CHAPTER 2 : LINEARITY AND HYPERTEXT

The promise of hypertext?

Hypertext has been hailed as an invention of crucial importance because it offers radically new ways of structuring information. This advantage is in addition to those which result from the fact that hypertext is a form of electronic text, with all the possibilities this affords for flexible manipulation and rapid searching. For a field that is comparatively young, there is a surprising degree of unanimity concerning the characteristic which defines hypertext and offers this potential: it is simply the ‘non-linearity’ of the node and link structures of hypertexts. Hypertexts have been favourably compared to the supposedly fixed, linear formats of both paper-based and standard electronic texts (see, for example, Duncan, 1989; Cooke and Williams, 1989; Trigg and Irish, 1987; Beeman et al., 1987.) Linear structures are typically seen as being constraining for both author and reader and an inferior way of presenting textual information. Duncan (1989) suggests that "a hypertext system can be both dynamic and interactive in a way that linear text can not – the user can explore a knowledge base in ways not previously determined by the system." Cooke and Williams (1989) claim that the aim of hypertext design should be to enable users "to explore information freely, in multiple parallel paths, instead of being confined to a fixed path or structure." The potential for user driven exploration is also echoed by Baird and Percival (1989).

Orality and literacy – the medium shapes the message

The argument outlined above suggests an intimate and interactive relationship between the tools that we have as a species (i.e., first spoken and then written language) for expressing ideas and the nature of the ideas that are expressed. If this suggestion can be justified then the bolder assertion that hypertext may shape the nature of our conceptions becomes tenable. Hypertext might then be seen as both the next stage in the evolution of communications technology and the possible agent of a further increase in our intellectual and technological sophistication.

Anybody reading this book is likely to find it difficult to even conceive of an oral world that is not some variant of a literate world since the way that we think, and even our thoughts, have been shaped by literacy. However, by drawing evidence from a variety of perspectives, an understanding of oral modes of thought and expression is possible. An understanding of orality allows us to step outside our current literate perspective and thereby appreciate the impact of first script and then print on the organisation of knowledge.

The oral organisation of knowledge

At the beginning of this century, Milman Parry demonstrated that the Odyssey and Iliad contain a limited and repeated number of themes and were constructed from a library of well used expressions or clichés (see Parry, 1971). Subsequently, Lord (1960) showed that this is equally characteristic of contemporary oral epic poetry. This seemed anathema to many classical scholars – who considered the Homeric epics to be among the literary prizes of classical Greek culture – but these characteristics were inevitable when oral poetry’s generation and perpetuation are considered. The Iliad and Odyssey are a unique, written insight into the oral tradition which lasted from the fall of Troy (c1250 BC) until the emergence of the Greek alphabet in about 750 BC, but Homer is nearly as mythical a figure as Odysseus.

The Homeric epics may have been ‘fixed’ when writing was newly available but they are the product of a period when knowledge was transmitted via word of mouth and preserved only in human memory. According to Havelock (1963):

Thus the second ‘intellectual’ technology, following the development of language, was mnemonics – the art of memory. Ong (1982) summarises the techniques used by the oral epic poets:

The epic poem is typically a series of episodes which are centred around individual experience. The perspective is usually subjective and the events are emotional and dramatic. As Chaytor (1945) puts it:

With the gift of hindsight this view of oral performance seems obvious. Even for highly literate audiences of today, the way in which a speech is made, be it dramatic monologue, formal address or traditional story, is crucial for its success; only a trivial part of the the actor’s craft is concerned with learning the lines. If this is true today then it must have been of even greater significance when memory

and speech were the only devices available for preserving and disseminating knowledge.

Since oral expression relied on human memory so completely for its preservation, it should not be difficult to see why the expressions were clichéd, repetitive, thematic, subjective and experiential. The same qualities are characteristic of the vehicle that oral cultures traditionally use to reflect their views of the world and its workings – mythology.

Plato was among the first to appreciate the possible impact that literacy would have on Greek thought and expression. Only a small proportion of the ten volumes of The Republic are given over to statecraft, while a relatively large proportion are devoted to a critical review of contemporary Greek education, art and philosophy. Among his principal targets were the poets, and he went as far as to advocate their exclusion from his ‘Ideal State’ because ‘poetry cripples the intellect of the listeners’. Plato claimed that art and poetry appealed to the lower, less rational, part of human nature since they dealt primarily with the appearance of things. The following dialogue from Book 10 summarises his position:

Plato saw not only that the domination of poetry had to be broken to allow prose to develop, but that the oral mode of thought which had generated the poetic tradition would also have to be abandoned. This would be necessary if scientific rationalism with its implied analysis and classification of experience and rearrangement into sequences of cause and effect were to become a common mode of thought. It should be remembered that it was not only ‘art’ that was preserved in poetic form but any enduring cultural activity: commerce, history, education, law and civil or religious custom.

The character of writing

Many of McLuhan’s assertions concerning the deterministic nature of communication media are based on the fact that the Greek alphabet was an abstract, linear code in which atomistic units were assembled to make semantically meaningful messages. However, there are also important consequences resulting from the introduction of writing that are revealed from a higher level analysis; from the fact that it supports remote, asynchronous communication with a permanent record. The asynchronous aspect of written communication contributes to its detached and objective character. A written text is prepared in advance of its delivery and can be endlessly revised until the precise form intended by the author is reached, without the reader being aware of this process. This has resulted in a concern with exactitude of meaning and tone which is rarely appropriate with oral forms of expression. In stark contrast, oral performers who repeatedly correct themselves will appear less, not more, authoritative.

A speaker and listener, or audience, not only share a common physical context but the speaker can also make assumptions regarding the prior knowledge, or previous experience, of the audience. In contrast, the writer must ensure that his message is completely self contained in terms of information content and cannot rely on direct feedback from the audience or any other type of external support.

As a simple example, consider the situation in which an ‘expert’ is describing the operation of a piece of sophisticated machinery to a group of potential new users. He can not only use gesture to indicate particular features and determine whether the audience is attending to them but he can also engage in a dialogue to achieve further clarity. Additionally, the speaker can make assumptions about the audience’s previous experience and tailor the verbal explanation accordingly. The author of technical documentation is well advised to write so that as wide an ‘audience’ as possible can obtain all the information they require from the text, but this is not without risk – there is a danger of creating a text which, in attempting to be appropriate to every reader, is in fact appropriate to no one. (Admittedly the readers have the advantage of being able to re-read the text as often as they wish and they may well be able to inspect technical drawings.)

It was not merely a useful characteristic of writing that it could be explicit, rather than implicit, in its conveyance of meaning; it was an inherent property. Although the first texts were written recordings of oral statements (e.g., the Homeric epics), the potential of the writing for ‘prose form’ was gradually realised. Havelock (1980) ascribes this change to the preservative power of writing:

The written text eventually came to be autonomous, and readers and authors expected it to ‘stand alone.’ As Olson (1977) puts it, in oral discourse ‘the meaning is in the context’, while in contrast, in literate tradition, ‘the meaning is in the text’. Explicit text allowed the formulation and preservation of statements that could not only be counter to the reader’s previous experience or ‘common sense’ but could also be subject to critical analysis. This enabled a concern about ‘literal truth’; the separation of mythology and history (Herodotus); and the development of systems of logical inference through analysis of the relationships among statements (Aristotle). Writing is also a medium which supports private study rather than the communal organisation represented by a teacher addressing, and being questioned by, his students (e.g., Socrates). A text designed for private study will need to be self explanatory since there will be no opportunity to interrogate the author. In contrast, Olson (1977) points out that:

Until now we have considered the impact of literacy in a historical context and, however convincing the argument, most of the evidence is indirect – by definition, oral cultures leave no permanent record. Fortunately, an alternative approach is possible because literacy, even today, is not a universal experience.

Contemporary perspectives on literacy

As Scribner and Cole (1981) point out, the fundamental changes in modes of thinking associated with literacy have all been inferred on the basis of changes at the cultural and linguistic level. There has been comparatively little effort to determine whether contemporary non-literate individuals or societies think or conceive of the world in ways which differ from literate individuals or societies. This objection betrays an element of simplification over the distinction of literacy/non-literacy. It is possible to think of a society which is wholly oral composed of members which are non-literate or pre-literate. In a society which has minority groups in which literacy is rare, it is not entirely appropriate to consider their members as being unaffected by literacy. There are many ways in which the ideas, concepts and structures generated by literacy will have affected their lives and in turn may shape their cognitive processes.

However, with this caveat in mind, a small number of investigations have shown striking differences in the modes of thinking of oral and literate individuals. The most notable of these is probably the one undertaken by Alexander Luria, a Russian psychologist who gathered data in two remote Russian provinces (Uzbekistan and Kirghizia) during the period 1931—2. The aim of the research was to investigate the impact of the socio-economic and cultural changes resulting from the revolution on the cognitive development of individuals living in the less developed regions. Striking differences were found between the non-literate and those with only a moderate degree of literacy. The non-literate peasants appeared incapable of demonstrating abstract reasoning or formal logic and they apparently conceived of the world in familiar, concrete ways.

In one of Luria’s tests, the subjects were presented with drawings of four objects (a hammer, hatchet, log and saw) and were asked to group the objects according to their similarity. The literate and semi-literate subjects invariably placed the tools together in a single category with the log separate, while the non-literate did not appear to employ categorical thinking at all: to them, all four objects are intimately related. As Ong (1982) suggests:

This operational, or ‘real-life’, mode of thought is beautifully captured in the reply of a 25 year old illiterate:

Luria also attempted to determine the peasants’ capacity to employ formal deductive logic – a Greek invention that appeared after the adoption of the alphabet. The non-literate peasants’ responses to simple syllogisms revealed a marked tendency to reject the self-contained and arbitrary premises in favour of real life knowledge, as though they were answering a riddle. Thus in response to the syllogism ‘In the Far North, where there is snow, all bears are white. Novaya Zembla is in the far North and there is always snow there. What colour are the bears?’ a typical response was, ‘I don’t know. I’ve seen a black bear. I’ve never seen any others… Each locality has its own animals.’ (Luria, 1976, p.108). However, a semi-literate 45-year-old replied ‘To go by your own words, they should all be white.’ (Luria, 1976, p.114). The respondent produces the ‘correct’ reply but is dubious about the premises since they are not in accord with his own experience.

It has been suggested that written expression is, by nature, objective and detached compared to spoken language and that this characteristic enables an abstract perspective. Without accepting the technological determinism implied by this argument, it is undeniable that in current everyday usage language differs markedly depending on whether it is spoken or written. While written text is continuous, natural oral utterances have been found to be produced in spurts of about two seconds duration and six words in length. William James (1950) suggested this represented a single ‘perching’ of consciousness and later psychologists have seen this as reflecting a limitation of cognitive processing.

According to Chafe (1982), when we come to write "we have time to integrate a succession of ideas into a single linguistic whole in a way that is not available in speaking" (p.37). Chafe describes a revealing study in which verbal statements from the same individuals produced in oral and written contexts that varied in formality (dinner table conversations, lectures, private correspondence and academic papers) were compared. The oral samples were found to be composed of idea units linked with co-ordinating conjunctions (and, but, so, because, etc.) while the written samples contained examples of a variety of techniques for integrating associated, subordinate, or qualifying ideas into what would have been a single utterance in an oral delivery. The oral construction of complex units such as these is usually achieved only by those considered highly articulate (and that usually means literate). The written samples were also found to be much more likely to include use of the ‘passive voice’ – a device usually employed to distance the speaker from the subject being described. Similarly, the written material was less likely to contain first person references (I, we, my, etc.), references to the speaker’s mental processes (I felt…, I thought…, etc.), fuzziness (sort of, about, lots of, etc.), expressions of enthusiastic involvement (really good) and direct quotes (and then he said "…"). Thus, oral language has a tendency to fragmentation and involvement while written texts are more inclined to integration and detachment.

So far, we have considered the nature of oral communication and modes of thought and then looked at the way that literacy changes not only the ways in which things are said but also the way in which we think. We have seen this in both historical and contemporary contexts and related this to the inherent characteristics of the media concerned. If hypertext represents a major change in representation technology, might it not seem highly likely that it will also have a significant impact on the form and content of our ideas? Before attempting to answer this question we shall turn to the second major theme of this chapter – the extent to which linearity in texts is a constraint that needs to be overcome. We shall see that, while manuscript technology may initially have been constrained and constraining, this limitation was gradually overcome.

The evolution of writing – from record to resource

The historical evolution of literacy and ‘manuscript technology’ clearly shows the relationship between linearity and text. While early texts were undoubtedly ‘linear’ in terms of content and usage, the increasing sophistication of both technology and readers’ skills allowed this ‘limitation’ to be overcome. Indeed, in some respects it is possible to claim that the linear document was obsolete by the 13th century! In this light, the claims that hypertext has the potential to free the reader from the constraints of linear based documents appear to be unjustified.

In England, the origins of popular literacy can be traced to the centuries following the Norman invasion of 1066. A wealth of documentary evidence is available that gives a clear insight into both the important changes resulting from the availability of written records and increased literacy. It is all too easy to assume that the skills of reading and writing, and attitudes towards them, were much the same in the middle ages as they are today and this is far from true. The following quote from Clanchy (1979) suggests some of the practices required in an age when literacy was still restricted to an educated few while the population at large relied on oral memory:

The popular acquisition of writing during the late middle ages freed people from the need to memorise knowledge in order for it to be passed on to future generations, and freed knowledge from the requirement that it should be encoded in a form amenable to verbal memorisation – thus repeating a process that had transformed Greek society some 2000 years before. And, once again, writing made it possible for people to have objective knowledge of the world in addition to subjective knowledge:

Historians, like Clanchy, have tended to assume that prior to the Norman invasion there had been no significant evolution in the technology or practice of literacy since the demise of the Roman Empire some 500 years before. But although the document-based Norman administrative system significantly increased both the importance of literacy among lay circles and the stock of secular texts, the evolution of literacy and manuscript technology had begun much earlier.

The majority of written texts produced in the later Roman era were composed of continuous rows of upper-case characters without word gaps or punctuation – as was the common practice in earlier Greek inscriptions. This style is entirely phonetic rather than ideographic. While Latin was the language of every day speech as well as erudite discourse, it could be read and written by very few – the educated elite. Writing was a process of transcription, with each successive sound being captured and given a permanent visual form. Spelling mistakes were often the result of mispronunciation by the writer/dictator but the reader was aided by the strong inflexion/conjugation and tightly-controlled sentence construction of Latin.

Words committed to paper were intended for reading aloud – indeed most texts were letters, decrees, proclamations and so forth, and were typically read aloud in their entirety, thus recreating the original oral form. Hence, even prose had a strong rhythmic quality, and writing was phonetically biased. According to Saenger (1982):

The virtual impossibility of rapid, visual and silent reading had a number of consequences. The scribe making copies was obliged to read short sections aloud and then write them down or to work with a dictator. The texts which were used as references required the use of marginal flags to provide access cues to specific items contained within the unbroken text. Quintilian, in a text on the art of rhetoric, recommended allocating marginal signs to specific arguments in a speech to be committed to memory. The signs were then to be memorised using visual-spatial techniques.

When the Christian church spread to the western fringes of Europe in the 6th century, it took with it Latin and literacy. However, according to Saenger (1982), while the Celts, for example, were converted to Christianity without too much difficulty, they experienced difficulty when reading Latin scriptures and liturgical texts aloud – they had no previous experience of literacy or everyday use of spoken Latin. In order to help pronunciation, it became common for words to be written with separations. Word spacing allowed the weaker reader to recognise complete words and reduced the need to build them up by verbally pronouncing their constituent sounds. This process was also facilitated by the gradual introduction of miniscule characters. Word separation and a more distinctive set of characters enabled the transformation of reading from being a predominantly oral/aural activity to a visual one. An early consequence of this change was the development of visual transcription which allowed scribes to copy texts silently. However, while this may have silenced the mumblings that had previously been characteristic of the monastic scriptorium, reading outside remained largely oral. The requirement for more widespread silent, and consequently rapid, reading did not arise until a later period.

While literacy was preserved within the monasteries of Europe during the Dark Ages, faster reading offered no particular advantage since reading served largely liturgical and devotional needs. However, this changed radically following the rise of medieval scholasticism in the 12th century. Scholasticism was associated with a virtual information explosion of textual material:

There was not just an increase in the volume of textual material; there was an equally significant increase in the complexity of the content. This exposed the limitations of oral styles of reading: complex arguments require detailed consideration with frequent re-view and are not well suited to being read aloud to an audience.

Scholasticism also radically changed the ways in which texts were created. Prior to this period, authors had typically dictated to the highly skilled scribe who first recorded the words on wax tablets of fairly limited capacity and then made a fair copy on vellum or parchment – a slow and laborious process. Scholars like Thomas Aquinas and Albert Magus preferred to compose their long and increasingly complex works on parchment directly using a cursive script. The description of text usage given by Saenger (1982) suggests that academic reading and writing styles have hardly changed in 600 years and that the concept of hypertext would have been as welcome then as it is to many scholars today:

The presentation format of manuscripts evolved in a similar way over this period as their rôle changed from sacred devotional relics to sources of information to be studied visually and silently. In the early middle ages it was rare for text to be split into sections shorter than the chapter, and many manuscripts of the Old Testament were not even divided into chapters. During the later middle ages, not only were chapter divisions introduced into classical texts but sub-divisions were also added by university scholars. This enabled the construction of tables of chapter headings, running headings and alphabetical tables by subject to become standard features of academic text.

While manuscripts were often visually beautiful, this has often been mistakenly taken as some sort of evidence that they were of limited utility. The Domesday Book for example used vermilion paint for three different types of rubric – capital letters for the names of shires and other headings; shading for the initial letter of each paragraph and certain abbreviations; underlining for the names of places and tenants. In legal manuscripts letters of the alphabet were used in the margins to indicate relevant sections of the text and this system was later used to gloss literary texts. An even more remarkable example of manuscript craft and utility is the Canterbury Psalter produced circa 1147. On each page the calligrapher managed to display three versions of St Jerome’s Latin text of the Psalms in parallel columns (see Figure 4) In addition, there are interlinear Latin glosses and Anglo-Saxon and French translations. If hypertext can have a paper format, this must represent its ultimate visual expression; it is difficult to imagine the usage of such a document as being ‘constrained by the linear, paper-based format’.

The switch from oral to visual emphasis can also be seen in the increased importance of diagrammatic illustrations in the manuscripts of the late middle ages. The presence of diagrams in very early texts has been largely overlooked in favour of representational illustrations – which seems surprising since their origins apparently go back as far as Aristotle’s works according to Saenger (1982), and they appear to suffer less corruption as a result of successive copying. However, diagrams achieved much greater prominence with the rise of scholasticism. The use of schematic diagrams in scholastic texts is significant since it clearly shows yet again how the written form managed to overcome the apparent restrictions of a fixed format. According to Evans (1980):

Evans includes a page from a 14th century French manuscript to illustrate this point and his explanatory note gives a vivid impression of the figure’s expressive power:

Insert Figure 4 – the Canterbury Psalter – on this page

It was only a relatively small step to abandon continuous prose completely and represent the distinctio as a schematic diagram, e.g., a hierarchical classification with the concept successively broken down into sub-component terms. These abstract diagrams were often given figurative forms such as trees, ladders or wheels.

Clanchy (1979) contrasts the attitudes shown towards books and styles of reading before and after the Norman invasion. Before the invasion books and reading (and writing) were largely confined to monastic communities and monks were "expected to ruminate on a text which had been designated to them as a sacred task" since "books, with their precious and brightly illuminated words, were images which produced a state of mystical contemplation and understanding." In contrast, by the time of Edward I who ruled from 1272 to 1307, Dominican monks and lay Court clerks "like modern academics, required extensive libraries in which they could glance rapidly over a whole series of books, many of very recent authorship, in order to construct a wide-ranging argument."

However, early manuscripts were still regarded in highly personal terms by readers – as though they were some orator’s words made flesh. John of Salisbury, writing in the mid 12th Century, claimed that:

This feeling was supported by the hand-crafted nature of the object and the often rhetorical style of the content. In fact, the majority of manuscripts did not even have titles and were referred to by their ‘incipits’ or opening phrases. If a medieval reader was asked who had written a particular manuscript then he would probably have suggested that it was the scribe rather than the ‘author’ of the text. Notions of originality, copyright and plagiarism have largely appeared since the advent of printing. In manuscript culture, authors were often regarded more as reporters or collators of material from much earlier (Classical) times.

Ong (1982) claims that the use of titles was a major factor in altering the perception of books from "frozen utterances" to things or objects, since titles are a form of label. However, reading and writing in Medieval times had far more in common with speaking than they do today. Reading meant reading aloud for the majority of readers until comparatively recently and even literate individuals often preferred to have a letter or statement read aloud to them than to read it themselves. Reading aloud and dictation also allowed the non-literate to participate in the use of documents whilst they are excluded by today’s silent reading and writing. Until quite recently, undergraduate students were supposed to spend their time ‘reading for a degree’ in a given subject. With our current literate perspective, this evokes images of many hours spent by the students in private study poring over text books and learned journals. In the middles ages the meaning was quite different. The students would spend a good deal of their time listening, and probably taking dictation, while the lecturers read the texts out aloud.

The printed book

In the same way that an appreciation of orality is difficult for contemporary literate cultures, so it is also difficult (albeit to a lesser extent) to empathise with the world of manuscripts rather than printed books. Eisenstein (1979) considers the impact of printing on European culture at length and observes:

Conventional wisdom has recognised the importance of the development of printing in mid 15th Century Europe from an early stage, and Francis Bacon’s observation that:

is frequently quoted as proof. However, the significance of print technology is usually assumed to be connected with improved access to written knowledge, i.e., greatly increased literacy and plentiful supplies of relatively cheap texts. Although this is undoubtedly true, it has tended to obscure less obvious but equally important factors.

The advent of an ever-increasing range of clearly printed texts radically changed reading skills. The manuscript reader not only spent much less time reading but the choice of material was very restricted. Consequently the level of fluency achieved by the average reader would be considered modest by contemporary standards. Chaytor (1945) contrasts the contemporary reader’s skills with those probably achieved by a medieval reader:

Writing enabled lengthy and detailed descriptions to be made and the advent of printing significantly increased the importance of exactitude over manuscript description. Printing was the first mass production technology and books share the essential characteristic of any mass-produced item – uniformity. A printed book looks exactly the same and contains exactly the same content as any other book in the same edition. This was never true of any two manuscripts, even if they were produced by the same scribe. Errors were common and unlikely to be detected by readers who rarely had an opportunity to compare multiple copies of the same manuscript. This problem was most acute with illustrations which were subject to continuous corruption as they were repeatedly copied by successive scribes.

The introduction of printing changed the form of books as well as the way in which their content was regarded. Printing was associated with commercial publishing, and although the first printed books were exact copies of manuscripts this fidelity was soon replaced by a commitment to technical improvement in order to gain a competitive advantage. Well before the year 1500, printers had started to experiment with different typefaces, running heads, footnotes, tables of contents, figures and cross references. For example, Peter Schöffer – Gutenberg’s partner – printed a prospectus for his firm’s books which described them as being more readable, having more complete and better arranged indexes and as having received more careful proof reading and better editing.

As library and publishers’ catalogues grew larger, new methods of cataloguing became essential. The eighth century library at York had a rhymed book list and other catalogues were equally idiosyncratic. Although a standard alphabetic order was employed in the Greek library at Alexandria, the convention was used infrequently and somewhat unsystematically through the period of manuscript culture. Only with the advent of printing did it become general practice.

The adoption of a standard alphabetic ordering was an important aid in transferring the book from a repository of utterances for verbal recitation to a visual object for critical examination. Alphabetic ordering, along with Arabic page numbering, revolutionised indexing. In the era of papyrus scrolls and manuscripts, making a reference to a specific sentence or section was virtually impossible. While scrolls often had line numbers, they originated as a metric for the payment of scribes rather than for purposes of reference. Similarly with manuscripts, indexing was often not worth the effort since no two manuscripts were ever exactly the same in terms of page numbers or even content. According to Witty (1965), manuscripts with even basic alphabetic indexing appeared no earlier than the 14th century and this often involved ordering only to the level of the first syllable. Indexes often consisted of sentences drawn from the text ordered alphabetically according to a "catchword" as in a contemporary key-word-in-context (KWIC) index. Page numbering is another feature that is taken for granted today but only appeared after printing. The numbers were originally placed on each leaf rather than on each page and were used to help the binder get the leaves in the correct order.

As printing reduced the need to maximise page filling for reasons of economy, so the number of contractions/abbreviations reduced and the amount of white space

increased, thus allowing greater separation of words, sentences and the visual separation of paragraphs.

The term ‘index’ also reveals the influence of oral/rhetorical tradition on manuscripts since it was a shortened form of ‘index locorum’ or ‘index locorum communium’. These were the index of places or index of common places (mental images) where various arguments could be retrieved from memory. This alteration of practical meaning of the term illustrates the shift from verbal to visual orientation that accompanied printing. A second example is the use of figures. Accurate drawings were included in very early ‘original’ texts (e.g., Galen’s text on medicine) but were frequently omitted, for reasons of economy, when the works were copied. If they were copied, their value quickly diminished:

While printed illustrations, by way of woodcuts and subsequently more accurate engravings, were initially employed by printers with little apparent care (incorrect captioning, reversed images and the frequent use of a very limited range of images), their introduction assisted in the development of technical texts. Indeed, it is possible to argue that the introduction of accurately printed figures, the new concern with accurate description and the increased circulation of texts enabled modern science to develop in the 16th century. Similarly, Eisenstein (1979) has discussed the the impact of the increased visual order given to books through printing:

She goes on to argue, like Ong, that the order imposed on printed words was internalised and led to more systematic approaches to information in general:

This increasing bias towards visualization of knowledge can be seen in the dialectical approach to organizing knowledge favoured by Pierre de la Ramee (Petrus Ramus) in the 16th century. Ramus was an influential and, even by the standards of his time, highly controversial French academic who firmly believed that the teaching and learning of any discipline could be greatly improved through the application of systematic structuring. Ramus was a confirmed humanist and rejected the artificial memory techniques which, although originating in classical Greek and Roman times, were associated with both medieval monastic scholasticism and the occult. According to Yeates (1966):

For Ramus, the dialectical analysis of subject matter by way of a hierarchical class system is optimal because dialectical order is ‘natural’ to the mind. This approach can be clearly seen to be another memory system but one which is based on memorising a visual schematic representation. Ong has claimed that Ramus’ perspective was a consequence of the increased tendency for visual ordering fostered by the arrival of printing. However, it is possible to find evidence within manuscript culture of visual categorisation in two dimensional space – e.g., the manuscripts employing the ‘scholastic method’ described by Evans (1980).

Although his work was incredibly popular and influential at the time among the protestant merchantile classes, it never achieved academic respectability and can be seen as something of a dead end. A possible reason for this is suggested by Figure 5 which shows the life of Cicero according to Ramus. The main events in Cicero’s life are presented as a hierarchical tree. While the graphical presentation may have aided memorisation, the end hardly justifies the means. The first dichotomy contrasts his life and death and many of the lower order items are similarly vacuous. A major motivation for Ramus’ work was the need for memorisation as a cornerstone of education but this requirement became less important with the arrival of print culture and the school text book.

The implications for hypertext

We have seen how technological advances in modes of representation (i.e., from oral to literate transmission) can change not only what is said but can also shape the mental processes which generate the thoughts themselves. This sounds similar to the claims expressed at the very beginning of the chapter, i.e., that widespread hypertext representation may herald an era of pluralistic rather than linear thought. However, while hypertext might seem to represent a radical change in medium, it is not likely to result in a revolution in thought; it is a comparatively trivial change compared to the advent of popular literacy.

Although hypertext differs significantly from printed text in its ‘arbitrary’ structure, it shares many similarities for the reader. Hypertexts, despite their node and link structure, are still composed of units of text and there is no reason to believe that, at the paragraph level at least, these are read any differently from units of conventional paper or other electronic text. At a higher level of organisation, it is common for the reader of a hypertext to be frequently presented with alternative routes through the text. However, although the reader may be encouraged to make more active choices, this still results in a ‘serial’ route through the text since only one node can be accessed at a time.

This is not to say that popular use of hypertext will result in an unchanged conception of ‘text’ but that the changes might not be obvious ones. An appropriate parallel might be with the changes in attitude that accompanied the introduction and usage of printed texts. Although the first books (incunabula) were printed versions of manuscripts and therefore were largely the same in terms of content, the fact that they were printed rather than hand-written resulted in significant conceptual changes. Uniformity, order and repeatability are characteristic qualities of the printing process but they were soon characteristic of the way that knowledge was organized – soon after the introduction of printing there was an significant increase in alphabetical ordering, indexing, bibliographies and classificatory systems in a variety of disciplines.

What changes are likely to be associated with popular use of hypertext? Almost any suggestion is likely to be highly speculative at such an early stage, but by way of example consider the issue of textual authority. A printed text has always had an authoritative quality that is not granted to a speech – in general, people are more likely to believe what they read than what they hear, especially when the text is couched in official terms and printed in an appropriate style. In a similar way, almost irrespective of the content, a perfectly reproduced typeset document has an authority that is simply not available to a manuscript. The reasons for this are varied and largely historical and include: the rôle of the Bible as the prototypical book; the very long period when literacy was skill restricted to the educated elite and a period nearly as long when individuals learned to read (often the Bible) but not write; finally there is the restricted access to publishing imposed by financial considerations. At the level of the individual text the physical presence of the document confirms the enduring quality of writing over the ephemeral nature of speech and reinforces the separate identity of a specific text. This aspect has long been recognised by the publishing industry which has developed a variety of ways of adding authority to texts by selective use of impressive bindings, decoration, typography and weight of paper.

In complete contrast, an electronic text is strikingly neutral. It has little physical presence and its shape is fluid rather than fixed. The author or editor (and increasingly they are the same person) experiences the generation of an electronic text as an endless round of amendments, some major but the majority minor. The creation of a printed book traditionally consisted of a series of discrete drafts since the text soon becomes unreadable after a certain number of editorial comments have been written on the typescript. This contrast is maintained after the publishing date since the cost of re-setting a printed book ensures that revised versions are kept to a minimum; there is little cost associated with changing an electronic text.

The contrast is even stronger for the reader who experiences a printed text in much the same way on every occasion while the electronic text is constantly changing before his eyes – this impression is strongest with scrolling displays but is still true of paged displays unless the screen refresh is almost instantaneous. In addition, every text displayed on a given display system is likely to look the same since a common font (or restricted range of fonts) will probably be used for all. Hypertext compounds this weakening by allowing readers to either jump from text to text with ease or by presenting texts which are composed largely of inclusions from other documents. Whether these differences are strong enough to reduce the reader’s willingness to question the authority of the text remains open to empirical evaluation but there seem a priori grounds to believe that it could well be true.

Turning to the second theme of the chapter, we have seen how the fixed physical form of the printed text, far from still being a constraint, has largely been transcended. The skills that we have acquired for text handling in the broadest sense make serial reading a rare strategy outside of fiction. Our current experience with (reading and writing) text reflects two complimentary, organizing influences. The detached, linear mode of thought that results from alphabetic literacy and the text handling skills that develop with continuous interaction with sophisticated printed materials. These skills allow rapid silent reading, skim reading, reading tables and graphs and intensive use of indices and contents lists. The skills, enabled by the years spent in education, the profusion of printed material, the standardisation of print formats and the evolution of typography, have allowed us to overcome many of the supposed limitations of the written record.

Judging by the current generation of hypertext system interfaces, it would appear that the majority of designers are unaware of these developments and skills. Instead of attempting to provide information-rich screens which have been designed with due regard to typography and graphical layout, all too often the user is presented with comparatively little textual, and even less contextual, information and a number of choices concerning additional screens of information. In order to gain an informed impression of the hypertext’s coverage in terms of breadth and depth, it is therefore necessary to scan a considerable number of screens.

By way of example, consider a broadsheet newspaper like The Times which has achieved a presentation format that maximises information display with rapid access for the user. A quick scan of the complete paper can be completed in minutes. In contrast, to display the contents of a single page of this newspaper on a standard (24 line x 80 character) microcomputer screen would require 25 separate screens – to display the whole newspaper would require in the order of a 1000 screens. Even if the hypertext was displayed on a larger workstation screen, browsing would still represent a considerable task. Yet hypertext is frequently suggested as a way by which readers can browse large databases rather than employ structured search strategies.

Despite these current weaknesses, hypertext’s true potential surely lies in its ability to improve access to the growing corpus of electronic information. Current on-line databases hold enormous quantities of information but their interfaces are frequently too daunting for the untrained or casual user. They may well hold the information required but they frequently require a trained intermediary to select the appropriate database and formulate an appropriate search query. ‘Trawling’ with fuzzily defined targets is definitely not recommended. Hypertext could provide a supportive interface to these, and future, databases (cf. Marchionini and Shneiderman, 1988).

However, hypertext could be much more than a superior access mechanism for today’s text-based databases. In a multimedia guise within a teaching/learning context, it offers much better ways of presenting information compared to traditional printed texts. Consider the advantages of being able to build multimedia hypertexts for schools and students which include video and sound as well as text and graphics – as described in Chapter 1. Obvious applications include natural sciences, art, drama, music and social history but the possibilities are really only limited by the imagination.

Such applications may significantly improve the quality of our teaching/learning tools, hopefully resulting in increased breadth, complexity and integration of teaching and learning, but it is not clear how ‘immersion’ in hypertext will affect the way that we mentally structure our world. Linear argumentation is more a consequence of alphabetic writing than of printed books and it remains to be seen if hypertext presentation will significantly erode this predominant convention for mentally ordering our world.