1. Introduction
This paper focuses on two particles that are found in Icelandic conversation: the domestic form ha and the English borrowing what. In Íslensk orðabók ‘Dictionary of Icelandic’ (2007), ha is categorized as an interjection that has four different meanings or functions: (i) a request that the speakers repeat their words, what did you say? (ii) to express surprise, indignation, or doubt, (iii) a request for confirmation, and (iv) to ask for attention or understanding, especially in children’s language. The oldest documented use in written language can be traced back to the eighteenth century (The Written Language Archive). The borrowing what, on the other hand, is originally a question pronoun in English, but is not used as such in Icelandic. Instead, it is used as a response in conversation, marking the information in the prior turn as unusual or unexpected (see Gisladottir Reference Gisladottir2015:322). The oldest attested use in written Icelandic is from around 2004, when the borrowing started to appear in blogs, discussion threads, and on social media (see Icelandic Giga Corpus). By then, what had most likely become established in informal conversation as a part of youth slang vocabulary.
In some environments, ha and what are considered to be interchangeable, or at least to serve very similar functions. They could, so to speak, be viewed as two sides of the same coin. For example, according to Gisladottir (Reference Gisladottir2015:322), both items display surprise. The similarities between the two particles are also suggested in bilingual dictionaries: ha is explained in English as ‘what?’ or ‘pardon?’ (see Icelandic–English Dictionary 1990). However, as pointed out by Gisladottir (Reference Gisladottir2015:323), what is more limited in use and cannot be treated as a repair initiator as it is in English (see Schegloff Reference Schegloff1997, Thompson, Fox & Couper-Kuhlen Reference Thompson, Fox and Couper-Kuhlen2015:65). In addition, in her data, what is delivered with a rising intonation (what↑) while ha has a falling contour (ha↓) (see also Dehé Reference Dehé2015).Footnote 1
Gisladottir’s study of ha and what is based on 5.5 hours of conversation that were recorded in 2011 and 2012 and included 51 instances of ha as a repair initiator and 3 of what (Gisladottir Reference Gisladottir2015:310–313). From the time of recording, the use of what has most likely changed. For this reason, a new study of more expansive and more recent data might give a more detailed picture of the interplay between the two particles.
The aim of this study is to provide a sequential analysis of two lexical items that are used in Icelandic conversation: the native form ha and the English borrowing what. The analysis will shed some light on the process of borrowing non-referential words from one language to another and on the influence of the world language English on discourse practices in another language. The research questions are as follows: (1) What are the similarities and differences between the two items in terms of turn-position, sequential placement, and function?, (2) What can the comparison tell us about the borrowing of discourse particles from one language to another?
The paper is structured as follows. In Section 2, I will discuss particles as a category: their formal features and functions, and what this means for their integration into new languages. In Section 3, data and methods are presented, and in Section 4 I will present the analysis. The paper ends with a summary and discussion.
2. Domestic and borrowed particles
As many studies have shown (e.g. Friðriksson & Angantýsson Reference Friðriksson and Angantýsson2021, Jónsdóttir Reference Jónsdóttir2021, Jónsdóttir & Hilmisdóttir 2021, Hilmisdóttir Reference Hilmisdóttir, Tarsi, Mårtensson and Williams2023, Reference Hilmisdóttir2024), English has a major influence on the vocabulary in contemporary Icelandic, particularly the informal language of the younger generations. A considerable portion of the vocabulary used in informal conversations amongst peers, empirical data suggest (e.g. Hilmisdóttir 2023:54–57, Reference Hilmisdóttir2024), can be characterized as pragmatic borrowing. The notion of pragmatic borrowings is defined by Andersen (Reference Andersen2014:17–18) in the following words: ‘items that do not contribute to the propositional content of utterances, but act as constraints on the interpretation process due to their subjective, textual, and interpersonal pragmatic functions’. These borrowings include categories such as address markers (e.g. dude), greetings and farewells (e.g. love ya), swearwords (e.g. shit), politeness markers (e.g. please), hortatives (e.g. lets go), and, finally, particles (e.g. what), which is the topic of this paper.
Particles constitute a class of lexical items that are difficult to define as a category. They tend to have ‘little or no propositional context-independent meaning and serve a variety of meta-discursive functions’ (Bolden Reference Bolden, Robinson, Clift, Kendrick and Wesley Raymond2024:613). Their main function is to show relationships between units of talk and to negotiate the status of knowledge. Their function is shaped by their position in a turn, sequential placement and the prosodic packaging (e.g. Hilmisdóttir Reference Hilmisdóttir, Auer and Maschler2016a).
As studies of particles have shown, they comprise a complicated system. Each particle, or even a cluster of closely related particles, for example responses to information, have acquired their functions through the centuries. They have been shaped and molded by speakers engaged in social interaction. This raises the question of what happens when a particle is borrowed from its natural environment, a source language such as English, and used in a new context, in an Icelandic conversation. Is it even possible to borrow such a complicated phenomenon in its complete form, with all potential functions and connotations? What do the speakers of the receiving language take with them from English and what do they leave out? And what does this mean for the interplay between a borrowing and a domestic item that has a much longer history in the receiving language?
To borrow a pragmatic or discursive item is not a question of a simple transition from one language to another, and, often, a certain adaptation takes place in the receiving language (Hilmisdóttir & Peterson Reference Hilmisdóttir, Peterson, Peterson and Beers Fägersten2024:98). As pointed out by Andersen (Reference Andersen2014:18), pragmatic borrowings can be expected to go through semantic shifts regarding ‘discourse functions and associated speakers’ attitudes’. Therefore, in studies focusing on such borrowings, the main objective should be to ask whether and to what degree specific occurrences are adapted to the target language (see Peterson Reference Peterson2017:122). Does the borrowed item show the same speaker attitude in the source and receiving language? And does the item in question have the same range of functions in both languages?
In this paper, the intent is not to provide a comparison between the function of a particle in source and receiving languages. In other words, the aim is not to investigate whether what is used in the same environment in American or British English and Icelandic. Instead, the focus is language-internal. The aim is to conduct a sequential comparison between a domestic (ha) and a borrowed (what) word that are considered to serve similar functions, and to describe a potential distribution of labor between the two items as it appears in the present data.
Here, it is also worth bearing in mind that when a particle or a discourse structuring device is borrowed into a receiving language, it may not only affect a comparable domestic item. As Betz & Sorjonen (Reference Betz, Sorjonen, Emma Betz, Mondada and Sorjonen2021:27–28) argue, the new addition may also initiate a process of reorganization of a system, for example a system of responses. Hence, with what gaining ground as a resource to treat new information as new and perhaps unexpected, it may not only affect the way speakers use ha but also other particles that are used as newsmarks or change-of-state tokens, such as nú ‘oh’ and er það ‘really’ (Hilmisdóttir Reference Hilmisdóttir2016b; see also Heritage Reference Heritage, Maxwell Atkinson and Heritage1984 on change-of-state tokens in English). However, due to the limited scope of the study, such questions will be left unanswered.
3. Data and methods
The paper draws data from a range of interactions between Icelanders between the ages of 13 and 26. The conversations were recorded and transcribed for the research project Icelandic youth language: An empirical study of interactional resources, which received a three-year grant from the Icelandic Research Center (2018–2020). The project was reviewed and approved by the ethics committee at the University of Iceland and registered at The Data Protection Authority. All participants and guardians of minors signed an informed consent before the recording of conversations. The participants in the podcasts gave their written consent for use of the data for research purposes.
Table 1 shows an overview of the data. As the table shows, the data represent different types of interactions: everyday conversations, task-oriented conversation, semi-moderated interviews, and podcasts.Footnote 2 The aim is to include as much recently recorded material as possible, which explains the discrepancy in the various strands of data. However, the study is first and foremost a qualitative exploration, and the point is not to compare the use of ha and what in different genres of conversation.
Information about the data
The everyday conversations (IU-hve) are represented by two different recordings. Each conversation has three participants (mixed groups). Both conversations were recorded after school, in a home setting.
The group discussions (IU-sko) were recorded in eight grammar schools (13y) and eight upper secondary schools (18y) in four areas in Iceland, in total sixteen recorded conversations. The number of instances produced by the moderator is shown in parenthesis. Each group consists of three or four participants in mixed groups. The discussion is moderated by a researcher and loosely follows a set of questions that have been prepared in advance. The questions revolve around topics such as favorite school subjects, hobbies, plans for the future, and taste in music. The database is accessible to users who have access to CLARIN.Footnote 3
The videogame conversations (IU-tol) consist of five recorded conversations. Each conversation has two participants. In one conversation, two 13-year-old boys are sitting side-by-side in front of a television screen while playing an action game. In the remaining four conversations, two 15-year-old boys are playing computer games online. They are not located in the same room but meet via virtual reality and communicate through a headset.
Finally, the podcast data (IU-pod) consisted of two different recordings: an interview with two women aged about 20 regarding their private life, and a conversation between two women in their mid-20s about fashion, lifestyle, and entertainment. The podcasts are private productions and not part of a public service or backed up by broadcasting companies in Iceland. The interlocutors in the podcasts were all well-known individuals in Icelandic youth culture at the time of the recording (see also Hilmisdóttir Reference Hilmisdóttir2024).
As seen in Table 1, there are 172 occurrences of ha and 72 of what in the data. The occurrences are distributed unevenly in the recordings. However, the aim of this study is not to compare frequency of use in different settings and by different speakers. Instead, the focus is on a sequential analysis of cases that cast a light on the differences between ha and what, in what kind of environments they occur, and how they are treated by the coparticipants. The methodology employed in this study is CA, in particular the linguistic approach which has been referred to as interactional linguistics (e.g. Couper-Kuhlen & Selting Reference Couper-Kuhlen and Selting2017, Deppermann & Mondada Reference Deppermann, Mondada, Emma Betz, Mondada and Sorjonen2021:34–38). As pointed out by Deppermann & Mondada (Reference Deppermann, Mondada, Emma Betz, Mondada and Sorjonen2021:34), the main concern of CA is to ‘identify practices that members use to perform social actions through which knowledge is conveyed and negotiated and through which interaction is organized and relationships managed’. One way of approaching these issues is to focus on a certain form, or in this case two different forms that are considered functionally comparable in two different languages.
Comparison is the basis for all studies conducted within the interactional linguistics framework. This refers to a comparison not only between different forms, such as ha and what, but also between each occurrence of a particular form. Occurrences that accord with common formal and sequential patterns are referred to as ‘collections’ (see Schegloff Reference Schegloff and Sidnell2009, Clayman Reference Clayman, Robinson, Clift, Kendrick and Wesley Raymond2024). In the following section, I will present such categories for both forms: ha and what.
4. A sequential analysis of ha and what in the data
In total, there are 172 instances of ha in the data and 72 of what. The occurrences were divided into three main categories: (i) open-class repair initiator, (ii) responses to informings in the prior turn (i.e. after speaker-shifts), and (iii) instances in continuations or turn-completions (i.e. no speaker-shift).
Table 2 shows the distribution between the three categories. The two forms, ha and what, show different tendencies. While the majority of cases of ha (63%) are used as open-class repair initiators, there are no such cases of what in the data. Similarly, only ha occurs in restricted repairs (3%). In contrast, most cases of what (97%) occur as responses to events or information in a prior turn. Finally, a small category occurs as continuations or in turn-completions without speaker-change (7% and 3% respectively).
Distribution of ha and what in the data
In the following sections, I discuss each category by showing examples of use from the data.
4.1 In open-class other-repairs
As shown in Gisladottir (Reference Gisladottir2015:314) the interjection ha is the most common open-class repair initiator in Icelandic. In the present data, 63% of all ha are found in this environment.
In an open-class other-repair, a speaker initiates a repair by signaling that there is a problem in the prior talk without specifying it any further (see Dingemanse & Enfield Reference Dingemanse and Enfield2015:105). An example of this can be seen in (1). In this conversation, two 15-year-old friends, Svenni and Bogi, are playing video games online. They are not located in the same room but communicate through their computers. Bogi has Svenni’s permission to record the game. As the boys prepare to end the conversation, Svenni asks Bogi to stop the recording (line 1).
After a 1.9 second pause, Bogi responds to Svenni’s request with an instance of ha. The particle is delivered with a slightly prolonged vowel and a falling intonation contour, which is the typical format for Icelandic repair initiators (see Dehé Reference Dehé2015:199, Gisladottir Reference Gisladottir2015). Svenni treats this as an open-class repair initiator and responds by offering a repair: a partial repeat with a slight modification (see Gisladottir Reference Gisladottir2015:314). The repeat is produced with laughter in the voice. Again, after a long pause, Bogi accepts the repair and closes the sequence. The turn begins with a change-of-state token, ah, which indicates that he now remembers something he did not remember before (see Heritage Reference Heritage, Maxwell Atkinson and Heritage1984), followed by ókei já ‘okay yes’, which indicates that he accepts the requests. This turn marks the end of the recording.
In some cases, ha responds to problems that are caused by other issues than hearing and sometimes there is even a certain ambiguity regarding the nature of the trouble-source. The repair initiator may be a way to deal with lack of understanding or even disagreement in a strategic way (i.e. by using a pro-forma repair initiator). In (2), which is drawn from a semi-moderated group discussion, three students in an elementary school are telling the moderator their plans for the future. Before the excerpt begins, Regína has told the other participants that she does not know what she wants to study. Then, she announces what she wants to be as an adult (line 1).
Regína’s announcement in line 1 stands in contradiction to her previous statement that she has not made up her mind about her future vocation. Valgerður responds by pointing this out: that despite what she has said earlier she has made up her mind (line 3). Then, Regína produces an instance of ha, delivered with a falling intonation contour (similar to the instance in (1)). Valgerður treats this instance as an open repair initiator, and partially repeats the prior turn. However, this time, she excludes the discourse particle nú and, perhaps due to an overlap by Regína, cuts off before she reaches a transition relevance place. In the overlap, Regína agrees that she has made up her mind. However, she has obvious difficulties with the formulation, which indicate that the answer is not straightforward and that there is a need for further explanation (line 7). Then, Valgerður takes the floor and produces a candidate understanding (line 8). Regína accepts the suggestion, and, finally, the moderator closes the sequence with the phrase ég skil ‘I understand’ (line 10).
In this excerpt, ha is treated by the interlocutor as an open-class repair initiator requesting a repeat of prior talk. However, as the conversation unfolds, it becomes clear that the problem the ha-speaker is having is not caused by audial problems. Instead, ha suggests problems with understanding or accepting the claim that is being made.
As Table 2 shows, open-class repair is the largest category for ha with 63% of all occurrences. In contrast, there are no occurrences in the present data which suggest that what can serve the same function. Thus, the study supports the findings made by Gisladottir (Reference Gisladottir2015:323) that this activity, to initiate an open repair, seems to be reserved for the native form ha.
4.2 Restricted repair initiators
Sometimes ha occurs in a turn combined with partial repetition of the troublesome turn, as pointed out by Dehé (Reference Dehé2015:193). By repeating words from a prior turn, the speaker frames them as a trouble-source (Dingemanse & Enfield Reference Dingemanse and Enfield2015:102). As seen in Table 2, there are five instances of ha as a restricted repair initiator in the data (3%) but no instances of what. In these cases, ha occurs turn-initially along with a partial repeat of the prior utterance.
The excerpt in (3) shows an example of ha that is used as a restricted repair initiator. The excerpt is drawn from a conversation between two 13-year-old boys who are playing a video game during a break in the school schedule. They are sitting side-by-side in front of a television screen. At the beginning of the excerpt, Kalli makes a request formed as a directive (line 1).
Siggi responds with ha and repeats the last word in the turn, the preposition úr ‘out of’ (line 2). Simultaneously, he moves his character across the screen. His action represents a candidate understanding of Kalli’s request. In the following turn, Kalli confirms that Siggi has made the right interpretation (line 3).
In (3), ha is used as a repair initiator that questions the content of the prior turn. The repetition brings the problematic lexical item into focus. Such use seems also to be restricted to the domestic form ha.
4.3 In responses to informings
As seen in Table 2, 27% of the occurrences of ha and 97% of what are categorized as responses to an informing in the prior turn. It is in this environment that ha and what have been interpreted as interchangeable.
The following excerpt is drawn from a face-to-face, everyday conversation between three teenagers living in the capital region. Pétur, who is the youngest, is still at school in the neighborhood, while his brother Hjálmar and Guðrún attend an upper secondary school downtown. In (4), Pétur is telling the others about his problems with the teacher. He has received two ‘birds’, which is some kind of warning system or punishment that the other two are familiar with (line 1).
Halldór treats the information in Pétur’s turn as newsworthy, which is indexed by the response erþa:↓ (see Hilmisdóttir Reference Hilmisdóttir2016b). Receiving two ‘birds’ in one week is, as the three friends discuss later in the conversation, quite unusual. Pétur then continues and upgrades his story: he also received two ‘stop tickets’ and was sent to the principal (lines 3–4). This new piece of information is followed by a brief pause and an instance of ha, produced with a prolonged vowel and falling intonation, which can be described as ‘punched-up prosody’ (Wilkinson & Kizinger Reference Wilkinson and Kizinger2006).
In this case, ha has the function of registering an informing as newsworthy and unexpected. As can be seen, the two participants who are being informed treat this news as something that is potentially very sensitive or could be a source of further problems. For example, when Halldór asks whether their father knows about Pétur’s issues at school (line 7), and when he receives a negative answer, he responds with a swearword uttered with a soft voice (line 10). These details indicate that the stakes are high for Pétur and that the information he has just shared is not to be taken lightly. Thus, in this case, ha is used as a strong response. It marks the information in the prior turn as unexpected and worthy of attention and discussion. Potentially, there is more to say about the topic and Pétur is the only one that could provide more information. By using ha, the speaker shows interest and concern for the interlocutor.
In (4), ha is used as a stand-alone particle, but it can also occur turn-initially prefacing a question. In (5), which is extracted from a podcast hosted by two young women, Arna and Unnur are discussing an art performance that took place fifteen years earlier and had a huge impact on them. Unnur points out that the reason they remember the performance so well is that they were children at the time (line 1–2).
Arna confirms that the two women were children at the time by placing the event in time: this was in 2007 (line 3). Note the lengthy pause after the informing in line 3, to which Unnur responds by partially repeating Arna’s turn, prefacing the turn with an instance of ha. The particle is delivered with a prolonged vowel, emphasis, and falling intonation. Furthermore, ha is prosodically integrated with the remainder of the turn, without any pause.
In (5), ha expresses a clash of expectations regarding the information that Arna provided. The turn-initial particle is followed by a question which is almost a verbatim repeat of Arna’s prior turn. It is designed as a request for confirmation. The particle and the question are resources to show the speaker’s attitude towards the new information. Unnur does more than just acknowledging the informing provided by Arna: she shows that she thinks it deserves extra attention. In other words, in this case, ha does not convey doubt or uncertainty. Unnur receives the information without questioning Arna’s memory, which can for example be seen by her use of vá ‘wow’ with laughter in the voice (line 7).
This, however, is not the case in (6), which is extracted from the same podcast. In this excerpt, Arna is looking through the photos on her phone to see if she can find out what she was doing on her birthday five years ago, when she was 21. However, she has difficulty finding photos from that year (lines 1–2).
Arna shows a photo from 2016 and both women laugh and express their satisfaction with the way they look (lines 6 and 7). Then, Arna begins a new turn which is probably a continuation of a previous story, but it is overlapped by Unnur who initiates a sequence that aims at correcting. The repair-sequence begins with an occurrence of what, followed by a polar question (lines 10–11). By its form, the question is designed to confirm that the photo was taken at Arna’s birthday in 2016. However, by prefacing the question with what, Unnur suggests a problem. First, Arna gives tentative confirmation that this was indeed taken at her birthday (line 12). In this case, the hesitation is indexed by a prolonged vowel and rising intonation, a try-marked response. Then, Unnur initiates a turn that seems to be designed to show that this cannot be the case. In an overlap, Arna withdraws her confirmation (line 14) and, finally, closes the sequence by correcting her statement: the photo was taken in September (line 18).
In (6), what indicates that there is a factual error in the prior talk. However, this time the statement that needs to be corrected is not in the prior turn, but further back in the conversation (line 5). The problem that Unnur is responding to is that she knows the timing of the photo Arna is showing does not make sense and that Arna’s turn is false (see line 13). By using what, Unnur shows non-affiliation and disalignment.
In some cases, the trouble what responds to is hidden and more difficult to spot. In (7), what is used in combination with a turn-initial oh. Here, a moderator interviews three 19-year-old students. At this point in the interview, the moderator is asking them about their taste in music. Just before the excerpt begins, the group has heard a music clip, a well-known jazz song. Sigga recognizes the song and tells the others that she knows it very well and that she hears it quite often at work. Then, the moderator asks whether they know the name of the artist (line 1).
Hólmfríður and Sigga both give a negative answer (lines 3 and 4). The turn design in Sigga’s answer indicates a complication of some sort, which could be a reference to the fact that she has already told the moderator she knows the song very well. As pointed out by Hilmisdóttir (Reference Hilmisdóttir2011), turn-final polar responses are typically preceded by some kind of explanation that addresses a problem or disagreement. Also, the creaky voice at the end of Sigga’s turn may be an indication that the negative answer is somehow dispreferred. Following the negative answers, the moderator tells the participants the right answer: the singer is Björk, a world-famous artist that the students may feel they should recognize, and especially Sigga who knows the song.
The three participants respond each in a different way. Pétur responds with the particle combination heyrðu já ‘ah, yes’ (line 7), which suggests some kind of recollection. Hólmfríður, in contrast, responds with the particle ó ‘oh’ (line 8), which indicates that the answer is somehow unexpected (see Samtalsorðabók; see also Heritage Reference Heritage, Maxwell Atkinson and Heritage1984 on change-of-state markers in English). Overlapping Hólmfríður’s turn, Sigga also responds with the combination oh what, delivered with a step-up and a level intonation (line 9). The conversation then continues to evolve around the singer and the fact that she has participated in all kinds of projects and has many different musical styles (not shown).
By responding with what, Sigga marks the moderator’s answer as newsworthy and unexpected (see also line 12). Another possible interpretation is that this response shows disaffiliation. Sigga is acknowledging an answer that she feels she should have known, which may make ha a less appropriate option. By using what, she also indexes an attitude that this is somehow strange to her; not something she would have guessed. The short vowel and the glottal stop in this excerpt indicate that the response (oh) what, which occurs frequently in the data, may originate as a cut-off version of the formulaic phrase what the fuck (see also Ameka Reference Ameka1992:111 on swearwords as interjectional phrases). This, however, is difficult to demonstrate and will not be addressed in this paper.
In some cases, what is used as a response to unexpected events. In (8), two classmates, Siggi and Kalli, are sitting in front of a television screen playing videogames. When the excerpt begins, Siggi notices a hand grenade (line 1), that explodes and turns out to cause greater destruction than he expects (line 3).
After a short pause, Kalli responds with an instance of the swearword oh shit, and, thereby, also showing that the turn of events is somehow more dramatic than he had foreseen. Overlapping Kalli’s response, Siggi utters an instance of what, produced with a level intonation. After a one second pause, he uses the exclamation ómæga:d ‘oh my god’, before he finally poses a rhetorical question (line 10).
In (10), Siggi uses what as a response to the turn of events in a computer game. By so doing, Siggi marks that the destruction is unexpected.
In this section, the focus has been on ha and what, which occur as receipts for informings and as a response to non-verbal events that are taking place on a computer screen. Although ha and what share some traits and occur in similar environments, the two forms show some differences. While ha occurs in environments in which the interlocutors agree and there is no discrepancy about the correctness of the informing, what may convey skepticism, embarrassment, disappointment, or reluctance to acknowledge and accept.
4.4 In continuations or turn-completions
As Table 2 shows, both ha and what may occur in a turn-final position or as a continuation, that is, a stand-alone particle following the same speaker’s prior turn. The excerpt in (9) shows an example of the former. In this conversation, which is drawn from an interview with 19-year-old-students in the capital region, the moderator is asking participants about the students’ workloads. Before the extract begins, Lilja has told the group that she has made an informed decision not to have a job while at school. She then elaborates on the topic and explains the reasons (lines 1–2 and 4).
Lilja pauses briefly in the middle of the turn, right before the slot in which the interlocutors would expect her to clarify what she would not have time for (line 4). After the pause, and the particle þúveist ‘y’know’, which often occurs after turn-internal pauses (see Samtalsorðabók), Lilja completes her turn in an overlap with Addi. Both speakers contribute with an explanation or a candidate explanation. Lilja gives a general description, ‘everything else I do’, while Addi mentions mental health followed by a turn-final ha::↓, produced with a prolonged vowel and a falling intonation (line 7). Finally, Lilja continues with an elaboration on the topic, explaining how tough her first year at school was (lines 8–9).
The excerpt in (9) shows an instance of a turn-final ha. Unlike the cases thus far, the particle does not constitute a response to another speaker’s turn. Instead, ha is used as a particle that pursues some kind of acceptance or confirmation of the proposition in the ongoing turn. Note also that Lilja confirms with an overlapping já ‘yes’ (line 6). By providing a candidate explanation, Addi shows interest in Lilja’s story and that he understands what she is experiencing. Thus, in this case, ha occurs in a slot in which Addi is showing empathy and connection (see Avis Reference Avis1972 on Canadian eh). As Hjartardóttir (Reference Hjartardóttir2006:45) shows, these are also the two key functions of ha as tags in Icelandic conversation.
While ha is used in alignments and turns that show empathy towards the speaker of the prior turn, what shows disalignment and disagreement. Consider (10). In this excerpt, three teenagers are discussing Danish as a school subject and their own difficulties learning the language. All three participants consider themselves to be good at English, particularly Hjálmar, who has learned the language during his school years abroad. The group agrees that teaching of Danish, and language instruction in general, should start early. In Iceland, many students start learning English in fifth grade and Danish two years later, but this varies slightly from school to school. Guðrún, who claims not to know any Danish, informs the coparticipants that she probably started studying Danish in sixth grade (lines 1–2).
Halldór responds by making the assessment that this is earlier than he had expected (line 4). After a confirmation from Guðrún (line 5) and what seems to be the beginning of an elaboration of the topic (line 7), Halldór shifts the focus to English as a school subject. He points out that some schools do not start teaching English until the fifth grade. After a pause and an acknowledgement from Guðrún (line 11), Halldór uses what delivered in a whispering voice and ends in quiet laughter (line 12). Guðrún responds to Halldór’s what with another acknowledgement token but does not contribute anything further on the subject. In the following turn, Pétur returns to the previous topic, that is, Icelandic students’ poor knowledge of Danish (lines 14–15).
In (10), what is produced as a continuation, that is, the speaker uses the particle as a follow-up to his own turn. The turn in question is a factual statement. By producing what in a continuation, Halldór treats his informing as out of the ordinary, laughable, and therefore perhaps unacceptable. In other words, it is a way for him to show his somewhat negative attitude towards the late introduction of English in Icelandic schools.
A similar instance of what is seen in (11). The excerpt is drawn from an online conversation between two 15-year-old friends playing a videogame. In the game, the two boys run a football team, and in this excerpt, they are in the middle of a match. The match is commentated in English by a computer voice (lines 1, 3, and 5).
As the ball reaches the other end of the field, Valdi remarks that the player was off-side, which should have given their team the ball (line 2). However, as the game unfolds, it becomes clear that the player was not off-side, which Valdi notes in a rush-through (line 6). He then uses an instance of what, produced with a pitch step-up, prolonged vowel and level intonation, i.e. ‘punched-up prosody’ (see Wilkinson & Kizinger Reference Wilkinson and Kizinger2006, Thompson, Fox & Couper-Kuhlen Reference Thompson, Fox and Couper-Kuhlen2015:91). After a 0.8 second pause, Bogi asks how this could happen (line 9). Then, the other team gets the ball and throws it into the field again (line 10). Finally, Valdi indexes his negative attitude and disappointment by using the phrase hvernig í fokkanum ‘how the fuck’ (line 11).
The occurrence in (11) can also be described as a non-acceptance and non-alignment. By using what, the speaker shows that the turn of events was not according to his expectations, and that he is displeased with the result.
The comparison of ha and what in continuations and turn-completions shows that the two forms tend to occur in different environments. While ha shows alignment, interest, and perhaps empathy, what signals disappointment and a somewhat negative attitude, for example to index something as questionable and not fair.
5. Summary and discussion
In this paper, I have compared two forms that occur in Icelandic conversation: ha and what. The data consisted of everyday conversations, group discussions with students in grammar schools and upper secondary schools, conversations between boys playing computer games and podcasts, in total 28 hours and 36 minutes of recorded conversation. The research question was twofold: (1) What are the similarities and differences between the two forms in terms of turn-position, sequential placement, and function?, (2) What can the comparison tell us about the borrowing of particles from one language to another?
As summarized in Table 3, although the two forms have some similarities, they are different in many ways. First, regarding the position in the turn, the domestic ha shows more variation: it can occur as a stand-alone particle as well as in a turn-initial and turn-final position. The borrowed form what, by contrast, is mostly used as a stand-alone particle.
A summary of the main characteristics of ha and what.
Second, ha and what show both similarities and differences in terms of sequential placement. As in position in the turn, ha shows more variation than what. More than half of the instances of ha, or 63%, were used as other-repair initiators, while this was completely absent in the collection containing instances of what. Around 97% of the occurrences of what were used as responses to informings while the figure for ha was only 27%. Both forms occurred as stand-alone particles in continuations or as turn-final particles. However, such cases were rare: 7% and 3% respectively.
Third, when it comes to prosody, what shows more variation than ha. While all occurrences of ha were produced with a falling intonation, and thereby confirming a previous study by Gisladottir (Reference Gisladottir2015), what was executed with two prosodic patterns: rising or level intonation. What with level intonation, which is sometimes delivered with a glottal stop, may have some connection to a cut-off form of the formulaic phrase what the fuck (i.e. what-), but this is not possible to demonstrate conclusively. Both ha and what occurred with initial pitch step-up, extra volume, and stretched vowel, which are some of the cross-linguistic characteristics of indexing surprise (Wilkinson & Kizinger Reference Wilkinson and Kizinger2006, Thompson, Fox & Couper-Kuhlen Reference Thompson, Fox and Couper-Kuhlen2015:91).
Fourth and finally, even though ha and what sometimes seem to occur in similar sequential environments, the two particles were used in slightly different ways. The domestic form tends to occur in more positive contexts. Thus, ha may be described in terms of affiliation and empathy. By using ha the speaker also shows curiosity. It has an inquisite tone and conveys a clear interest in what the interlocutor is saying. In other words, it shows orientation towards the coparticipants. The domestic form ha is also very much embedded into its context. It is a response to something, and it gives the floor or attention back to the interlocutor. The borrowed form what behaves somewhat differently. It is often used in contexts which could be described as non-affiliative and disaligning. Sometimes, the particle may index a negative attitude or a reluctance to accept the unfolding of events. What is also less embedded into its context. For example, the boys playing videogames often use what in response to unfortunate events on the computer screen. Such occurrences may be surrounded by long pauses both before and after.
To summarize the results of the study, it seems quite clear that the two forms, ha and what, are not completely identical. They have clear differences in terms of the position in the turn, sequential placement, prosodic delivery and functional characteristics. The differences can be explained by the different origins. While the central and most common function for ha is to initiate repairs or direct attention towards the interlocutor, what may have spread into informal Icelandic youth language through social media, video clips, and gaming. What is still a very recent addition to informal conversational language, and, in the future, it may develop other functions. Perhaps it will even acquire the function of an other-repair initiator, but it is too early to say at this time.
Finally, the results show that even though recent borrowings from English may seem to be ‘redundant’ and that they are only replacing older domestic forms, a closer comparison of conversational functions may reveal functional differences. In the future, a further comparison between what in other Nordic or European languages may also show whether this borrowing behaves in a similar way in each language or whether each language shapes what in its own way.
Acknowledgements
The content of the paper was first presented at a NOS-HS exploratory workshop which was funded by the Nordic national research fund. My sincere thanks go to all the participants for the lively discussion on the use of what in the Nordic languages. I would like to thank Jón Símon Markússon for reading and commenting my paper at an early stage, and three anonymous reviewers for their insightful and constructive comments.
Competing interests
The author claims none.
Open access
