https://d2ng68cf36kxsk.cloudfront.net/store/public/post/1193/post_images/Slide372.JPG">

Evelyn Berezin, daughter of Jewish Bronx, New York, invented the first word processor for offices, the Redactron of 1971. In a male-dominated industry, she decided to found her own company to manufacture and market her machine. An early advertisment hailed “the death of the dead-end secretary.” When asked more than forty years later whether she anticipated that her creation would endanger women’s jobs, she said, “I’m embarrassed to tell you that I never thought of it — it never entered my mind.”⁴⁵⁴

The first word processor to reach a wide consumer market was WordStar for personal computers, launched in 1978. Word processing also meant that the broad masses of text-creators were required for the first time to learn some of the half-millennium long, arcane technical language of letterpress typesetting: fonts, point sizes, justification, and if they wanted to dig deeper, other obscurities such as leading, the spaces between the lines which had formerly been blocked out with lead in the type formes. These are all concepts related to the visual design that adds meaning in the surrounds of text.^§AS1.4.7c

One of the most profound accompanying changes was to take text and its creation in the direction of a less linear architectonics. As material practices, handwriting and typing had until then been constrained by resistant, linear temporality of the line on a page, grapheme after immovable grapheme. Changes and corrections created a mess on the page, and when things got too bad, there was no alternative to the pain of having to rewrite or retype the page.

In this respect, scribing and typing suffered the same disadvantages as speech. Where the linearity of speech required audible self-correction, the linearity of scribed or typewritten text required visible amendment. Typesetting and word processing make changes invisible. The page is always clean, and correction friction-free.

The text you see before you now has gone through many iterations, starting with messy fragments, notes and instructions to selves that have been written over, and over again. Words, phrases, and sentences have been moved around many times, rethought, reframed, and rephrased. The job of text work has changed from one dominated by resistant linearity to infinitely fluid, multilinear complexity.

This may not be a good thing in some respects. But perhaps the text you are reading is awkward for the jumping about in its writing, perhaps we were afforded too much scope for revision. The neatness of a draft may deceive. Nobel prize-winning novelist Günter Grass distrusted the illusion of the text looking finished after a first draft, preferring for his own writing the slow, methodical process of scribing and then typing the novel three and sometimes four times.⁴⁵⁵ Whatever the disadvantages and no matter how subtle the differences in the final text, the materiality of medium affects its meaning. This multilinear architectonics in digital text work makes it less than ever like the work of speaking.

Digital text adds to naming and counting^§MS1.3.2 a third vector of transposition into quantity: rendering. In writing, this vector runs: text > alphanumeric naming of graphemes (Unicode) + visual design > decomposition into pixels > recomposition for rendering on screens or on paper as digital print.

The fact that this vector of transposition uses the same technologies of naming and calculation as image^§AS1.4.6c and sound^§AS1.4.2d accounts for the deep multimodality of contemporary texts. Until photolithographic and then digital printing,⁴⁵⁶ realistic image and text were placed for practical purposes on different pages or sections in books. Until digital television, there wasn’t much text in video – but now on the business and sports channels, there is a cacophony of text-image overlays because this possibility is so readily available.

Digital video layers sound over moving image with far greater ease than film because, though analogue sound and film were fundamentally different technologies, an electromagnetic sound track is merely stuck beside the film image. Even though the juxtaposition of technologies is awkward, the processes of transposition of sound and image are roughly the same in analogue and digital moving image and not much changes with digitization. But layering text into digital image is very much easier, and so this affordance has pushed moving image into new domains of possibility.

And another hugely significant thing happens in the transition from analogue media to digital. Every rendering is a recalculation. In analogue print, each new page, by dint of economies of setup and scale, was the same as the previous. In digital print, whether that be printed bills sent in the mail or rendering of a video to a screen, every image is recalculated for its rendering. This affords enormous possibilities for variation, including in the case of screens, user interfaces which allow the reader-viewer to co-design the reading-viewing experience.

Digital media no longer have lock-step replication built into the processes of their manufacture. Every digital text, image, or sound rendered is a re-manufacture, on-the-fly. Between the source and its rendering are numbers that bear no visible or speakable resemblance either to the source or its rendered destination. This allows space for the re-manufacture to be modified in the using or by the user.

Which brings us to Charles Goldfarb. Trained as a lawyer and confounded by the labor-intensive processes of legal documentation, Goldfarb took a position in the IBM labs in Cambridge, Massachusetts. There in 1969, he and several colleagues created Generalized Markup Language (GML),⁴⁵⁷ the first conceptual departure from the five-century-old discourses and practices of typography. They retained the word “markup,” the instructions to the typesetter about the visual design of the page written around a handwritten or typed manuscript. But the fundamental principle of markup that Goldfarb and his colleagues established was new.

Using GML, tags would be inserted into the digital text specifying paragraphs, sections, headings, tables, lists, and the like. These tags did not indicate how these meaning functions were to look, or the form they were to take when rendered. Instead, there were to be separate “stylesheets” which rendered text in different ways depending on the end device, whether, for instance, that was a paper printer or a screen. So, re-manufacture was not reproduction. The redesign was not a replicant of the design. Or, to use terminology of this grammar, the tags indicate meaning functions; the stylesheets determine the particularities of meaning form as realized in variable media.

Following these foundational principles were Standard Generalized Markup Language (SGML, 1974),⁴⁵⁸ then Hypertext Markup Language (HTML, 1993),⁴⁵⁹ Extensible Markup Language (XML, 1996),^§MS3.3b and Microsoft Word XML (2003).⁴⁶⁰ This is how digital texts can today so easily move around from device to device, medium to medium.

A fundamental change in textual practice, this is often described as a change from visual markup established by typography to “structural and semantic markup.” In HTML, the tag <title> is semantic, and the tag <p> for paragraph is structural. But these distinctions are not so clear, because as well as naming a document, <title> is in a structural sense located in a particular place in a text for the purposes of meaningful rendering, and <p> is semantic to the extent that it defines the beginning of a new idea.

Starting with GML, these were systems for specifying text grammatically and using that grammar to manufacture the text. The tags are a meta-text that re-manufacture the text. Each tag is an executable instruction at the same time as it is a description.⁴⁶¹ This is what Goldfarb calls an “Abstraction support – that trait of an information representation (data format, notation, et al.) that allows its users to distinguish what they consider to be the ‘abstract’ information content of a document from the ‘style information’ that is used to render it.”⁴⁶² Here, in another way, the erstwhile linearity of texts is overcome, in markup that describes then renders information hierarchies, nesting, cross-referencing, and co-relational structures such as tables.⁴⁶³

Not that the path from SGML to subsequent digital markup has always been straightforward. The function/form distinction made so clearly in SGML became mixed up with textual–visual function, which included in early versions of HTML such horrors as the <blink> tag. It has taken many torturous revisions of HTML to return it to something closer to the founding principles of SGML.

Nor have we escaped typographical markup. Half a century after Goldfarb’s insights, we still use widely used back-to-the future formats such as word processor documents and Portable Document Format (PDF). These continue to insist on faithfully reproducing typographic markup in fixed pages, making them hard to render across different devices and resistant to textual intervention on the part of users.

---

⁴⁵³ John F. Cavanaugh, 1978, “Text Handling at the United States Government Printing Office,” Technical Communication 25(3):12–15.

⁴⁵⁴ https://www.nytimes.com/2018/12/10/obituaries/evelyn-berezin-dead.html?action=click&module=Well&pgtype=Homepage&section=Obituarieshttps://en.wikipedia.org/wiki/Evelyn_Berezin

⁴⁵⁵ www.theguardian.com/books/2010/aug/24/gunter-grass-last-autobiography; Matthew G. Kirschenbaum, 2016, Track Changes: A Literary History of Word Processing, Cambridge, MA: Harvard University Press.

⁴⁵⁶ Bill Cope, 2001, “New Ways with Words: Print and Etext Convergence,” pp. 1–15 in Print and Electronic Text Convergence, Volume 2.1: Technology Drivers across the Book Production Supply Chain, from the Creator to the Consumer, edited by B. Cope and D. Kalantzis, Melbourne: Common Ground.

⁴⁵⁷ Charles F. Goldfarb, Edward J. Mosher, and Theodore I. Peterson, 1970 [1997], “An Online System for Integrated Text Processing,” Journal of the American Society for Information Science 48(7):656–61; Charles F. Goldfarb, 1990, “A Brief History of the Development of SGML,” SGML Users’ Group, retrieved February 5, 2002 (www.oasis-open.org/cover/sgmlhist0.html).

⁴⁵⁸ Charles F. Goldfarb, 1991, The SGML Handbook, Oxford: Oxford University Press.

⁴⁵⁹ Tim Berners-Lee, 1990, “Information Management: A Proposal,” Geneva: CERN (www.w3.org/History/1989/proposal.html).

⁴⁶⁰ Liam Magee and James A. Thom, 2014, “What’s in a WordTM? When One Electronic Document Format Standard Is Not Enough,” Information Technology & People 27(4):482–511.

⁴⁶¹ Bill Cope and Mary Kalantzis, 2004, “Text-Made Text,” E-Learning 1(2):198–282.

⁴⁶² www.sgmlsource.com/infaqs.htm

⁴⁶³ Allen Renear, Elli Mylonas, and David Durand, 1996, “Refining Our Notion of What Text Really Is: The Problem of Overlapping Hierarchies,” Research in Humanities Computing 4:263–80; Dino Buzzetti and Jerome McGann, 2006, “Electronic Textual Editing: Critical Editing in a Digital Horizon,” pp. 53–73 in Electronic Textual Editing, edited by L. Burnard, K.O.B. O’Keeffe, and J. Unsworth, New York: Modern Language Association of America.

---

Kalantzis, Mary, and Bill Cope. 2020. Adding Sense: Context and Interest in a Grammar of Multimodal Meaning. Cambridge UK: Cambridge University Press. pp. 169–172. https://doi.org/10.1017/9781108862059

Charles Goldfarb’s Generalized Markup Language §AS1.4.7d