractures. The fracture initiation of bipolar percussion usually is formed by wedging, with the resultant flakes often being compression flakes. Cotterell and Kamminga (1987, 686) noted that hard hammer percussion generally formed conchoidal flakes while bending flakes were formed with softer impactors and during pressure flaking. Research by Pelcin (1997b) noted that for hard hammer percussion, an increase in platform width and a low exterior platform angle also creates bending flakes.
Redman (1998) examined the debitage attributes associated with hard and soft hammer percussion, based on the experimental knapping of bifaces. In discussing the attributes usually cited as signifying differences between impactor types, she lists the following: lipping, bulb, flake thickness, flake weight, flake length, crushed platform, platform width, flake curvature, platform angles, and manufacture stage. The bifacial reduction experiments were performed by three knappers using both impactor types to produce six biface blanks from Wyandotte chert. Redman (1987, 41) noted that due to the inherent difficulties in accurately measuring platform angle, this attribute was not recorded. The only flakes used for the analysis were complete flakes which accounted for about half of the assemblage. Redman argued that the results of the analysis suggested that categories of ‘hard hammer flake’ and ‘soft hammer flake’ “are, in a sense, meaningless” – greater variability was seen between the three different knappers rather than impactor. The only variables immune to idiosyncratic knapper difference were the bulb thickness, max thickness and midpoint thickness; while immune to knapper difference they were nevertheless only weak at distinguishing between impactor type (Redman 1998, 90-1).
Redman’s research has highlighted that knapping experiments can benefit from the use of multiple knappers. This, however, was not practical for the present project which is based on a single knapper; future research on Irish quartz technology could usefully follow from this project by examining the variability between knappers. This thesis broadly follows the framework used by Knutsson (1988a), whereby quartz was knapped, by a single knapper, using various techniques and methods, resulting in an experimental assemblage of cores and debitage products with which to analyse the case studies assemblages. Inizan et al. (1999, 30) have defined method, technique, and technical procedures as referring to
"any carefully thought out sequence of interrelated actions, each of which is carried out according to one or more techniques...Physical actions – a deft flip of the hand, the use of a hard or soft hammer, the interposition of a punch – are all examples of techniques...Technical procedures are short systematic sequences of actions involved in any kind of preparation, such as: the abrasion of an overhang [etc.]".
Before the experimental phase, a preliminary examination of the excavated assemblages from the case studies was conducted, providing a general overview of the material. The experimental knapping used three different techniques (hard and soft hammer direct percussion and bipolar percussion) and two supports (elastic and inelastic) giving five different technique/support combinations (Appendix A- 4); the elastic support means freehand, or against the leg as support, while inelastic means the use of an anvil, such as platform-on-anvil or bipolar-on-anvil (Knutsson 1988a). Eight cores were knapped with each technique, using four types of vein quartz and cobble quartz (the ‘cobble quartz’ is also vein quartz that has been transformed into cobbles). One chert block was knapped with hard hammer to provide a base comparison for the quartz. The cores were knapped until they either reached a small size, or until no further good flaking angles could be achieved. This knapping provided an experimental assemblage of 41 sets of cores and debitage products, comprised of the 10 cores for each quartz source type, plus one from the chert. The experimental assemblage was then analysed through a combination of debitage typological analysis and attribute analysis (cf. Andrefsky 2001). Debitage typological analysis concerns the structuring of the debitage into recurring types on an individual basis, such as recurring fracture types of flakes and flake terminations (cf. Knutsson 1988a), while attribute analysis concerns the assemblages as a whole by assessing attributes such platform attributes (platform width/thickness) (cf. Odell 1989). The details of the recording of the experimental knapping events are presented in the Chapter 6.
The variables that were recorded during the first phase of the experimentation are divided into input (Appendix A- 5) and output (Appendix A- 6) variables (see Ahler 1989). Appendix A- 7 and Appendix A- 8 outline the various attributes recorded for the cores and debitage products during the experiments. The core and debitage types were devised as a result of the experimentation. While Inizan et al. (1999) have suggested that for ‘debris’ no other term should be used, from the knapping experiments it was noted that the debris often formed into thicker and thinner pieces. Therefore, it was decided to subdivide the debris into pieces that were ≥3mm thick or <3mm thick, with the thinner pieces defined as slivers; the 3mm size was chosen after experimentation showed that this size appeared to create a suitable division. The subdivision of the artefacts into slivers and non-slivers could be done rapidly by setting a callipers to 3mm, with those that passed through the gap defined as slivers.
Measurements were taken with a 200mm Vernier callipers rounded to one decimal. Weights were recorded with two scales; a 3000g ±0.1g scale rounded to one decimal for the original block of quartz, and a 400g ±0.01g scale rounded to two decimals for the resultant cores and debitage. For the sieving of the bulk-collected smaller debitage into size grades, two calibrated sieves were used – a 5mm perforated metal plate sieve and a 1mm woven wire sieve. The same instruments were subsequently used for the archaeological material.
For the statistical analysis, a 95% confidence level was used throughout. The statistical analysis and random sampling was conducted with SPSS 15.0 (2006b). The principal statistics used for scale data were Analysis of Variance (ANOVA), Univariate General Linear Model (UNIANOVA) and Multivariate General Linear Model (GLM), which provide analysis of variance. The One-Way ANOVA procedure “produces a one-way analysis of variance for a quantitative dependent variable by a single factor (independent) variable. Analysis of variance is used to test the hypothesis that several means are equal”. The UNANOVA “procedure provides…analysis of variance for one dependent variable by one or more factors and/or variables. Using this General Linear Model procedure, you can test null hypotheses about the effects of other variables on the means of various groupings of a single dependent variable. You can investigate interactions between factors as well as the effects of individual factors” (2006b). For UNIANOVA, one category from each variable is used as the reference category for that variable with which to compare to the other categories.
In analysis of variance, post-hoc tests are used “to assess which group means differ from which others, after the overall F test has demonstrated at least one difference exists. If the F test establishes that there is an effect on the dependent variable, the researcher then proceeds to determine just which group means differ significantly from others” (Garson 2009b). Three post hoc tests are used in the following analysis – least significant difference (LSD), Tukey’s honestly significant difference (HSD), and Bonferroni. These three are seen on a scale of liberal (more likely to determine differences) to conservative (less likely to determine differences), from the liberal LSD to the conservative Bonferroni (Garson 2009b). Generally, the conservative Bonferroni results were used and quoted, except in certain instances where apparent anomalies in p values (significance) called for a reporting of the LSD test results. For scale data with non-normal distributions, the appropriate transformations were used to achieve normal distributions (see Shennan 1997).
For categorical data, Multinomial Logistic Regression and Generalised Linear Model (GZLM) were used for multi-categorical data and binary-categorical data respectively. Logistic Regression “can be used to predict a dependent variable on the basis of continuous and/or categorical independents and to determine the percent of variance in the dependent variable explained by the independents; to rank the relative importance of independents; to assess interaction effects; and to understand the impact of covariate control variables” (Garson 2009a). The GZLM is used for similar predictions, but in cases where the categories are binary responses, such as absence/presence of an attribute and so forth (Garson 2009a). While for analysis of variance with scale data, post hoc tests are used to identify which particular categories differ from others, for logistic regression the difference between particular categories are identified from the parameter estimates which provide the p value for each category compared to the reference category chosen.
The term ‘significant’ has a specific meaning in statistics, but is also a term used in non-statistical language. The term is used in both senses here, but when discussing statistical data, the term stands for statistically significant.Top of Page
In order to evaluate how people with different analytical skill levels recognise and categorise quartz artefacts Graeme Warren suggested that a quartz recognition experiment, similar to that undertaken by Lindgren (1998) investigating the success of participants at identifying retouch on quartz artefacts, could be designed and held during the 2008 excavations at Belderrig, with the staff and volunteers at the excavation participating. This experiment was subsequently run again during the WAC conference held in Dublin, in July 2008 with conference attendees invited to participate.
A selection of platform and bipolar cores and debitage were selected from the experimental assemblage and presented to the participants. The participants were not told that the artefacts were from the experimental assemblage, but instead were told that they were ‘Belderrig quartz’ – which the experimental artefacts were made from. The reason for avoiding a description of the artefacts as deriving from the experimental assemblage was in order to maintain the possibility that some pieces were not anthropogenic in origin – as one might anticipate in an archaeological assemblage but not in an experimental one. Furthermore, I was concerned that if the quiz had taken place with participants knowing that a definitive answer existed interpretations may have been affected by this knowledge. Both factors seemed very important in terms of testing an archaeological typology (where prior knowledge does not usually exist).
For the first experiment at Belderrig 30 artefacts were selected, and subsequently 20 of these were presented to the participants at the WAC quiz. The reason for excluding some artefacts was done after feedback from the first quiz suggested that during the conference a greater number of people would do the quiz if it was shorter; while this meant that less pieces were seen, this would be made up for by a greater range of people doing the quiz. The participants were asked to provide details concerning their archaeological experience and their experience with analysing lithics in general and quartz specifically. The participants were then asked to categorise the quartz artefacts, following a form provided, and also using the terminology in the Inizan et al. (1999) publication, which was provided for reference. This was requested in order to avoid problems of participants using different terminologies as discussed in Section 5.4. The participants’ responses were then analysed, with a focus on the effect of their self-assessed skill level on their ability to identify and classify the artefacts.Top of Page
Two strands of research have investigated thermally altered quartz – those investigating the heat treating of quartz to improve the knapping characteristics of quartz (e.g. Flenniken 1981; Leveillee and Souza 1981) and those investigating the recognition of quartz that has been burnt in hearths and so forth (e.g. Gonick 2003; Ballin 2008).
Leveillee and Souza’s (1981) experiments with heat treating quartz arose because numerous excavations from southern New England (Figure 4-2) had produced quartz artefacts which had been suggested as being heat treated. Their experiment involved slicing a large boulder of quartz and heating these slices in an oven, using variables of temperature and time; they then used a mechanical flaking apparatus to flake the heat treated quartz. Their results suggested that heat treatment positively affected the size of the flakes, and that this was more affected by temperature than time; they argued that this supported the hypothesis that heat treatment was used for lithics (Leveillee and Souza 1981).
Flenniken’s (1981) experiments with heat treating quartz were part of the larger project discussed in the previous chapter. He experimented with heat treatment as a possible ‘sub-system’; these experiments involved using x-ray diffraction analysis, which records “the loss of inter-granular molecular water between the crystals of the heat treated sample” – the results suggested that none of the material had been heat treated, probably because it was unnecessary: “heat treatment does not improve the flakeability of vein quartz...vein quartz is an example of a high utility (creates sharp edges when fractured), low energy (sharp edges are created with little energy out-lay) lithic material” (Flenniken 1981, 39, 42).
Gonick (2003) experimented with the effects of both burning and boiling on quartz and quartzite, looking at the evidence for fire cracked stones in the archaeological record, as well as investigating whether quartz or quartzite is more suitable for heating water – while the experiments showed that quartz was more suitable than quartzite (as it cracked less), the different fracture patterns between boiling and cracking were inconclusive, but discolouration showed that they were in fact heat fractured (Gonick 2003, 158).
Ballin (2008, 39) noted an under-acknowledgment of burnt quartz in Scottish lithic reports and undertook experiments to ascertain the character of it – the findings were reported briefly, but unfortunately, he has not published his quartz burning experiment results beyond the mention in the 2008 publication. Ballin’s experimentation with burning quartz suggested that burnt quartz is generally characterised by: “i) pitting and ‘peeled-off’ surfaces, ii) a dull and opaque appearance (where fresh quartz tends to be clear and vitreous), iii) various degrees of ‘granulation’ and disintegration, and iv) occasional areas with either a reddish or a pink hue”. Ballin suggested that burnt quartz can be identified on prehistoric artefacts, however not as easily as with flint.
During the literature review of Irish quartz finds, only one reference was made to burnt quartz (in Dolan and Warren 2006). It was decided that this project would focus on experimenting with burnt quartz, as experimentation with heat treatment as well would be beyond the time available. While technically quartz cannot be burnt (Menuge 2009a), this shorthand for fire altered or thermally altered is used here, as it has common currency in archaeological literature. The identification of burnt quartz can have implications for the identification of otherwise ‘invisible hearths’ (see Sergant et al. 2006), and other depositional traits. Sergant et al. (2006) devised an experiment to examine the identification of ‘invisible hearths’ on Mesolithic sites in Belgium. The experiments consisted of firstly depositing flint into a fire and then excavating the surrounding area to identify the extent to which artefacts were dispersed from the fire, as well as examining the fragmentation rate of the burnt artefacts. A second experiment left a trail of artefacts from the hearth, in order to examine at what point did signs of burning stop on artefacts; a third experiment placed smouldering charcoal over artefacts to see the effect of this type of lower heat burning. For the latter two experiments Sergant et al. (2006) commented that only at a high temperature (300 ºC) was heat damage on the flint noted, and that the smouldering charcoal had no effect as it was insufficiently hot. The temperature boundary for effective damage was “very abrupt: the critical point is probably in the order of less than 10 ºC because artefacts exposed to temperatures of 280-290 ºC do not show traces of heat damage” (Sergant et al. 2006, 1001).
Following from the results of Sergant et al., the present experiment on burnt quartz focused on placing artefacts directly into a fire and was not concerned with experimenting with lower temperatures. This involved burning with both green and dead wood; a series of quartz debitage and cores knapped from the four source materials from Belderrig mentioned previously was produced and artefacts chosen by seven size grades – ≥10<15mm, ≥15<20mm, ≥20<25mm, ≥25<30mm, ≥30<35mm, ≥35<40mm, and ≥40<50mm. The hearth was sieved using 5mm and 1mm sieves to retrieve the material. Due to time limitations, an excavation of the area around the hearth was impossible; instead a series of drop cloths covering 25m² were placed surrounding the hearth in order to determine the extent of artefact dispersal from the fire. In order to prevent fire damage to the cloths, a layer of sand was placed surrounding the hearth, laying over the edges of the drop cloths. Samples of the burnt and unburnt debitage were thin sectioned by Dr. Julian Menuge for analysis.Top of Page
The analysis of the case studies used the same methodology set out for the experimental assemblage above, using the experimental datasets for comparative analysis. One notable difference in the analysis concerns sequential breaks and fragments which were identified during the experimental knapping but subsequently excluded from analysis of the case studies; Chapter 6 outlines these terms and the reasons for their eventual exclusion. The respective case study chapters outline the sites’ geological and palaeoenvironmental context, along with a description of the excavations and analysis of the assemblages.
It should be noted at the outset that the two case study assemblages have substantially different contexts of artefact retrieval. The Belderrig component analysed was excavated from a small trench and consisted of a series of artefact scatters with few features noted, while the Thornhill assemblage came from a much greater excavated area, and consisted of numerous features such as pits, palisades, and buildings – while the Belderrig excavated are is interpreted as involving in situ knapping, no clear knapping floor was identified in the Thornhill assemblage.
Throughout the analysis of the case study assemblages I have used the term ‘community of practice’ sensu Dobres (2000), previously mentioned in Section 4.5 and Section 5.4. In lithic reports and writings based on them – and in almost all discourse concerning prehistoric technology (Dobres 2000, 40, 58, passim) – it is usual for the noun/pronoun to be omitted and for the passive tense to be used, such as in the following two examples: “chert and quartz working was being undertaken close to or at the Drummenny house, but flint knapping appeared to be carried out off-site” (Smyth 2006a, 158); “the pebbles were broken using direct percussion, whereby a pebble core was held in the hand and a series of flakes was struck from it using a hammerstone” (Milliken n.d.).
The term ‘community of practice’ is therefore used to avoid this omission of the noun/pronoun from the analysis, and to firmly embed the prehistoric people under consideration within the discourse. In using ‘community of practice’, the implications are that technology is a verb – technology is enacted and practiced rather than possessed. This practice is by communities, highlighting that even if individuals are alone at work they are always part of a social grouping. Therefore the term stands for both an individual within a group and the group itself. This does not imply that a group is homogenous but simply that, as social beings, people are always situated within a community within which “they are woven into webs of pre-existing social and material conditions, rules, values, hierarchies of knowledge, demands, and expectations that together weave overarching structures, or forms of involvement…within which their making and use activities unfold” (Dobres 2000, 129).Top of Page
This chapter began with a discussion of the embeddedness of the social sphere in the technical sphere, and I suggested that the position taken in this thesis is that any study of technology is a study of social beings and their relationships with other people and the world in which they dwell. The consequence of this perspective is that there is no social side to life to study, nor are they simply two sides of the same coin, but rather they dwell on the same plane – there is only a social reality, and, as Ingold (2000, 195) has put it, “one of the outstanding features of technical practices lies in the embeddedness in the current of sociality”. The consequence of this approach is that there is a need to frame the question of sociality differently, rather than seeing it as an attribute added on, a particular realm divisible from the whole, a part of life’s machine. This thesis maintains that an approach founded on the exploring of the chaîne opératoires of lithic traditions can place emphasis on past practices, and technical choices; the emphasis on action allows for the interpretation of the archaeological record as a result of dynamism. The main thrust of this thesis is the experimental work which will develop an assemblage of worked quartz to understand the fracture mechanics involved as well as the effects of burning, and to bring this understanding to interpret the archaeological assemblages. This series of experiments is highly artificial in that it is designed to tease out attributes that are apparent from differing knapping techniques; this will hopefully allow for an understanding of the resultant debitage in the archaeological record.
This series of experiments is not an attempt to recreate a prehistoric technology; my reasons for working this material are fundamentally different than that of the prehistoric communities, as are my goals; my social reality both constrains and enables my routine of experimentation, as it also did in the past. Instead, the aim is to bring to the archaeological material a suite of analytical techniques with which to draw out an interpretation of the evidence, while taking care to be cognisant of the complexity of the archaeological record, in regards to complexity of the lives that created it, and the complexity of the process that have thence transformed it. This complexity is added to especially in consideration of the overwhelming fact that the lithic technology under scrutiny was but one small part of the prehistoric communities’ technical milieu which was comprised mostly of organic materials.
One of the 'British' publications was authored by an American (Andrefsky), and another by a Norwegian (Ballin).[return].
The Irish publication was co-authored by a Briton (Finlay), who also co-authored the Finlayson et al. (2001) publication.[return].
Bisson (1990, 120) mentions in one line that flaking experiments were conducted on the raw material, but does not elaborate on the results, beyond that it showed that the quartz “has a large number of cracks and other flaws”.[return].
However, Woodman and Johnson (1996) had used the terms 'debris' and 'debitage' interchangeably.[return].
Festooned chert is a form of chert from the Irish midlands (Nevill 1958; Warren et al. 2009).[return].