Abstract
Imagined interactions (IIs), a cognitive process generating mental imagery of conversations, play an important role in interpersonal relationships. This study investigates the frequency of everyday topics of talk represented in imagined talk and the bivariate and multivariate associations with the features of II functions and II attributes. Participants (N = 264) reported their general tendencies using II features and general frequency of topics of everyday talk in IIs in a cross-sectional survey. Bivariate and multivariate correlations were estimated to investigate four exploratory research questions. The results identify consistent patterns of association between II features and everyday topics of imagined talk at two levels of abstraction of everyday talk; relationally centered topics emerge as thematic for understanding II usage and functions. Contributions to the theoretical and methodological framework of IIs are discussed.
Everyday talk—from gossip and reminiscing, to conflict and serious talk—shapes how individuals see themselves and their relational partners (Duck, Rutt, Hoy, & Strejc, 1991). Perceptions of everyday talk impact perceived qualities in family (Schrodt et al., 2007), friends (Ledbetter, Broeckelman-Post, & Krawsczyn, 2011), and romantic partners (Frisby, Byrnes, Mansson, Booth-Butterfield, & Birmingham, 2011). Everyday talk is not just something that happens between people, it is something that constitutes relationships (Goldsmith & Baxter, 1996). Recognizing the importance of everyday talk in perceptions of interpersonal relationships, the topics comprising everyday talk are part of intrapersonal processes including daydreaming and vignettes (Honeycutt, 2003) and cognitive scripts developed to pursue interactional goals in the next everyday interaction (Palomares, 2014).
In imagined interactions (IIs), individual identities and perceptions of relational partners are shaped through the social-cognitive construction and representation of imagined conversation (Honeycutt, 2003, 2008). The features of IIs, including both functions and attributes, recognize that this intentional and goal-directed process shares many similarities with everyday talk like timing, self-dominance, and level of detail, but with one important difference: IIs are created and sustained within the mind of one social actor. Preliminary developments of the II theoretical framework acknowledge these imagined conversations feature a variety of relational partners (Honeycutt, Zagacki, & Edwards, 1990), but investigations of topic variety often rely on recollection of recent II topics (Honeycutt, Vickery, & Hatcher, 2015). The purpose of the current study is to explore variations and frequency of the topics and types everyday (imagined) talk reported in IIs and their association with general reports of II functions and attributes.
Imagined Interactions
IIs function in various ways, allowing individuals to practice, replace, or relive everyday talk. Honeycutt (2003) outlines six common functions of imagined talk and mental imagery including rehearsal, or preparing for an upcoming conversation, and catharsis, or releasing tension and anxiety about interaction. Rehearsal and catharsis are two of the most frequently reported uses of IIs (Honeycutt et al., 2015). Additional functions include conflict-linkage, where imagined talk is used to keep conflict alive through the replaying of past or future disagreements (Honeycutt, 2010); relational maintenance, where IIs help individuals create and maintain close interpersonal relationships; and compensation, or the imagining a conversation that cannot take place due to current restraints in time, place, and space. IIs fulfill multiple functions (Bodie, Honeycutt, & Vickery, 2013); that is, an individual may have an II replaying a past conflict which helps an individual feel better and relieve anxiety about the issue while also preparing him or her for the next time this topic is raised in everyday conversation. Inherent in the fulfillment of multiple functions is the importance of IIs in understanding the self in relation to others; indeed, Honeycutt (2003) argues for the importance of IIs in developing, maintaining, and transforming identities and interpersonal relationships.
The theoretical framework of IIs also identifies common qualities used to describe IIs (Honeycutt, 2008). Attributes identify or characterize common descriptions of IIs, often comparing IIs to actual conversation: The timing of IIs relative to actual conversation results in IIs being classified as proactive (before a conversation) or retroactive (after a conversation; Honeycutt, 2003, 2008). IIs also feature emotional evaluations, classified as either positively valenced or negatively valenced based on the types of emotions represented in the II. When imagining dialogue and conversation, IIs can be either dominated by the social actor imagining the conversation (self-dominated) or dominated by the imagined conversational partner (other-dominated). IIs also vary in discrepancy compared with actual conversation, ranging from similar with little discrepancy to highly discrepant, where what is imagined does not correspond with what happened or will happen in actual conversation (Honeycutt, 2003). Specificity also ranges from general abstract IIs featuring vague mental or verbal imagery to highly specific, where an individual may imagine visual imagery of setting, location, and background features to verbal imagery of topics of conversation (Honeycutt et al., 1990). Frequency captures how relatively frequent or infrequent IIs are over a span of time (Honeycutt et al., 2015). Finally, there is variety, which includes variety in topics, partners, and imagined settings. Topics discussed in IIs might include idiosyncratic topics specific to a particular conversation (e.g., “discussing reservations for the beach vacation” or “fighting about unloading the dishwasher”) or topics may represent more general patterns like making plans, disagreement, or relationship talk which may occur in conjunction with II functions like rehearsal, conflict, or relational maintenance, respectively.
General patterns in II research reveal two common foci, one on extending and investigating II features in relation to a host of individual and relational concepts (Honeycutt, Choi, & DeBerry, 2009; Honeycutt, Zagacki, & Edwards, 1992; Madison, Rold, & Honeycutt, 2014), the other exploring, improving, and refining the theoretical framework of IIs (Bodie et al., 2013; Van Kelegom & Wright, 2013). Perhaps due the appeal and accessibility of the II framework as an investigative tool, current trends in the first research foci have investigated associations between IIs and night dreams (Eldredge, Honeycutt, White, & Standige, 2015) and the use of II features in parasocial relationships (Madison, Porter, & Greule, 2016). The current study, in contrast, focuses on connecting the theoretical framework of IIs back to the interpersonal counterpart of everyday topics, talk, and conversation by focusing on the topics of talk in IIs and how these topics of talk represent attributes like variety, specificity, and frequency, as well as describe IIs focusing on relational maintenance, rehearsal, and conflict, along with other features of IIs.
Everyday Talk
When investigating everyday talk in IIs, the inductive analysis of everyday conversation conducted by Goldsmith and Baxter (1996) generates a potentially useful framework. Across four mixed-method studies, the researchers identify 29 topics of talk recognizable in North American culture, including getting to know you talk, gossip, serious conversation, lectures, sports talk, breaking bad news, and relationship talk (Goldsmith & Baxter, 1996). In addition, their framework organizes specific topics of talk into clusters of similar types of talk (Goldsmith & Baxter, 1996). For instance, getting to know you talk and sports talk are both part of the superficial talk cluster because these topics are typically not deep and involving. Gossip, on the other hand, is an example of informal talk as it involves a speech genre where there is emotional and relational involvement with the topic, partners, and subject of the gossip. Other topics, like serious conversation, breaking bad news, and relationship talk are involving and serious; often, these topics have important individual and relational outcomes. Everyday talk, then, is a goal-directed activity (Palomares, 2014); specifically, one cluster of everyday talk groups together informal goal-directed topics of talk like asking for favors or asking out a potential romantic partner, while another cluster of everyday talk includes more formal goal-directed topics of talk like making decisions, getting or giving instructions, and class information talk(Goldsmith & Baxter, 1996). Other clusters acknowledge variations in the involved emotional nature of talk; clustering positive involving talk like reminiscing, making up, or love talk together; and negative involving talk like serious conversation and conflict together (Goldsmith & Baxter, 1996). Identifying both specific topics of talk and types of talk organized into clusters, Goldsmith and Baxter’s (1996) typology is important for organizing and identifying topics and clusters of talk in everyday relationships.
Explorations of this typology in interpersonal relationships reveals important perceptive differences and patterns. For instance, Ledbetter et al. (2011) found women report more frequency across everyday talk clusters in their same-sex friendships, consistent across various channels including both face-to-face conversation and telephone conversations. In addition, Ledbetter et al. (2011)’s original investigations included reports not only on face-to-face and telephone conversations but subsequent explorations focus on social media (online) conversations as well (Ledbetter & Keating, 2015). These everyday talk clusters were also frequent in online conversations (Ledbetter & Keating, 2015), suggesting that much like face-to-face conversation, online interactions also comprise small talk, gossip, relationship talk, and other types of everyday conversation. Prior studies have found that communication apprehension does not relate to rehearsal (Honeycutt et al., 2009); perhaps instead of individual trait-like differences explaining rehearsal, it may be that individuals who have more conversations with friends across all media platforms and in face-to-face synchronous conversations may be able to better rehearse common topics of everyday talk in their interactions. While prior investigations did not specifically examine text messages, if the findings suggest similar reports in the frequency of everyday talk across the platforms of face-to-face, telephone, and online talk (Ledbetter et al., 2011), then it is likely these topics of talk also occur over text messages. For IIs in particular, the role of online IIs (Honeycutt, 2003) may relate to typing, erasing, and revising text messages when individuals imagine enacting a particular type of talk and their partner’s reaction to the conversation over text.
Recognizing the importance of IIs in interpersonal relationships (Honeycutt, 2008), explorations of everyday talk have also investigated the role of interpersonal relationship quality to account for variations in talk. Topics of talk like catching up, joking around, and recapping the day’s events—topics comprising the informal cluster of talk—are frequent in stepfamilies (Schrodt et al., 2007). These topics of talk could potentially relate to the relational maintenance function: Madison et al. (2014) found associations between relational maintenance and partisan identity; building upon this logic, perhaps the II function of relational maintenance might be associated with deep, involving topics of talk like serious conversation and in-depth discussions where political identity and beliefs are discussed at length. Additional investigations into everyday talk and stepfamily suggest both individual effects and relational effects in the frequency of and satisfaction with everyday talk (Schrodt, Soliz, & Braithwaite, 2008). Perhaps individuals vary in how frequently they report everyday talk because of different levels of using IIs for self-understanding; it might be easier for some individuals to recall which topics are discussed if these individuals are using the same topics of talk in both their imagined and interpersonal interactions. Frisby et al. (2011) examined everyday talk in military and nonmilitary couples and found similarities in frequency of everyday talk in both types of couples, but military couples report higher emphasis placed on everyday talk. Compensatory IIs may be associated with the types of relational talk individuals desire with partners who are not physically available at the moment, such as talking about the relationship or love talk. Recognizing the role of stress in Frisby et al.’s (2011) investigation, there may be important connections with everyday talk and the catharsis function of IIs which helps to relieve anxiety and stress (Honeycutt, 2003). Imagining everyday talk and experiencing stress may also relate to conflict-linkage functions of IIs if those topics imagined are lectures, interrogations, or disagreements.
These potential associations between topics of talk and IIs lead to exploratory research directions guiding the current study. If individuals do vary in their reports of everyday talk from actual conversation (Frisby et al., 2011; Ledbetter et al., 2011; Schrodt et al., 2007, 2008), then individuals may vary in their reports of everyday talk in imagined conversation. This potential variation is represented in the first formal research question (RQ) of this study: RQ1: Which topics of everyday (imagined) talk are reported most often in IIs? RQ2: How are the topics of everyday (imagined) talk associated with all II functions and the II attributes of frequency and variety?
Mirroring the arguments made to suggest associations between the topics of everyday talk and IIs features, there should be similar patterns that emerge when investigating associations between clusters of everyday talk and II features. Rehearsal and proactivity are two frequently reported features recalled in past IIs (Honeycutt et al., 2015). Developing and planning lines of dialogue where primary goals can be met in conversation is an important part of the persuasive message production process (Dillard & Knobloch, 2011). As such, those clusters of talk that emphasize the goal-directed nature of everyday talk may be associated with rehearsal and proactivity, like imagining the clusters of formal goal-directed talk (includes persuasion, decision-making, and lecturing) and relationship talk cluster. Conflict IIs frequently involve retroactive rumination about past conversations (Honeycutt, 2010). Conflict may emerge out of recalling more deep talk, suggesting an association between the conflict-linkage function and this cluster of talk featuring serious talk and complaining, which could lead to disagreement in both real and imagined talk. Related to the general use of retroactivity, there is evidence to suggest that this attribute is associated with greater general conversational sensitivity (Honeycutt et al., 1992). It may be that individuals who engage in more retroactivity and, in turn, catharsis and self-understanding are recalling those goal-directed clusters. While most of these potential associations focus on more goal-directed and highly emotional clusters of talk, the everyday clusters of superficial, informal, and task-related types of talk should also be associated with II features. Compensatory IIs often occur with loved ones that individuals cannot talk to in that current moment (Honeycutt, 2003); having brief or frequent IIs where individuals imagine joking around with a loved one or recapping the day, or even having small talk may be associated with greater reports of compensatory IIs. In general, frequency, specificity, and variety should be positively associated with all clusters of everyday (imagined) talk. After all, being able to recall and report on the details of topics and types of everyday talk inherently represents frequency, specificity, variety. Again, still other associations may emerge, so two final RQs are posed. Each question focuses on one set of II features: RQ3: How are the clusters of everyday (imagined) talk associated with II functions? RQ4: How are the clusters of everyday (imagined) talk associated with II attributes?
Method
Participants
Participants (N = 264) comprised students enrolled in introductory, general education communication studies courses at a public institution in the southern United States. Only a small portion of students in the sample were majors (n = 25, 9.5%) or minors (n = 23, 8.7%) within the department as most participants were affiliated with other academic programs on campus (n = 216, 81.8%). The average age of participants was 20.09 years old (SD = 2.68 years, range 18–47 years old, Mode = 19 years old) and were predominantly women (n = 186, 70.4%). All class ranks were represented including freshmen (n = 96, 36.4%), sophomores (n = 76, 28.8%), juniors (n = 44, 16.7%), and seniors (n = 46, 17.4%), one graduate student (0.4%) and one non-degree-seeking student (0.4%). Participants reported their ethnic identity: The majority of participants identified as White (n = 205, 77.65%). As participants could select multiple ethnic identities, other ethnic and national identities were also represented including African-American or Black (n = 42, 15.9%), Asian-American or Asian (n = 12, 4.5%), Hispanic or Latino/a (n = 12, 4.5%), Native American (n = 4, 1.5%), Middle Eastern (n = 3, 1.1%), and Pacific Islander (n = 2, 0.7%).
Procedures
Participants voluntarily selected this study from a pool of institutional review board-approved and available research studies and received a small portion of class credit in exchange for participating in this study. Upon signing up for this study, participants received a link to an online survey questionnaire. The first page presented participants with an informed consent document describing the study; participants who acknowledged the consent language were directed to provide optional demographic data and complete the survey. Instruments were presented in random order and individual items were also randomized. Upon completion of the questionnaire, participants were given instructions on how to request course credit for completing the survey.
Instrumentation
Survey of IIs
The survey of imagined interactions (SII; Honeycutt, 2003) was used to operationalize II functions and attributes. The instrument includes 60 items representing the 6 functions and 8 attributes, with approximately 3 to 5 items representing each II feature. Items are scaled on a 7-point scale ranging from disagreement (NO!) to agreement (YES!) with a neutral midpoint (?). Reverse-coded items were recoded prior to analysis. The six functions and eight attributes were fit in two separate confirmatory factor analysis (CFA) models as part of the measurement model validity process, with missing items replaced with the mean prior to estimations (1.6% of the total possible values).
The six-factor correlated II function model initially revealed poor fit, χ2 (260) = 743.87, p < .001, comparative fit index (CFI) = .76, standardized root mean squared residual (SRMR) = .09. Low-performing items were identified based on factor weights and standardized residual covariance values above 1.96 and removed; most of these items were the reverse-coded items. The final six-factor correlated II function model achieved improved fit, χ2 (75) = 188.04, p < .001, CFI = .91, SRMR = .07. Reliability estimates generated suggested appropriate internal consistency fit for rehearsal (α = .74, n = 3 items) and relational maintenance (α = .72, n = 3 items), but estimates for catharsis (α = .55, n = 2 items), compensation (α = .69, n = 2 items), conflict (α = .32, n = 2 items), and self-understanding (α = .68, n = 3 items) were below conventional guidelines.
Similarly, the eight-factor correlated II attribute model also initially revealed poor fit, χ2 (532) = 1556.54, p < .001, CFI = .69, SRMR = .11. Following the same procedures for identifying items with poor performance, the removal of these items yielded acceptable fit for the eight-factor correlated model, χ2 (181) = 305.91, p < .001, CFI = .94, SRMR = .05. Reliability estimates yielded acceptable internal consistency for discrepancy (α = .73, n = 3 items), frequency (α = .87, n = 3 items), self-dominance (α = .72, n = 3 items), proactivity (α = .72, n = 3 items), retroactivity (α = .80, n = 4 items), and valence (α = .75, n = 3 items). Specificity (α = .61, n = 2 items) and variety (α = .66, n = 2 items) yielded reliability estimates slightly below conventional thresholds.
Everyday (imagined) talk instrument
The everyday talk instrument used in prior investigations was adapted to operationalize types of talk within IIs (Ledbetter et al., 2011), but all 29 types of talk from the original typology were included. The correlational cluster structure tested in their operationalization was also used to represent the clusters of everyday (imagined) talk, versus the original clusters in the Goldsmith and Baxter (1996) typology. When adapted for this current project, item wording was modified (“I have IIs where I imagine making small talk to pass the time and avoid being rude”) and the scaling for the 29 items was changed from a 5-point scale (strongly disagree to strongly agree) to the 7-point SII scaling (NO! NO; no; ?; yes; YES; YES!). Missing responses were replaced with the mean prior to estimations (1.5% of the total possible values). The initial CFA model was a six-factor correlated model, but this model generated fit statistics below conventional thresholds, χ2 (335) = 881.10, p < .001, CFI = .79, SRMR = .08. Low-performing items were identified based on low factor weights and standardized residual covariance values above 1.96 and removed. The final five-factor correlated model achieved acceptable fit, χ2 (80) = 194.51, p < .001, CFI = .91, SRMR = .06. Reliability estimates for informal talk (α = .79, 4 items) and formal task talk (α = .74, 4 items) were acceptable for internal consistency, but superficial talk (α = .59, 2 items), relational talk (α = .65, 2 items), and deep talk (α = .68, 3 items) were slightly below conventional thresholds.
Results
With a sample size of 264 and α = .05, the power to detect bivariate associations in this study was .37 for small effects (r = .10) and in excess of .99 for both moderate (r = .30) and large (r = .50) effects.
The first RQ sought to explore the general frequency of everyday (imagined) topics of talk. Table 1 reports the means and standard deviations for all 29 types of talk from the original taxonomy. Most topics of talk resulted in mean scores at or near the scale midpoint (4). Love talk was the most frequently reported topic of imagined talk (M = 5.12), and sports talk was the most infrequently reported topic of imagined talk (M = 3.38). Within the superficial talk cluster, current events talk was more frequently reported but was still below the scale midpoint. Almost all topics of talk within the informal talk dimension had scores above the scale midpoint; the most frequent informal talk dimension was catching up (M = 4.60) followed by getting to know you (M = 4.59); least frequent was gossip (M = 3.66). In addition to love talk, other topics of talk within the relational talk dimension with higher than average scores were relationship talk (M = 4.78) and reminiscing (M = 4.50). Within deep involving talk, serious talk was the most common (M = 4.85) and breaking bad news the least commonly reported topic of talk (M = 4.16). Making plans was the most frequent within informal task types of talk, while asking for favor was the least frequent within the same cluster. Additional analyses do not report the informal task dimension because the final measurement model removed all three items due to poor performance. In response to this first RQ, those topics of talk that commonly occur in goal-directed conversations with close relational partners were most common. Many of these topics have importance for emotional expression and evaluation, versus the less frequent informal and superficial topics of talk.
Frequency and Descriptive Data for Everyday (Imagined) Topics of Talk.
aItem removed from final measurement model.
bCluster removed from final measurement model.
*p < .05. **p < .01. ***p < .001.
The second RQ sought to explore associations between everyday topics of (imagined) talk and II features, notably the functions of IIs and two selected attributes. Table 1 reports the zero-order bivariate correlation coefficients between all everyday topics of (imagined) talk and selected II features. Generally, II features shared positive associations with everyday topics of (imagined) talk, with coefficients reflecting small and moderate effects. Catharsis IIs shared positive associations with not only love talk and relationship talk but also interrogation and sharing problems. Catharsis IIs shared positive associations with most topics of talk in the superficial and informal talk clusters, with the exception of getting to know you talk. These associations support the importance of catharsis in emotionally charged imagined conversations. Compensatory IIs and topics comprising relational and deep involving talk clusters did not share many statistically supported relationship; there were no supported associations between compensatory IIs and conflict, serious talk, making up, or relationship talk. Because compensatory IIs make up for a lack of desired interaction, the lack of an association with conflict talk makes sense: Individuals who miss another person would not be likely to miss potential conflict but simply miss talking and interacting. As expected, conflict IIs were positively and moderately associated with conflict talk, along with the other topics captured in the deep involving talk cluster. Conflict IIs, however, were not associated with superficial talk or most topics of informal talk, except for catching up and getting to know you talk. Perhaps it is when people imagine catching up or get to know you that differences and disagreements become apparent. Rehearsal IIs shared no associations with superficial talk like small talk but did share associations with love talk, serious talk, decision-making talk, and making plans. Planning for these encounters in IIs may serve an important relational purpose. Self-understanding IIs were associated with complaining and sharing problems in one cluster and love talk and relationship talk in another cluster. As self-understanding IIs were not associated with most informal or superficial topics of talk, these associations suggest a desire to understand themselves in relation to their current problems more than understanding themselves in typical and superficial imagined conversations.
The third RQ explores associations between clusters of everyday (imagined) talk and II functions. Table 2 reports the bivariate associations between II functions and clusters of everyday talk, with similar patterns observed at this level of abstraction. Like the second RQ, conflict IIs were not statistically associated with superficial and informal talk clusters. Compensatory IIs were not statistically associated with the deep talk cluster, and rehearsal IIs were not associated with the superficial cluster, but the rest of the coefficients were positively associated. Subsequent estimates of canonical (multivariate) correlational models between the clusters of everyday talk and II functions generated two significant roots (see Table 3). The first root, Λ = .47, F (30, 914) = 6.42, p < .001, revealed a moderate association between the two sets of variables representing clusters of everyday talk and II functions, r = .59, r2 = .34. The second root, Λ = .72, F (20, 760.46) = 3.96, p < .001, r = .48, r2 = .23, supported the association between clusters of everyday talk and II functions. Within the first root, lower reported frequency of all five clusters of everyday talk was associated with lower reported II usage on all six functions. The second root reveals a more nuanced association where higher reports of superficial and informal talk with reduced deep involving talk are associated with greater usage of conflict-linkage and relational maintenance and reduced compensation and rehearsal IIs. Here, reports of less emotionally involved imagined talk in IIs were associated with more frequent IIs representing higher stakes interactions with relational partners.
Correlations of II Features and Clusters of Everyday (Imagined) Talk.
Note. 18. Abbreviated presentation of relational maintenance label to improve table fit. 19. Abbreviated presentation of self-understanding label to improve table fit.
aII attribute.
bII function.
*p < .05. **p < .01. ***p < .001.
Canonical Correlation Results: Everyday (Imagined) Talk and II Functions.
Note. Loadings above .30 are presented in boldface. EDIT = everyday imagined talk.
Canonical Correlation Results: Everyday (Imagined) Talk and II Attributes.
Note. Loadings above .30 are presented in boldface. EDIT = everyday imagined talk.
The fourth and final RQ noted the potential associations between the clusters of everyday (imagined) talk and II attributes, extending beyond frequency and variety. The zero-order bivariate associations between the clusters of everyday (imagined) talk and II attributes revealed that variety, valence, specificity, and discrepancy were positively and moderately statistically related to all clusters of everyday (imagined) talk. The coefficients where no statistically supported relationship was observed were all focused on the superficial cluster of talk: Retroactivity, proactivity, and self-dominance shared no association with this cluster. The canonical (multivariate) correlation models between clusters of everyday talk and II attributes generated two significant roots (see Table 4). The first root, Λ = .41, F (40, 987.91) = 5.60, p < .001, r = .68, r2 = .46, supported the moderate association between clusters of everyday talk and II attributes, as did the second root, Λ = .75, F (28, 819.88) = 2.34, p < .001, r = .38, r2 = .14. Coefficients and canonical loadings within the first root suggest reduced reporting of all clusters of everyday (imagined) talk was associated with the reduced frequency in reporting of all II attributes. The second root revealed that less frequent reports of deep talk and more frequent reports of superficial and informal talk were associated with lower reports of retroactivity, self-dominance, specificity, and variety but increased reporting of positive valence IIs. These associations revealed the importance of variety and frequency in IIs: As individuals reported imagining less frequent goal-directed and involved talk, they also reported using fewer characteristics to describe their IIs.
Discussion
The explorations of four RQs reveal important patterns of association between II features and everyday conversation. The results of the four RQs are now discussed, and the contributions of these findings to the theoretical and methodological framework of IIs focusing on the role of everyday talk in representing the verbal and mental imagery of imagined conversations.
First, the frequency of everyday topics of talk suggests great variation in how people imagine the verbal imagery of their IIs. These associations are emphasized in the first and second RQs, focusing on the specific topics of talk imagined in IIs. Much like the taxonomy developed by Goldsmith and Baxter (1996), topics imagined in IIs mirror their conversational counterparts in discussions of serious issues, relational talk, and love talk. The patterns observed emphasis the importance of IIs in intimate relationships, as relationships are truly maintained and kept alive in cognition. Individual topics of talk within IIs represent serious, involved, and goal-directed talk more than superficial topics of talk. One reason for this explanation may rest in cognitive scripts (Honeycutt & Bryan, 2011; Baldwin, 1992). The scripts for small talk, sports talk, current events talk, and even more goal-directed informal talk like asking someone out on a date or making plans are embedded in everyday experiences, meaning that practicing, rehearsing, replaying, or rethinking these smaller moments may go unnoticed when we reflect on general tendencies in mental imagery or may not be worth the same amount of cognitive effort than imagining a higher stakes conversation captured in serious or involved talk. The higher stakes conversations, however, seem to play in important role in IIs. Involving, emotionally laden, and goal-directed talk comprising such topics as serious conversation, conflict, and relationship talk occurs frequently in IIs, which is important for the theoretical framework of IIs.
While imagined conversations are a cognitive and individually centered process, the features of IIs and conversational topics rehearsed and practiced impact interpersonal relationships. IIs have individual benefits in emotional awareness and cognitive appraisal in the self-understanding and catharsis functions, seen in the association between these functions and topics like sharing problems and serious talk. II functions also recognize the relationally minded emphasis of IIs: This focus is not only seen in relational maintenance and compensation but also in conflict-linkage and rehearsal. Preparing for a serious conversation and imagining discussing disagreements with intimate partners to help relationships thrive, as evidenced by the strong associations between these functions and love talk and serious talk. One of the contributions of these findings to the theoretical framework of IIs is the conceptual arguments where all II functions—not only compensation and relational maintenance—may serve important relational functions, supporting Honeycutt (2008)’s argument of IIs “as major wellsprings that create expectations for relationship development” (p. 315). Indeed, this wellspring is represented in all functions, not only compensation or relational maintenance.
The patterns of association observed between everyday (imagined) talk and II attributes suggest another contribution of these findings to II theory. Certainly, the statistically supported associations between II attributes present evidence of convergent validity; the moderate size of these effects is further evidence of this argument. Looking beyond intrarelationships of II attributes, these results contribute to the attribute of variety. One earlier extension of II theory found potential problems with variety in model fit (Bodie et al., 2013). The definition of II variety includes both topics and partners under the same construct (Honeycutt, 2003), whereas these results suggest the need to distinguish and refine variety to include topic variety and partner variety as potentially related but not isomorphic attributes. Items used to represent variety in the SII include items like “I have recurrent IIs with the same individual over the same topic,” a reverse-coded item representing both topic and partner variety in the same question. This item was removed from the final measurement model featured in this study due to poor fit, which supports this contribution. Further, these findings provide further evidence that frequency of II features, specifically topics of imagined talk, can be considered in general tendencies of IIs. This contribution may help future research build prior on qualitative explorations of frequency of II topics (Honeycutt et al., 2015).
Another aspect of II theory is extending and testing the II framework in relation to other cognitive processes. Specificity involves IIs with a greater attention to detail; previously, specificity was found to be associated with conversational sensitivity (Honeycutt et al., 1992). There may be some individual variation that could explain why the superficial dimension of talk does not share a statistical association with many features of IIs. The reason for this pattern could be related to individual differences in personality (Honeycutt, Pence, & Gearhart, 2012) or it could be a function of studying IIs at a general level of abstraction (Van Kelegom & Wright, 2013). Alternatively, this observation could be related to the relational context of IIs. If IIs are indeed common in intimate relationships (Honeycutt, 2003), then perhaps a different investigation of everyday (imagined) talk and newer developing relationships may find more patterns of association. If people are still getting to know someone, they may practice different cognitive scripts and rehearse all conversations, including the more everyday conversations, to produce positive relational outcomes and ensure success at developing intimate relationships.
The current study investigated the general reporting of IIs and the general frequency of topics of talk within these IIs, which is important for both the theoretical and methodological framework of IIs. Notably, the detail comprising the topics of talk provide potential future options for capturing specific topics of talk or even inspiring some revisited ideas in the frequency and specificity attributes. There are items in the original long-form SII that assess verbal imagery, but these items are not directly included in the II attributes that frequently appear in inferential analyses because they are scaled differently. 1 One of these original SII items presents additional challenges to researchers: As an open-ended question asking participants to report on II topics, the results are subject to some thematic or coding procedures to organize information. In addition, the shift from general reporting on II tendencies to recalling recent II episodes changes the framing and abstraction of IIs (Van Kelegom & Wright, 2013). Future research may wish to represent topic variety using items adapted from the everyday talk instrument, using these findings and prior interpersonal investigations as a guide for design and implementation (Frisby et al., 2014; Ledbetter et al., 2011; Schrodt et al., 2008). This recommendation for future adaptation with the goal of advancing II methodology is not without precedent: Some items on the long-form SII do adapt items from Hecht’s (1978) communication satisfaction questionnaire to investigate satisfaction in recent IIs (Honeycutt, 2003).
The methodological contributions of this study to the II framework also relate to some of the noted limitations of this study. Both instruments employed to explore the RQs yielded poor initial measurement model fit. For the everyday talk instrument, some of the removed items might relate to extending the instrument into the context of imagined everyday conversation or the removed items may simply work in the theorized taxonomy but not work as an instrumentation of everyday talk. Future research is needed to determine similarities and differences in this test and prior tests of the everyday talk instrument. This limitation is seen in the exclusion of the cluster of informal goal-directed talk in the latter RQs. Similar concerns emerge with the SII. In this investigations and other investigations using the SII questionnaire, measurement models initially suggest poor fit on common low-performing items, like the reverse-coded items (Bodie et al., 2013; Vickery, Keaton, & Bodie, 2015); those studies reporting reliability estimates also reveal patterns of estimates below conventional thresholds which suggests some item consistency issues (Carpenter, 2016; Honeycutt et al., 1992, 2012; Madison et al., 2014). Future research should consider exploring and integrating the 30 years of research and advancement in analytic techniques which have occurred since the initial scale development and factor analyses used to develop the SII (Honeycutt, 2003; Honeycutt et al, 1992–1993). Future research should use both theoretical and methodological frameworks to consider improving the structure and consistency of the SII. The findings generated from exploratory studies like this one should provide a strong foundation for future predictive and deductive research.
The primary purpose of this article involves testing and refining the theoretical framework of IIs, but these findings also have important implications in light of the secondary research foci present in II research on extending investigations of IIs into related conversational and relational contexts. First, recognizing recent trends in the intersections of IIs and political communication (Madison et al., 2014; Madison, Honeycutt, Covington, & Auter, in press), these findings suggest the micro-level behaviors and conversations that may relate to political identity and everyday political talk. Recognizing the general topics of talk included in this study such as serious conversation, lecture, and interrogation, these general topics may coincide with specific topics related to political communication: For instance, a serious conversation may involve discussing the formation and beliefs associated with voting; one partner may lecture another on recent news and events occurring in Washington, D.C., or one relational partner may interrogate the other to discuss voting beliefs and behaviors, or even lecture on why certain candidates may be better for local or regional appointments. Recognizing the role of cathartic and retroactive IIs in Madison et al.’s (in press) findings, perhaps there are associations with conflict discussions as well, where people revisit and continue to discuss the results of past elections based on news and current events talk related to past election outcomes. Based on the general frequency of such micro-level abstractions of talk, future research should investigate specific topics of talk including conflict, serious conversation, lecture, and interrogation, focusing specifically on everyday political conversations within these topics of talk.
In addition, this potential for future research reveals a limitation of the current study: Like explained in the literature view and rationale, this study considered the general classification of everyday talk into topics (or genres) and did not investigate specific topics within the general classification like political talk or other conversational contexts. Extending beyond political conversation, troubles talk, or sharing troubles is one type of everyday conversation that follows typical norms for turn-taking and presenting information (Jefferson, 1988). Troubles talk is likely associated with breaking bad news, serious conversation, or even relationship talk depending on the nature of the problem or troubles. Recognizing the importance of relationships in the landscape of IIs, future research should incorporate these findings to investigate specific types of problems and topics of talk when relational partners share their problems with others. Perhaps one reason for the relatively low frequency for breaking bad news (and removal from the measurement model) may be that individuals do not rehearse or plan for breaking bad news, but it is something that occurs when catching up with others. In other words, individuals may not be able to plan how to break bad news or seek support using IIs, but instead must react in real-time to others’ troubles and offer support. To unravel these potentials, future research should consider the specific problems represented in bad news and serious conversation such as discussing the stress and anxiety surrounding job rejections, infertility struggles, relationship breakups, or other current struggles being discussed in imagined and interactive conversations. The decision to investigate only general topics of talk, and not specific idiosyncratic topics of talk, may be seen as an area in need of future research but also a limitation of the current study.
Additional limitations should be noted to contextualize this study within the framework of social and cognitive research. While the sample size was appropriate for investigating moderate and large effects in a cross-sectional exploratory study, the sample comprised a convenience sample of college students. Future research, especially that research improving and extending theory, should consider other representative samples. Honeycutt (2008) notes that IIs are a “pervasive part of daily existence that occur across the lifespan” (p. 85). Having a different representative sample may impact the frequency and variety of topics and clusters of everyday talk in future conceptual or exact replications. Young adults may be more apt to engage in class talk from the original taxonomy in their everyday conversations on campus, but individuals not working in education or attending school will likely not imagine class talk at all. Yet, these limitations may not be as impactful or change the main findings of this study focusing on those goal-directed, relationally focused aspects of everyday (imagined) talk and IIs. Also related to this limitation may be the general focus and abstraction of IIs and everyday talk: Perhaps the challenges in everyday superficial and informal talk were because individuals were generalizing across a variety of relational partners in their IIs. Had participants be asked to focus on context- or relational-specific IIs, particularly relationships like workplace relationships or developing friendships or romantic relationships, those everyday topics and types of talk may have been more commonly reported because those sorts of talk help develop our relationships with others (Duck et al., 1991).
The goals of the current study emphasize the importance of exploring the association between everyday (imagined) conversation and the theoretical framework of II features. Through this investigation, the importance of the cognitive and intrapersonal process of IIs in developing, maintaining, and transforming the interpersonal relational process of conversation is revealed in bivariate and multivariate models. The connections between everyday (imagined) talk and everyday talk are emphasized, continuing to connect inter-and intrapersonal communicative practices and behaviors. These contributions to the II theoretical framework provide new ways to specifically consider the attributes of frequency and variety which continue to characterize and describe IIs in the body of research advancing this theoretical framework.
Footnotes
Acknowledgments
An earlier version of this manuscript was submitted to the International Communication Association annual conference. The author would like to thank David F. Wynot for his assistance with data collection.
Declaration of Conflicting Interests
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author received no financial support for the research, authorship, and/or publication of this article.