The Odyssey to Belgium

Presentation held in honor of the Notae Praehistorica 2012 in Bruges (Belgium) (text included)

An article on the subject is to be downloaded from: Notae Praehistorica 32

[slideshare id=15670136&doc=symposiumnotae2012odysseelbkdef-121217064103-phpapp02]

Lecture: An Odyssey along the river Meuse. New perspectives on old Dutch LBK research (1925-2001) by Luc Amkreutz, Piet van de Velde & Ivo van Wijk

Luc Amkreutz

Slide 1. The presentation will be held by Ivo van Wijk (archol), Piet van de Velde (Leiden University), who will be talking in a minute and Luc Amkreutz (NMA). We are part of a bigger collective that had a particular goal and in this respect the title already gives it away a bit: to present new perspectives on old Dutch LBK research.

In the following minutes Ivo and I will briefly present our project after which Piet van de Velde will highlight some of the details and results of the research on ceramics


Slide 2. The project is officially titled:The LBK revisited:

“forgotten research” into the early farming communities of the LBK

And it was funded by NWO, the national science foundation in the Odyssey programme, This programme particularly aimed to analyse and publish ‘forgotten research. That sounds a bit spooky, but what is meant with it is research that has been lying in depots and archives for many years without being published and that is worthwhile having another look at, because it may change or increase our understanding of past events.

Objective in respect to LBK.


Slide 3. In our case this meant that we came up with 14 bandkeramik sites that were excavated in the period between 1925 and 2001 and that were never properly investigated or published. It was our intention to draw together all of this information in order to see how valuable it would be to add these results to the knowledge we have on LBK settlement in Dutch Limburg. This in fact meant re-doing the whole analysis of the excavations ranging from the old museum campaigns from the 1920’s to more recent amateur excavations. The process actually consisted of three facets.

Difficult to get everything together and use a single system to inventorize what is there


Background research. Esp. the older excavations a lot of info is missing such as daily reports, information on field drawings and photographs, often make  do with correspondence and puzzling together what was excavated form the data that was available.


Third specialist research. While Ivo and myself mainly focused on getting the bigger picture on the excavation, field data, features and number in order, a range of specialists took another look at all the finds and here you see them and that we have many facets of the LBK covered.


The big question is of course…: WHY??

Why go through so much trouble to scrape together all that information, while there are also good, well published sites….well….


Slide 4. That’s because this is the picture that many researchers in the Netherlands and abroad have of Dutch LBK. They refer to the well-known-well published and ground-breaking research, mainly by Modderman in the 1950,s and 60’s with the classical excavations of Sittard, Geleen, Elsloo and Stein. The sites that formed the basis for much of the typochronology we still use for pottery and houses and that form the context for the earliest Neolithic occupation in the Netherlands.

That is fine, but also problematical, since it offers a very skewed perspective, because the current reality….


Slide 5. …is this. Here you see the information available for LBK sites, on the Graetheide plateau, around Maastricht as well as how it links and continues into Belgium. And it is from the cluster of black dots that range from almost complete settlements to isolated finds that we have investigated the best selection:


Slide 6. And here you see a number of the sites we studied, especially and importantly including those on the other side of the Meuse. Which in act means we were becoming a bit tired of trying to understand the LBK, only looking from the Graetheide area, east, into Germany, and nowadays would much rather also cross the Meuse and see what the Belgians have to offer for understanding our earliest Neolithic.


Ivo van Wijk

Slide 6: That’s what the Odyssey is all about: giving value to an enormous dataset of find spots. For the Odyssey we’ve selected 14 find spots on both sides of the river Meuse which we thought provide a nice example of Dutch LBK sites ranging from typical LBK sites like Geleen-Seipgensstraat or Stein-Haven which is located in the Meuse valley to the oldest known but never published sites at the Caberg in Maastricht. But also a-typical sites like Echt-Annendaal on the sandy soils. These sites hold a lot of data but also provided a huge logistical operation. All these sites which have been excavated between 1925 and 2001. all were excavated by different excavators all using their own methodology and administration. It took us almost a year just to standardize all the data and make it available for our specialists.

Slide 7. As most archaeologists know, data from one site provides enough information for a decent analysis. We aimed a little bit higher and took up the task of analyzing 14 sites. As you can see it meant we had to analyse over 40.000 finds ranging from pottery, lithics to botany samples. All were sent to our participating specialists who made excellent analysis of which one will be presented by our pottery specialist Piet van de Velde.


Slide 8. But all find analysis provided excellent results. Leading factor in our analysis was the Meuse. This river functions as a topographical divider between the Graetheide and Heeswater cluster. Our analysis focussed on find spots on both sites of the river from the two clusters. With our analysis we hoped to get a better understanding of the resemblances and differences in these clusters. And in a way we hope to see whether the Meuse has to be regarded as a divider or more as a combiner. In other words, do we still have to talk about the Graetheide and the Heeswater cluster or are these boundaries slowly fading.

One effort of getting a better grip on the Dutch sites of these clusters is the Lithic analysis by Marjorie de Grooth. I won’t get into detail since Marjorie will give a presentation on the Steentijddag in February.

Much of her analysis dealt with Comparing Raw material sources and Spatial relationships between settlement clusters and flint sources. I will show you her flint extraction map, not made for the project but certainly used as a framework for the odyssey. And because it is a beautiful map I will take every opportunity to show it to the public. I can’t take credit for it.

But the Odyssey is also all about transferring knowledge. Shown here is a master class we took at her house where we were warmly welcomed by her.


Slide 9. We also had a look at the Bandkeramic landscape where our sites were located. Because of the wide topographical range it gave us an opportunity to make some assessments about the Bandkeramic settlement system. Where were they located and is there a differentiation to be made between the various settlements. Therefore we first have to research and publish all these sites.

By doing this it will be possible to analyse all the different find spots and connect them.

Not only do we have gathered a lot of dots on the map but can we also talk about dots on pottery. Dots which were numerously counted by the team to provide a relative chronology based on decorated pottery Piet van de Velde is going to talk about.


Piet van de Velde

Slide 10: introduction

My contribution will be to one of the oldest subjects in archaeology: the old-fashioned analysis and chronology of pottery decoration, this time regarding the LBK in Dutch Limburg just across the border of Belgian Limburg and Hesbay.

Slide 11: definitions

Pottery decoration has been used as a means towards chronological ordering, in much of LBK research. Especially so since the conversion of carbon isotope contents to calendar dates is unreliable for the largest part of LBK presence in north-western Europe: a wiggle in the calibration curve smoothes all readings between ca 5200 and 5000 BCE. Moreover, climate and soils in this region virtually prevent the survival of wood, which might have served towards tree-ring datings. As an answer, Modderman developed a relative chronology scheme (1959, 1970) based on stratigraphy, house construction and pottery decoration, to which estimated calendar dates have been attached later. Modderman recognized several theoretical shortcomings in his scheme which in daily practice hardly count, however, as the scheme does what it was intended for: it provides a set of pegs to anchor the chronological aspects of our analyses. Notwithstanding that all, LBK archaeologists keep spending on 14C datings. Encouraged by Modderman, back in the ‘seventies I started the development of a pottery chronology for the LBK (Van de Velde 1979) to improve his schemes –thanks to the Odyssey project all available decorated LBK-pottery from Dutch Limburg has now been analyzed and ordered chronologically.

= Methodical considerations: The Modderman scheme rests upon a mixed bag of suppositions and observations, with a different foundation in almost every phase. That way, a check on or control of its results is hardly or not possible. Restriction to one field of discourse like pottery decoration would be much better in all respects. For, in the first place, on every LBK site scores of decorated sherds are found. Then, simulations as well as theoretical considerations indicate that by number at least 90% of the original pots is still represented and so should warrant representativity. Then, the mere 5% of any pot has been preserved on the average in LBK settlement debris is sufficient for present purposes. Luckily, sherds from the same vessel are rather easily recognized most of the time – and grouped as a Sherd Family.

Slide 12.  LBK pottery decoration

The shape, the looks of material culture in general and of pottery decoration in particular is never random but according to custom/tradition of a society  –it is a stamp of identity in and of that society. The decorative repertory of the LBK is limited to the algebraic group of the (four) geometrically possible developments[1] of a bow and a spiral each, either in a curvilinear or a rectilinear execution: THERE IS NO OTHER DECORATION IN THE LBK (refer also to Van Berg 1983.) Changes occur in the tool that is used to draw the decoration on the pot surface (alternatively simple, double, multiple dented spatula), in the way the main motifs are depicted and filled (lines, pointlets, hatches, pivots), and in the complexity of the rim decoration. There are many other attributes that do change with time and/or place as well, but those shown here are easiest visible and recorded. Change in any attribute may start randomly, though in the aggregate there is a standard pattern in the frequencies of occurrence over space and/or time, which in archaeological texts generally is referred to as the battleship curve. The parameters of change differ from variable to variable, and jointly they therefore make a tight frame to measure relative chronology.

Methodically, mobile archaeological goodies are gathered per stratigraphical unit: each set a find. Per find there will be a variable number of sherd families, each family consisting of a different number of sherds, and so only qualitatively usable. By counting the occurrence of, for instance, single, double, and multi-dented spatulae in finds quantification becomes possible, but smaller and larger finds cannot validly be compared this way. However, by standardizing the counts to the number of sherd families in the find, finds become comparable independently of their size, and can be used to construct a chronological scheme.

Slide 13: chronological scheme

The chronological scheme of Dutch LBK pottery (result of a Principal Component Analysis) has been arbitrarily divided into 20 phases.<EXPAND ON COMPONENTS, PERCENTAGES, RIM DECORATION, SPATULA TYPE> The graph is intended primarily to illustrate the trends, actual values may well be off by 5 – 10%.

Slide 14: objections

Reasons for these offsets are that the graph summarizes the 168 largest finds in Dutch Limburg, aggregating 23 excavations. Local variation is consequently not visible, but should be assumed. Apart from that, only very rarely the finds are sufficiently rich to justify or to validate the computation of parts per hundred / percentages for which at the very least 20 or 25 sherd families per find unit are required.

Slide 15: chronological schemes compared

In this graph some chronological divisions of the NW-LBK currently in use (Meier-Arendt, Modderman, Stehli) are compared with my proposal. The latter has been arbitrarily divided into 20 phases. This might suggest an exactitude of just over ten years per phase (225 yrs of LBK in the Netherlands / 20 = 11 yrs/phase), probably a little overdone. As an exercise, assuming equal length for each of Stehli’s Haus Generationen (as postulated by him TICK) then my phases are longer than one HG in the older part, and shorter towards the younger end. Stehli’s division is based on calendar years, mine on statistical quantification of change.

Slide 16: a ceramic date is but a ceramic date (TICK)

Slide 17: Dutch Limburg LBK chronology

The present graph explores the relative chronological positions of the Dutch LBK sites. The customary “Modderman phases” are shown on top and indicated by the alternating vertical green and white bars. The first and larger group of sites in the graph (18 excavations, including the Elsloo Cemetery) is geographically located on the right bank of the Meuse, mainly on the Graetheide. Together their chronological bars span the total length of 20 ceramic phases. The smaller group of five sites towards the bottom of the graph (“Caberg”) is probably of more interest to the present audience: these sites are all situated on the plateau on the Meuse left bank near Maastricht, 15 kms South of the Graetheide, and 25 kms North of the Place Saint-Lambert in Liège. It is here that the Hesbayan Geer (“Jeker” in the Netherlands) and the Limburgian Heeswater discharge into the Meuse. As the graph makes clear, the Caberg area has been occupied simultaneously with the Graetheide; it was also deserted by about the same time.

Of necessity this brings up the question of the chronological relations between the Dutch and the Belgian sites. Several Belgian colleagues have suggested that at least the Hesbayan LBK is an offshoot of the Dutch branch of the NW-LBK, yet proof is still outstanding. If Dutch excavations have been published rather sparingly, so are those on the Hesbayan side of the border. It is difficult, therefore, to compare both areas AT THE LEVEL OF THE INDIVIDUAL FINDS.

Slide 18: Dutch vs. Hesbayan LBK

What I could lay my hands on has been coded and entered into the computation of a comparative relative chronology of Dutch and Hesbayan sites, as shown here. Certainly, the chart is only suggestive not conclusive: e.g., the 22 find units from Darion-Colia entered here stand for a much larger set, and I have no idea how and why they have been selected. Then, Liege’s Place Saint Lambert, as listed here, represents all pottery excavated there, but the excavations did evidently not discover the entire settlement. Similar considerations pertain to the other sites as well, also to those in the Netherlands. Given these restrictions, the Hesbayan sites are clearly younger than their Dutch counterparts: the medians of the spreads are all situated considerably more towards the right (younger) border of the graph.

Slide 19: column chart Belgian and Dutch frequencies of finds

Perhaps clearer even, is this column chart which shows the distribution of the percentages of finds over a stretch of an arbitrary 23 phases (bear in mind that the younger phases are probably half as long in calendar years compared to the older ones.) The Dutch finds (blue) have maxima in phases 3 and between 9 and 13, roughly comparing to 1c and 2b-c respectively in the Modderman scheme; the dip in phase 7 signals the 1d/2a transition in Dutch Limburg. The Belgian finds (red bars) do not occur earlier than this transition (except one single sherd family from Oleye), then go for a bright maximum between  phases 10 and 16, to disappear from MY records simultaneously with the Dutch finds in phases 17-19. After that dip a short renaissance seems to occur in the Hesbaye, without apparent parallels to the North-East (the small blue bars all refer to single pots from graves in the Elsloo Graveyard.)

That is, on the limited evidence presented here, the question may be posed whether in the Hesbay LBK phase 2d is continued, or that a cautious onset of LBK-3a is visible (cp. Dohrn-Ihmig 1979, and nearer to home, Jadin 2003, and Burnez-Lanotte & Constantin 2001, 2010.)

Slide  20, 21, 22: references, and final remarks

Pottery decoration is only one aspect of what we are doing in this Odyssey Project: as shown on this sheet several more subjects are being followed up. Moreover, a more balanced synopsis will appear in the forthcoming Notae Praehistoricae. Also, a poster announcing our endeavors is shown in the lobby of this auditorium. A report on flint distribution and use, similar to mine on pottery decoration, will be pronounced by Marjorie de Grooth on the 2nd of February next in Leiden (NL) during the regular “Steentijddag”. The full, final Odyssey report will appear in the early spring of next year.

To finish, my colleagues in this project have expressed their interest in taking up, renewing and extending contacts with our Belgian co-fraters working on the LBK. As you understand, that wish is mine, too.



[1] I.e., translation, glide, rotation and reflection; see Bell 1966: Ch. 8, and Shepard 1954: 259-293.

Leave a Reply

Your email address will not be published.