Educational models of knowledge prototypes development

The computational theories of Human Information Processing (HIP) analyze problem solving through sequential processing stages (input decoding, processing, storage in memory, information retrieval, output encoding) and use cognitive structures such as a central unit which controls the information flow, one or more memory storage units, sensory devices and output peripherals (Baddeley 1992).

Representational models study the cognitive formats whereby data can be processed: frames or representations of objects in Minsky, patterns or representations of events in Rumelhart, scripts or representations of sequences of actions in Schank.

The earliest models of communication theory, like the one Broadbent developed in 1958 in Perception and Communication (a model which applied the analogy of computer-based processing of information to the study of attention) distinguish between short-term and long-term memory. In the following model, devised by Atkinson and Shiffrin in 1968, information is first processed in parallel by several sensorial storage units and then sent to a short-term storage unit (STM) with a limited capacity which, in turn, communicates with a long-term storage unit (LTM). Around the Seventies, more complex theories took the place of modal theories of information. One of these was Craik and Lockhart’s theory of Levels of processing (1972) which stated that the deeper the processing of some information, the better it will be remembered. Information processed at the superficial sensory level will produce relatively short-lasting traces; a phonological processing will produce a longer-lasting one, whereas a deep semantic processing will produce the most stable kind of learning, that is understood and meaningful. The prototype of expert systems was the General Problem Solver (GPS) devised by Newell, Simon and Shaw in 1958, which was in charge of solving complex problems by managing few simple heuristics. The latest generation of expert systems is based on the method of production systems, an example of which is the ACT (Adaptive Control of Thought) system, which Anderson devised in 1983. This model of the cognitive architecture describes the dynamics of the information flow inside the cognitive system and includes various types of memory (working memory, which manages the information currently being processed by the system; declarative memory, which contains propositions; and production memory, which regulates the actions carried out) as well as several types of processes (encoding, performance, storage, retrieval, execution, comparison).

The spatial diagrams were derived inductively from a previous research survey in which a group of university students produced spatial diagrams that could be brought down to only a few classes, namely diffusion, hierarchical and comparative models (Santoianni 2006). This emerging data encounters research on knowledge prototypes, which identifies the early phases in knowledge construction in three categories of spatial models (Lambiotte et al. 1989). These spatial models, having a temporal and sequential nature, have been used to derive spatial diagrams 1 and 2 in the form proposed. On the other hand, the spatial model with sequential and derivative characteristics had already been illustrated in the literature concerning the construction of conceptual and mind maps and was used to draw spatial diagrams 2 and 5 in our form (Novak and Gowin 2001; Buzan and Buzan 2003). To draw spatial diagrams 3, 4 and 6 we used simplifications of concepts used in the teaching of maths (Mason et al. 2009).

Logical relations concern the relations between concepts and syntactic structures, but also imply semantic structures. According to conceptual semantics—which is concerned with the forms of internal mental representation that constitute conceptual structure and with the formal relations between this level and other levels of representation—“the question at issue, then, is whether conceptual structure is somehow different from syntax … whether, when we enter the domain of meaning, the rules of the game should be changed” (Jackendoff 1999: 312).

The Grounded theory methodology concerns the discovery of theory from data. Generating a theory from data means that most hypotheses and concepts not only come from the data, but are systematically worked out in relation to the data during the course of the research. The hypotheses are generated by constantly comparing conceptualized data on different levels of abstraction, and these comparisons contain deductive steps (Glaser and Strauss 1967).