ql-code-block-container" spellcheck="false">

Elements can be structured hierarchically: <given name> and <surname> can both be child elements of <name> . Elements can have variable attributes. And a lot more stuff in an official specification that now runs to twenty thousand words.¹⁷⁸

This means that at the most basic level, XML can do things to address the fundamentals of meaning, things we have named in this grammar “instance,” “concept, “property,” “relation,” and together, “structure.” Unlike natural language, in XML these kinds of meaning are explicitly framed as such. There is no need to resort to context to work out the meaning of the same string of characters “cope,” which can in a database or on a screen refer to a mental state, a priest’s cloak (obscurely), a family, or one unambiguously identifiable person. Banks, employers, email recipients, and social media users need to know this unambiguously, and a string of characters is not enough.

This is why ontologies surround us these days, mechanizing our meanings with markup, a more structured and strategically unnatural shadow layer of meaning that is visible only in code. They reach us in stuff that is ordinary writing through XSLT “stylesheet transformations.” This means that the mention of a name in a diary entry looks different on my device from the way it looks in an alert from Facebook.

In the lingo of digital text, structure and semantics are kept separate from rendering or presentation. Or, we would say in the terms of our grammar, the forms of meaning are separated from their functions. This is not just theoretical. It is mechanical, a technology for flexible and on-the-fly rendering of texts across multiple devices and apps.¹⁷⁹

Here are some of the ontologies we live with, all expressible in XML and in practice too, mostly expressed in XML. For medicine, there are ICD and SNOMED.^§MS3.2a For chemistry, there is Chemical Markup Language.¹⁸⁰ For places, there is GeoNames,^§AS1.3.2b and for languages, Ethnologue.¹⁸¹ There are standards for the internet of things.¹⁸² For date and time, there is an ISO standard for which the W3C has a specification.^§AS1.3.1b For citation and bibliography, there are MARC, Dublin Core, and ONIX.¹⁸³ And for text itself, the mundane stuff of paragraphs and headings is transferrable across platforms and renderable across multiple devices, often through XML transformations: from Microsoft Word, to LibreOffice, and to the web as HTML.¹⁸⁴

And so on – there are ontologies for just about all the writable things in the world, even writing itself. In most domains, these ontologies have shaken down to one or two, and where there is more than one they are for somewhat different kinds of use. Specialist domains frequently get caught in the gap, and have to spend a good deal of time mapping equivalences or term-to-term “crosswalks.”¹⁸⁵

Ontologies are universal and translingual to the extent that their translation across different languages largely preserves their essential structures of meaning. ICD in English is not so very different from ICD in Spanish, or at least not as different as spoken English is from spoken Spanish.

Such translations are done in the ontology, not the point of rendering into text. When you select your preferred language at an ATM,^§AS2.2.2 the textual interaction which delivers cash from your bank account does so because a mixture of information you provide and information on your card has been marked up for universal banker’s meaning where “natural” language is a functionally trivial condescension to the person standing at the machine. The language of the interface is functionally detached from the ontologies of banking.

And the play of ontologies, though textually arranged in visualizable schemas, is not just text. Any digital object can be so marked up, not just text, but images, video, sound, code, or datasets – hence the fundamental multimodality of the structures of digital media, notwithstanding the textual form of the tags used for its structuring.

The underlying textual forms of ontology-driven meaning are not much like the grammar of spoken sentences, or speech transliterated onto the page, or even what is spoken by Siri. Rather, this is a spatialized text of instances, concepts ordered in hierarchies of dependency, arranged in relations of difference and similarity, and ascribed properties. You can do the grammar of structures and relations in the spontaneously speakable sentence, but not with the same degree of elaborate specificity, relative lack of ambiguity, and independence from context that is possible and necessary in digital text. The functions are parallel, but the affordances of the forms make text and speech more different than ever. Multimodal transpositions like Siri or visual renderings to screen are cleverer than ever because the distance traveled in the transposition is so great.

Each of the ontologies that govern our digital meanings has a ramshackle feel to it – lists of things of uneven significance, and lists of lists, texts only roughly arranged in supermarket order.^§MS0d But surely, they are no more ramshackle, their designs no more irreducibly unique in each categorical node and every relation of their structures, than the natural-historical and social-historical worlds to which they refer.

However, for all their messiness, the ontologies of digital text are no flash-in-the pan inventions of computer whizzes. They are thoroughly thought-through semantic schemas. They have their source in centuries of meticulous agonizing about the structures of meaning by scientists, doctors, librarians, geographers, and such like. Collective thinking has made them unnatural, and this is the source of their peculiarly modern efficacy.¹⁸⁶

The infrastructure of these ontologies has been created and is maintained in this strange, universal, stateless space of self-governing standards bodies. If there is an uber-government of digital meanings, a regulator of the framework for the making of meanings of meanings, since its founding in 1994 it has been the Word Wide Web Consortium, or W3C.

Karl Marx foreshadowed the withering away of the state with the coming of socialism. In the actually existing socialisms that followed the 1917 Russian Revolution, instead of withering, the state became overbearing. Today, neoliberal capitalism is actively withering the state. In its place, these peculiar and largely invisible structures of world governance are emerging. This is an era of (un)governance by ontology.

---

¹⁷⁶ Jon Bosak and Tim Bray 1999, “XML and the Second-Generation Web,” Scientific American 280(5):89–93.

¹⁷⁷ https://en.wikipedia.org/wiki/XML#History

¹⁷⁸ www.w3.org/TR/REC-xml/

¹⁷⁹ Bill Cope and Mary Kalantzis, 2004, “Text-Made Text,” E-Learning 1(2):198–282.

¹⁸⁰ P. Murray-Rust and H.S. Rzepa, 2003, “Chemical Markup, XML and the World Wide Web,” Journal of Chemical Information and Computer Sciences 43(3):757–72.

¹⁸¹ www.ethnologue.com

¹⁸² www.itu.int/en/ITU-T/publications/Pages/default.aspx

¹⁸³ Bill Cope and Mary Kalantzis, 2004, “Text-Made Text,” E-Learning 1(2):198–282.

¹⁸⁴ 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.

¹⁸⁵ Richard Vines and Joseph Firestone, 2011, “Interoperability and the Exchange of Humanly Usable Digital Content,” pp. 429–90 in Towards a Semantic Web: Connecting Knowledge in Academic Research, edited by B. Cope, M. Kalantzis, and L. Magee, Oxford, UK: Chandos.

¹⁸⁶ Bill Cope, Mary Kalantzis, and Liam Magee, 2011, Towards a Semantic Web: Connecting Knowledge in Academic Research, Cambridge, UK: Elsevier; Allen Renear, David Dubin, C.M. Sperberg-McQueen, and Claus Huitfeldt, 2002, “Towards a Semantics for XML Markup,” pp. 119–26 in Proceedings of the 2002 ACM Symposium on Document Engineering, McLean, VA.

---

Cope, Bill, and Mary Kalantzis. 2020. Making Sense: Reference, Agency and Structure in a Grammar of Multimodal Meaning. Cambridge UK: Cambridge University Press, pp. 313–316. https://doi.org/10.1017/9781316459645

Extensible Markup Language §MS3.3b