Discussion
What can we learn from these scenarios? Are they plausible, and can the degree of credibility be estimated? Is it possible to validate such as model at all? WELAS- SIMO is a simulation of a complex system of human—environment interactions and processes, which took place 6000 years ago; it is impossible to validate such a model with the same degree of reliability as a model of recent processes, because of the lack of real, “living” analogies of the model.
However, it is possible to check the general plausibility and perform some tests to build confidence. The basic properties are taken for granted and need not be validated, as they are evidenced unequivocally by archaeology and archaeobiology: people lived in wooden houses that were constructed in close vicinity of lakes and mires and lived on a diet comprising of cultivated and gathered plants and animal products of domestic and wild animals. Also, a general transferability of knowledge on recent ecosystem features to the reconstructed ecosystem of the 4th Millennium can be taken for granted. When it comes to the details in the data used, the difficulties arise; e.g. for the amount of calories that may be gathered on one hectare of primary deciduous forest (see Table 9.2), or for the rate of annual variation in the gathering success of the settlements. Such numbers are very difficult to verify, as modern analogies of hunter-gatherers especially in the biome of the mid-latitude deciduous forests are virtually non-existent and thus, ethnographic evidence is scarce. Thus, I regard these data as good estimations that are open to discussion. Here lies the first one of a number of advantages of the approach: the input data is transparent, and can be modified or supplemented quite easily, if necessary. If the validity of the database is generally agreed on, keeping the articulated uncertainties in mind, the next question concerns the plausibility of the different scenarios. With the presented approach, a broad variety of different assumptions and hypotheses can be displayed. This helps to reduce the danger of too narrow or fixed presumptions in the modeling process. But at the same time, the question arises how to evaluate the scenarios; is it possible to isolate the one “solution scenario”, giving an answer to “how were things REALLY working back then”? The answer must be “no”, because no validation method exists with the required degree of reliability. But can certain scenarios be ruled out for some reasons? Work load calculations are an important factor in this discussion. These are not included in WELASSIMO, because only independent data that is not used for the simulations can be used as a test for the scenarios. Furthermore, the temporal resolution of one year turned out to be unsuited to integrate work load in detail, which is relevant rather on a daily or even lower scale. To check whether the scenarios are in accordance with the available work force, I use data published by Kerig (2008). He could demonstrate that one person may prepare one ha of crop fields on already established plots in 29 days. Probably imposing a somewhat stronger limitation for field size was the labor associated with weeding and crop harvest; here, Kerig states that one worker with a silex-knife can process 0.9 ha during the harvesting season (Beginning of July—mid of September). If the available workforce of the households in WELASSIMO is assumed to be 2.5 full-working persons (two adults and one child), 2.5 ha can be seen as the potential maximum a household could handle. The largest fields in a permanent cultivation technique (for which Kerigs figures apply) are simulated to be 2 ha, if 100 % of the calories are provided by crop plants and NIC is assumed. IGC and IFH have higher yields per ha, which possibly decreases the associated work load— however, this might be neutralized due to the additional workload in these systems for manuring and intensive care. The work load associated with SC is described by Ehrmann et al. (2009) according to results obtained in a large-scale experiment in Forchtenberg, Germany. According to their research, 800 full working days are necessary for one hectare of fields in SC. With an assumed average yield of 3000 kg of cereal using SC, as found by Ehrmann et al. on soils of “average” quality, and 2.5 workers as above, 96 full working days per year are required in order to meet the calorie requirements discussed for scenario 1 (see above). Ehrmann et al. calculate with the value they found for “good” soils of 5000 kg/ha; thus, a total of 50 days would be needed with the other configurations unchanged. Even if a hypothetic proportion of 100 % crop plants is assumed, and parts of the work could be done outside of the time needed for field preparation and harvesting, the respective work load could be handled. To sum up, the work load associated with crop production would most probably not pose a severe limitation for any of the scenarios realized within WELASSIMO. To investigate on the combined work load of all subsistence practices together would require a simulation with daily resolution; this is aimed at in future projects. The work load associated with Hazelnuts is in detail discussed in Holst (2010). She states that the harvesting season for Hazelnuts lasts around 14 days in the end of September, so no conflict arises with crop harvest. In one hour, she could harvest 1.26 kg of Nutmeat (equivalent of 1400 Nuts with 0.9 g each). 2.5 workers could harvest at the very least 300 kg per season. With 6000 kcal per kg of nutmeat, this equals 1.800.000 kcal per group of 6 or 300,000 kcal per person. For the configuration as in Scenario 4 with 85 % vegetable foodstuff of which 80 % are gathered, the requirements are 496,000 kcal gathered. So these numbers do not seem realistic; but if the importance of gathered food is reduced to 50 %, the assumed 300 kg Hazelnuts per group of 6 are sufficient.As stated above, it has become evident in the modelling process that neither of the previous land-use models (SC, NIC and IGC) is capable of integrating all data with confidence.
This is better achieved with the alternative scenario of “Integrated Forest Horticulture (IFH)”, which is formulated above. The fact that seemingly contradictory positive evidence on crop husbandry in the wetland settlements in the early 4th Millennium, which led to the formulation of the hypotheses SC, IGC or NIC, is included without being contradictory, is a hint that IFH might be the closest one to the Neolithic “reality” of all 4 land-use methods discussed. Until now, the charcoal and secondary forest peaks detected by pollen analyses were mainly seen as a by-product of SC (e.g. Rosch et al. 2014); if this was true, the evidence of manuring of the crop plants as documented for the site Ho1A (Bogaard et al. 2013) as well as the evidence for permanent fields with a distinct weed flora (e.g. Maier et al. 2001; Maier 1999; Jacomet et al. 2004, p. 136) would require explanation— which is very hard to give, then. Yet if it is assumed that the origin of the weed remains and the evidence for manuring is indeed a permanent and intensive form of crop husbandry, as interpreted by the researchers (e.g. Bogaard 2002; Bogaard et al. 2013), an explanation for the peaks documented by palynology is needed. As shown in scenario 4, a targeted promotion and facilitation of secondary forest growth by the use of fire and the parallel application of intensive cultivation methods for crop husbandry would be in good accordance with the existing data and at the same time provide systemic benefits, which could be demonstrated above. Here, the second advantage of an agent-based simulation approach becomes evident: the loose ends of the observed system, which are provided by the involved disciplines, may be woven into a stringent net of explanation, which considers the most of the available data. In the discussion of the increased value of patches with fallows or young secondary forest vegetation in contrast to their “wild” environment, the term “landscape” is so illustrative that I want to highlight its significance and sharpen its scope. One might as well use the term “cultural landscape”, but this is actually a pleonasm; both denote a strip of land that has been anthropogenically shaped. In the case observed, the landscape offers a higher economic value than the “wild” environment (I set “wild” in quotation marks because most likely, even those areas were influenced by humans to a certain degree). This applies most drastically for the amount of calories from gathered plants, especially from hazelnuts. To a minor extent, also the quantity and quality of livestock fodder increases in these areas, and also the availability of suitable timber in the right dimensions for house construction is triggered by the initiation of secondary forest growth. It is likely, that even the density of certain wild animal species used for hunting is positively correlated with a certain degree of openness of the landscape. In spite of these advantages, I do not claim that IFH is depicting “true” Neolithic conditions without any doubt, and that the other hypotheses are wrong; to the contrary, it is highly probable that elements of SC, IGC and NIC have been realized as well, but not in an exclusive, but in an integrative manner. And even the IFH Scenario is simply a model that explains some data—many aspects are not included, others are generalized, rough assumptions are used: it has to be stressed that no model can ever display the full complexity of such a system.9.6.1 Resource Use as a Need for Relocation?
The main question named initially is, whether a necessity for a settlement relocation arises due to the land and resource use of the people. Which subsistence strategy might contribute to this need, if any? For crop husbandry, the only scenario simulated in WELASSIMO, where a settlement shift after one or two house generations arguably makes sense, is SC. After this time, the distance to new fields might have grown large enough to motivate a settlement shift. If the parameters are set to force the maximum number of 120 people in WELASSIMO as far away as possible in a system that is still in accordance with the data—exhibiting the general properties as described above—then, a maximum distance of 2.5 km is necessary, which is reached after 30 years (SC, 5 % animal products, 80 % of plant calories are cultivated).
In his investigation on current practices of SC in India, Pratap (2000, p.72) describes an example that daily walking distances to SC plots are normally 1-2 km. On the other hand, the same author states on p. 84 that “the axiom of settlement mobility, as a logical necessity in SC Systems, is itself an a priori assumption often overriding the complexity of real-world situations.” Furthermore, Saradindu (1967) observes that fields are shifted annually in much larger radii of 6-7 miles. This means that daily travelling distances of 2.5 km are not necessarily forcing households to shift their location. Considering the alternative explanations of findings— that were originally used to back the hypothesis of SC—as described in the IFH-scenario, I think, that the theoretical benefits of the SC-Hypothesis need not further be discussed, because there is positive bioarchaeological counterevidence against it as the dominant mode of husbandry. In the other scenarios 4, 3 and 2, no economic reason appears to be motivating a settlement shift (for 2 and 3, the same restrictions are valid as for scenario 1: they are rather theoretically relevant, because they do not explain all data). To the contrary: as described above, the economic value of a landscape would rather increase with ongoing duration of use through an increase in the amount of livestock fodder and especially the gathering value connected with hazelnuts. Furthermore, paths have formed and familiarity with the surroundings may be suspected. The only resource which causes slightly increasing distances for its extraction is timber; but distances are far less than for the other activities with a maximum of 200 m for 4 households after 20 years. From all this it seems plausible to assume a certain degree of spatial stability and commitment to the site—but nonetheless, in many cases, the contrary is evidenced: a highly dynamic settlement system with a frequent relocation of the sites (e.g. Ebersbach 2010). From the preceding considerations, it is highly implausible that the land use activity of a small community caused the observed relocation pattern. Even if the number of households is set to 20, this does not change the situation drastically. It is a different thing, however, when not a hypothetical determining necessity for a relocation, but the relative benefit of a settlement shift is considered. Even if the landscape around the settlement is something appraised that is to be maintained, low-distance relocations might nonetheless be assumed, as the appreciation of the economic area does not necessarily include the site itself. A similar situation is documented for the site of Ho1A: after the documented destruction of many of the houses in 3910, a number of new houses with the building year 3909 were constructed only a few hundred meters away in the site of Ho3 (Dieckmann et al. 2006, p.418; Billamboz 2006, p. 314). As the construction elements of the houses had a limited lifetime seldom exceeding the span of 10-20 years, a restoration or rebuilding was required quite frequently. Suitable timber in considerable numbers was thus needed in a temporal pattern of relatively high predictability. Additionally, the availability of fuel wood likely decreased in the proximity of the settlement in the course of several years. Furthermore, the specific site preferences might have caused a small-scale shift of the buildings, as lake levels of Lake Constance as well as of smaller lakes in the vicinity did fluctuate markedly and frequently. So if a house needed to be rebuilt for whatever reason, and a stand of suitable timber was growing adjacent to a suitable settlement location—maybe people rather carried their few belongings to a newly constructed house, than carrying (the timber of) their new house to their belongings, if all the economic benefits of the cultural landscape could be maintained. Drivers and benefits of the abandonment of a specific site and its relocation inside the cultural landscape may have been acts of convenience, of hygiene or of environmental forcing other than resource avail- ability—or a combination of those. The above discussions are valid especially for smaller settlements with a relatively low population density of the landscape; a different situation might arise, if more and larger settlements in a defined landscape need more resources and affect their environment more intensely; a simulation to deal with this situation is currently prepared for the sites located near the effluence of Lake Zürich into the river Limmat.9.7
More on the topic Discussion:
-
Conflictology -
Ecology -
Economy -
Finance -
History -
Law -
Medicine -
Philosophy -
Religious studies -