Overcoming Blanking: Verbal and Visual Features of Prompting in Theatre Rehearsals

In theatre rehearsals, actors can occasionally be seen getting stuck in the play text, which is called blanking. To overcome such textual difficulties and continue with the given text, a prompter can verbalize the line in question, thus contributing to an actor’s word search by prompting. The paper focuses on interactional practices by which prompters and actors interactionally resolve blanking situations. This study’s data comprises a case collection of 67 prompting situations, which are taken from a 200-h video corpus of a rehearsal process at a professional theatre. These cases demonstrate how theatre professionals organize prompting situations and how they negotiate/sanction prompting actions such as when there was no blanking but a dramatic pause or when the dramatic performance is interrupted due to a missing prompt. In addition to the audiovisual recordings, eye tracking data of the person prompting is also used to describe the coordination of the visual resources in the context of multimodal interaction analysis. The analysis suggests that prompting and blanking persons interactively resolve blankings with the help of verbal and visual markers.


Prompting in Theatre Rehearsals
When theatre professionals work together on plays, actors can get stuck in the play text during rehearsals, which is called blanking. To overcome such textual difficulties and continue in the given text, a prompter can verbally perform the line in question, thus contributing to maintaining the ongoing play activity. Although members I express my sincere thanks to the editors and the anonymous reviewers for their very helpful comments on earlier versions of this paper.
1 3 of a theatre rehearsal usually strive to keep the performance in progress, short interruptions of rehearsals are a recurring phenomenon and not an institutional problem per se. Rather, the director has the right and, in terms of the usually tight time management, the obligation to interrupt a play rehearsal to give stage directions for the next iteration. That means, interruptions, repetitions, and prompts are routine features of theatre rehearsals. Interruptions of the play, however, become a problem when they are not initiated by the director, i.e., when an external person interrupts the rehearsal or when actors forget their lines. For the latter case, prompting is an institutional solution to a typical problem (blanking) in theatre rehearsals.
Rehearsals are work phases in the institution of theatre. As part of an institutional workspace, they differ from domestic and public situations above all in the extent to which they can be planned, the spatial and temporal constraints, and the specific framing of participants. The study of institutional workplaces has a long tradition in sociology and linguistics in the course of so-called workplace studies. Empirical work on court proceedings (Atkinson & Drew, 1979), medical examinations (Heath, 1986), call centres (Mondada, 2008) or air traffic control (Arminen et al., 2014) provide new perspectives on social interactions in complex work environments. Since most workplace research on theatre rehearsal relies on ethnographic approaches (e.g., McAuley, 2012), interactional coordination in theatre rehearsals is rarely investigated. The most recent conversation analytic work on social interaction in theatrical settings tracks the developments of scenes across rehearsals (Hazel, 2018), investigates the embodied ways in which a script is acted into being (Lefebvre, 2018), or focuses on the practices by which actors project a change from the real world to the play world (Schmidt, 2018).
This study of theatre rehearsals situates itself within this body of research and describes rehearsal processes as an interplay of institutionally entrenched activities that are interactionally negotiated by the participants. More precisely, this paper highlights the interactional work that members of a theatrical rehearsal do to maintain the play word by collaboratively resolving blanking situations.

Word Searches as Interactional Achievements
Although promptings in rehearsals appear analogous to word searches in everyday interactions (Bolden et al., 2019), they can be distinguished from word searches in unscripted discourses, as during rehearsals, the following words have already been set in advance in the script. As there is no natural negotiation on the right to speak due to the turns' predetermined order, the play activity's progress depends mostly on the right person performing the correct text at the appropriate time. Unlike everyday word searches, the textbook determines the turn allocation system in a strict sense (who says what in which order?), leaving mostly the performance (how) to be the focus of the theatre production members (see Lefebvre, 2018). Still, participants of prompting situations must locally negotiate when a blanking has to be treated as such, when it has been successfully overcome, or when it is still underway. In that sense, blankings are similar to word searches in everyday interactions. Both processes are sometimes about remembering a sentence, a line, a name etc., but they can also be about finding a specific situation-appropriate (the 'right') word. However, while word searches in everyday interactions are communicatively negotiated processes in which the participants agree on the word they are looking for according to the situation (Bolden et al., 2019;Goodwin, 1995;Helasvuo, 2004), blankings are collaborative word searches for which the words that follow have been predetermined (in the textbook).
To describe the sequential organization of word searches, Dressel (2020: 39) suggests three phases: onset of the search, search-in-progress, and search resolution. Within this framework, it becomes apparent that participants in everyday interactions resolve word searches in the similar way in which rehearsal members overcome blankings through prompting. However, as the following analysis demonstrates, there are also significant differences due to the script-bound nature of play activities in theatre rehearsals.
In the onset phase, speakers in word searches and blankings display problems with the progress of the ongoing talk by discontinuing their current TCU. Word searches as self-initiated repairs are often projected by sound stretches and interjections like uh (Schegloff et al., 1977: 367), and they can be found in blankings as well.
When the speakers cannot immediately solve the word search on their own, they enter the search-in-progress phase, in which they perform either solitary or joint searches (Dressel, 2020: 40-43). For the word-searching person, this phase's main interactional task is to dissociate the word search from the previous utterances. This can be achieved by pauses and repetitions (Goodwin, 1980), prosodic features like lower pitch (Ogden, 2013), and projections through gesticulation such as forward gestures (Streeck & Jordan, 2009) by which participants can foreshadow the word they are looking for. Blankings and everyday word searches differ because of their epistemic asymmetry (Drew, 1991;Sidnell, 2012) and deontic authority (Stevanovic & Peräkylä, 2012). It is the institutional task of the prompter to compensate for the actors' lack of knowledge when necessary. Deontically, prompters act "in the service" of the actors, exclusively in cases where there is actually a blanking. As the analysis will demonstrate, faulty prompting is sanctioned by those present. This means that prompters may only become institutionally visible when actors have problems maintaining the progress of the play.
Thus, the search resolution, in which the sought-for element has been found and will be integrated into the turn, differs for blankings and everyday word searches. In word searches, members of an interaction can refuse to accept an item presented by a coparticipant. This is not the case for blankings, because the present data suggests that actors always accept the prompted elements by producing partial or full repeats (Robinson & Kevoe-Feldman, 2010) of the prompted elements. However, they can indicate that they have performed a dramatic pause instead when the prompter misreads the cues for blanking. So, unlike everyday word searches, blankings are not so much about verbalizing missing text elements as they are about identifying blankings and distinguishing them from dramatic pauses.
To conclude, blankings are similar to joint word searches, but the two differ concerning their responsibility over the sought-for elements. In addition, blankings occur in multiactivity situations (Haddington et al., 2014), when the prompter has to pay attention to the current activity and read along in the script simultaneously. Hence, the remainder of this paper focuses on the three phases suggested by Dressel (2020), but it will be slightly adjusted to describe the interactional requirements of blankings.

Data: Eye Tracking in a Devised Theatre Production
The data that forms the basis of the prompting case collection used in this paper is a 200-h video corpus of a rehearsal process at a professional theatre in Germany. For six weeks, all 31 rehearsals of a so-called devised theatre production were recorded. Devised theatre means that the whole script, all the characters, and the dialogues of those characters were developed during the rehearsals (Perry, 2010). As a result, the play's script, in which two doctors recount their experiences of performing surgeries on children, changed from rehearsal to rehearsal as new parts were added or moved to other scenes; giving the actors and the rest of the team more discretion and potential to add and change the script during rehearsals. Perhaps due to the script's ever-changing character, the participants relied on prompting as a method to ensure the progress of the rehearsal after a blanking or a miscuing. Since improvisation is also a frequent and central element of the rehearsal process in the non-devised theatre productions, it can be assumed that the practices identified in this paper do not apply only to the devised theatre. The biggest difference in various forms of theatre potentially lies in the double responsibility of the assistant director when s/he has to transcribe improvisations in addition to the task of prompting.
During the rehearsal process, a total of 67 prompting incidents were identified (Table 1). Promptings were counted as cases when the prompter verbalized script text during play rehearsals. Only in one case did the assistant director produce a prompt where the actor had not displayed a blanking or miscuing (excerpt 2). Of all prompting cases, only 14 cases follow a miscuing (correct text at the wrong time) by a performer. The majority of the promptings (52) occur after a blanking, which makes those instances the most prominent ones in the data collection. Promptings following a blanking can be distinguished by whether the prompt is initiated by the prompter (excerpts 1 and 5) or evoked by the performer (e.g., by a finger snap in excerpt 4). In only four of these evoked cases, the performer chooses to request script text by verbal non-script material (e.g., well now), which leads to short interruptions of the play activity (excerpt 3). In all of the cases presented in this paper, the actor is the only one on stage, and thus, he is the only person to possibly get blank. In prompting situations, the prompter organizes visual resources in a way that allows to coordinate the reading along simultaneously with the visual monitoring of the performance on stage. To get analytical access to her visual coordination, mobile eye tracking data (Tobii Pro Glasses 2, 50 Hz, calibrated to 1.5 m) is used in excerpts 4-6. Eye tracking glasses measure the timing of eye movement (e.g., the exact moment when the gaze reaches another person) and the direction of gaze targets (e.g., where a person looks). This allows a more detailed description with more reliable assumptions regarding the visual orientation of the participants. However, like every measuring instrument, eye tracking is not without measurement error when it comes to accuracy and precision (see Holmqvist et al., 2012: 45). For that, every relevant gaze target (e.g., the actor, the script) is regarded as an area of interest (AOI). That means that every gaze measured close to one of these AOIs is considered an AOI hit. Only AOI hits and gaze movements between them are annotated in the transcripts. This approach has two advantages. First, it circumvents to an extent the accuracy/ precision uncertainties. These are prominent when using eye tracking in the field because the distances between potential gaze targets often change as compared to the initial calibration distance. Second, it simplifies the annotation process. Instead of annotating "gazes at actor's right shoulder" or "gazes at actor's head," only the annotation "gazes at actor" or "gaze at script" is required. Eye tracking glasses also have a scene camera mounted at the center of the eyeglass bridge that records video in 1920 × 1080p resolution at 25 frames per second. This makes it possible to record the direction of the wearer's gaze and small head movements of the wearer. For this paper, a combination of the eye trackers scene camera and its measurements were used to describe the assistant director's visual coordination in prompting sequences. Since the actor found wearing the eye tracking glasses too intrusive and disruptive to the performance, only the production's assistant director wore the eye tracking glasses. Thus, the eye tracking data provided insights into the intra-and interpersonal coordination (Deppermann, 2014) of the prompter but not for the blanking person. To determine the gaze of the actor, external cameras were used.
Also, as one reviewer noted on an earlier draft of this paper, the focus on pupils or head orientation easily ignores the gestalt of the face, i.e., facial expressions such as moving eyebrows (Dix & Gross, forth.) or widening/narrowing of the eyes. These features, which are likely crucial for members of a theatre rehearsal to understand and recognize the different actions involved in a blanking, are largely missed in the present study. This methodological shortcoming results from a mixture of a focus on the coordinative work made visible by the eye-tracking glasses and technical limitations such as video quality and distance to the actors. Future work on prompting should take the gestalt principle of multimodal communication seriously and also consider facial expressions.

Analysis: Overcoming Blanking in Prompting Sequences
Blankings are the situations within theatre rehearsals in which actors fail to produce the next element of the script. As the following analyses will demonstrate, promptings are the members' methods to overcome these blankings. In order to prompt, the prompter must analyze the ongoing play activity for mistakes and cues to enter a prompt without disrupting the performance-in-progress. In this regard, online analysis of the performance is crucial for theatre prompters, since a falsely detected blanking can bring the rehearsal to a standstill.

Blanking as a Members' Problem
The following excerpt from the eighth rehearsal day features the phenomenon of blanking. It demonstrates that an actor's (ACT) blankings are accompanied by verbal and visual cues. The assistant director (ASD), who functions as a prompter in the data, organizes her multimodal resources to react quickly to various indications of potential blankings. In this excerpt, the participants work on a monologue of the actor's character (He portrays a doctor.). He is the only actor on the stage, meaning every instruction is designed only for him. In play activities during rehearsals, actors usually stand on the stage in front of those present (here: director, assistant director, assistant, and videographer) and perform script text that they embody with various multimodal resources (including, but not limited to vocal modulation, posture, and gestures). When no playing partners are available on the stage, the actors use those present at the rehearsal stage as an audience. They direct their performances toward the director most of the time, which allows the director to adjust the play with small nods, smiles, or short verbal instructions (on-the-fly instructions, Krug, 2020). The gaze is only annotated in the following excerpts whenever an actor is not gazing toward the director. In prompting sequences, this is most prominent when the gaze is used to mark a blanking, as in the following excerpt.

Excerpt 1: Blanking in a Prompting Sequence
At the beginning of the excerpt, the actor performs the script text and, in character, tells about the activities that children often talk about before he anesthetizes them (lines 001-003). After a short pause (l.004), he first continues to perform the play text before delivering a stretched hesitation signal with a high-rising final pitch (uh::, l.004) that is then followed by another short pause (l.005). Up to this point, the assistant director has been reading along in the textbook (Fig. 1). Now, she looks up at the actor after he stops to verbally continue his performance following the hesitation signal. At this point, grammar and script overlap, as "yes and then" is hearable as the beginning of the next line, and it is only at the hesitation signal that the actor departs from the script. So, it can be assumed that the assistant director is suspicious regarding a potential blanking and is about to expand her auditory monitoring to visual monitoring of the play activity. While she is looking at the actor, he turns his gaze toward her so that both establish a mutual gaze (Fig. 2). This look by the actor towards her has a confirmatory function as the assistant director now treats the actor's performance visually addressed to her as a blanking. While maintaining an eye contact, she prompts the correct play text (I tell them to, l.006) in the actor's origo in a syntactic form following his projected construction "yes and then" (l.004). Then without letting the assistant director complete the prompt, the actor displays his epistemic status that he knows and accepts the suggested text (right, l.007). He turns his gaze back to the director and integrates the prompted text into his performance (l.008). By that, it becomes clear to the assistant director that the actor has overcome his blanking and is going to continue to play his character. As a result, she turns her gaze away from him and gazes toward her textbook (l.008).
This excerpt shows the complexities of interactional prompting situations. To prompt the missing text, the assistant director needs to read along in the script. Within interaction research, reading is described as a lonely activity that is typically not directed toward others and can be interrupted and resumed at any time as soon as another activity becomes locally more important (Goffman, 1981: 85f.). In this relation, reading along as a special form of reading is an exception to this rule as it requires someone who reads aloud (Heuser et al., 2020), i.e., a reader who verbalizes what s/ he perceives visually as written text. In that sense, reading along affords others to visually reproduce these verbalizations in their text ("listening-to-where-the-reader-is-inthe-text," Heap, 1990: 62). For the assistant director, the actor's performance is structurally similar to reading along situations as she visually tracks his verbalizations of the play text in her copy of the script. Because blanking indications manifest verbally and via embodied conduct, the assistant director needs to alternate her visual attention between the actor's visual performance on the stage and his ongoing talk's position in the script. It means that in prompting situations, at least three activities (playing, reading along, and prompting verbally) are co-relevant for the prompter and therefore form a multiactivity. Even though the prompter cannot participate equally in two of these activities (playing and reading along), these activities require her visual modality, making them structurally incompatible. When the assistant director visually monitors the actor's performance, she cannot read along and vice versa. Since pauses, hesitations, and disfluencies are an integral part of blankings, the assistant director relies on the performer's visual and auditory cues. As long as the performer displays no verbal sign of hesitation, the prompter uses her visual resources to read along in the script. However, as soon as the performer produces a lapse in his utterance, which is linked to a hesitation signal, the assistant director stops reading and starts to visually monitor the performer's embodied conduct. Then she decides on whether there is a blanking or not. If she interprets the cues as indexical signs of a blanking, she verbalizes what she has read, thus performing a prompt. 1 The actor then repeats the prompted line, integrates it into his performance, and continues the play activity. Therefore, the timing of the visual resources plays a central role in displaying a blanking (e.g., the mutual eye contact between the prompter and the actor as given in this fragment). However, it is also vital to the multimodal coordination of simultaneously relevant activities. For this reason, mobile eye tracking is used in this paper for the fine-grained description of the assistant director's gaze (excerpts 4-6).
As excerpt 1 suggests, prompting is a complex multiactivity to be observed in a specific situational setting-a theatre rehearsal. For a prompting to occur, activities are required (here: the play activity) with a more or less fixed framework in which specific actions can be classified as wrong and others as right (here: by the script). Prompting differs in this respect from overlapping in interactions (Jefferson, 1973(Jefferson, , 1986Schegloff, 2000;Mondada & Oloff, 2011) and choral speech (Lerner, 2002(Lerner, , 2004Pfänder & Couper-Kuhlen, 2019). In overlapping or choric speech of everyday interactions, no participant can objectively determine which continuation of an ongoing interaction is correct. Instead, the interactants negotiate in the situational unfolding of their interaction about which actions are preferred and which are socially sanctioned. However, what the members involved in a prompting activity have in common with these interactional phenomena of simultaneous speech is their orientation toward the progressivity of interactional processes, which should be maintained if possible: "interactants are concerned with advancing in-progress activities through sequences" (Stivers & Robinson, 2006: 386). In contrast to everyday interactions, the successive units of the play activity are predetermined by the script. As the following analysis demonstrates, this is particularly evident when performers have problems keeping up the progressivity by themselves and get blank. In such cases, the assistant director (as prompter) uses the performer's verbal and visual cues and relies on the script to determine when and what to prompt.

Maintaining Progressivity
The orientation of the continuation of the ongoing play activity is particularly evident during rehearsals when actors maintain their performance posture even during blankings to keep things in play. Simultaneously, the others who are present during the rehearsals adjust their participation in a way that does not disturb the ongoing play activity (e.g., by drinking, eating, loud speaking, or leaving the room). As excerpt 1 has demonstrated, the assistant director also helps to maintain the play activity's progressivity. She achieves this by performing her reading activity parallel with the play activity, thereby organizing it as a reading along activity. She can thus check whether the actor is speaking the correct text according to the script. By reading along, she can both visually capture the written words and auditorily monitor what is being spoken in the context of the play. As soon as the progressivity of the play activity is endangered by a blanking (recognizable in the excerpts by verbal pauses along with hesitation signals in combination with mutual gaze between the actor and the assistant director), the assistant director transforms her reading activity into a prompting activity for which she uses verbal resources. In excerpt 1, the prompted text is bite-sized for the actor so that he only has to repeat the given words to continue the play activity. As the following two excerpts will display, timing is of the essence. Generally, the more the time passes without the actor producing script text, the more apparent the blanking becomes. However, if the prompter waits for too long and the actor has to explicitly request the prompt, the prompt is considered to have failed from a member's perspective.
The following excerpt demonstrates the first prompting situation during the observed production. The assistant director, who claims to have prompted before on another play, performs the text prompt after the pattern "pause in speaking and hesitation signal" without visual monitoring of the actor. The actor interrupts his play and reprimands the assistant director for "going in" too fast (l.008). The transcript excerpt shows that the assistant director can treat the pattern "pause in speech and hesitation signal" as an auditorily perceptible indication of a blanking but s/he always needs visual monitoring of the performance to determine whether the actor is really blanking or just producing a pause within the play.

Excerpt 2: No Prompting Needed
The assistant director responds to the pattern "pause in speech and hesitation signal" to assume a blanking in the actor's performance. This pattern occurs after a verbal pause by the actor (l.004), which is followed by a highly lengthened hesitation signal (the::n, l.005) with a rising pitch movement, which could indicate 1 3 Overcoming Blanking: Verbal and Visual Features of Prompting… textual uncertainty and is hearable as a request for a line. Consequently, the assistant director treats the pattern as an indication of blanking and prompts text (l.007). Unlike the previous excerpt, however, the assistant director does not turn away from her reading along and toward the actor (Fig. 3). Therefore, she cannot visually perceive that the actor does not look in her direction as the actor in the previous excerpt did. Thus, the actor is still engaged in his performance when the assistant director prompts text. As a result, her verbal utterance overlaps with the actor's next verbal action. Responding to that, the actor interrupts his performance (l.006) as she repeats the overlapped text (l.008). In this sense, the actor was likely to be blanking but he managed to find his own line immediately after raising it as a recognizable problem of remembering. The discussion that follows displays that the members consider the prompting here to be failed and demonstrates that she cannot wholly rely on this pattern, because blanking also involves visual resources. Thus, to check a blanking indication, the assistant director needs to consider not only the auditory cues but also the visual modes of participation in the play. In this excerpt, she does not do so, which leads to her interrupting the actor's performance. Since this is the first (and only) case of prompting without an actual blanking in the data, it could be understood longitudinally as a case of initial adaptation to the norms of prompting. Here, the norms are explicitly stated by the actor (give me at least two seconds, l.011-014), supported by the director (l.012), and ratified by the assistant director (l.013). Thus, the participants negotiate the conditions to avoid such interruptions of the play by prompting in the future. Therefore, the main task of the assistant director in prompting is revealed: to locally establish how and when to prompt, since the members treat unnecessary individual interruptions as unwanted actions during rehearsals.
The interrupted state of the play is presented here as the actor falls out of his play frame (see Goffman, 1974). He directs his body orientation toward the assistant director (Fig. 4), addresses her directly, and rebukes her with non-play speech material. The director and the assistant director also participate with nonplay material in this negotiation sequence, indicating their understanding of the briefly interrupted play activity. Thus, the prompting failed from the participant's In contrast, in excerpt 3, which happens only 10 s after the prior excerpt, the actor is blanking while looking at the assistant director. However, since she is reading along in the textbook throughout the sequence, she cannot perceive the actor's visual reorientation from the director to her. Moreover, her actions seem to be guided by the norms just negotiated (giving the actor at least two seconds to resolve the blanking on his own). As a result, the play comes to a standstill again, and the actor has to evoke the prompt with text material outside the play frame. Thus, the excerpt demonstrates that the actor offers a blanking display to the assistant director by gazing at her. The assistant director not using this resource is again related to her not changing her visual participation from reading along to visually monitoring the performance. The fragment also displays that the suggested two-second timeframe is not always effective and that the assistant director must rely on other practices to overcome blanking situations. The above-mentioned markers for blankings can also be observed in this excerpt: verbal pauses (l.003) and hesitation signals (uh, l.004), along with visual addressing of the assistant director by the actor (l.004, Fig. 5). The actor only resolves his maintained gaze toward the assistant director when the latter prompts the text, and the actor begins to integrate the prompted text into his performance (l.006, Fig. 6). In direct comparison with the previous blanking sequences, the actor does not prosodically mark his hesitation signal uh in this excerpt with rising pitch movement. The lengthening of the hesitation signal in this instance is also much shorter than in the previous instances. In excerpt 1, the prompting is successfully timed as the assistant director directs her gaze toward the actor in the next possible slot after the pattern "pause in speech and hesitation signal," however, in the current case, the actor fills this slot by explicitly evoking the prompting ("well now," l.004). This means that the actor calls for the prompt in that sequential position that the assistant director would use for her visual check of the auditory blanking cue. Her quick reaction to his demand in this excerpt shows that it is not a cognitive failure on her part (she is not attentive, etc.) but an interactional problem of interpersonal coordination (the actor has neither prosodically marked his blanking nor given the assistant director time to react).  The explicit verbal evoking of the prompting with non-play material (well now, l.004) is the actor's reaction to the assistant director's lack of mutual gaze. The latter's immediate response to this minimal request provides the needed text (l.005), which allows the actor to resume his performance (l.006-007). By providing the required play text so quickly, the assistant director clarifies that she must have been reading along. Thus, her late reaction concerning the blanking cue "speech pause and hesitation signal" is not because she is not focused on the joint project play rehearsal. It is, presumably-in addition to the described prosodic design and interactional organization of the prompt-also because of simultaneously coordinating two structurally incompatible activities (problem of intrapersonal coordination), each of which requires visual resources for participation. The assistant director takes this into account by subsequently adapting her visual modes of participation in the prompting multiactivity in such a way that from this point on, she usually treats the pattern "pause in speech and hesitation signal" prophylactically as indications of blankings, which she then checks visually. Thus, the play-external evocation of prompting in this and the next case remains the exception (see "Data: Eye Tracking in a Devised Theatre Production" section). Interestingly, it is not the two-second norm proposed by the actor that leads to a successful prompting (as shown in excerpt 1). Instead, a combination of multimodally organized cues (pause, hesitation, and orientation) enable promptings that do not disrupt the play's progress. These cues are accomplished in a two-step process in which the assistant director utilizes the circumstances where pauses, hesitations, and disfluencies are a direct result of the blankings of an actor. Therefore, she first scans the performance for pauses and hesitations, which she then verifies using gaze.

Anticipatory Organization of Multiple Activities in Blanking Situations
As the previous analyses have demonstrated, the assistant director's visual monitoring is crucial for promptings. In this regard, the following three excerpts present cases in which the assistant director visually monitors the actor's play activity after a verbal pause and a hesitation signal. Since this means that she has to pause her participation in her reading along activity for the time, the cases demonstrate the anticipatory organization of the assistant director's visual resources in prompting sequences. In this regard, the following excerpt 4 displays a case in which the actor evokes prompting with the help of a finger snap. This is followed by excerpt 5 in which the same is achieved when the actor freezes in his play. The last case of this paper (excerpt 6) also shows the freezing of the actor in the play but here the assistant director decides against prompting since she (correctly) assumes a dramatic pause. All three cases demonstrate that the assistant director not only analyzes the actor's vocal performance but also heavily relies on her visual interpretation of his performance. To increase the accuracy of the description of her visual coordination, these three fragments use data from the eye tracking glasses worn by the assistant director.
The following excerpt 4 is interesting because it indicates how much time may pass for the actor from his blanking to the assistant director's prompting when they are in mutual gaze. In this excerpt, the same pattern of verbal (hesitation signal, pause), prosodic (lengthening), and visual (mutual gaze) features can be observed where the actor and assistant director negotiate the blanking situation. Unlike the previous fragments, the actor evokes prompting not with linguistic material (well now) but through a finger snap.

Excerpt 4: Prompting After Verbal Pause and Finger Snap
After the actor performs a sentence of the script (l.001-002), he breathes in while changing his gaze from the director to the assistant director (l.003). In doing so, he audibly projects the continuation of his performance while visually displaying a blanking. When the actor produces no other script text and does not progress the play activity for a second, the assistant director looks up at the actor (l.004). As the eye tracking data suggest, both hold their mutual gaze for 0.5 s (Fig. 8) until the actor starts to produce a lengthened hesitation signal uh (l.005). Simultaneously, right after the actor's hesitation signal and with his finger snapping (l.005), the assistant director starts to prompt the missing text (l.006). The eye tracking data shows that she looks back at her script as soon as the actor starts to redirect his gaze toward the director (l.006) and repeats the prompted text (l.007). Thus, both participants display that the blanking situation has been overcome and the progressivity of the play activity has been maintained. Both participants give insight into their online analysis of the possible blanking incident. To prompt the text, the assistant director must have been reading along (Fig. 7) and memorizing the script. In addition, even a mutual gaze and a verbal pause of almost two seconds are not enough for her to safely assume a blanking. Conversely, it appears that she needs a further signal to prompt, in this case, the hesitation signal. Simultaneously, the actor treats the mutual gaze as a fundamental requirement of prompting situations. In cases where no such mutual gaze is established (see excerpt 3), he does not allow the assistant director two seconds reaction time but evokes prompting almost immediately.
Moreover, evoking the prompting by finger snapping or linguistic material is a practice with which he can call in promptings that are almost too late. This opens up the question of what function this embodied practice has in contrast with verbal methods of evoking prompting. It is noticeable that all the previously mentioned criteria for prompting (pause, hesitation signal, and mutual gaze) are produced here, but it appears as if the actor would still need to evoke the prompt by snapping his fingers. However, the finger snap could also be understood as an embodied version of a response cry (Goffman, 1978). That would mean that the finger snap would be an additional cue that he is blanking rather than an explicit request for a prompt. This assumption is supported by the temporal organization of the blanking situation. The assistant director's prompt occurs immediately after/with the actor's finger snap. This makes it unlikely that she is reacting to his snapping but to his multimodal display of blanking through a pause, hesitation, and mutual gaze.
As the next excerpt shows, the resource of mutual gaze can be omitted if another resource compensates it. This happens in the following case where the actor freezes in his performance instead of establishing an eye contact with the assistant director.

Excerpt 5: Prompting After a Freeze
As in the other cases, the assistant director reads along in the script while the actor performs script text until he pauses his verbal participation and produces a hesitation signal (l.005). Despite the conjunction and, which could also project the continuation of the play, the assistant director interprets the verbal pause and the following lengthened hesitation signal as a potential blanking and directs her gaze to the actor. In contrast to the previous instances, the actor here neither continues to speak the text nor visually orients toward her. Instead, he stops all play-related movements (e.g., gesticulation) and thereby freezes his performance (Fig. 9). It takes the assistant director 0.2 s to interpret this frozen performance as an actual blanking (Fig. 10). She begins to prompt (I still stand motionless there, l.007), which according to the textbook is the current line. Interestingly, the content of the prompted text also corresponds with the actor's current (motionless) gestalt. It would have been plausible if the assistant director treated this as part of the performance. Regardless, she may decide to prompt here because, during previous run-throughs of the scene, the actor has never frozen his performance at this point before. Toward the end of her prompting TCU, the actor dissolves his frozen gestalt and begins to integrate the prompted sentence into his performance (l.007). In doing so, his play utterance overlaps with her prompting. Since this indicates that the actor can now resolve the blanking situation independently, the assistant director breaks off her prompting and resumes her reading along activity (l.007). In this way, she minimizes the time in which she verbally competes with the play activity and thus reacts to the potential disruption to the performance that may result from her verbal involvement in the play activity. The actor also contributes to maintaining the progressivity of the play by beginning to integrate the prompted text into his performance as soon as he can do so, thus minimizing the time he must freeze his performance. Thus, freezing can be understood as a bodily practice of evoking a prompt. Unlike the finger snapping in the previous case, this is less obtrusive since the actor does not have to use material outside the play's frame. Therefore, evoking a prompt through performance freeze can be understood as a skillful way to display a blanking without leaving the performance frame.
While finger snapping is a visually and acoustically perceptible practice, performance freezing is characterized by a vacancy in the play's progression. Therefore, the assistant director must distinguish performance freezes as prompting indicators from dramatic pauses as theatrical effects, as the following fragment suggests. At the beginning of the excerpt, the actor performs script text (l.001). As the eye tracking data indicates, the assistant director follows the actor's utterances in the script. She maintains this mode of participation until the actor pauses his verbal performance and produces a hesitation signal (l.002). As has already been demonstrated in the previous cases, this represents a pattern for the assistant director with which she can identify a potential blanking and thus requires her visual monitoring. Although most of the lengthened hesitation signal uh is produced with a lower pitch, its onset is notably higher than the surrounding elements, which marks it as something noticeable and could indicate a potential blanking. Consequently, this triggers the assistant director's visual monitoring. Turning to the prompter's visual behavior, the eye tracking enables the reconstruction of the assistant director's very brief glances with which she treats this incident as part of the performance and eventually does not prompt. Right after the actor's hesitation signal following a short pause, she gazes away from her script and orients toward the actor. When her gaze arrives at the actor after 0.3 s, she can see that he is neither orientated toward her nor frozen in his play, but he raises his arms (Fig. 11). Within the first syllable of his next utterance (seine, "his," l.003), she treats the actor's performance as still ongoing and does not prompt. There are at least two explanations for her behavior. First, she has potentially memorized the script text while reading along in the first place and can decide whether the actor continued with the correct script text or not. Second, she possibly identifies the lowering of his arms as part of the performance that corresponds with the script's content (the decreasing saturation) and precedes the death of the patient in the narration.

Excerpt 6: No Prompting After a Dramatic Pause
The excerpt demonstrates that the assistant director puts her reading along activity on hold and switches to visual monitoring of the performance on stage as soon as she detects a pattern of a verbal pause and a hesitation signal. She resumes her reading along activity the moment she perceives an indication that the play's progressivity is not in danger (when the actor continues to perform).
All cases in this section demonstrate the importance of visual monitoring in resolving blanking situations. To identify a potential blanking as such, the prompter relies on prosodic features (e.g., hesitation signals, pauses) and visual cues (e.g., body orientation and movement variations). The latter is especially crucial since not every pause that follows a hesitation signal is necessarily a blanking. Since the prompter can either read along in the script or gaze at the actor, she only stops her reading along activity when she has a strong indication that prompting may be needed. To conclude, the prompter organizes her activities in an anticipatory way in which the reading along activity is synchronized with the actor's performance. The visual monitoring is used only to verify a blanking assumption based on verbal and prosodic features of the actor's performance.

Conclusion: Verbal and Visual Features of Promptings
Blankings are equivalent to joint word searches in a system where the following lexical elements are agreed upon in advance. To overcome blankings, members of a theatre production (notably actors and prompters) utilize the interactive practice of prompting. The practice is a relevant device in theatre rehearsals with which participants maintain the progressivity of the play in a way that actors do not need to leave the play world and enter the real world every time they have problems producing the subsequent elements of the script. As the analyses have displayed, prompting is a complex practice that entails verbal, prosodic, and visual features. These practices are part of a robust procedure of four phases (Fig. 12).
The first phase in this procedure is the blanking indication, where the prompter relies mainly on auditory monitoring of the performance. To decide if an element is part of the script, the prompter is reading along the script. That means that the prompter needs to monitor the actor's vocal performance and synchronize her reading with the performance's progression. As the analysis has demonstrated, the prompter scans the actor's performance for hesitation signals (äh 'uh') that follow or precede a verbal pause. Actors can use lengthening of the hesitation signal to further mark it as not being part of the play without leaving the play world altogether (excerpt 1). As the prompter relies on this prosodic organization in this first phase of the procedure, a prosodically unmarked hesitation signal has the chance of not attracting the prompter's attention. Thus, it can lead to a temporary Overcoming Blanking: Verbal and Visual Features of Prompting… termination of the play activity (excerpt 3). However, as excerpt 6 demonstrates, not every lengthened hesitation signal must indicate a blanking, but it can still be part of the play (a so-called dramatic pause).
In the second phase of the procedure, the prompter changes from auditory to visual monitoring of the performance to visually verify her assumption regarding a potential blanking. Consequently, she aborts her reading along activity for that time and cannot check if the performance still corresponds to the agreed elements of the script. For that reason, she only interrupts the reading along activity when she has a clear (auditory) indication of blanking. In the present data, four features were identified that contribute to the prompter's evaluation of the performance as a blanking or a dramatic pause: mutual gaze between blanking actor and prompter, performance freezes, finger snapping, and verbal utterances that explicitly demand the missing text. These practices show a gradient from less to more intrusiveness and are used by members of a theatre rehearsal as practices "outside the play world" to address the problems of progress within given play world.
Mutual gaze: As excerpts 1 and 4 demonstrate, the prompter decides to treat a lengthened hesitation signal as a blanking indication if the actor and the prompter establish or maintain mutual gaze. In that sense, excerpt 4 also demonstrates that the prompter treats mutual gaze without lengthened hesitation signals as insufficient to identify a blanking. In excerpts 2 and 3, where the prompter did not use visual monitoring and could not establish mutual gaze with the actor, the play was interrupted, and the members left the play world.
Performance freeze: Excerpt 5 introduced performance freeze as a method by which actors can embody a blanking without physically orienting toward the prompter. When actors freeze their performance, they pause every movement. During their freeze, the actors maintain their posture and thus their participation in the current play activity. However, a freeze temporarily puts the performance on hold and could easily terminate the play world situation.
Finger snapping: While mutual gazes and performance freezes can still be seen as part of the play, the actor figuratively already sets foot outside the play frame when he evokes prompting with the help of finger snapping (excerpt 4). Thus, finger snapping as a practice to evoke prompting can be seen as more escalating than the aforementioned techniques.
Verbal utterance: The most escalated practice considered in this case collection represents evoking a prompt via verbally demanding the missing text (uh well now). As in excerpt 3, the actor does this with the help of non-play text and thus briefly leaves the play frame. In a performance situation in front of an audience, this would dissolve the theatrical effect and break the so-called fourth wall-outcomes that prompting tries to avoid.
In the third phase of the blanking procedure, when the blanking has been identified as such, the prompter can decide to provide the missing element. Then the prompter verbalizes what she has been reading along. Finally, in the fourth phase, when the actor integrates the prompt into his performance, the prompter reorients herself toward her script, resumes her reading along activity, and the blanking has been overcome.
Sequentially, prompting could be described as a retro-sequence since it always refers to a "problem source" found in the performance. Thus, prompting means treating something previous as a problem (missing script elements, mistakes, etc.). This becomes especially evident in the onset phase of promptings. As Goodwin and Goodwin (1986: 54f.) describe for word searches in general, recipients tend to gaze at the word-searching person to display an ongoing orientation to the current activity. However, word searches in everyday interactions and rehearsals differ in the institutional nature of the prompting. It is the prompter's predefined role to deliver the following lines of a blanking actor. This is reflected when the actor turns toward explicitly toward the assistant director in blanking situations. Although other members are reading in the script as well, he specifically turns toward that one person whose institutional task is to help him in blanking situations. To accomplish the assigned task, the prompter utilizes the actor's gaze for interpersonal and intrapersonal coordination. Gaze in blanking situations has an interpersonal coordination function in that actors perceive that they are being gazed at and that prompters are "alerted" by their previous and/or current behavior. As an intrapersonal coordination resource, the prompter uses her gaze to distinguish whether an actor is genuinely blanking or he is just performing a dramatic pause.
Prompting is a process based on the division of labor with a structural epistemic and deontic asymmetry within the institution of theatre. It is designed so that the central activity of this institution, the scenic play, can be maintained even if a problem with its progress arises. While the distinction between play text and blanking indication is a crucial feature in blanking sequences, blankings seem to be more of a matter to be resolved by the prompter than the blanking actor himself. This is due to the actor's and prompter's divided responsibility for delivering the right lines, which is related to the way these two have access to the script. An actor is expected to memorize the script cognitively and to be able to perform their lines at the right time. Prompters are not expected to do this because they have the script in front of them. Their responsibility is to read along and be ready to step in if a blanking situation occurs. Thus, in the terminology of repair research (e.g., Bolden, 2011;Schegloff et al., 1977), promptings are self-initiated repairs to enable "resayings" (Schegloff, 2004) for overcoming a problem with an activity's progress. This setting-specific repair organization results from an institutionalized arrangement in which the prompter has access to the text elements that follow a particular dialogue, but the actor does not have such an access.
Different from word searches in everyday interaction, prompters are not supposed to be heard or seen. This property of prompting is already documented in excerpt 1 when the assistant director immediately pauses her prompting (as is displayed when she turns her gaze away from the actor and starts reading along the script again) as soon as the actor overlaps and displays that he can now perform the next line. Different from promptings, participants in word searches have the option to decline an offer made by a coparticipant. Instead, the present data suggests that actors always accept the prompted elements by creating partial or whole repeats of the prompted elements (Robinson & Kevoe-Feldman, 2010). However, when the prompter misinterprets blanking cues (i.e., when actors can resolve a problem of remembering a line by their own), actors can instead display that they have conducted a dramatic pause and that the prompting itself was not necessary.
The analyses demonstrated that prompting only works if the prompter closely monitors the performance for relevant cues to deliver the right missing words at the right time. For this, the prompter draws on both her professional knowledge and the knowledge acquired during theatre rehearsals. Related to the former, the prompter's knowledge allows her to recognize a freeze as something noticeable (Sacks, 1972). This becomes especially evident in the case of a dramatic pause, which can bear a strong formal resemblance to real blankings but does not require prompting. Regarding the knowledge acquired in the course of the rehearsals, excerpts 2 and 3 demonstrate a process of mutual adaptation of the participants. Since Schmidt and Deppermann (2023a) show that "routine coordination of actions is an effect of the accumulation of shared knowledge over joint interactional histories," it is quite conceivable that prompting is more likely to fail in chronologically earlier cases.
The present study builds on existing research on social interaction in theatre settings (Hazel, 2018) in that here, too, theatre is understood as a workplace setting rather than a performance venue. While other studies have looked at the embodied means by which a script is brought to life (Lefebvre, 2018), or how directors use embodied demonstrations and verbal descriptions to instruct actors (Schmidt & Deppermann, 2023b), the novelty of this study is that it focuses on a function that has received little to no attention in current research: the prompter. The prompter is interesting because s/he primarily has a coordinating function within the work environment. In the present study, this is especially evident in her handling of the different relevancies of the multiactivity configuration (Haddington et al., 2014), i.e., paying attention to the current performance and reading along in the script simultaneously in order to prompt when an actor goes blank. In the way the prompter prioritizes some of these activities, the central requirement to maintain the progressivity of the current performance becomes apparent. As a potentially central principle in theatrical rehearsals, this requirement is analytically sidelined if we focus exclusively on the interactions between actors and directors. Also, analyzing promptings allows us to recognize the boundaries of the performance, i.e., what is still considered "in play" for participants and when they are out of their play frame. Thus, the paper shows that seeming standstills in progressivity (e.g., in gestalt freezes) does not necessarily imply an interruption of a performance. Instead, actors and prompters work together to make blankings accountable to each other in order to overcome them through prompting.
Funding Open Access funding enabled and organized by Projekt DEAL. None.

Availability of Data and Material
No public release for data protection reasons.
Code Availability Not applicable.

Conflict of interest
The authors declare that they have no conflict of interest.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.