Structuring writing for reading: hypertext and the reading body.

by Paul ten Have, University of Amsterdam, version 2.1, March 1999 *

A note on three versions of this paper
An ethnomethodology of reading and writing
On 'explicating devices'
On texts: reading, writing, design
What is HTML?
Writing in HTML
Reading HTML
Guiding the reader in HTML 1: sections and documents
Guiding the reader in HTML 2: Notes and References


This paper examines some textual devices that writers may use to pre-structure the activities of their readers. HyperText Markup Language (HTML) is used as an 'explicating device' to explore how writers can provide reading instructions, and how these can be experienced by readers. Structuring devices like paragraphs and sections, and hypertextual elements like notes and references are investigated in detail. In this way, the paper aspires to contribute to 'an ethnomethodology of textual practices'.


The activity of reading a text is structured by the artful use of textual devices, such as paragraphs and foot or end notes. Learning to read involves learning to use such devices to structure one's reading. Learning to write involves learning to produce texts which contain devices to structure readings effectively. Reading and writing are basic activities in many cultures, while the repertoire of textual devices tends to be relatively specific to (sub-)cultures. Textual devices, then, can be analyzed as ethno-methods for reading and writing.

For members, these ethno-methods are natural resources for doing some core activities of the culture. They can be encountered as topics in situations of explicitly learning to do these activities, as in schools, but also in mastering technologies designed to support or ease the process of writing, as in word processing technology.

Analysts of ethno-methods, ethnomethodologists, can use such situations and experiences to study some procedural basics for the ordinary activities of reading and writing. The art of writing requires the proof of reading. Studying writing, and writing technologies, requires the study of reading as well.

Word processing technology can be used as an explicating device for studying writing as an activity for structuring reading. While not explicitly doing so, it provides the user with preferred models of structuring texts, which cannot be said to be 'culture-free'. Studying such a technology allows one to study a normative culture of writing. HyperText Markup Language (HTML), used to produce documents for the World Wide Web, readable with a Browser like Netscape, can be similarly used. It even has some special properties that seem to make it particularly fit for such an exercise.

In this paper, I will discuss some of my experiences of using HTML to make available ETHNO/CA NEWS on the Web, including the 'conversion' of papers to HTML, as well as my experiences as a reader of HTML documents. I would like to suggest that, by facilitating internal and external linkings of texts, HTML offers the possibility to renew considering writing as artfully pre-structuring reading. Furthermore, these explorations have led me to reconsider reading as an embodied activity, in contrast to some ideological treatments of hypertext which celebrate it as ultimate freedom.

A note on three versions of this paper

This exploratory essay has been produced in three different formats. Firstly, I have written a paper version, submitted to the conference organizers for pre-conference distribution. It has been revised for post-conference publication. Secondly, I have given an oral presentation, at the conference itself. And finally, I have made pre- and post-conference versions in HTML available on the World Wide Web. The first formats were written using a word processor. The second combined various media, such as overhead projection and hand outs, in addition to my voice. The third version, however, seems to me to be the most fit for the job, as it is written in HTML, the subject of this paper. This will allow the reader to check some of my remarks directly, by 'reading' the paper and playing around a bit with its possibilities.

An ethnomethodology of reading and writing

I want to place my essay explicitly in the tradition of Ethnomethodology.

In their foundational paper, 'On formal structures of practical action', Harold Garfinkel and Harvey Sacks write:

The notion of member is the heart of the matter. We do not use the term to refer to a person. It refers instead to mastery of natural language, which we understand in the following way.
We offer the observation that persons, because of the fact that they are heard to be speaking a natural language, somehow are heard to be engaged in the objective production and objective display of commonsense knowledge of everyday activities as observable and reportable phenomena. We ask what it is about natural language that permits speakers and auditors to hear, and in other ways to witness, the objective production and objective display of commonsense knowledge, and of practical circumstances, practical actions, and practical sociological reasoning as well. What is it about natural language that makes these phenomena observable-reportable, that is account-able phenomena? For speakers and auditors the practices of natural language somehow exhibit these phenomena in the particulars of speaking and that these phenomena are exhibited is thereby itself made exhibitable in further description, remark, questions, and in other ways for the telling.
The interests of ethnomethodological research are directed to provide, through detailed analyses, that account-able phenomena are through and through practical accomplishments. We shall speak of "the work" of that accomplishment in order to gain the emphasis for it of an ongoing course of action. The work is done as assemblages of practices whereby speakers in the situated particulars of speech mean something different from what they can say in just so many words, that is, as "glossing practices".
Garfinkel & Sacks (1970): 342
 Several remarks are in order. First, the subject of ethnomethodological studies is the work of creating a local social order as an accountably 'natural' one; this requires that members' competence can be made 'public'. Second, this work is intrinsically embedded in the use of 'natural language'. And third, there is more to 'the situated particulars of speech' than can be expressed 'in just so many words'. To these 'glosses' can be added a fourth remark which is an observation: 'members' are pictured as 'speakers and auditors'; so the competent use of 'natural language' is conceived of as speaking and hearing. In other words, 'conversation' is conceived of as the paradigm of social life: sociality is equated with talk-in-interaction.

In a way, then, this quote offers a charter for Conversation Analysis (CA), but at the same time may be read as 'outlawing' the ethnomethodological study of some non-vocal families of 'glossing practices', including writing and reading and its recent offspring Computer-Mediated Communication (CMC).

In studying talk-in-interaction, CA uses the sequentially localized character of turns-at-talk as an important resource for searching and proving its analyses of members' meanings. In their famous paper on 'turn-taking' in conversation, Harvey Sacks, Emanuel Schegloff and Gail Jefferson (1978) note an interesting 'methodological consequence' of their model, in that, because a speaker is expected to react to previous utterances, one can inspect subsequent utterances to see how their speakers have taken up these previous utterances, thereby possibly displaying their understanding of those utterances.

In the first place, of course, such understandings are displayed to coparticipants, and are an important basis for conversation's local self-correction mechanism. (...) But while understandings of other turn's talk are displayed to coparticipants, they are available as well to professional analysts, who are thereby afforded a proof criterion (and a search procedure) for the analysis of what a turn's talk is occupied with. Since it is the parties' understandings of prior turns' talk that is relevant to their construction of next turns, it is their understandings that are wanted for analysis. The display of those understandings in the talk in subsequent turns affords a resource for the analysis of prior turns, and a proof procedure for professional analyses of prior turns, resources intrinsic to the data themselves.
Sacks, Schegloff, Jefferson (1978): 45
 In some forms of CMC, one could use a similar strategy, in that, for instance, later contributions to a 'thread' in a 'news group' or 'discussion list' can be used to inspect at least some members' analysis of previous postings. In other, less 'interactive' formats, such supporting devices are absent. In that respect, they tend to be closer to more traditional forms of written social life, like journal or book publishing, where an author offers a text to an unknown audience which most of the time doesn't speak back, obliging the author to design his or her text on the basis of an image of the audience, depending on general conceptions, restricted occasions of audience feed back, and his or her own reading experiences.

In advocating or commenting on the recent emergence of 'Hypertext' or more largely 'HyperMedia', many people have stressed the difference between these electronic forms of textual presentation, and the more traditional printed ones as in books and journals (c.f. McHoul & Roe, 1996 for a critical analysis of these positions and Benschop, 1997, for a moderate example). For the moment, I will not take a firm position regarding this issue, taking it to be an 'empirical' one that needs to be researched by studying actual practices of reading and writing. This 'empirical' position does imply a 'pre-empirical' stand, however, by suggesting that the basic activities in both printed and electronic forms of textual exchange are 'reading and writing'. My general thesis is that studying reading and writing in the practical field of electronic 'publishing', and comparing these to the ones used in the field of printed exchange, helps in understanding the praxeological structures in both fields. The methodological concept I will use to gloss my efforts will be 'explicating device', which I will elaborate in a separate section.

On 'explicating devices'

I have coined the term 'explicating devices' to gloss a loose category of techniques for making the work of doing some social order accessible to ethnomethodological study(1). In the corpus of ethnomethodological writings, one can find a collection of terms used for similar purposes, whether acknowledged or not, i.e. 'breaching experiments', 'becoming the phenomenon', making 'transcriptions', and 'perspicuous settings'. Garfinkel commented on his 'breaching experiments' as follows:
The studies reported in this paper attempt to detect some expectancies that lend commonplace scenes their familiar, life-as-usual character, and to relate these to the stable social structures of everyday activities. Procedurally it is my preference to start with familiar scenes and ask what can be done to make trouble. (...)
A word of reservation. Despite their procedural emphasis, my studies are not properly speaking experimental. They are demonstrations, in Herbert Spiegelberg's phrase, as "aids to a sluggish imagination." I have found that they produce reflections through which the strangeness of an obstinately familiar world can be detected.
Garfinkel (1967): 37-8
 In order to escape some of the practical and ethical problems generated by such experiments, other ethnomethodologists started using a different strategy. In this the researcher studies his own sense-making work by putting himself in some kind of extra-ordinary situation. This may be a situation where routine sense-making procedures are bound to fail, or where one has to master a difficult and unknown task, or where one is instructed by a setting's members to see the world in a way that is natural for them but not for oneself. Mehan & Wood (1975) use the expression 'becoming the phenomenon', while Schwartz & Jacobs (1979) recommend strategies of becoming 'The Stranger' or 'The Novice'. Out of many possible examples I would like to mention David Sudnow's (1978) study of becoming a jazz piano player, and Lawrence Wieder's (1974) study of his being instructed in the use of 'the Convict Code' as a general interpretive and explanatory device in a half-way house for paroled addicts.

Many other ethnomethodologists, however, have used methods more closely resembling traditional fieldwork. It involves observing situated activities in their natural settings and discussing them with the seasoned practitioners, in order to study the competencies involved in the routine performance of these activities. To further this close study, or to be able to study these activities after the fact, recording equipment may be used. Examples of this kind of study can be found in Garfinkel's (1967) work on juries and coroners, Zimmerman's (1969) study of case-workers in a welfare agency, and Lynch's (1985) research on laboratory scientists. Quite often, these studies concern activities whose 'produced order' may be seriously contested, as in the above cases by defenders, clients and competitive researchers. Melvin Pollner (1979), in his study of interactions in traffic courts, has coined such terms as 'explicative settings' and 'explicative transactions' to indicate the relatively unusual 'explicativeness' of what can be observed there, especially for defendants observing the ways in which previous cases are being presented and handled.

In other words, an ethnomethodologist may use 'explicative devices' that are already a natural part of the setting under study, or he/she may work to produce them on purpose as 'experimental' ones. Furthermore, many if not most established research methods, ranging from survey research interviews, through ethnographic field work, to CA's transcription techniques, can be used as 'explicating devices' as well. In each case, the device can be used to assist the researcher is noticing the details of the work of producing a local social order.

For a further discussion of these issues, a few quotes from Garfinkel & Wieder's (1992) paper may be useful:

Ethnomethodologically, every topic of order*--every topic of order, logic, meaning, reason, and method--is eligible to be found as a phenomenon of order*. Every topic of order* offers to ethnomethodological study its candidacy to a search for a phenomenon of order* as an achievement in and as of practical action.
(ibid.: 180)
Since EM's methods are discoverable phenomena of order* in their own right, they are not methods as methods are "straightforwardly" understood. The fact that EM's methods are discoverable phenomena of order* in their own right is central to EM's treatment of methodogenesis, the relation of EM methods to EM knowledge, its position on phenomena of order* as prior to and independent of their EM study, the place of competence as a reflexive constituent of the "work" to which that competence is addressed, and more. These features are partially addressed in the unique adequacy requirement of EM methods and in EM's use of perspicuous settings.
(ibid.: 181)
In its weak use the unique adequacy requirement of methods is identical with the requirement that for the analyst to recognize, or identify, or follow the development of, or describe phenomena of order* in local production of coherent detail the analyst must be vulgarly competent in the local production and reflexively natural accountability of the phenomenon of order he is "studying."
(ibid. 182)
We found it useful to explain the term perspicuous setting by saying ethnomethodology is "embedded" in a local culture (..)and under that condition the analyst examines the various objects in that culture (..). A perspicuous setting makes available, in that it consists of, material disclosures of practices of local production and natural accountability in technical details with which to find, examine, elucidate, learn of, show, and teach the organizational object as an in vivo work site.
(ibid. 184)
To find a perspicuous setting the EM policy provides that the analyst looks to find, as the haecceities of some local gang's work affairs, the organizational thing that they are up against and that they can be brought to teach the analyst what he needs to learn and to know from them, with which, by learning from them, to teach them what their affairs consist of as locally described, locally questionable, counted, recorded, observed phenomena of order*, in and as of their in vivo accountably doable coherent detail for each another next first time.
(ibid.: 186)
I will not try to explicate these rather opaque phrases here, but just summarize what I take from them for my particular 'local' purposes. I will use HTML as an 'explicating device' in the sense that I have used HTML's technologies in actual practices of reading and writing. While carrying out my local tasks, I also have observed my doings and experiences in order to learn about those in an ethnomethodological sense. Furthermore, I have, in trying to do reading and writing in an environment constituted by HTML technologies, observed how I was confronted with various possibilities, impossibilities, alternate possibilities and effects, whether expected or not. In other words, I have encountered some structures of a work space constituted by HTML and related technologies. In other sections, I will report on some of these experiences and present some of the reflections they have inspired.

On texts: reading, writing, design

A common sense conception of 'reading' may be glossed as 'following a text'. The reader's eyes see the text word by word, and the reader recognizes the words as they are available in his personal 'lexicon'. By combining the words in the sentence (or another format, such as a list) and using his or her grammatical knowledge, the reader can interpret the sentence to 'see' its meaning. In other words, the eyes scan the text as a series of lines, left to right, and the 'mind' is used to interpret this information to arrive at an interpretation of words, sentences, paragraphs, etc. This conception is based, first, on a notion of linearity, and second, on a notion of 'knowing a language'.

On a more encompassing level, reading involves 'following' more complex forms of organization than 'just' the ways in which words can be seen to form a sentence. For stories, analysts have formulated general formats which story-tellers tend to use. Reading a story, then, would involve 'fitting' parts of the text in terms of such implicit formats, 'interpretive frames', or 'macro-structures'. Similarly, following a written argument, might be seen as involving using argumentative 'macro-structures'. Such ideas imply that complex texts are constructed on the basis of 'chunks' of text of various kinds, which may, or may not, be made 'visible' on the surface of the text as paragraphs, sections or chapters. Furthermore, these chunks can be organized in a linear fashion, often in an order like: introduction > complication > solution. Such a type of organization would then be available again and again on various levels, i.e. paragraphs, sections and chapters.

In other words, according to such views(2), reading involves more than recognizing and understanding sentences; it requires complex structuring operations. Effective texts would be those in which writers have used various devices to 'support' such structuring activities in a way that makes it easy for the reader to discern a text's structure. These devices include the visual marking of the chunks, one from the other, for instance by indenting a paragraph or separating one from another by a blank line. When the reading eye meets such a mark, the reader is reminded that he or she is crossing a border, so to speak. When a text uses such devices in a way that these 'formal' indicators seem to fit with its 'contents', the reader may experience it as being well-written, or maybe better: 'well-designed'. Running a bit ahead of my argument, I would already suggest that this image of the eye following the text line-by-line until it meets a visual chunk-mark is too simplistic. The eye's focal vision may be 'following' the text in this way, but in peripheral vision, the overall visual image of the page is used as a context for this activity, 'preparing' the focus to jump to a next paragraph, so to speak, and reminding the 'mind' that the current paragraph is drawing to an end (so 'what's the conclusion?').

To simplify my later discussion, I would like to restrict the concept of 'writing' to the practice of constructing sentences, paragraphs, etc., i.e. ordered text. I will use the concept of 'design' for the visualization of a text's 'structure', which is commonly indicated by concepts such as 'styling', 'doing the layout', etc. Complementary to this distinction, one might distinguish two kinds or 'reading', one primary, the ability to follow the text as such, i.e. words, sentences, paragraphs, etc., and one secondary, the ability to 'see' the way in which the text-as-designed is organized. This would involve things like being able to see where a new paragraph, section or chapter starts, which parts of the text are 'covered' by a title, etc. Such 'secondary' abilities seem, to a certain extent, to be specific to text genres. Reading a novel is different in many ways from reading a newspaper report, or a scientific paper. The latter, as I will argue later, requires rather special competencies. My general impression is that these secondary reading abilities are rarely explicitly included in instructive texts, although some suggestions can be inferred from some treatments of writing (cf. Becker, 1986; Gilbert, 1993).

Writing and designing texts, then, involves the ability to 'pre-structure' adequately the reading of the text in an unobtrusive, 'natural' way. In so doing, the writer/designer creates a 'reading path'(3), in the sense that the reader is provided with various 'instructions' as to how to move through the text, which is also a 'guided tour', in the sense that there are various instructions for how to interpret the various textual components, i.e. as 'title', 'introduction', 'list', etc.

In short, what I'm trying to get at is that:

For ethnomethodology, these considerations open up a hardly explored field of investigations, i.e. the situated practices of reading/writing texts(4). The purpose of this paper is to use HTML as an 'explicating device' for exploring the ethnomethodology of this field.

What is HTML?(5)

In simple terms, HTML or HyperText Mark-up Language, is a set of conventions used to publish documents on the World Wide Web, often called 'the graphic part of Internet'. Such documents can be stored on a Server, connected to the Internet, and be read by using a special program, a Browser, such as Netscape, on a PC (or a Mac), after a connection has been made with an Internet Provider. For instance, since the fall of 1996, I have maintained an electronic version of ETHNO/CA NEWS by sending a collection of documents written in HTML to one of my university's Web Servers located in Amsterdam. Anyone with a PC connected to the Internet can use a browser to go to the URL:, and from there one can explore the contents of this site. One can down-load files such as the bibliography of EMCA studies in the 1990's, or one can use one of the indications of other sites, so-called 'links', to visit other places, for instance the EMCA bibliography until 1989, made available by Yoshifumi Mizukawa on a server in Japan.

Hypertext Markup Language (HTML) was developed from Standard Generalized Markup Language (SGML). Both are used to describe document layout, rather than document content. In essence, an HTML-document consists of a combination of text and tags. The text is in ordinary ASCII-format, while the tags 'describe the layout' and function as 'instructions to the browser'. Characters that are not available in the restricted ASCII-set, such as accented letters, can be included as special codes (marked by '&' and ';')

Tags are always marked by '<' and '>'. Some of the main kinds are:

In most cases, the part of the text to which the instructions refer will be enclosed by two tags, an opening tag and a closing tag, which is the same as the opening one, except for a / preceding the instructions, for instance <p> starts a paragraph, while </p> closes it.

It should be noted that some of the usual formats for structuring even a plain ASCII text are ordinarily ignored by browsers and therefore cannot be used routinely. These include indents, additional spaces and carriage returns. When one includes multiple spaces, indents and/or carriage returns, these are 'converted' into single spaces. In other words, special formatting tags need to be used to emulate such functions. If one wants to preserve the formatting of a specific part of one's text, such as a transcript excerpt for CA, this can be done by enclosing that part within special tags for 'preformatted text', i.e. <pre> and </pre>.

Writing in HTML

An HTML-document can be produced in a variety of ways(6): In my own practice, I have often used these methods in various combinations. At first, I started with a document in a word processor, i.e. WordPerfect 6.1, and used macro's to convert various WP formatting codes into HTML tags. Then I added various additional tags, such a the required ones, and saved the document in ASCII format, i.e. as a DOS text. That worked all right, for instance, in making a bibliography, which I had produced over the years in WP, available in HTML. I could then import it in a text editor to add the final points, or in an HTML editor to have it checked. The present paper was first written in WordPerfect 6.1, then converted into HTML using Corel WordPerfect 7. Then I used the Page Composer that goes with Netscape's Communicator 4.01 to put the layout into shape and revise various details. Since that is a WYSIWYG-program (What You See Is What You Get), however, I was at times confused about what caused certain layout effects and therefore opened the file in a text editor to clarify the situation and fine-tune the tags. In short, HTML-editors and converters are easy to use, but text editors offer an inside look and optimal control.

Reading HTML

The basic requirements for reading from a screen are similar to those for reading a printed page, or a hand-written one for that matter. The eyes move from left to right, then search for the start of the next line, but now, when the eyes meet the bottom frame, the finger touches a key, such as "PgDn", instead of turning the page -- and that more often.

When reading a paper, a journal, or a book, one has various means for locating the page one is reading in its larger setting, i.e. the paper, journal or book in its entirety. There will be page numbers, possibly head or foot texts repeating a chapter title. Furthermore, one has the physical object in one's hand: feeling the thickness of the pages already read and those still ahead. It depends, of course, on whether the physical object is the same as the text unit, as with separately printed papers, and especially with books, or whether the text unit is included in a larger collection, as in a journal issue or a paper collection. But in any case, there are material possibilities for locating one's current reading within the context of the reading task as a whole.

When reading a text from a screen, however, getting a similar overview can be much harder. This is partly dependent on the reading 'environment', i.e. the program used to project the text. Word processing programs will have various devices to provide some orientation and to 'scroll' the text, or, to put it in terms of a reader's perspective, to 'travel' through the text. The current page number and the line the cursor is on can be projected at the bottom of the screen. But when no section boundary or title is accidentally available on just this screen image, no clear indications of 'where one is' are available. The 'travel options' provided are general ones, like 'next screen', 'previous screen', 'next page', 'previous page', 'top of document', and 'end of document'. There will also be 'search' facilities, which allow one to type a 'string' of characters, i.e. words, to be searched for. These possibilities to 'travel' the text, however, are largely indifferent to the typographical 'chunk marks' the writer may have provided. Furthermore, apart from scrolling by line, sentence, paragraph, screen or page, these facilities tend to leave the reader-user in the blind as to the 'distance' he or she has traveled. A special case is the ability to place a 'bookmark' (mark the metaphor!), either a temporary, anonymous one, limited to one per document, or a permanent, named one (in WordPerfect). But these seem to be hardly used.

When one reads an HTML-document in a browser, one has even fewr resources to get an 'overview' of the text one is reading. An HTML-document does not have 'pages', such as a document in a word processor has, and no indications similar to page numbers are provided(7). What is called  a page in Internet language is the complete file. So the only guide available is the right-hand scroll bar, which provides a rough indication of the position of the current screen between the top and the bottom of the document.

Taking Netscape Navigator as an example, the general possibilities for 'traveling' within a document are limited to scrolling per line, using the 'up' and 'down' arrow keys, per screen, using 'PgUp' or 'PgDn', and going to the start or the end of the document, using 'Ctrl-Home' and 'Ctrl-End', while extra scroll functions per line and per screen are available with the mouse on the right-hand scroll bar. Finally, one can use a 'search' facility, to search for a string. No other systematic 'browsing' functions for traveling within the document seem to be available. In short, apart from the writer-constructed extra facilities, which I will discuss later, reading an HTML-document largely consists of a line-by-line and screen-by-screen following of the text provided, lacking a general orientation to 'where one is' in relation to either the document's beginning or end.

Guiding the reader in HTML 1: sections and documents

The previous discussion suggests that a reader of an HTML document lacks some of the mundane means of orientation as to 'where one is', compared with his or her conventional colleagues reading a printed text. In the present section, I explore some of the ways this lack of orientation can be repaired. A first issue is the 'recognizability' of the text's structure. In printed formats, paragraphs can be marked in a number of ways, the most common one being 'first line indent' and 'blank line' between paragraphs. HTML supports the last method and ignores the first. This means that if an author uses the first method for his or her original text, and later converts it automatically, all paragraph markings will disappear.

Some writers use an extensive numbering system, to 'locate' their sections and paragraphs into an overall system. In printed texts, this may seem a bit overdone, but in HTML documents it has some virtues to overcome the lack of orientation I have noted. The papers in  SOCIOLOGICAL RESEARCH ONLINE are now systematically formatted with paragraph numbers specifically for this purpose. This structuring can be further assisted by the following technique, based on the 'internal links' facility, explicated earlier. This involves adding a 'name tag' to each section title and starting the document with a table of contents, where each section title is formatted as an 'internal link' referring to the corresponding 'name tag'. In such a case, the reader who opens such a document can choose to jump to any section. And when one wants to look up a specific but different section while being somewhere in the middle of the text, one can jump to the top of the document, make one's choice and jump to that section's top. After consulting that section, the browser's 'Back'-function should bring one back to the original position in the text (you might take a minute to practice this reading technique in the present document). In practice, this may not always work, as the number of recorded previous positions is limited. So one has to remember and again choose the original section. This format is often used for documents giving organizational information (as on my faculty's home page:, but it can be used in any setting.

An author can also choose to distribute the information over a number of separate documents, made available in the same directory of the server. Then 'local links' can be provided to go from one to the other, and, again, the 'Back' button can be used to return to previous documents. Using 'external links', an author can provide for a reader to leave the original document, directory or server altogether, to travel the WWW in its entirety. In such a case, the concept of 'author' becomes ambiguous, of course, as the currently unresolved legal discussions regarding copyrights and various liabilities demonstrate. The example, given earlier, of providing a link to the bibliography on Yoshifumi Mizukawa's home page in the context of my ETHNO/CA NEWS pages, is a case in point. Sites like my colleague Albert Benschop's SocioSite basically consist of an enormous collection of such links. (In fact, giving such links as I do here is a bit risky, since it provides you with an inviting opportunity to leave my text altogether).

The metaphor of 'surfing' is now generally used for reading (part of) a document here, and then using a 'link' provided in that document to pop up another, which may be on the same or a different server, and so on. It focuses on the user's 'freedom' to travel the Internet, i.e. the network of servers, documents and links, but it presents that network as a kind of 'nature', a non-human unpredictable field of unlimited possibilities which can only be conquered by brave and experienced travelers. In that sense, it hides the Internet's character as a collective product of a distributed population of servers, providers, designers, writers, etc. The other metaphor in use, 'browsing', suggests the improvised and ad hoc reading of a page here and a chapter there by a library user or bookshop visitor. This at least pictures the field in a more realistic fashion. But both metaphors focus on the user rather than on the structure of the field of exploration, i.e. the fact that it is pre-structured and littered with signposts, toponyms, guided tour advertisements, etc. In other words, the Web as a field offers the user choices from pre-given lists of possibilities. The experience that the number of possibilities is enormous may divert one's attention from the facts of pre-structuring and guidance. For the present occasion, I will -- in the next section -- explore one particular kind of, and context for, such pre-structuring and guidance: notes and references in scientific papers.

Guiding the reader in HTML 2: Notes and References

In this section, I will first discuss the use of notes and references in printed texts, i.e. scientific papers, followed by an exploration of how papers formatted in HTML can and do provide notes and references. In so doing, I will confront writers' and users' perspectives. My interest in these issues derives, firstly, from my experience as a reader and writer of papers in both formats. It can also be argued, however, that even in conventional printed formats, notes and references represent a 'hypertextual' element. It is, therefore, of special interest to observe how these elements are used in a format that is specifically built for 'hypertextuality'.

Looking at the current formats used in books and journals, and the editorial guidelines that are provided, three broad styles of formatting notes and references can be discerned(8):

These are 'idealizations', of course, of a more complex and varied set of practices. Each of these styles has its particular pragmatic as well as aesthetic properties. Writers, readers, editors and publishers will have their own preferences in these matters.

Journals often provide instructions for the formatting  to be used in submitted papers which express a preference for one or another style, and which is, of course, reflected in the format of the papers published. In Discourse & Society, for instance, it reads: "Essential notes should be indicated by superscript numbers in the text and typed at the end of the text. References in the text should read: Brown (1987: 6304), (...) All text references should be listed alphabetically after the notes, as follows (...)". And indeed, the papers quite often have no notes at all, or only a limited number (say 5). So this is a clear example of style B, as described above.

Social Studies of Science, another journal published also by SAGE, has the following instruction for using notes: "The endnotes should be typed, double-spaced on separate sheets, numbered consecutively. They should be referred to in the text by numerical superscripts." The instructions do not mention the possibility of using the (name, year: pages)-method and it does not seem to be used in the published papers. All papers have endnotes, some over a hundred, containing remarks, elaborations and references to the literature. So here, it is style A, that is recommended and used.

Human Studies has the following in its instructions (copied from the Kluwer Web-site):

Endnotes should be indicated by consecutive superior numbers in the text.
Material to be endnoted should be typed separately and included before References at the end of the manuscript. Endnotes are only to be used when the author frames them in a discursive way. References should not be given in complete form in endnotes. The author-date system described below is sufficient."
So this is another clear case of 'style B'.

The choices, these journals make, are not accounted for explicitly, but, in general, publishers seem to 'favor the Harvard system and like to minimize (end)notes -- mainly for clarity and ease of reading'. But as one of my informants wrote: "So there are no hard and fast rules -- it's a mixture of disciplinary tradition, publisher, editor and author preference."

Within my own field, conversation-analytic papers quite consistently use 'style B', in some cases with quite an extensive set of notes(9), while Mike Lynch, an ethnomethodologist working in the area of science studies, generally uses a version of 'style A'. As he wrote to me:

As you observe, reference styles are often dictated by journals. I've used both 'A' and 'B'. My first book employs long endnotes together with author-date citations. My preference is to put complete references in notes, and to have notes printed on the same page as the text. Many journal (and edited book) editors discourage long footnotes, especially in 'empirical' fields like sociology and psychology. History and philosophy journals and books tolerate (indeed encourage) long footnotes/endnotes. (...)
From a writer's point of view, style A is simple in the sense that anything to be added to the body of text is typed in notes numbered subsequently. When typewriters were still used, it was hard to add a note later to an already typed text, because one had to renumber the notes. With advanced word processing technology, such difficulties have been solved, although some users report difficulties with large numbers of endnotes, especially when placed at a non-default location. In my experience, writing in style B does not seem to be any harder, again especially when using a word processor. Author-date references can be added anytime in the body of the text, while a list of references can be compiled at the end. For the latter purpose, I use a 'macro' that collects all lines that have '19' in them in a separate file. In this way, all references to publications dated from 1900-1999 will be included as well as some other materials. From that file, I delete all spurious information, after which I alphabetize the remaining names-dates list. Then I can add the required bibliographical information from special bibliographical files, constructed over the years. The 'minimal notes' requirement seems useful as a stimulus to think about what one tries to get across and why. Quite often, material that in style A would be put in a note, will now be either included in the text, forcing one to link it to the ongoing argument, or deleted as basically superfluous.

Before considering a reader's point of view, I will discuss the ways in which notes and references can and are being used in HTML. The basic HTML technology, for doing the work notes and references are supposed to be doing in printed texts, is making 'internal links', as explained earlier. So many papers published in electronic form, such as those in SOCIOLOGICAL RESEARCH ONLINE, as well as the current one, have notes indicated in the body of the text by a number, which is formatted as an internal link connected with a 'name tag' at the start of the note text at the end of the paper. So when the reader clicks on the number, the text of the note is displayed at the top of the screen. After reading, clicking the browser's 'Back' button returns one to the original location in the text. In a similar fashion, author-date references in the text are formatted as internal links connected to name tags in the references list at the end. Clicking the name-date in the text, therefore, leads to a display of the full reference at the top of the screen, while the 'Back' button allows one to return to the original location. (You can check how this works for you now!)  In these cases, the author only provides for the move out-of-the text, while the reader has to find his or her way to 'go back'. An alternative technology is to use 'double cross tags', to the number and name tags mentioned above. This would involve adding, in the document itself, a 'return to the text' tag with the reference or note text, and a 'name tag' at the original location. Using this 'browser independent' technology with name-date references runs into difficulties, however, in those cases where the text refers to the same title more than once. (You can check this out in Rodino, 1997, for instance)

'Internal' hypertext technologies, like these, may be supplemented with 'external' ones. If one refers to a document that is available on the WWW, one can include an 'external link' to it. A reader can choose to 'go to' that particular document by clicking on its mention in the original text (or in its reference section). After reading this other document, the 'Back' button may be used to return to the original text, or one can consult the 'history' list of one's browser to do so (but see Tauscher & Greenberg, 1997). From the point of view of the original writer, providing external links may be a 'dangerous' option, as I suggested before, since it may lead to 'loosing' some of one's readers. The 'double crossing' technology, discussed above, is unlikely to be available in these cases.

Scanning some of the papers available on the net, I have to conclude that these technologies, especially the internal ones, are quite often not used. Many papers, for instance on personal home pages, have apparently been written originally for printed distribution. Some have been already published in printed form, or have been submitted for publication. In such cases, the original format is often just converted to HTML (often by another person than the original writer), without adding 'internal links'. When reading such a text from the screen, note numbers or author-date references are not linked to their target places in the text at all. One would have to go to the end of the file to look them up, but then it is not too easy to return to the original location (it can be done by using the browser's search facility and use the note number as the search string, but then one has to pass by all other places where that number is mentioned also). Of course, such papers can be printed and then read as intended, but then the electronic medium is 'just' used as a cheap distribution device for a basically non-electronic object. I have been told that some readers of online journals print the papers they are interested in, to read or store them in printed format. Editors-formatters of such journals may anticipate such an alternative use in their formatting style. On the other hand, reading a paper that has been fully formatted for from screen reading, including internal links, from a printed version, can also lead to strange experiences. One of the readers of an earlier version of the present paper printed it to read it on the train back home from his office. He reported that some of the links I provided irritated him, because their formatting suggested 'emphasis', rather than 'click here to know more'. These observations support the claim that there is no sharp division between printed and electronic texts, or plain texts and hypertexts. In the actual practices of reading, writing, and formatting, there are many shades, representing ad hoc selections from, and reactions to, the technologies available. As each format has its own conventions, however, a reader may be confused, or become irritated, when reading a text in a different format, based on different conventions, than one is used to. The readability of transcriptions in the conversation analytic tradition is a case in point.

These observations return us to some of the technical issues of writing and designing HTML-documents, or converting existing documents to an HTML format. As discussed previously, HTML documents can be made in a number of ways, using a text-editor, an HTML-editor or an HTML-converter. When one uses a text editor, all links have to be typed in full at their required location. An HTML-editor will provide a number of options to create tags with one or more clicks. And a converter should 'translate' all formatting information included in the original text into the new formatting language. In actual practice, there are various complications. This is especially true for the converters, which are most probably used to format existing papers in HTML.

The 'Internet Publisher', which goes with Corel WordPerfect 7, converts most existing formatting information correctly. Some formatting information included in ordinary word processing documents is not compatible with HTML-rules, however, so it is either deleted or changed. For my present purposes the most important is that footnotes are formatted as endnotes. This is, of course logical if one considers the complete document as 'a page'. A more important point is that one may have to prepare one's document for typical hypertext functions. One can, as I have done for the first electronic version of this paper, use the 'bookmark' function to create hypertext links from author-date mentionings in the text to the relevant items in the reference list. The Internet Publisher converts these links correctly into 'internal links' in the HTML-format.

The 'Internet Assistant', which can be added to MS-Word 7, works similarly to the 'Internet Publisher', except that both footnotes and endnotes are simply deleted during the conversion process. This is no small handicap for converting scientific papers using this program. Since bookmarks and hyperlinks to them can be made using Word 7, the solution seems to be to convert all footnotes or endnotes to marks and links before converting, which is a rather cumbersome procedure.

It is against this background of the use of notes and references as hypertext links, that I want to consider the situated work of reading such notes and references in printed texts. It is, indeed, the crucial point of this paper that HTML as an hypertext technology, has opened up for review some of the hidden aspects of mundane reading work. Notes and references 'invite' the reader to interrupt his or her current 'following' of the text, the regular scanning of the lines, as well as the related interpretive work, to jump to another element of the text, a side remark, a temporary diversion or a reference to a source. The reader may choose to ignore or postpone this 'invitation', as many say they tend to do. But even considering this choice may serve as an interruption of sorts. In my own experience, I think my decisions are strictly local, depending on the text, the occasion, my general interests at the moment, and the interest raised by the 'invitation', or not. But design features of the text and the 'invitation' also play a role.

Consider footnotes contrasted with endnotes. From a reader's point of view, there are important pragmatic differences between the two. When the reader meets a footnoted text, the number in the text invites the eyes to move to the bottom of the page. After having read the note, one most of the time has an overall memory of the location where one started to travel down, which one can re-find in a glance, using the number as a find indicator.

With endnotes, it's a different story. One first has to find the location of the notes in general, then look for the note with the given number. After reading the note, one has to re-find the original page, and the location on the page where one interrupted reading the body of the text. Quite often, the location memory will have been blurred by the more extensive search procedure and the reading of the note itself. Most readers will use a finger, to mark the text page, but this makes searching the page where the note is to be found more difficult. So when reading a text with many endnotes, one may use some other object as a bookmark, and possibly still another to mark the page that contains the current endnotes. When no separate bibliography is provided, subsequent mentionings of an item referred to repeatedly are often shortened using an op.cit. or ibid. abbreviation. This method requires even more work, as one has to search the previous notes until the first mentioning is encountered, or depend on one's memory.

In my own reading, when I look up an endnote, I often find myself scanning the texts of the next few notes, to see whether they are at all interesting to look up when their number will turn up in the text. What I find most frustrating is the experience of doing the work of looking up a note (and returning afterwards), only to find an obvious bibliographical reference. The numbers in the text, however, do not provide any information on the character or content of the note to which they refer. As one reader remarked, one can also scan the references list in order to be able to recognise which titles are referred to by using 'style B'.

When the eyes meet an author-date reference, the reader again has a choice, to follow up on the 'invitation' to check-out the reference, or to ignore it for the moment. The decision may depend on various considerations. On the one hand, the preceding text may present the reader with a kind of puzzle, such as a less probable claim, for which the reference might be a support, and therefore a sort of solution to the puzzle. Or the text may deal with a problem of particular interest to the reader, who would like to know which authors have more on it. Such might be positive motivations for following the invitation. When, on the other hand, the reference seems to be of an 'obligatory' kind, or the preceding text rather uninteresting, the invitation may be ignored. Still a different situation may emerge when one works in a specific field, and reads a text from that field. Quite often, one can recognize some references from the name and date information itself. I myself would not be tempted to interrupt my reading to check-out a reference like Garfinkel (1967), Schegloff & Sacks (1973) or Suchman (1987)!

In short, note numbers and author-date references present the reader with a dilemma: follow the invitation to interrupt the current reading or ignore it, with the possibility to check-out the extra information later. A number of considerations can be taken into account in this decision. One of these may be the ability to quickly return to the original location of reading. For this purpose, HTML-technologies offer the best opportunities, although these are not always used in documents published on the Web. In printed documents, essential footnotes and author-date references seem to offer the best opportunities for 'quick returns', the first because of the same page location, the second because they already provide some information that helps to inform the choice.


In this essay, I have explored some partly hidden aspects of the culture of textuality in (social) science. I have used HTML as an 'explicating device' to study practices of reading, especially those involved in 'hypertextuality', i.e. the reading of notes and references to the literature. It seems that the path out of the text, from the source to the target, is generally better marked than the path back, from the target to its source. Current word processing programs support both footnote and endnote formatting, as well as 'bookmarking' which can be used for building 'jumps' to bibliographical information. The latter works for electronic texts only, not for printed ones, of course. These programs, or their add-on's, allow conversion to HTML. MS Internet Assistant, the add-on for MS Word 7, however, deletes all notes; as a program designed for business purposes, it does not support this essential element in the culture of science.

This, then, has been an effort to do an ethnomethodological exploration of aspects of the technology/practice interface in the field of textuality. I have not done an extensive literature search for this paper, but my general impression is that some of the existing writings on HyperText, HyperMedia and Web publishing tends to be rather abstract and ideological, while others are utterly practical and technical. What I have tried to do is focus on the embodied work of 'living' hypertext formats, compared to more conventional ones. The abstract literature tends to focus on literary texts, while the practical is mainly oriented to the needs of organizations to inform its workers and/or clients. My interest was to see how science is done, bringing it to life in (hyper)texts.

In the abstract literature, there is a tendency to make a sharp contrast between two kinds of 'reading', one 'linear', as in conventional texts, one 'non-linear' as in some literary experimental works. Hypertext formats would be especially fit to allow such 'non-linear' reading, and would, in so doing, provide the reader with a new 'freedom'. As McHoul & Roe (1996) have argued, such contrasts do not make very much sense. I agree that the contrast between ordinary, 'linear' texts, and non-linear 'hypertexts' is overdrawn. In actual (hyper)texts, one sees a mixture of linear and non-linear elements. Furthermore, the reader is not 'enslaved' by a linear text; he or she can move about. The various techniques an author can use to structure his or her document, are not just orders to be followed. They can also be used as means of orientation for those who prefer to 'browse' or 'scan', i.e. to construct their own selective path. The 'freedom' offered the hypertext reader, on the other hand, is a pre-structured one, providing for a choice to follow one or the other 'reading path'. I have suggested that within the separate texts, which are the elements in the hypertext network, the reader may lack the means of orientation he or she needs to 'move about' at will. So in a sense such texts offer rather less than more possibilities to leave the pre-structured path.

Finally, I want to work-up the phrase that reading, as well as writing, is an embodied activity. I have written about the eyes scanning the lines, the fingers striking a key, moving and clicking the mouse, and also marking the page. There is one bodily contrast between conventional and electronic reading that needs to be stressed, however. Screen reading requires a relatively fixed body position. A book or journal can be laid on a desk, held in the hand, moved about at will, allowing the body to change its reading position frequently. The screen for most personal computer users, however, is always 'there', forcing the body, and especially the head, into a stiffness that readily becomes uncomfortable.


* An earlier version of this paper, 'Structuring writing for reading: HTML as an explicating device', was presented at the International conference: ETHNOMETHODOLOGY: EAST AND WEST , Waseda University, Tokyo, Japan, August 21-23, 1997.  I would like to thank Leslie Jarmon (Indiana University, U.S.A.), Christoph Meier (Universität Giessen, Germany), and Stuart Peters, (Sociological Research Online, University of Surrey, U.K.), for interesting reflections and helpful comments on a draft version, and the conference audience for its attention and support.

1. In earlier work (i.e. Ten Have, 1990), I have argued that by focussing on the procedural basis of social order, provided by mostly unnoticed common sense methods, ethnomethodology has a peculiar methodological problem, which I glossed as 'the invisibility of common sense'; explicating devices can be seen as techniques for finding practical solutions to this problem, by making common sense methods 'visible', at least to a certain extent.

2. This is, of course, a very crude rendering of current views on reading and understanding texts. One essential element that should be added is 'knowledge of the world', but there are many other elements and complications that have been explicated in the relevant literatures. The summary in the text is one only 'for the sake of the (my local) argument'.

3. A concept coined by Landow (1992: 7), cited in McHoul & Roe (1996).

4.  See Alec McHoul (1982), who has been among the first to explore this area in any serious way; cf. also: Anderson (1978), Atkinson (1983), Morrison (1981).

5. There are numerous books on HTML, mostly oriented to instructing (would-be) writers of HTML documents, such as Heslop & Budnick (1995), or the very extensive Lemay (1997), but a lot of information is, of course, also available on the WWW itself.

6. Since I rely on my own experience, I can only explain the programs I have used, which are mostly Windows-based. For references to various HTML-editors, etc. see the sources mentioned in the previous note, and: Mag's Big List of HTML Editors.

7. Up-to-date word processors, like Corel WordPerfect 7, or Word 7, support reading HTML-documents in a way similar to a browser; the available functions are a mixture of ordinary word processor and browser ones; page numbers are provided, for instance, but the 'Back' button does not work as it does in a browser. I will not discuss these differences in detail.

8. The first part of this section is partly informed by E-mail correspondence with Teun A. van Dijk (editor Discourse & Society), Michael Lynch, George Psathas (editor Human Studies), and Jane Price (senior production editor journals, with Sage Publications).

9. Schegloff (1996), for instance, has 62 endnotes, taking more than 11 pages, for a paper which has 62 pages of body text; but this is an exploratory paper.


Anderson, D.C. (1978) 'Some organizational features in the local production of a plausible text', Philosophy of the social sciences 8: 113-35

Atkinson, J.M. (1983) 'Two devices for generating audience approval: a comparative study of public discourse and texts'. In: K. Ehlich & H. van Riemsdijk, eds. Connectedness in sentence, discourse and text. Tilburg: Katholieke Hogeschool Tilburg: 199-236

Becker, H.S., with P. Richards (1986) Writing for social scientists: How to start and finish your thesis, book, or article. Chicago: University of Chicago Press

Benschop, A. (1997) 'Lineaire en Hypertekst' [Linear and hypertext; in Dutch only], online only.

Garfinkel, H. (1967) Studies in ethnomethodology. Englewood Cliffs, N.J.: PrenticeHall

Garfinkel, H., H. Sacks (1970) 'On formal structures of practical action'. In: J.C. McKinney & E.A. Tiryakian, eds. Theoretical sociology: perspectives and developments. New York: Appleton-Century-Crofts: 338-66

Garfinkel, H., D.L. Wieder (1992) 'Two incommensurable, asymmetrically alternate technologies of social analysis'. In: G. Watson, R.M. Seiler, eds. Text in context: studies in ethnomethodology. Newbury Park, etc. Sage: 175-206

Gilbert, N. (1993) 'Writing about social research'. In: Gilbert, N., ed. Researching social life. London: Sage: 328-44

Have, P. ten (1990) 'Methodological issues in conversation analysis', Bulletin de Méthodologie Sociologique, Nr. 27 (June): 23-51

Heslop, B., L. Budnick (1995) HTML Publishing on the Internet, for Windows. Chapel Hill, NC: Ventana

Lemay, L. (1997) Teach yourself web publishing with HTML 3.2 in 14 days. Second professional reference edition. Indianapolis: Publishing

Landow, G. P. (1992) Hypertext: The convergence of contemporary critical theory and technology. Baltimore: The Johns Hopkins University Press. [cited in McHoul & Roe (1996)]

Lynch, M. (1985) Art and artifact in laboratory science: a study of shop work and shop talk. London: Routledge & Kegan Paul

McHoul, A.W. (1982) Telling how texts talk: essays on reading and ethnomethodology. London: Routledge & Kegan Paul

McHoul, A., P. Roe (1996) 'Hypertext and Reading Cognition'. In: Gorayska, B. and Mey, J. L. (eds), Cognitive Technology: In search of a humane interface. Elsevier: 347-59

Mehan, H., H. Wood (1975) The reality of ethnomethodology. New York: Wiley

Morrison, K. (1981) 'Some properties of "Telling Order Designs" in didactic inquiry', Philosophy of the Social Sciences 11: 245-62

Pollner, M. (1979) 'Explicative transaction: making and managing meaning in traffic court'. In: Psathas, G., ed., Everyday language: studies in ethnomethodology. New York: Irvington 229-55

Rodino, M. (1997) 'Breaking out of Binaries: Reconceptualizing Gender and its Relationship to Language in Computer-Mediated Communication', Journal of Computer-Mediated Communication 3(3)

Sacks, H., E.A. Schegloff, G. Jefferson (1978) 'A simplest systematics for the organization of turn taking for conversation'. In: J.N. Schenkein, ed., Studies in the organization of conversational interaction. New York: Academic Press: 755 (1974)

Schegloff, E.A. (1996) 'Turn organization: one intersection of grammar and interaction'. In: Ochs, E., E.A. Schegloff, S.A. Thompson, eds., Interaction and Grammar. Cambridge: Cambridge University Press: 52-133

Schegloff, E.A., H. Sacks (1973) 'Opening up closings', Semiotica 8: 289327

Schwartz, H., J. Jacobs (1979) Qualitative sociology: a method to the madness. New York: Free Press

Suchman, L. (1987) Plans and situated action: the problem of human-machine communication. Cambridge: Cambridge University Press

Sudnow, D. (1978) Ways of the hand: the organization of improvised conduct. London: Routledge & Kegan Paul

Tauscher, L., S. Greenberg (1997) 'How people revisit web pages: empirical findings and implications for the design of history systems' (online only)

Wieder, D.L. (1974) Language and social reality: the case of telling the convict code. The Hague: Mouton

Zimmerman, D.H. (1969) 'Record-keeping and the intake process in a public welfare agency'. In: S. Wheeler, ed. On record: files and dossiers in American life. New York: Russell Sage: 319-45

Back to Personal Home Page Paul ten Have