A theoretical approach to learning with hypertext

Technology in the form of computer-based educational resources has been on the rise in schools, private educational institutions, and universities for the last 20 years. Hypertext is applied in many different areas (e.g. Journalism, Sales,) and owes its popularity to the ability to present complex topic areas directly and comprehensively. Hypertext structures, like that of the CING, promise to be advantageous for learners’ knowledge acquisition (e.g., Spiro & Jehng, 1990; McKnight, Dillon, & Richardson, 1991; Whalley, 1993; Gerdes, 1997; Tergan, 2002). Previous studies of CING usability with students from the University of Chemnitz comparable to the subject group, however, found that its hypertext structure posed problems for some users (Heller, 2004a) (hypothesis 7).

Hypertext systems enable authors to present information on a content area in a networked form with the help of information nodes and links. Its use is often self-guided and influenced by individual interests and [learning] goals and can be considered advanced learning (Spiro, Feltovich, Jacobson, & Coulson, 1991; Jacobson & Spiro, 1994). This form of self-guided usage of networked information presentation, which the CING requires, is likely to have an impact on learners’ application of the program in their learning and will be discussed in more detail in the “Theories of autonomous learning” section, below.

3.3.1 Hypertext

The basic concept of hypertext, from its earliest discussions (Bush, 1945) onwards, has been to divide a content area into its individual information units and present it by means of interrelating nodes and linkages between the units, electronically in a networked format (see Tergan et al., 2000; Tergan & Lechner, 2000; Jacobs, 2004). The presentation can take the form of a branched or dynamic information display. This complex structure of information

enables users to place information in individual contexts as they establish (comprehension) links between the different units they access on their own (see Kuhlen, 1991; Krems, 1999). Networked structures were originally designed as a form of information presentation and (easy) information access for complex domains (Conklin, 1987; Kuhlen, 1991) before they were further developed for educational purposes (Whalley, 1993).

According to Nielsen (1994), the simplest distinction between traditional text and HT is the type of sequence. While traditional text is sequential, i.e., linear with “a single linear sequence” that defines “the order in which the text is to be read” (Nielsen, 1994, p. 1), hypertext is non-sequential, i.e., non-linear (Gerdes, 1997; Tergan, 2000; Tergan & Lechner, 2000; Jacobs, 2004), as “there is no single order that determines the sequence in which the text is to be read” (Nielsen, 1994, p. 1).

The material in a hypertext is represented by the hypertext’s nodes or links (see e.g., Figure 8), which are the units of information the hypertext can hold, for instance, in a content page (Nielsen, 1994) (e.g., the CING content page Use of Perfect). These nodes are connected to other nodes in the HT through “pointers” (Nielsen, 1994) or “links” (Kuhlen, 1991; Rouet, Levonen, Dillon & Spiro, 1996; Tergan, 1997; Tergan & Lechner, 2000), which are not fixed in advance but rely on the content of each node (see McAleese, 1990). This means that a node can contain one or several links depending on its content’s relevance or interrelation to other nodes in the HT (Nielsen, 1994) (Figure 19).

Figure 19: “Simplified view of a small hypertext structure” (Nielsen, 1994, p. 1)

Movement along these links and nodes in the network of a hypertext “is often referred to as browsing or navigating” (Nielsen, 1994). It is called “backtracking” if users move back

to previously visited pages. This form of navigation is supported by most hypertext systems (Nielsen, 1994, p. 3) and takes the form of browsers that help navigate the largest hypertext available, the World Wide Web (WWW) (e.g., MS Internet Explorer, Mozilla Firefox, Opera). This non-sequential structure means that there is not one but several ways to read the information included in a hypertext, and the “individual reader determines which [nodes] to follow at the time of reading the text” (Nielsen, 1994, p. 2) and what information to access.

3.3.2 Types of hypertextual structures

Given their representation of “different levels of abstraction, specificity, and significance of the contents of a subject area” (my translation; Tergan, 2002, p. 102), there are different types of hypertext structures available to authors (Gerdes, 1997). Contemporary hypertexts, for example, have a rather modular information structure where paragraphs are coherently interrelated by external links (e.g., the CING navigation bar) or links are integrated into the text, such as “clickable” words containing links to background or other related information. As Figure 20 shows, the basic types of hypertext structure (node-link/unstructured and structured hypertext) are often distinguished with regard to their representation of structure (Gerdes, 1997).

Figure 20: Basic classification of hypertext structures (Gerdes, 1997, p. 26) into node-link hypertexts and structured hypertext.

In unstructured hypertexts, also called Knoten-Link-Hypertexte [node-link hypertexts] (Gerdes, 1997), all nodes are interconnected (Jonassen & Grabinger, 1990) and the connections are based on “associative relations,” that is, not on hierarchical or equivalent relations (Kuhlen, 1991, p. 104). Gerdes (1997) even goes so far as to say that nodes in an unstructured hypertext are related on the grounds of “somehow” existing connections between each other that lack any explanation (Gerdes, 1997, p. 21).

Structured hypertexts, on the other hand, are organized in a linear, hierarchical, or networked fashion and rely, for their top structure, on semantic or pragmatic principles of connection (Kuhlen, 1991; Tergan, 1997; see also Jonassen, 1986), or on a topical hierarchy (Jacobs, 2004) based on the content area they present (see Figure 20). Farkas and Farkas (2002, p. 327) term the hierarchical hypertext structure “multi-path structure,” similar to Storrer’s (2000) category of “multi-sequential” (multi-sequentieller)hypertext.

Figure 21: Models of basic patterns of hypertexts structures (Gerdes, 1997, p. 27) with a) linear, b) hierarchic and c) network structure

While some fields (such as psychology, see Lawless & Kulikowich, 1998) are too complex in their structure for easy presentation and learning in HT, English grammar makes for a more finite and hierarchical content structure and seems appropriate for a presentation in hypertext, as is demonstrated by the CING.

In the CING, the major content areas of English grammar (e.g., the verb phrase or tense and aspect) and their subtopics (e.g., present tenses or simple present) are further divided into sub-subtopics (e.g. since + present perfect). While book grammars impose a linear structure on the complex system of English grammar, the CING was designed to represent the structure of English grammar content in a way that would help learners to comprehend it as a system where information has more than relevance to previous and following information but is interconnected with other information. The program was meant to alleviate confusion about topic allocation by presenting the relevant topics of a grammar topic easily accessible in the index or on the topic page itself (see Figure 23). This is to help learners understand the existing relations between grammar topics in a way that a book grammar cannot.