Considering the Context to Build Theory in HCI, HRI, and HMC: Explicating Differences in Processes of Communication and Socialization With Social Technologies

our research can be outpaced by developments in the modern technological landscape. To address this issue, we often focus our inquiries conceptually rather than technically through an affordance-based approach (Evans et al., 2016; Flanagin, 2020; J. Fox & Gambino, 2021; J. Fox & McEwan, 2017; Gambino et al., 2020; Rodríguez-Hidalgo, 2020; Sundar et al., 2015). An affordance-based approach engages with the concept underlying a feature or use, such as recordability or publicness, to establish a generalizable effect across features, platforms, and media. We strongly endorse the social affordance-based approach to the study of social technologies, and it is our aim in this manuscript to demonstrate the complementary utility of contextual factors of digital HMC through the lens of extant communication and social psychology theories.

Here, we advance a third perspective that incorporates both the former and the latter. Namely, that communication scholars are uniquely positioned to build HMC theories through consideration of the relationship between contextual factors in HMC and those in theories of communication and relationships.

We focus on the differences between the HMC and interpersonal context, advancing from description to theoretical potential. We first present our broader argument by engaging with processes of socialization, drawing heavily on social learning theory (Bandura, 1977, 1989). We then examine processes of interpersonal communication, focusing on contextually driven differences in goal structure that underlie message production (Dillard & Solomon, 2000).

We often define our research categorically, based on broad contextual differences in the persons (e.g., interpersonal communication, small-group communication) or the content of the communication act (e.g., health communication, political communication).

we argue that explication of the HMC context in relation to the interpersonal corollary is necessary to provide meaningful and nuanced explanations for groups of findings, and, ultimately, to build theories of both HMC and interpersonal communication.

In the case of long-term socialization, we focus our discussion on social learning theory and how consideration of the digital context provides ground for novel inquiries. In the case of message production, considering the context of digital HMC both provide a novel method of studying the interpersonal phenomenon of goal structures and allow for a valid assessment of the specific aspects of the framework relevant to the conditions of HMC.

Hence, if the context of digital HMC is constituted by social regularities that are qualitatively or quantitatively different from their interpersonal corollary, it follows that different goal structures should arise and, ultimately, different messages will be produced.

Proper conceptualization of goals in digital HMC requires consideration of core tenets of goals in communication. For example, in most interpersonal communication situations, humans are producers or receivers (Shannon & Weaver, 1949). This remains true in our modern technological landscape, and computers most often serve as a mediator (CMC). Advances in technologies have complicated this formula. Computers are often now considered distinct sources (Sundar & Nass, 2000), and, in HMC, researchers have suggested that computers can serve as active receivers as well (Guzman & Lewis, 2020).

According to Sundar (2008), machines are believed to have less fallible “memory” and

are capable of gathering and processing larger amounts of data than humans.

Because machines lack experiential capacities, such as emotions (Gray et al., 2007), as well as the ability to make social judgments (Sundar & Kim, 2019), secondary goals pertaining to such capacities that are common in interpersonal communication should be less relevant in HMC. Examples of such secondary goals include, but are not limited to, avoiding face threats, relationship maintenance, and impression management (Meyer, 2009). These secondary goals are premised on the target’s inner experience (e.g., face, social judgments, well-being), which machines lack (Gray et al., 2007). Therefore, they should be activated less frequently during digital HMC.

in an experiment where participants were interviewed by either a faceless computer system or a human, participants disclosed more sensitive information, with greater detail, to the computer interviewer (Pickard & Roster, 2020).

In another study where participants interacted with a chatbot designed to make small talk, researchers found the chatbot induced deep self-disclosure from participants during 3 weeks of use (Y.-C. Lee et al., 2020). In their follow-up interviews, participants expressed how carefree they were when answering the chatbot’s sensitive questions, often making reference to the nonjudgmental or feelingless nature of the chatbot.

reduce the cognitive burden of people.

In addition to suppressing these secondary goals common in human communication, digital HMC may trigger novel secondary goals that are not frequently considered during common acts of human communication. First, because machines are deficient in their ability to understand and contextualize human communication, secondary goals related to understandability or efficiency may be triggered. For example, Muresan and Pohl (2019) found that users of Replika reported limitations in its conversational capabilities, and users were therefore concerned about the degree to which a machine would understand them.

digital machines are often high in recordability (e.g., digital or digitized messages are often stored in a database). Such a high level of recordability may trigger concerns of privacy and confidentiality when the communication involves the disclosure of personal or sensitive information. This may lead persons to consider a secondary goal of information protection, leading to less breadth and depth in self-disclosure.

Specifically, we expect a simpler goal structure in HMC (i.e., fewer secondary goals) as compared with interpersonal, human communication.

Communication is context dependent. When our communication partners are digital machines, rather than humans, how do we change our processes of message production? To answer this question, it is beneficial, if not imperative, to understand the impact of context on message production. Originating from the view of message production as a goal-driven process (for a review, see Meyer, 2021), and in an effort to provide a more comprehensive and useful understanding of the message production process, Dillard and Solomon (2000) conceptualized communication context “in terms of perceived empirical regularities in social reality (i.e., social densities) and the configurations of interpersonal goals that follow from them (i.e., goal structures)” (p. 167)

In sum, contextual factors influence message production and design by dictating the primary and secondary goals. As such, variations of social affordances and broader contextual factors of digital HMC may trigger or inhibit certain goals, leading to differences in both the content of the primary and secondary goals, as well as the overall complexity of the goal structure.

With the primary and secondary goals of relational maintenance and development suppressed in such interactions, we expect to observe fewer relational talks, discussions on complex issues, and lower self-disclosure depth which may result in a person developing less social-emotional skills, particularly those related to narratives and emotions.

On a more micro level, with considerations for a machine as a social other, such as the machine’s well-being, judgments, face, availability (i.e., secondary goals) suppressed, we expect higher directness and lower politeness in messages sent to machines than to humans under the same primary goals. We also expect people to engage in digital HMC with fewer temporal and spatial constraints.

individuals of low cognitive complexity or low cognitive resources may handle digital HMC better than human communication. Following, they may choose to achieve goals through digital HMC over human communication, or they may choose to engage more frequently with computers.