Metaliterary Layers in Finnish Literature

Front matter

Studia Fennica Litteraria

THE FINNISH LITERATURE SOCIETY (SKS) was founded in 1831 and has, from the very beginning, engaged in publishing operations. It nowadays publishes literature in the fields of ethnology and folkloristics, linguistics, literary research and cultural history.

The first volume of the Studia Fennica series appeared in 1933. Since 1992, the series has been divided into three thematic subseries: Ethnologica, Folkloristica and Linguistica. Two additional subseries were formed in 2002, Historica and Litteraria. The subseries Anthropologica was formed in 2007.

Part Ii: Case Studies

THE DEAD END OF WRITING. AESTHETIC AND

PSYCHOLOGICAL REFLEXIVITY IN TIMO K. MUKKA'S

BEYOND THE EPIC AND THE NOVEL. LITERARY TRADITION

AND ITS IMPOSSIBILITIES IN FINLANDIA BY HANNU SALAMA... 89 Samuli Hägg

THE METAFICTIONAL GAMES IN THE FICTION

Mika Hallila THE NOVEL IS A CULTIVATED MONSTER. METAFICTIONALITY IN JUHA K. TAPIO'S NOVEL FRANKENSTEININ MUISTIKIRJA... 125

PART III: SPECIAL CASES Outi Oja CONTEMPORARY FINNISH PROSE POEMS DISCUSS THEIR

Kaisu Rättyä “I GOT TO TELL YOU A STORY"- METAFICTIONAL FEATURES IN FINNISH YOUNG ADULT FICTION IN THE 1980S AND 1990S 153

Merja Sagulin FROM INDIVIDUALISM TO PARTNERSHIP. METANARRATIVES

Kristina Malmio DOUBLE PLAY. IDENTITY, STATUS AND REFINEMENT IN FINNISH SELF-REFLEXIVE POPULAR LITERATURE

IN THE 1910S AND:T9208: «10030 saa RUE eed be aaendees 182 AFTERWORD

*THE TIMES THEY ARE A-CHANGIN”. AFTERWORD

Foreword

During the past few decades, metaliterary phenomena have attracted increased attention in Finnish departments of literature. The first articles on the metaliterary phenomena in the context of metafiction studies were published in the 1980s, and the first book-length investigations of Finnish metaliterary phenomena came out in the early 1990s. However, metaliterary layers and dimensions were not central issues in the study of Finnish literature in the 1980s and 1990s. It was not until after the turn of the new millenium that metaliterary layers and dimensions have gained a more significant position in the study of Finnish literature. Indeed, the first decade of the 21* century has seen the publication of approximately ten book-length Finnish studies on the topic.

Usually such studies have dealt with post-war Finnish literature, that is, Finnish literature from the late 1940s to the present. This is also the focus of this study. Yet we also recognize that the metaliterary point of view is fruitful also when studying older Finnish literature, in particular 19" century Finnish literature. Thus, it would be interesting to examine how later Finnish literature has commented on the early canon of Finnish literature, that is, on works of J. L. Runeberg, Elias Lónnrot, Zacharias Topelius and Aleksis Kivi. While these issues are beyond the focus of this collection, we hope that our anthology encourages researchers to study such topics in more detail.

We should like to thank the writers for their participation in this collective endeavour. Thanks are also due to PhD Esa Penttilá for checking the English language of the articles. Finally, we would like to express our gratitude to the Finnish Literary Society for including our book in its international publication series.

Joensuu, 11 August 2008

Samuli Hágg Erkki Sevänen Risto Turunen

On the Study of Metafiction and Metaliterary Phenomena

The Emergence of Metaliterary Concepts

Within the Western academic world, the study of metaliterary phenomena became a significant trend in the 1980s, albeit the concept of metaliterature and its sub-concepts were launched somewhat earlier, in the 1960s and early 1970s. It might be thought that metaliterary study is chiefly concerned with metafiction; however, the first studies that utilized the concept of metaliterature and its sub-concepts did not deal with fiction or prose but with drama and poetry. Linda Hutcheon (1985, 4) remarks that, in the United States, the study of metafiction was initiated by Robert Scholes (1967; 1970) at the turn of the 1960s and 1970s. One could add to this that at least Lionel Aber and Heinz Schlaffer had explored metaliterature before this; in the early 1960s, the former (see, Aber 1963) had published a book on metadrama, and three years later the latter (see, Schlaffer 1966) had published an article on metapoetry. All the same, in the 1970s and 1980s it was the study of metafiction that seemed to gain pride of place in the academic interest in metaliterature. That interest was active in the United States and France, in particular; in the United States, for example, Robert Alter (1975) aroused an influential discussion on the critical potentialities of metafiction, and in France Jean Ricardou (1973) and Lucien Dällenbach (1977) published their investigations on French nouveau roman and its metafictional devices. The 1980s was a turning point in this development; during that decade the concept of metafiction and its sub-concepts found their way to the departments of literary studies in different countries thus changing in this way the study of metafiction into a truly international phenomenon that partly exceeded the boundaries of the Western world (Hallila 2006, 113-114). When compared with this development, investigations on metadrama and metapoetry have chiefly functioned as side roads in the study of metaliterature.

Also in Finland the concept of metaliterature and its sub-concepts were widely applied in the 1980s. In the early 1980s, Eino Maironiemi (1982, 24—26) discussed the metafictional traits in Hannu Salama's novels, Jaana Anttila (1983) studied Italo Calvino's novel Se una notte d'invorno un viaggiatore (If on a Winter' s Night a Traveler, 1979) “as a book about books" and Pekka Tammi (1983) reflected on the phenomenon of self-conscious

Erkki Sevánen fiction at a general level. Yet, those articles, with the exception of Tammi, had only a slight connection to the theoretical investigations of metaliterature. The time for a more profound theoretical understanding of metafictionality came a few years later; in this respect, it was important that the Finnish departments of literary studies used Linda Hutcheon's and Patricia Waugh's systematic theoretical investigations on metafiction as course-books and reference material. In particular, Narcissistic Narrative: The Metafictional Paradox (1980/1985) and The Poetics of Postmodernism: History, Theory, Fiction (1988) by Hutcheon, a Canadian theorist, as well as Metafiction: The Theory and Practice of Self-Conscious Fiction (1984) by Waugh, a British researcher, were relevant in this connection. Several subsequent foreign, that is, non-Finnish, studies of metafiction have also used these books as their point of departure or as their sources (Hallila 2006, 113— 115). This indicates that the Finnish study of metafiction has maintained a close relationship with the comparable international study; first and foremost, it has been based on Anglo-American theoretical models, although it has, to a smaller extent, received ideas from France and Germany as well.

In the context of this book, it is relevant to take into account mainly those Finnish researches whose object of study is Finnish or “domestic” literature — and not "foreign" literature. In the study of Finnish metaliterary texts, this community of researches has not concentrated on elaborating theoretical ideas but on analyzing concrete texts and on applying generally accepted theoretical views in their analyses. So far, Mika Hallila's doctoral thesis Metafiktion käsite (The Concept of Metafiction, 2006) — that utilizes both non-Finnish and Finnish metafictions as its material — is the only book-length theoretical investigation on metaliterature in Finnish. In recent years, some Finnish researchers (Malmio 2005a; Oja 2004 and 2005; Peltonen 2005) have also published theoretically accentuated articles on metaliterature, which enables one to conclude that at present metaliterary phenomena seem to attract increasing attention in the Finnish departments of literary studies.

So far Finnish researchers have usually considered metaliterary texts from a formal-structural perspective, without placing them systematically into wider cultural and societal contexts in the same way as foreign studies of metafictionality used to do until the 1990s. In this respect, the clearest exception is Kristina Malmio's doctoral thesis Ett skrattretande (for fall (2005b), whose ambiguous title translates into English as *A Laughable Decay". In her study, Malmio discusses the metaliterary traits of the Finnish popular literature of the 1910s and 1920s; the material of her study consists of two detective novels, a love story, a humoristic play, and a collection of causeries. When investigating her material, Malmio is not only utilizing the theories of metaliterature, but she is also interpreting and explaining her material by means of cultural and sociological concepts and theories — thus showing that the study of metaliterary phenomena can obtain a profounder view of its object by systematically taking into account the cultural and societal contexts of metaliterature.

Definitions of the Concepts of "Metafiction" and “Metaliterature” What then are these phenomena called *metafiction", *metafictionality" and "metaliterature"? As far as metafiction is concerned, standard definitions tend to equate it with “narcissistic”, “self-conscious” or “self-referential” fiction. For example, Hutcheon begins her book Narcissistic Narrative by stating that

‘metafiction’, as it has now been named, is fiction about fiction — that is, fiction that includes within itself a commentary on its own narrative and/ or linguistic identity. (Hutcheon 1985, 1.)

Defined in this way, narrative self-consciousness or self-reflexivity would be the hallmark of metafiction; that is, metafiction presents a story, on one hand, and comments on the presentation of that story, on the other. Similarly, Waugh grants a central position to the idea of narrative self-consciousness in her own definition of metafiction:

Metafiction is a term given to fictional writing which self-consciously and systematically draws attention to its status as an artefact in order to pose questions about the relationships between fiction and reality. (Waugh 1984, 2.)

Actually, this definition combines two ideas. According to it, metafiction refers to itself and makes itself visible as a linguistic and narrative entity, and in this way it reflects upon the nature of fiction and reality. In other connections, Waugh, however, tends to think that the latter idea does not self-evidently characterize metafiction: “The lowest common denominator of metafiction is simultaneously to create a fiction and to make a statement about that fiction” (Waugh 1984, 6). Thus, metafictions do not always deal with questions that concern the relationship between fiction and reality, but Waugh emphasizes the fact that they necessarily refer to themselves and speak about themselves.

Hutcheon’s and Waugh’s definitions are applicable only to certain metafictions or to certain aspects of metafictionality. They cannot do full justice to the multiplicity of the phenomenon of metafictionality — despite the fact that certain literary dictionaries have adopted fairly similar definitions. For example, Routledge Encyclopedia of Narrative Theory (2005), edited by David Herman, Manfred Jahn and Marie-Laure Ryan, begins its definition of metafiction in a way that is completely in accordance with Hutcheon’s and Waugh’s definitions:

Metafiction is a term first introduced by narrative theorist and historian Robert Scholes to indicate the capacity of fiction to reflect on its own framing and assumptions. (Routledge Encyclopedia of Narrative Theory 2005, 301.)

Also here metafiction and metafictionality are, primarily, comprehended in terms of narrative self-reflexivity or self-consciousness. Although in their later books Hutcheon (1985, 52-54) and Waugh (1984, 4, 62) slightly widen their view of metafictionality, they do not make a clear-cut analytical difference between metafictionality and narrative self-consciousness.

The distinction between object language and metalanguage offers a point of departure for a wider understanding of metafictionality. This well-known distinction comes from philosophy, mathematics, logic and linguistics. In the 1920s and 1930s, David Hilbert, a German mathematician and philosopher, and Alfred Tarski, a Polish logician and philosopher, introduced this division, and some years later Louis Hjelmslev, a Danish linguist, elaborated it for the study of natural languages. The idea of metalanguage seemed to be part of the Zeitgeist of the day, since, besides the three pioneers, also Rudolf Carnap in Germany and Bertrand Russell in Great Britain worked on it in the 1930s (Historisches Wörterbuch der Philosophie 1980, 1301—1302). The concept of metalanguage is also mentioned in Waugh's (1984, 4) book, but it does not have a constitutive meaning in her thinking about metafictionality.

According to the distinction made by Hilbert and Tarski, object language can be characterized as a first-order language that speaks — in the case of mathematics and logic — about mathematical and logical entities or objects, in other words, about numbers and correct inferences; metalanguage, in turn, is a second-order language that speaks about the first-order languages of mathematics and logic. The difference between these two languages is not sharp, and they can share some parts in common. In the case of natural languages, this is more obvious than in mathematics, logic and other formal languages. In the 20" century, linguists elaborated formal metalanguages, by means of which they described the structure and properties of natural languages; these formal metalanguages were not entirely independent of natural languages, but they had only certain parts in common with them. However, Roman Jakobson (1960) has pointed out that the daily use of natural languages includes a clear-cut metalinguistic dimension as well. According to his list of the functions of natural languages, the metalinguistic function is one of the six basic functions of natural languages — besides referential, expressive, conative, poetic and phatic functions. When the speakers of a natural language utilize the metalinguistic function of their own language, they, for example, speak about the meanings and correctness of the speech acts produced by themselves.

Likewise, when a narrative or a fiction speaks about real or fictional states of affairs and events, it is operating as a first-order narrative or fiction. Subseguently, when a narrative or a fiction refers to itself and speaks about its own status as a narrative or fiction, it is operating as a metanarrative or metafiction (cf. Prince 1982, 115—128). Yet, this situation represents only one type or dimension of metafictionality; we can call it self-reflexive or self-conscious metafictionality. Mark Currie (1995, 1-5) points out that, in addition to this, metafiction may also speak about other concrete fictions and literary works or about fictions and literature in general (see also Oja 2004, 12—13). In this way, we have two further types or dimensions of metafictionality: intertextual metafiction refers to other fictions and literary works and comments on them, whereas general metafiction reflects upon guestions that concern the nature of fictional and literary work at a general level.

It should be noticed that the differences between these three types of metafictionality are analytical, that is, in literary practises they do not necessarily occur as separate. Concrete metafictions often contain elements of all of these three types, even if a certain type or dimension is dominating in them. This being the case, perhaps Italo Calvino's novel Se una notte d'inverno un viaggiatore (If on a Winter's Night a Traveler, 1979) in the first instance represents self-conscious metafictionality, for it constantly refers to itself and comments on its own narrative and communicative structure. Anna Makkonen (1991) has shown that in Finnish literature Marko Tapio's novel Aapo Heiskasen viikatetanssi (Aapo Heiskanen's Scythe Dance, 1956) contains, among other things, similar features, although they are not as visible or explicit as in Calvino's novel. As for intertextual metafictionality, Umberto Eco's Il nome della rosa (The Name of the Rose, 1980) and Pirkko Saisio's (alias Jukka Larsson's) Viettelijà (Seducer, 1987) can be seen as instances of it; the former refers to and transforms Arthur Conan Doyle's detective stories and Jorge Luis Borges's short stories, whereas the latter has the biblical story about the last days of Jesus as its subtext. For Currie (1995, 3), David Lodge's satirical novel Small World (1984) has a clear-cut general dimension, because, as he says, it critically describes a literary community and at the same time implicitly reflects upon its own status as a fiction. In contemporary Finnish literature, Kari Hotakainen's Klassikko (A Classic, 1997) has perhaps a rather similar character; on one hand, it offers a satirical and comical representation of current commercialized literary institution, and, on the other, it parodies popular genres such as confession and diary literature and autobiography.

Sometimes Hutcheon (1985, 52-54, 74) and Waugh (1984, 13, 70-71) seem to use the concept of self-conscious fiction in a broad manner or as an umbrella concept. In these connections, this concept does not only contain texts that refer to themselves and reflect upon their own status as literature; in addition, it includes texts that comment on other texts, literary conventions and different conceptions of literature. In this use, the above-mentioned three types or dimensions of metafictionality — self-conscious, intertextual and general metafictionality — are all instances of the self-understanding or self-reflexivity of fiction. It cannot be denied that even today certain researchers favour Hutcheon's and Waugh's way of using the concept of literary self-consciousness; this can be seen, for example, in the Finnish study of metafictionality and in some articles of this book. However, in a more detailed use of the concepts, it is useful to speak about the above-mentioned three types or dimensions of metafictionality.

By means of the three-part distinction at issue, it is possible to show that certain metafictional novels are hardly self-conscious at all. This holds, for example, for Väinö Linna's trilogy Täällä Pohjantähden alla (Under the North Star, 1959-1962), which deals with the history of Finnish society from the 1880s to the 1950s. This realist novel by Linna is strongly mimetic; it is even, in part, based on Linna’s own research work in historical archives. What is important here is the finding that it describes society in a way that does not bring out its own status as a linguistic and narrative entity — and nor does its narrator break its narrative frames, that is, the narrator does not show that he is constructing a story. On the contrary, the novel gives the impression that it is the historical reality itself that manifests itself in the characters and events described by the novel. Yet, at the same time Linna's novel contains a clear metafictional dimension, for it constantly presents critical comments on the 19" century Finnish literature, whose picture of Finnish society it characterizes as "distorted" or elitist (cf. Nummi 1993). In this way, Linna's realistic novel possesses a metafictional dimension without narrative self-consciousness.

Linna's novel might be an exception, for concrete metafictions usually contain elements of different types or dimensions of metafictionality. Due to this feature, they also, more or less and in their own way, practise literary criticism and theorize on literature. Formerly it was thought that it is the task of book reviews, literary criticism and literary theory to function as a metadiscourse in relation to literature, but the study of metafiction has taught us that literary works themselves can partly carry out this function as well. Currie (1998, 51—70) even wishes to use in this connection the term "theoretical fiction”, which, he continues, suits to characterize these features in the novels called metafictions. As such, the concept of theoretical fiction is appropriate here; yet, when using it we should not equate metafiction with theoretical fiction, since fictions can be theoretical in different ways. In contemporary literature, Milan Kundera, for instance, is a highly theoretical author, whose novels are rich with metafictional features. Yet, in his novels Kundera does not theorize only on literature but also on philosophical themes such as death, immortality, identity, sexuality, irrationality, the meaning of historical events, and European culture; because he utilizes narrative form as well as essayistic reasoning when dealing with these themes, his novels could also be called “artistic essay novels" (cf. Saariluoma 1998). Thus, both metafictionality and essayistic reasoning may characterize theoretical fictions, which remain hidden in Currie's suggestion.

The concept of metaliterature obviously includes similar ideas and distinctions as the concept of metafiction does. If this presupposition is accurate — and so far nothing seems to undermine it — one can say that metaliterary works are, in the first instance, self-conscious, intertextual or general by nature (cf. Oja 2004, 13). By using the word "reflexion" we can also say that metaliterary reflexion contains these three analytical types or dimensions. When a literary work reflects upon literature, it can point to and comment on itself, or activities like these can orient themselves to other concrete literary works or to literary conventions and traditions and different conceptions of literature.

Also in drama and poetry, metaliterary devices have made literary works more theoretical and more conscious of literary traditions. The theoretical dimension of metaliterature is accentuated clear-cutly in the German terminology concerning metapoetry. Outi Oja (2004, 7—8) points out that German researchers have often used the term “poetological poetry” (poetologische Lyrik) as a synonym for the term "*metapoetry" (Metalyrik), which indicates that they regard metapoetry and the theoretical study of poetry as kindred phenomena.

More about the Features and Devices of Metafictionality

In the 1970s and 1980s, theorists and researchers of metafictionality usually shared the idea that metafictionality has to be considered as a textual phenomenon in literature. Therefore, they continued, it can be studied empirically by means of narratological and linguistic methods, which are capable of reaching it more or less exhaustively. A thought like this was included, among other things, in Hutcheon's and Waugh's investigations as well as in Gerald Prince's (1982) narratology. Of these three theorists, it is perhaps Hutcheon who has inspired the study of metafiction most widely.

In her books about metafiction, Hutcheon mainly speaks about self-conscious metafiction, in relation to which she elaborates two fundamental distinctions. Some literary texts are, she writes, self-conscious at the level of their linguistic constitution or at the level of their use of language, while other literary texts prove to be diegetically self-conscious; the latter ones are metafictional at the level of their story. On the other hand, some literary texts display their metafictional features overtly, while in others metafictionality remains covert or hidden. In the latter case, reseachers can, with the help of information provided by the texts at issue and by means of additional information, reveal the metafictional nature of those texts. By utilizing these distinctions, Hutcheon elaborates four types of metafiction: diegetically overt metafiction, diegetically covert metafiction, linguistically overt metafiction and linguistically covert metafiction (Hutcheon 1985, 7).

Hutcheon does not comment on this typology in detail, but obviously it is reasonable to think that in practice metafictional texts may contain elements of all these four types. Hence, the typology in question should be regarded as analytical, albeit Hutcheon herself avoids a characterization like this. At a more concrete level, she concentrates on considering which devices are typical of metafictions, and she even presents a list or diagram of these devices. In this connection, it is not possible to present and analyze the entirety of that list; instead, we may bring up two devices mentioned by Hutcheon, namely parody and mise en abyme. By explicating them, one can gain a more concrete view of how fictions change into metafictions.

Parody is, for Hutcheon, not only a characteristic device of metafictions but also of novel as a literary genre, for it is since the days of Don Quixote (1605-1615) that Western novels have frequently utilized it. And just as Cervantes's novel mocked romances of chivalry and their conventions, subsequent metafictional novels make fun of other sub-genres of novel or deal with their conventions playfully, that is, in an ironic-parodic style. In recent literary culture, detective stories, fantastic stories and realistic novels, for example, belong to such parodied sub-genres. Subsequently, Umberto Eco, Vladimir Nabokov, Alain Robbe-Grillet and Dorothy L. Sayers have an ironic-parodic relation to detective stories, whereas Jorge Luis Borges and Italo Calvino transform fantastic stories for metaliterary purposes and John Barth, John Fowles and the representatives of the French nouveau roman, among others, appraise the conventions of literary realism critically.

Hutcheon (1985, 52, 73—74, 154) tends to think that in parodic novels like these metafictional devices operate, in the first instance, at the diegetical

Erkki Sevánen level, and they are overt or covert by nature. To this remark one has to add, following Hallila (2005, 100), that parody can operate on the linguistic level as well. A good example of this possibility is Väinö Linna's realistic novel Tuntematon sotilas (The Unknown Soldier, 1954; in English 1957), which deals with the war between Finland and the Soviet Union in 1941-1944. In Linna's novel, the common soldiers, who are the actual protagonists of the story, often mockingly cite phrases, tropes and sentences which originate from the Finnish patriotic-nationalistic literature of the 19" and early 20" centuries, that is, from the sublime poetry and epic of J. L. Runeberg and from warlike and heroic songs, poems and stories. During the war, the official Finnish propaganda leant massively on literature like this, but already in the final stages of the war, and especially soon after it, texts like these proved largely obsolescent. In Linna's novel, the officers also cite this literature, but, unlike the soldiers, they do it seriously and without an ironic-parodic element; in these cases, the narrator of the novel usually adopts an ironic position on the officers indicating that he views the events of the war from a perspective which is close to that of the common soldiers. Thus, by elaborating ironic-parodic devices such as these Linna's novel outlines a critical view of the previous Finnish patriotic-nationalistic literature, and for the same reason it obtains an easily recognizable intertextual and metaliterary dimension (cf. Nummi 1993).

When dealing with mise en abyme phenomena, Hutcheon (1985, 53—56) often regards them as devices that represent diegetically overt metafiction. From this standpoint, mise en abyme can be regarded as a textual structure or fragment that, in a miniature size, repeats the main theme or thesis or event of the whole text to which it belongs as a part or component. However, as Makkonen (1991, 20) emphasizes, in addition to these possibilities a mise en abyme structure can also contradict or question the main thesis of the text; then it, in a way, relativizes the truth of the thesis. On the other hand, mise en abyme fragments and allegories are kindred phenomena, for both of them represent things by means of similes; due to this state of affairs, Hutcheon (1985, 55—56) regards allegories as long mise en abyme fragments.

The two distinctions — overt/covert and linguistic/diegetical — have been utilized frequently in subsequent studies concerning metafiction and metaliterature. However, sometimes these studies have replaced the distinction overt/covert with the distinction explicit/implicit, which has the same meaning and which, instead of the overt/covert distinction, has long been a part of the vocabulary of literary studies (see, for example, Oja 2004, 17; Reinfeldt 1997, 247). Subsequent researchers have also completed the distinctions made by Hutcheon. Traditional narrative theory taught us that concrete fictional texts can be considered as narratives that contain the dimension of story or diegesis and the dimension of discourse; in the fictional world of a narrative, for instance, the events, states of affairs as well as the characters' acts and dialogue belong to the dimension of story, whereas the presentation of the story and the narrators' speech are situated on the dimension of discourse. A division like this forms a background for the distinction linguistic/diegetical in Hutcheon's theory, for obviously the diegetical mode is situated on the dimension of story, whereas the linguistic mode seems to belong both to the dimension of story and to the dimension of discourse. In this way, Hutcheon's distinction ignores the concrete literary text itself and its metafictional potentialities. In her own view of metafictionality, Liisa Saariluoma (1992, 24) expressly takes into account concrete literary texts when she writes that metafictionality also contains all those devices that draw readers' attention to fiction's artificial nature, that is, to its status as a linguistic and narrative entity; even the typographical devices used in literary texts can, then, possess a metafictional function.

Even if the distinctions made by Hutcheon have proved to be fruitful and useful, they are not in every respect unproblematic. In particular, the concept of covert or implicit metafiction is vague. This concept seems to suggest that a text does not function as a metafiction unless readers draw, on the basis of the text in question and by leaning on their own conception of literature and on their own understanding of language, conclusions in which they connect to that text a metafictional dimension. Thus, in a case such as this metafictionality would not be a purely empirical or linguistic feature of a text. Similarly, one could ask in what sense mise en abyme structures really comment on the texts whose components they are. Mise en abyme structures are undoubtedly empirical features of those texts, but shouldn't we say that usually it is expressly readers who notice analogies or incongruencies between a mise en abyme structure and its textual surroundings? Is it actually readers who make a mise en abyme structure comment on its textual surroundings? Outi Oja (2005, 106—107) seems to have something like this in her mind, when she says that not all of the mise en abyme structures are necessarily metafictional or metaliterary — that is, as purely textual phenomena some of the mise an abyme structures do not themselves make statements on their textual surroundings. Conversely, an intertextual metafiction contains references to other texts and comments on them, but in literary communication it does not function as a metafiction, unless readers recognize these references and comments (Plumpe & Werber 1993, 25).

These examples and questions tell us that in the study of metafictionality it is not reasonable to ignore the levels of reading and interpreting entirely. Currie explicitly takes these levels into account in his own conception of metafictionality. He even thinks that “a literary text and its reading are inseparable", wherefore metaliterary reflexion *is as much a function of reading as an inherent property of a text" (Currie 1995, 10). Thus, in Currie's conception, metaliterary reflexion seems to be one possible dimension in literary communication. A conception like this manifests a more general tendency in contemporary narrative theory and research. Due to its formalist and structuralist roots, classical narrative theory and research usually passed over the questions of reading and interpreting, whereas contemporary theorists and researchers are more apt to consider narratives in relation to readers (see, for example, Fludernik 1996). However, despite this shift, metafictionality is not solely dependent on readers' interpretations; at least some of its manifestations are explicitly and empirically visible in the surface of literary texts. Pentti Haanpää's novels, for example, contain a number of rather direct quotations from the Bible, whose textual fragments they often put into an ironic connection thus relativizing and questioning the truthvalue of the Bible's lessons (for more details, see Koivisto 1998). This means that metafictionality can, at least in part, be studied adeguately without taking into account the level of reading and interpreting.