To understand these models and their potential contributions fully, it is useful to consider the emergence of SIP in its historical theoretical context.
The theoretical landscape prior to SIP
Although it would not be until the mid-1990s that the Internet was available to the general public, CMC started its adoption in a few professional, educational, and recreational venues in the late 1970s and 1980s. As it appeared in high-tech businesses and proprietary services like CompuServe and Prodigy, in universities and dial-up bulletin-board systems (see Rapaport, 1991), researchers, journalists, and early adopters began to ask how CMC might change communication and how its changes might affect the social processes to which it may be put (e.g., Hiltz & Turoff, 1978; Johansen, Vallee, & Spangler, 1979; Rheingold, 1993).
The new medium was text-based, rather than multimodal. That is, compared with FtF communication, and even to telephone conversations, there were no nonverbal cues accompanying the written messages in CMC. Early on, research considered this characteristic a likely culprit that would make CMC devoid of social and interpersonal richness. The SIP and hyperpersonal models would turn this concern on its head, so to speak, first by addressing how users overcome the lack of nonverbal cues in making their messages sufficiently personal over time, and later, by explaining the actual advantages that accompany the flexibility of communicating via language and only language.
Although CMC sent written messages across vast distances almost instantaneously, it also featured the ability to âstore-and-forwardâ messages asynchronously. Most commonly seen in email, and now text messaging and social network systems, asynchronous communication means that one individual can post a message and it is retained in the CMC system until its intended reader(s) examine it at another time, at their convenience. This characteristic, too, departed from FtF and telephone messaging. Some observers suggested that asynchronous messaging would make it impossible for communicators to make coherent sense of a series of messages and responses (see, e.g., McGrath, 1990). The hyperpersonal model of CMC, in particular, would argue how asynchronous communication and the perceived control over message construction can actually be advantageous and facilitate more desirable messages and enhanced communicative control (Schouten, Valkenburg, & Peter, 2007).
When CMC was new and first being studied, then, much was expected, but much doubt about its interpersonal potential also accompanied the early theories and research about the new medium. Would CMC simply facilitate communication without any particular change, and obviate the need to schedule (or travel to) FtF meetings (Rockart & DeLong, 1988; cf. Vallee et al., 1975)? Would communicating with others remotely, without nonverbal cues, dehumanize its users (see Orcutt & Anderson, 1977)?
To answer these and more middle-range questions about the effects of interactive media, researchers appropriated established theories from teleconferencing research and developed original theories focusing on CMC per se for the purpose of predicting and explaining the likely effects of online interaction, primarily in large and small groups. These positions, as it turned out, were consistent with many positions in nonverbal communication research. The earliest theoretical positions argued that since the vast majority of our emotional expressions relied on the exhibition and detection of nonverbally encoded messages, text-based messaging without nonverbal cues must therefore lack socioemotional expression. Some theorists went farther to suggest that CMC, without the warmth of natural human communication, would lead users to antagonism and hostility with one another. Empirical research, primarily experimental, seemed to support these predictions.
The assumptions and propositions of the âcues-filtered-outâ approaches to CMC (see Culnan & Markus, 1987) have been summarized in numerous publications. Their essential position is that nonverbal cues facilitate a number of functions related to identifying who others are, what their individual personalities are, how they express emotion, and what their utterances mean. As though nonverbal cues have a unique capacity to accomplish these functions, the general argument of these approaches is that CMC, without nonverbal cues, is impoverished or is incapable of supporting these communicative functions (for review, see Walther, 2010). Small group interaction research long held that in order to be successful, groups need to support both task and socioemotional communication. Task communication is the facts, opinions, ideas, and arguments that group members need to surface in order to inform their decision making. Socioemotional messages convey charisma, humor, agreement, and interpersonal regard, and are also considered critical in order for groups to have effective and satisfying conversations. Research concerned itself with the question of whether and how well CMC could support both these aspects of small group interaction.
Empirical support for these positions often involved experiments comparing small groups communicating by CMC or FtF methods for a limited time, and content analysis of transcripts with which to compare verbal communication in both settings. Many such experiments appeared to support the cues-filtered-out perspective.
At about the same time as these theoretical and research-based interpretations of the effects of CMC appeared in the management, information systems, engineering, and psychology literature, stories of a quite different nature appeared, sometimes in academic outlets, and at other times in the popular press. Anecdotes described shy youngsters who found friends online, and remained online, who had never experienced the kinds of best friends as they did in cyberspace. Spontaneous romances arose via text, surprising their participants with emotional intensity, and appalling their friends. Case studies of high-tech firms showed that internal networks were being used as much for play as for work (Ord, 1989; Steinfield, 1986), for exchanging movie reviews as a hobby as much as distributing parking rules to employees. The cases did not fit the theories, and the experiments were at odds with the anecdotes. Although anecdotal examples should not be definitive scientifically, they seemed to reflect the experiences of a growing number of CMC users, while the theories and research seemed to map on to the suspicions about the medium among those who had not engrossed in it as much.
One other departure from the cues-filtered-out approach appeared in a theory of managerial media selection. It was originally referred to, also, as the social information processing theory of CMC. The term, in this case, came from work by Salancik and Pfeffer (1978), who had argued that managersâ perceptions of organizational artifacts are influenced by the social information oneâs coworkers generate in regard to them. That is, managers perceive things due in part to the communication reflecting othersâ perceptions of those same things. Fulk, Steinfield, and colleagues (1987) applied this approach to organizational membersâ perception of the richness of email. They argued, and later demonstrated empirically, that workers viewed emailâs expressiveness based not only on their own apprehension of emailâs capacity, and not only due to its actual features, but to a significant extent their perceptions were affected by the opinions and email-oriented behaviors of other individuals who shared strong sociometric ties to one another. Fulk and colleagues soon renamed their model a social influence theory of CMC (e.g., Fulk, Schmitz, & Steinfield, 1990).