268
Documenta Praehistorica XLVI (2019)
Introduction
What we call the Neolithic is shorthand for several
historical processes on different time and spatial
scales. Nevertheless, the Neolithic is not just a con-
struct, it is real and has some kind of downward cau-
sality on all the historical processes that make it.
The historical processes behind the Neolithic are a
result of the formation and development of a relati-
vely stable and resilient assemblage of human-mate-
rial relationships which develops in an increasingly
structured, organized and consistent social world
(Robb 2013). The Neolithic assemblage originated in
the Near East, where by 9500 cal BP people had
domesticated all the major crops and animals. They
started to make and use new things, including pot-
tery, figurines, polished stone axes and houses, be-
gun to live in villages and practice new rituals.
What is a proper scale to study the Neolithic? Behind
the long-term directionality and near irreversibility
of the process is the great local variability seen in
Neolithic and Copper Age settlement dynamics
in the Western Carpathian Basin and Eastern Alps
ABSTRACT – The paper tackles the spatio-temporal patterns of Neolithic and Copper Age settlement
dynamics in the Western Carpathian Basin and Eastern Alps with spatially explicit use of radiocar-
bon dates. It focuses on the spatial process of spread, movement, aggregation and segregation in the
time frame between 8500 and 5000 cal BP. The distribution of Neolithic and Copper Age sites in the
study area is clustered and patchy. The first Neolithic settlements appear as isolated islands or en-
claves which then slowly expand to fill neighbouring regions. After 6300 cal BP the study area expe-
rienced a significant reduction in the extent of settlement systems, associated with the Late Neolithic
to Copper Age transition.
IZVLE∞EK – ∞lanek se posve≠a prostorsko-≠asovnim vzorcem neolitske in bakrenodobne poselitve za-
hodne Panonske ni∫ine in vzhodnih Alp s pomo≠jo prostorsko eksplicitne rabe radiokarbonskih da-
tumov. Analizira prostorske procese ∏irjenja, gibanja, zdru∫evanja in razdru∫evanja v ≠asovnem loku
med 8000 in 5000 pred sedanjostjo. Neolitska in bakrenodobna najdi∏≠ se pojavljajo v neenakomer-
nih skupkih. Prve neolitska poselitev se pojavi v obliki izoliranih otokov ali enklav, ki se po≠asi ∏iri-
jo v okolico. Po 6300 pred sedanjostjo lahko zaznamo upad poselitve, ki ga povezujemo s prehodom
iz poznega neolitika v bakreno dobo.
KEY WORDS – Neolithic; Copper Age; Carpathian Basin; East Alps; settlement patterns; radiocarbon
dating; demography; modelling
KLJU∞NE BESEDE – neolitik; bakrena doba; Panonska ni∫ina; vzhodne Alpe; poselitveni vzorci; radio-
karbonsko datiranje; demografija; modeliranje
Neolitske in bakrenodobne poselitvene dinamike
v zahodni Panonski ni/ini in vzhodnih Alpah
Dimitrij Mleku/ Vrhovnik
Department of Archaeology, Faculty of Arts, University of Ljubljana, Ljubljana, SI
Institute for the protection of the cultural heritage of Slovenia, Ljubljana, SI
dmlekuz@gmail.com
DOI> 10.4312\dp.46.16

Neolithic and Copper Age settlement dynamics in the Western Carpathian Basin and Eastern Alps
269
ses and transformations in the established Neolithic
societies (Hofmann, Gleser 2019).
The aim of the paper is to approach the process of
Neolithic and Copper Age settlement of the Western
Carpathian Basin and Eastern Alps based on the avai-
lable radiocarbon data. It applies spatially explicit
use of radiocarbon dates to understand spatiotem-
poral trends. We are interested in the spatial pro-
cess of the spread, movement, aggregation and seg-
regation in the time frame between 8500 and 5000
cal BP.
The settlement dynamics proxies that revealed these
processes’ dynamics are based on the temporal fre-
quencies of radiocarbon-dated archaeological sites,
which are represented as summed probability densi-
ties (SPDs). The underlying assumption is that the
number and distribution of radiocarbon dates in
time and space indicate the existence of settlement
systems and reflect demography, as more people and
more settlements result in more activity and more
radiocarbon dates.
This is an explorative study. Its goal and focus are
to identify large spatio-temporal patterns in the pro-
cess of Neolithic settlement in the area around the
Eastern Alps and not to test mono-causal explana-
tions for dynamic processes of cultural change. In
this way, it is an open-ended study without definite
explanations.
Materials, methods and assumptions
The study area covers around 170 000km
2
and en-
compasses the western part of the Danube water-
shed above the Danube – Sava confluence. This in-
cludes the western part of the Carpathian basin, Eas-
tern Alps and north-eastern section of the Dinaric
Alps.
The study area was divided into grid cells over which
we summarized spatial variables. Hexagon cell shapes
were chosen as regular hexagons are the closest
shape to a circle that can be used in a tessellation.
Hexagons have reduced edge effects and have iden-
tical neighbouring cells, each sharing one of the six
equal length sides. Furthermore, the distance be-
tween centres is the same for all the neighbours.
A database of 141 sites with available absolute dates
from the Neolithic and Copper Age was compiled
for the study area. The observed mean distance be-
tween sites is around 11km, while the expected
the archaeological record across Europe. On the
broad scale, we can see a process of the movement
of groups of people and new material assemblage
from the southeast towards northwest over a span
of several millennia, which resulted in a uniform and
coherent thing we call the Neolithic (Robb 2013;
2014).
By about 9000 cal BP this assemblage had spread
to south-eastern Europe. The Neolithic assemblage
spread rapidly from the Aegean through the Balkans,
along the northern coast of the Mediterranean, and
across the Northern European Plain. The spread in
other areas was slower. There are regions which did
not become Neolithic for up to a millennium after
their initial contact with farmers.
The spread of the Neolithic assemblage was first es-
timated to be around 1 kilometre per year, covering
the distance between the Levant and Scotland in
about 3000 years (Ammerman, Cavali-Sforza 1984).
More recent research has refined this picture sub-
stantially (e.g., Gkiasta et al. 2004; Bocquet-Appel
et al. 2009; Fort 2015). Recent research has also de-
monstrated that this was not a uniform, ‘wave of ad-
vance’. Most archaeologically demonstrable move-
ments of people seem to be leap-frog migrations in
which small groups leave their community to es-
tablish enclave settlements in suitable environments.
This is best seen in percolation of the LBK settle-
ments in the river valleys of Central and Western
Europe and the spread of Impressa settlements over
the Mediterranean.
With the recent development of AMS dating and accu-
mulation of data, it is possible to access the dynam-
ics of spread in much finer temporal and spatial
resolutions. A large quantity of AMS radiocarbon
data – each individually dating a single event of the
end metabolism of an organism – transforms into a
new quality, allowing us to glimpse larger spatial
and temporal patterns. This radiocarbon ‘Big Data’
allows us to approach the Neolithic as a set of local
historical trajectories, each with its own speed, tem-
po and rhythm. It enables us to change the narrative
of gradually spreading Neolithic assemblage to a se-
ries of regional or local responses and actions be-
hind the larger process. In this way, the Neolithic
becomes less a uniform process, driven by a single,
perhaps evolutionary principle (Shennan 2018), but
instead a true historical development. The Neolithic
also gains temporal depth. Instead of a narrative of
the spread of a formed Neolithic assemblage, we can
begin to appreciate the complexity of social proces-

Dimitrij Mleku/ Vrhovnik
270
mean distance is around 20km, indicating the clus-
tered distribution of dated sites. This also dictated the
spatial resolution of the study. Grid cell diameter was
chosen to be 25km, with each grid cell covering ap-
prox. 520km
2
. In this way, the diameter of grid cell is
aproximately a double mean distance between sites.
The study area is covered with 320 grid cells. Due to
the highly clustered distribution, only 77 grid cells
are occupied with sites, forming several distinctive
clusters. Most of the grid cells are occupied with only
one site (43 grid cells, with a median one site per
grid and a third quartile of two sites per grid cell),
with the densest grid cell occupied by nine sites.
At this resolution, we assume that each grid cell re-
presents the area of a regional settlement system
(Kowalewski 2016) or settlement cluster (Parkin-
son 2002.397–398). A regional settlement system is
defined here as interacting interdependent groups
of people. It contains several (or several tens of) set-
tlements and communities, tied manly with an ex-
change of various kinds into “regionally-integrated
social networks” (Parkinson 2002.395)
A database of 815 radiocarbon dates from Neolithic
and Copper Age contexts between around 8000 and
5000 cal BP was compiled for all sites in the study
area. Neolithic and Copper Age contexts were de-
fined on their material assemblage (presence of hou-
ses, pottery, domestic animals, and plants); pragma-
tically this means that they were already assigned to
one of the regional Neolithic or Copper Age cultures
(LBK, Star≠evo, Lengyel, etc.) by the authors of the
original publications. In order to uphold the quality
of the database, all problematic dates (dates that
seem too early or too late for a given context) and
dates with standard deviations greater than 100 years
were discarded, resulting in 750 dates being used in
the analysis (see Appendix at http://dx.doi.org/10.
4312/dp.46.16).
The settlement proxies used in the study are based
on the temporal frequencies of radiocarbon-dated
archaeological sites, which are represented as sum-
med probability densities (SPDs). This proxy assumes
that the temporal frequencies of dates in a given site
indicate relative human population size and density
of occupation at the site. The SPDs are mainly used
Fig. 1. Study area with sites and grid used in the study.

Neolithic and Copper Age settlement dynamics in the Western Carpathian Basin and Eastern Alps
271
in demographic studies, while here they are used in
a slightly more general way as indicators of grid cell
occupation and therefore the existence of regional
settlement system in a grid cell.
There are several potential issues associated with the
use of summed probability densities, which are sum-
marized by Alan N. Williams (2012), mainly being
problems of sample size, intra-site sampling, tapho-
nomic loss and calibration effects.
Two fundamental assumptions of the method are
that the radiocarbon dates used in these analyses
are associated with occupation events; and that the
number of dates from a region represents the occu-
pation events in the region. The first assumption is
based on the logic of the selection of archaeological
samples for dating. The second assumption is not
necessarily true, as radiocarbon samples are not col-
lected randomly between and within sites, and the
process is heavily biased by sampling intensity and
history of research. The collection of radiocarbon
dates is always driven by specific research interests,
and consequently the number of dates coming from
different phases on the same site may often be a con-
sequence of the research questions being asked.
However, this bias is to some degree offset by aggre-
gation of data. The working assumption of summed
probability analysis is that a sufficiently large regio-
nal sample of radiocarbon dates will counteract any
problems at the site level, and that multiple small
non-systematic samples from a large assemblage of
sites constitute a quasi-random sample of regional
trends in occupation (Williams 2012.580).
In order to address this bias, the radiocarbon dates
are binned (or aggregated) within grid cells. Radio-
carbon dates are first binned into grid cell phases
and then sorted in decreasing order within each grid
cell phase (Shennan et al. 2013; Timpson et al.
2014). The dates within a given grid cell phase were
further subdivided into bins if the difference be-
tween two adjacent dates was greater 200 radiocar-
bon years. The dates are first calibrated and summed
within bins, with a bin sum normalized to the area
of 1, and the resulting bin sums are then summed
and normalized to produce the final SPD curve for a
Fig. 2. A number of radiocarbon dates per grid cell. Values are log10 scaled.

Dimitrij Mleku/ Vrhovnik
272
grid cell. This procedure controls for research bias
when it comes to the frequency of samples per site
or site phase, but it does not control for the bias
stemming from the different regional histories of
research.
All analysis was performed in an R statistical environ-
ment (R Core Team 2018), using the rcarbon pack-
age for radiocarbon calibration (using the IntCal13
radiocarbon curve; Reimer et al. 2013) and SPD ana-
lysis (Bevan 2018) and sp package for spatial ana-
lysis (Bivand et al. 2013).
For each grid cell, a normalized summed calibrated
radiocarbon probability distribution was calculated.
The number of radiocarbon dates varies from one
per grid cell (9 grid cells) to 88 radiocarbon dates
per grid cell with a median value of four dates per
grid and third quartile at eleven dates per grid cell.
The ranges were calculated on the basis of the high-
est probability density and are the shortest ranges
that include 95% of the probability in the probabi-
lity density function.
The lower 95% range endpoint date was taken as
the start of the Neolithic at a particular grid cell. This
was then used to estimate the spread of the Neoli-
thic across the study area using kriging interpolation
(see Brami, Zanotti 2015).
Kriging is a two-stage geostatistical method which
begins with analysis of the gathered data to estab-
lish the predictability of values from place to place.
This results in a graph known as a semivariogram
which models the difference between a value at one
location and the value at another location accord-
ing to the distance and direction between them (Chi-
lés, Delfiner 2012.147–150). Based upon these, it
estimates values at those locations which have not
been sampled. The technique uses a weighted aver-
age of neighbouring samples to estimate the un-
known value at a given location. Weights are opti-
mised using the semivariogram model, the location
of the samples and all the relevant inter-relation-
ships between known and unknown values. The
technique can also asses the uncertainty of the pre-
dictions.
Kriging data in our study consists of grid cell cen-
troids with the date for a beginning of the Neolithic
occupation, calculated using the procedure described
above. Grid cells with only one radiocarbon date
were excluded from the interpolation. The result of
kriging is an interpolated surface with values for the
earliest estimated date of Neolithic settlement with
a spatial resolution of 12.5km.
This data was used to compute the direction and
speed of the spread of the Neolithic. The aspect and
slope for 12.5km large grid cell were computed on
a smoothed surface. The slope is in this study is de-
fined as the rate of change between adjacent cells,
expressed as the time to traverse from each cell to
its neighbours, while aspect is defined as the direc-
tion of maximum slope from each cell to each of its
neighbours. Slope and aspect were visualized as a
vector field, with the size of each vector indicating
the speed and direction of spread.
SPDs were also used for crude demographic estima-
tion, which is the most common use of summed cali-
brated radiocarbon probability distributions. In most
of the palaeodemographic sites in studies, SPDs are
summed together to an empirical SPD that is treated
as a proxy for demographic dynamics. Therefore, it
is a number of sites and extents of activity at a parti-
cular site that provide a proxy for demographic
growth. Empirical SPDs are compared to theoretical
growth curves to test the statistical significance of
the empirical SPD curve (Shennan 2009; Por≠i≤ et
al. 2016; Blagojevi≤ et al. 2017).
In this study the normalized SPDs for each grid cell
are summed together. SPDs were thus aggregated or
binned over grid cells. This approach offsets bias in
the selection of regional research histories. Thus, a
grid cell with one site has the same weight as a grid
cell with many sites, as we assume that the difference
in a number of sites is a direct result of sampling
bias. The assumption is that each grid cell (and there-
fore local settlement system) has the same maximum
population (which is of course not necessarily true).
In this way, SPDs provide only a dynamic compo-
nent, an indication of a change in settlement inten-
sity over the grid cell, while the number of grid cells
provides the main proxy into overall demographic
dynamics.
Although this is an explorative study, we compared
the empirical SPD curve against the theoretical null
model of population growth. The null model assumes
that the underlying population was stationary. Stati-
stically significant positive local deviations from the
null model (peaks) occur between 6860 and 6180
cal BP, while significant negative local deviations
(dips) appear at 8000–7630 cal BP, 5880–5730 cal
BP, 5450–5390 cal BP and 5350–5260 cal BP.

Neolithic and Copper Age settlement dynamics in the Western Carpathian Basin and Eastern Alps
273
We also the compared empirical SPD curve for the
study area with the SPD created from the subset of
dates from the SE Alps area. The idea was to evalu-
ate regional variations in settlement trends (see
Timpson et al. 2015; Crema et al. 2016), and to test
whether differences between curves are statistically
significant, possibly indicating different settlement
trends in the SE Alps area. We found significant po-
sitive local deviations (higher settlement intensity in
the SE Alps area) at 6390–5940 cal BP, 5880–5780
cal BP and 5600–5390 cal BP and significant nega-
tive local deviation (lower settlement intensity in SE
Alps) at 7810–6850 cal BP.
Another similar, even simpler graph is the number
of grid cells occupied at a particular time. This esti-
mate gives the extent of the settlement system and
can provide insight into the spatial dynamics in
terms of the expansion and contraction of regional
settlement systems. It was constructed by counting
grid cells where there is an indication of occupation
(with 95% probability) for every century, and sum-
marized in a graph.
Results
The patchy distribution of occupied grid cells re-
flects the uneven density of Neolithic sites in the
study area (Figs. 2–3). Grid cells are agglomerated
into several contiguous clusters, two in Slavonia, a
large one stretching across the SE Alps, across West-
ern and Central Transdanubia, and a third in the
Vienna basin. There are also some curious gaps, an
especially large one in the Alps, but also smaller gaps
in the middle reach of the Sava (Posavina) and Dra-
va rivers (Podravina), parts of Southern and Central
Transdanubia and Styrian basin.
This is probably a result of research bias, as most of
the new dates are from recent research, especially in
relation to the Slovenian and Hungarian motorway
construction programme. However, it also reflects a
deeper pattern, as Neolithic sites seem to avoid hilly
and mountainous terrain.
When we plot each grid cell with the dates of earli-
est occupation (Fig. 3) it can be noted that the ear-
Fig. 3. The earliest appearance of the Neolithic settlement in a grid cell.

Dimitrij Mleku/ Vrhovnik
274
liest grid cells are concentrated in the SE edge of
the study area (mainly Slavonia around 8000 cal
BP), but isolated grids cells with very early dates are
spread all over the study area. It seems that within
500 years after the first appearance of the Neolithic
in Slavonia, Neolithic sites can be found all over the
study area, except the Alps. Thus we have the earli-
est appearance of Neolithic settlements after 8040
cal BP in Slavonia (Sopot: Krznari≤ πkrivanko 2011),
then after 7830 cal BP in the Vienna basin (Brunn
am Gebirge: Stadler, Kotova 2010), after 7780 cal
BP in the Budapest area, and after 7590 cal BP in
western Transdanubia, at the edge of the SE Alps
(Szentgyörgy-Pityerdomb: Bánffy 2004). There seem
to be two possible corridors of expansion from the
Slavonian core area, one along the Danube and the
other on along Drava River and then along the east-
ern edge of the Alps.
The first Neolithic thus appears as isolated islands or
enclaves of Neolithic settlements which then slowly
expand to fill neighbouring regions. However, there
are some areas, especially the SE Alps west of the
Mur River, which are consistently settled much later
than their neighbours.
The spatio-temporal pattern of the 2000-year long
process of the formation of Neolithic settlement sys-
tems in the study area is clearly visible on the map
of the estimated age of the arrival of the Neolithic
(Fig. 4).
The core area for the spread of the Neolithic is that
between the Sava and Drava. From the origin in Sla-
vonia, the Neolithic expands in two prongs, one along
the Danube and the other along the Drava, Mur and
eastern foothills of the Alps. This expansion is in the
form of several very early enclaves with a much ear-
lier appearance of the Neolithic than the surround-
ing areas, such as those enclaves along the Danube,
Vienna basin and Western Transdanubia. Those en-
claves are limited one or two grid cells, and might in
some cases reflect the research bias. What we see is
a very crude remnant of a string of small communi-
ties stretching along expansion corridors.
There are also some backwater areas with much later
Neolithic occupation. The most prominent being the
area of the Alps and the smaller area around Bala-
ton lake. While those small backwater areas are most
probably the result of research bias, the Alps area
Fig. 4. Isochrone map of the estimated age of the beginning of the Neolithic, result of a kriging interpola-
tion. The contour interval is 100 years.

Neolithic and Copper Age settlement dynamics in the Western Carpathian Basin and Eastern Alps
275
does not seem to be an artefact. A large number of
dates from the SE Alps indicate the relatively late ar-
rival of the Neolithic with rapid expansion along ri-
ver valleys.
Dense isochrones indicate the existence of a statio-
nary border, most prominently on the edges of Car-
pathian Basin and the Alps, along the lower course
of the Mur river, where the Neolithic expansion to-
ward the west halted for almost 500 years with a
stationary border, and more than 1000 years with a
stationary border on the western edge of the Vienna
basin toward the Alps.
The distance and shape between isochrones encode
the rhythm, tempo and direction of the process,
which can more clearly be visualized as a vector field
(Fig. 5). The overall speed of the process seems to be
quite rapid. The study area was crossed in a direction
from SE to NW in around 200 years, as the 370km
distance between Sopot in Slavonia and Brunn am
Gebirge in the Vienna basin was covered in a span
of around 210 years, which gives an average speed
of Neolithic expansion of about 1.7km per year. Thus
is a speed of enclave colonization over the study area
that reflects the high mobility of early Neolithic com-
munities.
The local speed of expansion was estimated to be
from 0.025 to around 5km per year, with the medi-
an speed around 0.15km per year. The local speeds
estimated in this study indicate other processes, a
relatively slow expansion around core regions and
enclaves that filled the landscape.
The estimated speed of expansion is the highest in
the areas of no data, such as the Alps and middle
reach of the Sava, where it slows down when en-
counters Alpine foothills, once it reaches the area
where we have more data. This points to significant
gaps in the data.
The general direction of expansion is mostly from
the core areas and enclaves toward surrounding re-
gions. Even so, it looks that the main direction of
spread is from SE to NE.
Although the spatial resolution is quite low, it seems
that the main corridors of expansion are the river
valleys of Danube, Drava, and the Sava.
Fig. 5. Direction and speed of the spread of the Neolithic, based on the estimated age of the beginning of
the Neolithic (Fig. 4) visualized as a vector field.

Dimitrij Mleku/ Vrhovnik
276
Expansion along the Mur and Drava Rivers slows
down until an over 500-year long standstill of the
stationary border when settlements reach the foot-
hills of the Alps. However, after the border was bre-
ached it expands very rapidly into the hilly fringe of
SE Alps. This expansion happens at roughly the same
time as the expansion of the Neolithic along the Sava
River into the SE Alps, and might be a part of the
same process.
Based on the analysed data we can identify at least
two processes behind the pattern. The first is the
establishment of enclaves which happened in the
first 500 years and then spread relatively slowly
from there. In some areas, especially on the western
fringe of the Carpathian Basin, we can observe the
formation of a stationary border for almost 500 to
1000 years, followed by quick spread into the Alpine
foothills.
The general SPD curve constructed from the SPD
curves for each grid cell thus reflects the settlement
and demographic dynamics in the study area (Fig.
6). The curve shows a rapid increase from 8000 to
7500 cal BP with another push after 7000 cal BP
when the curve reaches a peak at around 6300 cal
BP. After 6500 and especially after 6000 there is a
pronounced dip in the curve, with small increase
and local peak just before 5500 cal BP followed by
a slow decrease until the end of time frame. Main
peak and dips are statistically significant.
This curve might overrepresent the earliest dates due
to the research bias, as re-
search strategy is usually fo-
cused mainly on the oldest
and the earliest dates and con-
texts. Nevertheless, the SPD
curve reflects some trends, the
most interesting being the ra-
pid decline after 6300 cal BP.
The fast rise and peak are
consistent with the Neolithic
demographic transition model
(Bocquet-Appel 2011), which
postulates fast growth at the
border, followed by a drop a
few centuries layer. The same
pattern is found in other re-
gions all over Europe (Shen-
nan et al. 2013).
More interesting are regional
differences in the process.
The curve for the SE Alps rises rapidly just after
7000 cal BP. Most of the growth in the study area
between 7000 and 6300 cal BP can be attributed to
the expansion and growth in the SE Alps area in this
period. There is also proportionally less decline than
elsewhere after 6300 cal BP, where especially after
6000 cal BP the SE Alps contribute most of the va-
lue to the overall curve. Those differences from the
study area are statistically significant
Another estimate shows the number of occupied
grid cells at 100-year intervals (Fig. 6). This is a si-
milar although simplified estimate of the extent of
Neolithic settlement in the study area. The curve
shows a steady increase in the number of occupied
grid cells starts around 8000 cal BP and reaches a
peak around 6500 cal BP. After 6300 cal BP, begin-
ning of the Copper Age in the study area, the curve
experiences fast decline with some fluctuations
after 6000 cal BP. Overall it seems that the extent of
the Copper Age settlement systems is approximately
half that of the maximum extent of Neolithic settle-
ment around 6500 cal BP in the study area.
In contrast to the study area, the SE Alps experien-
ces different dynamics. Fast expansion into the SE
Alps starts just after 7000 cal BP and reaches a peak
at around 6500 cal BP, like the curve for the over-
all study area. It looks as if the main contribution
to the overall extent of settlement after 7000 cal BP
can be attributed to the expansion into the SE Alps.
When, after 6500 cal BP the curve experiences a no-
table and rapid drop, the reduction in the SE Alps is
Fig. 6. SPD curve based on the Neolithic dates from the study area (dark)
and a subset of dates from the SE Alps (light) and 200-year rolling means
(oranges).

Neolithic and Copper Age settlement dynamics in the Western Carpathian Basin and Eastern Alps
277
not as significant as in the
overall study area. After the
drop stabilises at around 6000
cal BP, the SE Alpine area con-
tributes a large number of grid
cells to the overall study area,
as up to half of the grid cells in
the study come from the area
of the SE Alps.
This might be exacerbated by
the research bias, as this is the
period of the appearance of
Neolithic in Slovenia, where a
lot of dating effort was fo-
cused. The Late Neolithic has
received much less focus else-
where. However, even consi-
dering this research bias, the
area of the SE Alps experien-
ces different dynamics than the rest of the study
area.
The spatial pattern of this process is clearly shown
in a sequence of settled grid cell maps at 500-year
intervals (Fig. 7). Neolithic settlement starts as sparse
isolated grid cells in Slavonia, along the Danube,
Bosnia and at the eastern edge of the Alps. Between
7000 and 6500 cal BP we can observe a process of
expansion around already established grid cells. The
first clusters of grid cells are formed in Slavonia, in
the area between the Sava and Drava and at the east-
ern edge of the Alps, between the rivers Balaton and
Mur.
The time slice between 7000 and 6500 cal BP is
marked by expansion into the SE Alps, with a fur-
ther process of expansion in other areas. This is also
the period where we can observe the abandonment
of the first grid cells. This process continues after
6000 cal BP, with continuous expansion into the SE
Alps and extensive abandonment of grid cells in the
lower reaches of the Sava, Drava and Danube. The
general decline in the settled grid cell density conti-
nues toward 5000 cal BP.
Discussion
Alasdair Whittle in his discussion of long-term and
large-scale processes suggests three interweaved
processes behind the formation of European Neoli-
thic settlement patterns. There is the first phase of
primary agricultural colonization, followed by the
second phase of internal infilling and continued ex-
ternal expansion, followed in turn by the final phase
of ‘packing’ (Whittle 1987.34).
The picture painted here is a bit more intricate. Com-
plex spatio-temporal processes can be decomposed
into three basic processes, spread, then movement,
and aggregation or segregation (O’Sullivan, Perry
2013). Although the present study observes these
processes at a very low spatial and temporal resolu-
tion it is still possible to appreciate the complexity
and identify the main components. The spread pro-
cesses include growth, diffusion and percolation,
and they all refer to the expansion of a common
boundary or fronts of a phenomenon, such as the
expansion of a gas into a vacuum, forest fire or
spread of animal species in a new environment
(O’Sullivan, Perry 2013.133–168). Movement refers
to the spread of individual entities, and can be seen
as the secession of shifts which relocate an entity
(single molecule of gas, fire, or individual animal or
human) from one location to another. These walks
can be random (as in case of isolated gas molecules)
or, more often, influenced by the environment or
other entities (O’Sullivan, Perry 2013.97–131). Ag-
gregation and segregation are two facets of the same
process, driven by a tendency of similar elements to
group together in space or dissimilar elements to se-
parate in space (O’Sullivan, Perry 2013.57–95).
The process of the formation of Neolithic settlement
systems in the study area was not a swift, uniform
transition that established stable Neolithic settlement
system in the course of a few centuries. It was not
an even diffusion of Neolithic settlements, filling the
Fig. 7. Number of occupied grid cells by century from the study area
(dark) and SE Alps (light).

Dimitrij Mleku/ Vrhovnik
278
landscape of the study area. Instead, as already
argued by Marek Zvelebil (2001.1), it was a complex
interaction of several processes with their own dyna-
mics and time depth, which included both movement
and contact, combining in a very complex and long
historical trajectory, embedded in the existing social
and historical conditions. In this sense, the social
context of the agricultural transition in the study
area had structure and agency. The formation of Neo-
lithic settlement systems in the study area lasted se-
veral millennia and included lives over tens of gene-
rations. The process was probably experienced more
as continuity than one of disjuncture and change
(Hofmann, Gleser 2019).
The spread of Neolithic settlements was part of a
wider phenomenon, the Neolithisation of Europe.
However, the spread of specific material assemblage
associated with Neolithic was not uniform. Instead,
the edge of the process observed in a study area has
a ragged, swirly, pixelated border (Robb 2014.33).
The Neolithic material assemblage percolated along
different paths, with the establishment of pioneer
communities at the front. The discontinuous spread
and establishment of enclaves point to the key role
of movement and personal and group mobility in
the process. Spread involved the movement of small
groups or even individuals (see Zvelebil 2001.2A).
These movements usually do not have a single ori-
gin point, creating a perplexing pattern of “migra-
tions without a homeland” (Robb 2014.658–659).
What drives such groups onward is poorly known
(but see Hofmann 2016). Movement seems to fol-
low the natural corridors in a landscape, especially
river valleys. Here, the Danube, Sava, Drava, and
Mur seem to be main lines along which Neolithic
communities moved forward. Rapid enclave move-
ments tend to halt when they encounter either dif-
ferent environments (Alpine foothills) or dense for-
ager settlements (possibly Alpine foothills and the
Balaton Area; see Bánffy 2006.130–136). The result-
ing frontiers lasted a long time, and may include the
movement of individuals, families or small groups of
people across the border.
The third process is the aggregation or segregation
of Neolithic settlements, which created a patchy land-
scape of a structured, organized and consistent Neo-
lithic social world surrounded by untamed, wild
landscape. There are two general factors behind the
process, one is the environment while the other is
demographic and social. Initial enclave colonization
targeted specific environmental niches. After the for-
mation of initial or core settlements, a gradual pro-
Fig. 8. The extent of Neolithic and Copper Age set-
tlement systems in the study area in 500-year time
slices. Dark grid cells indicate occupation while
light indicates abandoned grid cells.
cess of aggregation continues as a slow infill of the
landscape around initial settlements, creating Side-
lungskammern of Neolithic settlements. The move-
ment of people and things between settlements and
patches connected them in the social landscape,

Neolithic and Copper Age settlement dynamics in the Western Carpathian Basin and Eastern Alps
279
prone to shifting patterns of interaction and integra-
tion. There are several centripetal and centrifugal
forces leading to either integration or fissioning of
groups. These changes are reflected in a settlement
pattern and may be a driving force behind other
processes, such as movement and spread.
The dynamics of the Neolithic spread and the mo-
vement of individuals and small groups were deter-
mined to an extent by social processes in already
settled regions. This is especially pronounced in the
case of the secondary expansion of Neolithic settle-
ment systems into the SE Alps.
Around 6700 cal BP there is a pronounced change
in the settlement systems in the Balkans with the ap-
pearance of stratified tell sites, large nucleated settle-
ments and large cemetery grounds (Hofmann, Gle-
ser 2019.24–29). In the study area, this process is
very well documented with the dynamics in Alsón-
yék-Bátaszék in south-west Hungary, where the set-
tlement that formed in the Star≠evo phase experien-
ces sudden large-scale expansion around 6800 cal
BP with the erection of settlement with 122 houses
and cemetery with around 2300 graves. It is just one
of several substantial Lengyel culture sites in the
neighbourhood which include both cemeteries and
settlements. Alsónyék-Bátaszék became a large aggre-
gation of people, with a population that suddenly in-
creased almost fifty-fold. This aggregation stayed in
place for only one generation, followed by an equal-
ly fast dispersal (Osztás et al. 2012; 2013; 2016).
This process coincides with the Neolithic expansion
into the SE Alps, especially the area of modern Slo-
venia, which started after 7000 BP. It is marked by
a relatively fast expansion along the Sava River, es-
tablishment of settlements in the river valleys and
plains. This is followed by the expansion along Dra-
va and Mur river valleys into the Alps. This process
of expansion into the Alpine river valleys continues
for almost 500 years. The same pattern of breach of
long-standing frontiers is also visible elsewhere in
the study area.
A resurgence of Mesolithic ancestry in the Late Neo-
lithic has already been noted all over Europe, al-
though in some places this process was limited. Ge-
netic signatures associated with European hunter-
gatherers (mitochondrial U-haplotypes) reappear in
central Europe during the 7
th
and 6
th
millennium BP
(Haak et al. 2015; Bollongino 2013; Fu et al. 2016;
Lipson et al. 2017). The possible origins of this re-
surgence are currently not yet clear, however, it
might be associated with the expansion of Neolithic
communities into previously marginal areas, new
contacts with Mesolithic hunter-gatherer communi-
ties that could have been accompanied by increased
genetic exchange with more central areas.
On the other hand, it seems that by around 6600
cal BP, tell and nucleated sites which previously cha-
racterized most of the Carpathian Basin were sud-
denly abandoned. The transition from Late Neolithic
to Copper Age is marked by a change from nucleat-
ed to a dispersed settlement pattern. In the whole
Carpathian Basin previously nucleated sites were
replaced by smaller, flat settlements, largely charac-
terized by shallow single-layer occupation deposits,
along with a change from intramural burials to large
extramural cemeteries (Parkinson 2002.391–394;
Bori≤ 2015.157). This seems to be a wider process
that occurred almost simultaneously over the study
area.
This process of segregation can be detected all over
the study area. Initial Neolithic settlements in Lahi-
nja river valley and Krupsko polje in Bela Krajina,
Slovenia targeted fertile soils and were established
soon after 7000 cal BP. In the mid-6
th
millennium
BP there is an expansion from core areas into the
drier Karst hinterland, with new sites that were oc-
cupied less intensively and for shorter periods and
the formation of enclosed upland sites (Budja 1995;
Mason 1995). However, initial settlements, such as
Moverna vas, were not abandoned.
The pattern of smaller dispersed settlements in the
Early Copper Age, despite possible research biases
regarding the visibility of small dispersed sites, could
suggest a drop in population levels, even if the num-
ber of individual sites increases. However, the demo-
graphic decline did not affect all areas equally, but is
much more pronounced in core areas of Slavonia,
while newly settled areas peripheral areas seem to
experience much less severe declines.
Attempts to explain these discontinuities by simple
boom-boost cycles of population dynamics (caused
by climate change which affected subsistence practi-
ces, ultimately lowering reproductive success; Shen-
nan 2009; 2013; 2018) seem overly simplistic and
theoretically impoverished. If the Neolithic was a hi-
storical process (in contrast to an evolutionary epi-
sode) the explanations must take into account the
nature of social interaction and the way it is stabi-
lized by the use of durable material resources and
symbols. Material resources fix the way individuals

Dimitrij Mleku/ Vrhovnik
280
interact, behave and move, and dictate new skills, ha-
bits and actions. They impose new physical techni-
ques, training and disciplines, making individuals be-
come productive members of a specific assemblage.
The spatial segregation processes that mark the tran-
sition from Late Neolithic to Copper are obviously
connected to increased residential mobility, as re-
flected in the dispersed settlement pattern and occu-
pation of new areas with newly founded settlements.
It is difficult to identify the mechanisms behind the
centrifugal forces which caused the segregation of
previously dependent and closely-knit communities
at a larger regional level (Bori≤ 2015.189–193).
It might be the result of a restructuring of a Neolithic
assemblage which becomes destabilized with the
introduction of new components such as copper me-
tallurgy and the growing importance of domestic cat-
tle and pastoral economy (Orton 2012). After all,
assemblages are precarious composite entities that
just about hold together because all their parts hap-
pen to be in the right places, doing the right things
to achieve this. Adding and swapping new elements
in an assemblage can cause non-linear transitions to
occur (DeLanda 2006.10–11).
Conclusion
The paper approached large spatio-temporal trends in
the formation and change of regional settlement sys-
tems in the Western part of the Carpathian Basin area
around the Eastern Alps in the Neolithic and Copper
Age. We were interested in the spatial processes of
spread, movement, aggregation and segregation in
the time frame between 8500 and 5000 cal BP.
The distribution of Neolithic and Copper Age sites in
the study area is clustered and patchy. The first Neo-
lithic thus appears as isolated islands or enclaves of
Neolithic settlements which then slowly expand to
fill neighbouring regions.
The core area for the spread of the Neolithic is that
between the Sava and Drava. From the origin in Sla-
vonia, the Neolithic expands in two prongs, one
along the Danube and the other along the Drava,
Mur and eastern foothills of the Alps. This expansion
is in the form of several enclaves with much earlier
appearance of the Neolithic than surrounding areas,
such as ones along the Danube, Vienna basin and
Western Transdanubia.
There are also some backwater areas with much later
Neolithic settlement. The most prominent being the
area of the Eastern Alps. We identified the existence
of stationary borders, most prominently on the edges
of Carpathian basin and the Alps, along the lower
course of the Mur River, where the Neolithic expan-
sion toward the west halted for almost 500 years.
However, once the border was breached it expands
very rapidly into the hilly fringe of SE Alps. Fast ex-
pansion into SE Alps starts just after 7000 cal BP and
reaches a peak at around 6500 cal BP, which is also
the period of the maximum extent of Neolithic settle-
ment systems in the study area.
After 6300 cal BP study area experiences a signifi-
cant reduction in the extent of settlement systems,
associated with the Late Neolithic to Copper Age
transition. This was a significant decrease in the ex-
tent of settlements system, but not all areas were af-
fected to the same extent.
Appendix is available at
http://dx.doi.org/10.4312/dp.46.16

Neolithic and Copper Age settlement dynamics in the Western Carpathian Basin and Eastern Alps
281
Ammerman A. J., Cavali-Sforza L. L. 1984. The Neolithic
transition and the genetics of population in Europe.
Princeton University Press. Princeton (NJ).
Bánffy E. 2004. The 6
th
millennium BC boundary in
western Transdanubia and its role in the central Euro-
pean transition (The Szentgyörgyvölgy-Pityerdomb set-
tlement). Instituti Archaeologici Academiae Scientiarum
Hungaricae. Budapest.
2006. Eastern, Central and Western Hungary – varia-
tions of Neolithisation models. Documenta Praehisto-
rica 33: 125–142. https://doi.org/10.4312/dp.33.13
Bevan A., Crema E. R. 2018. rcarbon v1.2.0: Methods for
calibrating and analysing radiocarbon dates.
https://CRAN.R-project.org/package=rcarbon
Bivand S., Roger, Pebesma E., and Gomez-Rubio V. 2013.
Applied spatial data analysis with R. Springer. New York.
Blagojevi≤ T., Por≠i≤ M., Penezi≤ K., and Stevanovi≤ S.
2017. Early Neolithic population dynamics in the Eastern
Balkans and the Great Hungarian Plain. Documenta Pra-
ehistorica 44: 18–33. https://doi.org/10.4312/dp.44.2
Bocquet-Appel J.-P. 2011. The Agricultural Demographic
Transition During and After the Agriculture Inventions.
Current Anthropology 52(S4): 497–510.
https://doi.org/10.1086/659243
Bocquet-Appel J.-P., Naji S., Vander Linden M., and Koz-
łowski J. K. 2009. Detection of diffusion and contact zones
of early farming in Europe from the space-time distribu-
tion of 
14
C dates. Journal of Archaological Science 36 (3):
807–820. https://doi.org/10.1016/j.jas.2008.11.004
Bollongino R., Nehlich O., Richards M. P., Orschiedt J.,
Thomas M. G., Sell C., Fajko∏ová Z., Powell A., and Burger J.
2013. 2000 years of parallel societies in Stone Age central
Europe. Science 342: 479–481.
https://doi.org/10.1126/science
Bori≤ D. 2015. The End of the Vin≠a World: Modelling the
Neolithic to Copper Age Transition and the Notion of Ar-
chaeological Culture. In S. Hansen, P. Raczky, A. Anders,
and A. Reingruber (eds.), Neolithic and Copper Age be-
tween the Carpathians and the Aegean Sea. Chronolo-
gies and Technologies from 6
th
to 4
th
millennium BCE.
Habelt Verlag. Bonn: 177–237.
Brami M., Zanotti A. 2015. Modelling the initial expansion
of the Neolithic out of Anatolia. Documenta Praehistori-
ca 42: 103–116. https://doi.org/10.4312\dp.42.6
Budja M. 1995. Neolithic and Eneolithic settlement pat-
terns in the Bela krajina region of Slovenia. In A. Aspes,
(eds.), Settlement patterns between the Alps and the
Black sea from the 5
th
to 2
nd
millennium B.C. Museo
Civico di Storia Naturale di Verona. Verona: 119–127.
Chilés J.-P., Delfiner P. 2012. Geostatistics: Modeling Spa-
tial Uncertainty. Wiley. New York.
Crema E. R., Habu J., Kobayashi K., and Madella M. 2016.
Summed Probability Distribution of 
14
C Dates Suggests
Regional Divergences in the Population Dynamics of the Jo-
mon Period in Eastern Japan. PLoS ONE 11(4): e0154809.
https://doi.org/10.1371/journal.pone.0154809
DeLanda M. 2006. A new philosophy of science. Conti-
nuum. London.
Fort J. 2015. Demic and cultural diffusion propagated the
Neolithic transition across different regions of Europe.
Interface 12(106): 1–8.
https://doi.org/10.1098/rsif.2015.0166
Fu Q. and 63 co-authors 2016. The genetic history of Ice
Age Europe. Nature 534(7606): 200–205.
https://doi.org/10.1038/nature17993
Gkiasta M., Russell T., Shennan S., and Steele J. 2004. Neo-
lithic transition in Europe: The Radiocarbon record revis-
ited. Antiquity 77(295): 45–62.
https://doi.org/10.1017/S0003598X00061330
Haak W. and 38 co-authors. 2015. Massive migration from
the steppe was a source for Indo-European languages in
Europe. Nature 522: 207–211.
https://doi.org/10.1038/nature14317
Hofmann D. 2016. Keep on walking: The role of migra-
tion in Linearbandkeramik life. Documenta Praehisto-
rica 43: 235–251. https://doi.org/10.4312/dp.43.11
Hofmann D., Gleser R. 2019. The fifth millennium: the
emergence of cultural diversity in 13 central European
prehistory. In R. Gleser, D. Hofmann, (eds.), Contacts,
Boundaries & Innovation. Exploring developed Neoli-
thic societies in central Europe and beyond. Sidestone.
Leiden: 13–42.
Kowalewski S. A. 2008. Regional Settlement Pattern Stu-
dies. Journal of Archaeological Research 16(3): 225–285.
https://doi.org/10.1007/s10814-008-9020-8
Krznari≤ πkrivanko M. 2011. Radiokarbonski datumi uzo-
raka sa Sopota. In M. Dizdar, (eds.), Panonski prapovijes-
ni osviti – Zbornik radova posve≤en Korneliji Minichrei-
References

Dimitrij Mleku/ Vrhovnik
282
ter uz 65. obljetnicu ∫ivota. Institut za arheologiju. Zag-
reb: 209–225.
Lipson M. and 56 co-authors. 2017. Parallel palaeogeno-
mic transects reveal complex genetic history of early Euro-
pean farmers. Nature 551(7680): 368–372.
https://doi.org/10.1038/nature24476
Mason P. 1995. Neolitska in eneolitska poselitev v Beli
krajini: naselje v Gradcu in izraba prostora v ≠asu od 5.
do 3. tiso≠letja BC. Poro≠ilo o raziskovanju paleolita,
neolita in eneolita v Sloveniji 22: 183–201.
O’Sullivan D., Perry G. L. W. 2013. Spatial simulation:
Exploring Pattern and Process. Willey. New York.
Orton D. 2012. Herding, Settlement, and Chronology in
the Balkan Neolithic. European Journal of Archaeology
15(1): 5–40.
https://doi.org/10.1179/1461957112Y.0000000003
Osztás A., Bánffy E., Zalai-Gáal I., Oross K., Marton T., and
Somogyi K. 2013. Alsonyek-Bataszek: introduction to a
major Neolithic settlement complex in south-east Trans-
danubia, Hungary. Bericht der Römisch-Germanischen
Kommission 94: 7–21.
https://doi.org/10.11588/berrgk.1938.0.37149
Osztás A., Zalai-Gaál I., and Bánffy E. 2012. Alsónyék-Bá-
taszék: a new chapter in the research of Lengyel culture.
Documenta Praehistorica 39: 377–396.
https://doi.org/10.4312/dp.39.27
Osztás A. and 13 co-authors. 2016. Coalescent communi-
ty at Alsónyék: the timings and duration of Lengyel buri-
als and settlement. Bericht der Römisch-Germanischen
Kommission 94: 179–282.
Parkinson W. A. 2002. Integration, interaction, and tribal
‘cycling’: The transition to the Copper Age on the Great
Hungarian Plain. In W. A. Parkinson, (eds.), The archaeo-
logy of tribal societies. International Monographs in Pre-
history. Ann Arbor: 391–438.
Por≠i≤ M., Blagojevi≤ T., and Stefanovi≤ S. 2016. Demogra-
phy of the early Neolithic population in central Balkans:
population dynamics reconstruction using summed radio-
carbon probability distributions. PLoS One 11(8): e016
0832. https://doi.org/10.1371/journal.pone.0160832
R Core Team. 2018. R: A language and environment for
statistical computing. https://www.R-project.org/.
Reimer P. J. and 29 co-authors. 2013. IntCal13 and Ma-
rine13 Radiocarbon Age Calibration Curves 0–50,000
Years cal BP. Radiocarbon 55(4): 1869–1887.
https://doi.org/10.2458/azu_js_rc.55.16947
Robb J. 2013. Material Culture, Landscapes of Action, and
Emergent Causation: A New Model for the Origins of the
European Neolithic. Current Anthropology 54(6): 657–
683. https://doi.org/10.1086/673859
2014. The Future Neolithic: A New Research Agenda. In
A. Whittle, P. Bickle, (eds.), Early Farmers: The View
from Archaeology and Science. Oxford University
Press. Oxford: 21–38.
Shennan S. 2009. Evolutionary Demography and the Po-
pulation History of the European Early Neolithic. Human
Biology 81(2–3): 339–355.
https://doi.org/10.3378/027.081.0312
2013. Demographic continuities and discontinuities in
Neolithic Europe: evidence, methods and implications.
Journal of Archaeological Method and Theory 20(2):
300–311. https://doi.org/10.1007/s10816-012-9154-3
2018. The first farmers of Europe: An evolutionary
perspective. Cambridge University Press. Cambridge.
Shennan S., Doweny S. S., Timpson A., Edinborough K.,
Colledge S., Kerig T., Manning K., and Thomas M. G. 2013.
Regional population collapse followed initial agriculture
booms in Mid-Holocene Europe. Nature Communications
4: 2486. https://doi.org/10.1038/ncomms3486
Stadler P., Kotova N. 2010. Early Neolithic settlement from
Brunn Wolfholz in lower Austria and the problem of the
origin of (western) LBK. In J. K. Kozłowski, P. Raczky
(eds.), Neolithization of the Carpathian basin: northern-
most distribution of the Star≠evo/Körös culture. Polish
Academy of Arts and Sciences. Kraków–Budapest: 325–
348.
Timpson A., Colledge S., Crema E., Edinborough K., Kerig
T., Manning K., Thomas M. G., and Shennan S. 2014. Re-
constructing regional population fluctuations in the Euro-
pean Neolithic using radiocarbon dates: a new case-study
using an improved method. Journal of Archaeological
Science 52: 549–557. 
https://doi.org/10.1016/j.jas.2014.08.011
Whittle A. 1987. Neolithic Settlement Patterns in Tempe-
rate Europe: Progress and Problems. Journal of World
Prehistory 1(1): 5–52.
Williams A. N. 2012. The use of summed radiocarbon pro-
bability distributions in archaeology: A review of methods.
Journal of Archaeological Science 39(3): 578–589.
https://doi.org/10.1016/j.jas.2011.07.014
Zvelebil M. 2001. The agricultural transition and the ori-
gins of Neolithic society in Europe. Documenta Praehi-
storica 28: 2–26. https://doi.org/10.4312/dp.28.1