The pragmatic development of (jag) hoppas / hoppas jag in Swedish

Hanna Lehti-Eklund, Department of Finnish, Finno-Ugrian and Scandinavian Studies, University of Helsinki

1. Introduction

The aim of this study is to sketch the pragmatic and semantic development of the Swedish clause jag hoppas (att) + complement ‘I hope (that) + complement’ from the seventeenth century until present-day Swedish, particularly its connection with the construction(s) hoppas / hoppas jag in present-day Swedish. I will focus on the question whether this connection can be depicted with reference to a grammaticalization process, so that hoppas / hoppas jag can be regarded as discourse markers.

In present-day Swedish, hoppas is a deponent verb meaning ‘wishing and to some extent believing in a favourable development’ (Teleman et al. 1999, part 2: 554; Svensk ordbok 1986: 473). According to The Dictionary of the Swedish Academy (Svenska Akademiens Ordbok 1932), the main meaning of hoppas is ‘to expect that something one wishes is going to happen, to have a hope of something.’

The main uses of the verb hoppas with the subject I in present-day Swedish which are listed in dictionaries are shown in examples (1)–(3) (cf. Svenskt språkbruk 2003: 492). According to The Grammar of the Swedish Academy (Teleman et al. 1999, part 2: 555), hoppas occurs in the following constructions: hoppas {(att) [complement]} where the complement can be a nominal dependent clause starting with the complementizer att ‘that’ (1a) or an infinitive or infinitive phrase (2). The complementizer can, however, be left out. As Swedish is a V2-language, inversion is used if the predicate is preceded by one or more constituents. An initial adverbial, for example, causes an inverted word order as in (1b):



I hope that we will meet again soon.’




‘Therefore I hope that we will soon meet again.’

If the main clause and the infinitive phrase share the same subject, the complement is usually an infinitive phrase, as in (2), rather than a dependent clause:



I hope to be able to come.’

A third possibility is to combine hoppas with a prepositional phrase with the preposition :



I hope for a continuation.’

The combinations of hoppas with an infinitive or a prepositional phrase shown in (2) and (3) are not dealt with in this study.

In vernacular Swedish, hoppas is also used with the first-person singular pronoun in constructions and functions that are not found in dictionaries. Placed initially in an utterance leaving out the subject I, hoppas can be combined with a dependent clause without the complementizer att as in (4a).



Hope I won’t freeze to death.’

(, accessed 18.3.2011)

Example (4a) is from a blog where it is placed above a picture of a scantily dressed girl with a bare belly. The utterance serves as a comment on the message of the picture, which, together with the smiley, creates a comic effect.

The phrase hoppas jag ‘I hope’ can also be used as a comment clause, as in (4b), or placed finally, as in (4c) (cf. Biber et al. 2000: 197):



Now we are, I hope, on our way into a goal-oriented school.




‘Cold feet – no more !! (I hope)’

Only inverted word order is possible in the last two examples, and the subject cannot be left out. The variation between jag hoppas as in (1a) and hoppas jag (1b, 4b, 4c) is dependent on the construction’s position in an utterance (cf. Karlsson 2005: 129). In (4b) the explicit semantic meaning of hoppas is toned down, so that hoppas jag implicitly expresses the writer’s doubt about the move into a goal-oriented school. The construction’s loose connection with the utterance is shown by it being parenthesized. The function of hoppas jag in (4c) as an independent utterance or afterthought is emphasized by its position after the exclamation marks and the use of brackets. It is also to be noted that the utterance in (4c) has a non-prototypical form, consisting of an NP and an AdvP.

It should also be noted that to a certain extent the use of hoppas (jag) in (4a)–(4c) seems to resemble discourse markers, both formally and functionally. In discussing modal particles in German, Burkhardt (1994: 133) points out that they express the speaker’s subjective evaluation of aspects of the communicative situation and convey his or her opinions and preferences. On the other hand, the usage in the examples also resembles stance adverbials which ‘convey the speaker’s attitude or value judgment about the proposition’s content’ (Biber et al. 2000: 854; Foolen 2003). Formally hoppas / hoppas jag are placed in examples (4a–4c) in the same positions as discourse markers that frequently appear utterance-initially, medially and finally (Brinton 1996: 33). 

The earlier history of the construction is relevant for this paper with respect to the use of the complementizer. During the Early Old Swedish period (1225–1375), hoppas could take either an oblique or a personal subject, but the personal construction dominated (Falk 1997: 167). Parallel ways of construing the subject were also typical of other verbs expressing attitude, such as ångra ‘repent’ (Sundman 1985). An Old Swedish example of the oblique construction is shown in (5a) and of the personal construction in (5b). During the Early Old Swedish period, the oblique and personal variants of the subject I and hoppas + complement could occur either with the complementizer att or with a zero-complementizer (cf. Lindqvist 1912; Falk 1997: 106).



I hope he will come quickly to us.’

RK 1: 3681 (Söderwall)



I hope to God that what is said here of his job we will hear elsewhere.’

Bil 887 (Söderwall)

Hoppas and a number of other verbs lost the ability to take an oblique subject during the Early Old Swedish period (cf. Falk 1997). However, ever since the Old Swedish period, the main clause and the complement have been used with or without the complementizer att ‘that’, as can be seen in two letters from the Early Modern Swedish period (1541–1725). In (6a), with the zero-complementizer, the courtier Johan Ekeblad writes about young hares to his brother Claes. Six months later, he uses the complementizer in another letter to the same person (6b):



I hope we shall fight with them, god willing.’

(JE 1651)



‘But I hope that some rogues will not have reason to be pleased with the circumstances.’

(JE 1652)

The main topic of this study is this relation between the construction jag hoppas (att) + complement shown in examples (1a)–(1b) and the constructions hoppas / hoppas jag in examples (4a)–(4c). The study suggests that the main clause + complement has developed or is on the way to developing into discourse markers with different functions in various positions, namely in initial (i), medial (ii) and final (iii) position.

During the seventeenth and eighteenth centuries, a change appears in the use of jag hoppas (att) + complement in the first attestation of hoppas in initial position without the subject I, as well as hoppas jag in medial and final position. My analysis will draw on data from this period and Present-Day Swedish. By analysing its meaning, function, and position in an utterance, this study sheds light on the pragmatic development of the construction as far as this is possible within the limitations of a historical dataset. This implies making inferences about speakers’ actual intentions.

Two questions in particular will be addressed:

  1. Can hoppas (jag) be seen as a discourse marker in present-day Swedish when it is used as in (4a)–(4c)?
  2. Can the process from (1a)–(1b) to (4a)–(4c) be regarded as an instance of grammaticalization?

With respect to the first question, it is important to consider the fact that the temporality of speech is closely related to the temporality of action. Chafe (1998: 236) reminds us that language is a temporal process, pointing out that ‘language is affected by time on a moment-to-moment scale as people talk’. The position of the construction in an utterance is highly relevant in considering the various functions of (jag) hoppas / hoppas jag and their development. According to whether (jag) hoppas / hoppas jag occurs in initial, medial or final position in an utterance, it realizes different functions. In each of these functions the construction is likely to develop in a different direction.

In recent years, there has been an ongoing discussion about whether the emergence of pragmatic markers from clauses similar to jag hoppas ought to be seen as instances of pragmatization/pragmaticalization, lexicalization or grammaticalization (cf. Brinton 1996, Wischer 2000, Brinton & Traugott 2005, Fischer 2007). Proposals for linking grammaticalization with construction grammar have also been put forward (cf. Diewald 2006, Rostila 2006, Brinton 2008). This discussion will be considered in the present analysis. Features noted in the analysis are the integration of the main clause and the complement and the role of the complementizer att as a boundary marker between the main clause and the complement.

The outline of the paper is as follows: section 2 discusses earlier literature on discourse markers and grammaticalization. The data used in the analysis will be described in section 3. The analysis appears in section 4, and the results are discussed in section 5.

2. Theoretical background

2.1 The semantic and syntactic characteristics of jag hoppas / hoppas jag

2.1.1 (Jag) hoppas / hoppas jag as a discourse marker

Certain kinds of discourse marker typically develop from predicates expressing propositional attitude, for example, I think or I mean (Schiffrin 1987, Palander-Collin 1999). One aspect which is important for this development is that predicates expressing propositional attitude, knowledge or perception such as know or think are not semantically integrated with their complements (cf. Cristofaro 2003: 99, 122). According to Cristofaro’s cross-linguistic classification, hope as a predicate belongs to desideratives which are semantically more closely integrated with the complement, but despite this, hoppas seems to show a similar kind of development as the other predicates. Biber et al. (2000: 369) place hope among mental verbs along with know, think, and see. One question is whether the desiderative semantic meaning of hoppas influences the pragmatic development of jag hoppas in such a way that the result is a construction sharing features with discourse markers and stance adverbials.

The views on how discourse markers are defined and which constructions can be classified among them vary considerably. In this study, “discourse marker” will be used as a broad term covering discourse markers (Lenk 2003), pragmatic markers (Brinton 1996, 2008), and pragmatic particles (Östman 1995). Both semantic, pragmatic, and structural features are considered in defining the current status of hoppas / hoppas jag.

Discourse markers developed from clauses with first-person pronouns and predicates are not generally regarded as prototypical markers because they originate in clauses, whereas prototypical discourse markers, such as well or oh, are short (Östman 1995: 99; Brinton 2008: 241). In her study of English comment clauses such as I mean and I say, Brinton (2008) regards them, however, as pragmatic markers because they exhibit many of their typical features. Brinton bases her categorization on the following points: Discourse markers lack propositional content, have interpersonal or textual functions, are syntactically mobile, and their omission causes pragmatic problems but does not make discourse ungrammatical.

Biber et al. (2000: 1086) define discourse markers functionally as signalling transition in conversation or signalling ‘an interactive relationship between speaker, hearer, and message’. Biber et al. (2000: 1046, 1075) include “utterance launchers” such as You know and I mean in the category of discourse markers and note that stance adverbials shade into discourse markers. Hakulinen et al. (2004: §795, §1601) regard modal particles as one category among particles. In Finnish, they can also be placed in various positions in an utterance.

There is a great deal of cross-linguistic evidence that predicates expressing propositional attitude, knowledge, or perception have frequently developed into discourse markers both with first- and second-person pronouns. Examples of the development of predicates expressing propositional attitude include English methinks, I think, I mean, Swedish jag menar ‘I mean’, jag tycker / tycker jag ‘I think’, jag tror / tror jag ‘I believe’, Estonian ma arvan ‘I think’ (Saari 1986, Schiffrin 1987, Hopper & Traugott 1993, Thompson & Mulac 1991, Palander-Collin 1999, Karlsson 2003, Keevallik 2003, Karlsson 2005, Fischer 2007, Brinton 2008). Examples of discourse markers that have developed from predicates expressing knowledge or perception include English I know, you know, I gather, I see, you see, Swedish jag vet / vetja ‘I know / do I know’, du vet / vetdu ‘you know / do you know’, hordu ‘do you hear’, serdu ‘do you see’, Finnish kuule ‘listen’, kato ‘look’, tietsä ‘do you know’ (Östman 1981, Hakulinen & Seppänen 1992, Hakulinen et al. 2003, Lindström & Wide 2005, Joronen 2007, Brinton 2008). Whether hoppas as a desiderative verb has developed similar discursive functions will be addressed in this study. 

In the next section, the relation between the main clause and the complement will be briefly discussed with reference to their degree of integration.

2.1.2 The role of the complementizer att

In their influential paper on I think, Thompson & Mulac (1991) suggested a so-called “matrix clause hypothesis”, which proposed a zero-complementizer and the mobility of I think in an utterance as an indicator of grammaticalization. They concluded that the development of I think from a matrix clause with a complement into an utterance-final pragmatic marker consisted of three stages. The first stage, I think that C, with C as a complement, would be followed at the second stage by I think C with the deletion of the complementizer that. At the final stage of the development, the pragmatic marker I think would appear utterance-finally: U I think. Thompson & Mulac’s study was based on modern English data.

This hypothesis has found support in some diachronic studies, but others disagree. Palander-Collin (1999: 204-207), who studied the use of methinks in Middle and Early Modern English, found that the hypothesis only applies to methinks in clause-initial, -medial and -final positions without that. In the diachronic study of comment clauses by Brinton (2008: 249), the results do not support Thompson & Mulac’s hypothesis. The syntactic sources are more varied and the development as such is more complex. As regards variation between complementizer and zero-complementizer in English, Rissanen (1991: 287) has pointed out that ‘zero may well have been the unmarked object link throughout the history of English’.

As shown above, there is no diachronic evidence in Swedish that the matrix clause followed by a zero Jag hoppas han kommer ‘I hope he’s coming’ could have evolved from the variant Jag hoppas att han kommer ‘I hope that he’s coming’ with the complementizer att. The variation between using the complementizer or zero could rather be seen as parallel ways of expression showing a difference in genre rather than a diachronic development. Studies of present-day Swedish conversations between children and adults show that whereas children often use att, adults usually leave out the complementizer (Waldmann 2008: 196). Even though diachronic variation between the use of the complementizer and zero in Swedish has not been studied systemically, Lindqvist (1941: xxivff.) concludes on the basis of the Swedish New Testament of 1526 that it is typical of vernacular Swedish not to use a complementizer. The zero complementizer could therefore be considered a feature typical of spoken language.

Even though Thompson & Mulac’s assumption (1991) that leaving out the complementizer represents a later stage cannot be verified in the data with respect to hoppas and att, the frequent omission of the complementizer att seems to imply a closer connection between the main clause and the complement, and as such is also important for the development of jag hoppas / hoppas jag.

The syntactic relation between the matrix clause and the complement is discussed by Givón (1990: 517), who proposes an iconic relation in complementation, so that a high level of syntactic integration responds to a close semantic bond between the complement taking a predicate and the complement. He claims that there is a weak bond between verbs of cognition and utterance and the complement. On the semantic scale in Givón (1995: 125), say and know represent the end of the scale of minimal clause integration with the epistemic hope as the fifth verb. Both Cristofaro (2003: 99) and Givón thus place hope near the end of the scale of minimal clause integration, but Givón regards the verb as less integrated than Cristofaro does.

The relation between matrix clauses and complements in spoken language has been analysed by Laury (2006), who studied the integration of Finnish että-clauses, with että corresponding to att in Swedish and that in English. In a traditional analysis, the clauses functioning as complements have been regarded as objects of the matrix clause. Laury (2006: 318), however, argues ‘that Finnish että-clauses are not best accounted for as subordinate clauses’, because the syntactic integration between the predicate and the complement is loose and the main semantic content is expressed in the complement. According to Laury, the complement-taking predicate ‘opens an evidential/ epistemic frame’. Instead of an integrated constituent, the complement can be regarded as an addition in a chain. Similarly, Fischer (2007: 7) argues that Middle English I trowe and I woot occur as separate, independent clauses before the complement.

Considered in the light of Laury’s arguments, the complementizer att can be regarded as a non-obligatory boundary marker between the complement-taking predicate hoppas and the complement.



I hope that I don’t freeze to death
CTP (boundary marker) complement

In my analysis, hoppas in example (7) is, in Laury’s view, regarded as a predicate opening an epistemic frame that is followed by a non-integrated complement.

2.2 Arguments for grammaticalization or lexicalization

The difficulty of placing the process from main clauses to discourse markers on a cline of grammaticalization can be seen in the disagreement in the literature on how to define this process: it has, for example, been discussed as an instance of pragmaticalization, lexicalization, or grammaticalization (cf. Wischer 2000, Brinton & Traugott 2005: 62-88, Fischer 2007). The views on parameters of grammaticalization vary to some extent depending on the object of the study. I now turn to parameters developed in studies discussing the grammaticalization or lexicalization of main clauses into discourse markers with pragmatic or conversational functions.

The disagreement seems to result in part from different interpretations of the term “grammar”. In linguistics, grammar is still often regarded as having a written language bias, so that phenomena typical of spoken language are not included (cf. Linell 2005: 79). Discourse markers as such are not automatically regarded as part of the grammar as they typically belong to spoken, not written language. Brinton & Traugott (2005: 139) point out that the reason why discourse markers do not fit into the general patterns of grammaticalization is that they do not belong to grammar proper, functioning pragmatically and having an extra-sentential position. Aijmer (1996: 2), for example, used this argument when distinguishing between pragmaticalization and grammaticalization. She claims that grammaticalization is concerned with the development of grammatical forms and constructions from words or phrases, whereas collocations such as I think are examples of pragmaticalization. Brinton & Traugott (2005: 139), however, consider discourse markers as ‘belonging to grammar rather than lying outside it’. Although one suggestion made by Schiffrin (1987: 328) as to what is typical of a discourse marker is that ‘it has to be syntactically detachable from a sentence’, this does not mean that the markers do not form part of syntax or grammar. In his study of the typical positions of discourse markers in Swedish conversations, Lindström (2006: 112) concludes that ‘grammar /…/ is inseparable from what is going on in interaction’. His analysis reveals, for example, how the syntax of spoken Swedish is motivated by language use and how discourse markers, depending on their interactional functions, are systematically positioned utterance-initially or utterance-finally.

According to Brinton (2008: 242-243), the relevant parameters of grammaticalization of comment clauses are decategorialization, acquisition of pragmatic or discourse functions, increased subjectivity, mobility, bleaching, fusion, coalescence, persistence and layering. Brinton & Traugott (2005: 67, 138), who discuss the similarities and differences between lexicalization and grammaticalization in detail, regard the development of discourse markers from clauses as instances of grammaticalization. Their list of parameters is similar to Brinton’s. Both also mention cross-linguistic parallels, which Hopper & Traugott (1993: 28) call “typological generality”, as an additional argument for grammaticalization (Brinton & Traugott 2005: 28, Brinton 2008: 244).  

The most important parameter of grammaticalization is decategorialization, which implies a change of grammatical category from a subject + verb clause to a unit. The concept of fusion refers to the development of a frozen unit with a fixed form which, for example, in the case of comment clauses, indicates that there is no variation as regards tense and no optional modifiers, as in I dunno (Brinton 2008: 242). Coalescence, as in phonological attrition, is not as common in the grammaticalization of matrix clauses as in other cases; English examples such as you know > y’know or look you > look’ee can be mentioned (Brinton & Traugott 2005: 138; Brinton 2008: 242). Subjectification implies a development towards expressing relations or stance from the subject’s viewpoint. The development of Middle English epistemic parentheticals such as I gesse or I trowe may be mentioned as examples of a semantic change from a propositional to an expressive or an interpersonal meaning (Brinton 1996: 253-254). Replacing propositional meaning, the grammaticalized markers acquire pragmatic or discourse functions in interaction, in that, as the meaning of intention of the verb mean, for example, is weakened, I mean is increasingly used to mark repair or to exemplify. Persistence and layering refer to the fact that traces of the original meaning are evident after grammaticalization and that older forms exist beside newer ones (Brinton 2008: 244).

Brinton & Traugott (2005: 136-140) and Brinton (2008: 242-244) argue that discourse or pragmatic markers such as I think are the result of grammaticalization rather than lexicalization. Brinton & Traugott point out that results of lexicalization usually belong to major lexical categories, such as nouns or adjectives, and not discourse markers. Brinton (2008: 243) stresses that subjectification, intersubjectification and generalization of meaning are involved in grammaticalization, whereas lexicalization usually leads to more specific meanings. In her work, pragmaticalization is regarded as a process within grammaticalization. Brinton & Traugott admit, however, that both grammaticalization and lexicalization share the characteristics of fusion and semantic demotivation.

Other researchers have put forward arguments for lexicalization instead of grammaticalization or a combination of both. Wischer (2000) states that there are several similarities between lexicalization and grammaticalization, such as the unidirectionality of semantic bleaching and phonetic attrition. In discussing the development of methinks, she concludes that both processes are involved. Fischer (2007), who studies the development of English and Dutch parentheticals, also compares the processes of grammaticalization and lexicalization. Parenthetical phrases like I think, she argues, were instances of lexicalization rather than grammaticalization. She considered them to be formulaic tokens with a more epistemic, evaluative sense than the original phrases.

In recent research, grammaticalization has also been linked with Construction Grammar. According to Diewald (2006), constructional approaches where there is no sharp line between syntactic structures and lexical components are well-suited to gradual grammaticalization processes. Comparing lexicalization and grammaticalization, Rostila (2006) points out that the construction grammar perspective provides one explanation for the difficulty of keeping the two processes apart. He assumes that since the units resulting from lexicalization and grammaticalization can be processed holistically as different types of construction, ‘their internal structure can be ignored, and therefore it can also be expected to fuse and coalesce’ (Rostila 2006).

Construction Grammar may also offer solutions to some of the other problems mentioned above. It includes pragmatic meaning alongside semantic meaning and recognises “core” and “peripheral” grammatical patterns, such as discourse markers, on an equal footing (Fried & Östman 2004: 12). A constructional approach might also explain the cross-linguistic similarities in the development of main clauses into fixed units with discourse functions (cf. Brinton 2008: 254–256).

3. Data

The data analysed in the present study is both diachronic and synchronic in order to trace the historical development and to show present-day usage of (jag) hoppas / hoppas jag. Expressing stance through hoping is a verbal action conveying speaker or writer involvement or showing the speaker’s perspective on what is being spoken or written about (Fitzmaurice 2003: 110). Hoping may also be a dialogical action addressed to someone, a participant in a dialogue or the recipient of a letter (cf. Nordlund 2005). Linguistic instances of (jag) hoppas / hoppas jag are therefore to be found in genres showing addressee-oriented involvement, such as private letters, or genres including dialogue, such as plays. Nevalainen (2004: 183) sees personal letters as sharing features with spoken language; thus they can be regarded as being closer to comedy or fiction than literate genres.

The data for this study has therefore been selected to meet the following criterion: it consists of written or spoken dialogic genres or genres showing addressee involvement where hoppas is used in the present tense together with the subject jag ‘I’. In Swedish, the written genres which are closer to spoken language become more frequent from the sixteenth century onwards. The first bigger collections of private letters and diaries in Swedish go back to the seventeenth century, with the first plays appearing in the eighteenth. The diachronic corpus comprises texts from the sixteenth to the eighteenth centuries representing these genres, and the present-day Swedish data on conversations, literary dialogues and texts on the Internet, as examples of computer-mediated communication, are typical hybrid forms combining features of spoken and written language (Georgakopoulou 2003).

3.1 Corpora used in this study

The analysis of the earlier development of jag hoppas / hoppas jag is based on a corpus from the sixteenth to the eighteenth century extracted from an electronic Lund corpus of Early Modern Swedish texts (Fornsvenska textbanken, nysvenska annexet), consisting of private letters (JS), a biography (AH), a play (CG) and other fiction (JR) (Table 1). Three private letter corpora (JE, EOMS, MSEO) have been compiled by using the editions of the letters in volumes published in the nineteenth century.

Even though hoppas was used in Old Swedish, it appears very rarely in data before the sixteenth century. It could probably be found in private letters from the Old Swedish period (14th to 15th centuries), but this would need a separate study. Searching other than dialogic genres of the Lund corpus failed to provide any instances of jag hoppas (hoppas jag). [1]

From the seventeenth century onwards, hoppas seems to appear regularly in the data. Table 1 shows the total instances of jag hoppas / hoppas jag in the various texts and their frequency in the data counted as instances per 1,000 words. The searches have been made with the strings hop/håp in order to capture all the spelling variants.

Table 1. Occurrences of jag hoppas / hoppas jag in the diachronic data. Mean frequencies (/1,000), with absolute numbers in brackets.



Mean frequency




Anna Wasa’s letters (AV)





Johan Ekeblad’s letters to his brother Claes Ekeblad (JE)



not available


Agneta Horn Beskrivning över min vandringstid (AH)





Eva Oxenstierna’s letters to her husband Magnus Stenbock (EOMS)





Magnus Stenbock’s letters to his wife, Eva Oxenstierna (MSEO)





Jon Stålhammar’s letters to his wife (JS)




around 1710

Johan Runius’ prose in Swedish (JR)





Eva Oxenstierna’s letters to her husband Magnus Stenbock (EOMS)





Carl Gyllenborg’s play Swenska Sprätthöken (CG)









It is obvious that the construction is not very common in the diachronic data, appearing most often in the letters by Magnus Stenbock to his wife Eva Oxenstierna.

The Present-Day Swedish quantitative data consists of 437 instances of jag hoppas / hoppas jag, 429 of them in the electronic Parole corpus at the Språkbanken, University of Gothenburg, which consists of 19 to 19.4 million words and comprises mainly fiction and newspaper texts. Eight instances have been found in Swedish spoken-language corpora from Finland and Sweden. The Sweden Swedish spoken-language corpus used is the core corpus of the GRIS project (Grammar in Conversation: A Study of Swedish). The Finland Swedish corpora Svestra and SAM at the Department of Finnish, Finno-Ugrian and Scandinavian languages, University of Helsinki, consist of everyday conversations, group interviews and radio conversations.

As hoping is not very often expressed in everyday conversations, the quantitative present-day Swedish data has been complemented by collections of instances of (jag) hoppas / hoppas jag on the Internet in order to illustrate contemporary usage more comprehensively.

3.2 Methodological considerations

The calculation of frequencies and the analysis of jag hoppas / hoppas jag are based on two features in the data, the position of the construction in an utterance and the occurrence/non-occurrence of the complementizer att.

Swedish syntax is relatively fixed, with the finite verb always placed in the second position in main clauses. Main clauses are divided topologically into three fields, the front field, middle field, and end field (Teleman et al. 1999, part 4: 6). An utterance can be expanded with a pre-front and post-end field reserved especially for discourse markers and the third position in the middle field reserved for items like modal particles (Lindström 2008: 70, 95). In Figure 1, the examples from (1a)–(1b) and (4a)–(4c) are placed in a linear model of a Swedish clause showing the positions of (jag) hoppas / hoppas jag.

Expanded clause

Pre-front field

Front field

Middle field

End field

Post-end field

Clausal base




1a. Jag




(att) vi snart ses igen

that we see each other soon again

1b. Därför Therefore





4a. hoppas


jag inte fryser ihjäl

I don’t freeze to death



i morgon










4b. Nu






hoppas jag

hope I

på väg in i en målstyrd skola

on our way to a goal-oriented school



4c. Kalla fötter – no more!!

Cold feet – no more!!

(hoppas jag) (hope I)

Figure 1. The linear structure of a Swedish expanded clause with (jag) hoppas / hoppas jag.

In the analysis, the instances are divided into initial (i), medially-positioned (ii), and final (iii) ones. The first two examples function as a starting-point for the analysis of the development of (i) the initial hoppas C (C = complement) with a zero-complementizer and no subject (example (4a)).

In the middle field (ii), the medially-positioned hoppas jag (4b) takes the third position, which is typically reserved for negation, stance adverbials and discourse markers. In the analysis, this construction is shortened to U1 hoppas jag U1, with U marking the utterance where hoppas jag is positioned. In the last example (4c), (iii) the final hoppas jag is positioned after an utterance as U hoppas jag, as a post-completion which Lindström (2006: 111) claims is typical of verb-first declaratives functioning as “responsive parenthetical contributions” such as tycker jag ‘I think’.

4. From jag hoppas ‘I hope’ to hoppas / hoppas jag ‘hope / I hope’

4.1 The diachronic development

Table 2 shows the instances of the initially, medially and finally positioned (jag) hoppas / hoppas jag in the diachronic data from the sixteenth to the eighteenth centuries. The first three columns are reserved for the initially positioned (jag) hoppas, with instances with the complementizer in the first column, without it in the second, and without subject in the third. The instances that are positioned medially and finally are shown in columns four and five.

Table 2. Jag hoppas / hoppas jag. Total number of instances with and without the complementizer att in the diachronic data.









jag hoppas att C /
X hoppas jag att C

jag hoppas C /
X hoppas jag C

hoppas (att) C

U1 hoppas jag U1

U hoppas jag




2 2





















1697–1701 EOMS







1697–1706 MSEO







1700–1708 JS







around 1710 JR







1713 EOMS







1740 CG





















In the diachronic data, examples of (jag) hoppas C without the complementizer att become more frequent starting from the eighteenth century. One reason may be that texts representing genres that are dialogic or show addressee involvement become more frequent. Another reason could be that the grammaticalization process is starting at this time. In Table 2, the first examples without the complementizer att are in private letters (EOMS, MSEO, JS) and a play (CG).

There are only a few instances of hoppas C without subject in Table 2. Of the five cases, two occur with and three without the complementizer. In the diachronic data, the medially (ii) and finally (iii) placed hoppas jag only occur in the play (CG) in the data, which indicates that both constructions typically occur in spoken language genres.

Table 3 shows the variation between the same categories in the present-day Swedish data, both in the written Parole corpus and the spoken language corpora GRIS, Svestra and SAM. The categorization is the same as in Table 2, but the cases with the initial hoppas are divided into those with and without the complementizer in order to see whether leaving out att could be regarded as a precondition for the development of the marker hoppas.

Table 3. Jag hoppas / hoppas jag. Total number of instances with and without the complementizer att in the present-day Swedish data.







jag hoppas att C /
X hoppas jag att C

jag hoppas C /
X hoppas jag C

hoppas att C

hoppas C

U1 hoppas jag U1

U hoppas jag



186 71







1 -






Svestra & SAM

3 2







190 73







43.5 16.7






The frequencies in the first column are nearly the same as in the diachronic data, whereas the remaining columns show a development that is not yet present in Table 2. The ‘original’ construction of subject + verb+ complementizer + complement has not disappeared, which is clear from the fact that it comprises 43.5 per cent of all constructions. The type jag hoppas C and its variant with the inverted word order have, however, gone down from 43.9 to 16.7 per cent, whereas the type hoppas C with or without complementizer variants represent 31.7 per cent (cf. 5.1 per cent in the diachronic data). Especially noteworthy is the comparison between the third and fourth columns. It shows that there are more instances of hoppas C than jag hoppas C and jag hoppas att C, which means that the initial hoppas followed by a complement seems to be a well-established construction. The medially positioned instances are still rare, but the finally positioned instances have increased in the present-day Swedish data.

Sections 4.2, 4.3 and 4.4 discuss the development of (jag) hoppas / hoppas jag in the three positions in more detail.

4.2 hoppas C

The first instances of hoppas C without the subject I in my data are in letters from the beginning of the eighteenth century (examples (8) and (9)) with variation in the word order in the complement. In both examples, writing from a battlefield, the writers express to their wives hopes concerning the addressee or themselves. In (8), Magnus Stenbock hopes that his wife has received his letters, and in (9) Jon Stålhammar, after stating that he has reason to thank God and his king for his well-being, ends by hoping that he will cope in the war.



I hope my Angel has got them.’

(MSEO 1704)



I hope with God’s help I will cope’

(JS 1704)

In example (8), the word order in the complement is SVO, but in (9) it is ASV with hopass followed by an adverbial of manner (medh Gudhz hielp) before the subject and the predicate. The meaning of (9) could be translated by ‘I hope that I will cope with help from God’. Variation in the word order of eighteenth-century complement clauses is still quite common in Swedish (Pettersson 2005).

There are also some instances with hoppas C without subject but with complementizer from the end of the seventeenth century and the beginning of the eighteenth:



‘I have now written several letters from here; I hope that you have received them.’

(JS 1706)

Parallel with the first instances of hoppas C, earlier examples of the construction with the subject I (jag hoppas C and hoppas jag C) are attested, both with and without the complementizer att. In (11a) and (11b), there are examples of jag hoppas / hoppas jag without complementizer:



‘Finally I hope the Russians don’t come so soon again’

(MSEO 1700)



‘And I hope my dearest Angel is persuaded that all that comes from him to me can be no other than pleasant’

(EOMS 1701)

During the eighteenth century, the word order in Swedish is still somewhat unstable so that main clauses can occur after an adverbial or a conjunction both with and without inversion (Alving 1916: 57; Pettersson 2005: 188ff.). The adverbial in (11a) would nowadays be followed by inversion: Enfin hoppas jag. As the V2-rule does not apply after conjunctions, the inversion in (11b) after och would not apply.

Even though I have not systematically collected data from the nineteenth and twentieth centuries, searches in the corpus of the The Dictionary of the Swedish Academy (Svenska Akademiens Ordbok) show that the construction hoppas C seems to become more frequent during this time. This is also confirmed by the statistics presented in Table 3. As hoppas can still be used as a predicate in a main clause, the variation between hoppas C and jag hoppas C still exists today. However, very often hopefulness is expressed by hoppas C alone, for example, on the Internet. There are not many instances of hoppas in the conversational data, but one prototypical instance from the end of a radio show is illustrated in (12) with the reporter’s final words.



‘still fresh. pt.hh eh: (0.2) hope we see each other then. (.) bye’


On the Internet, hoppas C often occurs at the beginning of an insert where the writer comments on something, as a reaction to something written in a blog, on a home page or on a discussion forum. In these cases, the utterances consisting of hoppas C are addressed to a specific recipient. This is illustrated by examples (13a) and (13b). The complements are often short, consisting of a subject, a predicate and an object, a complement, or an adverbial.

Example (13a) is from a discussion forum for parents on small children. The mother, Jemima (changed name), writes that she is worried about her young son, who is sick and eats very little. B starts her insert by reacting to Jemima’s text with an interjection (Usch), an address and an utterance expressing a hope of improvement in the situation. After this, B continues with her own reminiscences of a sick child:



‘He wakes up 3 times a night and is very hungry, of course because he hardly eats during the day. Blä! Had to write this out of me am so very tired of this now!’


‘Usch Jemima. Hope he gets better soon.’

Example (13b) is a reader’s reaction to a young woman’s blog in which the subject changes from fashion to decoration.



:) hiih. all is fine with me


I hope all is all right with you.’

In (13b), the short comment only includes two utterances, in which the writer comments on her own situation with the first one and in the next one expresses the hope that the blog writer is also well.

The variation in Present-Day Swedish between the construction hoppas C and the main clause type, with subject and predicate, can be observed especially in comparing the usage in headlines and that in the body of the text. Hoppas with a short complement suits headlines demanding short texts as in (14a), whereas the articles in the written news or coverage use the main clause, as in (14b), from a football news page on the Internet:



Hope we can maintain this level!’



‘We are having a good day and I hope we can maintain this level up, B A says’

The construction hoppas C that appears for the first time at the beginning of the eighteenth century seems to have been established during the succeeding 300 years as a typical way of marking the subject’s hopeful stance towards a proposition, especially in spoken language or speech-related genres.

4.3 Medially-placed hoppas jag

The first instances of the medially-placed hoppas jag occur somewhat later than hoppas C, in the middle of the eighteenth century. Its function resembles that of a modal or a stance marker, with a comment function shared with other comment clauses (Brinton 2008: 241). By using hoppas jag, the speaker expresses his attitude towards what is said and tones down the proposition.

There are three examples of the medially-placed hoppas jag in the play Svenska sprätthöken by Carl Gyllenborg from 1740, which are the only ones in the diachronic data. In examples concerning the future, as in (15), the original meaning of hoppas ‘hope’ is expressed in the construction.



‘and there I shall be like my father, I hope, as long as I live’

(CG 1740)

An interesting aspect of the medially-placed hoppas jag is its syntactic mobility. Example (15) shows that the middle field position, shown in Figure 1, is not the only one. The position in (15) after the constituents in the end field and before any possible dependent clauses, shown in Figure 2, seems to reoccur in the data. A modern example of hoppas jag in a similar position figures in (16), where hoppas jag is used to tone down the reproach.



‘You have not forgotten, I hope, that Mollo was from K’


Front field

Middle field

End field

hoppas jag

Dependent clauses

och däri

and there

skall jag

I shall

wara Far lik

be like my father

hoppas jag

I hope

så länge jag lefwer

as long as I live



har inte

have not



hoppas jag

I hope

att Mollo var från K

that Mollo was from K

Figure 2. The linear structure of a Swedish clause with hoppas jag.

This position resembles to a great extent the post-end field position, as hoppas jag is placed after what Teleman et al. (1999, part 4: 6) call “the inner clause”, that is, after the front, middle and end field but before the heavy dependent clauses, which are usually placed after the end field (Teleman et al. 1999, part 4: 468).

Another already established position is the medial one, in which, for example, modal particles are placed in Swedish (Lindström 2008: 69, 95). Example (17), where hoppas jag stands in the middle field together with the negation icke ‘not’, illustrates this position. Here hoppas jag moderates the author’s strong statement:



‘Some rotten apples will, I hope, not come between us.’

(Strindberg 1882; Svenska Akademiens Ordbok)

The following example (18) from a letter from 1892 shows that, in a very literary style, hoppas jag can also be placed within an NP as a free addition called an annex by Teleman et al. (1999, part 4: 456):



‘Your sensible, on the surface somewhat morbid, but, I hope, beneath the surface healthy Scandinavian protoplasm’

(RYDBERG Brev 3: 130; Svenska Akademiens Ordbok)

Syntactically the medially placed hoppas jag is mobile and can be placed both in the middle field and at the end of the end field but before dependent clauses. Even other positions occur to some extent. The construction is used to play down or moderate what is being said.

The next section discusses the uses of the finally placed hoppas jag.

4.4 Final hoppas jag

The first instance of hoppas jag in final position is from a letter from 1652 and two more instances occur in Gyllenborg’s play (1740). In the diachronic data, there are only 3 finally-positioned instances, whereas, with altogether 22 instances, this position is much more frequent in the present-day data.

In example (19), Johan Ekeblad writes to his brother of an ointment that he has found which makes his foot feel better. Here the construction has the original semantic meaning of hoppas ‘hope’.



‘I have, however, now found a sort of ointment that helps me a lot and will be better, I hope

(JE 1652)

In the play, the construction is used after utterances that are replies to a previous speaker’s comment or question. In (20), Sophia starts her turn by not agreeing with the baron’s statement. She ends her negative comment by downtoning it with hoppas jag. Continuing her turn with a second utterance, she addresses Mrs Lotta, who is also on the scene, posing a question to her. Even though Gyllenborg ends the first utterance in Miss Sophia’s turn after hoppas jag with a comma and not a full stop, I have classified this instance of hoppas jag into the category of utterance-final occurrences because the two utterances, a declarative and a question, in Miss Sophia’s turn are syntactically complete and addressed to different people.


BARONSTADIG OmjagintetbedragermigSyster,lärerhanfälla

‘BARON ROBUST  If I am not deceived Sister, he will express the same verdict of you as you of him.’

FRÖKENSOPHIA Intetialthoppasjag,FruLotta,FruLotta,harI

‘MISS SOPHIA  Not in everything I hope, Mrs Lotta, Mrs Lotta, have you not walked yourself tired yet.’

(CG 1740)

The final hoppas jag can be regarded as a grammatical extension of the previous utterance placed in the post-end field. This can be seen in the inverted word order. Pragmatically, it forms a retrospective orientation towards what has just been said (cf. Karlsson 2005, 125). 

In the Present-Day Swedish data, the position after a responsive utterance is frequent, as in examples (21a) and (21b). In these examples, both from literary dialogue, the utterance preceding hoppas jag confirms what the previous speaker has said (21a) or answers a previous question (21b). In literary examples the various interactive functions occurring in conversations are difficult to spot; in the data for this study, the speaker seems to use the construction as an epistemic marker by softening or downtoning what has just been said.



‘You will never love me, Robban said

No and nor anyone else I hope, Kalas said.’




‘How is Lovely Godiva getting on in the competition?

Fine I hope.


In (21b) the final hoppas jag expresses the subject’s view as an afterthought like tycker jag ‘I think’ or tror jag ‘I believe’, which Lindström (2008: 115) calls extended discourse markers.

The question – answer sequence can also be posed by the same person, as in example (22) from the Internet, in which the writer comments on a Swedish radio programme on its home page.



‘When will your radio programme be shut down? As soon as possible I hope.’

The utterance preceding the final hoppas jag can have various structures from clauses to phrases, such as the adverbial phrase in (22) or a nominal phrase as in the following headline from a discussion forum on building houses on the Internet (23):



‘This summer’s project, I hope.’

In written language, there are several ways to show that although hoppas jag syntactically forms a coherent continuation of the previous utterance, it also stands apart as a separate entity. In (4c) hoppas jag is separated from the utterance by exclamation marks and parentheses. In (24a), the construction is preceded by a full stop and surrounded by parentheses, and in (24b) by three dots. Both are examples of blog headlines. In spoken language, pauses or intonation would probably have similar functions:



‘Well now. (I hope)’

(, accessed 2.5.2011



‘Short holiday… I hope

(, accessed 2.5.2011)

5. Discussion

In the present study, three paths of development for the main clause jag hoppas / X hoppas jag have been sketched. A tentative outline of the development from a main clause + complementizer + complement into hoppas / hoppas jag in three different positions is suggested below. The data analysed above shows that the constructions (i) hoppas C and (ii, iii) hoppas jag, which first appear in the eighteenth century, have become more frequent in present-day Swedish. It is, however, difficult to say whether this is the result of certain genres becoming more common or whether a grammaticalization process started at that time. Probably both reasons are relevant for the development.


jag hoppas (att) C
subject + verb


jag hoppas C
subject + verb


hoppas C
emerging DM/SA


X hoppas jag att C
verb + subject


U1 hoppas jag U1
emerging DM


X hoppas jag att C


U hoppas jag
emerging DM

The matrix clause hypothesis suggested by Thompson & Mulac (1991) seems to concern the first cline to a certain extent. They assumed that the three-stage development of the pragmatic marker I think would demand a deletion of the complementizer at the second stage. Even though the use of the complementizer att in my data varies at first, it is increasingly left out later. Although there are instances of hoppas att C in Present-Day Swedish, they are outnumbered by the instances of hoppas C.

One of the questions posed in this study was whether the constructions hoppas / hoppas jag could be regarded as discourse markers. Across various languages, the overall change to pragmatic and interactional functions seems to be typical of main clauses consisting of verbs expressing propositional attitude or knowledge and first-person subjects. Semantically, hoppas as a desiderative verb differs to some extent from these as it shows a certain kind of modal sense. Givón (1990) and Cristofaro (2003) regarded hope and similar desiderative verbs as more integrated with the complement than verbs expressing propositional attitude and knowledge. As a predicate opening an epistemic frame and followed by a complement, hoppas shows a development that to a great extent is similar to that shown by verbs like think or know, but not quite.

According to the categorization by Brinton (2008), discourse markers lack propositional content, have interpersonal or textual functions, are syntactically mobile, and their omission causes pragmatic problems but does not make discourse ungrammatical. Semantically, hoppas both in hoppas C and in hoppas jag has maintained some of the verb’s desiderative and modal sense, but it also has some discursive functions in which the original meaning of hoppas has bleached to some extent. The functions of the constructions hoppas / hoppas jag in initial, medial and final position are somewhat different because of their position in the utterance. 

The initially placed hoppas seems to have preserved the modal epistemic sense to some extent. It can still be used as an epistemic frame for the complement and as such meets some of the criteria of stance adverbials ‘being markers [...] of the speaker’s attitude to what is said’ (Biber et al. 2000: 1046). Some researchers, however, include modal particles expressing the opinions and preferences of the speaker among discourse markers (Burkhardt 1994: 133; Hakulinen et al. 2004: §795, §1601). The initial hoppas C also seems to have some discursive functions, which can be seen most clearly in the examples from the Internet, where it is used as a reaction to a previous contribution. The uses of the initial hoppas C could be summed up as showing features of both stance adverbials and discourse markers.

Placed in the middle of or after an utterance, hoppas jag can comment in different ways on what has been or is being said; for example, by softening or playing down statements made or inferences that can be drawn. These usages seem to fill the function of what Brinton (2008: 241) calls pragmatic markers or Lindström (2008: 115–116) expanded discourse markers. According to Lindström, a finally positioned tycker jag ‘I think’ can, for example, change the truth value of the previous utterance by marking it as a personal opinion. These uses of hoppas jag thus resemble other similar discourse markers.

As regards the criteria diagnostic of grammaticalization, hoppas / hoppas jag can be related to some parameters (section 2.2) but not all. The main clause consisting of a subject and predicate is decategorialized into a marker expressing stance. Both hoppas and hoppas jag appear in the present tense and do not take any other subjects. Whether hoppas in this case functions as a verb any longer is doubtful. The syntax of the initial hoppas C is also characterized by dropping the initial subject I. Leaving out the subject could perhaps be regarded as a criterion of grammaticalization similar to coalescence.

As most of the examples are from written texts, it is not possible to analyse the exact pragmatic and discursive functions of these constructions. The examples show, however, that by expressing hoppas C the writer/speaker always involves the addressee. The use of the construction is subjectified as it only concerns stance or opinions expressed by a first-person pronoun subject (Brinton & Traugott 2005: 29). Given that the other constructions expressing hope with jag hoppas (att) are still in use, the constructions also show layering and persistence.

One of the features characterizing comment clauses is mobility (Brinton 2008). The medially-placed hoppas jag can be placed in several positions in an utterance, with only one of them in the middle field being typical of stance adverbials. 

For the development of the constructions, their positions in an utterance are important. Language happens in time; by using jag hoppas / hoppas jag utterance-initially, medially or finally, we as speakers or writers express linguistically and pragmatically slightly different communicative purposes, which is relevant for the change the constructions have gone through in the last three hundred years.


[1] The searches were made in the Chronicle of Olaus Petri (Olaus Petris krönika, 1530s), the Chronicle of Peder Swart (Peder Swarts krönika, 1560) and Samuel Columbus’s Mål-roo eller Roo-mål (1675).


AH = Agneta Horn. 1959. Beskrivning över min vandringstid. (=Nordiska texter och undersökningar, 19), ed. by Gösta Holm. Stockholm: Almqvist & Wiksell.

AV = Anna Wasa’s letters.

CG = Carl Gyllenborg. [1740] 1959. Swenska Sprätthöken. (=Svenska litteratursällskapets klassikerutgåvor), ed. by L. Breitholz. Uppsala.

EOMS = Eva Oxenstierna to Magnus Stenbock. 1913–1914. En brefväxling II [Correspondence, vol. II], ed. by C. M. Stenbock. Stockholm: P. A. Norstedt & Söner.

Fornsvenska textbanken, nysvenska annexet [Old Swedish Text Bank, Supplement of Present-Day Swedish].

GRIS = Grammar in Conversation: A Study of Swedish.

JE = Johan Ekeblad. 1965. Johan Ekeblads brev till brodern Claes Ekeblad 1639–1655. (=Acta Universitatis Gothoburgiensis. Nordistica Gothoburgiensia, 2), ed. by S. Allén. Göteborg: University of Gothenburg.

JR = Johan Runius. 1870. Prosastycken på svenska. (=Svenska författare utg. av Svenska Vitterhetssamfundet, XVII:3.) Stockholm.

JS = Jon Stålhammar. [1700–1708] 1903. Jon Stålhammars brev till hustrun Sofia Drake. (=Karolinska krigares dagböcker, VII.). Lund.

MSEO = Magnus Stenbock to Eva Oxenstierna. 1913–1914. En brefväxling II [Correspondence, vol. II], ed. by C.M. Stenbock. Stockholm: P. A. Norstedt & Söner.

Parole =

SAM = Audio-recorded conversations collected in the 1990’s in the project Swedish Conversations in Helsinki. Department of Scandinavian Languages and Literature, University of Helsinki.

Svestra = Audio-recorded conversations collected in 1999–2001 in the project Finland Swedish Conversation Strategies. Department of Scandinavian Languages and Literature, University of Helsinki.


Aijmer, Karin. 1996. “I think – an English modal particle”. Modality in Germanic Languages: Historical and Comparative Perspectives (=Trends in Linguistics. Studies and Monographs [TiLSM], 99), ed. by Toril Swan & Olaf Jansen Westwik, 1–47. Berlin: Mouton de Gruyter.

Alving, Hjalmar. 1916. Det grammatiska subjektets plats i den narrativa satsen i svenskan. PhD dissertation, Uppsala.

Biber, Douglas, Stig Johansson, Geoffrey Leech, Susan Conrad & Edward Finegan. 2000. Longman Grammar of Spoken and Written English. Third impression. Harlow: Pearson Education.

Brinton, Laurel J. 1996. Pragmatic Markers in English. Grammaticalization and Discourse Functions. Berlin: Mouton de Gruyter.

Brinton, Laurel J. 2008. The Comment Clause in English. Cambridge University Press.

Brinton, Laurel J. & Traugott, Elizabeth C. 2005. Lexicalization and Language Change. Cambridge: Cambridge University Press.

Burkhardt, Armin. 1994. “Abtönungspartikeln im Deutschen: Bedeutung und Genese”. Zeitschrift für germanistische Linguistik 22: 129–151.

Chafe, Wallace. 1998. “How Time Affects the Shape of Language”. Senri Ethnological Studies 45: 235–250.

Cristofaro, Sonja. 2003. Subordination. Oxford: Oxford University Press.

Diewald, Gabriele. 2006. “Context types in grammaticalization as constructions”. Constructions.

Falk, Cecilia. 1997. Fornsvenska upplevarverb. Lund: Lund University Press.

Fischer, Olga. 2007. “The development of English parentheticals: A case of grammaticalization?”. Tracing English through Time. Explorations in Language Variation, ed. by Ute Smit, Stefan Dollinger, Julia Hüttner, Gunther Kaltenböck & Ursula Lutzky, 99–114. Vienna: Braumüller. Also

Fitzmaurice, Susan. 2003. “The grammar of stance in early eighteenth-century English epistolary language”. Corpus Analysis: Language Structure and Language Use (=Language and Computers, 46), ed. by Charles Meyer & Pepi Leistyna, 107–131. Amsterdam: Rodopi.

Foolen, Ad. 2003. “Pragmatic particles”. Handbook of Pragmatics online, ed. by Jef Verschueren & Jan-Ola Östman. Amsterdam: Benjamins.

Fried, Mirjam & Jan-Ola Östman. 2004. “Construction Grammar. A thumbnail sketch”. Construction Grammar in a Cross-Language Perspective. (=Constructional Approaches to Language, 2), ed. by Mirjam Fried & Jan-Ola Östman, 11–86. Amsterdam/ Philadelphia: Benjamins.

Georgakopoulou, Alexandra. 2003. “Computer-mediated conversation”. Handbook of Pragmatics online, ed. by Jef Verschueren & Jan-Ola Östman. Amsterdam: Benjamins.

Givón, Talmy. 1990. Syntax. Vol. 2: A Functional-Typological Introduction. Amsterdam/Philadelphia: Benjamins.

Givón, Talmy. 1995. Functionalism and Grammar. Amsterdam/Philadelphia: Benjamins.

Hakulinen, Auli, Leelo Keevallik Eriksson & Jan Lindström. 2003. “Kuule, kule, hördu. Projicerande partiklar i finska, estniska och svenska”. Grammatik och samtal. (=Skrifter utgivna av institutionen för nordiska språk vid Uppsala Universitet, 63), ed. by Bengt Nordberg, Leelo Keevallik Eriksson, Kerstin Thelander & Mats Thelander, 199–218. Uppsala: University of Uppsala.

Hakulinen, Auli, Maria Vilkuna, Riitta Korhonen, Vesa Koivisto, Tarja Riitta Heinonen & Irja Alho, eds. 2004. Iso suomen kielioppi. Helsinki: Suomalaisen kirjallisuuden seura.

Hakulinen, Auli & Eeva-Leena Seppänen. 1992. “Finnish kato: from verb to particle”. Journal of Pragmatics 18: 527–549.

Hopper, Paul. 1991. “On some principles of grammaticization”. Approaches to Grammaticalization. (= Typological Studies in Language, 19), vol. I, ed. by Elizabeth Closs Traugott & Bernd Heine, 17–35. Amsterdam: Benjamins.

Hopper, Paul J. & Elizabeth Closs Traugott. 1993. Grammaticalization. (=Cambridge Textbooks in Linguistics.) Cambridge: Cambridge University Press.

Joronen, Piia. 2007. Ärsyttävä kieli. Puhekieleen kohdistuvat affektiset kannanotot Ilta-Sanomien verkkokeskusteluissa. University of Joensuu. Master’s thesis in Finnish.

Karlsson, Susanna. 2003. “Interactional uses for an epistemic marker: the case of ‘jag tycker’/‘tycker jag’ in Swedish”. Melbourne Papers in Linguistics and Applied Linguistics 3(1): 5–23.

Karlsson, Susanna. 2005. “Modalitet i interaktion. En studie av jag tycker och tycker jag. Samtal och grammatik, ed. by Jan Anward & Bengt Nordberg, 119–138. Lund: Studentlitteratur.

Keevallik, Leelo. 2003. From Interaction to Grammar. Estonian Finite Verb Forms in Conversation. Acta Universitatis Upsaliensis. Uppsala.

Laury, Ritva. 2006. “On subordination, Finnish-style: Questioning the category of finite clausal complements in spoken Finnish”. A Man of Measure : Festschrift in Honour of Fred Karlsson on his 60th Birthday, ed. by Mickael Suominen, 310–321. Turku: The Linguistic Association of Finland.

Lehti-Eklund, Hanna. 2005. “Altså hoppas jag min Engell kan nu komma – om hoppas och grammatikalisering”. Översättning och etik, ed. by Paula Rossi, 143–150. Oulu: Uleåborgs Universitet.

Lenk, Uta. 2003. “Discourse markers”. Handbook of Pragmatics, ed. by Jef Verschueren & Jan-Ola Östman.

Lindqvist, Axel. 1912. Förskjutningar i förhållandet mellan grammatiskt och psykologiskt subjekt i svenskan. Lund & Leipzig: Gleerup & Otto Harrassowitz.

Lindqvist, Natan. 1941. Nya testamentet i Gustaf Vasas bibel under jämförelse med texten av år 1526. Stockholm: Svenska Kyrkans Diakonistyrelses Bokförlag.

Lindström, Jan. 2006. “Grammar in the service of interaction: Exploring turn organization in Swedish”. Research on Language & Social Interaction 39(1): 81–117.

Lindström, Jan. 2008. Tur och ordning. Introduktion till svensk samtalsgrammatik. Stockholm: Norstedts.

Lindström, Jan & Camilla Wide. 2005. “Tracing the origins of a set of discourse particles: Swedish particles of the type you know”. Journal of Historical Pragmatics 6(2): 211–236.

Linell, Per. 2005. The Written Language Bias in Linguistics: Its Nature, Origins and Transformations. (= Routledge Advances in Communication and Linguistic Theory.) London & New York: Routledge.

Nevalainen, Terttu. 2004. “Letter writing”. Journal of Historical Pragmatics 5(2): 181–191.

Nordlund, Sara. 2005. “Vet du som dialogiskt drag i en brevväxling från sent 1800-tal”. Språk och stil 15: 5–25.

Palander-Collin, Minna. 1999. Grammaticalization and Social Embedding: I THINK and METHINKS in Middle and Early Modern English. (=Mémoires de la Société Néophilologique, 55.) Helsinki: Société Néophilologique.

Pettersson, Gertrud. 2005. Svenska språket under sjuhundra år. Lund: Studentlitteratur.

Rissanen, Matti. 1991. “Syntax”. The Cambridge History of the English Language. Vol. III: Early Modern English 1476–1776, ed. by Roger Lass, 187–331. Cambridge: Cambridge University Press.

Rostila, Jouni. 2006. “Storage as a way to grammaticalization”. Constructions 1.

Saari, Mirja. 1986. “Jag menar”. Xenia Huldeniana. (=Meddelanden från Institutionen för nordiska språk och nordisk litteratur, B 10): 329–337.

Teleman, Ulf, Staffan Hellberg & Erik Andersson. 1999. Svenska Akademiens grammatik. Stockholm: Svenska Akademien and Norstedts.

Schiffrin, Deborah. 1987. Discourse Markers. (=Studies in Interactional Sociolinguistics.) Cambridge: Cambridge University Press.

Sundman, Marketta. 1985. “Från mik angrar till jag ångrar. Om förhållandet mellan satsdelskategori och semantisk roll”. Folkmålsstudier 29: 85–123.

Svenska Akademiens ordbok.

Svensk ordbok. 1986. Gothenburg: Språkdata and Esselte Studium.

Svenskt språkbruk. 2003. Svenska språknämnden and Norstedts Ordbok.

Söderwall, K. F. 1884–1918­. Ordbok öfver Svenska Medeltids-språket. Lund.

Thompson, Sandra A. & Anthony Mulac. 1991. “The discourse conditions for the use of the complementizer that in conversational English”. Journal of Pragmatics 15: 237–251.

Waldmann, Christian. 2008. Input och output. Ordföljd i svenska barns huvudsatser och bisatser. Lund: Lund University Press.

Wischer, Ilse. 2000. “Grammaticalization versus lexicalization. ‘Methinks’ there is some confusion”. Pathways of Change: Grammaticalization in English (=Studies in Language Companion), ed. by Olga Fischer, Anette Rosenbach & Dieter Stein, 355–370. Amsterdam/Philadelphia: Benjamins.

Östman, Jan-Ola. 1981. You Know: A Discourse-Functional Approach. (=Pragmatics and Beyond, 2(7).) Amsterdam: Benjamins.

Östman, Jan-Ola. 1995. “Pragmatic particles twenty years after”. Organization in Discourse (=Anglicana Turkuensia, 14), ed. by Brita Wårvik, Sanna-Kaisa Tanskanen & Risto Hiltunen, 95–108. Turku: University of Turku.