The prototypical pointing gesture (cf. pure pointing, Kendon, 1980; pointing-out, Lakoff, 1987; canonical pointing, Langacker, 2008) is a bodily movement toward a target in order to direct someone’s attention to it (e.g., Eco, 1976, p. 119; Clark, 2003; Cooperrider, 2023; Cooperrider et al., 2018; Kendon, 2004; Kita, 2003). The target of pointing is presumed to be visual and present in the speech situation; if the pointing gesture is co-produced with speech, it is taken for granted that what is pointed to is in some sense identical to what is simultaneously referred to in speech (cf. Clark et al., 1983). An extended index finger is often considered to be the origin of pointing and is connected to the target via an imaginary line or trajectory (e.g., Enfield, 2009; Kita, 2003; McNeill et al., 1993). However, the preference for pointing with the index finger is not universal (e.g., Cooperrider et al., 2018; Wilkins, 2003), and pointing comprises a much broader range of bodily actions involving the thumb, hands, extended arm, head, face (e.g., lip-pointing), and objects (e.g., Cooperrider, 2023; Kendon, 2004). Moreover, even within cultures in which the gesture prototypically takes the form of an extended index finger, other forms can also be used, potentially revealing functional differences (cf. Kendon, 2004). The gesture is usually characterized by a “post-stroke hold” (Kita et al., 1998) or “stasis” (Cooperrider, Forthcoming), a brief visual suspension of the gesture from a dynamic to a static position before being retracted or beginning the next gesture (cf. Kendon, 1980, 2004; Bressem and Ladewig, 2011).

Whereas prototypical pointing is used to make a reference to entities that are physically present in the immediate space of the interaction, such as an object, a person, a location, or a direction (e.g., Clark, 2003; Kendon, 2004; Kita, 2003), pointing can also be directed toward a seemingly empty space to provide new references (e.g., abstract deixis, McNeill, 1992; McNeill et al., 1993; Deixis am Phantasma, Stukenbrock, 2014). In addition, the pointed-at object can metonymically represent the intended (discourse) referent (e.g., Haviland, 2000; Levinson, 2006). As noted by Cooperrider (2014), such metonymic pointing is relatively well attested in the ethnographic literature on pointing, leading researchers to suggest that this type of metonymy “may be a pervasive feature of pointing in real-world settings” (p. 3).

While some studies of pointing gestures have addressed their deictic referential function (e.g., Clark, 2003; Kendon, 2004; Kita, 2003), other studies have revealed interactional practices that are accomplished by pointing (e.g., Bangerter, 2004; Cooperrider, 2011, 2014, 2016; Enfield et al., 2007; Goodwin, 2003; Healy, 2012; Hindmarsh and Heath, 2000; Holler, 2010; Mondada, 2007, 2012, 2014; Streeck, 2017; Yasui, 2023). For example, Mondada (2007) observed a work meeting involving a team of agronomists and computer scientists, and noted that the speakers used pointing gestures directed toward maps and other documents as a turn-taking device: In the turn-initial position, pointing indicated incipient speakership; in the pre-turn-initial position, pointing could be used as a claim for the next turn before the prior turn had been completed.

The trajectory of a pointing gesture can single out one of the participants in a conversation. As prototypical pointing is presumed to bring the recipient’s attention to a pointed-at entity (e.g., Kita, 2003), the pointing at the addressee raises some questions, such as under what circumstances the speaker would request the addressee to pay attention to themselves, or whether the addressee is the true target of the pointing. Several studies have shown that a pointed-at co-participant can stand in an (apparent) metonymic relation of speaker for utterance (e.g., Ishino, 2009); for example, when the speaker cites what the conversational participant has just said (e.g., Bavelas et al., 1992). Other studies have shown that English-speakers may use a pointing gesture to indicate agreement with a pointed-at person (e.g., Healy, 2012). Some scholars (e.g., Bavelas et al., 1995; Enfield et al., 2007) have suggested that, in the course of interaction, the gesture can be used to specify the addressee of an utterance in order to elicit their response. However, pointing at the addressee appears to be a frequent phenomenon in dyadic face-to-face interactions (as the current study attests), in which singling out a co-participant as the intended recipient of a certain action is irrelevant.

The current study explores interactional practices, beyond indexing the reference, that are accomplished by index-finger pointing directed at the addressee by Hebrew speakers. The following sections first introduce the data and methodology (section 2). Subsequently, in order for the target phenomenon to emerge as a distinct one, several cases in which pointing at the addressee is used for deictic referential functions are provided (section 3). Non-referential pointing at the addressee—the cases in which the relationship between what is pointed at and what is said is not straightforward—is then elaborated on (section 4). Following this, the findings are discussed, showing that these pointing gestures appear to differ functionally from the canonical case (section 5). Finally, the study is summarized, and concluding remarks are provided (section 6).

The data are drawn from the Haifa Multimodal Corpus of Spoken Hebrew (Maschler et al., 2024), which comprises approximately 22 h of video recordings of naturalistic conversations among friends and family members that were collected between 2016 and 2023. Informed consent for the collection and publication of the data was obtained from all of the participants. The participants were filmed in natural settings, including their homes, cafés, and workplaces, during casual conversations of approximately 30 min to 2 h. Following the setup of the camera and recording device, the researcher exited the environment, allowing the participants to interact freely without providing any instructions. This design facilitated uninstructed dialogue, enabling the participants to discuss the topics of their choice and fostering authentic social interaction. The present study is based on approximately nine and a half hours (571 min) of talk from 13 conversations—involving 30 speakers in total (nine dyadic, two triadic, and two quadratic conversations)—that were recorded between 2017 and 2021.

Tokens of addressee-directed index-finger pointing gestures were searched manually, excluding those used for indexing reference, such as when they were coordinated with utterances that included indexing the second person or when the pointed-at co-participant stood in an (apparent) metonymic relation of speaker for utterance or performer for action (see Section 3). The collection of addressee-directed index-finger pointing gestures, produced without any verbal reference to the addressee and not analyzed as metonymic pointing, comprised 81 tokens. These findings reveal that non-referential pointing directed at the addressee occurred at least every 5.3 min on average; ambiguous cases, such as those involving indexing the second person while also occurring in the contexts found to be associated with non-referential pointing, were not included in this count.¹

To elaborate on the actions that were accomplished via such gestures, I employ the methodologies of interactional linguistics (Couper-Kuhlen and Selting, 2018) and multimodal conversation analysis (e.g., Goodwin, 2018; Mondada, 2016). The analysis will consider the position of the gesture within turn and sequence, the accompanying talk, other bodily conduct, the surrounding environment, and the semiotic properties of the gesture—all of which may combine to contextualize the practices accomplished by the gestures at issue.

From the morphological perspective, the addressee-directed pointing gestures were identified by “movement toward” (Eco, 1976, p. 119) the addressee, using an extended finger. Usually, the index-finger morphology was used; however, in rarer cases, pointing at the addressee was accomplished using other fingers, such as the ring finger, while the little finger was also extended (one token), and the little finger (two tokens). These ring-finger and little-finger pointing gestures can be considered as ad hoc forms of pointing that in both cases were driven by “biomechanical ease” (cf. Cooperrider, 2024). The ring-finger pointing appeared when the speaker held a glass in her hand; therefore, it seems that, from the physiological perspective, the index finger (together with the middle finger) was preferred for holding the glass. The little-finger pointing occurred in an interaction involving four participants, and seemed to be influenced by the participants’ seating arrangement. This gesture was produced using the little finger of the right hand when the speaker turned to the participant sitting next to him from the right side. It seems that, in this condition, performing the gesture using the little finger took less effort since the speaker only needed to move the finger slightly to the right; had he pointed using his index finger, he would have had to move his entire hand.

When the index-finger morphology was used, the index finger was clearly protruded more than any other finger (in some cases the middle finger was also extended), with the thumb and other fingers remaining flexed. Flexing and extending varied in degree. Usually, index-finger pointing occurred with pronated orientation, but six times it occurred with supinated orientation. The gesture was produced by pivoting the arm from either the shoulder or from the elbow. Although such morphological differences may have an impact on the interaction between participants (cf. Holler, 2010; Kendon, 2004) and reveal further form-function correlations, the wide morphological variety found in the data revealed only few examples of each type, thus making meaningful quantitative comparisons difficult to accomplish.

Referential pointing occurs when the reference to the pointed-at participant is part of the propositional content. In such cases, the gesture is often coordinated with an utterance that includes indexing the second person. Such pointing is illustrated in Excerpt 1, which is taken from a conversation between two friends, Sigal and Orly. Prior to the segment shown in Excerpt 1, Orly told Sigal that she would like to go the United States, but her visa had expired.

Coordinated with the index-finger pointing at Sigal (Image 1), Orly asks Sigal whether she has a valid visa (lines 1–2). Orly’s utterance includes a verbal reference to Sigal in the form of the second person dative pronoun lax “to you” (line 1); thus, such pointing can be considered to be referential. However, this is an ambiguous case, since the pointing occurs in the context of eliciting a response, in which the gesture was also found to be used without any verbal reference to the second person, as will be shown in Section 4.

Referential pointing can also occur in cases in which the addressee stands in a metonymic relation of speaker for utterance or speaker for action. As in prototypical pointing a pointed-at object is usually under a particular description (Clark, 2003, p. 247), the content of the addressee’s utterance or the action that they perform is often evaluated by a pointer in a verbal component that is co-produced with the pointing gesture.

Excerpt 2 illustrates metonymic pointing in which the addressee represents the content of her previous talk. The example is taken from a conversation between two friends, Naomi and Kelsey. Prior to the segment in Excerpt 2, Kelsey told a story about a couple, friends of hers, who had a big celebration of their marriage proposal as if the proposal was happening in real time. However, the actual proposal had already been made previously and the couple had even set a date for the wedding. Kelsey and Naomi try to understand why they needed to have such an event. Kelsey says that maybe it had something to do with the ring (lines 1, 3); Naomi points at Kelsey with her index finger (Image 2) as she utters ze mamash muzar “that is really weird” (line 4), evaluating the story that Kelsey told as being extremely odd. In this case, the discourse deixis (Cornish, 2011, 2012) is accomplished not only gesturally, but also verbally by deploying the relative pronoun ze “this.”

In Excerpt 3, the index-finger pointing occurred when the pointer treated the addressee’s way of behaving with ridicule. The example is taken from a conversation involving a couple, Alon and Hillel. Alon is sitting with their baby cradled in his arms. After 37 min of conversation, their friend, Einav, joins them. We enter the interaction after Einav has interacted with Hillel and Alon for about six and a half minutes.

Hillel addresses Alon via a question, also attracting his attention via touch (Image 3), expressing surprise that Einav did not notice [that a camera and a recording device were in the room] (lines 1–5). Alon confirms this via naxon “right” (line 6) and Einav asks what it is that she did not notice (line 7). Hillel explains that a camera and a recording device are in the room, and points to these objects (lines 8, 10–11). Einav conveys surprise via ma? “what?” (line 13), while Hillel and Alon gaze at her and smile. Hillel and Alon then start laughing, and Hillel points at Einav with his index finger (Image 4). By pointing at Einav, Hillel metonymically spotlights her behavior, locating it as the cause of laughter (cf. Yasui, 2023) and evaluating it verbally as great (line 15).

In Excerpts 2 and 3, the pointed-at participants metonymically represented what they had said previously or were currently doing, respectively. Such examples of target-referent metonymy can be characterized as involving a “chain of indicating” (Clark, 2003, p. 264), namely static structures that can be examined link by link, in which the speaker indicates the addressee which, in turn, indexes the referent—utterance or action. However, Cooperrider (2014) argues that, from a cognitive perspective, the target-referent metonymy may be better characterized as being driven by compression (Fauconnier and Turner, 2002), which is a feature of conceptual integration. In this case, the pointed-at participants and their actions or utterances are connected via a conceptual relation.

The detailed description and distribution of various types of referential pointing directed at the addressee will be addressed in future research. This paper will now shift its focus to non-referential pointing, which constitutes the central concern of the current study.

The contexts in which the non-referential index-finger pointing directed at the addressee were employed appeared to be diverse. In the majority of the cases (N = 41), the gesture occurred in contexts that shared a broad sense of opposition: some were disaffiliative contexts associated with dispreferred actions, such as disagreement, disconfirmation, and repair, while in others speakers conveyed information that was (assumed to be) contrary to the addresses’ expectations. In these contexts, speakers typically established and maintained a convergent status and stance of epistemic authority (Heritage, 2012). In other contexts, recipients displayed cues of disengagement in the interaction (N = 15) or speakers attempted to elicit a (minimal) response from the addressee (N = 13). Additionally, some occurrences followed the completion of a cognitive process (N = 7), typically related to remembering, while others were associated with interruptions and discourse suspension (N = 5). Sometimes it was challenging to delineate among these categories, as some occurrences fit into more than one category.² In what follows, I will illustrate these contexts.

The study showed that non-referential pointing directed at the addressee occurred in various contexts. In most of the cases (N = 41), the gesture occurred in contexts involving an action that did not support or endorse the co-participant’s stance or point of view. Some of them were disaffiliative contexts associated with dispreferred actions, such as disagreement, disconfirmation, and repair, while in others, the speakers conveyed information that was (assumed to be) contrary to the addresses’ expectations. In these contexts, the speakers typically produced and maintained a convergent status and stance of epistemic authority. Additionally, the gesture occurred when the addressee conveyed cues of lack of engagement in the interaction (N = 15), when they were indicating accomplishment of a cognitive process (N = 7), when the speakers attempted to elicit a (minimal) response from the addressee (N = 13), and when discourse suspension or interruption occurred (N = 5). The analysis of the examples shows that these contexts occasionally had an overlapping nature, making delineation among them challenging.

The pertinent question that emerges is why the gesture in question appears across these various contexts. It can be assumed that, similar to prototypical pointing, usually defined as a bodily movement toward a target in order to direct someone’s attention to it (e.g., Clark, 2003; Cooperrider, 2023; Cooperrider et al., 2018; Eco, 1976; Kendon, 2004; Kita, 2003), the non-referential index-finger pointing gesture directed at the addressee may primarily function as a mechanism for capturing attention. However, while the prototypical function that is commonly attributed to pointing gestures entails redirecting a listener’s attention to a referent (e.g., Cooperrider, 2023) that is most often presumed to be visual and present in the speech situation, the index-finger pointing at the addressee that was examined in this study mainly serves to attract the addressee’s attention to the pointer’s upcoming utterance. Interestingly, research in experimental psychology has shown that recipients are most likely to notice deictic gestures while interpreting speech (e.g., Langton and Bruce, 2000). The heightened awareness of these gestures suggests that they may play a crucial role in directing the recipient’s focus, thereby reinforcing the hypothesis that these gestures function as an attention-drawing device.

Existing research has revealed various motivations for capturing interlocutor’s attention, with this function being assigned to a range of devices (e.g., Aijmer, 2010; Atkinson, 1979; Brinton, 2001; Keenan and Schieffelin, 1976), whether verbal (e.g., address terms, locating directives or notice verbs such as look or listen, interrogatives, demonstratives, and imperatives) or non-verbal (e.g., pointing, touch, snapping gesture, (single) handclaps, the Raised Index Finger gesture, and throat clearing). One of the motivations for using attention-drawing devices may be a speaker’s sense that they are not being listened to (e.g., Aijmer, 2010; Romero Trillo, 1997). This motivation was particularly evident in contexts where the addressee conveyed cues of disengagement (Excerpts 7 and 8), in which the addressee-directed index-finger pointing gesture was employed effectively as an attention-drawing device. In these cases, the pointer sought positive evidence of the addressee’s attention in order to be reassured of their involvement in the interaction. Regaining attention was also relevant in the context of displaying the accomplishment of some cognitive process (Excerpts 9 and 10). In these cases, hesitations or disfluencies frequently arise, involving temporary suspension of flowing speech (e.g., Kosmala, 2021), thus potentially distracting the recipient’s attention.

The addressee-directed pointing gestures share additional features with other gestural and verbal attention-getters. The study showed that, similar to particular attention-getters, the gesture in question can be employed as an abrupt method of interruption (Excerpt 11) or as an attempt to elicit a (minimal) response from the addressee (Excerpts 12 and 13). The participants used the gesture to enter or intrude into the discourse in order to convey the content that they had just recalled or to request clarification or confirmation. In a similar vein, the Raised Index Finger gesture (Inbar, 2022; cf. Uskokovic and Talehgani-Nikazm, 2022) and various notice verbs (e.g., look and listen; Aijmer, 2010; Keenan et al., 1987), in addition to drawing attention to the message, were used to take the floor or to interrupt (e.g., Aijmer, 2010; Brinton, 2001).

Another reason for the use of attention-getting devices has been attributed to the need to emphasize an important part of the utterance in order to ensure that the recipient understands the message correctly (Romero Trillo, 1997; see also Brinton, 2001; Keenan et al., 1987). Bavelas et al. (1995) noted that speakers gesture toward another person to emphasize part of their speech, while Yasui (2023) observed that speakers may produce a pointing gesture to focus on one particular part of their utterance. Some scholars have pointed out that the need to draw attention to an utterance could be motivated by disagreements and that verbal attention-drawing devices could be used in argumentative contexts (e.g., Aijmer, 2010; Brinton, 2001). In fact, in the majority of the cases, the index-finger pointing at the addressee occurred in contexts that involved performing an action that did not align with or support the co-participant’s stance or perspective.

Moreover, this study has shown that disaffiliation was typically conveyed in tandem with establishing and sustaining the status and stance of epistemic authority. The association between epistemic authority and both verbal and non-verbal attention-drawing devices has been established in the literature. For example, Fairclough (2001) observed that speakers can assert their authority via the frequent use of the verbal attention-drawing device look. The connection between epistemic authority, opposition, and pointing has been highlighted in relation to the Raised Index Finger gesture (Inbar, 2022).

Another question to be addressed is what, nonetheless, is driving this direction of pointing. What motivates a speaker to point at the addressee in dyadic interactions, especially when the addressee’s reference is not indicated in speech? It can be assumed that this direction of pointing can be prompted by an appeal shared by all contexts, in that a favorable response or consideration is being sought from the addressee, rather than merely providing information. This appeal can encompass requests for engagement, clarification, suspension, or acknowledgment of the speaker’s perspective. The aim of engaging the addressee in the interaction brings non-referential pointing at the addressee closer to the category of summons (e.g., Pillet-Shore, 2018; Schegloff, 1968, 2002), designed to invite or prompt a response from a co-participant. However, the relation between the gesture and the category of summons warrants a more thorough examination.

This study focused on non-referential index-finger pointing at the addressee, which is a gesture that indexes the addressee for interactional purposes that extend beyond merely indicating a reference. The study revealed the contexts in which non-referential pointing at the addressee occurs in Hebrew talk-in-interaction. To elaborate on its pragmatic functions, the analysis was grounded in a detailed examination of the examples from each context, including the identification and analysis of the multimodal gestalts of which the gesture is part, as well as an exploration of the gesture’s formation and semiosis. The study proposed that the gesture may primarily serve as a means of capturing attention, showing that it shares several characteristics with other verbal and non-verbal attention-drawing devices. Revealing these characteristics and their interrelationships makes it possible to better understand the phenomena of non-referential pointing, pointing directed at the addressee, and attention-drawing devices in general.

The recurrent use of the gesture in particular contexts may plausibly lead to expectations from the addressee concerning other persistent aspects of the information provided, action accomplished, or stance taken by the pointer in such contexts. Consequently, these meanings may become conventionalized and recognizable via this gesture. For example, being frequently produced in contexts where a speaker opposes a co-participant’s stance—while also typically maintaining convergent status and a stance of epistemic authority—the gesture can be reanalyzed as a cue whereby conveying a negative or epistemic stance is recognized. By elaborating on such cases, this study contributes to the growing body of knowledge about multimodal stance-taking (e.g., Andries et al., 2023; Inbar, 2022; Inbar and Maschler, 2023; Newman et al., 2023; Shor and Marmorstein, 2022).

Furthermore, it has been indicated that both gestural and verbal deictic elements may evolve into stance markers via discourse deixis (e.g., Inbar, 2022; Shor and Inbar, 2019). For example, Heine and Kuteva (2002) note that, in various languages, demonstratives provide fertile ground for grammaticalization processes that can lead to the creation of various grammatical elements, including focus particles. These studies reinforce the conceptual link among deixis, focus, and stance. This connection is also highlighted herein.

The study showed that the addressee-directed pointing gesture may entail the blending of two or more semantic or functional categories to form a multifaceted hybrid sign. Moreover, the gesture can be accounted for from different perspectives, while various factors (e.g., morphological differences, seating arrangements, the number of participants) may have an impact on its employment. To move toward a more generalized analysis of the phenomenon and to develop distributional rules, the analysis should be expanded by considering new data. Finally, further research is needed to explore potential cultural differences, variations in genre and gender, and a more detailed examination of the similarities and differences among other verbal and non-verbal means used to attract attention in interaction.