2. DEFINITIONS OF EMPATHY

The definitions of empathy vary significantly across fields, often with conflicting tones even within one discipline. Empathy is generally used to refer to the cognitively complex process in which one is able to stand in the shoes of another person: to understand their perspective, emotions and the situation they are in. For a long time, empathy was considered a distinctively human ability, but lately it has been suggested that some primates, especially chimpanzees, are able to behave in an empathetic manner, too (e.g. Koski & Sterck Reference Koski, Sterck, Sterck and Begeer2010). It has been proposed that the ability to empathize requires the ability to distinguish between the self and the other, and it would thus require a theory of mind. An empathetic individual understands that other individuals have a consciousness, thoughts and feelings different from their own (Hoffmann Reference Hoffman2000, Preston & de Waal Reference Preston and de Waal2002, Tomasello Reference Tomasello2003, Blair Reference Blair2005). (For a critical discussion on the theory of mind, see Zahavi Reference Zahavi2014:99–102.)

A crucial question has been whether empathy should be described as a primarily cognitive or affective phenomenon. Many psychologists have defined empathy as the ability to enter the other person's mind, as ‘the intellectual or imaginative apprehension of another's condition or state of mind’ (Hogan Reference Hogan1969), or as a process of ‘entering another's world without prejudice’ (Rogers Reference Rogers1975:4).

If we conceived of empathy as an ability to understand the other person's perspective, we could easily detect a display of empathy in the following excerpt:

1. (1)

   

In this example, Kirsi is starting to share a troublesome experience: she is explaining to her friends why she has not been able to participate in the events that are being talked about. She describes her experience with generic noun phrases (‘own personal things that like, in process’ ll. 1, 4) together with the so-called zero person construction (‘will take a long time’). These constructions are grammatically unspecified with regard to whose experience is being talked about, although the context makes it clear that Kirsi is sharing her own experience. With these constructions she can, however, open up her own experience so that it is recognizable to the other recipients (Laitinen Reference Laitinen and Helasvuo2006).

Saila responds to Kirsi's turn twice with the particle joo (ll. 2, 6). Sorjonen (Reference Sorjonen2001) suggests that when joo is used as a response to an affiliation-relevant utterance, such as the one Kirsi produces on lines 1 and 4, it merely registers the prior utterance as understood but does not affiliate with it. The joo response is thus, in this context, a linguistic resource with which the participant registers the prior turn and shows acknowledgment of what has been said without treating the utterance as one to be affiliated with (Sorjonen Reference Sorjonen2001:132). If empathy were defined as the ability to understand the other person's perspective, Saila's joo responses could be seen as empathetic. As a matter of fact, if we accepted this as the full definition of empathy, any utterance that claimed an understanding of the other person's turn in a coherent manner would fall under the notion of empathy.

Within psychology, the research line that stresses the affective component of empathy has been more influential than the one focusing on the cognitive side of the process. Rather than seeing empathetic experience as a process of understanding the other's perspective, it has underlined the centrality of ‘feeling with the other’. According to Stotland (Reference Stotland1969), when emapthizing, the recipient reacts in an emotional manner because he or she perceives another person experiencing a certain emotion. This definition made the concept too wide, however, as it made it cover such emotions as pity and schadenfreude (‘malicious pleasure’). Since the 1980s, careful conceptual differences have been established between different kinds of emotional reactions, especially with regard to whether an emotion involved in an empathetic experience is automatic, whether it is cognitively processed or whether it can be seen to be primarily oriented towards the self or the other (SEP 2010; Ruusuvuori Reference Ruusuvuori2005:205–206). Today, several leading psychologists define empathy as a multi-level process, which requires emotional involvement in the other's feelings (Preston & de Waal Reference Preston and de Waal2002, Hoffman Reference Hoffman2000, Singer Reference Singer2006). In this view, empathy consists of a cognitive representation of the other's feelings but it also contains an emotional component (Blair Reference Blair2005, Preston Reference Preston and Woodruff2007, Preston et al. Reference Preston and Woodruff2007).

An important part of this multidimensional process is the primary, automatic orientation to another person's emotional state, which is commonly referred to as emotional contagion. Emotional contagion is based on the functioning of mirror neurons, a sequence of events in which people transfer their emotions to others (Rizzolatti & Graighero Reference Rizzolatti and Graighero2004, Arbib Reference Arbib2012). This kind of emotional experience is based on an automatized copying of the others' emotions. People can end up feeling the same emotions simply because of their physical proximity and the neurological processes thus activated: a smiling face makes one smile, laughter catches on easily, and euphoria and panic spread quickly. Unlike cognitively processed empathy, emotional contagion does not require an understanding of the other person as the source of the emotion; on the contrary, people often think that these emotions originate within themselves (Scheler Reference Scheler1973:22). Emotional contagion has been used to explain, for instance, the process in which a baby starts to cry as a reaction to another baby's crying (see Darwall Reference Darwall1998). The underlying assumption is that children of this age are not yet capable of distinguishing between the self and the other (Hoffman Reference Hoffman2000:65; see also Zahavi Reference Zahavi2014), but that they start to imitate the emotional reactions of their surroundings (Zlatev Reference Zlatev, Racine, Sinha and Itkonen2008:222–224; Hari & Kujala Reference Hari and Kujala2009:466).

In this article, we adopt the definition of empathy as a multidimensional process in which all of the dimensions listed above are intertwined (see also Preston & de Waal Reference Preston and de Waal2002a, Hoffman Reference Hoffman2000, Preston Reference Preston and Woodruff2007). That is, the empathetic process includes an affective and a cognitive dimension as well as the dimension of emotional contagion. We find our position to be in line with that of Eisenberg (2010; see also Eisenberg & Fabes Reference Eisenberg and Fabes1990), who defines an empathetic phenomena as follows:

We thus understand empathetic phenomena as reactions to the other's emotional state or condition that involve both an affective and a cognitive dimension. Some kind of affective reaction stemming from the other's situation is always a requirement for a reaction to be empathetic. However, there also needs to be an understanding of the affect stemming from the other's situation; any display of such an understanding is what we refer to as the cognitive side of the process. A less conscious orientation to the other (‘mirroring’) can, however, also function as a dimension of the affective side of the empathetic process. In this sense, we will operate with the same dimensions as discussed earlier in this section: (i) emotional contagion, (ii) the cognitive dimension, and (iii) the affective dimension.

What is central to our approach is that we are dealing with a multidimensional phenomenon where one of these dimensions can be foregrounded in a given context, and that the orientations to the dimensions can be attested in different combinations. Here, we agree with the conversation analyst Kupetz (Reference Kupetz2014:27), who states:

She also underlines an important fact that comes up clearly in our own data as well: the empirical correlates for the affective and cognitive sides of empathy are not attestable as dichotomies, but are constantly renegotiated in the unfolding of the sequences (Kupetz ibid.).

While we rely on an affect-oriented definition when analyzing the displays of empathy in interaction, we also want to underline the role of empathy as a crucial element of the larger intersubjective machinery at play in all interaction. In addition to analyzing empathy as observable affective displays, we see empathy as a key mechanism in understanding other people and co-operating in an intelligible manner (Zlatev Reference Zlatev, Racine, Sinha and Itkonen2008:223; Hrdy Reference Hrdy2009; Zahavi Reference Zahavi2014). Our linguistic ability is primarily based on joint activity, as well as our ability to understand the other individual's perspective, intentions and emotions; and the core mechanism that enables this is our ability to empathize with the other (Zlatev Reference Zlatev, Racine, Sinha and Itkonen2008:223). Joint activity can thus be seen to rely on empathetic processing, irrespective of whether we actually display empathy or not. In this sense, empathy can be seen as a key element in joint activity and language use: language and interaction ‘force’ us to be empathetic to a certain degree, whether we show any signs of our relating to the others or not.

When empathy is approached from this perspective, it comes close to the definitions of intersubjectivity (compare the way the concept is defined in this special issue). To highlight this larger understanding of empathy is to suggest that empathy does not only involve displays of affective and/or cognitive involvement, but is also a key mechanism in the way that participants manage joint activities and joint agency (see Enfield Reference Enfield2013:35). Although this idea needs to be developed further in another setting, we will briefly come back to the ways in which the construal of the ground (including the relationship between the participants) is related to the sharedness of the ongoing action (see Section 4).

3. THE DIMENSIONS OF EMPATHY IN LIGHT OF CONVERSATION

We will now analyze an excerpt of conversational data to investigate the ways in which the dimensions of empathy are present in everyday interactional encounters and how this can elucidate our understanding of the language–empathy interface. Some observations have been made about empathic ‘devices’ in previous studies on conversation. Couper-Kuhlen (Reference Couper-Kuhlen2012:126) has pointed out in her study on the affiliative responses in complaint stories that prosody is a ‘primordial’ way of expressing empathy. Kupetz's (Reference Kupetz2014) in-depth study of empathy describes various response types that can be seen to display empathy in the affect-laden tellings of personal experiences. (On empathetic responses, see also Ruusuvuori Reference Ruusuvuori2005, Reference Ruusuvuori2007; Hakulinen & Sorjonen Reference Hakulinen and Sorjonen2012; Voutilainen Reference Voutilainen2012.) While we find all these studies highly relevant and will adopt their sequential approach, we want to underline in our analysis the different dimensions of empathy and how they can be foregrounded and oriented to in various ways at different phases of interaction.

In the following analysis, we focus on activities and gestures and, especially, on the grammatical structures that are used when empathy is expressed. As Stivers (Reference Stivers2008) has noted when studying how storytellers convey stance and how their hearers respond to it and convey affiliation, no response type can be said to be affiliating as such (also Couper-Kuhlen Reference Couper-Kuhlen2012:114): what is crucial is the sequential context of the response. This is certainly true about grammatical structures involved in expressing empathy as well (see Kupetz Reference Kupetz2014:22). In other words, even though we are analyzing certain grammatical structures or response types, we are not claiming that they are empathic as such. Our aim is rather to make some observations about the sequential contexts in which the different grammatical and prosodic features can be used to orient to the different dimensions of empathy.

Our data consist of a one-hour of video-recorded face-to-face conversation between two middle-aged Finnish women, Kati and Taru, who are celebrating New Year's Eve at Taru's.Footnote ² They are sitting in armchairs, drinking champagne and talking about various matters. In the middle of the conversation, and at the beginning of the extract we are focusing on here, there is a troubles-telling sequence where Taru starts to complain about a strange rash on her hand. We will analyze this single extract segment by segment, as we want to show in detail how the interaction unfolds; that is, instead of pointing out parts of the conversation where a certain dimension of empathy can be detected, we will go through the entire extract highlighting the complex ways in which the dimensions come into play in different sequential contexts.

At the beginning of the extract, Kati starts to talk about how she does not like to eat crayfish in wintertime because her skin gets very dry (ll. 1–7):Footnote ³

1. (2)

   

At the beginning of the extract, the women realize that they both experience dryness of hands in the wintertime. Kati delivers her observation about her dry hands (ll. 1–3), and Taru produces an overlapping turn in which she says that she shares the experience (l. 4), followed by Kati's specification that her hands are ‘dry and bad’ (l. 5). Taru then repeats that she has been wondering what she has on her hands: she looks down to her hands and says that they feel very dry (ll. 6–7). This turn is followed by a 0.5-second pause, which is followed by Kati's joo (l. 9), a recognition of what has been said. Kati does not encourage Taru to elaborate on her rash, but goes back to her own initial worry. At this point she uses an extreme case formulation (Pomerantz Reference Pomerantz1986) I have all , like it peals off: she is still talking about the dryness of her hands, now saying that all of her skin is dry.

What becomes clear later in the conversation is that Taru is initiating a troubles-telling sequence by calling attention to her rash. Here we refer to Jefferson (Reference Jefferson, Atkinson and Heritage1984, Reference Jefferson1988), who has shown the intricate ways in which troubles-telling sequences are organized, calling attention to the ways in which the participants delicately navigate through these sequences by moving from business-as-usual to trouble, and then back to business-as-usual. According to Jefferson (Reference Jefferson1988:420), troubles-telling consists roughly of the following segments: A) Approach, B) Arrival, C) Delivery, D) Work-Up, E) Close Implicature, and F) Exit. In the extract seen above, Taru is producing turns that are ambiguous in their trouble-implicativeness; in fact Kati either fails to recognize the approach or, due to the ambiguous nature of Taru's turn, chooses to react to it in a neutral way, still keeping the possibility open for a ‘neutral’ business-as-usual (Jefferson Reference Jefferson1988:422–423). (For a more detailed analysis of the same conversation as a troubles-telling sequence, see Etelämäki & Visapää Reference Etelämäki, Visapää, Etelämäki and Couper-Kuhlen2014:481–486.)

Nevertheless, when examining this extract from the perspective of empathy, the way in which the women mimic each other's movements becomes a focus of interest. When Kati first talks about her dry hands, she is looking down at them, moving them and rubbing them softly. Taru almost immediately starts to simulate these movements: she lifts her hands, starts to rub them and looks down at them. She is thus imitating Kati's movements, and then starts to talk about the irritating rash in her own hands.

This could be used as a textbook example of mirroring, which might or might not lead to emotional contagion. As discussed previously, the mirror-neuron system can make us mimic each other's movements, expressions and emotional states. In the example discussed above, mirroring seems to have an interactional consequence: the women first imitate each other's hand movements, and then end up negotiating whose experience will be the focus of talk. Although emotional contagion is hard to track through linguistic and interactional analysis – we cannot know whether the participants are truly emotionally synchronized –, we suggest that the ways in which participants imitate each other's behavior, body postures, prosody and linguistic structures can be a fruitful starting point for observing the synchronization involved in all empathetic understanding. (On resonance, see Du Bois & Giora Reference Du Bois and Giora2014; also Schegloff Reference Schegloff1998.)

As the conversation continues, Taru describes her rash more specifically and announces the trouble (ll. 11–14) (see Jefferson Reference Jefferson1988:423–424). Here, Kati drops her own project, leans over to Taru and produces an announcement response (l. 15), which displays empathy and thus commits her as a troubles-recipient (see Jefferson 1998:425):

1. (3)

   

We can thus see how the troubles-telling sequence is arrived at and responded to in an empathetic way. We suggest that Kati's turn shows an empathetic orientation to the trouble at hand in both a cognitive and an affective manner. First of all, the lexical choices in Kati's turn are affective as such, but it is the prosody in her turn – the turn is produced with a whispering sound – and the way in which she leans closer to look at Taru's hand that suggest affective involvement (see Jefferson Reference Jefferson1988:420). The cognitive orientation to the fact that this affect stems from the other's situation can best be explained by analyzing the pronouns. By using the demonstrative tää in line 13, Taru offers her rash as an observable and interpretable object (Etelämäki Reference Etelämäki2009:34); it is at this point that Kati leans forward to look at it more closely. With the pronoun tää, Taru also indicates that she has primary access to the symptoms on her hand: when she refers to her hand and her worry with tää and tämmöst, she marks it to be her subjective experience which she is sharing with the other (Etelämäki ibid.). When Kati then produces her response ‘it is quite bad oh my god it's bad’ (ll. 15–16), she uses the pronoun se, which construes the hand as known enough for present purposes (Etelämäki Reference Etelämäki2009:34). The worry about the hand is now a shared, known referent for both of the interactants, and Taru's rash has become the shared object of concern.

Taru continues to describe the symptoms more carefully and thus moves to the exposition phase of the sequence (ll. 18–19). According to Jefferson (Reference Jefferson1988), the exposition is typically followed by ‘affiliative turns’ (‘with expressions of “empathy”’, p. 42). Here, however, Kati gives advice: Taru should go and see a doctor (l. 20).

1. (4)

   

Kati delivers her advice by using an imperative construction käy lääkäris näyttämäs (‘go show it to the doctor’, l. 20). With the imperative, she can be seen as imposing her own perspective to suggest an action that Taru controls (see Lauranto Reference Lauranto2013:181). She does not adopt Taru's perspective but rather gives advice, which instantiates an asymmetry between the advice giver and the recipient (see Heritage & Lindström Reference Heritage, Lindström and Locher2012b:170). Her turn could, in other words, be seen as having a pro-social, sympathetic motivation that aims to help the other, but it cannot be considered to be empathetic.

Taru responds by saying that she has already visited the doctor but that he or she didn't ‘have anything to say’ (ll. 21–22). Kati responds to this turn with a subjectless third person construction eikä tiedä (‘and doesn't know’) (l. 23). After the advice seen in the previous excerpt, she is now producing an ‘affiliative response’ to Taru's exposition (Jefferson 1998:428). Jefferson considers this to be typically the emotionally heightened segment of the troubles-telling sequence. Kati's turn is interpreted in an affective way due to its sequential position, and through linguistic analysis, we can see how she adopts Taru's perspective. Semantically, she reformulates Taru's turn: she produces a turn with an anaphoric person reference where (an implied) third person ‘does not know what's wrong’. The fact that she uses a subjectless third person construction echoes the perspective of Taru's turn: the subject of Kati's turn is interpreted to be the same as that of Taru's; Kati has thus syntactically adopted Taru's perspective. Although Kati's turn is syntactically similar to Taru's and adopts the same subject, she changes the tense from past to present, thus marking her affiliation's relevance for the ongoing moment. Her turn thus orients both to the affective and cognitive dimensions of empathy. Also, the fact that Kati designs her turn to echo Taru's turn can be seen as an example of the synchronization of grammatical structures (see also Du Bois & Giora Reference Du Bois and Giora2014), which is possibly related to emotional contagion.

After this, Taru reformulates what she has said before: although she went to the doctor, it is not exactly that the doctor didn't have a diagnosis but that she has not remembered to call for the results (l. 24–25):

1. (5)

   

What becomes of special interest here are the body positions and the gaze of the women. Throughout her lengthy exposition (ll. 24–33) Taru mostly looks down, changing her position, touching her hand and arm, whereas Kati sits like a Sphinx, leaning back, looking at Taru without a break (see Sacks Reference Sacks, Atkinson and Heritage1984; Rossano Reference Rossano and Stivers2012:310; see also Ruusuvuori & Peräkylä Reference Ruusuvuori and Peräkylä2009), not moving or reacting in any way. In line 28, Taru's reaction shows that a response is indeed missing: after 3.7 seconds, she raises her gaze from her hands to quickly look at Kati. When Kati still does not move or say anything, Taru continues to describe her symptoms. After another long pause (1.4 seconds) in line 31, Taru again takes the turn, this time describing how she feels like wrapping her hand in something woolen. Here, Kati produces an overlapping turn toivottavast toi ei liity nyt jotenkin siihen ‘hopefully it is not connected to’, a diagnostic turn that she does not finish. Taru shows, however, with her response nii mm m (l. 36) that she orients to shared knowledge; she claims to recognize the scenario implied in Kati's turn, and the women are thus diagnosing the rash together. The lexical choice of Kati's next turn kammottavaa ‘dreadful’, as well as the way in which the turn is prosodically produced, displays strong affect and suggests that the shared knowledge refers to something unpleasant. As the pronoun toi in Kati's turn (l. 34) construes a shared access to the referent and as Taru does not deny the diagnosis and affect displayed in Kati's turns, both the cognitive and affective dimensions of empathy are oriented to.Footnote ⁴

4. THE LINGUISTIC CONSTRUAL OF THE GROUND AND THE DEGREE OF EMPATHETIC SHARING

Next, we look at the same conversation from a cognitive grammar perspective, paying attention to the construal of the ground. In CG, ground is the term for the speech event, including its immediate physical circumstances (place and time) as well as the participants and their shared knowledge (Langacker Reference Langacker1987, Reference Langacker1999). We are especially interested in whether the ground is symmetrically or asymmetrically construed, since it has been suggested that the more symmetrical its construal, the more it offers a place for a shared, affective perspective, as the participants (speaker and recipient) merge as a single conceptualizer (Etelämäki & Visapää Reference Etelämäki, Visapää, Etelämäki and Couper-Kuhlen2014).

The process where linguistic expressions are linked to the ground is called grounding. Languages have a set of elements for this purpose: grounding elements. Grounding elements are the means for specifying the relation between the expression and the speech situation (Langacker Reference Langacker1999:219–220). These include tense markers, pronouns, demonstratives and particles. In what follows, we focus on the use of demonstratives (tämä ‘this’, se ‘it’), tenses and pronouns and the ways in which they construe the relation between the participants, the utterance and the speech event.

While mainstream CG analyzes the grounding of individual NPs (ungrounded dog vs. grounded the dog) or finite clauses (ungrounded to speak vs. grounded he speaks) produced by a single speaker who construes the ground by him/herself, our starting point is that the ground is dynamically co-construed (Etelämäki et al. Reference Etelämäki, Herlin, Jaakola and Visapää2009; Etelämäki & Visapää Reference Etelämäki, Visapää, Etelämäki and Couper-Kuhlen2014: 493–495; also Langacker Reference Langacker, Visapää, Kalliokoski and Sorva2014). What is of interest then, from the perspective of empathy and ‘feeling together’, is to what degree the participants conceptualize the ground and themselves as part of the expression, and how they open up places for other interlocutors.

The two other concepts of CG we need to bring up are the notions of subjectivity and objectivity. These concepts call attention to the role of observation by highlighting whether an entity has a role as an object or a subject of observation. When an entity is construed objectively, it is the focus of attention; when it is construed subjectively, it remains implicit, and has a role as part of the observation process. Langacker (Reference Langacker2008:77) illustrates this with an analogy: when one wears glasses, the glasses are part of one's visual system and therefore subjectively construed and non-salient. The glasses become salient, however, when taken off the nose, looked at and thus construed objectively.

Similarly, the ground can be construed objectively or subjectively in a linguistic expression. The ground gets objectified when deictic expressions that make specific reference to the speech event are put ‘onstage’, that is, when they are profiled in the expression. Expressions that focus the attention to the aspects of the ground – such as the participants (especially first and second person pronouns), the place of the speech event (e.g. here) or the time of speaking (e.g. the temporal adverb now) – construe the ground objectively. The ground can also be activated with less explicit grammatical markers – grounding predications (in English, e.g. tense, articles and demonstratives). Here, some aspect of the ground functions as the reference point of the utterance without being explicitly mentioned; it is thus construed more subjectively. The third group is formed by non-deictic elements that include no conventionalized indexical reference to the speech situation (e.g. table, between); this is what is referred to as the maximally subjective construal of the ground (Langacker Reference Langacker1999:219–220)

When the ground is objectively construed by, e.g. first-person and second-person personal pronouns, the conceptualization of the participants becomes asymmetrical: with the pronoun I, it is the currently speaking participant that profiles him/herself in the ground as the focus of attention. Similarly, even the more subtle ways of referring to the ground always suggest an asymmetrical construal of the ground (Etelämäki & Visapää Reference Etelämäki, Visapää, Etelämäki and Couper-Kuhlen2014). Conversely, the less a turn is grounded, the less an expression posits grammatical asymmetries between the participants; that is, the construal is open for a fleeting ‘shared’ perspective, and can thereby invite shared understandings of the evaluated situations.

What we would like to suggest is that there could be a correlation between the linguistic construal of the ground and the degree of empathetic sharing between the interactants. On the basis of our observations, the more symmetrical the ground between the participants, the more the participants are ‘feeling together’. The construal of the ground could thus provide one solution to the me–us problem pointed out by Enfield (Reference Enfield2013), a solution that shows how the separateness and boundedness of the participants is navigated through grammatical means.

Let us revisit the data. From the perspective of construal, the first phase of the conversation can be seen in lines 1–14. In the entirety of this extract, both participants profile themselves as part of their turns using first person pronouns (e.g. mä ‘I’ in Taru's turns in lines 6, 11, 13; mu(l)(l)on ‘I have’ in Kati's lines 2, 9, and Taru's lines 4, 6, 11, 12). They thus explicitly objectify themselves in the ground. They both also objectify their hand as part of the ground, putting it onstage as the focus of attention; however both are markedly talking only about their own hands. From the perspective of grounding, the participants can be seen as construing themselves in a maximally asymmetrical way; both underline their own perspective on the matter. This is further supported by the use of the demonstrative pronoun tämä in Taru's turn in line 13, as well as Kati's nää (plural from of tämä) in line 3; Etelämäki (Reference Etelämäki2009) has shown that tämä typically indexes speaker's primary access to the referent.

The way in which the participants profile themselves in their expressions changes in line 15. Here Taru's hand becomes the primary object of concern for both participants. In lines 15–16, Kati produces a turn (se on ihan pahana herrajjumala ku se on paha ‘it is quite bad oh my god it's bad’), where the ground is more symmetrically organized. Here, the object of concern is referred to with the pronoun se, which indexes that the referent is symmetrically accessible to both participants (Etelämäki Reference Etelämäki2009). After this, however, Kati uses the imperative construction to give advice to Taru: with the imperative verb, she profiles the interaction of the speaker and hearer onstage, but does this markedly from her own perspective. In Taru's response ‘I have been to the doctor’, she profiles herself in the utterance, contesting Kati's suggestion by saying that she has already done what Kati told her to do.

The perspectives merge again in line 23, when Kati produces the turn eikä tiedä ‘doesn't know’ as a response to Taru's turn (l. 22) eikä se osannu sanoo mitään ‘s/he wasn't able say anything’. Here, Kati builds the interpretation of her anaphoric utterance on what Taru has said before, and the personal reference of the turn is interpreted with respect to the ground construal of Taru's turn. Interestingly, though, Kati uses the present tense unlike Taru. This positions the temporal reference of the utterance somewhat differently with respect to the ground, objectifying the present moment and thus marking the affiliation's relevance to the ongoing action.

After this, Taru produces a longer telling about visiting the doctor, and when Kati reacts to this (ll. 34, 37), she construes her turns again in a symmetrical fashion: first by using the pronoun toi ‘that’ in toivottavasti toi ei nyt liity siihen and then, in line 37, by producing the adjectival phrase kammottavaa which can be seen to be maximally subjective in Langackerian terms. As this free-standing phrase (kammottavaa ‘dreadful’) is grounded only by the larger indexical setting of the ongoing action, it construes the ground in a maximally symmetrical way, thus offering a place for shared perspective and affect (see Etelämäki & Visapää Reference Etelämäki, Visapää, Etelämäki and Couper-Kuhlen2014:492–493).

Our analysis of the ways in which the participants construe the ground and its connections to the ways in which they are ‘feeling together’ needs to be elaborated further. We hope to have shown, however, how the way in which the sequences unfold has its reflection also in the grammatical organization of the ground. Whereas empathy can undoubtedly also be displayed with grammatical structures that profile the participants in an asymmetrical position (e.g. ‘I feel for you’, ‘I think your situation is horrible’), our data invites us to see how moments of empathetic sharing are often built on a symmetrical construal of the ground.