School of Engineering and Applied Science
Princeton, NJ 08544-5263 USA
Paper presented at the annual meeting for the Society for American Archaeology, Minneapolis, May 1995. Please do not circulate or quote until publication in final form.
Before going further, let me locate this process in space and time. Geographically, we are concerned with an area which reaches from the middle Danube valley where Hungary, Austria, and Slovakia come together north to the lower Vistula and Oder rivers and west to the Paris Basin, an area of approximately 750,000 square kilometers. The period that concerns us here lies between 5500 and 4000 B.C. in recalibrated radiocarbon years. In actuality, the period between 5500 and 5000 B.C. saw the establishment of farming-based communities throughout this area by the Linear Pottery culture, for which several phases can be identified. The next millennium saw the consolidation of agricultural settlement by post- Linear Pottery cultures such as the Stroke-Ornamented Pottery, the Rössen, and the Lengyel. These later groups continue many of the characteristics of the Linear Pottery culture and can be said to be part of a single tradition.
In the millennia that preceded agriculture in this region, postglacial climatic conditions and vegetation had been established, and by the sixth millennium B.C., central Europe enjoyed somewhat wetter and warmer conditions than today, by about 2 degrees centigrade. The forested terrain sheltered modern forest fauna along with bands of postglacial foragers.
In order to understand how agriculture came to central Europe, it is important to know something of the geography of this region. I prefer to simplify the very complicated patchwork of hills, mountains, plains, and streams into two major landscape zones which have relevance for the study of early European farmers. These are the upland basins drained by the major river systems of central Europe and the flat lowlands of the North European Plain. I am putting aside the mountain chains like the Carpathians, Sudetens, and Harz, and the glacial outwash plains of central Poland and Niedersachsen, for these became of interest to European farming peoples only later. The upland basins of interior central Europe had generally served as traps for wind-blown dust during the last glaciation, which formed the fertile loess soils, while the North European Plain is covered with thinner soils which had been moved around quite a bit by glacial action. In the upland basins, streams formed a dendritic pattern separated by dry watersheds. On the North European Plain, the drainage was the result of glacial action: the bogs and streams that formed in meltwater valleys and kettle lakes, connecting with meandering little rivers and the broad floodplains of major streams like the Oder and Vistula.
Within the upland basins, there was one habitat that was of greatest interest to the early farming populations. This was the valleys of the smaller streams which drained patches of the loess. Loess is fertile but dry, and these stream valleys were oases of moistness from runoff from the adjacent watersheds and from upstream. Early farming populations settled in these habitats along the smaller rivers and creeks. In the lowlands of the North European Plain, there was also one very important habitat. This was among the chains and clusters of lakes left in meltwater valleys and dead-ice features that interrupt patches of ground moraine in several parts of the plain. In some respect, these features are analogues of the upland creeks, in that they are moist habitats in the midst of drier areas of fertile soil.
The question is, how to characterize the alternative. Terms such as "colonization" and "migration" have been applied, but these are generally imprecise. A colonization implies a certain degree of intentionality and directedness, with a clear core area from which the colonizing population originated and colonies which they founded, perhaps some distance away. Migration is coming back into favor as a valid archaeological explanation, as Dean Snow's article on Iroquoian migration in the January 1995 American Antiquity indicates, but it implies large-scale and sustained movement of a population. While V. Gordon Childe and his contemporaries saw the earliest European farmers as people who frequently relocated their settlements due to exhaustion of the soil's fertility, such a view is not sustained by work done in the last three decades.
Perhaps a more precise term, although slightly opaque, is demic diffusion: "demic" because it involves people and "diffusion" in the proper meaning of dispersal, rather than the archaic anthropological usage as a hypothesized means of transmitting a cultural trait. Indeed, one might speak of a "Neolithic diaspora", in which farming communities dispersed across central Europe between 6000 and 5000 BC. While an individual Neolithic farmer may have spent his entire life within a 50 kilometer radius, without relocating, the cumulative effect of Neolithic population movements brought agriculture from the middle Danube valley almost to the North Sea and Baltic Sea within only a few centuries. There was no intentionality behind this movement: Neolithic Hungarian farmers did not decide to move en masse to the lower Rhine valley. Instead the small incremental movements of Neolithic farmers resulted in the occupation of the desirable riverine habitats across central Europe within this relatively short period.
I've suggested that we need to consider the causes of the Neolithic diaspora in central Europe to lie in the goals and aspirations of individual households as they mature through their developmental cycle. As offspring establish their own households, for example some may find that relocation from the immediate environs of their parental household to be an attractive social option. Proving that such social factors underlay the Neolithic settlement of central Europe is difficult, if not impossible, but with the dismissal of measurable factors such as soil depletion and population growth we need to look towards such explanations.
How, you may ask, might such tiny, local factors have resulted in such a rapid spread of agricultural settlements? One explanation I would like to propose is that the Neolithic settlement of central Europe belongs to a category of phenomena called "complex adaptive systems" which are only in the last decade being subjected to systematic study. In such systems, individual elements spontaneously organize themselves into a pattern and acquire collective properties. But the result is not merely complicated, but rather it is dynamic, sometimes disorderly, and almost has a life of its own.
Complex adaptive systems have several important characteristics. The first is that they are composed of many agents. Each such system is a network of many agents working in parallel and making separate decisions. Moreover, these agents make mistakes and learn. Second, the results of individual decisions are not averaged away and forgotten, but may be magnified in impact as a result of other decisions and may decide the direction the system takes. This may cause the system at times to rush forward, at times to retreat.
Many complex adaptive systems are characterized by what economists would call "increasing returns". While conventional economic theory is based on an assumption of diminishing returns, in the last two decades some economists have embraced the existence of situations of positive feedback resulting in increasing returns and multiple equilibria. In such increasing returns systems, it is often the case that small chance events and decisions early in the history of the system can determine its course. Moreover, these initial decisions are based on limited experience and may cause a potentially ineffficient or sub-optimal technology or product to be chosen and to persist as the system evolves. The economist W. Brian Arthur refers to this as lock-in by historical events: ". . . increasing returns can cause the economy gradually to lock itself in to an outcome not necessarily superior to alternatives, not easily altered, and not entirely predictable in advance."
This sketch does not do justice to the richness of recent scholarship on complex adaptive systems, but with it as background, I would like in the remaining available time to consider the Neolithic settlement of central Europe as a complex adaptive system with increasing returns but locked-in to a subsistence and settlement system which, while successful, did not turn out to be the best way to exploit the central European environment.
First, it is relatively easy, I believe, to view the residential units of the Neolithic longhouses, characterized as "households", as the effective decision-making units of Neolithic society. Hundreds of longhouses have been found in the 65 years since they were first noted by Buttler at Köln-Lindenthal and Jazdzewski at Brzesc Kujawski, and countless more have been eradicated by erosion, development, and several millennia of cultivation. These longhouses are of virtually-identical dimensions from one end of central Europe to another, suggesting domestic units of constant size. These households were the agents, the individual adaptive units of this complex system: its grains of sand, its snowflakes.
Second, although the system did require certain inputs, it did not require a powerful engine to drive it forward. There is no evidence of soil depletion, nor is there any reason to assume that population pressure caused the explosive dispersal of agricultural communities across Europe. Instead, the developmental cycle of the individual households would have provided the system with a certain "metabolism" which, I would maintain, was sufficient to propel it forward. Spontaneously self-organizing systems often suddenly rush forward as small incremental behavior of individual agents tips them at a critical point and their effects are magnified. Thus individual local decisions about household relocation echoed throughout this web of self-interested agents to result in the agricultural settlement of almost all of central Europe within a matter of centuries.
Let us make a connection to the economic model of increasing returns. Conventional economists would argue that agricultural systems are subject to diminishing returns caused by limited amounts of fertile land. Brian Arthur points out that knowledge-based aspects of the modern economy, such as high technology, are largely subject to increasing returns. The initial large investment is rewarded by progressively cheaper unit costs and rapidly accumulated experience and knowledge in making the process function more efficiently. I would argue that given the condition of effectively unlimited fertile land at the beginning of the Neolithic in central Europe, the introduction of agriculture to this area was very much a knowledge-based process. Accumulated experience, after a demanding initial investment, with adapting the cultigens and livestock to central European habitats, using a standard settlement format, led to progressively greater understanding of soils, climate, landforms, plants, and animals which enabled the Neolithic farmers to rapidly disperse across the continent. "Costs" in mistakes, labor, and start-up effort would have fallen sharply as knowledge increased. Moreover, there would have been an incentive to share information, which would have further magnified the effects of individual decisions and chance events.
A cost in such situations, however, is the potential for inefficiency as conservative but generalized technologies and methods are "locked-in" by initial choices. The Early Neolithic settlement system of concentrations of household settlements within particular microregions may have been a good choice for frontier agriculturalists but may not have been the most sustainable system over time. After about 4000 B.C., there is considerable dispersal of settlement in many parts of this area, and such nucleation of settlement is not seen again for several millennia. Moreover, it appears that the complex of domesticated plants and animals used at the earliest farming sites was not yet the integrated "mixed farming" system that characterizes later European prehistory. In particular, there are inbalances such as the very heavy concentration on domestic cattle and a low reliance on domestic pigs which later prehistoric communities in this area leveled out. Yet, for about 1500 years, it was the system of choice in central Europe, and it worked.
Earlier attempts to search for "prime movers" or "engines" behind this diaspora have been unsatisfactory. I have argued instead that viewing this process as an example of a complex adaptive system in which small decisions by individual agents, increasing returns, and the metabolism of the household developmental cycle result in sudden dramatic changes is a profitable avenue for research. Even after several years of sporadically gathering material on complex systems, my own thinking on this issue is only starting to come into focus, but I expect to pursue it and welcome constructive suggestions.