THE GENRE OF TROLLS

THE GENRE OF TROLLS

The Case of a Finland-Swedish Folk Belief Tradition

Camilla Asplund Ingemark

Akademisk avhandling
som med tillstnd av
Humanistiska fakulteten vid bo Akademi
framlggs till offentlig granskning i Auditorium Armfelt, Arken,
Fabriksgatan 2, bo, fredagen den 28 januari 2005 kl. 12

bo Akademis Frlag bo Akademi University Press

bo 2004

THE GENRE OF TROLLS

The Case of a Finland-Swedish Folk Belief Tradition

Camilla Asplund Ingemark

bo Akademi University Press

bo 2004

bo Akademi University Press

& Camilla Asplund Ingemark 2004
Typographic design and prepress: Pr Sandin
Cover: Tove Ahlbck
Cover illustration: Emma Rnnholm
Printed at Ekens tryckeri AB, Ekens
ISBN 951-765-222-4

CIP Cataloguing in Publication

Asplund Ingemark, Camilla
The genre of trolls: The case of a Finland-Swedish folk belief tradition /
Camilla Asplund Ingemark. bo : bo Akademi University Press, 2004.
Diss.: bo Akademi University.
ISBN 951-765-222-4

PREFACE
I have greatly enjoyed writing this thesis, not least because of the many stimulating discussions I have had with colleagues and friends along the way.
Naturally, I have also incurred many debts of gratitude, the creditors of
which I hope I have faithfully listed below. I sincerely apologize for any
omissions or oversights.
The first set of thanks goes to my supervisor, Professor Ulrika WolfKnuts, who has encouraged me from the very start. Her unfailing devotion
to her students is remarkable, and I am grateful that I have been able to
benefit from it. She has read every draft of my dissertation, quite regardless
of what condition it was in, with speed and acumen, and with many annoying questions as a result, but I do not doubt that these have made the
manuscript more easily legible and the arguments more convincing. Her
knowledge of Finland-Swedish folk belief and the religious situation in
19th-century Ostrobothnia has been particularly valuable in the preparation
of the thesis, and she has liberally shared her insights with me during the
years.
I also owe Dr Lena Marander-Eklund many thanks. When I worked on
my M. A. thesis she functioned as my supervisor for a term, and during this
time she managed to introduce me to no less than two of the theories I am
utilizing in this book: Lotte Tarkkas theory of intertextuality, and Charles
Briggs and Richard Baumans theory of genre. I guess neither of us realized
in what direction these theories would take my work, but that is the charm
of doing research, after all. During the years she has also readily supplied
me with whatever archive material I have needed, and I am grateful for this
as well.
Moreover, I wish to express my gratitude to those colleagues who have
assisted me during my stays abroad: Professor Inger Lvkrona, the Department of Ethnology at Lund University, who took care of me for a term in
Lund; at the School of Celtic and Scottish Studies, Scottish Ethnology
Section at the University of Edinburgh, I benefitted from the generosity of
Dr Margaret Mackay who acted as my supervisorI am especially grateful
for the advice on finding English translations of internationally well-known
hymns, which posed a real problem for me. I also thank Dr John Shaw
who invited me to give a speech at the seminar of the School of Scottish
vii

Studies, and Dr Neill Martin for his encouragement. Jan Adams and Marie
Hamilton assisted me with many practical matters, for which I am grateful.
My fellow Ph.D students at the School deserve a special mention as well: I
enjoyed our post-seminar pub visits to Sandy Bells.
A number of scholars have kindly commented on various stages of the
manuscript. Dr Sven-Erik Klinkmann, bo Akademi University, gave
many erudite comments on an early draft of my chapter on intertextual
theory, and I confess I have not been able to follow up on all of them.
Dr Laura Stark, University of Helsinki, made provocative readings of
chapters 4 and 5, and also accepted the task of acting as preliminary examiner of the whole text. Once again, it has not been possible for me to take
all her feed-back into consideration, but the finished product has definitely improved because of it. Dr Martina Bjrklund, Section for Russian
Language and Literature, bo Akademi University, scrutinized my discussions on Bakhtin with zeal and enthusiasm, and I have heeded much of
her advice on formal matters as well.
Members of the folkloristic seminar at bo Akademi University, as well
as of the joint seminar of the science of religion and folklore, have given
many useful contributions during the years, both in terms of the structure
of the text, and of its contents. The discussions have always been characterized by knowledge, skill and grace, and the post-seminars afterwards have
been pleasant occasions. I am grateful for the generosity and patience that
have been accorded me.
I also want to thank the members of The Graduate School for Cultural
Interpretations, too numerous to mention individually, who have given
feedback on my presentations at the meetings of the school. It has always
given me food for thought, and I extend my sincerest thanks for the effort
expended in trying to improve my thesis. The social gatherings arranged in
connection with these meetings have given me the opportunity to get to
know colleagues in the whole of Finland better, and this has been a gratifyingand indubitably plannedspin-off effect.
A number of people have sent copies of archive material to me when I
needed it most, and for this I thank them: Dr Susanne sterlund-Ptzsch
and Dr Carola Ekrem, both providing me with material from the Folklore
Archives of the Swedish Literature Society in Finland, and M. A. Sofie
Strandn who copied records from the Folklore Archives at bo Akademi
University. Living abroad, I have been dependent on the kindness of my
viii

colleagues to obtain much of my research material, and I truly appreciate

the enthusiasm and rapidity with which it has been put at my disposal.
Elizabeth Nyman accepted the task of correcting my English before the
book went to print, and I thank her for this. Dr Pr Sandin kindly took
care of the technical editing of the text, for which I am very grateful. Professor Charles Lock, Department of English at the University of Copenhagen, acted as my second preliminary examiner, giving important corrections to the text. He is also an inspiration in his extensive and innovative
research on Bakhtin.
This study could not have been carried out without the generous financial support of the following: The Committee for Folklore of the Swedish
Literature Society in Finland; The Research Institute of bo Akademi
University; Waldemar von Frenckells Foundation; Chancellor Lars Erik
Taxells Fund, bo Akademi University; The Swedish-Ostrobothnian
Association; The berg Fund, The Swedish Foundation for Culture; The
Graduate School for Cultural Interpretations; and The Victoria Foundation.
I thank bo Akademi University Press for accepting my thesis for publication; Inger Hassel and Kristina Toivonen guided me in the practicalities
of finding a printer, Tove Ahlbck designed the book cover on the basis of
the excellent drawing made by Emma Rnnholm, and Anne Andersson
took care of the CIP cataloguing of the dissertation. I appreciate the work
of all of you.
Finally, I wish to thank my family and friends. My parents, Bengt and
Kristina Asplund, always encouraged me to read and write, and this is where
it got me. I have greatly enjoyed the journey. My sister, Linda Asplund,
has shared my interest in the bizarre. My grandmother, Birgit Asplund,
came to the rescue when I needed information on the parish of Vr, her
native parish. My uncle and aunt, Bror Rnnholm and Margareta WillnerRnnholm, have invited me to stay in their home every time I have been in
bo, and kept me sober and down-to-earth with the voices of experience.
In Finland, Olivia Granholm, Susanna stman, Viveca Rabb and Anette
Johansson have been agreeable companions. In Sweden, Martina and David
Finnskog, Henrik Gerding and Rebecka Randler, Elisabet and Anders
Gransson, Kristian Gransson and Maria Mellgren, Oskar Hagberg and
Shirley Nslund, Mi Lennhag, Bjrn Levander, Pr Sandin, Kristiina Savin
and Jonas Hansson, Aron Sjblad, Claes Schuborg and Karin Staffans,
Joachim Walewski, Per stborn and others have guided my thoughts to
ix

other things than intertextuality, Bakhtin and folk belief, for which I am
indeed grateful. My in-laws, Ingrid Ingemark, Thomas Dellans, Anna
Ingemark and Peter and David Milos have eased the load with their kindness and great humour.
My husband, Dominic Ingemark, has managed the impossible: to be
supportive, inspiring and a source of many insightful comments, without
being allowed to read the manuscript. I am also grateful for the forbearance
with which he has tolerated my frequently late nights of work, and my trips
to Finland at occasionally very inopportune moments.

Lund, November the 25th, 2004

Camilla Asplund Ingemark

TABLE OF CONTENTS
Preface

vii

Introduction

1.1
1.2
1.3
1.4
1.4.1
1.4.2
1.4.3
1.4.4
1.4.5
1.5

Statement of Purpose
Delimitations and Definitions
Trolls in the History of Research
Intertextuality in the History of Research
Intertextuality
Interdiscursivity
Intergenericity
Cultural Intertextuality
Subjective Intertextuality
Method

1
6
13
21
22
30
33
37
38
42

Material and Context

46

2.1
2.2
2.2.1
2.2.2

46
47
47

2.2.4
2.2.5
2.3
2.3.1
2.3.2

General Considerations
The Sources
The Rancken Collection (R) and Its Contributors
The Collections of the Swedish Literature Society
in Finland (SLS)
The Collectors of the Swedish Literature Society and
Their Interviewees
Printed Sources
On the Principles of Transcription and Translation
Context
Historical and Social Context
Religious Context

Description of the Troll Tradition

86

3.1
3.1.1
3.1.2

The Conditions of Encounter

The Place and Time of the Encounter
Womens Encounter with the Troll

86
86
88

2.2.3

50
55
65
70
71
71
76

xi

3.1.3
3.1.4
3.2
3.2.1
3.2.2
3.3
3.3.1
3.3.2
3.3.3
3.3.4
3.3.5
3.3.6
3.4
3.4.1
3.4.2
3.4.3
3.4.4
3.4.5
3.4.6
3.5

Mens Encounter with the Troll

Childrens Encounter with the Troll
The Troll and Its World
The Troll and Its Abilities
The Dwelling and Possessions of the Troll
Interaction between the Realms
Conflicts
Tension-Filled Tolerance
From Tolerance to Conflict
Tolerance and Conflict
From Conflict to Tolerance
The Troll and Christianity
Breaking the Contact
Men Dissociate
Women, Children and Animals Dissociate
Trolls Dissociate
Impersonal Phenomena Dissociate
Protective and Apotropaic Measures
The Fateful Encounter
Encountering the Troll

90
92
94
94
96
98
98
106
112
115
117
118
120
120
124
126
127
129
132
135

Intertextuality as Ideological Critique

137

4.1
4.2

Blindness and Illumination

The Outline of an Intertextual Network

143
176

Intertextuality as Social Critique

181

5.1
5.2

The Sins of an Exorcist

Intertextuality, Interdiscursivity and Power

181
206

Genre, Parody, Chronotopes and Novelization:

The Wonder Tales of Johan Aln

218

Genre
Parody
Chronotopes

219
233
245

6.1
6.2
6.3

xii

6.4
6.5

Novelization
Integrating the Perspectives

248
250

The Problems of Unfinalizability and Dialogue

253

7.1
7.2
7.3
7.4
7.5

Introduction
The Terrors of Unfinalizability
Halting Unfinalizability?
Aborted Dialogues
Unfinalizability, Dialogue and Stories of the
Supernatural

253
257
269
274

Discussion

280

Abbreviations and Bibliography

286

275

Appendix A: The Recorded Narratives of Johan Aln

301

Appendix B: Select Narratives of Johan Aln

302

Appendix C: Select Troll Narratives

314

xiii

1 INTRODUCTION
1.1 Statement of Purpose
The supernatural tradition that is the object of this study, the Swedish troll
tradition in Finland as documented in archived material collected between
the 1850s and 1925, has previously received little scholarly attention; apart
from a few articles (e.g. Lnnqvist 1996), this is the first monograph produced on the subject, and I will therefore begin my inquiry with a description of the folklore of trolls for the benefit of readers with a comparative
interest. The troll in the Swedish oral tradition in Finland is a supernatural
creature primarily associated with hills and rocks in the forest, but apart
from that, it is difficult, if not impossible, to give a good definition of the
troll. It is often a solitary being, but it may also live with others of its kind.
Judging by the sources at my disposal, it is not chiefly an empirical being
I have not been able to find reports of sightings or personal experience
stories, apart from a legend told in the first person singular, but that seems
to be more of a narrative strategybut this impression may well be false
due to the haphazard nature of collection. Perhaps reports of sightings
never happened to be recorded, even though they existed. Any definite
conclusion cannot be drawn on the basis of the recorded material alone.
My basic research problem can be thus formulated: how do the performers, of whose narratives we have some form of transcript, construct the image of the troll, and how is the relation between man and troll represented
in the texts? These questions recur in many guises throughout the thesis,
and I find them important because they imply an examination of the world
view of the narrators, and of what it means to be human in a world also
inhabited by extra-human forces. The description of the troll tradition is
divided into the following sections, roughly corresponding to the temporal
frame of encounter: 3.1 The Conditions of Encounter focuses on the
time and place of the encounter, and on the agent traversing the boundary
between this world and the otherworld. The conditions and distinguishing
characteristics of womens, mens and childrens encounters with the troll
are also considered. 3.2 The Troll and Its World discusses the appearance and abilities of the troll, its world and surroundings. 3.3 Interaction
between the Realms describes the relations between man and troll, both
hostile and friendly. The attitude of the troll to Christianity is explored as
Statement of Purpose

well. Finally, 3.4 Breaking the Contact contemplates the agent effecting
the dissociation of the human and supranormal sphere, and the means
through which it is achieved, protective and apotropaic measures included.
A special study of a peculiar form of encounter, here called the fateful encounter, is appended to this chapter. Individual records will be quoted as
examples. Hence chapter 3 deals with the construction of the image of the
troll, and of the relationship between man and troll, on a descriptive level.
In chapters 4, 5, 6 and 7, the problem of the construction of the image of
the troll and the representation of the relation between man and troll is
approached on three levels. On the first level, explored in chapters 4 and 5,
I examine the texts and discourses out of which the portrait of the troll is
woven. In other words, I am undertaking a study of the web of intertextual
relations between different troll texts, between troll texts and other folklore
texts, and between troll texts and Biblical narratives. Other scholars have
addressed the problem of the relationship between religion and folklore
(see e.g. Bringus 1997; Granberg 1971; Wolf-Knuts 1991; Wolf-Knuts 2000;
Dundes (1999) is exceptional in that the author discusses the Bible as folklore), and my contribution to this debate centres on the more wide-ranging
implications of my research approach. For example, I will argue that scholars need to pay attention to the ways in which Christianity influences folk
narrative and folk belief beneath the ostensibly pre- or extra-Christian surface of traditional stories, because religion helps to shape these narratives
from within by furnishing intertexts for them, from the Bible for instance.
I will be adapting the theory of intertextuality proposed by Julia Kristeva
and reworked by the Finnish folklorist Lotte Tarkka (for definitions of
terms utilized in this thesis, see also chapter 1.2; for discussions of concepts, see chapter 1.4). In her pioneering essay Le mot, le dialogue et le
roman, Kristeva construed any text as a mosaic of quotations, any text is
the absorption and transformation of another (Kristeva 1978: 8485): a
writer constructs his text in relation to an earlier literary corpus. Tarkka
aligns herself with this definition in stating that intertextuality refers to the
idea of the text as a meeting point of different texts, where different points
of view intermingle and collide. By the same token, intertexts are the other
texts giving the individual text its meaning (Tarkka 1993: 171). For this
reason, no text is simple and uncomplicated, it has many layers that a conscientious analyst should be aware of and strive to discover. Tarkka does

Introduction

not say this explicitly, but it is the logical and methodological consequence
of the theory, and in her own research, she abides by this rule.
I have chosen Tarkkas theory since it is one of the few extensive elaborations of intertextuality in Nordic folklore research, and because it is particularly adapted to the needs of folkloristic scholarship. It is also a moderate
approach, respecting the research traditions of folkloristics as well as those
of literary theory. In other words, it is a balanced view of folkloristic intertextuality, but nevertheless with some in-built deviations from earlier praxis.
One of the most fascinating aspects of Tarkkas theory is her conception
of metaphor and metonymy. Metaphor links two separate spheres, likened
because of their similarities and contrasted due to their differences, while
metonymy mediates between the oppositions of the metaphor (Tarkka
1994: 293294); Tarkkas use of these terms differs from the conventional
one, even though it does have an affinity with established definitions.
Metaphor usually denotes a word or phrase employed instead of another, as
a comparison intended to achieve a more striking effect, while metonymy
commonly implies the substitution of a word for another that is intimately
connected with it, such as the use of the crown to refer to the monarchy.
Tarkkas notion of metaphor is an extension of the common usage, but that
of metonymy represents a significant reworking of the concept. The link
to common usage is nevertheless present in a similar stress on the proximity created between the poles of the metaphor through metonymy.
Through the use of common themes and epithets, a series of metaphors is
created, constituting levels of world view (ibid.). In my analysis, I will apply the notion of levels of metaphorical relations to two groups of texts on
abduction and the exorcism of trolls, collected in the parish of Vr in
Ostrobothnia.
In my view, the singular achievement of Tarkka in devising her theory is
that she is able to provide the researcher with a powerful tool for investigating the narrators network of mental associations. I will use it to gain
new and exciting insights into Swedish-language folk culture in Finland at
the turn of the last century. Intertextual relations will be examined in terms
of agreement, inversion or reversal, and negation of the intertext.
On the second level, dealt with in chapter 6, I intend to investigate the
generic components out of which the image of the troll might be constructed, and how the manipulation of these elements can change the

Statement of Purpose

conception of the troll.1 The objects of analysis are two texts from the repertoire of a single narrator, the carpenter Johan Aln hailing from the village of Rejpelt in the parish of Vr. The definition of genre utilized here
is that presented by Charles Briggs and Richard Bauman in their article
Genre, Intertextuality, and Social Power (1992). Genres are viewed as
generalized or abstracted models of discourse production and reception
mediated through the relationship with prior discourse (Briggs & Bauman
1992: 147). Through genre, narrators may shape speech into ordered, unified and bounded texts with strong social and historical associations,
though the invocation of genre also renders texts fragmented, heterogeneous and open-ended because of the dependence on other discursive formations and contextual factors for the interpretation, production and reception of discourse (Briggs & Bauman 1992: 147149). Briggs and Bauman
emphasize the role of the narrator in shaping and reconfiguring genres, and
they introduce the notion of intertextual gaps, which can be minimized or
maximized, to describe the process of connecting an utterance to a generic
model. Minimization of the distance between texts and genres makes the
discourse maximally interpretable through reference to generic precedents,
while maximization is associated with various motives for distancing oneself from textual precedents (Briggs & Bauman 1992: 149). In contrast to
many earlier contributions to the folkloristic debate on genre,2 Briggs and
Bauman focus on how genres actually work, not on how they should be defined or on their source-critical value, whether they are useful or deplorable
concepts, whether emic or etic categories should be used, or whether generic designations ought to be employed in the classification of folklore in
tradition archives. This is of particular import in an analysis of the intertextual constitution of genre.
The question of genre is also actualized in relation to parody, of which
the two texts are specimens. Parodies are sometimes cited as prime examples of intertextuality (Dentith 2000: 56) due to their overt connection to
another, or several other, texts or to a genre. My hypothesis is that these
1

Here I am using the adjective generic to refer to genre, as Charles Briggs and Richard
Bauman have done.
2 See e.g. Abrahams 1976a; Abrahams 1976b; Honko 1968; Honko 1971; Honko 1976;
Honko 1981; Honko 1989; Ben-Amos 1976a; Ben-Amos 1976b; Ben-Amos 1992; Klintberg
1981; Alver 1967; Dgh 1976; Dgh & Vzsonyi 1976; Lthi 1976; von Sydow 1971a; von
Sydow 1971b; von Sydow 1971c.

Introduction

narratives are parodies of the genre of the wondertale, and that this entails
some fundamental changes to their structure, e.g. to the chronotope of the
stories. Mikhail Bakhtin regarded the chronotope, the intrinsic connectedness of temporal and spatial relationships that are artistically expressed in
literature, as a formally constitutive category of literature with a profound
impact on genre and generic distinctions, and on the image of man
(Bakhtin 1986a: 8485). In narratives of trolls, the image of the supernatural is subject to a certain number of restrictions; for example, trolls
should not marry humans and live happily ever after, nor should they be
allowed to live permanently in the human world. A change in these chronotopes constituting the image of the troll may influence genre. One form
of change of the chronotope is novelization, a term coined by Bakhtin to
denote the transference of novelistic features, such as indeterminacy,
openendedness and contact with the present, to other genres where they
are usually absent (Bakhtin 1986a: 7). I believe these features are also to be
found in folklore, particularly in jocular tales, a genre which was the speciality of this performer (for a selection of his narratives, see Appendix B).
Novelization might illuminate the process of the reconfiguration of genres
spoken of by Briggs and Bauman. One aspect of the novel stressed by
Bakhtin in his book on Dostoevsky is the introduction of the unfinalizable
hero into the novel. Unfinalizability refers to the indeterminacy and openendedness of a character, to a character who is evolving, outgrowing his
former bounds. In the case of Johan Alns tales, this indeterminacy is a
positive value.
On the third level, discussed in chapter 7, the relationship between man
and troll is scrutinized with the aid of two of Bakhtins favourite concepts,
unfinalizability and dialogue. This constitutes a reconsideration of Bakhtins
notion of unfinalizability in the context of narratives of the supernatural,
viewed as a genre. Hence I am broadening the scope of the inquiry into
generic concerns by reviewing the texts included in my material in the light
of the concept of unfinalizability, which I think might be useful in explicating the construction of texts depicting encounters with the supernatural.
Stories of such encounters appear to rely on the indeterminacy and unfinalizability of the supranormal beings for suspense and for the efficacy of the
narrative; from the perspective of the characters, though, unfinalizability is
not necessarily an unequivocal blessing. Therefore, there is a conflict between the demands of the story/genre and those of the characters, and it is
Statement of Purpose

the tension and interplay between these points of view that I wish to elucidate. In his early works (Bakhtine 1984; Bakhtin 1993), Bakhtin pondered
the relation between self and other in terms of finalization, the bestowal of
form, rather than unfinalizability, which he regarded as the consequence of
poor art. These early formulations of the problem of intersubjectivity are a
fair complement to the later theories of polyphony and dialogue with their
unbridled celebration of indeterminacy. Thus I will analyze the role of unfinalizability on the one hand and finalization on the other in the construction of the relationship between man and troll, taking the differing requirements of the genre and of the characters into account.
An important aspect of Bakhtins conception of unfinalizability is the assumption of a dialogical position in relation to the hero on the part of the
author. Such a position entails the adoption of a very open attitude to the
characters which allows them to develop freely within the narrative. Hence
dialogue is employed in a restricted sense, as it implies a willingness to listen unconditionally to the other. The performer of a narrative of trolls may
be presumed to exhibit such an attitude to his supranormal characters by
permitting them to be themselves, and this hypothesis needs to be verified, but I also want to examine whether the human characters in the text
engage in dialogue with the troll in this specific sense, and what it might
suggest for the interpretation of the relationship between humans and
supernatural creatures.
To summarize, I hope to demonstrate how theories of intertextuality and
genre taken together may serve to highlight the creation of images of the
supernatural in narrative. Finally, I want to stress that the interpretations
presented in this thesis are my own, based on my individual store of knowledge, and that others, including the narrators I am studying, might well
see, or have seen, other connections than I am able to perceive. Similarly,
the conclusions I draw on the basis of these interpretations are my own,
and they are custom-made for the material I am utilizing. Hence, even
though I believe they have some degree of general applicability, I also think
they have to be tested in each individual case.

1.2 Delimitations and Definitions

The present study is geographically, linguistically and temporally delimited;
it is primarily focused on the provinces of Nyland (Uusimaa), Southwestern
6

Introduction

Finland and Ostrobothnia, as well as the land Islands, in other words, on

those areas in Finland with Swedish-speaking inhabitants.3 This is because
it is the Swedish-language tradition of trolls in Finland I am examining. In
terms of time the investigation spans some seventy years, from the 1850s to
1925 (for the methodological implications of this fact, see chapter 1.5).
So how is a troll to be defined? The best answer to that question might
be that it cannot be defined (cf. Stattin 1992: 1819), but this has not stopped
scholars from trying. Elisabeth Hartmann makes a distinction between the
Eastern Scandinavian (Danish and Swedish) and the Western Scandinavian
(Norwegian) conceptions of trolls. She characterizes the Norwegian trolls
as solitary and fictitious beings, basically synonymous with the riese (she
uses the German spelling and not the Norwegian rise), which she regards
as a purely aetiological being. The Scandinavian forms of riesenthe
Norwegian jutul (sing.), jtnar (plural), gygr (fem.), the Swedish jttar
(plural) and the Danish kjmper (plural) sharply distinguish themselves
from empirical beings, according to Hartmann. She divides the conception
of the riese into two parts, one based on faith, the other entirely fictive, and
these intermingle in actual practice (Hartmann 1936: 4749, 51). The
Norwegian legend troll is generally of great stature and grotesquely ugly
(Hartmann 1936: 52). The Eastern Scandinavian trolls, on the other hand,
are social, empirical beings corresponding to the huldrefolk in Norwegian
folk belief. This is especially true of the Danish and South Swedish conception of trolls. In the former case, the term trold is rarely utilized, since
bjrgfolk is preferred to indicate a group of beings taking an intermediate
position between Swedish trolls on the one hand, and Norwegian huldrefolk
and Swedish vttar on the other. Hartmann likens the trolls of Southern
Sweden to those of Danish tradition, and identifies them with vttar. In
the north of Sweden, from Dalecarlia and Hlsingland northward, the
Eastern Scandinavian tradition reigns, and the limits of this tradition area
Swedish-speakers have been living in modern-day Finland at least since the 12th14th
centuries when the land was incorporated into the kingdom of Sweden (for the latest discussion of this issue, see Ivars & Huldn 2002). In the following I will be using Swedish
place-names when referring to areas inhabited by Swedish-speaking people, or to places
with a Finnish-speaking population but having Swedish names as well. It may also be
noted that the extent to which Finland-Swedish narrators were fluent in, or even knew,
Finnish should not be overestimated. Therefore a concentration on Swedish-language traditions is in order.

Delimitations and Definitions

are the same as those of the conception of vittra (the Norwegian huldra).
In Western Sweden the Eastern Scandinavian and Western Scandinavian
traditions blend into each other, even though they retain the grotesque
traits of the Norwegian tradition. The Central Swedish conception of trolls
exists in Gtaland, Southern and Central Svealand, Smland, Northern
Blekinge, stergtland, Eastern Vstergtland, Vstmanland, Nrke,
Sdermanland, Uppland, Gstrikland and Swedish-speaking Finland.
Hartmann describes the Central Swedish trolls as a group of creatures
largely corresponding to the Danish bjrgfolk and the Norwegian huldrefolk,
as well as to the conception of vttar. The latter were regarded as more
peaceful and well-disposed neighbours, while the trolls received a more
ominous, dangerous stamp since they lived in the wilderness and in the
mountains. The evil-mindedness of the trolls was evidenced in their abductions of cattle and humans alike. The Central Swedish trolls dwell in
mountains, those in Southern Sweden inhabit mounds, and Norwegian
trolls live in the high mountains. The looks of the troll may be pictured in
various ways, but its ugliness is a common feature. In Central Sweden the
troll can be of human height, whereas the Northern and Western parts of
the country favour huge trolls (Hartmann 1936: 6065).
Jan-jvind Swahn essentially agrees with Elisabeth Hartmann in his article in the Swedish Nationalencyklopedin. He notes that the term troll has
different significations in Swedish (including Finland-Swedish), Danish
and Norwegian tradition, and that the trolls of folk belief, and therefore of
legends, were envisioned as anthropomorphic, collective supranormal beings inhabiting mountains, barrows or woods in Central and South Swedish
folklore. They were thought to live their lives much as humans did, and in
some legends they live in peace with man, though they are mostly depicted
as thieving and dangerous. Conceptions of changelings and abductions
were the most important ones connected to them, and these motifs can also
be found in the Celtic world; this was one of Elisabeth Hartmanns main
points. Belief in trolls must be distinguished from that in witches. The
trolls of folktales are identical with giants, and their role is to serve as
supernaturally great and dangerous opponents to the heroes. The solitary
trolls associated with impressive natural formations in legends tend to have
characteristics drawn from both legends and folktales (Swahn 1995: 431).
With regards to the Norwegian trolls, Virginie Amilien describes them
as supernatural creatures, big and ugly, functioning as the opponents of the
8

Introduction

human heroes; the troll is the typical image of evil in Norwegian narratives.
Since the Christianization of Norway it has been related to the Devil. The
troll is a symbol of not only the power of evil, but also of the forces of
nature. The distinction between the troll and the jutul rests on the latters
mythical association: the jutul is linked to the past, connoting a temporal
reference, and its importance in legends springs from this fact. The tusse
makes rare appearances in folk narratives. Originally a creature of cosmogonic significance, the Old Norse thurs, the race to which Ymer belonged,
turned into an evil and naive personage, often depicted as a short man with
a white beard. The rise has preserved its link to giant dimensions. In contradistinction to Hartmann, Amilien argues that the rise has progressively
distinguished itself from other supranormal beings, but unfortunately she
does not expand on the subject. In later Norwegian tradition the gyger may
be the wife of the troll, playing the part of a secondary opponent, or she
may be the principal opponent, great and terrible. Employed synonymously with gyger, the hulder in one sense of the word is a man-eating, horrible giantess. The term may also designate a creature haunting the hills
and woods, or the family of the subterranean people, the huldrefolk, which
are viewed more positively, although the latter can function as both helpers
and opponents (Amilien 1996: 3542). Since Amilien restricts herself to
folktales, the relation between trolls and huldra and huldrefolk remains
indistinct, and it is difficult to compare her opinion on the subject with
Hartmanns.
If we juxtapose these accounts of the idiosyncrasies and mores of the
trolls to what can be gathered from my own material, the following observations can be made. Very little can be said of the size of the troll (cf. chapter 3.2.1): only rarely is it described as terribly large (SLS 31, 141: 111) or
something to that effect (SLS 65: 45). Most of the time its physique is not
mentioned at all. Thus, one cannot claim with certainty that trolls are huge
and ugly. They might be humanlike as in Central Sweden, the traditional
area to which Swedish-speaking Finland also belongs, in which case there
might be no need to specify their appearance. The trolls usually live in hills
in the woods; in that respect the definitions agree with my material. They
can be social as well as solitary beings, one of the few traits demarcating
them from the r, which is generally solitary. The forest is the home of
both the troll and the r, and both are equally notorious for abducting
humans or their cattle. However, the r is not commonly associated with
Delimitations and Definitions

the practice of changing babies; here the Devil (SLS 166: 687689; SLS 37:
7071), the earthdwellers (SLS 333: 208210) and the brownies (SLS 166:
728729) are the trolls fellow culprits, in addition to nameless creatures
(Finlands 1931: 147148).
No personal experience narratives of trolls have been recorded, but this
might be the result of the vicissitudes of collection. If we relate this fact to
the rest of the material on supernatural beings, we can see that personal
experience stories are rather scarce in general. There are a few narratives
about encounters with the r, for instance, but the number of such texts is
marginal. Their inclusion in the extant collections may be as much attributed to chance as to any conscious design, and it is possible that collectors
were biased against, or simply uninterested in, personal experience narratives. Accordingly, I cannot make any pronouncement on the empirical or
fictitious nature of the troll in the Swedish tradition in Finland.
I have chosen not to make any wide-ranging comparisons with Finnishlanguage and international traditions (concerning the former, see note 3 in
this chapter); it is my hope that anyone interested in this aspect of the troll
tradition in the Swedish-speaking districts in Finland will find what he
needs in order to draw his own conclusions.4 Similarly, I have declined to
consider the extent of belief in trolls; as Linda Dgh and Andrew Vzsonyi
have shown us, belief is a volatile condition (Dgh & Vzsonyi 1976), and
the contextual information is too scant to give any indication of belief or
non-belief in any case.
A choice I regret I have been forced to make, but that I nevertheless
deem necessary, is the limited attention I have been able to devote to intertexts from the field of wider folk belief, in chapters 4 and 5 in particular. I
have felt it more urgent to point to relations with religious tradition, which
is the prime contribution of this thesis to the study of folk belief, than to
provide an exhaustive account of the belief context to which the material
examined belonged.
In the present study the terms narrative, text, intertextuality and intertext, dialogue and discourse will be liberally employed. By narrative I mean
the narration of a series of events, involving a process of communication in
which the narrative is told by a performer to a recipient using verbal means
4

He and his will be employed to denote any anonymous person, and both men and women are included.

10

Introduction

(cf. Rimmon-Keenan 2001: 2). Story, and to some extent text, are employed synonymously with narrative. Text is also understood to be characterized by the connectedness of its components and the concepts underlying them (cohesion and coherence). It is constructed by the performer,
on conscious as well as unconscious levels, and its production is related to
the surrounding situational, social and cultural context. The text is a system in which each component is vital for the functioning of the system (cf.
Bjrklund 1993: 21), and in addition, it is an intertextual phenomenon
connecting communicative speech to other types of anterior or synchronic
utterances. The text is therefore a productivity, implying that: 1) it redistributes language, i.e., it changes and transgresses both linguistic and logical
categories; 2) it is an intertextuality, i.e., a permutation of texts: within the
space of the text several utterances drawn from other texts cross and neutralize one another (Kristeva 1978: 52). The intertexts are consequently the
utterances absorbed into and transformed in the text (Kristeva 1978: 8485).
One detail in the explication of intertextuality above is objectionable,
however, and that is the notion of intertexts neutralizing each other. Then
the tension, the dialogue between the utterances constituting the text would
disappear, and a significant component of its productivity would vanish.
Kristevas inspiration in devising the concept of intertextuality, Mikhail
Bakhtin, used the word dialogue in a number of different, but related
senses; I will only refer to those relevant to my own study (see also chapter
1.4.1). Firstly, dialogue exists within the word, as any word we utter has
been pronounced by others before us, imbuing it with the views, shared
thoughts, value judgements and accents of others. Our own appropriation
of the word enters into complex interrelationships of association, dissociation and intersection with those alien elements, which influences the actualization of the word (Bakhtin 1986a: 276). Secondly, there is dialogue
between points of view or discourses within the same utterance, hybridization. By this Bakhtin meant the fusion of the discourse of the author
with the discourse of the narrator, the implicit author or the characters
within a single proposition, so that the person from whose point of view
the text is structured cannot be pin-pointed (Bakhtin 1986a: 301308). The
conception of perspectives or discourses in dialogue has been assimilated
into folkloristic research (see e.g. Tarkka 1994: 251, 263265, 295). Thirdly, Bakhtin construed the relation between speaker and listener as a dialogue. He calls this form of dialogue addressivity, which he defined as the
Delimitations and Definitions

11

orientation of the speaker to the response of the listener; this anticipated

answer shapes the utterance. Addressivity requires an active understanding
on the part of the listener, who must absorb the utterance into his own
conceptual system, and thereby construct new interrelationships, consonances and dissonances with the utterance (Bakhtin 1986a: 280282). In
chapter 7 I will be using dialogue chiefly in this sense.
Within folkloristic research yet other meanings have accrued to dialogue.
Lotte Tarkka has spoken of a dialogue between genres (Tarkka 1994: 265,
267291, 295), and of dialogue on the thematic level, e.g. a symbolic dialogue between humans and supernatural creatures in ritual (Tarkka 1994:
251, 260261, 266272, 274287, 295). In dialogical anthropology, the word
has been applied to the interaction between interviewer and interviewee as
well (Vasenkari 1999; Vasenkari & Pekkala 2000). More generally, dialogue
has been applied to almost any form of linguistic exchange, but one peculiar characteristic of dialogue in comparison with other similar terms might
be worth mentioning. Unlike dialectic, for example, dialogue does not imply
the fusion of thesis and antithesis in a synthesis; dialogue has no end point,
no real resolution. It continues beyond the boundaries of any particular exchange (Morson & Emerson 1990: 4950).
Discourse is a much-used term in contemporary cultural research, and
this dissertation is no exception; the nuances of my own usage of the word
mainly derive from Michel Foucaults, Norman Faircloughs and Mikhail
Bakhtins definitions of it. Foucault employs it in three senses: firstly, it
represents the general domain of all statements (discourse without an article
in English); secondly, it refers to an individualizable group of statements (a
discourse); and thirdly, it signifies a regulated practice accounting for a certain number of statements (Foucault 1999: 106). I will be using the word in
all these senses, though chiefly in the first and second ones. Faircloughs
conception of discourse is related to Foucaults; for Fairclough, a discourse
is a specific way of constructing a subject matter or area of knowledge
(Fairclough 1992: 128), while discourse is language use as a form of social
practice (Fairclough 1992: 63).
Bakhtin, or rather his translator, deploys discourse somewhat differently;
sometimes it refers to a voice, as in double-voiced discourse, a designation
that will be utilized in chapter 6, and sometimes it denotes a method of
using words presuming authority, a usage that is due to the meaning of the
original Russian word slovo (Mills 2002: 78). Discourse may also be de12

Introduction

fined by its context of occurrence, for example the discourse of religion that
will be spoken of in chapters 4 and 5 (cf. Mills 2002: 9). In structuralist and
post-structuralist research, the word has connoted a move away from the
reflectionist view of language as an unproblematic vehicle of communication and representation, to a conception emphasizing language as a system governed by its own rules and constraints influencing the thoughts and
expressions of individual subjects (Mills 2002: 8). All these associations
have influenced my use of the word.
Thus, the heterogeneity of the current usage of the term is visible in my
own employment of it, and even though this might lead to some confusion,
I have not found it meaningful to substitute it with other labels of my own
invention, since that would only contribute to an unnecessary multiplication of technical terms. Nevertheless, I believe the different senses of the
word will be fairly easy to determine when interpreted in relation to the
context of use.

1.3 Trolls in the History of Research

In this survey of prior research I will concentrate exclusively on the
folkloristic literature on trolls. For an account of Nordic folk belief research
in general until 1975, see Velure (1976).
The first large-scale effort to present and analyze the troll tradition in
Scandinavia was made by Elisabeth Hartmann in Die Trollvorstellungen in
den Sagen und Mrchen der skandinavischen Vlker, published in 1936. It is
an ambitious work in that she attempts to cover the whole of Scandinavia,
with the exception of Iceland, the Faroes and Finland; these areas are
nevertheless treated as well, albeit in passing. The explicit aim of her study
is to examine the troll as it appears in legends and fairy tales respectively in
order to identify their distinctive characteristics. She then proceeds by
describing the legends and tale types associated with the troll, in the former
case chiefly narratives of changelings and abductions (bergtagning), and in
the latter case she orders the material according to the status of the troll as
opponent or helper. Hartmann believes that some of the Scandinavian
legends of changelings and abductions originate in actual emotional
experiences (Hartmann 1936: 77, 134), whereas the fairy tales are purely
fictive in their description of the troll (Hartmann 1936: 143). Thus she
thinks the conception of the troll as evidenced in legends is tangibly close
Trolls in the History of Research

13

and vivid, while the troll in fairy tales is stiff and unrealistic. However,
both genres have some common traits, e.g. the huge frame of the troll, its
predilection for living in mountains and its enmity for man (Hartmann
1936: 143). Mrchen fictions and tales for children hold the middle ground
between these genres, as their image of the troll is more or less similar to
that of the legend (Hartmann 1936: 176).
In addition to the primary research problem, Hartmann also considers
other points, such as the geographical origin or age of a given legend or
tale type, quite in line with the methodology of the historic-geographical
school, and, perhaps most importantly, the emergence of the main motif of
a story. Dreams and hallucinations constitute a frequently employed explanation (Hartmann 1936: 52, 77, 100101, 114, 116, 118, 121122, 127, 133134,
165), and in that respect it is a rather von Sydowian study with its predilection for psychologistic interpretations.
Hence Hartmann presupposes the specialization of genresone purely
fictitious, the other both fictitious and based on experiencebut unlike her
latter-day sympathizers (e.g. Honko 1989: 13), she does not ponder its
communicative functions; the meanings of the narratives she scrutinizes do
not exist as a factor worth taking into account. Notwithstanding, I have
followed in Hartmanns footsteps on one point, and that is in viewing the
relationship between man and troll in terms of conflict and tolerance,
rendered as Der [Mrchen]Troll als bernatrlicher Gegner and Der
[Mrchen]Troll als bernatrlicher Helfer in Hartmanns work. As can be
seen from my clarification in brackets, it is only the troll in fairy tales that
is dealt with in this manner. I have chosen to consider all texts pertaining
to trolls from this point of view, and I have also paid attention to the
possible changes occurring in the relationship between man and troll in the
course of a narrative in order to highlight the complexity of humansupernatural interaction.
Bengt Holbek studies the story of Hobergsgubben, the old Hoburg man,
in Danish folklore from a genre point of view in the article On the Borderline between Legend and Tale (1991; for a summary of the narrative,
see chapters 3.3.2, 3.3.3). He begins by rejecting Waldemar Liungmans
statement that the story originated in Denmark in the Middle Ages, and
he presents both internal and external evidence to support his claim; he
believes that the narrative reached Denmark through the translation of a
Swedish chapbook, that is through print. Many elements of the story are
14

Introduction

not to be found in Danish folklore, and the distribution pattern of the recorded oral variants is different from the common one (Holbek 1991: 180
182). Once the literary derivation of the text has been established, Holbek
contemplates what the folk narrators have done with the story, how they
have turned it into folklore, and why they have chosen to integrate that
specific narrative into their repertoires. All chapbooks did not enter folklore, after all. As for the first two questions, people retained those features
that were compatible with tradition, while transforming those that were
not. Examples of the latter are the figures of St. Peter and the Virgin Mary
who might be replaced by representatives of the church, as well as the
drummer who has been combined with the image of St. Peter or God himself, which is more in accord with tradition (Holbek 1991: 183186). The
third question does not receive much treatment.
Holbek then moves on to the topic of genre, and asserts that both tale
and legend, the genres to which the story might belong, serve as instruction
in right and proper conduct, but the tale creates a fictional world in which
interpersonal relations on the family level are treated, and the legend determines the order and the boundaries of the human world in opposition to
the chaotic world outside it. However, the narrative of the old Hoburg
man does not fit squarely into either category. The chapbook, a rationalist
and rather tiresome creation according to Holbek, is supposed to be fictitious, but there are elements in it with their roots in legend tradition. Thus,
the status of the printed source is ambiguous, and the oral versions have
followed suit. Some of the latter have not been completely faithful to the
literary text, and two distinct tendencies can be discerned in the development of the stories. Some stress the narratives identity as a tale, often
linking it with tales of the stupid ogre. It is viewed as entertainment, and
the troll is thought to deserve the treatment it gets. Nevertheless, the troll
is duped by reference to powers associated with the legend and the beliefs
of the community.
Others add further legend motifs to the text, frequently connecting it
with real barrows or hills in the landscape. The peculiarity of the otherworld is prominent, and man becomes the defender of his community
against alien intrusion. The happy end of the story brings it closer to
the tale in this respect, and Holbeks conclusion is that the narrative is
permanently poised on the borderline between legend and tale (Holbek
1991: 187191).
Trolls in the History of Research

15

Holbek broaches a subject that will be prominent in my own thesis,

namely the question of intergeneric dialogue or intergenericity (for the
terms, see 1.4.3; cf. chapter 6 for analysis). My theoretical framework is
different, but Holbek is an important precedent in several ways. Firstly, he
points to the phenomenon of intergeneric dialogue itself, and demonstrates
that the generic ambiguity of texts does not have to be resolved, and that a
text need not necessarily be inserted into one, unambiguous slot in the
genre system. Secondly, Holbek indicates the import of norms and values
in the construction of the tale and the legend as genres, and I will be
touching on that topic as well, but from the perspective of distorted norms
and values in parodic narratives.
Bo Lnnqvist briefly surveys the relation between physical appearance
and cultural barriers in the article Troll och mnniskor (Trolls and Humans, 1996). The form of the body has been used as a criterion for distinguishing between the normal and the abnormal, the human and the nonhuman in many contexts for several centuries, as his material stemming
from the 17th century onward shows. One of these contexts is the supranormal tradition of the 18th and 19th centuries, and especially the notion of
changelings. The changeling embodied what the human child should not
be like, possessing corporeal traits that were the object of derision in folk
culture. More generally, the physical attributes of supernatural beings could
be both positive and negative, and the meanings ascribed to them depend
on the situation, and the time and place at which they are encountered.
These traits might be turned into the inverse of the normal, substituting
oneeyedness for twoeyedness, an animal limb for a human one, etc. Apart
from such a conceptual delimitation, supranormal creatures are also subject
to a territorial demarcation, relegated as they are to the forest and the wilds,
the water or the subterranean world, separate from human habitation
(Lnnqvist 1996: 152155). The ambiguity between human and animal characteristics is an important one, recurring in traditional abusive terms, for
example. Lnnqvist stresses the duality of these features, and wonders
whether they are losing their ambivalence, becoming wholly negative, in
contemporary culture. In other words, is the nuanced perception of these
attributes giving way to increasing aggression resulting in greater humiliation for the victim (Lnnqvist 1996: 157)? Thus, it is the symbolic properties of trolls and changelings that are in focus, and they are regarded as manifestations of a more general cultural pattern.
16

Introduction

In her doctoral dissertation Le troll et autres cratures surnaturelles dans les

contes populaires norvgiens (1996), Virginie Amilien ponders the construction of the image of the troll and its world through the ages, and the symbolism and associations of the figure. In exploring what one might call the
metaphorical association between the domain of the troll and the world of
the dead, Amilien focuses on the traits common to both realms. One such
characteristic is the location of these worlds, in the mountains. The dead
were thought to inhabit specific mountains in the landscape, whereas trolls
dwelt in the mountains of the imagination. Timelessness and the absence
of spatial specificity define both realms. The Christian division of the world
of the dead into Paradise, Purgatory and Hell recurs in the description of
the otherworld of the troll in which these three categories intermingle. According to Amilien, the image of Paradise is present in the fertility of the
earth in the domain of the troll, as well as in the arduousness of the passage
to that realm (cf. the narrow path), visited by an elect few, but since the
world of the troll is often reached through a descent involving physical
mortification, it may also be linked to the image of Hell or to Purgatory.
The brilliant light encountered at the end of the journey, however, once
again associates it with Paradise (Amilien 1996: 108117). In this study I
consider the paradisical associations of the otherworld as well, but I will
point to other reasons for doing so, and the conclusions are somewhat
different.
Amilien employs a longue-dure perspective on her material, tracing the
evolution of the troll from the Old Norse sagas to modern folktales and
contemporary popular culture. In Old Norse literature the troll was connected with combat, and female trolls in particular were regarded as powerful and vicious. With the introduction of Christianity all supernatural creatures were denigrated, but only the troll was assimilated into the image of
the Devil. Amilien illustrates the influence of the Bible on folk narratives
with one, lucid example, tale types AT 300303, The Dragon Slayer. The
attributes of the dragon, beast, troll or rise acting as opponent in these
types are drawn from the Revelation of St. John the Divine. The many
heads of the troll, the horns it is occasionally endowed with and the crowns
adorning its heads are to be found in the description of the beast in Revelation 13: 1, and the ability of the troll to regenerate when not all of its heads
have been lopped off in one stroke parallels the wondrous resuscitation of
the beast in Revelation 13: 3 (Amilien 1996: 135, 142144, 146). I intend to
Trolls in the History of Research

17

make a similar, more extended and systematic investigation of such relationships between folklore and the Bible.
Amilien continues her exploration of the symbolic properties of the troll
by analyzing the image of supernatural beings in relation to the Old Norse
conceptions of the dead, the soul, and fate. She contends that supranormal
creatures seem to incarnate the image of the dead, oscillating between the
notions of revenants and of the souls of the dead (Amilien 1996: 164, 185,
220). The troll may be viewed as the embodiment of destiny as well, since
it goads the protagonist into action and gives him a mission in life through
its depredations. Simultaneously, it has traits in common with the fylgja,
the ancient tutelary spirit protecting a man or clan, and with the hamr,
which could detach itself from the body, representing the ancestral spirit.
The fylgja manifests itself in dreams, and the encounter with the troll occasionally begins with the hero inexplicably falling asleep, making the coming
of the fylgja/troll possible. The physical form of the fylgja is reminiscent of
the appearance of the female trolls, and the ambivalence of the fylgja,
which was both good and bad, is reflected in the malice of the male troll
and the benignity of a younger, female troll in some narratives; Amilien
believes the protective aspect of the troll has been suppressed by the
Church. When the troll is destroyed, its soul, i.e., the fylgja, passes into the
hero, who is reborn, concretely and symbolically. The hamr is the vital
principle giving the hero a new personality, and it is a symbolic image of
the power which is being embodied in him. For the first time he is becoming a person, an adult. The helper is like unto the hugr, the impersonal,
active manifestation of the soul. It is exterior to man, just as the helper is,
and it is connected with the souls of the dead. The hugr appears when it
sees fit, sometimes to the detriment of the individual it has supposedly
come to assist. This facet corresponds to those instances when the hero is
being aided against his will. There are also elements of the fylgja in the
image of the helper (Amilien 1996: 180185, 202207, 222223). In other
words, Amilien argues that the Old Norse concepts have survived in the
Norwegian tales of the troll.
Moreover, Amilien contemplates whether it is possible to isolate any
characteristics peculiar to the troll, and she settles on five traits usually
linked to trolls alone: firstly, particular sounds; the troll often arrives noisily, and it shouts rather than talks. Secondly, it can have more than two
eyes, or only one eye. Thirdly, it loves beauty, especially human, female
18

Introduction

beauty. Fourthly and fifthly, it is associated with two specific ways of dying, either at the sight of the rising sun, or in an explosion because of
angerit literally bursts into pieces (Amilien 1996: 255).
Amilien also considers whether there is anything typically Norwegian in
the folktales, and she concludes that the attributes of supranormal creatures
in the texts are not exclusively Norwegian; rather the national character lies
in the combination of wondrous attributes and the rigorous functions imposed by the narrative. The omnipresence of the supernatural and its close
liaison with the everyday is the true mark of Norwegianness, according to
Amilien (Amilien 1996: 258).
The principal fascination of Amiliens work is her manner of blending
the research problems posed in contemporary folkloristics with the old
question of survivals of ancient cultural conceptions in 19th- and early
20th-century folklore. In contrast to the representatives of the survivalist
approach, however, she does not view the historical evolution of this relic
as a degeneration of ancient forms, but as an adaptation to an existing historical context producing a culturally viable tradition. Each stage of evolution is given its due, and the Old Norse conceptions are not valorized
simply because they are the oldest. Similarly, she does not stoop to reductionism, confining the world of the troll to a feeble reflection of the ancient
kingdom of the dead, for instance, but emphasizing their connotative and
associative resemblances.
The implications of her analysis are interesting to deliberate. Why did
these ancient concepts survive, i.e., what function did they fill in later periods? Were they considered functional in wonder tales only, or did they persist in other contexts as well? How did they fit into the overall culture of
each era? These questions might be difficult to answer, but they certainly
deserve to be posed.
Knut Aukrust has studied the relationship between trolls, churches and
St. Olaf in an article with that title, Troll, kirker og St. Olav (1997).
St. Olaf occupied a special place in Norwegian folklore, something occasionally frowned upon by the ecclesiastical authorities. The saint also represented law and order for the peasantry which referred to him in disputes
with the authorities (Aukrust 1997: 235237).
The Christianization of the country effected by St. Olaf was not accomplished without opposition, and in the folk tradition the human, pagan
adversaries have been replaced by equally pagan, but supernatural creatures,
Trolls in the History of Research

19

such as trolls, jtnar and other beings. The troll emerges as the most important opponent of Christianization, and the slaying of the troll is another
way of expressing St. Olafs mission to give the people new norms and a
new religion. Yet the combat with the troll has cosmological overtones as
well, since it is associated with the building of churches. In this context the
trolls are not merely opponents, but also the actual builders of the church.
The church spoken of may be a local church, but it is often the magnificent
cathedral in Trondheim, Nidarosdomen, which was erected on the supposed site of Olafs grave. The construction of such an edifice must have been
carried out by divine or supernatural means, and the idea of the churchbuilding troll suits this line of popular reasoning perfectly (Aukrust 1997:
243245).
The story, which is well-known in Scandinavian folklore (see Klintberg
2002: 135136), runs as follows: A troll offers to complete a building project, or volunteers to build a church from scratch in return for either the
sun and moon, or one of the saints vital body parts. St. Olaf is desperate
enough to enter the bargain, and must then find out the trolls name in
order to evade a grisly fate. At the last moment he hears a lullaby sung by
the trolls wife in which the trolls name is mentioned, and St. Olaf can
save his hide. The narrative may be interpreted as a representation of the
disclosure and annihilation of, and the victory over trolls and supranormal
powers. Therefore, the erection of the church is a sign of conquest, geographically and cosmologically. The church appears as a sacred site, and as
an intersection with other planes of existence. The struggle is crucial in the
construction and establishment of sacred space. The forces of chaos, the
enemies of God, are symbolized by the troll; St. Olafs vanquishment of
the troll echoes the triumph of God over primordial chaos in the Creation
(Aukrust 1997: 245250).
The shrine is the centre of a cosmological landscape consisting of earth,
sky and the subterranean world. Sacred space furnishes the link between
these three levels: the cross and the steeple reach toward the sky, the foundation wall extends into the netherworld, into the realm of the dead. One
might add that the churchyard becomes the new habitation of the dead,
making the realm of the dead a Christianized sphere. Hence the church
creates a bridge between what comes before life on earth and what comes
after (Aukrust 1997: 250).
Like most other recent contributions to the scholarly literature on trolls,
20

Introduction

Aukrust scrutinizes the symbolism of the stories of St. Olaf and the trolls,
connecting them with the creation of sacred space and the victory of good
over evil.
In addition to these scientific works, there are a number of books aimed
at the general public which are worth mentioning, for instance Jan-jvind
Swahn and Bo Lundwalls 1984 book Trollen, deras liv, land och legender
(The Trolls, Their Life, Land and Legends), Olav Bs Trollmakter og godvette (Magic Powers and Godvette, 1987), and Ebbe Schn and Elisabeth
Nymans Troll (Trolls, 1997).
To summarize, my own work will develop three aspects already touched
upon in prior scholarship: firstly, I will expand Hartmanns perspectives on
conflict and tolerance between man and troll to include other genres than
fairy tales, and to focus on the ratio between these opposites in the narratives, and how it affects the relationship between man and troll. Secondly,
I will discuss the issue of intergeneric dialogue raised by Holbek in order to
highlight the generic constitution of the image of the troll. Thirdly, I will
launch a more systematic and sustained analysis of the connections between
troll narratives and Biblical stories, a relation briefly considered by Amilien
in her dissertation.

1.4 Intertextuality in the History of Research

In the following I briefly outline the history of the concepts and perspectives I employ in my analysis of the Swedish troll tradition in Finland. My
focus is on levels of intertextuality as they have been defined, explicitly and
implicitly, in literary and folkloristic scholarship. I do not intend to give a
complete overview of intertextual theories, nor do I endeavour to go into
every detail of the works I mention in my account. For good introductions
to intertextuality I refer the reader to Graham Allens Intertextuality (2000),
Michael Worton and Judith Stills Intertextuality: Theories and Practices
(1990), the collection Intertextuality edited by Heinrich F. Plett (1991), and
Influence and Intertextuality in Literary History edited by Jay Clayton and
Eric Rothstein (1991), for example. The early development of folkloristic
intertextualism is described in my Intertextuality in Nordic Folklore Research
(2001), in which I include particulars not touched upon here.
In an ascending order of abstraction, I discuss (1) intertextuality proper,
i.e., the notion of each text being an intersection of texts where at least
Intertextuality in the History of Research

21

one other text can be read (Kristeva 1978: 84); (2) interdiscursivity, which
moves from the relation between texts to relations between discourses,
here taken in the sense of specific ways of constructing a subject matter
(Fairclough 1992: 128), with attendant institutional settings and prevalent
power relations; (3) intergenericity, which takes the inquiry one step further, pertaining to the larger whole of genres and genre systems, and (4)
cultural intertextuality seeking to uncover more vague affinities between
cultural forms, synchronically and diachronically. Finally, we return to the
individual subject, whose identity has been said to be constituted by the
intermingling of all previous levels in his mind. This aspect of intertextuality may be labelled (5) subjective intertextuality.
Like any division of research trends into various categories, it is difficult
to achieve absolute consistency. In my presentation I have therefore adopted the following principles: if a scholar has contributed to the understanding of several aspects of intertextuality, I have split my account of his work
and placed it under the appropriate headings. Occasionally I have decided
to sort a scholars work into another category than the most immediately
obvious one; this I have done in order to bring out an important consequence of the theory or study.
1.4.1 Intertextuality
Most advocates of intertextuality have not confined themselves to the textual level alone, but this level does figure in many investigations as a component in the analysis of broader concerns. The concept was introduced by
Julia Kristeva in 1969 in her article Le mot, le dialogue et le roman; she
drew her inspiration from Saussurean linguistics and Bakhtinian dialogism,
and I therefore briefly present the thoughts of the latter which substantially
contributed to the innovative aspects of the theory. Nevertheless, in this
context I focus exclusively on those facets of Bakhtins work which are
directly relevant for the development of the conception of intertextuality;
the reader wishing to find more information on other features of Bakhtins
oeuvre has many valuable commentaries to consult.5 I will also introduce
e.g. Bjrklund 2000; Clark & Holquist 1984; Dentith 1995; Emerson 2000;
Hirschkop & Shepherd 1989; Hirschkop 1999; Holquist 1990; Lock 2001a; Lock 2001b;
Morson 1986; Morson & Emerson 1990; Vice 1997.

5 See

22

Introduction

other concepts elaborated by Bakhtin in the course of this thesis, and discuss them in their respective contexts.
I have already presented Bakhtins understanding of dialogue in some
detail in chapter 1.2, and I will not repeat it here. However, the very notion of intertextuality itself, the conception of the text as a mosaic of quotations, is foreshadowed by Bakhtins analyses of the roots of the works of
Rabelais and Dostoevsky in carnival and Menippean satire (Bakhtin 1968;
Bachtin 1991). In other words, texts may absorb the characteristics of other
genres and cultural forms, and be transformed by them as well as re-model
them in their turn. In this respect, Bakhtin is more concerned with interdiscursivity and intergenericity than with intertextuality, though he would
have employed none of these terms to describe his preoccupations.6
Kristeva produced her work in the intellectual climate of the group involved in the avant-garde journal Tel Quel, which boasted associates such as
Roland Barthes, Michel Foucault, Jacques Derrida and other prominent
theorists (Allen 2000: 3035). Taking the Bakhtinian word as a point of
departure, she described the word not as un point (un sens fixe), mais un
croisement de surfaces textuelles, un dialogue de plusieurs critures: de lcrivain, du destinataire (ou du personnage), du contexte culturel actuel ou antrieur (Kristeva 1978: 83).7 Kristeva operates with a fusion of writer and
addressee, and of addressee and cultural context. Through the word, the
text is situated within history and society which are viewed as texts read by
the author and into which he inserts himself by rewriting them. Diachrony
gives way to synchrony as the writer assimilates and reaccentuates anterior
texts (Kristeva 1978: 83). All texts are considered mosaics of quotations,
being the absorption and transformation of other texts. As a consequence,
intersubjectivity disappears and is replaced by intertextuality (Kristeva
1978: 85), signalling the emergence of an entirely textualized universe.
Grard Genette has devoted much effort to developing the analytical
6 In

this context I would like to point out that I do not regard Bakhtin as an intertextualist, though I acknowledge his partial predilection for problems related to the research field
subsequently given that label. Nevertheless, much of his work is difficult to subsume under this heading, and especially the later, psychoanalytical orientation of intertextuality is
hard to reconcile with Bakhtins more pragmatic view of things.
7 a point (a fixed meaning), but an intersection of textual surfaces, a dialogue of several
writings: of the writer, of the addressee (or the character), of the contemporary or anterior
cultural context (my translation; cf. Kristeva 1980).
Intertextuality in the History of Research

23

vocabulary of intertextual theory, or transtextuality as he prefers to call it,

defined as tout ce qui le [i.e., the text] met en relation, manifeste ou secrte, avec dautres textes (Genette 1992: 7).8 He divides transtextuality into five types: intertextuality, reduced to quotation, plagiarism and allusion;
paratextualitytitles, headings, prefaces, illustrations; metatextuality, i.e.,
commentary; architextuality, the generic framework, and perhaps the most
important, at least in this context, hypertextuality, the relation uniting the
hypertext, the text studied, with an anterior text, the hypotext. In his use of
the concept, Genette restricts himself to obvious hypertextuality, where the
whole of the hypertext is derived from the whole of the hypotext, due to
his refusal to accord the reader too prominent a role in the interpretation of
hypertextual relationships (Genette 1992: 13, 1819). The hypertext can be
created on the basis of the hypotext either through transformation or
imitation, the latter being a more complex and indirect procedure. An imitation is essentially another story inspired by the anterior text, while a
transformation merely transfers the same story to another setting (Genette
1992: 14). The rigid definition of hypertextuality utilized in practice limits
the applicability of the concept, especially since many texts have more than
one hypotext, and these may be less easily discernible than Genette would
hope for.
He has also elaborated a taxonomy of intertextual techniques, which are
far too numerous to mention here in their totality; some of the most common are condensation (contraction of the text), amplification (substantial
additions) and substitution (Genette 1992: 341342, 375, 384). Once again,
his categories are keyed to the study of effortlessly identifiable hypotexts,
and are more difficult to apply in less evident cases. His work has not figured much in folkloristic discussions of intertextuality, but I will be applying some of his terms where appropriate, in a rather unsystematic fashion.
Susan Stewart was one of the first to apply the theory of intertextuality
to folklore in her study of nonsense (Stewart 1979) in which she viewed
nonsense as generated in relation to a standard of common sense, the aim
being to scrutinize the transformative operations utilized in moving from
one domain to another. The concept of intertextuality is employed to describe the interdependence of different provinces of meaning, and to estab8

all that places it [the text] in relation, manifestly or secretly, with other texts (my
translation).

24

Introduction

lish the common-sense world as a platform for developing other domains

of meaning (Stewart 1979: 1617). Stewart elaborates an intertextual construct, proceeding from realism, which is most faithful to the everyday,
common-sense world, via myth, science fiction and fantasy transposing it
to another world while still being dependent on it, to irony splitting reality
into two separate spheres, the normal and the ironic, and metafiction stressing the cultural nature of signification and interpretation. The last level of
textuality, metafiction, bears a close resemblance to nonsense since it is embedded in an impossible context, criticizing fiction from within fiction itself (Stewart 1979: 1921). The intertextual construct serves to situate nonsense in the field of other utterances and textual practices, highlighting its
affinities with other ways of manipulating common-sense assumptions.
Stewart emphasizes the social context of the interpretation of intertextual relationships, citing biography, the traditional stock of knowledge at
hand and the concept of society in general as determining factors (Stewart
1979: 16). However, she does not discuss the genesis and definition of the
term more broadly.
Intertextuality received a significant folkloristic elaboration in 1993 when
the Finnish scholar Lotte Tarkka presented her comprehensive theory on
the subject in the paper Intertextuality, Rhetorics, and the Interpretation
of Oral Poetry. Like Kristeva, Tarkka construed the text as a meeting
point of different texts. She describes the subject (the writer or performer), the receiver (the reader or listener), and the cultural context, history
and reality as the focal points in the construction of meaning. The subject
is also a receiver, creating the text in relation to already existing texts
(Tarkka 1993: 171). Unlike Kristeva, Tarkka thus regards the receiver (the
addressee) as an empirical being, not as a purely discursive entity; Kristevas
textualized notions of history and society are replaced by more folkloristically oriented definitions. Context and text become different aspects of the
production of meaning, and where one ends and the other begins is not
self-evident (Tarkka 1993: 178). The act of performance links the text to
social and cultural reality, as well as to the performing subject, and it narrows the span of the web of intertextuality.
In a later study, Tarkka examines the relation between texts in terms of
the processes of metaphor and metonymy; here metaphor refers to the differences and similarities between a pair of opposites representing distinct
conceptual spheresfor example the human village and the supernatural
Intertextuality in the History of Research

25

forestwhich are connected by metonymy, combining elements of the

same conceptual order into a sequence, and bridging the distance between
the poles of the metaphor. In linguistic parlance, metaphorical relations are
paradigmatic, whereas metonymic ones are syntagmatic. As instances of
metonymy Tarkka mentions the concepts of dialogue and communication,
and rituals involving communication with the otherworld (cf. chapter 1.1 on
the difference between Tarkkas notions and common usage). Intertextual
relations, generated by the use of common themes and epithets, for instance, and forming an intertextual universe, are conceived as a series of
metaphors, each constituent being comparable but not identical to the
others (Tarkka 1994: 293294; cf. Tarkka 1998b). This approach facilitates
an investigation of networks of association in which similar themes recur in
a variety of texts, often representing different genres. I will return to this
topic in the discussion of intergenericity.
In her dissertation Magic, Body and Social Order (1998), Laura Stark-Arola
adopts Lotte Tarkkas notion of the intertextual universe as an organic
whole in which a single text receives its meaning only in relation to other
texts. She also discusses macro-texts, i.e., broader cultural traditions such as
ritual descriptions of love magic, lempi-bathing and instances of women
being perceived as polluting men, which gain meaning through their interrelation (Stark-Arola 1998: 67, 73). Nevertheless, the most important contribution made by Stark-Arola is perhaps the delineation of a concrete,
intertextualist research method, something that has been largely missing in
intertextual scholarship. Her deliberations on this point therefore deserve
some attention.
She takes her point of departure in the act of reading the whole corpus
of texts to be analyzed, saying that the researcher forms a pre-comprehension of recurrent correspondences, homologies and analogies among and
between the texts. A similar differentiation of the meanings contained in
the texts, applicable to the larger corpus, is made as well. This preliminary
picture constitutes the basis for the next step in the process of interpretation, the employment of the method of intertextual abstraction, as she calls
it, which is glossed as the crystallization of the common denominators of
texts sporting the same theme into hypothetical generalizations like the
core motifs she treats in her study. The farm house is an example of such a
motif, organizing the relations between domains (homevillage) and persons (members of the householdoutsiders); core motifs are specifically de26

Introduction

ployed to organize other cultural concepts, symbols and relations. Additionally, the scholar examines relations of opposition and exclusion; the force
of the female genitalia, female vki, should not come into contact with
mens travel gear, for instance, since that would ruin them or the horse in
some way (Stark-Arola 1998: 6768, 2324, 224230).
The emphasis on an understanding arrived at through the reading of a
corpus of texts in its entirety necessitates a substantial research material, as
a single text or a very small number of texts are deemed inadequate for the
production of a reliable interpretation of cultural thought. Hence a sizeable corpus is primarily needed for the identification of key texts, those
texts which will throw light on all texts involved. Such key texts may be
ones overtly articulating the assumptions remaining tacit in many other
records, as in Stark-Arolas case (Stark-Arola 1998: 68).
The scholarly interpretation of the semantic systems extracted from the
intertextual universe is worked out through reference to various contextual
frames, consisting of textual context, performance context, social context,
cultural context, folk belief context, genre context and inter-genre context.
Only the last two are labelled intertextualwith a broader conception of
intertext, all but the first, which is rather intratextual, could be regarded as
intertextual (Stark-Arola 1998: 6970).
Another seminal figure in the history of intertextuality is Michael Riffaterre. His version of the concept, presented in Semiotics of Poetry (1978) and
several articles, differs markedly from the ones dealt with thus far. He
defines an intertext as one or more texts which the reader must know in
order to understand a work of literature in terms of its overall significance
(as opposed to the discrete meanings of its successive words, phrases and
sentences) (Riffaterre 1990: 56). Intertextuality then becomes the network
of functions forming and regulating the relations between text and intertext. Riffaterre distinguishes between theme and intertext: the former is a
variant of a motif, and knowledge of it is not necessary for the interpretation of a text (Riffaterre 1990: 57, 61); it does not always constitute an intertext, but an intertext can simultaneously be a theme. This delimitation
of intertext is not congruent with the views of many intertextualist researchers. Lotte Tarkka, for example, has successfully analyzed themes as intertexts, and Ann Helene Bolstad Skjelbred has investigated the articulation
of the same theme in a diachronic body of material (see 1.4.4). In my own
work I have also regarded themes as intertexts, since I do believe the
Intertextuality in the History of Research

27

various formulations of a theme have bearing on the interpretation of

specific instances of it.
In Riffaterres vocabulary, the term intertext is reserved for texts which
must be adduced in order to comprehend the ungrammaticalities of the
text, the unintelligible, seemingly unmotivated departures from logic and
accepted usage. The presence of these intertexts is signalled by a connective, a word or phrase occurring in both the text and the intertext, linking
them to each other. The connective represents a problem in the text, but it
also furnishes the answer to that problem in the intertext. Riffaterre argues
that the combination of the sign systems of the text and intertext creates a
new entity accounting for the uniqueness of the text (Riffaterre 1990: 61,
5758); thus he draws a conclusion diametrically opposed to that of most
other scholars who deny the absolute originality of texts because of their
inclusion in an intertextual system.
Furthermore, Riffaterre differentiates between meaning, which is the
result of the first, syntagmatic reading, i.e., the reading of the words and
phrases of the text in succession, and significance, which springs from the
second, retroactive reading. Meaning is plural, but unlike most proponents
of intertextuality, Riffaterre does not cherish its multiplicity. The aim of
interpretation is to arrive at the one, unitary significance of the text, summarized in the matrix, a single word from which the structure of the whole
text can be derived. This word is not present in the text; it is detectable
only in the form of its variants, the ungrammaticalities evident in the text.
Consequently the discovery of the matrix calls for a reinterpretation of the
multifarious meanings of the text in order to achieve a uniform explication
(Riffaterre 1978: 26, 13, 19).
Riffaterre champions the non-referentiality of poetry, since the ungrammaticalities of the text require a reading not according to reality, but in relation to other texts. The habit of interpreting poetry as a representation of
reality must be overcome to arrive at the significance of the text. The function of the second reading is therefore to make this transition from mimesis and meaning to semiosis and significance. Riffaterre identifies two types
of intertextual representation of reality; the first creates its representation
by referring to an intertext at odds with reality, the second depicts reality
by negating an intertext conforming to that reality (Riffaterre 1984: 143).
His rejection of referentiality thus parallels Kristevas disregard for context,
although the arguments for doing so are slightly different.
28

Introduction

Lauri Honko has picked up on Riffaterres version of intertextuality in

his Textualising the Siri Epic (1998), though he dispenses with many of
Riffaterres assumptions, for example the non-referentiality of poetry; being
an ardent defender of the fundamental importance of context and performance (Honko 1998: 151), Honko has no sympathy for context-free approaches. What Honko and Riffaterre do share is the emphasis on intertextuality at the reception of a text. Riffaterres exegesis of French poetry is
ostensibly oriented to the reading to be made by the receiver, and Honko
places special theoretical weight on intertextual interpretation when the
singer is internalizing a specific epic on the one hand, and during the performance of an epic on the other, when the audience is creating coherence
in the text by referring the story to a set of intertexts, which might not be
the same for all participants (Honko 1998: 167, 145, cf. 399).
Other common features are the stress on the individuality of the text,
which Honko seems to feel tends to be downplayed in some accounts of
intertextuality (Honko 1998: 34), and the notion of a shared sociolect
(Riffaterre 1984: 160, n. 2) or pool of tradition (Honko 1998: 69 et passim)
consisting of thematic, poetic, performatory and other traditional models,
elements and rules. The individual performer then selects and adjusts components of this collective, intertextual store, and it is this application of the
concept that Honko finds most rewarding. The pool of tradition as organized through a singers personal tradition system becomes less disorderly
than the presumed collective one, and it is possible to discern how the material is retained in the mind, namely as prearranged units and orderings of
plot, but remaining open to editing and novel combinations of elements
(Honko 1998: 7071, 9293, 154155). The concept of the pool of tradition is
intimately connected with another notion: for each separate epic present in
the singers mind Honko posits a mental text, a pretextual template incorporating storylines, textual elements such as episodic patterns, images of
epic situations and multiforms, generic rules and contextual frames, e.g. remembrances of earlier performances. However, the term should not be
reduced to mean merely fixed wordings stored in the memory and subsequently used in performance, since that would greatly diminish its explanatory power. Its force lies in explicating the mechanisms behind the otherwise rather mysterious composition and performance of extended folklore
forms, like the long oral epic, by providing a distinct but flexible framework within which to develop the narrative. The totality of mental texts
Intertextuality in the History of Research

29

constitutes an intertextual network within the tradition system of the performer, and it vouches for the existence of a wider array of materials than is
actually employed in performed epics. Thus, the notion is very much linked
to the individual, and Honko doubts its relevance for larger social groups
and tradition areas, which do not furnish the same kind of thick corpus as
the multiply recorded repertoire of a single singer represents (Honko 1998:
9499).
A concept somewhat akin to Honkos pool of tradition is the notion of
ethnopoetic or ethnocultural substrate advanced by Lauri Harvilahti who
defines it as those devices by means of which the singer gives clues, i.e.
uses specific registers and markers in order to enable the interpretation of
the discourse during the flow of the performance (Harvilahti 2000: 68). It
is a common, essentially intertextual repository of poetic diction, prosody,
modes of performance, musical styles and traditional meanings that can be
utilized in various contexts (Harvilahti 2003: 125; cf. Harvilahti 2001). Both
Honkos and Harvilahtis terms are convenient for designating the store of
intertextual expressions out of which the individual texts under study are
constituted.
1.4.2 Interdiscursivity
The identification of intertextuality with mere source-hunting within a
paradigm of influence, implying unimaginative dependence on other texts
and authors, in later applications of the theory of intertextuality led Kristeva
to abandon the term intertextuality in favour of transposition, which better
expressed the important point that intertextuality involves a transposition
from one sign system to another, resulting in a rearticulation of the thetic
position, the enunciative and denotative position. As an example Kristeva
refers to her study of the medieval French romance whose sign system
sprang from the redistribution of the sign systems of the carnival, courtly
poetry and scholastic discourse (Kristeva 1985: 5960). Put in these terms,
intertextuality comes closer to a notion of interdiscursivity or intergeneric
dialogue.
Norman Fairclough has utilized the concept of interdiscursivity or constitutive intertextuality to describe the relation between different discursive
structures. Interdiscursivity denotes the constitution of texts out of elements (types of convention) of orders of discourse, defined as the totality of
30

Introduction

discursive practices within an institution or society (Fairclough 1992: 85, 43).

He regards interdiscursivity as applicable to many levels, e.g. to the societal
order of discourse, the institutional order of discourse, the discourse types
(a term used for any kind of convention) and elements constituting discourse types, such as genre and discourse. A discourse is a specific way of
constructing a subject matter or area of knowledge (Fairclough 1992: 124
128).
The concept draws on both Julia Kristevas theory of intertextuality and
Michel Foucaults and Michel Pcheuxs work on discourse, with the
crucial addition of the possibility of discursive as well as social change.
Fairclough situates his scholarly preoccupations within the field of critical
linguistics, and his version of it combines a concern for stringent discourse
analysis with questions of power relations and ideologies, the constitution
of social subjects and systems of knowledge and belief (Fairclough 1992: 12).
I will be applying Faircloughs conception of interdiscursivity in my examination of the relationship between folklore and religious tradition as
manifested in narratives of trolls, but I will not be addressing all the issues
raised by his standpoint. There is one problem with the notion of interdiscursivity though, and that is the difficulty of distinguishing the intertextual and interdiscursive level. It is often virtually impossible, and any
description of the effects of intertextual and interdiscursive relations on
discourse must take both dimensions into account.
In Ulrika Wolf-Knuts investigation of intertextual relations between the
Biblical story of the Creation and Fall of Man and folklore narratives about
women receiving aid from the Devil at the birth of their babies (Wolf-Knuts
2000), the latter is regarded as the inverted version of the former, constituting a return of sorts to a paradisical state in which women can give birth
without pain (Wolf-Knuts 2000: 100101). Thus, Wolf-Knuts identifies
specific Biblical intertextshere New Testament texts are mentioned as
wellto the folklore texts which she, referring to the work of Manfred
Geier (Geier 1985), who adopted Grard Genettes term (Genette 1992, originally published in 1982), calls palimpsests or a montage of texts. Following
Lotte Tarkka, she views each text as dependent on other texts to acquire its
meaning. She also considers the link between the Biblical and the folklore
texts as a relation between discourses, ecclesiastical and popular, even
though she does not employ the concept of interdiscursivity. Furthermore,
Wolf-Knuts points to the occasionally very elusive nature of intertexts: all
Intertextuality in the History of Research

31

of them are not even texts, they may be images or merely motifs actualized
in diverse situations. The connection between folklore and the Bible is
construed as just such a vague relationship; the motifs exist in peoples
minds and they can be used, reworked and inserted into new contexts
(Wolf-Knuts 2000: 9192). Wolf-Knuts work can be read in conjunction
with my own, as the subject matter and points of view are fairly similar.
Ole Marius Hylland has analyzed narratives about Elvis in terms of intertextuality (Hylland 2002), and he isolates four groups of narratives, each
more complex than the preceding one, and with diverse strategies for handling intertextual relations. The first group is constituted by biographies,
which must necessarily rely on anterior expositions of the subject, but still
distinguish themselves enough from these to merit publication and attention. The second group of narratives is produced by Elviss fans, making
their individual selection from all that can seen, heard, read, visited and
bought in relation to Elvis. The fans construct their narratives by actively
creating a personal relationship to Elvis merchandise, services and expressions connected to his person. The third group consists of parodies and
ironic treatments of Elvis and his fans, who function as the parodied intertexts. As Hylland notes, parody is a very intertextual practice, something I
will also endeavour to demonstrate in my own examination of parody in
chapter 6. The fourth group is the one commenting on all the others: the
academic dissertation or the artwork. Finally, Hylland gives an example of
a text blending all four types of narratives, a Gospel of Elvis complete with
scientific commentary and replete with irony, written by a historian and
scholar of culture (Hylland 2002: 145). Thus, Hylland ends up straddling
wider territory than his usage of the term intertextuality suggests, tacitly
incorporating issues of interdiscursivity and intergenericity, since the intertexts going into the production of the narratives represent a host of discourses and genres, and particularly in the case of the narratives of the fans,
they are building blocks in the constitution of the intertextualized subject,
as the fans define themselves and their lives according to the narratives
about Elvis.
The same can basically be said of Anne Eriksens article on the narratives
about a Norwegian thirteenth-century historical personage, Mindre-Alv
Erlingsson (Eriksen 2002). Applying Lotte Tarkkas notion of a web of
intertextual relations as the locus of meaning, and putting special weight on
Mikhail Bakhtins conception of utteranceshistorically situated, finalized
32

Introduction

wholes bounded by a change of speaking subjects and able to elicit a response (Bakhtin 1986b: 71, 76)uttered by voices in a dialogue, Eriksen analyzes the texts about Erlingsson in the context of national and local history. Scholars concentrating on Erlingssons position in national history
have adopted various perspectives, but all of them have related their vision
of him to the discourses prevalent within the discipline of history. On the
local level, Erlingsson is associated with sites in both stfold and Vestfold.
In stfold local historians have attempted to insert local history into the
larger frame of national history; the material used is comprised of historical
sources. In Vestfold oral tradition is the only available source, as historical
data confirming Erlingssons connection to Vestfold do not exist. Hence
historians in stfold and Vestfold orient their narratives in relation to different discourses, one to scholarly discourses, the other to orally transmitted folk discourses. Simultaneously, the inclusion of Mindre-Alv in a
folkloristic scholarly discussion spawns links to yet other professional discourses and creates a dialogue with new intertexts (it might be noted that
the historians of Vestfold practice the intertextual technique of amplification as designated by Genette, for they fill in the lacuna of the historical
sources with texts from the oral tradition). Eriksen draws the conclusion
that there is no tradition about Mindre-Alv, but rather a network of
voices and utterances in dialogue. The notion of tradition is completely
replaced by the concepts of voice and dialogue (Eriksen 2002: 149166).
1.4.3 Intergenericity
Recent folkloristic examinations of the relationship between genres, intergenericity, find their source of inspiration in Mikhail Bakhtins late work on
speech genres, The Problem of Speech Genres (1986b), in which he defined genres as relatively stable types of utterances (Bakhtin 1986b: 60).
I will focus on only one aspect of his treatment of genre, namely his distinction between primary or simple genres and secondary or complex genres.
Secondary genres, like the novel or drama, absorb and digest various primary genres which are altered as they enter into complex ones. Bakhtin
links primary genres, e.g. rejoinders in a dialogue or letters, to an immediate
relationship with actual reality and the real utterances of others, while a
primary genre absorbed into a novel for instance loses this connection with
reality and becomes part of a literary event (Bakhtin 1986b: 6162).
Intertextuality in the History of Research

33

Lotte Tarkka implicitly modelled her conception of the epic universe, a

symbolic network and intertextual space generated by the fluctuation of the
same images, texts and fragments of texts in the whole of oral poetry,
forming the synthetic level of the genre system, on Bakhtins notion of secondary genres. The genre of the epic turns into the point of departure for
the analysis of intergeneric dialogue, and the primary genres absorbed into
the epic through direct reference or allusion become components of this
epic universe, representing different and complementary perspectives
(Tarkka 1994: 251). Therefore, the creation of an intertextual relation, by
various means, between the epic and another genre leads to the incorporation of the latter into the epic. This enables a view of the absorbed genre as
a naturalized part of the epic on the one hand, and as distinct enough to be
located outside the epic universe on the other, since the epic universe engages in dialogue across its borders (Tarkka 1994: 265). To connect intergeneric dialogue in this way to a secondary genre has its advantages and
liabilities; the resulting organic definition grounds the mixing of genres in
something concrete, the epic as a genre, but this act might also imply a bestowal of primacy on secondary genres, despite Tarkkas protestations to
the contrary when she writes: The intertextual universe is this hybrid text
in its totality and it is in itself primary and prior to its individual fragments
(Tarkka 1993: 173). If the locus of intertextuality is a secondary genre, it
seems to deprive primary genres of the status as intertextual. The epic universe is constructed through a dialogue of metaphor and metonymy (see
1.4.1) forming levels of world view.
Charles Briggs and Richard Bauman combine Bakhtins perspective on
genre with Julia Kristevas notion of intertextuality, resulting in a theory
centred on the relation between generic frameworks and individual texts,
presented in their article Genre, Intertextuality, and Social Power (1992).
They define genres as generalized or abstracted models of discourse production and reception, and propose to regard genres as tools for shaping
speech into ordered, unified and bounded texts with strong historical associations on the one hand, and fragmented, heterogeneous and open-ended
through their dependence on other discursive formations and context for
the interpretation, production and reception of discourse on the other. The
process of connecting a text to a generic model unfailingly creates an intertextual gap, since the text cannot fit completely into the generic framework.
These intertextual gaps can be minimized and maximized; minimization
34

Introduction

reduces the distance between text and genre, while maximization increases
it. The former is associated with traditional renderings, whereas the latter
is more in line with idiosyncratic, avantgarde ones (Briggs & Bauman
1992: 147149). Bakhtins secondary genres come to the fore in Briggs and
Baumans discussion of tall tales, a genre moving from one type of generic
intertextuality (linked to the personal experience narrative) to another (the
hyperbolic tall tale). The connections to several sets of generic features or
to a mixed genre, or both, enable multiple strategies for the manipulation
of intertextual gaps, as well as an ideological rearticulation of the constituent genres and their mutual relations (Briggs & Bauman 1992: 154). In my
study of parody in chapter 6, I will be utilizing this point of view. The relationship between generic intertextuality and power pointed out by Briggs
and Bauman (Briggs & Bauman 1992: 159) has also sparked an interest
within Nordic folklore research.
Inger Lvkronas study of erotic folklore, Suktande pigor och finurliga
drngar (Yearning Maids and Clever Farmhands, 1996) is a meditation on
precisely the power over discourse and genre. She employs Lotte Tarkkas
theory of intertextuality in order to make comparisons between various
traditional erotic texts and their respective messages concerning women
and their place in society. She views intertextualization as a method for
highlighting the consistency of the texts message, i.e., the overt meaning
determined by the power relations portrayed in the relationship between
the subject and the object of the text. In contradistinction to message,
meaning is defined as being multiple and not manifested in the structure of
the text (Lvkrona 1996: 122). Briggs and Baumans notion of generic intertextuality and its link to ideology, politics and power is more vital to
the analysis however, and it is used in conjunction with feminist theories
in an investigation of the construction of gender and cultural identity, as
well as of the creation, reproduction and legitimization of gender hierarchies (Lvkrona 1996: 112113, 103106). Lvkrona argues that traditional
erotica exhibit a male perspective, but stresses that women had the possibility to contest it by identifying with negative stereotypes in the texts.
These stereotypes were censured socially, yet allowed for female empowerment, along with the subversive power of laughter (Lvkrona 1996: 156158,
162166; cf. Asplund 2001 for a longer discussion of Lvkronas article).
Consequently, womens reinterpretations of the male message belong to
the realm of meaning.
Intertextuality in the History of Research

35

Laura Stark-Arolas account of the construction of the genre context and

the inter-genre context is also relevant to this discussion. Both are devised
by indicating the similarity of semantic structures; concerning the former, a
multitude of texts of the same type, be they ritual descriptions, incantations
or narratives, are assembled in order to show that they form a class of texts
because of their resemblance to each other in terms of form, function and/
or message. As for the latter, the similarities are observed across classes,
types and genres (Stark-Arola 1998: 6970). An inter-genre macro-text is
created by the existence of common poetic elements and themes in texts
belonging to different genres (cf. Tarkka 1994: 292, 294). North Karelian
lempi-bathing incantations, for example, share many elements with wedding songs and lyric songs about marriage, and the inter-genre macro-text
thus generated addresses the themes of pairing and marriage. These intergenre macro-texts and the poetic elements constituting them are, in essence, a system of referentiality, invoking a larger, but implicit world of images and symbols (Stark-Arola 1998: 188). At this point, intertextuality has
indeed been transformed into intergenericity.
Anna-Leena Siikala briefly utilized Briggs and Baumans concept of intertextual gaps in an article on emotions and their expression in the culture
of the province of Savo in Finland (Siikala 1998). She noted that lyric singers from Savo tended to break the rules of tragic songs by inserting elements of humour into the text, thus producing a maximization of the intertextual gap always present in the linking of a text to a generic model, through
the abrupt change in style (Siikala 1998: 170). She returned to the topic in
two studies of the Southern Cook Island korero, history, which she considered to be an extensive intertextual network (Siikala 2000a, 2000b); the
korero is truly a secondary genre as Bakhtin envisioned it, functioning both
as a kind of metadiscourse in which the content is mediated using varied
strategies, and as a metagenre comprised of the same contentual matter and
the same functional field though represented and performed in diverse
forms. Hence the korero furnishes generic models for the concrete instantiations of history, ranging from the tua taito (old narratives), the poetic pees,
the mastery of which characterizes the specialist, called the tumu korero, genealogies, and the ura performances incorporating song and dance (Siikala
2000b: 221; 2000a: 353). The proper performance of the korero is invested
with great authority and has social and political implications because of the
link to leadership and land rights through genealogical knowledge. The
36

Introduction

korero is a superdiscourse, a marked form of culturally significant discourse

apprehended as an historically established understanding of the social hierarchy and the power relations within the community. It is a basis and generic force in the social processes creating and legitimizing the order of a society (Siikala 2000a: 366). Due to the status of the korero, the tumu korero
also assumes a specific professional habitus distinguished by respect for the
ancestors and dedication to historical accuracy (Siikala 2000b: 224).
Siikala provides some examples of the ways in which intertextual gaps are
manipulated by different persons in various contexts. Tumu koreros strive to
minimize the intertextual gap between performance and generic model,
whereas commoners asked to give an account of a korero maximize it, thereby disclaiming it as a performance. The commoner does not master the
generic models appropriate to the korero, and is reluctant to perform it in
earnest (Siikala 2000b: 240). Another example is the transformation of a
handwritten manuscript, a puka papaanga containing the genealogical information and epic tradition pertinent to the korero, into an oral performance which was then recorded in writing. In the first part of the narrative, the performer minimizes the intertextual gap between his text and its
generic models, which is a strategy for creating textual authority through
adherence to prior, authoritative discourses mediated by the performers
father and other experts. Toward the end of the narrative, the narrator
maximizes the intertextual gap as the text turns into a performance of his
own life story. The reproduction of prior discourse becomes less important, and the personal experiences and evaluations of the narrator take
centre stage. The text is not a conventional korero, but rather a personal
perspective on world history (Siikala 2000a: 354, 359360, 362). Siikalas
research points to the potentially immense social significance of the mastery of generic models and of the ability to manipulate intertextual gaps in
the appropriate way.
1.4.4 Cultural Intertextuality
There is only one entry under this heading, namely Ann Helene Bolstad
Skjelbreds investigation of diachronic intertextuality in Fortellinger om
huldra fortellinger om oss (1998). Focusing on emotion and its expression
in narratives, Skjelbred tried to discover relationships of meaning and connotation (cf. Tarkka 1993: 171), and a form of thematic continuity in the
Intertextuality in the History of Research

37

disparate material constituted by the folklore of preindustrial, agrarian

Norway on the one hand, and the media texts of contemporary Norwegian
society on the other. The concept of thematic field, a version of Elliott
Orings ideological field (Skjelbred 1998: 2223), is the guiding principle in
the construction of an intertextual network spanning two centuries and
various narrative complexes. Encounters with the supernatural, with huldra
in the old days, and with angels and aliens today, is one such complex, actualizing the issue of whether we should believe our own perceptions and
experiences. These narratives may be used as a way of discussing the value
of sensory experience, and of defining where experience ends and hallucination begins, by including these themes in their structure. Thus, the intertextual relationship is based on neither local nor temporal situatedness, nor
on association with a specific genre, genre system or medium, but solely on
the presence of an underlying existential problem that is the same. Due to
the disconnectedness from a relatively uniform social and cultural context,
and the dissimilarities in the creatures inhabiting the narrative world, this
type of cultural intertextuality is an illuminating example of how cultural
intertextuality in general works. It is a more elusive form of intertextuality
which does not rely on common verbal expressions or textual structures in
order to create a link between texts. In principle, the theme can be the only
common denominator in otherwise dissimilar texts. Because of this, the
intertextual connection is more difficult to demonstratethere is no unambiguous proof of its existencebut the notion of cultural intertextuality
opens to an examination of the wider coherence of culture, without overstressing that coherence. The differences in cultural context spoken of
above are therefore bridged by this cultural coherence over a certain period
of time, and in this sense it is still a cultural intertextuality. This phenomenon can also be discerned in Adrienne Mayors study of an ancient analogy to the modern legend of the Choking Doberman in which she adduces
similar social and psychological circumstances for the emergence of similar
narratives, even though she does not discuss it in terms of intertextuality
(Mayor 1992: 253268).
1.4.5 Subjective Intertextuality
The role of intertextuality in creating subjectivity has been explored by several researchers. Here I will restrict myself to those conceiving of it in ex38

Introduction

plicitly intertextual terms; therefore some theories of the subject, such as

Lotte Tarkkas, will not be considered here (but see Asplund 2001: 73).
Julia Kristeva has extended her discussion of intertextuality into the domain of the preverbal through her conception of the semiotic and the symbolic, presented in her doctoral dissertation La rvolution du langage potique
(1985). The semiotic is composed of the drives preceding the acquisition of
language and subjectivity as they are inscribed into language, the symbolic.
After the acquisition of language, the semiotic and the symbolic are part of
the same whole, and both constitute subjectivity. Hence all signifying systems contain elements of both, though art, and especially poetry, is more
open to the energy of the semiotic (Kristeva 1985: 22). The revolution in
poetic language referred to in the title of the thesis is the eruption of the
semiotic into the texts of poets like Mallarm, and the linguistic and social
revolution caused by it. Kristeva distinguishes between phenotext, which is
symbolic, and genotext, the inscription of the semiotic into the symbolic
(Kristeva 1985: 8386). Her prime concern is the movement back and forth
from one to the other in a text. The semiotic is associated with a nonverbal, atemporal and non-spatial receptacle of drives called the chora, anterior to the formation of subjectivity (Kristeva 1985: 2230), while the break
into language and identity is named the thetic, involving a separation of
subject and object (Kristeva 1985: 4143).
This aspect of Kristevas work has not received much attention within
the field of folkloristics, but it must be included here to do justice to the
notion of intertextuality. For a good introduction to this facet of her work,
see Smith (1998).
Roland Barthes approached the problem of subjectivity from a more prosaic angle. In his analysis of Honor de Balzacs short story Sarrasine,
entitled S/Z (1976), Barthes identified five codes constituting the text and,
by extension, Balzac as a subject. Barthes attempted to demonstrate that
none of the features of the story had its origin in Balzac the Author-Gods
unique, individual mind, imbuing the text with august authorial intention,
but in the tissue of quotations drawn from the innumerable centres of culture as he put it in The Death of the Author (Barthes 1977: 146). Balzac
is transformed from a humanist subject into an intertextualized one in the
process, even though Barthes denies the text itself the status of intertextual;
the classic text, the class to which Sarrasine belongs, is limited in its plurality (Barthes 1976: 13), and is connotative rather than truly intertextual, or
Intertextuality in the History of Research

39

so Barthes claims. Nevertheless, Barthes manages to release the allegedly

restricted plurality of Balzacs text and to indicate its intertextual construction, however flawed it might seem compared to modern and postmodern
texts with their highly conscious intertextuality.
From a folkloristic point of view, the most interesting code Barthes
mentions is the cultural code or code of reference which represents a collective, anonymous voice springing from human knowledge and wisdom. It
frequently assumes scientific or moral authority, and might acquire proverbial dimensions (Barthes 1976: 23). The idea of this kind of traditional
wisdom and shared presumptions is well suited to the study of folklore, but
this facet of Barthess work has remained in the background within folkloristic scholarship. Notwithstanding, his sustained examination of the constitution of the intertextualized subject is an important antecedent to such
investigations within our discipline, despite the fact that it is missing in the
bibliographies.
One such study is Anne Leonora Blaakildes A Vision of Twenty-FirstCentury Folkloristics (1998), in which she pondered the impact of intertextuality on individual identity. Blaakilde connects intertextuality with the
mixing of genres, and particularly with the mixing of genres within the
subject, basing her ideas on Mikhail Bakhtins theory of speech genres. The
points of intertextuality can be located in what she calls cultural tropes, expressing cultural difference (Blaakilde 1998: 109). The encounter between
the folklorist and the interviewee turns into a dialogue of cultural difference
as well as similarity in which each participant has his own constellation of
these characteristics. The dialogical nature of the encounter furnishes a
centre of communicative coherence while allowing for disparate tendencies.
The intertextualized subjecthere the interviewer as well as the intervieweeis made up of specific discourses and genres, and the construction
of that subject in discourse is thus the natural and necessary starting point
for any interpretation of cultural tropes, and by extension, of anything else
in the world, including academic discourses (Blaakilde 1998: 113114).
In essence, Blaakilde has viewed the intertextualized subject as similar to
a secondary genre in Bakhtins sense, absorbing and digesting primary genres; the subject is created in relation to other persons, and the texts and discourses of the surrounding world. The subject is not self-sufficient, complete and unchangeable in itself. It is always developed in interaction with
the outside world. The addition of the concept of cultural difference gives
40

Introduction

certain nuances to the picture of the workings of this absorption of genres,

emphasizing both the cultural and individual aspects of this process. The
intersubjective, dialogical encounter as an arena for further intertextualization of subjects and for the production of knowledge and interpretations is
described as the ground for agency (Blaakilde 1998: 109), since dialogue
cannot be passive; if it is, it degenerates into monologue (cf. Emerson
2000: 229). Therefore it demands some degree of agency to continue.
All scholars presented here have their own conceptions of the subject
and subjectivity, but a number of common features may be discerned. All
seem to agree on the produced and processual nature of the subject, which
is constantly being reassessed and remoulded. Furthermore, subjectivity is
characterized as a relational position, articulated in relation to other people,
objects or phenomena. Subjectivity also creates the point of view from
which the individual observes the world; this is especially prominent in
Barthess analysis, where the constitution of the perspective of the author,
Balzac, is determined by the intertextual codes generating his subjectivity
(cf. Bjrklund 1993: 242; Lvkrona 1996: 160166).
t

To summarize my own position vis--vis this research history, I will be applying Julia Kristevas conception of texts as mosaics of quotations, and be
utilizing Lotte Tarkkas notion of metaphor and metonymy to describe the
relationship between texts linked in this fashion. I will also adopt Norman
Faircloughs concept of interdiscursivity to delineate the relationships between texts on the level of discourse, and use Mikhail Bakhtins, and
Charles Briggs and Richard Baumans insights into the workings of genre
in order to examine the effects of generic framing on narrative. The previous study most in tune with my own is that of Ulrika Wolf-Knuts who
explores the relation between folklore and Biblical stories as well, but my
work will serve to elaborate the implications of that relationship and provide other theoretical tools for investigating them. The rest of the research
literature has also been important for the development of my understanding of intertextuality and of how it functions.

Intertextuality in the History of Research

41

1.5 Method
The choice of material has been guided by a number of practical and methodological considerations. Thus, in my investigation I have restricted myself to trolls alone, i.e., to supranormal creatures explicitly designated by
that name, either in the text itself or in the headline supplied by the collector. This is primarily to keep the corpus manageable; examining the folklore of all supernatural beings is a task of considerable proportions, and
exciting though it might be, I cannot embark on such a project here.
A special problem in selecting the corpus of study, however, has been to
discriminate between trollgumma and trollkring (old troll woman) in the
sense of witch and the use of the same terms in the sense of troll. My pragmatic solution to this thorny issue has been to check if these labels alternate with troll or a synonym for witch in the text, and if the former is the
case, I have included the record in my material. This method is not foolproof, of course, but there was no other option.
The selection of the period of study has also been dictated by rather
pragmatic concerns: the earliest archival text on trolls to be found in the
Finland-Swedish folklore collections was recorded in the 1850s, while 1925
is the date of the last recordat least as far as I knowto be noted down
manually. All texts were collected using a fairly uniform fieldwork methodology, described in chapter 2.2. A comparison of audiotape recordings
and handwritten documents is a task in itself, which I cannot perform in
the context of this thesis (for such a study on the nightmare, see Danielsson
1992). Besides a consistency in the fieldwork methods used in the collection of the material, my choice is also conditioned by the scrutiny of intertextual relations which I have opted to analyze from a roughly synchronous
perspective in order to achieve some uniformity in the intertexts invoked.
A consideration of intertexts changing over time might have offered an intriguing peek at the life of tradition from a diachronic point of view, but
then we would have had to contend with the different nature of the sources
as well.
Regarding chapters 4, 5, 6 and 7, the material has been selected employing the following criteria. In chapters 4 and 5 I have limited myself to the
tradition of a specific parish, the parish of Vr, and its immediate neighbours. Lotte Tarkka stresses the importance of working with material from
a single community (Tarkka 1993: 173) to ensure that the intertextual rela42

Introduction

tions found spring from the same network of associations, and the parish of
Vr is exceptionally rich in (recorded) supernatural traditions. Vr may
also be said to represent the average Ostrobothnian parish in many respects: chiefly rural, socially homogeneous, prone to religious revivals and
emigration. The choice of the parish of Vr as the object of a case study is
mainly dictated by the availability of material and the existence of other
studies on the folklore and history of the community. For the discussion in
chapters 4 and 5 I wanted a corpus of texts belonging to the same tale or
legend type, and Vr was the parish with the largest number of such texts.
The selected parish also had to furnish a substantial amount of other folklore material, and once again Vr was a good choice. In addition, I considered it important to have other studies on the parish to consult, especially in the reconstruction of the historical, social and religious context. A
more personal motivation is that my paternal grandmother was born in
Vr, and therefore it has pleased me to do my research on her native parish. However, if some other community had proven more useful for my
purposes, I would have opted for that instead. Vr is not a romanticized
and exoticized Finland-Swedish Dalecarlia or Karelia, representing the
home of the genuine, Finland-Swedish folk, at least not to me.
Two groups of texts, both of which consist of narratives of abduction,
though the second adds the motif of the banishment of the trolls, have
been singled out as particularly suitable for an intertextual analysis as they
represent the nearest equivalent to a thick corpus (Honko 2000: 1517) to be
found in my material. Lauri Honko glosses this term as the activity of
producing thickness of text and context through the multiple documentation of expressions of folklore in their varying manifestations in performance within a biologically definable tradition bearer, community or environment (Honko 2000: 17). Collected within the same community, the
texts studied in these chapters may indeed be viewed as part of living tradition systems maintained by individuals and groups having the possibility
of social exchange (Honko 2000: 16). Nevertheless, the thickness of the
material is mainly restricted to the textual level, while contextual thickness
is lacking; the latter was not a priority during the time of collection.
In chapter 6 I have concentrated on the repertoire of a single narrator,
another form of thick corpus, though with the same reservations as in the
previous case. As for chapter 7, the selection of texts has been guided by
the research problem posed, i.e., what effects the indeterminacy of the troll
Method

43

in stories of the supernatural has on the relationship between man and

troll. A number of records from the entire Swedish-speaking area have
been singled out for analysis on the basis of their illustrative character.
Concerning the choice of material in general terms, I have restricted
myself to prose narrative genres. I have also perused the extant records of
ballads, proverbs and riddles referring to trolls, but these are so scant that
they scarcely contribute anything to the investigation, and I have therefore
decided to exclude them.
The theories I use also constitute my method to a large extent, but a
number of comments on my concrete method of analysis may be made. It
is quite akin to those used by Laura Stark-Arola and Lotte Tarkka; in the
course of my work, I have read approximately 10 000 records9 as published
in certain volumes of Finlands svenska folkdiktning (The Swedish Folklore
of Finland), mainly legends of the supernatural, wonder tales and jocular
tales (Finlands 1917, 1920, 1931)I have read the cited volumes in their entirety. In addition to these, I have made goal-oriented searches in other
parts of this collection. Taking the troll stories as a point of departure, I
have scanned these three volumes on the basis of two criteria in the case of
texts intended for my intertextual investigations in chapters 4, 5 and 6: 1)
the presence of verbal expressions similar to the ones utilized in troll narratives (the epithets of Tarkkas theory), and 2) the employment of similar
themes (the core motifs in Stark-Arolas method, the common themes in
Tarkkas). When I have discovered linguistic or thematic similarities using
the method of intertextual abstraction, I have consulted the original records, as the published versions generally do not reproduce the vernacular.
I have also gone over the narratives in Ranckens second collection, going
straight to the original records (R II; see chapter 2.2.1).
In assessing the relevance of potential intertexts thus identified, I have
striven to maintain a rather close fit between the verbal expression and/or
theme of the troll text and the intertext. I have rejected intertexts in which
the similarity to the primary material was too tenuous and superficial.
9

The concept of the record, in its capacity as a text in particular, is not wholly unproblematic. Where does one record/text end and another begin? Especially in the light of intertextual theory, the boundaries of the text become permeable and fuzzy (cf. Tarkka 1993:
171). In this specific context I mean a text demarcated as a single whole in the manuscript,
occupying a page of its own and/or given an individual heading or number, or, in the case
of longer segments, parts dealing with the same topic or theme.

44

Introduction

However, it is difficult to draw an exact line between probable and improbable intertexts, and the final decision remains subjective, as indeed all interpretation does (cf. Tarkka 1993: 172173; Skjelbred 1998: 2122, referring to
Ingwersen 1995: 7790; Wolf 2002: 1628). Nevertheless, I have also tried to
corroborate my interpretations through reference to existing knowledge
about the beliefs, customs and contextual factors prevalent in the parish of
Vr at the time of collection.
Regarding chapter 7, my method has been somewhat different as the research problem has changed. Michael Holquist has stated that Bakhtins
works do not furnish any ready-made methods which can be applied directly to a research material, but that [a]n immersion in Bakhtins thought
will indeed transform the way one reads, but only after some time has
elapsed, and in ways that are not predictable (Holquist 1990: 107108). It
took me three years to see the utility of Bakhtins notions of unfinalizability
and finalization in my own work, and the key to this realization was the
same as Ann Helene Bolstad Skjelbreds in her examination of cultural
intertextuality: emotion. I had noted the sometimes very palpable sense of
fear of the supernatural described in some troll narratives, and I wondered
whether there was any significant reason for this other than the basic
human fear of the unknown. I started studying the situations in which this
fear appeared in the stories, and came to the conclusion that it was the uncertainty about the intentions and possible actions of the troll that provoked this reaction. Framing it as an instance of unfinalizability was principally occasioned by Bakhtins glossing of the concept as the unobstructed
unfolding of a character in narrative; this seemed to help to explain the behaviour of the troll in the narratives. The position of the author, or, in my
case, the narrator in relation to unfinalizable characters was also important;
narrators of troll texts seemed to strive for the creation of an image of the
troll dominated by its unpredictability which I interpreted as a consequence
of its unfinalizability. Then I began to look for evidence of a potential unfinalizability of supranormal characters in other texts, and of a possible
promotion of this trait on the part of the narrators in their construction of
their texts.

Method

45

2 MATERIAL AND CONTEXT

2.1 General Considerations
My primary material consists of 123 records made by many different collectors in various parts of those districts in Finland housing Swedish-speaking
inhabitants. In addition to these texts, I have cited 72 other records which
constitute my secondary material. The largest part of the material stems
from the Folk Culture Archives of the Swedish Literature Society in Finland in Helsingfors, some records belong to the Rancken Folklore collection currently deposited at the Department of Folklore at bo Akademi
University. Finally, 26 texts are drawn from printed sources, predominantly publications dealing expressly with folklore and folk culture.
The geographical distribution of records from Nyland (Uusimaa) and
Southwest Finland is almost even, but Ostrobothnia has yielded more than
fifty per cent of the texts. The land Islands are poorly represented in my
material, having contributed only two records. The primary corpus consists of prose narratives of trolls, supernatural beings mostly inhabiting forests, hills and rocks.
It is questionable whether it is possible to make any pronouncements on
the actual distribution and vigour of troll traditions in the Swedish-speaking
areas in Finland on the basis of the amount of records alone. Collection
was often guided by the personal interests of each fieldworker, and the
entire scope of traditions in a community was hardly covered by even the
most industrious of collectors. In addition, some parishes received more
attention than others. The politics of collection probably had a profound
influence on the formation of the collections now extant, and the lack of
records from a particular area cannot be equated to a lack of tradition
(Bergman 1981: 2223; Wolf-Knuts 1991: 3437; cf. Lilja 1996: 182).
In the following I will give an account of the ideology constructed and
sustained in and through the collection of Swedish folklore in Finland. All
translations from the Swedish are mine unless otherwise stated. I will also
provide a brief description of the context in which the records were made,
focusing on the material from the Ostrobothnian parish of Vr, which
will be the object of some in-depth studies in chapters 4, 5 and 6.10
10 In this thesis I refer to the sites of collection as parishes, even though this administrative
unit was replaced by the municipality in 18651868.

46

Material and Context

2.2 The Sources

2.2.1 The Rancken Collection (R) and Its Contributors
The material collected in Ostrobothnia on the initiative of Johan Oskar
Immanuel Rancken (18241895) is one of the earliest sources of troll narratives. As a senior master, and later headmaster, at Vasas secondary school,
he was in a perfect position to inspire his students to see the wonder and
importance of Swedish folklore to a country then in the process of constructing a national identity, at that time solely envisioned in terms of the
Finnish language and Finnish folklore (see Honko 1980 and Anttonen
2003 for an account of Finnish nation building). The rhetoric of the Swedish-speaking intelligentsia favoured a suppression of the Swedish language, which was perceived as an obstacle on the way to real nationhood.
The Swedish language had to die in order to allow the Finnish language to
grow (Andersson 1967: 120; Wolf-Knuts 1997: 33). J. O. I. Rancken, credited by Otto Andersson as the first official advocate of Swedish traditions
in Finland (Andersson 1967: 123), believed that the Swedish culture in
Finland bridged the gap between Finnish tradition and traditions in Sweden (Wolf-Knuts 1997: 33).
Rancken wanted to carry out historical, comparative research in a patriotic spirit, and folk culture was the key to the successful completion of such
inquiries, as it was more archaic and genuine. Ulrika Wolf-Knuts has
identified the key words in Ranckens writings, and recurrent ones are peculiarities (egendomligheter) and the peculiar (det egendomliga), whereby he
meant the typical (det typiska), the unique (det unika) or the non-universal
(det icke-universella), but also the odd (det underliga) and the original (det ursprungliga). It was of prime importance to collect these peculiarities, as
they were being stunted by urbanisation and education, the latter activity
being one in which Rancken himself was deeply involved in (Wolf-Knuts
1997: 3536, 4043). Rancken was not, however, negligent of urban culture,
but encouraged the collection of urban traditions as well. Variants were
mandatory for comparative research, and Rancken found it desirable for
the collector of folklore to be well-acquainted with the district in which he
worked. He thought himself unsuitable in this respect, as he was not a
native of Ostrobothnia (Wolf-Knuts 1997: 32, 34, 36). Also the images in
folklore, interpreted by Wolf-Knuts as attitudes, norms and values, lay
within Ranckens field of interest (Wolf-Knuts 1997: 43).
The Sources

47

Rancken tried to entice his pupils into an appreciation of folklore by

giving them essay assignments on local traditions (Andersson 1967: 129130;
Bergman 1981: 7; Hggman 1992: 7576; Wolf-Knuts 1997: 43). Three of the
texts I will cite stem from this kind of homework; one such essay was submitted by Odo Sandelin in January 1887 when he attended the eighth form.
The text is composed in standard Swedish. Sandelin comes across as hostile to folklore and condemns it in no uncertain terms. He connects superstition to the spiritual state of different religions; in pagan religions superstition prospers most, a bit less in Catholicism, and least of all in his own
faith, Protestantism. He writes:
Ostuderadt och oupplyst folk har en besynnerlig bengenhet att stta tilltro till en hel
hop underliga, oftast otnkbara och frnuftslsa historier och de i hg komma dem s
beundransvrdt vl. Dessa sitta de sedan om vinterkvllarna vid spiselelden och bertta
t sina barn, som sedan med sin rika, lifliga fantasi oppfatta dem mycket mer frunderliga n de i sjelfva verket berttats. S fortgr det frn generation till generation, alltid
med sm frndringar och tillsatser vid berttandet. Vrt folk har lnge sttt fjettradt i
okunnighetens bojor. P senare tider har likvl en stigande folkbildning brjat lossa
dessa bojor och jemsides med den har vidskepelsen och ftron brjat frsvinna eller tminstone minskas. S t. ex. i Kronoby kommun. Hr finns numera hgst f traditioner frn lnge sedan svunna dar. Kronoby r ocks i intellektuellt hnseende en af landets frmsta socknar. Endast af gammalt folk fr man hra dessa sgner; de unga tro
dem ej och endast gra spe af dem. (R I 86: 12)
Unstudied and uneducated people have a peculiar propensity to give credence to an entire host of strange, often unthinkable and senseless stories, and they recall them so admirably well. In the winter evenings they consequently sit by the fireside and tell these
to their children who, with their lively imagination, interpret them as far more wondrous than they really have been told. Thus it continues from generation to generation,
always with small changes and additions in the telling. Our people have long stood fettered in the bonds of ignorance. Recently the rise of education has, however, begun to
undo these bonds and alongside it superstition has begun to disappear or at least diminish. Thus [it is] e.g. in the municipality of Kronoby. Nowadays there are very few
traditions here from times long since past. Kronoby is also intellectually one of the
most prominent parishes in the country. You only get to hear these legends from the
old folk; the young do not believe in them and only ridicule them.

In addition to Sandelins essay, I have also referred to Matts Leander

Forsns essay Skrock, vidskepelse m.m. bland allmogen i Kronoby
(Superstitions etc. among the peasant folk in Kronoby) (R I 74), and Edv.
Ketos Skrock och vidskepelse i min hembygd (Superstitions in My Home
District), also from the parish of Kronoby (R I 78).
48

Material and Context

The rest of the texts that I have utilized from this collection spring from
the extensive collecting activities of Jakob Edvard Wefvar (18401911) who
collected for many other employers as well. Wefvars records are made in
the vernacular, though not using the phonetic script called landsmlsalfabet
which had not been invented at this point. Jakob Edvard Wefvar was born
in the parish of Munsala and studied under Rancken before entering the
University of Helsingfors, called Kejserliga Alexandersuniversitetet (the Imperial Alexander University) while the country was under Russian rule. He
continued collecting folklore whenever he was not prevented by illness, and
later in life he became known as a lay preacher and teacher (Hggman 1990:
139152. For more details see Dahlbacka 1984; Hggman 1990; Hggman
1992; Wolf-Knuts 1991).
As for the technical aspects of collection, Rancken preferred prose narratives to be recorded verbatim in order to bring out the peculiarities of the
narrators language. Ulrika Wolf-Knuts deems Wefvar willing to comply
with these instructions, but she finds it improbable that he would have
been able to follow them to the letter, as the texts do not reflect real
speech, even if they are in the vernacular. She does however indicate the
possibility of the narrator employing a specialized style of speech in storytelling (Wolf-Knuts 1991: 2224; Wolf-Knuts 1997: 32), what Dell Hymes
and other scholars following in his footsteps have labelled a register (Hymes
1989: 440). Ann-Mari Hggman in her turn has noted on perusal of
Wefvars posthumous documents that his field notes were often imperfect
and vague drafts, while the submitted manuscripts were clean copies and
edited to some degree (Hggman 1992: 81).
Wefvar seems to have been able to create a good rapport with the performers; his own background as the son of a crofter placed him more on a
par with the people he interviewed, his personality and his long stays in
their midst during the collecting trips endeared him to his hosts, and he
succeeded in developing a trusting relationship with them (Wolf-Knuts
1991: 2223). I think the great interest in folk life he exhibited might have
contributed to this circumstance, as well as his proneness to tell stories of
his own (for examples suggesting this interpretation see Wolf-Knuts 1991:
22; Hggman 1990: 153). Generally, Wefvar has provided the name, profession and domicile of the performer in the clean copy of the record (cf.
Wolf-Knuts 1991: 24), with some exceptions. At times he mentioned only
geographical provenance, which applies to eight records, six of which were
The Sources

49

made in the parish of Vr. One gives as the place of provenance the parish
as a whole (R II 151), while the rest were collected in the villages of Karvsor
(R II 11), Rejpelt (R II 70; R II 120; R II 204) and Kaitsor (R II 178). The
remaining two were recorded in the parish of Gamla Karleby (R II 420)
and the village of Dagsmark in the parish of Lappfjrd (R II 291) respectively. In the parish of Sideby he met Niklas Teir, the son of a lay assessor,
who provided him with one story (R II 394). Similarly, in the village of
Dagsmark he collected two stories, one from Henrik Lilljans (R II 295) and
the other from Karl Gustaf Lng (R II 305), the latter having also been
known as a ballad singer, as were Maria and Sofia Bergstrm from the village of Hrkmeri in the same parish. The Bergstrms jointly contributed
one story (R II 336), as did E. Granars from the same village (R II 338). In
Skaftung Wefvar came across Robert Emholm, another legend narrator
and ballad singer nick-named Rox-Robert, who gave him one narrative
(R II 339). The painter Peter Ragvals in the parish of vermark supplied
three stories, one of them in writing, but all texts still in the vernacular
(R II 325; R II 327; R II 328). In the parish of Vr Wefvar visited Greta
Mrtens who supported herself on life annuity (R II 32; R II 76), Anna
Kull (R II 19; R II 121), Johan Svens (R II 67), Majs Svens, merely eleven
years old (R II 199) and the carpenter Johan Aln in the village of Rejpelt
(R II 27; R II 58; see Appendix A). He also called on Sigfrids in the village
of Kaitsor, said to be supported by life annuity (R II 175), as well as on
Jakob Grnback who narrated one story (R II 133). In the village of Karvsor
Wefvar met the former bakers apprentice Mickel Bygdn (R II 46), and in
the village of Lotlax he encountered the shoemaker Svendlin (R II 188). In
the village of Karvat in the neighbouring parish of Oravais the carpenter
Erik Kock gave him one story (R II 36), while Maria Holstre from the village of Kimo supplied him with one narrative (R II 28a). Finally, Johan
Mattsson Palkis from Wefvars home village of Linnusper in the parish of
Gamla Karleby contributed one text cited here (R II 138). Unfortunately,
one text is illegible toward the end and I have not been able to decipher the
name of the performer (R II 427).
2.2.2 The Collections of the Swedish Literature Society in Finland (SLS)
The Swedish Literature Society in Finland was founded in 1885 in an
attempt to promote and legitimize Swedish culture in Finland at a time
when the political and linguistic situation was frustrated (Steinby 1985: 14
50

Material and Context

18). The struggle between the Finnish and Swedish language parties started in the 1870s and continued until the turn of the century, when Russification strategies implemented by the Russian government intensified to such
a point that the formerly opposing parties thought it best to unite their
forces rather than to squabble among themselves. There was also a cultural
dimension in addition to the political, namely the historical connection to
Sweden, which was emphasized (Wolf-Knuts 1991: 24, 34). From the 12th
14th centuries until the war in 18081809, Finland was part of Sweden, even
though the Swedish supremacy was often contested by Russia.
In the year 1887 Ernst Lagus proposed that the collection of folklore
should be included in the Literature Societys field of activity, and a decision was made to send out scholarship holders for this purpose. A list of
instructions was accordingly compiled to suggest the kind of folklore to be
collected, as well as how the collectors were to proceed in their work. The
instructions were published in the Societys publication series Frhandlingar
och uppsatser (FU) (Negotiations and Essays) no. 3 (FU 3: 101106), and
they were also distributed as an offprint. I will just mention a few points
pertaining to my material. The collectors were recommended to visit more
isolated areas, as the relics of antiquity (fornminnen) were considered to
flourish at their fullest and in their most original forms there. Elderly
people with good memories were preferred as informants. Time, place and
the name of the narrator were to be carefully documented, and the transcription to be made p den munart, som framstllaren anvndt, med strsta noggrannhet i afseende uttalet och ordalagen (in the dialect used by
the presenter, with the greatest accuracy concerning pronunciation and
wording (FU 3: 103104). Initially variants were tolerated, and it was emphasized that nothing was too trivial to be recorded (FU 3: 104). The list of
types of folk poetry and folk customs to be observed might have had a restricting effect on the nature of the material in case collectors focused exclusively on the items of the list, and failed to see other genres. Ulrika
Wolf-Knuts has nevertheless pointed out that Mrten Thors for example,
who has contributed several texts to this investigation, submitted records of
other genres than those requested by the Swedish Literature Society (WolfKnuts 1991: 34). The great bulk of variants amassed in the archive was a
thing of concern for Lagus already in 1895, and the next year he suggested
that folklore collection be diverted to those genres less represented in the
collections, such as superstitions, riddles and proverbs. Later, in 1908, the
The Sources

51

collection was entrusted to kompetenta fackmn (competent professionals),

a measure leading to an increasing specialization among the collectors, and
also fostering a more critical distance to the material at the collection stage,
which was indeed what Lagus had hoped to achieve (Bergman 1981: 2223).
The collectors were dissuaded from combining prose texts. Suitable occasions for collection were mentioned in the instructions as festivities and
winter evenings by the hearth in the case of folktales, and normal speech
situations for proverbs. These were considered the natural performance
contexts ensuring authenticity. Present-day folklore research, however,
emphasizes that the collector influences the situation in natural contexts
simply by being present; there is no such thing as participant observation
unaffected by the observer. In interview situations the instructions stress
that the interviewees personality, the time and the place, and the narrators
expressive powers must be taken into account (FU 3: 105) before fixing on a
fieldwork strategy. The interview was, in practice, the most common
method of collection, and interviewers often quoted previously collected
tradition by way of illustrating what they were looking for. They were also
quick to explain why they were roaming the countryside to dispel any suspicion harboured by informants (Bergman 1981: 28, 30, 39).
Collectors from the peasant class were probably better equipped to create
a relationship with the locals, as they came from similar cultural and social
backgrounds. It is hard to say to what extent the difference in social status
between the rural population and the many elementary school teachers functioning as collectors has moulded the material obtained. Notwithstanding,
not only the social position of collectors and informants influenced their
relationship in the field encounter, but also the dominance of the fieldworker in the interview situation. The unequal power relations in an interview should not be overlooked as an aspect having an impact on the encounter (see Vasenkari 1999: 58, 6566). Nevertheless, many other factors
may affect the state of affairs.
From 1908 onwards the scholarship holders were mostly academics and
the social gap grew even more noticeable, but a small, but diligent band of
local collectors still carried out unpaid fieldwork in the communities where
they lived. Problems in the interaction with the narrator arose partly out of
the insecure political context, people were suspicious of strangers, or strong
religious feelings made people less inclined to pass on sinful things like folk
belief (Bergman 1981: 2829, 32; Wolf-Knuts 1991: 3536).
52

Material and Context

The majority of fieldworkers stressed the imminent death of tradition,

and they perceived their work as a rescue operation aimed at saving valuable ancient folk memories for posterity, and they were looking for genuine folklore and folk culture (cf. Lilja 1996: 27). Education was seen as a
threat to the survival of tradition, while the collectors themselves frequently
worked as teachers and therefore contributed to the documentation of
those traditions their profession was said to obliterate. Contemporary
tradition was not held in very high regard. Neither widespread traditions
nor common ones were alluring objects of inquiry, and these were often
disregarded, as were beliefs the collectors themselves embraced. In coming
to a new place, the collectors were frequently informed in advance of previous activities theresometimes they even knew which informants other
collectors had visitedand when the project of supplementing the archive
was initiated, many fieldworkers avoided consulting former interviewees. Occasionally they went as far as to refuse travelling through parishes previously explored; they only wanted new, hitherto undiscovered
material.
As a result of the efforts at geographic and textual supplementation, texts
of diverging provenance were combined. The procedure is well attested in
song transcriptions, but Anne Bergman does not exclude the possibility of
its application to prose material as well. Many collectors appear to have
entertained the idea of an original form that could be reconstructed if different variants were fused (Bergman 1981: 3032, 36, 40), a thought also in
currency among scholars of the historic-geographical or the Finnish school,
inspired by Kaarle Krohn, professor of Finnish and Comparative Folklore
Research in Helsingfors (cf. Dundes 1974 on the devolutionist premise in
folkloristic theory). Folktales were often recorded as examples of vernacular speech, and the collectors were interested in both folklore and linguistics. The record is nevertheless regarded as a more accurate description of
the collectors language than of the narrators; it does not provide any unmediated access to the speech of the interviewee. Source criticism is further complicated by the propensity of the Societys examiners to correct
examples of the vernacular (Bergman 1981: 3233).
The material was to be submitted as clean copies, neatly ordered and
supplied with a table of contents. This work was time-consuming, and
the internal order of the records was probably disrupted. The Literature
Society advocated separate note pads for different genres, wherefore the
The Sources

53

narrators train of thoughts cannot be followed in the manuscripts. The

severe scrutiny to which the records were subjected might have tempted some collectors to edit their material in order to find favour with
their employer (Bergman 1981: 37; Hggman 1992: 90; Wolf-Knuts 1991:
2627).
In my analysis of the recorded texts, I cite the informant as the narrator
of the given story, in order to avoid confusion. To my mind, there are two
possible ways of approaching the problem of the respective contributions of
collector and informant. The first is through an analogy with unique Old
English manuscripts based on oral traditions, namely regarding the scribe
as a (re)performer (Doane 1991: 8081, 89). Notwithstanding, it should be
noted that there are several important differences between the Old English
chirographer, a scribe involved in the activity of handwriting (Doane 1991:
106), and the latter-day collector. The text produced by the chirographer,
though deviating from its original, whether written or oral, still draws
from the same sources and conforms to the same canon as an oral text; it is
not removed from the oral context, and it is created for an audience capable
of receiving it, or hearing it, as an oral text. It is a valid performance in its
own right, concrete and unique (Doane 1991: 81, 83, 89). This is not
necessarily true of the collectors text. Many fieldworkers, hailing from
another social context, might not have shared the tradition of their interviewees, and their reperformance may not be traditional in Doanes sense.
Likewise, the audience of the textthe examiners of the Swedish Literature
Society, for instancewas perhaps unable to really hear it as an oral text.
Still, there are some common features as well. Doanes description of the
oral written text is also a fairly accurate definition of the recorded text:
it is a product of voice, of a voiced performative situation; its origin is not
text but voice, and its destination is as a visual trace of a material event that
once existed in another register. The record/text that results is the product
of a writer listening to an outer voicehis own or anothersrather than
an inner one. Thus meaning is intertwined with two intentions, that of the
instigator of the text, the speaker, and that of the designator, the writer, in
a process that is less cooperative than it is mutually interventionist (Doane
1991: 88). In the case of the folklore record, intervention is more onesided
as it is chiefly exercised in the editing of the text. This scenario is probably
valid for some of the records in which the collector has rephrased the utterance of the performer.
54

Material and Context

Another possibility is to regard the folklore record as dialogic and

polyphonic, i.e., as containing many perspectives and voices, like Inger
Lvkrona has done in her study of infanticides in 18th-century Sweden,
based on the relations recorded in judgement books (Lvkrona 1999: 36
37). From this point of view, the voices of the collector and the narrator
intertwine in the recorded text, producing a multilayered narrative featuring contradictory as well as concordant tendencies. This alternative
seems more appropriate when the collector has endeavoured to retain as
much as possible of the narrators performance, but due to the difficulties
linked to the fieldwork methods of the time, taking the story down by
hand, and the revision required before submission to the commissioners,
has introduced his own voice into the text. Since this is unavoidable because of the general, intersubjective nature of fieldwork, and not an error to
be skirted at all costs, according to the findings of both reflexive folkloristics and the more recent dialogical approach borrowed from anthropology
(Ehn & Klein 1999: 10; Vasenkari 1999: 51, 56), any recorded folklore text is
subjected to the same conditions. Text and reality are created in intense
interaction with other people, including the interviewer (Ehn & Klein
1999: 11, 79; Vasenkari 1999: 58), and the transcription of the text is an analytical act, as Barbro Klein once put it (Klein 1990).
2.2.3 The Collectors of the Swedish Literature Society
and Their Interviewees
The very first collection submitted to the Swedish Literature Society in
1882 (SLS 1) contained records made by anonymous Ostrobothnian students. I have used two narratives (SLS 1, 3; SLS 1, 11), the first of them
performed by the shoemaker Anders Westerlund in Nykarleby.
H. Sthl sent in a collection of folklore from Ostrobothnia in 1889, and I
have employed one record from it (SLS 10: 598).
Mrten Thors (18621922) was educated as a teacher. He carried out collection in his home district Oravais in 1891 (SLS 22), and in 1892 and 1893
he also extended his activities to Vr (SLS 28, SLS 37). His records were
made using phonetic script according to the presentation of the informant,
Thors states. In the beginning of his fieldwork period in Vr he seems to
have encountered some problems in his attempts at establishing a rapport
with the inhabitants, but these difficulties appear to have vanished after a
while. In his report to the Society he says:
The Sources

55

Vrbon r frbehllsam, reserverad mot frmlingen och ytterst rdd fr att komma i tidningarna. Efter lngre bekantskap och frtroligt umgnge blir han ngot meddelsammare. Dock berttar han hgst ogrna under arbetstid. Sndagseftermiddagen r
den lmpligaste tiden, och man kan d gra en rtt god skrd, om man lyckats ta
honom p rtta sttet. (FU 8: xxvi)
The inhabitant of Vr is reticent, guarded toward the stranger and extremely afraid of
appearing in the newspapers. After a longer acquaintance and intimate intercourse he
becomes slightly more communicative. Yet he is most reluctant to narrate during work
hours. Sunday afternoon is the most suitable time, and you may get a pretty good harvest, provided you have managed to tackle him in the right way.

The collections were praised as good and valuable in the evaluation carried
out by the Literature Society. Thors consistently omitted the names of the
narrators and more specific details of their residence. Ulrika Wolf-Knuts
suggests this might be due to the generality of the narratives, but Thors
could also have had other motives for leaving the provenance of the texts
less meticulously specified (Wolf-Knuts 1991: 2829). Here I have utilized
sixteen records drawn from Thors three collections (SLS 22, 4; SLS 22, 10;
SLS 22, 11; SLS 22: 1617; SLS 22, 21; SLS 22, 26; SLS 28, 3; SLS 28, 12;
SLS 28, 19; SLS 37, 3; SLS 37, 5; SLS 37, 6; SLS 37, 8; SLS 37, 11; SLS 37,
28; SLS 37, 98).
Karl Petter Pettersson (18571912) was the elementary school teacher in
Nagu, later in Ini, and he also functioned as a parson, the chairman of the
local government committee and as the postmaster of the community. Of
his work as a collector of folklore Anders Allardt writes: Det vilar en kck
omedelbarhet ver hans stil, och en del av hans sagouppteckningar ro synbarligen frgade av hans personliga gemyt (A spirited immediacy suffuses
his style, and some of his folktale records are visibly coloured by his personal disposition) (Allardt 1920: 364365). In other words, Allardt thought
that Pettersson directly influenced the language and form of the records he
made. In 1890 Pettersson donated two collections of folklore from Nagu to
the Swedish Literature Society (SLS 21; SLS 31), whence four texts originate (SLS 21, 8; SLS 21, 29; SLS 31, 141; SLS 31, 146). In 1893, he submitted
another collection containing a record I have utilized (SLS 33: 201207).
J. A. Nygren submitted a collection of miscellaneous texts recorded in
18921894 to the Society. One entry, copied from the songbook of the
singer Jakob Lassus in the village of Kerklax in the parish of Maxmo in
1892, will be quoted in this investigation (SLS 45: 136137).
Emil Norrback delivered a collection from Sideby, Ostrobothnia, cover56

Material and Context

ing a five-year period from 1892 to 1896 to the Literature Society (SLS 56).
His records, ordered thematically according to the supernatural beings
appearing in them, are written in standard Swedish and provide no information on informants or details of provenance. The form of the records
indicates a method of compilation of material from various oral sources,
which have been woven into a whole by the collector himself, and it is probably Norrbacks voice we hear in the text included here (SLS 56: 151153).
In 1895 J. Torckell was granted a scholarship for carrying out fieldwork
on the land Islands, and in the field he also encountered narratives of
trolls. The text used here is in normalized language lacking references to
any narrators. Torckell gives vent to a slightly deprecating attitude to folk
belief in his summaries and comments (SLS 59: 4849).
According to Ann-Mari Hggman, Johannes Dahlbo (18501923) from
Prtom in Ostrobothnia was one of the first contributors to the folklore
collections housed in the archives of the Swedish Literature Society. An
elementary school teacher by profession, he lived and taught in the parish
of Korsholm from 1894 onwards (Hggman 1992: 88). His collection from
1898 (SLS 65), here represented by eight records, is intriguing particularly
because of his brackets at the end of the texts, where he might ponder the
lines of transmission, sketch the linguistic context (the co-texts) etc. The
fisherman Johan Berg from Sundom narrated two stories of trolls (SLS
65: 45; SLS 65: 47), one of them a retelling of a personal experience he had
heard narrated by Bata Svarfvar (SLS 65: 45). Arstu Jucka from Solf was
another of Dahlbos more communicative interviewees, and here I have
utilized one of his narratives (SLS 65: 49). His wife and son, Anna Maja
Nordbck and Isak Johansson Nordbck, jointly contributed the sole explicitly racy story found in my material (SLS 65: 44). Prtom informants supplied several small items; one was conveyed by Dahlbos grandmother (SLS
65: 48), who is not named, one is attributed to the elementary school teacher Hagman (SLS 65: 48). Dahlbos uncle, Gabriel Norrback who earned his
livelihood as a farmer, posthumously got one of his stories immortalized in
his nephews collection (SLS 65: 4344). A story narrated by Beata Eriksdotter Norrback has also been included (SLS 65: 89). All texts are in standard Swedish.
Erik Finne submitted a collection from the parishes of Pedersre and
Purmo, Ostrobothnia, in 1899 (SLS 71). Regarding his personal fieldwork
principles he states:
The Sources

57

Vid antecknandet af traditionen har jag ltit flere personer bertta om samma sak, d.v.s.
samma tradition, hvar efter jag gjort ett uppkok p hvad som slunda serverats. Genom ett dylikt tillvgagende tror jag mig ha vunnit den frdelen, att detaljerna tydligare
framst n annars hade varit fallet. Fr frigt har jag s samvetsgrant som mjligt frskt tergifva alt, hvad fr mig berttats, utan att hvarken tillskarfva eller frntaga. (FU
14: ii)
In the recording of the tradition I have made several people tell about the same thing,
i.e., the same tradition, and after that I have made a concoction of the things thus
served. By such a procedure I believe myself to have gained the advantage of the details
emerging more distinctly than would otherwise have been the case. As for the rest I
have tried to reproduce everything that has been told me as conscientiously as possible,
without either adding or subtracting.

In the record incorporated into my material (SLS 71: 3234), Finne is true
to his principles in the respect that he has avoided crediting any informants,
but the text is still rendered in the vernacular.
Filip Sundman was also collecting folklore in 1899 (SLS 72). Like many
other collectors, he had no great confidence in the ability of folklore to survive in the modern age, and in his report to the Swedish Literature Society
he writes:
Af mitt arbete erhll jag det intrycket, att det r hg tid att samla detta material, om
det icke redan r ngot fr sent. Huru mycket har inte fallit i glmska under de senaste
decennierna? Intresset fr allt gammalt har svalnat betnkligt, beroende dels p bildningens framtrngande, dels p kadt arbete vid jordbruket. De unga lr sig icke mera
ngot sdant, och de gamla lmnas ofta i sticket af minnet. Gldjande var dock att se
den beredvillighet, hvarmed de stodo till min tjnst, samt det vnliga bemtande, som
p ytterst f undantag nr, fveralt kom mig till del. (FU 14: iii)
From my work I have gained the impression that it is high time to collect this material,
if it is not already somewhat too late. How much has not fallen into oblivion during
the last few decades? The interest in everything ancient has precariously cooled, due
partly to the advancement of education, partly to the increased workload in agriculture.
The young no longer learn such things, and the old are often betrayed by their memory.
It was, however, gratifying to see the readiness with which they were at my service, and
the kind reception that, with very few exceptions, was accorded me everywhere.

The record in standard Swedish used here was made in the parish of Pojo
in the beginning of July following the dictation of an informant named
Sderbergskan, whose first name is not given (SLS 72: 36).
Two years later, in 1901, Torsten Stjernschantz (18821953), later curator
of the Academy of Arts Ateneum, did some fieldwork for the Literature
58

Material and Context

Society (Allardt 1920: 366) when he collected a narrative regarding trolls

(SLS 80: 4647). This specific text is rendered in standard Swedish, but
some of the other ones are in the vernacular. Stjernschantz does not name
his interviewees, nor does he specify the village where the story was collected. Of his working conditions he complains:
Kimito r annars ett mycket otacksamt arbetsflt fr upptecknaren av folklore, ity att
bland annat de talrika bruken, Dalsbruk, Bjrkboda och Skinnarvik, genom infrande
af en myckenhet ls frmmande befolkning samt inkpande av stora delar af n, infrt nya tiders seder och uppfattning, hvarigenom den gamla allmogekulturen rkat i
glmska. (FU 16: xxxiii)
Kimito is otherwise a very thankless work environment for the recorder of folklore, because the numerous works, for instanceDalsbruk, Bjrkboda and Skinnarvikhave
introduced the customs and ideas of modern times through the introduction of a multitude of unknown drifter residents and the purchase of large parts of the island. As a
result the old peasant culture has fallen into oblivion.11

H. R. A. Sjberg did fieldwork in the parish of Replot in 1905, and one of

the folk narratives he recorded has been cited in my analysis. The narrator
is the crofter H. Gammal (SLS 98, 46: 72).
G. E. Lindstrm collected Swedish-language folklore in the predominantly Finnish-speaking parish of Hausjrvi in the late 1870s, and one text
from the village of Kllarhult springs from this collection (SLS 166e, 2).
The name of the narrator is not mentioned. A record from the village of
Hstble in the parish of Sjunde has also been cited (SLS 166: 687689),
as well as another record from the same parish (SLS 166: 728729).
Henrik Kullberg submitted a collection of folklore from Nyland in 1912,
and one text, narrated by the cottar Johan Stark, hailing from Virble in the
parish of Strmfors, has been used (SLS 208: 678679). Two years later
Kullberg conducted fieldwork in Nyland again, interviewing Johan Stark
once more, and I have utilized a record made on this occasion (SLS
228: 8993).
Several folk high schools were engaged in the collection of folklore, and
in the 1890s pupils of the school in Kronoby, inspired by their teacher
Johannes Klockars, a diligent local collector, made transcripts of folklore,
which were subsequently submitted to the Swedish Literature Society (SLS
220). Karl Viktor Ulfves recorded a narrative from the parish of Lappfjrd
11 The English translation is slightly adjusted to make the text more intelligible.

The Sources

59

entitled En jgares fventyr (The Adventures of a Hunter) in the manuscript (SLS 220: 240242). The language is normalized and the informant
is not mentioned, unless it is indeed the collector himself. An anonymous
collector has donated an account of changelings under the heading Vidskepelse (Superstition), and Sideby is acknowledged as the place of recording (SLS 220: 6769). The Literature Society has made an addendum,
suggesting that the modest collector might be Emil Norrback. Yet another
text, ostensibly based on a tradition to be found in the parish of Korsholm
(SLS 220: 167168) is also included in this collection. Likewise, a record
from the parish of Lappfjrd has been utilized (SLS 220: 248).
The poet Jacob Tegengren (18751956), too, was interested in folklore,
and he worked as a local collector for the Swedish Literature Society (WolfKnuts 1991: 31). He visited the parish of Korsns and the island of Replot in
19121913 (SLS 215, 160; SLS 215, 248; SLS 215, 249; SLS 215, 250) and
finished two collections from Vr in 1922 (SLS 324: 292; SLS 324: 296; SLS
324: 299; SLS 333: 208210; SLS 333: 220221; SLS 333: 223); ten entries in
total concern us here. Tegengren used standard Swedish, and did not normally supply the narrators name, although exceptions to this rule do exist.
V. E. V. Wessman (18791958) was an unremitting collector of folklore.
His fieldwork methods and biography have been carefully analyzed by Gun
Herranen (Herranen 1986: 213230). As Wessmans experience as a collector grew, he became more sensitive to the importance of the narrator, and
having delivered merely his name in connection with the transcribed text
during his first years as a scholarship holder, he later appended miniature
biographies of the storytellers to the collection (Herranen 1986: 221222). In
the summer of 1909 this re-orientation was manifest in the brief notes on
the personal character of some of the narrators; two folktales derive from
this collection (SLS 137). Both represent the same taletype and are narrated
by the former tenant crofter Grnholm in Backa (SLS 137 I, 1) and Otto
Nylund, a dependent tenant in Snappertuna (SLS 137 II, 1). Wessman
reports that Grnholm was called GammelDrosi (Old Drosi) after the
name of his croft and that he was over 80 years old, bedridden for many
years. When he was young his repertoire included a large number of tales
and songs, but by the time Wessman visited him he had forgotten most
of them. Otto Nylund is described as a very talkative old man of 78 years.
His memory was said to have been incredible in his younger years.
Wessman wrote: What he once heard narrated, he remembered without
60

Material and Context

difficulty for years on end, thus (?) people tell that when he returned home
from church, he was able to recite the parsons sermon word for word.
Whether this statement is literally true or a simile for his perfect memory is
hard to tell. People used to call him the Bishop, but he was also considered an inveterate liar, which Wessman suspected might be occasioned by
his superstition (SLS 137, Smrre notiser angende berttarna (Minor
Remarks on the Narrators), no pagination).
The more elaborate biographies reach their peak in the 1911 collection; a
few years earlier Wessman had encountered Berndt Strmberg (18221910),
the blind master storyteller, whose narratives touched on trolls as well.
Gun Herranen, who has devoted much of her efforts as a folklorist to
explore every available detail of Strmbergs life and tales, claims that
Strmberg was unique among Swedish-speaking narrators in Finland.
Wessman recorded 120 folktales from him, and previous to his visits other
collectors had documented some of his stories, but to no great extent. The
texts are often long and complicated, yet easy to follow and stringent.
When Wessman found Strmberg in 1909 he was already 87 years old and
lived in a fishing croft in Leksvall in the parish of Ekens. Gun Herranen
deems the extraordinary wealth of details a characteristic feature of
Strmbergs style (Herranen 1995: 156157; cf. Herranen 1984, 1987, 1993).
Here I have used seven of his tales (SLS 202 Sagor II, 1; SLS 202 Sagor II,
8; SLS 202 Sagor II, 15; SLS 202 Sagor II, 19; SLS 202 Sagor II, 28; SLS
202 Sagor II, 61; SLS 202 Sagor II, 66). The same collection (SLS 202)
contains another story of trolls, narrated by the female elementary school
teacher Sandholm (SLS 202 Sagor I, 8). Wessman compares her version
with Strmbergs in the notes accompanying the text, though it seems he
did not record Strmbergs tale in full. In 1915 Wessman did fieldwork in
the southwestern archipelago, and on this occasion he recorded a text in
the village of Ut in the municipality of Finnby (SLS 255: 175176).
Wessman primarily regarded folklore as an element of social intercourse,
Herranen claims, and he stressed the importance of a trusting relationship
to the folk. He felt no need to be thought of as a learned man, and he
pretty much let people believe whatever they wished of him. Ideologically,
his was a quest for genuine folklore, collected in a linguistically and socially
homogeneous environment. In keeping with a general tendency among
collectors of the period, he discarded traditions inspired by literate culture
(Herranen 1986: 221222, 224, 226227). In 1917, when the country was on
The Sources

61

the verge of a civil war, he travelled the Ostrobothnian countryside recording narratives, 12 of which are included in my corpus (SLS 280). Karolina
Grannas from Hrkmeri in the parish of Lappfjrd, 78 years old at the
time, contributed one story of trolls (SLS 280: 362). Robert Hannus, a 72year-old Hrkmeri resident, recounted one story (SLS 280: 357), while Olga
Nummelin, 79 years old and hailing from Sideby, told a similar one (SLS
280: 379). Anders Ek from Krklax in the parish of Maxmo, 82 years old,
narrated a story retelling the supernatural experience of a work mate (SLS
280: 503504). One text from Vr lacks information on the informant
(SLS 280: 635636), which is rather unusual for a collector like Wessman.
In the parish of Solf, Sofia Snfs, 69 years old (SLS 280: 131) and Isak Snfs
(SLS 280: 132) from the village of Munsmo, narrated one story each, as did
Johanna Berg, 84 years old, in the village of Rimal (SLS 280: 136), Albertina
Hellman, 86 years old (SLS 280: 129) and Eva Sund, 87 years old, from the
village of Sundom (SLS 280: 130). Johan Grnlund, farmer in Taklax in
the parish of Korsns, 69 years old (SLS 280: 295) and Anders Rovhk, 81
years old, living in Frjns in the parish of vermark (SLS 280: 312) also
told him one story each. A year later, in 1918, he was recording folklore in
the parish of Ekens when he met Alma Sundstrm in Skld, who related
a narrative for him (SLS 290: 493).
As a rule, Wessmans records of folktales are in the vernacular, reproduced with painstaking faithfulness, while other texts, such as the ones
extracted from SLS 280, are rendered in normalized language with occasionally inserted dialectal expressions explained in notes at the bottom of the
page, a practice he evidently considered superfluous in the transcripts in the
vernacular, in which the notes mostly relate to other extant versions, in
manuscript or in print. For the longer narratives, the notes are thus recruited for exercises of scholarly comparison, whereas in the case of shorter
ones, the annotations form the running commentary of a linguist, even
though the words selected for annotation seem somewhat puzzling in
hindsight. A possible explanation for Wessmans divergent transcription
methods is that the texts comprising SLS 280 were something of a byproduct of the collecting trip; the real purpose was to collect single words
in the vernacular, not entire stories (SLS 280, Wessmans field report).
Axel Olof Freudenthal (18361911) was a prominent figure in the Swedishspeaking circles of the time; he strove for the protection of the Swedish
vernaculars and Swedish culture in Finland (Steinby 1985: 6166), and he
62

Material and Context

made his first advance for the cause by joining Illegala nylndska afdelningen
(The Illegal Division of Nyland) at the Imperial Alexander University,
legalized in 1868. Later he was one of the founders of the Swedish
Literature Society. He was also active in the Svenska landsmls-freningen i
Finland (The Swedish Society for the Vernaculars in Finland) from 1874
onwards (Bergman 1981: 8; Steinby 1985: 13; Wolf-Knuts 1991: 19). He
donated some records made between 1860 and c. 1900 to the Swedish
Literature Society, and the collection was filed as SLS 213. I have utilized
one transcript taken down in the vernacular, but still lacking any attribution to a narrator, and it ends in mid-sentence (SLS 213, 184).
Gabriel Nikander (18841959) was professor of Nordic Cultural History
and Ethnology at bo Akademi University. Prior to his appointment he
had worked as a teacher at a folk high school, among other things, and he
was a highly esteemed collector for the Literature Society (Steinby 1985: 161
162). In 1911 he collected folklore in Nagu, and an item attributed to an
H. I. in the record has been included in my material (SLS 192a: 141). Unfortunately this person was accidentally left out of the list of informants, as
far as I can tell. In 19121913 he collected folklore in Kyrksltt, and one text
has been used in my analysis (SLS 217: 542). In the summer of 1913 he was
doing fieldwork in Kimito when he recorded a narrative from Stava
Sderstrm, born in 1855 and living in the village of Pvalsby at the time of
collection (SLS 226: 150151), and he also documented an item collected
from Mathilda Gustafsson, born in 1875 and living in Ini (SLS 226: 462).
On the same field trip he visited Pargas, where he collected one story (SLS
226: 171172). His contemplations in the field report recall those of Torsten
Stjernschantz:
Arbetet hr var svrt. Redan p 1870-talet trodde Fagerlund sig samma konstatera, att
vidskepelsen nra nog utrotats i Korpo och Houtskr, och p 1890-talet sade Elmgren
detsamma om Pargas.12 Deras ord bra tydas s, att folkseden i dessa sjfartsidkande
socknar sedan lnge befunnit sig i upplsning och att endast fragment av folktron kvarleva. D hrtill kommer att folkbildningen har endast kommit halvvgs mot i Nyland,
12

L. W. Fagerlund published his Anteckningar om Korpo och Houtskrs socknar

(Notes on the Parishes of Korpo and Houtskr) in the 1870s, a work characterized by
Otto Andersson as a very significant contribution to the knowledge of the Swedish
people and Swedish folklife in Finland. Sven Gabriel Elmgren wrote a Beskrivning fver Pargas socken (A Description of the Parish of Pargas), by Andersson considered important, but unfortunately marred by serious factual errors (Andersson 1967:171, 133134).
The Sources

63

och sledes misstron mot frmlingen, som efterforskar folktrons relikter, r synnerligen
stor, fljer hrav att sdant arbete r svrt i dessa trakter. Jag har drfr antecknat talrika materiellt etnografiska notiser, d sdana om folktron icke str att erhlla. En del
allmnt frekommande vidskepelser har jag inte antecknat. (SLS 226:525)
Work here was hard. Already in the 1870s Fagerlund thought himself [able to] establish the same, [the fact] that superstition has been almost entirely eradicated in Korpo
and Houtskr, and in the 1890s Elmgren said the same of Pargas. Their words should
be thus construed that folk customs in these seafaring parishes have long been in a state
of disintegration, and that merely fragments of folk belief survive. In addition, as adult
education has progressed at only half the rate as in Nyland, and suspicion of the stranger inquiring into the relics of folk belief is therefore particularly great, such work is
consequently difficult in these areas. Therefore I have made notes on numerous materially oriented ethnographic items, since ones on folk belief are not in my power to
obtain. Some generally circulating superstitions I have not recorded.

That Nikanders comments concern previous scholarship is perhaps not all

too surprising, as he is one of the competent professionals enlisted by the
Swedish Literature Society, and we become well aware that he was not
spared the problem of encountering suspicion among his presumptive informants. In this instance adult education is actually credited with a positive influence on the folk, making them better equipped to understand the
inherent value of ancient folk belief and the importance of preserving it for
the nation, science and later generations, a mission education had not yet
fulfilled in Kimito. Like many other collectors (cf. chapter 2.2.2), Nikander
ignored very common tradition, and the records are made in standard
Swedish.
Jakob Edvard Wefvar collected for the Swedish Literature Society as
well, and here I have used one text from the parish of Vr, recorded from
the rope-maker Kastell living in the village of Rejpelt (SLS 275: 8788). The
narrative was documented some time in the 1870s.
Vivi Peters (18931945) was one of the few female collectors. She followed in her fathers, K. P. Petterssons, footsteps as she submitted her first
collection to the Swedish Literature Society in 1916. She extended the Societys collecting activities to new fields and areas of research, such as food
and hygiene. Her premier collection was well received, which spurred her
to go on. Ernst Lagus, otherwise so severe in his criticism, held her in high
regard. The 1918 collection (SLS 320) contains a record from Fina Lilja, a
crofters daughter and party cook from Finn on the island of Korpo (SLS
320: 80). Lilja was born in 1846.
64

Material and Context

The pupils of the folk high school in Vr delivered a collection to the

Literature Society in 1922, probably inspired by Jacob Tegengren, the local
collector (Wolf-Knuts 1991: 31). One of the texts, signed by Elna Kllbacka
(SLS 338: 2122), has bearing on my investigation, and it is intriguing to
note that it refers to a local tradition recorded by many other collectors as
well. In the same year, the folk high school in Nrpes, too, submitted a
collection to the Society (SLS 319), whence I. N.s essay Vidskepelse i min
hembygd (Superstition in my Home District) is taken (SLS 319: 3132).
Felix Andersson was collecting in Storpellinge in 1925 when he recorded
a story of a troll (SLS 374: 1012). No informants are acknowledged by
name and the language is normalized, though some quotations framed in
direct discourse have a vernacularized tinge. In one instance he has added
alternative phrasings in the margin, which might imply that he heard the
story performed several times.
The notes on folk belief compiled in 1852 by Mrten Lassus, a senior
juryman living in Vr, were incorporated into the collections of the Swedish Literature Society in the form of typed copies (SLS 299); of these I
have cited six stories (SLS 299: 3132; SLS 299: 33; SLS 299: 3334; SLS
299: 3435; SLS 299: 3536; SLS 299: 4748). The document was, in other
words, produced long before the Society even existed, and accordingly does
not conform to its rules of disposition. Incidentally, it is the only collection
of folklore thought to have been instigated by Ranckens appeal in the paper Ilmarinen in 1848.
2.2.4 Printed Sources
I have also made use of printed sources. Wilhelm Sjberg and Jacob
Tegengren contributed one and two entries respectively, concerning names
and legends of stones and rocks in the parish of Replot in the former case,
and the parish of Vr in the latter, published in the periodical Budkavlen
issued since 1922. Sjbergs article is included in the very first issue (Budkavlen 1922: 3941), Tegengrens appeared a couple of years later (Budkavlen
1923: 86; Budkavlen 1924: 85). Two items in Sjbergs text refer to trolls,
while Tegengrens articles contain one item each that have been employed
in my analysis. Tegengren was likewise involved in the publication of
Bygdeminnen, a series in three volumes running between 1909 and 1912,
containing folklore collected by students at the folk high school in Nrpes.

The Sources

65

There are four narratives in these volumes that I have quoted in my research (Bygdeminnen 1909: 3334; 1910: 4142; 1912: 5657). The first one was
collected by Alfred Lassfolk in the village of Yttermark in the parish of
Nrpes, the second by K. J. Valsberg in vermark, the third by Else
Tegengren from Lena Stenlund, farm mistress in Yttermark, and the
fourth by John Ahlbck in Korsholm.
L. W. Fagerlunds Anteckningar om Korpo och Houtskrs socknar
(Notes on the Parishes of Korpo and Houtskr) alluded to by Nikander
was printed in 1878 in the periodical Bidrag till knnedom af Finlands natur
och folk (Contributions to the Knowledge of the Nature and People of
Finland). It is an extensive examination touching on many aspects of the
history and culture of the parishes in question. Included therein is a tale
from Houtskr, Trollgobbin och vallgutin (The Old Troll and the
Shepherd Boy) (Fagerlund 1878: 169178). Axel Olof Freudenthal appended a story m in tjrng, som vast bjrgteji (On a Woman Who Was
Abducted) as a linguistic sample in his monograph on the Vr vernacular,
Vrmlet (Freudenthal 1889: 197). Both Freudenthal and Fagerlund employed phonetic script and the language of their samples evokes vernacular
speech patterns, but most other printed texts are in standard Swedish, and
they have been edited to suit the tastes of a reading educated public. This
is obvious in some texts, as the collectors have polished the material into
a finely honed piece of romantic poetic description (see e.g. Hembygden
1912: 120).
Budkavlens predecessor Hembygden is a rich source of folklore as well.
Rafael Karsten published an essay on Kvarlevor av hednisk tro bland
Finlands svenskar (Relics of Heathen Folk Belief Among the Swedes
of Finland) in 1910, where he quoted a story of trolls (Hembygden 1910:
145). In his conclusion, the author stresses the importance of collecting
beliefs and exhorts the journals readership to engage in this work. He
exemplifies:
De freteelser som vanligen g under benmningen skrock och vidskepelse br han ej
hna och bele ssom betydelselst nonsens, utan att han br ska att frst dem. De
utgra som vi sett fullt naturliga yttringar av primitivt tankeliv, och nutidsmnniskan
har anledning att tillvarataga och bedma dem med samma pietet som vriga minnesmrken av en frgngen tid. (Hembygden 1910:149)

66

Material and Context

He should not deride and sneer at the phenomena usually subsumed under the term
superstition13 as meaningless nonsense, but rather try to understand them. They
constitute, as we have seen, entirely natural manifestations of primitive thought, and
modern man has reason to preserve and judge them with the same piety afforded other
monuments of times past.

Folklore is used as a means of shedding light on primitive thought and a

time long past.
Uno Sandvik was one of Hembygdens local collaborators, and in 1911 he
submitted a record on trolls from Terjrv, Ostrobothnia (Hembygden 1911:
114). The folk high schools were contributors of prime importance in the
collection of folklore (Steinby 1985: 171); Svenska sterbottens Ungdomsfrbunds folkhgskola (The Folk High School of the Youth Association
of Ostrobothnia) delivered a small collection of texts to Hembygden in 1912.
Two of these are devoted to trolls; one was sent in by Edith Smeds from
the parish of Solf, and one by Edvard Hellman hailing from the same
parish (Hembygden 1912: 120121). J. Kaustinen, a local collector in Vr,
supplied some particulars on traditions of trolls in his parish in the same
issue (Hembygden 1912: 2021). The next year Edit Hkans from the village
of Petsmo submitted a text on trolls, although the heading designates them
as earth-dwellers (Hembygden 1913: 105). Jacob Tegengren was a diligent author in Hembygden too, and he contributed a story of trolls in the 1916 issue
(Hembygden 1916: 6263). Wilhelm Sjberg did not fail to appear either,
and published Sgner om Strkgobbin (Legends of the Old Strk
Man), narrated by Alfred Gdda, a fisherman from Panike in the parish of
Replot, in Hembygden 191718: 122123.
Nylndska afdelningen (The Division of Nyland) at the Imperial
Alexander University, referred to above in its illegal form, had funded
collection of folklore by granting scholarships to fieldworkers from 1860
onwards, and in 1882 the decision on the publication of the collections thus
obtained was made. The series, entitled simply Nyland, swiftly became the
model for both publishing and collecting folklore in the Swedish circles in
Finland. The first volume appeared in 1884, the last at the turn of the
century (Andersson 1967: 151152, 172173). In this study the volumes of
folktales, tomes 2 and 6, are most useful. Once again we encounter Jakob
13

The Swedish words skrock and vidskepelse both translate as superstition, and it seems unnecessary to replicate the English equivalent for the sake of accuracy in translation.
The Sources

67

Edvard Wefvar, who collected in Nyland as well, and in the parish of Karis
he recorded the tale of Krin Trtjla (Catherine with the Wooden
Skirt) from the lips of Edla Theilenius from Mangrdsby, born in 1846
(Nyland 1887, 19: 1517). Another Karis resident, Johan Bckstrm, was also
visited by Wefvar (Nyland 1887, 26: 2627). G. E. Lindstrm conducted
fieldwork in Svartbck in the parish of Sjunde, where he recorded a narrative from Lindholm, whose first name he does not mention (Nyland 1887,
77: 9093). All these texts are published in the vernacular utilizing phonetic script, which is not the case with Gustav bergs record from Lappom in
the parish of Strmfors (Nyland 1887, 180: 209211). Isak Eriksson Smeds,
who was later lecturer in Swedish in Joensuu, has contributed one record,
made in the vernacular, to the investigation (Allardt 1920: 363; Nyland 1887,
114: 137138). Adolf Backman, an editor, collected two stories of Hobergsgubben (The Old Man of Hoberg), one from vitsble in the parish of
Mrskom, the other from Lomble in Borg (Nyland 1896, 25: 19; Nyland
1896, 26: 2021). Both of the texts are in the vernacular. In Embom in the
parish of Liljendal Isak Alexius Bjrkstrm, a freeholder, interviewed the
dependent tenant Vilhelm Vilhelmsson, a younger man, whom he considered so noteworthy a narrator that he merited being mentioned by name
in the travel report (Allardt 1920: 316; FU 9: xxxiv; Nyland 1896, 129: 151154).
L. W. holm, who was the headmaster of a folk high school, recorded a
tale of trolls in the parish of Tenala in 1893 (Nyland 1896, 141: 179182).
The conception of folklore entertained by the Division is aptly illustrated
by G. A. bergs introduction to the second volume of Nyland:
Publikationens syfte r icke i frsta hand att ska stadkomma en samling roande frtljningar, utan dess uppgift r att t fderneslandet rdda terstoden af den rika folkdiktning, som under rtusenden lefvat hos vr folkstam, som fljt den slkte ifrn slkte
och i skiftande bilder afspeglar hela dess forna vrldsskdning, att med ett ord rikta
den kulturhistoriska vetenskapen med nytt material. Mycket r nmligen utur dem att
hmta fr den, som vill studera historien i dess innersta grund, som vill lra knna folkets anda och skaplynne, och flja hela gngen af dess inre utveckling. Ty folksagan
tillter oss mngen blick in i lngst hnsvunna tider, den ger en trogen och lefvande
bild af vra frfders seder och levnadsstt, och sprider fver forntiden ett ljus, hvilket
icke alltid str att vinna ifrn skriftliga urkunder vra folksagor frtjna ett bttre
de n att fr alltid begrafvas i glmskans natt. Men skola de rddas undan frstrelsen, br detta ske snart. De ro nmligen p vg att d ut eller frdrfvas under inflytande af en ny tid och nya frhllanden, och endast i aflgsnare bygder lyssnar man nnu
med begrlighet till dessa frklingande ljud, hvilka en gng voro hela folkets egendom
och den frsta nringen fr vra fders bildning. (Nyland 1887: viiviii)

68

Material and Context

The purpose of this publication is not primarily to try to accomplish a collection of entertaining narratives, but its task is rather to salvage the remains of the plentiful folk
poetry, for millennia flourishing within our tribe, that has accompanied it from generation to generation and in varying images reflecting the whole of its ancient world view,
for the native country, in one word to provide the discipline of cultural history with
new material. For there is much to find therein for someone who wants to study history in its innermost core, [someone] who wants to get to know the spirit and creative
genius of the folk, and trace the entire process of its inner growth. For the folktale permits many a glimpse of times long since past, it gives a true and lively picture of our
ancestors customs and ways of life, and sheds light on ancient times not always possible to gain from written documents our folktales deserve a better fate than to be
forever buried in the darkness of oblivion. But if they are to be saved from annihilation,
it must happen soon. For they are about to be extinguished or corrupted under the influence of a new age and new conditions, and only in remote areas do people still avidly
listen to these sounds dying away, which were once the property of the entire people
and the first nourishment for the education of our forefathers.

These apocalyptic words of warning resonate through much of the folkloristic literature of the timeI have already reviewed some examplesand
we also find another recurring motif, the notion of folktales as an instrument for divining certain particulars of a history long lost and shrouded in
mystery.
Janne Thurman appealed to similar sentiments in his introduction to an
essay on Ngra hednaminnen i pargasbons diktning (Some Heathen
Relics in the Poetry of the Inhabitant of Pargas), which also contained
specimens of folklore collected by the author himself, and one of the entries specifically relates to trolls (Thurman 1891: 111112). As a preamble
Thurman mentions the positive and negative effects of education on the
intellectual life of the folk and then continues:
Den som nskar hra prof p allmogens urgamla folkdiktning och vill rdda hvad som
numera rddas kan, han mste begifva sig till civilisationens utmarker, till trakter,
hvarest nghvisslans glla ljud ej nnu bortdrifvit sj- och skogsjungfrun, gastar och
andra spken. P sdana stllen r man nnu i tillflle att f hra vra frn hedenhs
bevarade valltar skalla genom dalen och att se folket samladt p sndagseftermiddagen
till vitter sagoskmtan. Lyckas man tillvinna sig allmogemannens frtroende, kan man
f taga plats i hans krets och anteckna mngen ldrig sgen, som tljes vid dylika tillfllen. Men det r svrt att fvervinna folkets misstroende. fven i de mest aflgsna bygder drager det sig fr att t en frmling gifva sina sagor, gtor och sin gudatro. Det tror
nmligen, att de bildade anse dessa sagor och snger barnsliga, och att de skratta t dess
uppfattning af djvulen och mnga af freteelserna i naturen Afsikten med denna
uppsats r ingalunda att nedstta den allmoge, ur hvars skte jag samlat nedanstende;
lngt drifrn, jag vill blott visa, huru i vr folkdiktning finnas mnga kta fornnordiska
The Sources

69

prlor, vrda att hopsamlas och bevaras. De hafva visserligen under rhundradenas lopp
filats af och slitits, men det oaktadt ro de fr oss, ssom alla minnen frn vr folkstams
barndom och ungdom, dyrbara och kra. (Thurman 1891: 104105)
He who wishes to hear specimens of the ancient folk poetry of the peasantry and wants
to rescue what nowadays may be rescued, he must go to the outskirts of civilization, to
places where the shrill call of the steam whistle has not yet banished the mermaid and
maiden of the forest, ghosts and other spectres. In such places you are still in a position
to hear our herdsmans songs preserved from times immemorial echo through the valley
and to see the people assembled to a learned jest of tales on Sunday afternoon. If you
manage to earn the farmers confidence, you can take your place in his circle and record
many an ancient legend told on such occasions. But it is hard to conquer the suspicion
of the folk. Even in the remotest districts they hesitate to give their tales, riddles and
faith to a stranger. For they believe that the literati deem these tales and songs childish, and that they laugh at their conception of the Devil and many other phenomena in
nature. The purpose of this essay is absolutely not to disparage the folk in whose
midst I have collected the following; far from it, I only wish to show how many genuine treasures from the Old Norse, worthy of being collected and preserved, [still] exist
in our folk poetry. In the course of the centuries they have certainly become eroded and
hackneyed, but regardless of that they are, like all monuments of our tribes childhood
and youth, precious and dear to us.

This excerpt spans almost all the themes occupying the minds of the collectors and researchers of the period; ideology and conceptions of folklore
intermingle with the practical problems facing a fieldworker. Folklore is
treasures from the Old Norse, eroded and hackneyed of course, but still
evidence of a glorious past, our tribes childhood and youth. The gaze
was turned toward times immemorial, and it is the historical dimension
that legitimizes collection, which is obstructed by the tendency of the folk
to distrust the collectors intentions. This is not wholly astonishing, as the
upper classes did not have too high an opinion of peasant culture earlier on
(Andersson 1967: 120). Old Norse culture was important as an ideological
point of reference in the late 19th and early 20th centuries, and furnished a
uniting factor in the rallying of Finland-Swedish resources (Lnnqvist
2001: 235244).
2.2.5 On the Principles of Transcription and Translation
Concerning the transcription of the Swedish original, I have quoted passages verbatim retaining the spelling used by the collector in cases where
the record is made in the vernacular, but not employing phonetic script.
Diacritic marks are reproduced according to the manuscript, to the extent I

70

Material and Context

have been able to discern them; some texts are rather difficult to read.
Records in phonetic script have been simplified by converting phonetic
notation into conventional letters, as all readers might not be familiar with
this specific script. Vowels in the zone between e and have generally
been transcribed as , and sounds between u and o as o, unless this
procedure would impede a reasonably fluent reading of the text. Ng is
written ng, except in front of a g. In the phonetic script i stands for
both i and j, and here I have normalized the transcription according
to the rules of standard Swedish. Double letters, notated as a single vowel
or consonant underlined in the manuscript, have been reproduced in line
with standard orthography. Again, diacritics, though quite rare, have been
supplied in accordance with the source. Letters in the beginning of a
sentence are not capitalized in the phonetic script, and I have retained this
feature.
As for the English translations, all are mine unless otherwise stated. In
general I have tried to preserve the expressions and the grammar of the original, and sometimes the English rendition will suffer for it. Nevertheless,
I believe this is the best way to preserve at least part of the colour of the original, which would otherwise be lost to the English-speaking reader. Unfortunately my command of English-language dialects is not such that I
would have been able to translate texts in the vernacular into English ones.
Instead I have had to content myself with translating them into standard
English, employing contractions and other more informal features to indicate the vernacular status of the texts.

2.3 Context
In this section I will give an account of the historical, social and religious
context in which the texts analyzed in three of the main chapters were
narrated. The focus lies on the South Ostrobothnian parish of Vr where
these narratives were collected.
2.3.1 Historical and Social Context
Vr borders on the parish of Oravais in the north, Hrm in the east,
Isokyr (Storkyro) in the south, Vhkyr (Lillkyro) in the southwest,
Kvevlax in the west and Maxmo in the northwest (Lnnqvist 1972: 53). The
settlement is concentrated to the Vr river valley and consists of seventeen

Context

71

villages: the villages of Rejpelt, Jrala, Andiala, Lomby, Bergby, Koskeby,

Miemois, Mkip, Rki, Llax and Tuckur in the south and east, the
villages of Lotlax, Palvis, Bertby, Kovik, Karvsor and Kaitsor in the north
and west. In addition, two groups of outlying villages are situated in the
west and in the east; the former, comprising the villages Kalap, RukasKomossa and Kaurjrv-Aknus, is Swedish-speaking, whereas the latter, constituted by the villages of Pettersbacka, Murto, Vesiluoma and Ruthsland,
is Finnish-speaking (Lnnqvist 1972: 5354).14 The neighbouring Swedishspeaking parishes of Oravais and Maxmo belonged to the parish of Vr
until 1859 and 1872 respectively, when they were officially separated from
Vr. Contacts have been especially lively between the villages of Kaitsor
in Vr and Karvat in Oravais, and between Palvis and Lotlax in Vr and
Krklax in Maxmo (Lnnqvist 1972: 5354; kerblom 1963: 145146).
The parish has been connected by roads to Vhkyr, Isokyr, Maxmo
and the gulf of Bothnia since the 17th centuryin the north Vr stretches
out to the sea. In the 1850s and 1860s a main road to Ylihrm was built,
but the present road to Vasa, the county capital, was not finished until the
1940s. During the time of collection, the parishioners travelled to Vasa via
Vhkyr, often to sell their produce at the marketmostly butter, meat,
pork and tallow; in 1821 the parish, like many others, lost the right to arrange fairs of its own (Lnnqvist 1972: 54, 57; Talve 1997: 116; kerblom
1963: 119120).
From the 1860s to the 1910s the population amounted to 70008000 persons, most of whom were occupied with farming. In 1876 there were 624
landowners in the parish, and 802 tenants. Four years later 73% were occupied in farming, and the corresponding ratio in 1901 was 66%. The rest
supported themselves through pensions, life annuity, public office, the poor
rota of the parish (a form of poor relief), social benefits (not specified),
trade, hunting and fishing, shipping, industry and handicraft, or communications (not specified). In 1901, tenants, cottars and dependent lodgers still
constituted a large group (Lnnqvist 1972: 62). Some parishioners found
temporary work in Sweden, a situation which became increasingly common
from the 1860s onward. The period was also characterized by accelerating
emigration, first to Sweden in the 1870s and later to North America from
14

The form and spelling of the names of these villages vary considerably in the records
and the research literature, and in quotations I have retained the spelling of the original.

72

Material and Context

the 1880s onwards. Between 1893 and 1910, 17.1 % of the parishioners had
left Finland; from the whole county of Vasa, more than 40 000 people emigrated to non-European countries in 18931901. Emigration reached its
peak in 19011913. The emigrants were mostly of lower social status: dependent lodgers, cottars, peasant boys and girls (Lnnqvist 1972: 5960;
kerblom 1963: 9, 13, 18).
Any large-scale industrialization or mechanization of farming methods
were not discernible during the period studied. The shoe factory, founded
in 1908, and the Hllns sawmill, established in 1890, were the only major
employers in the parish. However, local ventures were fairly numerous.
Several dairies and co-operative dairies were set up in 18861902, and various steam mills and saw-mills from the 1880s onwards. The farming guild
of Vr was instituted in 1902. By the 1890s Vr could boast many shops,
including specialized ones, in the different villages. The landless workers
employed in industry furnished the principal customer base, since they
were paid in cash (Lnnqvist 1972: 6566, 6364; kerblom 1963: 7684,
101, 122134).
Finland got its Education Decree in 1866 which meant that public schools
were established for children, making education freely available to the masses for the first time (Nykvist 1979b: 316; Talve 1997: 317). However, the
parish of Vr was already running 21 permanent village schools in 1862.
The number of pupils amounted to 730, and lessons were given during 22
weeks each year. Some parishioners also attended Anders Svedbergs renowned school in Munsala, and several of the village teachers sat in on his
classes in order to improve their own teaching and educational level. By
1886 the number of schools had increased to three elementary schools with
97 pupils and 22 primary schools with 770 pupils. Twenty-one children attended school in Vasa (Lnnqvist 1972: 64; Nykvist 1979b: 315; kerblom
1963: 266). In 1898 municipalities were enjoined to establish schools if a
minimum of 30 children had been registered for education, and since the
maximum distance to school was set at five kilometres, many new schools
were created. The immediate background for this decree was the intensification of Russification implemented by the Russian authorities at the
time, and a good education was perceived as the best bulwark against such
attempts (Nykvist 1979b: 321). At the end of the period, in 1925, the parish
of Vr had thirteen elementary schools (kerblom 1963: 275294). The
folk high school was founded in 1907 (Lnnqvist 1972: 64; Nykvist 1979b:
Context

73

331; kerblom 1963: 301); as mentioned earlier (chapter 2.2.3), its pupils
donated a collection of folklore to the Swedish Literature Society in 1922.
Apart from the schools, the parishioners desire for learning could be satisfied by the libraries and papers as well, and various non-profit organizations. The first library was housed in the vestry of the parish church following a petition by the dean in 1852. The collection consisted of religious
and other useful and educational writings. Children were frequent users of
the facilities, but their enthusiasm was not entirely appreciated, since they
failed to keep the books whole and clean. The library was later split in two
and moved to the elementary schools in the villages of Koskeby and Kovik.
Donations of books were received in 1882 and 1892, and in 1888 a large purchase of works by Zacharias Topelius was made. From the 1880s onwards,
six libraries were created in other parts of the parish with the financial aid
of Svenska folkskolans vnner (Friends of the Swedish elementary school).
By the turn of the century some 900 tomes were available in the libraries of
the parish.
Newpaper subscriptions became increasingly common in the 1880s: in
1882 the parishioners of Vr had taken out 50 subscriptions, two years later
the number had doubled to 117. The Ostrobothnian student nation in
Helsingfors promoted public reading cottages in the beginning of the 1890s,
and Vr received its own reading cottage in 1891. Being a large parish, the
situation was different in various parts of Vr. In 1892, reading was said to
be quite common a pastime in the southern and middle parts, but in the
north no such intellectual interests could be discerned. In the villages of
Tuckur, Kovik and Karvsor the act of subscribing to a newspaper was
thought to be such an enterprise that it required collective action, and the
inhabitants founded an association, constituted by half of the population,
for the purpose. No papers other than a religious one were read in Lotlax
and Palvis, but in Bertby and Kaitsor not even that was available. Data
from 1893 state that the most popular papers were Weckobladet (80 subscriptions), Wasa tidning (30 subscriptions) and Wasabladet (14 subscriptions).
Twenty other journals and papers were in circulation as well, in one to four
copies per issue. In 1900 Wasa Posten had 301 subscribers, and 444 in 1910,
when Wasabladet was ordered in 58 copies (Lnnqvist 1972: 64; Dahlbacka
1987: 137138, n. 22; kerblom 1963: 308310).
The promotion of decency and enlightenment among the youth was a
prime motive in the establishment of an association for young men in 1881.
74

Material and Context

The move was an important and apparently beneficial one to judge by the
complaints filed against the male youth at the parish and municipal meetings in the previous decades. The regulations of the parish of 1797, 1828 and
1841 were made more stringent in 1846, at the request of the dean, and the
next year the young men were prohibited to ride to and fro outside the
church during service on Michaelmas, as they thereby failed to attend mass
and had trampled children and old people. In 1867 the farm masters in the
northern parts of the parish started patrolling the countryside on weekends
and holidays due to the undisciplined behaviour of the youth, and in 1890
the county sheriff wanted to ban weddings because of the exaggerated consumption of liquor at the feasts and all the trouble resulting from it. Other
problems at weddings were the wild shooting, reckless riding and gatecrashing practised on those occasions. The creation of the association for
young men was thus a welcome event, and the associations founder, Johan
Hagman, a teacher in the village of Koskeby, summoned the members
once a fortnight to hear lectures given by himself or other speakers. Singing, declamation, debates and games were also included in the programme
of the association. It was succeeded by the youth association of Vr, open
to both sexes, which was founded in Koskeby in 1891. This association
strove to educate the youth by giving courses in various subjects, by maintaining a library and by arranging sport activities. During the same decade,
in 1896, a youth association in Rejpelt was added, occupying itself with
publishing a handwritten journal, training a choir and organizing other
events. The youth associations of Rki and Keskisnejden were established in 1908; the former launched debates once a week and was much
focused on temperance. The association of Lotlax was instituted the year
after. The programme was varied; in 1913 the branch received two libraries,
and the facilities were eagerly utilized by the members. The next year the
association arranged a sports competition. The association of Kovik became a combined youth association and temperance society in 1910. The
local temperance society launched its civilizing enterprise in 1906, and the
housewives got the opportunity to improve their domestic skills when they
joined the Martha organization in 1903. The Vr branch was known as
the most active in terms of participation in courses arranged by the organization (Lnnqvist 1972: 6465; kerblom 1963: 107, 317327).
Thus, the parish of Vr was socially homogeneous and mostly agrarian,
with quite many social activities on offer.
Context

75

2.3.2 Religious context

One of my principal arguments in the following chapters is that folklore
and religious traditions are not as separate as we are sometimes led to believe, and I attempt to demonstrate that narratives of trolls and Biblical
texts are part of the same network of associations. In order to substantiate
this claim, however, it is imperative to elucidate the ways in which the
Bible and other religious writings reached the rural population. The Bible
itself was an expensive article until the beginning of the 19th century, when
the Evangelical Association in Sweden, financially supported by the British
and Foreign Bible Society, started publishing cheap editions of the Bible
(Pleijel 1967a: 3739). These Bibles also found their way to the Swedishspeaking areas in Finland, and other editions were utilized as well. The
version of the Bible in use during much of the period under study was the
Bible of Charles XII, sanctioned in 1703, which is a very slightly altered
variant of the Reformation Bible of 1541 commissioned by Gustavus Vasa
Rex; this explains the archaic language of the translation, which was archaic
even in the 16th century (Olsson 2001: 5859, 40). Toward the end of the
period, the translation of 1917 came into circulation, but it is uncertain
whether it had any real impact on the adult population: people tend to be
reluctant to abandon their old, beloved translation.15
The Bible of 1703 in its 19th-century form has been called Swedens first
Bible for individual reading; previous versions were mainly intended for use
in service. A number of factors contributed to this development, in Sweden
as well as in Finland: new forms of production were introduced, distribution was intensified, the economy of the masses changed, literacy increased,
and religious revivalism encouraged reading of the Bible (Olsson 2001: 62
65).
Apart from individual reading, the message of the Bible was disseminated
in other ways: the Gospels and Epistles were read in church during service,
and as long as regular attendance was still a custom, these were commonly
15

It should be noted that the Swedish-speakers in Finland have always employed the
Swedish translations of the Bible; while Finland belonged to Sweden this procedure was a
matter of course, but even after the incorporation into the Russian realm and the Declaration of Independence, it remained the case. Most of these Bibles were also printed in
Sweden, although a limited number of printings in Swedish were produced in Finland
(Huldn 1991: 344345). Other translations than those authorized by the Church might
have been used as well (see Lindstrm 1991:208210 on unauthorized translations).

76

Material and Context

known (Pleijel 1967a: 44). The rewritten material found in the hymnal and
in prayers has been decisive in this regard. The old Swedish hymnal of 1695
was officially employed until 1886, when the second general Finnish synod
accepted a new one. However, the latter never attained the same status as
the former, since it neglected the needs of the revivalist movements, and it
was superseded by the songbooks of these movements in practical use
(Nsman 1979: 119). Some of these songbooks, such as Sions snger (The
Songs of Zion) and Sionsharpan (The Harp of Zion), were well-known in
Vr (Dahlbacka 1987: 287). The importance of Martin Luthers catechism,
reinforced by the parish catechetical meetings, should not be underestimated either (Pleijel 1967a: 13; Pleijel 1967b: 90; Olsson 1967: 113). Though
many may have loathed the catechism and the effort involved in memorizing it, it was nevertheless ingrained in peoples minds, as were the expositions of it. The exposition of Archbishop Svebilius, originally published in
1689, was particularly popular; the chapter of bo adopted it as a coursebook for the youth in 1759, and it retained its dominant position well into
the 20th century (Pleijel 1967b: 94; Nykvist 1979b: 309). We know that
Svebilius was cited by the folk, as attested by Janne Thurman in his article
on pagan memories in the poetry of the inhabitants of Pargas (Thurman
1891: 110). In the parish of Vr Svebilius exposition was utilized in confirmation classes and in the instruction of children (Wolf-Knuts 1991: 42).
Catechetical meetings were instituted by Bishop Johannes Gezelius,
senior in 1673. It was a matter of personal honour to perform well at the
examinations, and the proficiency of the participants was meticulously
checked. The meeting was usually held on a large farm, and everyone was
expected to be able to read from a book and to recite the catechism by
heart. The grades were entered in the parish register and in special reading
slips. Afterwards the clergy expounded a passage from the catechism, and
inquired into the moral state of the village (Nsman 1979: 45). Catechetical
meetings were still practised at the time of collection (Wolf-Knuts 1991: 42),
and confirmation classes and the examinations carried out in connection
with them, first decreed in 1763, were likewise common (Nsman 1979: 46,
7576). As a consequence of the prohibition against private religious meetings issued in 1726, catechetical meetings and visits to the sick were basically the only forms of direct contact the parishioners had with the parish
priests. The repeal of this law in the new Church Law of 1870 resulted in
an intense activity on the part of both laymen and professionals. The clergy
Context

77

started giving Bible classes in the villages, and laymen organized edifying
meetings (kerblom 1963: 158).
The knowledge of religious matters exhibited by the parishioners, both
young and old, was praised by the ecclesiastical inspector in 1804. His protocol states that the catechetical meetings arranged in the parish had attracted many visitors, children and youths in particular. The reading skills of
the parishioners were likewise deemed good and admirable in general, and
the dean expressed his satisfaction with the fact that the youth had begun
studying Svebilius exposition of the catechism. By the 1850s the population
increase turned the holding of catechetical meetings into a burden for the
clergy, and in 1877 the dean found the situation untenable. Literacy had
declined drastically, and he proposed an introduction of elementary schools,
but the parishioners rejected the idea.
Confirmation classes were organized annually in the autumn and in the
spring, and only those who had attended it during both terms were admitted to the Communion, the protocol of 1804 reports. In 1856 the terms
were specified as two weeks in the autumn and two weeks in the spring.
The vicar and the curate alternated with the assistant vicar so that both
boys and girls received instruction from them in turn. In addition, the
parish clerk taught them hymns one hour each day. The same year the
dean put forward the idea of creating Sunday schools, and the parishioners
accepted his suggestion. In the summer months the children tended to
forget what they had learned during the winter, and the autumn was mostly spent recovering the knowledge lost. The dean also thought the youth
idled the Sundays away by practising indecency, and considered Sunday
school a more edifying pastime. He exhorted the teachers to base their instruction in reading on correct spelling and to ensure that no additions,
omissions or transpositions were made when reading by rote, and that the
children understood what they had read. The best way of examining their
apprehension of the meaning of a text was to ask them to render it in their
own words (kerblom 1963: 262266). To what extent his injunction was
heeded, and if it was, how well it worked is difficult to tell.
Notwithstanding, the authority of the church dwindled in the 19th century, and the church was no longer capable of supervising the celebration of
Communion and participation in catechetical meetings (Raittila 1969: 106).
Simultaneously the clergy lost much of its worldly power in the separation
of ecclesiastical and municipal adminis