Role of six turn-initial demonstrative and emotive particles in Lithuanian

1 Introduction

Various discourse particles have been extensively studied in individual languages (Haselow 2012, 2013, 2015, Degand 2014, Ruskan 2019, Šinkūnienė et al. 2020) as well as cross-linguistically (Aijmer 2019, Borthen and Karagjosova 2021) in terms of their functional diversity, scope, clause peripheries, and (inter)subjective dimension. The proliferation of studies into discourse particles and emerging trends of research in this area are connected with the high frequency of discourse particles in spoken discourse and their special functions in managing various modes of spoken interaction. Discourse particles are short, uninflected words (e.g. English well, Polish no, and Lithuanian jo), and their elimination from discourse would hamper communication between the speaker and the addressee by reducing smoothness and expressivity of interaction.

One recent aspect of the investigation of particles across languages is the positional distribution of discourse particles in a clause/utterance or turn and their cognitive as well as discursive distinction. In a volume on final particles in English, German, Norwegian, and other European as well as Southeast Asian languages (Hancil et al. 2015), the final position of discourse particles in an utterance/turn has been given special attention by illustrating its functional and cognitive distinction from other positions. In this position, discourse particles perform a variety of functions, such as strengthening the illocutionary force of an utterance, linking to the preceding discourse, or performing a turn-giving function (Haselow 2012, 2013). Exploring the class of final particles in English, Haselow (2013) stresses the importance of identifying the host of a final particle. Drawing on Conversation Analysis, the scholar refers to a turn-constructional unit (TCU) as a host of final particles, defined as ‟any stretch of speech that represents an independent contribution or unit of talk” (Haselow 2013, 389). Since final particles as well as other types of discourse particles occur mostly in spoken register, this type of host allows us to capture ‟structures that naturally arise in on-line speech production, but which do not represent well-formed syntactic units” (Haselow 2013, 388–9).

The growing number of studies into final particles and insights into their cognitive and functional properties has brought to the fore a number of studies into turn-initial particles (Heritage and Sorjonen 2018). As claimed by these authors (2018, 2–3), discourse particles, including also those in turn-initial position are used as “tools for constructing actions that are housed in turns at talk, and located in sequences of emerging interaction”. Across languages, thoroughly explored turn-initial particles are English well and oh (Heritage 2015, 2018), Spanish bueno and pues (Raymond 2018), French voilà (Mondada 2018), Estonian no (Keevallik 2018), Polish no (Weidner 2018), Russian nu (Bolden 2018), as well as Finnish siis, eli(kkä) and nii(n) et(tä) (Sorjonen 2018). The cross-linguistic inventory of turn-initial particles includes response particles as well as those from other semantic classes (e.g. demonstrative or presentative voilà in French and expressive oh in English). Heritage and Sorjonen (2018, 12) note that “most particles, including those in turn-initial position, have general semantic and procedural meanings that are broad, variable and capable of local particularization.”

Among the most important properties of turn-initial particles is their relationship to both a previous turn and an emerging turn, which can result in the backward and/or forward-looking orientation of a particle. For example, English oh is characterised by a backward-looking orientation, as it usually marks a reaction to a previous turn, whereas well ‟in projecting a departure from the expectations established in a preceding turn, is inherently forward looking” (Heritage 2018, 182). Turn-initial Danish åltsa is both backward and forward-looking, as it elaborates or explains some part of the preceding turn and simultaneously ‘projects the expansion’ of the emerging turn (Heinemann and Steensig 2018, 446). According to Heritage (2015, 89), the movement between the turns that the initial particle indicates can be unmarked or marked: ‟In ‘unmarked’ movement, next turns are congruent with the understandings, expectations and projections that were established in the previous turn, or sequence of turns. In ‘marked’ movement, there are departures from some of these understandings, expectations and projections (Heritage 2013). Turn-initial particles are often implicated in these departures <…>.” The distinction between the unmarked and marked movement between the turns could be illustrated in the following examples:

In (1), Speaker B accepts Speaker’s A invitation to go to the movies by responding with the conventional marker yes. Thus, the expected reply from Speaker B shows the congruency between the two turns. By contrast, in (2), the turn-initial particle well, used to introduce a negative response, marks that Speaker’s B answer does not align with Speaker’s A expectations. This creates an incongruence between the two turns of the conversation and demonstrates the marked movement between the turns in which a turn-initial particle plays an important role.

As a number of studies in the volume on turn-initial particles across languages show turn-initial particles display a variety of marked movements between the turns. As illustrated in (2), the particle well is used in ‘dispreferred’ (Heritage 2018, 176) or ‘expanded’ responses, which highlight “the fact that the respondents, in departing from one or another of the constraints on their actions set by the questions, are frequently privileging their own concerns, agendas and projects over those of their questioners” (ibid. 179). The turn-initial particle no in Polish indexes some incongruence between the speaker’s epistemic stance and the addressee’s knowledge communicated in the preceding turn (Weidner 2018). The turn-initial particle åltsa in Danish (Heinemann and Steensig 2018) marks departures from the smooth transition between the turns, i.e. progressivity, by explaining, elaborating, or specifying some information in the preceding turn so that further communication and interaction could continue. It must be noted that functional variation of turn-initial particles depends on their sequential positioning, which can be first position, second, and third positions (Heritage and Sorjonen 2018, 9).

In Lithuanian, turn-initial particles have been analysed in individual studies that deal with the general functional profile of the particles, but not necessarily at the beginning of a turn. For instance, the response particles na, nu ‘well’ (Šolienė 2020) and the demonstrative particles tai ‘so’ and va ‘here/there’ (Ruskan 2019, Šinkūnienė et al. 2020, 24), which may occur turn-initially, were explored in terms of their functional variation or translation correspondences. However, there are no studies exploring specifically discourse particles in turn-initial position in a talk and their relationship to a previous and emerging turn. Moreover, in cross-linguistic studies as well as in Lithuanian investigations, discourse particles have been considered in interactional contexts in spoken discourse, whereas their role in interactional contexts in fiction has not been given enough attention. In their latest study into intensification in fiction, Ebeling and Hasselgård (2020, 313) claim that fiction represents hybrid nature and in order to properly study this discourse, a distinction should be drawn between dialogue and narrative. Since turn-initial particles are mostly found in spoken interactions, it is assumed that in fiction they most likely occur in the literary dialogue.

The present study focuses on the demonstrative particles tai, va, and štai and the emotive particles o, ogi, and nagi and their potential to occur in turn-initial position, as shown below:

In the context above, turn-initial tai ‘so’ expresses the speaker’s approval of the addressee’s explanation that for the renovation of the parking lot it is necessary to pay. The representatives of both demonstrative and emotive discourse particles are taken into consideration, since the demonstrative particles may become close in meaning to emotive particles. In the Lithuanian Grammar (Ambrazas 2006, 434), the particles va and štai belong to the class of demonstrative particles, the function of which is to point to persons, objects, and phenomena. But the two particles also occur in the group of intensifying-emphatic particles, alongside gi, jau, kad, and net (Valeckienė 1988, 194, Ambrazas 2006, 436). As noted by Ambrazas (2006, 436), intensifying-emphatic particles bear resemblance to emotive particles and interjections. The class of emotive particles includes juk, be, ogi, vai, nagi, o, and ei (Valeckienė 1988, Ambrazas 2006). Thus, it is important to study the emphatic contexts of the use of the demonstrative particles and their relationship to emotive particles and identify those particles that occur most frequently in turn-initial position in spoken discourse and literary dialogue.

The aim of the present study is to identify the functions of the demonstrative turn-initial particles tai, va, štai and the emotive particles o, nagi, and ogi in spoken discourse, namely in spontaneous private conversations and literary dialogues in fiction, by exploring the relationship of the particles to a previous and emerging turn. Moreover, the study aims to disclose the distribution of the particles used in turn-initial positions in spoken discourse and literary dialogue and establish whether the two discourses reveal any differences in the use of the turn-initial particles.

2 Data and methods

The data for this study have been retrieved from the Corpus of Spoken Lithuanian (more than 320,000 words, of which 225,000 words are morphologically annotated), which contains subcorpora of spontaneous and prepared speeches. The former include informal conversations among family members, friends, as well as institutional spontaneous conversations taking place in public places (shop, bank, hairdresser, pharmacy, market, etc.); the latter include speeches delivered in academic, theological, and mass media communication (Kamandulytė-Merfeldienė 2017). The present study is based on the data obtained from the subcorpus of spontaneous private conversations (121,788 words), which display natural interactions between interlocutors and allow us to identify functional variation of discourse particles in turn-initial position. The morphological annotation of this subcorpus as well as contextual information provided facilitate the identification of turn-initial particles and their functions. For the sake of illustration of the contexts of use of the turn-initial particles in spoken interactions, a number of examples have also been obtained from the subcorpus of spoken language (447,396 words) in the Corpus of the Contemporary Lithuanian Language (CCLL). The data for the study of the particles in fiction have been retrieved from the subcorpus of fiction (15,765,554 words) in the CCLL.

At first, the overall frequencies of tai, va, štai, o, nagi, and ogi (both raw and normalised per 10,000 words) in the corpora are presented. It should be noted that in the subcorpus of fiction, which is a part of the non-annotated Corpus of the Cotemporary Lithuanian Language, the frequencies of tai and o include their use as conjunctions or other markers^[1] alongside their particle use. To establish the proportions of the particle use of tai and o in the subcorpus of fiction, 100 random examples of each marker have been extracted and the percentage of their particle use has been counted. After the presentation of the overall frequencies of tai, va, štai, o, nagi, and ogi in the corpora, the percentage of turn-initial use of the particles is provided and discussed. For the calculation of the turn-initial occurrences of tai, va, štai, o, nagi, and ogi, 100 random examples of each marker in each discourse type (spontaneous private conversations and fiction), if available, have been selected. The presentation of the quantitative findings is followed by a functional analysis of the turn-initial particles in spoken Lithuanian and fiction.

In spoken discourse, where ‘speakers generally take turns’ (Denison 2020, 210), the particles are considered to occur in turn-initial position when they occur at the beginning of a TCU and are used in response to a previous turn, as in example (3) illustrated in Section 1. In fiction, the particles are considered to occur at the beginning of a turn in a literary dialogue, where two or more characters are having a conversation and taking turns, as in the following example:^[2]

Cases when the particles co-occur with other discourse particles or other extra-clausal units were considered as instances of turn-initial use as well (5); however, stand-alone cases of the particles were discarded (6):

In (5), the particle va is preceded by the response particle nu and emotive ai. However, the presence of the two particles does not affect the function of va. In (6), va functions as a marker of agreement but it does not preface any turn and was left out from further analysis.

3 Frequencies of the particles and their turn-initial distribution in the corpora

The overall distribution of tai, va, štai, o, nagi, and ogi in the corpora (Table 1) shows that the most frequent particles in spontaneous private conversations are demonstratives tai, va and the emotive particle o, which proves their interactional nature and tendency to occur in informal conversations. In fiction, the most common occurrences are also displayed by o and tai; however, as was mentioned in Section 2, the CCLL, part of which is the subcorpus of fiction, is not annotated, and the frequencies of o and tai alongside their particle use include other uses. In fact, the proportions of the particle use of tai and o (Table 2) show that in fiction these markers are used less frequently as particles than in spoken discourse. In spontaneous private conversations, the particle tai constitutes 76% of the overall use of this marker, whereas in fiction it accounts for only 16%. Similarly, the particle use of o in spoken discourse comprises 64% of the total, while in fiction it makes up 28%.

These distributional differences of tai and o in the two corpora could be explained by the heterogeneous nature of the subcorpus of fiction, which includes monologues as well as dialogues that contain other means of coding expressivity and discourse management than discourse particles. Moreover, as illustrated by Jucker (2021) in his study into the features of orality in performed fiction (plays, TV series, and films), written fiction (novels), and face-to-face conversations in English, features associated with spoken language, such as discourse markers, response markers, hesitators, are much less frequent in written fiction, as evidenced by the data from narrative fiction in the Corpus of Contemporary American English than in spontaneous spoken discourse, as illustrated by the data from the Santa Barbara Corpus of Spoken American English. Even in performed fiction, as shown by the examples from the Corpus of American Soap Operas, TV Corpus, Movies Corpus, and Sydney Corpus of Television Dialogue, elements typically found in spoken discourse are less frequent than in spontaneous conversational exchanges. Thus, the differences between the frequencies of the particles in the two corpora examined in Lithuanian could be ascribed to the ‘scripted language’ of fiction, which is devoid of spontaneity and the online process of planning one’s speech (Jucker 2021, 342).

The particles occurring predominantly in fiction but not in spoken discourse are the emotive particles nagi and ogi and the demonstrative particle štai, which sheds light on their less interactional character and more particularised meanings as well as on discourse preferences of individual particles. The opposite situation is held for the particle va, which is among the most frequent particles in spontaneous private conversations but occurs infrequently in fiction. The scarcity of va in fiction stresses its interactional nature that could be attested only in spontaneous conversations involving natural interactions between interlocutors. The overall distributional differences of the particles show that spoken discourse and fiction contain a different set of particles to manage interactions and/or create expressivity.

The analysis of 100 random examples of tai, va, and o in the subcorpus of spontaneous private conversations shows that the most frequent turn-initial particles are o and tai, followed by the demonstrative va (Table 3). The fact that o and tai show a preference for turn-initial particle use in spoken discourse clearly correlates with their overall higher frequencies as particles in spoken interactions, as illustrated in Tables 1 and 2. The higher frequencies of turn-initial o and tai may also be determined by their ideational meanings (i.e. o marks contrast, and tai denotes result), which show a relationship between a previous and following clause. Thus, in spoken interactions, these particles also display a tendency to occur at the beginning of a turn and mark a relationship to a previous turn. The particle va, however, is more mobile and less constrained by the initial position. It may appear within a phrase as well as in the left and right periphery of the clause/utterance (Ruskan 2019). The greater mobility of the particle va may correlate with its fewer occurrences in turn-initial position.

In fiction, the demonstrative particles tai and va as well as emotive o reveal far fewer occurrences in turn-initial position, which confirms their expected distributional differences in fictional language (dialogues in narratives) and spontaneous conversations. The scarcity of the turn-initial particles o, tai, and va in fiction confirms the less interactional nature of literary dialogues and their scripted nature, as discussed in the study by Jucker (2021). The less frequent or even non-occurring particles in spoken discourse, such as štai, nagi, and ogi, can be used turn initially in fictional dialogues, which suggests a different character of interactions in fiction, as compared to conversational exchanges in spoken discourse. The most frequent turn-initial particles used by fictional characters for emphasis and expressivity are nagi and ogi. Alongside tai, the demonstrative particle štai is among the least frequent turn-initial particles. As shown in the functional analysis presented in Section 4, štai is used by the speaker for the purposes of illustration and explanation but not as an interactive device in dialogue.

Thus, the distribution of turn-initial tai, va, štai, o, nagi, and ogi in fiction and spontaneous private conversations highlights differences in conversational exchanges in fictional language and spoken discourse. Ample evidence of the use of turn-initial particles in spontaneous private conversations could provide clues to those Lithuanian writers who aim to imitate spoken discourse in their narratives or scripts.

4 Functional variation of the particles

This section discusses the functional variation of the turn-initial particles in spontaneous private conversations and fiction (literary dialogues). First, functions of the demonstrative particles tai, va, and štai in turn-initial position are presented; then, functions of the emotive particles o, nagi, and ogi are illustrated. A summary of the functions of turn-initial particles is provided in Section 4.6.

5 Conclusions

The present study sheds light on the inventory and role of turn-initial particles in spoken Lithuanian, namely in spontaneous private conversations, and in fiction, and thus extends the scope of studies dealing with discourse particles in Lithuanian (Petit 2010, Panov 2019, Šolienė 2020, Šinkūnienė et al. 2020) as well as investigations of turn-initial particles across languages (Heritage and Sorjonen 2018). The study unfolds how the beginning of a turn in spoken interactions and literary dialogue determines the functional variation of the demonstrative particles tai, va, štai, and the emotive particles o, nagi, and ogi.

The analysis of the quantitative and qualitative parameters of the demonstrative and emotive particles shows that turn-initial position in spoken interactions is occupied mostly by the particles o, tai, and va, whereas in a literary dialogue, the emotive particles nagi and ogi are used most frequently turn-initially. The demonstrative particle štai occurs turn-initially rather rarely, as it more frequently occurs as a device for illustration and explanation within the speaker’s turn rather than an interactive particle across the turns. The distribution of the turn-initial particles in the two discourses (spoken language and fiction) proves that the inventory of turn-initial particles in spoken and literary interactions differs. These findings are also in line with Jucker’s study (2021), which shows significant differences in the distribution of discourse markers, interjections, and other elements typical of spoken language in spontaneous conversational exchanges and written fiction in English.

Used turn-initially, the particles tai and va display various responses to a previous turn and frame an upcoming turn. The particle tai may express the speaker’s conclusion on the basis of the arguments in the previous turn, agreement or disagreement with the addressee, willingness to help the addressee or specification of a previous turn. The turn-initial demonstrative particle va performs an attention-getting function and expresses shared experience or disagreement with the addressee. However, unlike the demonstrative particle tai, in individual cases it breaks the continuity of conversation and starts a new topic, highlighting the unexpectedness of the emerging turn. The meaning of emphasis ascribed to the demonstrative particles in the Lithuanian Grammar becomes salient in the context of social interaction that naturally marks agreement and disagreement between the interlocutors and outpour of various emotional responses (dissatisfaction, surprise), which brings these particles close to interjections expressing various emotions. The overlap of discourse structuring and expressive functions of tai and va reveals the multifunctional profile of the two particles. Their high overall frequencies in spoken interactions also throw light on their potential to function as fillers that facilitate the on-going talk in conversation, like particles nu, na, and reiškia (Šinkūnienė et al. 2020, Šolienė 2020).

The most frequent turn-initial particle o also reveals a multifunctional profile in spontaneous private conversations by performing discourse structuring and expressive functions. One of its dominant uses is prefacing a question with the intention of eliciting an answer from the addressee and showing the speaker’s genuine interest in the details of the answer. Like the particle va, o may start abruptly a new topic in the conversation, thus disrupting the progressivity of the emerging turn. However, it should be noted that unlike the particles tai and va, the particle o is less speaker-oriented and more addressee-oriented, since it does not focus on the speaker’s realisations or expressions of emotions but more on eliciting a response from the addressee. In fiction, the emotive particles nagi and ogi tend to occur in directives or emphatic responses to questions and focus on the speaker’s emotional state of dissatisfaction or irritation. They are not as multifunctional as the demonstrative particles or the emotive particle o. It should be noted that the multifunctionality of the turn-initial particles correlates with their overall frequencies and broad general semantic meaning, as illustrated by tai, o, and va.

The limitations of the study pertain to the small number of sequences of turns available in the Corpus of Spoken Lithuanian, which does not allow for a more comprehensive contextual analysis of the particles. However, the contextual information available makes it possible to identify the most prototypical functions of the particles in turn-initial position, determined by the interactional context of conversation in spoken discourse or literary dialogue.