Altogether, 893 fi nd numbers (including 1460 items or fragments, potsherds included)3 were col-lected from the castle deposits, in addition to a considerable quantity of zooarchaeological material.
From the town context, ca. 3750 sherds of ceramic vessels were collected, in addition to extremely few other fi nds (Fig. 3). So a statistical comparison cannot be undertaken.
Sherds of ceramic vessels form the majority of the fi nds both from the castle area (786 frag-ments, i.e. 55%), and the town context. During the period in Northern Europe, ceramic vessels form the predominant group both for food preparation and tableware, as the number of glass vessels and metal cooking utensils is almost negligible, at least among the archaeological fi nds (see e.g. Steppuhn 2003 for glass, Stephan 1982, 65–67 for a comment on the situation in Northern German towns).
In Estonia, the 13th-century contexts from towns and castles have revealed a remarkable share of imported pottery. In West Estonian towns, this forms an absolute majority of the fi nds (e.g. Russow 2006, 147, 205), while in South Estonian towns, such as Tartu and Viljandi, local products remained in use, although in somewhat smaller numbers, for the entire medieval period. The analysis of the ceramic complex from the castle of Viljandi (Fig. 4) fi ts into this general picture. Our study revealed that over three quarters of the ceramic complex at the castle site also consisted of locally produced vessels. This is not so much a surprise on the background of information from North German towns where local prod-ucts formed a considerable share of the ceramic complex until the 14th century (Schäfer 2008, 440–446).
3 Stored in the Museum of Viljandi, VM 10922: 1270–1690 and VM 11041: 1065–1504.
The share of imports in the castle context was signifi cant (ca. 22% based on sherd count), but many of the sherds were extremely fragmented and the number of vessels present might not have been great (Fig. 4). Another noteworthy issue is the existence of vessels that imitate the overall shape of stoneware jugs, but are made of local clay and probably in a local workshop (Fig. 5: 4, 7, 10). The predominant amount of fragments of such imitations known so far in Viljandi originates from the castle.
In case of the context from Pikk Street, the share of imports was almost non-existent: there were only seven sherds of imported wares recognized in the context chosen for in-depth analysis (0.2%). A comparison to the overall picture from the earlier context investigated within Viljandi town area (Haak and Russow 2013, Table 1) makes it clear that the early glazed wares (Fig. 5: 5, 8) and, to a lesser ex-tent, Paffrath-style globular pots (Fig. 5: 1) are far more common at the castle than in the town area (cf.
Fig. 4), while Siegburg proto- and near-stoneware (Fig. 5: 2, 9) has been found in both. Finds classifi ed as Lower South Saxony products (Fig. 5: 3), more numerous in the town area, most likely are refl ecting a somewhat later dating, as these were mostly used from the 1290s onwards. These are present in later
Find type Castle context: fi nd
numbers Castle context:
Brick fragments, glazed 42 61 0
Debris of bone and antler
working 30 174 0
Iron items 29 31 6
Horeshoe nails 28 28 1
Fragments of window glass 21 35 0
Hauberk fragments 21 41 0
Vessel type (Russow 2006) Dating Sherd count from the analysed castle
South Lower Saxony 1250–1325 9 1 7
Grey near-stoneware from
South Lower Saxony 1275–1350 6 1 38
Glazed redwares from North Germany and South
Scandinavia 1225–1325 13 3 –
Paffrath-type globular pots 1200–1250/75 76 4 5
Local ware 1225–1350 596 >10 >10000 Castle and from the town area.
Datings according to usage in West Estonian towns (Russow 2006). Consultations on imports by Dr Erki Russow (Institute of History, Tallinn University).
Abb. 4: Importiertes und lokal hergestelltes Keramikgeschirr aus den analysierten Kontexten von der Burg Viljandi und vom Stadtgebiet. Datierungen laut
144
assemblages from the castle, but their relatively small number in the castle context gives additional hint that its accumulation had stopped by the early 14th century.
Thus, the composition of the ceramic assemblage shows remarkable differences, especially between the town and castle complexes. A remarkable bias was caused by the pottery workshop in the vicinity of the Pikk Street site, resulting in a large quantity of sherds of local vessels at that site. Comparison with the whole town area shows that different patterns of consumption of ceramic vessels can be traced both within the town area and between town and castle. The differentiation of ceramic imports is clearly the highest in the castle area, yet local vessels, and even local imitations of stoneware drinking vessels were nonetheless used.
On the basis of these results, it is important to note that on the one hand, local pots (decorated with straight and wavy lines, Fig. 5: 6) were still considered suitable for food preparation in the castle among the imports. The tradition of wheel-made cooking pots, decorated with this pattern, reached Viljandi County probably in the late 12th century (Haak 2010) and continued into the 14th century, but was clearly on the decline after that. During the 13th century, sherds of such vessels were widespread in Fig. 5: Imported vessels and imitations of imports from Viljandi Castle (photo by H. Helves, courtesy
of Museum of Viljandi). 1 – Paffrath-style globular pot; 2, 9 – Siegburg proto-stoneware; 3 – yellow near-stoneware from South Lower Saxony; 4, 7, 10 – local imitation of stoneware; 5, 8 – glazed redware from South Scandinavia; 6 – local pottery with ornamentation of wavy and straight lines.
Abb. 5: Importiertes Geschirr und Nachahmungen der Importe aus der Burg Viljandi (Foto:
H. Helves, mit Genehmigung des Museums von Viljandi). 1 – Kugeltopf Paffrather Art;
2, 9 – Siegburger Protosteinzeug; 3 – gelber Faststeinzeug aus dem südlichen Niedersachsen;
4, 7, 10 – lokale Nachahmung vom Steinzeug; 5, 8 – glasierte rote Irdenware aus Südskandinavien;
6 – lokale Geschirrkeramik mit wellenförmiger und geradliniger Verzierung.
1
2 3
4
5 6
7
8 9
10
the town area, strongly represented in the castle complex and absolutely predominant in the Pikk Street assemblage.
There are several fi nds that may refl ect the period of construction of the castle. Fragments of brick and tile were present both in the castle and town contexts, but glazed ones were only found in the castle and are rather unlikely to be used elsewhere at that time. The same goes for fragments of profi le bricks, found in the fi ll just above the discussed context. Fragments of window-glass (and tin window frames) can also be considered as luxury items during the late 13th century, and these were missing from all the known town contexts of 13th and early 14th-century dating. There was also a glass fragment with traces of painting. Thus, our data seems to support the assumption that more decorative building materials, such as glazed brick or painted glass windows, were used in the castle. The existence of details of construc-tion in the deposit also shows that the castle was still in construcconstruc-tional stage at the time these strata had formed. This impression is reinforced by a large share of iron nails in the context (178 items), the origin of which cannot be traced for certain, but as the strata rest on a clay fi ll contemporary with the founda-tion of the eastern wing of the castle, scaffolds for construcfounda-tion or reused wood may be an opfounda-tion.
Some of the fi nds can be seen as indicators of specialised handicraft production in the cas-tle. On the basis of the already mentioned bone and antler working debris, bone and antler specialist H. Luik has suggested that antler parts of crossbows might have been produced at the site (Haak et al.
2012, 324–325, 328). In addition to that, there are numerous iron rings (1.4–1.8 cm in diameter), which may have been used for repairing hauberks.4 Therefore, it seems likely that at least two weapon-related handicrafts took place in the castle at that time.
Finally, some of the fi nds can be connected to specialised consumption and might indicate long-distance trade. In addition to stoneware already discussed, mention should be made of amber fragments, which are rather rare among fi nds from the medieval period. These might originate from present-day Lithuania and Poland. Also among the few coins, at least one originates from Germany or Holland5 and is certainly uncommon in Livonia. There are several items that may be seen as box fragments (decora-tions, handle, etc.), which is also an item connected to travelling and long-distance contacts. To sum up, there are several items that look far too luxurious to be associated with the castle builders or less-well-off handicraft persons.
The item collection from the chosen context seems manifold in its composition, thus it is likely to have been redeposited, refl ecting several activities that took place in the larger castle area during the peri-od in question. It therefore cannot be interpreted as originating from a single event, e.g. the fi rst habitation stage in the yard of the newly-built castle, as suggested earlier (cf. Haak 2004, 109 and Haak 2005, 92).