2.1 Data collection, data types and speaker characteristics

Any prosodic and interactional analysis will benefit from having as many examples of the phenomenon involved as possible, and for this reason the choice of field sites was based on hypotheses regarding where most speakers of the local dialect were to be found. The four residential areas are all characterised by having a high percentage of immigrants and descendants of immigrants from non-western countries (Danish housing and planning authority 2019), which is a demographic characteristic we know from previous studies as an indicator of where new urban dialects have developed (see Kotsinas Reference Kotsinas1988, Quist Reference Quist2000, Svendsen & Røyneland Reference Svendsen and Røyneland2008, Bodén Reference Bodén2010, Freywald et al. Reference Freywald, Mayr, Özçelik and Wiese2011). For the data collection, I chose to go as far and as wide as possible rather than limiting myself to predefined types of interaction or controlled groups of speakers. I contacted clubs, schools, social media fora, and personal networks to get permission to record people in their everyday activities. The result is a dataset of very mixed data types, as can be seen in Table 1. Please note that the categories of settings and social activities are loosely defined and cannot be treated as variables that can be correlated with the frequency of Strong Finals. Nevertheless, the overview gives a sense of where to look and which hypotheses to form for a possible future distribution analysis. Generally, there are four types of data.

• Video and audio recordings of speakers who each provided background information on their age, geographical upbringing, and linguistic association.

• Self-recordings made by speakers who then subsequently provided background information for the other speakers in the recordings.

• Recordings collected by Mette Vedsgaard Christensen. Available background information: geographical location of the recordings.

• Media data from Odense used for comparison of acoustic analyses, but not for interactional analyses.

All the speakers gave their verbal, informed consent that the data could be used for research, played for researchers, and represented in written form when anonymised.

One criterion applying to the speakers taking part was that they were born and raised in Aarhus and spoke Danish as their main everyday language. Some reported speaking additional first languages at home, for example Arabic, Kurdish, or Turkish, but they all reported Danish to be their main language. This contradicts a general understanding of all urban dialect speakers as bilingual. The majority of the speakers self-report as having an immigrant background, but Strong Finals are also used by speakers with no immigrant background, including one adult club worker. Across all groups of speakers, some use Strong Finals and some do not. All speakers in the dataset grew up in multilingual urban areas, mostly in the areas where the recordings were made.

2.2 Coding procedure: impressionistically and acoustically

The phenomenon was registered initially at an open data session with students of Linguistics at Aarhus University. The Strong Finals immediately caught everyone’s attention and fascination, and they were described impressionistically as having ‘a certain swing’ and ‘a sound of double stress’. After an initial acoustic analysis of a small part of the dataset (Azulay Reference Azulay2017), the subsequent discovery procedure in the entire dataset followed a process of auditory and acoustic inspection ‘done reflexively through repeated comparison of what can be heard with what can be identified in acoustic records’, as recommended by Walker (Reference Walker2012:456). Potential candidates were selected intuitively, and acoustic analyses of individual words showed that the final syllables all met at least two of the three criteria described below in Section 3.1, ‘Three acoustic characteristics’. In total, one-third of the examples, including all examples shown in this paper, have been tested on linguists working with Danish Talk-in-InteractionFootnote ⁴ to make sure they were all evaluated as notably different from Standard Danish.

2.3 Levels of analysis: acoustic measurements and transcriptions of interaction

Examples are shown with different levels of context. For the prosodic description in Section 3, the utterances with Strong Finals are transcribed in IPA (following Heegård Petersen et al. Reference Heegård Petersen, Juul, Pharao and Maegaard2021), and acoustic measurements of pitch and intensity as well as syllabic duration are shown in diagrams made in the acoustic software Praat (Boersma & Weenink Reference Boersma and Weeninkn.d.). For the interactional analyses in Section 4, the wider interactional context is shown by including preceding and following utterances by all speakers taking part in the activity. Stress, lengthening, and intonation movements are shown with the transcription conventions of Conversation Analysis (see Appendix). All names of speakers and locations are pseudonymised.

3.1 Three acoustic characteristics

In Standard Danish we would expect mandag and tirsdag to be produced with unstressed dag syllables, as transcribed in Den Danske Ordbog (DDO, The Danish Dictionary):

DDO’s description is based on careful speech, whereas in spontaneous speech we would expect the dag syllables to be produced with even less material, as unstressed syllables in Standard Danish spontaneous speech are often heavily reduced (Schachtenhaufen Reference Schachtenhaufen2013). A probable production in spontaneous speech could be [ˈmanˀd] [ˈtiɐ̯ˀsd], omitting the vowel part of the dag syllables. Compared to Standard Danish careful or spontaneous speech, the dag syllables in Strong Finals are produced with significantly more material. In Figures 1a–1b, syllabic duration is shown in seconds on the x-axis, pitch (fundamental frequency over time) is marked with a solid line on the y-axis, and intensity (volume over time) with a dashed line. Pitch span is based on the speaker’s pitch range in the surrounding couple of minutes. The acoustic measurements of mandag and tirsdag produced as Strong Finals show that the dag syllables are characterised by the following acoustic criteria:

1. i. longer duration than the first syllable

2. ii. non-falling intonation, i.e. either rising or flat pitch

3. iii. equal or higher intensity compared to the first syllable

Figure 1a.

Mandag ‘Monday’ (line 1, example 1) produced as Strong Final.

Figure 1a Long description

A line graph displays the fundamental frequency (F0) in Hertz and intensity in decibels over time in seconds for the word 'Mandag Monday'. The x-axis represents time in seconds, ranging from 0 to 0.663 seconds. The y-axis on the left represents the fundamental frequency (F0) in Hertz, ranging from 150 to 350 Hertz. The y-axis on the right represents intensity in decibels, ranging from 50 to 80 decibels. The graph shows two distinct sections labeled 'man' and 'da:', with durations of 0.29 seconds and 0.37 seconds, respectively. The F0 values start at around 250 Hertz, rise to approximately 350 Hertz, and then fluctuate before stabilizing around 300 Hertz. The intensity values start at around 60 decibels, rise to approximately 80 decibels, and then fluctuate before stabilizing around 70 decibels. All values are approximated.

Figure 1b.

Tirsdag ‘Tuesday’ (line 3, example 1) produced as Strong Final.

Figure 1b Long description

A line graph displays the fundamental frequency (F0) in Hertz and intensity in decibels over time in seconds. The x-axis represents time in seconds, ranging from 0 to 0.6982 seconds. The left y-axis represents the fundamental frequency (F0) in Hertz, ranging from 150 to 350 Hertz. The right y-axis represents intensity in decibels, ranging from 50 to 80 decibels. The graph shows two data lines: a solid line for F0 and a dashed line for intensity. The solid line fluctuates around 250 Hertz, while the dashed line shows peaks and troughs between 50 and 80 decibels. The graph is divided into two segments labeled 'tigs' and 'da:', with durations of 0.32 seconds and 0.38 seconds, respectively. All values are approximated.

Altogether, the perceived ‘double stress’ in Strong Finals is realised by increasing the three acoustic features duration, pitch, and intensity for syllables which in Standard Danish would be unstressed.

Figure 1c shows acoustic measurements of ‘Thursday Friday’, which follow the Standard Danish pattern of heavily reduced dag syllables.

Figure 1c.

Torsdag fredag ‘Thursday Friday’ (line 4, example 1) produced with Standard Danish reduction of final syllables.

Figure 1c Long description

The line graph shows two data lines representing the fundamental frequency (F0) in Hertz on the left y-axis and intensity in decibels on the right y-axis over time in seconds on the x-axis. The solid line indicates the fundamental frequency, which fluctuates between approximately 150 and 350 Hertz. The dashed line represents the intensity, ranging from around 50 to 80 decibels. The x-axis spans from 0 to 0.712 seconds. Below the graph, a table segments the time into intervals labeled with syllables: tps, da, fra, and da, with corresponding time durations of 0.27 seconds, 0.1 seconds, 0.25 seconds, and 0.1 seconds respectively. The graph illustrates variations in pitch and intensity during the spoken phrase.

3.2 Realisation in different syllabic and sentential contexts

The impressionistically categorised examples of Strong Finals were found both in mono-, bi- and trisyllabic words and in different syntactic positions. Examples (2–5) below (see also Figures 2–5) show some of the different variation and combinations, each with orthographic and phonetic transcription, glossing, translation, and acoustic illustrations.

Figure 2.

sån [sɒːn] ‘like.this’ (from example 2) pronounced as a Strong Final.

Figure 2 Long description

A line graph with two data lines representing different measurements over time. The x-axis ranges from 0 to 0.73999 seconds, and the y-axis on the left ranges from 150 to 350, while the y-axis on the right ranges from 50 to 80. The solid line represents one set of data points, and the dashed line represents another set of data points. The solid line starts at around 150, rises slightly, and then fluctuates around 250 before rising sharply towards the end. The dashed line starts at around 250, rises and falls in a wave-like pattern, and then fluctuates around 70 towards the end. Below the graph, a table shows three categories labeled 'ja', 'lob', and 'sp:n' with corresponding values of 0.12, 0.11, and 0.51 respectively. All values are approximated.

Figure 3.

før [fɶːˈɐ̯] ‘before’ (from example 3) pronounced as a Strong Final.

Figure 3 Long description

A line graph illustrates the relationship between fundamental frequency (F0) in Hertz and intensity in decibels over time in seconds. The x-axis represents time in seconds, ranging from 0 to 0.638 seconds. The left y-axis represents the fundamental frequency (F0) in Hertz, ranging from 150 to 350 Hertz. The right y-axis represents intensity in decibels, ranging from 50 to 80 decibels. The graph features a dashed line indicating intensity and solid dots representing F0 values. The dashed line shows a peak in intensity around the middle of the time range, while the solid dots indicate a relatively stable F0 with slight variations. The graph is divided into three segments along the x-axis, labeled with phonetic symbols and corresponding time values: gifd at 0.18 seconds, fE: at 0.21 seconds, and v at 0.24 seconds. All values are approximated.

Figure 4.

forskellige [fɒsˈgεlˈiː] ‘different’ (from example 4) pronounced as a Strong Final.

Figure 4 Long description

A line graph illustrates the relationship between F0 in Hertz and Intensity in decibels over a time span of 0.7325 seconds. The x-axis represents time in seconds, ranging from 0 to 0.7325 seconds. The y-axis on the left represents F0 in Hertz, ranging from 150 to 250 Hertz, while the y-axis on the right represents Intensity in decibels, ranging from 50 to 80 decibels. The graph features a dashed line representing F0 and a solid line with dots representing Intensity. Below the graph, a table lists four segments labeled 'træ', 'fø', 'sgε', and 'li:', with corresponding durations of 0.20 seconds, 0.1 seconds, 0.18 seconds, and 0.26 seconds respectively. The dashed line shows fluctuations in F0, peaking around 250 Hertz and dipping to around 150 Hertz, while the solid line with dots shows variations in Intensity, peaking around 70 decibels and dipping to around 50 decibels. All values are approximated.

Figure 5.

du [du] ‘you’ and tuneser [tu ˈne ˈsɐ] ‘Tunisian’ (from example 5) pronounced as Strong Finals.

Figure 5 Long description

A line graph displays the pitch frequency (F0 in Hertz) and intensity (in decibels) of the pronunciation of the words mandag and tirsdag by Zuuz. The x-axis represents time in seconds, ranging from 0 to 1.878 seconds. The y-axis on the left represents the pitch frequency (F0) in Hertz, ranging from 150 to 250 Hertz. The y-axis on the right represents the intensity in decibels, ranging from 50 to 80 decibels. The graph shows two lines: a solid line representing the pitch frequency and a dashed line representing the intensity. Below the graph, a table lists the syllables of the words mandag and tirsdag with their respective durations in seconds. The syllables are ken, a, du, en, tu, ne, and se, with durations of 0.17, 0.1, 0.27, 0.30, 0.27, 0.22, 0.23, and 0.31 seconds respectively. The table highlights the syllables en, tu, ne, and se in gray. All values are approximated.

In these examples, the Strong Final prosody is constituted by two or three of the three acoustic criteria. Monosyllabic Strong Finals are all lengthened and have either rising intonation, like (2) sån Footnote ⁵ and (5) du, or an extra syllabic boundary, like (3) før,Footnote ⁶ with two peaks in the intensity curve. The examples with monosyllabic words may appear similar to final lengthening in Standard Danish, as described by Tøndering (Reference Tøndering2010); but the difference is that Strong Final monosyllabic words are realised with rising intonation or an extra syllabic boundary, whereas Standard Danish final lengthening is just that, lengthening, with no rise or syllabic boundary. All the monosyllabic examples in the dataset were impressionistically identified as Strong Finals by Danish-speaking linguists. Bi- and trisyllabic Strong Finals appear to receive stress on both the ultimate and the penultimate syllable. The stress on the ultimate syllable is realised either by means of lengthening and increased volume, as in (1) tirsdag, by means of lengthening and rising intonation, as in (4) forskellige Footnote ⁷ and (5) tuneser,Footnote ⁸ or by means of all three characteristics, as in (1) mandag.

Syntactically, Strong Finals can occur at possible main clause endings as in (1–3) and (5) tuneser ‘Tunisian’, at subordinate clause endings as in (4), or mid-sentence but before a phonetic reset realised as a short pause as in (5) kender du (0.3) ‘do you know’. No correlations between syntactic environment and the way of realising Strong Finals have been found.

3.3 Intonation in Standard Danish

The main work on Danish intonation is Nina Grønnum’s (Reference Grønnum1992) intonation model, which predicts that, everything else being equal, we can expect falling intonation over the course of a sentence. As shown above, Strong Finals are characterised by the opposite, that is, non-falling intonation typically at the end of clauses. Grønnum’s work is mostly based on read-aloud speech, and is supplemented by John Tøndering’s (Reference Tøndering2008, Reference Tøndering2010) work of intonation in spontaneous talk in task-oriented speech. The main difference between read-aloud and spontaneous speech, Tøndering finds, is the existence of ‘final lengthening’ in spontaneous data. Tøndering defines final lengthening as ‘linguistic sounds which, all else being equal, have longer duration in final position than in non-final position’ (Tøndering Reference Tøndering2010:179, my translation). The durational aspect of Strong Finals may be connected to Tøndering’s final lengthening. Tøndering shows no examples in context, so no studies exist yet as to whether they appear in the same interactional environments. Nicolai Pharao has studied the acoustics of consonant reductions in Danish and notes that final lengthening, as described by Tøndering, may inhibit the tendency for reduction in spontaneous speech (Pharao Reference Pharao2010:51). In most cases, this is exactly what happens to vowels in Strong Finals: a usually unstressed and reduced syllable is produced with lengthening instead of reduction of the vowel. However, we have no clues yet that the origin or spread of Strong Finals can be related to the reduction in Standard Danish varieties.

Looking at whether specific syntactic utterance types are connected to specific intonation patterns, Grønnum & Tøndering (Reference Grønnum and Tøndering2007) compare spontaneous talk to read-aloud speech, and find that declarative questions are produced with steep falling intonation, wh- and inversion questions with falling but less steep intonation, and declarative questions with no fall, that is, their global contour is level (Grønnum & Tøndering Reference Grønnum and Tøndering2007:1229). The only difference between read-aloud and spontaneous speech is found in a subgroup of declarative questions which in spontaneous talk show steep falling intonation, and thus have the same intonation as declarative statements. No utterance types are always produced with rising global contour, but Grønnum & Tøndering show that some types are typically produced with a rise in post-tonic position. This ‘last rise’, as they call it, is seen after declarative statements and after declarative questions (Grønnum & Tøndering Reference Grønnum and Tøndering2007:1232). Note that Grønnum & Tøndering’s post-tonic rise is not the same phenomenon as this article’s Strong Finals, where the rise is on the actual tonic syllable, not in a post-tonic position.

Grønnum’s, Tøndering’s, and Pharao’s acoustic descriptions are based on speakers of Copenhagen Standard Danish; and even though Tøndering and Pharao use non-scripted speech, their data comes from a task-oriented experimental setting, namely the DanPASS corpus of map task dialogues (Pharao Reference Pharao2010), and might not accurately reflect patterns of speech in natural interaction.

3.4 Intonation in the regional standard

Danish is often described as one of the most standardised languages in Europe (see Pedersen Reference Pedersen, Auer, Hinskens and Kerswill2005, Kristiansen Reference Kristiansen2009), a language in which the only geographical variation left is regional differences in intonation. Grønnum (formerly Thorsen) describes the difference between Copenhagen Danish and Aarhus Danish intonation as being opposite tone-curves for stressed syllables followed by an unstressed syllable: in Copenhagen, the tone rises on the unstressed syllable, whereas in Aarhus, the tone falls (Thorsen & Jul Nielsen Reference Thorsen and Jul Nielsen1981:1), so in Standard Aarhus speech, we would expect a falling tone on the unstressed dag syllables. Kyst (Reference Kyst2008) describes an additional feature specific to the Aarhus regiolect: stød (a kind of laryngealisation: Pharao Reference Pharao2010:47) in monosyllabic words can be realised as ‘two-tonal’ words with falling pitch (Kyst Reference Kyst2008:238), i.e. similar to a prolonged diphthong. This would predict that a monosyllabic word like før (‘before’) in (3) could be pronounced [ˈfɶɐ↘] or lengthened [ˈfɶːɐ↘]. However, Kyst’s description does not account for lengthening with an extra syllabic boundary, nor does it describe monosyllabic two-tonals with rising intonation. Kyst’s two-tonal stød might be related to Strong Finals in a way that has not yet been studied, but stød is not the same as Strong Finals. In future studies it may be relevant to look at existing regional differences that might explain why Strong Finals are found in Western and Southern urban dialects in Aarhus and Odense, but not in the Eastern urban dialect in Copenhagen.

3.5 Summing up the prosodic description

What listeners perceive as ‘a certain swing’ or ‘double stress’ in Strong Finals is obtained through a combination of two or three acoustic features: (i) lengthening of a syllable, (ii) increased volume on a syllable, and (iii) non-falling intonation, i.e. flat or rising intonation. It is possible to say that the phonological rules of stress in Standard Danish and in the regional standard are altered in the urban dialect in certain contexts. It is important to note here that by far the most common prosodic pattern in the urban dialect is Standard Danish, and that Strong Finals simply seem to be an extra prosodic resource available to the speakers to use in certain contexts. In the following, I describe five interactional contexts in which Strong Finals are often used.

4.1 Listing items or events

This category of actions represents almost half of all examples of Strong Finals in the collection, a total of 129 examples out of 265. The category consists of two subtypes: listing items and listing events. Below is an example of listing items (6), followed by an example of listing events (7). Both subtypes are treated by recipients as a way of creating and resuming lists, and in both examples we can see a recipient treating it as such by choosing a place right after a Strong Final to insert a question about something they need clarified before the speaker can continue their list.

Creating a list is an action that requires a construction making it recognisable as a list for the recipients (Jefferson Reference Jefferson and Psathas1990), and in the examples below I show how Strong Finals support this action. On an abstract level, the examples may be described as [item A↗ and item B↘] or [event A↗ and then event B↘], where A elements are realised as Strong Finals and B elements are realised with prosodic step-down, a feature common on last elements of lists (Selting Reference Selting2007). Examples in the data suggest that there may be any number of A elements before the B element.

In example (6) (the same as example (1) in the Introduction) Zuuz is telling club worker Martin about his weekly football practice, and I will argue that by producing the item ‘Monday’ (line 1) as a Strong Final, Zuuz projects the beginning of a lengthy list of the type of item which could be called ‘weekdays’.

Zuuz’s Strong Final ‘Monday’ (line 1) ends a declarative sentence which makes it a syntactically possibly completed turn. Seen as such, Martin’s next turn, ‘do you have it as an elective course or what’ (line 2), may simply be the beginning of a new sequence, a question–answer adjacency pair. But looking at Zuuz’s next turn (lines 3–4), we see that he treats Martin’s activity as a sequence whose function is to clarify a minor detail before the listing can continue. Zuuz delivers a minimal response, ‘yes’ (line 3), and repeats his initial telling with the same wording as in line 1, only now expanded with the additional weekday item ‘Tuesday’. ‘Tuesday’ is produced as a Strong Final in direct extension of ‘Monday’ without prosodic reset or conjunctional ‘and’, which gives the impression that ‘Monday Tuesday’ is now one connected item. By designing his turn in line 3 as a repetition of line 1’s prosody and wording, Zuuz seems to insist that he is redoing his original project, that it was a list, and that there are more items in the list. He then continues with two more items belonging to the same class (weekdays), ‘Thursday Friday’ (line 4), again produced in close connection as one connected item, but now with falling intonation. This completes the item list, which, at least in retrospect, appears to have had the form [item A↗] in line 1, but in lines 3–4 is changed to [items A-A↗ items B-B↘].

A similar form is found in lists of events. Example (7) shows multiple Strong Finals used for events in a story, until finally the story reaches its end, and the last event is produced with falling intonation. The excerpt comes from a recording of Melissa and her high-school friends Dorit and Ann talking about a recent running competition in which Dorit beat the school record. Melissa has disclosed that she cheated and did not finish the run, and Dorit has asked how that was possible. In the excerpt, Melissa is enacting a sequence of events on the day of the competition.

The last word of the first event, sån ‘like.this’ (line 1), is produced as a Strong Final and accompanied by a large gesture (see Figure 6), supporting the function of projecting that the turn is not finished. Disregarding this, Ann now self-selects and asks a question about absence marking (line 2). The registration of all absences is an important issue for Danish high-school students, as the percentage of hours of absence influences how many final exams they have to take. Ann draws attention to her utterance by using the interjection eow,Footnote ⁹ which is described by Zachariassen & Nielsen (Reference Zachariassen and Lykkegaard Nielsen2022) as an interjection indicating problems or solutions to problems, and which might serve as a misplacement marker. We do not know whether Melissa did not hear Ann’s question or whether she chooses to ignore the interruption, but either way, she continues her storytelling in an overlap with Ann’s question (line 3). Ann treats the non-answer as a need for repair, but she repeats only the overlapped part ‘did they mark absence’ (line 4), which makes it possible to interpret the question as a simple yes/no question instead of the original how question. Melissa treats Ann’s question as a yes/no question and delivers the minimal response ‘yes’ (line 6). Like Zuuz above, Melissa treats Ann’s question as a request to clarify a minor detail before the listing can go on. After the minimal response, Melissa restarts the story with ‘I ran like up here↗’ (line 7), with a slight re-framing of words but with the same Strong Final prosody as in line 1, and then moves on to the next and final event in the sequence: ‘and then I ran like across↘’ (line 9). It is produced with ‘and then’ and with falling intonation, and it is followed by a punchline produced laughingly (line 10), making it a possible conclusion of the story (Kjærbeck & Asmuß Reference Kjærbeck and Asmuß2005). Taken together, the falling intonation and the laughter indicate that this is the final event.

Figure 6.

Still photos of Melissa’s gestures in example (7).

Figure 6 Long description

The image contains four still photos of Melissa's gestures in different positions, labeled as Line 1, Line 4, Line 7, and Line 9. In Line 1, Melissa is pointing with her right hand while standing. In Line 4, she is seated and pointing with her right hand. In Line 7, she is also seated and pointing with her right hand, but from a slightly different angle. In Line 9, she is seated and pointing with her right hand, with an arrow annotation indicating movement or direction. The photos are arranged side by side, showing the progression of her gestures.

The recipients’ actions show us that they orient towards Strong Finals as an indication that there is more to come. Ann does this by inserting a sequence with a misplacement marker, and the other friend, Dorit, does this by providing the response token hm (lines 5 and 8), serving as a ‘continuer’ (Schegloff Reference Schegloff and Tannen1982, not to be confused with Christensen’s ‘continuer intonation’, 2010), and thereby supporting Melissa’s storytelling project.

Summing up, Strong Finals together with other pragmatic, semantic, and gestural means support the construction of lists of items or events. In (6), Zuuz introduced an item in a class (weekdays) that has clear co-members (other weekdays) which makes listing relevant. In (7), Melissa described an event that was clearly not the entire story that was projected, together with a gesture. Both examples have a flavour of insisting on telling a story, as they are both followed by a repetition when resuming the storytelling after an inserted sequence.

A note on try-marking is due here. There is a further aspect to the interactional construction of lists: not only should the recipients know that the A item or event projects a lengthy list, but it may also be crucial that recipients recognise the references to things, people, and events in the list. In order to achieve recognition, speakers can use ‘try-markers’. Try-markers is a format originally described by Sacks & Schegloff (Reference Sacks, Schegloff and Psathas1979:18ff) for the reference to persons as the ‘use of such a recognitional as a first name, with an upward intonational contour, followed by a brief pause’. In examples of lists in my corpus, Strong Finals seem to function as try-markers by inviting recipients to indicate that they recognise what is being talked about. For example, in (7), Melissa’s explanation of the running route is formulated with deictics accompanied by large arm gestures (see pictures), and Dorit’s ‘hm’ (lines 5 and 8) may be seen as confirmation-of-recognition of the individual locations on the route. The form of this Strong Final differs from the form described by Sacks & Schegloff for (American) English try-marking as ‘an upward intonational contour’ which, judging from the cases shown here, amounts to a step-up in pitch. Strong Finals do not include step-up, as described above in Section 3.1. However, both Strong Finals and Sacks & Schegloff’s try-marking are prosodically marked, which helps to emphasise the word containing the marking, making it more relevant to interactional uptake (Selting Reference Selting1996). It is, of course, difficult in these cases to distinguish a continuer use of ‘hm’ from a confirmation-of-recognition use, and I will not proceed further with an attempt to disentangle the two here. I just wanted to point out that Strong Finals may have this ‘second-order’ function (Sacks & Schegloff Reference Sacks, Schegloff and Psathas1979:16) of locally inviting displays of recognition.

4.2 Projecting a description of N

The second most frequent action supported by Strong Finals is projecting a description. I call this ‘projecting a description of N’ because in many cases the Strong Final falls on a noun phrase. Two syntactic formats are common in this action: (i) a noun phrase in syntactic extra position ending with a Strong Final and followed by a subject copy, often a pronoun, as exemplified in (8), and (ii) a declarative main sentence ending with a Strong Final and followed by an adverb, often der (‘which’), initiating a subordinate clause, as exemplified in (9).

If we generalise pronouns and light adverbials as light elements (the two have been shown to constitute similar syntactic characteristics in Danish Talk-in-Interaction: Brøcker et al. Reference Brøcker, Hamann, Jørgensen, Lange, Mikkelsen and Steensig2012), the examples may be described on an abstract level as: [N↗ light element-initiated sentence↘].The Strong Finals in type (i) occur in a place where the turn for syntactic and semantic reasons cannot be finished yet, as in (8) where ‘my family in Zimbabwe’ constitutes only a noun phrase and not a full syntactic sentence. Strong Finals in type (ii), on the other hand, end a unit which constitutes a full syntactic sentence, and in (9) it is even followed by a pause. However, both type (i) and type (ii) Strong Finals are treated as signs that the turn is not finished yet, as we shall see in longer excerpts of (8–9) below.

Excerpt (8) is from a recording of a group of friends, a club worker, and the researcher, Ditte, talking about ethnic and national backgrounds. 14-year-old Muaaht describes his family’s background as refugees using a Strong Final when introducing the description.

After Muaath’s Strong Final in line 1, Ditte aligns as a listener by delivering a continuer ‘yeah’ (line 2), and Muaaht then proceeds with the description of the N element from line 1 (line 3). Muaaht leaves his turn open-ended with a word-searching ‘eh’ followed by a pause, and then produces a ‘yeah’ with falling intonation (line 5), which may serve as prosodically ending the turn without semantically and syntactically specifying where his family fled to. After a pause, Rahman self-selects for the next turn (line 6). This shows that Rahman treats falling intonation and a pause as a possible ending enabling him to take the floor, whereas Ditte treated Strong Final intonation as a place to align as a listener.

Excerpt (9) is from another youth club. Here Zuuz is explaining his family relations to a club worker by drawing a family tree with geometric symbols representing different family members.

In line 2, Zuuz delivers a syntactically and semantically possibly completed turn followed by a pause (line 3), but the club worker does not take up the turn. Instead, as described, the club worker aligns as a listener with her gaze and her silence. The excerpt is from the beginning of a sequence with almost two minutes of Zuuz telling and drawing, and throughout the two minutes, the club worker follows Zuuz’s telling and his drawings closely, providing continuers like ‘yeah’ and ‘mm’ from time to time but mostly remaining silent, as is also the case in the excerpt.

Summing up, Strong Finals on noun phrases (Ns) can project a subsequent description or elaboration. Strong Finals are used when Ns are part of possibly completed turns, and when Ns are located in a place where the turn is structurally incomplete. We see that recipients orient towards Strong Finals on N elements as a projection that there is ‘more to come’ by aligning as listeners.

4.3 Stating conditions

Strong Finals in connection with conditional clauses are, like subcategory (i) of noun phrases in Section 4.2 above, part of a syntactically unfinished format which, in and by itself, indicates that there is more to come. In the category of Strong Finals described in this section, the format is a conditional clause, i.e. an ‘if-then’ or ‘when-then’ construction. An abstract representation looks like this: [if A↗ then B↘] or [when A↗ then B↘], and examples in the data suggest that there may be any number of A elements before the B element. Interactionally, ‘if-then’ constructions have been described (for instance by Lerner Reference Lerner1991) as a projecting device: upon hearing the ‘if’ component (which in these cases ends with a Strong Final), a recipient will know that a ‘then’ component must come before the syntactic structure is complete. Couper-Kuhlen & Selting (Reference Couper-Kuhlen and Selting2017) treat this form of syntactic projection under the general term ‘pre-posed hypotactic clauses’ and show how these clauses create turn-construction units which are connected as compounds (Couper-Kuhlen & Selting Reference Couper-Kuhlen and Selting2017:454). In (10) below, a Strong Final supports this function. The excerpt comes from a cooking class in which Said is whipping eggs for an omelette, and club worker Rasmus dares him to drink the raw eggs.

Said responds to Rasmus’s question with a firm rejection: he is never going to drink the eggs, with emphasis on ‘never’ (line 2). Rasmus does not accept the answer but asks again ‘nej’, with rising intonation functioning as a sort of ‘are you sure’ question. Then follows a rather long pause before Said begins a story which functions as an argument for rejecting the dare. He is not a coward; in fact he has already participated in an egg-drinking dare in the past. Said rushes through the story with no pauses (marked with =) between the syntactic units, and ends the story in a conditional clause with a Strong Final ‘if’ part (line 7) and a falling intonation ‘then’ part (line 8). The ‘then’ part repeats the drikke det (‘drink it’) part of Rasmus’s initial question (line 1), and this repetition may support a concluding function showing that Said feels that this question has been answered sufficiently, and that Rasmus does not need to pursue it further because a thorough argument has been presented as to why Said does not have to take the dare.

As pointed out by Lerner (Reference Lerner1991, citing Sacks 1992 [Reference Sacks1964]), conditional clauses are, along with other types of turn-constructional compounds, an obvious basis for the joint production of utterances: they are cases in which one syntactic unit is produced by more than one speaker. Excerpt (11) includes such joint production: here one speaker produces the ‘if’ part and another the ‘then’ part. The excerpt comes from a self-recording by high-school students Charlie and Lizz, who are outside at a party talking about how they would both like to be more outgoing and how difficult it is to get to know new people at a party.

In the first four lines, Charlie describes herself in specific, concrete examples: ‘I’m not good […] I don’t know […] I do manage […]’. Then Lizz takes the floor and describes a hypothetical general scenario rather than a personal one: ‘If you’re together with someone you know’ (line 6). The personal pronoun du (‘you’) is used here to refer to an unspecified person, as a synonym of the more formal man (‘one’) (Jensen Reference Jensen2007). Lizz begins her turn in recognitional overlap with the previous utterance, that is, at a place where she projects the ending of Charlie’s utterance (Jefferson Reference Jefferson1993). She prefaces the utterance with det lige præcis det (≈ ‘that’s it’, line 5), a format which may be used to express agreement, confirmation, and acceptance, but in a way in which the speaker claims to have epistemic authority (Olesen Reference Olesen2019). In this way, Lizz claims to know as least as much about the situation as Charlie. Lizz’s hypothetical, general scenario is formatted as an ‘if’ sentence with a Strong Final, and when she takes a short breath (line 7), Charlie takes the turn and produces the then component as a possible completion: ‘then I hadn’t done it’ (line 8). This collaborative construction of a compound unit seems to be emblematic for the interaction that follows it; and after the excerpt, the description of the scenario continues in the form of collaborative storytelling.

Summing up, in both (10) and (11) the syntactic structure of a conditional clause in itself projects a following ‘then’ part, and this projection is strengthened or focused on by the Strong Final in the first part of the conditional clause. However, (11) shows that this ‘more’ may be provided by either participant.

4.4 Asking a question

For listings (Section 4.1) and stating conditions (Section 4.2), there is usually an inherent element of self-selection for the next turn, in the sense that they are syntactically and/or semantically constructed in ways that makes them unfinished. The action described in this section – asking a question – is on the other hand usually inherently other-selecting and thus unlikely to be produced with a ‘continuer marker’. Only eight examples are found in the dataset, but the category is included here to show the diversity of constructional settings in which Strong Finals occur. Based on Grønnum & Tøndering (Reference Grønnum and Tøndering2007; also our Section 3.3), we would expect questions to have a falling intonation, with wh-questions having rather less steep falls than inversion questions. Declarative questions may be produced with level intonation, but no question types are described with rising intonation in Standard Danish. The eight examples of questions, three wh-questions and five inversion questions, are produced with rising intonation. There is no apparent pattern between steepness of rise and syntactical and/or interactional categories.

In the following, I show two examples of questions where Strong Finals support the action of requesting an answer which is more elaborate than just a confirmation or rejection. Example (12) is from a recording in a youth club where Baqir is bombarded with questions from club worker Rasmus about an exotic vacation from which he has just returned. Muaaht and six other boys are listening carefully to Baqir’s stories. The excerpt begins with Muaaht making an observation (line 1) which prompts Rasmus to ask one more question. This question is produced with a Strong Final.

Rasmus’s question is formulated with a third-person subject, ‘he’ (line 2), as if Muaaht or someone other than Baqir has been selected to answer, but Baqir self-selects to answer (line 3). Rasmus’s question is a yes/no interrogative, but Baqir’s answer is not a type-conforming yes or no (Raymond Reference Raymond2003). Instead, he treats the question as an invitation to storytelling and formulates his reply with ‘also’ (ogs). Baqir’s turn can be seen as a non-answer as it ignores the speaker selection and does not conform to the type format. The speakers do not epistemically treat Rasmus’s interrogative as a question to which only the selected speaker has the knowledge to provide the answer (Heritage Reference Heritage2012). Instead, Baqir treats Rasmus’s interrogative as an invitation to storytelling.

In the next example (13), we see a similar case of an interrogative treated as a non-epistemic question. In the homework club five schoolmates are sitting around a table gossiping about people they know from school. Before the extract, Aziz has repeatedly tried to attract the attention of the other participants by calling their names, but without success. Ray and Yussuf are having a heated conversation about someone who allegedly would like to have sex with Yussuf. Dennis and Sivan are paying close attention to this interesting topic. The extract begins with Yussuf’s sceptical enquiry into the truthfulness of the story (line 1).

In lines 3–4, Ray and Dennis answer Yussuf in an overlap, both giving an assessment of the person in question, and the question–answer sequence is now completed. Aziz now self-selects as a speaker and introduces a new referent: ‘a Tunisian’ (line 5). In contrast to the earlier attempts to attract the attention of the participants by calling out their names, this time Aziz succeeds in taking the floor. Aziz does this by multiple linguistic and interactional means. Firstly, he begins at a recognitional onset, that is, in slight overlap (Jefferson Reference Jefferson1986). Secondly, he speaks loudly so he can be heard despite the overlap. Thirdly, he introduces a new referent, ‘a Tunisian’, which is unrelated to the previous topic. Fourthly, he formulates his turn as a question, appointing a recipient ‘you’ to deliver a second part, and finally he divides the question into two parts: ‘do you know’ + ‘a Tunisian’, each produced as Strong Finals. Stivers & Rossano (Reference Stivers and Rossano2010) describe a number of the above forms as ‘Response-Mobilizing Features’: interrogative morphosyntax, recipient-tilted epistemic asymmetry, and interrogative prosody. Stivers & Rossano’s third criterion can be regarded as equivalent to the use of a marked prosody, here the Strong Finals. After the excerpt, the conversation continues to focus on the Tunisian, and Aziz has succeeded in changing the topic, even though the new topic is far less interesting than the original topic.

Summing up, (12) and (13) both contain interrogatives that are treated as non-epistemic questions. Instead, the utterances are treated as an invitation to further engagement. The interrogative structures here seem to function as a response-mobilising feature rather than as questions.

4.5 Announcing reported speech

In this section I describe a fifth and final action in which Strong Finals support projection of ‘more to come’: before reported speech. In this category, the speaker has already given semantic clues that their project is not completed by using formulations like ‘he said to me’, and the added Strong Final supports building up to a story’s climax. Selting (Reference Selting2017) shows that many aspects of prosody can support this function of displaying and managing affectivity in story climaxes. The collection contains relatively few examples, but the category is included because I find it interesting that the formulations are so similar. They all consist of a pronoun subject + a verb synonymous to ‘speak’ – either ‘ask’, ‘tell’, or ‘say’. Examples of announcing reported speech are found across different locations, settings, age of speakers, and year of recording. Rathje (Reference Rathje, Gregersen, Parrott and Quist2011) discusses the lack of definitions of the concept ‘reported speech’ or ‘quotations’ and calls for a definition which takes into account the fact that quotations are always the responsibility of the quoter, not the quoted, and that they are always enactments and not replications of actual actions in the past. Additionally, Rathje points out, the quoted element can be thoughts, feelings, or facial expressions, as well as speech. A broader definition of quotatives, such as Rathje’s, might reveal more examples in my dataset that could be included in the collection.

Below I show two examples in which Strong Finals and reported speech support the build-up to a story’s climax. The build-up can be successful in that it makes a recipient respond as in (14), or it can be unsuccessful in summoning the recipients as in (15), where the speaker himself illustrates the climax by laughing at his own utterance. Example (14) is taken from a self-recording of Gözde telling her friend Bilge about a time when she owed a lot of money to the bank. The excerpt contains three Strong Finals, two in line 1 and one in line 2, which all contribute to the build-up to the story climax.

Lines 1–2 describe the event in which the reported speech occurred and has similar traits to the event lists in Section 4.1, except here the list consists of a time stamp, ‘this year’ (line 1), an event, calling the bank (line 1), another event, asking the banker (line 2), and then finally an element produced with falling intonation, here the reported speech (line 3). Bilge responds with awe, recognising that the amount Gözde pays every month makes this a story worth telling. Bilge’s response is prosodically marked with loudness, extra exhalation, and the lengthening of syllables, and semantically marked by the use of ‘wallah’, an interjection-like word often used for emphasis.

In the following example (15), the reported speech builds up to a climax which is not recognised as such by the recipients. But we can see that it was intended as a climax because the speaker himself laughs at his own utterance. The example comes from a homework club: Ray and Yussuf are mocking Amir (present but not in the excerpt) telling three other friends (ditto) about a time when Amir left a football camp all of a sudden.

Before the excerpt, Yussuf has been listening to music and singing along, adding Amir and London into the lyrics, while Amir has repeatedly claimed to have nothing to do with London. When Ray starts mocking Amir for missing out on this summer’s football camp (lines 1–2), Yussuf immediately picks up this information and uses it for a lengthy story. In lines 5–7 Yussuf quotes himself quoting Amir: I told them that Amir told me that he was going to London. The Strong Final falls on the part introducing Amir’s quote. It is not a direct quote, as it is formulated ‘he is going’ (a direct quote would be ‘I am going’), but it still falls within the definition of reported speech. Yussuf’s self-quote (line 6) is formulated with a lot of syntactic extras: ‘Amir’ in extra position and the resumptive subject ‘he’, which are usually directly connected (Brøcker et al. Reference Brøcker, Hamann, Jørgensen, Lange, Mikkelsen and Steensig2012), are split by an adverbial phrase ‘yesterday’, the time setting is presented twice, first as ‘yesterday’ and then specified as ‘yesterday evening’, and this specification is produced with a Strong Final. Amir’s part of the quote (line 7). is formulated with a drop in pitch: ↓han skal til London↘ (‘he is going to London’). The changes of intonation, the Strong Final, and the pitch drop illustrate that the story has reached a climax. None of the recipients – Ray, Amir, and the other friends – react to Yussuf’s story, but Yussuf himself treats the reported speech as the story’s climax by repeating it laughingly in line 9.

Summing up, Strong Finals in the constructions with reported speech can support the action of focusing on the reported speech and showing that this part of a storytelling process is leading up to a climax in the story.

4.6 Interactional common traits

Strong Finals can be found in structurally unfinished formats (stating conditions and announcing reported speech), structurally open-ended formats (listings and descriptions), and structurally finished formats (interrogatives). Interactionally, they share the common trait of projecting more to come, but differ as to whether this ‘more’ could or should be provided by someone other than the speaker. The prosodic pattern has previously been described as ‘continuer intonation’, ensuring that the speaker gets to keep the turn (Christensen Reference Christensen2012). But the functionality is more complex than this: it can also be used to assign turns to other speakers. In the examples above, the projected ‘more’ is often of greater significance to the activity; for example, it may be an abstract conclusion of a lengthy argument, an initiation of a new story or topic, or a build-up to the climax of a story. Recipients may treat Strong Finals as designed to either align as listeners or to contribute. Strong Finals are sometimes treated as try-markers and responded to with a short response token or a clarifying inserted sequence, and other times treated as a recruitment to engage in a collaborative storytelling or activity.

The examples often come from multi-speaker settings in which getting, keeping, and assigning the floor are important parts of turn negotiations. Certain interactional formats make Strong Finals a particularly relevant resource.

• In descriptions, Strong Finals are used to make a point that a lengthy list is coming, as in (6) Monday, or that an elaboration of a description is coming, as in (8) Refugees and (9) Triangle.

• In storytelling, Strong Finals are used when an order of events is crucial in terms of understanding the story, as in (7) School run, or when a concluding remark or a climax of the story is on its way, as in (10) Spin-the-bottle, (11) Party, (14) Debt, and (15) London.

• In questions, Strong Finals are used as a response-mobilising feature to invite to more than an epistemic answer, as shown in (12) Sun tan, and may additionally support a speaker in moving on to a new topic, as in (13) Tunisian.

Strong Finals can assist speakers in getting the floor or changing the topic of conversation, but they can also have the opposite effect of alerting recipients that this is the place to interrupt if they want to add something before a lengthy stretch of talk, as in (6) Monday and (7) School run.