2 Research question, data and methods
2.1 Research questionnext section
The main purpose of this article is to consider textile production in its social context. How and where were textiles and clothes made and by whom? Was cloth production already specialized and related to an extensive trade network or was it a craft that mainly took place at the household level?
To do so, it is necessary to reconstruct how textiles and clothes were made. It may be possible to identify indications for production other than for domestic consumption in Early Medieval society by the assessment of the degree of specialization in textile production.
There are several approaches to a contextual study of textile production. First, it is possible to study textile products with a view to understanding how the cloth was made. Second, one can ascertain the degree of specialisation needed to produce the textiles and the way these textiles would have been valued by the people using them. Third is the study of the tools used to produce textiles, their development and distribution within a settlement, which may point to locations where different parts of the production process took place. A comparison between different settlements might even give information about the relative importance of textile production at these sites. Lastly, an evaluation of the access to the raw materials for textile production, like wool and dyestuffs, and indications for overproduction may give a view on the role of a settlement in the textile trade. This type of information may be acquired with a landscape-centred approach where well-documented bone spectra from different sites are available. The focus of this article is however, the textiles themselves.
The textiles discussed in this article have all been found in settlements in the Netherlands. Most of these finds were uncovered in the dwelling mounds (or terpen) of a predominantly rural society in the north of the country. Few burials have been found at these sites and textile remains are predominantly found in settlements. A small percentage was found in major centres of a quite different character, such as Dorestad and Middelburg.
Fig. 1 Locations of sites mentioned in this article: 1. Middelburg, 2. Dorestad, 3. Hogebeintum, 4. Ezinge, 5. Dokkum, 6. Aalsum, 7. Oostrum, 8. Leens, 9. Ezinge, 10. Rasquert , 11. Westeremden, 12. Stadt Wilhelmshaven.
Geographically and culturally the terpen differ from the towns of Middelburg and Dorestad. The terpen are considered to have more ties to Scandinavia, northern Germany and Anglo-Saxon England. Dorestad and Middelburg, situated on the edge of the Merovingian and Carolingian empire, are likely to have been more influenced by the regions in the south.
The textiles uncovered in these settlements have properties that make them worth treating as a separate find category among the body of textile finds from the Netherlands, in contrast to the textile fragments that have survived in cemeteries through the corrosion of metal artefacts. Although the cemeteries offer much better information for a chronological framework for textiles in use and for the reconstruction of clothes, they do not lend themselves easily to the examination of the entire process of production and use of textiles. Textiles found in settlements, on the other hand, contain information not only about how a fabric was spun, dyed and woven. They also give information about finishing processes and about how a fabric was put together, sewn, used and repaired. It is possible to cover a wide range of questions about the process of textile production on the basis of the often very large pieces of textile from the settlements.
Table 1. Early medieval sites in the Netherlands with textile remains. The second column lists the entire habitation period of the site (after
Taayke 1996; Knol 1993). The third column shows the period to which the textiles may be assigned. This is based either on associated excavated material (*) or to the fact that the majority of finds from a site is dated in this period (**).
There are, however, some disadvantages. Firstly, the textiles from the settlements are often poorly dated. This is related to the way many of these textiles have been recovered. The habitation of the earliest terpen dates back to c. 600 BC. After the third century AD a decline in population commenced, followed by a phase of scarce occupation. Population increased only after the fifth century and the terpen have been gradually raised up to their present heights. At the end of the nineteenth century the soil that had accumulated for centuries was discovered as a valuable fertilizer and therefore groups of diggers methodically dug away large parts of the mounds. These people sometimes had an eye for antiquities but as they dug straight from the top down, they could collect objects dating over 1000 years apart in one single day (Knol et al. 2005). As a result there may be textiles in the dataset spanning approximately the period from 500 BC to 1500 AD. On the other hand, scientifically excavated sites like Dorestad, Middelburg and older excavations at Zinge, Leens and Westeremden provide datable material (fig.1).
Fig.2 The textiles from Cornjum (object nr. FM 120-411) were still in their original packaging from excavations conducted in the early 1900s (collection Fries Museum). Scale in cms.
In some cases textiles are assigned to a period of several centuries, based on the fact that most other finds from these sites date from that period (table 1, **). However, some textiles can theoretically be dated to anywhere within the long period of habitation of a site. This makes it impossible to use this dataset as a whole to create a chronological framework for textiles in the Early Middle Ages.
A second disadvantage is that the dataset mainly consists of woollen textiles. During the Middle Ages people wore clothes made from animal fibres, such as wool, or plant fibres, like linen. Degradation of these fibres is caused by micro-organisms, oxidization and other chemical processes in the soil. Linen fibres break apart by a process of hydrolysis, which occurs under acid conditions. Wool on the other hand, like leather and fur, may dissolve completely in alkaline conditions and is much better preserved in acid soil (Huisman 2009). Since soils are generally acid or alkaline, it is more likely that only one of the two types is preserved at any one site. In every settlement that has been examined, soil conditions were acid, which means that the preserved textiles are made of wool. Information about linen could come from the cemeteries, which will be published in the next few years.
Lastly, it must be considered that the textiles found in refuse layers in settlements are literally refuse. The fragments are generally heavily worn, re-used and finally discarded as rags, making it difficult to ascertain their original function.
The dataset consists of 440 fragments from 265 different textiles from 31 sites (table 1). Of these textiles, 80 have already been published in some detail. The others have not been analysed until now. Some were still untouched and in their original wrapping from the early 1900s (fig. 2), or were still adhering to the clay from which they had been recovered. As a consequence, these textiles had to be cleaned with demineralised water and dried flat before analysis could take place.
The textiles may be divided into woven fabrics (226) and others like ropes, cords, braids and felt (39). The finds vary in size from small scraps of a few square centimetres to large pieces approximately 40 x 35 cm in size. The textiles are probably not only the remains of people’s clothes but may also have been used for household needs such as bedding or sacking (see 4.1 for more information about the function of the textiles).
Seventy eight percent of the textiles can be assigned to a period of several centuries within the Early Middle Ages. These finds will be presented in this paper via discussion, tables and graphs grouped per site in chronological order. The other 22% will be treated as a separate group as it is not certain whether they are Early Medieval or older.
2.3.1 Some central concepts: skill and quality
The following discussion focuses on the concepts of skill and quality. The Oxford English Dictionary defines the noun ‘skill’ as: 1 the ability to do something well; expertise or dexterity. 2 a particular ability. Production of textiles was commonplace in Early Medieval society. Manufacture for daily clothes, household furniture, and sailing clothes must have taken up a considerable share of everyday life. From a study of working time conducted in 1760, it is known that up until the eighteenth century women in south Sweden worked at least eight months a year on textile production for domestic needs only (Andersson 2003). The techniques to make textiles were therefore widespread and embedded in the entire society. Skill in Early Medieval textile production must be conceived as the ability to make textiles of a higher level of skill, using refined techniques or expertise. Skill does not necessarily lead to finer textiles, there are many other reasons for which a textile may be valued. The quality of a textile and the way it was valued depends on its properties in relation to the use of the textile. It is often a combination of several properties, which makes a fabric suitable for a specific use. In cases of clothing it is obvious that a fine and technically complex fabric is very suitable and valuable. Fabric for sailing clothes, on the other hand, would have required other properties. A suitable sail cloth may very much resemble a fabric that in terms of clothing would be seen as of average quality (Cooke et al. 2002). The appearance of a textile may also be of significance because it can be just as important in signalling social status as the form or shape of a garment (Hammarlund et al. 2008). These properties are generally not easy to measure using standard analytical methods for archaeological textiles because these methods may describe textiles as technically similar while the naked eye conceives differences (Hammarlund 2005).
2.3.2 Chaîne opératoire
An important research question concerns differentiating domestic craftsmanship from production for a textile market via the analysis of textiles. A useful tool in this regard is the concept of chaîne opératoire, or operational sequence, which considers a production process as a sequence of (interrelated) actions influenced by technical possibilities and personal and cultural choices (Skibo & Schiffer 2008). This implies a study of the individual steps in the process of textile production, which makes it possible to evaluate every action and to discern the general mode of production. In each step one can ask whether producers had access to the same materials, skills and tools in the making of the final product. Although this is a somewhat technical approach, it is useful for possibly distinguishing ordinary textiles from specialized products made by people with specific skills or tools.
2.3.3 The social modes of production
Olausson (1997) has pointed out that it may be possible to gain greater insight into social complexity by studying how production was organised. Her model, developed for Neolithic Scandinavia, has previously been successfully applied to textile production by Andersson (2003, 2007). Olausson recognises five levels of production in terms of specialisation. Table 2 shows the characteristics of the first four levels.
Household production can be present in any kind of society but if it is possible to identify the product of a specialist, this should indicate a higher level of social complexity. The products of these types of specialised production must surely have different characteristics. A craftsman, producing for his patron, will make a product that will heighten the status of this patron. His products will therefore have to show the time and effort the craftsman has put into them. An independent craftsman would not have the time to elaborate on his work in the same way. He would have to be efficient to make his money, suggesting that his products could therefore be characterised as efficient and standardised, requiring a minimum of production time and there should be little evidence of errors (Olausson 1997).
Table 2. Characteristics of the different modes of flint production defined by Olausson (1997) and applied on archaeological textiles by
Following this model an assessment of the textiles is needed regarding how they were made and whether there are indications of a specific type of production or specialization as pointed out in table 2. This is reflected in the details of the process of textile production, from producing yarns (collecting fibres and the fineness of the spinning), through the careful weaving, to the way the cloth was sewn into a garment. It is also useful to consider whether the quality of the different steps in the process is the same or not. For instance, has the same level of craftsmanship been applied at each stage of the entire process of making textiles, or has part or parts of the process been of different levels? Lastly it is important to analyse how textiles were used after production and how they were possibly valued by the person or people using them. Therefore it is important to ask what they were used for, how they were sewn and how much effort was put into repairing them until they were finally discarded. It can be argued as to whether this is part of textile production, but these questions need to be taken into account when analysing the use, repair and value of a fabric.
2.3.4 Theory turned into practice
The textiles in the dataset are analysed to ascertain the degree of specialisation in the production process. To do so, every step of the production process has been considered. Analyses of spinning, weaving and needlework have been conducted by the author. These analyses involved describing the techniques used and the quality or craftsmanship visible in the thread, weave or stitching. Sections 3 and 4 include a discussion of the variables used in measuring these qualities. Fibre and dye analyses of several samples were done by P. Walton Rogers at the Anglo-Saxon Laboratory. The following sections will present the results of these analyses and a discussion of fibre processing, spinning, dyeing, weaving, finishing processes, sewing and repairing.