Many of the affordances associated with computer-based learning environments are due to their interactive nature. Such interactive learning environments do not merely present information like a book or an instructional video; instead, they permit learners to actively engage in order to learn.

Although at first glance it would seem clear what the terms “interaction” and “interactive” mean in the context of learning environments, there is, in reality, little agreement regarding these terms in the corresponding literature. There are almost as many different definitions as articles that attempt to illustrate these terms (see, e.g., Betrancourt 2005; Kennedy 2004; Kettanurak et al. 2001; Mayer 2001; McMillan 2002; Rafaeli 1988; Wagner 1994). Whereas some characterizations of interaction detail how such an interchange between learners and a learning environment might occur (e.g., control over pace, having influence on what will be presented and how this is done), we propose a rather general definition for the purpose of this Special Issue. Similar to Wagner (1994), we define interactions in the context of computer-based learning as a process in which the “actions” executed by learners and their learning environment are mutually dependent on each other. The mutual dependence of actions can be due to technical features of learning environments (e.g. problem-solving interfaces) or to the fact that cooperation with others is afforded by the technology. Examples of interactivity that were considered in this Special Issue are dialoguing, controlling, manipulating, searching, and navigating in learning environments with multiple presentation modes (Moreno and Mayer 2007); sequencing, controlling contents, and representation format in hypermedia environments (Scheiter and Gerjets 2007); pacing (i.e., control over the continuation) of a video-based model; controlling over appearance and task selection with respect to the model (Wouters et al. 2007); predicting, reacting to reflection, or self-explanation prompts in example-based environments (Atkinson and Renkl 2007; Wouters et al. 2007); problem solving with intelligent tutorial assistance (Koedinger and Aleven 2007); acting in virtual inquiry worlds (Nelson and Ketelhut 2007); and computer-meditated dialogic argumentation (Clark et al. 2007).

Although the forms of interactivity considered in these articles involve some overt action, all authors share the conviction that it is not the overt action per se that fosters learning but the cognitive processes that can be elicited by the interactivity. There is also agreement that interactivity can distract the learners from deeply processing the learning materials (see also the commentary of Kalyuga, this issue). For example, the possibility of learners’ selecting their own information in a computer-based learning environment may lead to disorientation, a phenomenon typically referred to as lost in hyperspace (Scheiter and Gertjets 2007). Hence, it is important to consider which aspect of interactivity is crucial for learning. In addressing this issue, we can leverage three theoretical stances that emerge from research on learning which are especially relevant when discussing the issue of interactivity: active responding, active processing, and focused processing. The first two stances were introduced by Mayer and colleagues (e.g., Robins and Mayer 1993) whereas the latter (i.e. focused processing) is a differentiated version of the active processing stance.

The active responding stance considers visible, open learning activities as necessary for effective learning. The quintessential example of this stance is Skinner’s (1954) operant conditioning model in which the immediate reinforcement of correct reactions or answers is seen as crucial. Today, there are “more modern” models of the active responding stance under the labels of constructivism (in a Piagetian sense) or socio-constructivism (referring to Vygotsky) which emphasize open behaviors such as manipulating learning objects, joint problem-solving activity, or actively engaging in academic discourse (e.g., Roschelle and Teasley 1995; Quintana et al. 2006; Stahl et al. 2006). With respect to the issue of interactivity, the active responding stance emphasizes that it is the “visible” exchange between the learners and their learning environment that fosters knowledge acquisition. None of the contributions to this Special Issue, however, adopts this position. Based on their research results and cited literature, all of the contributing authors acknowledge that the cognitive processes induced by interactivity are crucial.

The active processing stance is adopted by most cognitively oriented researchers focused on learning and instruction. It can also be characterized more generally as a constructivist position because it assumes that knowledge cannot be imparted on the learners but has to be actively constructed by information processing in the working memory. The contributions in this Special Issue all acknowledge that interactivity helps learning if it leads not only to overt action but to active processing of the learning contents. In some contributions, this stance—or a slight variation of it—is implicit, while others make explicit reference to it (e.g., Moreno and Mayer 2007: knowledge construction view; Wouters et al. 2007: cognitive interactivity).

We propose a differentiated version of active processing as a third stance in this Introduction: focused processing. The active processing stance does not explicitly specify that it may be crucial that the learners' active processing is related not only to the learning contents but to the central concepts and principles to be learned (e.g., mathematical theorems, physics laws). Unfocussed processing, as it may occur in hypermedia learning, for example, may cost time. If learning time is restricted, reduced learning outcomes may result (see also the commentary of Kalyuga, this issue). Other examples of active but unfocussed processing refer to the acquisition of “seductive details” (Garner et al. 1989) instead of central information in text learning. Mandl et al. (1993) found that vocational students can acquire “malprioritized concepts” in the domain of economics when working with computer-simulated jeans manufacturing. In this case, many learners were so preoccupied with the jeans not selling that they focused on keeping their stock rather empty. Unfortunately, this strategy was counterproductive to the original goal of the task, which was learning to maximize profits.

With regard to interactivity, the focused processing stance implies that the interactive features of a learning environment should not only elicit active processing of learning contents but also focus attention on (a) central concepts and their interrelations, and (b) central principles and their application in problem solving (Wittwer and Renkl, submitted for publication). In all of the contributions to this Special Issue, the authors elaborate on instructional techniques or instructional problems that fit well within a focused processing stance, for example: cueing in modeling environments (Wouters et al. 2007), guided activity principle (Moreno and Mayer 2007), principle-based self-explanation prompts (Atkinson and Renkl 2007), appropriate assistance giving in the form of central information (Koedinger and Aleven 2007), distraction problems and serendipity effects in hypermedia environments (Scheiter and Gerjets 2007), scripting dialogic argumentation for focused exchange (Clark et al. 2007), and ineffectively implemented inquiry learning caused by teachers’ misconceptions such as “hypothesis is a guess” (Nelson and Ketelhut 2007).

We hope that this introduction provides an interesting perspective for reading the following articles by highlighting several theoretical stances on interactivity. We also hope that this introduction spurs your curiosity about this Special Issue and its various contributions.