Content uploaded by Lasse Lukas Herskind
Author content
All content in this area was uploaded by Lasse Lukas Herskind on Jul 21, 2023
Content may be subject to copyright.
Mesolithic Semiotics?
A Quantitative Reassessment of the Portable Art
of Mesolithic South Scandinavia
MA dissertation
Lasse Lukas Platz Herskind
201809436
Master’s Degree Programme in Archaeology
Aarhus University
ECTS: 30
F23, 2023
Supervisor: Professor Felix Riede
Title:
Mesolithic Semiotics?
- A Quantitative Reassessment of the Portable Art of Mesolithic South Scandinavia
MA dissertation produced by:
Lasse Lukas Platz Herskind
Student ID: 201809436
At:
Spring Exam 2023
Prehistoric Archaeology
Department of Archaeology, School of Culture & Society
Aarhus University
Supervisor:
Professor Felix Riede
Department of Archaeology, School of Culture & Society
Aarhus University
Characters:
Dissertation: 176.359
Summary: 6.519
TABLE OF CONTENTS
Page
List of figures................................................................................................
i
List of tables..................................................................................................
ii
List of Supplementary Material..............................................................
ii
Summary.......................................................................................................
iii
1. Introduction.............................................................................................
1
1.1 Introducing the knowledge gap............................................................
1
1.2 Research aim.........................................................................................
4
1.3 Outline of the dissertation.....................................................................
4
2. Stand der Forschung..............................................................................
7
2.1 Overview of the South Scandinavian Mesolithic..................................
7
2.1.1 Chronological framework...............................................................
8
2.1.2 Regional variations – South Scandinavia as a ‘contextual area’...
11
2.2 The portable art of the South Scandinavian Mesolithic.......................
14
2.3 Interpretative approaches to prehistoric art.......................................
21
2.3.1 Interpretations of European Upper Palaeolithic markings.............
21
2.3.2 Interpretations of South Scandinavian Mesolithic portable art......
24
3. Theory........................................................................................................
28
3.1 The resurgence of semiotic hypotheses.................................................
28
3.2 A semiotic hypothesis applied to the portable art of Mesolithic South
...........Scandinavia..........................................................................................
33
3.2.1 Theoretical assumptions....................................................................
36
4. Materials & Methods..............................................................................
38
4.1 Utilising Tomasz Płonka’s catalogue......................................................
38
4.1.1 Temporal information.......................................................................
39
4.1.2 Spatial information...........................................................................
40
4.1.3 Coding the motifs...............................................................................
41
4.2 A methodology for discerning semiotic structure..................................
42
4.2.1 K-skip-n-grams – solving the issue of nonlinear corpus data..........
43
4.2.2 O/E ratio – a custom measure of statistical significance.................
45
5. Analysis & Results..................................................................................
48
5.1 Descriptive Statistics...............................................................................
48
5.1.1 Representativity.................................................................................
48
5.1.2 Descriptive statistics for the individual motifs..................................
57
5.2 Testing for semiotic structure.................................................................
67
6. Discussion.................................................................................................
74
6.1 Limitations of this dissertation...............................................................
74
6.2 Take-home messages from the descriptive statistics.............................
74
6.3 A (near) refutation of the semiotic hypotheses.......................................
77
6.3.1 The strong semiotic hypothesis..........................................................
78
6.3.2 The weak semiotic hypothesis...........................................................
79
6.4 Alternative interpretations: Maglemose drift, Ertebølle politics?........
84
6.5 Future studies..........................................................................................
87
8. Conclusions..............................................................................................
89
References.....................................................................................................
92
i
LIST OF FIGURES
Page
1.1
Examples of ‘exquisitely’ ornamented portable art objects............................
3
1.2
Flowchart of the dissertation..........................................................................
6
2.1
South Scandinavian Mesolithic date ranges and chronozones......................
11
2.2
Newell et al.’s proposed Maglemosian social territories................................
12
2.3
Regional variability of different Ertebølle artefact types..............................
13
2.4
‘Fringe’ motif on an amber pendant from Limfjorden....................................
16
2.5
Zigzag motif on an antler artefact from Holmegård IV.................................
17
2.6
‘Sheaf of grain’ motif on an antler axe from Eskelund....................................
17
2.7
Examples of South Scandinavian Mesolithic portable art.............................
20
2.8
Examples of portable art objects interpreted as ‘rituo-magic’.......................
25
2.9
The conflicting interpretations of Płonka and Vang Petersen........................
26
2.10
Nash’s interpretation of an Ertebølle ‘Kula ring’............................................
27
3.1
The antler pick from Uggerløse, argued by Marshack as notational.............
29
3.2
Research historical overview of interpretative approaches..........................
32
3.3
Comparison between Upper Palaeolithic, Mesolithic, and Bronze Age ‘art’..
34
4.1
Map of this dissertation’s regional analytical units.......................................
41
4.2
Example of motif attribution and the skipgrams generated..........................
45
5.1
Chronological distribution of ornamented artefacts.....................................
49
5.2
Spatial distribution of portable art attributed to the Maglemose period......
50
5.3
Spatial distribution of portable art attributed to the Kongemose period......
50
5.4
Spatial distribution of portable art attributed to the Ertebølle period..........
51
5.5
Overall regional distribution of ornamented artefacts..................................
51
5.6
The frequencies of different materials bearing ornamentation.....................
52
5.7
The frequencies of different artefact types bearing ornamentation..............
52
5.8
‘Chronology’ frequencies across the other four variables..............................
53
5.9
‘Region’ frequencies across the other four variables......................................
54
5.10
‘Material’ frequencies across the other four variables...................................
55
5.11
‘Artefact type’ frequencies across the other four variables.............................
56
5.12
Number of motifs per artefact for each culture complex................................
58
5.13
Overall motif frequencies of motifs on portable art objects............................
59
5.14
Bivariate representation of motifs across the five categorical variables......
61
5.15
Grid of motif attributions for the three cultures, ordered chronologically....
63
5.16
Spatio-temporal plot of the motifs with strongest Maglemose attribution...
64
5.17
Spatio-temporal plot of the motifs with strongest Kongemose attribution...
65
5.18
Spatio-temporal plot of the motifs with strongest Ertebølle attribution.......
66
5.19
Motif co-occurrences with the highest frequencies.........................................
68
5.20
Triple motif co-occurrences with the highest frequencies..............................
69
5.21
The most significant motif co-occurrences by O/E ratio................................
72
5.22
Heatmap of motif co-occurrences...................................................................
73
6.1
Heatmap by O/E ratio, including culture-specific motif combinations.........
80
6.2
The spatial distribution of the ‘Ertebølle cluster’............................................
81
ii
LIST OF TABLES
Page
1
Translation of Płonka’s chronological divisions.................................................
39
2
Translation of Płonka’s categories of contextual information quality...............
40
3
Translation of Płonka’s regional divisions.........................................................
41
4
The k-skip-n-gram levels performed in the analysis..........................................
45
5
Number of ornamented artefacts per site...........................................................
49
6
Frequencies of individual motifs.........................................................................
57
7
Number of motifs per artefact for each culture complex....................................
58
8
Frequencies of motif co-occurrences...................................................................
67
9
Frequencies of triple motif co-occurrences.........................................................
67
10
Frequencies of quadruple motif co-occurrences.................................................
67
LIST OF SUPPLEMENTARY MATERIAL
1
Data
2
R code for descriptive statistics
3
R code for k-skip-n-gram analysis
4
Data table for two-motif co-occurrences
5
Data table for three-motif co-occurrences
6
Data table for four-motif co-occurrences
7
Tomasz Płonka’s motif classification
iii
Summary
This dissertation contributes to understanding the function of portable art from the Meso-
lithic of South Scandinavia. Through a methodologically novel approach, it is investigated
whether the markings of this period functioned as a semiotic system. After refuting these ‘se-
miotic hypotheses’, alternative explanations are preliminarily suggested, based on the quan-
titative and data-driven analyses.
The South Scandinavian Mesolithic is a slice of human prehistory in which postglacial cli-
matic developments allowed for a diversification of its hunter-gatherer societies. New tech-
nologies and subsistence strategies emerged in tandem with climatic developments, which in
this area were particularly characterised by an inundation of vast areas of landmass. Follow-
ing traditional archaeological taxonomy, the South Scandinavian Mesolithic is divided into
three culture complexes: Maglemose (ca. 11000-8400 BP), Kongemose (ca. 8400-7400 BP),
and Ertebølle (ca. 7400-5950 BP). These three consecutive culture periods are distinct from
each other in several ways but do also share many traits. One such shared trait is the activity
of engraving patterns and motifs on objects of everyday use, such as antler axes, bone dag-
gers, amber pendants, etc. This ‘portable art’ has been known to archaeology since the 19th
century and has been subject to much research, yet the underlying purpose behind this orna-
mentation remains difficult to demonstrate unequivocally.
Approaches to interpreting such prehistoric art are diverse, particularly concerning Euro-
pean Upper Palaeolithic markings. These can be crudely divided into different ‘schools of in-
terpretation’, namely ‘decorative’, ‘rituo-magic’, ‘socio-political’, ‘weak-’ and ‘strong semi-
otic’. The semiotic hypotheses (following Paul Bouissac’s proposal of ‘weak’ and ‘strong’ sub-
divisions) imply that the markings were encoded with information in the form of notational
or proto-writing systems. This avenue of interpretation gained popularity in the 1970’s and
80’s, owing much to the work of Alexander Marshack; a strong proponent for Upper Palaeo-
lithic markings being lunar calendars, his claims were however met with scholarly scepti-
cism. Recent years have seen a marked resurgence of publications reminiscent of Marshack’s
work, arguing for semiotic structure in Upper Palaeolithic European markings. Yet, this re-
surgence of semiotic hypotheses does not critically engage with the unsettled debates of the
70’s and 80’s. The present dissertation positions itself within these research historical devel-
opments, recognising that the semiotic hypotheses have not been applied in the context of
the South Scandinavian Mesolithic. It is argued that this absence is due to a stasis in South
Scandinavian Mesolithic research, as well as a modern preconception of this corpus of porta-
ble art being poor in quality – but in fact, this period and region is argued to have promising
iv
potential for the emergence of such semiotic functions. The central purpose of this disserta-
tion is thus to test the semiotic hypotheses for the portable art of Mesolithic South Scandina-
via. While remaining critical of existing claims for prehistoric semiosis, Paul Bouissac’s call
for not disregarding these hypotheses a priori is heeded.
The difficulty of assessing the presence of semiotic structure in prehistoric art pertains
mostly to designing an appropriate methodology. This dissertation takes a methodologically
novel approach, adopting ‘k-skip-n-gram’ (‘skipgram’) models from the field of computa-
tional linguistics in order to establish the mutual relationship between individual motifs. By
customising the skipgram method, all co-occurrences of two, three, and four motifs, respec-
tively, are generated. The statistical significance of each co-occurrence is evaluated by com-
paring their observed (O) and expected (E) frequencies. This method can effectively identify
whether, and which, motifs were recurrently engraved on the same artefact. Prior to this
testing of the semiotic hypotheses, extensive descriptive statistics are presented concerning
the portable art objects and their associated motifs. The purpose of these descriptive statis-
tics is to identify biases as well as patterns across the variables of chronology, spatiality,
quality of contextual information, material, and artefact type. In combination, these two ana-
lytical parts qualify a subsequent discussion of the semiotic hypotheses’ likelihood. These
analyses are rendered possible by the contribution of Tomasz Płonka, who in his monograph
‘The Portable Art of Mesolithic Europe’ provides a comprehensive catalogue of ornamented
objects from the European Mesolithic along with a motif classification scheme. In utilising
Płonka’s work as the empirical foundation, this dissertation provides a quantitative and data-
driven contribution to a research field which has been predominantly qualitative and empiri-
cally selective.
This dissertation finds that statistically significant motif co-occurrences are uncommon, and
that statistically significant co-occurrences which are limited to a single archaeological cul-
ture are rarer still. Moreover, the portable art objects are characterised by considerable het-
erogeneity across all variables. On these grounds, it is rendered unlikely that the portable art
of Mesolithic South Scandinavia functioned within a semiotic system on any societal scale.
Moving beyond the dissertation’s primary aim, it is preliminarily suggested that the data
suggests random cultural drift, most markedly in the Early Mesolithic. This indicates an ab-
sence of selective pressure and qualifies the argument that Early Mesolithic ornamentation
was governed by the individual preferences of the specific engraver. Additionally, one excep-
tion for this seemingly random structure is demonstrated for a specific cluster of motifs from
the Ertebølle period. This cluster of five interrelated motifs suggests an independent phe-
nomenon, linked to the increasing societal complexity of the Late Mesolithic. Rather than a
v
semiotic system, it is argued that the underlying cause for this Ertebølle anomaly is more
likely to be of a ‘socio-political’ nature, suggesting an inter-territorial diffusion of motifs. Fi-
nally, it is argued that two central issues are 1) the ambiguity of classifications of portable art
motifs, and 2) the coarse chronological resolution of the portable art objects. These two per-
sisting problems must be addressed in order to further qualify this dissertation’s findings.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
1
Chapter 1
Introduction
The omnipresence of prehistoric art has qualified this specific activity as a universal human
phenomenon. Extending back to at least the African Middle Stone Age (Henshilwood et al.
2009), it has been documented in all corners of the world, from Australian rock faces (e.g.,
David & Lourandos 1998: 201–2) to the caves of Patagonia (e.g., Onetto & Podestá 2011).
These engraved, painted, and sculpted attestations are regarded as the most promising prox-
ies for understanding the early developments of human cognition, abstraction, and symbol-
ism (Henshilwood et al. 2002; Dutkiewicz et al. 2018; Tylén et al. 2020). The scholarly
recognition of anthropogenic ornamentation from the deep past happened in the 19th cen-
tury, contemporarily with the recognition of human antiquity itself (Lyell 1863). Since these
early beginnings, this field of study has continuously been occupied with the central question
of why the ornamentations were made.
This dissertation is concerned with the anthropogenic ornamentations from the South Scan-
dinavian Mesolithic (ca. 11.000–6.000 BP). In this slice of time and space, dramatic postgla-
cial landscape changes paved the way for a diverse and abundant natural environment (Bai-
ley 2008; Astrup 2018). Archaeological recoveries allow us an insight into the hunter-gath-
erer societies which inhabited this Early to Middle Holocene world. This archaeological rec-
ord includes artefacts which bear distinct traces of engraved patterns, motifs, and composi-
tions – predominantly geometric, but human and animal representations do also occur. This
phenomenon of ornamentation on portable objects are most commonly referred to as ‘porta-
ble art’. The purpose of this dissertation is to generate a novel contribution to understanding
the purpose and function of these engravings.
1.1 Introducing the knowledge gap
The study of prehistoric art is a vibrant field of research, and also one in which general
agreement is unattainable. The term ‘prehistoric art’ is itself somewhat ambiguous, as it
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
2
arguably presupposes an artistic function (e.g., as discussed by Bowyer 2002; Bednarik
2003; Coleman 2005), and other terms, such as prehistoric ornamentation, palaeoart, visual
culture, and visual expression, are used interchangeably. The arguable failure to devise an
unproblematic, generic term is telling of the fundamental difficulties pertaining to the study
of this specific human behaviour. Throughout this dissertation, the usage of generic concepts
such as ‘prehistoric ornamentation’ and ‘portable art’ is not meant to presuppose any specific
purpose, but only refers to “… all intentional carvings that have no relation to the manufac-
turing and use of the object” (Larsson & Molin 2017a: 87).
Prehistoric ornamentation has been suggested as being artistic expressions (Grosse 1894;
Halverson 1987; De Smedt & De Cruz 2011), magically equipped (Hultén 1939; Lewis-Wil-
liams et al. 1988; Vang Petersen 2021), socio-political markers (Conkey 1978; Nash 1998;
Sauvet et al. 2018), as well as information encoding (Marshack 1972a; b; Taylor 2021; Bacon
et al. 2023). The semiotic avenue of interpretation – hypothesising that information was en-
coded in the ornamentations, governed by a degree of syntactical rules – is among the most
contentious. A pioneering effort was made in the life work of Alexander Marshack (1972a; b,
1985, 1991a; b) who argued strongly for many Upper Palaeolithic incisions being counting
tallies of lunar months. Yet, sharp criticism of his (lack of) scientific approach (Rosenfeld
1971; King 1973; Lewis-Williams et al. 1988; d’Errico 1989; Robinson 1992; Davidson 1993;
Elkins 1996) has meant that these claims remain questionable. Recent years, however, have
seen a resurgence of testing similar semiotic hypotheses for Upper Palaeolithic material
(Sauvet & Wlodarczyk 2008; von Petzinger 2017; Bacon et al. 2023). This study area has po-
tential for sensational discoveries, but the demonstration of prehistoric semiotic systems re-
mains a difficult and ambiguous endeavour.
To the knowledge of the author, no studies have yet considered the semiotic hypotheses in
the context of the South Scandinavian Mesolithic corpus of engravings. Though competent,
most research on this topic is of a qualitative and empirically limited nature (e.g., Andersen
1981; Vang Petersen 2021). Some large-scale systematic analyses have been performed, how-
ever, (Nash 1998; Płonka 2003), but without explicitly engaging with the possibility of semi-
otic governing principles. This absence is undoubtedly due to this specific corpus’ seemingly
sparse and haphazard style of ornamentation. Of course, such observations are heavily influ-
enced by our modern ways of seeing, but it is exactly this type of contemporary qualitative
assessment that has led to the labelling of Mesolithic portable art as “… sparse and poor, of-
ten of a rather coarse character, without much care… As a whole, a hasty, random frippery
without independent worth or bearing” (Brøndsted 1957, translated from Danish by author).
Such strong-worded judgements are made in implicit comparison to the earlier and much
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
3
more famous Magdalenian animal depictions found in Western European caves such as Alta-
mira, Lascaux, and elsewhere (see e.g., Bahn & Vertut 1997). Contradictorily, the implicit
comparison also pertains to the later prehistory, e.g., the impressive visual culture of the
Bronze Age, and thus also reflects a progressivist view of a ‘simpler’ stone age. In this strange
way, the portable art of the Mesolithic does indeed fall between two stools, chronologically
speaking. Some ornamented artefacts of the South Scandinavian Mesolithic are, however,
‘exquisitely’ decorated (fig. 1.1), and have consequently received more scholarly attention
(e.g., Andersen & Trolle 1997; Nash 2001a).
Figure 1.1: Examples of ‘exquisitely’ ornamented portable art objects from the Mesolithic of
South Scandinavia. Left: antler axe from Værebro Å (CC-BY-SA, Arnold Mikkelsen, Na-
tionalmuseet). Right: antler axe from Skærbæk (CC-BY-SA, Lennart Larsen, Danmarks
Oldtid, Nationalmuseet). Edited by author.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
4
1.2 Research aim
In contrast to the predominance of qualitative approaches this study will consider the en-
tirety of South Scandinavian portable art, bringing a quantitative and data-driven contribu-
tion to the research field. The recognition that no one has investigated the semiotic hypothe-
ses in the context of South Scandinavian Mesolithic portable art is the raison d'etre of the
present dissertation. The objective here is not, however, to prove true that Mesolithic motifs
functioned within a semiotic system of encoded information. In fact, I would from the offset
agree with the – entirely unspoken – assumption that they probably were not. Yet, given the
recent resurgence of semiotic hypotheses, this dissertation will follow Paul Bouissac (1994a)
in his persistence that such an assumption should in fact be properly tested, rather than dis-
regarded a priori:
” … a sound methodological approach would consist of first trying to find ways to eliminate
some of the competing hypotheses. This is why the 'strong' semiotic hypothesis, which pro-
poses to construe palaeoart from its earliest records as a set of writing systems, should be
considered first, because it is the one which can be the most easily proven false.”
(Bouissac 1994a: 365)
Bouissac’s notion of the ’strong’ and ‘weak’ semiotic hypotheses, and how best to assess
them, are explicated in Chapter 3. Apart from testing these hypotheses, the dissertation will
also perform general descriptive statistics to query the corpus of materials for patterns across
different variables. This dissertation therefore builds on and expands earlier analyses such as
that by George Nash (1998) who limited his examination to a selection of 100 engraved arte-
facts. The descriptive statistics will be relevant in addressing the semiotic hypotheses, and
equally importantly, central to the subsequent broader discussion of how to understand this
phenomenon. A demonstration of semiotic structure would be a significant step towards un-
derstanding the function of the engravings – but, as this approach has not yet been applied
to this part of prehistory, a negative result will be as beneficial a contribution – since in that
case, as Bouissac argues, the most easily falsifiable hypothesis can then be put to rest.
1.3 Outline of the dissertation
A replicable and quantitative assessment of the semiotic hypotheses for this corpus of prehis-
toric ornamentation necessitates a novel methodology. This section outlines how such an
evaluation is rendered possible (see also fig. 1.2). To begin with, the South Scandinavian
Mesolithic is introduced and qualified as a relevant ‘contextual area’ (Richter et al. 2012), fit
for independent analysis. The scope of the study extends to the entirety of the Mesolithic,
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
5
encompassing the Maglemose, Kongemose and Ertebølle culture complexes, situated within
the early to mid-Holocene pollen zones Preboreal, Boreal and Atlantic. The core area of study
is present-day Denmark, with the inclusion of Southern Sweden, Northernmost Germany,
and the Baltic coast of Poland (see fig. 4.1). A quantitative approach to this study area is ren-
dered possible by the work of Tomasz Płonka (2003), whose comprehensive catalogue con-
tains the vast majority of known Mesolithic portable art objects. While twenty years old, this
publication is the to-date most thorough synthesis of Mesolithic portable art. Płonka’s pro-
posed motif classification (2003: 315–23, Supplementary Material 7) is also adopted for the
present dissertation. As will be discussed, such ‘typologies’ will always be a problematic en-
deavour for this type of material culture, yet Płonka’s contribution is undoubtedly the most
exhaustive, and by this token, offers the best angle of entry for a quantitative approach to
this material.
The dissertation will outline the research history of interpretative approaches to prehistoric
art of Mesolithic South Scandinavia, as well as the European Upper Palaeolithic. It is demon-
strated that these two fields of study are to some extent uncoordinated, as the interpretative
approaches deriving from Continental Europe have not been sufficiently emulated by South
Scandinavian scholars. This particularly concerns the semiotic hypotheses, which are re-
viewed in Chapter 3. It is demonstrated that, while critical issues regarding methodology and
basic assumptions remain unresolved, it is necessary to explicitly investigate the potential
presence of semiotic information systems for the South Scandinavian Mesolithic as well.
In the first part of the analysis (Chapter 5.1), the portable art objects and their engraved mo-
tifs are presented in comprehensive, descriptive statistics, considering the most relevant var-
iables: quality of contextual information, chronology, spatiality, material, and artefact type.
Following the descriptive statistics, the basis for rejecting or validating the presence of semi-
otic structure will be conducted by investigating motif co-occurrences. This is attempted by
adopting a method from the field of computational linguistics, namely ‘k-skip-n-grams’ (or
‘skipgrams’, see Chapter 4.2). Generating skipgrams allows for a recognition of motif co-oc-
currences at different levels (two-, three-, and four-motif co-occurrences). By calculating the
observed (O) and expected (E) ratio of the co-occurrences, a custom measure of statistical
significance is attained, which will help qualify which co-occurrences were potentially mean-
ingful to the Mesolithic people who produced them. Finally, the results of the analyses are
discussed and synthesised with current archaeological knowledge about the South Scandina-
vian Mesolithic period.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
6
Figure 1.2: Flowchart of the dissertation
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
7
Chapter 2
Stand der Forschung
This chapter sets out the empirical and theoretical framework on which the dissertation is
based. It is divided into three parts: Chapter 2.1, a brief, chronological overview of the South
Scandinavian Mesolithic, including reflections on regional variation; Chapter 2.2, an intro-
duction of the portable art of the South Scandinavian Mesolithic; and Chapter 2.3, the inter-
pretative approaches previously proposed for prehistoric art, outlining the major strands of
research history. This includes both prehistoric ornamentation outside the study area (focus-
ing on the European Upper Palaeolithic), as well as within.
2.1 Overview of the South Scandinavian Mesolithic
The ‘Mesolithic’ is a time-transgressive cultural denominator, with varying use in different
parts of the world. In Europe, it refers to the hunter-gatherer cultures of the early to middle
Holocene, from the last breaths of the Weichselian glaciation to the onset of agriculture. In
this period, climatic conditions generally improved, and hunter-gatherer communities
adapted to a changing world, which for the most part meant an increase in available food-
stuffs, both in terms of plant, animal and marine subsistence (Jochim 2011: 133). In other
words, the Mesolithic was a period of postglacial diversification and an ecological “Golden
Age” for foragers (Graeber & Wengrow 2022: 258). The Mesolithic societies of Southern
Scandinavia were later to adopt agriculture than their southern neighbours, and the period
in this particular region thus dates to approximately 11.000-6000 BP (see fig.2.2). This area
of the Mesolithic world has had “an importance disproportionate to its limited geographical
extent” (Larsson 1990: 257), due to favourable preservation conditions that characterised
much of the North European lowlands, as well as a relatively long research history. Still,
much remains unknown about the life of these hunters, fishers, and foragers.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
8
2.1.1 Chronological framework
As our period of interest represents the earliest part of the present interglacial, climate and
landscape development is a crucial aspect in understanding this period. Drastic vegetational
development, recognised through palynology (Iversen 1960), meant dynamically varying
landscape types. While the early Mesolithic was a coherent and heavily forested landmass,
iso- and eustatic processes had by the late Mesolithic transformed the region into a highly
marine landscape. This entails that Mesolithic research of this area is haunted by a very dif-
ferential site availability, as the early Mesolithic landscape has been inundated to a great ex-
tent (Christensen 2001; Astrup 2018: 106:67; Bailey et al. 2020). The three Mesolithic cul-
ture groups of the region are cursorily presented in the following.
The Maglemose culture (ca. 11000-8400 BP) inhabited the thin Preboreal birch and pine
landscapes which gradually developed into more dense forests. Complementing a wide sub-
sistence strategy, the Maglemosians are known to have hunted large mammals such as au-
rochs and elk (Jessen et al. 2015) with bow and arrow – the arrowhead being the characteris-
tic microlith (Sørensen 2006; Sørensen et al. 2013: 40–41). The typological changes in the
microlithic industry are used to further subdivide the period (e.g., Sørensen 2012). The
toolkit of bone and antler artefacts see a diversification from earlier periods, including leis-
ters, axe sockets, chisels (Price 1985; Larsson 1990), as well as the distinctive barbed bone
points (Blankholm 2008: 112). The scarcely known settlements, comprising of a small num-
ber of huts (Andersen et al. 1982; Sørensen 1987), suggest seasonal mobility, and the winter
settlements remain largely undiscovered (Price 1985: 348–49; Larsson 1990: 277; Sørensen
et al. 2018).
Archaeological recoveries from the Maglemosian have mostly been made in inland peat bogs,
which at the time of occupation were lake environments. The core area is the Åmose region
of Western Zealand (Mathiassen 1943). The Maglemose is thus traditionally understood as
an ‘inland’ culture, however characterised by high mobility (Schilling 2003). As mentioned,
the potential Maglemosian coastal settlements would be deeply inundated and therefore
practically inaccessible to archaeological recovery (Christensen 2001). This is a significant
taphonomic bias, and the argument that significant coastal adaptations occurred only in the
middle/late Mesolithic is therefore equivocal (Astrup et.al. 2021). In present day Norway,
however, isostatic uplift has meant that the early Mesolithic coastline now lies inland, and no
strong patterns of coastal settlement have been recognised there (Schmitt 2007; Bailey
2008; Breivik 2014). Yet, recent measurements of 13C values (Fischer 2001: 5; Boethius &
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
9
Ahlström 2018) have put it beyond any doubt that some degree of coastal subsistence was
practiced by the Maglemosians – but it remains uncertain to which degree.
The Kongemose culture (ca. 8400-7400 BP) differs from the Maglemose with regards to
toolkit as well as settlement pattern. Though many artefact types remain in use, flint blades
appear and microliths are replaced by trapezoid and rhombic/oblique arrowheads (Bokel-
mann 1999; Blankholm 2008: 112). This corresponds to the supra-regional ‘trapeze-horizon’
(or ‘blade-and-trapeze horizon’), originally proposed by J.G.D. Clark (1958; Biagi & Starnini
2016; Sørensen 2017: 142; Gronenborn 2017), in which blade and trapezoid lithics look to re-
place existing armatures on a continental scale. This phenomenon, however, remains largely
unexplored. At the time of the Maglemose-Kongemose transition, the climate changed con-
siderably, from the Boreal dry climate to the warmer and more wet Atlantic, perhaps due in
part to abrupt climate events (Alley & Ágústsdóttir 2005). The rise in sea level seems to have
been particularly rapid at this point (Sørensen 2017: 22), yet Astrup (2018: 106:159) con-
cludes in his recent, comprehensive review that it remains uncertain whether this had signif-
icant societal impact. The Kongemose settlements are primarily coastal, with the emergence
of the first shell mounds (Andersen 2000: 370). Once again, due to the problem of eustacy, it
is uncertain to what degree this is a true reflection of changes in settlement pattern.
The Ertebølle culture (ca. 7400-5950 BP) is situated within the warm Atlantic chronozone of
the middle Holocene, variously known as the Altithermal or the Holocene Optimum. By this
point, iso- and eustatic changes had altered the landscape to roughly resemble that of the
present day, i.e., a landscape with many bays, inlets, and islands. Chief characteristics of the
Ertebølle are 1) more permanently settled (semisedentary) coastal settlements and 2), argua-
bly, increased ‘societal complexity’. The recognition of the Ertebølle coastal adaptations owes
much to the favourable preservation conditions at the shallow waters off the modern-day
coasts (Bailey et al. 2020). Inland sites are also known and are popularly hypothesised as
seasonal camps for the acquisition of furs and other forest resources (Andersen 1995) . δ13C
and δ15N signatures from Ertebølle human bones support a highly marine diet, demonstrat-
ing a consistent economic relation to the coast (Fischer et al. 2007) . Increasing sedentism is
recognised through archaeological documentations of long and consistent occupation se-
quences at sites such as Tågerup (Karsten & Knarrström 2001, 2003), and Tybrind Vig (Ras-
mussen 2013), and further supported by the presence of kitchen middens, centuries-long ac-
cumulations of primarily oysters. Other well-preserved finds within these calcareous shell
mounds reveal that many everyday activities, as well as human burials, were at times closely
connected to these grand anthropogenic structures (Astrup et al. 2021).
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
10
The argued increase in social complexity of the Ertebølle is a contentious topic. The some-
what synonymous notions of ‘complex’ (Price & Brown 1985) , ’affluent’ (Sahlins 1972: 1;
Price 1985), and ‘delayed-return’ (Woodburn 1982) hunter-gatherer societies was initially a
critique of the progressivist, evolutionary paradigm that all prehistoric foraging societies
were egalitarian in nature (Price & Gebauer 2005: 146; Petersen & Meiklejohn 2007: 185). In
turn, this binary division between ‘simple’ and ‘complex’ hunter-gatherers has been critiqued
as itself being progressivist, or colonial, in nature (Layton & Rowley-Conwy 2013; Graeber &
Wengrow 2022: 128–30; Elliott & Warren 2023), and “[leading] to debates over definitions
rather than the investigation of variability” (Gamble 1991: 4, see also Warren 2005; Price
2020). However, that marked societal changes occurred, or accelerated, in the Ertebølle pe-
riod is unquestionable. Central examples follow:
o Although not entirely Ertebølle novelties, delayed-return economic strategies look to
have intensified, e.g., in the form of fish traps and weirs (Pickard & Bonsall 2007)
and a generally heavier exploitation of faunal resources such as furs and antler (Row-
ley-Conwy 1995; Layton 2001; - see also Maring & Riede 2019, demonstrating possi-
ble Ertebølle management of wild boar).
o Following the preceding Kongemosian, the Ertebølle culture complex shows a widen-
ing artifact repertoire – one example being the Limhamn axes from late Ertebølle
Zealand and Scania (Sørensen 2007), but most notable is the introduction of pottery
in the form of pointed base vessels and ‘blubber lamps’, seemingly with a different ge-
ographic origin than that of the succeeding neolithic pottery (Blankholm 2008: 115–
16; Povlsen 2013; Dolbunova et al. 2023).
o Several factors point to incipient social stratification and territoriality: marked re-
gional differences in material culture (Blankholm 2008: 113), the emergence of pres-
tige items (Klassen 1999), violent interhuman encounters (Karsten & Knarrström
2001: 171; Thorpe 2003: 155–59), and the emergence of cemeteries (Larsson 1995).
These factors may all be related to a supposed increase in population density during
the Ertebølle (Price & Brown 1985: 16; Price & Gebauer 1992: 108).
The Ertebølle period ended with the onset of the Neolithic at around 5950 BP, although
only after approximately thousand years of having neighbouring neolithic communities
in Northern Germany (Blankholm 2008: 126–27; Gron & Sørensen 2018). This delayed
transition to agriculture has been studied intensely, yet the cause is not broadly agreed
upon. The prevailing theory is that the Atlantic climate of the Ertebølle was ideal for a
hunter-fisher-gatherer lifestyle and did therefore not warrant economic turnover – but
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
11
that when the transition occurred, it was seemingly a result of both northward migra-
tions and active involvement by the indigenous hunter-gatherers (Sørensen 2014: 263).
Figure 2.1: Approximate date ranges for chronozones and South Scandinavian Mesolithic
culture complexes.
2.1.2 Regional variations – South Scandinavia as a ‘contextual area’
In order to assess the function of the portable art, a highly relevant question is how culturally
interconnected the South Scandinavian area was, or put another way, how different from
each other the regional ‘ethnic’ groups were, supposing there were any such. This may be re-
flected in the spatial distribution of artefact types, assuming that regional groups had region-
ally distinct material correlates. As Bergsvik (2003: 300) notes, it is then necessary to con-
sider the socio-historical circumstances that could have formed such groupings. These ques-
tions are difficult to answer as the find representativity varies geographically, and especially
considering that we are dealing with approximately 5000 dynamic years of prehistoric cul-
ture. Applying the concept of ‘contextual areas’ (sensu Richter et al. 2012) to the South Scan-
dinavian Mesolithic evidence, it is demonstrated that this period and region qualifies as a
‘contextual area’, relevant as an analytical unit of study. The notion of ‘contextual area’
“…serves as a methodological instrument to propose analytic, temporal-spatial units which
comprise cultural, anthropological, and environmental arguments.” (Richter et al. 2012: 20).
Spatial variability within the Maglemosian technocomplex is difficult to chart, perhaps due
to the “… sparse settlement [, which] would not have encouraged the formation of groups
with clear distinguishing features” (Larsson 1990: 288–89). In his analysis of the Maglemose
settlement patterns at Holmegårds Bog (Zealand), Schilling (2003: 357) notes that these in-
land lakeside settlements were only revisited sporadically, and that the human groups must
have employed geographically large social territories. Several studies have attempted to infer
these proposed social territories through material culture proxies. Newell et al. (1990) have
suggested distinctive ‘band’-level territories based on the spatial distribution of personal or-
naments such as amber, shell and tooth pendants (fig.2.2).
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
12
Figure 2.2: Newell et al.’s proposed ‘tribe’ and ‘band’ level social territories of South Scandi-
navia, ca. 9550-7868 BP (late Maglemose/early Kongemose), based on the study of per-
sonal ornaments. From Newell et.al.1990, figs. 130 & 131, edited by author.
Highly characteristic of the Maglemose culture are the barbed bone points, extensively dated
and analysed by Jensen et al. (2020). These barbed points are only known from easternmost
Jutland, Zealand, and Scania, with a heavy representation on Zealand – yet, the authors note
that this may simply reflect the geographic extent of soil conditions conductive for preserva-
tion of osseous material (Jensen et al. 2020: 3). The main finding of this seminal article is a
radiocarbon hiatus from approximately 10.300–9.700 cal BP, and the marked differences in
barbed bone point morphology before and after this hiatus. Another relevant observation is
David’s (2003: 656) who argues, based on bone and ambler assemblages, that the North Ger-
man site of Hohen Viecheln may represent a southern geographical limit for the early Magle-
mosian technocomplex. In summary, territorial divisions of Maglemose populations remain
hypothetical at best. This is perhaps due to the limited possibilities for synchronic analyses
but could also imply that there were in fact no substantial regional differences in material
culture, due to the sparse and mobile settlement structure. Diachronic studies are more ac-
cessible, and Maglemosian variability is thus best documented temporally (see also Gendel
1984).
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
13
As with the Maglemose culture, Zealand is the core area of archaeologically known
Kongemose sites (Sørensen 2017: 94). In other areas, such as Scania, Jutland and Northern
Germany, Kongemose assemblages vary to a considerable extent – and in the particular case
of Northern Germany, it has long been discussed whether the Kongemose concept even ap-
plies (Hartz 1985, 2009: 401; Sørensen 2017: 104–8). Compared to the preceding and suc-
ceeding culture complexes, the Kongemosian has been studied less intensely (Larsson 1990:
289), and the cultural taxonomy itself is perhaps in need of a critical revision (Friman 1996;
Reynolds & Riede 2019). In the Ertebølle period, the regional differences become more
clearly pronounced, and are well supported archaeologically. Areas of South Scandinavia
which are not all well represented by Maglemose and Kongemose assemblages show marked
increases in Ertebølle habitation. These include the Limfjord area, the eastern coast of Jut-
land, and the eastern coast of Zealand. Concurrently, aspects of material culture become re-
gionally distinct. This is commonly interpreted as a result of increased population sites and
more sedentary lifestyles, factors resulting in a need to make territorial claims. Increased ter-
ritoriality has been argued on the basis of flake axe typology (Vang Petersen 1984: 17, see
fig.2.3), pottery (Jennbert 2011) the placement of human burials (Grøn 2020), and ornamen-
tations on wooden paddles (Andersen 2011: 11). The Great Belt (Storebælt), separating East-
ern and Western Denmark, has been suggested as a cultural dividing line (Ballin 2017: 333).
Figure 2.3: Regional variability of different Ertebølle artefact types. Left: T-axes and Lim-
hamn axes (Vang Petersen 1984: 15), Centre: Flake axe morphology in North Eastern Zea-
land (Vang Petersen 1984: 17), Right: Pointed base pottery vessels (Sørensen 2015: 3). Ed-
ited by author.
However, such regional differences only pertain to a select few artefact types; broadly, the
Ertebølle technocomplex covers the entirety of South Scandinavia, i.e., Denmark as well as
Southernmost Sweden and Northernmost Germany (Vang Petersen 1982: 186), and there is
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
14
evidence of Late Mesolithic mobility and exchange (Vang Petersen 1990; Larsson & Molin
2017a; Bennike et al. 2021). It is interesting to note, on the basis of the above summary, that
throughout the Mesolithic, Eastern Denmark has been seemingly more culturally intercon-
nected with Southern Sweden than with Western Denmark (Vang Petersen 1982: 186; Pe-
tersen 1993: 49, 2006; Sørensen 2007).
In summary, the Maglemose technocomplex appears to cover a large geographical area, ex-
ceeding South Scandinavia, with limited regional variability. If social territories existed, the
evidence suggests that these were large, fluid, and without strong material correlates. The
Kongemose technocomplex sensu stricto appears limited to Eastern Denmark and Southern
Sweden, while the picture remains unclear in West Denmark, Northern Germany, and the
Baltic South Coast. Increased regionalisation is strongly supported in the Ertebølle, but not
in terms of fundamental spatial variability – the Ertebølle culture concept remains fitting for
the entire South Scandinavian region. When considering these observations in tandem with
the ecological and economic temporal changes reviewed in Chapter 2.1.1, it is evident that, in
the Mesolithic, the approximate area of South Scandinavia was a unit of broadly correspond-
ing continuities and discontinuities. The area’s distinctive environmental changes, shifting
by inundation from continental landmass to coastline-dominated landscape, is the additional
factor that qualifies the argument of Mesolithic South Scandinavia as a ‘contextual area’.
2.2 The portable art of the South Scandinavian Mesolithic
As there are no caves or rock shelters and hardly any protruding bedrock in the South Scan-
dinavian area, there is a natural absence of parietal art. Portable art, however, is abundant
compared to other regions of Mesolithic Europe (Płonka 2003: 158). The high number of
South Scandinavian ornamented artefacts compared to the rest of Europe is undoubtedly to
some degree a result of a long research history, as well as suitable conditions for preservation
of osseous material offered by the North European lowlands (primarily in the form of peat
bogs and submerged sites at shallow waters). The engravings onto Mesolithic artefacts were
one of the observed phenomena that contributed to making the Danish Mesolithic an in-
tensely studied period from the onset (Müller 1896). Posing a central issue, many of these
artifacts were recovered with inadequate documentation in the 19th century, or in other ways
are stray finds lacking contextual information (Płonka 2003: 18–20). However, the amount
of material allowed for a motif classification by Clark as early as 1936 (Clark 1936: 47). He
later proposed a modified version (Clark 1975: 159), and both Andersen (1981: 47), Nash
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
15
(1998: 78), and Płonka (2003: 315–25) have contributed with suggestions for motif classifi-
cations.
The portable art consists of primarily geometric motifs (figurative depictions of animals and
humans are also known (see fig .2.8 and the Ystad axe (number 5) in fig. 2.7), engraved onto
artefacts of bone, antler, amber and stone. A flint blade or flake was most likely used as the
engraving tool, and often, the surface of the object was polished smooth prior to engraving
(Andersen 1981: 47). In some cases, it has been demonstrated that the engraved grooves
were filled with black resin, accentuating the motifs (Toft & Petersen 2015: 205). The orna-
mented artefacts are most often functional objects (Clark 1936: 162; Larsson & Molin 2017a:
401) such as axes, harpoons, daggers, etc. Artefacts of more ambiguous functions, such as
shafts, pendants and stone plates, are also sometimes ornamented. Wooden objects bearing
ornamentation have also been recovered, underlining how little we know about the overall
ornamentation of the period. As Sørensen (2017: 141) stresses, the decorated wooden pad-
dles (see fig.2.9) have an ornamental style that fundamentally differs from the portable art
otherwise known from the period. It is quite possible that wood and other more perishable
materials, possibly including body ornamentation, were used as media. It can therefore
safely be assumed that what is archaeologically recovered and available for analysis repre-
sents only the tiniest fraction of what actually was.
In cases where the archaeological context of an ornamented object is known, no strong pat-
terns seem discernible; some are known from settlement floor contexts, and a few are be-
lieved to have been deliberately deposited in lakes or bogs (Nash 2001a: 234). Portable art is
also known from refuse layers (e.g., Larsson 2000: 36), yet Toft (2005: 617) advocates scep-
ticism concerning this ‘refuse layer’ default definition. Through early Mesolithic intra-site
analyses, he tentatively proposes that ornamented artefacts may have often been deposited
in specific ways and in connection with other socially important artefacts, as well as burials.
In a few cases from the late Mesolithic, portable art objects have been deposited as grave
gifts in inhumation graves, examples being Strøby Egede (Petersen 1990: 23), Fannerup F
(Rasmussen 1989), and Skateholm I & II (Larsson 2000: 45). Yet, statistically qualified in-
ferences concerning the deposition of portable art objects remain absent. Another observa-
tion that has not (to the authors knowledge) been quantitatively demonstrated is the notion
that Kongemose ornamentation is especially ‘complex’ i.e., bearing more motifs per artefact
on average than Ertebølle and Maglemose objects (Sørensen 2017: 135). A final example of
somewhat unsubstantiated claims is that while some artefact types are seldomly or never or-
namented, others, such as the slotted bone daggers, are in almost all cases decorated (Lars-
son 1990: 286, 2000: 37). The Late Mesolithic slotted bone daggers are most often made
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
16
from Elk (Alces Alces) bone, and the Danish finds are thus considered exotica (most likely
from Sweden), as this species was absent from the Danish islands at this point (Petersen
2015: 1:141). Some motifs have received far more scholarly attention than others. The three
types that have been studied most extensively are presented in the following, namely ‘fringe’,
or ‘comb ornaments’, zigzag motifs, and the so-called ‘sheaf of grain’.
The ‘fringe- ‘, or ‘comb ornament’ (‘frynseornament’) seems to be strongly associated with
amber pendants (Toft & Petersen 2015). The motif is characterised by Toft as “…either a ver-
tical band or line stroke with multiple shorter or longer parallel horizontal or oblique line
strokes hanging from them” (Toft 2017: 263). This corresponds to Płonka’s type B3, with B2,
B10, and B11 arguably being its variations (Toft & Petersen 2015: 206). Yet, Toft finds that
fringe ornaments with long perpendicular lines (B3) occur only on amber pendants, whereas
fringes with short perpendicular lines also occur on bone and antler artefacts (Toft 2017:
265). The amber pendants in question, and thus also the B3 motif, have been stylistically
dated broadly to the Maglemosian by Mathiassen (1959: 187-189). Toft’s research supports
the Maglemose attribution (Toft 2006: 109, 2017: 262–66), but concedes that the stylistic
dating remains uncertain and could be overruled by future direct dates.
Figure 2.4: Fringe- or comb motif on an amber pendant from Limfjorden. The engravings
are redrawn, argued to imitate the black resin fill. From Toft & Petersen (2015, fig. 6).
The zigzag motif appears often on decorated Mesolithic objects, as well as on objects of older
date. The motif comes in a plethora of variations, making it particularly difficult to catego-
rise. Płonka’s type A24 may be considered the default, simple zigzag, but A13, A14, A21-29,
B15-16, C12, F20, F51-F55, H3, and I5 (optionally, see Supplementary Material 7) all argua-
bly qualify as zigzag variations. Several papers (Płonka & Kowalski 2017; Vang Petersen
2018, 2021) have presented qualified arguments that zigzag engraving is a Northern Euro-
pean tradition that extends back into the Palaeolithic. Based on a sample of 100 artefacts,
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
17
Nash (1998: 49) shows that zigzag motifs are chronologically evenly distributed across the
South Scandinavian Mesolithic.
Figure 2.5: Antler artefact from Holmegård IV, bearing zigzags (A24) as well as the vari-
ant A28 (zigzag with perpendicular barbs). CC-BY-SA, John Lee, Nationalmuseet.
The ‘sheaf of grain’, or ‘wheatsheaf’ motif (Płonka’s types C4-C6) has perhaps received most
attention of all, as it is the motif that has been demonstrated to occur within the most limited
spatio-temporal range, namely the early Ertebølle (tentatively dated) in Eastern Jutland and
neighbouring islands such as Fyn and Langeland (Andersen 1981: 48). Its name is due to the
motif’s resemblance to a bundle of grain, possibly related to Ertebølle connections with
agrarian societies to the South (e.g., Nimura 2013: 31). However, whether the motif is actu-
ally a depiction of a sheaf of grain is entirely conjecture, and this popularly used term may
thus be misleading.
Figure 2.6: The ‘sheaf of grain’ motif on an antler axe from Eskelund.
From Andersen (1981: 9).
Clark (1975: 136) noted that the portable art must be one of the most promising proxies for
inferring social territories and networks of the South Scandinavian Mesolithic – and this ap-
proach has also since been the preferred angle of entry in studies of this material (Nash
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
18
1998; Terberger 2006; Toft 2006, 2017; Larsson & Molin 2017a; Płonka & Kowalski 2017).
The advantage of this approach is that it allows for sidestepping the difficult questions re-
garding the purpose of the ornamentations. Concerning Maglemosian regionality in motif
distributions, one gets the same impression of large and fluid territories as was discussed in
the previous subchapter. Płonka & Kowalski (2017) take the wide distribution of zigzag mo-
tifs across Late Palaeolithic/Early Mesolithic Northern Europe to imply high mobility and a
“free flow of information” (Płonka & Kowalski 2017: 163). Toft (2017: 263–65) argues for a
Maglemose social territory based on the abovementioned fringe ornaments with long teeth
(B3), covering the Western part of Denmark and extending out into the now submerged Dog-
gerland. Toft (2006) and Terberger (2006) have also engaged with Newell et al.’s (1990) pro-
posed Maglemose band areas (see fig. 2.2), but discussions regarding a North/South division
between present day Fehmarn Belt remain unresolved. While no spatial variability in
Kongemose motifs has been suggested, Andersen (1981: 20) argues for a division in Ertebølle
motif repertoire between East and West Denmark, divided by the Great Belt. This is however
solely based on the spatial distribution of the ‘sheaf of grain’ motif, and contradictorily, Nash
includes the western shore of Zealand in his proposed ‘Kula ring’ exchange area (discussed
further in Chapter 2.3.2).
For decades, Clark’s contributions (1936, 1975) have constituted the framework for investiga-
tions into the South Scandinavian portable art, but more recent work has taken its place as
the key reference. Płonka’s (2003) monograph on European Mesolithic portable art – consti-
tuting the empirical foundation of the present dissertation – is to be regarded as the most ex-
tensive study of this material to date. Płonka’s (2003) monograph, along with Nash’s (1998)
‘Status, Exchange, and Mobility…’, are the most ambitious in terms of both theoretical appli-
cation and quantification. It is worth noting that Clark, Nash, and Płonka are all scholars
from outside South Scandinavia, and that among scholars native to the region, only a hand-
ful are actively engaged with Mesolithic portable art. Most constructive of all is perhaps
Płonka’s catalogue of (nearly) all portable art objects from the entirety of Mesolithic Europe
(Płonka 2003: 217–325). The markings on all these artefacts are attributed to motif catego-
ries of his own design (see Supplementary Material 7 and fig. 4.2). These motif attributions
are solely based on qualitative assessment, and the unavoidable subjectivity is palpable.
Płonka’s interpretations must thus of course not be assumed to necessarily reflect Mesolithic
realities. That being said, his laborious contribution constitutes a much-needed platform
from which to conduct further research with a more holistic overview of this difficult mate-
rial. Not many, however, have made use of this platform in the twenty years that have passed
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
19
since its publication, no doubt due to the general stasis in research on South Scandinavian
portable art.
In summary, the investigation of the portable art of Mesolithic South Scandinavia is charac-
terised by three main issues:
I. The lack of research output with quantitative and statistical dimensions,
II. the difficulties and subjective biases of creating classifications, owing to the marked
heterogeneity of the motifs, and
III. the uncertain dating of many of the artefacts. In the absence of well-defined stratigra-
phies or direct radiocarbon dating, ornamented objects are often dated on the basis
of their motifs (e.g., Hartz 2009: 497). Such dating based on motif morphology is un-
certain, and studies of other domains of material culture have in recent years demon-
strated the fragility of typologies and taxonomies (von Petzinger & Nowell 2011; e.g.,
Reynolds & Riede 2019; Matzig et al. 2021; Tsirintoulaki et al. 2023).
A quote from Lars Larsson serves as a fitting closing remark for this chapter, summarising
the high degree of uncertainty that persists in the study of Mesolithic portable art: “Attempts
to find chronologically related motifs or combinations of motifs have been made without
achieving convincing results” (Larsson 2000: 33).
Figure 2.7 (next page): – Examples of South Scandinavian Mesolithic portable art.
1: Antler shaft, Silkeborg (Müller 1918: 1:9), 2: Amber pendant, Unknown (west coast of
Jutland) (CC-BY-SA, John Lee, Nationalmuseet), 3: Wood paddle, Tybrind Vig (Bailey et al.
2020: 66), 4: Antler axe, Carstensminde (Sørensen 2017: 138), 5: Antler axe, Ystad (Larsson
2017: 64), 6: Antler axe, Magleø (CC-BY-SA, Arnold Mikkelsen, Nationalmuseet), 7: slotted
bone dagger (detail), Skellingsted Bro (Jensen 2001: 153), 8: slotted bone dagger,
Kongemose (Sørensen 2017: 139), 9: bone mattock-head, Ryomå (CC-BY-SA, Lennart
Larsen, Danmarks Oldtid, Nationalmuseet). Redesigned by author.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
20
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
21
2.3 Interpretative approaches to prehistoric art
The following review introduces the main avenues of interpretation of prehistoric ornamen-
tation. I focus on the Upper Palaeolithic and Mesolithic, strictly from a European perspec-
tive, echoing the unfortunate eurocentrism that has been prevalent in this field of research
(David & McNiven 2018: 1). Yet, an overview of the global developments of prehistoric orna-
mentation is far beyond the scope of the present dissertation, and thus, we stay within these
limits. In an attempt to clarify this complex field of study, the opposing interpretations are
ordered into five arguable ‘schools of interpretation’, namely ‘decorative’, ’rituo-magic’, ’so-
cio-political’, ’weak semiotic’, and ’strong semiotic’. These five ‘schools’ are devised by the
author and do not claim to encompass all viewpoints – those falling outside them are also
briefly touched upon. For a visual summary of these research historical interpretative devel-
opments, see fig. 3.2.
2.3.1 Interpretations of European Upper Palaeolithic markings
19th century discoveries of Pleistocene ornamentation (e.g., Lartet & Christy 1875) opened
new fields of academic inquiry. From the onset, these prehistoric objects were deemed com-
parable to contemporary non-western art, and both naturally fell under the category of
‘primitive art’ (Moro Abadía 2006: 121; Moro Abadía & González Morales 2013). Famously,
the accumulating evidence of Palaeolithic art was chronologically organised by Breuil (1907).
Through observations of stylistic traits, he developed a four-stage chronology of European
cave art, progressing from ‘crude’ to ‘naturalistic’ (Moro Abadía 2006; Moro Abadía & Gon-
zález Morales 2013). This stylistic approach was further developed by Leroi-Gourhan and his
contemporaries (e.g., Raphael 1945; Leroi-Gourhan 1958, 1982; Laming-Emperaire 1969),
who applied structuralist principles – seeking to demonstrate internal structure – to the pre-
historic markings. (Moro Abadía & Palacio-Pérez 2015: 657). The work of Breuil and Leroi-
Gourhan is key literature for scholars of this field, and their pioneering research remains in-
fluential. Yet, today, in what has been labelled the ‘post-stylistic era’ (Moro Abadía & Gonzá-
lez Morales 2007), it has been argued that their stylistic approach rested on a progressivist
assumption of unilinear evolution – that the ‘quality’ of the art must naturally have improved
with time (Floss 2018). The focus on style also gave the prehistoric ornamentations a place in
art history, for better or worse (Bouissac 1994a: 349; Tedesco 2007).
The ‘decorative’ school of interpretation is synonymous with the notion of ‘art for art’s sake’.
By this is understood that prehistoric parietal and portable art was produced with the main
purpose of artistic expression or aesthetic pleasure. This was a common and often implicit
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
22
assumption in the research field’s early decades (e.g., Grosse 1894; Munro 1906). Yet, it
gradually fell into disfavour and has remained so, although proponents have always existed
(Schlette 1969: 239; Halverson 1987; De Smedt & De Cruz 2011). The reason for its unpopu-
larity from processualism onwards seems to be the consensus that humans of later prehis-
tory were more complex and purposeful than that (e.g., Lewis-Williams 1987: 79). However,
one could contend that performing art for art's sake is just as highly complex a behaviour as
any other. Without delving further into this debate, suffice it to say that the arguments for
‘art for art’s sake’ seem to be largely based on disproving the opposing viewpoints. Thus, this
school of interpretation can be regarded (perhaps problematically) as the ‘default’ interpreta-
tion; if no purpose can be demonstrated, there was perhaps none, and the engravings were
produced for their visual aesthetics and the pleasurable responses they elicited.
A school of interpretation which has retained its popularity is the ‘rituo-magic’, often in the
form of sympathetic hunting magic arguments (Reinach 1903; Cartailhac & Breuil 1906;
Guthrie 1984; Lewis-Williams 1997; Lewis-Williams & Clottes 1998; Keyser & Whitley 2006;
Palacio-Pérez 2010). This idea originates in the abundance of Pleistocene animal depictions
coupled with ethnographic observations – the comparison between modern and prehistoric
hunter-gatherers was early on deemed a promising inferential method. The core idea of the
‘magic’ approach is that the produced images are to be understood as an integral part of rit-
ual activity, perhaps with mythological dimensions (Porr 2010). Moreover, such magic would
have had a practical function in the literal sense, as it would have been instrumental to hunt-
ing success, warding off evil, or similar purposes. A complementary branch of hypotheses
within the ‘rituo-magic’ sphere argue that cave art was a product of ‘altered states of con-
sciousness’ (Lewis-Williams et al. 1988; Solomon 2019). The manifold ethnographic demon-
strations of the complex ritual dimensions of hunting (Keyser & Whitley 2006: 6–8; Will-
erslev 2007) provide support to the ‘hunting magic’ hypotheses – but only if uniformitarian
assumptions between present and prehistoric hunters are credible, which will always remain
an ambiguous bridging argument (Wobst 1978; Francfort et al. 2001).
Contrary to the above is the ‘socio-political’ avenue of interpretation, arguing that markings
were used to signal individual or group affiliations, or perhaps even intra-group social hier-
archies (Sauvet et al. 2018). A socio-political function of parietal art could also be the literal
or symbolic markings of social territories of human groups (Djindjian 2010: 360). Two
things should be noted: 1), that this is entirely different from the method of inferring social
territories from material indicators, in which no specific function of the engravings is as-
sumed, and 2), that a socio-political usage overlaps somewhat with the ‘weak semiotic’
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
23
hypothesis (see Chapter 3), as this would necessitate a broadly agreed-upon system of mark-
ings.
In the above, the ‘decorative’, ‘rituo-magic’, and ‘socio-political’ schools of interpretation
have been briefly outlined, but the discourse concerning the purpose of prehistoric art is of
course more complicated than that. The notion of symbolism is perhaps one of the most per-
sistent suggestions of what the prehistoric art functioned as – a highly relevant concept, but
also one that might obscure more than it clarifies. Symbolism is a broad term, used as a su-
perordinate for many other interpretations. Motifs being symbols means that they func-
tioned as conceptual placeholders referencing something else. This pertains both to 1), fig-
urative depictions of animals, which sometimes are so simplified and stylised that they might
be understood as a symbol for, rather than a depiction of, the given animal, and 2), geomet-
ric motifs, which may be an abstract reference to something concrete. Lorblanchet (1989)
proposes that such symbolism was fine-tuned over a significant period of time. The ‘symbol-
ism’ superordinate encompasses hunting magic as well as the branch that considers (some
of) the markings to be representations of socio-political structure, origin myths (Laming-
Emperaire 1970) or the ‘weak’ semiotic hypotheses, to which we turn in Chapter 3. The capa-
bility of symbolism has often been used to infer cognitive competences of prehistoric hu-
mans, and the question of when symbolism first occurred is thus a classic question (Bur-
dukiewicz 2014).
An overarching distinction, proposed by Wiessner (1983) helps organise the often opposing
viewpoints: the markings might have been of the ‘assertive’ (“personally-based and which
carries information supporting individual identity” (Wiessner 1983: 258)), or the ‘emblemic’
style category (“formal variation in material culture that has a distinct referent and that
transmits a clear message to a defined target population” (Wiessner 1983: 257)). These con-
cepts are equivalent to ‘social interaction’ and ‘information exchange’, respectively (Barton et
al. 1994: 187). The assertive category does not necessitate symbolism but may include deco-
ration with a primarily aesthetic purpose or passive social conventions; the ‘information’ it
carries is only to be understood in the soft, social sense of the word. The emblemic category,
however, is information encoding and is thus synonymous with the semiotic hypotheses and
arguably also the ‘socio-political’ branch.
Other interpretations of prehistoric markings do not readily fit into the ‘schools of interpre-
tation’, some examples being that the markings were intended as ‘performative’, storytelling
devices (Feruglio et al. 2019), or as visual priming for hunting and gathering (Mithen 1988;
Hodgson 2003). In what could perhaps be said to be a response to the structuralist
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
24
approach, a recent surge of research on Pleistocene art attempts to understand the process in
which the art was made. In a critique of the ocularcentrism of all the abovementioned ap-
proaches, this line of inquiry attempts to understand the dialogical process in which the
markings were created. Such thinking falls within the ‘Ontological Turn’ in archaeological
theory (Wisher 2022: 36). The interpretations produced through this framework underline
the importance of the act itself of creating the markings, as such actions helped situate the
humans within their environment and its materiality. Thus, an important aspect of the mark-
ings was the act itself and its ontological relevance (Nyland & Stebergløkken 2020; Needham
et al. 2022). Prehistoric art has also received much attention from scholars outside the field
of archaeology, only occasionally in collaboration with archaeologists. Interested neighbour-
ing fields include psychology (Pettitt et al. 2020), linguistics (Miyagawa et al. 2018), neuro-
science (Ramachandran & Hirstein 1999), and the philosophy of art (Gosden 2001; Currie
2011, 2016; Stejskal 2021). As would be expected, linguistic approaches favour the semiotic
hypotheses, while neuroscientific and especially art-philosophical contributions tend to fa-
vour the ‘art for art’s sake’ interpretation.
2.3.2 Interpretations of South Scandinavian Mesolithic portable art
While the South Scandinavian portable art has been known for long and studied intensely,
most scholars are remarkably cautious when it comes to the question of interpretation –
more so, I would argue, than their colleagues studying the Pleistocene continental European
material. In many articles on the subject, the question of why is only dealt with in a few sen-
tences among the concluding remarks (e.g., Andersen 1981: 36; Toft & Petersen 2015: 213;
Larsson & Molin 2017b: 401). This is not due to scholarly ineptitude, but a testament to the
remarkable difficulty of interpreting this specific past human activity – a “tenuous” exercise,
as Nash (2001b: 19) calls it. Some exceptions (e.g., Nash 1998; Vang Petersen 2021; Płonka
2021), however, address the question of purpose head on.
The oldest scholarly assumption regarding the function of the South Scandinavian Mesolithic
material is that the geometric motifs were decorative patterns inspired by everyday objects
(Müller 1918: 1:4; Clark 1936). This line of thought is echoed in Płonka’s recent contribution
(2021), in which he argues that variants of the crosshatching motif (see fig. 2.9) are represen-
tations of fishing nets. The increasing use of net patterns during the Kongemose and Erte-
bølle periods, he argues, reflects a parallel increase in the ontological importance of a life-
style based on marine subsistence.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
25
The ‘rituo-magic’ interpretation has also often been suggested for the South Scandinavian
corpus (Hultén 1939; Jensen 2001: 128–32, 213–21; Sørensen 2017: 141; Vang Petersen
2021). Indeed, it seems to be almost present day consensus – Sørensen points to “… an in-
creasing level of agreement concerning the interpretation of the art as an expression of a reli-
gious symbolic language” (Sørensen 2017: 135). A few artefacts bearing stylised zoo- and an-
thropomorphic motifs are recurrently used to strengthen this viewpoint, two central exam-
ples (see fig. 2.8) being the antler shaft from Åmosen, which is argued to be a ‘magical’ hunt-
ing scene of sorts (Petersen 1982), and the long bone from Ryemarksgaard, suggested by
Nash (2001a) to represent ‘altered states of consciousness and the afterlife’.
Figure 2.8: The two portable art objects most often used to argue for a ‘rituo-magic’ pur-
pose. Top: the aurochs bone from Ryemarksgaard, depicting five humans next to three
parallel zigzags (CC-BY-SA, Lennart Larsen, Danmarks Oldtid, Nationalmuseet). Bottom:
antler shaft from Åmosen, dubbed ‘Deer Killer’, as it arguably depicts a human (right)
hunting a deer (left) (CC-BY-SA, Roberto Fortuna & Kira Ursem, Nationalmuseet. Rede-
signed by author.
Similarly, Vang Petersen (2021) suggests that some motifs were inspired by observations of
fauna; zigzags may have been representations of snakes, oval patterns the patterned skin of
red deer fawns, and hexagonal motifs the patterns on turtle shells. He goes on to argue that
the imitation of these animal patterns was for apotropaic (protective) purposes, as the ‘magi-
cally equipped’ traits of these animals were by imitation transmitted to the bearer of the or-
namented object.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
26
Figure 2.9: Płonka’s (2021: 596) and Vang Petersen’s (2021: 71–72) respective, and conflict-
ing, interpretations regarding the same motif types. Redesigned by author.
The above figure (2.9) demonstrates a relevant dichotomy: the interpretations put forward
by Płonka (2021) and Vang Petersen (2021) are contradictory, as the same motifs (G23 and
G31) are explained as representing fishing nets and animal patterns, respectively. While
Płonka argues for these motif’s function as referencing subsistence economy, Vang Petersen
believes they are magically protective. Neither of the propositions have been met with sharp
criticism, nor has one of them been acclaimed as a solid answer. This dichotomy showcases
the interpretative difficulties that pertain to portable art.
The argument that the engravings reflect socio-political signalling has also been proposed for
the South Scandinavian Mesolithic corpus, but in a fundamentally different way than is the
case for Pleistocene parietal art. Since no parietal art exists in the South Scandinavian Meso-
lithic, the notion of staking claim to a geographical area by ornamenting a certain place is
not directly applicable. Instead, the ornamented, portable objects have been used to infer
networks and geographical connectivity (Larsson & Molin 2017a). Andersen (2011: 10) is a
proponent of the ‘socio-political’ function, in proposing that the Ertebølle paddle ornamenta-
tions functioned as “individual or group identifiers”. Another case for a ‘socio-political’ pur-
pose can be found in George Nash’ seminal work, ‘Status, Exchange and Mobility: Portable
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
27
Art in the Scandinavian Mesolithic’ (1998), in which the portability of the decorated objects
is used to suggest that their purpose was indeed to be moved. Drawing a direct parallel to
Malinowski’s ‘Argonauts of the Western Pacific’ (1922), Nash suggests that the ornamented
objects circulated along the Danish coasts in an exchange network that facilitated prestige
and complex social structures. This attempt at an interpretation is a bold one, and Płonka re-
ceived it with scepticism, arguing that “the existence of such a system in the period of inter-
est is not sufficiently supported by evidence not the least because late Mesolithic art of South
Scandinavia may be distinguished into a number of local styles suggesting that art objects
were usually produced and used locally” (Płonka 2003: 178). Terberger (2006) also some-
what disagrees with Nash, arguing that the regional distribution of the ’sheaf of grain’ motif
might reflect one social territory rather than exchanges across several. Furthermore, Nash’
‘Kula ring’ proposal is perhaps contrary to the previously mentioned observation that the
Great Belt constituted a cultural dividing line in the Ertebølle period, and that generally,
Mesolithic Zealand seems most strongly connected to Southern Sweden. Nash’s proposal of
an Ertebølle ‘Kula ring’ is thus not widely accepted at present.
Figure 2.10: Nash’s (1998: 65) map of the distribution of the ‘sheaf of grain’ motif, and re-
lated theory that portable art circulated in a ‘Kula ring’ type network. Edited by author.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
28
Chapter 3
Theory
In this chapter, the semiotic branch of hypotheses is independently addressed, demonstrat-
ing that these have seen a resurgence in recent years. Based on this review, the dissertation’s
theoretical foundation is outlined.
3.1 The resurgence of semiotic hypotheses
This subchapter is devoted specifically to the ‘semiotic hypotheses’, as this branch is central
to the present dissertation. As mentioned, the discourse on Pleistocene art as ‘symbols’ led
scholars to investigate to what extent and in which ways the past markings might have been
symbolic. This paved the way for arguments that the symbol system was more intricate than
previously expected, and that the markings’ primary purpose might have been that of infor-
mation encoding (Płonka 2003: 177).
An intrinsic element of the semiotic approaches is that they necessarily fall under the cate-
gory of structuralism, which is “… based on the idea that [human] actions follows similar
laws to those governing linguistic structures” (Leroi-Gourhan 1968, as cited by; Moro Abadía
& Palacio-Pérez 2015: 663–64). With regards to prehistoric art, the approach was spear-
headed by Raphael, Laming-Emperaire and Leroi-Gourhan, and structuralist approaches
have since been adopted in various versions by several of the authors cited in this and the
previous chapter. As structuralism in essence attempts to recognise internal, structural order
(Layton 2006), the testing of a semiotic hypothesis necessitates a structuralist methodology.
The hypothesis that some Palaeolithic markings were in ways semiotic in fact originates in
those very first publications on the subject – specifically, Lartet & Christy’s (1875) report, in
which parallel lines were tentatively suggested to be tallies recording killed prey (d’Errico
1995: 163). Alexander Marshack (1972a; b, 1991a) revived these thoughts in the 1970’s and
80’s, his main claim being that Upper Palaeolithic markings were (in some cases) used for
time-keeping in the form of lunar calendars. He made the same proposition for an object in
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
29
this dissertations’ study area, namely the Maglemosian antler pick from Uggerløse (Mar-
shack 1991b: 358–59; Hayden 2021: 5)
Figure 3.1: The antler pick from Uggerløse, argued by Marshack to be calendric notation.
From Marshack (1991b: 358–59), edited by author.
Several scholars followed Marshack’s pioneering approach, mostly regarding Upper Palaeo-
lithic material (Dewez 1974; Tratman 1976; Dalmeri 1985), but Marshack’s work was also
heavily discussed and often harshly critiqued (Rosenfeld 1971; King 1973; Lewis-Williams et
al. 1988; d’Errico 1989; Robinson 1992; Davidson 1993; Elkins 1996; Płonka 2003: 180). In
part, the criticism concerned the methodology, which were primarily in the form of micro-
scopic analyses to clarify the direction and succession of individual engraved lines, and
whether the engravings were performed with one or several tools. But most importantly,
sceptics argued that Marshack’s hypothetico-deductive method rested on loose assumptions,
cherry-picking of data, and some irrelevant 19th-century ethnographic analogies (e.g., Mar-
shack 1985). Marshack, in turn, defended his findings (Littauer et al. 1974; Marshack &
D’Errico 1989). These discussions are a central piece of research history, in which no agree-
ment was achieved. While scholars within this field hardly ever reject the existence of weak
semiotic systems entirely, it is evident that research suggesting its existence is almost invari-
ably met with scepticism.
Interestingly, recent years have seen a marked re-emergence of semiotic approaches (Azéma
2008; Sauvet & Wlodarczyk 2008; Azéma & Rivère 2012; Rappenglück 2015; von Petzinger
2017; Miyagawa et al. 2018; Igarashi & Floss 2019; Taylor 2021; Bacon et al. 2023). Com-
mon for these recent studies is that they are concerned with Upper Palaeolithic European pa-
rietal art, all arguing that the markings in question were governed by some degree of internal
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
30
semiotic structure. These studies are in part a result of the accumulation of data (Bouissac
1994a: 350, b: 99–100). Compilations of databases (e.g., Dutkiewicz et al. 2020), including
the one that is applied for this dissertation (Płonka 2003), have made it possible to make in-
vestigations based on a larger empirical foundation. The development of computational
methods has also been instrumental in this resurgence, however more advanced computa-
tional methods have yet to be applied to prehistoric ornamentation. Moreover, methodologi-
cal frameworks (Overmann 2013; Hayden 2021) have given scholars the tools to argue more
strongly for the existence of Palaeolithic semiotic systems.
The most recent and perhaps most noteworthy example of the resurgence of semiotic hy-
potheses is the work of Bacon et al. (2023). This study proposes that lines, dots, and ‘Y’ signs,
in combination with an animal representation, are notational tallies that inform the viewer
of the lunar month in which the given animal gives birth. The analysis spans geographically
and temporally wide within the European Upper Palaeolithic, and therefore suggests that
this was long-lived communication system, and “the first known writing in the history of
Homo sapiens” (Bacon et al. 2023: 1). A comments to this very recent article has already ap-
peared (Magli 2023), critiquing both the statistical methods as well as the logical assump-
tions. It is uncanny how closely this present discourse resembles the one sparked by Mar-
shack’s work some 50 years ago. In the wave spearheaded by Marshack, the notational inter-
pretation was most often made for one or a few artefacts (Marshack 1991a; d’Errico & Cacho
1994; d’Errico 1995). In the most recent wave of semiotically oriented studies, which is invar-
iably based on quantitative analyses based on larger sets of data, the interpretation is pro-
posed more broadly. Similar, however, is the type of hypothetico-deductive reasoning; a
preexisting hypothesis is tested and demonstrated to fit the data. In this methodology, the
vast spectrum of other possible explanations is often hastily rejected or not dealt with at all.
This was the critique that Marshack had to defend against, yet the recent publications do not
actively engage in this debate between Marshack, Elkins, d’Errico and others in the 1970’s,
80’s, and 90’s. Instead, basic assumptions which ignore this unsettled debate are presented,
e.g.: “It is generally accepted that regular series of marks on Upper Palaeolithic material cul-
ture represent notational counting…” (Bacon et al. 2023: 3). Although it is encouraging to
see contemporary scholars’ willingness to engage with the semiotic hypotheses, it is a cause
of concern that unresolved issues and debates remain unaddressed.
On the basis of this and the previous chapter, fig. 3.2 summarises the research historical
developments regarding interpretations of Upper Palaeolithic European and South
Scandinavian Mesolithic (shown with green) ornamentation. Visualised are only the
publications mentioned in the above chapters which argue ‘strongly’ for a given
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
31
interpretative ‘school’. The vast majority of publications on this topic cannot be said to argue
strongly for one interpretation, and the plot is thus a result of an extensive literature review
on the author’s part. As this is a large research field, I do not claim this visualisation to be
exhaustive, but I do believe that it is representative for the general research historical
developments. It is important to emphasise that the different schools of interpretation are
not necessarily mutually exclusive. Arguments can be constructed that support the idea of,
e.g., magical or mythological meanings ordered within a semiotic system. Yet, such
arguments are rare (Larsson 2000: 33), as scholars tend to either support one ‘school’
strongly or consider the different interpretations as more or less equally likely. Fig. 3.2
demonstrates three things which are highly relevant for the present dissertation: firstly, it is
a visualisation of the resurgence of semiotic hypotheses within the last 15 years or so (the
bottom cluster in the ‘weak’ semiotic column), dominating the discourse on European Upper
Palaeolithic markings. The preceding wave of ‘weak’ semiotic proponents, led by Marshack
through the 1970’s, 80’s and 90’s, is also distinctly visible. Secondly, there is an absense of
ardent proponents for the ‘strong’ semiotic hypothesis. Though it has undoubtedly been
tested, no peer-reviewed publications (to the knowledge of the author) have presented strong
arguments and claims for a prehistoric writing system proper, which is implied by the
‘strong’ semiotic hypothesis. The last, and perhaps most relevant observation that fig. 3.2
offers, is the absence of proponents for the semiotic hypotheses regarding South
Scandinavian Mesolithic portable art. Recall Chapter 2.3.2 and note that this absence of
proponents is in fact because the semiotic hypotheses have not been tested for this material
at all.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
32
Figure 3.2: The publications mentioned in this dissertation which argue strongly for one
‘school’ of interpretations, numbered and ordered chronologically.
1: Grosse 1894, 2: Reinach 1903, 3: Munro 1906, 4: Cartailhac & Breuil 1906, 5: Müller
1918, 6: Hultén 1938, 7: Raphael 1945, 8: Laming-Empaire 1959, 9: Schlette 1969, 10: Mar-
shack 1972a, 11: Marshack 1972b, 12: Dewez 1974, 13: Tratman 1976, 14: Conkey 1978, 15:
Guthrie 1984, 16: Dalmeri 1985, 17: Halverson 1987, 18: Lewis-Williams et al. 1988, 19:
Marshack 1991, 20: d’Errico & Cacho 1994, 21: d’Errico 1995, 22: Lewis-Williams & Clottes
1998, 23: Lewis-Williams 1997, 24: Nash 1998, 25: Jensen 2001, 26: Nash 2001, 27: Azéma
2008, 28: Sauvet & Wlodarczyk 2008, 29: Porr 2010, 30: Djindjan 2010, 31: De Smedt &
De Cruz 2011, 32: Andersen 2011, 33: Azéma & Rivère 2012, 34: Rappenglück 2015, 35:
Sørensen 2017, 36: von Petzinger 2017, 37: Sauvet et al. 2018, 38: Miyagawa et al. 2018,
39: Igarashi & Floss 2019, 40: Vang Petersen 2021, 41: Taylor 2021, 42: Bacon et al. 2023.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
33
3.2 A semiotic hypothesis applied to the portable art of the South Scandina-
vian Mesolithic
Based on the above review of existing strands of theory, this dissertation’ theoretical angle is
outlined. The recent resurgence of semiotic hypotheses, encouraging yet problematic, pri-
marily engages with European parietal art of the Upper Palaeolithic, but the parietal art of
the Scandinavian Bronze Age has also been suggested to have functioned as a semiotic sys-
tem (Tilley 1991). South Scandinavian Mesolithic portable art – a subfield with a long re-
search history and a relatively abundant archaeological record – has not yet been subjected
to testing of these semiotic hypotheses. It is relevant to briefly reflect on the reasons for this
absence. The Upper Palaeolithic and Bronze Age markings have several things in common
with Mesolithic portable art. To name a few similarities: in all three cases, there are distinctly
different motif types; both figurative and geometric motifs occur, and these are often seen in
combination; and the ornamentation of all three periods are characterised by a sequence-less
organisation of motifs onto its medium (see fig. 3.3). By this token and considering that the
Mesolithic case lies chronologically in between the two others, it seems equally relevant to
test the semiotic hypotheses on the Mesolithic corpus. If the proponents of European Upper
Palaeolithic semiotic systems are right in their claims, then the Mesolithic – being of a
younger age – would conceivably also be utilising notational systems. Granted, the notion of
unilinear evolution is rightly in disfavour, and knowledge can be lost. However, as the Meso-
lithic is characterised by technological diversification and population increase, it follows logi-
cally that Mesolithic societies would be more likely to have a semiotic system than those of
the Upper Palaeolithic.
I will argue that there are three major reasons for the absence of semiotics-testing on Meso-
lithic material: firstly, the ‘art’ of other periods, e.g., the Upper Palaeolithic and Bronze Age,
is arguably more aesthetically pleasing than that of the Mesolithic, generally speaking. This
is of course also a subjective claim but consider fig. 3.3 below; these three examples are com-
parable regarding subject matter, all depicting a horse or deer in relation to other motifs. The
Mesolithic Ystad motif (the artefact is also shown in fig. 2.7, image 5) is quite representative
of Mesolithic engravings in general, being more disorganised and ‘haphazard’-looking than
what is known from the other periods. Thus, the subjective preferences of modern viewers
probably affect which material has been most intensely investigated. Secondly, pertaining to
Upper Palaeolithic example, the sensationalism of ‘the oldest art’ or ‘the oldest writing’ has
likely resulted in the earlier material receiving more attention. The third and final reason
why the semiotic hypotheses remain untested for the Mesolithic material is the fundamental
differences in the practice of archaeological research conducted by scholars in South
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
34
Scandinavia and Continental Europe, respectively. There are not many actively researching
Danish experts in the South Scandinavian Mesolithic. Those that have specifically concerned
themselves with the period’s portable art are even fewer in number. Though these few ar-
chaeologists have contributed greatly to the field, they have evidently not adopted the ap-
proaches developed by their continental colleagues, neither with regards to theoretical
frameworks nor the application of data-driven approaches. As mentioned, the three most
ambitious and holistic contributions have been made by Clark, Nash & Płonka, all of whom
have non-Scandinavian academic backgrounds.
Figure 3.3: A comparison between the markings from three different periods of prehistory.
1: Horse from Lascaux, Upper Palaeolithic (Pruvost et al. 2011: 18629), 2: Ystad, Mesolithic
(Larsson 2017: 64), and 3: detail of Swedish Bronze Age rock art (Kristiansen 2012: 81)
Bouissac’s (1994a) seminal article ‘Art or Script? A Falsifiable Semiotic Hypothesis’ consti-
tutes a central part of the theoretical foundation of the present dissertation. Specifically, two
aspects of Bouissac’s text are important:
o The distinction between a ‘weak’ and a ‘strong’ semiotic hypothesis. In the weaker
version, the notion of symbolism is retained – motifs/visual designs reference real-
world phenomena, about which information is encoded. In a ‘strong’ semiotic system,
the motifs have lost their iconicity and function more in the sense of written language
as we use it today (Bouissac 1994a: 353). It is here important to note that the above-
mentioned attempts to understand prehistoric markings invariably belong to the
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
35
‘weak’ semiotic category, as is also visualised in fig. 3.2 – be it Marshack’s lunar cal-
endars or Bacon et.al.’s phenological ditto.
o The argument that semiotic structure, especially the ‘strong’ variant, should ideally
always be considered first, as it is the easiest to falsify. Though an archaeological rec-
ord will always be fragmentary, the present data availability possibly allows for test-
ing of semiotic structure. Bouissac proposes:
“…first to identify semiotic boundaries across both space and time; second, to construct
types from the systematic study of morphological variations; and third, to heuristically
scan strings of signs in whatever directions sequences can be construed, in search of recur-
ring combinations of types.” (Bouissac 1994a: 365).
As has been demonstrated in the previous chapter, the South Scandinavian region is within
the Mesolithic to be regarded as a ‘contextual area’, significantly interconnected, with many
shared culture traits. A testing of semiotic structure for the portable art of this region may
therefore be particularly relevant. As many of my peers within this field, I remain sceptical of
the validity of existing semiotic interpretations, but – following Bouissac – recognise the ne-
cessity of testing the semiotic hypotheses explicitly. The analytical design of the present dis-
sertation is therefore to work with the null hypothesis that there exists no significant semi-
otic relationship between the motifs, and that the motifs were therefore not a part of a semi-
otic information system.
o H0 = Null hypothesis: analysis reveals no internal, semiotic structure. The demon-
stration that the Mesolithic motifs are entirely independent of one another would
strengthen the likelihood of alternative interpretations for this region, which may
then be discussed on the basis of the EDA.
o H1a = The presence of ‘weak’ semiotic structure: some motifs are co-occurring to a
statistically significant degree. Such statistical significance should be further qualified
by other observations, such as spatial patterning, coherence regarding artefact type,
or similar factors. This would strengthen the argument that the motifs’ function is to
be understood in combinations with each other, as parts of some notational system
referencing real-world phenomena.
o H1b = The presence of ‘strong’ semiotic structure: all, or most, motifs can be demon-
strated to function within a semiotically structured system (perhaps with synchronic
or diachronic variations), strengthening the argument that the motifs are a part of a
(proto-)writing system.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
36
3.2.1 Theoretical assumptions
This section outlines how this dissertation’s hypotheses can be assessed. At the most basic
level, it is assumed that the cognitive capacities for developing semiotic systems were present
by at least Upper Palaeolithic times, and therefore unquestionably also in the early to middle
Holocene. Neither a negative nor a positive result of the following hypothesis testing will
therefore reveal anything about the cognitive capacities of the creators of the ornamentation,
but rather whether any societal need for a semiotic system was present at the time.
Some underlying assumptions regarding the Weak Semiotic hypothesis (H1a) can be derived
from Hayden’s (2021) recent attempt at categorising types of ‘record keeping’ (roughly
equivalent to Bouissac’s weak semiotic category) and their material correlates. Relying on
ethnographic observations, he argues that diagnostic features can indeed be discerned and
provides a checklist of how record keeping systems in prehistoric societies can be qualified. A
central point of interest for this dissertation is what he labels “Different marks and adjunct
symbols” (Hayden 2021: 8). He argues that, for instance, a distinct motif in association with
uniform marks may be a symbol and its count. From this can be derived a theoretical as-
sumption which is applied in this dissertation: if two motifs co-occur to a significant degree,
this might be an indicator of the marks being consistently used in a form of weak semiotic
system, as ‘record keeping’ or something similar. For the testing of the weak semiotic hy-
pothesis, the general assumption is that only some of the motifs had a semiotic function, and
that these may be found through co-occurrence analysis and subsequent discussions regard-
ing other factors, such as the spatio-temporal extent of the phenomenon. This approach does
not consider the possibility that a semiotic purpose existed for one specific object, e.g., that
one Mesolithic person decided to count lunations. Such a scenario is not inconceivable but
demonstrating this is perhaps too difficult a task in any case, for the reasons listed in the fol-
lowing paragraph. This dissertation therefore assumes potential semiotic structure to be a
societal phenomenon, collectively practiced by more than a single or a few individuals. This
sidesteps the issue that Marshack was criticised for, namely selecting for analysis only those
artefacts which offhand look semiotic to the modern investigator.
Hayden also notes ethnographic instances of tally sticks and similar notational objects that
are not used in combination with adjunct symbols. In most cases, this is in the form of one or
more series of parallel lines, which is indeed one of the most common motif types in the
Mesolithic South Scandinavian data. This type of notation must necessarily be investigated
through rather arduous methods, such as counting the number of lines on a given artifact.
This manual method is not attempted here, as the aim of the present dissertation is a
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
37
quantitative and data-driven contribution. It might, however, be relevant, as such counts
could potentially reveal interesting patterns, and as this method has been applied by some of
the studies mentioned above (e.g., Marshack 1991b; Bacon et al. 2023). However, I would ar-
gue that such manual counting will invariably be an ambiguous and subjective exercise; pre-
historic ‘art’ objects, parietal or portal, are more often than not taphonomically affected (i.e.,
engravings are worn down) and without clear structuring (as exemplified in fig. 2.7). There-
fore, arguing that one has made a truly objective count of the lines on such an artefact is
problematic. I would argue that the testing of inter-coder consistency is needed for assessing
the validity of such a count (as in Dutkiewicz et al. 2020: 11), and this is too extensive a task
for the present dissertation. Yet, it should ideally be tested, and is therefore relevant for fu-
ture studies of the South Scandinavian Mesolithic material. In summary, the possibility that
an artefact with only one motif, such as parallel lines (A1), was ‘weakly’ notational (in that
their original purpose was that of a ‘tally stick’, as exemplified by Hayden (2021: 4) cannot be
refuted on the basis of this dissertation. However, the presence of adjunct symbols, one vari-
ant of weak semiotic system, will be tested in the following analyses.
Concerning the ‘strong’ semiotic hypothesis (H1b), a purely quantitative approach such as
that of the present dissertation will potentially produce more unambiguous results. In this
case, we may generally assume that the whole corpus, or most of it, functioned as a semiotic
system. This is where the quantitative data accumulation, as presented in the following chap-
ter, is not only beneficial but necessary. The theoretical assumptions for demonstrating
‘strong’ semiotic structure, following Bouissac’s definition, are: 1) statistically significant tri-
ple or quadruple combinations of motifs, as this can arguably be regarded as the minimum
for ‘non-symbolic’ proto writing. 2), these should, as with the ‘weak’ semiotic hypothesis, be
qualified by other observations, such as spatio-temporal patterning or other consistencies.
Yet, it must be underlined that a demonstration of significant motif co-occurrences would
not in itself qualify the semiotic hypotheses. Such significant co-occurrences would only indi-
cate non-randomness, and not the underlying reason. The findings will therefore need to be
synthesised in the discussion with the inclusion of other relevant observations.
In the case of an acceptance of H0 (that no semiosis can be discerned), this dissertation can
be considered a critical response to the ‘semiotic resurgence’. In this scenario, the question of
the function of South Scandinavian Mesolithic portable art remains unanswered, but at least
a case for a dismissal of the semiotic hypotheses will have been made for this specific area of
study. If the analyses on the other hand allow for a rejection of H0, the dissertation lends
support to and places itself amid the resurgence of semiotics in prehistoric art research.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
38
Chapter 4
Materials & Methods
Forming the first part of the analysis, descriptive statistics are presented, giving basic sum-
maries of the compiled data. The descriptive statistics are subdivided into two sections: 1),
‘Representativity’, forming the basis from which representational issues can be evaluated,
and 2), ‘Descriptive statistics for the individual motifs’, quantifying and visualising Płonka’s
motif categories across the relevant categorical variables. Subsequently, the methodology for
discerning semiotic structure is presented in Chapter 4.2. All analyses were performed in the
open-source R programming environment (v.4.2.2; R Core Team 2021); specifically the anal-
yses concerning skipgrams were executed using the Quanteda package (Benoit et al. 2018,
for a detailed description, see R: Create n-grams and skip-grams from tokens. 2023). All
code is attached in Supplementary Material 2 and 3. Maps were produced using QGIS
(v3.22.3, Białowieża), and image editing was performed in GIMP (v2.10.30). All analyses
were thus performed in open-source and free accessible software environments and all data
used are published alongside this dissertation.
4.1 Utilising Tomasz Płonka’s catalogue
The analyses performed in the present dissertation are based on the catalogue compiled by
Tomasz Płonka (2003) in ‘The Portable Art of Mesolithic Europe’. Płonka’s monograph is
concerned with the European Mesolithic in its entirety. To utilize Płonka’s catalogue for the
purposes of the present dissertation, the information for artefacts deriving from the study
area were manually transcribed into Microsoft Excel. These data (Supplementary Material 1)
consist of several categorical variables, as well as binary data in the form of presence/ab-
sence coding of motif types on each artefact. Some of Płonka’s categorical variables have
been modified, as described in the following, to fit the purposes of the present dissertation.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
39
4.1.1 Temporal information
Płonka chronologically divides his data into chronozones. The large geographical area that he
is concerned with, however, has different dates for these chronozones, and he concedes
(Płonka 2003: 12) that the divisions are somewhat arbitrary. For the purpose of this disserta-
tion, the chronozone division is translated into the corresponding archaeological cultures of
South Scandinavia reviewed in Chapter 2.1.1. Płonka sets his Boreal/Atlantic division to
6900/6500 cal BC (ca. 8900-8500 BP; (Płonka 2003: 12). This corresponds well to the cul-
tural chronology applied in this dissertation, namely that the Maglemose/Kongemose transi-
tion is set to approximately 8400 BP. Thus, Płonka’s Boreal/Atlantic and all earlier subdivi-
sions are attributed to the Maglemose culture. Płonka also subdivides the Atlantic
chronozone into early and late. It is not clearly defined where this division is set, but from
the well-dated artefacts, a dividing line in calendrical years can be pinpointed: the antler axe
from Argus Bank has the latest date range among the artefacts in the ‘early Atlantic’ group, at
6010-5560 cal BC. The earliest dates in Płonka’s ’late Atlantic’ group are the four antler ob-
jects from Helsingør (Bergmansdal), set to 5570-5380 cal BC. The division is thus approxi-
mately around the 5500s BC (7500 BP), which also corresponds closely enough to the
Kongemose/Ertebølle transition at around 7400 BP. In summary, Płonka’s ‘early’ and ‘late’
Atlantic translates to the Kongemose and Ertebølle culture complexes, respectively, while
everything older is Maglemosian in association One issue appears, however: some typologi-
cally dated artefacts in the catalogue are only broadly dated to the Boreal or Atlantic. Those
ascribed to the Boreal can be translated as Maglemosian, but for those ascribed broadly to
the Atlantic, the attribution to either Kongemose or Ertebølle remains uncertain. Thus, a
fourth category, labelled ‘Kongemose/Ertebølle’ is necessary. The translation is visualised in
table 1 below.
Table 1: This dissertation’ translation of Płonka’s chronological divisions.
The colour attributions on the right are consistently used throughout.
Płonka’s chronological divisions
This dissertations’ translation
PB
Preboreal
Maglemose
PB/BO
Preboreal/Boreal transition
BO1
Early Boreal
BO
Boreal
BO1/BO2
Early/Late Boreal transition
BO2
Late Boreal
BO/AT
Boreal/Atlantic transition
AT1
Early Atlantic
Kongemose
AT
Atlantic
Kongemose/Ertebølle
AT1/AT2
Early/Late Atlantic transition
AT2
Late Atlantic
Ertebølle
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
40
As addressed in Chapter 2.2, chronological attribution is a general issue for this artefact
group. As an example, an antler axe from Langø is described in the original text (Broholm
1928: 165) as Maglemosian, but catalogued by Płonka as belonging to the Late Atlantic phase
(corresponding to Ertebølle (Płonka 2003: 258). The reasons for such incongruities are
rarely transparent, and it is very probable that at least some of the objects in this disserta-
tions’ data are incorrectly dated. Vang Petersen (2013) has argued for a revised date for some
of the amber figurines which are also included in Płonka’s catalogue. He argues that these
are Late Glacial rather than Maglemosian in date, but once again, on typological grounds.
Well-dated objects do, however, also exist. Płonka (2003: 18) categorises the quality of con-
textual information for each object (table 2). Categories 1 (“Grave finds or finds dated di-
rectly using the radiocarbon method”) and 2a (“Finds recovered during excavation work …
on sites dated by pollen samples and radiocarbon method”) are the only ones not dependent
on typological dating. Recent studies (von Petzinger & Nowell 2011; Matzig et al. 2021;
Tsirintoulaki et al. 2023) have demonstrated that chronological assumptions based on typo-
logical traits are often faulty, and this is likely to be the case for portable art motifs as well.
This dissertation does not engage with the debates concerning typological re-dating of spe-
cific artefacts, but instead, for the sake of consistency, retains the dates as attributed by
Płonka. Duly mindful of the uncertainties of typological dating, objects in categories 1 and 2a
are in the following analyses categorised as having ‘good’ contextual information.
Table 2: This dissertations’ translation of Płonka’s categories of contextual information
quality.
4.1.2 Spatial information
The coordinate for each individual artefact is not given by Płonka, so this information was
retrieved and catalogued manually. This was possible through the use of the Danish ‘Fund &
Fortidsminder’ data base, as well as the source literature indicated in Płonka’s catalogue. In
several cases, a single archaeological site has yielded more than one ornamented artefact.
The 483 artefacts derive from 240 sites; many are excavated archaeological sites, while oth-
ers refer to lakes, bogs etc., where a given object was recovered with no further find context.
Płonka catalogues the artefacts in regions, but in his analysis generally suffices to consider
Płonka’s categories of contextual information quality
Translation
1
Grave finds or finds dated directly using the radiocarbon method
Good
2a
Finds recovered during excavation work on sites dated by palynology and C14
Good
2b
Finds recovered during excavation work on sites dated by the typology of finds
Poor
3
Surface finds from sites with deposit of a single archaeological culture
Poor
4
Surface finds from sites with deposit of more than one archaeological culture
Poor
5
Stray finds lacking archaeological context
Poor
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
41
the large, analytical divisions of Europe (with South Scandinavia belonging to the ‘Northwest
Province’). This dissertation utilises different analytical division of South Scandinavia (fig.
4.1). Based on the research historical review given in Chapters 2.2 and 2.3, the four subdivi-
sions ‘East Denmark’ ‘West Denmark’, South Sweden’, and ‘North Germany & Poland’ is ap-
plied as analytical units in the descriptive statistics.
Table 3: This dissertation’s translationof Płonka’s regional divisions.
Figure 4.1: map of this dissertations’ spatial analytical units.
4.1.3 Coding the motifs
Płonka’s classification consists of 256 motif types, divided into 9 groups (groups ‘A’ to ‘I’,
Supplementary Material 7). Of these, 230 are present in the South Scandinavian material.
For all artefacts, binary presence/absence coding of these motifs was performed. This binary
table forms the basis of many of the following analyses. The 230 motifs included in the data
are all geometric, with the exceptions of the categories anthropomorphic, zoomorphic, and
relief. These three categories are quite distinct, firstly because ‘relief’ is really a technique ra-
ther than a type of motif, and secondly because the anthropo- and zoomorphics could be
considered umbrella terms for different variants, i.e., representations of distinct animals.
These three categories are nonetheless retained, and the low frequency of these three motif
types does not warrant further subdivision. Sculpted figurines depicting animals are placed
in this zoomorphic category along with two-dimensional engravings. In a few instances,
Płonka adds a question mark when attributing a motif to an artifact, to stress that this
Płonka’s regions
Translation
Bohuslän
South Sweden
Öland
Scania
Västergötland
Amager
East Denmark
Lolland-Falster
Møn
Zealand
Funen
West Denmark
Jutland
Langeland
Mols
Mors
Vendsyssel
Mecklenburg
North Germany
& Poland
Pomerania
Rügen
Schleswig-Hol-
stein
(Unknown)
No Context
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
42
interpretation is somewhat dubious. Question marks also occur in the ‘Piece’ column, which
denotes the artefact type on which the ornamentations occur. In the following analysis, this
category is modified slightly, attributing the values ‘-‘ and ‘?’ to the ‘fragment’ category,
meaning objects that cannot be attributed to an artefact class. In all instances where a ‘?’ is
added to a motif attribution, the question mark is disregarded.
Finally, do recall that at just over twenty years old, Płonka’s catalogue could benefit from an
update, as a number of artefacts have since been recovered and recorded – both from well-
excavated sites such as Tågerup (Karsten & Knarrström 2003: 102–18), Syltholm (Sørensen
2016: 6, 2020: 407), and Hjarnø (Skriver et al. 2018: 211–12) as well as loose finds tenta-
tively suggested to be Mesolithic in date (e.g., Borup 2008). The exact number of such addi-
tions to the Płonka catalogue is hard to quantify, however, and a comprehensive update of
this catalogue – desirable as it is – lies beyond the scope of this dissertation.
4.2 A methodology for discerning semiotic structure
Following the descriptive statistics, different methodological approaches are applied for the
analyses concerning semiotics. The central question is: what methodological approach is
needed to falsify the weak and strong semiotic hypotheses respectively? The methodology
follows the logical order proposed by Bouissac (1994a: 365): The first step, “to identify semi-
otic boundaries across both space and time” has been taken, in the sense that the South
Scandinavian region has been demonstrated to qualify as a contextual area. The second, “ to
construct types from the systematic study of morphological variations”, has been per-
formed by Płonka, whose classification is adopted here. The third, and last, of Bouissac’s
steps, “to heuristically scan strings of signs in whatever directions sequences can be con-
strued, in search of recurring combinations of types”, is what this part of the analysis at-
tempts. Bouissac does not specify further how this ‘heuristic scan’ ought to be designed, as
there is evidently a myriad of potentially relevant approaches, depending on the data in
question. In this dissertation, based on the theoretical framework concerning the semiotic
approaches to prehistoric art, a methodology derived from the broader field of linguistics
seems relevant. This dissertation tests the weak and strong semiotic hypotheses, respec-
tively, by performing k-skip-n-gram analyses at different levels, incorporating a custom
measure of statistical significance. As k-skip-n-grams is a method far removed from what is
normally applied in archaeological data analysis, an explicit description is required.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
43
4.2.1 K-Skip-N-Grams – solving the issue of nonlinear corpus data
N-grams, used traditionally in computational linguistics, natural language processing, DNA
sequencing, as well as other domains (El-Nasan et al. 2003; Tomović et al. 2006; Xu et al.
2019) is a method of identifying adjacent sequences of words (or other types of items). Used
for example in the Google NGram Viewer, it can probabilistically model n number of contig-
uous (adjacent) items and identify their frequencies of usage.
By identifying n-grams in a given corpus of text, it might be demonstrated that, e.g., ‘happy
holidays’ is an unlikely common 2-gram, indicating the presence of a specific concept. Such
statistically significant co-occurrences are referred to as ‘collocations’ (Gries 2017: 15–17). N-
grams are, however, bound by the limitation that the method only detects sequences contigu-
ously, i.e., in linear text (Sidorov 2013; Andresen & Zinsmeister 2017: 4). This dissertations’
data of interest has no such linearity, as motifs on an artefact are placed in such ways that
any sequential ordering is indiscernible to the contemporary interpreter (see, for example,
fig.2.7). Płonka’s – and consequently, this dissertations’ – motif cataloguing is therefore
simply ordered alphabetically after the motif name (A1, A2, etc.). Consequently, a pure n-
gram approach to this material would generate misleading results, as the adjacency of motifs
would only reflect the alphabetical cataloguing system.
This issue can effectively be solved through ‘k-skip-n-gram’ analysis. K-skip-n-grams, or
skipgrams, are variants of n-grams with the additional possibility to skip k items. The k rule
addition has the main purpose of circumventing data sparsity, as described by Guthrie et.al.
(2006), the seminal text concerning this method. For the purpose of the present dissertation,
however, the method can effectively fix the problem of dealing with non-sequential data
structures. Skipgrams allow n-grams to be formed by skipping words up to a certain limit, k.
Thus,“4-skip-n-gram results include 4 skips, 3 skips, 2 skips, 1 skip, and 0 skips (typical n-
grams formed from adjacent words)” (Guthrie et al. 2006: 1222). By allowing these skips,
Examples of n-grams:
“John kicks the green ball”
1-grams: (John), (kicks), (the), (green), (ball)
2-grams: (John, kicks), (kicks, the), (the, green), (green, ball)
3-grams: (John, kicks, the), (kicks, the, green), (the, green, ball)
4-grams: (John, kicks, the, green), (kicks, the, green, ball)
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
44
further n-grams are formed which might reveal syntactical structures that traditional n-gram
models might not capture – such as the 2-gram (kicks, ball) in the below example.
As mentioned, the individual items need not be separate words, but can be letters, syllables
or other things. It can also be applied to prehistoric motifs, although to the author’s
knowledge this has not been previously attempted. In treating individual motifs as items and
each ornamented artefact (bearing two or more motifs) as distinct ‘sentences’, this method
can effectively assess the frequency of co-occurrences of not only two, but three or more mo-
tifs. For the purposes of the present dissertation, the k value (skip distance) needs to be set
high enough for all motifs on each respective artefact to be paired into a skipgram. As the
highest number of individual motifs on a single object is 15 (the antler axe from Værebro Å,
shown in fig. 1.1), the k value for all skipgram analyses is consistently set to 13. This is dis-
similar to the normal usage of skipgrams, such as in analyses of modern language corpuses,
where it would rarely make sense to work with k values set to more than four; in any modern
language, word proximity correlates with these words’ mutual relations (Guthrie et al.
2006). Thus, generating 18-skip-2-grams from a modern language 20-word sentence would
provide all possible bigrams of the sentence – however, this would likely have no meaningful
value, as too many syntactically meaningless and thus irrelevant bigrams would be gener-
ated. This is not the case when applying the method to prehistoric, non-sequential motifs,
since (as explained above) ‘word’ proximity only reflects the alphabetical cataloguing. Fig.
4.2 is a visualisation of this analytical process, with Płonka’s motif attributions providing the
basis for generating skipgrams. For generating 3-grams, only artefacts with three or more
motifs are applicable (which corresponds to 168 objects out of the total 483, see table 4).
Similarly, at least four motifs per artefact are required for 4-gram analysis. In this way, the
data set becomes smaller at each analytical step.
Examples of k-skip-n-grams:
“John kicks the green ball”
1-skip-2-grams: (John, kicks), (John, the), (kicks, the), (kicks, green), (the, green), (the, ball), (green, ball)
2-skip-2-grams: (John, kicks), (John, the), (John, green), (kicks, the), (kicks, green), (kicks, ball), (the, green), (the, ball),
(green, ball)
3-skip-2-grams: (John, kicks), (John, the), (John, green), (John, ball), (kicks, the), (kicks, green), (kicks, ball), (the, green),
(the, ball), (green, ball)
2-skip-3-grams: (John, kicks, the), (John, kicks, green), (John, kicks, ball), (John, the, green), (John, the, ball), (John,
green, ball), (kicks, the, green), (kicks, the, ball), (kicks, green, ball), (the, green, ball)
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
45
Table 4: The skipgram levels performed in the analysis
Figure 4.2: Płonka’s motif attributions, exemplified by a section of the antler pick from Ka-
lundborg. The skipgrams (at the level of two-motif co-occurrences) generated from this ex-
ample are shown at the bottom. From Płonka (2003: 503), redesigned by author
4.2.2 O/E ratio – a custom measure of statistical significance
As explained above, this dissertation’s application of n-gram analysis is fundamentally differ-
ent to how this method is usually applied, due to the fundamentally different nature of the
data in question. Therefore, a custom measure of significance is designed for the purposes of
this analysis, on the basis of Gries’ (2017: 208–28) practical examples of co-occurrence anal-
ysis in corpus linguistics. The observed motif co-occurrence values are compared to the ex-
pected values, which are calculated in the following way:
Skipgram analyses performed
Model
Purpose
For testing hypothesis
# of objects applicable
13-skip-2-grams
Co-occurrences of two motifs
Weak Semiotic
265 (55%)
13-skip-3-grams
Co-occurrences of three motifs
Strong Semiotic
168 (35%)
13-skip-4-grams
Co-occurrences of four motifs
116 (24%)
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
46
P(a) = a / T
P(b) = b/ T
P(a|b) = P(a) * P(b)
E = P(a|b) * T
In the above formula, T is the total number of objects (in this case, a constant of 483),
a is the individual frequency of one of the two motifs in the given co-occurrence,
b is the individual frequency of the other motif in the given co-occurrence,
P(a) is the individual probability of one of the co-occurrence’s motifs occurring,
P(b) is the individual probability of the other of the co-occurrence’s motifs occurring, and
E is the expected frequency for the specific co-occurrence.
The observed frequency of the given co-occurrence, O, is divided by the expected frequency,
E, and this produces a ratio, O/E. If the observed frequency is equal to the expected fre-
quency, then the ratio will be 1. If the observed frequency is lower the expected, the O/E ratio
is less than one. The higher O is, compared to E, the more unlikely is it that the given colloca-
tion reflects randomness. This statistical measure therefore allows for recognising whether
any (and which) motif combinations on the ornamented Mesolithic objects are statistically
likely to have been a meaningful collocation to the Mesolithic engravers. In sum, the O/E ra-
tio is calculated as follows:
O/E = O / ((a / T) * (b / T) * T)
The formula is exemplified using an example from the corpus: Motifs C1 and B10 co-occur 13
times. C1 occurs a total of 125 times, and B10 occurs 50 times. The O/E ratio of the co-occur-
rence of C1 and B10 is then calculated in the following way:
P(a) = a / T = 125 / 483 = 0.259
P(b) = 50 / 483 = 0.104
P(a|b) = P(a) * P(b) = (125 / 483) * (50 / 483) = 0.027
E = P(a|b) * T = ((125 / 483) * (50 / 483)) * 483 = 12.9
O / E = 13 / (((125 / 483) * (50 / 483)) * 483) = 1.0
Thus, the probability of C1 and B10 co-occurring is 0.027 or 2.7%, and the expected number
of co-occurrences of C1 and B10 is 12.9. Dividing the observed frequency, 13, by the expected
frequency gives us the O/E ratio of 1.0. In this example, the value is almost exactly 1, which
tells us that the observed frequency is exactly what would be statistically expected.
In the case of higher-order n-gram co-occurrences, additional motif probabilities, P(c) and
P(d) are simply added to the equation. Note that in the calculations of O/E ratios for higher-
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
47
order n-grams, the expected values will be diminishingly small. There is a significant degree
of uncertainty when calculating the statistical significance of observations when the expected
frequency is this low. The O/E ratio values are therefore only meaningful in the sense that,
the higher the O/E ratio, the more interesting the collocation. In other words, for higher-or-
der n-grams, the threshold of 1 is not applicable. O/E ratios were calculated for all generated
skipgrams at the n levels of 2, 3, and 4 (co-occurrences of two, three, and four motifs). The
corresponding data tables for each of these levels are available as Supplementary Material 4,
5, and 6.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
48
Chapter 5
Analysis & Results
5.1 Descriptive Statistics
In the first part of the analysis, the five categorical variables ‘Context’, ‘Chronology’, ‘Region’,
‘Material’, and ‘Artefact Type’ are considered. Then, the individual motifs, as designated by
Płonka, are quantified, and charted across those same five variables. These descriptive statis-
tics are followed by a testing of the ‘weak’ and ‘strong’ semiotic hypotheses respectively. This
subchapter is an exhaustive exploratory analysis, designed to find general patterns and po-
tential biases in the data. Bear in mind, however, that the dissertation’s primary purpose,
testing whether weak or strong semiotic structure can be inferred from the data, is addressed
in section 5.2.
5.1.1 Representativity
Approximately two thirds of the objects (317/483) are from a ‘poor’ context, i.e., they are not
grave finds or from excavations dated through radiocarbon or palynology. Note that some of
these artefacts with ‘poor’ context may in fact be safely anchored in time, e.g., in the case that
they were excavated from layers with diagnostic lithics. Still, it is necessary to remain aware
of the uncertainties regarding chronological attributions. Table 1 shows how many orna-
mented artefacts are known from a given site (a ‘site’ encompassing both archaeologically ex-
cavated sites as well as locations for loose finds). Sites with only one ornamented object
make up ≈ 77%, and sites with several objects must be said to be very rare. Among the sites
with the highest concentration of ornamented objects are Kongemosen (n=18), Ageröd 1
(n=25), Øgårde (n=25) and Sværdborg 1 (n=31). While the objects from Kongemosen are at-
tributed to the Kongemose culture, the other three are Maglemosian. Concerning the chrono-
logical representativity (fig. 5.1), the data is spread as evenly as one might have expected,
given that the Kongemose covers a shorter time span and has been studied less intensely.
The chronologically uncertain group ‘Kongemose/Ertebølle’ is quite small but recall that the
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
49
chronological uncertainties extend beyond this group, given the large share of portable art
objects with poor contextual information.
Table 5: Number of ornamented artefacts per site
Figure 5.1: The chronological distribution of ornamented artefacts across the three archae-
ological cultures
The following three maps (figs. 5.2 -5.4, see also fig. 5.5) show the spatial distribution of the
ornamented artefacts for each culture period. The spatial distribution is rather heavily cen-
tered on East Denmark, specifically Zealand and, even more specifically, the Åmose area of
Western Zealand. This is especially true for Maglemosian and Kongemosian objects. Regard-
ing Ertebølle, a large number of objects appear along the coast of Central Eastern Jutland.
The dots with bold circumferences denote sites with good contextual information – note in
this regard that in Western Denmark, there are no ornamented objects with good context
that are attributed to the Maglemose or Kongemose complexes.
# of artefacts per site
Artefacts per site
Count
1
178
2
19
3
13
4
7
5
3
7
2
8
1
9
3
10
2
18
1
25
2
31
1
Total # of sites
232
Average # per site
2.1
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
50
Figure 5.2: Spatial distribution of ornamented artefacts attributed to the Maglemosian
Figure 5.3: Spatial distribution of ornamented artefacts attributed to the Kongemosian
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
51
Figure 5.4: Spatial distribution of ornamented artefacts attributed to the Ertebølle culture
Figure 5.5: Bar chart showing the overall regional distribution of ornamented artefacts
The figure below (fig. 5.6) shows that the material bearing ornamentations is predominantly
antler and bone, but with amber having a significant frequency as well. Flint and other stone
types are sparsely represented, especially considering the vast amount of lithic material
known from this period. Wooden objects are undoubtedly underrepresented due to tapho-
nomy.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
52
Figure 5.6: Bar chart showing the frequencies of different materials bearing ornamenta-
tion
As for the ‘Artefact Type’ variable (fig. 5.7), heterogeneity is the key word; a broad spectrum
of the artefact classes characteristic of Mesolithic technology has been ornamented occasion-
ally. By the same token, it seems clear that (antler) axes have had a special role. Of the 39
categories of artefacts many only occur once or a few times.
Figure 5.7: Bar chart showing the frequencies of different artefact types bearing ornamen-
tation
In the following, all bivariate combinations of the five variables are presented. Note that sev-
eral plots in figs. 5.8-5.11 show the same data but with switched axes. Regarding fig. 5.8:
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
53
o Context by Chronology: Regarding objects with ‘good’ contextual information, we see
that Kongemose is underrepresented. The distribution of ‘poor’ context objects shows
a larger share of Ertebølle, suggesting that many loose finds are attributed to the Er-
tebølle phase.
o Region by Chronology: Maglemose dominance in East Denmark, likewise for Erte-
bølle in West Denmark. Kongemose ornamentation is almost exclusively an Eastern
Danish and Southern Swedish phenomenon (corresponding to the general core
Kongemose area).
o Material by Chronology: Antler is slightly more used during the Ertebølle, and bone
more frequent in Maglemose, but not to an overwhelming extent. Ornamented wood
is yet only known from Ertebølle contexts.
o Artefact Type by Chronology: This plot firstly shows that artefact repertoires
changed to some degree during the Mesolithic; e.g., mattock-heads are primarily a
Maglemosian phenomenon, and ulna daggers were predominantly produced during
the Ertebølle (Płonka 2003: 39, 109). Axes were more frequently ornamented in the
Fig. 5.8: Grid of bar charts showing chronology frequencies across the other four variables.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
54
Ertebølle, whereas shafts and pendants look to have been ornamented to an equal ex-
tent throughout the Mesolithic. It is interesting to note that harpoon heads seem to
have been ornamented almost exclusively in the Ertebølle period. However, perhaps
too few ornamented harpoon heads are known for this observation to be unambigu-
ous.
The following observations can be made regarding the small multiples of the spatial variable
‘Region’ (fig. 5.9):
o Context by Region: Western Denmark has the highest number of objects with poor
context, which makes sense in tandem with the ‘Context by Chronology’ chart; West-
ern Denmark has a strong presence of Ertebølle objects, in part due to many loose
finds being attributed to this phase. South Sweden has the highest proportion of well-
dated objects.
o Chronology by Region: East Denmark is strongly correlated with Maglemose, while
Fig. 5.9: Grid of bar charts showing region frequencies across the other four variables.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
55
West Denmark is dominated by Ertebølle objects.
o Material by Region: A roughly even distribution, except for the following: orna-
mented amber is not known from Sweden, which on the other hand has the majority
of flint and (other) stone. Ornamented wood is only known from Western Denmark –
however, the recent excavations at Syltholm (Eastern Denmark) have documented at
least one piece of ornamented wood (Sørensen 2020: 407), and the Western Danish
majority may have more to do with preservation conditions than anything else.
o Artefact Type by Region: All of the most frequently occurring artefact types (from
axes to harpoon heads) are distributed evenly across the regions. From there on, re-
gional variations can be discerned: figurines are only known from Western Denmark
and Northern Germany & Poland; long bones, picks and sleeves are nearly only orna-
mented in Eastern Denmark; and (flint) cores, flakes, and plates are almost exclu-
sively a Swedish phenomenon.
Fig. 5.10: Grid of bar charts showing material frequencies across the other four varia-
bles.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
56
Regarding the material variable (fig. 5.10, previous page):
o Context by Material: Roughly equal distributions, except in the case of amber, which
as mentioned is nearly always loose finds.
o Artefact Type by Material: Unsurprisingly, axes and shafts are exclusively made of
antler, pendants are nearly always of amber, daggers are of bone, etc.
The final figure of this series of bivariate visualisations (fig. 5.11) shows the artefact type
across the other four variables. Compared to the previous ones, this figure is rather more
chaotic, but this clearly demonstrates that across the entire South Scandinavian Mesolithic,
the selection of artefacts for ornamentation does not seem to have been of great importance.
Rather, all objects used in everyday activities could potentially be ornamented. Antler axes
from the Kongemose and Ertebølle periods have significant relative frequencies, however,
echoing what has been demonstrated above, namely that these were quite frequently orna-
mented in the late Mesolithic.
Fig. 5.11: Grid of bar charts showing chronology frequencies across the other four varia-
bles.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
57
5.1.2 Descriptive statistics for the individual motifs
For this next part of the analysis, we turn to the motifs, their individual frequencies, and
their distribution across the five variables. The total number of motifs, as designated by
Płonka, observed within the South Scandinavian Mesolithic is 230. Table 6 shows how often
a given motif occurs – note here that 123 are unique within the South Scandinavian Meso-
lithic, i.e., they only occur once. The frequencies follow an exponential pattern, with higher
frequencies being gradually rarer. The three most frequently occurring motifs are exception-
ally common, at 136, 125, and 98 occurrences respectively. Fig. 5.12 (next page) shows the
number of motifs per artefact, or what might be called complexity. The line graph shows that
the Maglemose objects are characterised by the least ‘complexity’ on average, mainly due to
the very high frequency of objects bearing only one motif. The Kongemose objects score a
significantly higher average number of motifs per artefact, at 4.1. This is in accordance with
Sørensen’s (2017: 135) previous assertions.
Table 6: Frequencies of individual motifs
The motifs which only occur once or twice are without much analytical
value for the purposes of the present dissertation, as their spatio-tem-
poral mapping and co-occurrences with other motifs cannot be argued
to be representative. The threshold at >2, which will be applied in the
following analyses, is arbitrary. Fig. 5.13 (page 59) shows all 68 motifs
occurring three times or more, i.e., those that are part of the following
analyses. Fig. 5.13 also shows Płonka’s stylised drawings of each motif.
These stylised motif images will be used recurrently in the following,
but recall that internal motif variability is high, and the stylised repre-
sentations are indeed highly stylised. This especially goes for the cate-
gories ‘ANT’ (anthropomorphic), ‘ZOO’ (zoomorphic), and ‘RELIEF’,
which have the greatest internal variability. Fig. 5.13 shows that A1, C1,
and G1 are the most frequent, and we might arguably add A24 to this
‘most common’ category. Another relevant calculation, though not visu-
alised, is the number of distinct motifs occurring in each of the three
cultures: 74 motifs occur in the Maglemosian, 126 in the Kongemosian,
and 119 motifs are known on artefact attributed to the Ertebølle period.
Recall that this is out of a total of 230.
Motif frequencies
Frequency
Count
1
123
2
39
3
19
4
6
5
4
6
7
7
2
8
3
9
2
10
1
11
2
12
1
13
1
14
1
16
2
17
1
18
1
19
1
21
1
22
2
23
1
26
1
28
1
34
1
36
1
49
1
50
1
69
1
98
1
125
1
136
1
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
58
Table 7: Number of motifs per artefact for each culture complex
Figure 5.12: Line plot showing the number of motifs per artefact for each culture complex.
This visualisation is complementary to table 7 (above).
# of motifs per artefact
# of mo-
tifs
Maglemose
(n=205)
Kongemose
(n=87)
Ertebølle
(n=168)
Total
(n=460)
1
110
22
69
201
2
39
16
38
93
3
21
10
17
48
4
13
4
14
31
5
9
7
7
23
6
6
9
5
20
7
3
5
6
14
8
0
4
4
8
9
0
3
3
6
10
0
1
2
3
11
0
0
3
3
12
1
0
0
1
13
1
3
0
4
14
0
0
0
0
15
0
0
1
1
Average
2.1
4.1
2.8
2.7
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
59
Figure 5.13: Overall frequencies of motifs on portable art objects. Only included are those motifs which occur at least three times.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
60
The following figure (5.14) subsets the motifs to those only occurring >9 times (n=25), and
plots these in a grid of stacked bar charts, one for each of the five variables.
o Motifs by Context: Most motifs show an even distribution between ‘good’ and ‘poor’
contexts. Some motifs, however, are exclusively or almost exclusively found in poor
contexts, those being ANT, ZOO, C5, and B3, with a few other motifs perhaps show-
ing the same tendency.
o Motifs by Chronology: Most of these most frequent motifs show a roughly even dis-
tribution across the three culture complexes, but some motifs seem to correlate
strongly with a single culture. Most prominent are the Maglemose ‘types’ A24, B10,
B2, and B3, and the more Ertebølle-specific motifs G1, C5, and D15. None of the mo-
tifs look to be purely Kongemosian in attribution, yet F5, E1, and B6 are more
strongly represented in this period. The chronological variable is of central im-
portance, and chronological mappings of the motifs occurring less than 10 times will
thus also be considered in the following analysis steps.
o Motifs by Region: Once again, while most motifs are evenly distributed across the an-
alytical regions, some show strong tendencies towards a single region. Most notable
are C5, mostly known from Western Denmark, and Eastern Danish motifs being E1
and B6 (which, as noted above, are Kongemose-correlated).
o Motifs by Material: The motifs do generally not show strong connections to a specific
material, with a few exceptions: all C5 motifs are found on antler, and all B3 are
found on amber. I1 are also mostly seen on amber objects.
o Motifs by Artefact Type: I1 and B3 are quite pendant-specific, and C5 is solely found
on axes. This evidently fits with the above observation, and it can be concluded that
B3 is related to Maglemose amber pendants, and the same goes for I1, though this
motif also occurs on antler and bone artefacts in the Ertebølle period. Likewise, C5 is
only found on Ertebølle antler axes. The ZOO motif is strongly correlated to amber
figurines but do also occur on antler objects. Recall that carved animal figurines are
included, and that amber is the material which is most frequently sculpted.
Fig. 5.14 thus demonstrates that only a few of these most frequent motifs show distinct pat-
terns across the five variables, the most prominent ones being C5 and B3. Motif B3 is closely
connected to both a specific context, chronology, material, and artefact type, namely loose
finds of amber pendants attributed to the Maglemose. This supports Toft’s (2017: 262–66)
findings. Interesting to note is also that, while motifs B2 and B10 are quite similar to B3, they
do not have this strong correlation with amber pendants. All three motifs are, however,
strongly connected to the Maglemose culture complex.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
61
Figure 5.14: Grid of stacked
bar charts for motifs with
frequencies =>10, shown by each
of the five categorical variables
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
62
The following plot (fig. 5.15) is of particular importance and should be given more careful at-
tention. We are once again considering all motifs occurring three times or more (as in fig.
5.13). There are 68 such, but two of these (C8 and G9) each occur on an artefact dated to the
unspecific ‘Kongemose/Ertebølle’ chronological category and are therefore excluded. Thus,
fig. 5.15 shows 66 motifs with Płonka’s stylised representation and motif name, as well as the
motif’s overall frequency indicated in brackets. The bars show each motif’s percentual distri-
bution across the three cultures, and the grid is ordered as a sequence from the most Magle-
mose-specific to the most Ertebølle-specific motifs. In this way, the figure demonstrates that
all three cultures have motifs that are specific to that culture, however these culture-specific
motifs are almost invariably motifs with low overall frequencies. Many motifs are evenly dis-
tributed across the culture periods and note that these are generally also the ones with more
overall occurrences.
B14, B3, B10, and B2 are quite similar motifs, in that they all consist of a long line with at-
tached small strokes to one side. The variability between these motifs is that the strokes may
be long or short, or perpendicular, oblique, or chevron-like, but the similarities are signifi-
cant, and all these motifs are solely or strongly linked to Maglemosian contexts. Interesting
to note, however, is that morphological similarity does not necessarily correlate with chro-
nology; D6 and D7 are both motifs characterised as “a triangle without base, … encountered
singly or in a row” (Płonka 2003: 211), but with different internal fill. These motifs are evi-
dently very similar but are known from the Maglemose and Ertebølle periods, respectively.
The following three pages (figs. 5.16-5.18) are spatial mappings of the same 66 motifs shown
in fig. 5.15. 22 motifs are shown per figure, shown in the same order as in fig.5.15. Thus, the
ordering of the motifs on the three following map pages are ordered broadly by chronological
attribution – but note that several of the motifs are not strongly linked to any specific cul-
ture. Regarding fig. 5.16, showing the motifs with the strongest Maglemosian affiliation, a
relevant observation is that most of the motifs show large spatial distributions, even many of
those with low frequencies. Only motifs A14, B14, F3, and E19 are spatially very limited. This
arguably also pertains to B1, which only occurs on Zealand within the Maglemosian, but this
motif (a solitary, straight line) is so generic that it probably does not have particular im-
portance. B3 is strongly associated with West Denmark. The motifs mapped in fig. 5.17 in-
clude those which are most strongly affiliated with the Kongemose, and the Kongemosian
core area of Zealand and Scania is unsurprisingly strongly represented. Most interesting are
perhaps E1 and B6, which are strongly associated with Zealand while also having higher fre-
quencies. Fig. 5.18 shows a strong Ertebølle trend that the east coast of Jutland becomes
highly represented, and that this is the case for almost all Ertebølle motifs.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
63
Figure 5.15: Grid of motif attributions to the three cultures, including Płonka’s stylised rep-
resentations. Purple = Maglemose, red = Kongemose, yellow = Ertebølle. The bars indicate
relative attribution to the cultures. The overall frequency of each motif is given in brackets.
Objects attributed to the unspecific ‘Kongemose/Ertebølle’ category are excluded here.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
64
Fig. 5.16: Spatio-temporal distribution of the 22 motifs with the strongest Maglemose attribution
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
65
Fig. 5.17: Spatio-temporal distribution of the 22 motifs with the strongest Kongemose attribution
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
66
Fig. 5.18: Spatio-temporal distribution of the 22 motifs with the strongest Ertebølle attribution
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
67
5.2 Testing for semiotic structure
Following the above descriptive statistics, we now turn to the testing of the dissertation’s
main question – whether semiotic structure of any kind can be inferred. Table 8-10 summa-
rise the frequencies of respective motif co-occurrences at each level of skipgram analysis (2-,
3-, and 4-grams, which are synonymous with ‘motif co-occurrence’, ‘triple motif co-occur-
rence’ and ‘quadruple motif co-occurrence’). In these tables, ‘frequency’ indicates the co-oc-
currence frequency, and ‘count’ denotes how many distinct motif combinations correspond
to that frequency. In all three tables, frequencies of 1 are dominating, as would be expected
when generating skipgrams, especially with the high-set k value of 13. Table 8 shows the fre-
quencies of motif co-occurrences at the level of 2-grams. The notable observation here is that
a few combinations are high in frequency, two of them occurring 44 times – which is quite a
lot, as this means that both of these combinations occur on approximately 9% of all objects
in the corpus. Table 9 shows, as expected, a higher overall number of generated skipgrams
compared to the 2-grams, as well as a significantly higher count of frequencies of 1. This is
expected simply because the skipgram method will generate many more combinations when
the n value is increased. This trend is continued in table 10, showing all generated quadruple
co-occurrences. The most important demonstration is undoubtedly that recurrent quadruple
co-occurrences are so incredibly rare, with only four combinations occurring more than
three times (as indicated by the two bottom rows of table 10).
Table 8: Frequencies of
motif co-occurrences
Table 9: Triple motif
co-occurrences
Table 10: Quadruple motif
co-occurrences
Frequency of
13-skip-2-grams
(n=1543)
Frequency
Count
1
1224
2
154
3
59
4
32
5
20
6
14
7
8
8
6
9
5
10
5
12
6
13
2
14
1
15
2
16
1
22
1
28
1
44
2
Frequency of
13-skip-3-grams
(n=4422)
Frequency
Count
1
4094
2
222
3
56
4
27
5
9
6
8
7
4
8
1
19
1
Frequency of
13-skip-4-grams
(n=8466)
Frequency
Count
1
8335
2
111
3
16
4
3
5
1
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
68
The above tables demonstrate a scarcity of triple- and especially quadruple motif co-occur-
rences, which already at this point suggests that the deliberate combination of three or four
motifs was not a common Mesolithic practice. This preliminary interpretation of the data can
however not immediately be made for the combination of only two motifs, as significantly
more combinations show higher frequencies. These motif combinations with the highest fre-
quencies are show below (fig. 5.19). This plot is subsetted to include combinations with fre-
quencies higher than 5 (refer to table 8). It shows that the two combinations occurring 44
times are A1-C1 and C1-G1, with A1-G1 being the third highest at a frequency of 28. The in-
clusion of stacked colours, indicating chronological attribution, also allows for a relevant ob-
servation, namely that almost none of these combinations are attributed solely to a specific
culture complex. From this can be gathered that, if these motif combinations were in fact
meaningful collocations, they must have been so through an extended period of time. This
argument, however, is not sufficiently qualified by this figure alone.
Figure 5.19: Motif co-occurrences with the highest frequencies, stacked by chronological at-
tribution.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
69
For comparison with the above plot, fig. 5.20. shows the most frequent triple co-occurrences.
This plot is subsetted to include combinations with frequencies above three (refer to table 9).
It is revealed that the exceptional triple co-occurrence with a frequency of 19 pertains to the
combination A1-C1-G1, echoing fig. 5.19. Recall that these two bar plots are based on the
same corpus and are thus representations of the same ornamented artefacts. This implies
that the 19 objects on which A1, C1, and G1 co-occur are all necessarily also represented in
the bars A1-C1, A1-G1, and C1-G1 of fig. 5.19. Putting aside this relatively frequent triple co-
occurrence, all others occur eight times or less, and it is not immediately obvious from this
visualisation whether they reflect randomness. Noting the chronological attributions, how-
ever, some combinations are in fact limited to a specific culture complex, though most seem
evenly distributed chronologically. The culture-specific combinations might potentially re-
flect temporally limited phenomena of meaningful collocations, and it is necessary to further
explore this possibility.
Figure 5.20: Triple motif co-occurrences with the highest frequencies, stacked by chrono-
logical attribution.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
70
The two above stacked bar plot (figs. 5.19 and 5.20) allow for some relevant observations:
they illustrate that the same motifs recur in many bars, most prominently A1, A24, B10, C1,
G1, and H1. This trend is even more pronounced for the triple motif co-occurrences in fig.
5.20, with almost all bars including at least two of these most prominent motifs. Recall from
Chapter 5.1.2 these are exactly the motifs with the highest individual frequencies, and this is
likely the only reason why these motifs dominate the figures. As these co-occurrences might
not be statistically significant, the plot must also not be taken to reflect meaningful Meso-
lithic phenomena. We will therefore now consider the generated skipgrams based on their
individual O/E ratios rather than their raw frequencies. In this way, a demonstration of
somewhat rarer, but more statistically significant (and thus potentially meaningful) combi-
nations is rendered possible.
As accounted for in Chapter 4.2.2, the O/E ratio has been calculated for all generated
skipgrams – this includes both 2-, 3-, and 4-grams. These data tables (Supplementary Mate-
rial 4, 5, and 6) are in descending order by O/E ratio and have been subsetted to only include
co-occurrences with frequencies of three or above – as the inclusion of all would generate too
massive data tables, and as co-occurrence frequencies of two or less are in any case the least
analytically relevant.
Fig. 5.21 shows three bar charts, showing motif combinations at each of the three levels (two-
, three-, and four-motif co-occurrences) which have the highest O/E values (indicated at the
bottom of each bar). For the top and middle bar chart, 27 combinations are included, which
is an arbitrary threshold. For the lowest bar chart, only 20 combinations are shown, as these
2o are the only four-motif combinations generated by the skipgram analysis with frequencies
of three or more.
o The stacked bar plot at the top shows the most significant two-motif co-occurrences,
and it is immediately evident that more of these combinations are linked to a specific
culture complex than is the case in fig.5.19. This is in itself an indication that O/E ra-
tio values is a fitting method for finding more meaningful combinations than consid-
ering raw frequencies alone. The chronological attributions are skewed, however:
Only one combination (A24-A14) is Maglemose-specific, two are Kongemose-specific
(B10-F38 and H1-D33), and ten are unique for the Ertebølle period (I1-I5, I1-I13, I5-
I13, C13-12, C12-C4, C5-C4, C5-C12, C5-C13, C5-D7, and H1-RELIEF). The combina-
tions which are not culture specific hint at a gradual, temporal replacement which re-
sembles the pattern visualised in fig. 5.15. Recall once again that chronological attrib-
utions are uncertain, and direct datings of the artefacts in question would potentially
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
71
alter these plots to show either stronger or weaker patterns of culture-specific attrib-
utions of motif combinations.
o The middle bar chart of fig. 5.21 shows the most significant triple motif co-occur-
rences. In comparison to the bar chart at the top, the culture-specific combinations
are fewer, and the impression of gradual replacement is more marked. Yet, the Erte-
bølle-specific combinations still dominate. Also contrary to the bar chart at the top is
that the motifs with the highest individual frequencies – A1, C1, and G1 – now domi-
nate, as they occur in almost all of the triple collocations. Maglemose-attributed or-
namented objects stand out even less than in the top bar chart, and Kongemose ob-
jects blend more significantly with both Maglemose and Ertebølle. Many of the Erte-
bølle-specific collocations consist of the same motifs as have been discussed above,
C1, C4, C5, C12, and C13. However, D7 does not occur as a collocation with C5.
o The bottom bar chart shows the most significant quadruple motif co-occurrences. As
mentioned, only 20 four-motif combinations occur three times or more (see table 10).
In this plot, the chronological divisions blend yet more significantly. The Ertebølle
cluster, clearly distinct in the two- and three-motif collocations, is not strongly repre-
sented here. Only three Ertebølle artefacts are engraved with the same four motifs.
In summary, fig. 5.21 suggests that some motifs were indeed combined on an artefact more
than once within a specific culture period, however such recurrent combinations are only
known in low frequencies, the highest being five. In spite of the low co-occurrence frequen-
cies, the O/E ratios lend support to the possibility that these might be non-random Meso-
lithic phenomena. Thus, motifs do overall not seem to co-occur with each other in significant
ways, though with some exceptions. The most interesting co-occurrences, both with regards
to chronological attribution as well as O/E ratio, seem to be from the Ertebølle period, and
this particular demonstration will be investigated further in the discussion chapter. Moreo-
ver, it is evident that the potential for meaningful collocations decrease correspondingly to
the order of skipgrams in question. In other words, two-motif co-occurrences seem more
likely to reflect Mesolithic deliberation than combinations of three or four motifs.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
72
Figure 5.21: Most significant motif combinations for the three levels of skipgram analysis.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
73
As fig. 5.21 has demonstrated that two-motif co-occurrences show the most promising poten-
tial, these are further analysed by visualising them as a heatmap (fig. 5.22). This visualisation
is based on the data table of two-motif co-occurrences (Supplementary Material 4), the cell
colours indicating O/E ratios. Important to note is that co-occurrences with the three most
frequent motifs, A1, C1, and G1, do not score particularly high O/E’s, once again indicating
that co-occurrence frequencies alone are not a sufficient measure. The combination ‘A1-C1’,
for example, has an observed frequency of 44 and an expected frequency of 48.99, and there-
fore looks to represent a random co-occurrence. This heatmap will also be revisited in the
upcoming discussion chapter.
Figure 5.22: Heatmap of motif co-occurrences of =>3, with O/E ratios as values. Note that
the x- and y-axis are identical, and that the order of motifs corresponds to fig. 5.15, reflect-
ing the best possible chronological ordering of motifs.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
74
Chapter 6
Discussion
6.1 Limitations of this dissertation
As has been discussed previously, the prevailing issues concerning this material are substan-
tial: most critically, the poor contextual information of the majority of objects negatively af-
fects the overall chronological resolution, but it also problematises interpretations regarding
whether ornamented objects were deposited in particular ways, whether ornamented objects
were associated with specific people, and similar questions. The second significant bias con-
cerns the motif classification: this dissertation utilises Płonka’s classification uncritically.
However, the designation of patterns to specific motif categories is very much a subjective
exercise, and disagreements between scholars is not uncommon (e.g., Terberger 2006, foot-
note 5; Toft & Petersen 2015: 210). This dissertation does not solve these issues, nor is it im-
mune to the problems they pose. The above analyses have primarily focused on motif co-oc-
currences, while other potentially relevant factors, such as engraving technique and motif
composition on artefacts, are not considered. Also not considered are intra-site analyses to
further investigate depositional patterns, and neither from what animals the decorated bone
and antler artefacts derive. In other words, while this dissertation’s methodological approach
is novel, there are many other relevant angles of entry to this material.
6.2 Take-home messages from the descriptive statistics
The descriptive statistics presented in Chapter 5.1.1 are designed to identify patterns in the
portable art objects across the five variables ‘Context’, ‘Chronology’, ‘Region’, ‘Material’, and
‘Artefact Type’. Regarding these five variables, the most noteworthy observation is perhaps
how few patterns are discernible. Very few sweeping statements can be made about the cor-
pus as a whole, but there are some relevant observations. First of all, the activity of engraving
portable objects seems to have been a constant of the South Scandinavian Mesolithic. The
ornamented objects are rather evenly distributed across the three cultures, and nothing
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
75
indicates a hiatus in ornamenting objects at any given time. Furthermore, the spatial distri-
bution of ornamented artefacts within the Maglemose, Kongemose, and Ertebølle (figs. 5.2-
5.4) closely reflect the general site distributions for the respective cultures. This qualifies the
argument that ornamentation of portable artefacts was a common behaviour across the spa-
tio-temporal span of the South Scandinavian Mesolithic. The denser clusters, such as the
Åmose area of East Denmark, are most likely the result of taphonomic bias, reflecting differ-
ential preservation conditions. This is also the most probable explanation for the scarcity of
finds from Western Jutland, Northernmost Germany, and the Polish Baltic coast. It is en-
tirely possible that, if direct radiocarbon dates were obtained for a large enough portion of
the corpus, a temporal hiatus might be recognised, similar to what has been demonstrated
for the Maglemosian bone points (Jensen et al. 2020). At present, the temporal resolution is
far too poor for such questions to be investigated, however.
The material types bearing ornamentation are naturally present across the study area. This
goes especially for antler and bone. Ornamentation occurs predominantly on antler, bone,
and amber (≈ 90%), most often in the form of antler axes, antler shafts, amber pendants,
bone daggers and bone knives. Chronological differences in ornamented artefact types are
discernible, but I would argue that this primarily reflects the changing artefact repertoires
from early to late Mesolithic. A few observations from each of the culture complexes are po-
tentially relevant:
o The Maglemosian activity of ornamenting long bones (n=11) is unique for this earliest
Mesolithic phase (e.g., the aurochs bone from Ryemarksgaard, fig. 2.8). These bones,
predominantly aurochs metacarpals or metatarsals, were in almost all cases scraped
and polished before being engraved, and it is unknown whether they had a practical
function (Płonka 2003: 53–54), but it has been suggested that these functioned as
clubs (Andersen & Mortensen 2022). If not, the ornamentation of non-functional ob-
jects looks to be a Maglemosian phenomenon, which is perhaps indicative of a differ-
ent purpose.
o A relatively high proportion (≈ 16%) of ornamented Kongemose objects are daggers,
most of these being slotted bone daggers (e.g., the daggers from Kongemose and
Skellingsted Bro shown in fig. 2.7). This has been noted before (Larsson 1990: 286,
2000: 37), but is here quantified (fig. 5.11). The frequent ornamentation of this dis-
tinct artefact type, coupled with the observation that slotted bone daggers from the
Danish region might have been Late Mesolithic exotica from present day Sweden,
suggest that this phenomenon deserves further investigation.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
76
o Beginning in the Kongemose period, but culminating in the Ertebølle, is the high fre-
quency of ornamented antler axes. ≈ 36% of all ornamented Ertebølle objects are ant-
ler axes. Akin to the Kongemose predilection for ornamenting slotted bone daggers,
this might be a slight indication that the type of artefact had significance. Yet, as or-
namentation is consistently seen across many different artefact types, such an argu-
ment remains conjecture.
Concerning the exploratory analyses of distinct motifs (as designated by Płonka), the hetero-
geneity continues to be evident: table 6 shows that 123 motifs out of the total 230 only occur
once. Recall that these ‘unique’ motifs, as well as those occurring only twice, are not analyti-
cally meaningful for the present dissertations’ methodology and have thus not been consid-
ered in the analyses. The observation that so many (53%) of the motifs are unique, however,
is very much a relevant observation in the pursuit to understand the purpose of the portable
art as a whole. It might reflect several things: perhaps it is a result of rigorous classification
by Płonka, where a motif’s internal variability is so great that Płonka has had to make the
subjective decision to divide what was in Mesolithic reality one motif into several motif
types. It might also be an indicator that the engravers did not follow strict dogmas, and the
overall motif heterogeneity in fact reflects to some extent the creativity or individuality of the
individual engraver – and thus, that the classification of motifs is in itself a misleading en-
deavour. The variability (or ‘complexity’), however, varies over time, with the most drastic
changes occurring between the Maglemose and Kongemose periods; Maglemosian ornamen-
tation is the least complex, with 74 distinct motifs and an average of 2.1 motifs per artefact
(primarily due to a high number of objects bearing only one motif). By these measures, early
Mesolithic ornamentation is the most homogenous. This fits with the general image of the
Maglemosian technocomplex as having very limited spatial variability due to lower popula-
tion numbers and high mobility. This seems to change altogether in the succeeding
Kongemose period, which is characterised by the highest number of distinct motifs. Bearing
in mind that it is also the period with fewest known ornamented objects, Kongemose orna-
mentation is the most ‘complex’, both in terms of average number of motifs per artefact (4.1)
and overall heterogeneity (126 distinct motifs). The objects from the Ertebølle period have an
average of 2.9 motifs per artefact and 119 distinct motifs. As this is also based on a larger
number of objects, it is evident that the complexity decreases somewhat from Kongemose to
Ertebølle. It is noteworthy that the Ertebølle ornamentation does not score highest in this
complexity measure, given that the period is renowned for its general increase in societal
‘complexity’. However, as the following discussion chapter (6.3.2) construes, Ertebølle motif
combinations show the best potential of being meaningful collocations – and this is perhaps
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
77
to be regarded as an alternative measure of complexity. Other highly relevant demonstra-
tions from the descriptive statistics include:
o The zigzag motif (A24) is characteristic for Maglemose – this stands in contrast to
Nash’ (1998: 49) finding, based on a sample of 100 artefacts, that zigzags are tempo-
rally evenly distributed. Zigzags, in different variations, do occur throughout the Mes-
olithic, and therefore the motif should not a priori be typologically attributed to the
early Mesolithic, even if the Maglemosian majority is substantial.
o B3 is one of the very few motif types that are strongly correlated to a specific artefact,
namely amber pendants. This is not a novel finding of the present dissertation (Toft
& Petersen 2015), but it is further supported here.
o Kongemose-specific motifs are present in the data (E2, F16, F38), but these are
known in low frequencies, and a demonstration of Kongemose-specific ‘Leitmotifs’ is
thus not quantitatively supported.
o While the variability of motifs is highest in the Kongemose period, the Ertebølle pe-
riod has more motifs which are limited to that culture complex. By both these
measures, Maglemose is the ‘simplest’.
o The motifs A1 (short, parallel lines), C1 (long, parallel lines), G1 (crosshatching), and
A24 (zigzag) are very frequent compared to the other motifs. These four are spatio-
temporally omnipresent, yet, while A1 and C1 are temporally evenly distributed, the
frequency of A24 decreases during the Mesolithic, while G1 has an opposite, increas-
ing trend.
6.3 A (near) refutation of the semiotic hypotheses
There is no one methodology of testing for semiotic structure. Indeed, if there were, scholars
would not have argued for decades about the, e.g., Indus Valley and Vinca ‘scripts’, and
whether they represent writing systems or not. Due to the complexities of inferring semiotic
structure, these discussions regarding prehistoric art will continue in the decades to come –
especially since ‘Rosetta Stone’ scenarios, in which new key finds render possible complete
and unambiguous deciphering must be assumed near impossible. By this token, the method-
ologies and analyses applied in the present dissertation do not facilitate a definite answer to
the question of whether semiotic structure of any kind existed within the portable art of Mes-
olithic South Scandinavia. Further studies with different, complementary methodologies are
necessary, and these will need to be evaluated and discussed by scholars within the fields of
archaeology, linguistics, and anthropology. The present dissertation does, however,
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
78
constitute a robust and data-driven starting point from which such further studies can be un-
dertaken. Heeding Bouissac’s call to “heuristically scan strings of signs in whatever direc-
tions sequences can be construed, in search of recurring combinations of types”, the analyses
show both interesting patterns and lacks thereof, rendering possible an assessment of the
‘weak’ and ‘strong’ semiotic hypotheses, respectively.
6.3.1 The Strong Semiotic Hypothesis
The theoretical criteria for supporting a presence of ‘strong’ semiotic structure were outlined
in Chapter 3.2.1. These include the demonstration of recurrent triple or quadruple motif
combinations, and that such collocations should ideally be with a limited spatio-temporal
distribution and further strengthened by other observations from the descriptive statistics.
Based on these analyses, the following observations serve as indices that there existed no
‘strong’ semiotic structure within the portable art of Mesolithic South Scandinavia:
o The overall diversity of motifs: ≈70.4% motifs, as designated by Płonka, occur just
once or twice, which strongly indicates a lack of systematisation, and instead suggests
creative freedom of the individual engraver.
o The minimal ornamentation of most objects: ≈43.7% of the ornamented objects only
bear one motif, and ≈63.9% have two or less.
o The internal heterogeneity of a given motif type: A given motif, represented in a styl-
ised way by Płonka, shows great variability when observing the motif on the actual
artefacts. Once again, this is indicative of a lack of systematisation.
o The lack of systematic selection of media: A diverse array of objects have been orna-
mented, and only a few strong correlations between artefact type/material and motif
exist, these being motif B3 for amber pendants and the ‘sheaf of grain’ for antler
axes).
o The scarcity of potentially meaningful triple or quadruple collocations: The low fre-
quencies and dispersed chronological attributions of (nearly) all statistically signifi-
cant triple and quadruple co-occurrences indicate strongly that the purposeful combi-
nation of three or more specific motifs was not a Mesolithic practice. This effectively
renders highly unlikely that ‘strong’ semiotic structure existed.
In summary, the presence of ‘strong’ semiotic structure must be considered extremely un-
likely for the entire corpus, and highly unlikely for a spatio-temporally limited part of the
corpus. But as is evident from the review of theoretical research history (see fig. 3.2), no
scholars really argue for the presence of strong semiotic structure in the Palaeo- or
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
79
Mesolithic. This strong rejection is therefore not sensational, but rather very much expected.
Following Bouissac, however, and in the wake of the recent revival of semiotic interpreta-
tions of prehistoric art, no hypothesis should be disregarded a priori.
6.3.2 The Weak Semiotic Hypothesis
Considering the ‘weak’ semiotic hypothesis, confident conclusions such as the one above are
more difficult to arrive at. As explicated in Chapter 3.2.1, the following discussions assume
weak semiotic structure to be a societal phenomenon, rather than something practiced by a
single or a few individuals. Under this assumption, the ‘weak’ semiotic hypothesis looks to be
a highly unlikely explanation for the entirety of the corpus, for the same reasons as listed in
the above. However, given the relatively large spatio-temporal extent of the study, it is more
likely that a weak semiotic system would have emerged intermittently and locally. This possi-
bility is explored in the following, based on the analyses presented in the previous chapter.
In fig. 6.1 below, the heatmap of co-occurrences is revisited, additionally marking all combi-
nations that are specific to one culture complex. Recall that chronological attributions are in
many cases uncertain, and therefore the figure would likely look different if objects were
dated absolutely. At present, this visualisation reflects the best chronological clarity obtaina-
ble. Note that here is a particularly significant cluster of the Ertebølle-specific motifs C4, C5,
C12, C13, and D7, most combinations of which are highly significant (indicated by the deep
red colouring of the given cell, denoting a high O/E value). It is also evident that, although
many combinations are specific to the Maglemose and Kongemose periods, only very few are
statistically significant (white or weak red colouring). Therefore, no motif ‘clusters’ can be
demonstrated within the Maglemose or Kongemose. Apart from the Ertebølle cluster, signifi-
cant combinations from the Ertebølle are the combinations of I1, I5, and I13, as well as H1-
RELIEF. These two observations are not only motif combinations, but pertain to methods of
ornamentation: The ‘I’ group of Płonka’s classification is ‘drilled patterns’ (Płonka 2003:
214), and the ‘RELIEF’ motif type primarily refers to the ornamented wooden paddles of the
Ertebølle. But it is the Ertebølle cluster of motifs C4, C5, C12, C13, and D7 which unequivo-
cally has the most promising potential of reflecting a semiotic system, as it is five motifs
showing interrelated statistical significance of co-occurrence. This is further supported by
the following map (fig. 6.2), which shows that the cluster is also spatially limited. Another
factor which adds to the probability that they are meaningful is that all objects are antler
axes, except for one antler shaft (Ydernæs 6 (Fiskerhuset)).
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
80
Figure 6.1: Heatmap filled by O/E ratio, sorted by chronological attribution from lower left
to upper right. Motif combinations limited to one specific culture complex are marked with
a small, coloured box around the corresponding cell. Purple: Maglemose, red =
Kongemose, Yellow = Ertebølle. The culture-specific combinations which also have particu-
larly high O/E ratios are shown in large boxes. The box with motifs C4, C5, C12, C13, and
D7 is a cluster of co-occurring motifs.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
81
Figure 6.2: The spatial distribution of co-occurrences for the five motifs of the Ertebølle
cluster. The coloured patches show the known extent of each individual motif. The pie
charts denote artefacts, and their colours denote which of the five motifs are present on the
object.
1: Kolindsund, 2: Eskelund, 3: Horsens “Loddentot”, 4: Stensballe Sund, 5: Torpegård, 6:
Bogø Nor, 7: Korsør, Glasværk, 8: Skellingsted Bro, 9: Ydernæs 6 (Fiskerhuset), 10: Revlen.
First of all, fig. 6.2 strongly opposes the notion that the Great Belt constituted a cultural bar-
rier in the Ertebølle period, as has otherwise been proposed (Andersen 1981: 20; Ballin 2017:
333). Rather, the spatial distribution of both individual motifs and their mutual co-occur-
rences suggest a cohesion between Zealand and the Eastern coast of Jutland, as well as the
central Danish islands of Fyn and Langeland. While Zealand seems to have been a particular
cultural locus in both the Maglemose and Kongemose cultures, it appears (at least when only
considering the portable art) to remain so in the Ertebølle period, but with the inclusion of
the coast of East Jutland. Recall that this coastline was newly formed in the Late Mesolithic,
providing new and untapped affordances regarding marine subsistence.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
82
From a qualitative perspective, C12 and C13, as well as C4 and C5, are very similar motifs.
Perhaps we are not dealing with five discrete motifs, but rather three related motifs with a
significant degree of variability. But, importantly, note that C6 is also a version of the sheaf
of grain, bearing close resemblance to C4 and C5, but that C6 does not co-occur significantly
with any of the motifs of the above shown ‘Ertebølle cluster’. This might in fact serve as sup-
port for Płonka’s rigid motif attributions. Fig. 6.2 echoes the findings of Andersen (1981) and
Nash (1998), who both attributed particular importance to the ‘sheaf of grain’ motif
(Płonka’s C4, C5, and C6). The present analysis has further demonstrated that the ‘sheaf of
grain’ motif is not only noteworthy in its own right, but also when it comes to statistically sig-
nificant (and thus potentially meaningful) motif co-occurrences. As discussed in Chapter
2.3.1, Nash’s ‘Kula ring’ proposal is not widely accepted. Yet, this dissertation’s finding lends
support to Nash’s conclusion to some extent – primarily in the sense that something anoma-
lous and potentially novel occurred at this very point in time and space, and that this is re-
flected materially in the portable art.
The central questions are, then, 1), whether this Ertebølle motif cluster did have a semiotic
purpose, and 2), whether the present findings support or oppose Nash’s interpretation of a
‘Kula ring’ of status-driven, intergroup exchange. Recall that Nash’s interpretation falls un-
der the ‘socio-political’ school of interpretation, but also that a ‘socio-political’ purpose could
potentially coincide with the semiotic hypotheses. An answer to these questions might be in-
ferred from the spatial distributions visualised in fig. 6.2: In the above map, the locus is the
Åmose area of Western Zealand. The axe from Skellingsted Bro is also the only object bear-
ing all five focal motifs. All five motifs are spatially limited, and they all converge only in
Western Zealand, represented by the Skellingsted Bro axe. More than anything else, this is
indicative of several small groups interacting with their immediate neighbouring groups. If
the five motifs were semiotic collocations with agreed-upon meanings in a specific area,
these motifs would be expected to have more similar spatial distributions. Therefore, the Er-
tebølle cluster does not appear to reflect a local emergence of semiotic meaning. Rather, it
seems like all five motifs are parts of a regional system, each with a different core area, seem-
ingly with an eastern trend for motifs C12 and C13, and a western trend for C4, C5, and D7.
Following this line of thinking one may conjecture that the heartland of motif D7 might have
been roughly the area around the Bay of Aarhus but that the motif also diffused to Western
Zealand. Through the same process of neighbour-to-neighbour contact, the inverse occurred
with motif C13: its core area was Zealand, and it diffused to the Bay of Aarhus and Lange-
land, respectively.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
83
This interpretation comes with some substantial reservations, however. First, it is only based
on merely ten artefacts. If these ten objects were produced within the same human genera-
tion, the argument would be strong, but that is unlikely to be the case. They could have been
engraved several centuries apart, and this underlines the crucial importance of obtaining di-
rect radiocarbon dates for more Mesolithic portable art objects. Furthermore, it must be as-
sumed that only a fraction of Mesolithic portable art objects is known to us, and that there-
fore the spatial extent of both individual motifs and their co-occurrences were potentially
larger than what is known archaeologically.
In summary, the weak semiotic hypothesis has been deemed potentially applicable for a spe-
cific cluster of Ertebølle motifs. Yet, in considering its spatial distribution, the cluster seems
more likely to reflect local styles or preferences of motifs, and that motifs in this case dif-
fused to the immediate neighbouring groups. The reason for the statistical significance of the
motifs of the Ertebølle cluster is thus not likely due to them being semiotic collocations but
suggests that these five motifs were particularly prevalent in inter-group motif diffusion.
This interpretation is not entirely consistent with either Płonka’s, Terberger’s, or Nash’s in-
terpretations. Płonka (Płonka 2003: 178) suggested that “…late Mesolithic art of south Scan-
dinavia may be distinguished into a number of local styles suggesting that art objects were
usually produced and used locally”; this is equivalent to Terberger’s argument, that a specific
motif’s distribution could reflect the territory of one regional ‘band’. These interpretations
are not supported by the spatial distribution in fig. 6.2, given that all five motifs of the Erte-
bølle cluster overlap considerably. The interpretation is more in accordance with Nash’s
‘Kula ring’ notion, but this dissertation has not engaged enough with this socio-political hy-
pothesis to discuss it further here – this project does not necessarily support nor oppose
Nash’s suggestion that portable art functioned in the context of prestige- or status-related ex-
change. Yet, Andersen’s interpretation, that motifs served as group identifiers, seems highly
unlikely for the Ertebølle motif cluster, as there is too much overlap in the spatial distribu-
tion of motifs for them to be indicators of individual or group affiliation. On the whole, the
‘weak’ semiotic hypothesis must be rendered unlikely for the Ertebølle motif cluster, assum-
ing that such weak semiosis should be collectively known and practiced on local or regional
level.
In light of the above, this dissertation serves as a strong case for the argument that the porta-
ble art of Mesolithic South Scandinavia did not have a semiotic function. Therefore, the se-
miotic hypotheses are rejected in the case of this contextual area – the ‘strong’ semiotic
seems irrefutably rejected to the author, while the ‘weak’ semiotic is rendered highly un-
likely, with the reservation that other methodologies should be applied to further support
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
84
this claim. If a semiotic function for Mesolithic or even older ornamentation is to be widely
accepted, I would argue that it needs to be demonstrated on a societal level, such as at-
tempted by the present dissertation. Such an attempt was not made by Marshack or his con-
temporaries some decades ago, and the scholars engaged in the recent ‘semiotic resurgence’
have made this effort without receiving general acceptance by their peers. This dissertation
has placed itself in the midst of this semiotic resurgence and demonstrated that, for the
South Scandinavian Mesolithic material, a semiotic purpose is not the most likely explana-
tion.
6.4 Alternative interpretations: Maglemose drift, Ertebølle politics?
Through the above analyses and discussions, the main question of this dissertation has now
been addressed. The following chapter will venture slightly beyond the purpose of the disser-
tation, and address the question: If the semiotic explanations are not the most likely, which
are? The considerations and findings of all the preceding chapters will underpin this discus-
sion, but it should be underlined that the following arguments are preliminary and necessi-
tate studies in their own right. Recall the theoretical assumption that if no (proto-) writing
existed at a given time and space, we must assume that this was not because of a lack of ca-
pabilities, but because no need for it existed at the time. Conceivably, such a need might have
arisen in the Ertebølle, considering the general ‘complexification’, and increasing contact to
outside groups (see Chapter 2.1.1). The demonstration of the significant Ertebølle motif clus-
ter does in itself indicate a more ‘complex’ portable art phenomenon, regardless of it being a
semiotic system or not. The opposing schools of interpretation are, in the author’s view, just
as complex human behavioural phenomena.
Based on the significant differences between early and late Mesolithic portable art, it seems
likely to the author that different purposes existed at these different points in time, and thus
that more than one ‘school’ of interpretation is the correct interpretation of the corpus as a
whole. The Ertebølle motif cluster indicates the emergence of a new purpose for the orna-
mentations – arguably a socio-political one, given the distinct local and regional spatial pat-
terning of the motifs in question. Disregarding this anomaly, I will in the following argue
that neutral cultural transmission in motif morphology is strongly indicated for the corpus as
a whole. This is first and foremost indicated by a very noteworthy pattern in the skipgram
analyses (Chapter 5.2), namely the dominance of Kongemosian artefacts. Considering that
ornamented artefacts are most scarcely represented in the Kongemose, a surprisingly large
proportion of the co-occurrences are found on Kongemose objects. There are two evident
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
85
reasons for this: 1), Kongemose objects have a higher average of motifs per object, and as a
result, more combinations are produced from these objects when they are read as sequence-
less sentences per the skipgram methodology, and 2), it reflects that most motifs occurring in
the Kongemose also occur in the two other periods – as the Kongemose lies chronologically
in the middle. This is indicative of gradual replacement of motifs over time, rather than sud-
den changes in motif repertoire. Figs. 5.15 and 6.1 are also clear visualisations of this gradual
replacement/development trend. Therefore, an argument can be made that the overall pat-
terns of this portable art corpus resemble neutral cultural transmission. By this is meant that
individual motif morphology was seemingly not subject to selective pressure, i.e., ‘individual
motifs’ were never, or perhaps only exceptionally, strictly perceived and copied as such by
the Mesolithic engravers themselves. Thus, the Mesolithic practice of ornamenting objects
might be, broadly speaking, a long process of ‘cultural drift’ variation, governed by random
copying (Koerper & Stickel 1980). Apart from the gradual replacement of motifs over time,
an equally strong argument for the ‘random copying’ process is the overall frequencies of in-
dividual motifs (table 6): as demonstrated by Bentley et.al (2004), random, cultural drift can
be inferred from the relative frequency distribution of individual variants accumulated over
time, which will be distributed as “… a power law, with a small number of the variants attain-
ing very high frequencies but most occurring only very few times” (Shennan 2011: 1073).
Such an exponential frequency distribution is exactly what is demonstrated for Płonka’s
Mesolithic motifs, with motifs A1, C1, G1, and A24 being very frequent while the great major-
ity occurs only three times or less.
This notion of ‘random copying’ should ideally be further supported by attaining a more fine-
grained chronological resolution as well as a critical revision of Płonka’s classification. The
demonstration of neutral cultural transmission and an apparent ‘randomness’ would not de
facto disprove any school of interpretation. Yet, I would argue that it most strongly supports
the ‘decorative’, in the sense that it most readily fits with Wiessner’s (1983) ‘assertive’ cate-
gory, implying passive social conventions, as discussed in Chapter 2.3.1. Thus, rather than
any strict, social conventions implied by the other schools of interpretation, it seems more
likely that subjective, culture-specific aesthetic preferences were the driving force behind the
continuous tradition of ornamenting objects. That decorative, or ‘aesthetic’ incentives were
the primary drivers behind ornamentation is most often argued for markings from a much
deeper prehistory (Hodgson 2000, 2006; Currie 2011; Tylén et al. 2020), and, as reviewed,
often causes a stir when suggested for art of the Upper Palaeolithic onwards. Nevertheless, it
seems to be the most likely (yet still preliminary) interpretation for this Mesolithic corpus,
when considered as a whole. This suggestion is thus also contrary to what is most commonly
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
86
proposed for the South Scandinavian Mesolithic portable art, namely ‘rituo-magic’ or ‘socio-
political’ interpretations (see fig. 3.2). Indeed, it has previously been argued that “… it is in
the portable art that "art for art's sake" finds its most plausible examples, for in many cases
one may simply be seeing the decoration of functional objects” (Bahn 1987: 72). Apart from
the corpus passing resemblance to neutral cultural transmission, other indices pointing to a
‘decorative’ purpose include the general observations that 1), the selection of media does not
seem to have been of importance, 2), the majority of motifs are unique, and 3) internal motif
variability further supports the general absence of strict social conventions. Of course, aes-
thetics preferences are also very much culture-specific (Jacobsen 2010; Bao et al. 2016), and
the spatio-temporal mapping of specific motifs might therefore in fact reflect the spatio-tem-
poral extent of culture-specific aesthetic traditions – which are known to be potentially very
long-lived (e.g., Sreenathan et al. 2008). These arguments for a ‘decorative’ purpose, re-
flected by cultural drift, fit most readily with the portable art of the Early Mesolithic (Magle-
mosian, perhaps extending into the Kongemosian). Maglemosian objects are characterised
by the least significant motif co-occurrences. Moreover, the demonstration that Maglemo-
sian ornamentation has the fewest distinct motif types fits with the ‘random copying’ as-
sumption that the degree of innovation (i.e., the occurrence of novel motifs) depends on the
effective population size (Shennan 2011: 1073), which is commonly assumed to increase from
Early to Late Mesolithic. Maglemosian societies were, as reviewed in Chapter 2, likely part of
a broader and scarcely populated North European tradition (Płonka & Kowalski 2017). In
summary, it is argued that the portable art of the Maglemosian reflects decorative traditions
shared across large territories, characterised by slow, gradual, and random change.
As discussed, the above interpretation does not fit for the entire corpus: the demonstrated
Ertebølle motif cluster most strongly implies a ‘socio-political’ function, though the exact na-
ture of such socio-politics remains indiscernible. Thus, it seems very likely that this small
subset served a different purpose than the remainder of the corpus. The reason for this ex-
ceptional cluster of regionally limited, but interrelated motifs is most probably a reflection of
the general societal developments of the Ertebølle period, with its overall increase in ‘com-
plexity’ and regionally distinct material correlates. Thus, there are indications that the porta-
ble art served different purposes in the Maglemose and Ertebølle, respectively, with the
Kongemosian technocomplex representing an enigmatic transition phase in the middle. The
relative ‘richness’ of Kongemosian ornamentation seems peculiar when considering that
these objects cluster more densely in a smaller area, namely Zealand and Southernmost Swe-
den. Contradictorily, also recall the notion that the Kongemose may be part of a continental
‘trapeze horizon’. It seems evident that a better understanding of the Kongemose
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
87
phenomenon would unlock a better understanding of the Mesolithic as a whole, including its
portable art.
6.5 Future Studies
This dissertation provides the foundation for further quantitative studies of this material,
which can be angled in many other ways than what has been done here. The atypically quan-
titative approach is an attempt to provide a large-scale, data-driven perspective to a research
field that has been overwhelmingly qualitative and selective regarding which objects to in-
vestigate. However, some form of qualitative assessment is necessary for gaining a holistic
understanding of this heterogenous material, but the issue of subjective assessment is a con-
stant – especially when it is solely WEIRD (Western, Educated, Industrialised, Rich and
Democratic, see Killin & Pain 2022) scholars performing such assessments. One way to by-
pass this subjectivity bias could be a standardised measurement of ‘complexity’ (as in e.g.,
Tamariz & Kirby 2015; Karjus et al. 2022). This approach could test whether motifs become
gradually simpler from Early to Late Mesolithic, due to ‘learning’ and ‘reproducing’ processes
of cultural transmission chains, or conversely, if motifs become more complex over time, due
to new purposes arising in tandem with societal changes. In any case, such a study could fur-
ther clarify whether temporal changes in motif morphology was a result of selective pressure.
While this dissertation has utilised Płonka’s classification uncritically, future studies should
ideally critically assess Płonka’s categorisation. This should include inter-coder consistency,
to test the reliability of Płonka’s motif categories. As Płonka’s contribution is the most com-
prehensive and thorough, peer scholars investigating Mesolithic portable art ought to refer-
ence his categorisation, but also engage with it critically so it can be improved. Though some
motifs definitely seem like distinct ‘tokens’, this is not readily apparent for most motif ‘types,
as designated by Płonka. Therefore, and particularly for this material, the archaeological
habit of devising rigid classifications should perhaps itself be fundamentally questioned. If
researchers are intent on recognising the Mesolithic realities of engraving and observing
these marks, such a ‘typological’ frame of mind is probably to some extent a hindrance rather
than an aid.
This dissertation has also underlined the potential of using portable art motifs to infer Meso-
lithic social territories, as has been attempted previously (Nash 1998; Terberger 2006; Toft
2006, 2017; Larsson & Molin 2017a; Płonka & Kowalski 2017). The available data regarding
this material, compiled by Płonka, is potentially sufficiently large for inferring important
spatio-temporal motif variations. However, a greater chronological resolution is needed
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
88
before synchronic conclusions can be made. This chronological resolution is evidently the
most substantial issue, and one that should ideally be addressed first. As has been under-
lined several times, the research on this corpus of portable art is impaired by the overall poor
chronological resolution. The bulk of ornamented artefacts are loose finds without contex-
tual information, and chronological attribution is often based on typology, which is undoubt-
edly imprecise, at times dramatically so (Philippsen et al. 2019; Manninen et al. 2021;
Orłowska & Osipowicz 2021). An unknown portion of the present dissertation’s data must
thus be assumed to be imprecise as well. If direct radiocarbon dates were obtained for a stra-
tegic selection of this material, it could potentially provide many answers, such as the dura-
tion of individual motifs, temporal differences in intensity of ornamentation, and the demon-
stration of either gradual or abrupt changes in motif repertoires.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
89
Chapter 7
Conclusions
This dissertation is the first study explicitly addressing the question of semiotics for the port-
able art of Mesolithic South Scandinavia. This specific corpus of prehistoric engravings is a
relatively abundant archaeological record, owing to optimal preservation conditions and a
long history of Mesolithic research in the area. The South Scandinavian Mesolithic has here
been demonstrated to constitute a ‘contextual area’, fit for independent analysis, and by this
token, serves as a favourable place to investigate why portable art was produced. Further-
more, Tomasz Płonka’s (2003) extensive classification and catalogue provides an invaluable
contribution which facilitates quantitative and data-driven research on this heterogenous
material.
Engaging critically with theoretical movements of Continental Europe, this dissertation
demonstrates a resurgence of semiotic hypotheses in prehistoric art research and underlines
the unresolved issues that persist for this avenue of interpretation. Nevertheless, following
Bouissac’s (1994a) call for not disregarding any hypothesis a priori, it is tested whether the
portable art of South Scandinavia was governed by semiotic structure. This is done by adopt-
ing a methodology from the field of computational linguistics: by generating k-skip-n-grams
(skipgrams), motif co-occurrences are quantified, and the statistical significance of all co-oc-
currences is evaluated by calculating the O/E (observed/expected) ratio. This analysis, cou-
pled with a comprehensive presentation of descriptive statistics, have qualified the following
findings:
The descriptive statistics show a general heterogeneity across all variables. With only few ex-
ceptions (Płonka’s B3 and C5), individual motifs are not strongly correlated with a specific
region, material, or artefact type. Although several motifs seem to have been present
throughout the Mesolithic, a minority of motifs only occur within a specific culture complex.
The Kongemose and Ertebølle are the most ‘complex’, in terms of greater motif diversity and
overall richness of ornamentation. Thus, portable art is an additional group of material cul-
ture which can be said to ‘complexify’ in the Ertebølle, with its beginnings in the Kongemose.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
90
The skipgram analyses have clearly demonstrated that, concerning the vast majority of the
corpus, specific combinations of motifs do not look to have been of importance to the Meso-
lithic engravers. Most co-occurrences which are statistically significant, as indicated by the
O/E ratio values, are too low in frequency and too dispersed chronologically to qualify an ar-
gument that they reflect meaningful phenomena. Thus, this dissertation renders the semiotic
hypotheses highly unlikely on the whole and thereby constitutes an indirect critique of recent
years’ resurgence in claims for prehistoric semiosis. While the ‘strong’ semiotic hypothesis
seems irrefutably rejected, the presence of a ‘weak’ semiotic system cannot be fully refuted
on the basis of this dissertation, though it is rendered unlikely. It has been demonstrated
clearly that the corpus of portable art as a whole shows no signs of semiotic structure, and
that motif combinations do generally not seem to have been of specific importance. Instead,
this dissertation has qualified this alternative interpretation: Maglemosian ornamentation
was not governed by any dogmatic rules on a societal scale. As almost no patterns can be dis-
cerned, it is most likely that Early Mesolithic ornamentation strongly suggests ‘cultural drift’,
which most readily fits with the ‘decorative’ argument of culture-specific aesthetic traditions.
This interpretation could also apply for the later Mesolithic periods, although the increasing
‘complexity’ of Kongemose and Ertebølle ornamentation problematises this interpretation
for the corpus as a whole. One specific cluster of Ertebølle motifs has been demonstrated to
be statistically significant, and the mutual co-occurrences of these five motifs are likely to
have been meaningful in the Mesolithic. It is argued that this anomalous Ertebølle cluster is
likely to reflect a change in function of the portable art. This cluster most strongly suggests a
‘socio-political’ function within the ‘Kula ring’ area of Central Denmark, a phenomenon pre-
viously demonstrated by George Nash (1998). Thus, this dissertation presents a furthering of
Nash’ argument, demonstrating a spatio-temporally limited pattern of regionality and neigh-
bour-to-neighbour contact with portable art as its material correlate – not only regarding
specific motifs, but also motif combinations.
Two central issues persist in the research on this material: 1), the evident subjectivity and
ambiguity in the classification of portable art motifs, and 2), the coarse chronological resolu-
tion of the ornamented objects. Płonka’s contribution provides the most comprehensive
overview to date, and scholars ought to consistently refer to his work when analysing the
portable art of Mesolithic Europe. However, his motif classification should not be taken to
reflect Mesolithic reality but needs to be continuously reassessed and perhaps exploratively
tweaked. The second, and perhaps most pressing, issue is the coarse chronological resolu-
tion, hindering a more detailed understanding of the portable art phenomenon. Therefore,
direct radiocarbon dating of ornamented objects is the best possibility for future
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
91
clarification. Such dating is entirely possible as well, as the great majority of ornamented ar-
tefacts are of bone or antler, lying dormant in South Scandinavian archaeological storage fa-
cilities. This is the most promising avenue for further research on the portable art of the
South Scandinavian Mesolithic.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
92
References
ALLEY, R.B. & A.M. ÁGÚSTSDÓTTIR. 2005. The 8k event: cause and consequences of a major
Holocene abrupt climate change. Quaternary Science Reviews 24: 1123–49.
ANDERSEN, K., S. JØRGENSEN & J. RICHTER. 1982. Maglemose hytterne ved Ulkestrup Lyng.
Vol. 7 (Nordiske fortidsminder, Serie B). København: Det Kongelige Nordiske
Oldskriftselskab.
ANDERSEN, S.H. 1981. Ertebøllekunst. Nye østjyske fund af mønstrede Ertebølleoldsager.
Kuml 29: 7–59. https://doi.org/10.7146/kuml.v29i29.107185.
—. 1995. Ringkloster: Ertebølle trappers and wild boar hunters in eastern Jutland. A survey.
Journal of Danish archaeology 12: 13–59.
—. 2000. ‘Køkkenmøddinger’ (Shell Middens) in Denmark. A survey. Proceedings of the Pre-
historic Society 66: 361–84. https://doi.org/10.1017/S0079497X00001857.
—. 2011. Ertebølle canoes and paddles from the submerged habitation site of Tybrind Vig,
Denmark, in J. Benjamin, C. Bonsall, C. Pickard & A. Fischer (ed.) Submerged pre-
history: 1–14. Oxford: Oxbow Books.
ANDERSEN, S.H. & M.F. MORTENSEN. 2022. Maglemosekøller af urokseknogler. Aarboeger
for Nordisk Oldkyndighed og Historie 2018: 7–16.
ANDERSEN, S.H. & A.L. TROLLE. 1997. En mønstret pragtøkse fra ældre Ertebølletid. Kuml 41:
9–28.
ANDRESEN, M. & H. ZINSMEISTER. 2017. The benefit of syntactic vs. linear n-grams for lingui-
stic description., in Proceedings of the Fourth International Conference on Depen-
dency Linguistics: 4–14. Depling.
ASTRUP, P.M. 2018. Sea-level change in Mesolithic Southern Scandinavia: long-and short-
term effects on society and the environment. Vol. 106 (Jutland Archaeological Soci-
ety Publications). Jutland Archaeological Society in cooperation with Moesgaard Mu-
seum.
—. 2021. A drowned Mesolithic shell midden complex at Hjarnø Vesterhoved, Denmark and
its wider significance. Quaternary Science Reviews 258: 106854.
https://doi.org/10.1016/j.quascirev.2021.106854.
AZÉMA, M. 2008. Representation of movement in the Upper Palaeolithic: An ethological ap-
proach to the interpretation of parietal art. Anthropozoologica 43: 117–54.
AZÉMA, M. & F. RIVÈRE. 2012. Animation in Palaeolithic art: a pre-echo of cinema. Antiquity
86: 316–24.
BACON, B., A. KHATIRI, J. PALMER, T. FREETH, P. PETTITT & R. KENTRIDGE. 2023. An Upper
Palaeolithic Proto-writing System and Phenological Calendar. Cambridge Archaeo-
logical Journal, 1–19. https://doi.org/10.1017/S0959774322000415.
BAHN, P.G. 1987. Comment to: Art for Art’s Sake in the Paleolithic by John Halverson. Cur-
rent Anthropology 28: 72–73.
BAHN, P.G. & J. VERTUT. 1997. Journey through the ice age. Berkeley: University of Califor-
nia Press.
BAILEY, G. 2008. Mesolithic Europe: overview and new problems, in G. Bailey & P. Spikins
(ed.) Mesolithic Europe: 357–72. Cambridge: Cambridge University Press.
BAILEY, G., S.H. ANDERSEN & T.J. MAARLEVELD. 2020. Denmark: Mesolithic coastal land-
scapes submerged, in G. Bailey, G. Galanidou, H. Peeters, H. Jöns & M. Mannenga
(ed.) The Archaeology of Europe’s Drowned Landscapes, 35: 39–76 (Coastal Re-
search Library). Springer, Cham.
BALLIN, T.B. 2017. Rising waters and processes of diversification and unification in material
culture: the flooding of Doggerland and its effect on north-west European prehistoric
populations between ca. 13000 and 1500 cal BC. Journal of Quaternary Science 32:
329–39.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
93
BAO, Y., T. YANG, X. LIN, Y. FANG, Y. WANG, E. PÖPPEL & Q. LEI. 2016. Aesthetic Preferences
for Eastern and Western Traditional Visual Art: Identity Matters. Frontiers in Psy-
chology 7: 1596.
BARTON, C.M., G. CLARK & A. COHEN. 1994. Art as Information: Explaining Upper Palaeo-
lithic Art in Western Europe. World Archaeology 26: 185–207.
https://doi.org/10.1080/00438243.1994.9980272.
BEDNARIK, R.G. 2003. ‘Prehistoric’ ‘art’: a response to Bowyer. La Pintura 29: 9.
BENNIKE, O., T.Z.T. JENSEN, A.J. TAUROZZI, M. MACKIE, L. CLAUDI-HANSEN & B. MAGNUS-
SEN. 2021. A Mid-Holocene reindeer antler from Regstrup, Sjælland, Denmark. Dan-
ish Journal of Archaeology 10: 1–8.
BENOIT, K., K. WATANABE, H. WANG, P. NULTY, A. OBENG, S. MÜLLER & A. MATSUO. 2018.
quanteda: An R package for the quantitative analysis of textual data. Journal of Open
Source Software 3: 774. https://doi.org/10.21105/joss.00774.
BENTLEY, R.A., M.W. HAHN & S.J. SHENNAN. 2004. Random drift and culture change. Pro-
ceedings of the Royal Society of London. Series B: Biological Sciences 271. The Royal
Society: 1443–50.
BERGSVIK, K.A. 2003. Mesolithic ethnicity – too hard to handle?, in L. Larsson, H. Kindgren,
K. Knutsson, D. Loeffler & A. Åkerlund (ed.) Mesolithic on the move. Papers pre-
sented at the Sixth International Conference on the Mesolithic in Europe, Stockholm,
2000: 278–89.
BIAGI, P. & E. STARNINI. 2016. The Origin and Spread of the Late Mesolithic Blade and Tra-
peze Industries in Europe: Reconsidering J. G. D. Clark’s Hypothesis Fifty Years Af-
ter. Interactions, changes and meanings. essays in honour of igor manzura on the
occasion of his 60th birthday edited by stanislav ţerna and blagoje govedarica ki-
shinev 2016, 33–45.
BLANKHOLM, H.P. 2008. Southern Scandinavia, in G. Bailey & P. Spikins (ed.) Mesolithic Eu-
rope: 107–31. Cambridge: Cambridge University Press.
BOETHIUS, A. & T. AHLSTRÖM. 2018. Fish and resilience among Early Holocene foragers of
southern Scandinavia: A fusion of stable isotopes and zooarchaeology through Bayes-
ian mixing modelling. Journal of Archaeological Science 93: 196–210.
https://doi.org/10.1016/j.jas.2018.02.018.
BOKELMANN, K. 1999. Zum Beginn des Spätmesolithikums in Südskandinavien. Geweihaxt,
Dreieck und Trapez, 6100 cal BC. Offa 56: 183–92.
BORUP, P. 2008. Storfangeren. Skalk 5: 16–17.
BOUISSAC, P. 1994a. Art or script? A falsifiable semiotic hypothesis. Semiotica 100: 349–68.
https://doi.org/10.1515/semi.1994.100.2-4.349.
—. 1994b. Introduction: A challenge for semiotics. Semiotica 100: 99–108.
https://doi.org/10.1515/semi.1994.100.2-4.99.
BOWYER, W., J. 2002. On defining prehistoric ‘art’. La Pintura 29: 8.
BREIVIK, H.M. 2014. Palaeo-oceanographic development and human adaptive strategies in
the Pleistocene–Holocene transition: A study from the Norwegian coast. The Holo-
cene 24: 1478–90.
BREUIL, H. 1907. L’évolution de l’art pariétal des cavernes de l’âge du Renne (Congrès inter-
national d’anthropologie et d’archéologie préhistoriques). Monaco: Imprimerie de
Monaco.
BROHOLM, H.C. 1928. Langøfundet. En Boplads fra den ældre Stenalder paa Fyn. Aarboeger
for Nordisk Oldkyndighed og Historie 1928: 129–90.
BRØNDSTED, J. 1957. Danmarks Oldtid. 2nd ed. Vol. 1. København: Gyldendal.
BURDUKIEWICZ, J. 2014. The origin of symbolic behavior of Middle Palaeolithic humans: Re-
cent controversies. Quaternary International 326: 398–405.
https://doi.org/10.1016/j.quaint.2013.08.042.
CARTAILHAC, É. & H. BREUIL. 1906. La caverne d’Altamira à Santillane près Santander
(Espagne). Vol. 1. Monaco: Imprimerie de Monaco.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
94
CHRISTENSEN, C. 2001. Kystbosættelse og havniveauændringer i stenalderen, in O.L. Jensen,
S.A. Sørensen & K.M. Hansen (ed.) Danmarks jægerstenalder-status og perspekti-
ver: beretning fra Symposiet" Status og perspektiver inden for dansk mesolitikum".
Vordingborg 1998: 183–93. Hørsholm: Hørsholm Egns Museum.
CLARK, J.G.D. 1936. The Mesolithic Settlement of Northern Europe. Cambridge University
Press.
—. 1958. Blade and trapeze industries of the European Stone Age., in Proceedings of the Pre-
historic Society, 24: 24–42. Cambridge University Press.
—. 1975. The earlier Stone Age settlement of Scandinavia. London: Cambridge University
Press.
COLEMAN, E.B. 2005. Aesthetics as a Cross-Cultural Concept. Literature and Aesthetics 15:
57–78.
CONKEY, M.W. 1978. Style and information in cultural evolution: toward a predictive model
for the Paleolithic, in C.L. Redman (ed.) Social archaeology: beyond subsistence and
dating: 61–85. New York: Academic Press.
CURRIE, G. 2011. The Master of the Masek Beds: Handaxes, Art, and the Minds of Early Hu-
mans, in E. Schellekens & P. Goldie (ed.) The Aesthetic Mind: Philosophy and Psy-
chology: 9–31. Oxford: Oxford University Press. https://doi.org/10.1093/ac-
prof:oso/9780199691517.003.0002.
—. 2016. Aesthetic Explanation and the Archaeology of Symbols. British Journal of Aesthet-
ics 56: 233–46. https://doi.org/10.1093/aesthj/ayw059.
D’ERRICO, F. 1989. Palaeolithic Lunar Calendars: A Case of Wishful Thinking? Current An-
thropology 30: 117–18. https://doi.org/10.1086/203721.
—. 1995. A New Model and its Implications for the Origin of Writing: The La Marche Antler
Revisited. Cambridge Archaeological Journal 5: 163–206.
https://doi.org/10.1017/S095977430001502X.
D’ERRICO, F. & C. CACHO. 1994. Notation versus Decoration in the Upper Palaeolithic: a
Case-Study from Tossal de la Roca, Alicante, Spain. Journal of Archaeological Sci-
ence 21: 185–200. https://doi.org/10.1006/jasc.1994.1021.
DALMERI, G. 1985. L’arte mobiliare dell’abitato tardopaleolitico-mesolitico di Terlago
(Trento). Preistoria alpina 21: 21–31.
DAVID, B. & H. LOURANDOS. 1998. Rock art and socio-demography in northeastern Austral-
ian prehistory. World Archaeology 30: 193–219.
DAVID, B. & I.J. MCNIVEN. 2018. Introduction: Towards an Archaeology and Anthropology of
Rock Art, in The Oxford Handbook of the Archaeology and Anthropology of Rock
Art: 1–21. Oxford: Oxford University Press.
DAVID, E. 2003. The contribution of the technological study of bone and antler industry for
the definition of the early Maglemose culture., in L. Larsson, H. Kindgren, K.
Knutsson, D. Loeffler & A. Åkerlund (ed.) Mesolithic on the Move. Papers presented
on the sixth international conference on the Mesolithic in Europe, Stockholm, 2000.
Oxford: Oxbow Books.
DAVIDSON, I. 1993. Review of: The Roots of Civilization: The Cognitive Beginnings of Man’s
First Art, Symbol and Notation by Alexander Marshack. American Anthropologist
95: 1027–28. https://doi.org/10.1525/aa.1993.95.4.02a00350.
DE SMEDT, J. & H. DE CRUZ. 2011. A Cognitive Approach to the Earliest Art. The Journal of
Aesthetics and Art Criticism 69: 379–89. https://doi.org/10.1111/j.1540-
6245.2011.01482.x.
DEWEZ, M.C. 1974. New hypotheses concerning two engraved bones from la grotte de re-
mouchamps, Belgium. World Archaeology 5: 337–45.
https://doi.org/10.1080/00438243.1974.9979578.
DJINDJIAN, F. 2010. Functions, meaning and symbolism of European Upper Palaeolithic ani-
mal representations., in J. Clottes (ed.) Pleistocene Art of the World, 5: 1807–16.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
95
DOLBUNOVA, E. et al. 2023. The transmission of pottery technology among prehistoric Euro-
pean hunter-gatherers. Nature Human Behaviour 7: 171–83.
https://doi.org/10.1038/s41562-022-01491-8.
DUTKIEWICZ, E., S. WOLF & N.J. CONARD. 2018. Early symbolism in the Ach and the Lone
valleys of southwestern Germany. Quaternary international 491: 30–45.
DUTKIEWICZ, E., G. RUSSO, S. LEE & C. BENTZ. 2020. SignBase, a collection of geometric
signs on mobile objects in the Paleolithic. Scientific Data 7: 364.
ELKINS, J. 1996. On the Impossibility of Close Reading: The Case of Alexander Marshack.
Current Anthropology 37: 185–226.
ELLIOTT, B. & G. WARREN. 2023. Colonialism and the European Mesolithic. Norwegian Ar-
chaeological Review, 1–19. https://doi.org/10.1080/00293652.2023.2182232.
EL-NASAN, A., S. VEERAMACHANENI & G. NAGY. 2003. Handwriting recognition using posi-
tion sensitive letter n-gram matching., in Seventh International Conference on Docu-
ment Analysis and Recognition, 2003: 577–82. IEEE.
FERUGLIO, V., C. BOURDIER, M. DELLUC, P. MORA, N. AUJOULAT & J. JAUBERT. 2019. Rock
art, performance and Palaeolithic cognitive systems. The example of the Grand Panel
palimpsest of Cussac Cave, Dordogne, France. Journal of Anthropological Archaeol-
ogy 56: 101104. https://doi.org/10.1016/j.jaa.2019.101104.
FISCHER, A. 2001. Mesolitiske bopladser på den danske havbund–udfordringer for forskning
og forvaltning. Danmarks Jægerstenalder–status og perspektiver, Hørsholm: Hørs-
holms Egns Museum, 59–74.
FISCHER, A., J. OLSEN, M. RICHARDS, J. HEINEMEIER, Á.E. SVEINBJÖRNSDÓTTIR & P. BENNIKE.
2007. Coast–inland mobility and diet in the Danish Mesolithic and Neolithic: evi-
dence from stable isotope values of humans and dogs. Journal of Archaeological Sci-
ence 34: 2125–50.
FLOSS, H. 2018. Same as it ever was? The Aurignacian of the Swabian Jura and the origins of
Palaeolithic art. Quaternary International 491: 21–29.
https://doi.org/10.1016/j.quaint.2016.12.044.
FRANCFORT, H.P., R. HAMAYON & P.G. BAHN. 2001. The concept of shamanism: uses and
abuses (Bibliotheca Shamanistica). Budapest: Akadémiai Kiadó.
FRIMAN, B. 1996. Does the Kongemose Culture exist? About the concept of culture. Archaeo-
logia Polona 34: 143–63.
GAMBLE, C. 1991. The Social Context for European Palaeolithic Art. Proceedings of the Pre-
historic Society 57: 3–15. https://doi.org/10.1017/S0079497X00004849.
GENDEL, P.A. 1984. Mesolithic social territories in Northwestern Europe (BAR International
Series 218). Oxford: BAR Publishing.
GOSDEN, C. 2001. Making Sense: Archaeology and Aesthetics. World Archaeology 33: 163–
67.
GRAEBER, D. & D. WENGROW. 2022. The Dawn of Everything: A new history of humanity.
London: Penguin Books Ltd.
GRIES, S.T. 2017. Quantitative Corpus Linguistics with R: A Practical Introduction. 2nd ed.
New York: Routledge.
GRON, K.J. & L. SØRENSEN. 2018. Cultural and economic negotiation: a new perspective on
the Neolithic Transition of Southern Scandinavia. Antiquity 92: 958–74.
GRØN, O. 2020. Late Mesolithic marine territoriality in coastal southern Scandinavia.
Archäologische Informationen 43: 277–88.
GRONENBORN, D. 2017. Migrations before the Neolithic? The Late Mesolithic blade-and-tra-
peze horizon in Central Europe and beyond., in H. Meller, F. Daim, J. Krause & R.
Risch (ed.) Migration und Integration von der Urgeschichte bis zum Mittelalter, 17:
113–27.
GROSSE, E. 1894. Die Anfänge der Kunst. Freiburg : J.C.B. Mohr.
GUTHRIE, D., B. ALLISON, W. LIU, L. GUTHRIE & Y. WILKS. 2006. A Closer Look at Skip-gram
Modelling., in Proceedings of the Fifth International Conference on Language
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
96
Resources and Evaluation: 1222–25. Genoa: European Language Resources Associa-
tion (ELRA).
GUTHRIE, R.D. 1984. Ethological observations from Palaeolithic art. Contribution de la Zool-
ogie et d’Ethologie á l’Interpretation de l’Art des Peuples Chasseurs Préhistoriques.
Editions Universitaires Fribourg Suisse, Saint-Paul Fribourg 1984: 35–74.
HALVERSON, J. 1987. Art for Art’s Sake in the Paleolithic. Current Anthropology 28: 63–89.
HARTZ, S. 1985. Kongemose Kultur in Schleswig-Holstein? Offa 12: 35–56.
—. 2009. Towards a new chronology of the Late Mesolithic in Schleswig-Holstein., in Chro-
nology and Evolution Within the Mesolithic of North-West Europe: Proceedings of
an International Meeting, Brussels, May 30th-June 1st 2007: 395–416.
HAYDEN, B. 2021. Keeping count: On interpreting record keeping in prehistory. Journal of
Anthropological Archaeology 63: 101304.
https://doi.org/10.1016/j.jaa.2021.101304.
HENSHILWOOD, C.S. et al. 2002. Emergence of modern human behavior: Middle Stone Age
engravings from South Africa. Science 295: 1278–80.
HENSHILWOOD, C.S., F. D’ERRICO & I. WATTS. 2009. Engraved ochres from the middle stone
age levels at Blombos Cave, South Africa. Journal of Human Evolution 57: 27–47.
HODGSON, D. 2000. Art, perception and information processing: an evolutionary perspec-
tive. Rock Art Research 17: 3–34.
—. 2003. Seeing the ‘Unseen’: Fragmented Cues and the Implicit in Palaeolithic Art. Cam-
bridge Archaeological Journal 13: 97–106.
https://doi.org/10.1017/S0959774303000064.
—. 2006. Understanding the origins of paleoart: The neurovisual resonance theory and brain
functioning. Paleoanthropology 2006: 54–67.
HULTÉN, E. 1939. Magiska ornament i mesoliticum? Fornvännen, 193–225.
IGARASHI, J. & H. FLOSS. 2019. Signs associated with figurative representations Aurignacian.
Examples from Grotte Chauvet and the Swabian Jura. Quaternary International
503: 200–209. https://doi.org/10.1016/j.quaint.2017.12.048.
IVERSEN, J. 1960. Problems of the Early Post-Glacial Forest Development in Denmark. Dan-
marks Geologiske Undersøgelse IV. Række 4: 1–32.
https://doi.org/10.34194/raekke4.v4.7005.
JACOBSEN, T. 2010. Beauty and the brain: culture, history and individual differences in aes-
thetic appreciation. Journal of Anatomy 216: 184–91. https://doi.org/10.1111/j.1469-
7580.2009.01164.x.
JENNBERT, K. 2011. Ertebølle pottery in southern Sweden - A question of handicraft, net-
works and creolisation in a period of neolithisation. Bericht der Römisch-Germani-
schen Kommission 89: 89–110.
JENSEN, J. 2001. Danmarks Oldtid. Stenalder 13.000-2.000 f.Kr. København: Nordisk For-
lag.
JENSEN, T.Z.T. et al. 2020. An integrated analysis of Maglemose bone points reframes the
Early Mesolithic of Southern Scandinavia. Scientific Reports 10: 1–12.
https://doi.org/10.1038/s41598-020-74258-8.
JESSEN, C.A., K.B. PEDERSEN, C. CHRISTENSEN, J. OLSEN, M.F. MORTENSEN & K.M. HANSEN.
2015. Early Maglemosian culture in the Preboreal landscape: Archaeology and vege-
tation from the earliest Mesolithic site in Denmark at Lundby Mose, Sjælland. Qua-
ternary International 378: 73–87. https://doi.org/10.1016/j.quaint.2014.03.056.
JOCHIM, M. 2011. The Mesolithic, in S. Milisauskas (ed.) European Prehistory: A Survey:
125–51 (Interdisciplinary Contributions to Archaeology). New York: Springer.
https://doi.org/10.1007/978-1-4419-6633-9_6.
KARJUS, A., M.C. SOLÀ, T. OHM, S.E. AHNERT & M. SCHICH. 2022. Compression ensembles
quantify aesthetic complexity and the evolution of visual art. arXiv.
http://arxiv.org/abs/2205.10271. https://doi.org/10.48550/arXiv.2205.10271.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
97
KARSTEN, P. & B. KNARRSTRÖM. 2001. Tågerup – Fifteen hundred years of Mesolithic occu-
pation in western Scania, Sweden: a preliminary view. European Journal of Archae-
ology 4: 165–74.
—. 2003. The Tågerup excavations. Lund: Riksantikvarieämbetet.
KEYSER, J.D. & D.S. WHITLEY. 2006. Sympathetic Magic in Western North American Rock
Art. American Antiquity 71: 3–26. https://doi.org/10.2307/40035319.
KILLIN, A. & R. PAIN. 2022. How WEIRD is Cognitive Archaeology? Engaging with the Chal-
lenge of Cultural Variation and Sample Diversity. Review of Philosophy and Psychol-
ogy 2022: 1–25. https://doi.org/10.1007/s13164-021-00611-z.
KING, A.R. 1973. Review of: The Roots of Civilization: The Cognitive Beginnings of Man’s
First Art, Symbol and Notation by Alexander Marshack. American Anthropologist
75: 1897–1900.
KLASSEN, L. 1999. Prestigeøkser af sjældne alpine bjergarter: En glemt og overset fund-
gruppe fra ældre stenalders slutning i Danmark. Kuml 13: 11–52.
KOERPER, H.C. & E.G. STICKEL. 1980. Cultural drift: a primary process of culture change.
Journal of Anthropological Research 36: 463–69.
KRISTIANSEN, K. 2012. Rock art and religion: the sun journey in Indo-European mythology
and Bronze Age rock art. Adoranten 2012: 93–115.
LAMING-EMPERAIRE, A. 1969. Pour une nouvelle approche des sociétés préhistoriques. Anna-
les 24: 1261–69. https://doi.org/10.3406/ahess.1969.422133.
—. 1970. Système de pensée et organisation sociale dans l’art rupestre paléolithique.
L’homme de Cro-Magnon. Anthropologie et Archéologie 1968: 197–212.
LARSSON, L. 1990. The Mesolithic of Southern Scandinavia. Journal of World Prehistory 4:
257–309. https://doi.org/10.1007/BF00974882.
—. 1995. Man and sea in Southern Scandinavia during the Late Mesolithic. The role of ceme-
teries in the view of society, in A. Fischer (ed.) Man & Sea in the Mesolithic - Coastal
Settlement Above and Below Present Sea Level: 95–106 (Oxbow Monograph 53). Ox-
ford: Oxbow Books.
—. 2000. Expression of art in the Mesolithic societies of Scandinavia, in A. Butrimas (ed.)
Prehistoric art in the Baltic region: 31–61. Vilnius: Vilnius Academy of Fine Arts.
—. 2017. Animals and animal depictions in the early Holocene of Northern Europe: Based on
an antler adze with deer depictions., in V. Brieske, A. Dickers & M.M. Rind (ed.) Tiere
und Tierdarstellungen in der Archäologie: 63–73.
LARSSON, L. & F. MOLIN. 2017a. Decoration of bone and antler artefacts as an indication of
Mesolithic networks (finds from a central Swedish Late Mesolithic site). Самарский
научный вестник 6: 85–91.
—. 2017b. Symbols in the Late Mesolithic : Ornaments on bone and antler from Strandvägen,
Motala, in Central Sweden, in M. Margarit & A. Boroneant (ed.) From hunter-gather-
ers to farmers. Human adaptations at the end of the Pleistocene and the first part of
the Holocene. Papers in Honour of Clive Bonsall: 397–408. Editura Cetatea de
Scaun.
LARTET, E. & H. CHRISTY. 1875. Reliquiae Aquitanicae: Being Contributions to the Archaeol-
ogy and Palaeontology of Périgord and the Adjoining Provinces of Southern France.
London: Williams & Norgate.
LAYTON, R. 2001. Hunter-gatherers, their neighbours and the nation state, in C. Panter-
Brick, R. Layton & P. Rowley-Conwy (ed.) Hunter-gatherers: An interdisciplinary
perspective: 292–321. Cambridge: Cambridge University Press.
—. 2006. Structuralism and Semiotics, in C. Tilley, S. Kuechler-Fogden & W. Keane (ed.)
Handbook of Material Culture: 29–42. London: SAGE Publications Ltd.
LAYTON, R. & P. ROWLEY-CONWY. 2013. Wild things in the north? Hunter-gatherers and the
tyranny of the colonial perspective. Anthropologie (1962-) 51: 213–30.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
98
LEROI-GOURHAN, A. 1958. La fonction des signes dans les sanctuaires paléolithiques. Bulletin
de la Société préhistorique française 55: 307–21.
https://doi.org/10.3406/bspf.1958.3664.
—. 1968. L’expérience ethnologique. Encyclopédie de La Pléiade 1968: 1816–25.
—. 1982. The Dawn of European Art: An Introduction to Palaeolithic Cave Painting. Cam-
bridge: Cambridge University Press.
LEWIS-WILLIAMS, D. 1987. Comment to: Art for Art’s Sake in the Paleolithic by John Halver-
son. Current Anthropology 28: 79–80.
LEWIS-WILLIAMS, D.J. & J. CLOTTES. 1998. The mind in the cave—The cave in the mind: Al-
tered consciousness in the Upper Paleolithic. Anthropology of Consciousness 9: 13–
21.
LEWIS-WILLIAMS, J.D. 1997. Harnessing the brain: vision and shamanism in Upper Palaeo-
lithic Western Europe. Beyond art: Pleistocene image and symbol 23: 321–42.
LEWIS-WILLIAMS, J.D. et al. 1988. The Signs of All Times: Entoptic Phenomena in Upper Pal-
aeolithic Art [and Comments and Reply]. Current Anthropology 29: 201–45.
LITTAUER, M.A., F.D. MCCARTHY & A. MARSHACK. 1974. On Upper Paleolithic Engraving.
Current Anthropology 15: 327–32.
LORBLANCHET, M. 1989. From man to animal and sign in Palaeolithic art. Animals into art
1989: 109–43.
LYELL, C. 1863. The Geological Evidence of the Antiquity of Man. London: J.M. Dent and
Sons.
MAGLI, G. 2023. Statistics in the Study of Palaeolithic Art: The Case of the Alleged ‘Phenolog-
ical Calendar’. SSRN Electronic Journal. https://doi.org/10.2139/ssrn.4395377.
MALINOWSKI, B. 1922. Argonauts of the Western Pacific. New York: EP Dutton & Co.
MANNINEN, M.A., V. ASHEICHYK, T. JONUKS, A. KRIISKA, G. OSIPOWICZ, A.N. SOROKIN, A.
VASHANAU, F. RIEDE & P. PERSSON. 2021. Using Radiocarbon Dates and Tool Design
Principles to Assess the Role of Composite Slotted Bone Tool Technology at the Inter-
section of Adaptation and Culture-History. Journal of Archaeological Method and
Theory 28: 845–70. https://doi.org/10.1007/s10816-021-09517-7.
MARING, R. & F. RIEDE. 2019. Possible wild boar management during the Ertebølle period. A
carbon and nitrogen isotope analysis of Mesolithic wild boar from Fannerup F, Den-
mark. Environmental Archaeology 24: 15–27.
MARSHACK, A. 1972a. Cognitive Aspects of Upper Paleolithic Engraving. Current Anthropol-
ogy 13: 445–77.
—. 1972b. Upper Paleolithic Notation and Symbol: Sequential microscopic analyses of Mag-
dalenian engravings document possible cognitive origins of writing. Science 178:
817–28.
—. 1985. A Lunar-Solar Year Calendar Stick from North America. American Antiquity 50:
27–51. https://doi.org/10.2307/280632.
—. 1991a. The Taï plaque and calendrical notation in the Upper Palaeolithic. Cambridge Ar-
chaeological Journal 1: 25–61.
—. 1991b. The Roots of Civilization: the Cognitive Beginnings of Man’s First Art, Symbol
and Notation. New York: Moyer Bell Limited.
MARSHACK, A. & F. D’ERRICO. 1989. On Wishful Thinking and Lunar ‘Calendars’. Current
Anthropology 30: 491–500.
MATHIASSEN, T. 1943. Stenalderbopladser i Aamosen (Nordiske Fortidsminder 3). Køben-
havn: Nordisk Forlag.
MATZIG, D.N., S.T. HUSSAIN & F. RIEDE. 2021. Design space constraints and the cultural tax-
onomy of European Final Palaeolithic large tanged points: a comparison of typologi-
cal, landmark-based and whole-outline geometric morphometric approaches. Jour-
nal of Paleolithic Archaeology 4: 27.
MITHEN, S.J. 1988. Looking and learning: Upper Palaeolithic art and information gathering.
World Archaeology 19: 297–327.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
99
MIYAGAWA, S., C. LESURE & V.A. NÓBREGA. 2018. Cross-Modality Information Transfer: A
Hypothesis about the Relationship among Prehistoric Cave Paintings, Symbolic
Thinking, and the Emergence of Language. Frontiers in Psychology 9.
https://www.frontiersin.org/articles/10.3389/fpsyg.2018.00115.
MORO ABADÍA, O. 2006. Art, crafts and Paleolithic art. Journal of Social Archaeology 6: 119–
41. https://doi.org/10.1177/1469605306060571.
MORO ABADÍA, O. & M.R. GONZÁLEZ MORALES. 2007. Thinking About ‘Style’ in the ‘Post-Sty-
listic Era’: Reconstructing the Stylistic Context of Chauvet. Oxford Journal of Ar-
chaeology 26: 109–25. https://doi.org/10.1111/j.1468-0092.2007.00276.x.
—. 2013. Paleolithic Art: A Cultural History. Journal of Archaeological Research 21: 269–
306. https://doi.org/10.1007/s10814-012-9063-8.
MORO ABADÍA, O. & E. PALACIO-PÉREZ. 2015. Rethinking the Structural Analysis of Palaeo-
lithic Art: New Perspectives on Leroi-Gourhan’s Structuralism. Cambridge Archaeo-
logical Journal 25: 657–72. https://doi.org/10.1017/S0959774315000086.
MÜLLER, S. 1896. Nye stenalders former. København: Det Kongelige Nordiske Oldskriftsel-
skab.
—. 1918. Oldtidens kunst i Danmark. Vol. 1. København: CA Reitzel.
MUNRO, R. 1906. Man as Artist and Sportsman in the Palæolithic Period. Proceedings of the
Royal Society of Edinburgh 25: 92–128.
https://doi.org/10.1017/S0370164600008361.
NASH, G. 2001a. Altered States of Consciousness and the Afterlife: A Reappraisal on a Deco-
rated Bone Piece from Ryemarksgaard, Central Zealand, Denmark. BAR Interna-
tional Series 937: 231–40.
NASH, G.H. 1998. Status, Exchange and Mobility: Portable Art in the Scandinavian Meso-
lithic (BAR International Series 710). Oxford: Archaeopress.
—. 2001b. Expressing sexuality, fertility and childbirth in art: engendering south Scandina-
vian Mesolithic portable artefacts, in L. Bevan (ed.) Indecent Exposure - sexuality,
society and the archaeological record: 19–42. Glasgow: Cruithne Press.
NEEDHAM, A., I. WISHER, A. LANGLEY, M. AMY & A. LITTLE. 2022. Art by firelight? Using ex-
perimental and digital techniques to explore Magdalenian engraved plaquette use at
Montastruc (France). PLOS ONE 17: e0266146. https://doi.org/10.1371/jour-
nal.pone.0266146.
NEWELL, R.R., D. KIELMANN, T.S. CONSTANDSE-WESTERMANN, W.A.B. VAN DER SANDEN & A.
VAN GIJN. 1990. An Inquiry Into the Ethnic Resolution of Mesolithic Regional
Groups: The Study of Their Decorative Ornaments in Time and Space. Leiden: E.J.
Brill.
NIMURA, C. 2013. 'Five feet high and rising’: Environmental crises and the art of prehistoric
Scandinavia. (ed.)E.M. Van der Milt & J.M. Jimenéz Tough Times: The Archaeology
of Crisis and Recovery (BAR International Series), 23–41.
NYLAND, A.J. & H. STEBERGLØKKEN. 2020. Changing perceptions of rock art: storying prehis-
toric worlds. World Archaeology 52: 503–20.
https://doi.org/10.1080/00438243.2021.1899042.
ONETTO, M. & M.M. PODESTÁ. 2011. Cueva de las Manos: An outstanding example of a rock
art site in South America. Adoranten 2011: 67.
ORŁOWSKA, J. & G. OSIPOWICZ. 2021. Accuracy of the typological classifications of the Late
Glacial and Early Holocene osseous projectile points according to the new AMS dates
of selected artifacts from Poland. Archaeological and Anthropological Sciences 14: 8.
https://doi.org/10.1007/s12520-021-01483-1.
OVERMANN, K.A. 2013. Material scaffolds in numbers and time. Cambridge Archaeological
Journal 23: 19–39.
PALACIO-PÉREZ, E. 2010. Cave Art and the Theory of Art: The Origins of the Religious Inter-
pretation of Palaeolithic Graphic Expression. Oxford Journal of Archaeology 29: 1–
14. https://doi.org/10.1111/j.1468-0092.2009.00337.x.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
100
PETERSEN, E.B. 1982. Jæger og bytte, et 8–9000 år gammel billede fra Åmosen. Museet for
Holbæk og Omegn 1981: 47–54.
—. 1990. Nye grave fra jægerstenalderen: Strøby Egede og Vedbæk. Nationalmuseets ar-
bejdsmark 1990: 19–33.
—. 1993. Ældre stenalder, in S. Hvass & B. Storgaard (ed.) Da klinger i Muld... 25 års arkæo-
logi i Danmark: 46–49. København: Det Kongelige Nordiske Oldskriftselskab.
—. 2006. Cultural and social landscapes of Mesolithic Vedbæk, Denmark., in C.-J. Kind (ed.)
After the ice age: 15–31. Stuttgart: Konrad Theiss Verlag.
—. 2015. Diversity of mesolithic vedbæk. Vol. 1 (Acta Archaeologica 86). Oxford: Wiley.
PETERSEN, E.B. & C. MEIKLEJOHN. 2007. Historical Context of the Term ‘Complexity’. Acta
Archaeologica 78: 181–92. https://doi.org/10.1111/j.1600-0390.2007.00105.x.
PETTITT, P., L.-E. MEYERING & R. KENTRIDGE. 2020. Bringing science to the study of ancient
senses - archaeology and visual psychology. World Archaeology 52: 183–204.
https://doi.org/10.1080/00438243.2020.1909932.
PHILIPPSEN, B., L. IVANOVAITĖ, K. MAKHOTKA, F. SAUER, F. RIEDE & J. OLSEN. 2019. Eight
New Late Pleistocene/Early Holocene AMS Dates from the Southeastern Baltic. Ra-
diocarbon 61: 615–27. https://doi.org/10.1017/RDC.2018.153.
PICKARD, C. & C. BONSALL. 2007. Late Mesolithic coastal fishing practices, in B. Hårdh, K.
Jennbert & D. Olausson (ed.) On the Road: Studies in Honour of Lars Larsson: 176–
83. Lund: Almqvist & Wiksell.
PŁONKA, T. 2003. The portable art of Mesolithic Europe (Acta Universitatis Wratislaviensis
2527). Wrocław: Wydawn. Uniw. Wrocławskiego.
—. 2021. Net patterns in Mesolithic art of north-western Europe, in D. Borić, D. Antonović &
B. Mihailović (ed.) Foraging Assemblages, 2: 595–601.
PŁONKA, T. & K. KOWALSKI. 2017. Symbolic artefacts in the Late Palaeolithic and early Meso-
lithic in Northern and Central Europe, in T. Płonka & K. Kowalski (ed.) Rusinowo.
The symbolic culture foragers in the Late Palaeolithic and the Early Mesolithic.
Wrocław: University of Wrocław: 163–81.
PORR, M. 2010. Palaeolithic Art as Cultural Memory: a Case Study of the Aurignacian Art of
Southwest Germany. Cambridge Archaeological Journal 20: 87–108.
https://doi.org/10.1017/S0959774310000065.
POVLSEN, K. 2013. The introduction of ceramics in the Ertebølle Culture. Danish Journal of
Archaeology 2: 146–63.
PRICE, T.D. 1985. Affluent Foragers of Mesolithic Southern Scandinavia, in T.D. Price & J.A.
Brown (ed.) Prehistoric Hunters-Gatherers: 341–63.
—. 2020. Mesolithic Complexity: A Complicated Question. Mesolithic Miscellany 28: 3–15.
PRICE, T.D. & J.A. BROWN. 1985. Aspects of Hunter–Gatherer Complexity, in T.D. Price &
J.A. Brown (ed.) Prehistoric Hunters-Gatherers: 3–20.
PRICE, T.D. & A.B. GEBAUER. 1992. The final frontier: Foragers to farmers in southern Scan-
dinavia, in T.D. Price & A.B. Gebauer (ed.) Transitions to agriculture in prehistory:
97–116.
PRICE, T.D. & A.B. GEBAUER (ed.). 2005. Smakkerup Huse: a late mesolithic coastal site in
Northwest Zealand, Denmark. 1st ed. Aarhus: Aarhus University Press.
PRUVOST, M. et al. 2011. Genotypes of predomestic horses match phenotypes painted in
Paleolithic works of cave art. Proceedings of the National academy of sciences 108:
18626–30.
R CORE TEAM. 2021. R: A Language and Environment for Statistical Computing. (version
4.2.2). Vienna: R Foundation for Statistical Computing. https://www.R-project.org/.
R: Create n-grams and skip-grams from tokens. 2023. Accessed May 24. https://search.r-
project.org/CRAN/refmans/quanteda/html/tokens_ngrams.html.
RAMACHANDRAN, V. & W. HIRSTEIN. 1999. The Science of Art: A Neurological Theory of Aes-
thetic Experience. Journal of Consciousness Studies 6: 15–51.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
101
RAPHAEL, M. 1945. Prehistoric cave paintings (Bollingen Series 4). Washington: Pantheon
Books.
RAPPENGLÜCK, M.A. 2015. Possible Astronomical Depictions in Franco-Cantabrian Paleo-
lithic Rock Art, in C.L.N. Ruggles (ed.) Handbook of Archaeoastronomy and Ethno-
astronomy: 1205–12. New York: Springer.
RASMUSSEN, G.H. 1989. Okkergrave fra ældre stenalder på Djursland. Kuml 36: 31–42.
https://doi.org/10.7146/kuml.v36i36.110860.
RASMUSSEN, K.L. 2013. Radiocarbon dates from Tybrind Vig, in S.H. Andersen (ed.) Tybrind
Vig: Submerged Mesolithic settlements in Denmark: 363–64. Højbjerg: Jysk
Arkæologisk Selskab.
REINACH, S. 1903. L’art et la magie à propos des peintures et des gravures de l’âge du renne.
L’Anthropologie 14: 257–66.
REYNOLDS, N. & F. RIEDE. 2019. Reject or revive? The crisis of cultural taxonomy in the Eu-
ropean Upper Palaeolithic and beyond. Antiquity 93: 1368–70.
https://doi.org/10.15184/aqy.2019.156.
RICHTER, J., T. HAUCK, R. VOGELSANG, T. WIDLOK, J.-M. LE TENSORER & P. SCHMID. 2012.
“Contextual areas” of early Homo sapiens and their significance for human dispersal
from Africa into Eurasia between 200 ka and 70 ka. Quaternary International 274:
5–24. https://doi.org/10.1016/j.quaint.2012.04.017.
ROBINSON, J. 1992. Not counting on Marshack: A reassessment of the work of Alexander
Marshack on notation in the Upper Palaeolithic. Journal of Mediterranean Studies
2: 1–17.
ROSENFELD, A. 1971. Alexander Marshack: Notation dans les gravures du paléolithique supé-
rieur. Antiquity 45: 317–19. https://doi.org/10.1017/S0003598X00069969.
ROWLEY-CONWY, P. 1995. Meat, furs and skins: Mesolithic animal bones from Ringkloster, a
seasonal hunting camp in Jutland. Journal of Danish Archaeology 12: 87–98.
SAHLINS, M. 1972. Stone age economics. Chicago: Aldine-Atherton.
SAUVET, G. & A. WLODARCZYK. 2008. Towards a Formal Grammar of the European Palaeo-
lithic Cave Art. Rock Art Research 25: 165–72. https://doi.org/10.3316/ie-
lapa.485811180638079.
SAUVET, G., R. BOURRILLON, D. GARATE, S. PETROGNANI, O. RIVERO, E. ROBERT & G. TO-
SELLO. 2018. The function of graphic signs in prehistoric societies: The case of Canta-
brian quadrilateral signs. Quaternary International 491: 99–109.
https://doi.org/10.1016/j.quaint.2017.01.039.
SCHILLING, H. 2003. Early Mesolithic settlement patterns in Holmegård Bog on South Zea-
land, Denmark. A social perspective., in L. Larsson, H. Kindgren, K. Knutsson, D.
Loeffler & A. Åkerlund (ed.) Mesolithic on the Move. Papers presented at the Sixth
International Conference on the Mesolithic in Europe, Stockholm 2000: 351–58. Ox-
ford: Oxbow Books.
SCHLETTE, F. 1969. Ornament oder Symbol? Zu den Anfangen ornamentaler gestaltung.
Jahresschrift fur Mitteldeutsche Vorgeschichte 53: 215–40.
SCHMITT, L. 2007. The West Swedish Hensbacka from an Anthropological Point of View, and
Recent Developments Concerning the Final Drainage of the Baltic Ice Lake, in M. Ko-
busiewicz & J. Kabaciński (ed.) Studies in the Final Palaeolithic Settlement of the
Great European Plain: 139–50. Poznan: Polish Academy of Sciences.
SHENNAN, S. 2011. Descent with modification and the archaeological record. Philosophical
transactions of the royal society B: Biological sciences 366: 1070–79.
SIDOROV, G. 2013. Non-linear construction of n-grams in computational linguistics. México:
Sociedad Mexicana de Inteligencia Artificial.
SKRIVER, C., P.M. ASTRUP & P. BORUP. 2018. Hjarnø Sund–all year, all inclusive. A sub-
merged Late Mesolithic coastal site with organic remains. Danish Journal of Archae-
ology 7: 195–217.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
102
SOLOMON, A. 2019. Rock Arts, Shamans, and Grand Theories, in B. David & I.J. McNiven
(ed.) The Oxford Handbook of the Archaeology and Anthropology of Rock Art: 565–
86. Oxford: Oxford University Press. https://doi.org/10.1093/ox-
fordhb/9780190607357.013.51.
SØRENSEN, L. 2014. From Hunter to Farmer in Northern Europe - Migration and Adaptation
During the Neolithic and Bronze Age. Acta Archaeologica 85 1: 1–276.
https://doi.org/10.1111/j.1600-0390.2014.00917.x.
—. 2015. Hunters and farmers in the North – the transformation of pottery traditions and
distribution patterns of key artefacts during the Mesolithic and Neolithic transition in
southern Scandinavia, in J. Kabaciński, S. Hartz, D.C.M. Raemaekers & T. Terberger
(ed.) The Dąbki Site in Pomerania and the Neolithisation of the North European
Lowlands (c. 5000–3000 calBC): 385–432 (Archaeologie Und Geschichte Im Ost-
seeraum). Verlag Marie Reidorf.
SØRENSEN, M. 2006. Teknologiske traditioner i Maglemosekulturen. En diakron analyse af
Maglemosekulturens flækkeindustri, in B.V. Eriksen (ed.) Stenalderstudier. Tidligt
mesolitiske jægere og samlere i Sydskandinavien: 19–76. Højbjerg: Jysk Arkæolo-
gisk Selskab.
—. 2012. The arrival and development of pressure blade technology in Southern Scandinavia,
in P.M. Desrosiers (ed.) The emergence of pressure blade making: from origin to
modern experimentation: 237–59. Boston: Springer.
SØRENSEN, M., T. RANKAMA, J. KANKAANPÄÄ, K. KNUTSSON, H. KNUTSSON, S. MELVOLD, B.V.
ERIKSEN & H. GLØRSTAD. 2013. The first eastern migrations of people and knowledge
into Scandinavia: evidence from studies of Mesolithic Technology, 9th-8th Millen-
nium BC. Norwegian archaeological review 46: 19–56.
SØRENSEN, M., H. LÜBKE & D. GROß. 2018. The Early Mesolithic in Southern Scandinavia
and Northern Germany, in N. Milner, C. Conneller & B. Taylor (ed.) Star Carr: A per-
sistent place in a changing world, 1: 305–29. York: White Rose University Press.
SØRENSEN, S.A. 1987. A Maglemosian Hut at Lavringe Mose, Zealand. Journal of Danish Ar-
chaeology 6: 53–62.
—. 2007. Limhamn axes in Denmark, in B. Hårdh, K. Jennbert & D. Olausson (ed.) On the
Road: studies in honour of Lars Larsson: 184–87. Lund: Almqvist & Wiksell.
SØRENSEN, S.A. 2016. Syltholm: Denmark’s largest Stone age excavation. Mesolithic Miscel-
lany 24: 3–10.
SØRENSEN, S.A. 2017. The Kongemose Culture (Nordiske Fortidsminder, Serie B). Odense:
University Press of Southern Denmark.
SØRENSEN, S.A. 2020. Ritual depositions in the coastal zone: A case from Syltholm, Den-
mark, in A. Schülke (ed.) Coastal Landscapes of the Mesolithic, 1st ed.: 394–414.
London: Routledge.
SREENATHAN, M., V.R. RAO & R.G. BEDNARIK. 2008. Palaeolithic cognitive inheritance in
aesthetic behavior of the Jarawas of the Andaman Islands. Anthropos 103: 367–92.
STEJSKAL, J. 2021. Aesthetic Archaeology. Critical Inquiry 48: 144–66.
https://doi.org/10.1086/715981.
TAMARIZ, M. & S. KIRBY. 2015. Culture: Copying, Compression, and Conventionality. Cogni-
tive Science 39: 171–83. https://doi.org/10.1111/cogs.12144.
TAYLOR, B. 2021. Lunar Timekeeping in Upper Paleolithic Cave Art. Praehistoria New Series
3: 215–32.
TEDESCO, L.A. 2007. Introduction to Prehistoric Art, 20,000–8000 B.C. The Metropolitan
Museum of Art, The Met’s Heilbrunn Timeline of Art History. https://www.met-
museum.org/toah/hd/preh/hd_preh.htm.
TERBERGER, T. 2006. Mesolithic group territories of the Northern lowlands in discussion., in
C.-J. Kind (ed.) After the Ice age. Settlements, subsistence and social development in
the Mesolithic of central Europe, 78: 221–35. Stuttgart: Konrad Theiss Verlag.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
103
THORPE, I.J. 2003. Anthropology, archaeology, and the origin of warfare. World Archaeol-
ogy 35: 145–65.
TILLEY, C. 1991. Material Culture and Text: The Art of Ambiguity. London: Routledge.
TOFT, P.A. 2005. What goes where? Intrasite studies of depositions of Maglemosian art and
flint picks of the Maglemose and Kongemose cultures., in S. Mccartan, R. Schulting,
G. Warren, P. Woodman & P. García-Argüelles (ed.) Mesolithic Horizons. Papers
Presented at the Seventh International Conference on the Mesolithic in Europe, Bel-
fast: 614–20. Oxford: Oxbow Books.
TOFT, P.A. 2006. Hvor går grænsen? Et studium af sociale territorier og ornamentik i Magle-
mosekulturen, in B.V. Eriksen (ed.) Stenalderstudier. Tidlig mesolitiske jagere og
samlere i Sydskandinavien: 101–19. Højbjerg: Jysk Arkæologisk Selskab.
—. 2017. Regionalism and Mobility in Early Mesolithic art?, in M. Sørensen & K.B. Pedersen
(ed.) Problems in Paleolithic and Mesolithic Research: 259–74 (Arkæologiske Stud-
ier 12). København: The Saxo Institue, University of Copenhagen & Museum of
Southeast Denmark.
TOFT, P.A. & E.B. PETERSEN. 2015. Five Thousand Years of Decorated Amber Pendants from
the Danish Mesolithic. Kunde Neue Folge 64: 197–218.
TOMOVIĆ, A., P. JANIČIĆ & V. KEŠELJ. 2006. n-Gram-based classification and unsupervised
hierarchical clustering of genome sequences. Computer Methods and Programs in
Biomedicine 81: 137–53. https://doi.org/10.1016/j.cmpb.2005.11.007.
TRATMAN, E.K. 1976. A late upper palaeolithic calculator (?), Gough’s Cave, Cheddar, Somer-
set. Proc. Univ. Bristol Spalaeol. Soc. 14: 123–29.
TSIRINTOULAKI, K., D.N. MATZIG & F. RIEDE. 2023. A 2D Geometric Morphometric Assess-
ment of Chrono-Cultural Trends in Osseous Barbed Points of the European Final Pal-
aeolithic and Early Mesolithic. Open Archaeology 9: 20220276.
TYLÉN, K., R. FUSAROLI, S. ROJO, K. HEIMANN, N. FAY, N.N. JOHANNSEN, F. RIEDE & M. LOM-
BARD. 2020. The evolution of early symbolic behavior in Homo sapiens. Proceedings
of the National Academy of Sciences 117: 4578–84.
VANG PETERSEN, P. 1982. Jægerfolket på Vedbækbopladserne. Kulturudviklingen i Konge-
mose-og Ertebølletid. Nationalmuseets Arbejdsmark 1982: 179–89.
—. 1984. Chronological and regional variation in the late Mesolithic of eastern Denmark.
Journal of Danish archaeology 3: 7–18.
—. 1990. Eksotiske faunarester i Kongemose- og Ertebølletid : et resultat af udveksling?
Hikuin 16.
—. 2013. Amber pendants, bears and elks. Kunde 64: 219–37.
—. 2018. Ravfigurer og zigzagmønstre – jægerkunst fra istiden. Nationalmuseets Arbejds-
mark 2018: 136–51.
—. 2021. Zigzag lines and other protective patterns in Palaeolithic and Mesolithic art. Qua-
ternary International 573 (The Shigir Idol in the Context of Early Art in Eurasia):
66–74. https://doi.org/10.1016/j.quaint.2019.09.029.
VON PETZINGER, G. 2017. The First Signs: Unlocking the Mysteries of the World’s Oldest
Symbols. New York: Atria Books.
VON PETZINGER, G. & A. NOWELL. 2011. A question of style: reconsidering the stylistic ap-
proach to dating Palaeolithic parietal art in France. Antiquity 85: 1165–83.
https://doi.org/10.1017/S0003598X00061986.
WARREN, G. 2005. Complex arguments..., in N. Milner & P. Woodman (ed.) Mesolithic Stud-
ies at the Beginning of the 21st Century: 69–80. Oxford: Oxbow Books.
WIESSNER, P. 1983. Style and Social Information in Kalahari San Projectile Points. American
Antiquity 48: 253–76. https://doi.org/10.2307/280450.
WILLERSLEV, R. 2007. Soul Hunters: Hunting, Animism, and Personhood Among the Sibe-
rian Yukaghirs. University of California Press.
WISHER, I. 2022. The Origins of Visual Culture? Psychological Foundations of Upper Palaeo-
lithic Figurative Cave Art in Northern Spain. PhD thesis, Durham University.
Mesolithic Semiotics?
MA dissertation / Aarhus University
Lasse Lukas Platz Herskind
104
WOBST, H.M. 1978. The archaeo-ethnology of hunter-gatherers or the tyranny of the ethno-
graphic record in archaeology. American antiquity 43: 303–9.
WOODBURN, J. 1982. Egalitarian societies. Man, 431–51.
XU, L., G. LIANG, C. LIAO, G.-D. CHEN & C.-C. CHANG. 2019. k-Skip-n-Gram-RF: A Random
Forest Based Method for Alzheimer’s Disease Protein Identification. Frontiers in Ge-
netics 10. https://doi.org/doi: 10.3389/fgene.2019.00033.
