‘Trade beads and bead trade’ - unplugged

The material analysed by Callmer in ‘Trade beads and bead trade’ comprises 296 bead assemblages from Viking Age burials from across Scandinavia, each with at least ten beads recorded. These graves include a total of 13,395 beads, each of which are studied, recorded and in many cases drawn (Figure 7.1).

In order to analyse this material, Callmer devises a strictly formal system of classification, based on a combination of traits. He first distinguishes fifteen basic groups of beads, defined by material and production techniques (Figure 7.2): glass beads produced either as wound, folded, blow, drawn tubes or composite mosaics (A-K), beads made from cut glass (Q), faience (R) or different types of stone (S-V). The number of beads assigned to these groups varies greatly. Four groups occur very sparsely, namely C (folded glass beads), D (blown glass beads), U (amethyst beads) and V (jade beads), while three others are very rare: H, J, and K (4-15 beads recorded). The vast majority of beads are spread among the remaining eight groups.

EXSOO®E5apQ00OQg080®®@

WÖislefteaSaolfioilSf

oe>O0O8e®oS>s^®®araa^

POaQoögQC««^®"¹?®

p•• WoO p«t QC30ga8 ® »Ö._D¡0

Figure 7.1 Examples of Viking Age beads classified according to Callmer’s principles. Illustrated are a selection from six out of a total of 18 general bead groups: wound beads (B), folded (C), blown (D), segmented, drawn beads (E), cut, drawn beads (F) and composite mosaic beads (G). Not shown are the common (A) monochrome wound beads, rock crystal (S) and carnelian (T), and eight less common groups of stone or glass beads. After Callmer 1977, color plate III.

Every bead is subsequently classified among 27 types of shape, 4 classes of length/diameter ratios, 11 classes of diameter, 3 classes of translucence, 24 color variations, and 3 classes of ornamental patterns, which are again subdivided into 106 specific traits. In theory, this classification allows for millions of unique bead variants. Out of this number, 406 individual trait combinations are found among the assemblages recorded, while further ones are noted in other data-sets, leading to a total of 614 individually defined types.

The types thus separated allow for a detailed and nuanced classification of the material. By present standards a dataset of this magnitude is a large but manageable material for analysis. Yet for the purpose of hand calculations, it was prohibitively complex. Callmer therefore determines that the ‘necessary systematization of the beads must be worked out in steps’ (Callmer 1977, 33). He first reduces the 614 classes of beads to 54 general classes or ‘bead groups’ (Callmer 1977, 78). Each group comprises beads made in the same technique and range of colors, and which are considered relevant in terms of chronology and provenance.

Figure 7.2 Basic division of Viking Age glass and stone beads according to Callmer’s classification.

He then uses a general impression of the material to separate - on eye, so to speak - 18 preliminary (‘postulated’) groups of bead assemblages (also called ‘bead groups’), that is, not a classification of the bead types, but of the graves in which they occur (Callmer 1977, 56). Twelve basic groups are labelled bead group I to XII while six intermediate subgroups are called bead group I/II, II/III, etc. The groups are defined according to presence or abundance of seven general classes and 23 specific types of beads. These initial groups are assessed in relation to 62 types of metal ornaments found with the bead assemblages in order to assess chronological trends.

Six of the ‘bead groups’ (group I, IV, VII, VIII, XI and XII) are characterized by a predominance of various types of wound beads. Most of these are supposedly hand crafted in small batches in Scandinavia. Six other groups defined by dominance (>49 percent) of resp. segmented beads (group II and IX), small cut beads (Group III and VI) and stone beads (Group V and X).

The 18 bead groups are then examined by means of a graph analysis (Figure 7.3). Here links (lines) between these edges (bead groups) indicate

Figure 7.3 Callmer’s (1977) graph analysis of beads associated in Scandinavian Viking Age burials. To enable a paper calculation the 584 individual types of beads are lumped into 18 groups prior to formal graph analysis. Lines between these groups indicate strong (‘primary’) relations (double lines), or weaker (‘secondary’) relations (single lines). The diagram is interpreted to show that these groups can be further merged as 9 revised groups (‘Bead Periods’), assumedly chronological phases. After Callmer 1977, fig. 6.

A complex beadwork 153 common bead types in bead groups, resp. strong (‘primary’) relations (double lines), or weaker (‘secondary’) relations (single lines). The intermediate groups are shown tied closely with the main groups. As the number of groups is deemed ‘too large’, they are merged with these (group I/II is merged with II, etc.) on the assumption that this will not affect the chronological analysis (Callmer 1977, 65). Group V and X (assemblages with >49 percent stone beads) are found to display a mixed chronology, and are split and recombined with other groups. Group XI, a small group with only three assemblages, is merged with group VIII (both >49 percent wound beads). In this way nine revised groups or ‘Bead Periods’ are defined, which are proposed to correspond to chronological phases (Figure 7.4).

This step-wise procedure allows for a detailed comparison with the archaeological material. Yet each step in the process implies a reduction in the amount of relational data considered in the next step. The initial 18 groups are thus separated with reference to broad trends, without reference to a detailed comparison, while the later reduction to nine ‘Bead Periods’ is justified by pragmatic considerations, only partially supported by the relations visualized in the graph analysis. It is conspicuous that the groups eventually divide up the entire set tidily without any remaining outliers or groups. In

Figure 7.4 The distribution of the main types of beads according to the nine ‘Bead Periods’. Note non-normal peak distributions of‘A’ (monochrome wound beads, phase I & XII), E (segmented, drawn beads, phase II and IX) and ‘F’ (small cut beads, phase III and VI). After Callmer 1977, fig. 8.

some steps, for example when bead group V, X and XI are abandoned, it could have been relevant to revise the parameters and restart the analysis in a new iteration. Instead, the initial groups and group names are retained wherever possible, presumably because each change would involve a timeconsuming manual revision of records. The intricate steps with repeated sub-definition and re-working of the analytical groups makes the process of analysis difficult to follow.

Callmer proposes that the nine Bead Periods eventually defined represent a chronological sequence. The principal dynamic behind the topology of the network is supposedly a successive introduction and abandonment of types according to changing fashions. He summarizes the assumptions in what might well be called an agent-based model:

Since we may assume an extra-Scandinavian provenance for at least part of the bead material in Scandinavia in the Viking Period, we may suggest that beads were diffused to Scandinavia probably primarily, to the large centres of diffusion, in bags from the workshops. Beads could be acquired either regularly or irregularly by certain agents of further diffusion or directly diffused by the original agent. As a consequence of these assumptions and presumptions it is likely that a quantity of beads locally acquired is characterized by the workshops dominating the total supply of beads at a given time or, if there are stable regular exclusive relations between certain agents of diffusion hinted at above, a locally acquired quantity of beads could be characterized only by a part of the total available supply of beads.

(Callmer 1977,56)

In accordance with the periods defined, the main trends of the process may be summarized thus:

• • Bead Period I: At the beginning of the Viking Age, North European beadmakers produced wound beads, using a dark blue glass mass as the most abundant raw material.
• • Bead Period II: Early in the Viking Age masses of segmented beads began to be imported from the Middle East, in particular dark blue or silver and golden ‘metal foil’ beads.
• • Bead Period III: Somewhat later, the vogue for segmented beads was superseded by masses of small cut, tubular beads, also imported from the Middle East.
• • Bead Period IV: Subsequently beads cut from cornelian and rock crystal came into fashion, and were similarly imported.
• • Bead Period VII: Imported beads go out of fashion, and there is a return to wound beads, now predominantly made from green or turquoise glass.
• • Bead Period VIII: Stone beads come into fashion again.
• • Bead Period VI: (Very) small cut beads are in vogue again.

• • Bead Period IX: Eventually there is a revival of imported segmented, now mostly from yellow or green glass.
• • Bead Period XII: A final return to wound beads, many either red or transparent.

Some aspects of this synthesis are puzzling. First, it appears conspicuous that the beads can apparently be assigned to phases that are generally less than 30 years’ length, in a situation where other artefact chronologies are much broader. Second, no other dynamic except chronological succession is seen to have a significant impact on the associations of the materials. This is partially mitigated in Callmer’s subsequent regional analysis, which uses the Bead Periods to chart the relative abundance of finds in regions across Scandinavia as an index of the intensity of exchange. Third, the resulting distribution of the main types of beads according to the Bead Periods (Figure 7.4) shows a curious, non-normal distribution with two peak occurrences for each of the main groups, including ‘A’ (monochrome wound beads, Bead Period I & XII), ‘E’ (segmented, beads, Bead Period II and IX) and ‘F’ (small cut beads, Bead Period III and VI).