Texts as Cultural Artefacts: Theoretical Challenges to Empirical Research on Utterances and Texts

Abstract

This chapter presents a theory of texts often used in Scandinavian “sakprosa” research. In the chapter, the constitutive features of any text are presented. It is shown how the text as a cultural artefact has the quality of and at the same time being a unique utterance and instantiating a text norm. In the text theory presented, text is part of a text culture, to which other text sharing similar norms belong. Thus, a “text” is conceived as an utterance that competent participants in a specific cultural time and space assign a certain cultural value.

Introduction: Textual Studies in Sakprosa Research

In the last 50 years, extensive theoretical discussions on how to understand what a ‘text’ is have taken place amongst Nordic text researchers and discourse analysts. These discussions took off when text linguistics was established as an active Nordic academic tradition in the 1970s (e.g., Enkvist 1974; Fossestøl 1980; cf. Berge 1993 for critical remarks). Towards the turn of the century, systemic functional linguistics (SFL) became a new paradigm in the Nordic countries for researching (also multimodal) texts (e.g., Halliday and Hasan 1989), and this tradition has also been debated and challenged (e.g., Berge 2012a; Ledin and Machin 2019a, b). In relation to and dialogue with these and other research traditions, we have developed a theory of texts as cultural artefacts, which we will elaborate on in this chapter.

There has been and still is comprehensive empirical Nordic studies on texts mediated through several semiotic resources and in different cultural and situational contexts (e.g., Melander and Olsson 2001; Berge 2007). Such studies encompass, for example, how texts are to be understood in societal and institutional contexts and how texts influence historical events and lead to cultural changes. In other words, this text research has been both synchronically and diachronically oriented. Furthermore, it tends to focus on the widely distributed, used, and read texts, that have had profound and lasting effects on the development and structuring of the Nordic societies’ cultural mentalities, political traditions, and institutions.

The notion used to label this research is sakprosa (Englund et al. 2003; Tønnesson 2012; cf. Blikstad-Balas and Tønnesson 2020, where the notion is revisited). The notion is similar to what in colloquial English is called ‘non-fictional prose’, in colloquial German Sachprosa, in some text theories ‘subject oriented texts’ (Schröder 1991), and in other traditions ‘faction’. A common denominator of this Scandinavian word and concept is that it does not include or cover texts that may be characterized as ‘belles lettres’, that is, texts belonging to aesthetics or fiction, such as novels, poetry, and theatre plays. This literature is conventionally named skjønnlitteratur (similar to ‘belles lettres’) in the Scandinavian languages and is often associated with an aura of quality, taste, and advanced intellectual and aesthetic aspirations. Skjønnlitteratur has often been given a privileged position in mother-tongue education.

The Nobel Prize in literature given yearly by the Swedish Academy with few exceptions favours authors and texts belonging to this specific literary practice. In addition, twentieth-century literature research in the Nordic countries, including text-historical research, has focussed on conventional ‘belles lettres’. The Nordic text research on non-fictional texts (sakprosa), therefore, represents what may be called a new paradigm in textual studies, but, importantly, it also accounts for aesthetic dimensions of texts. Tønnesson (2004), for example, analyses how historians communicate, using the Bakhtinian idea of voices and modelling texts as musical scores.

Critique has been raised that the notion of sakprosa is too wide and cover a too vast field of written texts (cf., Tønnesson 2012). Another criticism is that the notion implies a naïve understanding of the relation between a text and its extensional references, so that such texts directly represent the outer world or a subject matter. It has also been argued that the notion is too closely related to paper-based media, such as books, and consequently does not focus on, for example, digitalization and new social media.

In this chapter—partly addressing this critique—we outline a theory of texts as cultural artefacts. We depart from what we take to be the constitutive features of any text. We will stress how the text as a cultural artefact has the quality of at the same time being a unique utterance and instantiating a text norm. Furthermore, a text is part of a text culture, to which other text sharing similar norms belong. Thus, we conceive of ‘text’ as an utterance that competent participants in a specific cultural time and space assign a certain cultural value. We will also discuss basic principles for the structuring of texts.

These constitutive features are first illustrated in an analysis of children’s utterances, which allows us to pinpoint texts as a major means for acting in and experiencing the world. We then exemplify them in an historical analysis of how the text cultures of news and public opinion evolved in seventeenth- and eighteenth-century Scandinavia. We will point to examples from empirical work that has been carried out in Nordic research projects and Ph.D. theses. Our main argument is that it is by delimiting the object of analysis, namely elaborating a theoretical understanding of texts as cultural artefacts, the research field of sakprosa gains a coherent and sound epistemology. We conclude by discussing the text theory we propose in relation to text linguistics and systemic functional linguistics to further clarify our standpoints.

Constitutive Features of Texts as Cultural Artefacts

To pinpoint the basic qualities of texts, we argue that the text as a cultural artefact is constituted by six features. We introduce them below and then exemplify them in the following sections.

• 1. All texts are unique-situated utterances. In an utterance, meaning is created and conveyed in communicative co-operation between an actual utterer and an addressee in a unique situation. Utterances are, therefore, representations of intersubjective intentionality (Berge 2012b; Ledin 2013). The intentionality of any utterance may be oriented towards and model an addressee representing another human being, but it may also be oriented towards and model the utterer him-/herself as the addressee, as in auto-communication, in what Lotman (1990: ch. 2) in his cultural semiotics approach calls an “I-I language”.

• 2. All utterances and, therefore, all texts are mediated semiotically, using more or less elaborated sign systems, such as a lexicogrammar, drawing, mathematical notation, and photos. Texts may be oral as well as written. These semiotic resources are realized in different types of materiality, such as sound waves, clay, stones, sticks, parchment paper, digital screens, as has often been noted when looking at the phylogeny and history of writing (e.g., Ledin 2015).

• 3. The third constitutive feature is that all utterances and, therefore, all texts have boundaries providing them an outer form. Accordingly, a text has a contextual configuration representing its distinctiveness and difference from other configurations. For the child taking a sheet of A4, the boundaries of the paper are often taken as constituting an outer form, the bonds of communication, whereas social media naturally sets up more complex configurations, where, for example, on Facebook, posts and comments are marked as for their outer form and possible responses, and where a given paradigm of emojis is included. This parallels Bakthin (1986: 71), who in his dialogical theory argues that utterances have “an absolute beginning and an absolute end” that give texts a wholeness and invite “other’s active responsive understanding”.

• 4. The fourth constitutive feature is that utterances have a more or less developed inner structure. If we return to the Facebook page, the contextual configuration is prefabricated with ready-made posts and comments displayed as boxes. Within this outer form, a post will be written and given an inner structure. Obviously, posts are written differently, so the inner structure can vary, but something must be written for communication to occur.

• 5. The fifth constitutive feature of any text is that it is an instantiation of a text norm. Text norms may be constitutive or directive (Searle 1969; Berge 1990). Constitutive norms are norms that define the necessary features of any activity, such as the rules than constitute chess as chess, or a doctoral disputation as a doctoral disputation. In this way, constitutive norms define the necessary features of relevant utterances in a text culture. Utterances that are construed according to a text norm are by the competent participants ascribed value as text. Utterances that are not construed according to a text norm are by the competent participants considered as a potential text, or even a non-text. Consequently, the realm of utterances is much wider than the realm of texts. Directive norms target how the text should be construed and specify what is to be considered as culturally adequate, in terms of good, original, unclear, and so on.

• 6. The sixth constitutive feature is that texts that are instantiations of the same text norm together constitute a specific text culture, such as religion, belles lettres, the juridical system, the news, politics, and science (cf., Berge 2012b, where text cultures are similar to semiospheres in Lotman’s 1990, theory). A text culture is characterized by norms that have evolved in social interaction and have become shared. Consequently, constitutive text norms are necessary for the formation of a text culture, as they define the text culture as such. Directive text norms can be more or less developed or explicit. In the formation of a text culture, the directive norms may be unclear, as was the case when the text culture of public opinion was established in different countries, such as Sweden (Gustafsson 2009) and Denmark-Norway in the eighteenth century (Berge 1991, 2015).

These six constitutive features outline, we argue, an epistemology where the object of analysis for text research is delimited. The differentiation between utterances and text, as well as between the realm of utterances and text cultures, and the methodological consequences for text research are summarized in Table 1.

Table 1 The levels of description of utterances, text, and constitutive principles

Utterances are unique and individual acts of meaning making. All utterances are immediately and directly accessible for ideographic methodological observation, as we can observe them as oral performances such as political speeches, inscriptions, where a sign system (pertaining to drawing, using the alphabet, or mathematical notation, etc.) is, using available technologies (a pen, a keyboard, an app), materialized (on a paper, a screen, parchment, etc.) (Ledin 2013). An utterance is a cultural artefact, simply because it depends on cultural and semiotic resources that are historically evolved. Artefacts arise when humans modify physical material to represent the outer world and reflect on it. Artefacts are, as Cole and Derry (2005: 4) put it, “simultaneously ideal (conceptual) and material”.

Texts are instantiations of text norms that constitute text cultures. Consequently, texts are only accessible for systematic nomothetic research indirectly by having access to informants that are competent in creating utterances using the relevant text norms intuitively or explicitly. Norms constituting which utterances that are considered texts are often tacit amongst competent participants in a text culture, and only accessible for systematic research by using different strategies for making explicit the norms using the competent participants intuitive judgments as a methodological strategy. These methods should be considered as reconstructive (Habermas 1984).

The third level of textual research is the general level where theoretical models of text constitutive principles are formulated, which is precisely what we do in this section. This level is based on abductive reasonings on how utterances and texts are constructed and formed based on different semiotic resources. It is important to note that utterance-/text-constitutive principles are not the same as grammatical principles in an everyday language. Any text in any everyday language may be translated to any other everyday language without changing the way the text is structured, that is, its configuration and inner structure.

From Utterance to Text: The Growth of Normative Sensitivity in Children’s Scribbling and Writing

In this section, we show how textual features emerge as part of the ontogenesis of writing and the child’s appropriation of cultural norms. In Fig. 1, we meet a drawing made by Espen, four years old, in which he mediates and reinterprets his relationship with his father. Figure 2 displays an attempt by Ella, who has just turned six years old, to write her autobiography.¹

Fig. 1

A drawing by the four-year-old Espen, portraying the relationship between him and his father

Fig. 2

Two folded A4 sheets used by the six-year-old Ella to write her autobiography

Both these artefacts are unique utterances, where the children also engage in auto-communication to make sense of their self and their identities and social relations (text feature 1). They are mediated semiotically, where the sign system used by Espen in Fig. 1 is drawing, or scribbling, whereas the older Ella in Fig. 2 combines drawing with alphabetic writing, and where this is materially expressed by using sheets of A4 paper and felt pens (text feature 2). The children draw upon and use material and semiotic resources historically developed in culture. The utterances have their origins in family settings. They are made at home and collected by the children’s parents.

The outer form that gives the contextual configuration (text feature 3) and makes the text apt for “the responsive understanding of others” (Bakthin 1986: 71) is for Espen the bonds of the A4 paper, within which he portrays his relationship with his father. Ella deploys a more elaborated contextual configuration. She has folded A4 papers to signal that she is in fact writing a book (in Fig. 2, the sheets of A4 paper are unfolded by us). But Ella also uses writing to meta-communicate. The first page reads (with misspellings not rendered in the translation): ‘Hello. I will tell a story about myself.’ (HEJ. JAG SKA BERETA OM MEJ SELV.) This—together with the folded A4 sheets—codes what may evolve as an autobiographical genre, a personal narrative with a chronological structure. It can be noted that in Scandinavia it is common that children use Hej and Slut (‘Hello’ and ‘The end’) to mark outer form, or “an absolute beginning and an absolute end” of the utterance (Bakthin 1986: 71), and this also happens in young children’s school writing (Ledin 2013). As for the actual narrative and inner structure (text feature 4), there is just one sentence (found on the next sheet): ‘I was born one day’ (JAG FÖDES EN DAG.)

Espen’s drawing draws heavily on the unique context of his family. Here, to get a sense of the inner structure (text feature 4), we have to know that Espen has put himself, drawn in blue, in the middle and his father (also indexed by his car to the left) in yellow at the bottom. The red figure at the top is an elk, with a tree and a sun to its left—here Espen indexes that he is telling a story to his father, or the theme of what he wants to say. The inner structure of the drawing is thus based on a triptych.

As for these utterances to be an instantiation of a text norm (text feature 5), it is clear, as we have said, that Ella has a sense of the text norm of autobiography. She tells the story of herself since her birth. She has—being six years old—a good sense of this genre tradition and its constitutive text norms, but developing the inner structure, using the linear medium of writing, and aiming for a book, is more than she can handle. Obviously content generation is a problem. She was certainly born, but then it becomes hard for her to remember her early years, which hampers the development of the narrative.

Whereas Ella’s utterance qualifies as (an attempt to write) a cultural text, it can be discussed if and how Espen instantiates a text norm (text feature 5). His drawing is embedded in his unique family context, where his father often tells him spooky stories. The relation to his father is semiotically foregrounded, but the roles are reversed, so that Espen positions himself as a self-confident narrator of such a story. This is indexed by Espen’s broad smile, to be compared with his father’s lips going down, signalling that he is somehow uneasy. Therefore, the utterance both draws on and reverses the traditional positions within his local family context. For someone competent in the domain of children’s drawings, where the child and his family are most often foregrounded, such as a parent or preschool teacher, the utterance could be ascribed culture value and viewed as an important semiotic tool for, in this case, visualizing fantasies and anxieties. Such interpretative work by the part of the reader, that is, reconstructing a context of culture, is necessary to relate the utterance to a text norm and ascribe its value.

The general text culture (text feature 6) to which Ella’s text in Fig. 2 belongs could, using Miller and Shepherd (2004), be characterized as depending on exhibitionism and voyeurism, where semiotic mediation is about personal feelings and beliefs being communicated and confirmed by others, and where participants consequently have access to others’ private and even intimate life. Facebook and Instagram are two of many platforms being designed for this. Ella has a sense of this exhibitionism and try out this cultural norm—she also draws herself to the right of the written text on both sheets. Twentieth-century autobiographies were written mostly by old and successful men looking back at their lives, but nowadays we can meet autobiographies written by six-year-olds.

As for Espen, his drawing is part of the social and material world of his unique family, where he creates a cultural artefact to reflect and imagine. His drawing allows him to make a representation of and ponder family life and invent events that have never existed—like him reading for his father (cf. Cole and Derry 2005).

The Development of Constitutive Text Norms: Historical Evolvement of the Text Culture of Written News and Political Opinion

We now turn to how public news and opinions were created in the Nordic countries in the seventeenth and eighteenth centuries and how this provoked profound changes in norms for how written and printed news and political opinions should be structured (Berge 1998, 2015)

In Fig. 3, we see the first printed Dano-Norwegian newspaper called Den Danske Mercurius. It was established in 1666 on the initiative from the King Frederik III, but written by the former tutor Anders Bording, and financed by a county in Norway. The newspaper was published in the quarto format and was four pages long. The quarto format was discursively related to colloquial local and international topics and reading practices, and the merchant class and the kings’ officials were model readers. The picture shows the edition of March 1, 1677, informing the citizens of the twin kingdoms that the king’s most trusted official, and consequently the most powerful man in the kingdoms after the king, Peder Griffenfeld, was overthrown.

Fig. 3

The first page of Den Danske Mercurius from March 1, 1677

What is peculiar is that the newspaper was written in alexandrine verse form, refined especially in the French classical tragedies of Corneille and Racine. This intertextual normativity indexes that it was the absolute king’s voice that was mediated in the printed verses where local and international news were presented as small bulletins. The text norms drawn upon reveal that the newspaper was a comprise between the king’s information control of local communication in the realm and the merchants and officials demand of reliable facts in their professions and everyday activity. As a result, Den Danske Mercurius could not be characterized as communicating propaganda, nor did it mediate ‘fake news’. Still, the newspaper presented news from the angle of the sovereign king, and it was warranted by the absolute monarchy’s economical and ideological interests. Not surprisingly, it could not satisfy the merchants and officials demand for information on the political and economic realities.

The eighteenth century witnessed economic changes related to mercantilism which led to the development, growth, and establishment of a new and dynamic social class which built its position in the society on trade and investments: the bourgeois. This social class’ environment was the city, and the members of the class met at the square, the coffee houses, and the stock market. Trading and economical investments were risky business and demanded access to accurate information on local and international events.

It was this development which established the ground for a new commodity called ‘news’, which written or in printed form led to a new industry for information sale, called “newspapers”, “journals”, “avvisi”, “gazetta”, and so forth. Together the newspapers established a new text culture called ‘the press’, where the text-producing participants were called “journalists”. In many European countries, this need of accurate information went hand in hand with trade and economic growth. The advancing bourgeois class came to challenge the sovereigns’ monopoly on information and distribution of information in their realms.

An extreme example of this normative instability is how political opinion developed. In Sweden a quite wide freedom of the press was granted in 1766 (Gustafsson 2009). An even more radical reform was passed in Denmark-Norway in 1770 (Berge 1991, 1998, 2015), creating an international sensation. Not only was it now possible to inform the citizens of actual incidents locally and internationally from a potentially objective perspective, but it was also—for the first time in modern European history—possible to express in written and printed public opinions on the autocratic political system, as well as criticize it. Actually, the royal rescript abolishing censorship expressively invited the citizens to criticize and in this way “inform older times’ delusions and prejudices” (Berge 2015).

In Fig. 4, we see one of the over 1000 pamphlets published during this freedom-of-writing period in Denmark-Norway (Berge 1991). It is called “Samtale imellem Einar Jermonsøn og Reiar Randulvsøn paa Opland i Aggerhuus-Stift i Norge” (‘Dialogue between Einar Jermonsøn and Reiar Randulvsøn at Opland in Aggerhuus diocese in Norway’). In the pamphlets the authors tried to achieve something rhetorically new, as they wanted to criticize the power elite and call for political action. This called for the enunciation of two speech acts. First, a declarative act where the actions of people with power are defined and presented as more or less ‘corrupt’. This act is supported by narratives of the power elites’ behaviour in order to warrant the authors’ utterance. Second, a directive act encouraging the reader to take political action, that is, to accept the authors claims as the basis for political intervention.

Fig. 4

The Norwegian pamphlet “Samtale imellem Einar Jermonsøn og Reiar Randulvsøn paa Opland i Aggerhuus-Stift i Norge”, published in 1771

The pamphlet in Fig. 4 was published in Copenhagen in 1771. It was written by two Norwegian peasants that had been told by the vicar that the King allows them to ‘print books, without that the bishop or other superiors’ censor them or investigate the identity of the author’.² They ask the printer to give the pamphlet to its main addressee, or model reader, which is the King himself. Moreover, the authors swear to God that the book is ‘trustworthy’.

The pamphlet’s general topic is that the Kings’ officials in the peasants’ diocese are corrupt. It informs the King that the peasants are ripped off by these officials in the diocese, in the same way as the King himself is ripped off by international money lenders.³ Since the peasants and the King share the same experience and mutually support each other, and thus are allied by an ancient social contract, they request the King to fire and punish his officials in the Norwegian diocese.

A striking fact is that the pamphlet is written in the peasants’ local Norwegian dialect. It is the first known example of Norwegian used as written language for serious purposes since the Lutheran reformation in 1536. Still, the use of Norwegian is not consistent throughout the text. When the peasants formulate their direct request to the King at the end, they use a more formal register in Danish, indexing the language of and phrases from the Danish Lutheran translation of the Bible. Consequently, the language of the pamphlet should be characterized as heteroglossic.

Heteroglossic hybridity also characterize the pamphlets’ textual structure. The authors use a elaborated frame composition as the dominant inner structure to achieve their political purpose. The frame structure is presented in Table 2.

Table 2 The frame structure of the pamphlet in Fig. 4

We hear the voices of the participants through six different frames (I–VI), where the King is positioned as the texts’ model reader in frame I. The King’s officials’ corruption in the peasant’s diocese is documented in narratives in frame VI. The author employs the traditional and popular genre ‘didactic dialogue’ as a normative resource, thereby indicating that an informed person will communicate to an uninformed person about important facts. This takes place in frames II and V. In frame II, the uniformed peasant Einar is telling Reiar what the local vicar has told him in the church last Sunday. In frame III (indirectly) and frame IV (directly), the vicar tells the peasants that the King has published a decree telling them that the citizens of the realm now are free to tell the truth and inform prejudices of all possible subjects. The introduction of this topic in frame IV gives Reiar and Einar in frame V the possibility to inform themselves, the actual readers of the pamphlet, and the model reader, the King, that the officials of the diocese are corrupt. In frame VI, Reiar and Einar identify and narrate stories about the officials in the diocese and their corrupt activities.

This rather complicated way of structuring political utterances in the completely new text culture of public opinion is typical of the ways the pamphlets of the first freedom of writing period in Europe were organized. The elaborate frame structure in most of these pamphlets reveals that text norms adapted to this new semiosphere of public opinion were not clear-cut or established. As also Gustafsson (2009) points out for the contemporary situation in Sweden, a wide variety of traditional genres are drawn upon, such as prayers, odes, and so on, to achieve the political purposes. In this way, the Dano-Norwegian freedom of writing period is characterized by text structures being tested out. In the longer term, this led to a recognizable text culture based on clear-cut constitutive text norms for political debates, where also regulative text norms were established as for the structuring of texts.

Comparisons with Other Text Theories Used in Sakprosa Research

The view of texts as cultural artefacts that we have outlined is influenced by and must be compared to other theories. Here we will first discuss text linguistics, which since its establishment in the 1970s has profoundly influenced how we conceive of texts, not least in the Nordic countries and continental Europe. Second, we look at systemic functional linguistics (SFL), where seminal work was conducted in the UK and Australia. SFL has, since the turn of the millennium, become a sort of grand theory for the study of text and social context. We discuss these theories to pinpoint our epistemology. We also bring up these traditions since our theory of texts as cultural artefacts is developed in dialogue with them.

The core question of text linguistics concerns what constitutes textuality, and the answer is based on the notions of cohesion and coherence. In Enkvist’s (1974) seminal paper on basic questions of text linguistics, cohesive ties between clauses and sentences are distinguished, such as reference and conjunction, and an important insight is that cohesion by itself does not ensure coherence (cf. Fossestøl 1980). Cohesion does not make a text semantically meaningful by itself. Enkvist notes that it is hard to delimit a non-text, since we as human have a capacity for inventing one or another context where an utterance makes sense. And this in contrast to grammar, where a non-grammatical clause can easily be constructed.

Important in text linguistics is that the text becomes the primary object of enquiry and that texts are treated as meaningful—coherent—wholes. Halliday and Hasan’s (1976) classic work on cohesion in English provides a huge taxonomy of cohesive ties and chains in the same vein as Enkvist, where the linear structure of texts is the focus of the analysis. Once again it is determined that a text “is not a grammatical unit” (ibid., p. 5). Thus, we must, as Berge (1990) explains, separate language norms (that constitute separate languages) from text norms (that qualify utterances as texts). Even if this text linguistic tradition had its heydays in the late twentieth century, it continues to have impact, not least in mother tongue instruction and teacher training.⁴

Another important strand of text linguistics aims at uncovering a delimited set of text types (see Ledin 1999, for an overview), where many cognitive models has been proposed, such as Werlich (1976). Werlich’ basic text types are description, narration, exposition, argumentation, and instruction. These are seen as idealized principles for text structuring used by humans when dealing with their environment. In description our perception is spatially oriented, and in narration temporally, and this is reflected in language. In Scandinavia, Berge (1990) simplifies Werlich’ typology and suggest four text types, each based on a specific type of conjunction. These are the descriptive text type based on additive conjunction, the explicative text type based on implicative conjunction, the argumentative text type based on contrastive conjunction, and the narrative text type based on temporal conjunction.

A later and influential typology is suggested by Adam (1992). Here the relations between propositions are foregrounded and form sequences stating explicit text structures. The text types are similar to Werlich’s. Adam’s sequences have been used in Scandinavian text historical research (e.g., Nord 2008, departs from the instructional sequence when exploring garden books from the seventeenth century until present). It can also be noted that Werlich, Berge, and Adam see the text types as prototypical. We should not expect a text to display features from only one text type, but several—texts are in this sense heterogenous.

As Virtanen (1992) makes clear the issue of text types contains two levels. On the one hand it has to do with the purpose of discourse—Virtanen labels this ‘discourse type’. On the other hand, it has to do with language, or text-internal characteristics, as in Adam’s sequences. This level is labelled ‘text type’. The distinction is an important one and allows for discourse types to be realized by different text types. The argumentative discourse type (where the purpose is persuasive) can in principle be realized by any text type, or combination of text types. As we all know, a narrative, or a long description or explanation, might well serve a persuasive purpose.

Distinctions that resemble Virtanen’s model has been constitutive for the teaching on how to write texts resembling sakprosa as well as research on sakprosa in Scandinavia, as documented in the Wheel of Writing. It is a theoretical construct of writing in Norwegian schools (Berge et al. 2016). Here, writing is construed as both an act (similar to Virtanen’s text types) and as a purpose (similar to Virtanen’s discourse type). The six writing acts are to convince, to interact, to reflect, to describe, to explore, to imagine, and seen as existing independently of the six purposes of writing, which include, for example, persuasion, knowledge organization, and identity formation. So, acts are related to purposes in a flexible way, and the model elaborates how combinations of acts and purposes might be semiotically mediated.

So far, we have pointed out how research on sakprosa has been and still is influenced by text linguistics and its tools and analytical distinctions. What text linguistics lack, we argue, is a notion of text that takes in cultural norms and thereby distinguishes utterances from texts. In the text linguistic and cognitive paradigm texts most often come out as tied to psychological processes (Berge 1993), as in the widespread work of Beaugrande and Dressler (1981: 3), who defines “TEXT” (upper case in original) as a “COMMUNICATIVE OCCURRENCE which meet seven standards of textuality”, where cohesion and coherence are the first two standards, and others include, for example, informativity and acceptance.

A text thus becomes a unique linguistic occurrence of communication, coded by universal cognitive operations. In other words, a text is conceived of as meeting criteria—or standards—for cognitive processing. Differences between texts are consequently explained by individual differences between actors as for the cognitive processing. This means that actors are stripped off culture and social context, and that the text as a cultural artefact disappears.

Returning to Figs. 1 and 2, the children’s texts by Ella and Espen, we have argued that they—like all texts—must be seen as utterances. These utterances are cultural artefacts, where materiality is shaped in ways possible in the actual culture and conceptual worlds are set up in communication. They are contextually configured with an outer form, within which an inner structure unfolds. This also explains the wholeness of utterances and texts, or coherence. Coherence, we argue, does not, as text linguistic theory proposes, arise as a function of universal cognitive processes, but is inherent in the utterance as a cultural artefact. When an utterance instantiates a text norm, it comes to be valued as text, which is the case for Ella’s utterance in Fig. 1, where the genre of autobiography is evoked and tried out.

The impact of systemic functional linguistics (SFL) on all sorts of discourse studies has been enormous, and it has certainly influenced research on sakprosa. Here, we will discuss the SFL notion of text and pay special attention to SFL approaches to multimodality and educational linguistics. SFL defines text as “any instance of living language that is playing some part in a context of situation” (Halliday and Hasan 1989: 31). This means that dialogue and spoken interaction becomes fundamental for making meaning, where the context of situation unfolds over time. In written texts—or text as a product—the context is encapsulated in and can be recovered from the text (ibid., p. 11). Important here is that texts are made of meanings and that these meanings are systemic and formalized in the three metafunctions. These three strands of meaning are ideational (construing experiences), interpersonal (construing social relations), and textual (construing coherence).

The metafunctions resonate with the context of situation, where ideational meanings relate to field of discourse (what is going on), ideational to tenor (who are taking part), and textual to mode of discourse (the role assigned to language). This means that the communication itself, the social situation, is seen as a metafunctional configuration, and that the organization of semiotic resources also reflect the metafunctions, a view that in SFL is coded in models of stratification.

So SFL wants to understand and describe the systemic resources that humans use when making meaning through texts. It is a universal semantic theory, often labelled social semiotics, which means that texts can be coded in different semiotic systems. The most influential work here is Kress and Van Leeuwen (1996) attempt to construct a visual grammar. Kress and van Leeuwen developed their model from the first edition of Halliday’s Introduction to systemic functional grammar (1985) and made this grammar applicable to images.

The basis for this is to develop metafunctionally organized systems where meanings are formalized as choices and then analyse texts according to the choices made (cf. Jewitt et al. 2016: 49). So, for example, in Halliday (1985), the ideational metafunction is formalized in a systemic network of processes, such as material and mental processes. Kress and van Leeuwen (1996: 73) adopt and revise this network, where we, just as in language have material and mental processes, where the former can be agentative or not and the latter just as in language code propositions. This means that the ideational metafunction, formalized in systemic networks, not least for processes, is presented as the entry point for studying any semiotic mode or material. So, for example, when Ravelli and McMurtrie (2016) develop what they call SDA (Spatial Discourse Analysis), a similar systemic network for processes recurs when analysing architecture and buildings (p. 31) (cf. Ledin and Machin 2019a).

Kress and van Leeuwen’s visual grammar has had an immense impact, also on Nordic research on sakprosa. The detailed analyses of images appealed to linguistics and sparked all sorts of historical and multimodal analyses of the evolution of different texts and genres: encyclopaedias, textbooks, the weekly press, garden books, and handbooks of different sorts (see Melander and Olsson 2001, for such examples). There are many textbooks which, in a simplified and pedagogic manner, explains Kress and van Leeuwen’s work, for example, Björkvall (2009) in a Nordic context. An overview of this tradition and a discussion of multimodal sakprosa research in the Nordic countries are found in Ledin et al. (2019).

The grammatical approach is characteristic for different SFL analyses and focuses on how systemic resources are used in texts. An ‘outside-in’ approach is taken, where texts are divided into discrete semantic units pertaining to the metafunctions. This approach risks to shatter the wholeness of texts, and, as in text linguistics, its qualities as a cultural artefact disappears (Berge 2012a). Also, since the metafunctions link language to context, both are rendered as metafunctional configurations, and actual social contexts and actors might well disappear. As Ledin and Machin (2019a) argue, SFL tend to become a ‘grand theory’ for researching any semiotic system in more or less the same way, regardless of their materiality, sociohistorical origins, the rhetorical exigences in a certain context, and so on.

To be fair, this critique does not apply to all SFL approaches. There is also an ‘inside-out’ analysis, where systemic resources used and foregrounded in the actual text are pinpointed. Still, if we go to Martin and Rose (2008), a genre model with a stratification of context is proposed, in which different outer levels are realized in inner ones in a strict ‘outside-in’ fashion. Genre is an outer layer in this model, which means that social purposes and generic structures are reified. Genre is then realized through register, discourse semantics, and lexicogrammar, with the metafunctions cutting through and resonating in all levels (Martin and Rose 2008). As for developing a genre pedagogy, there might be advantages with modelling cultural resources in this way, as the reification makes culture visible.

For our purposes, in developing our theory of texts as cultural artefacts, this structuralist view of genre, and of texts as being determined by pre-given systemic networks, is not productive. Following Bakthin (1986), we see every text as an utterance, existing in a unique context of situation, where linguistic patterns and generic structures are more or less stable. The eighteenth-century pamphlet in Fig. 4 exemplifies this. It is an utterance taking political stance, so in this sense, in recognizing it as expressing a political opinion, we can talk about a very general constitutive text norm. But the pamphlets produced in this era showed extreme variation as for different textual patterns and did not follow clear-cut regulative norms.

Also, the inner (and generic) structure of texts exists within the outer form of a text, such as in Ella’s utterance in Fig. 1. Here, a mutual intentionality is coded in the general frame of ‘this is me telling you about myself’. The outer form creates a wholeness that is not accounted for in SFL. Furthermore, both SFL and text linguistics models mostly linear structures, for example, conjunction on a local level or the staging of the text on a global level. As we have shown in the analysis of the pamphlets documented in Table 2, frame structures, in which different voices are positioned, are a fundamental textual and communicative means (Berge 2012a; Ledin and Machin 2019b).

Final Remarks on Epistemological Status of the Theory of Texts as Cultural Artefacts

In this chapter, we have outlined a theory of texts as cultural artefacts. In developing this theory, we have drawn upon different other theories, for example, cultural semiotics (e.g., Lotman 1990), dialogism (e.g., Bakthin 1986), and cultural psychology (e.g., Cole and Derry 2005). In order to position and explain the theory, we have related it to text linguistics and systemic functional linguistics, two traditions that have had an immense impact on text research in the Nordic countries and elsewhere and that we both have worked within. Our aim is not to criticize these traditions per se but to point to how we conceive of, for example, coherence as a function of the qualities of an utterance as a cultural artefact and not, as in text linguistics, as a function of universal cognitive processes, or how we view unique utterances with a contextual configuration and shared intentionality as being the fundamental means of communication, not grammar or genres, as in systemic functional linguistics.