2.1 Theoretical foundations of person-centered theory

Person-centered theory is grounded in a version of psychological constructivism. The strongest theoretical strain in psychological constructivism is personal construct theory, originally formulated by Kelly (1955/1991; also see Chiari and Nuzzo 1996; Neimeyer 2001; Raskin 2002). Personal construct theory postulates that knowledge about people and social life is embedded in bipolar symbolic structures, called interpretive or personal constructs. Personal constructs (a) are a function of a person’s unique experience, (b) constantly produce and reproduce daily realities, (c) are the basic structures through which people evaluate and anticipate social interactions, and (d) become increasingly more differentiated, abstract, and integrated with age and social experience (Delia, Kline and Burleson 1979). The operational mechanism of these personal construct structures is cognitive complexity, usually assessed via the Role Category Questionnaire (Crockett 1965). The move toward construct structures as the organizing epistemic principle for interpersonal communication and thus also person-centered theory is significant because that shift moved the theory toward a cognitive theory of communication with a distinct focus on the individual as the primary unit of analysis.

Applegate (1980) first formulated the hierarchical structure of interpersonal interaction by fusing Delia’s constructivist approach with Bernstein’s (1971) sociolinguistic code theory. Burleson (1982) subsequently adapted person-centered message characteristics to supportive encounters that define roles and expectations for helpers and emotionally upset people who experience mild-to-moderate daily stress. Perhaps the most important contribution of person-centered theory to social support and supportive communication is not its actual conceptualization. Supportive messages that express compassion and validate emotions should be perceived as more beneficial. Indeed, this notion was articulated by Carl Rogers (1959, 1995). What is so ingenious about the theory is to arrange supportive messages in progressive order. Notably, Carkhuff’s (1969) Empathy Scale, developed in response to Roger’s work and used by most US psychotherapists, reflects a similar hierarchical arrangement of empathic responses. Yet what makes the person-centered hierarchy so unique for interpersonal communication is its distinct focus on tangible messages. Because of its utility, the person-centered hierarchy of support messages has been influential in supportive communication.

2.2 Support for person-centered theory

People easily discern more and less person-centered messages (High and Dillard 2012). Studies reporting these findings relied on hypothetical scenario designs and preformulated comforting message rating formats. Two interaction studies detected similar results: Samter (1984, 1985) found that cognitively complex participants who comforted a confederate feigning distress used more sophisticated comforting messages which were evaluated as more sensitive than participants who used less sophisticated messages. Jones and Guerrero (2000) found that distressed people reported feeling better after having received highly person-centered comfort from a confederate helper.

Although people consistently rank supportive messages on the basis of person-centered characteristics, there are moderating factors, including sex, gender orientation, and ethnic group (Burleson 2003). A recent iteration of person-centered theory has integrated these moderators into a dual-process theory of supportive message outcomes, which proposes that the impact of supportive communication is a joint function of message quality and how recipients process messages (Bodie and Burleson 2008). When messages are processed more extensively, they appear to have a greater impact on long-lasting outcomes that assist in future coping (for a review see Bodie 2013a).

3.1 Person-centered and position-centered dimensions of supportive messages

Our first trajectory concerns the basis of person-centered theory: Bernstein’s sociolinguistic code theory. We wonder whether Bernstein’s two codes, namely the restricted code (position-centered talk) and the elaborated code (person-centered talk) are two distinct functional dimensions or whether they are merely two anchors of one dimension that captures differences in supportive talk. According to Bernstein, when people learn a language they also learn a taken-for-granted stock of knowledge that regulates everyday social interactions in unique social groups (e.g., the family, a work team, a group of friends). Using 1950s British social class structure, Bernstein differentiated between a restricted code used by working class people and an elaborate code spoken by middle class people. The restricted code (re)produces talk that focuses on social roles as they are defined by the social group. The elaborated code focuses on explicit expressions of unique motivations, feelings, and intentions (Atherton 2011). Bernstein’s code theory, which resembles Bourdieux’s notions of habitus (Harker and May 1993), remains influential in continental sociolinguistics and education (see Sriprakash 2011), and has morphed into a social realist theory referred to as Legitimation Code Theory (Maton 2011).

It is worth exploring whether the two codes actually serve different relational functions: Position-centered talk may effectively express the unique, shared reality of two or more people who have developed a close-knit relationship. In that relationship, intimate information “lives on” in people’s everyday idiomatic talk. Details of past shared experiences become socially-shared markers that characterize the intimate nature of the relationship between people. Elaborated talk, on the other hand, may be most effective when shared experiences (i.e., relational turning points) do not yet exist. In that case, people are more inclined to provide a detailed narrative. Atherton (2011) applies these functional code differences to family conversations. He states,

Within the comforting context, person-centered and position-centered talk may thus be two message dimensions that serve similar yet somewhat different secondary functions above and beyond the primary goal of ameliorating difficult emotions. Specifically, sophisticated person-centered talk legitimizes and validates the emotional experiences of the upset person in talk. Yet people may possess emotion knowledge that ranges from what is culturally and normatively expected when offering help (e.g., people usually cry when they grieve the loss of a loved one) to idiosyncratic and unique emotion knowledge about the support recipient (e.g., She never cries when she’s sad), that privileges the unique relationship, and that also gets expressed in more and less sophisticated person-centered messages.

The expression of feelings in position-centered speech is typically accomplished through nonverbal channels, particularly the extent to which people effectively use nonverbal immediacy behaviors when expressing support. These behaviors convey the communicator’s willingness to engage in interaction. Immediacy cues include direct eye contact and expressions of vocal and facial warmth. They also include listener-adapted cues, such as backchanneling cues that signal understanding (e.g., headnods). More and less position-centered talk may reflect the (un)skilled use of idiosyncratic knowledge unique to the relationship. So for instance, position-centered talk may range from relational knowledge that consists of scripted, role-based behavior (e.g., normative expressions of sympathy between a sales clerk and a customer) to knowledge that reflects intimate, relational information a supporter possesses about the recipient (e.g., knowing that a recipient is sensitive to touch).

Each message dimension may contribute in slightly different ways to an upset person’s affective improvement, and much like there is a range in the extent to which people effectively use person-centered talk, we would also expect variations in people’s ability to use position-centered support more or less effectively. In short, we propose an integrated, rather than a distributive model of supportive messages: It may not be the case that people use either person-centered or position-centered messages. Rather, it might be the case that people use both person centered and position centered messages in more and less skilled ways. For example, it may well be that a highly position-centered message consists of few words, but highly immediate nonverbal cues that express compassion.

There is initial evidence that person-centered and position-centered talk may be somewhat conceptually and operationally distinct, yet related dimensions of supportive communication. Consider that person-centered studies have put a premium on verbal message exchange as the primary vehicle through which support is conveyed, but empirical research on “being there” shows that the mere presence of another person affects well-being (Dakof and Taylor 1990; Woodgate 2006). Current work on nonverbal expressions of support suggests further that these non-verbal features may be important components of the “feeling better” process (Jones and Wirtz 2006). Furthermore, it seems that nonverbal and verbal expressions of support play different roles in the support process. Jones and Wirtz (2001) found that more and less person-centered messages are primarily responsible for differences in perceived and actual well-being in stranger relationships, whereas immediacy cues other than those that are normatively appropriate in these relationships did not significantly moderate the relationship between verbal person-centered messages and perceived as well as actual well-being. Lastly, the dual-process theory discussed above suggests ways in which those mechanisms may differ. Whether more and less person- and position-centered forms of support play out in interpersonal relationships that vary in function and closeness remains to be empirically seen.

3.2 Person-centered conversations

Our second trajectory concerns the methodological paradigm of person-centered theory. The refreshingly eclectic philosophical framework of constructivism stands in stark contrast to the methodological determinism that has guided empirical research on person-centered theory over the past 25 years. Even though the exponents of constructivism have argued for the application of research methodology that can unfold those interpretive processes which (re)produce social reality and against a research model that stresses variable analysis (Delia 1977; Delia, O’Keefe, and O’Keefe 1982), it is undoubtedly the latter that has dominated person-centered scholarship.

People have little difficulty discerning more and less sophisticated, preformulated person-centered messages such as those presented in Table 1. But it is still an empirical question whether competent supporters actually do talk the sophisticated person-centered talk when asked to help. Some evidence suggests that they do. For example, Samter (1984, 1985) found that participants used what was subsequently coded highly person-centered support in actual conversations. But Metts and colleagues (1995) found that when support providers responded to the feigned distress of confederates, only 4.2% of their responses were highly person-centered. It seems imperative that we examine the presence and nature of person-centered talk in actual conversations. An expedition into person-centered features of actual conversations sparks interesting questions: How many person-centered messages does it take to turn a person’s sadness into no sadness (if we assume that negative emotion is not the absence of positive emotion; see Watson and Tellegen 1985)? Does one highly person-centered message “make up” for two or three moderate messages? Or, in line with Gottman’s (1999) famous 5 : 1 ratio, does it take 5 HPC messages to make up for one LPC message?

Evidently, people do not talk in messages; people have supportive conversations with other people they know well, and it is these conversations and relationships that sustain peoples’ sense of perceived support and that ultimately contribute to well being (cf. Lakey and Orehek 2011). The conversation is the site where the helper can “express, elaborate, and clarify relevant thoughts and feelings” (Burleson and Goldsmith 1998: 260). Burleson and Goldsmith (1998) presented a theory of supportive conversation. In line with stress buffering models which assume that support is most beneficial in direct response to a stressor (for a review see Uchino et al. 2012), the supportive conversation is a process of facilitating reappraisals whereby emotional experiences result from appraisals which, in turn, are about how events are evaluated in the context of personal goals and needs. Comforting works by discursively constructing appraisals over the course of a supportive conversation. To date, the theory remains untested. The problem is that it is difficult to examine the extent to which supportive talk causes emotional change.

Conversational linkages are important in counseling psychology, and a process that could be applied to appraisal talk is motivational interviewing (MI; Miller and Rollnick 2013). In line with appraisal theory, MI postulates that the person is responsible for changing emotion, behavior, and/or thought. However, change is difficult because people possess ambivalent motivations (e.g., wanting to succeed in class, yet also wanting to socialize with friends). These motivations are resistant to change because both motivations are rewarding (and potentially costly). Therefore, direct persuasion is often met with reactance. A first step for the helper is to facilitate an assessment of where the recipient is in terms of wanting to change. MI stipulates five cognitive change stages, namely recontemplation (“I don’t think I have a problem”), contemplation (“Yeah, something’s not right”), preparation (“I think I’m going to do something about that”), action planning (“Here’s what I’m gonna do”), and actual implementation (I’m doin’ it”). Facilitating this assessment ultimately involves the use of person-centered support which ought to fulfill four goals: (a) express empathy, (b) encourage help recipients to experience and express their discrepant motivations, (c) roll with resistance (i.e., accepting resistance to change), and (d) foster self-efficacy. Rolling with resistance is entirely in line with appraisal theory, which suggests that the helper is there to facilitate, not “fix” or resolve problems that may have contributed to stressful and upsetting emotions.

But how do we encourage people to use these skills in everyday supportive encounters? We propose mindfulness as a cognitive, motivational mechanism for person-centered talk. Conceptually, both mindfulness and person-centered support stress intention, attention and awareness of present relational realities (Shapiro et al. 2006). Mindfulness is defined as awareness of and attention to mental stimuli as they present themselves to us in the here and now in a dispassionate, nonjudgmental way (Brown, Ryan and Creswell 2007). Numerous experimental studies showed that mindfulness improves people’s mental and physiological health. Mindfulness may play an important role in appraisal talk (Brown, Ryan, and Creswell 2007; Strohsahl and Wilson 1999). Most helpers experience some anxiety or pressure when they are asked to help (Jones 2011). Fears of not providing the “proper” support, expectations to fix problems, distorted views of support as a fast-acting remedy – all of these preoccupy the helper who is therefore anything else but mindful. Mindfulness may permit person-centered supporters to first become aware of, attentive to, and ultimately let go of their own fears before they engage in person-centered talk (i.e., validating feelings). Mindfulness also provides helpers with fundamental intentional action readiness or a perspective that focuses on other-centered kindness and compassion, and thus the opposite of self-centered performance anxieties (Shapiro et al. 2006). The inclusion of mindfulness deviates from previous assumptions implicit in person-centered theory that the helper ought to attend virtually exclusively to the distressed person through the process of validating emotion. The pressures and expectations that may lead helpers to simply not “be there” or to offer glib comfort may dissipate by fostering simple mindful attunement.

Mindfulness, coupled with motivational interviewing, may provide us with a richer conceptualization of what person-centered talk could look and feel like. It could also help clarify what is meant by conceptualizations of “being there.” Lastly, similar to the many automatic mechanisms we develop over time (e.g., stopping at a red light), mindfulness could easily turn into an effective cognitive heuristic that activates a person’s use of person-centered messages and behaviors. In short, mindfulness may be all that’s needed to activate person-centered constructs and thus the use of person-centered strategies.

3.3 Steady as she goes: Capitalizing on good emotions with person-centered messages

The last trajectory conceived of supportive messages as most efficacious in buffering immediately experienced stressors. Our next trajectory explores the utility of everyday person-centered talk above and beyond its immediate stress-buffering benefits. We propose that person-centered support may contribute significantly to main effects. The main effects model (viz. the stress buffering model) stipulates that a person’s relatively stable perceived support, which is sustained through ordinary, nonstressful social interactions, generates a main effect for health (Lakey and Orehek 2011).

Person-centered support may play an important role in ordinary talk and when we share good news. Of course, we experience mundane (and good) things far more frequently than bad things. However, sharing a day filled with good things may sound trivial at first blush: catching the bus, having a good conversation, a project well done. Nevertheless, these mundane things fill our daily conversations. How people turn to others to share good news is called capitalization (Gable and Reis 2010). Because a person’s perceived support is usually relatively stable and because well-being is a function of how people deal with their emotional experiences, it makes sense to explore the impact of person-centered support in response to everyday minor hassles (i.e., the bad news) and everyday minor positive events.

Supportive communication and capitalization differ in at least one fundamental way: Beneficial supportive communication minimizes negative emotions, whereas successful capitalization support maximizes, or capitalizes on positive emotions. People maintain optimal well-being when they experience, on average, a ratio of three positive emotions to one negative emotion (Frederickson and Losada 2005). In addition, negative and positive events are regulated by two independent processes, namely appetition and aversion. The former deals with rewards and positive emotions, whereas the latter deals with threats and negative emotions (Carver 1996). Positive and negative emotions also serve different functions: Negative emotions orient a person to actions that aim to minimize these emotions.

Frederickson’s (1998) “broaden and build” theory contributes to capitalization and proposes that positive emotions function to broaden the scope of interpersonal construct structures (thought-action repertoires; Frederickson 2001), social bonds, and support resources. Conveying positive events, be they minor or more substantial, has important personal and relational consequences. Positive emotions enhance “broad-minded coping” (i.e., examining the “big picture of a problem, generating multiple solutions), which ultimately reinforce long-term coping abilities and stress resilience. Thus, positive emotions contribute to what Frederickson labels “up-ward spirals” in a person’s well-being (Frederickson and Joiner 2002; Tugade and Frederickson 2004).

The vast amount of mundane and good news is conveyed in close relationships (Gable et al. 2004), and how close others respond to this news is significantly associated with personal well-being and relational satisfaction (Gable, Gonzaga and Strachman 2006). To measure the effects of responses on the capitalizer (the recipient) Gable and her team adapted the active/passive-constructive/destructive accommodation response matrix designed as part of the investment model by Rusbult and colleagues (e.g., Rusbult, Drigotas and Verette 1994). Active-constructive responses are person-centered in nature. In fact, Gable and Reis (2010) described these responses as expressing involvement, excitement or enthusiasm about the good news. These responses also tend to ask open-ended questions that encourage capitalizers to elaborate on the implications and consequences of the positive event. Interestingly, McCullough and Burleson (2012) report initial results of a study that tested specific message features of celebratory support. Two active-constructive features in particular, namely acknowledging feelings and offers to celebrate, were viewed as most effective.

Two response categories tend to reflect low person-centered responses. Passive-destructive responses minimize the relevance of the positive event by not acknowledging it at all, by dismissing any positive emotions associated with the event, and by immediately changing the topic. Additionally, active-destructive responses are those that reflect an active and attentive listener, yet one who will dismiss the positive event or who will point out the negative sides of the event or who will attempt to reframe the event in less favorable ways. Passive-constructive responses are closest to MPC messages because these responses may implicitly indicate a positive attitude yet the responder says very little. Indeed, this response differs from the active-constructive response in level of involvement. Passive-constructive responses do not ask questions or comment on the personal meaningfulness of the event to the capitalizer. As expected, active-constructive responses to positive event disclosures are usually positively associated with personal and relational well-being, whereas the other three response styles are not (Reis et al. 2010).

Capitalization and broaden-and-build theory are part of a growing area of scholarship on positive psychology, which is concerned with the pursuit of understanding optimal human functioning and human strength (see Lopez and Snyder 2011). From the perspective of person-centered theory, it might be most insightful to examine the impact of PC messages on positive event disclosures. A first question concerns the perceived and real impact of PC messages on positive and negative event disclosures. It seems that LPC messages might be more relationally and personally harmful in response to positive event disclosures than in response to negative event disclosures. A distressed person who just shared his or her frustrations will likely feel only worse in response to low person-centered support, but a happy person who just shared his good news will likely not only feel less happy but also frustrated and sad. In other words, in addition to lower levels of positive emotion, capitalizers who receive LPC support may also experience negative emotions. Thus, the personal and relational consequences of LPC responses to good news are more detrimental than LPC responses to bad news. Second, the presence of person-centered support is a relational resource that contributes significantly to one’s perceived social support, which is itself positively associated with personal well-being. So, people who consistently receive HPC support may be better able to sustain adverse events. A final research question concerns the extent to which people actually use high PC messages in response to good news.

3.4 Person-centered supportive communication: Reprieve

Let us summarize the development of PC since its inception in 1980. Early work in supportive communication cast PC as a cognitive construct best operationalized in messages. PC was originally labeled “listener-adapted persuasive strategies” (O’Keefe and Delia 1979: 231), a slight oversimplification that was revised quickly to acknowledge that messages (like people) are more or less “complex” (O’Keefe and Delia 1982). The fact that research on PC in the context of support emanated from persuasion may explain why PC is operationalized in messages. Table 1 shows nomothetic person-centered messages in the form of declarative statements, imperatives, and exclamations with the goal to ameliorate emotional upset, a goal that is analogous to attitudinal change in the context of persuasion (see Bodie 2013a).

Conceptualizations of person-centered comfort have since extended beyond the use of such assertions. As discussed earlier, Burleson and Goldsmith’s (1998) theory of cognitively induced reappraisal has shifted the focus from examining perceived differences in PC messages to the complex discursive choices supporters and recipients make during talk. This shift certainly complicated things for PC, a point that did not elude Burleson. Indeed, in one of the last conversations the second author had with Burleson (who passed away in 2010), Burleson admitted that PC is a “messy” construct and perhaps not so easily captured in preformulated messages. Burleson (2003: 582–583) comments in more detail in a chapter:

PC support has evolved from a mere message on paper into a skill enacted by a helper through the use of a “class” of specific behaviors, or rather “a repertoire of behavioral strategies and tactics” (Burleson 2003: 580). We would like to push this evolution a bit further and argue that what PC looks like in conversation is likely very different than what it looks like in single-shot messages, a point that indeed is reflected already in the model presented by Burleson and Goldsmith (1998). PC is no longer a function of a message but a characteristic behavior of a person and a property of social interaction. This conjecture is somewhat reflected in transcripts from Jones and Wirtz (2006), which show that trained confederates used questions to explore the recipient’s emotions and other aspects of the problematic event. In the HPC example, for instance, the confederate has 14 conversational turns, 10 of which are questions (71.4%). In contrast, the LPC example contains 8 questions over 24 total turns (33%); and, of course, most of the questions from the LPC helper challenge emotions rather than seek to explore them. So there are clearly qualitative differences and differences in types of questions that constitute high and low PC support.

4.1 The beginnings of a research program

The shift from person-centered messages to person-centered helpers is subtle but significant. While the focus on messages largely removes the support provider from systematic scrutiny, the focus on people and the ways in which they enact emotional support opens up a rich set of theoretical and empirical questions. The first link is simple: Producing messages with higher levels of PC and supportive (or active) listening require significant cognitive processing. Evidence for the connection between PC support and skilled listening comes from research that people with higher levels of interpersonal cognitive complexity (ICC) are more adept at producing HPC messages (Burleson 2011). In addition, ICC has been linked to listening comprehension (Beatty and Payne 1984), tendencies to remember conversations (Neuliep and Hazleton 1986), and more sophisticated understanding of supportive communication (Bodie et al. 2011).

The importance of providing a nuanced account of supportive listening is corroborated by findings that people report a preference for interacting with engaged helpers (Jones and Wirtz 2007) and consistently rate listening as an important component of social support (for review see Bodie, Vickery and Gearhart 2013). Support providers who are more attentive and conversationally responsive elicit more detailed disclosures from distressed others (Miller, Berg and Archer 1983) and are more likely to provide appropriate responses to those disclosures (Clark 1993). Consequently, stressed individuals seek out support from others whom they view as particularly good listeners. If good listening is so crucial to support provisions, it behooves us to determine the profile of good listening.

4.2 Mapping the behavioral correlates of supportive listening

In line with personal construct theory, we wanted to examine the personal constructs people associate with good listening and what specific behaviors are associated with these constructs. We sought to uncover components of implicit theories of listening – those cognitive representations central to how listening works and the behaviors that are associated with these attributes (Bodie et al. 2012; Bodie, Vickery and Gearhart 2013). Our theoretical approach was informed by an assumption common to social cognition, namely that people are naïve scientists. Kruglanski’s (1990) lay epistemic theory is an inferential theory about knowledge formation processes and proposes that people validate their hypotheses on the basis of evidence that is either a result of logical (if-then), or probabilistic and statistical inferences (half of all Americans vote). Applied to our context, we wanted to examine listening attributes (e.g., what listening is) and the behavioral indicators that are associated with these attributes (e.g., what listeners do) (Bodie et al. 2012). Much like personal constructs, attributes are beliefs about what an object is and behavioral indicators are beliefs about what an object does (Pavitt and Haight 1985).

In a set of three studies we found that people view competent listeners as possessing five attributes, namely attentiveness, understanding, responsiveness, friendliness, and the ability to sustain conversational flow (Bodie et al. 2012). Attributes, such as intelligence, confidence, humor, and clarity were not highly related to listening competence. These five attributes become salient when judging others as good or bad listeners. Our studies also revealed a range of specific behavioral indicators that are associated with these five attributes and that are thus relevant to supportive conversations: (a) eye contact is primarily associated with attentiveness, (b) smiling and laughing with friendliness, (c) verbal and physical composure with conversational flow, and (d) asking questions with understanding and responsiveness. It seems that people have implicit expectations or mental representations about good listening and subsequently “look for” certain kinds of behaviors that fulfill these expectations.

People’s implicit theories of listening easily integrate with person-centered theory. The five listening attributes map onto explications of “supportive people” who possess the motivation and ability to acquire (a) knowledge about the feelings and emotional states of others; (b) knowledge about human emotion and its dynamics; and (c) knowledge of specific nonverbal, linguistic, and rhetorical resources through which supportive interactions can be realized in specific message strategies (Burleson and Kunkel 1996). The characteristics of supportive listeners mirror these general characteristics: listeners “get the meaning” of their conversational partners; as part of recognizing their knowledge, they “attend to verbal cues” and “offer feedback,” and “make a conscious effort” expressing a “willingness” to listen to the partner.

What our research on implicit theories of listening also suggests is that listening and its attributes/behaviors are expected. Expectations are usually defined as normative behavioral scripts that define social interactions (Jones and Guerrero 2001). People expect to see certain supportive listening behaviors (i.e., establishing eye contact, asking questions) that signal conversational engagement, responsiveness, attentiveness, and understanding. The absence of these supportive listening behaviors may have negative implications for conversational partners and the relationship. The power of expectations is imbedded in the philosophical framework of person-centered theory: To be person-centered is to accommodate the other person. By incorporating principles of accommodation we can further advance the second trajectory outlined above, namely to move the study of supportive communication from a focus on individuals producing and processing messages to dyads co-constructing problems and coping solutions.

4.3 Linking effective supportive listening to outcomes

Carl Roger’s (1959, 1995) philosophy of active-empathic listening permeates work on verbal person centeredness. However, recommendations for “active listening” are extrapolated from counseling (Hutchby 2005; Nugent and Halvorson 1995) with little direct evidence of its effectiveness in interpersonal relationships (Cramer 1987; Gottman et al. 1998). Therefore, we conducted a study to addresses whether active listening behaviors influence important outcomes of informal supportive conversations (Bodie and Vickery 2012). Supportive behaviors contribute to three immediate outcomes, namely generating emotional awareness and changing affect, solving a problem, and reinforcing the relationship between helper and recipient (Goldsmith et al. 2000). To explore the impact of active listening behaviors on these three outcomes, we analyzed 171 supportive conversations. Of particular relevance to our argument are the results of the relative importance of specific active listening behaviors to perceived emotional awareness and affect change.

Active listening includes a range of observable behaviors, including paraphrasing, reflecting feelings, assumption checking, and asking questions, each of which is mentioned by Burleson (2003) as an important behavior in the perception of helpful emotional support. Paraphrases are content and relational summaries that signal understanding (Weger, Castle and Emmett 2010). These summaries are typically prefaced with short introductions that indicate their speculative nature (It seems like …), but these linguistic hedges are not necessary. What is important for competent paraphrasing is that active listeners accurately detect and effectively articulate the upset person’s feeling state (Hutchby 2005). One strategy to assure accuracy is to follow the paraphrase or reflection with assumption checking. These “check-outs” or “tags” are short questions that determine whether the listener has accurately captured the meaning of the discloser’s response (Does that fit for you?) (Baldwin 1987). Finally, active listeners also engage with questioning in the form of open questions which shift the conversation in particular directions (Healing and Bavelas 2011).

We found that perceived emotional awareness is a function of how well a listener paraphrases and reflects feelings, suggesting that emotional awareness is primarily communicated through summary statements that show understanding of a discloser’s content and feelings. On average, the verbal behaviors were 3.31 times more important to the prediction of emotional awareness than were the nonverbal behaviors. For affect change, open questions, check outs, paraphrasing, eye contact, and facial expressions contributed substantively. Although the overall effect was small for any given behavior, on average, the verbal behaviors were stronger predictors of affect change (2.72 times stronger on average) than were nonverbal behaviors, mirroring results from the model predicting emotional awareness.

In line with two other studies (Bodie and Jones 2011; Jones and Guerrero 2001), the results of this study clearly point to the superiority of linguistic listening cues. We now have evidence from three studies suggesting that verbal person-centered cues are (a) more important contributors to impressions of others as “good listeners” (Bodie et al. 2012; Jones and Guerrero 2001) and (b) rated as more important by third-party observers who rated a supportive listener (Bodie and Jones 2011). However, the somewhat arbitrary “verbal” and “nonverbal” labels may not explain these results. Similar to lay epistemic views of listening, Clark (1996) suggests that both verbal and nonverbal listening behaviors contribute jointly to discourse. When we examined our set of supportive conversations, our research team found that conversations with fewer listening indicators (e.g., less eye contact, fewer direct verbal displays of understanding), did not flow as smoothly. In these conversations, stories also were not told as coherently, and disclosers were more likely to repeat themselves.

But why do verbal behaviors contribute more strongly to listening outcomes than do nonverbal behaviors? Clark suggests that contributions to discourse are achieved in two phases, the presentation phase and the acceptance phase. In the presentation phase, one partner offers a proposition which can then be accepted (fully, partially or not at all) by the other partner. As part of the acceptance phase (and quite reminiscent of motivational interviewing we discussed above), listeners can engage in a range of behaviors that provide (in)valid evidence of understanding and thus contribute more (or less) to explicit acceptance of what was said (or lack thereof). Clark lays out four types of positive evidence of understanding, including (1) displays (e.g., using immediacy cues, and verbal statements acknowledging emotions), (2) exemplifications (e.g., paraphrasing, using iconic gestures), (3) assertions (i.e., generic backchannel responses), and (4) presuppositions (i.e., uptaking or initiating the next turn). As noted previously, these types sustain our claim that listening cannot be viewed as independent of the speaker; it is jointly construed to generate shared meaning.

Whether active listening techniques promote beneficial outcomes in supportive conversations (and stress-buffering situations) and how active listening functions for main effects models of support is another empirical question. Our research suggests that traditional listening theories need to be revised. These theories are usually based on stage models of information processing and contain a sequential or parallel set of procedures through which all listeners supposedly traverse when making sense of aural information (for reviews see Bodie et al. 2008; Imhof 2010). These procedures consist of sensing, interpreting, and responding; some models also include processes such as “hearing” prior to sensing and “evaluating”prior to responding (Burleson 2011). Arguably, it is difficult to empirically verify these models; sensing is a theoretical attribute, not a behavioral indicator. In addition, these theories imply that good listening is guided by information-processing tropes that put a premium on comprehension: If a message has been understood, the person has listened effectively (for a review see Bodie 2013b). We may want to move away from the traditional cognitive-structural listening troika (i.e., sensing, responding, interpreting) toward more integrative theories that conceive of listening as a interactive process unique to the human condition, and that are grounded in a socially-shared reality, in our case the reality of the supportive encounter.

4.4 Extending Supportive Listening

Bodie et al. (2008) present an organizational framework from which a meaningful and comprehensive set of listening theories could be launched. Their framework incorporates preconditions for listening (person factors, listening context), the actual listening process (internal mental and overt behavioral process), and listening outcomes (knowledge acquisition, relationship building, affective outcomes). The empirical evidence we accumulated thus far taps some factors and culminates in four points we must attend to when conceptualizing active listening in the context of social support: First, emotion regulation likely plays an important role for supportive listeners. Listeners called on to provide emotional support often have to set aside their own agenda to problem solve or otherwise “fix” the other’s dilemma. Second, the supportive listener not only has to “interpret” information, but he or she also must draw upon shared experience to discover common themes or habits of action, and situate the listening response in the here and now. In short, the supportive listener has to be “mindful”. Third, scholars often invoke the construct as an explanatory mechanism through which support occurs (Lipari 2010). Clearly, there is no dearth in “listening-is-good” claims. But just how does good listening get to matter so much and how does good listening actually work? Lastly, as we pointed out above, supportive listening is what Clark (1996: 212) calls a joint construal. Indeed, constructivism may serve useful for formulations of active listening as a joint construal problem because Constructivism explicates how support provider and recipient co-create meaning and reality.